EZH2 deletion in early mesenchyme compromises ...

3 downloads 0 Views 12MB Size Report
Dec 1, 2016 - Chemicals Limited, Seven Hills, NSW, Australia). Mineralized ... insulin (Austr47219; Royal Adelaide Hospital Pharmacy, Ade- laide, SA ...
The FASEB Journal article fj.201600748R. Published online December 1, 2016. THE

JOURNAL

• RESEARCH •

www.fasebj.org

EZH2 deletion in early mesenchyme compromises postnatal bone microarchitecture and structural integrity, and accelerates remodeling Sarah Hemming,*,† Dimitrios Cakouros,*,†,1 John Codrington,‡ Kate Vandyke,†,§,{ Agneiszka Arthur,*,† Andrew Zannettino,†,§ and Stan Gronthos*,†,1,2

*Mesenchymal Stem Cell Laboratory and §Myeloma Research Laboratory, School of Medicine, Faculty of Health Sciences, and ‡School of Mechanical Engineering, University of Adelaide, Adelaide, South Australia, Australia; †Cancer Theme, South Australian Health and Medical Research Institute, Adelaide, South Australia, Australia; and {South Australia Pathology, Adelaide, South Australia, Australia

In this study, we examined the functional importance of EZH2 during skeletal development and homeostasis using the conditional deletion of Ezh2 (Ezh2fl/fl) in early mesenchyme with the use of a Prrx-1-cre driver mouse (Ezh2+/+). Heterozygous (Ezh2+/2) newborn and 4-wk-old mice exhibited increased skeletal size, growth plate size, and weight when compared to the wild-type control (Ezh2+/+), whereas homozygous deletion of Ezh2 (Ezh22/2) resulted in skeletal deformities, reduced skeletal size, growth plate size, and weight in newborns, and 4-wk-old mice. Ezh22/2 mice exhibited enhanced trabecular patterning. Osteogenic cortical and trabecular bone formation was enhanced in Ezh2+/2 and Ezh22/2 animals. Ezh2+/2 and Ezh22/2 mice displayed thinner cortical bone and decreased mechanical strength compared to the wild-type control. Differences in cortical bone thickness were attributed to an increased number of osteoclasts, corresponding with elevated levels of the bone turnover markers cross-linked C-telopeptide-1 and tartrate-resistant acid phosphatase, detected within serum. Moreover, Ezh2+/2 mice displayed increased osteoclastogenic potential coinciding with an upregulation of Rankl and M-csf expression by mesenchymal stem cells (MSCs). MSCs isolated from Ezh2+/2 mice also exhibited increased trilineage potential compared with wild-type bone marrow stem cells (BMSCs). Gene expression studies confirmed the upregulation of known Ezh2 target genes in Ezh22/2 bone tissue, many of which are involved in Wnt/BMP signaling as promoters of osteogenesis and inhibitors of adipogenesis. In summary, EZH2 appears to be an important orchestrator of skeletal development, postnatal bone remodelling and BMSC fate determination in vitro and in vivo.—Hemming, S., Cakouros, D., Codrington, J., Vandyke, K., Arthur, A., Zannettino, A., Gronthos, S. EZH2 deletion in early mesenchyme compromises postnatal bone microarchitecture and structural integrity, and accelerates remodeling. FASEB J. 31, 000–000 (2017). www.fasebj.org ABSTRACT:

KEY WORDS:

mesenchymal stem cells



skeletal development



differentiation



epigenetics



conditional

knockout The primitive mesenchyme, marked by its expression of paired-related homeobox-1 (PRRX1), is responsible for early limb bud development and is thought to give rise

to bone mesenchymal stem/stromal cells (BMSCs) within the newly formed skeleton (1). Furthermore, BMSC lineage determination is driven by key regulators

ABBREVIATIONS: 2D, 2 dimensional; 3D, 3 dimensional; a-MEM, a-minimum essential medium; Alk phos, alkaline phosphatase; AMSC, amnion-

derived MSC; BFR, bone formation rate; BFR/BS, bone formation rate/bone surface; BMSC, bone marrow–derived stromal/stem cell; BS, bone surface; BV, bone volume; CFU-F, colony-forming unit-fibroblast; Col II, collagen type II; Ct.Th., cortical thickness; CTX, cross-linked C-telopeptide; E, embryonic day; EZH2, human enhancer of zeste homolog 2 (gene); EZH2, human and mouse enhancer of zeste homolog 2 (protein); Ezh2, mouse enhancer of zeste homolog 2 (gene); Ezh2+/+, Tg.Prx-1 Cre+:Ezh2 wt/wt; Ezh2+/2, Tg.Prx-1 Cre+:Ezh2 fl/wt; Ezh22/2, Tg.Prx-1 Cre+:Ezh2 fl/fl; H3K27, histone 3 lysine 27; H&E, hematoxylin and eosin; hMSC, human mesenchymal stem cell; FBS, fetal bovine serum; LVDT, linear variable differential transformer; MAR, mineral apposition rate; M-CSF, macrophage colony stimulating factor; Me3, trimethylation; micro-CT, microtomography; MSC, mesenchymal stem cell; N.Adip/Mar.Ar, number of adipocytes in marrow area; N.Ob/B.Pm., number of osteoblasts on bone perimeter; OCN, osteocalcin; OPN, osteopontin; PgC, polycomb group; PrC, polycomb repressive complex; Prrx, paired related homeobox promoter; RANKL, receptor activator of nuclear factor-kB ligand; ROI, region of interest; Ror, receptor tyrosine kinase-like orphan receptor; Runx, runt-related transcription factor; Sox9, sex-determining region SRY of the Y chromosome 9; Tb.Sp., trabecular spacing; Tb. Th., trabecular thickness; TRAP, tartrate-resistant acid phosphatase; TV, tissue volume 1 2

These authors contributed equally to this work. Correspondence: School of Medicine Faculty of Health Sciences, University of Adelaide, Adelaide 5005, SA, Australia. E-mail: stan.gronthos@adelaide. edu.au

doi: 10.1096/fj.201600748R This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

0892-6638/17/0031-0001 © FASEB

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

1

of lineage-associated transcription factors, including Runtrelated transcription factor (RUNX)-2)/core-binding factor (CBF)-A1, peroxisome proliferator-activated receptor (PPAR)-g2, and sex-determining region SRY of the Y chromosome (SOX)-9, which are essential for dictating osteogenic, adipogenic, or chondrogenic lineage specification, respectively (2–4). The precise role of epigenetic regulation in relation to these and other genes involved in BMSC maintenance and cell fate determination have yet to be fully determined. The epigenetic regulation of stem cell differentiation and self-renewal has been mainly focused on embryonic stem cells with fewer studies in adult stem cells (5). Approximately 28 known histone modifications have been associated with post-translational modification and remodelling of nucleosomes. Many of these modifications are known to regulate stem cell differentiation by regulating gene expression and dictating lineage-specific gene expression (6, 7). To date, the single most important epigenetic complex in stem cell differentiation is the polycomb complex, which was first discovered in Drosophila (8). Polycomb group (PcG) proteins assemble into 2 polycomb repressive complexes: PRC-1 and -2. PRC2 contains the histone 3 lysine 27 methyl transferase, enhancer of zeste (EZH) 1 and EZH2, embryonic ectoderm development, (EED) and suppressor of zeste 12 (SUZ12) (9). Histone 3 lysine 27 me3 (H3K27me3) is the physiological substrate for PRC1, which acts as a docking site for the chromodomain protein, CBX, present in PRC2 (9, 10). The Pc proteins cause compaction of nucleosomes, chromatin compaction, and mediate long-range chromatin interactions (11, 12). PcG proteins are highly expressed in stem cells, binding promoters of transcription factors and developmental genes to inhibit differentiation to maintain an immature stem cell phenotype (13, 14). EZH2 maintains the multipotency of hematopoietic, muscle, and neural stem cells (15–17) and is also involved in mediating tissue-specific differentiation (18–22). In mouse, EZH2 promotes adipogenic differentiation in peripheral preadipocytes through repression of Wnt1, -6, -10a, and -10b (18). Human studies have shown that CDK1 inhibits EZH2 methyltransferase activity via the phosphorylation of T48, resulting in the promotion of osteogenesis of human MSCs (19). Thus, EZH2 down-regulation or inhibition, leads to a decrease in H3K27me3 along osteogenic genes such as RUNX2, and an increase in RUNX2 expression, which is necessary for BMSC osteogenic lineage specification. In agreement with these findings, we have reported that EZH2 expression decreases during osteogenesis and is essential in repressing osteogenic differentiation of MSCs by acting on the osteogenic master regulatory gene, RUNX2, and its downstream bone gene target, osteopontin (OPN) (22). Conversely, we found that EZH2 promotes adipogenesis via an indirect mechanism, most probably through inhibition of the Wnt pathway, a known repressor of adipogenesis (18, 23). The in vivo function of EZH2 in skeletal development has been difficult to examine, mainly because of the embryonic lethality exhibited by EZH2-deficient mice (24, 25). However, conditional knockout of EZH2 in uncommitted mesenchymal cells resulted in multiple defects in skeletal patterning and 2

Vol. 31

March 2017

bone formation, including shortened forelimbs, craniosynostosis, and clinodactyly (7, 26). In the present study, we examined the in vivo function of EZH2 in regulating trabecular and cortical bone microarchitecture, strength, and remodelling, using Ezh2 conditional knockout mice driven by the Prrx1 promoter. MATERIALS AND METHODS Generation of EZH2 conditional knockout mouse Ezh2tm1Tara (Ezh2fl/fl) mice from the Mutant Mouse Resource and Research Centre (University of California, Davis, Davis, CA, USA) on a mixed background were backcrossed to 7 generations to achieve a C57BL/6 genetic background. The Cre-loxP system (1) was used to conditionally ablate EZH2 in limb bud mesenchyme by crossing Ezh2fl/fl mice with paired related homeobox 1 promoter (Prrx1)–derived enhancer (Prrx1-Cre transgene B6). Cg-Tg(Prrx1-cre+/2)1Cjt/J (Prrx-1Cre+) mice (The Jackson Laboratory, Bar Harbor, ME, USA). Mice carrying a transgene, in which exons encoding the SET domain of Ezh2 are flanked by loxP sites, are described in Hirabayashi et al. (27). Previously described Ezh2 and Prrx-1 Cre genotyping primers were used to detect knockdown (1, 7). The SA Pathology Animal Ethics Committee-approved the following protocols: EZH2 breeding colony (BC06/12), Prrx1-breeding colony (BC70a/12), and the EZH2 development study (103c/13). The Adelaide University Animal Ethics Committee approved the M-2013-143 protocol.

