Modulation of appositional and longitudinal bone growth in the rat ulna by applied static and dynamic force

Bone Vol. 29, No. 2 August 2001:105–113

Modulation of Appositional and Longitudinal Bone Growth in the Rat Ulna by Applied Static and Dynamic Force A. G. ROBLING,1 K. M. DUIJVELAAR,2 J. V. GEEVERS,2 N. OHASHI,1 and C. H. TURNER3 1

Department of Anatomy and Cell Biology, Indiana University School of Medicine, Indianapolis, IN, USA Maastricht University School of Medicine, Maastricht, The Netherlands 3 Department of Orthopedic Surgery and Biomechanics and Biomaterials Research Center, Indiana University School of Medicine, Indianapolis, IN, USA 2

Key Words: Bone growth; Bone modeling; Growth plate; Static loading; Dynamic loading; Rat ulna; Histomorphometry.

Appositional and longitudinal growth of long bones are influenced by mechanical stimuli. Using the noninvasive rat ulna loading model, we tested the hypothesis that briefduration (10 min/day) static loads have an inhibitory effect on appositional bone formation in the middiaphysis of growing rat ulnae. Several reports have shown that ulnar loading, when applied to growing rats, results in suppressed longitudinal growth. We tested a second hypothesis that load-induced longitudinal growth suppression in the growing rat ulna is proportional to time-averaged load, and that growth plate dimensions and chondrocyte populations are reduced in the loaded limbs. Growing male rats were divided into one of three groups receiving daily 10 min bouts of static loading at 17 N, static loading at 8.5 N, or dynamic loading at 17 N. Periosteal bone formation rates, measured 3 mm distal to the ulnar midshaft, were suppressed significantly (by 28 – 41%) by the brief static loading sessions despite normal (dynamic) limb use between the daily loading bouts. Static loading neither suppressed nor enhanced endocortical bone formation. Dynamic loading increased osteogenesis significantly on both surfaces. At the end of the 2 week loading experiment, loaded ulnae were approximately 4% shorter than the contralateral controls in the 17 N static and dynamic groups, and approximately 2% shorter than the control side in the 8.5 N static group, suggesting that growth suppression was proportional to peak load magnitude, regardless of whether the load was static or dynamic. The suppressed growth in loaded limbs was associated with thicker distal growth plates, particularly in the hypertrophic zone, and a concurrent retention of hypertrophic cell lacunae. Negligible effects were observed in the proximal growth plate. The results demonstrate that, in growing animals, even short periods of static loading can significantly suppress appositional growth; that dynamic loads trigger the adaptive response in bone; and that longitudinal growth suppression resulting from compressive end-loads is proportional to load magnitude and not average load. (Bone 29:105–113; 2001) © 2001 by Elsevier Science Inc. All rights reserved.

Introduction Bone tissue has the capacity to adapt its mass and architecture to meet the prevailing mechanical loading environment. Mechanical loads borne by the vertebrate appendicular skeleton during normal functional use (e.g., locomotion) are dynamic, with constantly changing strain magnitudes, rates, and orientations. Given the dynamic nature of the strains typical of normal functional use, it follows teleologically that bone adaptation should be driven by signals derived from dynamic rather than static loading. Numerous in vivo experiments conducted over the last several decades have shown that bone is sensitive to, and responds anabolically to, dynamic loading.4,13,25,27,35 However, the few experiments addressing the effects of static loading on bone adaptation have produced less consistent results, including suppressed bone formation,10 no response,10,12 and bone loss.12 Using a preparation in which steel pins were surgically implanted transcortically through the metaphyses of growing and mature rabbit tibiae, Her˘t et al.10 showed that continuous bending (24 h/day, 7 days/week, 56 weeks) of the diaphysis resulted in an overall suppression of periosteal apposition at midshaft, but endocortical bone formation was unaffected by the static load. In a subsequent experiment13 using the same model, dynamic loading (1–3 h/day, 7 days/week, 4 weeks) resulted in a significant increase in bone formation on both periosteal and endocortical surfaces. Later, Lanyon and Rubin12 studied the differential effects of static and dynamic loading in the mature turkey ulna, using a surgical preparation similar to that of Her˘t. They showed that a continuous (24 h/day, 7 days/week, 8 weeks) compressive axial load had no effect on periosteal apposition but was associated with a significant amount of bone loss on the endocortical and haversian surfaces, a result that mimicked disuse. In the same experiment, turkey ulnae that were exposed to dynamic loading of the same magnitude for ⬍2 min/day exhibited significant increases in periosteal bone formation. Her˘t10 and Lanyon and Rubin12 used continuous static loading in their experiments, which may also have simulated immobilization. The application of short-duration, daily bouts of static loading might better reveal the effects of static loading on appositional dynamics, without the potentially complicating effects of immobilization. It should be noted that other investigators have found positive osteogenic effects associated with static loading3,8 and, in some cases, dose responses have been reported.15,16 In light of the

Address for correspondence and reprints: Alexander G. Robling, Ph.D., Department of Anatomy and Cell Biology, Indiana University School of Medicine, 635 Barnhill Drive, MS 5045, Indianapolis, IN 46202. E-mail: [email protected] © 2001 by Elsevier Science Inc. All rights reserved.

