This article is mentioned in:

Introduction

Chondrocyte differentiation is a pivotal event during bone formation in most mammals, and it comprises of two processes (1). Intramembranous ossification is an essential process during which mesenchymal stem cells directly differentiate into osteoblasts and is responsible for the development of irregular bones, such as the skull bone, clavicle and parts of the jaw (1). Endochondral ossification is the second important process of chondrocyte differentiation, responsible for the formation of long bones and vertebrate skeleton (2), which is controlled by a series of events (3). These events are initiated by the establishment of a chondrogenic template and culminated by its replacement with the coordinated activity of osteoblasts, osteoclasts and endothelial cells (3,4). Chondrocyte maturation ultimately leads to hypertrophy, which is an essential contributor to longitudinal bone growth (5,6). However, abnormal chondrocyte hypertrophy is involved in the pathogenesis of osteoarthritis and skeletal dysplasia (7). Meanwhile, histone deacetylase 4 (HDAC4) is known as a member of a large family of HDAC that alter gene expression by catalyzing the removal of acetyl groups from core histones, which are responsible for chondrocyte hypertrophy and bone formation (8,9). However, the regulatory mechanism underpinning HDAC4 function remains largely unclear. Therefore, a better understanding of the mechanism underlying HDAC4-mediated regulation of endochondral ossification in chondrocyte hypertrophy is necessary for the improved management of skeletal growth diseases.

A previous study demonstrated that HADC4 plays an important role in the formation of the skeletal system (10). Knockout of HDAC4, a repressor of matrix metalloproteinase-13 (MMP-13) transcription, results in increased MMP-13 expression in vitro and in vivo (11). MMP-13 plays a critical role in endochondral ossification and bone remodeling (12). Chen et al (13) reported that HDAC4 inhibits the expression of MMP-13 and prevents chondrocyte hypertrophy by inhibiting the activity of runt-related transcription factor-2 (Runx2), a transcription factor essential during endochondral bone formation. Wnt5a is a secreted glycoprotein belonging to the Wingless/integrase 1 family of secreted glycoproteins; it is highly conserved among species and plays key roles in processes regulating embryonic development and postnatal tissue homeostasis (14). Wnt5a-knockout mice reportedly demonstrate perinatal lethality primarily due to respiratory failure and present extensive developmental abnormalities and the Wnt5a signaling pathway can regulate fundamental cellular processes, including cell proliferation and survival (15). Therefore, HDAC4 is considered as a central regulator of chondrocyte hypertrophy and growth plate development, and the complete deletion of HDAC4 is lethal (9,10,16–18). To investigate the roles of HDAC4 in postnatal skeletal development, the present study constructed a mouse strain with conditional HDAC4-knockout in collagen type 2α1 (Col2α1)-expressing cells.

Materials and methods

Generation of floxed HDAC4 animals

The animal experiments and protocol were approved by the Institutional Animal Care and Use Committee of Rhode Island Hospital (Providence, United States of America). To investigate the physiological role of HDAC4 in hypertrophic chondrocytes, the HDAC4 gene was deleted in Col2α1-Cre mice using the Cre/loxP gene targeting technique (six female mice, mean weight 20 g, from University of Texas Southwestern Medical Center, mice transgenic for Cre in Col2α1-expressing chondrocytes). Col2α1-Cre mice were mated with HDAC4fl/fl (six male mice, mean weight 20 g, provided by Dr Olson, University of Texas Southwestern Medical Center) animals to obtain HDAC4fl/−, Col2α1-Cre mice (9). Mice transgenic for Cre in Col2α1-expressing chondrocytes (Col2α1-Cre) have been previously reported (19). These mice were subsequently interbred with HDAC4fl/fl animals, and their offspring (HDAC4d/d, Col2α1-Cre) and HDAC4fl/fl animals were analyzed at postnatal (P) days 2, 4, 6, 8, 10 and 21 after birth. All mice were housed in a cage with free access to tap water and food. The housing environment was set at 22±2°C and 60±10% humidity in a 12-h light/dark cycle. The health and behavior of mice were monitored and recorded on a daily basis. A total of 40 mice were divided into two groups (HDAC4fl/fl group and HDAC4d/d; Col2α1-Cre group; n=20 for each group). The mice were euthanized by an overdose of carbon dioxide according to the American Veterinary Medical Association (AVMA) guidelines (20). The experiment endpoint was set when mice were postnatal 21 days old. The mice are first put into the euthanasia chamber and CO2 was directed into the top of the euthanasia chamber. Weep holes were drilled into the euthanasia chamber lid to allow room air to escape as CO2 filled the chamber. One-hundred percent compressed CO2 gas was used to euthanize mice. When the flow rate displaced no ≤30% of the chamber volume/minute for apparent death, death was confirmed by absence of breathing or heartbeat. In this way, the mice had no pain, no nerve reaction and no signs of bleeding. The growth progress of mice was recorded at postnatal (P) days 2, 4, 6 and 8, 10 and 21.

