Introduction
Bisphosphonate is one of the most effective drugs currently available for suppressing bone resorption. Naturally, combination therapies with other antiresorptive or formative agents have been investigated: PTH [
1‐
3], vitamin D [
2,
4], estrogen [
5‐
7], and other agents [
8]. Risedronate, a pyridinyl (amino) bisphosphonate, significantly reduces the risk of hip fracture among elderly women with confirmed osteoporosis and if combined with estrogen or raloxifene, produces greater gains in bone mass in comparison to single-agent treatment [
9]. Oral administration or intake from food of vitamin K
2, on the other hand, has been shown to prevent the occurrence of fractures in Japanese women [
10,
11] and was reported to prevent bone loss partly through the improved bone formation in animal studies [
12]. It was also reported that vitamin K
2 (MK-4) inhibited bone resorption [
13]. A more recent cellular study reported that the osteoprotective action of vitamin K
2 is through steroid and xenobiotic receptor (SXR)/pregnane X receptor (PXR)-modulated Msx2 gene transcription [
14]. Such an entirely different pharmacological action of vitamin K
2 from other drugs would make it worth studying combinatory administration with bisphosphonate. Limited reports of trabecular bone implied that the combined treatment is more efficacious in osteoporotic rats [
15,
16], while others have reported otherwise [
17,
18]. Therefore, the efficacy of their combinatory use was further investigated in ovariectomized (OVX) ICR mice to clarify the effect on the cortical bone and strength. We tried to separate the effect of vitamin K
2 on matrix from that on mineral and to compare with the effect of risedronate by lowering the experimental vitamin K
2 intake level to ~100 μg MK-4/kg/day, which is at the dietary level.
Materials and methods
Experimental animals
The Animal Care Committee of Kanagawa Dental College approved the entire experimental protocol. Nine-week-old ICR mice were purchased from Japan Clea (Tokyo, Japan). All animals were kept under local vivarium conditions (temperature 23.3°C, humidity 55% and a 12-h on/off light cycle).
Sixteen-week-treatment experiment
Fifty-nine, 9-week-old, female ICR mice were either ovariectomized (n = 43) or sham-operated (n = 8). After a month, during which all mice were fed with conventional rodent food pellets, the ovariectomized mice were divided into six groups. In addition to the OVX group (n = 8), five groups (n = 7) received medication, which was switched at the 8-week midpoint. In the K to R group, mice were treated with MK-4 for 8 weeks and then with risedronate for eight more weeks. R to K mice were treated with risedronate first and then with vitamin K2. K to WO, R to WO, and R/K to WO mice received either vitamin K2, risedronate, or both for 8 weeks, and then the drug(s) was withdrawn. Except in the OVX groups during K and R/K period, which received pellets containing 50 μg/100 g vitamin K2 (MK-4), all animals received the conventional rodent food. Both the conventional rodent pellets (CE-2) and the vitamin K2 pellets were prepared by Japan Clea with MK-4 kindly provided by Eizai (Tokyo, Japan). Calculated from the average 6 to 7 g a day consumption of the ration, the pellets were prepared so that the animals received ~100 μg MK-4/kg/day, which is at a dietary level. During the R or R/K period, mice received 0.25 mg/kg of daily oral risedronate after 2-h fasting. They were fed after another 2-h fasting. The femurs were excised from mice euthanized after the 16-week therapeutic period and were preserved at −80°C for microfocused X-ray computed tomography (micro-CT) and peripheral quantitative computed tomography (pQCT) analyses and confocal Raman spectroscopy.
Eight-week-midpoint experiment
To examine femurs at the 8-week midpoint, total of 42 ICR mice were treated under similar conditions to those of the 16-week experimental animals. They underwent either sham surgery (n = 9) or an ovariectomy (n = 33). OVX groups include control OVX (OVX, n = 9), OVX treated with risedronate (OVX-R, n = 8) or vitamin K2 (OVX-K, n = 8), and the concomitant administration (OVX-R/K, n = 8).
Microfocused X-ray computed tomography
Using MCT-CB 130F (Hitachi Medico, Tokyo, Japan), three-dimensional imaging data of the distal epiphyseal region of the femur, between 1.5 to 2.75 mm proximal to the growth plate, were obtained. The spatial resolution was set to 7 µm with the voxel size of 17.8 × 17.8 × 17.8 (µm), and the tube voltage and current were 60 kV and 100 µA, respectively. The resolution was set to medium (200 projections each), and slice thickness and increment were set to 20 µm. A morphological analysis was carried out using TRI 3D BONE (Ratoc System Engineering, Tokyo) for such parameters as BV (mm
3), bone volume; BS (mm
2), bone surface; BV/TV (%), bone volume fraction; Tb.Th (μm), trabecular thickness; Tb.N (1/mm), trabecular number; Tb.Sp (μm), trabecular separation; Tb.Spac (μm), trabecular Space; FD, fractal dimension [
19]; and structural model index, SMI [
20].
