cGMP-dependent protein kinase-2 regulates bone mass and prevents diabetic bone loss

in Journal of Endocrinology
Restricted access

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.

Downloadable materials

  • Supplemental Figure 1: Activity of purified wild type and mutant PKG2 in vitro, presence of the prkg2RQ transgene in founder mice, and PKG2 protein level in POB membranes. (A,B) N-terminally Flag-tagged wild type and mutant PKG2 were affinity-purified from transfected 293T cells, and the proteins were analyzed by SDS-PAGE and Coomassie Blue staining (A). Kinase activity was measured with increasing concentrations of a synthetic peptide in the presence of 10 µM cGMP as described in Methods (B). (C) Presence of the prkg2RQ transgene was detected in tail DNA by PCR using primers F2/R2 shown in Fig. 1C; one male and two female founder mice with sex-matched non-transgenic litter mates are shown. The Fe-9.1 line was further characterized. (D) PKG2 protein was assessed by Western blotting of POB membrane fractions from two mice per genotype, with caveolin-1 serving as a loading control.
  • Supplemental Figure 2: Proliferation, ERK and Akt/GSK-3β signaling, and nuclear β-catenin in POBs isolated from 8 week-old female Col1a1-prkg2RQ transgenic and wild type mice. POBs were isolated from 8 week-old female transgenic mice and their wild type litter mates (n= 3 mice per genotype) and cultured as described for male POBs in Fig. 2. (A) Metabolically-active cells were quantified by measuring tetrazolium (MTS) reduction to formazan as described in Methods. (B) Serum-deprived POBs were treated with vehicle or 100 µM 8-CPT-cGMP (+ cGMP) for 10 min, and Western blots of cell extracts were analyzed for ERK1/2(pTyr204), Akt(pSer473), GSK3β(pSer9), and PKG2, with β-actin serving as a loading control. (C-E) Blots were obtained as in panel B, with independent POB isolates from three mice per genotype; they were analyzed by densitometry scanning using ImageJ. (F,G) Nuclear localization of β-catenin was determined by immuno¬fluo¬res¬cence staining of POBs isolated from three female mice per genotype. Some cells were treated with 100 µM 8-CPT-cGMP (+ cGMP) for 1 h (F). Some cells were pre-treated for 1 h with 4 mM N(ω)-nitro-l-arginine methyl ester (L-NAME) to inhibit NO production, and then were transferred to fresh medium with 4 mM L-NAME and incubated overnight prior to staining for β-catenin (G). Graphs show means ± SEM, *p/#p<0.05, **p/##p<0.01, and ***p/###p<0.001 for the indicated comparisons.
  • Supplemental Figure 3: Differentiation of POBs and BMSCs isolated from 8 week-old male Col1a1-prkg2RQ transgenic and wild type mice. POBs and BMSCs were isolated from eight week-old male transgenic (TG, n=5) mice and their wild type (WT, n=4) litter mates. (A, B) POBs were plated at high density and switched to differentiation medium after reaching confluency. After 14 d, cells were stained for ALP activity by staining with a colorimetric substrate (A). Mineralization was assessed after 21 d by Alizarin Red staining (B). Staining intensity was quantified by densitometry scanning. (C-E) Bone marrow mononuclear cells were plated at 4 x 105 cells/cm2, and adherent BMSCs were switched to osteoblastic differentiation medium for 14 d. BMSC colonies were assessed for mineralization by Alizarin Red staining (C). The relative mRNA abundance of osteocalcin (bglap), runx2, and actb (D) or prkg2 (E) was quantified by qRT-PCR and normalized to three housekeeping genes, as described in Fig. 2I. Data were calculated according to the ∆∆Ct method, with the mean of the WT group for each gene assigned a value of one. Data represent means ± SEM; *p<0.05 and **p<0.01 for comparison to WT.
  • Supplemental Figure 4: Micro-CT analysis of cortical parameters of 8 week-old male and female mice: Comparison of Col1a1-prkg2RQ transgenic and wild type littermates. In eight week-old male (A) and female (B) mice cortical bone area fraction, cross-sectional thickness and tissue mineral density (TMD) were measured by micro-CT at the mid-tibia as described in Methods. Data represent means ± SD (males: n=8 WT and n=10 TG; females: n=11 WT and n=10 TG).
  • Supplemental Figure 5: Calceine labeling of cortical bone in 8 week-old male mice and expression of Wnt-/osteoblast-, and RANKL-/osteoclast-related genes in tibiae from 8 week-old female mice. (A) Mineralizing surfaces (MS/BS), mineral apposition rates (MAR), and bone formation rates (BFR) were measured at endocortical surfaces between 5 and 0.25 mm proximal to the femoral growth plate in 8 week-old male Col1a1-prkg2RQ transgenic mice and wild type littermates. Data show the means ± SD of n=4 WT and n=5 TG mice; *p<0.05 and **p<0.01. (B,C) RNA was extracted from tibial shafts of 8 week-old female mice, and the relative abundance of (B) Wnt-/osteoblast- and (C) RANKL-/osteoclast-related transcripts was quantified by qRT-PCR and normalized to three different housekeeping genes (18S, hprt, and b2m). The mean of the WT group for each gene was assigned a value of one, as described in Fig. 5F. Gene names: Ctnnb1 (β-catenin), bglap (osteocalcin), alpl (alkaline phospha¬tase), spp1 (osteopontin), ccnd1 (cyclin D1), actb (β-actin), tnfsf11 (RANKL), tnfrsf11b (osteoprotegerin), acp5 (tartrate-resistant acid phosphatase) and ctsk (cathepsin K). Data represent means ± SEM from n=7 wild type and n=10 transgenic females per genotype. *p<0.05 for the comparison between WT and TG mice.
  • Supplemental Table 1: Primers Used for Quantitative RT-PCR