105

8756-3282/01/$20.00 PII S8756-3282(01)00488-4

106

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

Bone Vol. 29, No. 2 August 2001:105–113

Figure 1. (A) Schematic diagram of the rat ulna loading model. The right distal forelimb is held between upper and lower aluminum cups (shown in hemisection), which are fixed to the loading platens. When force is applied to the upper platen (large arrows), the preexisting mediolateral curvature of the ulnar diaphysis becomes accentuated and translates most of the axial load into a bending moment (small arrow), which is maximal near the midshaft. (B) After sacrifice, the radius and ulnae were sectioned in the coronal plane at the distal end, in the parasagittal plane at the proximal end, and in the transverse plane near (2 mm distal to) the diaphyseal midshaft. (C) The proximal and distal sections revealed the growth plate, metaphysis, and epiphysis; the diaphyseal sections revealed fluorochrome labeling (not shown; see Fig 2) on endocortical and periosteal surfaces.

inconsistencies regarding the osteogenic potential of static loading, we sought to elucidate the effects of static loading on appositional bone growth, using a nonsurgical ulna loading preparation in which both periosteal and endocortical surfaces could be examined without complications arising from injury or trauma. Specifically, we were interested in addressing the effects of daily short-duration (10 min/day) static loads on periosteal and endocortical bone formation. One of the side effects associated with the rat ulna loading model32 is longitudinal growth suppression, which has been hypothesized previously to be proportional to average load.18 The design of our experiment also permitted us to investigate the effects of average load on longitudinal growth suppression in the rat ulna. This was accomplished by applying the same average load in two different waveforms (one static, one dynamic). We predicted equal longitudinal growth suppression in ulnae loaded at 17 N of dynamic force and 8.5 N of static force in light of their equivalent average load, and greater growth suppression in ulnae loaded at 17 N of static force, given its greater (⫻2) average load. We also predicted that the suppression of longitudinal growth in loaded ulnae is associated with a reduction in growth plate thickness and a change in chondrocyte populations. In summary, we tested the following hypotheses: (1) Short (10 min/day) bouts of static loading, applied to the growing rat ulna at a peak force of 17 N or 8.5 N, suppresses periosteal and endocortical bone formation; 17 N of dynamic loading enhances bone formation on both surfaces. (2) Longitudinal growth suppression of the ulna is proportional to average load and is associated with decreased growth plate thickness and altered physeal chondrocyte populations. Materials and Methods Twenty-nine growing male Sprague-Dawley rats were purchased from Harian Sprague-Dawley (Indianapolis, IN) at a mean body

mass of 192 ⫾ 15 g. The rats were received 1 week before the experiment began (acclimation period) and were housed two per cage at Indiana University’s Laboratory Animal Resource Center (LARC), where they were provided standard rat chow and water ad libitum daily during the acclimation and experimental periods. Daily body mass measurements were collected from the beginning of the acclimation period through the day of sacrifice. During the experimental period, all rats were subjected to a daily 10 min session of axially applied compressive end-loading of the right ulna while under general anesthesia (ethyl ether inhalation). The ulnar loading model is a nonsurgical preparation that applies mechanical force to the ulna through the olecranon and flexed carpus32 (Figure 1A). The natural curvature of the ulnar diaphysis translates the axial load into a bending moment in the middiaphysis that produces tension in the lateral cortex and compression in the medial cortex—a strain distribution similar to that resulting from normal ambulation in vivo.19 Force was applied to the ulnae by an electromagnetic vibrator with feedback control,7 or by an open loop, stepper motor-driven spring linkage with an in-line load cell.33 Both systems were modified to accommodate the morphology of the rat forearm and wrist. All procedures performed in this experiment were in accordance with the Indiana University Animal Care and Use Committee guidelines. Experimental Design On day 1 of the experimental period, the rats were assigned randomly to one of three groups (n ⫽ 9 or 10). The three groups differed in the nature (static vs. dynamic) and magnitude (17 N vs. 8.5 N) of the force applied to the right forearm. Group 1 received 1200 cycles of a 17 N dynamic load, applied as a haversine waveform at a frequency of 2 Hz. In the growing male Sprague-Dawley rat of this body mass, 17 N elicits a compressive strain of approximately 3500 ␮ε on the medial surface of the

