Does simvastatin stimulate bone formation in vivo?

Background Statins, potent compounds that inhibit cholesterol synthesis in the liver have been reported to induce bone formation, both in tissue culture and in rats and mice. To re-examine potential anabolic effects of statins on bone formation, we compared the activity of simvastatin (SVS) to the known anabolic effects of PTH in an established model of ovariectomized (OVX) Swiss-Webster mice. Methods Mice were ovariectomized at 12 weeks of age (T0), remained untreated for 5 weeks to allow development of osteopenia (T5), followed by treatment for 8 weeks (T13). Whole, trabecular and cortical femoral bone was analyzed by micro-computed tomography (micro CT). Liquid chromatography/mass spectrometry (LC/MS) was used to detect the presence of SVS and its active metabolite, simvastatin β-hydroxy acid (SVS-OH) in the mouse serum. Results Trabecular BV/TV at T13 was 4.2 fold higher in animals treated with PTH (80 micro-g/kg/day) compared to the OVX-vehicle treated group (p < 0.001). However, the same comparison for the SVS-treated group (10 mg/kg/day administered by gavage) showed no significant difference (p = NS). LC/MS detected SVS and SVS-OH in mouse serum 20 minutes after gavage of 100 mg SVS. A serum osteocalcin assay (OC) demonstrated that neither bone formation nor osteoblast activity is significantly enhanced by SVS treatment in this in vivo study. Conclusions While PTH demonstrated the expected anabolic effect on bone, SVS failed to stimulate bone formation, despite our verification by LC/MS of the active SVS-OH metabolite in mouse serum. While statins have clear effects on bone formation in vitro, the formulation of existing 'liver-targeted' statins requires further refinement for efficacy in vivo.


Background
Statins, potent compounds that inhibit cholesterol synthesis in the liver by blocking 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA), have been reported to induce osteoblast activity and lead to bone formation, both in tissue culture and in rats and mice [1][2][3][4][5][6]. These findings, if confirmed and extended to include humans, have the potential to introduce immediate new treatment options to patients with low bone mass and osteoporosis. Such treatment is readily available, since an estimated 3 million patients in the US alone take statins daily to treat hyperlipidemia with few side effects [7]. Osteoporosis is increasingly recognized as an eminent public health problem, as the mean age of populations in most Western countries increases and costs related to osteoporosis are expected to grow six-fold by the year 2040 [8]. Data from the National Health and Nutrition Examination Survey, using a definition of osteoporosis developed by the World Health Organization, estimate that in the United States up to 20% of Caucasian women over 50 have osteoporosis and up to 50% have low bone mass. Non-caucasian women and men tend to have lower rates of osteoporosis but contribute up to 25% of the fractures annually [9].
The only bone anabolic agent known to increase bone mineral density in both animal models and clinical settings is human parathyroid hormone 1-34 (PTH), which has recently been approved by the FDA in patients with advance osteoporosis [10]. Bone-related basic and clinical research has been searching intensively for alternative anabolic agents for the treatment of osteoporosis.
Due to the potential clinical impact of statins as a treatment option for patients with low bone mass and osteoporosis, it is essential to evaluate its anabolic activity in several in vitro and experimental animal models. We have recently reported the anabolic action of PTH in an ovariectomized (OVX) Swiss-Webster mouse model, which like the rat model, displays a substantial amount of trabecular bone in its femoral metaphyses and a quantifiable loss of bone following OVX [11]. The aim of this study was to investigate the effect of simvastatin on the microarchitecture of bone in an established mouse model of osteoporosis, and compare its anabolic action with PTH. We also verified the presence of systemically administered simvastatin in the mouse serum, as this has previously not been reported and is likely to be of critical importance when evaluating the effect of simvastatin on bone.

Methods
Animals 72 eleven-week-old virgin female Swiss-Webster mice were purchased from Taconic Farms (Germantown, NY) and maintained at the animal research facility at the Beth Israel Deaconess Medical Center. Animals were fed Purina Formulab Diet containing 1% Ca (Formulab, Richmond, IN) and water ad libitum throughout the experiment. Mice were sacrificed by CO 2 inhalation.

