| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institute of Food Nutrition and Human Health, Massey University, Palmerston North, New Zealand
Submitted 5 August 2004 ; accepted in final form 25 February 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
disuse osteoporosis
Previous workers have suggested, on the basis of extensive work with the tail suspension model, that local unloading or reloading causes a temporary inhibition of bone formation resulting principally from a concerted decrease in osteoblast numbers and rate of calcium accumulation (6, 9, 10, 23). Local changes in mechanical loading are thought to initiate bony changes that are limited to the unloaded limb via changes in perfusion pressure and arterial pressure causing alteration to the degrees of shear stress in the walls of the bony canaliculi (5), bringing about local changes in the levels of nitric oxide, prostaglandin E2 (24), and osteocalcin (20) and in local resistance to parathormone via increased levels of insulin-like growth factor I (13).
Later workers postulated that osteoporotic change in the tail suspension model that resulted from an increase in bone resorption, i.e., osteoclastic activity (28), could occur only if the animals became stressed and lost, or failed to gain, weight (17). These more general effects would be expected to bring about a generalized reduction of bone mineral density (BMD). In this event, there would be differing relationships between bony area and BMD in loaded and unloaded bones according to whether bony resorption was accompanied by cessation of mineralization.
Quantitative assessment of such local and general contributions to the genesis of immobilization osteoporosis is difficult using the tail suspension model because it does not allow for direct comparison of the effects of unloading and loading on functionally equivalent structures; i.e., both hindlimbs are rendered non-weight bearing. Moreover, the general elevation of the hindquarters that is incumbent in tail suspension influences a number of the vascular dynamics that have been implicated in the control of osteogenesis (8). Thus venous drainage is increased and arterial perfusion pressure reduced in both hindlimbs (5).
Again, venous perfusion pressure and pulse pressure are reported to be variably increased in the head and forelimbs, a result that has been linked in some cases to hyperostotic change in these bones (5). The contribution of changes in the rate of venous perfusion to modulation of bone dynamics has been confirmed by experiments where unilateral ligation of the femoral vein brought about ipsilateral increases in bony mineralization (2). Similar effects arising from impedance of venous drainage may complicate the interpretation of studies where unilateral unloading is achieved by strapping of the limb to the torso (14).
The purpose of the present study was to investigate the onset and amelioration of effects of mechanical unloading by unilateral splintage on the growth, mineralization, and mechanical properties of bones in the ipsilateral and contralateral limbs of 3-mo-old male Sprague-Dawley rats by using a novel splint design that prevented weight bearing but maintained the limb in a functional position without compression. This splint was designed to replicate the complete unloading that would be experienced during non-weight-bearing external splintage of a fracture in humans.
Cortical and trabecular bone may be affected differently by immobilization (14). In this study, the effect of mechanical unloading on a defined site of predominantly cortical bone (14), i.e., the central area of the metaphyses, was evaluated in view of the known differences in the elastic properties of cancellous and cortical bone (1, 4), the adaptation of cortical bone on the anterior surfaces of limbs to withstand tensional force and posterior cortical bone to withstand compression force (21), and the site-specific variation in response of cortical and cancellous bone to unloading (3).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The rats were housed separately in shoebox cages and maintained at a constant temperature (22 ± 2°C) and light-dark cycle (12:12-h day-night) in a dedicated room in the small animal unit at Massey University. All rats were given ad libitum access to a casein-based semisynthetic diet (Table 1) and deionized water. The diet was formulated based on AIN-93M but modified in terms of oil used. We routinely use corn oil instead of soybean oil because we require a neutral fat as a background source. Various kinds of fats have been investigated with regard to bone effects but corn oil is often used as a reference oil in the diets because it is predominantly omega-6. The vitamin and mineral mix were formulated by Crop and Food Research, New Zealand, according to National Research Council nutrient requirements for laboratory animals (AIN-93M) to ensure optimal levels of trace minerals to support bone health. These are formulated without calcium because we routinely modify the calcium of our diets according to protocol (11).
|
The study was approved by Massey University Animal Ethics Committee (02/47).