Bone Vol. 29, No. 2 August 2001:105–113

ulnar midshaft,19 which is greater than peak compressive strains normally encountered during unrestricted running (1200 ␮ε) or jumping down from a 30 cm platform (2500 ␮ε).19 Putting these strains into a human physiology context, peak compressive strains in the adult human tibial shaft during vigorous exercise reach approximately 1200 ␮ε2. Group 2 was subjected to 10 min of static loading at 17 N. Group 3 received 10 min of static loading at 8.5 N, which elicits a compressive strain of approximately 1300 ␮ε on the medial surface of the ulnar midshaft.19 In both static loading groups, force was applied and released at a rate approximating 1 N sec⫺1. Left ulnae were not loaded and served as controls for the right (loaded) ulnae. All animals were allowed normal cage activity between the daily loading sessions. Loading was administered on days 1–5 and 8 –12 of the experimental period. On days 5 and 12, the rats were given an intraperitoneal injection of calcein (7 mg/kg body mass; Sigma, St. Louis, MO) after loading. All animals were sacrificed on day 16. Tissue Processing and Preparation Immediately after sacrifice, the right and left radius and ulna were dissected out of the forearm en bloc (interosseous membrane and annular ligament remained intact). The maximum proximodistal lengths of the ulnae (olecranon to styloid process) were measured with digital calipers to the nearest 0.1 mm. Subsequently, the proximal and distal ends of each ulna/radius were cleaved from the diaphyses 7 mm from each end (Figure 1B). The three portions (proximal end, distal end, diaphysis) were immersed in 10% neutral buffered formalin for 48 h, dehydrated in graded alcohols, cleared in xylene, and embedded in methylmethacrylate. Using a diamond-embedded wire saw (Histo-saw; Delaware Diamond Knives, Wilmington, DE), transverse thick sections (⬇70 ␮m) were removed from the diaphyseal blocks at a point 2 mm distal to the ulnar midshaft, and were mounted unstained on standard microscope slides. This diaphyseal location has been reported to show the greatest response to loading.19 The proximal and distal blocks were sectioned longitudinally at a thickness of 5 ␮m using a Reichert-Jung 2050 Supercut microtome (Reichert-Jung, Inc., W. Germany). The distal blocks were sectioned in the coronal plane. Five sections were cut from the region midway between the ventral and dorsal surfaces and were mounted on charged microscope slides. The proximal blocks were sectioned in the parasagittal plane. Five sections from the point midway between the medial and lateral surfaces of the olecranon were mounted on charged slides. The thin sections from the proximal and distal portions, which revealed the epiphysis, growth plate, and metaphysis at both ends of the ulna, were stained with either MacNeal’s tetrachrome or toluidine blue (Figure 1C). Data Collection and Analysis All slides were read on a Nikon Optiphot fluorescence microscope at ⫻125 magnification, and primary data were collected using the BIOQUANT 98 (R&M Biometrics, Nashville, TN) image analysis system. From the transverse thick sections of the ulnar diaphysis, the following primary histomorphometric data were collected at the endocortical and periosteal surfaces: total perimeter (B.Pm); single label perimeter (sL.Pm); double label perimeter (dL.Pm); and double label area (dL.Ar). From these primary data, the following first- and second-order derived quantities were calculated for both surfaces: mineralizing surface (MS/ BS ⫽ [1⁄2sL.Pm ⫹ dL.Pm]/B.Pm; %); mineral apposition rate (MAR ⫽ dL.Ar/dL.Pm/7 days; ␮m/day); and bone formation rate (BFR/BS ⫽ MAR ⫻ MS/BS ⫻ 3.65; ␮m3/␮m2 per year).23 On the proximal and distal thin sections, the following mea-

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

107

surements and counts were made from the ulnae: growth plate area (GPl.Ar); growth plate (proximodistal) height (GPl.Ht); hypertrophic zone (proximodistal) height (Hp.ZHt); number of proliferating chondrocyte lacunae (NPr.Ce.Lc); and number of hypertrophic chondrocyte lacunae (NHp.Ce.Lc). GPl.Ar was measured by manually tracing the growth plate borders in BIOQUANT. The metaphyseal and epiphyseal borders of the growth plates were defined by the extent of toluidine blue staining of cartilage. GPl.Ht and Hp.ZHt were measured and averaged along eight equidistant test lines superimposed over the digitized image of the growth plate. The microscope stage was rotated for each section so that the test lines were aligned parallel to the long axis of the ulna. The metaphyseal borders used for measuring GPl.Ht and Hp.ZHt were the same as those used for GPl.Ar. The epiphyseal border used for GPl.Ht was the same as that used for measuring GPl.Ar. The hypertrophic-proliferative boundary used for measuring Hp.ZHt was defined as the point at which the flattened cell lacunae of the proliferative zone exhibited proximodistal height that was approximately equal to or greater than the transverse dimension. Hypertrophic cell lacuna counts were conducted within the hypertrophic zone as defined earlier. Proliferative cell lacuna counts were conducted in the region between the hypertrophic-proliferative boundary and the epiphyseal extent of the columnar arrangement of cells. Cell lacuna counts were normalized by growth plate area to control for differences in animal size and in sectioning plane. To separate the effects of treatment from individual differences in growth rates and hormonal influences, relative values, calculated by subtracting left limb values from right limb values, were derived for all of the variables. Right (loaded) vs. left (control) differences were tested for significance using paired t-tests. Differences among relative group means were tested for significance by analysis of variance (ANOVA), followed by Fisher’s protected least significant difference (PLSD) test for all pairwise post hoc comparisons. Results Three of the 29 rats died during the experiment from anesthesiarelated complications. Mean body mass increased by approximately 90 g over the experimental period (213 ⫾ 27 g on first load day; 305 ⫾ 16 g at sacrifice). No significant differences in initial or final body mass were found among groups. No evidence of swelling or limb-use impairment was observed in any of the animals during the experiment. Appositional Bone Formation Both static groups exhibited suppressed periosteal apposition in the loaded ulna (Figure 2), resulting in a 28%– 41% disparity between loaded and control limb periosteal BFR/BS and MAR (Table 1). Conversely, dynamic loading of the right (loaded) limb enhanced periosteal BFR/BS and MAR by 78% and 67%, respectively, when compared to the left (control) limb. Right vs. left differences in periosteal MS/BS were not significantly different for either of the 17 N groups, but the 8.5 N static group exhibited significantly lower periosteal MS/BS in the right ulna when compared to the left ulna. Post hoc comparisons among relative (right minus left) periosteal BFR/BS and MAR values revealed no statistical difference between the two static groups, but the 17 N dynamic group was significantly different from both static groups (Figure 3). Periosteal rMS/BS was significantly different among all three groups. Bone formation on the endocortical surface appears to have been neither enhanced nor suppressed by static loading (Figure 2). No significant right vs. left differences were found for any of the