Experimental protocol
The experimental protocol was approved by the Institutional Animal Care and Use Committee of the Beth Israel Deaconess Medical Center. The study design is shown schematically in Figure 1. The mice were randomly subdivided into five groups of bilaterally ovariectomized (OVX) and three groups of SHAM OVX animals, with 9 animals in each group. A group of each of OVX and SHAM animals was sacrificed one week postoperatively (T1/OVX and T1/ SHAM, respectively) and served as baseline controls. Four additional weeks were allowed to pass before initiation of treatment in the remaining groups in order to permit significant bone loss to occur in the OVX animals. At this time, one group of OVX (T5/OVX) and one group of sham OVX mice (T5/SHAM) were sacrificed to evaluate pretreatment bone loss.
An eight-week daily treatment (5 days a week) consisting of daily gavage of either 10 mg/kg of simvastatin (Apin Chemical Limited UK, Oxfordshire, England) or vehicle (VEH) only (0.5% methylcellulose in H 2 O) was administered to the T13/OVX/SVS and T13/OVX/VEH groups, respectively.
An additional control group of OVX mice received daily treatment (5 days a week) consisting of subcutaneous (sc) injections of 80 µg/kg per day of hPTH (1-34) (Advanced ChemTech, Louisville, KY, USA). All groups were sacrificed after completion of the 13 week study.

Micro-computed tomography analysis
For a detailed qualitative and quantitative 3-D evaluation, whole femoral bones were examined by a desktop µCT system (µCT 20, Scanco Medical AG, Bassersdorf, Switzerland) equipped with a 10 µm focal spot microfocus X-ray tube as a source [12]. For image acquisition, the specimen was mounted on a turntable shifted automatically in an axial direction over 216° (180° plus half fan angle on either side), taking 600 projections. To scan the entire femoral width (3.4-5.1 mm), including the femoral head, a total of 100-150 micro-tomographic slices were acquired at a 34 µm slice increment. CT images were reconstructed in 512 × 512 pixel matrices using a standard convolutionbackprojection procedure with a Shepp and Logan filter. Images were stored in 3-D arrays with an isotropic voxel size of 34 µm. A constrained 3-D Gaussian filter was used to partly suppress the noise in the volumes and the mineralized tissue was segmented from soft tissues by a global thresholding procedure [13]. Morphometric parameters were determined using a direct 3-D approach [14] in three different pre-selected analysis regions: whole bone (including the articular ends), secondary spongiosa in the distal metaphysis, and diaphyseal cortical bone. Parameters determined in the metaphyseal trabecular bone included bone volume density (BV/TV), bone surface density (BS/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp) and connectivity density. Cortical volume (TV), percent marrow volume (%MV) and cortical thickness (Ct.Th) were assessed in a 1 mm-thick slab in the mid-diaphysis.

Sample preparation
All sample preparation procedures were carried out in an ice-water bath. Stock solution of simvastatin was prepared in ethanol and of simvastatin hydroxy acid in methanolwater (90:10) at 4°C. Calibration standards were prepared fresh for each analysis, by dilutions in methanolwater (70:30). The doses of the standards at the respective points on the calibration graphs were 1, 10, 30, 60, 100, 150 ng. Recovery of simvastatin and of simavastatin hydroxy acid from mouse plasma following the extraction procedure below, was assessed by comparison of samples at three doses (30, 80 and 150 ng), which were used to spike 0.1 ml mouse plasma followed by extraction and injection into LC-MS, to similar doses directly injected to the LC-MS. The quantitation was based on integration of the area under the respective peaks in the UV spectra obtained at 240 nm (approximately the λ max for both simvastatin and simavastatin hydroxy acid). Each of the samples from the in vivo tests (0.15 ml), or the recovery test above, was diluted with 0.4 ml of water prior to extraction.