RNA extraction, cDNA synthesis, and real-time PCR We extracted RNA from passage 5 BMSCs by using Trizol (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer’s instructions. Generation of cDNA and real-time analysis was performed as previously described in triplicate (28). Total RNA from femora and tibia from newborn mice fixed in 95% ethanol were serially rehydrated, and bone marrow was flushed. Bones were frozen in liquid nitrogen and crushed with a mortar and pestle. Fragments were put into TRI reagent (Dricet-zol RNA MiniPrep Kit; Zymo Research, Irvine, CA, USA), sonicated for 30 s and centrifuged for 30 s. The RNA containing supernatant was then purified as per protocol for the Directzol RNA MiniPrep Kit (Zymo Research). RNA (500 ng) was converted to cDNA with SuperScript IV (Thermo Fisher Scientific). Real-time PCR primers (GeneWorks, Thebarton, SA, Australia) used in this study were: Ezh2 (NM_007971.2), (forward) 592 ACTGCTGGCACCGTCTGATG239, 592TCCTGAGAAATAATCTCCCCACAG239; M-csf (NM_007778.4), (forward) 59-2GCCAGGGGAAAGTGAAAGTT239, (reverse) 592CATGAGGAGACAGACCAGCA239; Rankl (NM_011613.3), (forward) 592 AGCCGAGACTACGGCAAGTA239, (reverse) 592AGTCCTGCAAATCTGCGT239; Opg (NM_008764.3), (forward) 592ATGAACAAGTGGCTGTGCTG239, (reverse) 592RGTAGGTGCCAGGAGCACATT239; Runx2 (NM_001289690.1), (forward) 592 CCTCTGACTTCTGCCTCTGG239, (reverse) 592TATGGAGTGCTGCTGGTCTG239; Opn (NM_001204201.1), (forward) 592 AGCAAACTCTTGCAAGCAA239, (reverse) 592GATTCGTCAGATTCATCCGAGT239; Ocn (NM_0010-32298.3), (forward), 592AAGCAGGAGGGCAATAAGGT239, (reverse) 592TCAAGCCATACTGGTGTGATAGC239; Pparg2 (NM_001127330.2), (forward) 592TTTTCCGAAGAACCATCCGATT239, (reverse) 592ATGGCATTGTGAGACATCCCC239; C/ebpa (NM_001287514.1), (forward) 592CAAGAACAGCAACGAGTACCG239, (reverse) 592CTCACTGGTCAACTCCAGCCA239; AdipQ (NM_009605.4), (forward) 592TGTTCCTCTTAATCCTGCCCA239, (reverse)

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

592CCAACCTGCACAAGTTCCCTT239; Hopx (NM_175606.3/ 001159900.1), (forward) 592GAGGACCAGGTGGAGATCCT239, (reverse) 592TCAAAACAGCCTGGGTAAGC239; Ror2 (NM_013846.3), (forward) 592AGGCCGTCATGTATGGAAAG239, (reverse) 592GATCATCTCCACCACGTCCT2 39; Myadm (NM_001093764.1), (forward) 592ATCAGCCTTTCCAGGTGTG239, (reverse) 592TGGTGGAGGATGACGTTGTA239; Mx1 (NM_010846.1), (forward) 592GCAGACGGAATATTGGGAGA239, (reverse) 592GACCAGGAAAGCCACATAGC239; Bmp2 (NM_007553.3), (forward) 592ACTTTTCTCGTTTGTGGAGC239, (reverse) 592GAACCCAGGTGTCTCCAAGA239; Wnt10b (NM_011718.2), (forward) 592 TCTCTTTCAGCCCTTTGCTCGGAT239, (reverse) 592ACAACTGAACGGAAGGAGAAGCCT239; Wnt5a (NM_009524.3), (forward) 592CAAATAGGCAGCCGAGAGAC239, (reverse) 592CTCTAGCGTCCACGAACTCC239; Fhl1 (NM_001077361.1) fwd 592gtgtccaaggatggcaagat239, rev 592gggctgatcctggtaagtga2 39; and Gapdh (NM_001289726.1), (forward) 592AGGTCGGTGTGAACGGATTTG239, (reverse) 592tgtagaccatgtagttgaggtca239.

Western blot analysis Murine BMSCs were seeded at 8.0 3 103 cells/cm2 and cultured in regular growth medium, as previously described. Whole-cell lysates (50 mg) were separated on SDS gels (1–4). Membranes were probed with anti-mouse EZH2 (Acc2, 1:1000; Cell Signaling Technology, Beverly, MA, USA), anti-rabbit H3K27me3 IgG (07-449; 1:1000; Merck Millipore, Bayswater, VIC, Australia) and anti-mouse b-actin IgG (1:2500; Sigma-Aldrich, Castle Hill, NSW, Australia) antibodies. Secondary detection was performed with anti-rabbit alkaline phosphatase (alk phos) conjugate (1:10,000; Merck Millipore) and anti-mouse alk phos (1:10,000; Merck Millipore) antibodies.

Microtomography Three-dimensional (3D) microarchitecture of newborn long bones was evaluated using microtomography (micro-CT; SkyScan 1076 X-ray Micro-CT; Bruker microCT, Kontich, Belgium). Whole newborn skeletons were scanned at 9 mm resolution, no filter, excitation 1767 ms, voltage 55 kV, current 120 mA, rotation step 0.5, and 2-frame averaging. Four-week-old tibiae and femora were scanned at 9-mm resolution, Al 0.5 mm filter, excitation 5890 ms, voltage 48 kV, current 110 mA, rotation step 0.6, and 2-frame averaging. Reconstruction of the original scan data was performed using NRecon (NRecon Reconstruction 64-bit, ver. 1.6.10; Bruker microCT) with a smoothing of 1, ring artifact of 8, and beam hardening of 30%. Reconstructed .bmp files were realigned in Data Viewer (DataViewer, 64-bit, ver.1.5.2; Bruker microCT) and volume of interest was subjected to 2D and 3D morphometric parameters calculated by a Comprehensive Tex Archive Network CT-analyzer (CTAn, ver. 1.15,+ CTVol, ver. 2.3: 32-bit version XP, Vista, Win-7 2D/3D processing, analysis, visualization; Bruker microCT) in 2D or 3D based on a surface-rendered volume model in 2D or 3D space. The bone parameters assessed were based on those of the American Society for Bone and Mineral Research, published by Parfitt et al. (29). Bone microarchitecture parameters from a previously published study were used (30). For the newborn tibia and region of interest (ROI) analyzed was defined as 40 slices under the proximal growth plate to a defined region of bone within the proximal epiphysis. The length of the bone was determined from the first slice of the epiphysis to the end of the proximal epiphysis. The trabecular ROI was defined manually to elude the cortex (31). The midpoint of the bones, determined by the length measurement at 20 slices above and below this point was analyzed (40 slices). The medullary space was subtracted from the cortical data set to remove Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

any trabeculae present. 3D images were generated using Avizo Fire 3D software (FEI, Hillsboro, OR, USA) or CTVol 3D software (Bruker microCT). Histomorphometric bone analysis Histomorphometric analyses were conducted on blinded coded slides with OsteoMeasurexp ver. 3.3.02 (Osteometric, Decatur, GA, USA) software on an Olympus BX53 Microscope (Olympus, Notting Hill, VIC, Australia) at 3200 magnification. An area directly under the primary spongiosa was used to analyze the trabeculae in the secondary spongiosa and tissue immediately adjacent to the cortex was excluded from trabecular analysis. An area of trabecular bone (0.72 mm2)-to-tissue volume (BV/TV, %), trabecular thickness (Tb.Th., in micrometers), trabecular spacing (Tb.Sp., in micrometers), and trabecular number (Tb.Th. per millimeter) were analyzed. A length of 3.6 mm of cortical bone was analyzed at the midpoint of the diaphysis determining cortical BV/TV and cortical thickness (Ct.Th, in millimeters). Analysis of the number of osteoblasts on trabecular bone (N.Ob/B.Pm. per millimeter) and adipocytes per marrow area (N.Adip/ Ma.Ar, in square millimeters) was analyzed in a region of the secondary spongiosa. Tissue immediately adjacent to the cortex was excluded from trabecular analysis. Biomechanical testing Femora were stored at 220°C and prepared as published in Arthur et al. (32). The 3-point bend testing was undertaken with a test resources mechanical test machine (800LE4; TestResources, Shakopee, MN, USA) (33, 34). The lower span width was 3.5 mm, the lower and upper contact anvils had a radius of 1 mm. Femora were placed posterior side down (on the 2 lower anvils), with the midpoint between the anvils located centrally along the diaphysis. A preloaded force of 0.5 N was applied, and the test was conducted in displacement control at a rate of 1 mm/min, and the load was measured with a 6111 N (25 lb) load cell. Displacement was measured from a linear variable differential transformer (LVDT; 65 mm) attached above the load cell. LVDT measurement corrected for compliance in the load line by calibrating with an aluminum test piece. Load displacement, flexural rigidity, and ultimate bending moment were calculated (32). Histology Paraffin-embedded femora were sectioned, and the resultant 5-mm sections were dewaxed and stained with hematoxylin and eosin (H&E) and imaged with NanoZoomer 2.0-HT scanner (Hamamatsu Photomics, Hamamatsu City, Shizuoka, Japan). In brief, fixed tibiae were decalcified, infiltrated, embedded in methyl methacrylate (Merck Millipore), mounted, and sectioned (35, 36). Longitudinal tissue sections (5 mm) were mounted on gelatin-coated Superfrost Plus slides (Thermo Fisher Scientific) and were stained with Safranin O/Fast Green, and or toluidine blue (32). Whole embryonic day (E)16.5, E17.5, newborn, and 4-wk-old tibias were stained with 0.3% Alcian blue (SigmaAldrich) and 0.1% Alizarin red S (Sigma-Aldrich). Immunohistochemistry Paraffin-embedded sections (5 mm) of 4-wk-old femora were dewaxed through histolene and ethanol solutions. After proteinase K (20 mg/ml) antigen retrieval, the sections were incubated at 37°C for 30 min in a humidity chamber. Sections were washed 3 times in 13 PBS, and then blocked in 0.3 v/v% H2O2 in