Experimental procedures. Rats from the splinted group were anesthetized, and a "MURFLEN" splint system was applied to one randomly selected hind leg (Fig. 1). The splint served to prevent weight bearing while maintaining the limb in extension without undue tension or compression by projecting rearward in such a manner as to prevent the extended leg being drawn into a functional weight-bearing position (Fig. 1). The proximal end of the splint was contoured to bear against the abdomen and groin, and the total length of the splint exceeded that of the limb by one-third. The splint was held in position by two longitudinally oriented tapes that were secured lengthwise to the skin of the medial and lateral aspects of the hindlimb by surgical Superglue and attached distally to a tie bar that was inset in the end of conical aluminium device (Fig. 1). Thus the ties prevented the limb from being withdrawn while the splint enclosed it entirely, protecting the securing tapes from interference. Prior trials have shown that, when applied, the splint did not interfere with general mobility but caused the leg to be trailed passively behind the animal.
|
Rats from the control group were similarly anaesthetized at the start of the experiment, but no splint was fitted. After this procedure they were also maintained in individual cages for the same time periods as the splinted animals, i.e., 7 wk, before being euthanized. This group formed a basis for comparison of the overall effects of splintage on total limb growth.
Immediately after completion of the 4-wk remobilization period, all rats were anesthetized, and both hind tibiae were again DEXA scanned. The animals were subsequently euthanized with 100% CO2 gas, and their femurs were removed for mechanical testing.
Anesthesia.
Anesthesia was induced with a mixture of 0.2 ml acepromazine + 0.5 ml ketamine + 0.1 xylazine + 0.2 ml sterile H2O, delivered intraperitoneally with a 25-gauge x 5/8-in. needle at 0.05 ml/100 g body wt. This gave a suitable level of anesthesia 5–10 min after injection that was maintained for 
2 h.
DEXA scans. BMD measurements were determined with a Hologic QDR4000 bone densitometer equipped with a pencil-beam unit (Bedford, MA). A daily quality control scan of a bone phantom was taken to check precision. The cumulative coefficient of variation (CV) for this was 0.44%. Regional high-resolution scans were performed using a 0.06-in.-diameter collimator with 0.0127-in. point resolution and 0.0254-in. line spacing. Rats were placed in a supine position on an acrylic platform of uniform 1.5-in. thickness for each of the two scans. Rats were positioned so that the femur was at right angles to the spine in the horizontal plane and the scanned tibia similarly at right angles to the femur.
The DEXA scans on rat bones were checked by conducting repeated scans of the same bone. The percent CV of BMD of rat tibial metaphyses was 1.0% with repositioning of the limb for successive measurements (n = 10). Similarly, the CV for area of tibial metaphyses was 2.5% (n = 10). These results are in the range reported by other workers (19)
Bone biomechanics. Bone strength assessments were conducted on the femur rather than the tibia because our experience had shown that the orientation of the triangular cross-sectional geometry of the midshaft of the tibia varied significantly between animals, necessitating variation in the direction of application of loading force, whereas the more constant and more circular midshaft geometry of the femur allowed for consistent anterior-posterior orientation of loading force.
The breaking strain, i.e., the percent deformation of the femur just before the time of breaking, and the breaking energy, i.e., the amount of energy required to break the femur, were determined using a Shimadzu Ezi-test (Kyoto, Japan) in the following manner. [These parameters were measured because a previous study has shown concordant reduction of bone mineral content (BMC) with strain (5a).]
Each femur was scraped clean of adherent flesh and stored in phosphate-buffered saline at –20°C. The bones were thawed to 23°C for biomechanical testing. The length of each femur was measured with an electronic caliper, and the midpoint was marked with a waterproof pen. The vertical and transverse radii at the midpoint of the femur were similarly determined. The femurs were subjected to a three-point bending test with the application point of the upper fulcrum positioned midway between the two supporting rods of the testing jig, which were spaced 15 mm apart. Load was applied at a constant deformation rate of 50 mm/min.
After biomechanical testing, each bone (along with any fragment material that was generated) was retained and dried overnight at 105°C in a forced-draft oven and subsequently weighed. The dry matter weights of the femurs from splinted and from unsplinted animals were used along with body weight to assess the overall effect of splintage on bodily growth.
Data manipulation and statistics. The DEXA scans of all hind limbs were examined so as to formulate an algorithm that consistently restricted the analysis to the region of calcified tibial metaphyses lying inner to the zone of cartilaginous growth and progressive columnar calcification that lay at the base of each epiphysis. The following procedure was used. The tibial bone map was visually inspected, and lines were placed at right angles to long axis of the tibia at the innermost points of the proximal and distal tibial epiphyses. The distance between the two lines was determined, and further lines were placed at right angles to the long axis of the tibia as follows: 1) at a point one-fourth of the length of the tibia from the most proximal point of the proximal epiphysis and 2) at a point one-third of the length of the tibia from the most distal point of the distal epiphysis. Thus the metaphyseal area was a consistent fraction that would increase in size with growth of the tibia.
