Brief communication: The effects of disuse on the mechanical properties of bone: What unloading tells us about the adaptive nature of skeletal tissue

The importance of mechanical loading in bone metabolism is well established. The crucial role it plays in skeletal health is aptly illustrated by the severe osteopenia that accompanies prolonged periods of immobilization, commonly referred to as disuse osteoporosis. Examining the effects unloading has on the structural properties of bone provides a unique opportunity to better understand the role mechanical loading plays in the adaptation and maintenance of this dynamic tissue. This study reports the results of a comprehensive macro- and microscopic analysis of all major long bones belonging to two individuals that experienced long-term immobilization, and a broad range of comparable individuals demonstrating normal mobility levels.

Long-term immobilization, resulting in a profound reduction of mechanical loading, dramatically affects bone turnover rates and compromises the structural and functional integrity of bone (Uhthoff and Jaworski,1978; Jaworski et al.,1980; Stout,1982; Dehority et al.,1999; Giangregorio and Blimkie,2002). Prolonged skeletal disuse results in decreased bone mass and subsequently disuse osteoporosis. There are various circumstances of unloading that demonstrate both temporary and permanent osteoporotic conditions, including extended bed rest (Bloomfield,1997), hypogravity (i.e., space flight) (Globus and Morey-Holton,2009; Bloomfield,2010), myopathies, denervating pathologies (Wakley et al.,1992), and spinal cord sectioning (Eser et al.,2004; Qin et al.,2010).

Currently, a mechanism based on threshold levels established during skeletal growth and development (Carter,1984; Lanyon,1984; Frost,1987), commonly referred to as the mechanostat, provides one explanation for why unloading results in such dramatic bone loss. This mechanism envisions a negative feedback loop wherein signals due to disproportion in mechanical loading are detected and corrected for via a series of set points delineating a normal threshold level. Strain outside this defined threshold range initiates a modeling or remodeling response to adjust bone strength accordingly. Strain above a higher remodeling threshold activates modeling to increase periosteal apposition, and remodeling to replace fatigue-damaged bone; collectively strengthening and maintaining bone. Strain below a lower remodeling threshold activates remodeling to decrease bone mass and subsequently increase the strain level.

Disuse-driven remodeling is characterized by the coupled, coordinated action of osteoclasts and osteoblasts to remove and replace bone from either the endosteal and intracortical surfaces or both. The result of each remodeling event is the formation of a new bone structural unit, also known as a secondary osteon. Remodeling in any form results in a net bone loss as a vascular space, or Haversian canal, is an essential osteonal component for fluid and nutrient movement. Disuse-remodeling demonstrates an even greater net loss as osteoclasts become uncoupled from osteoblasts (Alexandre and Vico,2011), resulting in large resorption spaces as substantially more bone is removed than is replaced. Therefore, disuse-remodeling results in marrow expansion as the endosteum is slowly resorbed via coalescing porous spaces that effectively trabecularize the cortical matrix (Burr,2010; Zebaze et al.,2010), reducing the metabolic expense of unnecessary bone tissue.

In addition to strain level, increased osteocyte apoptosis resulting from diminished intramedullary fluid pressure (Turner,1999), may also be a factor elevating remodeling events following disuse. Additionally, recent research on the role of sclerostin protein, a known bone forming inhibitor produced by the SOST gene, has found increased SOST mRNA levels in rodents subjected to hindlimb disuse (Robling et al.,2008). Sclerostin is presumed to reduce proliferation while elevating the frequency of cellular apoptosis (Sutherland et al.,2004).

Previous studies largely based on animal models have confirmed the overall effect immobilization has on bone mineral density (see Gross et al.,2010 for a review). However, many of these studies lack significant periods of disuse or histomorphometric investigation, limiting understanding of remodeling activity associated with permanent, long-standing disuse. Additionally, few studies have assessed the degree of intraskeletal variability in remodeling rates between skeletal elements. The goal of this study is to examine multiple indicators of bone metabolism within a sample of individuals with distinct activity level differences, subjects that suffered from long-term quadriplegia and those that had full limb mobility. This provides a unique investigation into remodeling associated with skeletal unloading rather than loading, allowing us to contribute to both established and recently proposed hypotheses of remodeling behaviors reflective of various loading regimes.