108

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

Bone Vol. 29, No. 2 August 2001:105–113

Figure 2. Double fluorochrome labeling in the ulnar diaphysis illustrates the osteogenic potency of dynamic loading (lower right panel) on the periosteal and endocortical surfaces, compared to the normal growth-related bone formation exhibited in the control limb (lower left panel). Static loading at both high (17 N) and moderate (8.5 N) magnitudes (upper and middle panels) suppressed bone formation on the periosteal surface— particularly in the caudal and lateral regions—when compared with control limbs, but had no detectable effect on the endocortical surface.

derived endocortical parameters among the static loading groups (Table 1). Conversely, endocortical MAR and BFR/BS were 2.5 and 3.1 times greater, respectively, in the loaded ulna of the 17 N

dynamic group when compared to the control side. Endocortical MS/BS was also significantly greater— by 26%—in the right ulna of the 17 N dynamic group.

Table 1. Summary of appositional bone formation in the ulnar diaphysis following 2 weeks of a daily mechanical loading stimulus Periosteal surface Load group (side) 8.5 N static (n ⫽ 8) Control (L) Loaded (R) 17 N static (n ⫽ 9) Control (L) Loaded (R) 17 N dynamic (n ⫽ 9) Control (L) Loaded (R)

MAR (␮m/day)

MS/BS (%)

Endocortical surface BFR/BS (␮m3/␮m2 per day)

4.06 (0.27) 2.65 (0.17)

}

99.7 (0.3) 89.7 (1.9)

}

1477.1 (98.0) 865.3 (54.3)

3.95 (0.18) 2.91 (0.27)

}

94.9 (2.2) 91.4 (3.5)

}n.s.

1356.9 (33.9) 973.1 (103.4)

b

3.50 (0.18) 5.86 (0.52)

}

96.0 (1.7) 98.6 (1.4)

}n.s.

1229.6 (70.5) 2128.3 (199.7)

b

c

a

b

Values presented are means and (standard errors). n.s., not significant (p ⬎ 0.05). a Paired t-test significant at ␣ ⫽ 0.05. b Paired t-test significant at ␣ ⫽ 0.01. c Paired t-test significant at ␣ ⫽ 0.001.

b

}

c

} }

MAR (␮m/day)

MS/BS (%)

BFR/BS (␮m3/␮m2 per day)

1.38 (0.32) 1.76 (0.39)

}n.s.

56.2 (3.0) 54.7 (5.5)

}n.s.

288.5 (68.7) 394.6 (101.2)

1.43 (0.40) 1.62 (0.35)

}n.s.

56.5 (6.5) 57.0 (7.8)

}n.s.

341.8 (98.8) 373.5 (84.5)

1.58 (0.44) 5.55 (0.55)

}

61.4 (5.1) 77.4 (4.1)

}

b

a

}n.s.

}n.s.

370.7 (109.5) 1530.9 (120.6)

}

c

Bone Vol. 29, No. 2 August 2001:105–113

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

109

in length was not significantly different between the 17 N dynamic and 17 N static groups; however, both 17 N groups exhibited significantly greater right vs. left length differences than the 8.5 N group (Figure 4). Measurements recorded from the distal ulnar growth plates revealed that the 17 N static and dynamic groups exhibited significantly thicker (95% and 170% thicker, respectively) growth plates on the loaded side when compared to the control side (Table 2). The 8.5 N group exhibited no significant right vs. left difference in total growth plate height or in any of the other measurements at the distal growth plate. In the two 17 N groups, the hypertrophic zone of the right distal physis was markedly increased in height— by three to five times in the static and dynamic groups, respectively— compared to the nonloaded side (Figure 5). This difference was also reflected in the normalized number of hypertrophic cell lacunae, which were nearly twice as numerous in the loaded side in both 17 N groups. Conversely, the number of proliferating chondrocyte lacunae per unit growth plate area was reduced by 45% and 61% in the loaded limbs of the 17 N static and dynamic groups, respectively. Post hoc comparisons among relative (right minus left) means for the distal growth plate measurements revealed significantly fewer proliferative cell lacunae per unit area in the two 17 N groups when compared to the 8.5 N group. The relative growth plate height and relative hypertrophic zone height were significantly different in all pairwise comparisons among the three load groups, with the 17 N dynamic and 8.5 static group exhibiting the greatest and least height, respectively. The proximal growth plate revealed no significant right vs. left differences in height measurements or in normalized cell lacuna counts for any of the groups, with the exception of hypertrophic chondrocyte lacuna number in the two 17 N groups (Table 2). Normalized hypertrophic cell lacuna counts were slightly greater (p ⬍ 0.05) in the control limbs than in the loaded limbs among these two groups. Discussion

Figure 3. Relative (right minus left) bone formation rate (A), relative mineral apposition rate (B), and relative mineralizing surface (C) were suppressed on the periosteal surface but unaffected on the endocortical surface in the two static loading groups. Dynamic loading was stimulatory on both surfaces. Asterisk indicates significant difference from 8.5 N static group at ␣ ⫽ 0.05; dagger indicates significant difference from 17 N static group at ␣ ⫽ 0.05 (Fisher’s PLSD). Error bars indicate ⫾ 1 standard error. See Table 1 for significance of right vs. left comparisons.

