Parts of this work were presented at the 26th Annual Meeting of the Society for Bone and Mineral Research, October 1–5, 2004, Seattle, WA, USA.
The authors state that they have no conflicts of interest.
GSK-3, a component of the canonical Wnt signaling pathway, is implicated in regulation of bone mass. The effect of a small molecule GSK-3 inhibitor was evaluated in pre-osteoblasts and in osteopenic rats. GSK-3 inhibitor induced osteoblast differentiation in vitro and increased markers of bone formation in vitro and in vivo with concomitant increased bone mass and strength in rats.
Introduction: Inactivation of glycogen synthase kinase −3 (GSK-3) leads to stabilization, accumulation, and translocation of β-catenin into the nucleus to activate downstream Wnt target genes. To examine whether GSK-3 directly regulates bone formation and mass we evaluated the effect of 603281-31-8, a small molecule GSK-3 α/β dual inhibitor in preosteoblastic cells and in osteopenic rats.
Materials and Methods: Murine mesenchymal C3H10T1/2 cells were treated with GSK-3 inhibitor (603281-31-8) and assayed for β-catenin levels, activity of Wnt-responsive promoter, expression of mRNA for bone formation, and adipogenic markers and alkaline phosphatase activity. In vivo, 6-month-old rats were ovariectomized (OVX), allowed to lose bone for 1 month, and treated with GSK-3 inhibitor at 3 mg/kg/day orally for 60 days. At the end of treatment, BMD was measured by DXA, bone formation rate by histomorphometry, vertebral strength (failure in compression), and the expression levels of osteoblast-related genes by real-time PCR.
Results: Treatment of C3H10T1/2 cells with the GSK-3 inhibitor increased the levels of β-catenin accompanied by activation of Wnt-responsive TBE6-luciferase reporter gene. This was associated with an increased expression of mRNA for bone sialoprotein (1.4-fold), collagen α1 (I) (∼2-fold), osteocalcin (1.2-fold), collagen α1(V) (1.5-fold), alkaline phosphatase (∼160-fold), and runx2 (1.6-fold), markers of the osteoblast phenotype and bone formation activity. Alkaline phosphatase mRNA expression paralleled alkaline phosphatase activity. The mRNA levels of collagens α1 (I), α1 (V), biglycan, osteonectin, and runx-2 increased on treatment with the GSK-3 inhibitor in rat femur compared with the OVX control. DXA analyses revealed significant increases in BMC and BMD in cancellous and cortical bone of OVX rats treated with GSK-3 inhibitor. This was associated with increased strength (peak load, energy, and stiffness) assessed by lumbar vertebra load to failure in compression. Histomorphometric analyses showed that 603281-31-8 robustly increased bone formation but did not exclude a small effect on osteoclasts (resorption).
Conclusions: An orally active, small molecule GSK-3 inhibitor induced osteoblast differentiation and increased markers of bone formation in vitro, and increased markers of bone formation, bone mass, and strength in vivo, consistent with a role for the canonical Wnt pathway in osteogenesis.
The canonical Wnt signaling pathway is highly conserved among different species and is crucial for the specification of cell fates during development, regulation of cell growth, differentiation, and apoptosis.(1–3) Wnt ligands are secreted glycoproteins that signal through a receptor complex comprising of seven transmembrane domain receptors, Frizzleds, and the co-receptor lipoprotein-receptor-related proteins 5 or 6 (LRP5/6).(4–11) The resulting signal is propagated and integrated through a downstream key signaling kinase, glycogen synthase kinase-3β (GSK-3β). GSK-3, a key regulator of glycogen metabolism, is a serine/threonine kinase that regulates cellular functions including cell proliferation and differentiation, cell adhesion, microtubule dynamics, and phosphorylation of initiation factors.(12) Molecular cloning of GSK-3 from skeletal muscle revealed two closely related isoforms, GSK-3α and GSK-3β, which are ubiquitously expressed in mammalian tissues. These proteins share 97% sequence similarity within their kinase catalytic domains.(13,14) GSK-3 is present in a protein complex consisting of axin, adenomatosis polyposis coli (APC) and β-catenin. In cells lacking Wnt signaling, GSK3-β phosphorylates β-catenin, inducing rapid degradation of β-catenin through the E3 ubiquitin/proteasome pathway. In contrast, stimulation of the Wnt pathway signaled through disheveled (Dsh) and GSK-3 binding protein inhibits GSK-3 and prevents the phosphorylation of β-catenin. This leads to the stabilization, accumulation, and translocation of β-catenin from cytoplasm to the nucleus and binding to the T-cell factor/lymphoid enhancer factor (TCF/LEF) family of transcription factors to activate transcription of specific genes.(1,2)
