INTRODUCTION

Top Abstract Introduction References

LOSS OF MECHANICAL LOADING (i.e., disuse) substantially reduces bone mass in the adult skeleton. The cellular activity that is responsible for this adaptation is locally mediated because adjacent portions of the skeleton remain unaffected (22, 27). In skeletons demonstrating active forming and resorbing surfaces (e.g., young rodents), immobilization rapidly diminishes formation and increases resorption to create a bone-deficient vs. a normal animal (25). In the primarily quiescent mammalian skeleton, disuse rapidly activates osteoclastic resorption (9, 37). Osteoblastic activity normally coupled with resorption at intracortical remodeling sites is also diminished (15, 26). Although osteoclastogenesis and the cellular events resulting in bone resorption are well studied (38), the pathway by which disuse induces the quiescent skeleton to initiate local osteoclastic activation is not understood.

The vascular system has the potential to affect bone cell populations at a variety of levels (45). Like muscle, bone is populated by vessels containing endothelial cells (2). In response to the metabolic needs of a tissue, endothelial cells regulate tissue blood flow by expressing vasoactive substances (1). In turn, many vasoactive factors (e.g., prostaglandins) are known to regulate bone cell activity in vitro (10, 42).

Bone, like other tissues, becomes hyperemic in association with elevated cellular activity (13, 34, 35, 40, 44). Although vasoactive factors are capable of initiating osteoclastic differentiation from marrow precursors, evidence is equivocal at the in vivo level as to whether blood flow is altered before heightened cellular activity. For example, bone blood flow after 7 days of immobilization has been reported to be elevated (36) or unchanged (39). The use of animal models in which both osteoblasts and osteoclasts are active at the onset of the experiment, different blood flow measurement techniques, and partial weight bearing may contribute to this discrepancy.

If tissue vasoregulation actively participates in initiating osteoclastic activity, then vasoregulation must occur before the onset of the cellular response. We have attempted to address this question at the in vivo level by undertaking an experiment with the avian model of disuse osteopenia. In this model, the left ulna of a 1-yr-old, skeletally mature animal is isolated from mechanical loading (33). As in mammals, intracortical resorption and formation occur only at a low level in young adult turkeys (31). Disuse rapidly induces osteoclastic activation, with preliminary data indicating that osteoclasts are evident on intracortical bone surfaces by 10 days, but not after 7 days, of disuse (5). By measuring bone blood flow at time points conservatively bracketing osteoclastic activation (7 and 14 days), our study determined whether bone blood flow is altered before disuse-induced intracortical resorption.

	MATERIALS AND METHODS

Twenty-four adult white Leghorn roosters (age 1 yr) were randomly assigned to intact-control, 7-days-sham-surgery, 7-days-disuse, and 14-days-disuse groups. Animals included in the disuse groups underwent the functionally isolated avian ulna procedure (33). In this procedure, two cutaneous incisions (2 cm) were made over both metaphyses of the left ulna. The ulna was directly visible at both incision sites, and a parallel-sided template was clamped to the bone (the template is U shaped to avoid disturbing the diaphysis). With use of the template as a guide, two parallel osteotomies were made at both metaphyses with an oscillating saw. Small (2-mm) wafers of bone were removed at either site, and the exposed diaphyseal bone ends were covered with methacrylate-filled delrin caps. In this way, the diaphysis retained its musculature, vasculature, and innervation but was isolated from loading via the articulating joints (Fig. 1). Sham surgeries used an identical surgical approach, with partial osteotomies performed through one-third of both metaphyseal cortices to disrupt the marrow and blood supply yet maintain the structural integrity of the bone (1 sham surgery animal was removed from the study because of a spiral fracture originating at the distal partial osteotomy site). Intact-control animals were age matched but did not undergo surgery. The right ulna of each animal served as an intact contralateral control. The protocol complied with National Institutes of Health and Canadian Council on Animal Care standards for the care of animals and was approved by the University of Calgary Health Sciences Animal Care Committee.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Scanned radiograph of intact left avian wing (A) with humerus (h), radius (r), and ulna (u) noted. B: ulna procedure isolates diaphysis from mechanical loading via parallel osteotomies performed at both metaphyses. C: 2-mm-thick bone wafers resulting from osteotomies are removed, and exposed diaphyseal bone ends covered with methacrylate-filled delrin caps (light gray areas). Middiaphyseal sites at which bone blood flow and intracortical porosity were determined are noted (arrows).

