INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

BONE IS A DYNAMIC TISSUE THAT serves both mechanical and metabolic functions. To fulfill its mechanical role successfully, bone displays the ability to adapt to the functional demands placed on it. This form-function relationship is demonstrated by the loss of bone mass associated with disuse (10, 33) as well as the hypertrophy of bone as a result of increased loading such as with exercise (1, 12). Bone adaptation is regulated at the cellular level and is dependent on the ability of bone cells to perceive and respond to local mechanical signals. However, the mechanisms involved in the adaptation of bone to its physical environment remain largely unknown. The loss of bone mass and microarchitectural deterioration that occur as consequences of normal aging suggest a disruption of this perception-response mechanism with senescence. The purpose of the present study, therefore, was to investigate the ability of bone to respond to its mechanical environment as a function of age.

Both human (1, 5, 7) and animal studies have shown that moderate exercise can add a significant amount of new bone to the growing skeleton. Animals studies using swimming (25), treadmill exercise (16, 18, 22, 26), or jump training (31) have almost universally demonstrated a beneficial effect of increased loading on bone mass and mechanical properties; however, investigations using senescent animals have been equivocal. Raab et al. (19) found that the long bones of 25-mo-old rats were as responsive to increased loading via treadmill exercise as those of 2.5-mo-old rats, although the mechanism of adaptation differed between the two groups. Furthermore, Umemura et al. (31) found that 8 wk of jump training increased the fat-free weight and diameter of tibias in 27-mo-old rats to the same extent as in young animals. To the contrary, investigations using in vivo mechanical loading models, such as the surgically isolated turkey ulna (20) and four-point bending of the rat tibia (30), suggest a decrement in the capacity of aged bone to detect mechanical strain. However, these in vivo loading models involve surgical manipulations and/or do not invoke the systemic hormonal responses or muscle forces on bone that are associated with traditional exercise regimens.

Previous studies have shown that the adaptation to mechanical loading is influenced by the magnitude of the applied load (21). This is supported by cross-sectional studies in humans that suggest that activities that impose high strain magnitudes or high strain rates, such as weightlifting (1, 3) or gymnastics (27), are more osteogenic than endurance-type activities. This is corroborated by animal studies that have shown that jump training (31) or running with additional weight (32) is more osteogenic than treadmill running without applied resistance. Therefore, we employed a unique resistance exercise model designed to progressively overload the hindlimbs of rats. Previous work with this model found a significant osteogenic response after 6 wk of resistance exercise training in 5-mo-old rats (34). We extend the use of this model to investigate the hypothesis that the responsiveness of long bones to increased loading varies as a function of age in the male Fischer 344 rat. We have shown that the long bones of male Fischer 344 rats display age-related changes, including periosteal and medullary expansion as well as decreases in cancellous bone volume, bone formation rate, and the apparent material properties of the bone tissue and thus are an appropriate model in which to study the effect of aging on bone metabolism (Buhl KM, Jacobs CR, Turner RT, Evans GL, Farrell PA, and Donahue HJ, unpublished observations).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals

Resistance Exercise

Data Collection

Three-point bending of femoral diaphysis. Biomechanical tests were performed by use of a servo-hydraulic testing machine (Enduratec, Minnetonka, MN). On the day of mechanical testing, femurs were thawed to room temperature while wrapped in saline-soaked gauze inside sealed plastic bags. The freezing and thawing of bones does not change their biomechanical properties (23). Specimens were kept moist throughout the testing procedures. Before breaking, each femur was weighed, its length measured with a micrometer, and its midpoint determined. Femurs were then placed on two supports that were positioned 16, 18, or 20 mm apart from each other depending on the age of the animal and thus femur length. A breaking force was then applied to the midpoint of the bone, and perpendicular to its long axis, by a crosshead moving at a constant rate of 0.1 mm/s. Force was applied to the anterior surface of the bone because this orientation was found to provide the most stability and minimize rotation during testing. Thus the bone was loaded with compression on its anterior surface and tension on its posterior surface. During testing, a computer-generated load-deformation curve was obtained.

