Mesenchyme-specific overexpression of nucleolar protein 66 in mice inhibits skeletal growth and bone formation

Abstract

Previous studies showed that nucleolar protein 66 (NO66), the Jumonji C-domain-containing histone demethylase for methylated histone H3K4 and H3K36 (H3K36me), negatively regulates osteoblast differentiation in vitro by inhibiting the activity of transcription factor osterix (Osx). However, whether NO66 affects mammalian skeletogenesis in vivo is not yet known. Here, we generated transgenic (TG) mice overexpressing a flag-tagged NO66 transgene driven by the Prx1 (paired related homeobox 1) promoter. We found that NO66 overexpression in Prx1-expressing mesenchymal cells inhibited skeletal growth and bone formation. The inhibitory phenotype was associated with >50% decreases in chondrocyte/osteoblast proliferation and differentiation. Moreover, we found that in bones of NO66-TG mice, expression of Igf1, Igf1 receptor (Igf1r), runt-related transcription factor 2, and Osx was significantly down-regulated (P < 0.05). Consistent with these results, we observed >50% reduction in levels of phosphorylated protein kinase B (Akt) and H3K36me3 in bones of NO66-TG mice, suggesting an inverse correlation between NO66 histone demethylase and the activity of IGF1R/Akt signaling. This correlation was further confirmed by in vitro assays of C2C12 cells with NO66 overexpression. We propose that the decrease in the IGF1R/Akt signaling pathway in mice with mesenchymal overexpression of NO66 may contribute in part to the inhibition of skeletal growth and bone formation.—Chen, Q., Zhang, L., de Crombrugghe, B., Krahe, R. Mesenchyme-specific overexpression of nucleolar protein 66 in mice inhibits skeletal growth and bone formation.

The majority of the mammalian skeleton is composed of 2 types of tissues: cartilage and bone. Cartilage is made by chondrocytes that are required for the longitudinal growth of bones. Bone forms through 2 distinct processes: intramembranous and endochondral ossification. In intramembranous ossification, bones form directly from the differentiation of mesenchymal progenitor cells, which then differentiate into cells of osteoblast lineage including preosteoblasts, osteoblasts, and osteocytes. However, in endochondral ossification, osteochondroprogenitors, which derive from mesenchymal progenitors, segregate into either chondrocytes that form a cartilage template or osteoblast precursors, which differentiate to form bone tissue that replaces the cartilage template. The progression through mesenchymal condensation, chondrocyte differentiation, and osteogenic-lineage commitment involves 3 key transcription factors: SRY-related HMG-box 9 (Sox9), runt-related transcription factor 2 (Runx2), and osterix (Osx). During endochondral ossification, Sox9 is required for mesenchymal progenitor condensation and expression of type II collagen (Col2) α1 (1–3). Runx2 has a dual role in the regulation of osteogenic-lineage commitment and the differentiation of mature chondrocytes (4–8). Runx2-null (Runx2^−/−) mice showed a complete absence of skeletal elements, markedly reduced formation of hypertrophic chondrocytes, and decreased expression of type X collagen (Col10) α1 (4, 5, 8). Osx is a C2H2-type zinc finger-containing transcription factor that is highly expressed in cells of the osteogenic lineage and is essential for osteoblastogenesis (9, 10). In Osx-null (Osx^−/−) mice, expression of bone formation markers, such as type I collagen (Col1), alkaline phosphatase (Alp), bone sialoprotein (Bsp), and Osteocalcin (Oc), was inhibited, and formation of both endochondral and intramembranous bones was abolished (9). Bone formation and growth are also controlled by a broad signaling network including IGF, bone morphogenetic protein (BMP), and members of the Wnt family (11–15). Previous studies showed that dysregulation of these signaling molecules results in abnormal bone growth (16–19).

In addition to the transcription factors and signaling molecules, the Jumonji-domain-containing histone demethylases have been shown to be of great importance in the control of osteogenic differentiation (20, 21). Nucleolar protein 66 (NO66) is a Jumonji C-domain-containing protein and is conserved in eukaryotes. It was initially described as a component of cell nucleoli (22) and reported to be associated with c-Myc (23). Previous studies showed that NO66 binds to Osx and inhibits its transcriptional activity in the promoter region of the Bsp gene, a downstream target of Osx (24). NO66 harbors histone demethylase activity for methylated H3K4 (H3K4me) and H3K36 (H3K36me), 2 marks of active chromatin (24). Overexpression of NO66 in COS7 cells reduced the staining intensity of both H3K4me3 and H3K36me3 (24). Moreover, NO66 was found to be recruited to the Polycomb Repressive Complex 2 during embryonic stem cell differentiation, leading to the loss of H3K36me3 and transcriptional silencing of previously active genes (25), highlighting an important role for NO66 in histone demethylation-mediated gene regulation. Nevertheless, to date, it remains unclear whether excess NO66 in vivo could disrupt gene regulation and organogenesis.

