Microdamage and Bone Remodelling (or bone remodelling responses to mechanical stimuli)
In addition to its role in calcium homeostasis, the bone remodelling process is essential for maintaining the structural integrity of the skeleton. Centered around the Basic Multicellular Unit (BMU), remodelling is tightly regulated by a both paracrine and endocrine factors. Central to this regulation is the RANK-RANKL-OPG system, which acts locally to enhance or inhibit the actions of osteoclasts, the bone-resorbing cells. Bone remodelling has been shown to occur in response to mechanical stimuli and microdamage. Microcracks, a commonly observed form of microdamage, occur in response to fatigue loading and are repaired via remodelling. Osteocyte apoptosis
The skeleton is a dynamic living system, which is continuously modified, or remodelled, throughout adult life. Remodelling, the resorption of bone by osteoclasts and subsequent reformation by osteoblasts, serves three main functions. Firstly it is the primary mechanism for the release of calcium, which plays an essential role in cellular signaling, into the serum, and for storage of excess calcium in the form of hydroxylapatite crystals. Secondly, bone can be remodelled in response to mechanical signals to give greater mass (and therefore strength) in regions of bone that are subjected to higher levels of stress. Finally, it is through remodelling that damage to the bone, such as fractures or, more commonly, microcracks are repaired. Remodelling for the sake of maintaining mineral homeostasis is systemically controlled and need not involve targeted remodelling. (Indeed, it has been proposed that calcium ion release from the bone may not occur principally through remodelling, but via parathyroid hormone (PTH) mediated release across inactive bone surfaces.)
Bone is formed during embryonic development and rapidly grows, with formation exceeding resorption, throughout childhood and into early adulthood. Following this, the skeleton enters a prolonged period where bone mass remains relatively stable. During this period, lasting from about age twenty to age fifty or sixty, remodelling of both cortical and trabecular bone occurs continuously resulting in a turnover of about 10% of the skeleton each year with no net effect on bone mass. The maintenance of skeletal mass is regulated through a balance between the cells that resorb bone and those that form bone. This balance often decreases with age, and for women, commonly ends at menopause. Clinical disorders in which bone resorption exceeds formation are common and include osteoporosis, Paget's disease of bone and bone wasting secondary to certain cancers. Osteoporosis is the most common of these, and drugs for osteoporosis generally function by
Bone is an anisotropic material composed of osseous tissue, which predominantly presents an inorganic component comprising of calcium phosphate, or in the chemical form, hydroxyapatite, which provides mechanical strength and contributes to 75% of bone weight. The organic constituent includes Collagen Type I, which contributes up to 90% of the organic material, along with other non-collagenous proteins, or ground substance, which include osteoclacin, glycoproteins and proteoglycans, which provide tenacity and durability. Collagen provides a framework for the deposition of hydroxyapatite during mineralisation, and forms the template for the formation of lamellae (Reid, 1986). Bones permit movement of an organism, by facilitating muscle attachments, allow for the application of load to a body, provide protection for internal organs, calcium homeostasis, and for the production of blood cells. Bone homeostasis is controlled by the coordination of various cell types: osteocytes, osteoclasts, which remove bone tissue, and osteoblasts, which replace bone tissue. These cells allow for maintenance of calcium and phosphorous levels in bone, thus maintaining the structure and function of bone. Following resportion of bone by osteoblasts, numerous products are released which stimulate the differentiation and proliferation of osteoblasts. These will be discussed in more detail further on in section 1.1.4.
In human bone, two types of tissue are observed- cortical and trabecular, which form the extracellular matrix. Cortical bone, which is highly calcified, is present on the outer layers of the diaphysis of bones, and provides mechanical strength and constitutes 80% of bone weight. Cortical bone is characterised physically by a series of lamellae surrounding blood vessels. Trabecular or cancellous bone is spongy and durable, attributed by a parallel lamellar system, more metabolically active and forms the majority of bone surface tissue and the ends of long bones (Adler, et al, 2000, Ashman et al, 1988). Periosteum, the external layer of bone, is comprised of a fibrous and vascular layer. This surrounds an osteogenic layer, the endosteum, which is responsible for new bone formation. Osteocytes inhabit the lacunae, which are cavities found mainly in osteonal tissue. Compact bone comprises of tightly packed osteons and consists of various canals, known as Haversian canals. Running between the lacunae and Haversian canal are canaliculi. These organelles are often a product of the removal of bone and function in allowing for fluid flow and transport of nutrients (Buckwalter et al, 1987, Hilka, 2003).
