- Email: [email protected]
Loading induces site-specific increases in mineral content assessed by microcomputed tomography of the mouse tibia
Bone 36 (2005) 1030 – 1038 www.elsevier.com/locate/bone
Loading induces site-specific increases in mineral content assessed by microcomputed tomography of the mouse tibia J.C. Frittona,b,T, E.R. Myersb, T.M. Wrighta,b, M.C.H. van der Meulena,b b
a Biomedical Engineering, Cornell University, Ithaca, NY 14853, USA Hospital for Special Surgery, Laboratory for Biomedical Mechanics and Materials, Research Division, 535 E. 70th Street, New York, NY 10021, USA
Received 2 September 2004; revised 14 January 2005; accepted 25 February 2005
Abstract Adaptation to mechanical loading has been studied extensively in cortical, but not cancellous bone. However, corticocancellous sites are more relevant to osteoporosis and related fracture risk of the hip and spine. We tested the hypotheses that adaptation in a long bone would be greater at cancellous than cortical sites and would depend on the term of daily in vivo cyclic axial loading. We applied compressive loads to the adolescent, 10-week old, male C57BL/6 mouse tibia to examine the skeletal response immediately prior to attainment of peak bone mass. Adaptation was quantified at the completion of either 2-week (n = 8) or 6-week (n = 12) loading terms by directly comparing volumetric bone mineral content between loaded and contralateral limbs by microcomputed tomography. The increase in mineral content was site specific with a greater response found in the corticocancellous proximal metaphysis (14%) than the cortical mid-shaft (2%) after 6 weeks of loading. Furthermore, bone volume fraction and average trabecular thickness of cancellous bone in the proximal tibia increased after 6 weeks by 15% and 12% respectively. Diaphyseal response was only evident proximal to the mid-shaft as indicated by an 8% increase in maximum principal moment of inertia. Both loading terms produced similar results for mineral content, volume fraction, and moments of inertia. Our finding that non-invasive loading increases the bone volume and fraction at a corticocancellous site by as much as 15% motivates exploring the use of mechanical loading to attain greater peak bone mass and inhibit osteoporosis. D 2005 Elsevier Inc. All rights reserved. Keywords: In vivo mechanical loading; Bone adaptation; Trabecular bone; Mice; MicroCT
Introduction Mechanical loading of the skeleton regulates the attainment of peak bone mass in the growing animal and produces adaptive changes in the adult. Load bearing during growth has a strong influence on the resulting adult structure [1]. Loss of normal load bearing leads to reduced appositional bone growth and bone mass deficit in adults [1 – 4]. Conversely, increased load bearing can result in increased bone mass in adults [5,6]. Gaining insight into this adaptive system will help elucidate basic mechanisms involved in mechanotransduction. * Corresponding author. Hospital for Special Surgery, Laboratory for Biomedical Mechanics, 535 E. 70th Street, New York, NY 10021, USA. Fax: +1 212 606 1490. E-mail address: [email protected] (J.C. Fritton). 8756-3282/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2005.02.013
The factors controlling attainment of peak bone mass in humans are particularly critical at corticocancellous sites, perhaps more so than in the cortical diaphysis. Corticocancellous regions such as the hip, spine, and distal radius experience the greatest adult bone loss and risk of fracture [7]. Furthermore, they may be more responsive to mechanical loads than diaphyseal cortical sites during growth [8]. Therefore, mechanical loading models of corticocancellous sites are required to understand the mechanical and genetic factors that impact fracture risk. One impediment to ascertaining adaptive responses at a corticocancellous site is that the complex mechanical structure exists in a dynamic environment of bone turnover. Previously, cortical adaptive formation responses have been measured with 2D histomorphometry of two or more bone labels given at short intervals within the term of controlled loading [9 – 11]. However, during growth, the cortical
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
adaptive response to loading includes site-specific inhibition of resorption [12 – 16]. At cancellous sites, increased short-term formation rates may not indicate a concomitant increase in bone volume fraction [17]. This lack of bone volume change suggests an increase in both osteoblastic and osteoclastic activity similar to that observed during high turnover. Quantitative microcomputed tomography (microCT) is a 3D technique useful in measuring the net mineral deposited or left unresorbed over the entire term of study. A net mineral measurement is suited for high turnover sites and high periods of turnover such as growth and adaptation. Many studies have focused on older animals to eliminate the potentially confounding effects of growth, but the adolescent animal is critical to delineating the factors that determine peak bone mass [18]. Higher loading magnitudes and rates can result in suppression of longitudinal growth of the rat ulna [13,14,19,20]. Additionally, loading paradigms observed in the adult may not apply to periods of rapid growth and mineral acquisition. Compared to the adult, lower mechanical loading thresholds are sufficient to produce responses in juveniles [21]. We tested the hypothesis that the volume of mineralized tissue would increase in both cortical and cortiocancellous sites and that the increase is site specific and depends on the term of controlled cyclic axial loading applied to the adolescent mouse tibia. Our approach extended that used for axial compression of the ulna [12,16]. The tibia was selected for the distinct advantage of a substantial cancellous bone volume subject to loading. Bone mineral content was compared between loaded and contralateral limbs by using quantitative microCT. We also examined the effects of short-term (2 weeks) and long-term (6 weeks) mechanical loading.
Materials and methods Study design To determine whether bone mass is increased in the adolescent mouse tibia due to mechanical loading, we studied the mineral response to a daily regimen of in vivo, controlled, axial loading in twenty male mice. Cyclic, mechanical loads were imparted with a device specifically designed for the mouse tibia. At the completion of either 2week (n = 8) or 6-week (n = 12) loading terms, bone mineral content and distribution were measured using quantitative microCT in both the loaded and the contralateral, control limb. The number of mice per group was based on a power analysis considered prior to the experiment. With 8 mice per group, there was over 90% power to detect an effect size of 0.9 for the primary factor of interest, loading, using a twofactor repeated-measures design (time as a between-subjects factor and load as a within-subjects factor, alpha = 0.05). We
1031
should, therefore, be able to determine a mean paired difference in mineral content that was approximately 90% of the standard deviation with 8 per group. Four mice were added to the 6-week group to account for potential loss due to repeated use of anesthesia. Power calculations were made using a Repeated Measure AOV panel (PASS 6.0, NCSS, Kaysville, UT). Animals C57BL/6J male mice (B6, Jackson Laboratories, Bar Harbor, ME) were received and acclimated to our facility for 3 weeks. All mice were 10 weeks old (T1 day) at the start of loading. Mice were caged in groups of four and fed ad libitum. The B6 mouse has relatively low peak bone mass and low bone formation rates during growth in the cortices of the femur and tibia and metaphyses of the femur compared with the C3H/HeJ strain [18,22,23]. However, the adult B6 mouse has an enhanced cortical bone formation response to loading [24,25]. Therefore, this strain may prove useful for studying the basis of mechanical loading in the attainment of peak bone mass. Our Institutional Animal Care and Use Committee approved all protocols. Mechanical loading protocol The loading scheme was similar to that of the axial ulna loading model [12,16] but carried out on the tibia with our loading device. Compression was applied at the ends of the left tibia using a small, electro-magnetic actuator (LA18-18000A, BEI Kimco Magnetics, San Marcos, CA). Load transmission was through a brass platen securing the foot and attached to the core of the actuator. Another brass platen attached to the stiff aluminum frame cupped the knee (Fig. 1). The frame was adjustable to accommodate slight differences in limb lengths. Interposed between the frame and the knee platen was a 100 N load cell (ELFS-T3E, Entran, Fairfield, NJ). A pre-load of 0.2 N was applied immediately prior to dynamic compression to maintain the initial position of the limb. Loads applied in this manner induce combined axial compressive and bending strains because of the natural curvature of the tibia.
Fig. 1. Schematic of loading apparatus to apply axial compression to the mouse tibia. Arrows indicate the direction of loading.
1032
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
The waveform was a load ramp of 0.075 s duration followed by a symmetric unload ramp of 0.075 s duration (Fig. 2B). The dwell, or rest period between each load cycle, lasted 0.100 s at the pre-load level of 0.2 N. Loading was applied at 4 cycles/second representing the stride frequency of the mouse [26]. Loading continued for 1200 cycles, yielding a daily (5 days/week) loading period of 5 min. Peak-to-peak load magnitude was 3 N. The actuator was driven with amplified signals (BOV-20-10M amplifier, KEPCO, Flushing, NY) produced by a digital waveform function generator (AFG320, Tektronix, Beaverton, OR). Tibial mid-diaphyseal mechanical strains were recorded from the periosteal medial surface in 10-week old anesthetized mice (n = 3) to relate applied loads to resulting strain. Standard surgical techniques were used to elevate the periosteum and bond a single element, foil gage (CEA015UW-120, Vishay Micro-Measurements, Raleigh, NC) aligned with the long axis. The medial cortex at the midshaft was chosen to minimize soft tissue dissection and to compare to results from previous investigators who have reported strains at this site for different modes of loading [10 –16,27]. Strain and load were amplified and digitally recorded (DaqView v7.2.30, DBK43A module, IOTech, Cleveland, OH). Gage location was verified on subsequent microCT images. Mice were repositioned multiple times to establish the range of strains applied during daily loading.
Loads ranging from 1 to 5 N resulted in linearly increasing mean tensile (+) strains from 250 to 1250 microstrain (Fig. 2). For the 3 N load, peak-to-peak mechanical strain averaged approximately 800 microstrain. The 40 N/s load rate resulted in a strain rate of about 11,000 microstrain/s. Loads of 6 N or greater tended to dislocate the knee joint. Peak locomotor strains in the ulna of running mice (+1676 microstrain) and rats ( 1200 microstrain) have been previously measured [14,26]. A recent report suggests lower magnitudes ( 300 microstrain) in the B6 tibia [28]. Bone adaptation to mechanical loading for 2 or 6 weeks was assessed in a separate set of 10-week old mice (n = 20). With the animal under isofluorane anesthetic (2.0% in 1 l/ min O2), loading was applied to the left tibia. Recovery from the anesthetic required approximately 10 min. Mice were allowed normal cage activity between loading sessions, and body mass was measured daily as an indicator of health. Intraperitoneal injections of calcein (20 mg/kg, Molecular Probes, Eugene, OR) and xylenol orange (90 mg/kg, Sigma-Aldrich, St. Louis, MO) were administered post load application at 10 days and 3 days before death, respectively. However, we experienced technical difficulty in precisely aligning histology sections between paired bones, so the labels were not used. Mice were killed by carbon dioxide inhalation 1 day after the last loading session. Methods of measurement
Fig. 2. (A) Calibration of mid-shaft tensile strain to applied compressive force at 5 load levels (n = 3 mice). The linear relationship was 256 microstrain/N. (B) Schematic of 1 s of the daily 5 min loading signal. Peakto-peak loading at 3 N resulted in calculated strain of approximately 800 microstrain on the medial mid-shaft surface of the tibia.