Post hoc comparisons of endocortical rBFR/BS, rMAR, and rMS/BS values revealed no statistical difference between the two static groups (Figure 3). The 17 N dynamic group was significantly different from both static groups in all of the relative endocortical parameters, with the exception of the comparison to the 17 N static group for rMS/BS. Longitudinal Growth Right (loaded) ulnae were significantly shorter than left (control) ulnae in all 3 groups (Table 2). The mean right vs. left difference

Our main objective in this study was to determine the effects of brief (10 min) daily bouts of static loading on periosteal and endocortical bone formation. Despite the fact that 99% of each loading day (23.8 h/day) the right limb was involved in normal ambulatory activity, we found that periosteal bone apposition was suppressed significantly by brief daily bouts of static loading. Thus, a potent inhibitory signal was created during the 10 min loading period, which essentially muted any otherwise osteogenic signals created during the remainder of the day. The two static load magnitudes appear to have had similar inhibitory effects on periosteal bone formation, which suggests that static load magnitude, at least in the range we used, is of minor importance in suppressing periosteal apposition. The main effect of loading on periosteal cell populations in all three load groups appears to have been on the rate of new bone apposition (MAR), with relatively little difference between loaded and control limbs in the number of active cells on the periosteal surface (MS/BS). The lack of effect in MS/BS is not surprising given that the animals were rapidly growing and, consequently, nearly the entire periosteal surface (⬇90%) of the ulnar midshaft was already a forming surface. The significantly lower MS/BS found in the loaded (compared to control) limb of the 8.5 N static group appears to be the result of a relatively high (⬇100%) forming surface in the control limb, rather than a substantially lower forming surface in the experimental limb. As expected, 17 N of dynamic loading created a potent osteogenic stimulus on the periosteal surface. In contrast to the results reported by Mosley et al.,17 we also found a significant osteo-

96.1 (6.0) 540.7 (67.2)

} 279.8 (19.4) 755.6 (51.3) a

} 686.0 (36.9) 553.2 (20.1)

}n.s. 1575.1 (70.6) 1624.4 (86.6)

}n.s. 64.4 (2.7) 64.1 (5.1)

}n.s. 189.0 (8.1) 197.4 (8.7) c

} 30.76 (0.18) 29.52 (0.16)

Values presented are means and (standard errors). KEY: GPl.Ht, total growth plate height (␮m); Hp.ZHt, hypertrophic zone height (␮m); L, total length of the ulna, (mm); NHp.Lc/GPA, number of hypertrophic chondrocyte lacunae/mm2 of total growth plate area (includes all zones); NPr.Lc/GPA, number of proliferating chondrocyte lacunae/mm2 of total growth plate area (includes all zones); n.s. not significant (p ⬎ 0.05). a Paired t-test significant at ␣ ⫽ 0.05. b Paired t-test significant at ␣ ⫽ 0.01. c Paired t-test significant at ␣ ⫽ 0.001.

a

}

571.0 (19.0) 1008.4 (168.6) c

}

2273.3 (95.7) 890.1 (118.8) c

}

c

}

a

}

540.5 (35.2) 906.9 (147.4) 2277.0 (105.6) 1250.8 (185.2) b

}

100.3 (5.1) 342.3 (57.8) b

} 279.3 (13.0) 543.6 (62.0) a

} 700.4 (22.4) 573.1 (53.8)

}n.s. 1454.7 (57.6) 1400.7 (84.9)

}n.s. }

196.2 (6.3) 64.5 (2.5) 241.2 (22.5) n.s. 73.4 (6.4)

} 30.83 (0.09) 29.62 (0.09)

c

} }n.s.

298.7 (30.1) 309.3 (19.2)

}n.s. 738.3 (52.6) 682.5 (33.5)

}

1685.3 (46.2) 1676.8 (132.4) n.s.

}n.s. 64.4 (4.4) 57.5 (2.8)

}n.s. 198.0 (6.8) 187.5 (9.3)

}

8.5 N static Control (L) Loaded (R) 17 N static Control (L) Loaded (R) 17 N dynamic Control (L) Loaded (R)

30.85 (0.15) 30.18 (0.14)

c

b

550.5 (54.2) 599.5 (77.1) 2135.7 (73.8) 2157.6 (196.2) n.s. 115.8 (15.8) 122.1 (22.1) n.s.

}

NPr.Lc/GPA

Distal growth plate

Hp.ZHt GPl.Ht NHp.Lc/GPA NPr.Lc/GPA

Proximal growth plate

Hp.ZHt GPl.Ht L

Load group (side)

Table 2. Summary of longitudinal growth and growth plate measurements in the ulna following 2 weeks of a daily mechanical loading stimulus

}n.s.

