SEARCH

SEARCH BY CITATION

Keywords:

  • functional adaptation;
  • microdamage;
  • osteocytes;
  • targeted remodeling

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Repetitive bone injury and development of stress fracture is a common problem in humans and animals. The Thoroughbred racehorse is a model in which adaptive failure and associated development of stress fracture is common. We performed a histologic study of the distal end of the third metacarpal bone in two groups of horses: young Thoroughbreds that were actively racing (n = 10) and a group of non-athletic horses (n = 8). The purpose of this study was to determine whether development of articular microcracks was associated with specific alterations to subchondral plate osteocytes. Morphometric measurements were made in five regions of the joint surface: lateral condyle, lateral condylar groove, sagittal ridge, medial condylar groove, and medial condyle. The following variables were quantified: hyaline cartilage width; calcified cartilage width; the number of tidemarks; microcrack density at the articular surface; blood vessel density entering articular cartilage; the presence of atypical bone matrix in the subchondral plate; bone volume fraction; and osteocyte density. Adaptation of articular cartilage was similar in both groups of horses. Vascularization of articular cartilage was increased in the group of non-athletic horses. Microcracks, which typically had an oblique orientation to the joint surface, were co-localized with blood vessels, and resorption spaces. Microcracking was increased in the condylar grooves of athletic horses compared with the other joint regions and was also increased compared with the condylar groove regions of non-athletic horses. Coalescence of microcracks also led to development of an intracortical articular condylar stress fracture in some joints and targeted remodeling of affected subchondral plate. The subchondral plate of the condyles in athletic horses was sclerotic, and contained atypically stained bone matrix with increased numbers of osteocytes with atypical morphology. However, osteocyte numbers were not significantly different between groups. We conclude that differences in site-specific microdamage accumulation and associated targeted remodeling between athletic and non-athletic horses are much greater than differences in subchondral osteocyte morphology. However, the presence of atypical subchondral bone matrix in athletic horses was associated with extensive osteocyte loss. Although osteocyte mechanotransduction is considered important for functional adaptation, in this model, adaptation is likely regulated by multiple mechanotransduction pathways.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Fractures develop because of monotonic overload, or because of repetitive injury and pathological weakening of bone tissue. In human runners, stress fractures are common in both track athletes and military recruits (Bennell et al. 1996; Milgrom et al. 2004). Stress fractures are also common in animals such as racehorses and racing greyhounds (Muir et al. 1999, 2006; Verheyen et al. 2006). The Thoroughbred racehorse is a useful model for studies of exercise-induced fracture development (Verheyen et al. 2006); an important feature of this model is that high-speed running induces particularly large cyclic strains to distal limb bones (Nunamaker et al. 1990).

Articular parasagittal fracture of the distal end of the third metacarpal/metatarsal bone (Mc-III/Mt-III), or condylar fracture, is a common exercise-induced injury in the racing Thoroughbred (Rick et al. 1983). Condylar fractures are usually complete, but incomplete condylar fractures also occur (Kawcak et al. 1995). Condylar fractures are site-specific and typically propagate from the articular surface of the distal end of Mc-III affected with microcracking of calcified cartilage and adapted subchondral bone (Riggs et al. 1999a; Radtke et al. 2003; Muir et al. 2006). However, it is unclear whether all condylar fractures are pathological and represent adaptive failure. These fractures arise from the palmar/plantar region of the condylar groove (Fig. 1) between the sagittal crest and the condyles of the distal joint surface (Riggs et al. 1999a; Radtke et al. 2003).

image

Figure 1. Photograph of an oblique frontal section through the palmar distal region of the distal end of a Mc-III bone, from a 3-year-old castrated male Thoroughbred. Regions-of-interest across the joint surface were: (1) lateral condyle; (2) lateral condylar groove; (3) sagittal ridge; (4) medial condylar groove; (5) medial condyle. These regions were established by identifying the point that was equidistant from the tip of the sagittal ridge and the bottom of the condylar groove as the interface between the condylar groove and sagittal ridge regions. The same distance abaxial from the bottom of the condylar groove determined the interface between the condylar groove and the condyle regions. Adaptation of the subchondral plate with coalescence of trabeculae develops at an early age in the Thoroughbred horse (black arrowheads). Calcified section after bulk-staining with basic fuchsin, bar = 10 mm. The hyaline articular cartilage has been removed by digestion with 0.1 m NaOH (Radtke et al. 2003). Reproduced from Muir et al. (2006), with permission from Elsevier.

