Aged bones have been considered to have reduced capacity to respond to changes in incident loading. By subjecting young and adult rats to increased loading and subsequent deconditioning, we observed quantitatively similar adaptive responses of bone in these two groups, but young skeletons adapted primarily through geometric changes and adult bones through increased volumetric density. Loss of the exercise-induced bone benefits did not depend on age.
Introduction: Aging has been shown to decrease the sensitivity of the mechanosensory cells of bones to loading-induced stimuli, presumably resulting in not only reduced capacity but also different adaptive mechanism of the aged skeleton to altered loading, as well as poorer capacity to preserve exercise-induced bone benefits.
Materials and Methods: Fifty young (5-week-old) and 50 adult (33-week-old) male rats were randomized into control and exercise (+deconditioning) groups. After a 14-week progressively intensified running program, one-half of the exercised rats (both young and adult) were killed, and the remaining rats underwent subsequent 14-week period of deconditioning (free cage activity). A comprehensive analysis of the femoral neck was performed using peripheral quantitative computed tomography and mechanical testing.
Results: In comparison with the controls, both young and adult exercised rats had significant increases in almost all measured parameters: +25% (p < 0.001) and +10% (not significant [NS]) in the cross-sectional area; +28% (p < 0.001) and +18% (p < 0.001) in bone mineral content; +11% (p < 0.05) and +23% (p < 0.001) in bone mineral density; and +30% (p < 0.01) and +28% (p < 0.01) in the breaking load, respectively. The skeletal responses were not statistically different between the young and adult rats. After the 14-week period of deconditioning, the corresponding exercised-to-controls differences were +17% (p < 0.05) and +10% (NS), +18% (p < 0.05) and +13% (p < 0.05), +2% (NS) and +2% (NS), and +11% (NS) and +6% (NS), respectively. Again, the response differences were not significant between the age groups.
Conclusion: Quantitatively, the capacity of the young and adult skeleton to adapt to increased loading was similar, but the adaptive mechanisms appeared different: growing bones seemed to primarily display geometric changes (increase in bone size), whereas the adult skeleton responded mainly through an increase in density. Despite this apparent difference in the adaptive mechanism, aging did not modulate the ability of the skeleton to preserve the exercise-induced bone gain, because the bone loss was similar in the young and adult rats after cessation of training.
Age is speculated to modulate the skeletal sensitivity to mechanical loading, because exercise interventions have been shown to induce significant bone gains in young (growing) individuals while hardly resulting in any increase in mature skeleton and only preserving, at best, the existing bone stock in the elderly people.(1–8) Human studies thus quite uniformly suggest that adolescence (puberty) provides a unique opportunity to intervene with loading exercise.(2,7,9–11) The reduced capacity of the aged skeleton to respond to changes in the loading environment has been attributed to decreased sensitivity of the mechanosensory cells of the bones to loading-induced stimuli,(12) although a very recent study on human bone cells found no evidence for the loss of mechanosensitivity with donor age.(13) The existing in vivo experimental data on the age dependence of the skeletal responsiveness to external loading is far more controversial, because studies have shown the responsiveness of the aged skeleton to be increased,(14) reduced,(15) or unaffected.(16,17)
In an attempt to summarize the apparent confusion regarding the modulatory effect of aging on the sensitivity of bones to respond to loading, it has been proposed that during the longitudinal growth period and particularly during puberty, increased loading can produce actual structural changes in bone (altered bone geometry), whereas additional bone acquired after skeletal maturity is probably deposited along the existing bone structures.(2,3,18) Furthermore, it has been assumed that the adolescence-obtained structural bone changes may, in turn, be at least partly preserved despite decreased loading,(19,20) whereas the maintenance of the additional bone acquired by loading after the skeletal maturity has been speculated to require persisting activity.(1,21–23) However, solid scientific evidence confirming or opposing these assumptions are lacking.
This paper completes our series of three individual experiments on the effects of various factors modulating the skeletal responsiveness to mechanical loading.(23,24) In the current study, 100 rats from two age groups (the 5- to 19- and 33- to 47-week-old correspondingly named as “young” and “adult”) were used to answer to three study questions. (1) Does the skeletal response to loading differ quantitatively between young and adult rats? (2) Are there qualitative differences in the adaptation of young and adult bones to increased mechanical loading? (3) Does age modulate the ability of the skeleton to maintain the loading-induced skeletal changes?