 

      Society for Endocrinology

Article Information

Metrics

All Time Past Year Past 30 Days
Abstract Views 160 160 83
Full Text Views 803 803 40
PDF Downloads 55 55 6

Altmetrics

Related Articles

Figures

  • View in gallery

    Characterization of osteoblast-specific PKG2RQ expression in transgenic Col1a1-Prkg2 RQ mice. (A and B) WT and mutant PKG2RQ enzymes were affinity-purified from transfected 293T cells, and basal kinase activity was measured with a synthetic peptide in the absence of cGMP (A) or in the presence of increasing cGMP concentrations (B). (C) Scheme of the Col1a1-Prkg2 RQ construct used for injection into fertilized mouse eggs; the location of the R242Q mutation in the rat Prkg2 cDNA is indicated by an asterisk. The positions of PCR primers used for genotyping (F1/R1) or detection of transgenic mRNA (F2/R2) are shown by arrows. (D) RNA was extracted from the indicated organs of a transgenic mouse, and transgene-derived mRNA was detected by RT-PCR using the F2/R2 primer pair shown in C. No PCR signal was obtained when reverse transcriptase was omitted (not shown); Gapdh served as a control for RNA quality. (E) RNA was extracted from bone, brain, and kidney of WT and transgenic (TG) mice, and Prkg2 mRNA was quantified by qRT-PCR using primers recognizing both rat and mouse transcripts. Data were normalized to β2-microglobulin (B2m) and calculated according to the ∆∆Ct method, assigning the mean of the WT group a value of one. Data are means ± s.e.m. from n = 10 mice per genotype for bone (5 males and five females) and n = 4 for brain and kidney (***P < 0.001). (F) Cell membranes were purified from primary osteoblasts (POBs) from WT and TG mice using a percoll gradient, and cGMP-stimulated PKG activity was measured (mean ± s.e.m. of three independent experiments; ***P < 0.001). (G and H) POBs from WT and TG mice were serum-starved and treated with 100 µM 8-pCPT-cGMP (+cGMP) for 10 min. The amount of PKG2 protein and phosphorylation of the PKG substrate VASP on Ser239 were assessed by Western blotting of whole cell lysates, with β-actin serving as a loading control (G). Results of three independent POB isolates per genotype were quantified by densitometry scanning (H). Data represent means ± s.e.m. (*P < 0.05, **P < 0.01, ***P < 0.001, and ### P < 0.001 for the indicated comparisons). (I) Immunofluorescence staining for PKG2 (green) in POBs isolated from WT and TG mice (nuclei counterstained with Hoechst 33342). (J) Immunohistochemical staining for PKG2 in tibial sections from WT and TG mice; top panels show osteoblasts on trabecular surfaces, bottom panels show megakaryocytes, which served as a positive control (bar = 20 µm). I and J are representative of POBs and bones from three mice per genotype.

