| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
McCaig Centre for Joint Injury and Arthritis Research, Faculties of 1Kinesiology, 2Medicine, and 3Engineering, University of Calgary, Calgary, Alberta, Canada T2N 1N4
Submitted 22 October 2003 ; accepted in final form 2 January 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
, 30 Hz, 5 days/wk for 3 wk. Left tibiae were the contralateral control condition. Mice were randomly assigned into one of two groups: continuous high-frequency (CT) stimulation for 100 s (n = 9), or 1-s pulses of high-frequency stimuli followed by 10 s of rest (RI) for 100 s (n = 9). Calcein labels were administered on days 1 and 21; label incorporation was used to histomorphometrically assess periosteal and endosteal indexes of adaptation. Periosteal surface referent bone formation rate (pBFR/BS) was significantly enhanced in CT (>88%) and RI (>126%) loaded tibiae, relative to control tibiae. Furthermore, RI tibiae had significantly greater pBFR/BS, relative to CT tibiae (>72%). The endosteal surface was not as sensitive to mechanical loading as the periosteal surface. Thus short-term high-frequency loading significantly elevated pBFR/BS, relative to control tibiae. Furthermore, despite the 10-fold reduction in cycle number, the insertion of rest periods between bouts of high-frequency stimuli significantly augmented pBFR/BS, relative to tibiae loaded continually. Optimization of osteogenesis in response to mechanical loading may underpin the development of nonpharmacological regiments designed to increase bone strength in individuals with compromised bone structures. mouse; bone adaptation; biomechanics
For a given duration of loading, the frequency of loading is proportional to the total number of loading cycles. As the total number of cycles increases, the strain threshold required to maintain bone mass decreases (20). When the number of load cycles approximates 100,000, the strain threshold decreases rapidly and nonlinearly. Cycle numbers approximating that order of magnitude are typically reached with high-frequency loading. Thus high-frequency loading could be useful for maintaining bone mass. After 28 days of disuse, Rubin et al. (27) found that bone formation rates were suppressed 92%, in the proximal tibiae of skeletally mature rats relative to control values. Ten minutes of weight bearing per day slightly curbed the observed suppression in bone formation rates. Conversely, 10 min/day of low-magnitude, high-frequency (90 Hz) stimuli normalized bone formation rates to control values.
In addition to the ability of high-frequency loads to maintain bone mass, high-frequency, low-magnitude loading can be a potent anabolic stimulus. Rubin and colleagues (26) measured a 34% increase in bone density in the proximal femoral region in sheep, relative to control animals after 20 min/day of low-level (0.3 g), high-frequency (30 Hz) stimuli. Similar results were reported for the proximal tibiae of C57BL/6 mice, after 10 min/day of 0.25 g high-frequency (45 Hz) stimuli (13).
One factor to consider when designing loading regimes to boost skeletal mass is that adaptive osteogenesis in response to mechanical loading saturates early. Rubin and Lanyon (28) found that osteogenesis in avian ulnae did not increase when the number of loading cycles per day was increased from 36 to 1,800 (50-fold increase). Similarly, Umemura and colleagues (40) showed that rats trained to jump 100 times per day did not significantly improve their hindlimb adaptive responses over rats trained to jump 40 times per day. These studies demonstrated that tissue sensitivity to mechanical loading can be proportional to 1(N + 1)-1, where N is the number of consecutive loading cycles (39).
Recently, it has been found that the insertion of rest periods within loading regimes can circumvent adaptive response saturation (21, 36). Rest periods can be long term or short term. Long-term rest periods entail separating a loading regime into a number of discrete bouts. Robling et al. (21) found that separating a 360-cycle loading regime into discrete bouts significantly increased osteogenesis and that increases in osteogenesis were positively correlated with the number of discrete bouts. Sixty cycles 6 times/day boosted endocortical lamellar bone formation rates 10-fold relative to control bone, whereas 180 cycles 2 times/day only increased bone formation rates 4-fold. Additionally, it has been reported that a recovery period of 8 h was sufficient to restore full mechanosensitivity to saturated bone cells (22).
