JOURNAL OF BONE AND MINERAL RESEARCH Volume 6, Number 7, 1991 Mary Ann Liebert, Inc., Publishers

A Histomorphometric Study of Cortical Bone Activity During Increased Weight-Bearing Exercise DIANE M. RAAB,' THOMAS D. CRENSHAW,2 DONALD B. KIMMEL,3 and EVERETT L. SMITH'

ABSTRACT To quantify cortical bone response to weight-bearing exercise, bone size, mineral content, and formation were measured at the femoral midshaft in swine. Bone formation was measured histomorphometrically on the periosteal, endosteal, and osteonal surfaces. Sedentary adult crossbred sows (3 years, 229 kg) were randomly assigned to basal (B, n = 6), control (C, n = 7), or trained (T, n = 7) groups. The basal and control groups did not exercise and were killed initially (B) or after 20 weeks (C). The trained group walked on a treadmill 20 minutedday at 5 km/h and 5% grade, 5 days/week for 20 weeks. Bone length, area, or fat-free dry weight was not different with time (B versus C) or with training (C versus T). Periosteal modeling was stimulated by walking. Periosteal formation surface and mineral apposition rate (MAR) were greater in trained than control femora. No effects of walking were measured on the endosteal surface. Intracortical remodeling was not affected by walking. The number of labeled osteons (22.4 cm-2) was not different among groups, but osteonal MAR was greater in trained (1.18 Fm/day) than control (0.96 p/day) femora. Walking for 20 weeks in the previously sedentary sows was not a sufficient stimulus to create differences in gross measures of bone size or mineral content but did increase periosteal and intracortical MAR. The primary effect of increased exercise appeared to be osteoblast activation.

INTRODUCTION upon exercise for increasing bone mass and preventing osteoporosis. Weight-bearing exercise prevents bone loss and increases bone mineral content and width in In animals, exercise increases bone mass, size, and breaking ~ t r e n g t h . ' ~Although -~l bone responds to changes in force and strain, the pattern and mechanism for the response is unclear. Frost has proposed the mechanostat theory as a paradigm to explain skeletal adaptation to mechanical usage (SATMU).('o,'llThe theory states that increased mechanical loading increases bone modeling and decreases bone remodeling. Modeling is bone formation and resorption activity occurring at different bone locations in a coordinated pattern to alter bone size and shape. Remodeling

A

TTENTION HAS RECENTLY FOCUSED

is the sequence of resorption followed by formation activity at the same bone site. Remodeling is bone turnover with little change in bone size or shape. Modeling occurs on the periosteal and endosteal surfaces and remodeling on the intracortical and endosteal surfaces. Frost's mechanostat theory has not been tested in nonsurgical models of mechanical loading. Surgical models, such as ulnar osteotomy and external loading via implanted pins, have consistently shown increased periosteal formation. (12-L6) These models provide valuable information because specific mechanical variables can be isolated and tested. However, the results from these models are not directly applicable to humans because of the surgical manipulation and the nonphysiologic stress distributions. Bones from an exercise model must be studied to examine the effect of exercise on modeling and remodeling activity.

'Biogerontology Laboratory, University of Wisconsin, Madison. 'Department of Meat and Animal Sciences, University of Wisconsin, Madison. 'Center for Hard Tissue Research, Creighton University, Omaha, Nebraska.

741

RAAB ET AL.

142

Changes in bone mass and size appear to be stimulated by an increase in mechanical loading. The implementation of exercise in the general population is limited by an individual’s motivation and risk of injury. A strenuous unsupervised exercise program may pose an increased risk of joint and soft tissue injury for an unfit adult. Walking is a safe, reasonable activity for a beginning exercise program, but it is unclear whether walking alone provides sufficient stimulus to induce positive changes in bone mass and The present study examined the effects of walking on femoral bone mass, size, and cell activity in previously sedentary adult animals. The crossbred sow model was chosen because swine are a common model for human physiology(19)and have a similar bone morphometry. The human and porcine femora have similar cross-sectional diameters (3.2 versus 3.5 cm) and areas (6.5 versus 4.8 cm’), and both demonstrate intracortical remodeling. ( 2 0 )

MATERIALS AND METHODS Adult crossbred (Duroc, Large White, and Landrace) sows were ranked by weight (199-258 kg) and blocked into seven groups of three sows. Pigs within each block were randomly assigned to one of three treatment groups. A basal group ( n = 6) was slaughtered at the start of the study to measure baseline bone properties. A control group ( n = 7) and exercise-trained group ( n = 7) were maintained for 20 weeks. For 8 months before the study the sows were maintained nongravid to allow bone cell activity to reach a steady state. All sows were maintained in a specific pathogen-free facility in standard 2.1 x 0.6 m pens. These pens provided acceptable space for sows to stand and lie down‘’’) but restricted forward movement to about 0.3 m. Sows were fed 2 kg/day of a standard corn-soybean meal diet calculated to contain 0.8% Ca and 0.7% P. All nutrients met or exceeded recommended levels.(22’Water was available ad libitum. To assess ovarian function, estrus cycles were monitored throughout the experiment, and the ovaries were removed and examined at death. Animal care and experimental procedure were approved by the University of Wisconsin Research Animal Resource Committee.