Determination of mechanical properties. After three-point bending, a cross section just distal to the fracture site was cut from the distal half of the femur by use of a diamond saw. The cross section was dehydrated by immersion in 100% ethanol for 24 h and then stained with AgNO₃. A digital image of the cross section was acquired with a charge injection device camera with a macro lens attached to a computer-based frame grabber. Images were imported into a customized version of NIH Image software (Inertial Image; van der Meulen, Stanford University), from which the cross-sectional area (CSA), medullary area (MA), and moment of inertia (MI; a measure of the distribution of mass around the bending axis) were determined. Cortical area (CA) was then calculated as CSA minus MA.

Histomorphometry. Histomorphometric analysis was performed by using a SMI-Microcomp semiautomatic image-analysis system, which consists of a Compaq computer interfaced with a microscope and image analysis system. A high-resolution color video camera, mounted on an Olympus BH-2 microscope, displays the image of the specimen on a color monitor. The movement of a pen on a graphics table superimposes a tracing on the image of the specimen on the video screen. Microcomp software is then used to calculate the line length and the area bounded by tracing.

Cortical bone measurements. Tibias were dehydrated in a series of increasing concentrations of ethanol. After dehydration, cross sections 60 µm thick were cut with a diamond saw just proximal to the tibia-fibula junction. Sections were then ground to a thickness of 15-20 µm on a roughened glass plate. The specimens were mounted in glycerol, and the fluorochrome labeling was observed under ultraviolet light. The following measurements were made: 1) CSA (in mm²), the area of bone and marrow space within the periosteal surface; 2) CA (in mm²), the area of bone excluding the marrow area; 3) MA (in mm²), the area bounded by the endocortical surface of the bone, which normally houses the bone marrow; 4) periosteal mineral apposition rate (PsMAR, in mm/day × 10³), calculated as the interlabel thickness divided by the labeling period and assumed to be proportional to osteoblast activity; and 5) periosteal bone formation rate (PsBFR, in mm²/day × 10³), calculated as the PsMAR multiplied by the mean double-label length and assumed to be proportional to osteoblast number.

Cancellous bone measurements. The proximal ends of the tibias were dehydrated in a series of increasing concentrations of ethanol and then embedded without demineralization in methyl methacrylate. The specimens were then sectioned at a thickness of 5 µm by use of a microtome (Leica Instruments, Nussloch, Germany), mounted in glycerol, and observed for fluorochrome labeling under ultraviolet light. A standardized sample site 1 mm below the growth plate was used. A total area of 3.14 mm² was measured to determine the following: 1) bone volume-to-tissue volume ratio (BV/TV), the fractional volume of tissue volume that contains cancellous bone; 2) double-labeled surface-to-bone surface ratio, the fractional surface of bone surface covered by double fluorochrome label; 3) trabecular thickness (TbTh, in µm), calculated as 2/(bone perimeter/bone volume); 4) trabecular number (TbN), calculated as (BV/TV)/TbTh; 5) trabecular separation (TbSp, in µm), TbTh × (tissue volume/bone volume); 6) mineral apposition rate (MAR, in µm/day), the mean distance between the two labels measured every 50 µm divided by the labeling interval; and 7) bone formation rate-tissue volume ratio (BFR/TV, in %/day), calculated as MAR × double-label perimeter and expressed as a percentage of tissue volume.

Statistical analysis. A two-way ANOVA was used to test the effect of exercise training on biomechanical properties and histomorphometry within each age group. A Bonferroni test was used to correct for multiple comparisons. P < 0.05 was used to determine significance. All data are expressed as means ± SE except for PsBFR and PsMAR, which are shown as individual values.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

One each of the adult Exer, old Exer, and old Con rats died during the training period. An additional old Exer rat was removed from the study due to illness. Thus the final number for each group was 8 in the old Exer, 9 in the adult Exer and old Con groups, and 10 animals in all other groups. Initial and final body weights of the rats are reported in Table 1. Young and adult rats maintained or gained weight over the course of the study, whereas body weight decreased in all old rats. No significant differences in either initial or final body weight between Exer and Con rats within each age group were found. However, there was a strong trend toward a lower body weight with exercise in all age groups.