MATERIALS AND METHODS

RESULTS AND DISCUSSION

Figure 1 presents mean plots of the four indicators of bone metabolism between all mobile individuals included in the study and each of the quadriplegic females. Compared to the individuals with normal mobility, the 54-year-old immobile female demonstrates lower OPD values outside the 95% confidence intervals for all skeletal elements (Fig. 1a). This is most likely a result of both reduced mechanical stimuli and cortical thinning resulting from endosteal resorption, as discussed below. The effects of reduced mechanical loading are further supported in the plot for On.Ar, which reveals that the long bones typically subjected to the greatest load (i.e., humeri, radii, femora, and tibiae) are larger in size within the immobile limbs compared to those with normal mobility (Fig. 1b). The fibulae and ulnae demonstrate osteon sizes well within the range of their mobile counterparts, though greater than the mean, reflecting their lesser role in bearing applied loads compared to their more prominent forelimb partners (i.e., the radius and tibia) (Wang et al.,1996; Birkbeck et al.,1997). In terms of cross-sectional geometry, both J (Fig. 1c) and % Ct.Ar (Fig. 1d) are generally lower than the 95% confidence interval of individuals with normal mobility, which most likely reflects increased endosteal resorption and/or reduced bone formation. However, as demonstrated in the plot for On.Ar, the fibula and ulna appear to be less affected as % Ct.Ar for these skeletal elements is clearly within the normal range.

When considering the plots for the 70-year-old quadriplegic female, most of the trends observed in the 54-year-old are similar. Some important contrasts are observable in respects to OPD (Fig. 1a) wherein several of the immobile long bones—particularly the fibulae, radii, and tibiae—demonstrate values well within the 95% confidence intervals of the normally mobile individuals. This rise in OPD compared to the 54-year old may reflect estrogen depletion characteristic of a post-menopausal state (Troen,2003; Kearns etal.,2008), wherein osteoclast proliferation is less inhibited, resulting in elevated bone turnover rates. Additionally, in terms of % Ct.Ar, the fibulae and ulnae demonstrate endosteal resorption well outside 95% confidence intervals compared to that observed in the younger 54-year-old quadriplegic (Fig. 1c).

With regards to mean distribution between the respective immobilized individuals and their age-matched groups of mobile individuals, trends are consistent with those noted above (Figs. 2 and 3). The primary difference is the increased 95% confidence intervals of the mobile individuals due to decreased sample size, which minimally dilutes differences observed when all mobile individuals are considered. For all four indicators of bone metabolism, the femora and humeri appear to be the most consistently affected skeletal elements with respect to long term disuse.

In addition to the potential histomorphometric differences between mobility groups, visual observations between OPD and On.Ar of all long bones appear to be negatively related. This presumed relationship compliments Frost's (1990) hypothesis, which postulates that as osteon density increases osteon size decreases. Additionally, the data are in agreement with previous histomorphometric investigations by Skedros et al. (2001) and van Oers et al. (2008), suggesting both experimentally and computationally that bone resorption is inhibited by strain-generated osteocyte signals. According to these studies, osteoclasts are inhibited in mechanically loaded bone, resulting in smaller osteon diameters in cortical regions subjected to high strain magnitudes. Presence of small, densely packed osteons in cortical regions experiencing strains above the normal remodeling threshold is presumably advantageous as the cement lines of secondary osteons are believed to assist in arresting the propagation of micro-cracks, decreasing the probability of catastrophic tissue failure (Burr et al.,1988; Currey,2002; Sobelman etal.,2004). By reducing the osteon diameter and increasing their concentrations, the efficiency in which transmitted energy is absorbed is also increased (Yeni et al.,1997).

CONCLUSIONS

Osteopenic changes in long bones that result from long-term immobilization exemplify the important role mechanical loading plays in bone metabolism. Although limited in sample size of individuals experiencing long-term disuse, this study expands our understanding of this relationship with a comprehensive analysis of both macro- and microscopic signatures of functional adaptation and maintenance of the human skeleton. Analyses of the major long bones of individuals experiencing drastic mobility differences have both confirmed previous hypotheses concerning skeletal responses to mechanical loading and shed new light on potential intraskeletal remodeling parameters associated with the strain environment.

Outcomes of this study suggest bones associated with long-term disuse have lower osteon population densities and larger osteon areas, reflecting dramatically lower remodeling rates and reduced strain levels. Regarding bone mass, immobilized skeletal elements demonstrate a reduction in the percentage of cortical area due to endosteal resorption, as predicted by the mechanostat. Differences in bone distribution and the percentage of cortical area present between the long bones of the immobilized individuals and their comparative sample of individuals with normal mobility are observed, particularly within the skeletal elements of the immobilized 70-year old. This observation suggests that bone modeling activity may be insignificant in the unloaded skeleton. A larger sample is required for quantitative statistical analysis, allowing for a more comprehensive study that accounts for a larger range of human variation.