NO/cGMP signaling is important for bone remodeling in response to mechanical and hormonal stimuli, but the downstream mediator(s) regulating skeletal homeostasis are incompletely defined. We generated transgenic mice expressing a partly-activated, mutant cGMP-dependent protein kinase type 2 (PKG2R242Q) under control of the osteoblast-specific Col1a1 promoter to characterize the role of PKG2 in post-natal bone formation. Primary osteoblasts from these mice showed a two- to three-fold increase in basal and total PKG2 activity; they proliferated faster and were resistant to apoptosis compared to cells from WT mice. Male Col1a1-Prkg2 R242Q transgenic mice had increased osteoblast numbers, bone formation rates and Wnt/β-catenin-related gene expression in bone and a higher trabecular bone mass compared to their WT littermates. Streptozotocin-induced type 1 diabetes suppressed bone formation and caused rapid bone loss in WT mice, but male transgenic mice were protected from these effects. Surprisingly, we found no significant difference in bone micro-architecture or Wnt/β-catenin-related gene expression between female WT and transgenic mice; female mice of both genotypes showed higher systemic and osteoblastic NO/cGMP generation compared to their male counterparts, and a higher level of endogenous PKG2 activity may be responsible for masking effects of the PKG2R242Q transgene in females. Our data support sexual dimorphism in Wnt/β-catenin signaling and PKG2 regulation of this crucial pathway in bone homeostasis. This work establishes PKG2 as a key regulator of osteoblast proliferation and post-natal bone formation.
Cyclic GMP is produced by two types of guanylyl cyclases (GCs): receptor GCs activated by peptide ligands such as C-type natriuretic peptide (CNP) and soluble GCs activated by nitric oxide (NO) (Hofmann et al. 2009, Hannema et al. 2013, Kuhn 2016). cGMP regulates ion channels, phosphodiesterases and two types of cGMP-dependent protein kinases (PKG1 and 2), encoded by two homologous genes (Prkg1 and 2) (Hofmann et al. 2009). PKG1 is widely expressed with high levels found in the vasculature, whereas PKG2 expression is limited to specific cell types in bone, intestine, brain and kidney (Vaandrager et al. 2005). PKG1 and 2 are both found in osteoblasts and chondroblasts (Pfeifer et al. 1996, Rangaswami et al. 2009).
The CNP/GC-B/PKG2 pathway regulates endochondral ossification, responsible for fetal skeletal development and post-natal longitudinal bone growth. Loss-of-function mutations or genetic ablation of CNP, GC-B or PKG2 cause dwarfism in humans and rodents, while increased CNP expression or a gain-of-function mutation in GC-B cause long bone overgrowth (Pfeifer et al. 1996, Chusho et al. 2001, Bartels et al. 2004, Bocciardi et al. 2007, Bonnet et al. 2010, Miura et al. 2012, Hannema et al. 2013).
The NO/soluble GC pathway regulates adult skeletal homeostasis: NO is required for the anabolic response of bone to mechanical stimulation and regulates bone remodeling downstream of estrogens and thyroid hormone (Watanuki et al. 2002, Rangaswami et al. 2010, Kalyanaraman et al. 2014, 2017, 2018a). Some strains of NOS3-deficient mice have reduced bone mass due to defects in osteoblast number and maturation, whereas other strains only show exaggerated bone loss after ovariectomy (OVX) and a blunted response to estrogens (Aguirre et al. 2001, Armour et al. 2001, Grassi et al. 2006). Treatment with NO-generating agents and NO-independent soluble GC activators prevents bone loss from estrogen deficiency and type 1 diabetes, respectively (Wimalawansa et al. 1996, Wimalawansa 2000, Jamal et al. 2004, Nabhan & Rabie 2008, Kalyanaraman et al. 2017, 2018b).
We have shown in vitro that PKG1 and 2 are both targets of NO and cGMP in osteoblasts and osteocytes, with largely distinct functions: PKG2 mediates pro-proliferative effects of insulin and mechanical stimulation, while PKG1 and 2 both mediate pro-survival effects of estrogens via distinct mechanisms (Rangaswami et al. 2009, 2010, Marathe et al. 2012, Kalyanaraman et al. 2018b). PKG2 regulates nuclear translocation of β-catenin, a transcriptional co-activator downstream of the Wnt co-receptor low-density lipoprotein receptor-related protein-5 (LRP5), which plays a central role in skeletal homeostasis (Rangaswami et al. 2012, Baron & Kneissel 2013, Kalyanaraman et al. 2017). NO and cGMP also regulate osteoclast motility and acid secretion via PKG1; however, some NO effects in bone are cGMP independent (van’t Hof & Ralston 2001, Wimalawansa 2007, Kalyanaraman et al. 2018a).
To examine the role of PKG2 in post-natal bone acquisition, independently of its function in endochondral bone formation, we established transgenic mice expressing a novel PKG2 mutant with increased basal kinase activity under control of the 2.3 kb Col1a1 promoter; the promoter is specific for cells of osteoblast lineage and has no detectable activity in osteoclasts (Dacquin et al. 2002, Kalajzic et al. 2002). The transgenic mice exhibited gender-specific increases in trabecular bone volume and Wnt/β-catenin-related gene expression in bone. We detected higher NO and cGMP concentrations in serum and osteoblasts from female compared to male mice, which at least partly explain the sexually dimorphic skeletal phenotype of Col1a1-Prkg2RQ transgenic mice. To avoid the influence of estrogen on NO/cGMP/PKG signaling, we then focused on male mice and found that male transgenic mice are protected from diabetes-induced bone loss.
Materials and methods
Antibodies against Akt, Akt(pSer473), Erk1(pTyr204), GSK-3β(pSer9), vasodilator-stimulated phosphoprotein (VASP) (pSer239), NOS-3(pSer1177), caveolin-1 and cleaved caspase-3 were from Cell Signaling Technology. Antibodies against PKG2 and β-catenin and FITC-labeled secondary antibodies were from Invitrogen. A second monoclonal mouse antibody against PKG2, and a β-actin antibody were from Santa Cruz Biotechnology. 8-(4-chlorophenylthio)-cGMP (8-pCPT-cGMP) was from BioLog. The IP3-receptor peptide GRRESLTSFG was synthesized by Eton Bioscience.