Overall, mineral adaptation was assessed as differences between the loaded and control tibiae of each animal. Mineral content and distribution were measured from a quantitative microCT scan (MS-8, GE Healthcare, London, Ontario, Canada) along the length of the tibia. The hardware of this system was specifically engineered to minimize the beam hardening associated with polychromatic X-ray to ensure uniformity of the beam at the detector array. Filtering of the low-energy portion of the X-ray spectrum with aluminum, acrylic, and a saline bath resulted in a flat field where the CT number within a uniform material is independent of spatial position. Scanner linearity was established with a phantom containing several densities of standard calibration material (Gammex RMI, Middleton, WI) and met the American Association of Physicists in Medicine standard requiring that the CT number not vary more than two standard deviations from the mean. The microfocus X-ray tube was driven at high energy (80 kVp and 100 AA; mean of 40 keV). Partial volume effects were addressed by using the smallest voxel isotropic resolution (11.5 Am) for the required field-of-view with 16-bit gray level precision. Noise reduction was accomplished with a large number of views (400) and increased frame averaging (7/view) and shutter exposure time (3s/frame), resulting in a total scan time of over 3 h per specimen. A calibration phantom including air, saline, and a mineral standard material (SB3, Gammex RMI) allowed calibration and
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
conversion of X-ray attenuation such that mineral density was proportional to grayscale values in Hounsfield Units (HU). Each hind limb was dissected, disarticulated at the hip, and positioned for scanning in a custom built acrylic holder with the same foot and knee geometry as the loading device. The holder, which was filled with saline for scanning, provided the same alignment for each tibia, so that direct comparisons between paired tibiae could be made. The hind limbs of a single animal were scanned on the same day whenever possible. Each scan resulted in a ‘‘stack’’ of approximately 1600 thin sections representing the entire length of the tibia. Left limbs were reflected to also appear as right limbs for the blinded observer. Digital reconstruction of ray projection to CT volume data was accomplished with a modified Parker algorithm. Bone mineral content (BMC, g), mineralized tissue (bone) volume (BV, cm3), and total volume (TV, cm3) were measured within volumes of interest. The ratio of bone volume fraction (BV/TV) was calculated for cancellous bone. Tibial length was determined as the distance between the most proximal and distal cross-sectional slices that contained bone. Appropriate functional parameters were calculated and averaged over diaphyseal and metaphyseal volumes of interest. Principal area moments of inertia (I max and I min), indicators of resistance to bending, were directly computed for every cross-section by integration of the product of each bone pixel area and the square of pixel distance about orthogonal axes aligned with the long axis of the tibia [29]. Axis orientation with respect to a standard reference frame was tracked because angular orientation of principal axes may also change with loading. Three-dimensional morphometry determinations of mean trabecular thickness and separation were directly computed by the distance transformation method [30]. Volumes of interest were chosen to isolate local sites responding to mechanical loading. Each entire tibia data set
1033
was used to calculate BMC and BV for the whole bone. The differential response was then examined along the entire length of the tibia at cross-sectional slices that were spaced 0.625% (¨115 Am) along the bone length. Based on this analysis and our hypothesis, two volumes of interest were defined: the proximal metaphysis composed of a mixture of cancellous and compact shell bone and the mid-diaphysis composed of purely cortical bone (Fig. 3). We chose the mid-diaphysis because significant adaptation occurs in ulna loading studies at this site and because the mid-diaphysis was a common site of strain measurement in other longbone adaptation studies [10 – 16,31,32]. We chose the proximal metaphysis because there is a large amount of cancellous bone. The proximal metaphysis consisted of the volume located within 7.5– 10% of the bone length from the proximal end, and the mid-diaphysis volume was centered about the mid-shaft (50%) and also spanned 2.5% of the bone length (¨0.46 mm, Fig. 3). To examine possible tibial expansion due to loading or growth at the proximal metaphysis, the total volume was measured including bone and marrow space but excluding the fibula (Fig. 3). To examine a volume of cancellous bone only, a final volume of interest with regular cylindrical geometry was defined within the endosteum of the proximal metaphysis (Fig. 3). The cylinder had a diameter of 1.24 mm and a height equal to that of the proximal metaphysis volume (2.5% of the length or ¨0.46 mm). To analyze diaphyseal cross-sectional moments of inertia along the length, volumes of interest at 25%, 50%, and 75% of length were examined (Fig. 3). All cross-sectional values were averaged from 40 consecutive, tibial microCT slices spaced 11.5 Am apart, representing 0.46 mm of length. Analysis of the moments of inertia excluded the fibula. Comparative mineralized tissue volume, moments of inertia, trabecular morphometry, and length measurements were made at a global threshold of 1100 Hounsfield Units (1100 HU = 0.43 g/cm3) to separate mineralized tissue from the marrow and soft tissues. A low threshold was chosen to
Fig. 3. Illustration of analysis volumes of interest (VOI). The proximal metaphysis volume of interest consisted of the volume located within 7.5 – 10% of the bone length and the mid-diaphysis volume of interest was centered about the mid-shaft. Diaphyseal volumes of interest at 25%, 50%, and 75% were analyzed for cross-sectional moments of inertia exclusive of the fibula. The cancellous volume of interest was cylindrical with diameter of 1.24 mm. All volumes of interest had length of 2.5% of the bone length (¨0.46 mm).
1034
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
were assessed graphically and by the Kolmogorov –Smironov test. Significant tests were conducted by two-factor repeated-measures analysis of variance with full interaction. The within-subject factor was loaded versus control limb, and the between-subjects factor was time, which was included to examine any differences between 2 and 6 weeks of loading. In addition, for comparison of properties measured at different locations, a three-factor analysis was performed to test variations with length. Statistical analyses were run using SYSTAT (SPSS Science, v. 8.0, Chicago, IL). The Type I error rate (alpha) was set at 0.05.
Results
Fig. 4. Lengthwise comparative analysis of bone mineral content (BMC) assessed by microCT of C57BL/6 mouse tibiae showing percent increases in mineral content for 2-week ( SD) and 6-week (+SD) loaded tibiae. All plotted cross-sectional values were spaced ¨3% along the bone length. Increases in mineral content were significant in the proximal metaphysis.
capture newly mineralized bone tissue. Bone mineral and moment of inertia determinations were accomplished using custom code written in MATLAB (MathWorks, v. 6.5, Natick, MA). The cancellous cylinder analysis was performed in Microview (GE Healthcare, v. 1.23). Analysis of data Data summaries are presented for each limb and as percent difference in the loaded limb relative to the contralateral limb. Each variable is summarized as mean and 95% confidence interval. Departures from normality
Comparison of mineral content indicated adaptation of the loaded tibiae compared with the contralateral bones, and the response was site specific. The response declined sharply along the entire length distal to the proximal metaphysis (Fig. 4). Analysis of volumes of interest chosen to isolate local sites responding to mechanical loading indicated that loading induced significantly greater increases in mineral content and volume in the proximal metaphysis than in the mid-diaphysis. Mineral content of the proximal metaphysis increased significantly by 9% after 2 weeks of loading and by 14% after 6 weeks of loading, whereas middiaphysis increases were only 2% for both terms (Table 1). Mineralized volume of the proximal metaphysis significantly increased by 8% after 2 weeks of loading and by 13% after 6 weeks of loading, while mineralized volume of the mid-diaphysis was unaffected by loading (Table 1). The mineral content response in both mid-diaphysis and proximal metaphysis regions was not significantly different between the 2-week and 6-week terms although the
Table 1 Mean values and 95% confidence intervals for mineralized tissue properties and the differences between loaded and control for whole tibia, and two subvolumes of interest Variable
Length (mm) (95% confidence interval) Whole bone mineralized tissue content (mg) Whole bone mineralized tissue volume (mm3) Proximal metaphysis mineral content (mg) Proximal metaphysis mineralized volume (mm3) Proximal metaphysis total volume (mm3) Mid-diaphysis mineral content (mg) Mid-diaphysis mineralized volume (mm3)
2 weeks of loading
6 weeks of loading
Loaded
Control
Difference (%)
Loaded
Control
Difference (%)
17.76 (17.38, 18.13) 33.52. (30.18, 36.86) 23.35. (21.29, 25.41) 1.48. (1.32, 1.63) 1.36. (1.24, 1.47) 1.91. (1.78, 2.04) 0.71. (0.63, 0.79) 0.42 (0.37, 0.46)
17.72 (17.36, 18.07) 32.99 (29.59, 36.39) 23.05 (20.82, 25.27) 1.36 (1.20, 1.51) 1.26 (1.13, 1.39) 1.82 (1.66, 1.99) 0.70 (0.61, 0.79) 0.41 (0.36, 0.46)
0.24 ( 0.13, +0.61) 1.7 ( 1.3, +4.6) 1.4 ( 1.2, +4.0) 9.4 (+1.5, +17.3) 8.1 (+1.5, +14.8) 5.0 (+0.4, +9.7) 1.9 ( 1.0, +4.7) 0.79 ( 2.1, +3.7)
18.00 (17.82, 18.18) 33.40. (31.16, 35.64) 22.93. (21.69, 24.18) 1.53. (1.38, 1.67) 1.34. (1.23, 1.45) 1.78. (1.73, 1.84) 0.71. (0.66, 0.76) 0.42 (0.39, 0.44)
17.95 (17.77, 18.13) 32.24 (30.43, 34.05) 22.17 (21.06, 23.28) 1.34 (1.20, 1.47) 1.18 (1.07, 1.30) 1.62 (1.53, 1.70) 0.70 (0.65, 0.74) 0.41 (0.39, 0.43)
0.27 ( 0.12, +0.65) 3.5 (+1.0, +6.0) 3.4 (+1.8, +5.0) 14 (+9.8, +19.0) 13 (+9.4, +17.2) 11(+7.4, +14.1) 2.0 ( 0.4, +4.3) 1.7 ( 0.2, +3.7)
Proximal metaphysis and diaphysis volumes of interest were 0.46 mm in length. Significance levels by ANOVA: .P < 0.05 versus control, -P < 0.05 versus 2 weeks of loading.