MATERIALS AND METHODS
Animals and sample preparation
A total of 100 3-week-old male littermates of the Sprague-Dawley rat strain were used in the study. During the first 2 weeks of the study, all rats were run on a flat-bed treadmill at a slow speed (10–20 cm/s) for 3 minutes/day for 3 days a week to let the animals to adapt to the treadmill running and to remove those animals refusing to run. At the beginning of the exercise intervention, the rats were 5 weeks old, with a body weight of 145 ± 21 g.
After the above described 2-week acclimatization period, the rats were randomly assigned into eight groups: four control groups (C1, C2, C3, and C4) and four exercise groups (EX1, EX1 + DC, EX2, and EX2 + DC; Fig. 1). The control groups consisted of 10 animals and exercise groups consisted of 15 animals. The animals were housed in cages (18 × 35 × 55 cm), five animals per cage, at 20°C with a light cycle of 12 h. They were fed standard laboratory chow and water ad libitum. The research protocol was accepted by the Ethics Committee for Animal Experiments of the University of Tampere.
The control rats moving freely in their cage were killed after 14 (C1), 28 (C2), 42 (C3), and 56 weeks (C4). The animals were killed with carbon dioxide inhalation, and body weights were measured. The right calf muscles (gastrocnemius, soleus, and plantaris) were dissected, and the wet weights were measured. Both femora were excised and stored: the right femur was stored at −20°C in small freezer bags wrapped in saline-soaked gauze bandages to prevent dehydration, and the left femur was placed in 70% ethanol solution. This kind of storage has been shown not to affect bone's biomechanical properties.(25,26) The killing and preparation procedures were similar for each rat in all study groups.
Exercise and deconditioning
All exercise groups (EX1, EX1 + DC, EX2, and EX2 + DC) were subjected to an identical, progressive exercise program for 14 weeks (Table 1); the training began at the age of 5 weeks in the EX1 and EX1 + DC groups and at the age of 33 weeks in the EX2 and EX2 + DC groups. The exercise program consisted of running on a rodent treadmill once a day, 4 days a week. During the first week of the program, the rats ran 5 minutes at a treadmill speed of 20 cm/s and an inclination of 5°. The program was progressive, and after the first week, the running time was increased to 10 minutes per session, which was maintained till the end of the program. The speed of the treadmill and the uphill inclination were gradually increased so that the speed of 30 cm/s was achieved in week 4 and an inclination of 30° in week 9.
Table Table 1. Progressive Exercise Regimen of the Study
After the exercise period of 14 weeks, the exercised animals in groups EX1 and EX2 were killed, and specimen were collected and stored as described earlier. The remaining exercise groups (EX1 + DC and EX2 + DC) underwent a deconditioning period of 14 weeks, during which the rats were allowed to move freely in the cage until death at the age of 33 and 61 weeks of age, respectively.
Femoral neck analysis
At the day of testing, the femora were slowly thawed at the room temperature and kept wrapped in saline-soaked gauzes except during measurements.
Peripheral quantitative computed tomography
The cross-sections of the femoral necks were scanned with a commercial peripheral quantitative computed tomography (pQCT) system Stratec XCT 960A with software version 5.20 (Stratec Medizintechnik GmbH, Birkenfeld, Germany). The bones were cut approximately at the midshaft, and the proximal part of the femur was inserted, with the femoral neck in an axial direction, into a specially constructed plastic tube for the measurement. The scan line was adjusted to the midneck using the scout view of the pQCT software. The cross-sectional image of the femoral neck was scanned with a voxel size of 0.092 × 0.092 × 1.25 mm3. Total cross-sectional area (tCSA), total bone mineral content (tBMC), and total bone mineral apparent density (tBMD) (that is, the BMC divided by the total volume within the periosteal envelope) at the femoral midneck were recorded as given by the pQCT software. Each bone was measured twice and the average of these measurements was taken as the outcome variable. In our laboratory, the CVrms was 3.9% for the tCSA, 2.2% for the tBMC, and 2.1% for the tBMD.