  • View in gallery

    Proliferation, Erk and Akt/GSK3/β-catenin signaling, survival, and differentiation of POBs isolated from male Col1a1-Prkg2 RQ transgenic and WT POBs. POBs were isolated from eight week-old male transgenic (TG, n = 5) and WT (n = 4) mice. (A and B) Cells were plated at 1.2 × 105 cells/cm2 and were counted 72 h later to calculate population doubling times (A). Metabolically-active cells were quantified by MTS reduction as described in Methods (B). Each data point represents the mean of three independent experiments performed with POBs from one mouse at passage 2 (lines show means ± s.d.; **P < 0.01). (C, D, E and F) Serum-deprived POBs were treated with vehicle or 100 µM 8-pCPTcGMP (+cGMP) for 10 min. Western blots of cell extracts were analyzed with phospho-specific antibodies for ERK1/2-(pTyr204), Akt(pSer473), or GSK3β(pSer9); β-actin served as a loading control. Bar graphs summarize results from three independent POB isolates per genotype. (G) Nuclear localization of β-catenin was determined by immunofluorescence staining of POBs isolated from three male mice per genotype. Bar graphs (D-G) show means ± s.e.m., *P/# P < 0.05, **P/## P < 0.01, and ***P/### P < 0.001 for the indicated comparisons. (H) POBs were serum-starved for 24 h or kept in normal growth medium; apoptosis was detected by immunofluorescence staining for cleaved caspase-3 (green, nuclei counterstained with Hoechst 33342). The percentage of cells staining positive for cleaved caspase-3 is shown below (means ± s.e.m. of three independent experiments; # P < 0.05 for comparison to WT in full serum, and *P < 0.05 for comparison to WT serum-starved). (I) RNA was extracted from confluent POBs cultured in differentiation medium for 14 day; osteoblast differentiation-related transcripts were quantified by qRT-PCR and normalized to three housekeeping genes (18S, hprt, and b2m). Data were calculated according to the ∆∆Ct method, with the mean of the WT group for each gene assigned a value of one. Gene names: Ctnnb1 (β-catenin), Col1a1 (collagen 1a1), Bglap (osteocalcin), and Actb (β-actin). (J) Bone marrow mononuclear cells from eight week-old male TG and WT mice (n = 5 mice of each genotype) were plated at 4 × 105 cells/cm2, and adherent BMSCs were switched to osteoblastic differentiation medium. After 14 day, BMSC colonies were stained for ALP activity; the number of ALP+ colonies per well was counted and expressed per 106 bone marrow cells plated (means ± s.e.m.; *P < 0.05 and **P < 0.01 for comparison to WT).

  • View in gallery

    Gender-specific differences in serum NOx and cGMP concentrations in mice. (A, B and C) POBs were isolated from 8 week-old male or female WT mice (n = 4 each) and were plated at passage 3 at equal density. Cells were transferred to fresh medium for 1 h prior to measuring NOx (indicating the sum of nitrate plus nitrite) in the medium by Griess reaction (A). NOS-3 phosphorylation on Ser1799 was assessed in cell extracts by Western blotting with a phospho-specific antibody (B). Intracellular cGMP concentrations were measured by ELISA (C). (D and E) Serum was obtained by cardiac puncture at the time of euthanasia from male and female WT and transgenic (TG) mice (n = 6 per group). NOx and cGMP concentrations were quantified by Griess reaction and ELISA, respectively. Data represent means ± s.d. (**P < 0.01 and ***P < 0.001 for the indicated comparisons).