Training protocol The sows walked on a motor-drived treadmill 5 days/week at 5% grade. Training time was increased from 5 minutes/day on day 1 to 20 minutes/day on day 8. Training began at 2.5 km/h and was increased 0.8 km/h in week 2, 3, and 4. In weeks 4-20, sows trained 20 minutes/day 5 days/week at 5% grade and 5 km/h. Heart rate (HR) was recored on an electrocardiograph (Burdick EK-5) from three electrodes held by rubber straps around the trunk at a level 2.5-8 cm caudal to the forelimbs. Resting HR was measured before the study and during weeks 10 and 19. It was measured in the evening while the sows appeared asleep. Exercise HR were measured during the last minute of exercise after 3, 5, 6, 10, 14, and 18

weeks of training. Heart rate data were analyzed using analysis of variance procedures appropriate for repeated measures and Scheffe posthoc pairwise comparisons. (23) Training intensity was calculated from exercise HR measurements as a percentage of maximal HR reserve capacity.(14)Maximal HR reserve capacity was the difference between the most recent resting HR and a maximal HR of 255 beats/minute estimated from preliminary measurements.(25)

Bone labeling Bone mineralization sites were labeled with fluorescent markers injected into the jugular vein. The control and trained sows received oxytetracycline injections (25 mg/kg of body weight) on 2 consecutive days at the start of the experiment. All sows received a double calcein (Sigma, St. Louis, MO) injection (15 mg/kg of body weight) beginning 21 days before slaughter. The calcein injection protocol was 2 days injection, 12 days off interval, 2 days injection, and 5 days off interval before slaughter.

Data collection At the time of slaughter the left and right femora were removed from each sow and cleansed of all soft tissue. Femoral length was measured from the distal condyle to the proximal tip of the greater trochanter, and the midpoint of each femur was marked. The intact right femora were sealed in plastic bags and frozen at -20°C for morphometric analysis. The left femur was sectioned with a band saw, and a 1 cm thick cross section from the midpoint was fixed in 70% alcohol for histomorphometric analysis. Morphometric Analysis: Bone mineral content (BMC) of the right femoral middiaphysis was determined by dualphoton absorptiometry (Lunar DP3, Madison, WI). Soft tissue was simulated by immersing the femora in a 10 cm deep water bath over a 1 cm thick plexiglass support. BMC was measured over a 33.02 cm region centered at the midpoint of the bone. Each femoral shaft was scanned twice for an average BMC, first in the frontal plane and then rotated 90” and scanned in the sagittal plane. When the region of interest was offset proximally and distally 0.5 cm, the test-retest reliability was 0.998 for BMC in four bones. The midshaft of the femur appears uniform in bone mineral content and width. The fat-free dry weight of the right femora was measured after extraction in ether for 10 days and dehydration at 80°C for 15 h. Cross sections from the middiaphysis 1 cm thick were photographed and digitized using the AutoCAD program (Autodesk Inc., Mill Valley, CA) to measure cortical cross-sectional dimensions. Histomorphometric Analysis: Cross sections 1 cm thick from the left femoral middiaphysis were sectioned into anterior, posterior, medial, and lateral subsections (Fig. 1). All subsections were dehydrated undecalcified, with processing times increased by a factor of 3 over those recom-

743

CORTICAL BONE AND EXERCISE

had completed formation before the oxtetracycline and calcein injections. The measurements of W.Th represent erestudv values. The mean osteonal bone formation period was calculated as W.Th divided by MAR."')

ANTERIOR

Statistical analysis

0 0 0 o o 0 0 0 0 0

o

'

POSTER1OR FIG. 1. Four quadrant subsections were cut from each femoral cross section. The area of individual subsections averaged 65 mm2, and the total cortical area averaged 4.8 cm2. The relative distribution of double-labeled osteons in control sows is shown by the small circles. Each circle represents four osteons.

mended for smaller bones,(26)and embedded in polymethyl methacrylate. The embedded blocks were sectioned on a modified milling machine at 500 pm, glued to a glass slide, and ground to approximately 60 pm (Buehler Ecomet 11) for analysis. Two sections 4 mm apart were analyzed from each quadrant for an average area of 130 mm2 per quadrant per sow. The active formation surface on the periosteum and endosteum was quantified as the percentage of double-labeled surface relative to the total bone surface (dLS/BS). The extent of intracortical formation was calculated as the number of double-labeled osteons in the intracortical region relative to the total bone area analyzed (N.dL.On/ B.Ar). Double-labeled endosteal and intracortical surfaces were defined as surfaces with two calcein labels. Doublelabeled periosteal surfaces contained both a tetracycline and at least one calcein label. Analysis of the periosteum was limited to the control and trained groups, since the basal group was not given a tetracycline injection. The mineral apposition rate (MAR) was calculated for the periosteum, endosteum, and osteon. Interlabel width (1rL.Wi) was measured between the midpoints of all double calcein bands on the endosteum and in the osteons. On the periosteum, IrL.Wi was measured from the middle of the tetracycline label to the center of the double calcein-labeled region. MAR was calculated as 1rL.Wi per 136 days for periosteal formation and IrL.Wi per 14 days for endosteal and osteonal remodeling. The tetracycline-calcein interval represented the MAR on the periosteum for the entire training period. The same measurement was not possible on the osteonal surface, where formation period was only 63 days, or on the endosteal surface, where most of the initial tetracycline label was resorbed. Osteonal wall thickness (W.Th) was measured in the posterior quadrant in a subsample of 10 femoral crosssections. From each bone, 100 completed osteons were measured at four equally spaced radii. All measured osteons