Extraction procedure
This procedure was based on the previously published protocol by Wu et al. [15], implementing minor modifications. Each solid-phase extraction cartridge (1 ml/100 mg) was conditioned with Methanol-water (50:50). The diluted sample (see Sample preparation) was loaded on to each cartridge, followed by low vacuum (1 min), and each cartridge was subsequently washed with 1 ml H 2 O, 1 ml 5% formic acid, 1 ml H 2 O and dried for 1 min. Each cartridge was then eluted with 1 ml of methanol-water (70:30) solution and with 1 ml of acetonitrile, followed by low vacuum (1 min). The combined eluates were evaporated to dryness in a SpeedVac at 45°C. The dry residues were reconstituted in 100 µl of ammonium acetate (10 mm, pH 4.0)-acetonitrile (30:70). The reconstituted extracts were transferred into autosampler vials and 50 µl of each sample were injected on to the LC column (equivalent to 75 µl of plasma sample).  T0  T1  T2  T3  T4  T5  Osteocalcin assay Blood was sampled prior to sacrifice to determine serum levels of osteocalcin (OC). Serum intact OC was measured by a two-site immunoradiometric assay (IRMA kit, Immutopics Inc., San Clemente, CA) against the mid-region C-terminal and amino-terminal position of mouse OC with a known sensitivity of 0.1 ng/mL. Intra-and interassay CVs were 2.3% and 4.4%, respectively.

Statistical Analysis
A power analysis was conducted in planning the experiments to determine the sufficient number of mouse femurs in this investigation. A total sample size of 72 animals (n = 9 per treatment condition) was required to provide 90% power (α = 0.05, β = 0.1) to detect a significant difference of 1 standard deviation (effect size = 1.0) between the Simvastatin, PTH, OVX, and SHAM treatments with respect to each trabecular and cortical bone parameter using one-way analysis of variance (ANOVA) with a two-tailed significance level of 0.05. Sample size calculations were determined using the nQuery Advisor software program (version 4.0, Statistical Solutions, Boston, MA).
All continuous variables were checked for normality using the Kolmogorov-Smirnov goodness-of-fit test and were found to follow a normal distribution closely [16]. Therefore, results are reported in terms of means ± standard deviation (SD). Treatment groups were compared with respect to trabecular and cortical bone parameters obtained by µCT by use of analysis of variance (ANOVA) with the Fisher least significant difference (LSD) method for post-hoc multiple testing [17]. A two-tailed value of p < 0.05 was established as the criterion for statistical significance. Statistical analysis was performed using the SPSS (version 11.0, SPSS Inc., Chicago, IL) and SAS (version 6.12, SAS Institute, Cary, NC) software packages.

Effects of OVX on bone loss
Consistent with the findings in our recent study [11], µCT analyses of femoral trabecular and cortical bone showed little change over the duration of the study in any of the structural variables for SHAM OVX Swiss-Webster mice (groups T1/SHAM, T5/SHAM, T13/SHAM). OVX mice demonstrated significantly lower trabecular bone volume density (BV/TV) compared to SHAM controls at 1 week (15 ± 6% vs. 26 ± 5%, P < 0.001) 5 weeks (5 ± 3% vs. 28 ± 12%, P < 0.001), and showed a reduction of 85% at 13 weeks (5 ± 4% vs. 34 ± 11%, P < 0.001, Table 1). In addition, trabecular thickness, spacing and number were significantly different in OVX mice compared to SHAM controls at 5 and 13 weeks (all P < 0.001). However, cortical parameters including total volume, marrow volume/ total volume, and thickness were not significantly different between OVX and SHAM controls throughout the 13week study course (all P > 0.20, Table 1).

Effect of PTH and SVS treatment on bone
At 13 weeks, trabecular bone volume density (BV/TV) in mice treated with PTH alone (T13/OVX/PTH) was 4.2fold higher than vehicle-treated animals (T13/OVX/VEH) ( Table 2). PTH treatment was associated with significant increases in Tb.Th and Tb.N as well as an expected decrease in Tb.Sp. In addition, significant differences in cortical %MV/TV and thickness were observed at 13 weeks (P < 0.01, Table 2).
In comparison, trabecular BV/TV in SVS-treated animals at 13 weeks (T13/OVX/SVS) was not significantly increased over that observed in vehicle-treated animals (T13/OVX/ VEH) ( Figure 2). Furthermore, none of the other trabecular bone parameters (Tb.Th, Tb.Sp and Tb.N) revealed a significant difference after 13 weeks of SVS treatment when compared to T13/OVX/VEH animals, (all P = NS, Table 2).

