CREB activation in hypertrophic chondrocytes is involved in the skeletal overgrowth in epiphyseal chondrodysplasia Miura type caused by activating mutations of natriuretic peptide receptor B

Abstract

Natriuretic peptide receptor B (NPRB) produces cyclic guanosine monophosphate (cGMP) when bound by C-type natriuretic peptide (CNP). Activating mutations in NPRB cause a skeletal overgrowth disorder, which has been named epiphyseal chondrodysplasia, Miura type (ECDM; OMIM #615923). Here we explored the cellular and molecular mechanisms for the skeletal overgrowth in ECDM using a mouse model in which an activating mutant NPRB is specifically expressed in chondrocytes. The mutant mice (NPRB[p.V883M]-Tg) exhibited postnatal skeletal overgrowth and increased cGMP in cartilage. Both endogenous and transgene-derived NPRB proteins were localized at the plasma membrane of hypertrophic chondrocytes. The hypertrophic zone of growth plate was thickened in NPRB[p.V883M]-Tg. An in vivo BrdU-labeling assay suggested that some of the hypertrophic chondrocytes in NPRB[p.V883M]-Tg mice continued to proliferate, although wild-type (WT) chondrocytes stopped proliferating after they became hypertrophic. In vitro cell studies revealed that NPRB activation increased the phosphorylation of cyclic AMP-responsive element binding protein (CREB) and expression of cyclin D1 in matured chondrocytes. Treatment with cell-permeable cGMP also enhanced the CREB phosphorylation. Inhibition of cyclic adenosine monophosphate (cAMP)/protein kinase A pathway had no effects on the CREB phosphorylation induced by NPRB activation. In immunostaining of the growth plates for the proliferation marker Ki67, phosphorylated CREB and cyclin D1, most signals were similarly observed in the proliferating zone in both genotypes, but some cells in the hypertrophic zone of NPRB[p.V883M]-Tg were also positively stained. These results suggest that NPRB activation evokes its signal in hypertrophic chondrocytes to induce CREB phosphorylation and make them continue to proliferate, leading to the skeletal overgrowth in ECDM.

Introduction

Longitudinal growth of long bones occurs in the growth plates located between the epiphysis and metaphysis. The growth plate is a cartilaginous tissue having a highly organized structure with horizontal zones of chondrocytes at different stages of maturation (1–4). The resting zone at the epiphyseal end of the growth plate contains quiescent chondrocytes, and is also called the reserve zone or stem cell zone. When these chondrocytes start cell division, they enter the proliferating zone, where flattened chondrocytes proliferate unidirectionally to form orderly parallel columns and produce specific extracellular matrix (ECM) proteins including type II collagen (Col II) and aggrecan. The cells away from the epiphyseal end exit the cell cycle and differentiate to become prehypertrophic chondrocytes with an increase in cell size (5). Subsequently, the cells mature to hypertrophic chondrocytes with a bigger cell size and produce type X collagen (Col X). Finally, the cartilaginous matrix is mineralized and invaded by vessels, and the cells undergo apoptosis or transdifferentiate into osteoblastic cells. Both the proliferation and hypertrophy of chondrocytes contribute to the longitudinal growth of the bone, and various molecules including transcription factors, soluble factors and ECMs regulate these processes. Defects in these factors are often associated with skeletal dysplasias (1–4).

Natriuretic peptide receptor type B (NPRB), which is encoded by the NPR2 gene in humans and is also known as guanylyl cyclase (GC)-B, functions as a homodimeric receptor-type GC that catalyzes the synthesis of the intracellular signaling molecule cyclic guanosine monophosphate (cGMP) (6–8). When bound by its cognate ligand C-type natriuretic peptide (CNP), NPRB shows an enhanced GC activity to produce an increased amount of cGMP. There is another type of receptor for CNP, natriuretic peptide clearance receptor (NPRC), which lacks the intracellular domain and serves to clear the ligands of the natriuretic peptide family including CNP (6–8).

Results from both mouse studies and analyses of human diseases have provided evidence for a significant role of the CNP/NPRB signaling pathway in skeletal growth. Transgenic overexpression of CNP accelerated, while the knockout of its gene Nppc inhibited, the skeletal growth in mice (9–11). Loss-of-function mutations in both alleles of the human NPR2 gene cause acromesomelic dysplasia, Maroteaux type (OMIM #602875) characterized by severe dwarfism (12), which resembles the phenotype of Npr2-knockout mice (13). In addition, it has been revealed that heterozygous loss-of-function mutations cause idiopathic short stature with mild skeletal defects (OMIM #616255) (14–16).

In 2012, we reported the first family of the autosomal-dominant skeletal overgrowth disorder caused by an activating mutation of NPRB (17). The affected members of the family exhibited extremely tall stature and macrodactyly of the great toes. The mutation identified in this family was a conserved valine-to-methionine substitution at codon 883 (p.V883M; c.2647G>A) in the catalytic domain of human NPRB (17). In vitro transfection experiments and detailed enzymatic analyses revealed a gain of function of the NPRB mutant (17,18). Since then, two more families were reported to have NPRB-associated skeletal overgrowth (19,20), and the condition has been established as a disease entity and named epiphyseal chondrodysplasia, Miura type (ECDM; OMIM #615923). The mutations identified in the second and third families were an arginine-to-cysteine substitution at codon 655 (p.R655C; c.1963C>T) and an alanine-to-proline substitution at codon 488 (p.A488P; c.1462G>C), respectively, and both were confirmed to be gain-of-function-type mutations (19,20). These findings clearly indicate that activation of NPRB leads to cartilage-mediated skeletal growth, although the underlying cellular and molecular mechanisms are not fully understood. In our previous work, we generated a transgenic mouse line expressing the NPRB[p.V883M] identified in the first family of ECDM, specifically in chondrocytes. These mice exhibit skeletal overgrowth similar to the phenotype of human ECDM and can be used as a disease model (17). In the current study, utilizing this mouse model and chondrocytic cell models, we clarify the cellular and molecular mechanisms by which the activated NPRB signaling leads to the skeletal overgrowth in ECDM.

Results

Skeletal overgrowth of NPRB[p.V883M]-Tg mice is attributable to thickening of the hypertrophic zone of the growth plate

Consistent with our previous report (17), both male and female NPRB[p.V883M]-Tg mice exhibited postnatal skeletal overgrowth (Fig. 1A). We also confirmed increased cGMP content in the cartilage from NPRB[p.V883M]-Tg mice compared with that from wild-type (WT) littermates (Fig. 1B), which was attributable to the gain of function of NPRB[p.V883M] (17,18). In immunohistochemical staining of growth plates using an antibody raised against amino acids 171–250 of human NPRB and able to react with both human mutant NPRB derived from the transgene and the murine endogenous NPRB, signals were detected predominantly in the plasma membrane of hypertrophic chondrocytes in both NPRB[p.V883M]-Tg mice and WT littermates, and the spatial staining pattern was similar between the genotypes (Fig.1F). The absence of ectopic expression of NPRB supports the suitability of NPRB[p.V883M]-Tg as a model for ECDM.