The results were analyzed with procedures appropriate for a split-plot, two-factor randomized block design using the General Linear Model Program (SAS, 1982). The main plot variables included treatment groups and weight blocks; the subplot variables included femoral quadrants and the interaction of treatment by quadrant. Appropriate transformations were performed on data not normally distributed. An arcsin transformation was imposed on dLS/ BS and N.dL.On/B.Ar, and a logarithmic transformation was imposed on periosteal MAR. Differences among treatments were evaluated with preplanned comparisons to identify differences over the course of the study due to changes in age and environment (basal versus control) and due to training (control versus trained). Orthogonal comparisons were used to compare the differences among quadrants. The two surfaces most affected by bending in the sagittal plane were contrasted (anterior versus posterior) and those affected by bending in the frontal plane were contrasted (medial versus lateral). The sum of the effects from bending in the frontal plane was contrasted to the sum of effects from the sagittal plane (anterior + posterior versus medial + lateral). Differences were considered significant at P < 0.05.

RESULTS All animals appeared healthy throughout the 20 week study. The sows tolerated the exercise well, with no observed problems in body temperature regulation, injuries, or changes in mobility. The average initial ages (2.99 0.08 years, mean f SEM, standard error of the mean) and body weights (228.9 f 1.O kg) were similar among groups. All groups had similar reproductive histories at the start of the study. The basal, control, and trained sows averaged the same number of litters (3.3, 4.3, and 3.6 0.2) and a similar time postpartum (10.9, 8.0, and 8.6 k 0.3 months). After the 20 week period, the control sows were significantly older than the basal sows (Table 1) and were a greater number of months postpartum. Body weight among groups did not change during the study. All sows, except one in the control group, had regular (21 day) estrus cycles. Visual examination of the ovaries at slaughter confirmed that multiple ovulations occurred in previous months.

*

Heart rate response to training The heart rate training response was evaluated at rest and during exercise (Fig. 2). Although resting HR on the pretest was not significantly different between groups, the control sows had a 12% higher HR than the trained sows. An analysis of covariance, with initial HR as the covariate

744

RAAB ET AL. TABLE1 . Sow SAMPLE CHARACTERISTICS AND FEMUR MORPHOMETRY~

~

Basal Control Trained SEM

6 7 7

3.03b 3.48b 3.29 0.08

224 229 224 2.3

27.25 27.76 27.68 0.20

328 347 338 4.6

4.83 4.85 4.79 0.08

18.46 17.86 17.72 0.28

asow data for age and body weight (BW) are at death. Femoral data represent total length, fatfree dry weight (FFW), middiaphyseal cortical area, and bone mineral content (BMC). bDifference between marked values within a column ( P < 0.05).

.-kC2

A

190 170

ad

n

Y

150 Q,

c

Q

U

' c

tJ

i 80

>

length and fat-free weight were similar for all groups. Within the cortical midshaft region, there were no differences in either cortical cross-sectional area or BMC. The similarity in femoral dimensions and total body weight suggests the groups contained sows of similar stature and that training did not affect bone morphometry as measured by these techniques.