Generation of Col1a1-Prkg2 RQ transgenic mice
A 2.3 kb segment of the mouse type Ia collagen (Col1a1) promoter, including the transcription start site and a short 5′-untranslated sequence (Kalajzic et al. 2002), was amplified by PCR using the following primers: 5′-attCCTAGGCTGCCTCTGCTTCTGTTCCA-3′ and 5′-aatgcggccgcATGCCACGTGTAAAGG-3′ (including a NotI fusion site). A rat Prkg2 cDNA encoding PKG2 with a R242Q mutation (hereafter referred to as Prkg2 RQ or PKG2RQ) was generated by site-directed mutagenesis (Zhao et al. 2005). We did not add an N-terminal epitope tag to avoid problems with enzyme mis-localization (Vaandrager et al. 1996), and we found that C-terminal tags reduce kinase activity. The Col1a1-Prkg2 RQ transgenic construct was generated by fusing the 2.3 kb Col1a1 promoter with the entire coding sequence of the rat Prkg2 RQ cDNA, followed by a SV40 polyadenylation sequence. The construct was sequenced, and purified DNA micro-injected into pronuclei of fertilized eggs from C57BL/6 mice. Presence of the transgene in offspring was detected by PCR analysis of tail DNA using the primers 5′CGAGCCGAAAGAGTCTACA3′ (F1) and 5′GTCTTGAAGTCCTCGCGCTCATGG3′ (R1). Mice hemizygous for the Col1a1-Prkg2RQ transgene were bred with WT C57BL/6 mice from Harlan (Cumberland, IN, USA).
All mouse experiments were approved by the Institutional Animal Care and Use Committee of the University of California, San Diego. Mice were housed in a temperature-controlled environment with a 12-h light/darkness cycle and ad libitum access to water and food (Teklad Rodent Diet #8604). To induce insulin deficiency (type 1 diabetes), 6-week-old WT (n = 18) and transgenic (n = 16) male litter mates were randomly assigned to receive two intraperitoneal injections of streptozotocin (STZ, 100 mg/kg/d) or vehicle on two consecutive days (after 4 h of fasting). Twelve days later, glucose was measured in tail vein blood and mice with blood glucose concentrations ≥15 mM were considered diabetic; three STZ-injected mice with glucose <15 mM were excluded. Double calcein labeling was performed by intraperitoneal injection of calcein (25 mg/kg) at 7 and 2 days before killing by CO2 intoxication and exsanguination.
Primary osteoblast cultures and proliferation assays
Primary osteoblasts (POBs) were isolated from femurs and tibiae of 8-week-old mice and were cultured as described (Kalyanaraman et al. 2014). To induce differentiation, cells were plated at high density, and after confluency were switched to medium with 0.3 mM sodium ascorbate and 10 mM β-glycerophosphate. Each POB preparation was characterized by alkaline phosphatase (ALP) staining and mineralization capacity as described (Kalyanaraman et al. 2014); cells were used at passages 1–3 and counted using a T20 automated cell counter (BioRad). Metabolically active osteoblasts were quantified using a CellTiter 96 AQueous Cell Proliferation Assay Kit (Promega): after 16 h in medium with 0.1% FBS, cells were stimulated with 10% FBS for 24 h and tetrazolium (MTS) reduction to formazan was measured spectrophotometrically during the last 4 h.
Bone marrow stromal cell (BMSC) cultures
Bone marrow mononuclear cells were plated at 4 × 105 cells/cm2 in RPMI 1640 with 10% FBS and 10% horse serum (Joshua et al. 2014). After 7 days, adherent cells were switched to α-MEM supplemented with 10% FBS/10% horse serum, 0.3 mM sodium ascorbate and 10 mM β-glycerolphosphate to induce differentiation. ALP and Alizarin Red staining was performed 14 and 21 days later, respectively; colonies were counted and stained area was quantified by ImageJ (Joshua et al. 2014, Kalyanaraman et al. 2014).
PKG2 purification and activity assays
WT PKG2 and mutant PKG2RQ were expressed with N-terminal Flag epitope-tags in transiently-transfected 293T cells, purified on anti-Flag beads and eluted with Flag peptide (Kalyanaraman et al. 2014). Kinase activity was measured using the synthetic peptide GRRESLTSFG and [γ-32P]ATP in the absence and presence of cGMP (Zhao et al. 2005). POB plasma membranes were isolated by density-gradient centrifugation (Kalyanaraman et al. 2014).
After removing growth plates and bone marrow, bone shafts were snap-frozen and pulverized. RNA was purified with RNeasy Mini-Kit columns (Qiagen). Quantitative RT-PCR was performed with 300 ng of RNA as described (Kalyanaraman et al. 2017). All primers (Supplementary Table 1, see section on supplementary data given at the end of this article) were tested with serial cDNA dilutions. Genes of interest were normalized to three different reference genes: 18S rRNA, Hprt and B2m. The mean ∆CT obtained for control WT mice was used to calculate the fold change in mRNA expression in the transgenic group using the ∆∆CT method; for each mouse, the fold change in mRNA expression obtained using the different reference genes was averaged.
Western blotting, immunofluorescence and immunohistochemical staining
Western blots were generated using horseradish peroxidase-conjugated secondary antibodies detected by enhanced chemiluminescence (Rangaswami et al. 2009). POBs plated on glass coverslips were fixed in 4% paraformaldehyde, permeabilized with 1% Triton-X-100 and incubated with PKG2 antibody (1:100) or cleaved caspase-3-specific antibody (1:100 dilution), followed by a secondary FITC-conjugated antibody. Nuclei were counterstained with Hoechst 33342, and images were analyzed with a Keyence BZ-X700 fluorescence microscope.
Tibiae were fixed overnight in 10% formalin, decalcified in 0.5 M EDTA (pH 7.5) for 5 days and embedded in paraffin. Eight micrometer sections were de-paraffinized in xylene and rehydrated in graded ethanol and water. For antigen retrieval, slides were placed in 10 mM sodium citrate buffer (pH 6.0) at 80–85°C and allowed to cool to room temperature for 30 min. Endogenous peroxidase activity was quenched in 3% H2O2 for 10 min. After blocking with 5% normal goat serum, slides were incubated overnight at 4°C with anti-PKG2 antibody (1:100), followed by a horseradish peroxidase-conjugated secondary antibody. After development with 3-diaminobenzidine (Vector Laboratories), slides were counterstained with hematoxylin.