Here, we generated transgenic (TG) mice overexpressing a flag-tagged NO66 transgene driven by the Prx1 (paired related homeobox 1) promoter to study the in vivo role of NO66 in skeletogenesis. The Prx1 promoter was shown to be active in the mesenchyme of limb buds, the craniofacial area, sternum, and ventral rib cage of mouse embryos (26). We found that overexpression of NO66 in cells of Prx1-expressing mesenchymal lineage inhibits proliferation and differentiation of chondrogenic as well as osteogenic cells, resulting in retarded skeletal growth and bone formation in TG mice. Further in vivo and in vitro analyses revealed an inverse correlation between the level of NO66 and the activity of IGF1/IGF1 receptor (IGF1R)/protein kinase B (Akt) signaling pathway, which is important for cell growth and survival.

MATERIALS AND METHODS

Generation and genotyping of NO66-TG mice

Flag-tagged NO66 cDNAs (24) were subcloned into the Prx1-promoter vector, which was obtained from Dr. James Martin (Baylor College of Medicine, Houston, TX, USA), to generate a Prx1-NO66 plasmid (Fig. 1A). DNA was microinjected into fertilized B6D2 F1 eggs, which were then transferred into CD1 foster mothers. Progenies were crossed once with wild-type (WT) C57BL/6 mice to establish TG lines that were used for the study. The founders and offspring of TG mice were genotyped using PCR analysis with tail DNA and human growth hormone (Hgh) primers: 5′-TGTCTGACTAGGTGTCCTTC-3′, and 5′-GCAAGCAACTCAAATGTCC-3′. All experiments using mice were in compliance with standard care approved by the MD Anderson Cancer Center Institutional Animal Care and Use Committee.

Figure 1.

Generation of NO66-TG mice. A) Schematic of NO66 plasmid for microinjection. B) qPCR shows expression of NO66 mRNA in limbs of E14.5 WT, hemizygous embryos from TG line TG-1 (TG-1-Hemi), and hemizygous or homozygous embryos from TG line TG-2 (TG-2-Hemi and TG-2-Homo, respectively) (n = 3). C) qPCR shows expression of NO66 mRNA in calvaria of P1 WT and TG mice from lines TG-1 and TG-2 (n = 3). D) Paraffin section of hind limb of E14.5 embryo. E and F) Distal femur sections of E14.5 WT (E) and NO66-TG (F) embryos were stained using an anti-Flag antibody. G) Paraffin section of top head from P1 mouse. H and I) Head sections of P1 WT (H) and NO66-TG (I) mice were stained with anti-Flag antibody. (E), (F), (H), and (I) are higher-magnification staining images of the boxed regions in (D) and (G). The boxed areas in (E) and (F) indicate location of the perichondrium/periosteum. Blue indicates DAPI, and pink nuclear staining represents Flag-NO66. Yellow arrows and arrowheads in (E) and (F) show cells in femoral periosteum and adjacent connective tissue, respectively. White arrows and arrowheads in (H) and (I) point to cells in calvarial periosteum and brain, respectively.

RNA isolation and quantitative PCR

Total RNA of different murine tissues was isolated using TRIzol reagent (Invitrogen, Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s protocol. For quantitative PCR (qPCR), the total RNA was pretreated with TURBO DNase (Ambion Corporation, Naugatuck, CT, USA) to remove genomic DNA contamination and was then reverse transcribed into cDNAs using a high-capacity reverse transcription kit (Applied Biosystems, Foster City, CA, USA). A total of 50 ng cDNA and a gene-specific TaqMan primer probe (Applied Biosystems) were used in each PCR; each qPCR was performed in triplicate. Levels of mRNA expression were normalized by Hprt expression.

Histology

The limbs of mouse embryos at embryonic day 14.5 (E14.5) and pups at postnatal day 1 (P1), as well as heads of embryos at E16.5, were paraformaldehyde fixed, paraffin embedded, sectioned into 7 µm thicknesses, and then stained with Alcian blue following routine experimental procedures. To examine mineral deposition in bone, the undecalcified femoral and calvarial tissue sections were stained using von Kossa’s experimental approach and then counterstained with Nuclear Fast Red (Sigma-Aldrich, St. Louis, MO, USA) as described previously (9).