The femoral necks were subjected to compression testing using a Lloyd material testing machine (LR5K; J.J. Lloyd Instruments, Southampton, UK) according to protocol described previously in detail.(27,28) Briefly, the femur was cut approximately at the midshaft, and the proximal half of the femur was collected and mounted in a specially constructed fixation device for the compression test.(29) The specimen was then placed under the materials testing machine, and a vertical load was applied to the top of the femoral head using a brass crossbar until failure of the femoral neck. The breaking load (Fmax) of the femoral neck was determined from the load-deformation curve. In our laboratory, the CVrms of the Fmax was 6.0%.(27)
All data were analyzed using the SAS statistical program. Relative exercise and deconditioning effects (i.e., the percent difference between exercised/deconditioned and control groups) were both tested using the two-way ANOVA. The group assignment (exercised/deconditioned, control) and age (young, adult) were used as factors. The possible difference in the response to mechanical loading or deconditioning between young and adult rats was assessed using the log-transformed variables followed by the antilog-transformation of the parameter estimates. To eliminate the inherent bias arising when comparing experimental groups that differ in body weight and size (i.e., control versus exercised animals and young versus adult animals), all the data pertaining to the characteristics of the femoral neck were equalized in terms of the animals' apparent loading environment by adjusting the results with a representative of the lean body mass (here the weight of the calf muscle), the principal origin of natural loading of the skeleton.(30,31) In all tests, an α level less than 5% (p < 0.05) was considered significant.
Two rats were lost (in the group EX2 + DC) during the 56-week study period, leaving 13 animals in that group. Figure 2 shows the weight development curves of the rats in the different study groups. The results of the femoral neck pQCT and mechanical strength analyses are summarized in the Table 2. Percent differences between the young and adult exercised-deconditioned groups versus corresponding control groups are presented in Figs. 3 and 4A–4D.
Table Table 2. Femur Length, Calf Muscle Weight, and Raw and Muscle-Weight Adjusted Values of pQCT Measurements and Compression Test of the Femoral Neck of the Exercised (+Deconditioned) and Control Rates (mean ± SD)
After the 14-week exercise period, the body weights of the exercised animals were clearly lower in both the young (−9%, p < 0.05) and adult (−13%, p < 0.05) rats compared with the corresponding controls (Fig. 2). However, no significant difference was observed in the lengths of the femora between these groups (Table 2).
In the pQCT analysis of the femoral neck, higher values were observed in all recorded parameters of both the young and adult exercised rats compared with their controls; the tCSA was +25% (p < 0.001) and +10% (NS), the tBMC was +28% (p < 0.001) and +18% (p < 0.001), and the tBMD was +11% (p < 0.05) and +23% (p < 0.001) for the young and adult exercised rats, respectively (Figs. 3A–3C). Furthermore, the mechanical strength of the femoral neck (the maximum load, Fmax) was +30% (p < 0.01) and +28% (p < 0.01) higher in the exercised rats than controls in both the young and adult groups, respectively (Fig. 3D). However, no significant age-related effect (young versus adult) could be seen in any of the measured parameters (Table 2; Fig. 3).
The observed retardation in weight development in both young and adult exercised rats during the 14-week exercise intervention was fully compensated for by an accelerated weight gain during the detraining period, and accordingly at the end of the detraining period (at 28 and 56 weeks, respectively), no significant differences in body weights were observed between the exercised-deconditioned and control groups in young or adult rats (Fig. 2).
Subsequent to the 14 weeks of deconditioning, both the young and adult rats showed a similar, clear decrease in all measured bone parameters. Some of the exercised-induced benefits were still preserved, because the tCSA was +17% (p < 0.05) and +10% (not significant [NS]) in the previously exercised young and adult rats compared with their controls. The corresponding values were +18% (p < 0.05) and +13% (p < 0.05) for the tBMC and +11% (NS) and +5% (NS) for the Fmax, whereas the exercise-induced gain in tBMD had disappeared by the end of the detraining period (+2% and +1%, respectively, both NS). Similarly to the exercise effects presented above, no age-related difference was observed in the preservation of the exercise-induced bone gain (Table 2; Fig. 4).