Quantitation of NOx and cGMP
NO production was measured as the sum of nitrite and nitrate accumulation in the medium or serum, using a two-step colorimetric assay as described (Kalyanaraman et al. 2018b). cGMP concentrations were measured using an ELISA kit according to the manufacturer’s protocol (Biomedical Technologies).
Ethanol-fixed tibiae were analyzed according to established guidelines (Bouxsein et al. 2010), using a Skyscan 1076 (Skyscan, Belgium) scanner at 9 µm voxel size, and applying an electrical potential of 50 kVp and current of 200 mA, with a 0.5 mm aluminum filter as described (Kalyanaraman et al. 2017). Mineral density was determined by calibration of images against 2 mm diameter hydroxyapatite rods (0.25 and 0.75 g/cm3). Cortical bone was analyzed by automatic contouring 3.6–4.5 mm distal to the proximal growth plate, using a global threshold to identify cortical bone, and eroding one pixel to eliminate partial volume effects. Trabecular bone was analyzed by automatic contouring of the proximal tibial metaphysis (0.36–2.1 mm distal to the growth plate), using an adaptive threshold to select the trabecular bone (Kalyanaraman et al. 2017).
Femurs were fixed in 70% ethanol, dehydrated and embedded in methyl methacrylate and sectioned at the University of Alabama Center for Metabolic Bone Disease. Masson’s Trichrome-stained sections were used for static parameters, and unstained sections were used to assess fluorochrome labeling. Slides were scanned with a Hamamatsu Nanozoomer 2.0 HT Slide Scanning System (Hamamatsu Corporation), and image analysis was performed using the Nanozoomer Digital Pathology NDP view2 software. Trabecular bone was analyzed between 0.25 and 1.25 mm, and cortical bone between 0.25 and 5 mm proximal to the growth plate by an observer blinded to the genotype of the mice (Joshua et al. 2014).
GraphPad Prism5 was used for two-tailed Student t-test (to compare two groups) or one-way ANOVA with Bonferroni post-test (to compare more than two groups); P < 0.05 was considered significant. When analyzing bone micro-architecture in diabetic mice, we tested the following primary and secondary hypotheses by two-sided t-test: (i) Transgenic animals are protected from diabetic bone loss; (ii) WT diabetic animals have lower bone mass compared to non-diabetic WT animals and (iii) transgenic animals have higher bone mass compared to WT animals.
Generation of transgenic mice with expression of a mutant PKG2R242Q from the Col1a1 promoter (Col1a1-Prkg2 RQ )
To delineate in vivo functions of PKG2 in osteoblasts, we introduced a mutation in the first cGMP binding domain of PKG2, homologous to a gain-of-function mutation in PKG1 associated with human thoracic aortic aneurysms and dissections (Guo et al. 2013). In contrast to the R177Q mutation in PKG1 – which causes nearly complete activation and renders the enzyme largely cGMP-unresponsive (Guo et al. 2013) – the PKG2 R242Q mutation increased basal activity of the purified enzyme only about four-fold, to <10% of maximal cGMP-stimulated activity (Fig. 1A and B and Supplementary Fig. 1A). Both mutations are located in the first cGMP binding pocket and likely disrupt interaction between inhibitory and catalytic domains, but the relative cGMP affinities of the first and second cGMP binding pockets are reversed between PKG1 and 2 (Vaandrager et al. 2005) – the first cGMP binding site in PKG2 is a low affinity, rapidly-dissociating site, which may explain the less dramatic effect of the mutation in PKG2.
The mutant PKG2RQ was cGMP responsive with a similar cGMP affinity as the WT enzyme (Ka = 0.4 and 0.45 µM, respectively). In addition, the mutant enzyme’s affinity for a synthetic peptide substrate was similar to that of the WT enzyme (Supplementary Fig. 1B). We chose to express mutant PKG2RQ under control of the murine 2.3 kb Col1a1 promoter, because the latter is active in cells of the osteoblastic lineage, starting from committed mesenchymal stem cells through mature osteocytes (Dacquin et al. 2002, Kalajzic et al. 2002). One of several transgenic founder lines was further characterized (Line Fe-9, Supplementary Fig. 1C).
Osteoblast-specific PKG2RQ expression in Col1a1-Prkg2 RQ transgenic mice
PCR analysis of reverse-transcribed RNA extracted from multiple organs showed amplification of a specific, 210 bp band corresponding to the transgene-derived Prkg2 RQ mRNA in bone, but not in other organs (Fig. 1D); this band was not detected in the absence of reverse transcriptase, thus excluding genomic DNA contamination. Using a primer pair recognizing both mouse (endogenous) and rat (transgenic) Prkg2 transcripts, we found that the transgenic transcript was expressed ~35-fold higher in bone from transgenic mice compared to the endogenous Prkg2 transcript in WT mice. In contrast, WT and transgenic mice showed similar Prkg2 mRNA expression in brain and kidney, two organs with high endogenous PKG2 mRNA expression (Vaandrager et al. 2005) (Fig. 1E).
Since PKG2 associates with plasma membranes, whereas PKG1 is predominantly cytoplasmic (Vaandrager et al. 1996, Hofmann et al. 2009, Rangaswami et al. 2009), we measured PKG activity in purified plasma membranes of primary osteoblasts (POBs) isolated from WT and transgenic mice. cGMP-stimulated PKG activity was only about three-fold higher in plasma membranes from transgenic compared to WT mice, correlating with an increased amount of PKG2 protein (Fig. 1F and Supplementary Fig. 1D). We could not reliably determine basal PKG2 activity in the absence of cGMP using this kinase assay, because available PKG inhibitors are not effective against the mutant PKG2, and the synthetic peptide can be phosphorylated by other membrane-associated kinases. Using a second PKG2-specific antibody, approximately three-fold more PKG2 protein could be detected in whole cell lysates of transgenic compared to WT POBs (Fig. 1G and H). Correspondingly, phosphorylation of VASP on Ser239, a site specifically targeted by PKG1 and 2, was increased in transgenic compared to WT POBs, both in the absence and presence of cGMP (representing basal and maximal PKG activity, respectively) (Fig. 1G and H). Immunofluorescence staining of PKG2 in POBs from transgenic mice was considerably brighter than that of WT POBs (Fig. 1I). Immunohistochemical staining of bone sections showed a stronger signal in bone-lining osteoblasts in the transgenic compared to WT mice (Fig. 1J; transgenic mice had more bone-lining osteoblasts than WT mice; described below and shown in Fig. 5D). However, staining of endogenous PKG2 in megakaryocytes was similar between transgenic and WT animals (Fig. 1J). Thus, Col1a1-Prkg2 RQ mice express the mutant PKG2 in osteoblasts, leading to a moderate increase in both basal and cGMP-stimulated PKG2 activity.