60 40 . . . . . . . . . . . . . . . . . . . . . 0 2 4 6 8 10 12 14 16 18 20

Cortical bone formation with training

Averaged across all quadrants, periosteal dLS/BS (Table 2) was 27% greater ( P < 0.01) in the trained than the control sows. Periosteal MAR, averaged across quadTime (wks) FIG. 2. Sow heart rates during the 20 week training rants, was 76% greater ( P < 0.03) in the femora from the study. Values represent average + SEM of seven sows. trained than from the control sows (Table 2). Endosteal dLS/BS ( P > 0.4) and MAR ( P > 0.6), Exercise heart rate ( 0 trained) was measured after 19 minutes of continuous walking at 5% grade and 4.75 km/h in pooled across all quadrants, were not different among week 3 and 5 km/h in weeks 5, 6, 10, 14, and 18. The two groups (Table 2). After the 20 week period, the control lower lines are resting heart rates measured while sows ap- group had 47% less labeled surface than the basal group, peared asleep (V trained and A control). *Different from but this comparison was not significant. When tested by week 5 tested with Scheffe posttest ( P < 0.05). +Different analysis of covariance, age did not account for a signififrom week 6 tested with Scheffe posttest ( P < 0.05). #Difference between groups tested with analysis of covariance cant amount of variation in dLS/BS. The number of double-labeled osteons did not differ ( P < 0.05). among groups (Table 2). MAR did not differ with time (basal versus control). Trained sows, however, had a 23% to adjust for initial values, was used to test resting HR at greater ( P = 0.04) osteonal MAR than control sows. weeks 10 and 19. Resting HR was lower ( P < 0.002) in the Osteonal W.Th did not vary among groups and was 76 trained sows than the control sows at week 10; a similar 2.2 pm in femoral posterior quadrants. Given an average MAR of 1.2 pm/day in the posterior quadrants from but nonsignificant trend was observed at week 19. the bone formation The relative training intensities as a percentage of esti- sedentary sows and no off time,(28.z9) mated maximal HR capacity were 56% in week 3 and 60, period would be 63 days. Osteonal bone formation period 56, 53, 49, and 50% in weeks 5, 6, 10, 14, and 18, respec- in the femur from the adult crossbred sow is estimated as 2 tively. The initial increase in HR intensity was due to an in- and 3 months. crease in walking speed; thereafter a decrease in the relative exercise intensity was observed. Exercise HR (Fig. 2) Cortical bone formation within quadrants were significantly lower in weeks 10, 14, and 18 than in There were no significant interactions between quadrant week 5. Heart rates in week 14 were significantly lower than in week 6. Both the resting and exercise HR measure- and treatment for any histomorphometric measurement. ments suggest that a cardiovascular training response oc- Averaged across groups, regional differences in formation were seen between the individual quadrants (Table 3). curred primarily within the first 10 weeks of the study. Periosteal formation was not different between quadrants within the sagittal plane of motion (anterior versus Bone morphometry posterior) or the frontal plane of motion (medial versus Morphometric characteristics of the right femur were lateral). The combined activity in the sagittal plane (antenot different with age or with training (Table 1). Femoral rior + posterior surfaces) was greater than that in the

745

CORTICAL BONE AND EXERCISE TABLE 2. BONEFORMATION ON

THE

FEMORAL MIDDIAPHYSEAL SURFACES~

Formation surface Group

n

Basal Control Trained SEM

6 7 7

Mineral apposition rate

Periosteal b (dLS/BS)

Endosteal (dLS/BS)

Osteonal (no./cm’)

Periosteal (pm/day)

Endosteal (pm/day)

Osteonal (pm/day)

-

31.11 16.36 25.18 6.36e

23.2 19.6 24.4 2.6e

-

1.01 0.91 1.05 0.13f

1.11 0.96d 1.18d 0.08f

73.2% 93.48c 4.58e

0.59d 1.04d 0.1 le

”Values represent the means for treatment groups pooled across four quadrants. There were no significant interactions between groups and quadrants. bPeriosteal surface represents 14 sows, except that 1 control and 1 trained sow were each missing data from one of the four quadrants. CDifference between marked values within a column ( P < 0.01). dDifference between marked values within a column ( P < 0.05). ‘Standard error of the mean of the transformed data. ‘Standard error of the mean of the raw data.

TABLE3. BONEFORMATION WITHINQUADRANTS ON THE FEMORAL MIDDIAPHYSEAL SURFACES POOLEDACROSSTREATMENT GROUPS Formation surface Mineral apposition rate Group

n

Anterior Posterior Medial Lateral SEM

20 20 20 20

Periosteal a (dLS/BS)

Endosteal (dLS/BS)

Osteonal (no./cm’)

Periosteal (pm/day)

Endosteal (pm/day)

Osteonal (pmlday)

93.55 81.84c 13.37 78.40c 1.83e

20.77 21.43 21.74 31.54 1.25e

7.8b 56.7b.C 7.3d 17.6Cgd 0.6e

1.14 0.91c 0.73 1.00C 0.03e

0.93b 1.14b 0.92 0.97 0.04f

1.07b 1.24b 0.90d 1.12d 0.03f

aPeriosteal surface data represent 13 sows for the anterior and lateral quadrants and 14 sows for the posterior and medial quadrants. bsignificant difference between anterior and posterior marked values within a column ( P < 0.05). CSignificant difference, combined (anterior + posterior) versus (medial + lateral), within a column ( P < 0.05). dsignificant difference between medial and lateral marked values within a column ( P < 0.05). eStandard error of the mean of the transformed data. fStandard error of the mean of the raw data.

frontal plane (medial + lateral surfaces) in both percentage of double-labeled surface ( P < 0.04) and MAR (P < 0.0002). On the endosteum, combined across all groups, the quadrants were not different for dLS/BS. The MAR on the endosteum, however, was significantly greater on the posterior than the anterior surface ( P < 0.05). The intracortical remodeling rate was significantly different (P < 0.001) between quadrants. The posterior quadrant contained 63% of the total number of osteons; the anterior and medial quadrants each contained only 8%. The relative distribution of double-labeled osteons in control sows is shown in Fig. 1. MAR was greater (P < 0.03) in the posterior than anterior, lateral than medial, and the combined anterior + posterior than medial + lateral quadrants.

