The genetic influence on bone loss in response to mechanical unloading was investigated within diaphyseal and distal femoral regions in three genetically distinct strains of mice. One mouse strain failed to lose bone after removal of function, whereas osteopenia was evident in multiple regions of the remaining two strains but in different areas of the bone.
Introduction: It is well recognized that susceptibility to osteoporosis is, in large measure, determined by the genome, but whether this influence is systemic or site-specific is not yet known. Here, the extent to which genetic variations influence regional bone loss caused by disuse was studied in the femora of adult female mice from three inbred strains.
Materials and Methods: Adult C57BL/6J (B6), C3H/HeJ (C3H), and BALB/cByJ (BALB) mice were subjected to 15–21 days of disuse, achieved by hindlimb suspension, and six distinct anatomical regions of the femur were analyzed by high-resolution μCT.
Results and Conclusions: In B6 mice, the amount of disuse stimulated bone loss was relatively uniform across all regions, with 20% loss of trabecular bone and 10% loss of cortical bone. The degree of bone loss in BALB mice varied greatly, ranging from 59% in the metaphysis to 3% in the proximal diaphysis. In this strain, the nonuniformity of bone loss was directly related to the nonuniform distribution of baseline bone morphology (r2 = 0.94). In direct contrast with BALB and B6, disuse failed to produce significant losses of bone in any of the analyzed regions of the C3H mice. Instead, these animals displayed a unique compensatory mechanism to disuse, where the large loss of calcified tissue from the endocortical surface (−24%) was compensated for by an expansion of the periosteal envelope (10%). These data indicate a strong, yet complex, genetic dependence of the site-specific regulation of bone remodeling in response to a powerful catabolic signal. Consequently, the skeletal region of interest and the genetic make-up of the individual may have to be considered interdependently when considering the pathogenesis of osteoporosis or the efficacy of an intervention to prevent or recover bone loss.
THE SENSITIVITY OF the skeleton to changes in its mechanical environment is reflected by the rapid and site-specific loss of bone tissue after removal of function. Consequently, conditions such as bedrest, spinal cord injury, or space flight can be severely detrimental to the mass, architecture, and mechanical strength of the bone tissue, ultimately transforming specific sites of the skeleton into an osteoporotic state.(1)
Interestingly, the variability of bone loss between individuals when placed under conditions of disuse is highly variable, ranging from those who lose significant amounts of bone to those who remain largely unaffected. For instance, after 6 months of space flight, the decrease in trabecular BMD from the proximal tibia of astronauts ranges from 0% to 23%.(2, 3) While several environmental factors (e.g., differences in diet and activity levels during or even before space flight) inevitably contribute to this variability, it is unlikely that such environmental factors are the predominant cause for the disparity between subjects because of the similarity of these imposed factors in this specific population. Alternatively, genetic variations between individuals that give rise to large differences in skeletal mass and morphology(4, 5) may also account for the differences in the amount of bone loss. In fact, evidence from distinct strains of inbred mice suggests that genetic variations even influence the manner and magnitude of bone's sensitivity to altered demand—both for trabecular(6) and cortical bone.(7-10) Furthermore, the influence of genetics on rates of bone loss is not limited to disuse osteoporosis but extends to age-related and postmenopausal osteoporosis.(11)
Whereas these human and mouse studies indicate a strong genetic influence on bone's responsiveness to changes in its local and systemic environment, little is known about the extent by which genetic variations also influence the nature (cortical versus trabecular) and site-specificity (epiphyseal versus metaphyseal) of bone loss. If such as high degree of site-specificity exists, it raises the possibility that measurements are highly unique to that site,(12) thus confounding any overriding diagnostic interpretation for skeletal-wide remodeling.
Because of the strong site-specific effects of the genome on bone morphology observed in our accompanying paper, here we hypothesized that the response of bone to a catabolic stimulus would exhibit a similar interdependence on genetics and the specific anatomical location. Given that both genetically distinct human(13-15) and mouse(6) populations with high skeletal mass seem to be less susceptible to catabolic pressure, we further hypothesized that regions within a bone displaying a genetically defined large amount of bone will have a lower propensity to lose bone mineral.
MATERIALS AND METHODS
All procedures were reviewed and approved by the Institutional Animal Care and Use Committee. Three strains of mice, C57BL/6J (B6), BALB/cByJ (BALB), and C3H/HeJ (C3H), all female and 4 months of age, were used in this study.(6) Animals served either as age-matched controls (n = 8, n = 9, and n = 8 for B6, BALB, and C3H, respectively) or mechanical disuse animals (n = 8, n = 10, and n = 10, respectively). Disuse in mice was induced by hind-limb elevation through tail suspension.(6, 16) All mice were individually housed in standard cages (28 × 17 × 13 cm) and had access to standard rodent chow and water ad libitum. Control and experimental mice were killed after 15 (B6 and C3H) and 21 days (BALB). On death, right femurs were harvested and preserved in 70% EtOH.