Figure 1

Skeletal overgrowth in NPRB[p.V883M] was attributable to thickening of the hypertrophic zone of the growth plate cartilage. (A) The growth curves of naso-tail length in NPRB[p.V883M]-Tg mice (closed squares) and WT littermates (open squares) from 3 weeks to 8 weeks of age. Data are shown as the mean ± SD (n = 3). ^*P < 0.05 versus WT; ^**P < 0.01 versus WT. (B) Cyclic GMP content in the costal cartilage harvested from 9-day-old NPRB[p.V883M]-Tg mice and WT littermates. The cGMP content was standardized based on the DNA content in each sample. Data are shown as the mean ± SD (n =3). ^*P < 0.05 versus WT. (C) Representative images of HE-stained sections of the tibial growth plates from 4-week-old male NPRB[p.V883M]-Tg mice and WT littermates. P, proliferating zone; PH, prehypertrophic zone; H, hypertrophic zone. (D, E) Thickness of the proliferating zones (D) and prehypertrophic to hypertrophic zones (E) of growth plates was measured in the HE-stained sections. The thickness of the prehypertrophic to hypertrophic zones was increased in NPRB[p.V883M]-Tg mice compared to WT littermates. Data are shown as the mean ± SD (n = 3). ^*P < 0.05 versus WT. (F–H) Representative images of immunostaining for NPRB (F), Col II (G) and Col X (H). In (F), the utilized antibody could react with the transgene-derived human mutant NPRB as well as the endogenous mouse NPRB, and the signals were detected at the plasma membrane of hypertrophic chondrocytes in both WT and NPRB[p.V883M]-Tg mice.

To clarify the mechanisms of the skeletal overgrowth in NPRB[p.V883M]-Tg mice, we closely analyzed their growth plates. For this purpose, we harvested tibiae from the male mice at 4 weeks of age, when the naso-tail length became significantly longer in NPRB[p.V883M]-Tg mice. In hematoxylin-eosin (HE)-stained sections, there was no significant difference in the thickness of the proliferating zone where the flat chondrocytes formed columns (Fig. 1C and D). In contrast, the prehypertrophic/hypertrophic zone was significantly thicker in the NPRB[p.V883M]-Tg mice (Fig. 1C and E). Immunostaining for NPRB also demonstrated thickening of the hypertrophic zone in the growth plates of NPRB[p.V883M]-Tg mice (Fig. 1F). Immunostaining for Col II and Col X also suggested that the hypertrophic zone rather than the proliferating zone was thickened in the growth plates of NPRB[p.V883M]-Tg mice (Fig. 1G and H).

Increase in BrdU-positive cells in the hypertrophic zone of NPRB[p.V883M]-Tg mice

To analyze the cellular basis for the thickening of growth plate in NPRB[p.V883M]-Tg mice, we evaluated the proliferation of chondrocytes by in vivo BrdU (5'-bromo-2'-deoxyuridine) labeling assay. Since the mice were sacrificed 2 h after the injection of BrdU, positive staining with the anti-BrdU antibody indicated that the cells had been proliferating during those 2 h. We identified the proliferating, prehypertrophic and hypertrophic zones of growth plates in NPRB[p.V883M]-Tg and WT mice based on the cell morphology in the HE-stained sections (Fig. 2B), and then calculated the BrdU labeling index in each zone. In the growth plates of WT mice, about 15% of the cells in the proliferating zone were labeled with BrdU, but virtually no cells had positive signals in the hypertrophic zone. In contrast, in NPRB[p.V883M]-Tg mice, about 4% of the cells were positively stained for BrdU in the hypertrophic zone (Fig. 2A and C). These results suggest that some of the hypertrophic chondrocytes continued to proliferate in the NPRB[p.V883M]-Tg mice, although WT chondrocytes stopped proliferating when they became hypertrophic.

Figure 2

Increase in BrdU-Positive cells in the hypertrophic zone of NPRB[p.V883M]-Tg mice. (A) Representative images of in vivo BrdU labeling. BrdU was injected into 4-week-old male NPRB[p.V883M]-Tg mice and WT littermates. Two hours after the injection, tibiae were harvested and subjected to immunostaining with anti-BrdU antibody. P, proliferating zone; PH, prehypertrophic zone; H, hypertrophic zone; and PS, primary spongiosa. The arrows indicate BrdU-positive cells in the hypertrophic zone. (B) Cell morphology of proliferating (P), prehypertrophic (PH), and hypertrophic (H) chondrocytes in high power images of HE-stained sections. (C) BrdU labeling indexes were calculated in each zone by dividing the numbers of BrdU-positive nuclei by those of total cell nuclei. Data are shown as the mean ± SD (n = 3–4). (D) Representative images of TUNEL-stained sections of the tibial growth plates from 4-week-old male mice. The arrows indicate TUNEL-positive cells. (E) The calculated ratio of the numbers of TUNEL-positive cells to those of total cells in each zone. Data are shown as the mean ± SD (n = 3–4).

We also carried out TUNEL (TdT-mediated dUTP nick end labeling) staining of the growth plates to evaluate apoptosis, and found no significant difference in the rate of TUNEL-positive cells between WT and NPRB[p.V883M]-Tg mice (Fig. 2D and E).

Activation of NPRB re-increased the expression of cyclin D1 in matured ATDC5 chondrocytic cells

As a next step, we attempted to identify the molecules involved in the growth plate thickening in NPRB-Tg mice. The predominant expression of NPRB in the hypertrophic chondrocytes (Fig. 1F) and the increased BrdU-labeled cells in the hypertrophic zone of NPRB[p.V883M]-Tg mice (Fig. 2A and C) suggested that the NPRB-mediated signal exerts its effects on chondrocytes in a differentiation stage-dependent manner, and that the hypertrophic chondrocyte is likely to be its main cellular target. To screen for the molecules that mediate the effects of NPRB activation in the later stage of chondrocyte maturation, we took advantage of the murine cell line ATDC5, a commonly used cell model for endochondral ossification (21–23). When cultured in chondrogenic medium, ATDC5 cells reproduce the multistep chondrogenic differentiation consisting of mesenchymal condensation, proliferation, hypertrophy and mineralization in vitro (21–23). When chondrogenesis is induced as described in the Materials and Methods section, ATDC5 cells will become hypertrophic at around 6 weeks of culture, which will be followed by mineralization of ECMs (22,23).