Genotyping and PCR

Genomic DNA was isolated from skin tissue from the tail of the mice using the QIAamp DNA Micro kit (Qiagen GmbH) and used for PCR genotype identification of HDAC4d/d, Col2α1-Cre transgenic mice. Routine mouse genotyping was performed using PCR, using the DNA polymerase Hotstar from Qiagen GmbH. The following primer pairs were used for the Cre allele: Forward, 5′-ATCCGAAAAGAAAACGTTGA-3′ and reverse, 5′-ATCCAGGTTACGGATATAGT-3′. PCR was initiated for 15 min at 95°C, then 30 cycles of denaturation at 95°C for 30 sec, primer annealing for 40 sec at 55°C and a final extension step at 72°C for 1 min. The HDAC4 allele and target gene deletion were identified using the following primers: Forward, 5′-ATCTGCCCACCAGAGTATGTG-3′ and reverse, 5′-CTTGTTGAGAACAAACTCCTGCAGCT-3′. PCR was initiated for 15 min at 95°C, then 35 cycles of denaturation at 94°C for 40 sec, primer annealing for 40 sec at 58.5°C and a final extension step at 72°C for 1 min. The expected product sizes for the wild-type HDAC4, floxed and Cre alleles were 480, 620 and 328 bp, respectively.

Cell proliferation assay

Cell proliferation of chondrocytes was assessed using the in vivo bromodeoxyuridine (BrdU) assay as previously described (21). Briefly, mice were intraperitoneally injected with 25 µl/25 g body weight BrdU at P3 and P5. The animals were then euthanized at P8, P14, and P21 for specimen collection followed by immunohistochemistry (22).

Histology

After the animals were euthanized at P8, P14, and P21, their right knee joints (n=3) were harvested and immersed in 10% formalin for 72 h at room temperature. The specimens were processed without decalcification and embedded in a single block of Paraplast X-tra (Thermo Fisher Scientific, Inc.). Thereafter, 6-µm coronal sections were mounted on slides. Safranin O/Fast green staining for 2 h at room temperature was performed to evaluate the developmental growth plate and hypertrophic differentiation at P8 and P21, whereas Von Kossa staining for 2 h at room temperature was used to evaluate tissue mineralization at P8 and P14. Photomicrographs were obtained using a Nikon Ri1 microscope (Nikon Corporation) (23,24).

Whole-body staining and skeletal analysis

The mice were euthanized before the experiment began. The mineralization pattern of the skeleton was analyzed at P10 using whole-body staining as previously described (25). Briefly, mice were skinned under anesthesia, eviscerated and fixed in 95% ethanol for 1–2 days. Subsequently, acetone was used to remove fat. Skeletons were stained using Alizarin red S/Alcian blue for 3 days at 37°C and sequentially immersed in 1% aqueous KOH at room temperature until the skeletons were clearly visible. Mineralized bones were visualized through the staining (23).

Microcomputed tomography (µCT)

Three tibial tissues isolated from each group (8-day old mice) were fixed in 70% ethanol for 24 h at room temperature and prepared for high resolution µCT (SkyScan 1172). Images were obtained using the following parameters: 40 kV, 80 µA, Pixel size=8.5 µm, 2,000×1,332 matrix, six averages and a 0.5-mm aluminum filter. Images were reconstructed using a threshold of 0–0.09, beam hardening correction of 35, ring artifact correction of 7 and Gaussian smoothing (factor=1). From the reconstructed images of the tibia, the 637-µm region immediately proximal to the growth plate was examined for the trabecular bone. The volume images of each group were analyzed using CTAn software (v1.7.1, SkyScan, Burker) to determinate the trabecular bone microstructure and trabecular thickness (Tb.Th), trabecular separation (Tb.Sp).