Immunofluorescence

For detection of hypertrophic chondrocyte and osteoblast differentiation markers, paraffin-embedded mouse femoral and calvarial sections were deparaffinized and subjected to enzymatic digestion with hyaluronidase [2 mg/ml in phosphate buffer (pH 5.5); MP Biomedicals, Santa Ana, CA, USA]. Primary antibody against COL10 (1:30; a gift from Klaus von der Mark, Nikolaus Fiebiger Centre of Molecular Medicine, University of Erlangen-Nurenberg, Erlangen, Germany), COL1 (1:200; EMD Millipore, Billerica, MA, USA), or BSP (1:200; Acris Antibodies, San Diego, CA, USA) was used. For detection of TG Flag-tagged NO66 protein and endogenous Osx, tissue slides were boiled in 10 mM sodium citrate buffer (pH 6.0) for 15 minutes using a microwave oven. Primary antibody against Flag (1:200; Chemicon, Temecula, CA, USA) or Osx (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA) was incubated overnight at room temperature. The secondary antibodies, including Alexa Fluor 488 goat anti-mouse and -rabbit or 555 goat anti-rabbit, -rat, or -mouse (1:1000; Molecular Probes/Invitrogen, Life Technologies), were incubated for 1 hour at room temperature. Slides were then washed and mounted with antifade Gold with DAPI (Molecular Probes, Life Technologies) and analyzed under a fluorescence microscope.

5-Bromo-2′-deoxyuridine incorporation in mice

DNA replication was detected using 5-bromo-2′-deoxyuridine (BrdU) incorporation. BrdU solution (Invitrogen, Life Technologies) was intraperitoneally injected into pregnant female mice or newborn pups (100 μl/10 g body weight). Two hours later, the mice were humanely killed and analyzed using immunofluorescence as described previously (27).

Western blotting

Western blotting was performed using lysates of mouse femurs as well as C2C12 cells overexpressing control vector or NO66 plasmid, following experimental procedures described previously (28, 29). Briefly, the proteins in lysates were isolated using the NE-PER Nuclear and Cytoplasmic Extraction Kit (Pierce, Rockford, IL, USA), separated on SDS-PAGE, transferred to a nitrocellulose membrane, and immunoblotted using anti-Flag (Sigma-Aldrich), anti-total or phosphorylated Akt (p-Akt; S473) (Cell Signaling Technology, Danvers, MA, USA), anti-total or phosphorylated Yes-associated protein (p-YAP; S127) (Cell Signaling Technology), anti-IGF1R (Abcam Incorporated, Cambridge, MA, USA), anti-H3K36me3 (Abcam Incorporated), anti-histone H3 (Abcam), anti-GAPDH (glyceraldehyde 3-phosphate dehydrogenase; Abcam Incorporated), or anti-NO66 (Santa Cruz Biotechnology) antibody. This was followed by reaction with appropriate horseradish peroxidase-labeled secondary antibody. The signals were then detected by SuperSignal chemiluminescence reagent (Pierce).

Transfection, BrdU labeling, and immunostaining of NO66 in C2C12 cells

NO66 expression plasmid was obtained from Dharmacon, GE Healthcare Life Sciences (Lafayette, CO, USA). C2C12 cells were cultured in α-minimum essential medium supplemented with 10% fetal bovine serum and electrotransfected with the NO66 plasmid or control vector using the Invitrogen (Life Technologies) Neon Transfection System. After 24 hours, the cells were incubated with BrdU following BrdU assay protocol from Roche (Basel, Switzerland), or harvested for immunostaining of NO66 as well as Western blots, as described previously (24).

Statistical analysis

All statistical results are presented as the mean ± sd. The differences between groups were calculated using the 2-tailed Student’s t test, and P values <0.05 were considered statistically significant.