Genetic analysis has implicated the Wnt signaling pathway as an important regulator of bone mass. Activating mutations in LRP5 have been shown to result in a high bone mass phenotype associated with increased bone formation and no other obvious abnormalities in humans and animals, whereas inactivating mutations in LRP5 result in osteoporosis-pseudoglioma syndrome (OPPG).(15–19) In mice, targeted disruption of LRP5 or LRP6 leads to a low bone mass phenotype, whereas compound Lrp5 and Lrp6 mutations resulted in dose-dependent deficits in BMD and limb formation, suggesting their functional redundancy.(20,21) Several biochemical and expression studies have shown that manipulations of components of the Wnt pathway are associated with bone cell function.(3,22–25)
In view of the central regulatory role of GSK-3 in the Wnt pathway, modulation of GSK-3 activity/expression might be an useful approach to the treatment and prevention of clinical conditions of low bone mass. We have developed an orally active GSK-3 α/β inhibitor (603281-31-8) that showed a 500-fold selectivity window for GSK-3β over a panel of other kinases.(26) In this study, we investigated the effects of this GSK-3 inhibitor in pre-osteoblastic C3H10T1/2 cells and in vivo on bone using osteopenic rats. We show that GSK-3 inhibitor induced osteoblast commitment/differentiation in vitro and increased markers of bone formation in vitro and in vivo and a concomitant increase in bone mass and strength in vivo. These findings confirm the role of Wnt signaling pathway in bone and suggest that potent and selective inhibitors of GSK-3 could have use as bone anabolic agents.
MATERIALS AND METHODS
GSK-3 α/β dual inhibitor, 603287-31-8, was synthesized at the Lilly Research Laboratories (Indianapolis, IN, USA) as previously described.(26)
Animals and dosing regimen
Six-month-old virgin Sprague-Dawley female rats (Harlan Industries, Indianapolis, IN, USA) weighing ∼220 g were maintained on a 12-h light/12-h dark cycle at 22°C with ad libitum access to food (TD 89222 with 0.5% Ca and 0.4% P; Tekdad, Madison, WI, USA) and water. Bilateral ovariectomies (OVX) were performed on the rats except for the Sham-operated controls, and the OVX rats were randomized into treatment groups of six to eight rats. OVX rats were permitted to lose bone for 1 month to establish osteopenia before the initiation of treatments. GSK-3 inhibitor at a dose of 3 mg/kg/day was given to the OVX rats through gavage for 2 months. Age-matched Sham and OVX controls were treated with respective vehicles. 603281-31-8 (Eli Lilly and Co.) was prepared in 1% carboxymethylcellulose/0.25% Tween 80 vehicle (Sigma-Aldrich, St Louis, MO, USA). Animals were weighed every 2 weeks, and the dosing volumes were adjusted accordingly. The protocols were approved by the Animal Care Committee of Lilly Research Laboratories to ensure compliance with NIH guidelines. One day after the last dose, the animals were killed by CO2 inhalation. Femurs were removed and cleaned of soft tissue, and the right femur was fixed in 10% formalin for 48 h and stored at 4°C in 70% ethanol. From the left femur, the distal epiphysis including the growth plate was removed, and a subjacent 3-mm-wide band of the metaphyseal primary spongiosa was resected for mRNA analysis. Additionally, lumbar vertebrae L5 were removed and processed for biomechanical analyses.
In vitro protocols
Murine C3H10T1/2 mesenchymal cells (ATCC, Rockville, MD, USA) were plated as monolayer cultures and maintained in DMEM (Invitrogen, Carlsbad, CA, USA) containing 10% heat-inactivated FBS (hiFBS; Hyclone Laboratories, Logan, UT, USA). Medium was changed every 3 days. For mRNA analysis, cultures (4-T150 flasks/group) of cells were grown as described above to 80–90% confluence and switched to media containing 0.1% FBS overnight. C3H10T1/2 cells were treated with GSK-3 inhibitor at a concentration of 1.0 μM for 1, 3, 6, and 24 h.