To delineate the stage-specific effects of NPRB activation on gene expression, we cultured ATDC5 cells in chondrogenic medium for 0, 2, 4, 6 or 8 weeks and then transiently activated NPRB by treating the cells with 1 μM CNP for 48 h. RNA was extracted from these cells and subjected to real-time PCR analyses for the genes involved in chondrocyte proliferation and differentiation (Fig. 3). The expression of Sox9, the key transcription factor of chondrocyte differentiation (24), peaked at 2 weeks of culture and then declined in vehicle-treated cells, which confirmed that the cells underwent the chondrogenic process from mesenchymal condensation and proliferation to hypertrophic maturation. The expression of Runx2, which is involved in chondrocyte hypertrophy (25), remained high after 2 weeks of culture. Transient activation of NPRB by a 48 h treatment with CNP had no effects on the expression of Sox9 and Runx2 at all time points analyzed. The Npr2 gene encoding NPRB was highly expressed in the cells cultured in chondrogenic medium for 2 weeks and longer, while the expression of Nppc encoding CNP reached a maximum at 2 weeks of culture and rapidly decreased thereafter. We also analyzed the expression of Npr3 encoding NPRC, the clearance receptor for CNP and found that its expression pattern during the differentiation of ATDC5 cells was similar to that of Npr2. The transient activation of NPRB by a 48 h treatment with 1 μM CNP caused no significant change in the expression of Npr2, Nppc or Npr3 (Fig. 3).

Figure 3

NPRB activation in matured chondrocytic ATDC5 cells re-increased the expression of Ccnd1 encoding cyclin D1. To screen for genes whose expression was altered by NPRB activation specifically in mature chondrocytes, ATDC5 cells were cultured in chondrogenic medium for the indicated period of time and then were treated for 48 h with 1 μM CNP, to activate NPRB (closed bars), or vehicle (open bars). After the treatment, RNA was extracted and subjected to real-time PCR analyses. The copy number of the target cDNA in each sample was estimated by referring to a standard curve and was standardized by that of Gapdh. In vehicle-treated cells, the expression of Ccnd1 peaked at 4 weeks of culture and declined thereafter. In contrast, CNP-induced NPRB activation in matured ATDC5 cells at 6 and 8 weeks of culture up-regulated the Ccnd1 expression to a level similar to that at 4 weeks. Data are shown as the mean ± SEM (n = 3). ^*P < 0.05 versus the values at 0 week in each group; ^#P < 0.05 versus vehicle-treated cells; W, weeks.

Among the genes analyzed, Ccnd1 (Cyclin D1) encodes a key regulator of the G1-phase progression of the cell cycle (26), which is essential for cell proliferation. In vehicle-treated cells, the expression of Ccnd1 peaked at 4 weeks of chondrogenic induction and then decreased, and this decline might reflect that the cells exited the cell cycle to become hypertrophic. Interestingly, activation of NPRB by CNP up-regulated the Ccnd1 expression when it was added at 6 or 8 weeks of chondrogenic induction, to a level similar to that at 4 weeks (Fig. 3).

Regarding the genes for ECMs, the expressions of Col2a1, Col10a1 and Acan encoding Col II, Col X and aggrecan, respectively, were examined. In addition, the expressions of the genes for Indian hedgehog (Ihh), parathyroid hormone related protein (Pthrp) and its receptor parathyroid hormone 1 receptor (Pth1r) were analyzed. The treatment with CNP for 48 h up-regulated the expressions of Col2a1 and Pthrp at 2 weeks of culture and those of Pth1r at 2 and 6 weeks (Fig. 3). Addition of CNP to ATDC5 cells markedly increased the production of cGMP, confirming the activation of NPRB (Fig 4A).

Figure 4

NPRB activation in matured chondrocytic ATDC5 cells enhanced the phosphorylation of CREB at Ser133. (A) Cyclic GMP production in chondrocytic ATDC5 cells by CNP-induced NPRB activation. ATDC5 cells cultured in chondrogenic medium for 4 or 8 weeks were serum-starved for 24 h. After pre-incubation with 0.5 mm IBMX for 10 min, the cells were treated for another 10 min with 1 μm CNP to activate NPRB (closed bars) or vehicle (open bars) before cell lysates were harvested for cGMP measurement. The cGMP content was standardized based on the protein content in each sample. Data are shown as the mean ± SD (n = 3). (B) To screen for signaling molecules modulated by NPRB activation in mature chondrocytes, ATDC5 cells were cultured in chondrogenic medium for the indicated period of time. After serum-starvation for 24 h, the cells were pre-incubated with 0.5 mm IBMX for 10 min and then treated with 1 μM CNP to activate NPRB (+) or vehicle (−) for 30 min in the absence of serum. Total cell lysates were harvested and subjected to Western blotting. CNP-induced NPRB activation in the cells cultured for 6 and 8 weeks enhanced the phosphorylation of CREB. (C) Densitometry was carried out to evaluate the ratio of the intensity of the signals corresponding to phosphorylated CREB to that of total CREB in the cells at 6 and 8 weeks of culture shown in (B).

Activation of NPRB induced the phosphorylation of cyclic AMP-responsive element binding protein in matured ATDC5 cells

We also performed Western blotting analyses to screen for the signaling pathways that might mediate the effects of NPRB activation in hypertrophic chondrocytes. ATDC5 cells were cultured in chondrogenic medium for 2, 4, 6 or 8 weeks and then were subjected to 24 h starvation in serum-free media. Based on the gene expression profile shown in Fig. 3, fully matured ATDC5 cells after 6-week chondrogenic induction were considered as a model for hypertrophic chondrocytes. After a 10-minute pre-treatment with 0.5 mm IBMX, the cells were treated with 1 μM CNP or vehicle for 30 min before cell lysates were harvested. Immunoblotting was carried out to examine the phosphorylation status of cyclic AMP-responsive element binding protein (CREB), extracellular signal-regulated kinase 1/2 (ERK1/2), p38 mitogen-activated protein kinase (p38 MAPK), Akt and Smads. Among these molecules, we found that the CNP-induced activation of NPRB resulted in increased phosphorylation of CREB in ATDC5 cells at 6 and 8 weeks of chondrogenic induction (Fig. 4B and C). This is interesting because it is well established that CREB activates the transcription of the Ccnd1 gene (5,27). Thus, in matured ATDC5 cells after 6 weeks of culture in chondrogenic medium, NPRB-mediated signaling caused the activation of CREB, which led to a re-increase of Ccnd1 expression.

Activation of NPRB induced the phosphorylation of CREB and up-regulated Ccnd1 in primary chondrocytes

Since the experiments using ATDC5 cells suggested that activation of NPRB led to phosphorylation of CREB and increased the expression of Ccnd1, we confirmed these observations in primary chondrocytes derived from rib cage cartilage of WT mouse neonates. In the primary chondrocytes, a 10-minute treatment with CNP (10^-8 or 10^-7 M) after a pre-incubation with 0.5 mm IBMX increased the production of cGMP in a dose-dependent manner (Fig. 5C), indicating the activation of NPRB. The amount of cGMP produced by CNP stimulation was much larger in the primary chondrocytes than in ATDC5 cells, suggesting their higher sensitivity to CNP (Figs 4A and5C). The activation of NPRB by a 30-minute treatment with CNP resulted in the phosphorylation of CREB (Fig. 5A). Since it was previously reported that treatment with CNP inhibited FGF2-induced activation of ERK1/2, we also examined the effects of CNP on ERK1/2 phosphorylation. In our experiments, treatment with CNP alone had no effects on ERK1/2 phosphorylation in the primary chondrocytes (Fig. 5B).