Immunohistochemistry

BrdU and proliferating cell nuclear antigen (PCNA) are typical proliferation markers that can indicate the proliferative activity of chondrocytes. BrdU is a synthetic nucleoside used to detect cellular proliferation by incorporating the place of thymine and pairs with guanine (26). PCNA is a multifunctional protein present in the nuclei of eukaryotic cells and has a crucial role in DNA replication and DNA repair systems (27). Matrix metalloproteinase (MMP)-13 is marker associated with endochondral ossification and bone repair (28). Runx2 and osteoprotegerin (OPG) are typical mineralization markers. Runx2 is a transcription factor essential for osteoblast differentiation and chondrocyte maturation (29). OPG is a key protective factor of bony tissue and is responsible for osteoclast formation and postnatal bone growth (30). To determine the expressions of these markers, 6-µm coronal sections of right knee joints were placed on positively charged glass slides (Thermo Fisher Scientific, Inc.). The sections were dried on a hotplate to increase tissue adherence. Immunohistochemistry was performed using the 3,3′diaminobenzidine (DAB), streptavidin-peroxidase (SP) and the DAB Histostain-SP immunohistochemistry kit (Zymed; Thermo Fisher Scientific, Inc.). Sections were deparaffinized and rehydrated using xylene and an alcohol gradient. Endogenous peroxidase was blocked by treating the sections with 3% hydrogen peroxide in methanol (Sigma-Aldrich; Merck KGaA) for 30 min at room temperature. The sections were then digested with 5 mg/ml hyaluronidase in phosphate-buffered saline (PBS) (Sigma-Aldrich; Merck KGaA) for 20 min and incubated with specific antibodies against MMP-13 (1:500, cat. no. ab219620, Abcam), BrdU (1:100, cat. no. ab8955, Abcam), PCNA (1:10,000, cat. no. ab29, Abcam), Runx2 (1:1,000, cat. no. 192256, Abcam), OPG (1:100, cat. no. 203061, Abcam) and CD34 (1:2,500, cat. no. 81289, Abcam) at 4°C overnight. After treatment with biotinylated secondary antibody and SP conjugate (500 µl, Histostain®-Plus 3rd Gen IHC Detection kit, cat. no. D859673, Zymed; Thermo Fisher Scientific, Inc.) at 37°C for 30 min, the sections were developed in DAB chromogen (Zymed; Thermo Fisher Scientific, Inc.) and then counterstained with hematoxylin (Zymed; Thermo Fisher Scientific, Inc.) for 2 min at room temperature. Images were captured using a Nikon Ri1 light microscope (Nikon Corporation) at ×4 magnification. Percentages of positive staining for BrdU, PCNA, MMP-13, Runx2, OPG and CD34 in both the groups were semi quantified using the Image-Pro Plus 7.0 software (Media Cybernetics, Inc.) as described previously (31). Mean values from three independent measurements were used for statistical analysis (n=3).

In situ hybridization (ISH)

ISH was used to determine type X collagen (ColX) and Wnt5a expression levels as an indicator of HDAC4-downregualtion. ColX is a critical marker that represents terminal differentiation during chondrogenesis (32). Wnt5a a member of Wnt family that plays a vital role in regulating skeletal development (33). ISH was performed on paraffin sections prepared from tibial tissues at P8 as described previously (34). Accordingly, 10-µm sections were placed on positively charged glass slides (Thermo Fisher Scientific, Inc.) and subsequently dried on a hotplate to increase tissue adherence. These sections were then deparaffinized and rehydrated using conventional methods. Briefly, the tibial sections were incubated with proteinase K (cat. no. P8107S; New England BioLabs, Inc.) at 60°C for 15 min and then rinsed with diethyl pyrocarbonate (DEPC)-PBS for 5 min. Thereafter, the tibial sections were incubated with 5 mg/ml hyaluronidase at 37°C for 15 min and then rinsed with DEPC-PBS for 5 min. The sections were prehybridized for 1 h and hybridized for 16 h with a 0.5-µl probe (Col X and Wnt5a, from Department of Orthopedics, The Warren Alpert Medical School of Brown University) at 50°C. The sections were washed with 5X saline sodium citrate (SSC) for 5 min at room temperature and 4X SSC for 30 min at 50°C. Thereafter, the tibial sections were incubated with Tris-HCL-NaCl-EDTA buffer at 37°C for 10 min; washed with 2XSSC for 10 min, 0.2XSSC for 10 min and 0.1XSSC buffer for 10 min. Following this, the sections were incubated with a blocking buffer for 1 h at room temperature. Digoxin antibody (1:200, ca. no. ab420, Abcam) was added to the sections for 2 h at room temperature. The reaction was stopped with tris-EDTA (TE) buffer (pH 8.0) for 10 min. The sections were then washed with distilled water and counterstained with hematoxylin (Zymed; Thermo Fisher Scientific, Inc.) for 2 min at room temperature. Images were captured using a Nikon Ri1 light microscope (Nikon Corporation) at ×4 magnification.