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

3

methanol for 10 min at room temperature. Washed sections were incubated with horse anti-mouse blocking serum (PK-6200; Vectastain Elite Universal ABC kit; Vector Laboratories; Burlingame, CA, USA). Serum was removed, and 2 mg/ml of anti-rabbit H3K27me3 IgG (07-449; Merck Millipore) or 2 mg/ml rabbit serum (negative control) was added to sections and incubated overnight. Sections were washed and incubated with a universal biotinylated secondary antibody (Vectastain Elite Universal ABC kit; Vector Laboratories) for 45 min at room temperature. Sections were washed and incubated with an avidin biotin complex for 30 min, and after incubation, the sections were washed in the peroxidase substrate 3-amino-9-ethylcarbazole (AEC; SK-4200; Peroxidase substrate kit; Vector Laboratories). Sections were left for 10 min to allow color to develop, washed with H2O, counterstained with hematoxylin, and mounted with an aqueous mounting solution. Collagen type II (Col II) expression in paraffin-embedded cartilage pellets from Ezh2+/+ and Ezh22/2 was assessed. Paraffin-embedded sections (5 mm) were prepared and stained with anti-Col II monoclonal antibody, as previously described (37). Tartrate-resistant acid phosphatase staining Cells were stained for tartrate-resistant acid phosphatase (TRAP) activity with leukocyte acid phosphatase (387A-1KT; Sigma-Aldrich) (30). The cells were visualized and photographed with an CKX41 inverted microscope and a DP11 digital imaging camera (cellSens; Olympus, Notting Hill, VIC, Australia) at 3200 magnification. Calcein labeling and histomorphometric analysis Calcein (20 mg/kg) was injected intraperitoneally at 4 d and 24 h before harvesting of mice. Histomorphometric analyses were conducted on blinded coded slides with OsteoMeasurexp, ver. 3.3.02 (Osteometric) at 3200 magnification. Methacrylateembedded tibiae were sectioned, deplasticized, and coverslipped for analysis of calcein labeled bone formation. Mineral apposition rate (MAR; in micrometers per day) was evaluated as the mean distance between the calcein-labeled surfaces, divided by 3 d (the interval between labeling and the death of the animals). Bone formation rate on bone surface (BFR/BS) was derived with the formula BFR = MAR 3 MS/BS 3 d/100 (results expressed in cubic micrometers/square millimeters/ day). Analysis of bone turnover serum markers Blood serum samples were recovered after cardiac puncture, and the levels of TRAP, osteocalcin (OCN), and cross-linked C-telopeptide (CTX)-1 analyzed with a MouseTRAP ELISA (TRAP 5b/SB-TR103; Immunodiagnostics, Boldon, United Kingdom), Rat-Mid Osteocalcin EIA ELISA (AC-12F1; Immunodiagnostics) and RatLaps CTX-I EIA ELISA (AC-06F1; Immunodiagnostics), as specified by the manufacturer’s instructions. Bone marrow–derived MSC isolation and culture Mouse stromal progenitors were isolated from flushed femora and tibiae (30). Flushed bone marrow cells were cryopreserved and used for osteoclast assay. Cells were seeded for colonyforming unit-fibroblast (CFU-F) assays at a density of 1 3 105 and incubated for 7 d at 37°C in hypoxic conditions (5% O2, 10% CO2; Coy Laboratory Products, Grass Lake, MI, USA). At d 8, the CFUF were stained with an Alkaline Phosphatase Staining Kit (85L3R1KT; Sigma-Aldrich) and counterstained with toluidine blue. 4

Vol. 31

March 2017

Replicate cultures were seeded and cultured in normal growth medium in hypoxia for 2 wk. MSCs were subcultured up to passage 5 for differentiation and gene/protein expression studies. In vitro differentiation assays Normal growth conditions BMSCs were cultured in hypoxic conditions (5% O2, 10% CO2; Coy Laboratory Products) at 37°C in a-minimum essential medium (a-MEM; M4526, 500 ml; Sigma-Aldrich) supplemented with 20% (v/v) fetal calf serum (FCS; SAFC Biosciences, Melbourne, VIC, Australia), 2 mM L-glutamine (G7513-100 ml; SigmaAldrich), 1 mM sodium pyruvate (S8636; Sigma-Aldrich), 10 mM HEPES buffer (H0887-100 ml; Sigma-Aldrich), and 50 U/ml penicillin–50 mg/ml streptomycin (P4333-100 ml; Sigma-Aldrich).

Osteogenic differentiation aMEM supplemented with 20% (v/v) FCS, 2 mM L-glutamine, 1 mM sodium pyruvate, 10 mM HEPES buffer, 50 U/ml penicillin, 50 mg/ml streptomycin, 100 mM L-ascorbate-2-phosphate (Wako Pure Chemical Industries, Richmond, VA, USA), 10 nM dexamethasone (RAH688A; Royal Adelaide Hospital, Adelaide, SA, Australia), and 4 mM KH2PO4 (Asia Pacific Specialty Chemicals Limited, Seven Hills, NSW, Australia). Mineralized bone matrix (mineral) formation was identified with Alizarin red (A5533-25G; Sigma-Aldrich) staining of mineral deposits (38). Extracellular calcium was measured by Calcium Arsenazo III (TR29226; Thermo Fisher Scientific) in triplicate samples and normalized to DNA per well with a PicoGreen dsNDNA quantitation kit (P11496; Thermo Fisher Scientific) (38). Adipogenic differentiation aMEM supplemented with 20% (v/v) FCS, 2 mM L-glutamine, 1 mM sodium pyruvate, 50 U/ml penicillin–50 mg/ml streptomycin, 100 mM L-ascorbate-2-phosphate, 1 mM dexamethasone, and 60 mM indomethacin (I8280; Sigma-Aldrich), and 1 mg/ml insulin (Austr47219; Royal Adelaide Hospital Pharmacy, Adelaide, SA, Australia). The medium was sterilized with a 0.2 mm bottle cap filter (CLS431161-48EA; Sigma-Aldrich) and cultured on BMSCs for up to 14 d, with the medium changed twice weekly. Lipid formation was assessed by Nile Red 25 ng/ml (N3013; Sigma-Aldrich) staining (38). Quantitation of lipid was assessed by Nile Red fluorescence staining, normalized to DAPI 600 mM (D1306; Thermo Fisher Scientific), stained nuclei per field of view in triplicate wells (38, 39). Chondrogenic differentiation MSCs (1–5 3 106) were pelleted by 600 g centrifugation in 10 ml polypropylene tubes. Pellets were cultured in DMEM high glucose (Sigma-Aldrich) supplemented with ITS+Premix (354352; BD Biosciences, San Jose, CA, USA), L-ascorbate-2phosphate100 mM, penicillin 50 U/ml, streptomycin 50 mg/ml, 25 L-glutamine 2 mM, dexamethasone 10 M, and bovine serum albumin 0.125% with the presence or absence of TGF-b3 (recombinant TGF-b3, 243-B3-002; R&D Systems, Minneapolis, MN, USA) (37). Osteoclast differentiation Flushed cultured BM cells were seeded in 96-well chamber slides or 96-well plates onto slices of dentine (South Australian

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

Museum, Adelaide, SA, Australia) (30). In brief, BM cells were cultured at 3.1 3 105 cells/cm2 in 200 ml a-MEM supplemented with 75 ng/ml of M-CSF (GF053; Merck Millipore) and receptor activator of nuclear factor-kB ligand (RANKL; GF091; Merck Millipore). Medium was replaced every 3 d and the formation of TRAP+ multinucleated osteoclasts was assessed after 6 d of osteoclast induction. Cells were stained with leukocyte acid phosphatase (387A-1KT; SigmaAldrich), per the manufacturer’s recommendations. Osteoclasts with 3–5, 6–10, and greater than 10 nuclei were quantitated per well in duplicate wells. Cells were visualized and photographed with a CKX41 inverted microscope and a DP11 digital camera (cellSens; Olympus) at 3200 magnification. The formation of resorption pits on the dentine slices was assessed after 9 d of osteoclast induction. Dentine slices were incubated in 13 trypsin for 3 h, washed with PBS, and carbon coated for imaging by the Philips XL30 Scanning Electron Microscope (SemTech Solutions, North Billerica, MA, USA, at Adelaide Microscopy, University of Adelaide, Adelaide, SA, Australia), as described in Cantley et al. (40). Three representative fields of view were obtained with the SEM at 3500 magnification, and the surface area of total resorption was analyzed with image-analysis software (ImageJ, ver. 1.50e; National Institutes of Health, Bethesda, MD, USA; http://imagej.nih.gov/ij/).

Statistics Data analysis, graph generation, and statistical analyses were performed with Microsoft GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA). For differentiation and gene expression analysis, a paired or unpaired Student’s t test was used to compare Ezh2+/+ and Ezh2+/2 BMSCs treated with control, osteogenic, and adipogenic differentiation medium. For analyses to detect differences between Ezh2+/+, Ezh2+/2, and Ezh22/2 animals, a 1-way ANOVA, with Tukey’s multiple-comparisons test was performed. A 2-way ANOVA with Sidak’s multiplecomparisons test were used to determine the differences between nuclei groups in osteoclast assays and differences in alkaline-positive colonies. A value of P , 0.05 indicated statistical significance.

RESULTS Mesenchymal specific deletion of Ezh2 results in altered newborn hind limb trabecular bone formation and microarchitecture EZH2 has been shown to play a pivotal role in determination of human MSC fate (22). Because global EZH2 knockout mice are embryonically lethal (24, 25), we used a conditional knockout approach to examine the function of EZH2 on skeletogenesis and bone homeostasis. The generation of Prrx-1:Cre+-Ezh2wt/fl heterozygous (Ezh2+/2) and Prrx-1:Cre+-Ezh2fl/fl homozygous (Ezh22/2) mice were confirmed by genotyping (Supplemental Fig. 1A). Ezh22/2 mice were born at the expected Mendelian ratio; however, they died shortly after birth, with only 3 female mice surviving to 4 wk of age, whereas Ezh2+/2 mice were born at the expected Mendelian ratio (data not shown). Given that no male Ezh22/2 mice were detected after birth, we analyzed the skeletons and long bones of female newborns and 4-wk-old mice for this study. EZH2 deletion and associated H3K27me3 modification was Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