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

NHpc.Lc/GPA

110

Bone Vol. 29, No. 2 August 2001:105–113

Figure 4. Relative (right minus left) gross length of the ulna. The two 17 N groups (static and dynamic loading) exhibited approximately twice the suppression of growth as that measured in the 8.5 N group. Asterisk indicates significant difference from 8.5N static group at ␣ ⫽ 0.05 (Fisher’s PLSD). Error bars indicate ⫾1 standard error. See Table 2 for significance of right vs. left comparisons.

genic effect on the endocortical surface in the 17 N dynamic group. Our results regarding the suppressive effects of static loading on periosteal apposition in growing animals are consistent with those reported by Her˘t et al.,10 who subjected growing rabbit tibiae to continuous static loading for 13 months and found suppressed periosteal apposition at the midshaft. Although the static loading study by Lanyon and Rubin yielded no effect of static loading on periosteal apposition, their results are not inconsistent with ours. Their study was conducted on skeletally mature animals, which typically exhibit a quiescent periosteum. In the absence of a growing periosteum, it is not possible to suppress appositional growth. Thus, the maturity of the animal (and metabolic activity of the resident periosteal cells) plays a major role in the suppressive vs. null periosteal response to static loading. Other studies, however, have found positive osteogenic effects associated with continuous static loading of rabbit calvariae8 and dog femora.16 One often-cited explanation for the positive association is that the statically loaded skull and limb bones in those experiments also could have been exposed to a significant superimposed dynamic loading component arising from daily activities (e.g., mastication or locomotion), which could stimulate the cells to activate or increase bone production.12,14 This explanation is less likely to account for the osteogenic effects of static loading found by Chamay and Tschantz,3 who applied continuous static bending to the ulnae of adult dogs, immediately after which the entire limb was encased in a plaster cast to avoid superimposition of dynamic loads associated with locomotion. The ulna was bent 15°–30°, which was severe enough to deform the bone in the plastic phase. Consequently, significant damage to the tissue was observed histologically, and the investigators noted that bone formation occurred only in regions of the cortex exhibiting plastic slip lesions, which suggests that the osteogenic response might have arisen from an injury response, rather than a coordinated mechanical adaptation. To our knowledge, only two studies other than this one have addressed the effects of daily (discontinuous) static loading bouts on bone formation. Turner et al.35 applied a static bending moment to the tibiae of mature rats for 18 sec/day for 2 weeks. The endocortical bone formation rate measured in the static bending tibiae was not different from that measured in the control (nonloaded) side, a result consistent with Her˘t’s contin-

Bone Vol. 29, No. 2 August 2001:105–113

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

111

Figure 5. Photomicrographs of the distal ulnar growth plate, epiphysis, and metaphysis from rats subjected to a daily 10 min bout of loading for 2 weeks. The superimposed column toward the right of each photomicrograph denotes the borders used for measurements of the growth plate: total column height denotes growth plate height (GPl䡠Ht); the hatched region denotes the hypertrophic zone height (Hp䡠ZHt) and defines the metaphyseal and epiphyseal boundaries within which the number of hypertrophic cell lacunae (NHp䡠Ce䡠Lc) counts were collected; the horizontally striped region defines the proliferating zone boundaries, within which the number of proliferating cell lacunae (NPr䡠Ce䡠Lc) counts were collected. Note in the 17 N static and dynamic groups the large increase in the height of the growth plate, the majority of which is accounted for by an increase in the height of the hypertrophic zone, and the irregular-shaped lacunae with lack of columnar arrangement in the hypertrophic layer. Left (control) limb distal growth plates did not exhibit significant variation among loading groups for any of the measurements; therefore, the growth plate illustrated in the left panel can be considered as a control for each of the loaded growth plates to the right. The epiphysis is toward the top.

uous static loading experiments.10 Periosteal dynamics could not be evaluated in that experiment, however, because of limitations of the loading model. The other study, by McDonald et al.,15 employed Her˘t’s9 surgical model to investigate the effects of daily bouts (45 min/day for 2 weeks) of static bending on periosteal and endocortical bone formation in mature rabbit tibiae, using three different force magnitudes. They found a positive dose response between static load magnitude and new bone area, culminating in a 18% increase in new bone area among the highest (9.8 N) static loading group. However, these results should be considered in light of some limitations of the experiment, including: (1) internal controls were not employed (some animals were loaded in static bending in one limb and dynamic bending in the other); and (2) large portions of the endocortical surface were resorbed and replaced with woven bone (in one case, the entire thickness of a portion of the cortex was resorbed and replaced with woven bone). Thus, it is difficult to determine whether the changes in new bone area were mechanically adaptive or initiated by an injury response. A clear advantage to using the rat ulna loading model in studying the effects of diaphyseal bone adaptation is that: (1) the load is applied noninvasively; and (2) force is applied through

soft tissues at sites remote from those being examined histologically. Despite these advantages, the possibility that diaphyseal bone formation is a response to injury or trauma cannot be ruled out, because there is currently no sham loading configuration for the model. It is also unclear how much potential irritation/ inflammation effects have on sites located closer to the points of force application (e.g., growth plates). Another of our objectives in this study was to determine whether suppression of longitudinal growth in the rat ulna was proportional to average load. Contrary to our predictions, the data demonstrate that 17 N of dynamic loading and 17 N of static loading resulted in: (1) equal growth suppression; and (2) approximately twice the growth suppression observed in the 8.5 N static group. The results indicate that growth was suppressed in proportion to the peak load magnitude rather than to average load, and that the suppression was independent of the cyclic nature of the load. The longitudinal growth suppression observed in the two 17 N groups was associated with a number of changes in the distal ulnar growth plate, including an increase in distal growth plate height—particularly at the hypertrophic zone, and a concomitant accumulation of hypertrophic cell lacunae. The inverse associa-