Reverse transcription-quantitative (RT-q)PCR

The mRNA levels of HDAC4, PCNA, MMP-13, Runx2, OPG and osteocalcin were quantified using RT-qPCR. Total RNA was isolated using TRIzol (Life Technologies; Thermo Fisher Scientific, Inc.) from the sternum cartilage using the RNeasy isolation kit (Qiagen, Inc.). Accordingly, 1 µg total RNA was reverse transcribed into cDNA using the iScripTM cDNA synthesis kit (Bio-Rad Laboratories, Inc.), with reaction conditions: 25°C 5 min, 42°C for 30 min, 85°C for 5 min and kept at 4°C. Thereafter, 50 ng/µl of the resulting cDNA was used as the template to quantify the relative mRNA content using the QuantiTect SYBR Green PCR kit (Qiagen, Inc.) with the DNA Engine Opticon 2 Continuous Fluorescence Detection system (MJ Research; Bio-Rad Laboratories, Inc.). qPCR was initiated for 30 sec at 94°C, then 40 cycles of denaturation at 9°C for 30 sec, primer annealing for 3 sec at 60°C and a final extension step at 72°C for 30 sec. Primer pairs that were used for quantitative detection of gene expression are listed in Table I, and 18S rRNA was used as the internal control. The cycle threshold values for target genes were measured and calculated using a computer software (MJ Research; Bio-Rad Laboratories, Inc.). Relative transcript levels were calculated using the 2−ΔΔCq method where ΔΔCq=ΔCq E-ΔCq C, ΔCq E=Cqexp-Cq18S, and ΔCq C=CqCCq 18S (35).

Table I. Primer sequences for reverse transcription-polymerase chain reaction.

Table I.

Primer sequences for reverse transcription-polymerase chain reaction.

Mouse gene	Primer sequence, 5′-3′
HDAC4
Forward	ATCTGCCCACCAGAGTATGTG
Reverse	CTTGTTGAGAACAAACTCCTGCAGCT
MMP-13
Forward	GGACCTTCTGGTCTTCTGGC
Reverse	GGATGCTTAGGGTTGGGGTC
Runx2
Forward	CCGCACGCAAACCGCACCAT
Reverse	CGCTCCGGCCCACAAATCTC
PCNA
Forward	GCCGAGATCTCAGCCATATT
Reverse	ATGTACTTAGAGGTACAAAT
OPG
Forward	ATTGGCTGAGTGTTTTGGTGG
Reverse	CGCTGCTTTCACAGAGGTCA
Osteocalcin
Forward	ATGAGAGCCCTCAGACTCCTC
Reverse	CGGGCCGTAGAAGCGCCGATA
18s rRNA
Forward	CGGCTACCACATCCAAGGAA
Reverse	CGCTCCGGCCCACAAATCTC

Statistical analysis

Data from at least three separate experiments are presented as mean ± standard deviation. Differences between groups were analyzed using unpaired t-tests and P<0.05 was considered to indicate a statistically significant difference. Statistical analyses were performed using SPSS software v18.0 (SPSS Inc.).