RNA isolation, Northern blots, and quantitative real-time PCR
At the end of each treatment, tissue samples were removed from the animals and snap frozen in liquid nitrogen as individual samples of seven and eight animals per group. Total RNA was isolated from the metaphyseal primary spongiosa of rats treated with vehicle or treated with the GSK-3 inhibitor as previously described.(27,28) Samples were homogenized in Ultraspec-II reagent (Biotecx, Houston, TX, USA) using an LS 10–35 Polytron homogenizer (Brinkmann Instruments, Westbury, NY, USA) as recommended by the manufacturer. RNA was isolated from C3H10T1/2 cultures by adding Ultraspec-II directly to the culture flasks. The resulting cell lysates were passed several times through a 10-ml pipette before collection. Poly A+ RNA was isolated from total RNA using Oligotex resin (Qiagen, Santa Clarita, CA, USA) according to manufacturer's protocol and quantified by spectrophotometry. The absorbance at 260 nm was determined, and the 260/280 nm absorbance ratio was used to ensure RNA quality. Two micrograms of Poly A+ RNA were denatured in 0.04 M 3-(N-morpholino) propanesulfonic acid (pH 7.0), 10 mM sodium acetate, 1 mM EDTA, 2.2 M formaldehyde, and 50% formamide at 60°C for 10 minutes, size fractionated by electrophoresis through 1% agarose gels in 2.1 M formaldehyde and 1× 3-(N-morpholino) propanesulfonic acid, and transferred to nylon membranes (Brightstar-Plus, Ambion, Austin, TX, USA). The nylon membranes were air-dried, and the samples were cross-linked to the membranes by UV irradiation in a Stratalinker (Stratagene, La Jolla, CA, USA). Collagen α1 (I), bone sialoprotein (BSP) and GAPDH cDNA was used to generate radioactive probes using the Random Primer DNA labeling kit (Invitrogen, Carlsbad, CA, USA). DNA probes (25 ng) were labeled using [α-32P] dCTP (Amersham Pharmacia Biotech). The unincorporated nucleotides were removed by centrifugation through a Centricon-50 column (Amicon, Bedford, MA, USA). Prehybridization and hybridization were carried out at 48°C in NorthernMax buffers (Ambion, Austin, TX, USA) for 16 h. After hybridization, membranes were washed for 30 minutes at room temperature in buffer containing 2× SSC and 0.1% SDS and then for 30 minutes at 48°C in 0.2× SSC and 0.1% SDS. The membranes were exposed to Biomax MS X-ray film (Eastman Kodak Co.) at −70°C. Autoradiograms were quantitated by scanning laser densitometry (2400 Gel Scan XL; LKB, Piscataway, NJ, USA). Labeled bands were quantitated as densitometric units and normalized to that of GAPDH signals used as internal control. The data were expressed as fold-change over untreated controls.
Quantitative real-time PCR was performed to assess the expression levels of bone and adipogenic marker genes using ABI Prism Sequence Detection System 5700 (Applied Biosystems, Foster City, CA, USA). RNA samples were treated with DNaseI for 30 minutes at 37°C using a DNA-free kit (Ambion) to eliminate genomic DNA. First-strand cDNA was synthesized from 4 μg of total RNA with random hexamer primers using the SUPERSCRIPT II first strand synthesis kit (Invitrogen). Fluorogenic primer/probe sets to detect mRNA encoding alkaline phosphatase, biglycan, osteocalcin, BSP, collagen α1 (I), collagen α1 (V), osteonectin and procollagen C proteinase enhancer (PCPE), runx2, osteoprotegerin (OPG), RANKL, cebpa (CCAAT enhancer binding protein α), GPD1 (glycerol-3-phosphate dehydrogenase 1), Lpl (lipoprotein lipase), FABP4 (fatty acid binding protein 4), and Pparg (peroxisome proliferator activated receptor γ) were obtained from Applied Biosystems as Assays on Demand reagents (Foster City, CA, USA). Specific amplification reactions from the cDNA were carried out by two-step real-time PCR, and the relative quantities were obtained by generating a standard curve for each gene. The expression levels of genes were normalized to 18S ribosomal RNA used as internal control. All measurements were performed on three to five replicate samples.