Increased proliferation, Erk and Akt/GSK-3/β-catenin signaling in POBs from Col1a1-Prkg2 RQ transgenic mice
NO/cGMP activation of PKG is required for osteoblasts to proliferate in response to insulin or mechanical stimulation (Rangaswami et al. 2010, Kalyanaraman et al. 2018b). To examine whether expression of PKG2 RQ in osteoblasts affects proliferation, we compared multiple POB cultures that were independently established from 8 week-old Col1a1-Prkg2 RQ transgenic mice and their WT litter mates. The population doubling time of male transgenic POBs was 1.5 ± 0.1 day compared to 3.2 ± 0.9 day for WT cells (Fig. 2A). We also quantified metabolically-active cells after 24 h of serum stimulation by MTS reduction to formazan. Male transgenic POBs reduced 45% more MTS than WT POBs (Fig. 2B). For comparison, POBs isolated from female transgenic mice reduced 18% more MTS than POBs from female WT litter mates (Supplementary Fig. 2A).
PKG activation by NO/cGMP in osteoblasts leads to activation of ERK and Akt via Src: PKG2 directly phosphorylates and stimulates the phosphatase SHP-1, which de-phosphorylates Src on an inhibitory site (Rangaswami et al. 2010, 2012). In contrast, PKG2 inactivates GSK-3β both directly, via phosphorylation of a negative-regulatory site, and indirectly, via Akt activation (Rangaswami et al. 2012, Kalyanaraman et al. 2017). To assess effects of transgenic PKG2RQ expression on ERK, Akt, and GSK-3β phosphorylation, we compared three independent POB lines of male and female WT and transgenic mice (12 lines total). POBs from male transgenic mice showed significantly higher basal ERK and Akt phosphorylation, and higher cGMP-stimulated ERK, Akt, and GSK-3β phosphorylation compared to POBs from male WT littermates (Fig. 2C, D, E and F). In POBs from female transgenic mice, cGMP-induced ERK phosphorylation was also increased, but Akt and GSK-3β showed only a trend toward increased phosphorylation compared to female WT mice, and basal phosphorylation was the same between both genotypes in the females (Supplementary Fig. 2B, C, D and E). As described above, PKG2 expression was similar in male and female transgenic POBs (Fig. 1G and Supplementary Fig. 2B).
PKG2 directly and indirectly (via Akt activation) inhibits GSK-3β, thereby regulating β-catenin stability and nuclear translocation (Zhao et al. 2005, Rangaswami et al. 2012, Kalyanaraman et al. 2017). In POBs from male Col1a1-Prkg2 RQ transgenic mice, more nuclei stained positive for β-catenin, under both basal and cGMP-stimulated conditions, compared to cells from WT mice (Fig. 2G). In contrast, POBs isolated from female transgenic and WT mice showed similar nuclear β-catenin staining under both basal and cGMP-stimulated conditions (Supplementary Fig. 2F). However, the number of β-catenin-positive nuclei was almost twice as high in unstimulated female compared to male WT POBs (Fig. 2G and Supplementary Fig. 2F). Interestingly, when basal NO synthesis in female WT POBs was inhibited with N(ω)-nitro-l-arginine methyl ester (L-NAME), the percentage of β-catenin-positive nuclei was reduced to levels found in male cells (Supplementary Fig. 2G). These data confirm that nuclear localization of β-catenin is under control of NO (Kalyanaraman et al. 2017), and suggest a higher level of basal β-catenin signaling in female compared to male osteoblasts, which appears to obscure the PKG2 RQ effect on β-catenin in female cells (Supplementary Fig. 2F).
Apoptosis resistance and enhanced expression of osteoblast differentiation-associated genes in POBs and BMSCs from male Col1a1-Prkg2 RQ transgenic mice
POBs from male Col1a1-Prkg2 RQ transgenic mice showed enhanced survival after prolonged serum starvation: compared to WT cells, fewer transgenic POBs staining positive for cleaved caspase-3 (Fig. 2H). To examine osteoblast differentiation, POBs from WT and transgenic mice were plated at high density, and switched to differentiation medium after reaching confluency. Differentiated POBs from male transgenic mice expressed higher amounts of mRNA encoding β-catenin (Ctnnb1), collagen-1α (Col1a1), and osteocalcin (Bglap) compared to WT POBs, while expression of Runx2 and β-actin (Actb) was unchanged (Fig. 2I). However, POBs of both genotypes showed similar ALP activity and mineralization after 14 and 21 d, respectively (Supplementary Fig. 3A and B).
To examine the effect of PKG2RQ expression on osteoblast progenitors, we cultured BMSCs from male WT and Col1a1-Prkg2 RQ transgenic mice in osteoblastic differentiation medium. We found the same number and size of colonies staining positive for ALP activity, and the same degree of mineralization (Fig. 2J and Supplementary Fig. 3C). However, transgenic BMSCs expressed three-fold more osteocalcin (Bglap) and two-fold more Runx2 mRNA, respectively, compared to WT BMSCs, suggesting a modest increase in osteoblastic differentiation potential (Supplementary Fig. 3D). The ratio of transgenic to endogenous Prkg2 mRNA expression in BMSCs was similar to that found in bone (Supplementary Fig. 3E; compare to Fig. 1E).