Peripheral quantitative computed tomography
The distal metaphysis, 1.4 mm proximal to the growth plate and mid-diaphysis of femurs (5 mm proximal to the midpoint), was scanned by a Research SA+ pQCT model (Norland Stratec, Berkenfeld, Germany) with a tube voltage of 50 kV and a tube current of 550 µA using a voxel size of 80 × 80 × 46 (µm). The cortical bone was defined as the area of bone mineral density (BMD) > 690 mg/mm
3, while a threshold of 395 mg/mm
3 at the contour mode 1 was set to define trabecular bone in the bone marrow. Total BMD (mg/cm
3) and the content, BMC (mg/mm), were presented as metaphyseal mineral properties. In addition, the cortical thickness (CTh), cross-sectional moment of inertia (CSMI), and polar stress/strain index (pSSI), an index of strength [
21], were calculated.
Mechanical properties of femurs
The bone strength of the femoral diaphysis and distal epiphysis was evaluated using three-point breaking tests and compression tests using a MZ-500 s device (Maruto, Tokyo, Japan). The crosshead speed in the three-point breaking test and the compression test was 10 and 1.0 mm/min, respectively. In the latter, the distal epiphysis, approximately 3.0 mm thick, was compressed to 1.5 mm. The ultimate load (UL) and stiffness (s) were determined from the load–displacement curve and were converted to the material properties. Ultimate stress (US) was calculated by using the equation US = (UL × d × L)/(8 × CSMI), where d is the diameter at midshaft, and L is the support span at the bottom (10 mm). The elastic modulus, E, was calculated by using the equation E = (s × L
3)/(48 × CSMI).
Confocal Raman spectroscopic measurements
Confocal laser Raman microspectroscopy was used to examine the composition and relative amounts of the mineral and matrix produced in the tibia. Raman spectroscopy is particularly useful for bone analysis because it probes the molecular and ionic vibrations of the mineral and matrix component in unprocessed preparations that preserve these components. The innate, prominent vibrations were measured as described by Tarnowski et al. [
22]. Crystallinity was determined using the method reported by Yerramshetty et al. [
23] as the inverse of the width of the phosphate symmetric stretch band (PO
4
3−
ν
1 at 959 cm
−1) at half the maximum intensity value. A Nicolet Al.mega XR Dispersive Raman microscope system equipped with the OMNIC Atlμs™ imaging software program (Thermo Fisher Scientific, MA, USA), which enable to map a small area less than 1 μm
3 on the bony microsurface of the cortical bone on the video microscope stage control. A high brightness, low-intensity laser operating at 780 nm was used as the excitation source with a laser power of 35 mW. Each spectrum is the sum of ten 10-s measurements. The spectral resolution of the Almega XR under the conditions used was 3.85 cm
−1. For each femur, one averaged Raman image was acquired in the middle of the anterior cortical bone by the ten 10-s measurements.
Statistical analysis
All data values were expressed as the means ± standard deviation (SD). Unless otherwise mentioned, the group means for each parameter were determined for the 8-week midpoint experimental results and compared using a one-way analysis of variance (ANOVA), with the post hoc Tukey–Kramer test. Dunnett's multiple comparisons test was used for 16-week treatment groups with the OVX group as a reference. The probability values of p < 0.05 were considered to be statistically significant for all the comparisons. The Stat View software package (Stat View 5.0; Abacus Concepts, Berkeley, CA, USA) was used for all analyses.
Discussion
Generally, drugs targeting different functions are combined for multidrug therapy with the expectation of complementary action. For vitamin K, however, even the efficacy by itself is still controversial. Earlier, low concentrations of circulating vitamin K have been associated with bone fractures [
24] and with low bone mineral density [
25]. The undercarboxylated osteocalcin was associated with fracture risk [
26,
27], and its reduction by the vitamin K intake was reported without the effect on BMD [
28]. Furthermore, a vitamin K intake level of less than 109 μg/day has been reported to be a risk factor for hip fracture in women [
29]. Shiraki et al. treated postmenopausal patients with 45 mg/day MK-4 and reduced the new fractures to one third. Their lumber BMD was found to be significantly higher than that observed in the control women [
10]. In a more recent study, the combination of alendronate with 45 mg/day MK-4 was reported to be superior to alendronate monothrapy in decreasing undercarboxylated osteocalcin, increasing femoral neck BMD and decreasing the urinary deoxypyridinoline [
30]. In the animal studies, a much higher dosage of 30–50 mg MK-4/kg/day has been used, thus resulting in a significantly higher mineral content in cortical bone without bisphosphonate [
31]. However, the results are inconsistent among different animals or strains [
16‐
18,
32‐
34]. In the present study, we did not observe significant increase in BMD or BMC at the lower level of ~100 μg/kg/day unless MK-4 was followed by risedronate.