Although the results of DEXA scans are commonly expressed as BMD, as discussed above, the two contributing parameters, bone mineral content (BMC) and bone area, may be affected differently by splintage according to whether splintage influences growth, bone mineral resorption or deposition locally or generally. Therefore, both metaphyseal area and total BMC were determined as well as BMD and the relationship between the former two parameters subsequently assessed in each of the three limb types.
We used ANCOVA to compare the relationship between metaphyseal area and BMC in the various limb "types" because this method allowed for assessment of differences in both slope and intersect.
Yield stress was assessed across the central 15 mm of the metaphyses with the opposing fulcrum applied at the midpoint and with identical orientation of the shaft, i.e., with the femoral head at an angle of 90° to the line of stress.
Statistical testing was conducted using the SYSTAT statistical software package.
Nested ANOVAs were carried out on data pertaining to limb type, i.e., limbs that were immobilized in splinted animals, the free limbs of splinted animals, and the limbs of unsplinted animals. Relationships within nested ANOVAs were explored post hoc by Bonferroni testing. Residuals from multiple regression were examined graphically for uniformity of distribution before being tested for differences between limb types by one-way ANOVA with post hoc Bonferroni testing. ANCOVAs were similarly conducted on data pertaining to each of the three limb types, with the presence or absence of significant interaction terms being used to indicate the presence of significant slope differences.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
|
|
It is noteworthy that, 4 wk after splint removal, the BMD values in the control tibial metaphyses were significantly greater (df = 1, 71; F = 13.1; P = 0.001) on repeated-measures ANOVA than those recorded 4 wk earlier but that the BMD values in neither the immobilized nor the free tibial metaphyses of splinted animals were significantly different from those recorded earlier. Thus, whereas the differences in BMD in the three limb types after 4-wk mobilization followed the same pattern as those immediately after removal of the splint, they were accentuated by increases in the BMD in the limbs of the unsplinted (control) animals. Thus the effects of splintage on BMD were immediately apparent in the immobilized limb but became more apparent in both limbs with the passage of time.
Bone growth and total BMC (Table 3). At the time of the removal of the splint, the areas of both the free and immobilized tibial metaphyses of splinted animals were significantly lower than those of unsplinted animals on nested ANOVA (df = 1, 139; F = 5.67; P = 0.019) with no significant difference on post hoc Bonferroni testing between the areas of tibial metaphyses in the immobilized and in the free limbs of splinted animals. Thus splintage equally reduced the areas of the tibial metaphyses in both the immobilized and the free limbs of the splinted animals compared with those in unsplinted animals.
However, after 4 wk of mobilization, there were no longer any significant overall differences, on nested ANOVA, between the tibial metaphyseal areas of splinted and unsplinted animals. Similarly, post hoc Bonferroni testing showed no differences between the tibial metaphyseal areas of the immobilized and the free limbs of splinted and subsequently mobilized animals. The reestablishment of the growth of the tibial metaphyses of both of the hindlimbs of splinted animals 4 wk after the removal of the splint is shown by significant increases in their area relative to those at the time of splint removal. Thus both the immobilized limbs (df = 1, 34; F = 7.65; P = 0.009) and the free limbs (df = 1, 34; F = 6.31; P = 0.017) of splinted animals as well as those in unsplinted animals (df = 1, 71; F = 6.91; P = 0.011) increased significantly in area on repeated-measures ANOVA. Thus, after 4 wk of remobilization, whereas the local effect of splintage on BMD persisted, the general effect of splintage on growth was no longer apparent.
The total BMC of the tibial metaphyses immediately after splintage was significantly different on nested ANOVA (df = 1, 126; F = 7.03; P = 0.009) between splinted and nonsplinted animals (P = 0.001) with significant differences between the free and immobilized tibiae of splinted animals (P = 0.022) on post hoc pairwise Bonferroni comparison (Fig. 3). Thus BMC varied both locally with immobilization and bilaterally with splintage.
|
|
A plot of BMC vs. metaphyseal area 4 wk after splint removal (Fig. 5) showed that the relationship of metaphyseal area with the BMC in the tibial metaphyses of the free limbs of previously splinted animals was no longer noticeably different from that in the other limb types, i.e., control or immobilized limbs. This finding was confirmed by the lack of either a significant interaction term on ANCOVA of "limb type" with BMC or an effect of limb type on the variation of BMC with metaphyseal area. Thus the distinctive effect of splintage on the relationship between BMC and metaphyseal area in the free limbs of splinted animals was no longer present 4 wk after the removal of the splint.