Protein isolation and immunoblot analysis
β-catenin levels were measured in the whole cell lysates of C3H10T1/2 cells treated with the GSK-3 inhibitor. Cultures were seeded at 2.5 × 106 cells onto a T-25 flask in 0.5% serum containing media and allowed to attach overnight at 37°C. The following day, the cells were treated with different concentrations of GSK3 inhibitor ranging from 0.01 to 10 μM for 24 h. At the end of treatment, the cells were washed once with PBS. After PBS wash, the cell pellets were Dounce homogenized using a disposable microcentrifuge pestle in 300 μl freshly prepared lysis buffer containing 10 mM K2HPO4, pH 7.2, 1 mM EDTA, 5 mM EGTA, 10 mM MgCl2, 50 mM β-glycerophosphate, 1 mM Na3VO4, 2 mM dithiothreitol, 1% Triton X-100, and protease inhibitors (complete protease inhibitor tablet; Roche). Lysates were incubated on ice for 30 minutes and centrifuged at 14,000 rpm for 30 minutes at 4°C. The supernatant was transferred to a clean tube, and the total protein concentration was determined by the bicinchoninic acid (BCA) method (Pierce, Rockford, IL, USA)
Sample buffer was added to 5 μg of total protein in NuPage to a final concentration of 1× and separated on a 10% NuPage gel (Invitrogen). After electrophoresis, the protein was transferred to a nitrocellulose membrane. The nonspecific binding was blocked using 5% dry milk in PBS containing 0.5% Tween 20 (PBST) for 1 h at room temperature. The blots were incubated in mouse monoclonal β-catenin antibody (Transduction Laboratories, San Jose, CA, USA) at 1:1000 dilution in PBST containing 5% dry milk overnight at 4°C, with gentle agitation. As a control for equal loading of protein anti-mitogen-activated protein (MAP) kinase 1/2 extracellular receptor kinase ([Erk1/2]-CT, rabbit polyclonal antibody; Upstate Biotechnology, Waltham, MA, USA) was added at a dilution of 1:5000 in PBST. The next day, the blots were washed three times in PBST for 5 minutes at room temperature and incubated with horseradish peroxidase (HRP) conjugated secondary antibody (NEB anti-mouse and anti-rabbit) at 1:2000 plus 1:5000 Precision Strep Tactin HRP conjugate (BioRad 161-0380) in 5% dry milk in PBST for 1 h at room temperature. After incubation with the secondary antibody, the blots were washed three times in PBST for 5 minutes at room temperature. The blots were drained on a paper towel and reacted with 1:1 ECL mix (Amersham Pharmacia Biotech) for 1 minute at room temperature wrapped in Saran wrap and exposed to film. The levels of β-catenin were quantitated by scanning laser densitometry (2400 Gel Scan XL; LKB). The bands were quantitated as densitometric units, and the data are expressed as fold-change over untreated controls.
Transfection and luciferase reporter gene assays
The TCF reporter plasmid containing six copies of β-catenin DNA binding site TCF/LEFs upstream of luciferase reporter gene, TOPflash, obtained from Upstate Biotechnology was transiently transfected into C3H10T1/2 cells using Fugene 6 reagent (Roche Molecular Biochemicals). After transfection for 4 h, the cells were washed and resuspended in media containing 0.5% FBS. For reporter gene assays, cells were plated at a density of 50,000 cells/well in 96-well plates, and the experiments were initiated after serum withdrawal for 12–16 h. The cells were treated with different concentrations of GSK-3 inhibitor ranging from 0.01 to 10 μM for 24 h. After treatment, the cell extracts were assayed for luciferase activity using the luciferase reporter gene assay kit (Roche Biochemicals) as recommended by the manufacturer. Luminiscence was measured in Dynatech MLX Luminometer, and light integration was measured at 5 s as summed relative luminescence. All measurements were performed on eight replicate samples.
Detection of alkaline phosphatase activity
C3H10T1/2 cells were maintained in 10% hiFBS in DMEM. After trypsinization, cells were plated at 4000 cells/well in 96-well plates in 10% hiFBS/DMEM and incubated overnight at 37°C, 5% CO2, 95% air. The next day, medium was replaced with 5% hiFBS/DMEM. The cells were treated with different concentrations of GSK-3 inhibitor ranging from 0.016 to 10 μM in duplicate plates and incubated for 72 h with or without indicated concentrations of GSK-3 inhibitor. After treatment, media was removed from one of the plates, wells were rinsed with 100 μl PBS, and the plate was submitted to three freeze (−80°C)/thaw cycles. One-Step PNPP (200 μl; Pierce) was added, and the plate was incubated at 37°C and read at 405 nm after 30 minutes.(29) To take into account possible changes in cell number, 20 μl CellTiter 96 (Promega) was added to the wells of the duplicate plate, and the plate was incubated at 37°C for 1 h before reading at 490 nm. Under these conditions, cell number was linearly related to OD 490 nm. Therefore, a corrected OD for alkaline phosphatase activity was calculated by dividing the OD 405 nm for alkaline phosphatase assay by the OD 490 nm from CellTiter assay.