confirmed in BMSCs by Western blot, RT-PCR analyses, and immunohistochemical analysis of paraffin-embedded femora (Supplemental Fig. 1B–E). The phenotypic effects attributable to EZH2 deficiency in the mesenchymal stromal population were assessed in embryonic and postnatal skeletons. Supplemental Figs. 2 and 3 show distinct phenotypic differences in skeletal size and morphology in E17.5, newborn, and 4-wk-old Ezh2+/2 and Ezh22/2 mice. Specifically, Ezh22/2 mice exhibited craniosynostosis, absences of the third sacral vertebrae, and decreased size of fore and hind limbs. Notably, growth plate defects were observed in Ezh22/2 (Supplemental Figs. 2 and 4) in accordance with previous reports (7, 26). In contrast to the Ezh22/2 phenotype, heterozygous deletion of Ezh2 resulted in larger and heavier newborns (Fig. 1A, B). Ezh2+/2 mice exhibited larger skeletons and skulls, and the occipital bones were notably larger with more defined sutures compared with wild-type control skulls (Fig. 1C, D). Compared with wildtype controls, Ezh2+/2 newborn exhibited larger fore and hind limbs, as seen in the images of Alizarin red– and Alcian blue–stained newborns (Fig. 1E, F). Ezh2+/2 newborn mice exhibited larger tibial and femoral growth plates with smaller resting zones and larger proliferative and hypertrophic zones, correlating with increased bone length (Fig. 1G–K). Overall deletion of EZH2 affected skeletal, skull, and growth plate size and overall length of limbs. Histological analysis of bone microarchitecture revealed that trabecular bone was patterned throughout the Ezh22/2 newborn tibiae, whereas trabeculae were restricted to the metaphysis of the femora of Ezh2+/2 and Ezh2+/+ mice (Fig. 1L). Compared to Ezh2+/2 and Ezh2+/+ mice, micro-CT analysis revealed that Ezh22/2 mice displayed an increased BV/TV ratio within the tibiae of newborns (Fig. 1M). However, there were no differences between trabecular BV/TV for Ezh2+/2 and Ezh2+/+ newborns. Examination of trabeculae parameters in newborn revealed an increased number of trabeculae (Tb.N) in Ezh22/2 mice compared with wild-type mice, with no significant difference from Ezh2+/2 mice (Fig. 1N). Of note, no significant differences were evident between Ezh2+/2 and wild-type mice except for an increased TV, BV (data not shown), and longer bones. Overall, the newborn tibiae patterning defects altered the location and number of trabeculae throughout the long bones in Ezh22/2 animals. Mesenchymal-specific deletion of Ezh2 results in altered 4-wk-old hind limb trabecular bone formation and microarchitecture Ablation of EZH2 resulted in complete separation of tibiae and fibulae only in the Ezh22/2 newborn and 4-wk-old hind limbs (Supplemental Fig. 2K). The femora of Ezh22/2 mice showed differences in morphology of the epiphysis and diaphysis, which was considerably wider compared with Ezh2+/2 and Ezh2+/+ femora (Supplemental Fig. 3I). Femora of Ezh2+/2 mice were similar in morphology, while considerably larger in length compared with control mice (Fig. 2A, B).

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

5

Figure 1. Ezh2 deletion alters newborn limb patterning and trabecular bone microarchitecture. A) Photograph of Ezh2+/+ and Ezh2+/2 newborns. B) Weight of Ezh2+/+and Ezh2+/2 newborns. C ) 3D image of micro-CT–scanned Ezh2+/+and Ezh2+/2 newborns. D) 3D image of newborn skulls identifying differences in size and suture formation. Alizarin red staining of mineral and Alcian blue staining of cartilage in newborn (E ) fore limbs and (F ) hind limbs. G) Histological images of Safranin O–stained cartilage in the growth plate of newborn mice, depicting resting (i), proliferative (ii), and hypertrophic (iii) zones. Quantitation (continued on next page) 6

Vol. 31

March 2017

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

Similar to newborn tibiae microarchitecture, analysis of 4-wk-old femora revealed that trabecular bone was patterned throughout the entire Ezh22/2 femoral shaft, whereas trabecular cells were restricted to the metaphysis of Ezh2+/2 and Ezh2+/+ mice (Fig. 2C). Ezh22/2 femora exhibited an increased percentage of total trabecular bone proportional to tissue volume with an increase in the number of trabecular cells, when compared to Ezh2+/2 and Ezh2+/+ mice (Fig. 2D, E). Of note, no significant differences were evident between Ezh2+/2 and wild-type mice, except for an increased TV, BV (data not shown), and longer bones. Overall, 4-wk-old femora patterning defects altered the morphology, location, and number of trabeculae through the long bones in Ezh22/2 animals. Further analysis of the secondary spongiosa trabeculae in tibial sections stained with toluidine blue (Fig. 2F) revealed decreased BV/TV of Ezh22/2 mice compared with wild-type controls, even though more trabeculae were present throughout the whole length of the femora (Fig. 2H). Moreover, the secondary spongiosa trabeculae in Ezh22/2 mice were found to be significantly thicker, with increased trabecular spacing (Tb.Sp) and decreased cell number, when compared to Ezh2+/+ and Ezh2+/2 mice (Fig. 2G–J). Calcein labeling was used to determine whether the differences in trabecular thickness were related to changes in BFRs in 4-wk-old mice. Histological analysis of methacrylate embedded undecalcified tibial sections found 2 distinct fronts of calcein-labeled, newly formed secondary spongiosa trabeculae in Ezh2+/2 and Ezh22/2 mice, compared to control mice (Fig. 2K). Quantification of trabeculae labeled with calcein in the secondary spongiosa found increased MARs and BFRs on BS (BFR/BS) in Ezh2+/2 and Ezh22/2 compared to wildtype control mice (Fig. 2L, M). Histomorphometric analysis reveal a greater number of osteoblasts lining the trabecular bone perimeter (N.Ob/B.Pm) in Ezh2+/2 and Ezh22/2 mice, compared to control mice (Fig. 2N). Serum analysis of Ezh2+/2 and Ezh22/2 mice revealed higher levels of the osteoblastic activity marker OCN when compared to that in control mice (Fig. 2O). These findings suggest that EZH2 deletion affects the trabecular pattern, decreasing the number of secondary spongiosa in trabecular bone and increasing the spacing between thicker trabeculae, correlating with increased MAR, BFR/BS, number of osteoblasts on the bone surface, and expression of OCN. Mesenchyme-specific deletion of Ezh2 results in altered hind limb bone cortical formation and microarchitecture Micro-CT 3D modeling of a defined cortical bone region in the middle of the diaphysis demonstrated thinner cortical

bone in Ezh2+/2 and Ezh22/2 femora, compared to wildtype controls (Fig. 3A). Assessment of bone parameters found that Ezh22/2 mice exhibited decreased cortical BV/ TV, with both Ezh2+/2 and Ezh22/2 having a significant decreased Ct.Th, compared with Ezh2+/+ mice (Fig. 3B, C). Histological assessment of H&E-stained tissue sections of 4-wk-old femora revealed a disorganized patterning of the cortical bone in Ezh22/2 mice compared to Ezh2+/2 and Ezh2+/2 mice (Fig. 3D). Histomorphometric analysis of cortical bone confirmed a decreased cortical BV/TV and cortical thickness in Ezh22/2 and Ezh2+/2 mice compared with wild-type mice and between Ezh22/2 and Ezh2+/2 mice (Fig. 3E, F). Calcein labeling experiments revealed that newly formed cortical bone in the middle of the diaphysis exhibited distinct calcein labeled fronts in Ezh2+/2 and Ezh22/2 compared with wild-type control mice where the calcein-labeled bone fronts were close together (Fig. 3G). An increase in MAR and BFR/BS was observed in Ezh2+/2 and Ezh22/2 cortical bone when compared to wild-type controls (Fig. 3H, I). Both MAR and BFR/BS cortical bone parameters were significantly higher in the Ezh22/2 mice than in the Ezh2+/2 mice, associated with irregular bone mineralizing fronts. Histomorphometric analysis reveal a greater number of osteoblasts lining cortical bone parameter (N.Ob/B.Pm) in Ezh2+/2 and Ezh22/2 mice, compared to that in control mice (Fig. 3J). Biomechanical testing was subsequently used to assess bone structural integrity of 4-wk-old femora, following micro-CT analysis. The data showed that the femora of 4-wk-old Ezh2+/2 and Ezh22/2 mice exhibited decreased flexural rigidity or stiffness and yield moment, when compared to wild-type controls (Fig. 3K, L). Therefore, EZH2 deletion reduced cortical thickness and organization, despite increased rates of bone formation, leading to compromised bone strength and integrity. Mesenchyme-specific deletion of Ezh2 results in altered bone turnover and remodelling The decrease in cortical bone observed in Ezh2-deficient mice was in contrast to the increased rate of bone formation. This finding prompted us to examine osteoclastmediated bone resorption in Ezh2-deficient animals. To investigate whether EZH2 deletion affects bone resorption in 4-wk-old animals, we assessed the rate of osteoclastmediated bone resorption and found that Ezh2+/2 and Ezh22/2 mice had a significantly higher number of osteoclasts on the trabecular and cortical bone perimeter and increased levels of serum TRAP and CTX-1, compared with control mice (Fig. 4A–E). Ezh22/2 mice also expressed significantly more osteoclasts on trabecular bone and higher serum levels of TRAP and CTX-1 compared with Ezh2+/2 and control mice. These findings suggest that

of growth plate size (H ) and resting (I ), proliferative (J ), and hypertrophic (K ) zones. L) Toluidine blue–stained cross sections of newborn tibiae identified differences in trabecular patterning and tibial size. M ) BV/TV and (N ) trabecular number (Tb.N) within the newborn tibiae. Representative image of 5 Ezh2+/2, 5 Ezh2+/+, and 3 Ezh22/2 replicate mice. Data are means 6 SEM of results in 5 Ezh2+/2 or 5 Ezh2+/+ female and 3 Ezh22/2 female mice (B, D–F, G–J, L, M ). *P , 0.05. Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

7

Figure 2. Ezh2 deletion affects 4-wk-old hind limb size, trabecular patterning, and microarchitecture. A) Image of micro-CTscanned femora. B) Length of 4-wk-old femora calculated by micro-CT. C ) 3D-generated model of trabecular pattern in the femora. D) Micro-CT-quantitated percentage of BV/TV and (E ) number of trabeculae (Tb.N) in 4-wk-old femora. F ) Image of toluidine blue–stained growth plate and secondary trabecular spongiosa. Histomorphometric analyses of the secondary spongiosa: BV/TV (G), Tb.N (H ), trabecular thickness (Tb.th) (I ) and trabecular spacing (Tb.Sp) (J ). K ) Image of calcein labeled (arrows), newly formed trabecular bone in Ezh2+/+, Ezh2+/2, and Ezh22/2 4-wk-old methacrylate-embedded tibiae. Bone formation was determined by MAR. Scale bar, 10 mm. (L) and BFR/BS (M). N ) Number of osteoblasts lining the trabecular bone perimeter (Nb.ob/B.Pm). Levels of serum OCN. A, C, F, K ) Representative images of 5 Ezh2+/2, 5 Ezh2+/+, and 3 Ezh22/2 replicate mice. Data are means 6 SEM of results in 5 Ezh2+/2 or 5 Ezh2+/+ and 3 Ezh22/2 female mice (B, D, E, G–J, L–O). *P , 0.05.