Figure 5

NPRB activation enhanced the phosphorylation of CREB at Ser133 and increased the protein level of cyclin D1 in WT primary chondrocytes. (A, B) Activation of NPRB enhanced the phosphorylation of CREB (A) but not that of ERK1/2 (B). Chondrocytes isolated from the rib cages of WT mouse neonates were cultured for 14 days. After serum-starvation for 6 h, the cells were pre-incubated with 0.5 mm IBMX for 10 min and then treated for 30 min with CNP of the indicated concentrations or vehicle in the absence of serum, before cell lysates were harvested for western blot analysis. The densitometry data are shown as the mean ± SD of the values in three independent experiments. ^*P < 0.05. (C) CNP-induced NPRB activation increased the cGMP production in cultured primary chondrocytes. Chondrocytes isolated from WT mouse neonates were cultured for 14 days. After pre-incubation in serum-free medium containing 0.5 mm IBMX for 10 min, the cells were treated with CNP (10^-8 – 10^-7 M) or vehicle for another 10 min before the cell lysates were harvested for analysis. The cGMP content was standardized based on the protein content in each sample. Data are shown as the mean ± SD (n = 3). ^**P < 0.01. (D) CNP-induced NPRB activation increased the protein level of cyclin D1 in cultured primary chondrocytes. After serum-starvation for 24 h, the cells were treated with CNP (10^-7 M) or vehicle for 6 or 24 h before cell lysates were harvested for western blot analysis. The densitometry data are shown as the mean ± SD of the values in three independent experiments. ^*P < 0.05.

We also investigated the effects of NPRB activation on the gene expression in WT primary chondrocytes (Fig. 6). The cells were treated with CNP (10^-8 or 10^-7 M) or vehicle for a total of 48 h by replacing the stimulant- or vehicle-containing medium every 12 h, i.e. 4 times. The expression of Npr2 encoding NPRB itself was not changed by CNP treatment. In contrast, the expression of Npr3 encoding the clearance receptor NPRC was markedly elevated by CNP stimulation. This observation is consistent with a previous report in which the authors demonstrated that treatment with CNP increased the expression of Npr3 in the hypertrophic zone of growth plate of mouse tibiae in organ culture (28). The expression of Ccnd1 was increased by the treatment with 10^-8 M CNP, although the effect of 10^-7 M CNP was not significant, probably due to the accelerated clearance of CNP by NPRC (Fig. 6). Western blotting confirmed that CNP-induced NPRB activation increased the protein level of cyclin D1 in primary chondrocytes (Fig. 5D).

Figure 6

Effects of NPRB activation on gene expression in WT primary chondrocytes. Primary chondrocytes isolated from WT mouse neonates were cultured for 14 days and then treated with CNP (10^-8 – 10^-7 M) to activate NPRB or vehicle for a total of 48 h, replacing the media every 12 h. After the treatment, RNA was extracted and subjected to real-time PCR analyses. The copy number of the target cDNA in each sample was estimated by referring to a standard curve and was standardized by that of Gapdh. Data are shown as the mean ± SD (n = 3). ^*P < 0.05.

Among the genes analyzed, the expression of Col2a1 was also significantly increased by 10^-7 M CNP. In addition, the expression of Sox9, Pthrp and Ihh tended to increase by CNP treatment (Fig. 6). These results suggest that CNP-induced activation may influence both proliferation and differentiation of chondrocytes. However, considering that Ccnd1 is known as a direct transcriptional target of CREB (27) and that BrdU-positive cells were detected in the hypertrophic zone of growth plate in NPRB[p.V883M]-Tg mice (Fig. 2), we have focused on whether CREB and cyclin D1 are involved in the skeletal overgrowth of ECDM in the following analyses.

The presence of cells positively stained for Ki67, phosphorylated CREB and cyclin D1 in the hypertrophic Zone of NPRB[p.V883M]-Tg mice

The results of the cell studies described above suggest that activation of NPRB leads to the phosphorylation of CREB and up-regulation of Ccnd1 in mature chondrocytes. Therefore, we conducted immunostaining to examine whether the phosphorylation of CREB and expression of cyclin D1 were indeed enhanced in the hypertrophic chondrocytes of NPRB[p.V883M]-Tg mice.

In immunostaining for Ki67, a marker for cell proliferation, many cells in the proliferating zones and the primary spongiosa, and some cells in the prehypertrophic zones, were positively stained in both WT and NPRB[p.V883M]-Tg mice. However, in the hypertrophic zone, we detected some cells positively stained for Ki67 in NPRB[p.V883M]-Tg mice but not in WT mice (Fig. 7A). These observations were consistent with the results of BrdU-labeling experiments (Fig. 2A and C). We then immunostained the adjacent sections using the antibody against phosphorylated CREB and found a staining pattern similar to that with anti-Ki67. In both genotypes, abundant signals for phosphorylated CREB were observed in the proliferating zone. In the hypertrophic zone, some cells were positively stained for phosphorylated CREB in NPRB[p.V883M]-Tg, although the signals were very few in WT mice (Fig. 7A). We found that the signals for Ki67 and those for CREB overlapped in the same cells in the hypertrophic zone of NPRB[p.V883M]-Tg mice (Fig. 7B).

Figure 7

Immunostaining for Ki67 and phosphorylated CREB in the growth plates of NPRB[p.V883M]-Tg mice and WT littermates. (A) Representative images of the serial sections cut from the tibiae of 4-week-old male NPRB[p.V883M]-Tg mice and WT littermates and subjected to HE-staining and immunostaining for the proliferation marker Ki67 and CREB phosphorylated at Ser133. P, proliferating zone; PH, prehypertrophic zone; H, hypertrophic zone; and PS, primary spongiosa. The arrows indicate signals in the hypertrophic zones. (B) Higher magnification images of the hypertrophic zones of NPRB[p.V883M]-Tg shown in lower center and lower right panels in (A). Some signals for Ki67 and those for phosphorylated CREB are overlapping in the same cells (arrowheads).

As expected, in immunostaining for cyclin D1, most signals were detected in the proliferating zone and the primary spongiosa in both WT and the mutant mice. We also observed some positively stained cells in the hypertrophic zone of NPRB[p.V883M]-Tg mice (Fig. 8A). However, the signals for cyclin D1 were fewer than those for phosphorylated CREB, which might have been due to the previously reported instability of cyclin D1 protein (29,30). We detected cells positively stained for both cyclin D1 and phosphorylated CREB in the hypertrophic zone of NPRB[p.V883M]-Tg mice (Fig. 8B).

Figure 8

Detection of cyclin D1 expression in the hypertrophic zone of the growth plates in NPRB[p.V883M-Tg] mice. (A) Representative images of the immunostaining for phosphorylated CREB and cyclin D1 of the serial sections cut from tibiae of 4-week-old male NPRB[p.V883M]-Tg mice and WT littermates. Images of HE-stained sections are also shown. P, proliferating zone; PH, prehypertrophic zone; H, hypertrophic zone; and PS, primary spongiosa. The arrows indicate signals in the hypertrophic zones. (B) Higher magnification images of the hypertrophic zones of NPRB[p.V883M]-Tg shown in lower center and lower right panels in (A). The arrowheads indicate the cells positively stained for both phosphorylated CREB and cyclin D1.