Results

Generation of HDAC4, Col2α1-Cre mice

To investigate the function of HDAC4 in vivo, the mouse HDAC4 gene was deleted using homologous combination. The targeting strategy is presented in Fig. 1A. Wild-type, homologous and heterozygous mice were confirmed using qPCR. Product sizes for wild-type HDAC4, floxed and Cre alleles were 480, 620 and 328 bp, respectively (Fig. 1B). The HDAC4d/d, Col2α1-Cre mice survived until P21, were severely runted and smaller compared with HDAC4fl/fl mice at P2, P4, P6 and P8 (Fig. 1C). These results indicated that the development of HDAC4d/d, Col2α1-Cre mice was slower compared with that of HDAC4fl/fl mice.

Figure 1. Generation of HDAC4, Col2α1-Cre mice. (A) Mating of HDAC4fl/−, Col2α1-Cre and HDAC4fl/fl mice. (B) Representative PCR results of genomic DNA from wild-type and HDAC4 heterozygous and homozygous null animals. (C) Images of the HDAC4fl/fl and HDAC4 d/d, Col2α1-Cre offspring at postnatal days 2, 4, 6 and 8. HDAC4, histone deacetylase 4; Col2α1, collagen type 2α1; fl, floxed; d, down.

Shortened growth plates and disordered hypertrophic chondrocyte zones in HDAC4d/d, Col2α1-Cre mice

Chondrocyte hypertrophy is widely observed in endochondral ossification (5). During bone formation, chondrocytes were enlarged, and they secreted more ColX compared with the early stage of the ossification (5). Here, Safranin O/Fast green staining was performed on paraffin sections from HDAC4fl/fl and HDAC4d/d, Col2α1-Cre mice at P8 (femur and tibia) (Fig. 2A-C) and P21 (tibia; Fig. 2D). As shown in Fig. 2A2-C2 and Fig. 2A4-C4, HDAC4d/d, Col2α1-Cre mice had fewer prehypertrophic and hypertrophic chondrocyte zones compared with the HDAC4fl/fl mice (Fig. 2A1-C1 and A3-C3) at P8. In addition, the hypertrophic chondrocyte zone appeared smaller, and chondrocytes underwent degradation in the HDAC4d/d, Col2α1-Cre mice group (Fig. 2A2-C2 and A4-C4) compared with the HDAC4fl/fl mice (Fig. 2A1-C1 and A3-C3) at P8. As presented in Fig. 2D-2, HDAC4d/d, Col2α1-Cre mice showed decreased proteoglycans zones (marked by yellow arrows), premature ossification (marked by black arrows) compared with the HDAC4fl/fl mice (Fig. 2D-1) at P21. These results indicated that the shortened growth plates and disordered hypertrophic chondrocyte zones in HDAC4d/d, Col2α1-Cre mice compared with that of HDAC4fl/fl mice.

Figure 2. Safranin O/Fast green staining on paraffin sections of the femur and tibia tissues at P8 and P21. (A-C) Safranin O/Fast green staining showing femur and tibia tissue sections (magnification, ×4, ×10 and ×20 respectively) at P8. The HDAC4d/d, Col2α1-Cre mice had fewer cartilage tissues in the hypertrophy zone at P8 compared with those in the HDAC4fl/fl mice. Black arrows represent hypertrophic chondrocytes, and orange arrows represent chondrocytes degradation. Scar bars, 100 or 50 µm. (D) Safranin O/Fast green staining showing tibia tissue sections (magnification, ×4) at P21. The HDAC4d/d, Col2α1-Cre mice had fewer proteoglycans but more ossification at P21 compared with the HDAC4fl/fl mice. The yellow arrow represents zone of proteoglycans, the black arrow represents zone of ossification, Scale bars, 100 µm. P, postnatal day; HDAC4, histone deacetylase 4; Col2α1, collagen type 2α1; fl, floxed; d, down.