Proximal tibias and tibial shaft were stained for 4 days in Villanueva osteochrome bone stain for osteoid staining (Polysciences, Warrington, PA, USA) and dehydrated in a graded ethanol, defatted in acetone, and embedded in methyl methacrylate. Longitudinal sections of 210 μm thickness were cut using a diamond wafering saw (Buehler Isomet, Evanston, IL, USA) and further hand ground to 20-μm sections of proximal tibia metaphysis (PTM) and 30 μm of tibial shaft (TX). For PTM analyses, the measurements were performed on the entire marrow region within the cortical shell between 1 and 4 mm distal to the growth plate–metaphyseal junction using an Image Analysis System (Osteomeasure). Trabecular area, perimeter, single- and double-labeling surfaces, eroded surface, osteoid surface, labeling, and wall width were measured, and trabecular number, thickness, separation, mineralizing surface, mineral appositional rate, bone formation rate/bone volume (BFR/BV), surface reference (BFR/BS), and activation frequency were calculated. Osteoclast number was measured on the entire marrow region within the cortical shell between 0.67 and 2 mm (PTM) under ×20 magnification. The osteoclast number was normalized to trabecular bone surface. For analysis of cortical bone, TX, cross-sectional area, marrow area, eroded surface, single-and double-labeling surfaces, and labeling width were measured. These parameters were used to calculate the percent cortical bone area, marrow area, mineralizing surface, mineral appositional rate, and bone formation rate/surface reference (BFR/BS) as described previously.(30,31)
Biomechanical analyses of lumbar vertebrae
Excised L5 vertebrae were used to evaluate the biomechanical properties of the GSK-3 inhibitor–treated bones. Mechanical properties of the L5 vertebrae were analyzed after the posterior processes were removed, and the ends of the centrum were made parallel using a diamond wafering saw (Buehler Isomet). Vertebral specimens were loaded to failure in compression, using the materials testing device and analyzed using Test Works 4 software (model: 661.18c-01; MTS Corp., Minneapolis, MN, USA). The compressive load was applied through a pivoting platen to correct for possible nonparallel alignment of the faces of the vertebral body. Specimens were tested in a saline solution at 37°C after equilibration. Parameters measured from the load-displacement curve included ultimate load (Fu), stiffness, and energy (area under the curve). The modulus of toughness was calculated by normalizing energy by the area.
DXA (Norland, Fort Atkinson, WI, USA) scans were conducted ex vivo on all femurs, and BMD, BMC, and cross-sectional area were determined. The scan field was 2.5 × 4.5-cm with 0.1 × 0.1-cm resolution and scan speed of 10 mm/s.
Data are expressed as mean ± SD. Group differences were assessed by ANOVA with pairwise contracts examination. Fisher's protested least significant difference test (PLSD) was used to compare the differences between the groups, where the significance levels for the overall ANOVA was p < 0.05. Real-time PCR results were analyzed by Student's t-test, and p < 0.05 was considered statistically significant.
GSK-3 inhibitor increased the levels of functional β-catenin in C3H10T1/2 cells
The levels of β-catenin protein measured by Western blot analysis in the whole cell lysates of C3H10T1/2 cells treated with the GSK-3 inhibitor (603281-31-8) were increased at concentrations ranging from 0.01 to 10 μM after 24-h treatment (Fig. 1A). This increase was dose-dependent, with a maximum increase of 1.5-fold at 1.0 μM concentration of inhibitor. Activation of Wnt signaling was assessed by using 603281-31-8 in C3H10T1/2 cells transiently transfected with TCF-luciferase reporter plasmid, TOPflash (Fig. 1B). 603281-31-8 stimulated the TCF-luciferase reporter gene activity at concentrations ranging from 0.01 to 10 μM. A 2- to 7-fold increase in the reporter gene activity was observed at 24 h after treatment with the compound. These results indicate that inhibition of GSK-3 leads to the increase and functional activation of β-catenin protein in C3H10T1/2 cells.
GSK-3 inhibitor had a biphasic effect on markers of osteoblast differentiation and decreased proliferation in C3H10T1/2 cells
GSK-3 inhibitor (603281-31-8) was evaluated on the mRNA levels of established osteoblast marker genes in C3H10T1/2 cells. The cells were treated with 603281-31-8, and the expression levels of bone marker genes were assessed by Northern blot analysis or by quantitative PCR. As shown in Figs. 2A and 2B, 603281-31-8 had a biphasic effect on the mRNA expression of collagen α1 (I), BSP, osteocalcin, collagen α1 (V), and runx2. For collagens α1 (I) and α1 (V), BSP, and runx2, there was an initial increase at 1 h, a decrease at 3 and 6 h, and an increase at 24 h. The expression of osteocalcin decreased at 1–6 h but eventually increased or rebounded to above control levels at 24 h. Interestingly, the mRNA level of alkaline phosphatase, a marker of early osteoblast differentiation, was increased by 160-fold by 24 h with virtually undetectable levels at initial time-points (Fig. 2C). This was associated with increased enzymatic activity of alkaline phosphatase (Fig. 2D). A dose-dependent increase in the alkaline phosphatase activity was observed with highest response at 0.1–1.0 μM of GSK-3 inhibitor followed by a decrease at 10.0 μM. Because GSK-3 plays an important role in cell proliferation and differentiation, we evaluated the effect of the GSK-3 inhibitor (603281-31-8) on cell proliferation. As shown in Fig. 2D, treatment with 603281-31-8 decreased cell proliferation. This effect was evident at the lowest concentration of the inhibitor tested (0.016 μM) and pronounced at the higher concentrations of 3–10 μM. Together, these results show that treatment with GSK-3 inhibitor decreased cell proliferation and concomitantly increased markers of osteoblast differentiation.