8

Vol. 31

March 2017

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

the increased bone turnover rate in Ezh2+/2 and Ezh22/2 mice in vivo is related to an increase in number and activity of osteoclasts. In the absence of a sufficient number of replicate Ezh22/2 mice, bone marrow cells were isolated from only the long bones of Ezh2+/2 and Ezh2+/+ mice and cultured with M-CSF and RANKL to induce osteoclast differentiation in vitro. Ezh2+/2-derived bone marrow cells demonstrated a greater potential to form multinucleated TRAP+ osteoclasts compared with those of Ezh2+/+ mice (Fig. 4F, G). Moreover, differentiated osteoclasts from Ezh2+/2 mice formed a greater number of dentin resorption pits compared with differentiated osteoclasts from Ezh2+/+ mice (Fig. 4H). RT-PCR analysis of cultured BMSCs isolated from EZH2-deficient mice expressed higher levels of the pro-osteoclastic factors M-csf and Rankl, but not Opg, than did the control BMSCs. (Fig. 4I–L). Heterozygous deletion of Ezh2 in early limb bud mesenchyme enhances trilineage differentiation The role of EZH2 deficiency in MSC differentiation in vitro was assessed only in the Ezh2+/2 mice because of the limited number of viable Ezh22/2 mice. BMSCs from Ezh2+/2 mice formed a greater number of CFU-Fs, with a greater proportion of these colonies expressing the preosteogenic marker alk phos than BMSCs from control mice (Fig. 5A, B). When cultured under osteogenic inductive conditions, BMSCs from 4-wk-old hind limbs of Ezh2+/2 mice produced significantly more Alizarin red mineralized deposits and higher levels of extracellular calcium, compared with Ezh2+/+ MSC (Fig. 5C, D). Real-time PCR analysis found that Ezh2+/2 MSC expressed higher levels of osteogenesis-associated genes, Runx2, and the mature osteoblast markers Ocn and secreted osteopontin Opn, compared with wildtype controls (Fig. 5E–G). Histological assessment of the long bones of Ezh22/2 mice revealed significantly greater levels of marrow adipose tissue than in the Ezh2+/2 and Ezh2+/+ mice, suggesting that complete ablation of EZH2 early in murine MSCs promotes adipogenic differentiation in vivo (Fig. 5H, I). Furthermore, MSC isolated from Ezh2+/2 mice formed significantly higher levels of Nile Red and Oil Red O-stained lipid–containing adipocytes, compared with wild-type control MSCs, when cultured under adipogenic conditions (Fig. 5J, K). Confirmatory real-time PCR analysis found that Ezh2+/2 MSC expressed higher levels of adipogenesis-associated genes Pparg2, CCAAT/enhancer binding protein (C/EBPa), and adiponectin, compared to wild-type controls (Fig. 5L–N). Under chondrogenesis-inductive conditions, MSCs isolated from Ezh2+/2 mice formed chondrogenic pellets expressing higher levels of Col II, when compared with chondrogenic pellets from wild-type MSCs (Fig. 5O). Real-time PCR analysis found that Ezh2+/2 MSC expressed lower levels of Sox9 and Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

higher levels of Col2a1 and aggrecan compared with wild-type MSCs after 28 d of chondrogenesis induction (Fig. 5P–R). These findings suggest that deletion of Ezh2 in early mesenchyme accelerates the multilineage differentiation of postnatal MSCs and that Ezh2 heterozygous deletion increases the levels of MSCs and the proportion of osteogenic precursors. Assessment of known Ehz2 target genes in Ezh2-deficient mice We analyzed the gene expression patterns of known Ezh2 target genes regulated during human MSC osteogenic differentiation (41) in long-bone samples collected from wild-type and newborn Ezh22/2 mice. The data showed that most of these genes (with the exception of FHFL1) were significantly upregulated in the long bones of Ezh2-null mice compared with wild-type controls, correlating to an increase in the expression of osteogenesis inductive factors such as Bmp2 and Runx2 (Fig. 6). These studies also confirmed that Ezh22/2 mice exhibited elevated transcript levels of the canonical Wnt signaling molecule, Wnt10b, a reported inducer of osteogenesis and target of Ezh2 (7). Furthermore, a survey of other Wnt signaling molecules identified the noncanonical Wnt molecule, Wnt5a, as being highly expressed in the long bones of Ezh22/2 mice. Therefore, there appears to be a complex regulatory role for Ezh2 during skeletal development via the suppression of osteogenic promoting factors and known inhibitors of adipogenesis, associated with Wnt/BMP signaling (Fig. 7).

DISCUSSION In the present study, we observed skeletal and trabecular bone pattern phenotypes similar to those reported for Prrx-1-Ezh2 homozygous knockout mice (7, 26). However, this is the first study to compare the differential effects of the homozygous and heterozygous Ezh2 deletion in relation to bone strength, microarchitecture of cortical and trabecular bone, effects on osteoclastogenesis, bone turnover, and differentiation capacity of BMSCs. Heterozygous deletion of Ezh2 in early mesenchyme resulted in an overall larger skeleton from E17.5 onward, with increased bone mineralization in femora, tibiae, vertebrae, and the skull. The frontal, parietal, and occipital bones were larger, with enhanced mineralization and longer and accelerated development of the fore and hind limbs. These findings suggest that in Ezh2+/2 mice, bone development occurs earlier during embryogenesis and early postnatal development and at a faster rate than in wildtype mice. A higher number of clonogenic BMSCs were identified in Ezh2+/2 derived from the hind limbs, with a greater percentage of alk phos + fibroblast colonies compared to wild-type mice, suggesting increased

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

9

Figure 3. Ezh2 deletion affects cortical bone microarchitecture and bone strength. A) Micro-CT image of newborn tibias revealed a decrease in cortical bone thickness in Ezh2+/2 and Ezh22/2 4-wk-old femora. Micro-CT quantitated BV/TV (B) and Ct.Th (C ). D) H&E-stained cortical bone revealed thinner and disorganized bone. Histomorphometric analyses of the percentage of cortical BV/TV (E) and the Ct.Th (F). G) Image of fluorescent calcein labeled (arrows) of newly formed cortical bone in Ezh2+/+, Ezh2+/2, and Ezh2/2 4-wk-old methacrylate-embedded tibiae. H–J) Cortical bone (H ) MAR and (I ) BFR/BS. J ) Histomorphometric analysis identified a greater number of osteoblasts on cortical bone perimeter (N.Ob/B.Pm). K, L) Three-point bending tests revealed decreases flexural rigidity (stiffness) (K ) and yield moment (L) in Ezh2+/2 and Ezh22/2 femora point of catastrophic failure. Representative images of 5 Ezh2+/2, 5 Ezh2+/+ or 3 Ezh22/2 replicate mice (A, D, G) . Data are means 6 SEM of results in 5 Ezh2+/2, 5 Ezh2+/+, and 3 Ezh22/2 female mice (B, C, E, F, H–L). *P , 0.05.

expansion and maturation of the skeletal progenitor pool. Furthermore, culture-expanded BMSCs exhibited a greater osteogenic potential in vitro and in vivo and expressed higher levels of osteogenic markers. These findings are in agreement with previous studies showing EZH2 to be a negative regulator of human MSC 10

Vol. 31 March 2017

osteogenic differentiation (AMSCs) (7, 19, 22). Those studies found that inhibition of EZH2 promotes the expression of regulators EGR1, HEY1, MSX2, NKX3-2, RUNX2, OPN, and OCN, all critical for osteogenic differentiation and MSC lineage determination (7, 19, 22). Moreover, EZH2 was shown to function as a repressor

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

Figure 4. Ezh2 deletion increases osteoclast differentiation and activity. A, B) Histomorphometric analysis of number of osteoclasts on trabecular (A) and cortical bone (B) perimeters. C ) Micrograph of TRAP+ positive osteoclasts (arrow) lining the trabecular and cortical bone surface. Magnification, 3400. D, E ) Ezh2+/2, Ezh22/2 levels of TRAP and CTX-1 detected in serum compared with Ezh2+/+ serum. F ) Quantitation of the number of multinucleated TRAP+ osteoclast differentiated in vitro. G) TRAP+ multinucleated osteoclasts (arrows). H ) Differentiated osteoclasts formed more pits (arrow) in whale dentine samples than in Ezh2+/+ bone marrow cells. I–L) Ezh2+/2 BMSCs extracted from the hind limbs of 4-wk-old mice expressed higher levels of M-Csf and Rankl, whereas Opg expression was unchanged when compared with Ezh2+/+ BMSCs. C ) Representative image of 5 Ezh2+/2, 5 Ezh2+/+, or 3 Ezh22/2 replicate mice or (G (both images), H, top) representative images of 3 Ezh2+/2 and 3 Ezh2+/+ replicate mice. D, E ) Data are means 6 SEM of results (A, B, D, E) in 5 Ezh2+/2, 5 Ezh2+/+, and 3 Ezh22/2 female mice; or (F, I-L) 3 Ezh2+/2 and 3 Ezh2+/+ female mice. *P , 0.05.

of osteogenic differentiation through the direct methylation of genes critical for osteogenic lineage specification. More recently, bioinformatic analyses of gene expression and ChIPseq data sets of human MSC Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

osteogenic differentiation revealed several novel Ezh2 target molecules mediating MSC osteogenesis (41). Although some of these genes (summarized in Fig. 7) are known effectors of BMP-2 and canonical Wnt signaling,

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

11

Figure 5. Ezh2 deletion increases trilineage differentiation. A) MSCs extracted from the hind limbs had greater CFU-F potential than Ezh2+/+ control MSCs. B) Increased percentages of 100% alk phos+ fibroblast colonies were found in Ezh2+/2 mice vs. Ezh2+/+ control colonies. C ) Ezh2+/2 hind limb MSCs produced more mineral at 7 d of osteogenic induction than did Ezh2+/+ control MSCs. Magnification, 340. D) Osteogenic differentiated Ezh2+/2 MSCs produced more extracellular calcium when normalized to DNA (continued on next page)

12

Vol. 31 March 2017

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

Figure 6. Assessment of known Ezh2 target genes. After the bone marrow was flushed, total RNA was extracted from fixed femora and tibiae of newborn Ezh22/2 and Ezh2+/+mice. qPCR analysis found that Ezh22/2 bone samples expressed significantly higher levels of the osteogenic inductive factors, Runx2, Mx1, Bmp2, Wnt10b, and Wnt5a and several other putative Ezh2 target genes Hopx, Myamd, and Ror2, with the exception of Fhl1, when compared to Ezh2+/+ mice. Data are means 6 SEM of results in 3 Ezh22/2 and 5 Ezh2+/+ female mice. *P , 0.05.