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
Fig. 5. Box plots of regional change in average diaphyseal maximum crosssectional moment of inertia, I max. Increases were significant at the 25% cross-section and were similar after 2-week and 6-week loading terms. All cross-sectional values were averaged from 40 consecutive, tibial microCT slices spaced 11.5 Am apart, representing 0.46 mm of length. The horizontal lines of each box denote the lower quartile, median, and upper quartile values. The open circle marks the mean. The whiskers denote the extent of the rest of the data.
tendency was for longer term loading to result in larger increases. A similar site-specific response was observed for maximum principal cross-sectional moment of inertia (I max). I max was increased at the 25% volume of interest after 2 and 6 weeks of loading by 9% and 8% respectively and there was no difference in the response at 2 versus 6 weeks (Fig. 5). The difference in I max was not significant in the 50% and 75% volumes of interest for either loading term. I min and the orientation of principal axes were not significantly affected at any of the three locations after either loading term. The purely cancellous cylindrical subvolume also had significantly increased mineralized tissue volume (Table 2; Fig. 6). Bone volume fraction (BV/TV) of the cancellous subvolume was greater at both time points: 13% for the 2week group and 15% for the 6-week group (Table 2). The similarity in bone volume fraction increases due to the term of loading indicates that the term had no effect on the overall cancellous mineral response. Trabecular thickness was increased by 12% after 6 weeks of loading but was not significant after 2 weeks of loading indicating that the term of loading had an effect on trabecular structure. However, trabecular separation was not affected by loading. Trabecular separation did indicate age related modeling by an
1035
increase after the 6-week term relative to the 2-week term (Table 2). Modeling in the proximal metaphysis with age was also indicated by a decrease in the total volume at 6 weeks relative to 2 weeks. This volume decreased 7% in loaded and 11% in control limbs (Table 1). At each of the two time points, the loaded proximal tibia had a greater total volume compared with the control side. Additionally, the relative increase in total volume after 6 weeks of loading was significantly greater than after 2 weeks of loading, increasing from 5% to 11% (Table 1). Factors that could have confounded the measured outcomes include animal health, differences in limb usage, differences in tibial lengths, loss of mice, and departures from normal distributions in the data. The mean body mass of the mice was 21.4 g T 1.7 g when received at our facility. During the 3-week acclimation period, their mass increased to 23.8 g T 1.9 g at the start of loading, consistent with a period of rapid growth in the adolescent mouse [33]. All mice maintained or increased body mass over the experimental period, indicative of good health and tolerance to the anesthetic and mechanical loading regimen (at 2 weeks: 24.0 g T 1.8 g; at 6 weeks: 24.6 g T 1.8 g). No signs of lameness were observed between daily loading sessions. Tibial lengths were not significantly different between loaded and control limbs (Table 1). One animal in the 6week group did not complete the study for reasons unrelated to the mechanical loading protocol. The data from an animal in the 2-week group were lost due to an unrecoverable microCT scanning error. No variables were found to depart significantly from a normal distribution.
Discussion We hypothesized that the volume of mineralized tissue would increase in response to controlled cyclic axial loading applied to the adolescent mouse tibia and that the adaptation would be site specific. Indeed, the volume of mineralized tissue increased in the corticocancellous proximal metaphysis. Furthermore, the volume fraction of cancellous bone was increased in the proximal metaphysis. A significantly lower response was noted in the cortical mid-diaphysis than proximal metaphysis.
Table 2 Mean values and differences for bone volume fraction and morphometric measures within loaded and control proximal metaphyseal cancellous volumes of interest Variable
Bone volume fraction (BV/TV) (95% confidence interval) 3D direct trabecular thickness (Am) 3D direct trabecular separation (Am)
2 weeks of loading
6 weeks of loading
Loaded
Control
Difference (%)
Loaded
Control
Difference (%)
0.23 (0.21, 0.25).
0.20 (0.18, 0.23)
13 (+4.8, +21.4)
0.23 (0.20, 0.25).
0.21 (0.17, 0.24)
15 (+1.3, +28)
43.1 (41.6, 44.7) 204 (201, 208)
42.1 (40.6, 43.6) 204 (200, 208)
2.6 (+0.2, +4.9) 0.4 ( 2.4, +3.2)
46.5 (44.2, 48.7). 216 (209, 222)
41.7 (40.0, 43.3) 213 (205, 221)
12 (+5.6, +18)1.6 ( 2.2, +5.3)
The cylindrical volume of interest had diameter of 1.24 mm and height of 0.46 mm. Significance levels by ANOVA: .P < 0.05 versus control, -P < 0.05 versus 2 weeks of loading.
1036
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
Fig 6. Superior view of three-dimensional reconstructions of the proximal metaphysis (top) from paired tibiae of a 6-week loaded mouse. The proximal cut surface of the metaphysis is white. Mineral content of the loaded side is increased 22% over the control side. Bone volume fraction (BV/TV) of the cancellous volume of interest is 23.8% in the loaded and 19.3% in the control, a 23% increase with loading. Scale bar is 1 mm. Sections are ¨0.46 mm thick.
We further hypothesized that adaptation would depend on the term of loading. Most of our data do not support the loading-term dependent hypothesis; results from short-term (2 weeks) and long-term (6 weeks) mechanical loading were similar for most volumes of interest. This lack of effect is in agreement with previous work that has shown both cortical and cancellous formation reverting to control levels after the initial response to mechanical loading [10,17,34]. One explanation for this behavior is that the mechanosensitivity of bone cells diminishes in response to continued loading [34,35], though low anabolic activity of osteoblasts does not necessarily require loss in cellular ability to sense a signal. We did, however, detect two pieces of evidence in the proximal metaphysis in support of our loading term hypothesis. The total volume and trabecular thickness were increased more after 6 weeks of loading than after 2 weeks of loading. These increases could have resulted from increased osteoblastic activity, reduced osteoclastic activity, or both. Site-specific reductions in osteoclastic activity have been demonstrated in vitro and in vivo due to mechanical loads [12,15,16,36]. The failure of trabecular thickness increase to reach statistical significance after 2 weeks of loading may also indicate a detection limit of our measurement technique due to the shorter term combined with our level of loading (¨800 Aq). Our study was not specifically powered for this measurement. Tibial loading of 14-week old B6 mice at a high load magnitude (producing ¨2000 Aq) for 3 weeks also increased the trabecular thickness measured by microCT [28]. Comparing proximal metaphysis results from contralateral bones in the 2-week and 6-week groups showed that the total volume decreased and trabecular separation increased with age (Tables 1 and 2). We believe these changes represent growth-related modeling of the metaphysis. Our low mineral response at the tibial mid-diaphysis (Table 1; Fig. 4) contrasts with previous reports of increased formation due to axial ulnar loading in rodents and tibial
loading in mice. Ulnar studies have utilized peak strains (¨3000 Aq) that were based on vigorous activities such as falling from a ledge (e.g., [9,12 –16,34]). The aforementioned tibial study utilized lower strains (2000 Aq) [28]. The peak cortical strains induced in our study were even lower (¨800 Aq) and within the physiological range found in the fore limbs of mice and rats for activities such as running [14,26]. However, tibial strain during normal locomotion in the B6 may be lower than previously expected (¨300 Aq) [28]. Nonetheless, a site-specific threshold may need to be reached to produce a detectable mineral response and higher strains and rates of loading have been associated with increased cortical formation responses [13,14,21]. The lack of response in moments of inertia at the tibial mid-diaphysis also contrasts with a rat ulna study [34]. In the rat ulna, supra-physiological strains induced the largest response in the distal diaphysis that were greater in minimum (70%) versus maximum (23%) moment of inertia after 16 weeks of loading and were sustained along a large portion of the diaphysis [34,37]. We detected a smaller tibial increase in maximum diaphyseal moment of inertia (9%) in the proximal diaphysis (volume of interest at 25% of length). Minimum moments of inertia were not increased. In addition to the aforementioned strain magnitude and animal difference, the loading term and site were dissimilar to our study. There are important cortical and cancellous structural differences between fore and hind limbs. The rat ulna is more slender with a mid-diaphysis I max:I min ratio of approximately 6 [37], while the ratio in the B6 mouse tibia does not exceed 2 at the mid-diaphysis and 4 at 25% of length. The function of these two bones is different and the rodent ulna is complemented and closely associated with a substantial radius bone, while the fibula, which fuses with the distal tibia, has little load bearing capability. The metaphyses of the tibia also contain more cancellous bone than those of the ulna, which may also reflect differences in function. Our finding of increased bone volume fraction confirms the novel utility of non-invasive in vivo axial loading of the B6 tibia on cancellous bone of the proximal tibia [28]. Another promising avenue for studying cancellous bone formation has been demonstrated due to loading of 9-month old, male rat-tail vertebrae [17]. However, despite the formation gains, in vivo axial compression of the rat tail failed to increase bone volume fraction at the vertebral corticocancellous site. The different response may have age, species, site, loading term, and load related components. We chose to use B6 mice at 10 weeks of age to examine the skeletal response immediately prior to attainment of peak bone mass at 16 weeks [22]. B6 mice do not reach skeletal maturity in terms of cortical bone mechanical properties until the age of 20 weeks [38]. Additionally, in the rat-tail model, the vertebrae were clamped between loading sessions so that any effects of short daily loading periods were superimposed on a background of disuse. Our mice were allowed normal cage activity between loading sessions. Combining the
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
applied mechanical loading approach we have taken with hind limb suspension would provide a similar loading/ unloading regimen for a more direct comparison [2,39]. MicroCT has both strengths and limitations for assessing adaptation. This quantitative method provides a threedimensional assessment of net mineral response resulting from both osteoblastic and osteoclastic activity throughout the tibia. Traditionally, adaptation responses have been measured with two-dimensional histomorphometry using fluorochrome double labels. In contrast, quantitative microCT allows detection of differences in bone volume fraction, mass (g), and morphometric parameters [40]. The method could eventually be augmented with three-dimensional quantification of bone formation volumes [41,42]. The higher resolution of microCT compared with other Xray techniques is critical to localizing the mineral response in bony structures of small animals and facilitating paired analysis by allowing very accurate alignment of loaded versus contralateral bones. Longitudinal growth suppression, which affects alignment ability, was not observed in our study, in contrast to previous reports on axial compression of the rodent ulna [13,14,19,20,34,38]. The use of polychromatic X-rays precludes the elimination of all beam-hardening artifacts. All digital microscopy techniques such as microCT are also subject to partial volume and threshold selection errors. However, as detailed in the Materials and methods section, we addressed beam hardening and digitization errors by filtering of the low-energy portion of the X-ray spectrum, optimizing resolution, and calibrating with a mineral standard. Additionally, we posed our hypothesis to take advantage of a comparative analysis such that each loaded limb was paired against its contralateral control limb to minimize any artifacts. In conclusion, our results indicate that a site-specific bone mineral increase is generated in the proximal metaphysis by a daily regimen of controlled axial loading applied to the B6 mouse tibia prior to peak bone mass attainment. The response in the corticocancellous proximal metaphysis was greater than in the cortical diaphysis. Furthermore, the overall corticocancellous response saturated within 2 weeks of onset of loading while an increase in trabecular thickness was detected only after 6 weeks of loading. This loading model provides an opportunity to more fully understand the role of mechanical loading during the growth of skeletal structures. Elucidation of genetic and mechanical factors important to the attainment of peak bone mass in young adult life and maintenance into maturity will help to eliminate the inadequate accumulation and excessive loss that currently predispose individuals to osteoporosisrelated fracture.