GSK-3 inhibitor decreases the expression levels of adipogenic markers in C3H10T1/2 cells
C3H10T1/2 behaves as a pluripotent mesenchymal stem cell, with the ability to differentiate into osteoblasts, adipocytes, chondrocytes, and myoblasts. We therefore explored the effect of 603281-31-8 on markers of adipogenesis (Cebpa, GPD1, Lpl, FABP4, and Pparg). In contrast to the biphasic increase in osteoblastic markers, treatment with 603281-31-8 decreased expression of adipocyte-related genes (Fig. 3). These results are consistent with the anabolic effect of 603281-31-8 being the result of enhanced osteoblast commitment-differentiation rather than expansion of a progenitor population.
Effect of GSK-3 inhibitor on body weight, bone mass, and histomorphometric indices of bone formation/resorption
We next explored the in vivo effects of 603281-31-8 on body weight, bone mass, and indices of bone formation/resorption. We confirmed that OVX increased body weight above Sham levels by 1 month after surgery.(32) Treatment with GSK-3 inhibitor in OVX rats for 2 months did not significantly decrease the body weights (Table 1). DXA analysis of whole femora, distal femora, and midshaft showed significant differences in BMD of Sham and OVX rats (Fig. 4). OVX for 12 weeks caused significant decreases in the BMD at whole femur and distal femur (10–15%). After treatment with 603281-31-8 at a dose of 3 mg/kg body weight/day for 2 months, there was a significant increase in BMD at the whole femur, midshaft, and distal femora compared with OVX rats (p < 0.05). Furthermore, significant beneficial bone effects of the compound were observed even in the absence of differences between Sham and OVX at the midshaft. Histomorphometric analysis showed that OVX induced significant decreases in trabecular bone area, trabecular number, and mineral apposition rate and increases in trabecular separation, mineralizing surface, and BFR/BS compared with Sham controls in the proximal tibial metaphysis (Table 2). Treatment with 603281-31-8 at 3 mg/kg significantly increased trabecular bone area, thickness, and number and mineral apposition rate and decreased the trabecular separation compared with OVX controls. Similarly, in the tibial midshaft (Table 3) OVX significantly decreased the cortical area and increased the marrow area compared with Sham controls, whereas 603281-31-8 increased the cortical area and decreased the percentage of marrow area compared with OVX controls. 603281-31-8 significantly increased the periosteal labeled surface and periosteal BFR/BS even in the absence of differences between Sham and OVX. Additionally, resorption indices (osteoclast number and eroded surface) were significantly increased in OVX animals compared with Sham controls (Tables 2 and 3). Treatment with 603281-31-8 decreased the percent of eroded surface (35%) compared with OVX but did not significantly alter osteoclast number (Table 2). Together, these results suggest that GSK-3 inhibitor increased bone formation in OVX rats with a mild effect on resorption.
Table Table 1.. Body Weights of Rats Treated with GSK-3 Inhibitor
Table Table 2.. Histomorphometric Parameters of Proximal Tibial Metaphyseal Cancellous Bone
Table Table 3.. Histomorphometric Parameters of Tibial Midshaft
GSK-3 inhibitor improved vertebral strength in OVX rats
Biomechanical analysis of the lumbar vertebra L5 showed that OVX decreased the vertebral strength by 16% and reduced energy (work to failure) by 28% relative to Sham controls (Table 4). 603281-31-8 restored biomechanical properties to Sham levels by significantly increasing peak load, energy, and stiffness relative to OVX by 23%, 30%, and 14%, respectively. These data show that 603281-31-8 functionally improves trabecular bone quality as measured by vertebral strength (peak load) and toughness.