RESULTS

Generation of TG mice overexpressing NO66 transgene driven by the Prx1 promoter

NO66 is conserved between species and widely expressed in different cells and tissues (22). To examine whether excess NO66 could affect mammalian skeletogenesis, we overexpressed NO66 in cells of mesenchymal lineage driven by the Prx1 promoter (Fig. 1A). We generated 2 stable NO66-TG mouse lines, named TG-1 and TG-2. Because integration of NO66 transgene into the genome is random, each TG animal is hemizygous. The hemizygous mice generated from line TG-1 exhibited growth retardation, whereas the hemizygous mice from line TG-2 did not show overt abnormalities. However, we observed a similar growth retardation phenotype in homozygous TG-2 mice, suggesting that the inhibitory phenotype of mice overexpressing NO66 is dependent upon the level of transgene. Our qPCR assays demonstrated that NO66 expression in limbs of hemizygous TG-1 embryos was much higher than that in limbs of hemizygous or homozygous TG-2 embryos on E14.5 (Fig. 1B). We obtained similar results when we performed qPCR using total RNA from calvaria of hemizygous TG-1 and hemizygous and homozygous TG-2 mice on P1 (Fig. 1C). We, therefore, chose the TG-1 (hereafter renamed NO66-TG) mice for further analysis. Immunostaining with anti-Flag antibody showed that the Flag-tagged NO66-TG protein was present in chondrocytes, particularly cells in the perichondrium/periosteum as well as adjacent connective tissue, in femurs of E14.5 NO66-TG embryos (Fig. 1F). We also detected the TG protein in periosteal and stromal cells in calvaria of P1 NO66-TG mice (Fig. 1I). By comparison, we did not observe positive staining of Flag-tagged NO66 protein in the TG mouse brain (Fig. 1I), where the Prx1 promoter is inactive. The observed expression pattern of the NO66 transgene is consistent with the Prx1-promoter activity reported previously (26).

Skeletal growth in NO66-TG mice during development

The Prx1 promoter begins to be active in mouse limb mesenchyme around E9.5 (26). At E10.5, we did not observe overt differences between WT and NO66-TG embryos (Fig. 2A). However, we observed growth retardation in NO66-TG embryos with development. By P1, skeleton preparation demonstrated that the skulls and limbs of NO66-TG mice were significantly smaller than those of their WT littermates (Fig. 2B–G; P < 0.01), suggesting inhibited growth of membranous and endochondral bones. Moreover, we found that the sutures in skulls of those NO66-TG mice were more widely open than those in skulls of WT controls (Fig. 2C), implicating a delay in membranous bone formation. These results indicated that overexpression of NO66 in Prx1-expressing mesenchymal cells inhibits bone growth.

Figure 2.

Gross morphology. A) E10.5 WT and NO66-TG embryos are shown. B–E) Skeletons of WT (B and D) and NO66-TG (C and E) mice at P1 were stained with Alcian blue and Alizarin red. The arrow in (C) indicates an open suture in the skull of NO66-TG mice. F and G) Quantification of widths of frontal calvaria (F; brackets in B and C) (n = 3) and humeral lengths (G; brackets in D and E) (n = 6).

Formation of endochondral and intramembranous bones in NO66-TG mice

We then performed histologic analyses of long bones and calvaria of WT and NO66-TG mice at different development stages. In Alcian blue and von Kossa staining of femur sections of E14.5 mouse embryos, we observed much smaller hypertrophic zones in NO66-TG embryos than in WT embryos (Fig. 3A, D) with no primary ossification or mineralization (Fig. 3B, E). Also, we detected less staining for COL1, a marker of bone formation, in the femurs of E15 NO66-TG embryos than in those of controls (Fig. 3C, F). By P1, the distal femurs of NO66-TG mice contained disorganized chondrocyte columns with very few hypertrophic chondrocytes (Fig. 3J), reduced amounts of mineralized tissue (Fig. 3K), and markedly decreased staining for COL10 (Fig. 3L), a marker of hypertrophic chondrocytes. These results indicated that overexpression of NO66 inhibits hypertrophic chondrocyte differentiation and endochondral bone formation. In addition, in the presumptive calvaria of E16.5 NO66-TG embryos, we also observed much less mineralized tissue as well as bone formation markers, COL1 and BSP, when compared with those in WT controls (Fig. 3M–R). These observations demonstrated that overexpression of NO66 inhibits intramembranous bone formation. Together, the results of these experiments indicated that mesenchymal overexpression of NO66 in mice inhibits hypertrophic chondrocyte differentiation and endochondral as well as intramembranous bone formation.

Figure 3.