Seven trabecular and cortical regions within the diaphyseal and distal regions of the femur, representing both cortical and trabecular bone, were analyzed by high-resolution μCT,(17) as described in the accompanying paper. Cortical bone was analyzed from the metaphysis and from three 30-μm-long regions of the diaphysis: the midshaft (defined at 50% of the total femoral length) and two diaphyseal regions, one proximal of the mid-diaphysis at 40% of the femoral length and one distal at 60%. Trabecular volumes of interest included the epiphysis and metaphysis. A local thresholding procedure was used that is described in detail in our accompanying paper. The same threshold was used for both disuse and control bones, thus removing any bias in determining the relative amount of bone loss induced by mechanical unloading.
Cortical bone area (Ct.Ar) as well as areas of the endocortical (Ec.En) and periosteal envelopes (Ps.En) were calculated as averages along the length of each volume of interest. For all trabecular regions, bone volume fraction (BV/TV), trabecular separation (Tb.Sp), trabecular thickness (Tb.Th), trabecular number (Tb.N), connectivity density (Conn.D), the geometrical degree of anisotropy (DA), and the structural model index (SMI) were determined.(18-20)
Paired t-tests within groups established whether disuse resulted in significant weight loss. Two-tailed t-tests were used to determine whether the loss of weight bearing led to differences in bone morphometric indices between control and disuse mice in each of the three mouse strains. Two-way ANOVAs with “mouse strain” as one factor and the “level of weight bearing” as the second tested whether disuse produced differential effects across the strains (as indicated by statistically significant interactions between the two factors). Changes in bone morphological indices were presented as the difference in means between control and disuse mice and the 95% CI of the difference in means. For a given genetic strain, linear correlations were used to test for associations between the distribution of bone morphology (baseline bone quantity) and the distribution of bone loss across the evaluated sites within the femur. The significance level was set at 5%.
B6, C3H, and BALB mice subjected to disuse lost 5.8% (p > 0.05), 6.4% (p < 0.001), and 3.9% (p < 0.02) of their body mass over the course of the experimental protocol, respectively. The following comparisons between disuse and control mice are reported in the order of B6, C3H, and BALB mice, which have been previously labeled low, high, and medium BMD mouse strains, respectively.
Metaphyseal trabecular bone
Disuse failed to significantly alter any indices of metaphyseal trabecular bone quantity and architecture in B6 mice due to large interanimal variations (Table 1). Whereas trabecular bone volume (BV/TV) of the metaphyseal region in C3H mice was unaffected by disuse compared with controls, trabecular thickness was reduced by 13% (p < 0.02), and connectivity density was increased by 23% (p < 0.05). In contrast, in the metaphysis of BALB mice, disuse reduced BV/TV by 59% (p < 0.001), trabecular thickness by 26% (p < 0.001), trabecular number by 19% (p < 0.002), and connectivity density by 40% (p < 0.02; Fig. 1). Also, a change in the principal orientation of the trabeculae was indicated by a decrease of 7% (p < 0.02) in the degree of anisotropy (Table 1).
Table Table 1.. Mean Differences (±95% CIs of the Difference) in Trabecular Parameters of Bone Quantity and Microarchitecture Between Control and Disuse Animals of the Three Strains of Mice
Epiphyseal trabecular bone
The epiphyseal region of B6 mice was highly responsive to mechanical unloading, marked by a 21% decrease in BV/TV (p < 0.005), a 13% decrease in trabecular thickness (p < 0.006), and a 37% increase in SMI (p < 0.006; Fig. 2). In C3H mice, however, tail suspension failed to affect the quantity and architecture of trabecular bone. In the epiphyseal region of BALB mice subjected to disuse, the amount of bone loss (BV/TV) relative to controls was approximately one-half of the reduction in bone volume in the metaphysis (−27%, p < 0.0001). Trabecular thickness (−17%, p < 0.004), connectivity density (59%, p < 0.0001), and SMI (183%, p < 0.0001) were also significantly altered (Table 1).
Metaphyseal cortical bone
Disuse caused significant erosion of the metaphyseal endocortical surface in all three strains of mice; the difference in the endocortical envelope between control and disuse mice was 9% in B6 mice (p < 0.02), 24% in C3H mice (p < 0.002), and 15% in BALB mice (p < 0.02). However, cortical bone area was lost only in B6 (−8%, p < 0.008) and BALB mice (−18%, p < 0.001); no change was observed in C3H mice because the expansion of the endocortical envelope in disuse mice was compensated for by an increase in the periosteal envelope (10%, p < 0.02; Fig. 3).