  • View in gallery

    Increased trabecular bone mass in male, but not female Col1a1-Prkg2 RQ transgenic mice. Male (A, C and E) and female (B, D and F) transgenic (TG) mice and their WT litter mates were analyzed at eight weeks of age. (A and B) Body weight and tibia length are shown for each gender. (C and D) Tibiae were analyzed by micro-CT, with three-dimensional reconstruction of the trabecular bone at the proximal tibia below the growth plate shown. (E and F) Trabecular BMD, trabecular bone volume fraction (BV/TV), trabecular number (Tb.N), and trabecular thickness (Tb.Th) were measured by micro-CT at the proximal tibia. Data represent means ± s.d. (males: n = 8 WT and n = 10 TG; females: n = 11 WT and n = 10 TG). *P < 0.05 for the indicated comparisons.

  • View in gallery

    Increased bone formation and expression of Wnt-related genes in male Col1a1-Prkg2 RQ transgenic mice. (A and B) Trichrome stains of distal femur sections were analyzed in 8 week-old male mice and the thickness of the growth plate was measured in the center and at three equally spaced points on each side (n = 4 WT and n = 5 TG). (C) Eight week-old males received calcein injections 7 and 2 d prior to euthanasia, and trabecular calcein labeling was assessed by fluorescence microscopy. Mineralizing surfaces (MS/BS), MAR, and BFR were measured at trabecular surfaces between 0.25 and 1.25 mm proximal to the femoral growth plate. (D and E) Osteoblasts (D) were counted on femoral trabecular and endocortical surfaces, osteoclasts (E) on trabecular surfaces. Panels C, D and E show means ± s.d. of n = 4 WT and n = 5 TG mice; *P < 0.05 and **P < 0.01 by two-sided t-test. (F and G) RNA was extracted from tibial shafts, and the relative abundance of Wnt-/osteoblast- and RANKL- or osteoclast-related transcripts was quantified by qRT-PCR and normalized to three different housekeeping genes (18S, Hprt, and B2m). Data were calculated according to the ∆∆Ct method, with the mean of the WT group for each gene assigned a value of one. Gene names: Ctnnb1 (β-catenin), Bglap (osteocalcin), Alpl (alkaline phosphatase), Spp1 (osteopontin), Ccnd1 (cyclin D1), Actb (β-actin), Tnfsf11 (RANKL), Tnfrsf11b (osteoprotegerin), Acp5 (tartrate-resistant acid phosphatase) and Ctsk (cathepsin K). Data represent means ± s.e.m. from n = 6 mice per genotype. *P < 0.05 and **P < 0.01 for the comparison between WT and TG mice (by two-sided t-test).

  • View in gallery

    Male Col1a1-Prkg2 RQ transgenic mice are protected from diabetes-induced bone loss. (A and B) Six-week-old male mice were injected with vehicle or streptozotocin (STZ, 100 mg/kg/d for two days) to induce type 1 diabetes. Blood glucose was measured 12 days later (B), and only hyperglycemic (>15 mM) STZ-treated mice were included in further analyses. All mice received calcein injections 7 and 2 d prior to euthanasia at the age of 12 weeks. (C and D) Tibiae were analyzed by micro-CT imaging (n = 7 WT/control, n = 8 WT/STZ, n = 7 TG/control, n = 9 TG/STZ). Trabecular bone volume fraction, trabecular thickness, and BMD were quantified by micro-CT at the proximal tibia (C). Cortical bone area fraction, cross-sectional thickness, and TMD were quantified at the mid-tibia (D). (E) Endocortical calcein labeling was assessed at the tibia, with quantification of mineralizing surfaces (MS/BS), MAR, and BFR (n = 4–5 mice per group). Trabecular calcein labeling in diabetic mice was too weak to allow reliable quantification. Data in dot blots (B, C, D and E) represent means ± s.d.; *P < 0.05, **P < 0.01, ***P < 0.001 for the indicated pair-wise comparisons. (F) RNA was extracted from tibial shafts. Relative mRNA abundance of osteoblastic and Wnt-related genes was quantified by qRT-PCR, as was expression of cyclin D (Ccnd1) and Hprt. Gene expression was normalized to 18S and data were calculated according to the ∆∆Ct method, with the mean of the WT control group for each gene assigned a value of one (n = 6 mice per group; *P < 0.05, **P < 0.01, ***P < 0.001 for the comparison to WT control mice, # P < 0.05 for the indicated comparison).

PubMed

Google Scholar