Histologic examination of mice. A–L) Femur sections of embryos and pups were stained with Alcian blue (A, D, G, and J), von Kossa (B, E, H, and K), and antibody against COL1 (C and F) or COL10 (I and L). The red solid and dashed squares in (A), (D), (G), and (J) show hypertrophic zones in femurs, the arrows in (B), (E), (H), and (K) indicate mineralized tissue, the yellow arrows in (C) and (F) indicate COL1 (green) in periosteal regions of femurs, and the white arrows in (I) and (L) show COL10 (green) in hypertrophic zones of distal femurs. M–R) The sagittal sections of top heads from E16.5 embryos were stained with von Kossa (M and P) and antibody against COL1 (N and Q, green) or BSP (O and R, red). The staining signals are indicated by arrows.

Differentiation of osteoblast-lineage cells in NO66-TG mice

The osteoblast lineage consists of cells originating from osteoprogenitors or preosteoblasts, which then differentiate into mature osteoblasts and osteocytes to form endochondral or membranous bones. Differentiation of osteoblast lineage is accompanied by expression of several markers, such as Col1, Bsp, and Oc. Osx is a master transcription factor in osteogenesis and is essential for differentiation of osteoblast-lineage cells by activating expression of these marker genes (9, 10). Because we observed the decreased COL1 and BSP in bones of the NO66-TG mice (Fig. 3Q, R), we wanted to know whether the overall number of Osx-expressing osteoblast-lineage cells was affected. We performed immunostaining with anti-Osx antibody on tissue sections of femurs and calvaria from WT and NO66-TG mice. The results showed that the numbers of Osx-positive cells in the periosteal regions of femurs in E15.5 NO66-TG embryos were markedly lower than those in these regions in WT controls (Fig. 4A–D; Table 1; P < 0.01). We observed similar results in calvaria of NO66-TG mice at P3 (Fig. 4E–H; Table 1; P < 0.01). These data suggested that mesenchymal overexpression of NO66 in mice inhibits differentiation of osteoblast lineage in both endochondral and membranous bones.

Figure 4.

Immnuostaining of Osx. Femur sections of E15.5 embryos (A–D) and sagittal head sections of P3 mice (E–H) were stained with anti-Osx antibody. The yellow and white arrows indicate Osx-positive cells in the periosteal regions of femurs and calvaria, respectively. (C), (D), (G), and (H) are higher-magnification images of the boxed areas in (A), (B), (E), and (F), respectively (pink indicates Osx; blue represents DAPI). Scale bars, 100 μm.

TABLE 1.

Percentage of Osx-positive cells per area analyzed

	Mean percentage ± sd
Mice	E15.5 (femurs)	P3 (calvaria)
WT	46.5 ± 5.6	27.2 ± 3.2
NO66-TG	23.8 ± 4.5*	18.6 ± 2.2*

^*P < 0.01 (n = 3).

Proliferation of skeletal cells in NO66-TG mice

Among cells of osteoblast lineage, Osx-expressing osteoprogenitors or preosteoblasts are proliferating cells and located in the periosteal regions of long bones and calvaria. To test whether mesenchymal overexpression of NO66 could also affect the number of proliferating cells, we performed BrdU incorporation assay on femoral and calvarial sections of P1 mice. We found that the numbers of BrdU-positive cells in the periosteal areas of frontal calvaria and distal femurs of NO66-TG mice were markedly decreased when compared with those of WT controls (Fig. 5). This decrease is consistent with the observed decrease in Osx-expressing cells in these areas, suggesting a reduction in the number of osteoprogenitors or preosteoblasts. Moreover, we found that the numbers of BrdU-positive cells in proliferating chondrocytes and adjacent connective tissues in distal femurs of NO66-TG mice were also fewer than those of WT controls (Fig. 5H, I). By comparison, we did not observe a significant change in the number of BrdU-positive cells in NO66-TG brain, where the Prx1 promoter is not active (Fig. 5F, I; P > 0.05). These observations indicated that mesenchymal overexpression of NO66 in mice inhibits skeletal cell proliferation.

Figure 5.

BrdU incorporation. A, C, E, and G) Sagittal head (A and E) and distal femur sections (C and G) of P1 mice are shown. B, F, D, and H) Head sections in (A) and (E) and femur sections in (C) and (G) were stained with anti-BrdU antibody. The staining images shown in (B), (F), (D), and (H) are the corresponding boxed regions in (A), (E), (C), and (G) with a higher magnification (red indicates BrdU; blue represents DAPI). White arrows show BrdU-positive cells in frontal calvaria (white dashed boxes in B and F) and in periosteal regions of distal femurs (white dashed boxes in D and H), yellow triangles point to BrdU-positive cells in the brain of heads (B and F) and chondrocytes in growth plates (yellow solid boxes) of distal femurs (D and H), and the green triangles point to BrdU-positive cells in connective tissues adjacent to the periostea of femurs (D and H). I) Percentages of BrdU incorporation. Scale bars, 100 μm.