112

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

tion between longitudinal growth (ulna length) and growth plate height found in this study was unexpected and is inconsistent with a number of other in vivo loading studies that have shown a positive, proportional association between growth plate height and longitudinal growth rate28; i.e., shorter bones have been associated with thinner growth plates and vice versa.17,20,29 –31 A few in vivo loading studies have reported thicker growth plates associated with shorter long bones in overloaded limbs, but these effects were associated with a diagnosis of dyschondroplasia (osteochondrosis).5,6,11,22,36 We showed in a subsequent study,21 in which rats were sacrificed immediately after the last loading session, that the enlarged growth plates observed in this study were not the result of an accelerated growth reaction, or “growth recovery,” during the 4 day period between cessation of loading and sacrifice. The growth suppression observed in loaded limbs of the rats in our study could have been the effect of damage to the tissue;21 however, the lower load group (8.5 N) exhibited growth suppression as a result of loads normally encountered during natural ambulation in vivo.19 It remains unclear whether the changes in growth plate activity and longitudinal growth were the result of mechanical adaptation or a response to damage. The significant reduction in proliferating chondrocyte lacunae per unit area observed in the two 17 N groups appears to be an artifact of the expanded total growth plate area in these groups. If hypertrophic zone height is subtracted from total growth plate height, the remainders are not significantly different among groups (data not shown), which suggests that the proliferative and resting zones are not significantly different among groups. This also explains why the number of hypertrophic cell lacunae was not significantly different among the groups despite the fact that it was in this zone that the vast majority of the effect was manifest. Thus, our data do not address whether the proliferative zone was truly affected, but it is clear that the hypertrophic zone was. The growth-suppressive effects of loading appeared to be limited to the distal growth plate. With the exception of hypertrophic lacunae counts in the two 17 N groups, no detectable changes were observed at the proximal growth plate. The lack of effect at the proximal growth plate could be a function of growth plate cross-sectional area—the distal growth plate is smaller in cross section and therefore would develop more stress than the proximal growth plate for the same applied load. More likely, however, the disparity reflects differences in normal synthetic activity at the two bone ends. Human studies have shown that 85%–95% of longitudinal growth of the ulna occurs at the distal growth plate.24 If the rat ulna exhibits the same degree of disparity in proximal vs. distal growth rates, the proximal epiphysis might not have been engaged in sufficient growth during the 2 week experimental period to accrue any significant effects of loading. In conclusion, we found that brief (10 min) daily bouts of static loading applied to the growing rat ulna significantly suppressed periosteal apposition despite normal, presumably growth-promoting use of the limb between daily sessions. The degree of suppression was not related to static load magnitude in the range used. No effect of static loading was found on the endocortical envelope. Conversely, dynamic loading had clear anabolic effects on both periosteal and endocortical surfaces. Application of force via the rat ulna loading model resulted in significant longitudinal growth suppression, in proportion to peak load magnitude. Longitudinal growth suppression was associated with a number of changes at the distal growth plate, including increases in growth plate height, hypertrophic zone height, and hypertrophic cell lacuna number. The data suggest that bone adaptation to increased loads is driven by signals

Bone Vol. 29, No. 2 August 2001:105–113

originating from dynamic rather than static forces—a result consistent with in vitro1,26 and in vivo34,35 evidence for the role of fluid flow in bone adaptation.

Acknowledgments: The authors thank Mary Hooser, Diana Jacob, and Thurman Alvey for assistance with tissue processing. This work was supported by NIH Grants AR43730 and T32, AR07581.