Bone is formed from a number of different tissues, and therefore every individual bone of the skeletal system can be considered an organ. The main tissue type in bone is osseous tissue, which if a form of connective tissue. As such it consists of cells and fibres dispersed within a matrix. The extracellular matrix of osseous tissue has an organic phase and an inorganic phase. The organic phase, composed of water and collagen fibers (mainly type 1 collagen), which form a meshwork to provide the bone with elasticity and tensile strength. The inorganic phase of bone matrix, which is mainly composed of hydroxyapatite crystals, provides bone with stiffness and resistance to compressive stresses.
The Basic Multicellular Unit in Remodelling
The Basic Multicellular Unit (BMU) is a temporary grouping of osteoblasts and osteoclasts, which travels through cortical bone and along the surface of trabecular bone, dissolving old bone matrix and laying down new bone. Osteoclasts are large multinucleated cells derived from haematopoietic stem cells. They act by forming a sealed vacuole over the bone surface into which they secrete hydrogen ions and enzymes such as tartate-resistant acid phosphatase (TRAP) and caspase K. Osteoblasts are mononucleated cells derived from mesenchimel
The RANK-RANKL-OPG system
Osteoclast differentiation and activation is controlled by the RANK-RANKL-OPG signaling pathway, a signaling axis involving members of the Tumour Necrosis Factor (TNF) receptor family. Receptor activator of nuclear factor ?- (RANK) is a receptor expressed on the surface of osteoclasts and osteoclastic precursor cells. When activated it causes either differenciation of preosteoblasts into osteoblasts, or causes the activation of mature osteoclasts (Burgess et al 1999)
Mechanoresponsiveness of Bone Tissue
The observation that secondary osteons are aligned in the dominant loading direction of a bone, thus providing the bone architecture with additional strength in this direction, led to the hypothesis that the BMU is "steered" in the direction of the prevailing mechanical strain (Frost, 1987). The mechanism for this has been largely elucidated through the work of Klein-Nulend and colleagues (for review, see Klein-Nulend et al 2005). Osteocytes, which with their ramified morphology form a network throughout the bone matrix, are well suited to a mechanosensing role. Indeed, they have been shown to be the most sensitive of the bone cell types to mechanical stress in vitro, producing nitric oxide (NO) and prostaglandin E2 (PGE2) in response to pulsating fluid flow (PFF) (Klein-Nulend et al 1995a) and, to a lesser extent, to intermittent hydrostatic compression (Klein-Nulend et al 1995b). This suggests that the mechanism for the transduction of mechanical signals to biochemical signals in bone by osteocytes is quite similar to that of the mechanosensitive endothelial cells of the vasculature, which also respond to fluid shear stress through the release of NO and prostaglandins (see Klein-Nulend et al 2005). Indeed, the release of NO in both endothelial cells and osteocytes is mediated by the enzyme endothelial NO synthase (ecNOS) (see Klein-Nulend et al 2005). In osteocytes, inhibition of NO release in response to PFF also inhibited the release of PGE2, implying that NO is the principle biochemical mediator of mechanical effects on bone (Klein-Nulend et al 1995a).