mRNA expression of signaling molecules, transcription factors, and skeletal markers in NO66-TG mice

The commitment, proliferation, and differentiation of osteochondroprogenitors into mature chondrocytes and osteoblasts are coordinately controlled by 3 key transcription factors: Sox9, Runx2, and Osx (10, 18). Expression of these factors is differentially regulated by a broad signaling network including IGF1, BMP2, and Wnt/β-catenin signaling pathways (11, 13–15). To determine changes in mRNA expression of these factors, we performed qPCR assays using total RNA from limbs of E13.5 mouse embryos. The results demonstrated that the NO66 mRNA was highly expressed in limbs of NO66-TG embryos (Fig. 6A; P < 0.01). Expression of Sox9 and its target gene Col2 was also higher in limbs of NO66-TG embryos than that in their control littermates (Fig. 6A; P < 0.05). In contrast, expression of Igf1, Igf1r, Runx2, and Osx was significantly down-regulated, concurrent with decreases in mRNA expression of the bone formation markers Alp and Col1 as well as hypertrophic chondrocyte marker Col10 in limbs of NO66-TG embryos (Fig. 6A; P < 0.05). By comparison, expression of Bmp2 and transcription factor 7, a downstream effector of Wnt signaling pathway, was slightly but not significantly down-regulated in those NO66-TG embryos (Fig. 6A; P > 0.05). These results implied that overexpression of NO66 may inhibit the IGF1/IGF1R signaling, which is important for Runx2 and Osx expression. To further confirm whether mRNA expression of Igf1, Igf1r, Runx2, and Osx was also decreased in membranous bones, we performed qPCR using total RNA from calvaria of 1-month-old WT and NO66-TG mice (Fig. 6B). We again observed high-level expression of NO66 but significantly decreased expression of Igf1, Igf1r, Runx2, and Osx, along with decreases in expression of bone markers such as Alp, Col1, Bsp, and Oc in calvaria of the NO66-TG mice (Fig. 6B; P < 0.05).

Figure 6.

qPCR. The results of qPCR assays using total RNA from limbs of E13.5 mouse embryos (n = 4) (A) or calvaria of mice at 1 month old (n = 3) (B) are shown.

Changes in IGF1/IGF1R/Akt signaling pathway and level of H3K36me3 in NO66-TG mice

A previous study showed that the IGF1/IGF1R-induced osteoblast proliferation is mediated through both Akt and MAPK pathways (15). To further investigate whether the IGF1/IGF1R/Akt signaling cascade was affected by excess NO66 in limbs of NO66-TG mice, we performed Western blotting to examine the p-Akt using proteins from femurs of WT and NO66-TG mice at different developmental stages. We detected a high level of Flag-tagged NO66-TG protein but a markedly decreased level of p-Akt in femurs of both E17 and P1 NO66-TG mice when compared to those in their WT littermates (Fig. 7A, C; P < 0.01). It has been known that the Hippo signaling controls organ size by phosphorylating the transcriptional coactivator YAP (30). As a control, we did not observe a significant decrease in the level of the p-YAP in femurs of E17 and P1 NO66-TG mice (Fig. 7A, C; P > 0.05), even though the level of total YAP in these mice was lower than those in WT controls (Fig. 7A, C; P < 0.05). These results suggested that mesenchymal overexpression of NO66 in mice may inhibit the IGF1/IGF1R/Akt signaling pathway, an important signal in the control of chondrocyte and osteoblast proliferation (14, 15).

Figure 7.

Western blotting. A and B) The cell lysates (A) or nuclear extracts (B) were isolated from femurs of E17 and P1 WT and NO66-TG mice (n = 3) and then blotted with antibodies as denoted in the figure. IB, immunoblot. C and D) Quantification of images in (A) and (B).

Previous in vitro studies reported that NO66 has a histone demethylase activity that is specific for H3K4me3 and H3K36me3. Overexpression of NO66 in COS7 cells markedly reduced the staining intensity of H3K4me3 and H3K36me3, 2 marks of transcriptionally active chromatin (24). To test whether mesenchymal overexpression of NO66 in mice could alter these histone marks, we performed Western blotting to examine the level of H3K36me3 in femurs of WT and NO66-TG mice at different stages. We found that the level of H3K36me3 in femurs of NO66-TG mice on E17 and P1 was markedly lower than that in WT control littermates (Fig. 7B, D). These results suggested that mesenchymal overexpression of NO66 in mice alters the cellular histone H3 methylation state, which may contribute in part to the decreased mRNA expression of Igf1 and/or Igf1r.