Download figure to PowerPoint

Extensive functional adaptation of both the shaft and the condyles of the Mc-III bone develops over time in Thoroughbreds (Nunamaker et al. 1989; Boyde & Firth, 2005). Functional adaptation likely protects the Mc-III bone from fracture during normal loading. However, once joint microcracking has compromised the cortical shell of the distal end of the bone, the plate-like sagittally orientated trabecular structure of the distal metaphysis provides little resistance to condylar fracture propagation (Boyde et al. 1999; Currey, 2003; Muir et al. 2006). Accumulation and coalescence of microcracks in the joint surface and adjacent adapted subchondral plate of the condylar grooves (Muir et al. 2006) likely propagates growth of a critical crack in the cortical shell, enabling stress fracture development during normal racing activity. Remodeling of the subchondral plate targeted at microcrack repair may facilitate crack propagation by increasing porosity in the subchondral plate (Muir et al. 2006).

Although approximately 10–30% of remodeling may be targeted to microdamage repair (Burr, 2002), the targeting mechanism is not understood. Signaling from apoptosis of osteocytes adjacent to sites of microdamage may form part of the targeting mechanism (Verborgt et al. 2002; Noble et al. 2003). Profound loss of osteocytes from the navicular bone of horses, with associated microdamage accumulation and up-regulation of remodeling suggests that such a targeting mechanism may also exist in equine bone (Bentley et al. 2007). However, this relationship between up-regulation of remodeling and lower osteocyte density has not been found consistently in other skeletal sites in equine distal limb bone that are predisposed to fatigue injury and stress fracture (Da Costa Gomez et al. 2005). The purpose of the present study was to determine whether habitual athleticism in the racing Thoroughbred and the targeted remodeling of subchondral microcracks that develop naturally in this model are associated with site-specific disruption of the network of osteocytes in the subchondral plate of the distal end of the Mc-III. We hypothesized that targeted modeling of fatigue damage to the joint surface would be associated with osteocyte disruption and loss of osteocytes from the subchondral plate.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Horses

Distal thoracic limb pairs were transected at the level of the carpus and stored at −20 °C after humane euthanasia. Mc-III bones from 10 Thoroughbred racehorses that had been euthanatized because of an exercise-associated injury during racing were used. Age and racing history were obtained. Extensive adaptation of the distal Mc-III develops by the time Thoroughbreds undertake racing activity (Nunamaker et al. 1989; Boyde & Firth, 2005). A left or right Mc-III bone with an intact distal end was selected from each horse after radiographs were made. Condylar fracture was present in three of these horses in the contralateral limb. A group of eight horses that were not in race training and had not had a recent history of racing were also studied. These horses were euthanatized for reasons unrelated to the present study.

Specimen preparation

After thawing to room temperature, each Mc-III bone was transected at the distal metaphysis. An oblique dorso-proximal palmaro-distal frontal plane bone block of the distal end of the Mc-III (Riggs et al. 1999b; Muir et al. 2006), approximately 1 cm thick was then prepared using a band saw. The bone block of the joint surface was then divided into five pieces using sagittal plane cuts to create separate blocks of each of the five regions-of-interest: lateral condyle, lateral condylar groove, sagittal ridge, medial condylar groove and medial condyle (Fig. 1).

Bone blocks were fixed in 70% ethanol and bulk-stained in 1% basic fuchsin in a graded series of ethanols (80%, 90%, 100%) for a total staining time of 18 days under a vacuum (20–40 mmHg) (Burr & Hooser, 1995). This technique stains osteocyte lacunae and canaliculi and stains microcracks and diffuse matrix damage that existed before histologic sectioning, thus allowing them to be differentiated from unstained artifactual damage associated with section preparation (Burr & Hooser, 1995; Bentolila et al. 1998). This method also stains blood vessels within calcified cartilage and the hyaline articular cartilage. After embedding in polymethylmethacrylate, calcified oblique frontal plane sections were prepared 125 µm thick from the center of each block.