Despite the tendency for a lower body weight in the Exer rats vs. their controls, femur weight and length were not significantly different. However, the old Exer rats had a significantly smaller MA than their age-matched controls (Fig. 1). MA did not vary in the other age groups. Furthermore, the exercise training had no effect on the CSA, CA, or MI of the femoral diaphysis in any of the age groups.

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Medullary area of the femur at middiaphysis after 9 wk of resistance exercise in male Fischer 344 rats of 3 age groups: young = 4 mo, adult = 12 mo, and old = 22 mo. Exer, exercised rats; Con, control rats. All values are means ± SE. *Significantly different from Old Exer, P = 0.0006.

The mechanical properties of the femurs were determined by three-point bending. There was no significant effect of exercise on either the structural or apparent material properties of the femurs in any age group (Table 2).

Results of the dynamic cortical histomorphometry are shown in Fig. 2. Nine weeks of resistance exercise did not significantly alter PsBFR (Fig. 2A) or PsMAR (Fig. 2B).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Effect of exercise on osteoblast activity in cortical bone. Double fluorochrome labeling was used to measure periosteal bone formation rate (PsBFR, A) and periosteal mineral apposition rate (PsMAR, B) in the middiaphysis of the femur. Male Fischer 344 rats of 3 age groups (see Fig. 1) were trained for 9 wk with a final pack weight equal to 65% of body weight. Data are presented as individual values.

Results of cancellous bone histomorphometry are presented in Table 3. The Old Exer rats had a significantly decreased TbSp compared with their age-matched controls. In addition, there were strong nonsignificant trends toward a greater BV/TV and TbN in the old Exer animals. There were no significant differences in any parameters between the Exer and Con rats in either the young or adult groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A strong trend toward a lower body weight in the Exer rats was found in this study. This may have resulted from the exercise training itself or from differences in food consumption, or it may represent a stress response. Previous investigations have found significantly lower body weights in male rats in response to both voluntary and forced regular exercise. Furthermore, these investigations showed that the exercised rats had lower protein mass and fat mass than the sedentary rats (2, 14, 17, 28). A study using this resistance exercise model found a trend toward a decreased body weight gain in male rats after 6 wk of training (34). Although food consumption was not measured in the present study, other investigations have been equivocal finding either a decrease in appetite (17) or an increase in food consumption (2, 14, 28) with exercise training. A stress response due to handling, training, or the use of a brief footshock stimulus may have contributed to the trend toward a lower body weight in Exer rats. We did not incorporate any measures indicative of a stress response such as corticosterone levels in this study. However, we attempted to control for stress in the study design by handling the Con animals to the same extent as the Exer rats. The Con rats underwent the same training as the Exer rats. In addition, Con rats performed exercise sessions without a weighted vest throughout the training period and were matched for shock to an Exer animal (average shocks per session = 2.52).

The resistance exercise training resulted in a smaller femoral MA in the old Exer rats vs. their controls; however, this effect was not found in the tibia. Other investigations have reported an exercise-induced inhibition of the age-related increase in MA (11, 14, 35). In the present study, no labeling was observed on the endocortical surface; therefore, our current data do not allow us to determine the mechanism responsible for this phenomenon. However, Mosekilde et al. (14) attributed a decreased MA to a decline in endocortical remodeling rate with increased loading. No other significant changes in femoral geometry or in bone structural or material properties were found in any age group. The lack of a significant effect of exercise in the young and adult rats may reflect the low-intensity level of the exercise regimen used in this investigation. Whereas Mosekilde et al. (14) found low-intensity treadmill exercise to have no effect on femur biomechanics in 2-mo-old rats, other investigations using exercise protocols of moderate intensity have demonstrated significant increases in bone strength in young and adult animals (19, 25, 35). However, high-intensity exercise may be detrimental to the growing skeleton (8, 13). It is possible that the trend toward a lower body weight in the Exer rats opposed the effects of training on the mechanical properties of the femur; however, even after normalization for body weight, no significant differences were found (data not shown).