Co-treatment with inhibitors against CREB or cyclin D1 abolished the increase in chondrocyte proliferation by CNP treatment

To further investigate whether CREB and cyclin D1 mediate the skeletal overgrowth in ECDM, we next examined the effects of NPRB activation and the inhibitors against CREB and cyclin D1 on the proliferation of primary chondrocytes by MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)- 2H-tetrazolium, inner salt] assay. We used 666-15 as a CREB inhibitor (31) and imperatorin as a cyclin D1 inhibitor (32). The cell number after 2 day treatment was significantly increased in CNP-treated cells compared with in vehicle-treated cells, and this effect of CNP-induced NPRB activation was abolished by co-treatment with either the CREB inhibitor 666-15 or the cyclin D1 inhibitor imperatorin (Fig. 9). These findings support the notion that NPRB activation may promote chondrocyte proliferation through CREB phosphorylation and up-regulation of cyclin D1.

Figure 9

NPRB activation accelerated the proliferation of primary chondrocytes, which was abolished by co-treatment with either a CREB inhibitor or a cyclin D1 inhibitor. Primary chondrocytes isolated from WT mice were plated into 96-well culture plates (1 × 10⁵ cells/well) and cultured for 24 h. Then, the cells were added with either a CREB inhibitor (666-15; 10^-6 M), a cyclin D1 inhibitor (imperatorin; 10^-4 M) or the solvent (DMSO), which was followed by an addition of CNP (10^-7 M) or its corresponding vehicle. Two days later, the cell number in each well was evaluated by an MTS assay, represented by the corrected absorbance at optical density (O.D.) 490 nm. Data are shown as the mean ± SD (n = 3). ^*P < 0.05.

Activation of NPRB induces phosphorylation of CREB through cGMP production and independently of the PKA pathway

Our findings described above suggest that the activated NPRB signaling in the hypertrophic chondrocytes of NPRB[p.V883M]-Tg mice may result in the increased phosphorylation of CREB and the up-regulation of cyclin D1 in these cells, leading to their ectopic proliferation and thickening of the hypertrophic zone. To confirm that this process is mediated by cGMP, we examined the effects of 8-Bromo-cGMP, a cell-permeable analogue for cGMP, on the phosphorylation of CREB in WT primary chondrocytes. The treatment with 8-Bromo-cGMP enhanced the phosphorylation of CREB in a dose-dependent manner, implying that the effects of activated NPRB signaling on the CREB phosphorylation were mediated by the increased production of cGMP (Fig. 10A).

Figure 10

NPRB activation in primary chondrocytes induced the phosphorylation of CREB at Ser133 by cGMP production and not through the cAMP/PKA pathway. (A) Treatment of WT primary chondrocytes with the cell-permeable cGMP analogue 8-Br-cGMP induced the phosphorylation of CREB at Ser133. Chondrocytes isolated from the rib cages of WT mice were cultured for 14 days. After serum-starvation for 6 h, the cells were treated for 30 min with 8-Br-cGMP or vehicle. Total cell lysates were then harvested for western blotting. The densitometry data are shown as the mean ± SD of the values in three independent experiments. ^*P < 0.05 versus vehicle-treated cells. (B) NPRB activation did not lead to increased production of cAMP. WT primary chondrocytes were cultured for 14 days and serum-starved for 6 h. After pre-incubation in 0.5 mm IBMX for 10 min, the cells were treated for another 10 min with PTHrP (10^-8 M) as a positive control or CNP (10^-8 – 10^-7M) before the lysates were harvested for cAMP measurement. The cAMP content was standardized based on the protein content in each sample. Data are shown as the mean ± SD (n = 3). ^**P < 0.01 versus vehicle-treated cells. (C, D) Effects of a PKA inhibitor, H-89, on the CREB phosphorylation induced by NPRB activation (C) and PTHrP (D). After serum-starvation, WT primary chondrocytes were pre-treated with 10 μm H-89 or vehicle, and then the indicated stimulants or vehicle was added to the medium. Thirty minutes later, cell lysates were harvested for western blot analysis. The experiments were performed in the absence of serum. The densitometry data are shown as the mean ± SD of the values in three independent experiments. Co-treatment with H-89 had no effects on the CREB phosphorylation induced by CNP, although it suppressed the CREB phosphorylation induced by PTHrP.

It is well established that PTHrP produced by perichondrial cells and chondrocytes at the ends of long bones delays chondrocyte hypertrophy through PTH1R, expressed in proliferating and prehypertrophic chondrocytes, and keeps them in the proliferation stage, and this action of PTHrP is mainly mediated by the cAMP/protein kinase A (PKA)/CREB signaling pathway (4,27). Therefore, we examined the possibility that cGMP produced by NPRB activation induced the CREB phosphorylation through an increase in cAMP/PKA activity. In WT primary chondrocytes, the treatment with CNP (10^-8 to 10^-7 M), which markedly increased the production of cGMP, did not increase the production of cAMP (Figs 5C and 10B). Moreover, co-treatment with H-89, an inhibitor of PKA, had no effect on the CREB phosphorylation induced by CNP, although it suppressed the CREB phosphorylation induced by PTHrP (Fig. 10C and D). Thus, the effects of NPRB/cGMP on CREB phosphorylation in chondrocytes appear to be independent of cAMP/PKA pathway.

Discussion

The significance of NPRB-mediated signaling as a determinant of longitudinal growth of the skeletons has been established by the identification of related disorders. Inactivating mutations in both alleles of the NPR2 gene cause acromesomelic dysplasia, Maroteaux type, which is characterized by severe dwarfism and short limbs (12), and heterozygous inactivating NPR2 mutations are responsible for short stature with non-specific skeletal abnormalities (14–16). In contrast, activating mutations in NPR2 cause ECDM, a rare, autosomal dominant skeletal overgrowth disorder. The reported patients of ECDM exhibited tall stature up to or greater than +3 SD, long fingers, mild scoliosis and enlarged great toes. To date, three mutations, p.V883M, p.R655C and p.A488P, have been found to be responsible for ECDM. All of these three mutant NPRB proteins were shown to produce increased amounts of cGMP, even in the absence of its cognate ligand CNP, when exogenously expressed in cultured cells, and addition of CNP to these NPRB mutants further enhanced the production of cGMP (17,19,20). The clinical manifestation of ECDM resembles that of the patients with overexpression of CNP due to balanced translocations of chromosome 2q37.1 (33,34), confirming that NPRB mediates the effects of CNP on the growth plate. These lines of evidence indicate that NPRB activation accelerates the skeletal growth, although the underlying mechanism has remained obscure.