Detection of the active Simvastatin hydroxy acid metabolite in mouse serum after gavage administration
In this study, SVS was administered by gavage, and requires hepatic conversion to the metabolically active β-hydroxy acid (SVS-OH). Because SVS failed to have any effect on any static µCT parameter measured at either site in mouse femur, we conducted separate experiments designed to detect the presence of SVS-OH in mouse serum after SVS gavage by liquid chromatography/mass spectrometry (LC/MS).
Reverse-phase high performance liquid chromatography (RP-HPLC) tandem electrospray ionization mass spectrometry (ESI-MS) method published by Wu et al. [15] was used to analyze samples of mouse serum for the presence of SVS and SVS-OH. The identity of the eluted peak in RP-HPLC was established by the determination of the mass associated with it by ESI-MS. The retention times for SVS and SVS-OH were 19.5 ± 0.5 and 15.35 ± 0.20 min, respectively. Threshold of detection on the LC-MS at λ 240 nm (AUFS = 0.1), the most sensitive wavelength, for both SVS and the SVS-OH was ~1 ng. Calibration curves for both the SVS and the SVS-OH extracted from the area under the curve in the HPLC profile was linear in the range of 1-150 ng. These curves were used to quantitate the drug and metabolite levels in mouse plasma. Based on the spiking of mouse plasma ex-vivo with known amounts of drug and subjecting the samples for the extraction procedure, the efficiency of drugs recovery was estimated to be 60-75% for SVS and 70-85% for SVS-OH. Figure 4 is a representative example of the LC-MS analysis of a sample collected from a mouse 30 min after treatment with 100 mg/ kg SVS. In this sample both the RP-HPLC tracing and the ESI-MS analysis of the peaks corresponding to the drug and its metabolite confirm the presence of both in the serum.
Both residual SVS and its in situ-generated metabolite SVS-OH were detected in the mouse plasma as early as 10 min after gavage administration of 100 mg/kg of SVS (the shortest sampling interval). The levels of SVS and SVS-OH in the plasma 10 min after administration were 97.3 and 15.3 ng/µl, respectively. These findings establish the bioavailability of the drug following this route of administration.

Discussion
The aim of this study was to investigate the effect of Simvastatin on bone in an established mouse model of oste- oporosis. This Swiss-Webster outbred mouse strain utilized in this study has proven to be a valuable model for investigating the efficacy of both anabolic and anti-resorptive therapies for the treatment of osteoporosis. Here, we have used PTH as a positive control [11], and compared its effects to SVS. While PTH demonstrated the expected anabolic effect on bone, SVS failed to stimulate bone formation, despite our verification by LC/MS of the active SVS-OH metabolite in mouse serum.
The SVS-OH metabolite is first detectable in mouse serum by 10 minutes after gavage administration, and serum levels of SVS-OH continued to be observed by both massspectrometric and UV measurements through the last serum samples collected at 50 minutes post-gavage. Other experiments failed to detect SVS-OH at longer post-gavage timepoints of 3 and 24 hours. These data are in general agreement with studies in both rats and canines that calculate peak serum levels of SVS-OH at 40-60 minutes after gavage administration [18]. Thus, although we failed to document any bone anabolic activity of SVS in a mouse model of OVX-induced osteoporosis, LC/MS analysis of the SVS-OH metabolite in mouse serum document the efficacy of our gavage administration of SVS to experimental mice.
Extrapolating from mice to humans, the dose of SVS administered in this study was approximately 400-fold higher than the average dose given to humans. The results suggest that Simvastatin does not enhance bone formation in this mouse model in vivo, despite our ability to verify the presence of the SVS-OH metabolite in mouse serum by LC/MS analysis. Furthermore, the osteocalcin as- 34 ± ± ± ± 11 5 ± ± ± ± 4 5 ± ± ± ± 4 21 ± ± ± ± 11 say did not suggest any increase in osteoblast activity in SVS treated mice, consistent with its failure to increase bone density.
It is generally accepted that the hydroxy acid form of statins are potent inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, an enzyme that plays a critical role in cholesterol metabolism, where they block substrate accessibility to HMG-CoA reductase, effectively subverting cholesterol levels of HDL [19][20][21]. In addition to lowering cholesterol, statins seem to have a number of additional effects, such as the nitric oxide-mediated promotion of new blood vessel growth [22], protection against oxidative modification of low-density lipoprotein, as well as anti-inflammatory effects and a reduction in Creactive protein levels [23]. All statins limit cholesterol bi-osynthesis by inhibiting the committed step in the biosynthesis of isoprenoids and sterols [24].
Considering the enormous potential impact of statins on the treatment of osteoporosis, several groups have further investigated the skeletal effects of statin use. Several groups have failed to confirm the initially reported anabolic effect of systemically administered statins in vivo in rats [25][26][27]. One obvious criticism of the negative data in these studies is that they did not verify the actual presence of statin in the animal serum. This limitation may be of importance, as Crawford et al [27] could detect an increase in cortical bone when statin was administered locally. Our study demonstrated the presence of systemically administered the SVS-OH metabolite in the mouse serum by LC/MS analysis.