the present study identified Wnt5a as a putative target of Ezh2. Previous studies have shown that Wnt5a binds to receptor tyrosine kinase-like orphan receptor 2 (Ror2), to promote osteogenesis (42). Both Wnt5a and Ror2 gene expression can be induced by BMP-2, but their ability to promote osteogenesis appears to acting via a Smad-independent manner. Collectively, these findings imply that Ezh2 normally functions to suppress osteogenesis and known inhibitors of adipogenesis, which are associated with Wnt/BMP signaling (Fig. 7). In vivo, complete ablation of Ezh2 in early limb bud mesenchymal cells resulted in limb and cranial patterning defects related to altered expression of the Hox and zinc finger genes (7, 26); derepression and accelerated expression of developmental, bone-related extracellular matrix proteins; and cyclin-dependent kinase inhibitors critical for postproliferative cell growth arrest (7, 26). The present study identified that Ezh2 ablation resulted in trabecular bone patterning

throughout the entire hind limbs of newborn and 4wk-old mice, whereas the trabeculae in heterozygous and wild-type controls were restricted to the metaphyseal regions of the bone. These observations suggest that ablation of EZH2 not only affects the size of the limbs but also affects the localization of trabeculae within the hind limb bones and is highly dependent on the levels of EZH2 present during embryonic and postnatal development. In addition, structural differences in bone size and morphology can alter the mechanical loading on bone. Observed differences in trabecular patterning in EZH2 homozygous mice could be attributable to structural and morphological differences in femur architecture. Wolff’s law states that skeletal transformation is dependent on the exertion of pressures from outside the animal. The cells within bone senses loading strains through mechanotransduction, and these signals can translate into information that elicits a response within the signaled cells or surrounding

than did Ezh2+/+ control MSCs. E–G) Ezh2+/2 MSCs expressed higher levels of Runx2, Ocn, and Opn than did Ezh2+/+ control MSCs. H ) Image of the increase in marrow fat observed in Ezh22/2 compared with Ezh2+/+ and Ezh2+/2 mice. I ) Analysis of marrow fat in the secondary spongiosa revealed a significant increase in lipid present in Ezh22/2 vs. Ezh22/2 and Ezh2+/2 mice. J ) Nile Red– and DAPI-stained 7 d adipogenic differentiated MSCs revealed an increase in adipogenic potential of Ezh2+/2 BMSCs vs. Ezh2+/+-differentiated MSCs. Magnification, 320. K ) A significant increase in lipid was found when Nile Red–stained lipid was normalized to DAPI-stained nuclei. L–N ) Adipogenic differentiated Ezh2+/2 MSCs expressed higher levels of Pparg2, C/ebpa, and adiponectin (AdipQ) vs. Ezh2+/+ adipogenic differentiated MSCs. O) Col II and isotype control IB5-stained cartilage pellets. Differentiated cartilage pellets from Ezh2+/2 MSC expressed lower levels of Sox9 (P ) and higher levels of Col2a1 (Q and Acan (R), compared with Ezh2+/+ differentiated MSCs. C, H, J, O) Representative images of 3 Ezh2+/2and 3 Ezh2+/+ replicate mice. Data are means 6 SEM of results in 5 Ezh2+/2 or 5 Ezh2+/+ female and 3 Ezh22/2 female mice (A, B, I ) or of 3 Ezh2+/2and 3 Ezh2+/+ mice (D-G, I, K-N, P-R). *P , 0.05. Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

13

Figure 7. Proposed Ezh2 regulation of BMSC adipogenic and osteogenic differentiation. Ezh2-mediated gene suppression via H3K27me3 (K27) of the known adipogenic inhibitors Bmp, Smad1/5, Wnt10b, and Runx2; the osteogenic factors Runx2, Bmp, Smad1/5, Wnt5a, Ror2, and Mx1 (solid lines); and the proposed osteogenic inductive factors, Myamd, Fhl1, and Hopx (dotted line).

cells. We suggest that the differences in bone morphology of Ezh2+/2 and Ezh22/2 could result in altered mechanical loading on the femora promoting activation of osteogenic cells to initiate bone modeling or recruitment of osteoclast to facilitate with resorptive modeling (43). The skeletal defects described in heterozygous deletion of Ezh2 in early mesenchyme tissue is similar to that seen in human patients who carry autosomal dominant mutations in Ezh2 in a condition known as Weaver syndrome (44,46). The bones of people with Weaver syndrome grow and develop more quickly, both before and after birth, and adults are generally taller, display clinodactyly, a larger head (macrocephaly), and craniosynostosis (44). Their general overgrowth and advanced bone age suggests that the skeletal stem cell pool undergoes premature cellular senescence and maturation, as we have shown for cultured human BMSCs after knockdown of Ezh2 (47). Therefore, the Ezh2 heterozygous mouse could provide a promising murine model for Weaver syndrome, which could allow us to understand the role of EZH2 during skeletal aging. Through calcein staining of newly formed bone, we confirmed that loss of EZH2 accelerated bone formation, with increased trabecular and cortical bone MAR in heterozygous and homozygous Ezh2 knockouts. We suggest the reduced number of trabeculae and increase in trabeculae spacing in the secondary 14

Vol. 31 March 2017

spongiosa of Ezh22/2 is a consequence of altered trabecular patterning. However, an increase in the number of osteoblasts lining the bone surface was also present on the periosteal and endosteal surfaces of the cortical bone of Ezh2+/2 and Ezh22/2 mice, with an increase in MAR and BFR/BS values. Furthermore, EZH2-deficient mice expressed elevated serum levels of OCN, a marker of osteoblast activity, and BMSCs isolated from Ezh2+/2 mice exhibited higher osteogenic differentiation potential in vitro compared with BMSCs isolated from wildtype mice. Biomechanical testing of femora revealed that the Ezh22/2 bones were less stiff and exhibited less load before catastrophic failure, which could be attributable to thinner and disorganized cortical bone and to the abundance of marrow adipocytes present within the marrow cavity of the Ezh22/2 mice. The normal cortical lamellar bone pattern was absent in the Ezh22/2 mice, with cortical bone resembling woven bone with collagen fibers laid down in a disorganized manner. This disorganized bone structure in homozygous Ezh2 mice is similar to that seen in Paget’s syndrome, where osteoblast and osteoclast activities are increased, leading to formation of disorganized woven bone (48–50). The cause of Paget’s disease is still unknown; however, it is conceivable that EZH2 plays an important role in regulating osteoblast and osteoclast interaction in the disease.

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

It is well established that EZH2 acts as a repressor of Runx2 transcription, and inhibition of EZH2 allows the activation of Runx2, promoting osteogenic differentiation. To this end, transgenic mice overexpressing RUNX2 exhibited altered osteoblast maturation and increased cortical bone reabsorption and bone formation inducing high bone turnover rates (51). Enhanced expression of RUNX2 could explain the increased bone turnover rates observed in Ezh2 +/2 and Ezh22/2 mice. Furthermore, deletion of Ezh2 in MSCs may trigger elevated osteoclast activity, either through increased paracrine stimulation of osteoclast development or their recruitment to bone surfaces, which could counteract any increases in BFRs. This study suggests that the degree of osteoclast stimulation could be dependent on the levels of Ezh2 expressed in MSCs. This notion is supported by the increased number of TRAP+ multinucleated osteoclasts observed on the bone surfaces and associated elevation in serum levels of the resorption markers, TRAP and CTX-1, in 4-wk-old Ezh2+/2 and Ezh22/2 mice, compared with wild-type controls. Our studies also showed an increase in TRAP+ multi-nucleated osteoclasts in vitro isolated from Ezh2+/2 bone marrow cells compared with Ezh2+/+ control mice. Moreover, in the present study cultured Ezh2+/2 MSCs exhibited increased gene expression levels of the osteoclastinductive factors M-csf and Rankl; however, the levels of the RANKL decoy receptor, Opg, were unchanged compared with those of wild-type mice. This finding may involve the derepression of a secreted factor that activates osteoclasts when EZH2 expression or activity is suppressed in MSCs. We must note that complete ablation significantly increased TRAP and CTX-1 levels in the serum over the elevated levels seen in Ezh2 +/2 and lower levels present in control mice. We suggest that the increases in number of osteoclasts and in TRAP and CTX-1 levels are a result of the increased number of osteoblasts and BFR. This reciprocal increase in bone turnover and remodelling could be attributable to the bone’s trying to maintain bone homeostasis. Complete ablation results in bone remodelling and turnover exceeding bone formation and leading to an osteoporotic-like phenotype. The adipogenic potential of Ezh2 +/2 BMSC in vitro was found to be enhanced, correlating with the striking increase in bone marrow adipose tissue in situ, where the whole bone diaphysis was completely filled with fat-laden adipocytes in newborn and 4-wk-old Ezh2-deficient mice. However, these observations were contrary to studies showing that adipogenesis is decreased in human BMSCs and AMSCs after Ezh2 knockdown (7, 22) or when Ezh2 has been deleted specifically in preadipocytes in mice (18). Furthermore, inhibition of EZH2 methyltransferase activity by DZNep in an ovariectomized mouse model of osteoporosis reduced bone loss and adipogenic formation during osteoporosis in adult mice (23). These findings suggest that the function of EZH2 in adipogenesis is stage specific, Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

where deletion of Ezh2 in the early limb bud mesenchymal stage seems to result in a general nonspecific activation of BMSC maturation, whereas inhibition of EZH2 appears to inhibit adipogenesis in adultderived MSC populations and adult ovariectomized mice. Although the effect of EZH2 on adipogenesis has been shown to be caused by its ability to inhibit Wnt signaling, removal of EZH2 at the early mesenchyme stage may alleviate repression of adipogenic genes directly or the expression of another pathway that influences adipogenesis, resulting in increased adipocytes (18, 23). Furthermore, the chondrogenic differentiation capacity of Ezh2+/2 BMSCs was enhanced with increased expression of the more highly mature chondrogenic markers aggrecan and Col II. This accelerated chondrogenic potential was associated with the downregulation of the chondrogenic transcription factor Sox9 expressed at the prechondroblast stage, implying a possible depletion of chondroprogenitors or rapid maturation of chondroblasts. Therefore, we suggest that the increased skeletal size in Ezh2+/2 mice is a result of enhanced chondrogenic maturation. It is plausible that complete ablation of EZH2 impairs chondrogenic differentiation of the limbs similar to the impairments seen during neural crest cartilage development (21). Further studies investigating the mechanism of how EZH2 regulates endochondral ossification and intramembranous ossification will be critical for a greater understanding of how bone development is epigenetically regulated. A better understanding of the molecular pathways targeted by EZH2 activity at different stages of BMSC maturation, during BMSC maintenance and cell fate determination, may help identify novel pharmacological targets for the treatment of different bone diseases and possibly congenital bone disorders, as recently demonstrated in a preclinical model of osteoporosis (23). ACKNOWLEDGMENTS The authors thank the University of Adelaide’s Microscopy Service for their technical support with micro-CT imaging and analysis. The authors also thank David Haynes, Kencana Dharmapatni, and Julia Kuliwaba for the use of their laboratory facilities (Adelaide Medical School, University of Adelaide, Adelaide, SA, Australia). This study was partially supported by National Health and Medical Research Council Project Grant APP1046053 (to S.G.), Fellowship APP1042677 (to S.G.), Australia Postgraduate Award (to S.H.), and Dawes Top-up scholarship (to S.H.). D.C. and S.G. are co-senior authors. The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS S. Hemming, D. Cakouros, and A. Arthur performed the research; S. Hemming, D. Cakouros, and S. Gronthos wrote the paper; K. Vandye, J. Codrington, and A. Zannettino reviewed the manuscript; D. Cakouros, K. Vandyke,

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

15

J. Codrington, A. Zannettino, and S. Gronthos contributed reagents and analytic tools; and S. Hemming, D. Cakouros, K. Vandyke, J. Codrington, A. Arthur, A. Zannettino, and S. Gronthos analyzed the data.