Acknowledgments We thank Dr. Mathias Bostrom and the technical staff at HSS for surgical assistance, Mr. Michael Thornton for
1037
technical assistance with microCT, and Dr. Karl Jepsen for providing a moment of inertia algorithm. This work was financially supported by Cornell University graduate student fellowship and NASA GSRP #NGT5-50375 (JCF), Sigma Xi Grant-in-Aid of Research, NIH (P30AR046121 and S10RR014801), and Clark and Kirby Foundations.
References [1] Carter DR, van der Meulen MCH, Beaupre GS. Mechanical factors in bone growth and development. Bone 1996;18S:5 – 10. [2] Morey-Holton ER, Globus RK. Hindlimb unloading of growing rats: a model for predicting skeletal changes during space flight. Bone 1998;22:83S – 8S. [3] Rubin CT, Lanyon LE. Regulation of bone formation by applied dynamic loads. J Bone Joint Surg 1984;66A:397 – 402. [4] Uhthoff HK, Sekaly G, Jaworski ZF. Effect of long-term nontraumatic immobilization on metaphyseal spongiosa in young adult and old beagle dogs. Clin Orthop 1985;192:278 – 83. [5] Biewener AA, Bertram JE. Structural response of growing bone to exercise and disuse. J Appl Physiol 1994;76:946 – 55. [6] Woo SL, Kuei SC, Amiel D, Gomez MA, Hayes WC, White FC, et al. The effect of prolonged physical training on the properties of long bone: a study of Wolff’s Law. J Bone Joint Surg 1981;63A:780 – 7. [7] Riggs BL, Melton LJ. The worldwide problem of osteoporosis: insights afforded by epidemiology. Bone 1995;17:505S – 11S. [8] van der Meulen MC, Moro M, Kiratli BJ, Marcus R, Bachrach LK. Mechanobiology of femoral neck structure during adolescence. J Rehabil Res Dev 2000;37:201 – 8. [9] Lee K, Jessop H, Suswillo R, Zaman G, Lanyon L. Endocrinology: bone adaptation requires oestrogen receptor-alpha. Nature 2003; 424:389. [10] Srinivasan S, Agans SC, King KA, Moy NY, Poliachik SL, Gross TS. Enabling bone formation in the aged skeleton via rest-inserted mechanical loading. Bone 2003;33:946 – 55. [11] Turner CH, Akhter MP, Raab DM, Kimmel DB, Recker RR. A noninvasive, in vivo model for studying strain adaptive bone modeling. Bone 1991;12:73 – 9. [12] Hillam RA, Skerry TM. Inhibition of bone resorption and stimulation of formation by mechanical loading of the modeling rat ulna in vivo. J Bone Miner Res 1995;10:683 – 9. [13] Mosley JR, Lanyon LE. Strain rate as a controlling influence on adaptive modeling in response to dynamic loading of the ulna in growing male rats. Bone 1998;23:313 – 8. [14] Mosley JR, March BM, Lynch J, Lanyon LE. Strain magnitude related changes in whole bone architecture in growing rats. Bone 1997; 20:191 – 8. [15] Noble BS, Peet N, Stevens HY, Brabbs A, Mosley JR, Reilly GC, et al. Mechanical loading: biphasic osteocyte survival and targeting of osteoclasts for bone destruction in rat cortical bone. Am J Cell Physiol 2003;284:C934 – 43. [16] Torrance AG, Mosley JR, Suswillo RF, Lanyon LE. Noninvasive loading of the rat ulna in vivo induces a strain-related modeling response uncomplicated by trauma or periosteal pressure. Calcif Tissue Int 1994;54:241 – 7. [17] Kim CH, Takai E, Zhou H, von Stechow D, Muller R, Dempster DW, et al. Trabecular bone response to mechanical and parathyroid hormone stimulation: the role of mechanical microenvironment. J Bone Miner Res 2003;18:2116 – 25. [18] Sheng MHC, Baylink DJ, Beamer WG, Donahue LR, Rosen CJ, Lau K-HW, et al. Histomorphometric studies show that bone formation and bone mineral apposition rates are greater in C3H/HeJ (high density) than C57BL/6J (low density) mice during growth. Bone 1999;25:421 – 9.
1038
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
[19] Ohashi N, Robling AG, Burr DB, Turner CH. The effects of dynamic axial loading on the rat growth plate. J Bone Miner Res 2002;17:284 – 92. [20] Mosley JR, Lanyon LE. Growth rate rather than gender determines the size of the adaptive response of the growing skeleton to mechanical strain. Bone 2002;30:314 – 9. [21] Turner CH, Takano Y, Owan I. Aging changes mechanical loading thresholds for bone formation in rats. J Bone Miner Res 1995; 10:1544 – 9. [22] Beamer WG, Donahue LR, Rosen CJ, Baylink DJ. Genetic variability in adult bone density among inbred strains of mice. Bone 1996;18:397 – 403. [23] Sheng MH, Baylink DJ, Beamer WG, Donahue LR, Lau KH, Wergedal JE. Regulation of bone volume is different in the metaphyses of the femur and vertebra of C3H/HeJ and C57BL/6J mice. Bone 2002;30:486 – 91. [24] Robling AG, Turner CH. Mechanotransduction in bone: genetic effects on mechanosensitivity in mice. Bone 2002;31:562 – 9. [25] Akhter MP, Cullen DM, Pedersen EA, Kimmel DB, Recker RR. Bone response to in vivo mechanical loading in two breeds of mice. Calcif Tissue Int 1998;63:442 – 9. [26] Lee KCL, Maxwell A, Lanyon LE. Validation of a technique for studying functional adaptation of the mouse ulna in response to mechanical loading. Bone 2002;31:407 – 12. [27] Fritton SP, Rubin CT. In vivo measurement of bone deformations using strain gauges. In: Cowin SC, editor. 2nd edR Bone mechanics handbook, ch. 8. Boca Raton, FL’ CRC; 2001. p. 1 – 41. [28] de Souza RL, Matsuura M, Eckstein F, Lanyon LE. Unique evidence for load induced modification in trabecular bone architecture as well as increased cortical bone formation in a non-invasive axial loading of mouse tibiae. Osteoporos Int 2004;15S1:13. [29] Nagurka ML, Hayes WC. An interactive graphics package for calculating cross-sectional properties of complex shapes. J Biomech 1980;13:59 – 64. [30] Hildebrand T, Ruegsegger P. A new method for the modelindependent assessment of thickness in three-dimensional images. J Microsc 1997;185:67 – 75. [31] Robling AG, Li J, Shultz KL, Beamer WG, Turner CH. Evidence for a
[32]
[33] [34]
[35] [36]
[37]
[38]
[39]
[40]
[41]
[42]
skeletal mechanosensitivity gene on mouse chromosome 4. FASEB J 2003;17:324 – 6. Gross TS, Srinivasan S, Liu CC, Clemens TL, Bain SD. Noninvasive loading of the murine tibia: an in vivo model for the study of mechanotransduction. J Bone Miner Res 2002;17: 493 – 501. Body Weights for Selected Strains of Jax\ Mice. Available at: http:// jaxmice.jax.org/library/charts/bodyweights.pdf. Robling AG, Hinant FM, Burr DB, Turner CH. Improved bone structure and strength after long-term mechanical loading is greatest if loading is separated into short bouts. J Bone Miner Res 2002; 17:1545 – 54. Turner CH. Three rules for bone adaptation to mechanical stimuli. Bone 1998;23:399 – 407. Rubin J, Fan X, Biskobing DM, Taylor WR, Rubin CT. Osteoclastogenesis is repressed by mechanical strain in an in vitro model. J Orthop Res 1999;17:639 – 45. Robling AG, Hinant FM, Burr DB, Turner CH. Shorter, more frequent mechanical loading sessions enhance bone mass. Med Sci Sports Exerc 2002;34:196 – 202. Brodt MD, Ellis CB, Silva MJ. Growing C57B1/6 mice increase whole bone mechanical properties by increasing geometric and material properties. J Bone Miner Res 1999;14:2159 – 66. Simske SJ, Guerra KM, Greenberg AR, Luttges MW. The physical and mechanical effects of suspension-induced osteopenia on mouse long bones. J Biomech 1992;25:489 – 99. Muller R, van Campenhout H, van Damme B, van der Perre G, Dequeker J, Hildebrand T, et al. Morphometric analysis of human bone biopsies: a quantitative structural comparison of histological sections and micro-computed tomography. Bone 1998;23:59 – 66. Waarsing JH, Day JS, van der Linden JC, Ederveen AG, Spanjers C, De Clerck N, et al. Detecting and tracking local changes in the tibiae of individual rats: a novel method to analyse longitudinal in vivo micro-CT data. Bone 2004;34:163 – 9. Kinney JH, Ryaby JT. Resonant markers for noninvasive threedimensional dynamic bone histomorphometry with X-ray microtomography. Rev Sci Instrum 2001;72:1921 – 3.