Gender-specific differences in POB NO/cGMP generation and NO/cGMP serum concentrations
Previous workers showed that estradiol acutely increases endothelial NO production via membrane-bound estrogen receptor(ER)-α stimulation of NO synthase-3 (Russell et al. 2000, Mendelsohn & Karas 2010). Since POBs produce NO and cGMP constitutively (Rangaswami et al. 2009, Mendelsohn & Karas 2010, Marathe et al. 2012, Joshua et al. 2014), we hypothesized that female POBs, which contain higher ER-α copy numbers than males (Chaudhri et al. 2014), may produce more NO than male cells. Indeed, female POBs produced about 1.5-fold more nitrites and nitrates (stable NO metabolites) per hour per 106 cells compared to male POBs, and this correlated with increased phosphorylation of NOS-3 on Ser1179, a stimulatory site (Fig. 3A and B). Correspondingly, the intracellular cGMP concentration was 1.7-fold higher in female compared to male POBs (Fig. 3C).
Estrogen replacement therapy in postmenopausal women increases NOx and cGMP serum concentrations, and 24 h urinary cGMP excretion (Rosselli et al. 1995, Hayashi et al. 2000, Mueck et al. 2001). Moreover, 24 h urinary cGMP excretion in pre-menopausal women is higher than in age-matched men (Markovitz et al. 1997, Cui et al. 2009). However, to our knowledge, gender-specific differences in serum NOx and cGMP concentrations have not been examined. In eight week-old WT and Col1a1-Prkg2 RQ transgenic mice, we found that NOx and cGMP serum concentrations were 56% and 32% higher in female compared to male mice, respectively, with no significant difference between the two genotypes (Fig. 3D and E). Since PKG2 has a Ka for cGMP of about 0.4 µM (Fig. 1B) (Vaandrager et al. 2005), a change in cGMP concentration from 60 to 80 nM would be expected to increase PKG activity. These results suggest that females have higher endogenous PKG2 activities, at least in osteoblasts, but probably also in other tissues.
Increased trabecular bone mass in male Col1a1-Prkg2 RQ transgenic mice
Col1a1-Prkg2 RQ transgenic mice were born at the expected Mendelian frequency, were indistinguishable from their WT littermates at birth, and showed no obvious skeletal abnormalities on plain X-Rays at eight weeks of age. They had similar body weights and tibial lengths as WT mice, although the male transgenics showed a trend toward increased body weight (+4%, P = 0.1) (Fig. 4A and B). Computed tomography analysis of tibial bone micro-architecture revealed that eight week-old male transgenic mice had increased BMD (+20%), trabecular bone volume fraction (+47%), trabecular number (+27%), and trabecular thickness (+15%) compared to male WT littermates (Fig. 4C and E). In contrast, no apparent differences in trabecular micro-architecture were found between female transgenic and WT mice at this age (Fig. 4D and F). Importantly, Prkg2 mRNA and PKG2 protein expression were the same in bones and POBs from male and female Col1a1-Prkg2 RQ transgenic mice (data in Fig. 1F, G and H represent combined results from both sexes, but data were also analyzed separately). Cortical parameters (cortical bone area fraction, cross-sectional thickness and TMD) were the same in eight week-old transgenic and WT mice of both genders (Supplementary Fig. 4A and B).
Increased bone formation parameters in male Col1a1-Prkg2 RQ transgenic mice
Consistent with the normal length of transgenic tibiae, femoral growth plate thickness was the same in male transgenic and WT littermates (Fig. 5A and B). These results confirm lack of biologically-significant PKG2RQ transgene expression in growth plate chondroblasts, and are in keeping with absence of Col1a1 promoter activity in the chondroblastic lineage (Dacquin et al. 2002, Kalajzic et al. 2002).
To determine the effect of PKG2RQ expression on bone formation parameters, we performed histomorphometric measurements in the distal femur after double calcein labeling. Eight week-old male transgenic mice showed dramatically increased trabecular mineralizing surface (MS/BS, +91%), mineral apposition rate (MAR, +80%), and bone formation rate (BFR, +234%) compared to their male WT littermates, suggesting both increased osteoblast numbers and activity (Fig. 5C). Similar results were found on endocortical bone surfaces (Supplementary Fig. 5A). The number of osteoblasts on trabecular as well as endocortical surfaces was increased by 71% and 97%, respectively, in male transgenic animals, while there was no change in osteoclast numbers (Fig. 5D and E).
Gender-specific increases in the expression of Wnt-related genes in male Col1a1-Prkg2 RQ transgenic mice
Pharmacologic activation of the NO/cGMP/PKG signaling cascade regulates the expression of Wnt-related genes in osteoblasts in vitro (Rangaswami et al. 2012, Kalyanaraman et al. 2017). In the tibiae of male WT and Col1a1-Prkg2 RQ transgenic mice, mRNA expression of β-catenin (Ctnnb1), Wnt1, the Wnt co-receptor LRP5 (Lrp5), collagen-1α (col1a1), and alkaline phosphatase (Alpl) was increased two- to five-fold compared to male WT tibiae; Wnt3a and osteocalcin (Bglap) mRNA showed a trend toward increased expression, while osteopontin (Ssp1) and β-actin (Actb) mRNA were unchanged (Fig. 5F). Expression of the β-catenin target gene cyclin D (Ccnd1) was increased >2-fold in transgenic bone, and Fra-1 mRNA showed a trend toward increased expression (Fig. 5F), consistent with increased osteoblast proliferation. In contrast, in tibiae isolated from female mice, β-catenin mRNA was only 1.6-fold higher and the other mRNAs did not differ significantly between WT and transgenic littermates (Supplementary Fig. 5B). Expression of the osteoclast regulators receptor of activated nuclear factor-Κb ligand (RANKL; gene name Tnfsf11) and osteoprotegerin (OPG; Tnfrsf11b), and of two osteoclast-specific genes – tartrate-resistant acid phosphatase (Acp5) and cathepsin K (Ctsk) – were the same in WT and transgenic tibiae of both genders (Fig. 5G and Supplementary Fig. 5C).
Protection from diabetes-induced bone loss in male Col1a1-Prkg2 RQ transgenic mice
We recently demonstrated that the bone loss observed in male mice with type 1 diabetes is at least partly due to impaired NO/cGMP/PKG signaling in osteoblasts, and that restoring cGMP synthesis recovers depressed osteoblast functions in insulin-deficient animals (Kalyanaraman et al. 2018b). We focused on male mice to exclude the influence of estrogens on NO/cGMP signaling. To determine if increased PKG2 activity in osteoblasts is sufficient to protect Col1a1-Prkg2 RQ transgenic mice from diabetes-induced bone loss, we induced insulin deficiency by injecting 6 week-old male mice with streptozotocin (STZ). Mice were euthanized six weeks after STZ injections (Fig. 6A). Hyperglycemia was similar in STZ-treated transgenic mice and their WT litter mates (Fig. 6B).