Short-term rest periods entail separating consecutive load cycles with brief periods of rest. Srinivasan and Gross (35) found that insertion of 10-s rest periods between sequential 1-Hz loading cycles transformed a mild low-frequency loading regime into a potent anabolic stimuli by activating additional periosteal surfaces on isolated rooster ulnae. Similarly, Srinivasan et al. (36) found that insertion of 10-s rest periods within a low-frequency loading regime augmented periosteal osteogenesis in mice tibiae, relative to mice loaded continually for the same duration. Affirming that 10-s rest periods can enhance osteogenesis in low-frequency loading regimes, Robling and colleagues (22) found that rest periods of 
7 s do not boost osteogenesis, whereas 14 s of rest significantly augmented rat tibial bone formation rates (22).
Given that high-frequency loads can be osteogenic and that rest insertion can augment osteogenesis, our hypotheses were twofold. First, we hypothesized that high-frequency loading would be osteogenic to loaded bone in contrast to normal control bone. Second, we hypothesized that insertion of rest-periods in a continuous high-frequency loading regime would augment bone adaptation relative to a continuous high-frequency loading regime.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
Murine loader and loading regime. To apply the exogenous loads, a noninvasive in vivo murine loader was constructed after the design of Gross et al. (10) (Fig. 1). Loads were applied with a linear force actuator (model AFX10-200, Motran Industries, Valencia, CA) controlled by a 16-bit data-acquisition processor (model DAP 820, Microstar Laboratories, Bellevue, WA). DAPview (version 1.12, Microstar Laboratories) was used to concatenate customized waveform input to loader output. In both groups, loads were applied to right tibiae at 30 Hz for 100 s, 5 days/wk, for 3 wk. Short-term (100 s) loading periods minimized the potential adverse effects of chronic anesthesia. The RI group had loads applied in 1-s pulses followed by 10-s periods of rest (Fig. 1). That pulse and rest pattern was repeated for the 100-s loading duration. All loading was conducted under halothane anesthesia (5% in air for induction, 2% in air for maintenance). Induction to recovery took <5 min per mouse. Loads were calibrated with ex vivo strain gauges placed 1 mm proximal to the distal tibia-fibula junction on the lateral right tibial middiaphysis (model FLK 1-11, TML, Tokyo Sokki Kenkyujo, Tokyo, Japan) on 2 calibration mice (10 trials per mouse) to induce peak strain magnitudes of 800 µ (coefficient of variation = 14%). Strain gauges were attached with cyanoacrylate strain gauge adhesive (CN adhesive, TML, Tokyo Sokki Kenkyujo). Calcein injections (10 mg/kg) were administered intraperitoneally on days 1 and 21 (Fig. 2). Calcein is a calcium chelator and thereby adheres to regions of new bone mineralization. After the 3-wk loading protocol, mice were allowed 1 wk of normal cage activity to allow consolidation of newly deposited bone.
|
For a mechanical loading regime to maintain bone mass, a strain threshold (magnitude) must be surpassed, and there is a nonlinear dependence of that strain threshold on the total number of daily loading cycles (20). On the basis of an algorithm presented by Qin et al. (20) (Eq. 1), for our loading regime consisting of 3,000 (100 s · 30 Hz) daily loading cycles to maintain bone mass, the strain magnitude had to reach 589 µ per load cycle. For a given cycle number, osteogenesis was proportional to strain magnitude (29). Hence, we chose 800 µ because it was above the projected strain threshold (589 µ) required to maintain bone mass and thus was potentially osteogenic.
Waveform analysis. Insertion of rest periods within continuous stimuli altered the frequency content of the waveforms. To determine the power spectral density of the customized waveforms, waveform inputs were subjected to fast Fourier transformations (Matlab, Natick, MA) and subsequently plotted as periodograms.
Tissue processing. After successful completion of the 4-wk loading protocol, mice were euthanized via CO2 hypoxia. Tibiae were dissected and cleaned of adherent nonosseous tissue. A 350-µm section of the tibia was taken with a diamond-wafer saw (blade: 101.6 mm x 12.7 mm, arbor x 0.012, E. T. Enterprises, Calgary, AB, Canada; saw: Buehler Isomet, Lake Bluff, IL) corresponding to where calibration strain gauges were attached on calibration mice. Sections were subsequently ground to 50 µm with 1,500 grit sandpaper using 70% ethanol as a lubricant. Sections were mounted (Polymount, Fisher Scientific, Edmonton, AB, Canada) and viewed at x120 light microscopy with a mercury-vapor fluorescent light source (blue excitation, 400-800 nm). Digital images were captured (model DXC-950P, Sony, Tokyo, Japan), and assembled in a collage (version 7.0, Adobe Photoshop, San Jose, CA). All procedures were conducted under the ethical aegis of the University of Calgary Animal Care Committee.