Loading induces site-specific increases in mineral content assessed by microcomputed tomography of the mouse tibia J.C. Frittona,b,T, E.R. Myersb, T.M. Wrighta,b, M.C.H. van der Meulena,b b
a Biomedical Engineering, Cornell University, Ithaca, NY 14853, USA Hospital for Special Surgery, Laboratory for Biomedical Mechanics and Materials, Research Division, 535 E. 70th Street, New York, NY 10021, USA
Received 2 September 2004; revised 14 January 2005; accepted 25 February 2005
Abstract Adaptation to mechanical loading has been studied extensively in cortical, but not cancellous bone. However, corticocancellous sites are more relevant to osteoporosis and related fracture risk of the hip and spine. We tested the hypotheses that adaptation in a long bone would be greater at cancellous than cortical sites and would depend on the term of daily in vivo cyclic axial loading. We applied compressive loads to the adolescent, 10-week old, male C57BL/6 mouse tibia to examine the skeletal response immediately prior to attainment of peak bone mass. Adaptation was quantified at the completion of either 2-week (n = 8) or 6-week (n = 12) loading terms by directly comparing volumetric bone mineral content between loaded and contralateral limbs by microcomputed tomography. The increase in mineral content was site specific with a greater response found in the corticocancellous proximal metaphysis (14%) than the cortical mid-shaft (2%) after 6 weeks of loading. Furthermore, bone volume fraction and average trabecular thickness of cancellous bone in the proximal tibia increased after 6 weeks by 15% and 12% respectively. Diaphyseal response was only evident proximal to the mid-shaft as indicated by an 8% increase in maximum principal moment of inertia. Both loading terms produced similar results for mineral content, volume fraction, and moments of inertia. Our finding that non-invasive loading increases the bone volume and fraction at a corticocancellous site by as much as 15% motivates exploring the use of mechanical loading to attain greater peak bone mass and inhibit osteoporosis. D 2005 Elsevier Inc. All rights reserved. Keywords: In vivo mechanical loading; Bone adaptation; Trabecular bone; Mice; MicroCT
Introduction Mechanical loading of the skeleton regulates the attainment of peak bone mass in the growing animal and produces adaptive changes in the adult. Load bearing during growth has a strong influence on the resulting adult structure [1]. Loss of normal load bearing leads to reduced appositional bone growth and bone mass deficit in adults [1 – 4]. Conversely, increased load bearing can result in increased bone mass in adults [5,6]. Gaining insight into this adaptive system will help elucidate basic mechanisms involved in mechanotransduction. * Corresponding author. Hospital for Special Surgery, Laboratory for Biomedical Mechanics, 535 E. 70th Street, New York, NY 10021, USA. Fax: +1 212 606 1490. E-mail address: [email protected] (J.C. Fritton). 8756-3282/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2005.02.013
The factors controlling attainment of peak bone mass in humans are particularly critical at corticocancellous sites, perhaps more so than in the cortical diaphysis. Corticocancellous regions such as the hip, spine, and distal radius experience the greatest adult bone loss and risk of fracture [7]. Furthermore, they may be more responsive to mechanical loads than diaphyseal cortical sites during growth [8]. Therefore, mechanical loading models of corticocancellous sites are required to understand the mechanical and genetic factors that impact fracture risk. One impediment to ascertaining adaptive responses at a corticocancellous site is that the complex mechanical structure exists in a dynamic environment of bone turnover. Previously, cortical adaptive formation responses have been measured with 2D histomorphometry of two or more bone labels given at short intervals within the term of controlled loading [9 – 11]. However, during growth, the cortical
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
adaptive response to loading includes site-specific inhibition of resorption [12 – 16]. At cancellous sites, increased short-term formation rates may not indicate a concomitant increase in bone volume fraction [17]. This lack of bone volume change suggests an increase in both osteoblastic and osteoclastic activity similar to that observed during high turnover. Quantitative microcomputed tomography (microCT) is a 3D technique useful in measuring the net mineral deposited or left unresorbed over the entire term of study. A net mineral measurement is suited for high turnover sites and high periods of turnover such as growth and adaptation. Many studies have focused on older animals to eliminate the potentially confounding effects of growth, but the adolescent animal is critical to delineating the factors that determine peak bone mass [18]. Higher loading magnitudes and rates can result in suppression of longitudinal growth of the rat ulna [13,14,19,20]. Additionally, loading paradigms observed in the adult may not apply to periods of rapid growth and mineral acquisition. Compared to the adult, lower mechanical loading thresholds are sufficient to produce responses in juveniles [21]. We tested the hypothesis that the volume of mineralized tissue would increase in both cortical and cortiocancellous sites and that the increase is site specific and depends on the term of controlled cyclic axial loading applied to the adolescent mouse tibia. Our approach extended that used for axial compression of the ulna [12,16]. The tibia was selected for the distinct advantage of a substantial cancellous bone volume subject to loading. Bone mineral content was compared between loaded and contralateral limbs by using quantitative microCT. We also examined the effects of short-term (2 weeks) and long-term (6 weeks) mechanical loading.
Materials and methods Study design To determine whether bone mass is increased in the adolescent mouse tibia due to mechanical loading, we studied the mineral response to a daily regimen of in vivo, controlled, axial loading in twenty male mice. Cyclic, mechanical loads were imparted with a device specifically designed for the mouse tibia. At the completion of either 2week (n = 8) or 6-week (n = 12) loading terms, bone mineral content and distribution were measured using quantitative microCT in both the loaded and the contralateral, control limb. The number of mice per group was based on a power analysis considered prior to the experiment. With 8 mice per group, there was over 90% power to detect an effect size of 0.9 for the primary factor of interest, loading, using a twofactor repeated-measures design (time as a between-subjects factor and load as a within-subjects factor, alpha = 0.05). We
1031
should, therefore, be able to determine a mean paired difference in mineral content that was approximately 90% of the standard deviation with 8 per group. Four mice were added to the 6-week group to account for potential loss due to repeated use of anesthesia. Power calculations were made using a Repeated Measure AOV panel (PASS 6.0, NCSS, Kaysville, UT). Animals C57BL/6J male mice (B6, Jackson Laboratories, Bar Harbor, ME) were received and acclimated to our facility for 3 weeks. All mice were 10 weeks old (T1 day) at the start of loading. Mice were caged in groups of four and fed ad libitum. The B6 mouse has relatively low peak bone mass and low bone formation rates during growth in the cortices of the femur and tibia and metaphyses of the femur compared with the C3H/HeJ strain [18,22,23]. However, the adult B6 mouse has an enhanced cortical bone formation response to loading [24,25]. Therefore, this strain may prove useful for studying the basis of mechanical loading in the attainment of peak bone mass. Our Institutional Animal Care and Use Committee approved all protocols. Mechanical loading protocol The loading scheme was similar to that of the axial ulna loading model [12,16] but carried out on the tibia with our loading device. Compression was applied at the ends of the left tibia using a small, electro-magnetic actuator (LA18-18000A, BEI Kimco Magnetics, San Marcos, CA). Load transmission was through a brass platen securing the foot and attached to the core of the actuator. Another brass platen attached to the stiff aluminum frame cupped the knee (Fig. 1). The frame was adjustable to accommodate slight differences in limb lengths. Interposed between the frame and the knee platen was a 100 N load cell (ELFS-T3E, Entran, Fairfield, NJ). A pre-load of 0.2 N was applied immediately prior to dynamic compression to maintain the initial position of the limb. Loads applied in this manner induce combined axial compressive and bending strains because of the natural curvature of the tibia.
Fig. 1. Schematic of loading apparatus to apply axial compression to the mouse tibia. Arrows indicate the direction of loading.
1032
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
The waveform was a load ramp of 0.075 s duration followed by a symmetric unload ramp of 0.075 s duration (Fig. 2B). The dwell, or rest period between each load cycle, lasted 0.100 s at the pre-load level of 0.2 N. Loading was applied at 4 cycles/second representing the stride frequency of the mouse [26]. Loading continued for 1200 cycles, yielding a daily (5 days/week) loading period of 5 min. Peak-to-peak load magnitude was 3 N. The actuator was driven with amplified signals (BOV-20-10M amplifier, KEPCO, Flushing, NY) produced by a digital waveform function generator (AFG320, Tektronix, Beaverton, OR). Tibial mid-diaphyseal mechanical strains were recorded from the periosteal medial surface in 10-week old anesthetized mice (n = 3) to relate applied loads to resulting strain. Standard surgical techniques were used to elevate the periosteum and bond a single element, foil gage (CEA015UW-120, Vishay Micro-Measurements, Raleigh, NC) aligned with the long axis. The medial cortex at the midshaft was chosen to minimize soft tissue dissection and to compare to results from previous investigators who have reported strains at this site for different modes of loading [10 –16,27]. Strain and load were amplified and digitally recorded (DaqView v7.2.30, DBK43A module, IOTech, Cleveland, OH). Gage location was verified on subsequent microCT images. Mice were repositioned multiple times to establish the range of strains applied during daily loading.
Loads ranging from 1 to 5 N resulted in linearly increasing mean tensile (+) strains from 250 to 1250 microstrain (Fig. 2). For the 3 N load, peak-to-peak mechanical strain averaged approximately 800 microstrain. The 40 N/s load rate resulted in a strain rate of about 11,000 microstrain/s. Loads of 6 N or greater tended to dislocate the knee joint. Peak locomotor strains in the ulna of running mice (+1676 microstrain) and rats ( 1200 microstrain) have been previously measured [14,26]. A recent report suggests lower magnitudes ( 300 microstrain) in the B6 tibia [28]. Bone adaptation to mechanical loading for 2 or 6 weeks was assessed in a separate set of 10-week old mice (n = 20). With the animal under isofluorane anesthetic (2.0% in 1 l/ min O2), loading was applied to the left tibia. Recovery from the anesthetic required approximately 10 min. Mice were allowed normal cage activity between loading sessions, and body mass was measured daily as an indicator of health. Intraperitoneal injections of calcein (20 mg/kg, Molecular Probes, Eugene, OR) and xylenol orange (90 mg/kg, Sigma-Aldrich, St. Louis, MO) were administered post load application at 10 days and 3 days before death, respectively. However, we experienced technical difficulty in precisely aligning histology sections between paired bones, so the labels were not used. Mice were killed by carbon dioxide inhalation 1 day after the last loading session. Methods of measurement
Fig. 2. (A) Calibration of mid-shaft tensile strain to applied compressive force at 5 load levels (n = 3 mice). The linear relationship was 256 microstrain/N. (B) Schematic of 1 s of the daily 5 min loading signal. Peakto-peak loading at 3 N resulted in calculated strain of approximately 800 microstrain on the medial mid-shaft surface of the tibia.