Table Table 4.. Mechanical Test Analyses of L5 Vertebrae
GSK-3 inhibitor increased expression levels of osteoblast-phenotype/bone formation markers in rat bone
We examined the effects of 603281-31-8 on the mRNA levels of established markers of bone formation [alkaline phosphatase, biglycan, osteocalcin, BSP, collagen α1 (I), collagen α1 (V), osteonectin, procollagen C-proteinase enhancer, runx2, and resorption (OPG and RANKL)]. The expression levels were determined in distal femur metaphyseal RNA by real-time PCR after treatment of OVX rats with 603281-31-8 for 2 months. As shown in Fig. 5A, there was a significant increase in the expression levels of biglycan, collagen α1 (I), collagen α1 (V), osteonectin, and runx2 compared with OVX rats. There was a significant decrease in the OPG/RANKL ratio in OVX rats compared with Sham (Fig. 5B), consistent with the increase in osteoclast number as shown in Table 2. Treatment with 603281-31-8 resulted in nonsignificant increase in OPG/RANKL ratio compared with OVX (Fig. 5B). These results suggest that the increased BMD observed with GSK-3 inhibitor in OVX rats is likely mediated by an increase in bone formation with a subtle effect on resorption.
The important role of the canonical Wnt signaling pathway in bone raised the possibility that GSK-3 (a key protein kinase) might be a suitable target for enhancing bone formation. In this hypothesis, the inhibition of GSK-3 would stabilize β-catenin, leading to accumulation and translocation to the nucleus(33) and activating downstream Wnt pathway targets, resulting in bone formation. In this study, we directly tested this hypothesis and showed that a novel, orally bioavailable small molecule, GSK-3 α/β dual inhibitor, increases Wnt signaling, markers of cellular (osteoblast) differentiation in vitro, and bone mass and strength in vivo. These findings confirm the role of the Wnt signaling pathway in bone and suggest that potent and selective inhibitors of GSK-3 could have use as bone anabolic agents.
A number of GSK-3 inhibitors have been reported, including lithium, synthetic peptides such as hymenialdisene, paullones, indirubins, and maleimide compounds that display varying degrees of GSK-3 inhibition.(34–36) However, many of these compounds inhibit protein kinases other than GSK-3. The GSK-3 inhibitor used in this study is efficacious and shows a 500-fold selectivity window for GSK-3β over a panel of other kinases.(26) Our results showed in vivo efficacy of the GSK-3 inhibitor in a rat delayed-dosing intervention model. The anabolic effects of 603281-31-8 are in part explained by both cellular and molecular findings in vitro and in vivo. Our results showed that inhibition of GSK-3 by this compound increased the levels of β-catenin, and further downstream, activated the TCF/LEF artificial reporter system in pluripotent stem cell like C3H10T1/2 cells. This was accompanied by an increase in key markers of osteoblast differentiation/function, therefore making the use of these multipotent cells ideal to carry out a limited study of the effect of GSK inhibition and Wnt signal activation on proliferation, commitment, and differentiation. These murine mesenchymal cells have also been studied by others for effects of Wnt stimulation and LiCl inhibition of GSK.(22,36) Our results showed a decrease in proliferation, a biphasic increase in osteoblast markers, and a decrease in adipogenic markers, suggesting enhanced osteoblast commitment and differentiation rather than expansion of a progenitor population (proliferation) and are consistent with findings of others with Wnt stimulation.(15,23,37) The reason for the biphasic effect on osteoblast markers is not clear but a similar observation has been made with PTH in studies of primary calvarial osteoblasts.(38) We speculate that this may represent an initial suppression of osteoblast differentiation, maintenance of progenitors, and reprogramming to promote commitment to the osteoblast lineage. Overall, our findings confirm the previous reports supporting the notion that β-catenin is directly involved in controlling bone mass. Either a stabilized β-catenin mutant or treatment with Wnts that stabilize β-catenin may result in the induction of alkaline phosphatase and osteogenic differentiation in undifferentiated mesenchymal cells.(22,23) Furthermore, β-catenin signaling predisposes mesenchymal precursors to undergo osteoblast commitment in response to osteogenic cues.(24)
In addition to the focus on the osteoblast lineage and formation effects of Wnt signaling, two recent studies of genetically engineered mice showed a profound effect of β-catenin on resorption through stimulation of production of OPG, a major inhibitor of osteoclast differentiation.(25,39) In this study, we observed a trend toward a decrease (15%) in osteoclast number (nonsignificant) and a small but significant decrease (35%) in the eroded endocortical surface. There seemed to be a trend toward a small increase in OPG/RANKL ratio, but this effect was not significant. Our results show a strong coherence between histomorphometric and molecular measurements as indicators of osteoblast and osteoclast function. The pharmacological activation of the Wnt pathway in OVX rats (osteopenia model) through inhibition of GSK-3 has a primary and robust effect on bone formation and perhaps a small effect on osteoclasts. Despite differences between the genetic models(25,39) and our observations with pharmacological inhibitors in OVX model, a unifying role of Wnt pathway on osteoblasts and osteoclasts is indeed plausible. The high bone mass phenotype in human subjects with a G171V mutation in LRP5 is the result of increased osteoblast activity, with normal resorption parameters.