WT mice experienced diabetes-induced bone loss in both trabecular and cortical compartments (Fig. 6C and D), and this was associated with decreased bone formation parameters and decreased expression of osteoblast differentiation- and Wnt-related genes (Fig. 6E and F). Non-diabetic (vehicle-treated) male Col1a1-Prkg2 RQ transgenic mice showed higher trabecular and cortical bone volumes compared to their WT litter mates (Fig. 6C and D). The increase in trabecular bone was similar in 8- and 12-week-old mice, but findings in cortical bone were more pronounced at 12 compared to 8 weeks of age (compare Fig. 6C and D to Fig. 4E and Supplementary Fig. 4A). In the transgenic mice, STZ-induced diabetes did not significantly affect trabecular bone and cortical bone area fraction; only cortical thickness and TMD decreased, but most values remained at least as high as those measured in control WT mice (Fig. 6C and D). Similarly, bone formation parameters were elevated in transgenic control mice and remained high after induction of diabetes (Fig. 6E). The protection from bone loss and preservation of bone formation in diabetic Col1a1-Prkg2 RQ transgenic mice may relate, at least in part, to increased expression of osteoblast differentiation- and Wnt-related genes (including the β-catenin target gene cyclin D) (Fig. 6F).
Osteoblast-specific expression of a novel PKG2 mutant at a physiologically-relevant level
PKG2 deletions in humans and rodents cause dwarfism, and global PKG2 knockout mice exhibit a ~30% reduction in tibial and femoral length with profound growth plate abnormalities at 8 weeks of age; both genders appear equally affected (Pfeifer et al. 1996, Bonnet et al. 2010, Lipska et al. 2011). Eight week-old Col1a1-Prkg2 RQ transgenic mice had normal tibial length and growth plate architecture, consistent with specificity of the 2.3 kb fragment of the murine Col1a1 promoter for cells of the osteoblastic lineage (Dacquin et al. 2002, Kalajzic et al. 2002). In the mouse embryo, transcription from the 2.3 kb Col1a1 promoter starts day 14 post conception in ossification centers, where osteoblasts differentiate and initiate bone mineralization (Dacquin et al. 2002). We cannot exclude adaptive developmental changes in the transgenic mice, but the absence of bone deformities and the relatively subtle changes in micro-architecture argue against major developmental changes.
The novel PKG2RQ mutant had modestly increased kinase activity in the absence of cGMP and normal responsiveness to physiological increases in cGMP concentrations. We confirmed tissue-specific expression of the transgene, producing about three-fold increased PKG2 protein and cGMP-stimulated kinase activity in transgenic versus WT POB membranes. Phosphorylation of the PKG substrate VASP was increased in intact transgenic POBs to a degree similar to that seen when osteoblasts are mechanically stimulated or treated with estrogen (Rangaswami et al. 2010, Marathe et al. 2012). Thus, PKG2RQ activity was in a physiologically meaningful range, yet it had profound effects on osteoblast numbers and bone formation parameters. The PKG2 gain of function resulted in cell-autonomous increases in POB proliferation and resistance to apoptosis, explaining increased osteoblast numbers and BFR in vivo.
Transgenic mice with chronically elevated plasma CNP concentrations exhibit a prominent skeletal overgrowth phenotype from excess PKG2 activity in growth plate chondroblasts (Miyazawa et al. 2002, Kondo et al. 2015). Since CNP also increases intracellular cGMP concentrations in primary osteoblasts, it is surprising that these mice do not have increased BFRs and show decreased trabecular bone volumes in the femurs of 8 week-old females (data on transgenic males were not reported) (Kondo et al. 2015). Whether CNP stimulates or inhibits osteoblast proliferation is controversial (Hagiwara et al. 1996, Lenz et al. 2010). The reason for these inconsistencies is unclear, but cGMP pools generated by soluble versus receptor GCs may be compartmentalized and differentially activate PKG1 and 2 in different cell types.
Gender-specific differences in the skeletal phenotype of Col1a1-Prkg2 RQ transgenic mice
Male mice expressing the Prkg2 RQ transgene had increased trabecular bone volumes and BMD at two and three months of age, whereas this phenotype was not evident in two month-old female mice. Despite similar PKG2RQ expression in male and female POBs and bones, we found gender-specific differences in nuclear β-catenin and Wnt/β-catenin-related gene expression. The sexual dimorphism of Col1a1-Prkg2 RQ transgenic mice likely relates to cross-talk between ER-α, NO/cGMP and Wnt/β-catenin signaling pathways (Kouzmenko et al. 2004, Mendelsohn & Karas 2010): First, estrogens increase NO (and cGMP) concentrations in endothelial cells and osteoblasts via ER-α activation of NO synthase (Mendelsohn & Karas 2010, Marathe et al. 2012, Joshua et al. 2014). Urinary cGMP excretion over 24 h is higher in pre-menopausal women compared to age-matched men (Markovitz et al. 1997, Cui et al. 2009), but, to our knowledge, there has been no prior direct comparison of serum or tissue NO/cGMP concentrations between males and females. We found higher serum NO and cGMP concentrations in female compared to male mice, consistent with our previous finding that ovariectomy decreases serum cGMP concentration while estrogen supplementation increases it (Joshua et al. 2014). Second, PKG2 directly and indirectly (via Akt) regulates GSK-3β activity, causing β-catenin stabilization, nuclear translocation, and induction of gene expression (Zhao et al. 2005, Kawasaki et al. 2008, Kalyanaraman et al. 2017). Female WT POBs produced more NO and cGMP, and had more nuclear β-catenin compared to male POBs. The higher NO/cGMP activation of endogenous PKG2 may partly obscure the effects of PKGRQ expression in female osteoblasts. Third, PKG2 activation by NO/cGMP increases mRNA and protein expression of the Wnt co-receptor LRP5 (Kalyanaraman et al. 2017). Increased lrp5 expression likely contributed to the increased trabecular bone volume in male Col1a1-Prkg2 RQ transgenic mice, because transgenic mice over-expressing WT LRP5 show a similar increase in bone volume (Babij et al. 2003). Multiple reports indicate sexual dimorphism in Wnt/β-catenin signaling in bone, possibly caused by direct interactions between β-catenin and sex hormone receptors (Kouzmenko et al. 2004): (i) the low bone mass phenotype of LRP5 knockout mice and of transgenic mice over-expressing a Wnt inhibitor (secreted frizzled-related protein-1) is much more severe in males than females (Sawakami et al. 2006, Yao et al. 2010); (ii) the high bone mass phenotype of transgenic mice expressing WT or constitutively-active LRP5 mutants shows gender-related differences (Babij et al. 2003, Dubrow et al. 2007); and (iii) human studies suggest that associations between genetic LRP5 variants and BMD are stronger in males than in females (Ferrari et al. 2004, Kiel et al. 2007). To fully understand the molecular basis of these gender-specific differences requires further investigation.