Neither PsBFR nor PsMAR was significantly affected by the exercise training in any age group. These data are consistent with our mechanical analysis, which found no significant changes in femoral CSA with exercise. We believe that these findings are reflective of the low intensity of exercise used in this study. Pilot data from our laboratory using this exercise model showed a greater PsMAR and PsBFR in 2-mo-old male rats after 6 wk of exercise. In that study, rats lifted up to 140% of body weight vs. 65% in the present study, suggesting a role for exercise intensity. The length of the labeling period relative to the duration of training may contribute to our finding of no significant differences in PsBFR. We used a single double-labeling period at the end of the 9 wk of exercise. The possibility exists that PsBFR was increased with exercise earlier in the training but had returned to control levels on adaptation to the novel mechanical stimulus.

The exercise regime did result in a greater cancellous bone mass in the old rats as reflected in a significantly lower TbSp and strong nonsignificant trends toward a greater TbN and BV/TV. Previous investigations have demonstrated similar effects on cancellous bone in young and adult rats using treadmill running (24, 36), resistance exercise (34), and overloading by hindlimb immobilization (9). The mechanism responsible for this greater cancellous bone mass in the old Exer rats remains unclear, although several possibilities exist. Among these is minimodeling, characterized by formation in the absence of prior resorption, which has been demonstrated on trabecular surfaces of adult cancellous bone (4). However, the lack of a significant increase in either BFR or MAR suggests that new bone formation is not involved. Alternatively, we believe that the greater bone mass is the result of an exercise-induced inhibition of age-related bone loss and therefore represents a preservation of existing bone mass. This preservation of bone may be the consequence of a reduction in the activation of new remodeling units and/or a reversal of the age-related uncoupling of the remodeling process such that formation equals resorption. In the present study, staining revealed no osteoclasts on the cancellous bone sections from old rats. Clearly, more studies examining the cellular processes contributing to the adaptation to loading in senescent bone are needed.

We found that the long bones of old rats responded to a low-intensity resistance exercise protocol that did not elicit effects in the long bones of either young or adult rats, suggesting that the responsiveness to changes in the loading environment is altered as a function of age. This is supported by investigations that have demonstrated an equal or greater response to increased loading in senescent rats compared with their young counterparts (19, 31). Our finding of a greater responsiveness of old bones to alterations in loading environment may be related to differences in background activity as a function of age. The lower daily activity of old rats may place their bones in a state of disuse, contributing to a decreased bone mass with aging. As a consequence of this age-related bone loss, exogenous loads are distributed over a lesser bone mass, resulting in a greater strain per unit bone. Alternatively, as a result of a decreased basal activity in old rats, the low-level strains imposed in this study may have exceeded the adaptation threshold in the old rats but were within the range encountered during normal activity in young and adult rats and thus no adaptation occurred. A recent study by Mosley et al. (15) found that bone that has been functionally isolated is more responsive to subsequent high-intensity loading events than bone that has been under normal locomotor loading. They suggest that short periods of high-intensity exercise interspersed with sedentary or low-intensity periods may be more beneficial to bone.

In summary, our data show that the long bones of senescent rats are more responsive to low-intensity resistance exercise than either young or adult rats. Exercise resulted in a decreased MA and TbSp in old rats but had no effect on bone mechanical properties or bone formation rate. Thus, in old animals, the response to low-intensity resistance exercise appears to involve a preservation of bone mass rather than additional bone formation.