Overall, mineral adaptation was assessed as differences between the loaded and control tibiae of each animal. Mineral content and distribution were measured from a quantitative microCT scan (MS-8, GE Healthcare, London, Ontario, Canada) along the length of the tibia. The hardware of this system was specifically engineered to minimize the beam hardening associated with polychromatic X-ray to ensure uniformity of the beam at the detector array. Filtering of the low-energy portion of the X-ray spectrum with aluminum, acrylic, and a saline bath resulted in a flat field where the CT number within a uniform material is independent of spatial position. Scanner linearity was established with a phantom containing several densities of standard calibration material (Gammex RMI, Middleton, WI) and met the American Association of Physicists in Medicine standard requiring that the CT number not vary more than two standard deviations from the mean. The microfocus X-ray tube was driven at high energy (80 kVp and 100 AA; mean of 40 keV). Partial volume effects were addressed by using the smallest voxel isotropic resolution (11.5 Am) for the required field-of-view with 16-bit gray level precision. Noise reduction was accomplished with a large number of views (400) and increased frame averaging (7/view) and shutter exposure time (3s/frame), resulting in a total scan time of over 3 h per specimen. A calibration phantom including air, saline, and a mineral standard material (SB3, Gammex RMI) allowed calibration and
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
conversion of X-ray attenuation such that mineral density was proportional to grayscale values in Hounsfield Units (HU). Each hind limb was dissected, disarticulated at the hip, and positioned for scanning in a custom built acrylic holder with the same foot and knee geometry as the loading device. The holder, which was filled with saline for scanning, provided the same alignment for each tibia, so that direct comparisons between paired tibiae could be made. The hind limbs of a single animal were scanned on the same day whenever possible. Each scan resulted in a ‘‘stack’’ of approximately 1600 thin sections representing the entire length of the tibia. Left limbs were reflected to also appear as right limbs for the blinded observer. Digital reconstruction of ray projection to CT volume data was accomplished with a modified Parker algorithm. Bone mineral content (BMC, g), mineralized tissue (bone) volume (BV, cm3), and total volume (TV, cm3) were measured within volumes of interest. The ratio of bone volume fraction (BV/TV) was calculated for cancellous bone. Tibial length was determined as the distance between the most proximal and distal cross-sectional slices that contained bone. Appropriate functional parameters were calculated and averaged over diaphyseal and metaphyseal volumes of interest. Principal area moments of inertia (I max and I min), indicators of resistance to bending, were directly computed for every cross-section by integration of the product of each bone pixel area and the square of pixel distance about orthogonal axes aligned with the long axis of the tibia [29]. Axis orientation with respect to a standard reference frame was tracked because angular orientation of principal axes may also change with loading. Three-dimensional morphometry determinations of mean trabecular thickness and separation were directly computed by the distance transformation method [30]. Volumes of interest were chosen to isolate local sites responding to mechanical loading. Each entire tibia data set
1033
was used to calculate BMC and BV for the whole bone. The differential response was then examined along the entire length of the tibia at cross-sectional slices that were spaced 0.625% (¨115 Am) along the bone length. Based on this analysis and our hypothesis, two volumes of interest were defined: the proximal metaphysis composed of a mixture of cancellous and compact shell bone and the mid-diaphysis composed of purely cortical bone (Fig. 3). We chose the mid-diaphysis because significant adaptation occurs in ulna loading studies at this site and because the mid-diaphysis was a common site of strain measurement in other longbone adaptation studies [10 – 16,31,32]. We chose the proximal metaphysis because there is a large amount of cancellous bone. The proximal metaphysis consisted of the volume located within 7.5– 10% of the bone length from the proximal end, and the mid-diaphysis volume was centered about the mid-shaft (50%) and also spanned 2.5% of the bone length (¨0.46 mm, Fig. 3). To examine possible tibial expansion due to loading or growth at the proximal metaphysis, the total volume was measured including bone and marrow space but excluding the fibula (Fig. 3). To examine a volume of cancellous bone only, a final volume of interest with regular cylindrical geometry was defined within the endosteum of the proximal metaphysis (Fig. 3). The cylinder had a diameter of 1.24 mm and a height equal to that of the proximal metaphysis volume (2.5% of the length or ¨0.46 mm). To analyze diaphyseal cross-sectional moments of inertia along the length, volumes of interest at 25%, 50%, and 75% of length were examined (Fig. 3). All cross-sectional values were averaged from 40 consecutive, tibial microCT slices spaced 11.5 Am apart, representing 0.46 mm of length. Analysis of the moments of inertia excluded the fibula. Comparative mineralized tissue volume, moments of inertia, trabecular morphometry, and length measurements were made at a global threshold of 1100 Hounsfield Units (1100 HU = 0.43 g/cm3) to separate mineralized tissue from the marrow and soft tissues. A low threshold was chosen to
Fig. 3. Illustration of analysis volumes of interest (VOI). The proximal metaphysis volume of interest consisted of the volume located within 7.5 – 10% of the bone length and the mid-diaphysis volume of interest was centered about the mid-shaft. Diaphyseal volumes of interest at 25%, 50%, and 75% were analyzed for cross-sectional moments of inertia exclusive of the fibula. The cancellous volume of interest was cylindrical with diameter of 1.24 mm. All volumes of interest had length of 2.5% of the bone length (¨0.46 mm).
1034
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
were assessed graphically and by the Kolmogorov –Smironov test. Significant tests were conducted by two-factor repeated-measures analysis of variance with full interaction. The within-subject factor was loaded versus control limb, and the between-subjects factor was time, which was included to examine any differences between 2 and 6 weeks of loading. In addition, for comparison of properties measured at different locations, a three-factor analysis was performed to test variations with length. Statistical analyses were run using SYSTAT (SPSS Science, v. 8.0, Chicago, IL). The Type I error rate (alpha) was set at 0.05.
Results
Fig. 4. Lengthwise comparative analysis of bone mineral content (BMC) assessed by microCT of C57BL/6 mouse tibiae showing percent increases in mineral content for 2-week ( SD) and 6-week (+SD) loaded tibiae. All plotted cross-sectional values were spaced ¨3% along the bone length. Increases in mineral content were significant in the proximal metaphysis.
capture newly mineralized bone tissue. Bone mineral and moment of inertia determinations were accomplished using custom code written in MATLAB (MathWorks, v. 6.5, Natick, MA). The cancellous cylinder analysis was performed in Microview (GE Healthcare, v. 1.23). Analysis of data Data summaries are presented for each limb and as percent difference in the loaded limb relative to the contralateral limb. Each variable is summarized as mean and 95% confidence interval. Departures from normality
Comparison of mineral content indicated adaptation of the loaded tibiae compared with the contralateral bones, and the response was site specific. The response declined sharply along the entire length distal to the proximal metaphysis (Fig. 4). Analysis of volumes of interest chosen to isolate local sites responding to mechanical loading indicated that loading induced significantly greater increases in mineral content and volume in the proximal metaphysis than in the mid-diaphysis. Mineral content of the proximal metaphysis increased significantly by 9% after 2 weeks of loading and by 14% after 6 weeks of loading, whereas middiaphysis increases were only 2% for both terms (Table 1). Mineralized volume of the proximal metaphysis significantly increased by 8% after 2 weeks of loading and by 13% after 6 weeks of loading, while mineralized volume of the mid-diaphysis was unaffected by loading (Table 1). The mineral content response in both mid-diaphysis and proximal metaphysis regions was not significantly different between the 2-week and 6-week terms although the
Table 1 Mean values and 95% confidence intervals for mineralized tissue properties and the differences between loaded and control for whole tibia, and two subvolumes of interest Variable
Length (mm) (95% confidence interval) Whole bone mineralized tissue content (mg) Whole bone mineralized tissue volume (mm3) Proximal metaphysis mineral content (mg) Proximal metaphysis mineralized volume (mm3) Proximal metaphysis total volume (mm3) Mid-diaphysis mineral content (mg) Mid-diaphysis mineralized volume (mm3)
2 weeks of loading
6 weeks of loading
Loaded
Control
Difference (%)
Loaded
Control
Difference (%)
17.76 (17.38, 18.13) 33.52. (30.18, 36.86) 23.35. (21.29, 25.41) 1.48. (1.32, 1.63) 1.36. (1.24, 1.47) 1.91. (1.78, 2.04) 0.71. (0.63, 0.79) 0.42 (0.37, 0.46)
17.72 (17.36, 18.07) 32.99 (29.59, 36.39) 23.05 (20.82, 25.27) 1.36 (1.20, 1.51) 1.26 (1.13, 1.39) 1.82 (1.66, 1.99) 0.70 (0.61, 0.79) 0.41 (0.36, 0.46)
0.24 ( 0.13, +0.61) 1.7 ( 1.3, +4.6) 1.4 ( 1.2, +4.0) 9.4 (+1.5, +17.3) 8.1 (+1.5, +14.8) 5.0 (+0.4, +9.7) 1.9 ( 1.0, +4.7) 0.79 ( 2.1, +3.7)
18.00 (17.82, 18.18) 33.40. (31.16, 35.64) 22.93. (21.69, 24.18) 1.53. (1.38, 1.67) 1.34. (1.23, 1.45) 1.78. (1.73, 1.84) 0.71. (0.66, 0.76) 0.42 (0.39, 0.44)
17.95 (17.77, 18.13) 32.24 (30.43, 34.05) 22.17 (21.06, 23.28) 1.34 (1.20, 1.47) 1.18 (1.07, 1.30) 1.62 (1.53, 1.70) 0.70 (0.65, 0.74) 0.41 (0.39, 0.43)
0.27 ( 0.12, +0.65) 3.5 (+1.0, +6.0) 3.4 (+1.8, +5.0) 14 (+9.8, +19.0) 13 (+9.4, +17.2) 11(+7.4, +14.1) 2.0 ( 0.4, +4.3) 1.7 ( 0.2, +3.7)
Proximal metaphysis and diaphysis volumes of interest were 0.46 mm in length. Significance levels by ANOVA: .P < 0.05 versus control, -P < 0.05 versus 2 weeks of loading.