Histomorphometry. Images were analyzed histomorphometrically (Image J, Frederick, MD) in accordance with previously established methods (15). Our pilot data revealed that contralateral control tibiae (left) suitably represented the sham-loaded condition (right tibiae) and that chronic halothane anesthesia did not alter histomorphometrical indexes of bone adaptation (data not presented). For both the periosteal and endosteal surfaces, the following measurements were made: circumference, area, single-label surface distance (sLS.Pm), double-label surface distance (dLS.Pm), and double-label surface area (Ir.L.Ar). Those measures were used to calculate mineral apposition rate, [MAR = (Ir.L.Ar/dLS.Pm)/Ir.L.t.], where Ir.L.t. is the interlabel time period; mineralizing surface, [MS = (dLS.Pm + sLS.Pm/2)/BS], where BS is the endosteal or periosteal bone surface; and surface referent bone formation rate, BFR/BS = MAR·MS/BS. MAR represents the speed that new bone is deposited radially, MS represents the proportion of the surface with active bone formation, and BFR/BS represents the MAR, corrected to the proportion of the surface forming new bone. All analyses were done with the observer blinded to the randomized sample identity.
Statistics. Differences between control (left) and loaded (right) tibiae were assessed with Wilcoxon's signed-rank tests. Differences between CT and RI groups were assessed using Mann-Whitney U-tests. For all tests, P 0.05 was considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
Periosteal response. High-frequency loading significantly increased periosteal MAR, relative to control tibiae in the CT (>48%) and the RI (>91%) groups (Fig. 4). Additionally, the periosteal MAR in RI tibiae was enhanced relative to CT-group tibiae (>54%). However, that enhancement was not significant (P = 0.06). In contrast, loading did not enhance the periosteal MS relative to control tibiae, and there were no significant differences between tibiae in the CT and the RI groups (Fig. 5). BFR/BS is a measure of MAR corrected to the active surface of bone. Loading significantly enhanced the periosteal BFR/BS relative control tibiae in the CT (>88%) and the RI (>126%) groups (Fig. 6). Furthermore, loaded tibiae in the RI group had a significantly greater mean periosteal BFR/BS (>72%) relative to tibiae loaded in the CT group.
|
|
|
Endosteal response. The endosteal surface was less sensitive to loading than the periosteal surface. Tibiae in the CT group exhibited increases in endosteal MAR, MS, and BFR/BS relative to control tibiae, but those increases were not significant (Table 1). Conversely, tibiae in the RI group had significantly greater MAR, MS, and BFR/BS relative to control tibiae (Table 1). Endosteal indexes of bone adaptation were not significantly greater in tibiae loaded in the RI group relative to tibiae loaded in the CT group.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Periosteal vs. endosteal responses. In the present study, mechanical stimulation resulted in a periosteal expansion, and endosteal osteogenesis was not as predominant as periosteal expansion. That finding was consistent with previous reports (10, 16, 36). It has been found that BFRs on the endosteal and periosteal surfaces of C57BL/6 mice femora change with age (5). Thus our observed differences in endosteal and periosteal adaptation may pertain to age-related changes in surface activity. That premise, however, may be flawed because it has been shown that femoral periosteal and endosteal MAR and BFR pace each other until 22 wk of age (5). We assessed tibiae, and thus those conclusions remain equivocal. Alternatively, those differences may have been a result of different mechanical thresholds for the endosteal and periosteal surfaces, as proposed by Frost and others (8, 33).
When intracortical fluid pressure rises in response to mechanical loading, fluid potentially will flow toward the marrow or the periosteal surface (19). The endosteal surface provides an open circulation between marrow pressure and intracortical flow, whereas flow from the cortical bone to the periosteal surface is restricted by the periosteum (19). Potentially, the forced flow of fluid past the highly cellular periosteum more effectively stimulated bone cell recruitment. The endosteal surface lacks a periosteum, possibly making cell recruitment more difficult.