|
|
Breaking strain varied significantly with parameters of femoral size (df = 2, 39; F = 67.34; P < 0.0005), i.e., with femoral length (t = 4.11; P < 0.0005) (Fig. 6) and vertical diameter (t = 6.71; P < 0.0005), on multiple regression. The effect of splintage on breaking strain was explored by analysis of residuals from this regression. This showed an overall difference between the three limb types of borderline significance (df = 2, 137, F = 2.65; P = 0.074), whereas Bonferroni post hoc testing showed a difference of borderline significance between residuals from immobilized limbs and the limbs of unsplinted animals (P = 0.08) but not between free limbs of splinted animals and the limbs of unsplinted animals (P = 0.689). Thus the difference in gross breaking strain between the limbs of splinted animals and those of unsplinted animals was largely due to an effect of splintage on femoral size, perhaps with some local effect from immobilization.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The relative reduction of the metaphyseal area in both the immobilized and the contralateral limbs indicates a more general effect of splintage, perhaps acting via stress-induced glucocorticoid secretion, arresting growth, influencing collagen synthesis, and inducing osteoblastic apoptosis (22). The significant reduction in the bodily growth of splinted animals as represented by reduction in their body weight and femur dry tissue weight compared with those of unsplinted animals supports such an hypothesis.
The differing relationship between the metaphyseal area and BMC in the tibial metaphyses of the contralateral limbs of splinted animals compared with that in those of the immobilized limbs and the limbs of control animals also support the hypothesis that bony dynamics on immobilization are an outcome of local and general effects. Thus, whereas both metaphyseal area and BMC are relatively reduced in the immobilized limb, there is potential for ongoing mineralization in the contralateral limb where growth only is impaired. Because the data from this study pertain to the central metaphyseal area, the differences in the contralateral limb are unlikely to reflect previous findings (14) where, despite BMD remaining relatively constant during recovery from unilateral splintage, there was load-induced augmentation of cancellous bone formation. Rather, the differing dynamics may reflect differing levels of bony deposition and possibly resorption in anterior and posterior aspects of the cortex according to changes in the distribution of compression and tension in the contralateral limb (21).
The analysis of the changes in relationships between BMC and tibial metaphyseal area in the immobilized and free limbs of splinted rats after mobilization sheds light on the dynamics of mineralization during recovery from immobilization. Thus, although there is restitution of metaphyseal tibial area after 4-wk mobilization, there is incomplete reversal of the effect of splintage on mineral content. However, the similarity of slopes of the lines describing the relationship between BMC and metaphyseal area after mobilization for immobilized, free, and control limbs may indicate the return of synchrony between growth and recalcification. However, the significantly lower intersect point of the line for the data from immobilized limbs compared with those from the contralateral and control limbs is in keeping with a lasting "lag" effect from cessation of both growth and mineralization over the 4-wk period of mobilization as has been reported by other workers (14–16). A more protracted period of surveillance after splint removal would be required to evaluate the time required for full recovery.
The significant reduction of breaking strain and energy of both the immobilized and the free femurs of splinted animals compared with those of unsplinted animals indicates a general effect of splintage on these parameters that persisted for 4 wk after mobilization. However, the significant relationships between femoral size and breaking strain, and between femoral size and breaking energy, and the absence of highly significant effects when the effect of femoral size have been factored out, indicate that this results from a general effect of splintage on bone growth rather than a direct general effect on mineralization. Thus the lag effect on growth and mineralization referred to above has a lasting effect on growth and hence on elasticity and breaking strain.
In summary, our findings show that the pattern of changes in bone biometrics and mineralization associated with disuse osteoporosis in the central tibial metaphyses of young Sprague-Dawley rats is a complex outcome of local effects on calcification and general effects on growth as postulated by other workers (17). Moreover, the results demonstrated that the effects on growth indirectly influence the elastic properties of femur shafts via change in the relative area of the shaft that is subject to biomechanical testing.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
, Limbs of unsplinted rats; 
, immobilized limbs of splinted rats; 
, free limbs of splinted rats. Slope (P = 0.038) and intersect (P = 0.038) of the line for free limbs of splinted rats (short dashed line) were significantly different on ANCOVA from the limbs of unsplinted rats (top solid line).