DISCUSSION Treadmill walking generated a cardiovascular training response in these previously sedentary adult sows. The response was measured at both rest and exercise after 10 weeks of walking at 60% of maximal heart rate. Although low, a training intensity of 60% is considered a modest intensity for inducing a heart rate training response in hum a n ~ . ( *As ~ ) fitness improved, the relative training intensity decreased to 50% of maximum, and no further decreases in heart rate were observed. Although heart rate is not a direct measure of mechanical strain on bone, the cardiovascular changes were induced by increased weightbearing activity and may indirectly reflect changes in bone loading. Given that the cardiovascular adaptation occurred within the first 10 weeks and that periosteal active surface

746

remained elevated at 20 weeks, it appears that adaptation to increased exercise is slower in bone than in the cardiovascular system.

Bone formation on cortical surfaces

RAAB ET AL. the adaptation period may depend upon the relative change in loading caused by the exercise program. Regular increments in loading may be necessary to maintain a steady rate of increased formation.

Endosteal Surface Activity: The endosteum had no training-related differences in formation. The endosteal labeling was randomly distributed throughout the quadrants and had only one-fourth as much labeled surface as the periosteum. In past studies, the endosteal response to increased loading has been less consistent than the periosteal response. Several studies with growing animals reported a decrease in endosteal diameter after strenuous exercise.(9.35'Others rePeriosteal Modeling: Periosteal apposition was ported no endosteal response to loading in mature anigreater in trained than control sows. Exercise-trained sows mals. ( 1 2 . 1 5 , 1 6 1 Endosteal changes have been reported in had a 20% higher double-labeled surface (nearly 100%) models that alter force distribution by external loading(36) and a 76% higher mineral apposition rate compared with or osteotomy.'14)Differences in animal maturity, lack of controls. Periosteal formation was an activation of model- information on intensity of the mechanical load, and the ing with no evidence of prior resorption. These data sup- distribution of the load to the endosteum make it difficult port Frost's theory that increased mechanical usage in- to interpret the different endosteal responses. In this study, the forces created during walking increases modeling. The most effective adaptation to increased mechanical loading or strain is an increase in corti- creased periosteal bone formation but did not affect the cal diameter.(30)The present increase in periosteal forma- endosteum. The difference in response probably relates to tion agrees with previous reports of greater cortical diame- relative changes in strain on the two surfaces. Over 80% of the strains produced in weight-bearing activities are due to ters in exercise than control animal^.'^'.^^) Periosteal expansion has been consistently observed bending. (37) With bending, strain increases linearly from after ulnar osteotomy and with external load application. the neutral axis to the outer cortex. The highest strains are In these surgical models with altered strain distribution, produced on the outer periosteal surface and the lowest on the periosteal response is more robust than in exercise the endosteal surface. The strains created on the endosteal models and often includes areas of woven bone forma- surface during walking may not have been sufficient to tion.(12-16)These models have also reported regions of stimulate bone formation. periosteal resorption that may represent a modeling drift Osteonal Remodeling: Studies of external load applicato adapt the shape of the bone to the nonphysiologic load tion and osteotomy report either no ~ h a n g e ' ~ ~or. ' in~) distribution. with inThe increased periosteal apposition rate measured in the creased remodeling activation frequency(14.15.38.39) trained sows may have resulted from increased osteoblast creased strain rate and magnitude. The present study is the recruitment rather than increased osteoblast vigor. The re- first to examine intracortical remodeling rates after exerported periosteal MAR was the average of all mineral ap- cise. The activation frequency of osteonal remodeling at 20 positional activity over 20 weeks. It is possible that forma- weeks was unaffected by increased walking. tion was continuous with minimal off time(28.29) in the The present data and data from the surgical models do trained sows but intermittent with on and off time in the not support Frost's theory of depressed remodeling with control sows. The 20 week periosteal MAR for the trained increased mechanical loads. ( I 1 ) The surgical models are sows was similar to the 2 week MAR measured on the en- confounded by (1) potential stimulation of a regional acdosteal and osteonal surfaces. However, the periosteal rate celeratory phenomenon in response to the surgery,(I4)(2) for the control sows was only 60% of that on the other dramatic changes in the load distributions, which may surfaces. Intermittent activity may also account for differ- cause microdamage, and (3) the removal of normal loading ences between groups in active surface extent. Transiently patterns, which may result in regions of relative disuse. active surfaces incorporate label only when cells are formFrost suggests that remodeling is dramatically depressed ing new bone tissue. The observation of increased activa- above 50 microstrain and continues to be depressed as tion with no change in individual osteoblast vigor concurs strain level increases. Normal cage activity should have with others' findings.'L4.33) created strains well in excess of 50 microstrain. The small Although the study was designed to last longer than one increase in mechanical load and strain due to daily walking bone formation period in porcine cortical bone, the level may have been insufficient to further depress the activaof bone formation seen in the present study may have been tion frequency of intracortical remodeling. Future studies transient. It is not possible to determine a steady state need to test exercise at higher intensities and sample bone from a single time point study. Frost suggests that when over a time range. The effects of loading on activation of the increase in bone mass is sufficient for structural adap- remodeling may be cyclic,(14)and we may have missed the tation to the new loads, strain levels return to normal and effect because of our choice of sampling time point (20 the stimulus for formation is The duration of weeks, 1.5-2 sigmas). The increase in weight-bearing exercise in the present study produced differences in cortical bone formation on the femoral periosteal and osteonal surfaces. The differences can be interpreted relative to Frost's paradigm for skeletal adaptation to mechanical usage. Frost predicted increased modeling and depressed remodeling with increased mechanical loading.