Regarding the first objective of the study, we found that the magnitude of the skeletal response to increased mechanical loading did not differ between young and adult rats, because the treadmill training-induced benefits in the bone mass and strength were virtually identical in these two age groups of rats. In respect to the previous studies, the existing experimental in vivo data on the effect of age on skeletal responsiveness to mechanical loading is rather inconsistent: Raab et al.(16) reported a comparable skeletal response to exercise in young and old (2.5 and 25 months old, respectively) rats, but used a different running velocity in the two age groups, thus somewhat hampering valid comparisons. Umemura et al.(17) also reached a conclusion that the effects of exercise are not limited by age when comparing rats of 3, 6, 12, 20, and 27 months of age subjected to both jump training and running. However, although the parameters evaluated in this study were somewhat different to those generally used (again complicating comparisons with other studies), a trend for better response in the younger rats could be detected.
In contrast, Rubin et al.(12) showed, using their classic experimental model of externally loadable functionally isolated turkey ulna preparation, that a physical signal clearly osteogenic in the 1-year-old young adult skeleton was hardly acknowledged in the older (3-year-old) bone tissue. Similarly, Turner et al.(15) observed that both the periosteal and endocortical surfaces of the tibias of 19-month-old rats were significantly less responsive to mechanical loading than those of 9-month-old rats. However, the use of historical controls and inappropriate statistical comparisons diminish the strength of this study. In agreement with these two studies, Dehority et al.(32) used a completely opposite model to increased loading, skeletal unloading by hindlimb suspension, to show that the effects of nonweight bearing are prolonged and have a greater relative effect on bone formation in the adult than in the young growing rats. To add yet another dimension to this discussion, Buhl et al.(14) recently reported that 22-month-old male rats had a greater sensitivity to squatlike exercise than their younger counterparts (4- and 12-month-old male rats).
A simple quantitative comparison of the (magnitude of) responses between our young (5- to 19-week-old) and adult (8- to 11-month-old) rats may oversimplify the issue. Although the responses between the two age groups were quantitatively highly comparable (+30% and +28% in the failure load and total BMC in the young versus +28% and +18% in the adult, respectively), an apparent trend for difference in mechanisms of adaptation was observed: The young rats displayed striking increase in the total cross-sectional area (+25%, p < 0.001) while the volumetric bone density of the femoral neck showed only a moderate increase (+11%, p < 0.05), whereas a practically reciprocal/opposite response was observed in the adult rats (+10%, NS and +23%, p < 0.001, respectively). The fact that this apparent difference in the adaptation mechanisms to exercise in the young and adult skeletons did not reach statistical significance exemplifies the demanding nature of studies comparing difference in skeletal responsiveness between two or more study groups: Despite the precise methods (CVrms of the pQCT ranging between 2.1% and 3.9%) and the relatively large study groups used in this study (n = 10 animals in the control and 15 in the exercise groups), differences in response as obvious as those presented above still do not suffice to be statistically significant.
To summarize our findings concerning the effect of age on the skeletal sensitivity to increased loading, our results quite clearly show that quantitatively the responses are identical between young and adult rats. Regarding the qualitative aspects (i.e., the type) of the response, although we could not unambiguously corroborate the previous hypothesis that older bones are less capable of responding through geometrical changes,(18) our findings certainly provided support for the assumption that growing bones adapts to increased loading predominantly through geometrical changes (increased cross-sectional area), whereas the mature bone responds mainly through condensation of bone mineral within the existing bone structure, presumably because of lacking capacity to expand periosteally. In considering both of the above reviewed primary objectives, one also has to keep in mind that, although our adult rats were over 8 months old at the beginning of treadmill training, they still grew in size (as shown by the increase in the body weight, Fig. 2), and thus, they probably do not represent a true “old” counterpart to the young rats used in this study. It is possible that both a quantitative and qualitative difference in the osteogenic responsiveness, similar to studies by Rubin et al.(12) and Turner et al.,(15) could have been demonstrated with older animals that have a clearly reduced or even completely lack the ability to respond through periosteal expansion.