Histomorphometry

Morphometric measurements were made from the five regions-of-interest (Fig. 1, Muir et al. 2006) in each horse using 100× and 200× magnification and image analysis software (Image J, NIH, Bethesda, MD, USA). For each region-of-interest, the following features were studied: 1) articular cartilage width; 2) vascular ingrowth in the articular cartilage; 3) microcracking of the joint surface; 4) tidemark duplication; 5) adaptation of the subchondral plate; 6) osteocyte density of the subchondral plate and the epiphyseal trabecular bone. All data for each dependent variable were collected by a single observer (A.L.P. or S.J.S.).

Quantification of cartilage width and tidemark duplication

Using bright-field microscopy, hyaline and calcified cartilage areas were determined in each region. The average width of the hyaline cartilage layer (Hy.Cg.Wi, mm) and the calcified cartilage layer (Cl.Cg.Wi, mm) in each joint region was determined by dividing the area by the length of each region-of-interest. In each region-of-interest, digital photomicrographs of the calcified cartilage were captured. The number of tidemarks was also counted at four evenly spaced sites across each region-of-interest. The tidemark is the transition zone between mineralized and non-mineralized cartilage. Multiple tidemarks indicate advancement of calcified cartilage towards the joint surface.

Quantification of microcracking and blood vessels in calcified cartilage and subchondral bone

We considered a linear structure to be a subchondral microcrack if it was stained with basic fuchsin. Unstained cracks were considered processing artifacts. Stained microcracks were counted within each region and normalized to the length of the regional bone boundary to give a microcrack boundary density (N.Cr/B.Bd, n mm−1) in each joint region. Blood vessels that entered the calcified cartilage within each joint region were also counted and normalized to the length of the regional bone boundary to give a blood vessel boundary density (N.Ve/B.Bd, n mm−1). In addition, blood vessels that entered the hyaline cartilage within each joint region were also identified and normalized to the length of the regional cartilage boundary (N.Ve/Cg.Bd, n mm−1).

Quantification of subchondral bone adaptation

Initial examination of sections determined that subchondral bone adjacent to calcified cartilage often had minimal uptake of basic fuchsin by the cells and matrix constituents of bone tissue after bulk-staining. This area of atypical bone matrix was measured and normalized relative to the regional bone boundary (A-B.Ar/B.Bd, mm2 mm−1). In addition, the porosity of the subchondral plate was found to be decreased within approximately 3 mm of the bone-calcified cartilage interface. Therefore, within each region-of-interest across the joint surface, bone volume fraction (B.Ar/T.Ar, %) was determined in four image fields from the bone-calcified cartilage interface proximally into the subchondral bone for a total distance of 3.2 mm (Muir et al. 2006).

Osteocyte morphometry

Bulk-staining of bone with basic fuchsin highlights the non-mineralized matrix layer that lines lacunae and canaliculi, as well as cell membranes (Bentolila et al. 1998). Osteocyte morphology in bone sections was assessed using confocal laser microscopy (Bio-Rad MRC-1012 Laser Scanning Confocal Microscope, Bio-Rad, Hercules, CA, USA) using a 60× objective, as basic fuchsin fluoresces strongly in a krypton/argon laser with a 568-nm excitation and a 585-nm emission filter. Disruption of osteocytes is directly related to loss of basic fuchsin staining (Bentolila et al. 1998; Bentley et al. 2007). Representative Z-stack images consisting of three fields-of-view, 2 µm apart, were collected from the subchondral plate immediately below the calcified cartilage and in the more proximal trabecular bone of the epiphysis, approximately 10 mm from the joint surface. If atypical subchondral bone was found, Z-stack images were also collected from this tissue. Volumetric osteocyte number density (Ot.N/TV, n mm−3) was then determined (Colopy et al. 2004). A proportion of these cells and their associated lacunae were found to have atypical morphology indicative of osteocyte devitalization or death. If loss of fluorescent staining, a lacuna with a shrunken cell, or an empty lacuna was seen, the cell was counted as an atypical osteocyte. An atypical osteocyte density was then derived (A-Ot.N/TV, n mm−3) (Colopy et al. 2004).