21.

22.

REFERENCES 1. Logan, M., Martin, J. F., Nagy, A., Lobe, C., Olson, E. N., and Tabin, C. J. (2002) Expression of Cre recombinase in the developing mouse limb bud driven by a Prxl enhancer. Genesis 33, 77–80 2. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A. L., and Karsenty, G. (1997) Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89, 747–754 3. Bi, W., Deng, J. M., Zhang, Z., Behringer, R. R., and de Crombrugghe, B. (1999) Sox9 is required for cartilage formation. Nat. Genet. 22, 85–89 4. Tontonoz, P., Hu, E., Graves, R. A., Budavari, A. I., and Spiegelman, B. M. (1994) mPPAR gamma 2: tissue-specific regulator of an adipocyte enhancer. Genes Dev. 8, 1224–1234 5. Lee, T. I., Jenner, R. G., Boyer, L. A., Guenther, M. G., Levine, S. S., Kumar, R. M., Chevalier, B., Johnstone, S. E., Cole, M. F., Isono, K., Koseki, H., Fuchikami, T., Abe, K., Murray, H. L., Zucker, J. P., Yuan, B., Bell, G. W., Herbolsheimer, E., Hannett, N. M., Sun, K., Odom, D. T., Otte, A. P., Volkert, T. L., Bartel, D. P., Melton, D. A., Gifford, D. K., Jaenisch, R., and Young, R. A. (2006) Control of developmental regulators by Polycomb in human embryonic stem cells. Cell 125, 301–313 6. Kouzarides, T. (2007) SnapShot: Histone-modifying enzymes. Cell 128, 80.e1–802.e2 7. Dudakovic, A., Camilleri, E. T., Xu, F., Riester, S. M., McGee-Lawrence, M. E., Bradley, E. W., Paradise, C. R., Lewallen, E. A., Thaler, R., Deyle, D. R., Larson, A. N., Lewallen, D. G., Dietz, A. B., Stein, G. S., Montecino, M. A., Westendorf, J. J., and van Wijnen, A. J. (2015) Epigenetic control of skeletal development by the histone methyltransferase Ezh2. J. Biol. Chem. 290, 27604–27617 8. Denell, R. E. (1978) Homoeosis in Drosophila, II: a genetic analysis of polycomb. Genetics 90, 277–289 9. Margueron, R., and Reinberg, D. (2011) The polycomb complex PRC2 and its mark in life. Nature 469, 343–349 10. Vandamme, J., V¨olkel, P., Rosnoblet, C., Le Faou, P., and Angrand, P.-O. (2011) Interaction proteomics analysis of polycomb proteins defines distinct PRC1 complexes in mammalian cells. Mol. Cell. Proteomics 10, M110.002642 11. Francis, N. J., Kingston, R. E., and Woodcock, C. L. (2004) Chromatin compaction by a polycomb group protein complex. Science 306, 1574–1577 12. Bulger, M., and Groudine, M. (1999) Looping versus linking: toward a model for long-distance gene activation. Genes Dev. 13, 2465–2477 13. Aloia, L., Di Stefano, B., and Di Croce, L. (2013) Polycomb complexes in stem cells and embryonic development. Development 140, 2525–2534 14. Richly, H., Aloia, L., and Di Croce, L. (2011) Roles of the Polycomb group proteins in stem cells and cancer. Cell Death Dis. 2, e204 15. De Haan, G., and Gerrits, A. (2007) Epigenetic control of hematopoietic stem cell aging: the case of Ezh2. Ann. NY Acad. Sci. 1106, 233–239 16. Caretti, G., Di Padova, M., Micales, B., Lyons, G. E., and Sartorelli, V. (2004) The polycomb Ezh2 methyltransferase regulates muscle gene expression and skeletal muscle differentiation. Genes Dev. 18, 2627–2638 17. Sher, F., R¨ossler, R., Brouwer, N., Balasubramaniyan, V., Boddeke, E., and Copray, S. (2008) Differentiation of neural stem cells into oligodendrocytes: involvement of the polycomb group protein Ezh2. Stem Cells 26, 2875–2883 18. Wang, L., Jin, Q., Lee, J. E., Su, I. H., and Ge, K. (2010) Histone H3K27 methyltransferase Ezh2 represses Wnt genes to facilitate adipogenesis. Proc. Natl. Acad. Sci. USA 107, 7317–7322 19. Wei, Y., Chen, Y.-H., Li, L.-Y., Lang, J., Yeh, S.-P., Shi, B., Yang, C.-C., Yang, J.-Y., Lin, C.-Y., Lai, C.-C., and Hung, M.-C. (2011) CDK1dependent phosphorylation of EZH2 suppresses methylation of H3K27 and promotes osteogenic differentiation of human mesenchymal stem cells. Nat. Cell Biol. 13, 87–94 20. Kamminga, L. M., Bystrykh, L. V., de Boer, A., Houwer, S., Douma, J., Weersing, E., Dontje, B., and de Haan, G. (2006) The Polycomb 16

Vol. 31 March 2017

23.

24. 25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

38.

group gene Ezh2 prevents hematopoietic stem cell exhaustion. Blood 107, 2170–2179 Schwarz, D., Varum, S., Zemke, M., Scholer, A., Baggiolini, A., Draganova, K., Koseki, H., Schubeler, D., and Sommer, L. (2014) Ezh2 is required for neural crest-derived cartilage and bone formation. Development 141, 8672877 Hemming, S., Cakouros, D., Isenmann, S., Cooper, L., Menicanin, D., Zannettino, A., and Gronthos, S. (2014) EZH2 and KDM6A act as an epigenetic switch to regulate mesenchymal stem cell lineage specification. Stem Cells 32, 802–815 Jing, H., Liao, L., An, Y., Su, X., Liu, S., Shuai, Y., Zhang, X., and Jin, Y. (2016) Suppression of EZH2 prevents the shift of osteoporotic MSC fate to adipocyte and enhances bone formation during osteoporosis. Mol. Ther. 24, 217–229 Huang, X. J., Wang, X., Ma, X., Sun, S. C., Zhou, X., Zhu, C., and Liu, H. (2014) EZH2 is essential for development of mouse preimplantation embryos. Reprod. Fertil. Dev. 26, 1166–1175 O’Carroll, D., Erhardt, S., Pagani, M., Barton, S. C., Surani, M. A., and Jenuwein, T. (2001) The polycomb-group gene Ezh2 is required for early mouse development. Mol. Cell. Biol. 21, 4330–4336 Wyngaarden, L. A., Delgado-Olguin, P., Su, I. H., Bruneau, B. G., and Hopyan, S. (2011) Ezh2 regulates anteroposterior axis specification and proximodistal axis elongation in the developing limb. Development 138, 3759–3767 Hirabayashi, Y., Suzki, N., Tsuboi, M., Endo, T. A., Toyoda, T., Shinga, J., Koseki, H., Vidal, M., and Gotoh, Y. (2009) Polycomb limits the neurogenic competence of neural precursor cells to promote astrogenic fate transition. Neuron 63, 600–613 Cakouros, D., Isenmann, S., Hemming, S. E., Menicanin, D., Camp, E., Zannetinno, A. C. W., and Gronthos, S. (2015) Novel basic helixloop-helix transcription factor Hes4 antagonizes the function of twist1 to regulate lineage commitment of bone marrow stromal/stem cells. Stem Cells Dev. 24, 1297–1308 Parfitt, A. M., Drezner, M. K., Glorieux, F. H., Kanis, J. A., Malluche, H., Meunier, P. J., Ott, S. M., and Recker, R. R. (1987) Bone histomorphometry: standardization of nomenclature, symbols, and units; report of the ASBMR Histomorphometry Nomenclature Committee. J. Bone Miner. Res. 2, 595–610 Vandyke, K., Dewar, A. L., Diamond, P., Fitter, S., Schultz, C. G., Sims, N. A., and Zannettino, A. C. (2010) The tyrosine kinase inhibitor dasatinib dysregulates bone remodeling through inhibition of osteoclasts in vivo. J. Bone Miner. Res. 25, 1759–1770 Bouxsein, M. L., Boyd, S. K., Christiansen, B. A., Guldberg, R. E., Jepsen, K. J., and M¨uller, R. (2010) Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J. Bone Miner. Res. 25, 1468–1486 Arthur, A., Panagopoulos, R. A., Cooper, L., Menicanin, D., Parkinson, I. H., Codrington, J. D., Vandyke, K., Zannettino, A. C., Koblar, S. A., Sims, N. A., Matsuo, K., and Gronthos, S. (2013) EphB4 enhances the process of endochondral ossification and inhibits remodeling during bone fracture repair. J. Bone Miner. Res. 28, 926–935 Schriefer, J. L., Robling, A. G., Warden, S. J., Fournier, A. J., Mason, J. J., and Turner, C. H. (2005) A comparison of mechanical properties derived from multiple skeletal sites in mice. J. Biomech. 38, 467–475 Jepsen, K. J., Silva, M. J., Vashishth, D., Guo, X. E., and van der Meulen, M. C. (2015) Establishing biomechanical mechanisms in mouse models: practical guidelines for systematically evaluating phenotypic changes in the diaphyses of long bones. J. Bone Miner. Res. 30, 951–966 Vandyke, K., Dewar, A. L., Fitter, S., Menicanin, D., To, L. B., Hughes, T. P., and Zannettino, A. C. (2009) Imatinib mesylate causes growth plate closure in vivo. Leukemia 23, 2155–2159 McNeil, P. J., Durbridge, T. C., Parkinson, I. H., and Moore, R. J. (1997) Simple method for the simultaneous demonstration of formation and resorption in undecalcified bone embedded in methyl methacrylate. J. Histotechnol. 20, 307–311 Gronthos, S., Zannettino, A. C. W., Hay, S. J., Shi, S., Graves, S. E., Kortesidis, A., and Simmons, P. J. (2003) Molecular and cellular characterisation of highly purified stromal stem cells derived from human bone marrow. J. Cell Sci. 116, 1827–1835 Isenmann, S., Arthur, A., Zannettino, A. C., Turner, J. L., Shi, S., Glackin, C. A., and Gronthos, S. (2009) TWIST family of basic helix-loop-helix transcription factors mediate human

The FASEB Journal x www.fasebj.org

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

HEMMING ET AL.