Diaphyseal cortical bone
The three diaphyseal regions (distal, mid-, proximal) were compared between control and disuse mice in each of the three genetically distinct strains. In B6 mice, control mice had significantly greater diaphyseal bone area than their disuse counterparts at all three regions. This difference was 11% (p < 0.001) averaged across the three regions and consistent throughout the regions. The decrement in bone volume was accompanied by expansion of the endocortical envelope and contraction of the periosteal envelope, although only the decrease in the periosteal envelope of the distal diaphysis was statistically significant (−4%, p < 0.05). In C3H mice, there were no differences in cortical bone area or periosteal and endocortical envelopes in any of the three diaphyseal regions analyzed except for the endocortical envelope at the mid-diaphysis, which was 18% greater (p < 0.04) in disuse mice. In BALB mice, disuse significantly decreased the amount of cortical bone in the mid-diaphysis (−11%, p < 0.03) as well as the distal diaphysis (−11%, p < 0.02), but not in the proximal diaphysis (Table 2)Table 3..
Table Table 2.. Relative Changes in Cortical Bone Area (Ct.Ar), the Area Enclosing the Endocortical Envelope (Ec.En), and the Area Enclosing the Periosteal Envelope (Ps.En) Induced by Disuse
Table Table 3.. Synopsis of Disuse Related Changes in Bone Quantity in Different Trabecular and Cortical Regions of the three Strains of MICE
Interactions between genetic variations and change in morphology
Two-way ANOVA tested more directly whether genetic variations between the three strains influenced altered bone morphology associated with the removal of the mechanical environment. Significant interactions between the two factors were found for most indices, indicating that the site-specific magnitude of bone loss was indeed related to genetics (Table 1; Table 2).
Correlation between bone morphology and bone loss
For correlations between the magnitude of bone loss in the distinct femoral regions and “baseline” bone quantity, bone quantity was calculated as the relative difference in bone volume for control mice of a given strain to the average bone volume of the three strains at the specific site. In BALB mice, the strain that displayed the greatest femoral bone quantity, 94% (p < 0.004) of the variation in bone loss was accounted for by variations in existing bone quantity. No significant correlations between the difference in genetically determined bone quantity and bone loss were observed in B6 or C3H mice.
The influence of the genome on the manner and magnitude of the skeletal response to alterations in its mechanical environment was investigated by comparing the bone loss stimulated by disuse that occurred in three distinct strains of mice. Both B6 (low femoral bone quantity) and BALB mice (high femoral bone quantity) responded to the loss of weight bearing with significant reductions in bone quantity and architecture in diaphyseal as well epi- and metaphyseal regions of the distal femur, yet the degree of trabecular and cortical bone loss greatly depended on the specific region. In BALB mice, trabecular bone loss was twice as large in the metaphysis than in the epiphysis, similar to the 2-fold difference in cortical bone loss measured between the metaphysis and diaphysis. Tissue losses in B6 mice were more evenly distributed within the trabecular and cortical regions. The absence of altered tissue volume in high BMD C3H mice did not result from the absence of a response at the endocortical and periosteal envelopes but from a concomitant expansion of the two cortical envelopes. Whereas the large amount of bone loss in BALB mice led to the rejection of the hypothesis that high BMD skeletons lose less bone, an association between the amount of bone present at a specific site and the magnitude of the loss of bone tissue in this strain indicated the selective erosion of bone tissue from sites enjoying genetically determined greater bone morphology (relative to the other strains). Taken together, these data indicate that bone's plasticity to altered mechanical loading is strongly nonuniform within a bone and not only dependent on an individual's genotype but also on the specific site.
This study built on our accompanying paper that indicated a highly site-specific influence of genetics on bone morphology, thus stimulating the hypothesis that mechanosensitivity is also interdependently modulated by genetics and anatomical site. Prudence requires, however, to extrapolate any data collected from mice to human conditions with care because of obvious (e.g., size, metabolism) or more subtle differences (e.g., different type of bone, physeal closure occurs very late or never).(21) Despite mice being the model of choice in biomedical research in general and in skeletal genetics in specific, the effect of these differences on research data are still unknown, although most studies have observed similar effects in mice and humans both for the skeletal response to mechanical stimuli(22) as well as for gene-bone morphology interactions.(23) Furthermore, different experimental durations limited direct comparisons of absolute values between BALB mice and the other two strains that were subjected to 15 days rather than 21 days, yet these differential protocol lengths had likely a minimal effect on the primary aim of this study, which focused on relative spatial differences within a bone across strains.