Statistical analysis

Repeated-measures ANOVA and a Bonferroni post-hoc t-test were used to determine the effect of group (racing Thoroughbreds and non-athletic horses) on hyaline cartilage width (Hy.Cg.Wi), calcified cartilage width (Cl.Cg.Wi), number of tidemarks, blood vessel boundary density (N.Ve/B.Bd), bone volume fraction (B.Ar/T.Ar), volumetric osteocyte density (Ot.N/TV), and volumetric density of atypical osteocytes (A-Ot.N/TV). Student's t-test was used to determine whether horse age was significantly different between groups. These data are presented as mean ± SD. All error bars presented in figures denote 1 SD. The Kolmogorov–Smirnov test indicated that N.Ve/Cg.Bd and N.Cr/B.Bd data were not normally distributed. Therefore, the Mann–Whitney U and the Friedman ANOVA tests were used for analysis of these variables. Associations between dependent variables and horse age were examined using the Pearson r statistic (Muir et al. 2006). Correlations for N.Ve/Cg.Bd and N.Cr/B.Bd were examined using the Spearman R statistic (Muir et al. 2006). In the group of racing Thoroughbreds, dependent variables were also correlated with cumulative athletic activity, which was estimated from the number of races undertaken (Muir et al. 2006). Least-squares regression was also used to examine associations between N.Cr/B.Bd and Ot.N/TV, and B.Ar/T.Ar and Ot.N/TV (Vashishth et al. 2002). Results were considered significant at P < 0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Mean age of the racing Thoroughbreds was 4.4 ± 1.6 years (range of 3–8 years). Mean age of the non-athletic horses was 6 ± 1.9 years (range of 3–8 years); the age of one horse was not known. There was no significant difference in age between the two groups (P = 0.09). In the Thoroughbreds, the number of races undertaken ranged from 3 to 65.

Articular cartilage morphology

There were no significant differences in cartilage morphology between groups (P > 0.05) (Fig. 2). Hyaline cartilage in the sagittal ridge was thinner compared with the condyles (P < 0.001). Hyaline cartilage in the lateral condyle was thicker compared with all other joint regions (P < 0.05). Hyaline cartilage width was also lower in the condylar grooves than in the condyles. Adaptation of calcified cartilage followed a similar pattern. Calcified cartilage width in racing Thoroughbreds was not significantly different when compared with non-athletic horses (P = 0.08).

image

Figure 2. Effect of joint region on hyaline and calcified cartilage width (Hy.Cg.Wi and Cl.Cg.Wi, respectively). There were no significant differences in cartilage morphology between groups. (A) Hyaline cartilage was thinnest in the sagittal ridge and thickest over the lateral condyle. (B) The greatest difference in calcified cartilage regions was between the medial condyle and the lateral and medial condylar grooves. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. Joint regions with differing superscript letters are significantly different at P < 0.05. Bars denote one standard deviation.

Download figure to PowerPoint

Joint region did not have a significant effect on the number of tidemarks (P > 0.05). Overall, the number of tidemarks was increased in the non-athletic horses when compared with the group of racing Thoroughbreds (Fig. 3, P < 0.05).

image

Figure 3. Effect of joint region on number of tidemarks in articular cartilage. Joint region did not have a significant effect on formation of tidemarks, although the number of tidemarks was greater in non-athletic horses when compared with Thoroughbred racehorses. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. *Overall, tidemark formation was greater in the group of non-athletic horses (P < 0.05). Bars denote one standard deviation.

Download figure to PowerPoint

Subchondral microcracking and in-growth of blood vessels into calcified cartilage

Microcrack density varied significantly across joint regions in the racing Thoroughbred group (P < 0.001), but not in the non-athletic group (Fig. 4, P = 0.09). Microcrack density was increased in the condylar grooves of Thoroughbreds when compared with the other joint regions (P < 0.05) and was also increased when compared with the condylar grooves of non-athletic horses (P < 0.05). Microcracks involving the joint surface were typically oblique to the joint surface, at approximately 45º, and were orientated towards the interface between the condylar groove and the condyle regions. Development of a parasagittal linear defect in the condylar groove was also seen in some horses in this interface area, with collapse of the overlying articular cartilage and targeted remodeling of associated microcracks (Fig. 5).

image

Figure 4. Effect of joint region on subchondral microcrack boundary density (N.Cr/B.Bd). Joint microcracks were increased in both the lateral and medial condylar grooves of racing Thoroughbreds when compared with non-athletic horses. Joint microcracking in these regions in the Thoroughbred group was also increased when compared with other regions of the joint. Regional comparisons within the non-athletic group of horses were not significant. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. In the racing Thoroughbred group, joint regions with differing superscript letters are significantly different. Joint regions marked with an asterisk have increased N.Cr/B.Bd in the Thoroughbred group when compared with the non-athletic group of horses (P < 0.05).