(16) On the other hand, patients with autosomal dominant osteopetrosis type I (ADOI) caused by T253I mutation in LRP5 manifest an osteopetrotic phenotype, with lower numbers of osteoclasts in vivo,(40,41) but show no increase in bone formation markers.(42) This thus identifies different outcomes from two separate mutations in LRP5. The G171V mutation leads to increased bone through an osteoblast and bone formation effect, and the T253I mutation does so by reducing osteoclast activity. It was noted that serum OPG levels were 40% higher (nonsignificant) in T253I mutation patients.(43) It therefore seems that, depending on the nature and location of mutations in LRP5 (and perhaps the nature or mode of activation of the Wnt signaling), there can be varying outcomes on the osteoblast and osteoclast function. The cell signaling and transcription consequences differ in ways that might not be readily revealed by these methods. The two mutation sites on LRP5 are in different parts of the YWTD domain of the first propeller structure.(18) Although DKK-1 inhibition of LRP binding and activity results from the G171V mutation,(16) the consequences for LRP5/6 binding of the T253I mutation are not known. Given the complexity of the Wnt system, with its receptor complex of inhibitory and stimulatory components, its decoy receptors, the multiprotein inhibitory complex acting on GSK-3, and the necessary phosphorylation steps, it should perhaps be expected that subtly different mechanisms will emerge for the control of activation. Combined input from biochemical studies and mouse and human genetics has provided some important new clues.
Many previous findings have shown that osteogenic differentiation (and bone formation) are associated with expression of osteoblast phenotypic markers such as osteocalcin, osteopontin, collagen α1 (I), and alkaline phosphatase among others.(16,22–24) As expected the anabolic effect of the GSK-3 inhibitor in C3H10T1/2 cells was associated with increased expression of mRNA for collagen α1 (I) (2-fold), BSP (1.4-fold), osteocalcin (1.2-fold), runx2 (1.6-fold), collagen α1 (V) (1.5-fold), and alkaline phosphatase (∼160-fold) and a parallel increase in alkaline phosphatase activity. Interestingly, the effect on osteocalcin (1.2-fold) was rather marginal. Similarly, in vivo treatment with GSK-3 inhibitor increased the expression of some key osteoblast markers [biglycan, collagen α1 (I), collagen α1 (V), osteonectin, and runx2] but not osteocalcin and BSP. Although there is concordance in the direction of change in a majority of the markers in vivo and in vitro, the differential response with a few of the markers could be explained by the fact that the in vivo RNA data were obtained from whole metaphyseal bone after 2 months of treatment. This is supported by a very robust increase in biglycan, collagen α1 (I), collagen α1 (V), osteonectin, osteocalcin, BSP, and alkaline phosphatase in our preliminary studies of intact rats treated with GSK-3 inhibitor for 7 days (data not shown). Thus, future time-course studies after vivo treatment will help clarify this observation.
GSK-3 is involved in several signaling pathways such as the classical MAPK cascade, the phosphatidylinositol-3,4,5-trisphosphate [Ptdins(3,4,5)P3]-dependent, cAMP-dependent protein kinase A (PKA), and canonical β-catenin signaling pathways. Growth factors, insulin, amino acids, PKA agonists, and lithium chloride, among others, can inhibit GSK-3 activity by phosphorylating serine 21 (Ser21) in GSK-3α and Ser9 in GSK-3β near the amino terminus.(44) The canonical β-catenin signaling pathway can be stimulated in response to Wnt-1, 2, 3A, 8, and 8B or by inhibition of GSK-3.(3) Ectopic expression of β-catenin or activation of endogenous β-catenin by lithium chloride treatment has been shown to induce bone morphogenetic protein (BMP2)-mediated early osteoblast differentiation.(22) Our data provide in vitro and in vivo evidence that inhibition of GSK-3 stimulates the β-catenin signaling pathway to induce bone formation; however, the role of other signaling pathways that involves GSK-3 in bone formation is currently unknown.
In summary, we showed that, in undifferentiated C3H10T1/2 cells, inhibition of GSK-3 was associated with an increase in the levels of β-catenin, an increase in functional activation of downstream Wnt reporter gene, an increase in the expression levels of osteoblast and biochemical markers of bone formation, and a decrease in adipogenic markers and cell proliferation. In vivo, treatment of OVX rats with GSK-3 inhibitor increased bone mass, strength, and expression levels of markers of osteoblastic differentiation/function. Collectively, these data suggest that GSK-3 inhibitors constitute a promising family of compounds that can be used as potential therapeutic agents for treatment of osteoporosis.
The authors thank Dr Rachelle Galvin for critical review of the manuscript and for valuable suggestions.