Blood flow measurements were made by using colored microspheres. The roosters were anesthetized, and the aortic root was cannulated retrograde from the right carotid artery. Location of the cannula was confirmed by a pressure transducer. With use of a computer-controlled syringe pump, colored microspheres (4 × 10⁶ microspheres/kg, 15.5-µm diameter; Triton Technologies, San Diego, CA) in mixed suspension with saline and Tween 80 were injected into the aortic root over a 30-s period. Starting 10 s before the injection, a reference blood sample was drawn for 1 min at 3 ml/min from a second cannula placed in the left anterior tibial artery. After the animals were killed, the lungs, kidneys, and ulnae were removed for analysis.

Transverse cross sections (5 mm thick) were taken from the middiaphysis of the experimental and control ulnae. The 5-mm-thick sections were cleansed of marrow, weighed, decalcified (3 days in 10% HNO₃), and digested (1 days in 4 M KOH). The microspheres trapped within the bone samples were isolated by vacuum filtering the digested tissue through 7-µm-pore filters. The spheres were then visualized and directly counted by using an epifluorescent microscope (×40 magnification, 435-nm wavelength). Counting of microspheres was performed with the operator blinded to the identity of the samples.

The number of spheres trapped within each reference blood sample, lung, and kidney was determined by filtering the digested samples, eluting the dye contained within the spheres with N,N-dimethylformamide, and quantifying fluorescence with a spectrophotometer. Manufacturer calibration curves were used to convert absorbence readings to the number of spheres contained within each sample. Standardized flow (ml · min¹ · 100 g¹) was determined by dividing the number of spheres trapped within a tissue sample by the number of spheres contained in the reference blood sample and dividing this value by the mass of the tissue sample (19).

A 300-µm-thick cross section was also removed from each ulna adjacent to the proximal edge of the 5-mm section. These sections were hand ground to 100 µm by using 600-grit sandpaper and were microradiographed. Microradiographs were blinded and scanned at 1,200 dots/inch, resulting in an approximate resolution of 12 µm/pixel (a typical Haversian canal is 40-50 µm in diameter). An edge-detection algorithm was used to define intracortical porosity and endocortical envelope boundaries on the resulting gray-scale images (23). The range of pixel intensities within these boundaries was determined. On the basis of previous analysis, 1/ of the intensity range was used as a cutoff to define whether an individual pixel was included in the porosity (17). Area of porosity, number of porosities, and endocortical envelope area were determined by pixel summation.

Because the data were not normally distributed and group size was small, data were described by using median ± 95% confidence intervals (16). Nonparametric Wilcoxon tests (P = 0.05) were used to identify any differences between blood flow in the left (experimental) and right (control) ulnae of each group. Kruskal-Wallis one-way nonparametric ANOVA (P = 0.05) and Tukey nonparametric post hoc comparisons (P = 0.05) were used to determine whether blood flow within the left ulnae differed across treatment groups (all right ulnae were grouped together as controls in these analyses). For those measures in which left-right symmetry was anticipated (i.e., blood flow in the left and right ulnae of the intact control and sham surgery groups), the statistical power of the comparison (1  ) was estimated by using the mean-difference method (46).