Download figure to PowerPoint

image

Figure 5. Photograph of an oblique frontal calcified bone section through the palmar distal region of the medial condylar groove of a three-year-old male Thoroughbred racehorse. Formation of a large resorption space in the subchondral plate (black arrows) is co-localized with development of a parasagittal defect in the joint surface and collapse of the overlying articular cartilage. Calcified section after bulk-staining with basic fuchsin, bar = 300 µm.

Download figure to PowerPoint

Microcracks were usually co-localized with blood vessels penetrating calcified cartilage. In some racing Thoroughbreds, coalescence of oblique microcracks led to fracturing of the joint surface (Fig. 6A,B). In the condyles, coalescence of obliquely orientated microcracks deep within the subchondral plate was associated with an intense targeted remodeling response and formation of a large articular intracortical saucer-shaped stress fracture (Fig 6C,D).

image

Figure 6. Photograph of oblique frontal sections through the palmar distal region of the lateral condyle (A) and lateral condylar groove (B) of the distal end of the third metacarpal (Mc-III) of an eight-year-old Thoroughbred racehorse. Notice that oblique microcracks involving the articular surface (white arrowheads) are orientated towards the axial margin of the lateral condyle and the abaxial margin of the lateral condylar groove. Coalescence of these microcracks has led to fracture of the subchondral plate and the overlying articular cartilage at the interface between the lateral condylar and condylar groove. Multiple tidemarks are also evident. Microcracking of the joint surface was associated with formation of resorption spaces within the subchondral plate (white arrows). In another four-year-old Thoroughbred racehorse, coalescence of microcracking deep within the subchondral plate of the lateral (C) and medial (D) condylar regions led to development of large saucer-shaped stress fractures of the joint surface. Calcified section after bulk-staining with basic fuchsin, A,B, bar = 300 µm; C,D, bar = 600 µm.

Download figure to PowerPoint

Subchondral blood vessels often entered calcified cartilage in both groups of horses. The density of vessels was not significantly different between joint regions, although vessel density was increased in the group of non-athletic horses when compared with racing Thoroughbreds (Fig. 7, P < 0.05). The presence of vessels entering hyaline cartilage was much lower and was not influenced by group (P > 0.7), or joint region (P > 0.12).

image

Figure 7. Effect of joint region on the density of blood vessels penetrating the articular surface (N.Ve/B.Bd). Joint region did not have a significant effect on ingrowth of blood vessels in the joint surface, although N.Ve/B.Bd was increased in non-athletic horses when compared with racing Thoroughbreds. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. N.Ve/B.Bd was increased in the group marked with an asterisk (P < 0.05). Bars denote one standard deviation.

Download figure to PowerPoint

Adaptation of the subchondral plate

Atypical subchondral bone was increased in racing Thoroughbreds when compared with non-athletic horses (Fig. 8, P < 0.001). In non-athletic horses, the area of this tissue was small and did not vary between joint regions. In contrast, in racing Thoroughbreds, the area of bone was increased in the condyles when compared with the other joint regions, and was saucer-shaped (Figs 8–9, P < 0.05). Fracture of the joint surface of the condyle secondary to targeted remodeling of articular microcracks was usually localized in atypical subchondral bone (Fig. 9). Fracture of the joint surface developed because of coalescence of microcracks. Collapse of the subchondral plate and the overlying cartilage was also seen (Fig. 9).

image

Figure 8. Effect of joint region on the area of atypical subchondral bone tissue area (A-B.Ar/B.Bd). A-B.Ar/B.Bd was increased in racing Thoroughbreds when compared with non-athletic horses. Regional differences were found in the Thoroughbred group, but not in the group of non-athletic horses. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. A-B.Ar/B.Bd was increased in the group marked with an asterisk (P < 0.001). Within a group, joint regions with differing superscript letters are significantly different at P < 0.05. Bars denote one standard deviation.