747

CORTICAL BONE AND EXERCISE Osteonal MAR was increased by 20 minutes of walking per day. The possibility of increased activation of the osteoblast within a remodeling unit, without stimulation of the osteoclast, has not been discussed by Frost. As on the periosteum, the increased MAR could represent an increase in osteoblast recruitment, formation uninterrupted by customary periods of quiescence, or an increase in osteoblast vigor. A similar response to increased loading was reported in rat cancellous bone, where there was no significant change in labeled or eroded surfaces but MAR In contrast, no change in intracortical apposition rate was reported in the dog radius after ulnar osteotomy.(14)Further research is necessary to understand the effect of mechanical loading on both the osteoclast and osteoclast in remodeling.

Regional differences in bone formation Significant regional differences in bone formation existed for all sows regardless of group. The anterior and posterior surfaces had the greatest active surface, MAR, and number of double-labeled osteons. Similar regional variations in bone formation activity have also been reported in the canine femoral middiaphysis.(41)The high intracortical remodeling in the posterior quadrant may relate to the extensive muscle attachments along the linea aspera of the femur. The reason for high periosteal and endosteal activity in the anterior and posterior quadrants is unclear. Differences in regional strains within the quadruped femur have not been well described. The presence of significant regional differences in formation activity of a normal femur emphasizes the importance of a standard sampling site. Although active surface and MAR differed by region, there were no interactions between treatment groups and quadrants. In external loading studies, bone formation appeared greatest on the surfaces with the greatest increase in strain.(1s,38.39) Other studies have described regional differences in formation due to e x e ~ c i s e ( ~and . ’ ~ ~ osteotomy,(’6.42,43) but the differences were not statistically validated. Again, the higher forces applied in previous studies, resulting in greater strain magnitude and altered strain distribution, seem more likely to cause regional variations than the conditions of this study.

Theoretical calculation of morphometric properties Based upon histomorphometric measurements, the femoral cross-sectional diameter should have increased more in trained than control sows since they had a greater periosteal mineral apposition rate. To emphasize the difference in sensitivity between morphometric and histomorphometric measurements, the theoretical change in midshaft dimensions was calculated. This calculation assumed that all sows started with bones similar to the basal group, that there were no net changes in bone mass due to endosteal or intracortical activity, and that the femora were cylindrical. The cortical diameter after 20 weeks can be calculated from the average periosteal MAR (Table 4). The predicted diameters represent a 1.8 and 3.3% increase in femoral cross-sectional area for control and trained sows. The apparent discrepancy between actual and predicted values may relate to differences in bone cross-sectional area before training. Since periosteal expansion or MAR was directly measured, it is reasonable to assume that the bones expanded at different rates but that the initial bone areas were not identical among groups. Given the predicted increase in cortical area and assuming a constant bone mineral density, a 1.5% increase in bone mineral content is predicted due to training alone. This 1.5% change would be difficult to detect statistically in a cross-sectional design due to intraanimal variability and measurement error. Power calculations predict group sizes of 59 would be required to detect a 0.27 g/cm difference (P < 0.05) in bone density with a 0.74 standard deviation.(13)

Conclusions In previously sedentary sows, walking increased bone formation activity at both the periosteum and intracortical surfaces., The increased osteoblast activity occurred on bone surfaces involved in both modeling and remodeling. External loading experiments have shown that the primary mechanical stimuli that increase bone mass are increased strain magnitude and altered strain distribution.(12,13,38) Therefore, for an exercise program to be effective, the activity must create forces that cause strains above and different from normal daily levels. Thus, the change in bone activity and mass caused by an exercise

TABLE4. PREDICTED AND ACTUAL CROSS-SECTIONAL AREAAFTER20 WEEKSOF EXERCISE

MAR a (pm/day)

Growthb (pm per 20 weeks)

PDc (cm)

p-CAd @m2)

Changee

Group

(%)

a-CAf (cm2)

Basal Control Trained

0 0.59 1.04

0 165 29 1

3.534 3.550 3.563

4.83 4.92 4.99

0 1.8 3.3

4.83 4.85 4.19

aAverage periosteal mineral apposition rate for the group from Table 2. bPeriosteal expansion in 20 weeks = MAR x Z(surfaces x 140 days. CPeriostealdiameter of basal + growth (assume all sows were initially equal to basal group). dPredicted cortical area; assumes medullary diameter (2.518 cm) is equal for all groups and endosteal activity was minimal; = 7r(PD/2)’ - ~ ( 2 . 5 1 8 / 2 ) ~ . ePercentage change relative to basal group. fActual cortical area.