Protection from diabetic bone loss by osteoblast-specific PKG2RQ expression
Low bone turn-over osteoporosis from impaired osteoblastic bone formation is a common complication of type 1 diabetes (Napoli et al. 2016). We have linked defective bone formation to decreased NO/cGMP/PKG signaling in osteoblasts under hyperglycemic conditions and shown that Prkg1 and 2 are both transcriptionally downregulated in bones of insulin-deficient mice (Kalyanaraman et al. 2018b). Here, we provide evidence that increasing PKG2 activity in osteoblasts is sufficient to protect male mice from diabetic bone loss by restoring osteoblast proliferation and activity, in part by preventing downregulation of Wnt/β-catenin signaling in diabetic bone. Thus, PKG2 appears to be the major cGMP target in diabetic bone.
This is linked to the online version of the paper at https://doi.org/10.1530/JOE-18-0286.
Declaration of interest
The authors declare that they have no conflicts of interest with the contents of this article. The content of the manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
This work was supported by NIH grants R01-AR051300 and R21-AR065658 (to R B P); P01-AG007996 (R L S), P30-CA023100 (UCSD Gene Targeting Mouse Core), P30-NS047101 (UCSD Neuroscience Microscopy Shared Facility) and P30-AR04603 (University of Alabama, Birmingham, Center for Metabolic Bone Disease). N S was supported by the Deutsche Forschungsgemeinschaft (Project KI 1303/2-1).
Authors’ contribution statement
Study design: R B P and G R. Study conduct: G R, N S, H K, N W, J C, D E C, S M, and E C. Data collection: G R, N S, H K, and E C. Data analysis: G R, H K, R L S and R B P. Data interpretation: R B P, G R, H K and R L S. Drafting manuscript: G R and R B P. R B P and G R take responsibility for the integrity of the data analysis.
The authors are grateful to Ella Kothari, Jennifer Santini, Dana Smith and Dezhi Wang for their expert assistance.
AguirreJButteryLO’ShaughnessyMAfzalFde MarticorenaIFHukkanenMHuangPMaclntyreIPolakJ 2001 Endothelial nitric oxide synthase gene-deficient mice demonstrate marked retardation in postnatal bone formation, reduced bone volume, and defects in osteoblast maturation and activity. American Journal of Pathology 158 247–257. (https://doi.org/10.1016/S0002-9440(10)63963-6)
ArmourKEArmourKJGallagherMEGodeckeAHelfrichMHReidDMRalstonSH 2001 Defective bone formation and anabolic response to exogenous estrogen in mice with targeted disruption of endothelial nitric oxide synthase. Endocrinology 142 760–766. (https://doi.org/10.1210/endo.142.2.7977)
BartelsCFBukulmezHPadayattiPRheeDKvan Ravenswaaij-ArtsCPauliRMMundlosSChitayatDShihLYAl-GazaliLIet al. 2004 Mutations in the transmembrane natriuretic peptide receptor NPR-B impair skeletal growth and cause acromesomelic dysplasia, type Maroteaux. American Journal of Human Genetics 75 27–34. (https://doi.org/10.1086/422013)
BocciardiRGiordaRButtgereitJGimelliSDiviziaMTBeriSGarofaloSTavellaSLeroneMZuffardiOet al. 2007 Overexpression of the C-type natriuretic peptide (CNP) is associated with overgrowth and bone anomalies in an individual with balanced t(2;7) translocation. Human Mutation 28 724–731. (https://doi.org/10.1002/humu.20511)
BonnetCAndrieuxJBeri-DexheimerMLeheupBBouteOManouvrierSDelobelBCopinHReceveurAMathieuMet al. 2010 Microdeletion at chromosome 4q21 defines a new emerging syndrome with marked growth restriction, mental retardation and absent or severely delayed speech. Journal of Medical Genetics 47 377–384. (https://doi.org/10.1136/jmg.2009.071902)
CuiRIsoHYamagishiKOhiraTTanigawaTKitamuraAKiyamaMImanoHKonishiMShimamotoT 2009 Relationship of urinary cGMP excretion with aging and menopausal status in a general population. Journal of Atherosclerosis and Thrombosis 16 457–462. (https://doi.org/10.5551/jat.No083)
FerrariSLDeutschSChoudhuryUChevalleyTBonjourJPDermitzakisETRizzoliRAntonarakisSE 2004 Polymorphisms in the low-density lipoprotein receptor-related protein 5 (LRP5) gene are associated with variation in vertebral bone mass, vertebral bone size, and stature in whites. American Journal of Human Genetics 74 866–875. (https://doi.org/10.1086/420771)
GuoDCRegaladoECasteelDESantos-CortezRLGongLKimJJDyackSHorneSGChangGJondeauGet al. 2013 Recurrent gain-of-function mutation in PRKG1 causes thoracic aortic aneurysms and acute aortic dissections. American Journal of Human Genetics 93 398–404. (https://doi.org/10.1016/j.ajhg.2013.06.019)
HagiwaraHInoueAYamaguchiAYokoseSFuruyaMTanakaSHiroseS 1996 cGMP produced in response to ANP and CNP regulates proliferation and differentiation of osteoblastic cells. American Journal of Physiology 270 C1311–C1318. (https://doi.org/10.1152/ajpcell.1996.270.5.C1311)
HannemaSEvan DuyvenvoordeHAPremslerTYangRBMuellerTDGassnerBOberwinklerHRoelfsemaFSantenGWPrickettTet al. 2013 An activating mutation in the kinase homology domain of the natriuretic peptide receptor-2 causes extremely tall stature without skeletal deformities. Journal of Clinical Endocrinology and Metabolism 98 E1988–E1998. (https://doi.org/10.1210/jc.2013-2358)