Download figure to PowerPoint

image

Figure 9. Photomicrographs of oblique frontal sections through the palmar distal region of the distal end of the Mc-III bone. (A) In the lateral condylar region of this racehorse, a large saucer-shaped region of the subchondral plate consisted of atypical bone tissue with reduced uptake of basic fuchsin. (B) In this racehorse, a saucer-shaped line of increased uptake of basic fuchsin is evident at the margin of atypical bone (black arrows). (C) In this racehorse, large microcracks have developed in atypical subchondral bone. (D) In this racehorse, collapse of the subchondral plate has occurred, with thickening of the overlying hyaline articular cartilage. (E) In this non-athletic horse, the area of poorly stained atypical subchondral bone adjacent to the joint surface is much smaller than in the racing Thoroughbreds. Regions-of-interest were lateral condyle (A,C,D) and medial condyle (B,E). Calcified bone sections after bulk-staining with basic fuchsin, bar = 3 mm.

Download figure to PowerPoint

Overall, bone volume fraction in the subchondral plate was increased in racing Thoroughbreds when compared with non-athletic horses (Fig. 10, P < 0.001). In non-athletic horses, differences in bone porosity were found between subchondral bone adjacent to the joint surface versus more proximal fields in each joint region. In contrast, in the Thoroughbreds, such differences were only evident in the sagittal ridge. In the group of non-athletic horses, bone volume fraction in the sagittal ridge and the condylar grooves was not significantly different, whereas in the group of racing Thoroughbreds, these comparisons were different (P < 0.001).

image

Figure 10. Effect of joint region on bone volume fraction (B.Ar/T.Ar) at four levels from the calcified cartilage interface (Field 1) proximal to the epiphyseal side of the subchondral plate (Field 4) for a total distance of 3.2 mm (Muir et al. 2006). B.Ar/T.Ar was increased in racing Thoroughbreds when compared with non-athletic horses (P < 0.001). The effect of joint region and field were significantly different in the two groups of horses (P < 0.005), with the racing Thoroughbreds showing increased sclerosis of the subchondral plate. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. B.Ar/T.Ar was increased in the group marked with an asterisk (P < 0.05). Joint regions with differing superscript letters are significantly different at P < 0.05. Within a group and region, fields with differing letters are also significantly different at P < 0.05. Bars denote one standard deviation.

Download figure to PowerPoint

Osteocyte morphology in the subchondral plate and epiphyseal trabecular bone

Overall, osteocyte density was significantly decreased in the subchondral plate when compared with epiphyseal trabecular bone (P < 0.001, Table 1). However, there were no significant differences in osteocyte density between the two groups of horses (P = 0.21). Overall, osteocyte density was increased in the medial condyle when compared with the medial condylar groove.

Table 1.  Influence of joint region on osteocyte populations in the subchondral plate and epiphyseal trabecular bone of the distal end of the third metacarpal bone in the horse
Joint regionOsteocyte density, (Ot.N/TV, ×104 mm−3)Atypical osteocyte density (A-Ot.N/TV, ×104 mm−3)
Subchondral plateEpiphyseal trabecular boneSubchondral plateEpiphyseal trabecular bone
  1. Mean ± SD. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. Differences in Ot.N/TV and A-Ot.N/TV between the subchondral plate and epiphyseal trabecular bone areas in each joint region are as indicated. NS, not significant; *P < 0.05; **P < 0.01, ***P < 0.001.

Non-athletic horses
LC13.06 ± 5.47NS19.78 ± 6.1710.01 ± 3.45NS 4.76 ± 3.48
LCG12.70 ± 5.22*20.75 ± 3.2910.13 ± 4.30NS 6.96 ± 3.60
SR14.78 ± 4.91NS21.61 ± 2.83 9.28 ± 4.21NS 5.62 ± 3.79
MCG12.70 ± 3.84NS19.53 ± 4.84 9.40 ± 3.26NS 4.27 ± 3.34
MC16.24 ± 2.86NS19.90 ± 4.4611.60 ± 2.05** 4.27 ± 4.17
Thoroughbred racehorses
LC14.26 ± 4.40NS20.70 ± 6.1710.94 ± 5.28NS 7.91 ± 5.18
LCG14.94 ± 5.56***24.80 ± 4.1513.67 ± 4.44NS 9.86 ± 5.30
SR12.89 ± 2.98***23.33 ± 4.7511.23 ± 2.53NS10.31 ± 6.86
MCG10.74 ± 4.07*20.18 ± 4.3410.74 ± 4.03NS 9.98 ± 5.56
MC17.97 ± 4.30NS22.66 ± 2.4314.45 ± 6.00NS10.45 ± 4.37