Due to the mineralization of the extracellular matrix, bone tissue is stiff and bone loading produces very little mechanical deformation of cells. However, the porosity of bone tissue, created by the lacuno-canalicular network, means that slight strains, such as those experiences in everyday loading cycles, will create specific patterns of fluid flow within the bone matrix. This effect is shown in Figure 2 (from Klein-Nulend et al 2005) below. Burger et al (2003) described these canalicular flow patterns and how they affected the directional activity of the tunneling BMU via their effects on mechanosensitive osteocytes. Figure 3, taken from this paper, shows the change in canicular flow patters within the cutting cone of a remodeling BMU at maximum load, i.e.: during the heel strike of the walking cycle. As can be seen from this graph, at the tip of the cutting cone there is a reversal of flow at maximum loading, so that the net fluid flow at this point is zero. Thus, at maximal loading the osteocytes of the tip of cutting cone will not experience fluid shear stress and will not produce NO. Burger et al (2003) proposes that insufficient NO production by the osteocytes at this point of the cutting cone due to a lack of fluid shear stress will cause osteocytes apoptosis, in a similar manner to which a lack of basal NO production by blood vessel endothelial cells in areas of reduced blood flow results in cell death. The same paper also suggests that, as apoptotic endothelial cells release factors that attract phagocytic macrophages, apoptotic osteocytes attract phagocytic osteoclasts. In fact, the stimulation of local bone resorption by osteoclasts in response to osteocytes apoptosis was proven in a study done in 2005 by Gu et al, which will be described further in the following section of this review.
The model described above explains the mechanism for the initiation of osteoclastic activity at the tip of the cutting cone and the progression of the BMU in the direction of maximal strain. However, the question remains as to why the osteoclasts of the BMU do not continue to resorb the bone outward in all directions from this point. The fluid flow patterns through the cutting cone described by Burger et al (2003) can resolve this also. As shown in Figure 3, at the base of the cutting cone, which is also known as the reversal zone, there is unidirectional fluid flow during bone loading. Thus, osteocytes in this portion of the cutting cone will experience fluid shear stress, with those cells closest to the surface of the bone and furthest from the tip of the cutting cone experiencing the highest flow rate and producing high levels of NO and PGE2. The NO produced will protect the osteocytes against apoptosis, but in vitro studies have shown that it may also have the effect of promoting the withdrawal of osteoclasts from the bone surface. In 1991, MacIntyre et al showed that when applied to osteoclasts in vitro, NO, at a concentration of 30M, reduced the spread area of the osteoclasts and also reduced bone resorption, through osteoclast detachment, without causing osteoclast death. Based on this, Burger et al (2003) propose an overall model for the mechanism of the alignment of secondary osteons in the direction of mechanical loading, where well-stressed osteocytes at the base of the cutting cone release NO in concentrations sufficient to cause osteoclast detachment, while understressed osteocytes at the tip of the cutting cone undergo apoptosis and attract osteoclasts in the direction of loading. This model is illustrated in Figure 4 (Burger et al, 2003) below.
There are certainly other factors involved in the mechanotransduction of loading signals in bone. These factors are not included in the model suggested by Burger et al (2003) For example PGE2 produced by osteocytes in response to high levels of NO will have an anabolic affect through activation of nearby osteoblasts (Jee et al 1990). as the reduction of RANKL production and increase in
Microdamage in bone High impact, intensive exercise has been shown to engender microdamage in bones. Significant levels of microdamage can be found in the ribs of rowers (Warden et al. 2002), the long bones of race horses(Norrdin et al. 1998), the leg experiencing the highest loads in greyhounds running on oval tracks in a common direction (Muir et al. 1999; Tomlin et al. 2000) and associated with the ''march fractures'' of army recruits undertaking intensive training (McBryde 1975). While the existence of microdamage in a bone appears to decrease its stiffness, the overall strength or force to failure can be reduced dramatically (Carter & Hayes 1977). This phenomenon clearly has direct clinical relevance. Importantly, several studies have shown strong relationships between increased age of an individual and exponential increase in crack density(Schaffler et al. 1995). Burr and colleagues found that osteoclastic activity was over represented at sites of microdamage (Burr et