OVX/vehicle OVX/PTH OVX/SVS
Osteocalcin ng/ml * The exact mechanism by which statins perform their effects on bone cell function is a central issue. It has been hypothesized that statins mediate their effects by increasing the expression in bone of the growth factor bone morphogenic protein-2 (BMP-2) which in turn leads to osteoblast differentiation and bone formation [2]. This mechanism of action has been addressed by Baumann et al. in rats, where the anabolic effect of lovastatin on bone was at least partially due to their inhibiting effect on osteoclast development [28]. Similarly, in the mouse calvaria model Gasser et al. failed to detect any anabolic activity of statins in bone and suggested that the disruption of the capillary integrity and local bleeding might explain some of the previously reported bone responses [26].
Several studies in humans reported a decreased risk of fracture after statin treatment [3,29,30]. However larger human studies failed to confirm statins beneficial effect on bone. LaCroix et al. found the risk of hip and arm fracture to be similar among women who used statins and those who did not take lipid lowering medication [31]. Van Staa et al. examining 218,062 individuals with a fracture in the General Practice Database in the United Kingdom, and found no relationship between statin use and nonspine fracture [32]. Interestingly, a report by Meier [3] had used the exact same database as van Staa et al. [32], but sampled a much larger cohort (218,062 individuals with fracture) than Meier (3940), and in addition, selected an age and gender matched control patient for every fracture patient. The clinical data available at the moment does not allow the conclusion that statin treatment has a postive relationship with fracture risk.

Figure 4
Liquid chromatography/mass spectrometry (LC/MS) for SVS (M + Na + = 442 kD) and SVS-OH (M + Na + = 460 kD) of mouse serum 30 min after gavage with 100 mg/kg SVS. Similarly as described in the clinical trials, the early animal studies reported data supportive of the initial findings [1,[4][5][6]33,34]. These results however, were not substantiated in later studies in rodents, where SVS had no or little effect on bone mineral density [26]. However, none of the studies has verified the actual presence of the active SVS-OH metabolite in the animal serum, as shown in this study. In combination with the osteocalcin data, this report provides strong evidence that simvastatin does not have an effect on bone in an established Swiss Webster mouse model of OVX-induced osteoporosis.

Conclusion
In summary, the currently available data on the effects of statins on bone has done little to confirm the finding that statin therapy increases bone mass and thereby lowers fracture risk. Few of the studies in animals and humans have been able to show a predictable beneficial effect on bone mass and bone turnover. This report verifies the presence of the active SVS-OH metabolite in the mouse serum yet failed to demonstrate any beneficial effect on bone. While statins have clear effects on bone formation in vitro, the formulation of existing 'liver-targeted' statins may require further refinement for efficacy on bone formation in vivo.