39.

40.

41.

42.

43. 44.

mesenchymal stem cell growth and commitment. Stem Cells 27, 2457–2468 Fitter, S., Vandyke, K., Schultz, C. G., White, D., Hughes, T. P., and Zannettino, A. C. (2013) Plasma adiponectin levels are markedly elevated in imatinib-treated chronic myeloid leukemia (CML) patients: a mechanism for improved insulin sensitivity in type 2 diabetic CML patients? J. Clin. Endocrinol. Metab. 95, 3763–3767 Cantley, M. D., Fairlie, D. P., Bartold, P. M., Rainsford, K. D., Le, G. T., Lucke, A. J., Holding, C. A., and Haynes, D. R. (2011) Inhibitors of histone deacetylases in class I and class II suppress human osteoclasts in vitro. J. Cell. Physiol. 226, 3233–3241 Hemming, S., Cakouros, D., Vandyke, K., Davis, M. J., Zannettino, A. C., and Gronthos, S. (2016) Identification of Novel EZH2 targets regulating osteogenic differentiation in mesenchymal stem cells. Stem Cells Dev. 25, 909–921 Nemoto, E., Ebe, Y., Kanaya, S., Tsuchiya, M., Nakamura, T., Tamura, M., and Shimauchi, H. (2012) Wnt5a signaling is a substantial constituent in bone morphogenetic protein-2mediated osteoblastogenesis. Biochem. Biophys. Res. Commun. 422, 627–632 Pearson, O. M., and Lieberman, D. E. (2004) The aging of Wolff’s “law”: ontogeny and responses to mechanical loading in cortical bone. Am. J. Phys. Anthropol. 39 (Suppl.), 63–99 FORGE Canada Consortium. (2012) Mutations in EZH2 cause Weaver syndrome. Am. J. Hum. Genet. 90, 110–118

Ezh2 REGULATES MURINE SKELETAL DEVELOPMENT

45. Jalaguier, J., Montoya, F., Germain, M., and Bonnet, H. (1983) [Acceleration of bone maturation and dysmorphic syndrome in 2 siblings (Marshall-Weaver syndrome)]. J. Genet. Hum. 31(Suppl 5), 385–395 46. Mikalef, P., Beslikas, T., Gigis, I., Bisbinas, I., Papageorgiou, T., and Christoforides, I. (2010) Weaver syndrome associated with bilateral congenital hip and unilateral subtalar dislocation. Hippokratia 14, 212–214 47. Cakouros, D., Isenmann, S., Cooper, L., Zannettino, A., Anderson, P., Glackin, C., and Gronthos, S. (2012) Twist-1 induces Ezh2 recruitment regulating histone methylation along the Ink4A/Arf locus in mesenchymal stem cells. Mol. Cell. Biol. 32, 1433–1441 48. Hosking, D. J. (1981) Paget’s disease of bone. Br. Med. J. (Clin. Res. Ed.) 283, 686–688 49. Sabharwal, R., Gupta, S., Sepolia, S., Panigrahi, R., Mohanty, S., Subudhi, S. K., and Kumar, M. (2014) An insight in to Paget’s disease of bone. Niger. J. Surg. 20, 9–15 50. Yates, A. J. (1988) Paget’s disease of bone. Baillieres Clin. Endocrinol. Metab. 2, 267–285 51. Geoffroy, V., Kneissel, M., Fournier, B., Boyde, A., and Matthias, P. (2002) High bone resorption in adult aging transgenic mice overexpressing cbfa1/runx2 in cells of the osteoblastic lineage. Mol. Cell. Biol. 22, 6222–6233 Received for publication June 30, 2016. Accepted for publication November 22, 2016.

Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

17

EZH2 deletion in early mesenchyme compromises postnatal bone microarchitecture and structural integrity, and accelerates remodeling Sarah Hemming, Dimitrios Cakouros, John Codrington, et al. FASEB J published online December 1, 2016 Access the most recent version at doi:10.1096/fj.201600748R

Supplemental Material

http://www.fasebj.org/content/suppl/2016/12/01/fj.201600748R.DC1.html

Subscriptions

Information about subscribing to The FASEB Journal is online at http://www.faseb.org/The-FASEB-Journal/Librarian-s-Resources.aspx

Permissions

Submit copyright permission requests at: http://www.fasebj.org/site/misc/copyright.xhtml

Email Alerts

Receive free email alerts when new an article cites this article - sign up at http://www.fasebj.org/cgi/alerts

© FASEB Downloaded from www.fasebj.org to IP 163.1.29.14. The FASEB Journal Vol., No. , pp:, January, 2017

Supplementary Figure 1. Confirmation of the generation of Ezh2 conditional knockout mice. (A) Agarose gel depicting PCR analysis of genomic DNA for Ezh2 floxed alleles (450bp). Heterozygous mice expressing a wildtype (400bp) and floxed Ezh2 allele (450bp). Primers to Cre recombinase detecting Cre at 100bp. (B) Western Blot analysis of Ezh2 and H3K27me3 protein in Ezh2+/compared with wildtype Ezh2+/+ 4 week BMSC. (C) Ezh2 transcript levels assessed by RT-PCR in 4 week BMSC from Ezh2+/- and Ezh2+/+ mice. (D&E) Immunohistochemistry of 4 week old Ezh2+/+, Ezh2+/- and Ezh2-/- mice for presence of H3K27me3. Arrows identify proliferative chondrocytes and blood cells in the femora (D). Arrows identify osteocytes within the bone and osteoblasts lining the bone surface in the femora (E). Representative image of 3 Ezh2+/+or Ezh2+/- mice. (D) Representative image of 4 Ezh2+/+, Ezh2+/- mice and 3 Ezh2-/- independent mice. Graphs depict mean ± SEM of 3 independent mice per genotype (C) or representative of 3 independent mice per genotype (D). Statistical significance represented as * (p < 0.05).

Supplementary Figure 2. Ezh2 deletions affect the size and skeletal development in E17.5 and newborn mice. (A) Alizarin red and alcian blue staining of whole E17.5 embryos. (B) Images of the skeletal phenotype of Ezh2+/+, Ezh2+/-, Ezh2-/- mice. (C) Weights of female newborn mice. (D) Micro-CT Avizo generated 3D image of newborn skeletons assessing vertebrae and skull formation. (E) Alizarin red and alcian blue stained lumbar (L) and sacral (S) vertebrae in Ezh2-/- mice lacking third sacral vertebrae (S3). (F) Alizarin red and alcian blue stained skulls depicting differences in cranial bone patterning at E17.5. (G) AVIZO 3D imaging of the skulls revealing craniosynostosis in Ezh2-/- and increase in parietal and occipital bone size in Ezh2+/- newborns. Alizarin red staining showing differences in length of (H) E17.5 fore limbs, (I) newborn fore limbs, (J) E17.5 hind limbs, (K) newborn hind limbs. (L) Image J quantitation of newborn scapular length, (M) Humerus length, (N) Ulna length, (O) Radius length, (P) Femur length, (Q) Tibia length and (R) Fibular length. Representative images of 4 replicate embryos or newborn per genotype or representative image of 3 replicate embryos or newborn per genotype (A,B,D,E,F,G,H,I,J,K). Graphs depict mean ± SEM of 5 Ezh2+/- or Ezh2+/+ female and male and 3 Ezh2-/- female mice (C) or the mean ± SEM of 4 independent mice per genotype (D). Statistical significance represented as * (p < 0.05).

Supplementary Figure 3. Ezh2 deletion effects 4 week skeletal size and for and hind limb morphology. (A) Analysis of weights between Ezh2+/-, Ezh2-/- female mice and Ezh2+/+ control males and females. (B) X-ray images of 4 week old mice showing differences in skeletal size between Ezh2+/+, Ezh2+/-, Ezh2-/- mice. X-ray images of (C) 4 week old fore limbs comparing Ezh2+/- and Ezh2-/- mice compared with Ezh2+/+ control mice, and (D) humerus showing differences in size and morphology. (E) Tibial length differences in Ezh2+/- and reduced in Ezh2-/- mice. (F) AVIZO 3D imaging of the tibiae revealing differences in tibial length and fibular morphology. (G) Quantitation of femora lengths. (H) Micro-CT images of 4 week old femora. (B, C, D & F) are representative images of 5 Ezh2+/-, 5 Ezh2+/+ or 3 Ezh2-/replicate mice. Graphs depict mean ± SEM of 5 Ezh2+/- or 5 Ezh2+/+ female 3 Ezh2-/- female mice (A, E, & G). Statistical significance represented as * (p < 0.05).

Supplementary Figure 4. Ezh2 deletions affect the size of the growth plates and the cartilage zone size. (A) Safranin O stained newborn growth plates. Arrows identify resting, proliferative and hypertrophic cartilage in images. (B) Safranin O stained growth plates of 4 week old Ezh2+/+, Ezh2+/-, Ezh2-/-. Histomorphometric quantitation of newborn growth plate zones, (C) growth plate length, (D) resting cartilage zone length, (E) proliferative cartilage zone length, (F) Hypertrophic cartilage zone length. Histomorphometric quantitation of 4 week old growth plates (G) growth plate length, (H) resting cartilage zone, (I) proliferative zone and (J) hypertrophic zone. (A&B) are representative image of 5 Ezh2+/-, 5 Ezh2+/+ or 3 Ezh2-/- replicate mice. Graphs depict mean ± SEM of 5 Ezh2+/- or 5 Ezh2+/+ female 3 Ezh2-/- female mice (C-J). Statistical significance represented as * (p < 0.05).