Unlike the trabecular bone measured in the metaphyseal region, in which BALB mice showed a hypersensitivity to mechanical unloading, the diaphyseal region revealed that low bone quantity B6 mice were very responsive, showing a similar amount of bone loss to that observed in BALB's (despite being on a shorter protocol). In contrast, disuse failed to reduce bone volume at any cortical site of C3H animals, yet they lost three times as much bone from the endocortical surface of the metaphysis than B6 mice. A net loss of bone was avoided in these animals by stimulating a periosteal expansion, similarly to the expansion of the periosteal envelope during aging.(24) While the modulation of this compensatory mechanism is unclear at this point, preliminary μCT data from C3H mice of the same cohort (n = 4) indicate that the periosteal envelope does not change between 16 and 18 weeks of age. Thus, the genetically specific compensatory mechanism in C3H mice was most likely modulated by an expansion of the periosteal envelope rather than by inhibiting its contraction.
As described in the companion paper, bone quantity and microarchitecture at the six sites within the femur were also dependent on genotype and specific anatomical location, suggesting a possible relationship between baseline morphology and mechanosensitivity. In clinical studies of osteoporosis and aging, no consistent association between individuals with high or low BMD and high or low rates of bone loss has been established,(25-28) perhaps influenced by the timing of the baseline bone density measurements that did not coincide with the event of peak bone mass. Our data are consistent with the clinical studies indicating that, for a given skeleton, the propensity to resorb bone is not correlated with baseline bone mass (i.e., both low bone quantity B6 and high bone quantity BALB mice lose large amounts of bone). They suggest, however, that the manner by which bone is resorbed may be different between high and low bone quantity skeletons; thus, skeletons with large amounts of bone may enjoy a strategic advantage by removing tissue from sites where the genetic variations had placed a larger than normal amount of tissue. Extrapolated to humans, the data also suggest that the genes regulating bone morphology at a given site may overlap with the genes regulating mechanosensitivity only in some individuals (e.g., large bone mass, fast loser).
The wide spatial variations in bone's plasticity to disuse indicates that the genome enables the regulation of mechanosensitivity on a site-specific basis, with distinct genes or combination of genes modulating the response within each region within a bone. A strategy that enlists a distinct set of genes for each skeletal location would, however, add a great deal of complexity. Rigorous segregation studies will eventually elucidate the number of genes involved in the genetic manifestation of such a site-specific regulatory mechanism, but it is likely that gene-environment interactions also contributed to the observed site-specificity. Inevitably, the genetically regulated differences in baseline morphology within a bone and between mouse strains also produced distinct mechanical milieus within the femur when subject to load (or its removal). Our understanding of how bone senses the loss of functional weight-bearing is incomplete, but the loss of bone in disuse is related to the suppression of key regulatory signals that maintain homeostasis.(29) The distribution of these mechanical signals across a bone, independent of their identity (e.g., strain magnitude, strain gradients, or high-frequency strains), is nonuniform and related not only to differences in bone morphology between different intraosseous sites, but also to the imposed mechanical environment (e.g., as influenced by potential differences in muscle-bone interactions or body mass distribution between the strains). The nearly 5-fold greater endocortical envelope of the proximal metaphysis compared with the endocortical envelope of the mid-diaphysis in C3H mice reflects such a non-homogeneous distribution of bone tissue in addition to other morphological parameters, such as bone curvature or moments of inertia,(30) which will all affect the loading pattern at a given site. This hypothesis of strong gene-environment interactions is currently being tested by quantifying the distribution of candidate mechanical parameters within a bone through in vivo strain gauging and finite element modeling.(31)
In summary, this study suggests that genetic modulation of bone loss is strongly site-specific, hence greatly limiting our ability to categorize entire skeletons into varying degrees of responsivity based on their genotype. Although in this study a mechanical stimulus was used to generate catabolic pressure, it is likely that a similar interdependence applies to biochemical challenges (e.g., hormonal) that will ultimately transform the skeleton into an osteoporotic state. Thus, the optimal site of detection for osteoporosis may depend on an individual's genotype, and the definition of those genes controlling this site-specific relation may ultimately allow the identification of those at greatest risk to incur osteoporosis. In addition, preclinical evaluations of interventions designed to augment bone quantity and/or quality (or inhibit its demise) may have to consider the responsivity of the tissue, as well as region-specific responses.
This research was supported, in part, by the National Space Biomedical Research Institute, the National Aeronautics and Space Administration, and the Alberta Heritage Foundation for Medical Research. Expert technical support from Sherry Wang was greatly appreciated.