Accelerated vascular invasion and mineralization due to HDAC4-knockout

During hypertrophic chondrocyte mineralization, blood vessels are induced by hypertrophic chondrocytes that secreted vascular endothelial growth factor (36). Von Kossa staining was performed to assess mineralization and vascular invasion of the tissue in HDAC4fl/fl mice and HDAC4d/d, Col2α1-Cre mice at P8 (Fig. 3A and B) and P14 (Fig. 3C and D). As encircled in Fig. 3B1, vascular invasion of the growth plates in HDAC4d/d, Col2α1-Cre mice began to appear at P8, whereas HDAC4fl/fl mice had no vascular formation at P8 (Fig. 3A1). More importantly, HDAC4d/d, Col2α1-Cre mice developed an earlier secondary center of ossification (red circled area) and growth plates (yellow circled area; Fig. 3D and D1) compared with HDAC4fl/fl mice (Fig. 3C and C1) at P14. Moreover, the HDAC4d/d, Col2α1-Cre mice had thicker cortical (Fig. 3D2) and cancellous bones (Fig. 3D3) compared with HDAC4fl/fl mice (Fig. 3C2 and C3). These results indicated that the vascular invasion and mineralization were accelerated in HDAC4d/d, Col2α1-Cre mice compared with HDAC4fl/fl mice.

Figure 3. Mineralization and vascular invasion of tissues at (A and B) P8 and (C and D) P14 using Von Kossa staining. (B1, the red circled regions) vascular invasion and (D1, the red circled regions) secondary center of ossification were observed earlier in the HDAC4d/d, Col2α1-Cre mice compared with in HDAC4fl/fl mice (A1 and C1, the red circled regions). The HDAC4d/d, Col2α1-Cre mice had (B2 and D2) thicker cortical and (B3 and D3) cancellous bones compared with the (A2, C2, A3 and C3) HDAC4fl/fl mice. Yellow circled areas represent growth plates. Black circled areas represent cortical bones. Green circled areas represent cancellous bones. Boxes indicate area magnified in adjacent figures. Scale bars, 100 µm. P, postnatal day; HDAC4, histone deacetylase 4; Col2α1, collagen type 2α1; fl, floxed; d, down.

Abnormal and premature mineralization due to HDAC4 deletion

The detailed skeletal phenotypes of the HDAC4fl/fl and HDAC4d/d, Col2α1-Cre groups were analyzed using Alizarin red and Alcian blue staining at P10 (Fig. 4A and B). Alizarin red and Alcian blue staining of the whole skeleton of HDAC4d/d, Col2α1-Cre mice revealed obvious malformed and premature mineralization of the endochondral skeleton, including the sternum (Fig. 4D), xiphisternum (Fig. 4D), tibia (Fig. 4F) and intervertebral disk (Fig. 4H) compared with HDAC4fl/fl mice(Fig. 4C, E and G). Ectopic ossification of the endochondral cartilage prevented normal expansion of the rib cage (Fig. 4D) µCT analysis showing the HDAC4d/d, Col2α1-Cre mice (Fig. 4J) have more cancellous bone than HDAC4fl/fl mice (Fig. 4I) at P8. The Tb.Th was significantly greater in the HDAC4d/d, Col2α1-Cre mice compared with HDAC4fl/fl mice, as measured by µCT (P<0.05; Fig. 4K). Simultaneously, Tb.Sp was significantly reduced in the HDAC4d/d, Col2α1-Cre mice compared with that in HDAC4fl/fl mice (P<0.05; Fig. 4L). These results indicated that abnormal and premature mineralization occurs in the knockout mouse.

Figure 4. Premature ossification in HDAC4d/d, Col2α1-Cre mice. (A-H) Alizarin red and Alcian blue stained skeletons of the HDAC4fl/fl and HDAC4d/d, Col2α1-Cre mice at P10, the red boxes and arrow represent xiphoid process and sternum, the green boxes and arrow represent lumbar vertebrae, black arrows represent tibia cartilage. (I and J) µCT analysis showing the bone image of tibial tissues from the HDAC4fl/fl and the HDAC4d/d, Col2α1-Cre mice at P8, yellow boxes and arrow represent cancellous bone. (K and L) Tb.Th, trabecular thickness; Tb.Sp, trabecular separation. in the HDAC4fl/fl and HDAC4d/d, Col2α1-Cre mice were evaluated using µCT analysis. Each assay was performed in triplicates for each group. Scale bars, 200 µm. *P<0.05 vs. HDAC4fl/fl mice. P, postnatal day; HDAC4, histone deacetylase 4; Col2α1, collagen type 2α1; fl, floxed; d, down; Tb.Th, trabecular thickness; Tb.Sp, trabecular separation.