	RESULTS

Blood flow was symmetric between the left and right kidneys and lungs (power = 0.87; Fig. 2). Median (± 95% confidence interval) blood flow was the same for the left and right ulnae of the intact-control group (1.6 ± 1.2 vs. 1.0 ± 0.8 ml · min¹ · 100 g¹; power = 0.95) and was not altered by sham surgery (2.6 ± 2.3 vs. 2.1 ± 1.8 ml · min¹ · 100 g¹; power = 0.92). After 7 days of disuse, middiaphyseal blood flow in the experimental bones was significantly and highly elevated compared with contralateral bone blood flow (6.5 ± 5.2 vs. 1.0 ± 0.8 ml · min¹ · 100 g¹; P = 0.03; Fig. 3). At 14 days of disuse, experimental bone blood flow was still elevated compared with the intact contralateral bones but to a lesser extent (3.7 ± 2.9 vs. 1.0 ± 0.8 ml · min¹ · 100 g¹; P = 0.03). The low level of intracortical porosity evident in time 0 control bones (0.002 ± 0.009 vs. 0.003 ± 0.013 mm²) was not altered in the 7-days-sham-surgery (0.000 ± 0.002 vs. 0.000 ± 0.000 mm²) or 7-days-disuse group (0.000 ± 0.002 vs. 0.001 ± 0.001 mm²). However, intracortical porosity of bones exposed to 14 days of disuse was significantly elevated compared with contralateral ulnae (0.015 ± 0.02 vs. 0.002 ± 0.002 mm²; P = 0.03; Fig. 4). Intracortical porosity after 14 days of disuse was also significantly elevated compared with the intact left ulnae of the other three groups (P = 0.03). The median number of porosities was unchanged in the left and right ulnae of the control (0.5 ± 2.5 vs. 0.5 ± 2.5), 7-days-sham-surgery (0.0 ± 1.0 vs. 0.0 ± 0.0), and 7-days-disuse groups (0.5 ± 0.7 vs. 1.0 ± 0.5) but was significantly elevated at 14 days of disuse (4.0 ± 8.1 vs. 1.0 ± 1.5; P = 0.03). Endocortical envelope area was not altered within any group.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Blood flow for left and right kidneys and lungs. Values are medians ± 95% confidence intervals. Blood flow was symmetric in these tissues, indicating that colored microspheres were evenly distributed throughout the animal.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Middiaphyseal bone blood flow in left [experimental (Exp)] and right (control) ulnae (A). Values are medians ± 95% confidence intervals. Blood flow in left and right ulnae of intact control group was equivalent, suggesting that measurement bias was small. Sham surgery did not alter blood flow, indicating that influence of surgery on middiaphyseal blood flow was minimal. Blood flow was significantly elevated after 7 (7 d) and 14 days (14 d) of disuse (* P = 0.03). At 7 days of disuse (B), each animal demonstrated elevated blood flow in experimental (X) vs. control (C) ulna.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Intracortical porosity for each of the treatment groups (A). Values are medians ± 95% confidence intervals. Porosity was not elevated by sham surgery or 7 days of disuse compared with intact bones (1 animal within intact group had bilateral porosities, which accounted for increased confidence intervals). Intracortical porosity was elevated at 14 days of disuse (* P = 0.03 vs. contralateral and intact left ulnae of other groups). At 14 days of disuse (B), each animal demonstrated elevated intracortical porosity in experimental (X) vs. control (C) ulna.

	DISCUSSION

In the quiescent skeleton, disuse rapidly precipitates locally mediated osteoclastic activation. Our in vivo data are the first, to our knowledge, to demonstrate that hyperemia precedes disuse-induced intracortical resorption. Given this temporal relation, we propose that bone vasoregulation is potentially associated with the cellular events leading to bone resorption.

Functional isolation of the avian ulna requires invasive surgery, which has the potential, in itself, to alter blood flow. Sham surgery was used to control for this influence. By disrupting the metaphyseal blood supply with partial osteotomies, we attempted to alter the bone blood supply in a manner similar to that experienced in the disuse group while minimally altering the mechanical environment of the middiaphysis. No change in middiaphyseal bone blood flow was observed in the ulnae undergoing sham surgery. This result was not unexpected because metaphyseal and diaphyseal blood supplies are primarily independent (28, 41). Thus we conclude that the surgical procedure alone was not responsible for the blood flow alterations observed at 7 days of disuse. Confirmation of these data in a noninvasive immobilization model would further substantiate this observation. Because the cellular response to surgery is extremely rapid (14, 43), we assumed that a 14-days-sham-surgery group would not provide additional insight.

Previous studies with the avian model indicate that osteoclastic and osteoblastic activities in the contralateral ulna are not altered by the experimental procedure (18, 31). Because bone blood flow had not previously been assessed in this model, an intact-control group was included in the experimental design. We found that bone blood flow in the contralateral ulnae of the 7-days-sham-surgery, 7-days-disuse, and 14-days-disuse groups was unchanged compared with basal blood flow measured in the intact control group. These data emphasize that disuse-induced vasoregulation is locally, rather than systemically, mediated.