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
Fig. 5. Box plots of regional change in average diaphyseal maximum crosssectional moment of inertia, I max. Increases were significant at the 25% cross-section and were similar after 2-week and 6-week loading terms. All cross-sectional values were averaged from 40 consecutive, tibial microCT slices spaced 11.5 Am apart, representing 0.46 mm of length. The horizontal lines of each box denote the lower quartile, median, and upper quartile values. The open circle marks the mean. The whiskers denote the extent of the rest of the data.
tendency was for longer term loading to result in larger increases. A similar site-specific response was observed for maximum principal cross-sectional moment of inertia (I max). I max was increased at the 25% volume of interest after 2 and 6 weeks of loading by 9% and 8% respectively and there was no difference in the response at 2 versus 6 weeks (Fig. 5). The difference in I max was not significant in the 50% and 75% volumes of interest for either loading term. I min and the orientation of principal axes were not significantly affected at any of the three locations after either loading term. The purely cancellous cylindrical subvolume also had significantly increased mineralized tissue volume (Table 2; Fig. 6). Bone volume fraction (BV/TV) of the cancellous subvolume was greater at both time points: 13% for the 2week group and 15% for the 6-week group (Table 2). The similarity in bone volume fraction increases due to the term of loading indicates that the term had no effect on the overall cancellous mineral response. Trabecular thickness was increased by 12% after 6 weeks of loading but was not significant after 2 weeks of loading indicating that the term of loading had an effect on trabecular structure. However, trabecular separation was not affected by loading. Trabecular separation did indicate age related modeling by an
1035
increase after the 6-week term relative to the 2-week term (Table 2). Modeling in the proximal metaphysis with age was also indicated by a decrease in the total volume at 6 weeks relative to 2 weeks. This volume decreased 7% in loaded and 11% in control limbs (Table 1). At each of the two time points, the loaded proximal tibia had a greater total volume compared with the control side. Additionally, the relative increase in total volume after 6 weeks of loading was significantly greater than after 2 weeks of loading, increasing from 5% to 11% (Table 1). Factors that could have confounded the measured outcomes include animal health, differences in limb usage, differences in tibial lengths, loss of mice, and departures from normal distributions in the data. The mean body mass of the mice was 21.4 g T 1.7 g when received at our facility. During the 3-week acclimation period, their mass increased to 23.8 g T 1.9 g at the start of loading, consistent with a period of rapid growth in the adolescent mouse [33]. All mice maintained or increased body mass over the experimental period, indicative of good health and tolerance to the anesthetic and mechanical loading regimen (at 2 weeks: 24.0 g T 1.8 g; at 6 weeks: 24.6 g T 1.8 g). No signs of lameness were observed between daily loading sessions. Tibial lengths were not significantly different between loaded and control limbs (Table 1). One animal in the 6week group did not complete the study for reasons unrelated to the mechanical loading protocol. The data from an animal in the 2-week group were lost due to an unrecoverable microCT scanning error. No variables were found to depart significantly from a normal distribution.
Discussion We hypothesized that the volume of mineralized tissue would increase in response to controlled cyclic axial loading applied to the adolescent mouse tibia and that the adaptation would be site specific. Indeed, the volume of mineralized tissue increased in the corticocancellous proximal metaphysis. Furthermore, the volume fraction of cancellous bone was increased in the proximal metaphysis. A significantly lower response was noted in the cortical mid-diaphysis than proximal metaphysis.
Table 2 Mean values and differences for bone volume fraction and morphometric measures within loaded and control proximal metaphyseal cancellous volumes of interest Variable
Bone volume fraction (BV/TV) (95% confidence interval) 3D direct trabecular thickness (Am) 3D direct trabecular separation (Am)
2 weeks of loading
6 weeks of loading
Loaded
Control
Difference (%)
Loaded
Control
Difference (%)
0.23 (0.21, 0.25).
0.20 (0.18, 0.23)
13 (+4.8, +21.4)
0.23 (0.20, 0.25).
0.21 (0.17, 0.24)
15 (+1.3, +28)
43.1 (41.6, 44.7) 204 (201, 208)
42.1 (40.6, 43.6) 204 (200, 208)
2.6 (+0.2, +4.9) 0.4 ( 2.4, +3.2)
46.5 (44.2, 48.7). 216 (209, 222)
41.7 (40.0, 43.3) 213 (205, 221)
12 (+5.6, +18)1.6 ( 2.2, +5.3)
The cylindrical volume of interest had diameter of 1.24 mm and height of 0.46 mm. Significance levels by ANOVA: .P < 0.05 versus control, -P < 0.05 versus 2 weeks of loading.
1036
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
Fig 6. Superior view of three-dimensional reconstructions of the proximal metaphysis (top) from paired tibiae of a 6-week loaded mouse. The proximal cut surface of the metaphysis is white. Mineral content of the loaded side is increased 22% over the control side. Bone volume fraction (BV/TV) of the cancellous volume of interest is 23.8% in the loaded and 19.3% in the control, a 23% increase with loading. Scale bar is 1 mm. Sections are ¨0.46 mm thick.
We further hypothesized that adaptation would depend on the term of loading. Most of our data do not support the loading-term dependent hypothesis; results from short-term (2 weeks) and long-term (6 weeks) mechanical loading were similar for most volumes of interest. This lack of effect is in agreement with previous work that has shown both cortical and cancellous formation reverting to control levels after the initial response to mechanical loading [10,17,34]. One explanation for this behavior is that the mechanosensitivity of bone cells diminishes in response to continued loading [34,35], though low anabolic activity of osteoblasts does not necessarily require loss in cellular ability to sense a signal. We did, however, detect two pieces of evidence in the proximal metaphysis in support of our loading term hypothesis. The total volume and trabecular thickness were increased more after 6 weeks of loading than after 2 weeks of loading. These increases could have resulted from increased osteoblastic activity, reduced osteoclastic activity, or both. Site-specific reductions in osteoclastic activity have been demonstrated in vitro and in vivo due to mechanical loads [12,15,16,36]. The failure of trabecular thickness increase to reach statistical significance after 2 weeks of loading may also indicate a detection limit of our measurement technique due to the shorter term combined with our level of loading (¨800 Aq). Our study was not specifically powered for this measurement. Tibial loading of 14-week old B6 mice at a high load magnitude (producing ¨2000 Aq) for 3 weeks also increased the trabecular thickness measured by microCT [28]. Comparing proximal metaphysis results from contralateral bones in the 2-week and 6-week groups showed that the total volume decreased and trabecular separation increased with age (Tables 1 and 2). We believe these changes represent growth-related modeling of the metaphysis. Our low mineral response at the tibial mid-diaphysis (Table 1; Fig. 4) contrasts with previous reports of increased formation due to axial ulnar loading in rodents and tibial
loading in mice. Ulnar studies have utilized peak strains (¨3000 Aq) that were based on vigorous activities such as falling from a ledge (e.g., [9,12 –16,34]). The aforementioned tibial study utilized lower strains (2000 Aq) [28]. The peak cortical strains induced in our study were even lower (¨800 Aq) and within the physiological range found in the fore limbs of mice and rats for activities such as running [14,26]. However, tibial strain during normal locomotion in the B6 may be lower than previously expected (¨300 Aq) [28]. Nonetheless, a site-specific threshold may need to be reached to produce a detectable mineral response and higher strains and rates of loading have been associated with increased cortical formation responses [13,14,21]. The lack of response in moments of inertia at the tibial mid-diaphysis also contrasts with a rat ulna study [34]. In the rat ulna, supra-physiological strains induced the largest response in the distal diaphysis that were greater in minimum (70%) versus maximum (23%) moment of inertia after 16 weeks of loading and were sustained along a large portion of the diaphysis [34,37]. We detected a smaller tibial increase in maximum diaphyseal moment of inertia (9%) in the proximal diaphysis (volume of interest at 25% of length). Minimum moments of inertia were not increased. In addition to the aforementioned strain magnitude and animal difference, the loading term and site were dissimilar to our study. There are important cortical and cancellous structural differences between fore and hind limbs. The rat ulna is more slender with a mid-diaphysis I max:I min ratio of approximately 6 [37], while the ratio in the B6 mouse tibia does not exceed 2 at the mid-diaphysis and 4 at 25% of length. The function of these two bones is different and the rodent ulna is complemented and closely associated with a substantial radius bone, while the fibula, which fuses with the distal tibia, has little load bearing capability. The metaphyses of the tibia also contain more cancellous bone than those of the ulna, which may also reflect differences in function. Our finding of increased bone volume fraction confirms the novel utility of non-invasive in vivo axial loading of the B6 tibia on cancellous bone of the proximal tibia [28]. Another promising avenue for studying cancellous bone formation has been demonstrated due to loading of 9-month old, male rat-tail vertebrae [17]. However, despite the formation gains, in vivo axial compression of the rat tail failed to increase bone volume fraction at the vertebral corticocancellous site. The different response may have age, species, site, loading term, and load related components. We chose to use B6 mice at 10 weeks of age to examine the skeletal response immediately prior to attainment of peak bone mass at 16 weeks [22]. B6 mice do not reach skeletal maturity in terms of cortical bone mechanical properties until the age of 20 weeks [38]. Additionally, in the rat-tail model, the vertebrae were clamped between loading sessions so that any effects of short daily loading periods were superimposed on a background of disuse. Our mice were allowed normal cage activity between loading sessions. Combining the
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
applied mechanical loading approach we have taken with hind limb suspension would provide a similar loading/ unloading regimen for a more direct comparison [2,39]. MicroCT has both strengths and limitations for assessing adaptation. This quantitative method provides a threedimensional assessment of net mineral response resulting from both osteoblastic and osteoclastic activity throughout the tibia. Traditionally, adaptation responses have been measured with two-dimensional histomorphometry using fluorochrome double labels. In contrast, quantitative microCT allows detection of differences in bone volume fraction, mass (g), and morphometric parameters [40]. The method could eventually be augmented with three-dimensional quantification of bone formation volumes [41,42]. The higher resolution of microCT compared with other Xray techniques is critical to localizing the mineral response in bony structures of small animals and facilitating paired analysis by allowing very accurate alignment of loaded versus contralateral bones. Longitudinal growth suppression, which affects alignment ability, was not observed in our study, in contrast to previous reports on axial compression of the rodent ulna [13,14,19,20,34,38]. The use of polychromatic X-rays precludes the elimination of all beam-hardening artifacts. All digital microscopy techniques such as microCT are also subject to partial volume and threshold selection errors. However, as detailed in the Materials and methods section, we addressed beam hardening and digitization errors by filtering of the low-energy portion of the X-ray spectrum, optimizing resolution, and calibrating with a mineral standard. Additionally, we posed our hypothesis to take advantage of a comparative analysis such that each loaded limb was paired against its contralateral control limb to minimize any artifacts. In conclusion, our results indicate that a site-specific bone mineral increase is generated in the proximal metaphysis by a daily regimen of controlled axial loading applied to the B6 mouse tibia prior to peak bone mass attainment. The response in the corticocancellous proximal metaphysis was greater than in the cortical diaphysis. Furthermore, the overall corticocancellous response saturated within 2 weeks of onset of loading while an increase in trabecular thickness was detected only after 6 weeks of loading. This loading model provides an opportunity to more fully understand the role of mechanical loading during the growth of skeletal structures. Elucidation of genetic and mechanical factors important to the attainment of peak bone mass in young adult life and maintenance into maturity will help to eliminate the inadequate accumulation and excessive loss that currently predispose individuals to osteoporosisrelated fracture.