Decreased chondrocyte proliferation due to HDAC4 deletion via Wnt5a signaling pathway inhibition and increased chondrocyte hypertrophy

Immunohistochemistry was performed to determine the presence of BrdU, PCNA, MMP-13, Runx2, OPG and CD34 in HDAC4fl/fl and HDAC4d/d, Col2α1-Cre mice (Fig. 5A-L). The percentages areas positively expressing MMP-13, Runx2, OPG and CD34, were significantly higher in the HDAC4d/d, Col2α1-Cre mice compared with the HDAC4fl/fl mice (all P<0.05; Fig. 5M-R). Conversely, the expressions of typical proliferative markers, including BrdU and PCNA, were significantly lower in the HDAC4d/d, Col2α1-Cre group compared with the HDAC4fl/fl group (P<0.05; Fig. 5M-R). ISH confirmed increased ColX and decreased Wnt5a expression (red boxes and arrows) in the HDAC4d/d, Col2α1-Cre mice (Fig. 5S-Z). These results indicated that HDAC4 deletion reduce chondrocyte proliferation via Wnt5a signaling pathway inhibition and increased chondrocyte hypertrophy.

Figure 5. Decreased chondrocyte proliferation due to HDAC4 deletion. (A-L) Immunohistochemical staining of the tibial articular cartilage of HDAC4fl/fl and HDAC4d/d, Col2α1-Cre mice at P8 for determining the expressions of BrdU, PCNA, MMP-13, Runx2, OPG. and CD34 Positive signals (brown staining) are indicated by red arrows. Scale bars, 20 µm. (M-R) Semiquantitative analysis of positively expressed areas for BrdU, PCNA, MMP-13, Runx2, OPG and CD34. (S-Z) In situ hybridization on paraffin sections prepared for determining the expressions of Col X and Wnt5a in tibial tissues at P14. The HDAC4d/d, Col2α1-Cre mice had decreased expressions of BrdU, PCNA and Wnt5a but increased expressions of MMP-13, Runx2, OPG, Col X and CD34. Boxed areas in S-V indicate the regions shown in W-Z, respectively. Scale bars, 50 µm. *P<0.05 vs. HDAC4fl/fl mice. P, postnatal day; HDAC4, histone deacetylase 4; Col2α1, collagen type 2α1; fl, floxed; d, down; BrdU, bromodeoxyuridine; PCNA, proliferating cell nuclear antigen; MMP-13, matrix metalloproteinase-13; Runx2, runt-related transcription factor 2; Col X, type X collagen; OPG, osteoprotegerin.

Genetic changes associated with increased hypertrophy and mineralization and decreased proliferation in HDAC4d/d, Col2α1-Cre mice

qPCR analyses indicated that the mRNA levels of HDAC4 and PCNA were lower and those of MMP-13, Runx2, OPN and osteocalcin were higher in the HDAC4d/d, Col2α1-Cre mice compared with those of HDAC4fl/fl mice (all P<0.05; s. 6A-F). These results indicated that there are genetic changes associated with increased hypertrophy and mineralization and decreased proliferation in HDAC4d/d, Col2α1-Cre mice.

Discussion

The deletion of HDAC4 results in precocious lethality in mice (9); however, the mechanism underlying HDAC4-mediated postnatal regulation of growth plate development remains unclear. A previous study demonstrated that HDAC4 is a critical negative regulator of chondrocyte hypertrophy due its binding to and inhibition of Runx2 (37). To further investigate the role of chondrocyte-derived HDAC4 in postnatal skeletal development, the present study selectively ablated the HDAC4 gene in collagen type 2-expressing cells using the Cre/loxP gene targeting technique.

The findings of the current study demonstrated that HDAC4-deletion in Col2α1-expressing cells results in severely runted and slow-growing mice postdelivery, which is consistent with the result reported by a previous study (9). In addition, the expressions of typical proliferative markers BrdU and PCNA were significantly lower in the HDAC4d/d, Col2α1-Cre mice compared with in their HDAC4fl/fl control littermates, which suggested HDAC4-deletion is associated with decreased chondrocyte growth. ISH demonstrated that the HDAC4d/d, Col2α1-Cre mice had a downregulated expression of Wnt5a compared with that in the HDAC4fl/fl mice. Recent studies have also reported that Wnt signaling is one of the numerous signaling pathways involved in the regulation of diverse processes underpinning fundamental and normal development, such as apoptosis, proliferation, differentiation and polarity (38,39).