In the current study, to clarify the cellular and molecular mechanism for the skeletal overgrowth in ECDM, we analyzed the growth plate phenotype of NPRB[p.V883M]-Tg mice in detail. In this mouse model, human NPRB[p.V883M] is specifically expressed in cartilage under the control of the promoter and intronic enhancer of the Col11a2 gene (17). This promoter/enhancer sequence of the Col11a2 gene allows limitation of the expression of the transgene in chondrocytes (35,36). Moreover, the Col11a2 promoter possesses insulator activities, which contribute to the stable expression of transgenes in mice (37). We performed immunostaining using an antibody that detects both endogenous murine and transgene-derived human NPRB, and observed strong signals almost exclusively in hypertrophic chondrocytes in the growth plates of both WT and NPRB[p.V883M]-Tg mice. The spatial pattern and the intensity of the staining were quite similar between the genotypes (Fig. 1F). This finding confirmed that NPRB[p.V883M] was expressed in the same region of the growth plate as the endogenous murine WT NPRB protein, and also strongly suggested that the hypertrophic chondrocyte is the main cellular target for NPRB-mediated signaling in the growth plate.

It should be noted that NPRB protein was found to be localized at the plasma membrane of hypertrophic chondrocytes in the growth plates of 4-week-old mice (Fig. 1F), because previous in situ hybridization studies using 7- to 8-day-old mice detected the Npr2 mRNA in the proliferating and prehypertrophic zones (9,13). This discrepancy between the distribution of the Npr2 mRNA and that of the NPRB protein in the growth plate suggest that it may take time for NPRB to be translated from the mRNA and to be targeted to the plasma membrane. Alternatively, there may be some unknown mechanism by which the localization of NPRB at the plasma membrane is restricted to the hypertrophic chondrocytes. Interestingly, in a previous study by Nakao et al., immunohistochemical analysis using 2-week-old mice detected the signals for NPRB protein predominantly in prehypertrophic chondrocytes. The difference between their data and our data might be due to the difference in the ages of mice and/or that in the protocol for immunostaining (11).

Consistently with our previous study (17), the NPRB[p.V883M]-Tg mice exhibited skeletal overgrowth and an increased production of cGMP in the cartilage (Fig. 1A and B). Detailed histological analysis of growth plates in the current study revealed that the overgrowth of the long bones in NPRB[p.V883M]-Tg mice was attributed to thickening of the hypertrophic zone (Fig. 1C, E and F), which supports the notion that hypertrophic chondrocytes are the receptive cells for the signal evoked by NPRB activation. Accordingly, we attempted to identify the molecules that might mediate the effects of NPRB activation specifically in matured chondrocytes by an in vitro screening using chondrogenic ATDC5 cells, and found that the CNP-induced activation of NPRB caused the phosphorylation of CREB and up-regulation of Ccnd1 (Figs 3 and 4). The NPRB activation by CNP also increased the phosphorylation of CREB and the expression of Ccnd1 in the murine primary chondrocytes (Figs 5 and 6), although they contained the cells at various differentiation stages.

CREB is a member of the CREB/activating transcription factor (ATF) family of transcription factors playing critical roles in cellular responses to various external stimuli, and leading to cell proliferation, differentiation, apoptosis and survival (38). CREB is activated by posttranslational modifications, the main one of which is the phosphorylation at Ser133. As indicated by its name, CREB can be activated by accumulation of cAMP through PKA-mediated phosphorylation of Ser133 (39). PTHrP, which is known to be one of the main inducers of cAMP/PKA-mediated CREB activation in the growth plate cartilage, is produced by the perichondrium and early proliferating chondrocytes at the ends of long bones, and acts on PTH1R expressed in proliferating and prehypertrophic chondrocytes to keep them in the proliferative stage and delay further differentiation (27). Binding of PTHrP to PTH1R induces CREB phosphorylation, leading to the transactivation of Ccnd1 and thus cell cycle progression in proliferating and prehypertrophic chondrocytes (27). In our immunohistochemical analyses, most of the signals for BrdU, Ki67, phosphorylated CREB and cyclin D1 were observed in the proliferating zone of both WT and NPRB[p.V883M]-Tg mice (Figs 2, 7 and 8), which was likely to be attributable to PTHrP-induced signaling. In contrast, the phosphorylation of CREB and the expression of cyclin D1 in the hypertrophic chondrocytes of NPRB[p.V883M]-Tg mice appear to be consequences of NPRB activation. The observed overlapping of the signals for phosphorylated CREB with those for Ki67 and cyclin D1 in the hypertrophic chondrocytes of NPRB[p.V883M] strongly support the notion that CREB activation mediates the proliferation of these cells (Figs 7 and 8). The result of MTS assay using primary chondrocytes also suggests the involvement of CREB and cyclin D1 in the acceleration of chondrocytes proliferation induced by NPRB activation (Fig. 9).

Cell cycle regulatory molecules play critical roles in the control of chondrocyte proliferation and differentiation (5). Progression of the eukaryotic cell cycle is regulated by a family of cyclin-dependent kinases (CDKs), which heterodimerize with distinct cyclins. Cyclin D1 belongs to the D-type cyclins and controls the transition of G1 to the S phase of the cell cycle in complexes with CDKs 4 and 6 (26). The abundance of cyclin D1 has been shown to be increased by various growth factors, including insulin growth factor (IGF)-I and IGF-II (26). In the growth plate cartilage, the expression of cyclin D1 has been shown to be restricted in proliferating chondrocytes at both mRNA (40) and protein (41) levels. Cyclin D1-deficient mice exhibit growth retardation (42,43), suggesting its profound role in the growth plate cartilage. It has been reported that TGF-β as well as PTHrP activates cyclin D1 expression and facilitates proliferation in chondrocytes (27), and that the transcription factors ATF-2 (for both TGF-β and PTHrP) and CREB (for PTHrP) mediate these effects (27,44). Our results shown here suggest that an enhanced activation of NPRB by a gain-of-function mutation causing ECDM increases the expression of cyclin D1 in hypertrophic chondrocytes via the activation of CREB, although we cannot exclude the possibility of the involvement of additional transcription factors such as ATF-2. Since hypertrophic chondrocytes have a larger volume than authentic columnar proliferating chondrocytes, the proliferation of hypertrophic chondrocytes would result in substantial thickening of the growth plate. In both general and chondrocyte-specific Npr2-knockout mice, which show short limbs and dwarfism, the thinning of the hypertrophic zone of the growth plate was more prominent than that of the proliferating zone (11,13), suggesting that an NPRB-mediated signal plays a physiological role in the appropriate thickening of the hypertrophic zone.

Because of its potent effects on the skeletal growth, the CNP/NPRB pathway is now considered to be a promising therapeutic target for severe dwarfism, such as achondroplasia, the most common skeletal dysplasia caused by an activating mutation in fibroblast growth factor receptor 3 (FGFR3). Yasoda et al. reported that overexpression of CNP in chondrocytes prevented the shortening of the long bones in a mouse model for achondroplasia harboring an activating mutation of FGFR3 (10). Those authors also demonstrated that CNP inhibited the ERK1/2 phosphorylation induced by FGF2, and suggested that CNP rescued the dwarfism in murine achondroplasia by inhibiting the MAPK pathway of the FGFR3 signaling, although they did not examine the effects of the treatment with CNP alone in the absence of FGF2 (10). Their findings imply that the inhibitory impact on the MAPK pathway is likely to be one of the mechanisms by which CNP/NPRB signaling facilitates the longitudinal growth of the skeleton, especially in the conditions with activated FGFR3 signaling, as exemplified by achondroplasia. However, the CREB activation in hypertrophic chondrocytes induced by the NPRB-mediated signal that we have identified in this study appears to be independent of the FGFR-MAPK pathway, considering that FGFR3 is expressed in the proliferating and prehypertrophic chondrocytes but not in hypertrophic chondrocytes of the postnatal growth plate cartilage in mice (45). Thus, CNP/NPRB signaling might regulate the skeletal growth via multiple pathways.