Acknowledgments We thank Dr. Mathias Bostrom and the technical staff at HSS for surgical assistance, Mr. Michael Thornton for
1037
technical assistance with microCT, and Dr. Karl Jepsen for providing a moment of inertia algorithm. This work was financially supported by Cornell University graduate student fellowship and NASA GSRP #NGT5-50375 (JCF), Sigma Xi Grant-in-Aid of Research, NIH (P30AR046121 and S10RR014801), and Clark and Kirby Foundations.
References [1] Carter DR, van der Meulen MCH, Beaupre GS. Mechanical factors in bone growth and development. Bone 1996;18S:5 – 10. [2] Morey-Holton ER, Globus RK. Hindlimb unloading of growing rats: a model for predicting skeletal changes during space flight. Bone 1998;22:83S – 8S. [3] Rubin CT, Lanyon LE. Regulation of bone formation by applied dynamic loads. J Bone Joint Surg 1984;66A:397 – 402. [4] Uhthoff HK, Sekaly G, Jaworski ZF. Effect of long-term nontraumatic immobilization on metaphyseal spongiosa in young adult and old beagle dogs. Clin Orthop 1985;192:278 – 83. [5] Biewener AA, Bertram JE. Structural response of growing bone to exercise and disuse. J Appl Physiol 1994;76:946 – 55. [6] Woo SL, Kuei SC, Amiel D, Gomez MA, Hayes WC, White FC, et al. The effect of prolonged physical training on the properties of long bone: a study of Wolff’s Law. J Bone Joint Surg 1981;63A:780 – 7. [7] Riggs BL, Melton LJ. The worldwide problem of osteoporosis: insights afforded by epidemiology. Bone 1995;17:505S – 11S. [8] van der Meulen MC, Moro M, Kiratli BJ, Marcus R, Bachrach LK. Mechanobiology of femoral neck structure during adolescence. J Rehabil Res Dev 2000;37:201 – 8. [9] Lee K, Jessop H, Suswillo R, Zaman G, Lanyon L. Endocrinology: bone adaptation requires oestrogen receptor-alpha. Nature 2003; 424:389. [10] Srinivasan S, Agans SC, King KA, Moy NY, Poliachik SL, Gross TS. Enabling bone formation in the aged skeleton via rest-inserted mechanical loading. Bone 2003;33:946 – 55. [11] Turner CH, Akhter MP, Raab DM, Kimmel DB, Recker RR. A noninvasive, in vivo model for studying strain adaptive bone modeling. Bone 1991;12:73 – 9. [12] Hillam RA, Skerry TM. Inhibition of bone resorption and stimulation of formation by mechanical loading of the modeling rat ulna in vivo. J Bone Miner Res 1995;10:683 – 9. [13] Mosley JR, Lanyon LE. Strain rate as a controlling influence on adaptive modeling in response to dynamic loading of the ulna in growing male rats. Bone 1998;23:313 – 8. [14] Mosley JR, March BM, Lynch J, Lanyon LE. Strain magnitude related changes in whole bone architecture in growing rats. Bone 1997; 20:191 – 8. [15] Noble BS, Peet N, Stevens HY, Brabbs A, Mosley JR, Reilly GC, et al. Mechanical loading: biphasic osteocyte survival and targeting of osteoclasts for bone destruction in rat cortical bone. Am J Cell Physiol 2003;284:C934 – 43. [16] Torrance AG, Mosley JR, Suswillo RF, Lanyon LE. Noninvasive loading of the rat ulna in vivo induces a strain-related modeling response uncomplicated by trauma or periosteal pressure. Calcif Tissue Int 1994;54:241 – 7. [17] Kim CH, Takai E, Zhou H, von Stechow D, Muller R, Dempster DW, et al. Trabecular bone response to mechanical and parathyroid hormone stimulation: the role of mechanical microenvironment. J Bone Miner Res 2003;18:2116 – 25. [18] Sheng MHC, Baylink DJ, Beamer WG, Donahue LR, Rosen CJ, Lau K-HW, et al. Histomorphometric studies show that bone formation and bone mineral apposition rates are greater in C3H/HeJ (high density) than C57BL/6J (low density) mice during growth. Bone 1999;25:421 – 9.
1038
J.C. Fritton et al. / Bone 36 (2005) 1030 – 1038
[19] Ohashi N, Robling AG, Burr DB, Turner CH. The effects of dynamic axial loading on the rat growth plate. J Bone Miner Res 2002;17:284 – 92. [20] Mosley JR, Lanyon LE. Growth rate rather than gender determines the size of the adaptive response of the growing skeleton to mechanical strain. Bone 2002;30:314 – 9. [21] Turner CH, Takano Y, Owan I. Aging changes mechanical loading thresholds for bone formation in rats. J Bone Miner Res 1995; 10:1544 – 9. [22] Beamer WG, Donahue LR, Rosen CJ, Baylink DJ. Genetic variability in adult bone density among inbred strains of mice. Bone 1996;18:397 – 403. [23] Sheng MH, Baylink DJ, Beamer WG, Donahue LR, Lau KH, Wergedal JE. Regulation of bone volume is different in the metaphyses of the femur and vertebra of C3H/HeJ and C57BL/6J mice. Bone 2002;30:486 – 91. [24] Robling AG, Turner CH. Mechanotransduction in bone: genetic effects on mechanosensitivity in mice. Bone 2002;31:562 – 9. [25] Akhter MP, Cullen DM, Pedersen EA, Kimmel DB, Recker RR. Bone response to in vivo mechanical loading in two breeds of mice. Calcif Tissue Int 1998;63:442 – 9. [26] Lee KCL, Maxwell A, Lanyon LE. Validation of a technique for studying functional adaptation of the mouse ulna in response to mechanical loading. Bone 2002;31:407 – 12. [27] Fritton SP, Rubin CT. In vivo measurement of bone deformations using strain gauges. In: Cowin SC, editor. 2nd edR Bone mechanics handbook, ch. 8. Boca Raton, FL’ CRC; 2001. p. 1 – 41. [28] de Souza RL, Matsuura M, Eckstein F, Lanyon LE. Unique evidence for load induced modification in trabecular bone architecture as well as increased cortical bone formation in a non-invasive axial loading of mouse tibiae. Osteoporos Int 2004;15S1:13. [29] Nagurka ML, Hayes WC. An interactive graphics package for calculating cross-sectional properties of complex shapes. J Biomech 1980;13:59 – 64. [30] Hildebrand T, Ruegsegger P. A new method for the modelindependent assessment of thickness in three-dimensional images. J Microsc 1997;185:67 – 75. [31] Robling AG, Li J, Shultz KL, Beamer WG, Turner CH. Evidence for a
[32]
[33] [34]
[35] [36]
[37]
[38]
[39]
[40]
[41]
[42]
skeletal mechanosensitivity gene on mouse chromosome 4. FASEB J 2003;17:324 – 6. Gross TS, Srinivasan S, Liu CC, Clemens TL, Bain SD. Noninvasive loading of the murine tibia: an in vivo model for the study of mechanotransduction. J Bone Miner Res 2002;17: 493 – 501. Body Weights for Selected Strains of Jax\ Mice. Available at: http:// jaxmice.jax.org/library/charts/bodyweights.pdf. Robling AG, Hinant FM, Burr DB, Turner CH. Improved bone structure and strength after long-term mechanical loading is greatest if loading is separated into short bouts. J Bone Miner Res 2002; 17:1545 – 54. Turner CH. Three rules for bone adaptation to mechanical stimuli. Bone 1998;23:399 – 407. Rubin J, Fan X, Biskobing DM, Taylor WR, Rubin CT. Osteoclastogenesis is repressed by mechanical strain in an in vitro model. J Orthop Res 1999;17:639 – 45. Robling AG, Hinant FM, Burr DB, Turner CH. Shorter, more frequent mechanical loading sessions enhance bone mass. Med Sci Sports Exerc 2002;34:196 – 202. Brodt MD, Ellis CB, Silva MJ. Growing C57B1/6 mice increase whole bone mechanical properties by increasing geometric and material properties. J Bone Miner Res 1999;14:2159 – 66. Simske SJ, Guerra KM, Greenberg AR, Luttges MW. The physical and mechanical effects of suspension-induced osteopenia on mouse long bones. J Biomech 1992;25:489 – 99. Muller R, van Campenhout H, van Damme B, van der Perre G, Dequeker J, Hildebrand T, et al. Morphometric analysis of human bone biopsies: a quantitative structural comparison of histological sections and micro-computed tomography. Bone 1998;23:59 – 66. Waarsing JH, Day JS, van der Linden JC, Ederveen AG, Spanjers C, De Clerck N, et al. Detecting and tracking local changes in the tibiae of individual rats: a novel method to analyse longitudinal in vivo micro-CT data. Bone 2004;34:163 – 9. Kinney JH, Ryaby JT. Resonant markers for noninvasive threedimensional dynamic bone histomorphometry with X-ray microtomography. Rev Sci Instrum 2001;72:1921 – 3.