al. 1985; Mori et al. 1993), leading to the suggestion that removal of microdamage in bone is not achieved through a ''random'' stochastic remodelling process alone but that damage is actively targeted for removal by the osteoclast population (Burr & Martin 1993; Parfitt 2001). Recent experiments have demonstrated the timedependent migration of osteoclasts into regions of cortical bone containing experimentally induced microdamage in rats. These findings have strengthened the argument for the targeting of microdamage in bone (Verborgt et al. 2000; Noble et al. 2003). Since high levels of microdamage can increase fracture risk these data point to the importance of the bone effector cell targeting system in the maintenance of bone health. The mechanism by which microdamage is targeted in bone is unknown as is the reason for its reduced efficiency under conditions of ageing. While targeting of damage appears to be sufficient for maintenance of strength in healthy bone, osteoclast targeting of damage does not appear to work efficiently in older individuals (Frost 1960a; Wong et al. 1985; Schaffler et al. 1995; Mori et al. 1997; Norman & Wang 1997). More specifically damage is not efficiently removed from interstitial volumes which are representative of older regions of bone (Wong et al. 1985). Accumulation of microdamage is higher in females than males and there is a particularly marked increase in women over the age of 40 years (Schaffler et al. 1995). The reasons for the enhanced accumulation of microdamage under these circumstances is not entirely clear but is likely to be due to one of two broadly defined changes in the bone:
1. Change in the material properties of the bone leading to accumulation of damage that is beyond the scope of the repair system.
2. A problem associated with the cells involved in the sensing, signalling and/or repair of the bone.
In fact it is likely that both of these changes contribute to the accumulation of damage but overall, that any alteration in the material properties of the bone will not occur entirely independently of changes in the cell population that create, remodel and maintain it. Cells might lose their ability to produce the stimulus, their sensitivity to the targeting stimulus or their ability to respond to it, or there may be changes in detailed cell function such as synthesised collagen composition, resorptive enzyme activities or their numbers in bones.
The involvement of osteoclast activity in the removal of microdamage is demonstrated by its reduced repair upon treatment with high doses of bisphoshonates (Hirano et al. 2000; Mashiba et al. 2000, 2001; Li et al. 2001) due to their inhibitory effects on osteoclast function. In general, though, the reduced removal of microdamage can occur under conditions of increased or steady osteoclast activity such as during estrogen loss and age-related bone loss, perhaps pointing to the importance of target osteoclast activity in the removal process rather than general activity.
Microcracks accumulate naturally as a result of cyclic loading, day to day stress and strain and repetitive insult to the area of under stress. It has also been found that microdamage collects at points of diminished osteocyte integrity (Qiu et al, 1997, Vashishth, et al, 2000). The production of microcracks allows bone to take in energy without breaking (Schaffler et al, 1994). Microdamage manifests itself as a series of cracks ranging up to 400 m, depending on the nature or the crack and exceeding these values, the microcrack becomes a potential stress fracture. Two patterns of microcracks, linear and cross-hatched, have been observed. Linear microcracks were observed in both the central portion and near surfaces of trabeculae. (Wenzel et al, 1994). Larger cracks result from compression, and diffuse cracks tend to be considerably smaller, from 2-10 m, and run parallel to the loading direction (Martin and Burr, 1982). A build up of microcracks in bone is a factor in reduced strength and resistance to fracture in bone (Schaffler, et al, 1995). Force and loading on a bone has been found to elicit DNA damage in osteocytes proportional to the force applied (Noble et al, 1997).
Cracks can be observed on the interstitial bone between osteons, on channels and lacunae, and are often stopped in their tracks by collision with osteonal cement line (Boyde et al, 2003, Martin, 2003). Furthermore, it was hypothesised by Martin and Burr (1982) that Haversian canal adjacent to the crack will produce a new osteon, and essentially traps the microcracks within the lamellae of the osteon (Martin and Burr et al, 1982). Interaction of microdamage with cells can increase mechanical strain, and in doing so, triggers bone remodelling. Mechanisms of controlling such an action were considered by Parfitt, 1984, and included the incidents of fatigue damage, changes in cell membranes, hydrostatic pressure on cell membranes thanks to alterations in fluid flow, and stress potentials.