RAAB ET AL.

748

program depends not on the absolute amount of force applied to the bone, but upon the change in forces applied. The sows had sufficient room in their pens to shuffle forward and backward about 0.3 m; however, there was not enough room for a full-length stride. The sows rarely walked outside the pen; therefore, a daily 20 minute walk was a change in activity level for the trained sows. Although bone size and mechanical loading patterns are different between sows and humans, a similar change from a sedentary, shuffling life-style to a moderate walking program may stimulate bone formation in humans. This study suggests that it is not intense exercise but a change in exercise level that is necessary to stimulate periosteal bone formation. For very sedentary women, walking may be a reasonable and effective exercise for increasing bone. This study defined a specific activity (walking), frequency (5 daydweek), intensity ( 5 km/h), and duration (20 minutes). Future exercise and bone studies should also report such specific information to enable comparisons between studies and proper interpretation. These training parameters are already well defined for the muscular and cardiovascular systems and should be systematically investigated and established for the skeletal system.

7. McDonald R, Hegenauer J, Saltman P 1986 Age-related dif-

8.

9.

10.

11.

12.

ferences in the bone mineralization pattern of rats following exercise. J Gerontol 41:445-452. Raab DM, Smith EL, Crenshaw TD, Thomas DP 1990 Bone mechanical properties after exercise training in young and old rats. J Appl Physiol 68:130-134. Woo SL, Kuei SC, Amiel D, Gomez MA, Hayes WC, White FC, Akeson WH 1981 The effect of prolonged physical training on the properties of long bone: A study of Wolffs law. J Bone Joint Surg [Am] 63:780-787. Frost HM 1990 Skeletal structural adaptations to mechanical usage (SATMU). 1. Redefining Wolffs law: The bone modeling problem. Anat Rec 226:403-413. Frost HM 1990 Skeletal structural adaptations to mechanical usage (SATMU). 2. Redefining Wolffs law: The remodeling problem. Anat Rec 226:414-422. Rubin CT, Lanyon LE 1984 Regulation of bone formation by applied dynamic loads. J Bone Joint Surg [Am] 66:397-

402. 13. Rubin CT, Lanyon LE 1985 Regulation of bone mass by mechanical strain magnitude. Calcif Tissue Int 37:411-417. 14. Burr DB, Schaffler MB, Yang KH, Wu DD, Lukoschek M, Kandzari D, Sibaneri N, Blaha JD, Radin EL 1989 The effects of altered strain environments on bone tissue kinetics. Bone 10:215-221. 15. Churches AE, Howlett CR 1982 Functional adaptation of

bone in response to sinusoidally varying controlled compressive loading of the ovine metacarpus. Clin Orthop 168:265-

ACKNOWLEDGMENTS We gratefully acknowledge the assistance of Mary Checovich, Chris Grieshop, Mark Gahl, Steve Van Lannen, and Maureen Miller. This research was supported by a NIH Biomedical Sciences Support Grant (Project No. 881697), the National Institute of Biogerontology, and the College of Agriculture and Life Sciences. This work was previously presented in abstract form in J Bone Miner Res 4:s180, 1989 and Med Sci Sports Exerc 2 2 ~ 6 3 ,1990.

280. 16. Lanyon LE, Goodship AE, Pye CJ, MacFie JH 1982 Mechanically adaptive bone remodelling. J Biomech 15: 141154. 17. Sandier RB, Cauley JA, Hom DL, Sashin D, Kriska AM 1987 The effects of walking on the cross-sectional dimensions

of the radius in postmenopausal women. Calcif Tissue Int 41~65-69. 18. White MK, Martin RB, Yeater RA, Butcher RL, Radin EL 1984 The effects of exercise on bones of postmenopausal women. Int Orthop 7:209-214. 19. Tumbleson M 1986 Swine in Biomedical Research, Vols. 1-3.

Plenum, New York. 20. Smith RW, Walker RR 1964 Femoral expansion in aging

REFERENCES

21.

1. Chow R, Harrison JE, Notarius C 1987 Effect of two ran-

domised exercise programmes on bone mass of healthy postmenopausal women. Br Med J 292607-610. 2. Dalsky GP, Stocke KS, Ehsani AA, Slatopolsky E, Lee WC, Birge SJ 1988 Weight-bearing exercise training and lumbar bone mineral content in postmenopausal women. Ann lntern Med 108:824-828. 3. Krolner B, Toft B, Nielsen SP, Tondevold E 1983 Physical exercise as prophylaxis against involutional vertebral bone loss: A controlled trial. Clin Sci 64:541-546. 4. Smith EL, Gilligan C, McAdam M, Ensign CP, Smith PE 1989 Deterring bone loss by exercise intervention in premenopausal women and postmenopausal women. Calcif Tissue lnt

22.