In conclusion, by analyzing a mouse model expressing an activating mutation of NPRB in chondrocytes and cell models, we have revealed that activation of NPRB exerts its signal in hypertrophic chondrocytes to induce CREB phosphorylation and make them to proliferate, which leads to thickening of the hypertrophic zone of growth plates and skeletal overgrowth in ECDM. These findings might shed light on a novel mechanism by which the CNP/NPRB signaling pathway facilitates the longitudinal growth of long bones.

Materials and Methods

Animals

The animal experiments were carried out in accordance with the `Guidelines for Proper Conduct of Animal Experiments’ formulated by the Scientific Council of Japan. The protocols were approved by the Institutional Animal Care and Use Committee of the Research Institute, Osaka Women’s and Children’s Hospital (Permit number: BMR27-2). All mice used in the study were of the C57BL/6J strain.

The generation of transgenic mice expressing NPRB[p.V883M] (NPRB[p.V883M]-Tg) was previously described (17). In these mice, human NPRB[p.V883M] is specifically expressed in chondrocytes, using the p742Int vector (a kind gift from Dr Noriyuki Tsumaki, Kyoto University) containing the promoter and intronic enhancer of the Col11a2 gene (35,36). Genotyping was performed by PCR for amplification of the transgene-specific sequence, using genomic DNA extracted from tails and the following primer set: forward, 5ʹ-CTAGGCCTGTACGGAAGTGTTAC-3ʹ; reverse, 5ʹ-GTAATCTGGAACATCGTATGGGTA-3ʹ. Mice were maintained in a pathogen-free barrier facility with a 12 hour light, 12 h dark cycle and were fed standard mouse chow ad libitum.

Reagents

Recombinant CNP [32–53] (H-1296) was purchased from Bachem (Torrance, CA, USA). 8-Bromo-cGMP and 3-isobutyl-1-methylxantine (IBMX) were obtained from Tocris (Bristol, UK) and Wako Pure Chemical Industries (Osaka, Japan), respectively. H-89, which is an inhibitor of PKA, was purchased from Sigma-Aldrich (St. Louis, MO, USA). The CREB inhibitor 666-15 was obtained from Merck (Damstadt, Germany), and the cyclin D1 inhibitor imperatorin was purchased from Abcam (Cambridge, UK).

Growth curve of mice

Body lengths (naso-tail lengths) of the NPRB[p.V883M]-Tg mice and their WT littermates were measured under anesthesia every week from 3 to 8 weeks of age.

Histological analyses

Tibiae harvested from 4-week-old mice were fixed in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS) for 24 h and decalcified in 20% EDTA solution for a week (pH 7.4). The specimens were embedded in paraffin and cut into sagittal sections for analyses of growth plates. HE staining was performed according to the standard procedure. Maturation stages of chondrocytes were determined based on the cell morphology. Thicknesses of the proliferating, prehypertrophic and hypertrophic zones were measured using Leica Application Suite Software version 4.0 (Leica Microsystems Ltd., Heerbrugg, Switzerland).

Immunohistochemical analyses were performed using the following primary antibodies: anti-NPRB antibody (H-80; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-Col II antibody (Clone 2B1.5; Thermo Fisher Scientific Anatomical Pathology, Fremont, CA, USA), anti-Col X antibody (LB-0092; LSL Co., Tokyo, Japan), anti-Ki67 antibody (Abcam, Cambridge, UK), anti-phosphorylated CREB (Ser133) antibody (87G3; Cell Signaling Technology, Beverly, MA, USA) and anti-cyclin D1 antibody (Clone SP4; Thermo Fisher Scientific). After deparaffinization and rehydration, antigen retrieval was performed by incubating the sections in 10 mm citrate buffer (pH 6.0) at 98°C for 10 min for the anti-Ki67 and for 1 h for the anti-phosphorylated CREB. The sections were incubated in Target Retrieval Solution pH9.0 (Dako) at 98°C for 15 min for staining with anti-cyclin D1 antibody. For the antibodies against NPRB, Col II and Col X, antigen was retrieved by treatment with proteinase K [0.04 mg/ml in Tris-buffered saline (TBS)] at 37°C for 10 min. Endogenous peroxidase activity was quenched by a 60-minute incubation in Peroxidase Block solution of the ImmunoCruz Staining System (Santa Cruz Biotechnology). Non-specific binding was blocked by incubation in 5% goat serum in TBS at room temperature for 2 h, and then the sections were incubated with the primary antibodies at 4°C overnight. Normal immunoglobulin G (IgG) served as a negative control. For signal detection, the corresponding horseradish peroxidase (HRP)-conjugated secondary antibodies (Envision+ System-HRP Labelled Polymer; Dako) and 3, 3'-diaminobenzidine (Liquid DAB+ Substrate Chromogen System; Dako) were utilized. Sections were counterstained with hematoxylin (Wako Pure Chemical Industries).

TUNEL assay was carried out using the in situ Apoptosis Detection Kit (Takara Bio Inc., Otsu, Japan).

In vivo BrdU labeling assay

BrdU (100 μg/g body weight; Wako Pure Chemical Industries) and FdU (2'-deoxy-5-fluorouridine, 12 μg/g body weight; Wako Pure Chemical Industries) were intraperitoneally injected to 4-week-old mice as described previously (46). Two hours later, the tibiae were harvested for analysis. The specimens were fixed in 4% PFA, decalcified in 20% EDTA and then embedded in paraffin. Sagittal sections of growth plates were cut from these specimens and were subjected to immunostaining using anti-BrdU antibody (1:100; Merck). The sections were counterstained with hematoxylin. Serial sections were stained with HE to identify proliferating, prehypertrophic and hypertrophic zones based on the cell morphology. The BrdU labeling index was calculated by dividing the numbers of BrdU-positive nuclei by those of total cell nuclei.