Unfortunately, the continuing growth of the adult rats (despite being clearly at the descending phase of growth) also complicates the interpretation of our data regarding the third objective of this study: the possible difference in the ability of young and adult bones to maintain exercise-induced benefits after cessation of training. Theoretically, it has been speculated that the adolescence-obtained exercise-induced bone benefits would be better maintained than those obtained in adulthood, because the former is mostly geometrical (increased bone size/girth) in nature, and thus, unlikely to become abolished (i.e., bone “shrinkage”) because of decreased loading, which is the case in the corresponding extra condensation of bone mineral into mature bones.(18,20,33,), 34 However, no difference could be demonstrated in this characteristic between the two age groups. As shown in Fig. 4, the beneficial effects of exercise generally clearly diminished in response to cessation of exercise, although in both groups some benefit was still maintained after the 14 weeks of deconditioning. We pursued the topic of the maintenance of exercise-induced bone gain in detail in one of our previous studies(23); in that study, the total follow-up time was 56 weeks, and it was shown that the exercise-induced beneficial bone effect still evident after the first 14 weeks of deconditioning vanished during the subsequent 14-week period of free cage activity (no significant differences in any of the measured parameters of the femoral neck between the previously exercised and the control groups at either 28 or 42 weeks of deconditioning), suggesting that the exercise-induced bone benefits are eventually lost if exercise is completely ceased, thus prompting us to propose that continued training is probably needed to maintain the positive effects of youth exercise into adulthood.
In addition to the problems attributable to the continuous growth of the male rats, there is another difficult characteristic pertinent to the use of male rats in experimental osteoporosis research; that is, retardation in the gain of the body mass during training (opposed to no exercise-induced change or even a slight increase in body weight in females).(22,35,36) Given the virtually direct relationship between body mass and bone mass (strength),(27) a direct comparison of the rats in the different study groups in our study would be analogical to comparing the bones of a 40-kg female gymnast to those of a 60-kg woman living a contemporary life: without taking into account the different functional environments the bones act on, the bones of the gymnast naturally have lower absolute mass (BMC), smaller size, and thus, lower mechanical strength than those of the much heavier woman. However, given the fact that the primary role of the long bones is to bear skeletal loads(31,37,38) and that the natural loading is the primary regulator of their size, shape, geometry, and strength,(30,37) we felt that the only reasonable framework to compare the quality (mass, density, geometry, and strength) of the bones in this study was within this locomotive context. Thus, to eliminate the imminent bias introduced by difference in the body weights of the animals between both the young (−9%, p = 0.011) and adult (−13%, p = 0.001) rats (Fig. 2), we adjusted all the data pertaining to the characteristics of the femoral neck with the weight of the calf muscle, a good representative of the lean body mass and the principal origin of natural loading of the skeleton.(30,31)
The above described data normalization has been used previously. Biewener and Bertram(39) stated in 1994 that “normalization of data for body weight or size differences, in fact, is critical in avoiding bias in tests of the effects of exercise (mechanical loading) on the structural changes of the bone.” In other words, independent of the influence of physical activity, larger animals can be expected to have larger bones. The possibility of a size-dependent bias introduced by comparing experimental groups that differ in body mass has largely been overlooked in previous studies. Although still scarcely used, we have followed this line of thinking in all our recent experimental studies,(22,23,27,), (28,40,41) and accordingly, adjusted all the data pertaining to mechanical competence of the rat femora, first by body weight, and more recently, by calf muscle weight. In this way, we have made the bones of different animal groups comparable with each other in terms of their natural loading. However, to provide a better reflection of the absolute changes that took place in the bones of the young and adult rats, we included a table of the crude values for the readers to be assured that the reported differences in the response between the exercised and controls truly existed (Table 2). Regarding the skeletal site analyzed, the femoral neck was chosen as it is the most important skeletal site clinically and has been shown to be sensitive to changes in the prevailing loading in rodents.(42) We also have good previous experience in using this skeletal structure in the similar study designs(23,27,28) as well as a universally approved and well-validated methodology (pQCT and materials testing machine) for the analysis of the response.
In conclusion, this study showed that there were no quantitative differences in the skeletal response to loading in young and adult rats. However, an apparent trend for different mechanisms of adaptation to increased loading was observed: the young skeletons mainly adapted through geometrical changes (increase in bone size), whereas adult rats seemed to adapt mainly through increase in bone density. The ability of the bones to preserve the exercise-induced bone benefits did not seem to be related to age, because the loss of bone in the young and adult rats was identical after cessation of exercise.
We thank Elina Selkälä and Minna Vanhala for excellent technical assistance. This work was supported by grants from the Research Council for Physical Education and Sports, Ministry of Education, the Medical Research Fund of Tampere University Hospital, and the Research and Science Foundation of Farmos.