Vitamin K
2 has been known to be essential for the γ-carboxylation of osteocalcin [
35]. Therefore, the function was once assumed through activating osteoblasts and leading them to enhanced mineralization [
36]. The mice genetically deficient for osteocalcin, however, exhibited the gain in bone mass instead of loss [
37], suggesting that the osteoprotective action of vitamin K is mediated by some other pathways. Recent reports showed that vitamin K
2 activates osteoblastic transcription of extracellular matrix-related genes [
38] through steroid and xenobiotic receptor (SXR)/pregnane X receptor (PXR)-mediated Msx2 gene transcription in cooperation with the estrogen-bound ERα [
14].
According to the findings of our 8-week administration, only the MK-4 monotherapy at the dietary level resulted in cortical bone matrix formation and maturation without significant increase in BMD or BMC. It was shown that vitamin K
2 not only stimulates cortical bone matrix formation but also accelerates proline hydroxylation, which is a prerequisite for collagen cross-linking to achieve a mature collagenous matrix. Whether the enzymes involved in these processes are the target of vitamin K
2 or not is yet to be resolved. In addition, MK-4 alone provided significant effect in most of the structural parameters of femoral trabecular bone. On the other hand, risedronate, at 0.25 mg//kg/day, was certainly effective, alone or in combination with MK-4, in femoral cortical BMD, BMC, and some trabecular structural parameters in the 8-week treatment. Of note, however, the 8-week concomitant administration was no more effective than each effective monotherapy. This led us to investigate the sequential administration of the two drugs with the same total dosage. The resulting final mechanical properties at 16 weeks were significantly better than the OVX controls only in K to R group. Despite of beneficial effects of 8-week MK-4 pretreatment on the cortical bone matrix, enhanced degradation of collagen seems to occur if the MK-4 is discontinued. In other words, the elevated 16-week mineral/matrix ratio in K to WO is distinct from the lowest 8-week midpoint ratio in the OVX-K. In contrast, the K to R group retained a much lower mineral/matrix ratio at 16 weeks. Since the K to WO mineral value, the numerator, is lower than the K to R value judged from the cortical BMD, the higher mineral/matrix ratio in K to WO was derived from the denominator, the smaller matrix value. It suggests either the collagen degradation or the decreased synthesis after the MK-4 withdrawal. An elevated serum CTx value was observed in K to WO in the later 8 weeks (data not shown). During the later 8-week treatment in K to R, risedronate clearly prevented the increase in CSMI, which occurred in K to WO. The lack of such prevention as well as the lack of other beneficial effects found in K to R cortex, such as the higher/larger BMD, BMC, and thickness, would explain why no significant effect was detectable in K to WO by the three-point bending test. In the MK-4 treated pre-OVX rats, Iwamoto et al. reported the elevated eroded surface as well as the bone formation rate that remained high after the MK-4 withdrawal [
16]. The cellular mechanisms of the elevated collagen degradation therefore have to be confirmed in the future.
In the compression test, the ultimate stress parameter of K to WO as well as of K to R was significantly larger than the OVX control. This result was supported by the significantly better parameters of the trabecular structure in K to WO such as BV, BS, BV/TV, Tb.N, and Tb.Sp in comparison to the OVX controls. No such benefit was observed in R to WO and R/K to WO. The difference in the effect of MK-4 withdrawal on cortical bone and trabecular bone may be related to the distinct distribution of ERα vs. ERβ [
39] or the different ERα signaling pathways [
40], on the assumption that vitamin K and estrogen via the ERα cooperatively promote the osteoblast function through the Msx2 gene induction [
14].
Concomitant administration of risedronate and MK-4 is probably not recommended because R/K to WO was generally not beneficial except in the metaphyseal total BMC. In addition, R to WO but not R/K to WO cortical thickness and BMC are significantly higher at 16 weeks than the OVX control, resulting in the increased ultimate stress only in R to WO. Since OVX-R and OVX-RK at 8 weeks exhibited similar cortical thickness and BMC values, the negative effect of RK withdrawal is apparent. The continuous 16-week administration of risedronate and MK-4 (R/K to R/K) was not beneficial in any parameters tested, including the metaphyseal total BMC (data not shown). Although R to K also showed a significantly positive effect in metaphyseal total BMD and BMC, it is probably not recommended to follow R by K because none of the benefits available in the cortex of R to WO was seen in R to K. In conclusion, prior 8-week treatment with MK-4 followed by the 8-week risedronate prevented the shortcomings of individual drugs and significantly increased femur strength of ICR mice compared to the OVX controls.