Overexpression of NO66 inhibits cell growth and the IGF1R/Akt signaling in vitro

A previous study reported that NO66 was associated with c-Myc, and overexpression of NO66 in NIH3T3 cells facilitated cell growth (23). This in vitro observation seems opposite to the growth inhibitory phenotype that we observed in the NO66-TG mice. To further confirm this discrepancy, we performed in vitro BrdU assay by transiently overexpressing NO66 in C2C12 cells (Fig. 8). We found that most cells with excess NO66 did not incorporate BrdU (Fig. 8A–D), resulting in a markedly decreased percentage of BrdU-positive cells with overexpression of NO66 (Fig. 8E; P < 0.05). These results indicated that transient overexpression of NO66 in C2C12 cells inhibits cell proliferation, consistent with the decreased BrdU-positive cells observed in bones of NO66-TG mice. To test whether transient overexpression of NO66 in C2C12 cells could affect cellular IGF1R/Akt signaling, we performed Western blotting to examine the levels of IGF1R and p-Akt in those NO66- or control vector-transfected C2C12 cells. The results showed a high level of NO66 but low levels of IGF1R and p-Akt (Fig. 8F, G). By comparison, there was no significant change in the level of p-YAP (Fig. 8F, G; P > 0.05). These in vitro data demonstrated the inverse correlation between the level of NO66 and IGF1R/Akt signaling.

Figure 8.

Immunostaining and Western blotting. A–C) Immunostaining. The NO66-transfected C2C12 cells were labeled with BrdU and then stained with anti-BrdU (A, red) and anti-NO66 (B, green) antibodies. Cell nuclei were stained with DAPI (C, blue). D) Merged images of (A)–(C). Arrows indicate cells with excess NO66 (Scale bars, 100 μm). E) Percentage of BrdU-positive cells in control (CTRL) and NO66-transfected (NO66) groups (n = 3). F) Western blotting. The lysates of C2C12 cells overexpressing NO66 or control vector were blotted with antibodies as denoted in the figure. G) Quantification of images in (F). *P < 0.05; ** P < 0.01 (n = 3).

DISCUSSION

In this study, we generated a TG mouse model by overexpressing a flag-tagged NO66 transgene driven by the Prx1 promoter, to study gain of function of NO66 in skeletogenesis. Our results showed that high-level expression of NO66 in Prx1-expressing mesenchymal cells causes retarded endochondral bone growth. This includes decreased hypertrophic chondrocyte differentiation, delayed primary ossification, and decreased proliferation of skeletal cells. The markedly reduced bone formation markers observed in the presumptive calvaria of E16.5 NO66-TG embryos demonstrated that membranous bone growth was also delayed or inhibited by overexpression of NO66. Similarly, the decreased number of BrdU-positive skeletal cells and Osx-expressing osteoblast-lineage cells in bones of these mice could be a consequence of the mesenchymal overexpression of NO66. In the bones of NO66-TG mice or C2C12 cells overexpressing NO66, there was a tight correlation between decrease in cell proliferation and low levels of IGF1R and p-Akt. We therefore propose that the decrease in the IGF1R/Akt signaling cascade may contribute, at least in part, to the skeletal growth retardation of NO66-TG mice. Moreover, in the NO66-TG mice, there was also a correlation between decrease in Igf1/Igf1r mRNA expression and low level of H3K36me3, a mark of transcriptionally active chromatin. We speculate that alteration of histone H3 methylation state induced by excess NO66 may contribute in part to the decreased mRNA expression of Igf1 and/or Igf1r in bones of NO66-TG mice.

The longitudinal growth of mammalian bones requires coordinate proliferation and differentiation of chondrocytes in cartilage template. In our NO66-TG mice, we observed decreased BrdU incorporation in the proliferating zone of chondrocytes and reduced staining for COL10, a marker of hypertrophic chondrocytes. We therefore postulate that in long bones of NO66-TG mice, the decreases in chondrocyte proliferation and differentiation or maturation are likely responsible for the retarded longitudinal growth of endochondral bones in these mice. In addition, bone formation requires proper differentiation of osteoblast lineage, which consists of osteoprogenitors or preosteoblasts and mature osteoblasts and osteocytes. Previous studies showed that NO66 is a negative regulator of preosteoblast differentiation because knockdown of NO66 in MC3T3 preosteoblasts increased expression of osteoblast differentiation markers, Col1, Bsp, and Oc (24). In our current study, we observed decreased expression of these markers along with a decrease in the number of Osx-expressing cells in bones of NO66-TG mice, suggesting an inhibitory role of NO66 in osteoblast differentiation, which seems consistent with the previous in vitro observation. However, given that the percentage of BrdU incorporation in perichondrium/periosteum as well as proliferating chondrocytes was markedly decreased in bones of NO66-TG mice, we speculate that excess NO66 inhibits not only osteoblast differentiation but also commitment or differentiation of osteochondroprogenitors.