Wnt5a-knockout mice demonstrate perinatal lethality primarily due to respiratory failure and presented extensive developmental abnormalities (15). Wnt5a and its signaling pathway can regulate fundamental cellular processes, including proliferation and survival. These results suggest that Wnt5a is a potential target for treating proliferation-related diseases (40). The present findings suggested that HDAC4-deletion results in decreased chondrocyte proliferation partially due to the inhibition of Wnt5a expression.

The HDAC4d/d, Col2α1-Cre mice exhibited premature and disordered hypertrophic chondrocytes as well as increased levels of MMP-13 and type X collagen. MMP-13 and collagen type X are essential for maintaining chondrocyte hypertrophy and allowing ossification (41,42). It is universally accepted that hypertrophic chondrocytes implicate the final step in the maturation process of chondrocytes, which eventually results in apoptosis of chondrocytes and replacement of hypertrophic chondrocytes by bone tissues (43). Ossification begins when hypertrophic chondrocytes undergo programmed cell death, and the calcified cartilage is invaded by blood vessels, osteoblasts, osteoclasts and mesenchymal precursors, thereby forming primary and secondary ossification centers (3,4). The degradation and remodeling of the cartilage matrix is necessary for vascular invasion, whereas angiogenesis has been implicated as a crucial step in endochondral ossification (5). The present study identified that HDAC4 gene ablation can accelerate the mineralization and ectopic ossification of endochondral cartilages. The expressions of typical mineralization markers, including Runx2, OPN and osteocalcin, were increased in the HDAC4d/d, Col2α1-Cre mice. Therefore, it was hypothesized that premature and disordered chondrocyte hypertrophy in HDAC4-knockout mice accelerates the malformation of bones and results in runted mice.

Notably, the vascular invasion and secondary center of ossification of the growth plates began to appear at P8 in the HDAC4d/d, Col2α1-Cre mice, whereas that in the control group began to appear at P14. Vascular invasion of the primarily avascular hypertrophic chondrocyte zone of the growth plate is a prerequisite for endochondral bone formation and occurs in a sequence of events (44). Blood vessels supplying bones orchestrate the process of bone development and remodeling as well as regulate skeleton regeneration by delivering nutrients, oxygen, hormones or growth factors to bone cells (45). This process requires a dynamic and close interaction between hypertrophic chondrocytes and vascular invasion. The expressions of typical vascular formation markers, such as CD34 (a widely used marker of hematopoietic and endothelial progenitor cells), were reportedly increased in the tibia of the HDAC4d/d, Col2α1-Cre mice (46). The findings of the present study demonstrated that knockout of the HDAC4 gene accelerates endochondral bone formation by promoting vascular invasion.

There are some limitations in this experiment. The number of animals was small and no cell experiment was conducted. More research is needed on the HDAC4 pathway and in the future Wnt5a loss and gain of function studies in vitro and in vivo should be performed to confirm the regulating effect of the HDAC4-Wnt5a pathway. HDAC4-Wnt5a-β-catenin signaling activity should be analyzed as it is speculated that HDAC4-Wnt5a-β-catenin may be an indispensable role in the homeostasis of articular cartilage (47). In the present study, the HDAC4d/d, Col2α1-Cre mice displayed a marked phenotype characterized by decreased chondrocyte hypertrophy, early blood invasion and premature bone formation, leading to abnormal growth and runted mice. These results demonstrated that chondrocyte-derived HDAC4 is primarily responsible for the regulation of chondrocyte differentiation and endochondral bone formation.

Acknowledgements

The authors would like to thank Dr Olson (University of Texas Southwestern Medical Center) who provided the HDAC4fl/fl mice and technical support.

Funding

The current study was supported by The National Institute of General Medical Sciences, National Institutes of Health (grant no. P30GM122732), RIH Orthopedic Foundation and The National Natural Science Foundation of China (grant no. 81704103, mainly used for the author's study expenses in the United States).

Availability of data and materials

All data generated or analyzed during this study are included in this published article.

Authors' contributions

LW and HZ designed the study. GD, CX and XS performed the experiments. YS, PL and XWe collected the experimental data. MZ, SW, LG, XWa and NW analyzed the experimental data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The present study was approved by the Institutional Animal Care and Use Committee of Rhode Island Hospital.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References