Isolation of primary chondrocytes

Primary chondrocytes were isolated from 6- or 7-day-old C57BL/6J mice according to a previous report (47) with slight modifications. In brief, the cartilage tissue dissected from ribcages was minced, rinsed in PBS twice and then treated with 3 mg/ml collagenase (Wako Pure Chemical Industries) in Dulbecco’s modified Eagle’s Medium (DMEM, Sigma-Aldrich) for 45 min at 37°C in an incubator under 5% CO₂ in a Petri dish. After the incubation, the tissue fragments were agitated using a 10 ml pipette to detach soft tissues. The cartilage pieces were transferred to a new Petri dish containing 0.5 mg/ml collagenase in DMEM and were digested for 18 h at 37°C under 5% CO₂. After that, the digestion solution with residual cartilage was successively pipetted through 25-, 10-, 5-ml and Pasteur pipettes to disperse cell aggregates. The released cells were collected into a 50 ml tube through a 100 μm cell strainer. After being washed with PBS, the collected cells were inoculated into 6-well culture plates (2.5 × 10⁵ cells/well) or 96-well culture plates (1 × 10⁵ cells/well) and cultured in DMEM supplemented with 10% fetal bovine serum (FBS, Biosera, Kansas City, MO, USA), 100 units/ml penicillin, 100 μg/ml streptomycin, 4 mm L-glutamine (Thermo Fisher Scientific) and 50 μg/ml ascorbic acid (Sigma-Aldrich). The medium was replaced every 2 days until use for experiments.

Maintenance and chondrogenic induction of ATDC5 cells

ATDC5 cells were maintained in a 1:1 mixture of DMEM and Ham's F12 (DMEM/F12) medium (Sigma-Aldrich) supplemented with 5% FBS and 1% insulin-transferrin-selenium-G supplement (Thermo Fisher Scientific) at 37 °C in a 5% CO₂ atmosphere. For chondrogenic induction, cells were seeded into 6-well culture plates (5 × 10⁴ cells/well). Two days later, the medium was changed to alpha minimal essential medium supplemented with 5% FBS and ITS and the culture plates were sealed with adhesive tape to facilitate mineralization, as previously described (22,23). The medium was replaced every 3 days.

Assay for cGMP content in mouse cartilage

Costal cartilage was harvested from 9-day-old NPRB[p.V883M]-Tg mice and their WT littermates, snap-frozen in liquid nitrogen and stored at −80°C until use. The samples were added with 5% TCA and homogenized using a Biomasher (Nippi, Tokyo, Japan). After centrifugation at 1500 × g for 10 min at 4°C, the supernatants were collected to serve for cGMP measurement using an EIA kit (Cayman Chemicals, Ann Arbor, MI, USA). The cGMP concentration was standardized based on the DNA concentration in each sample.

Assay for cGMP and cAMP production in chondrocytic cells

ATDC5 cells were seeded into 6-well culture plates (5 × 10⁴ cells/well) and cultured in chondrogenic differentiation medium as described above. At 4 and 8 weeks of culture, the cells were serum-starved for 24 h. The cells were then incubated in serum-free DMEM containing 0.5 mm IBMX for 10 min and subsequently treated with CNP (10^-6 M) or vehicle (0.1% bovine serum albumin in PBS) for another 10 min before the cell lysates were harvested in 0.1 M HCl for cGMP measurement.

Primary chondrocytes isolated from rib cages of WT mouse neonates were cultured for 14 days as described above. After incubation in serum-free DMEM containing 0.5 mm IBMX for 10 min, the cells were treated with CNP (10^-8 – 10^-7 M) or vehicle for 10 min and then harvested in 0.1 M HCl for the measurement of cGMP and cAMP. PTHrP (hypercalcemia of malignancy factor fragment 1-34 amide human, Sigma-Aldrich) was used as a positive control to induce cAMP production.

The concentrations of cGMP and cAMP in the cell lysates were assayed using EIA kits (Cayman Chemicals) according to the manufacturer’s protocols and standardized based on the protein content of the total cell lysates extracted from identical wells.

RNA extraction and quantitative real-time PCR analyses

Total RNA was extracted using TRIzol Reagent (Thermo Fisher Scientific), treated with DNase (Qiagen, Tokyo, Japan), and then reverse-transcribed using random hexamers (Promega, Madison, WI, USA) and SuperScript II (Thermo Fisher Scientific). Real-time PCR was performed using TaqMan® Gene Expression Assays with the StepOne Plus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). To generate standard curves for real-time PCR, we cloned the amplicons of interest into a pT7-Blue vector (Novagen, Madison, WI, USA) and included serial 10-fold dilutions of the constructed plasmids in the assay. The copy numbers of the target cDNA were estimated by referring to the standard curves and standardized based on those of Gapdh in each sample.

Western blotting

Whole cell extracts were prepared in radioimmune precipitation (RIPA) buffer [1% Triton, 1% Na deoxycholate, 0.1% SDS, 150 mm NaCl, 10 mm Tris-Cl (pH 7.4), 5 mm EDTA, 1 mm orthovanadate, 1 mm NaF, and protease inhibitor cocktail (Complete^TM; Roche Diagnostics, Mannheim, Germany)]. Equal amounts of protein were subjected to SDS-PAGE and transferred to PVDF membranes (BioRad Laboratories, Hercules, CA, USA). After blocking with Blocking-one reagent (Nacalai Tesque, Kyoto, Japan) or BlockAce reagent (Dainippon Pharmaceuticals, Osaka, Japan), the membranes were immunoblotted with the following primary antibodies: anti-phospho CREB antibody, anti-CREB antibody, anti-phospho ERK1/2 antibody, anti-ERK1/2 antibody, anti-phospho p38 MAPK antibody, anti-p38 MAPK antibody, anti-phospho Akt antibody, anti-Akt antibody, anti-phospho Smads antibodies, anti-Smads antibodies, anti-cyclin D1 antibody (Cell Signaling, Beverly, MA, USA) and anti-Gapdh antibody (Santa Cruz Biotechnology). After incubation with an HRP-conjugated secondary antibody (Santa Cruz Biotechnology), the proteins were visualized using the enhanced chemiluminescence detection system (GE Healthcare, Buckinghamshire, UK). In some experiments, densitometry of the signals was performed using ImageJ software.

Assay for cell proliferation

Primary chondrocytes isolated from WT mouse neonates were plated into 96-well culture plates (1 × 10⁵ cells/well). On the next day, the CREB inhibitor 666-15 (10^-6 M), the cyclin D1 inhibitor imperatorin (10^-4 M) or the solvent (dimethyl sulfoxide; DMSO) was added to the cells, and 1 h later CNP (10^-7 M) or the corresponding vehicle was also added. The media containing the reagents were replaced 24 h later. Forty-eight h after the initiation of the treatment, we evaluated the cell number in each well by an MTS assay using a CellTiter96® aqueous one solution cell proliferation assay kit (Promega).

Statistical analysis

We used the Student’s t-test to compare data between two groups. To compare the data among more than three groups, we used one-way analysis of variance and the methods of Tukey–Kramer and Dunnett for post hoc tests.

Acknowledgements

We thank Dr Noriyuki Tsumaki for providing p742Int vector containing Col11a2 promoter/enhancer. We also thank Dr Wei Wang for technical assistance and Drs Noriyuki Namba and Kohji Miura for helpful discussions.

Conflict of Interest statement. None declared.

Funding

The Japan Agency for Medical Research and Development (17eko0109135h003 to K.O. and T.M.); Scientific Research from the Ministry of Education, Science and Culture, Japan [18H02780 to K.O. and T.M.]; the Foundation for Growth Science [2014-19 to T.M.].