Waveform analysis. The power spectral density of the RI waveform was dramatically different than that of the CT waveform. Insertion of 10-s rest periods substantially reduced the power of the 30-Hz frequency components and gave rise to frequency components near zero (Fig. 3). The new frequency components near zero had the greatest power of all frequency components in the RI group. Although the components near zero may have been responsible for the significant enhancement in adaptive osteogenesis relative to continuous stimuli, it was not likely. With such low frequencies, strain magnitudes would need to be much larger than the strains induced in the current study to elicit an adaptive response. Qin and coworkers (20) presented an algorithm demonstrating the relation between loading cycle number and the strain magnitude required to maintain bone mass (Eq. 1)
 |
where x represents the daily loading cycle number and where y is the resultant strain magnitude required for maintaining bone mass (20). On the basis of that algorithm, for a frequency of 1 Hz and equivalent loading duration of 100 s, the strain magnitude required to maintain bone mass would be 1,302 µ. For a frequency of 2 Hz, the strain magnitude must have reached 1,142 µ to maintain bone mass. Accordingly, strains higher than those values would have to have been applied for the low-frequency components to elicit an osteogenic response.
With use of that same algorithm, the RI waveform (3,000 total daily cycles) would have required strain magnitudes of >1,052 µ to maintain bone mass. The strains applied in the present study were well below that threshold (800 µ) and, furthermore, were patently osteogenic. That algorithm (Eq. 1) only applies to continuous stimuli, and one can conclude that rest insertion had a significant osteogenic potential despite the substantial reduction in load cycles. With rest insertion, fewer cycle numbers induced greater bone growth.
Osteogenic potential of high-frequency loading. In the present study, high-frequency stimuli were effective to mount an adaptive response to cortical bone. That was in contrast with studies reporting that high-frequency stimuli are osteogenic to trabecular, but not cortical, bone (25, 26). One critical difference between our study and those studies was that our peak strain magnitudes were 800 µ, applied for 100 s, whereas the peak strain magnitudes of Rubin and coworkers (25, 26) approximated 5 µ and were applied for 1,200 s. We may have observed a cortical osteogenic response because of our larger peak strain magnitude. A strain magnitude of 800 µ was well within the daily strain history of a bone (7). The stimulation frequency (30 Hz) was significant in that it was similar to the contractile spectra of muscle (12). To our knowledge, the present study is the first to report the osteogenic potential of short-duration (100 s), high-frequency loads to cortical bone.
Although the exact mechanism remains elusive for explaining why high-frequency loading can be a potent anabolic stimulus, some suggestions have been proposed. The sensitivity of high-frequency loading may be a function of the high number of loading cycles associated with high-frequency loading [Eq. 1; (20)]. Perhaps there is a preferential sensitivity of bone cells to higher frequency loading (30). Alternatively, the osteogenic potential of high-frequency loading may be explained by the indirect by-products of high-frequency stimuli such as fluid flow and intramedullary pressure (26).
Osteogenic potential of rest insertion. The data presented in the present study showed that the insertion of 10-s periods of rest between 1-s high-frequency pulses significantly augmented periosteal BFR/BS relative to a continuous, high-frequency stimulus. Other reports indicated that insertion of rest periods of 
10 s within low-frequency loading regimes significantly augmented osteogenesis in turkey ulnae (35, 36), mice tibiae (36, 37), and rat tibiae (22) and that those periods of rest must be >7 s to enhance osteogenesis (22). Srinivasan et al. (36) found that the insertion of 10-s rest periods within a low-frequency loading regime augmented periosteal MARs 
0.4 µm/day on skeletally immature C57BL/6 mice tibiae. Our results showed that the insertion of rest periods within a high-frequency loading regime also augmented MARs 0.4 µm/day on skeletally mature C57BL/6 mice tibiae.
Physically, our data supported current hypotheses that fluid flow affected bone adaptation. Experimental and mathematical studies have shown that bone pore fluid pressure relaxation was 1.5 s (32, 41). Thus, at a loading frequency of 30 Hz, there will not be sufficient time for full fluid relaxation between loading cycles. Consequently, load-induced fluid flow near osteocytes would be substantially reduced in the subsequent cycles beyond the first few load cycles. Srinivasan and Gross (34) calculated those reductions to be >45%. On the basis of their interpretations, that reduction was related to the forced flow of viscous interstitial fluids through canaliculi and consequent "inertia" of the fluid flow (36). Presumably, periods of rest between loading cycles would diminish inertial fluid flow effects and enhance fluid flow near the osteocytes in subsequent cycles.