References 1. Ajubi, N. E., Klein-Nulend, J., Alblas, M. J., Burger, E. H., and Nijweide, P. J. Signal transduction pathways involved in fluid flow-induced PGE2 production by cultured osteocytes. Am J Physiol 276:E171–E178; 1999. 2. Burr, D. B., Milgrom, C., Fyhrie, D., Forwood, M., Nyska, M., Finestone, A., Hoshaw, S., Saiag, E., and Simkin, A. In vivo measurement of human tibial strains during vigorous activity. Bone 18:405– 410; 1996. 3. Chamay, A. and Tschantz, P. Mechanical influences in bone remodeling. Experimental research on Wolff’s law. J Biomech 5:173–180; 1972. 4. Churches, A. E. and Howlett, C. R. Functional adaptation of bone in response to sinusoidally varying controlled compressive loading of the ovine metacarpus. Clin Orthop 168:265–280; 1982. 5. Duff, S. R. Effect of unilateral weight-bearing on pelvic limb development in broiler fowls: Morphological and radiological findings. Res Vet Sci 40:393– 399; 1986. 6. Duff, S. R. Histopathology of growth plate changes in induced abnormal bone growth in lambs. J Comp Pathol 96:15–24; 1986. 7. Forwood, M. R., Bennett, M. B., Blowers, A. R., and Nadorfi, R. L. Modification of the in vivo four-point loading model for studying mechanically induced bone adaptation. Bone 23:307–310; 1998. 8. Hassler, C. R., Rybicki, E. F., Cummings, K. D., and Clark, L. C. Quantification of bone stresses during remodeling. J Biomech 13:185–190; 1980. 9. Her˘t, J., Lisˇkova´, M., and Landa, J. Reaction of bone to mechanical stimuli. 1. Continuous and intermittent loading of tibia in rabbit. Folia Morphol (Praha) 19:290 –300; 1971. 10. Her˘t, J., Lisˇkova´, M., and Landrgot, B. Influence of the long-term, continuous bending on the bone. An experimental study on the tibia of the rabbit. Folia Morphol 17:389 –399; 1969. 11. Jussila, J. and Paatsame, S. Radiological changes in the distal epiphysis, epiphyseal cartilage and metaphysis of the radius and ulna in pigs. A preliminary report of a study with special reference to the leg weakness syndrome. Acta Radiol 319(Suppl.):121–126; 1972. 12. Lanyon, L. E. and Rubin, C. T. Static vs dynamic loads as an influence on bone remodelling. J Biomech 17:897–905; 1984. 13. Lisˇkova´, M. and Her˘t, J. Reaction of bone to mechanical stimuli. 2. Periosteal and endosteal reaction of tibial diaphysis in rabbit to intermittent loading. Folia Morphol (Praha) 19:301–317; 1971. 14. Martin, R. B. and Burr, D. B. Structure, Function, and Adaptation of Compact Bone. New York: Raven; 1989. 15. McDonald, F., Yettram, A. L., and MacLeod, K. The response of bone to external loading regimens. Med Eng Phys 16:384 –397; 1994. 16. Meade, J. B., Cowin, S. C., Klawitter, J. J., Van Buskirk, W. C., and Skinner, H. B. Bone remodeling due to continuously applied loads. Calcif Tissue Int 36(Suppl.):S25–S30; 1984. 17. Montufar-Solis, D. and Duke, P. J. Gravitational changes affect tibial growth plates according to Her˘t’s curve. Aviat Space Environ Med 70:245–249; 1999. 18. Mosley, J. R. and Lanyon, L. E. Strain rate as a controlling influence on adaptive modeling in response to dynamic loading of the ulna in growing male rats. Bone 23:313–318; 1998. 19. Mosley, J. R., March, B. M., Lynch, J., and Lanyon, L. E. Strain magnitude related changes in whole bone architecture in growing rats. Bone 20:191–198; 1997. 20. Nyska, M., Nyska, A., Swissa-Sivan, A., and Samueloff, S. Histomorphometry of long bone growth plate in swimming rats. Int J Exp Pathol 76:241–245; 1995. 21. Ohashi, N., Robling, A. G., Burr, D. B., and Turner, C. H. Mechanical loading diminishes longitudinal growth rate and changes chondrocyte populations in the growth plate. J Bone Miner Res 15(Suppl. 1):S477; 2000. 22. Paatsama, S., Rokkanen, P., and Jussila, J. Etiological factors in osteochondritis dissecans. An experimental study into the etiological factors in osteochondritis dissecans in the canine humeral head using overloading with and without

Bone Vol. 29, No. 2 August 2001:105–113

23.

24. 25.

26.

27. 28. 29. 30.

somatotropin and thyrotropin hormone treatment and mechanical trauma. Acta Orthop Scand 46:906 –918; 1975. Parfitt, A. M., Drezner, M. K., Glorieux, F. H., Kanis, J. A., Malluche, H., Meunier, P. J., Ott, S. M., and Recker, R. R. Bone histomorphometry: Standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res 2:595– 610; 1987. Pritchett, J. W. Growth plate activity in the upper extremity. Clin Orthop 268:235–242; 1991. Raab-Cullen, D. M., Akhter, M. P., Kimmel, D. B., and Recker, R. R. Periosteal bone formation stimulated by externally induced bending strains. J Bone Miner Res 9:1143–1152; 1994. Reich, K. M., Gay, C. V., and Frangos, J. A. Fluid shear stress as a mediator of osteoblast cyclic adenosine monophosphate production. J Cell Physiol 143:100 –104; 1990. Rubin, C. T. and Lanyon, L. E. Regulation of bone formation by applied dynamic loads. J Bone J Surg [Am] 66:397– 402; 1984. Seinsheimer, F. D., and Sledge, C. B. Parameters of longitudinal growth rate in rabbit epiphyseal growth plates. J Bone Joint Surg [Am] 63:627– 630; 1981. Simon, M. R. The effect of dynamic loading on the growth of epiphyseal cartilage in the rat. Acta Anat 102:176 –183; 1978. Smith, S. D. Effects of long-term rotation and hypergravity on developing rat femurs. Aviat Space Environ Med 46:248 –253; 1975.

A. G. Robling et al. Modulation of bone growth by static and dynamic loads

113

31. Steinberg, M. E. and Trueta, J. Effects of activity on bone growth and development in the rat. Clin Orthop 156:52– 60; 1981. 32. Torrance, A. G., Mosley, J. R., Suswillo, R. F., and Lanyon, L. E. Noninvasive loading of the rat ulna in vivo induces a strain-related modeling response uncomplicated by trauma or periosteal pressure. Calcif Tissue Int 54:241–247; 1994. 33. Turner, C. H., Akhter, M. P., Raab, D. M., Kimmel, D. B., and Recker, R. R. A noninvasive, in vivo model for studying strain adaptive bone modeling. Bone 12:73–79; 1991. 34. Turner, C. H., Forwood, M. R., and Otter, M. W. Mechanotransduction in bone: Do bone cells act as sensors of fluid flow? FASEB J 8:875– 878; 1994. 35. Turner, C. H., Owan, I., and Takano, Y. Mechanotransduction in bone: Role of strain rate. Am J Physiol 269:E438 –E442; 1995. 36. Walker, T., Fell, B. F., Jones, A. S., Boyne, R., and Elliot, M. Observations on leg weakness in pigs. Vet Rec 79:472– 479; 1966.

Date Received: August 4, 2000 Date Revised: January 25, 2001 Date Accepted: March 14, 2001