Human bone can tolerate strain levels of up to 2000 e in quantities from 4-10 million cycles prior to failure, and the greater the strain encountered by the bone, the more frequent the microdamage and the greater the loss of stiffness is experienced (Pattin et al, 1996, Schaffler et al, 1989). If strain exceeds 7000 e, it will be irreparably damaged, and bone will die (Burr et al, 1996). It has been theorized that microdamage to bone is one of the principal mechanism involved in the initiation of apoptosis, which is, however, confined to areas surrounding the microcrack. This function coincides with bone resorption by the BMU and the instigation of the bone remodelling process (Verborgt et al, 2000). According to Parfitt et al, 1996, this is referred to as targeted bone remodelling, and occurs as a result of the previously described cytokine cascade. The events occurring at the site of the cracks have been of great interest and the source of considerable hypothesising as to whether cells are necrotic or apoptotic. To consider this question, it was demonstrated in a study by Noble et al, 1998, which described an increase in osteocyte apoptosis following deformation of bone, preceding high rates of bone resorption. This also provided a platform to consider microcracks as a modulator of bone remodelling. Klein-Nulend suggested that levels of nitric oxide produced following microdamage are sufficient to induce apoptosis (Klein-Nulend et al, 2005). Turner et al, 2002, proposed that cells directly at the site of injury signal through a series of gap junctions within osteocytic processes, forming a syncytium, allowing cell-cell signalling.
Microdamage is a naturally occurring phenomenon in bone which is caused by fatigue/loading cycles that occur in normal activities such walking, running, or lifting heavy weights. It can be seen as an evolutionary compromise: as a skeleton large enough to support a heavy animals weight without failure would be metabolically expensive, it is more advantageous to have a light skeleton, which regularly undergoes failure but which can be repaired. Because they are surrounded by a hard mineral matrix, it is difficult to keep osteocytes alive within osseous tissue in vitro. As such, most studies involving microdamage up to this point have either used specimens of dead bone for the observation of crack initiation and propagation (Schaffler et al, 1989; Burr et al, 1998; O'Brien et al, 2003), or have examined microcracks formed in vivo in human and animal bones (Schaffler et al, 1995; Bur et al, 1985; Lee et al, 2002). Microdamage is generally observed as microcracks, of length ranging up to 500m (Wenzel et al, 1996). When microcracks extend beyond this length, they are termed microfractures and can lead to bone failure, i.e.: a stress fracture. A study done in by Schaffler, et al in 1995, using the femoral diaphesis from 28 individuals, found that 80-90% of microcracks in cortical bone occur in the interstitial spaces between osteons. This implies that the bone microstructure may have a role in preventing the growth of microcracks, if their propagation is stopped or slowed once they meet the cement lines of secondary osteons. showed that microdamage density increases exponentially with age. There are two general patterns of microcracks in bone; either linear or cross-hatched (Wenzel et al, 1996).
An in vivo study showing a positive correlation between loading, incidence of microdamage and remodeling spaces was done in 2002 by Lee et al. Three groups of adult sheep were used for the study; the first group had undergone an ulnar osteomy to increase loading on the proximal radius (Group O, n=1), a second group underwent Steinmann pinning of the ulna in an attempt to decrease loading on the radius (Group P, n=12), and a third group, which underwent a sham operation (Group C, n=11).
1. Microdamage in all models
2. Correlation between location/timing of cracks and resorption cavities
3. Increased microdamage & repair in group O
Discussion: Shows microdamage as naturally occuring and suggests that microcracks are stimulus for bone remodelling
Limitations: Loss of 4 from group O, no significance for formation, Group P no effect
Microdamage in bone represents a significant contributor to fractures in the elderly. Thus it is worrying that many common pharmacological therapies for osteoporosis may increase their occurrence through the repression of bone remodeling. A therapy which could target systemic remodeling while allowing remodeling to repair microdamage would be an ideal solution. The biochemical events surrounding bone remodeling following microdamage should therefore be fully explored.