23. 24. 25.

26.

44~312-321. 5. Bell RR, Tzeng DY, Draper HH 1980 Long-term effects of

calcium, phosphorus and forced exercise on the bones of mature mice. J Nutr 110:1161-1168. 6. Beyer RE, Huang JC, Wilshire GB 1985 The effect of endurance exercise on bone dimensions, collagen, and calcium in the aged male rat. Exp Gerontol 20:315-323.

27.

28.

women: Implications with osteoporosis and fractures. Science 145156-157. Curtis S 1988 Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching. Consortium Developing Guide for Care & Use of Agricultural A, Champaign, IL. National Research Council 1988 Nutrient Requirements of Swine, 9th ed. National Academy of Sciences, Washington, D.C. Snedecor G, Cochran W 1980 Statistical Methods. Iowa State University, Ames. Karvonen M 1957 The effects of training on heart rate. A longitudinal study. Ann Med Exp Biol Fenn 35307-315. Raab DM 1989 Cortical bone adaptation to weight-bearing exercise: A mechanical and histomorphometric analysis. Doctoral Dissertation, University of Wisconsin, Madison. Sheehan DC, Hrapchak BB 1980 Theory and Practice of Histotechnology, 2nd ed. C.V. Mosby Co., St. Louis, MO. Parfit AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR 1987 Bone histomorphometry: Standardization of nomenclature, symbols, and units. J Bone Miner Res 2595-610. Hori M, Takahashi H, Konno T, lnoue J, Haba T 1985 A classification of in vivo bone labels after double labeling in

CORTICAL BONE AND EXERCISE canine bones. Bone 6:147-154. 29. Keshawarz NM, Recker RR 1986 The label escape error: Comparison of measured and theoretical fraction of total bone-trabecular surface covered by single label in normals and patients with osteoporosis. Bone 7:83-87. 30. Currey J 1984 The Mechanical Adaptations of Bones. Princeton University Press, Princeton, NJ. 31. Steinberg ME, Trueta J 1981 Effects of activity on bone growth and development in the rat. Clin Orthop 15652-60. 32. Uhthoff HK, Jaworski ZFG 1985 Periosteal stress-induced reactions resembling stress fractures. A radiologic and histologic study in dogs. Clin Orthop 199:284-291. 33. Parfit AM 1990 Bone-forming cells in clinical conditions. In: Hall BK, ed. Bone, Vol. 1. The Osteoblast and Osteocyte. Telford Press, West Caldwell, NJ, pp. 351-429. 34. Frost H 1987 Bone “mass” and the “mechanostat”: A proposal. Anat Rec 219:l-9. 35. Matsuda J J , Zernicke RK, Vailas AC, Pedrini VA, PedriniMille A, Maynard J A 1986 Structural and mechanical adaptation of immature bone to strenuous exercise. J Appl Physiol 60:2028-2034. 36. Liskova M, Hert J 1971 Reaction of bone to mechanical factors. Part 2. Periosteal and endosteal bone apposition in the rabbit tibia due to intermittent stressing. Folia Morphol (Praha) 19:301-317. 37. Rubin CT, Lanyon LE 1982 Limb mechanics as a function of speed and gait: A study of functional strains in the radius and tibia of horse and dog. J Exp Biol 101:187-211. 38. O’Connor JA, Lanyon LE, MacFie H 1982 The influence of strain rate on adaptive bone remodelling. J Biomech 15:767781.

749 39. Hert J, Pribylova E, Liskova M 1972 Reaction of bone to mechanical stimuli. Part 3. Microstructure of compact bone of rabbit tibia after intermittent loading. Acta Anat (Basel) 82:218-230. 40. Jee WSS, Li XJ 1990 Adaptation of cancellous bone to overloading in the adult rat: A single photon absorptiometry and histomorphometry study. Anat Rec 227:418-426. 41. Marotti G 1963 Quantitative studies on bone reconstruction. 1. The reconstruction in homotypic shaft bones. Acta Anat (Basel) 52:29l-333. 42. Burr DB, Shaffler MB, Yang KH, Lukoschek M, Sivaneri N, Blaha JD, Radin EL 1989 Skeletal changes in response to altered strain environments: Is woven bone a response to elevated strain? Bone 10:223-233. 43. Goodship AE, Lanyon LE, McFie H 1979 Functional adaptation of bone to increased stress. J Bone Joint Surg [Am] 61~539-546.

Address reprint requests to: Diane M . Raab

Center for Hard Tissue Research 601 N. 30th Street, Suite 5740 Creight on University Omaha, N E 68131

Received in original form July 23, 1990; in revised form December 21, 1990; accepted January 16, 1991.

A histomorphometric study of cortical bone activity during increased weight-bearing exercise.

To quantify cortical bone response to weight-bearing exercise, bone size, mineral content, and formation were measured at the femoral midshaft in swin...
809KB Sizes 0 Downloads 0 Views