Biologically, one of the earliest events in bone cell mechanotransduction can be intracellular calcium signaling (3, 42). Elevations in intracellular calcium occurred 15 s after continual fluid stimuli (4).
Donahue et al. (3) found that periods of rest were required to "resensitize" bone cells to mechanical stimuli. Cellular refractory periods have been suggested to relate to cytosol calcium diffusion and intracellular calcium concentration refilling kinetics, cytoskeletal disorganization, and G protein-coupled resensitization (1, 3, 22).
Limitations. The data of the present study were limited because two samples from the left tibiae CT group in the two samples from the left tibiae RI group were fractured and rendered unusable during specimen preparation. Surrogate contiguous sections were not taken to replace those sections, because the geometry of tibia changed rapidly along the longitudinal axis of the tibia near our region of interest. Accordingly, the surrogate sections would be of different geometry and would not be directly comparable with the other sections. For endosteal MAR, MS, and BFR to be significantly greater in the CT loaded tibiae, relative to control tibiae, 12, 47, and 8 specimens would have been needed per group (Table 1).
Because our loading regimen was conducted under halothane anesthesia, loads were applied for a short duration (100 s) to minimize the potential adverse effects of chronic anaesthesia. Accordingly, our strain magnitudes were larger than those induced in other high-frequency studies (e.g., Ref. 13). Further research is warranted to investigate the effect of rest insertion in combination with high-frequency loading for lower strain magnitudes and for trabecular bone.
Clinical significance. Mechanical loading regiments designed to induce osteogenesis are an attractive therapy to combat age-related bone pathologies characterized by low bone mass. Some exercises, such as high-impact loading, associated with significant osteogenesis, however, could be pathological to a fragile skeleton. High-frequency vibrations are an attractive loading alternative for inducing osteogenesis, because they can be applied at low magnitudes and remain a potent osteogenic stimuli (e.g., Ref. 26). A vibrating platform has been designed for human use (6), and preliminary data from a prospective, randomized, double-blind, placebo-controlled clinical trial that used these platforms confirmed the anabolic potential of short-term, high-frequency stimuli in the human condition (31). Our present data also suggested that high-frequency loading regimes could be used in combination with the insertion of rest periods to augment osteogenesis.
Summary. In summary, the significance of our results was threefold. First, we demonstrated that short-term continuous high-frequency loading could be osteogenic to skeletally mature cortical bone. This osteogenesis could be directly related to the large number of loading cycles associated with high-frequency loading, preferential sensitivity of bone cells to high-frequency stimuli, or indirectly to the by-products of high-frequency loading such as intracortical fluid flow and intramedullary pressure perturbations. Second, and in congruence with previous research, we showed that, despite a 10-fold reduction in loading cycles, rest insertion with loading could augment bone adaptation relative to continuous stimuli. Enhanced osteogenesis with the insertion of rest periods could indirectly relate to biological mechanisms including cellular refractory periods, or physical mechanisms including enhanced intracortical fluid flow. Last, we demonstrated that rest insertion could be effective in combination with high-frequency loading regimes.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. L. Manske, C. R. Lorincz, and R. F. Zernicke Bone Health: Part 2, Physical Activity Sports Health: A Multidisciplinary Approach, July 1, 2009; 1(4): 341 - 346. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. von Stengel, W. Kemmler, W. A Kalender, K. Engelke, and D. Lauber Differential effects of strength versus power training on bone mineral density in postmenopausal women: a 2-year longitudinal study Br. J. Sports Med., October 1, 2007; 41(10): 649 - 655. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Srinivasan, B. J. Ausk, S. L. Poliachik, S. E. Warner, T. S. Richardson, and T. S. Gross Rest-inserted loading rapidly amplifies the response of bone to small increases in strain and load cycles J Appl Physiol, May 1, 2007; 102(5): 1945 - 1952. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T S Gross, S Srinivasan, and Y-X Qin Building bone mass through exercise: could less be more? Br. J. Sports Med., January 1, 2006; 40(1): 2 - 3. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