KalajzicIKalajzicZKaliternaMGronowiczGClarkSHLichtlerACRoweD 2002 Use of type I collagen green fluorescent protein transgenes to identify subpopulations of cells at different stages of the osteoblast lineage. Journal of Bone and Mineral Research 17 15–25. (https://doi.org/10.1359/jbmr.2002.17.1.15)
KalyanaramanHSchwappacherRJoshuaJZhuangSScottBTKlosMCasteelDEFrangosJADillmannWBossGRet al. 2014 Nongenomic thyroid hormone signaling occurs through a plasma membrane-localized receptor. Science Signaling 7 ra48. (https://doi.org/10.1126/scisignal.2004911)
KalyanaramanHRamdaniGJoshuaJSchallNBossGRCoryESahRLCasteelDEPilzRB 2017 A novel, direct NO donor regulates osteoblast and osteoclast functions and increases bone mass in ovariectomized mice. Journal of Bone and Mineral Research 32 46–59. (https://doi.org/10.1002/jbmr.2909)
KalyanaramanHSchwaerzerGRamdaniGCastilloFScottBTDillmannWSahRLCasteelDEPilzRB 2018b Potein kinase G activation reverses oxidative stress and restores osteoblast function and bone formation in male mice with type 1 diabetes. Diabetes 67 607–623. (https://doi.org/10.2337/db17-0965)
KawasakiYKugimiyaFChikudaHKamekuraSIkedaTKawamuraNSaitoTShinodaYHigashikawaAYanoFet al. 2008 Phosphorylation of GSK-3beta by cGMP-dependent protein kinase II promotes hypertrophic differentiation of murine chondrocytes. Journal of Clinical Investigation 118 2506–2515. (https://doi.org/10.1172/JCI35243)
KielDPFerrariSLCupplesLAKarasikDManenDImamovicAHerbertAGDupuisJ 2007 Genetic variation at the low-density lipoprotein receptor-related protein 5 (LRP5) locus modulates Wnt signaling and the relationship of physical activity with bone mineral density in men. Bone 40 587–596. (https://doi.org/10.1016/j.bone.2006.09.029)
KondoEYasodaAFujiiTNakaoKYamashitaYUeda-SakaneYKanamotoNMiuraMAraiHMukoyamaMet al. 2015 Increased bone turnover and possible accelerated fracture healing in a murine model with an increased circulating C-type natriuretic peptide. Endocrinology 156 2518–2529. (https://doi.org/10.1210/en.2014-1801)
MaratheNRangaswamiHZhuangSBossGRPilzRB 2012 Pro-survival effects of 17beta-estradiol on osteocytes are mediated by nitric oxide/cGMP via differential actions of cGMP-dependent protein kinases I and II. Journal of Biological Chemistry 287 978–988. (https://doi.org/10.1074/jbc.M111.294959)
MarkovitzJHLewisCESandersPWTuckerDWarnockDG 1997 Relationship of diastolic blood pressure with cyclic GMP excretion among young adults (the CARDIA Study): influence of a family history of hypertension. Coronary Artery Risk Development in Young Adults. Journal of Hypertension 15 955–962. (https://doi.org/10.1097/00004872-199715090-00005)
MiuraKNambaNFujiwaraMOhataYIshidaHKitaokaTKubotaTHiraiHHiguchiCTsumakiNet al. 2012 An overgrowth disorder associated with excessive production of cGMP due to a gain-of-function mutation of the natriuretic peptide receptor 2 gene. PLoS ONE 7 e42180. (https://doi.org/10.1371/journal.pone.0042180)
MiyazawaTOgawaYChushoHYasodaATamuraNKomatsuYPfeiferAHofmannFNakaoK 2002 Cyclic GMP-dependent protein kinase II plays a critical role in C-type natriuretic peptide-mediated endochondral ossification. Endocrinology 143 3604–3610. (https://doi.org/10.1210/en.2002-220307)
RangaswamiHSchwappacherRTranTChanGCZhuangSBossGRPilzRB 2012 Protein kinase G and focal adhesion kinase converge on Src/Akt/beta-catenin signaling module in osteoblast mechanotransduction. Journal of Biological Chemistry 287 21509–21519. (https://doi.org/10.1074/jbc.M112.347245)
RosselliMImthurnBKellerPJJacksonEKDubeyRK 1995 Circulating nitric oxide (nitrite/nitrate) levels in postmenopausal women substituted with 17 beta-estradiol and norethisterone acetate. A two-year follow-up study. Hypertension 25 848–853. (https://doi.org/10.1161/01.HYP.25.4.848)
SawakamiKRoblingAGAiMPitnerNDLiuDWardenSJLiJMayePRoweDWDuncanRLet al. 2006 The Wnt co-receptor LRP5 is essential for skeletal mechanotransduction but not for the anabolic bone response to parathyroid hormone treatment. Journal of Biological Chemistry 281 23698–23711. (https://doi.org/10.1074/jbc.M601000200)
WatanukiMSakaiASakataTTsurukamiHMiwaMUchidaYWatanabeKIkedaKNakamuraT 2002 Role of inducible nitric oxide synthase in skeletal adaptation to acute increases in mechanical loading. Journal of Bone and Mineral Research 17 1015–1025. (https://doi.org/10.1359/jbmr.2002.17.6.1015)
WimalawansaSJ 2000 Nitroglycerin therapy is as efficacious as standard estrogen replacement therapy (Premarin) in prevention of oophorectomy-induced bone loss: a human pilot clinical study. Journal of Bone and Mineral Research 15 2240–2244. (https://doi.org/10.1359/jbmr.2000.15.11.2240)
ZhaoXZhuangSChenYBossGRPilzRB 2005 Cyclic GMP-dependent protein kinase regulates CCAAT enhancer-binding protein beta functions through inhibition of glycogen synthase kinase-3. Journal of Biological Chemistry 280 32683–32692. (https://doi.org/10.1074/jbc.M505486200)