Osx is a master transcription factor that is essential for osteoblast-lineage differentiation and bone formation (9, 10). A previous in vitro study indicated that NO66 negatively regulates osteoblast differentiation by binding to Osx and inactivating its activity (24). We therefore hypothesize that the decreased Osx mRNA expression and reduced Osx protein staining observed in NO66-TG mice are likely responsible for the inhibited bone formation in these mice. Bone formation and Osx expression are controlled by a broad signaling network (10, 18, 19). In this study, we examined the mRNA expression for only a limited number of signaling molecules in bones of NO66-TG mice and observed consistent decreases in expression of Igf1, Igf1r, Runx2, and Osx in these mice. The decreased level of p-Akt along with the low level of Igf1r mRNA or IGF1R protein, which was observed in bones of NO66-TG mice or in C2C12 cells overexpressing NO66, demonstrated that excess NO66 inhibits the IGF1R/Akt signaling cascade. The IGF1R/Akt signaling is an important upstream signal that induces expression of Runx2 and Osx (11, 14, 15). Runx2 has been known to act upstream of Osx and control differentiation of hypertrophic chondrocytes and expression of COL10 (7–10). We speculate that the decreased IGF1R/Akt signaling in bones of NO66-TG mice could contribute in part to the decreased mRNA expression of Runx2 and Osx, and this decrease may be then responsible for the decreased number of hypertrophic chondrocytes and Osx-expressing cells observed in these mice. In fact, the phenotype of our NO66-TG mice resembles that of cartilage-specific Igf1r knockout mice, which exhibit growth retardation with disorganized chondrocyte columns, delayed primary ossification, and decreased chondrocyte proliferation (14). We propose that the reduction of IGF1R/Akt signaling cascade in bones of NO66-TG mice contributes in part to the bone growth retardation phenotype of these mice. Given that NO66 can directly bind to Osx and inactivate its activity, it is possible that excess NO66 represses the IGF1R/Akt signaling in parallel with NO66-induced inactivation of Osx. Nevertheless, it is likely that excess NO66 inhibits other signaling molecule(s), which may also contribute to the growth retardation of NO66-TG mice.

In addition, NO66 was reported to be associated with c-Myc and facilitate its cell proliferation activity (23). In this previous study, researchers stably overexpressed NO66 in NIH3T3 or non-small cell lung cancer cell line and observed an increase in cell proliferation, implicating an oncogenic feature of NO66. In contrast to this previous report, we observed an inhibitory role of excess NO66 in cell proliferation when we transiently overexpressed NO66 in C2C12 cells. Given that each of the experimental systems is different, it is challenging to provide a coherent explanation that would reconcile the different results.

In conclusion, our studies provide the first in vivo evidence that NO66 histone demethylase functions as an inhibitor of mammalian bone growth. NO66 also appears to be a novel repressor of the IGF1R/Akt signaling pathway, which is critical for cell growth and survival, and therefore may be a potential target in the treatment of growth retardation or dwarfism.

Glossary

Akt

protein kinase B

ALP

alkaline phosphatase

BMP

bone morphogenetic protein

BrdU

5-bromo-2′-deoxyuridine

BSP

bone sialoprotein

COL1

type I collagen

Col2

type II collagen

COL10

type X collagen

E

embryonic day

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

H3K36me

methylated H3K36

H3K4me

methylated H3K4

Hgh

human growth hormone

IGF1R

IGF1 receptor

NO66

nucleolar protein 66

Oc

osteocalcin

Osx

osterix

P

postnatal day

p-Akt

phosphorylated protein kinase B

Prx1

paired related homeobox 1

p-YAP

phosphorylated Yes-associated protein

qPCR

quantitative PCR

Runx2

runt-related transcription factor 2

Sox9

SRY-related HMG-box 9

TG

transgenic

WT

wild-type

YAP

Yes-associated protein