The colored-microsphere techniques used in this study were adapted from methods we developed to measure ligament blood flow in rabbits (7). The underlying assumption of both radioactive- and colored-microsphere techniques is that the spheres are evenly mixed and distributed throughout the animal in proportion to blood flow (8). The major sources of error with microspheres are methodological (e.g., uneven mixing, counting errors) and stochastic (i.e., random variability in microsphere distribution) (4). Because we observed symmetric blood flow in the lungs, kidneys, and ulnae of the intact-control group, we concluded that the microspheres were evenly distributed through the animal and that our counting techniques were minimally biased.

Relative error in estimating bone blood flow has been experimentally determined (24). When 100-150 microspheres are contained within a bone sample, error ranges between 7 and 9%. Error increases to 15% as the number of spheres decreases to 50 per sample. These estimates were determined by using radioactive microspheres in which the number of spheres within each sample was determined indirectly (analogous to our determination of blood flow within lung and kidney samples). An advantage of the colored-microsphere technique is that the number of spheres within low blood flow samples are more accurately determined. On the basis of pilot studies, we estimate that <2% of the spheres contained within bone samples were lost during processing (quantified by trapping and refiltering fluids used during the initial filtering process). In the present study, intact-control bone samples contained 21 ± 15 microspheres, whereas the disuse bone samples contained 79 ± 63 microspheres. Precision for samples with microsphere counts up to 200 (determined by repeatedly counting the same sample) was >98%. Even tripling these error estimates would not appreciably alter the sixfold elevation in blood flow we observed at 7 days of disuse.

In this study, the area of intracortical porosis after 7 days of disuse was not different from control values, but intracortical porosity was significantly elevated by 14 days of disuse. Our study did not identify specific osteoclast or osteoblast activity, but instead it quantified the tissue level result of the imposed disuse. As a result, we are unable to definitively state that increased porosity at 14 days of disuse arose solely from elevated osteoclastic activity because it is possible that depressed osteoblastic activity contributed to the bone deficit. However, under normal conditions, minimal osteoclastic and osteoblastic activity are present in the adult ulna (31, 32). Indicative of a quiescent skeleton, control bones within all groups demonstrated minimal intracortical remodeling activity. Only after 14 days of disuse did we observe elevated porosity. These data correspond with preliminary data from the same model in which undecalcified thin-section histomorphometry was used to determine that osteoclastic activation was evident by 10 days, but not by 7 days, of disuse (5). Intracortically, depressed osteoblastic activity would have been evidenced by increased size of porosities with no change in the number of porosities (i.e., an acute uncoupling of resorption and formation at previously active intracortical remodeling sites). Instead, we observed an increased number of porosities, suggesting that disuse precipitated the differentiation and recruitment of osteoclasts to intracortical sites.

The mechanism by which the observed hyperemia was achieved will require further investigation. However, it is reasonable to assume that, like muscle (12), altered bone blood flow was achieved by an endothelial cell-mediated process. Endothelial cells are sensitive modulators of local flow states and accomplish this process by expressing vasoactive substances (11). Interestingly, many of these substances (e.g., endothelins and nitric oxide) are known to initiate osteoclastic differentiation or act as paracrine regulators of bone cell activity (3, 30). As a result, this study can be viewed to provide indirect support for the concept that endothelial cells may influence bone cell activity (6, 45).

It is also intriguing to consider the stimulus for disuse-induced hyperemia. Blood flow is commonly elevated when the metabolic needs of the tissue are heightened (20). Extrapolating to bone, disuse would presumably decrease the metabolic demand of the cells within the tissue and would therefore result in decreased bone blood flow. Alternatively, mechanical loading may serve a vital role in ensuring that cells within bone (particularly osteocytes) receive sufficient metabolic exchange (29). It is apparent that loading of bone serves to enhance diffusion through the tissue (21). Loss of loading may therefore deprive osteocytes of the metabolic supplies to which they have become accustomed. We hypothesize that a physiological response to this deprivation (e.g., hypoxia) may be responsible for the vasoregulation observed in this study.

In conclusion, we have examined cortical bone blood flow alterations in response to disuse. Hyperemia was locally mediated, with no contralateral or sham-surgery effects observed. Bone blood flow was highly elevated by 7 days of disuse and remained elevated, although to a lesser extent, at 14 days of disuse. Pronounced intracortical porosity was evident by 14 days, but not at 7 days, of disuse. We conclude that bone hyperemia precedes the onset of disuse-induced intracortical resorption and either precedes or is coincident with cellular events responsible for osteoclastic activation.