Overall, the density of atypical osteocytes was increased in racing Thoroughbreds when compared with non-athletic horses (P = 0.01). Atypical osteocyte density was also increased in the subchondral plate when compared with epiphyseal trabecular bone (P = 0.001, Table 1). Joint region did not have a significant effect on atypical osteocyte density (P = 0.12). No significant relationship was found between osteocyte density and bone volume fraction, or between osteocyte and microcrack densities.

Confocal microscopy revealed extensive diffuse damage was present in the bone matrix of the subchondral plate, with loss of osteocytes and loss of osteocyte connectivity in both groups (Fig. 11). These changes were less evident in epiphyseal trabecular bone. In subchondral bone with atypical tissue staining, confocal microscopy again revealed extensive diffuse damage with disruption to the osteocyte network (Fig. 11). Overall, osteocyte density and atypical osteocyte density in atypical subchondral bone in the racing Thoroughbred group were 13.52 × 104 mm−3 and 13.45 × 104 mm−3 respectively; more than 99% of osteocytes had atypical morphology.

image

Figure 11. Confocal Z-stack photomicrographs of the subchondral plate in oblique frontal calcified bone sections through the palmar distal region of the distal end of the Mc-III bone. (A) In the lateral condyle of this non-athletic horse, many osteocytes are visible in the field-of-view, with an extensive network of canaliculi. (B) In the medial condyle of an 8-year-old racing Thoroughbred, fine linear microcracks within the bone matrix are seen (arrows). (C) In the medial condylar groove of the same horse as (B), obvious linear and diffuse damage (arrowheads) is seen, with extensive uptake of basic fuchsin stain in the bone matrix. Pronounced osteocyte loss is also seen, with many empty lacunae within the field-of-view (arrows). (D) In atypical bone in the lateral condyle of this 3-year-old neutered male racing Thoroughbred, again extensive diffuse damage is seen (arrowheads) with loss of osteocytes and loss of connectivity of the remaining cells. Calcified bone sections after bulk-staining with basic fuchsin, bar = 25 µm.

Download figure to PowerPoint

Influence of horse age and cumulative athletic activity on functional adaptation of the joint surface

In the both groups of horses, few significant correlations with age were found (Table 2). Similarly, in the group of racing Thoroughbreds, few significant correlations with cumulative racing activity were found (Table 2).

Table 2.  Influence of horse age and cumulative athletic activity in joint adaptation
VariableJoint region
LCLCGSRMCGMC
  1. LC, lateral condyle; LCG, lateral condylar groove; SR, sagittal ridge; MCG, medial condylar groove; MC, medial condyle. Hy.Cg.Wi, hyaline cartilage width; Cl.Cg.Wi, calcified cartilage width; N.Ve/B.Bd, blood vessel density entering calcified cartilage; N.Ve/Cg.Bd, blood vessel density entering hyaline cartilage; N.Cr/B.Bd, microcrack density; B.Ar/T.Ar, bone volume fraction in the subchondral plate immediately adjacent to calcified cartilage; Ot.N/TV, subchondral plate osteocyte density; A-Ot.N/TV, atypical subchondral plate osteocyte density. Significant correlations are indicated in bold type. Correlations were not performed for empty cells because of lack of variance in the dependent variable. Similar B.Ar/T.Ar results were obtained for the three remaining more proximal fields. No significant correlations were identified for Ot.N/TV and A-Ot.N/TV in the trabecular region-of-interest.

Non-athletic horse age
Hy.Cg.Wir = 0.77, P = 0.04r = 0.07, P = 0.88r = 0.02, P = 0.97r = 0.46, P = 0.30r = 0.83, P = 0.02
Cl.Cg.Wir = 0.20, P = 0.67r = 0.65, P = 0.11r = 0.71, P = 0.07r = 0.39, P = 0.39r = 0.10, P = 0.83
No. of Tidemarksr = −0.09, P = 0.84r = 0.28, P = 0.55r = −0.54, P = 0.21r = 0.16, P = 0.73r = −0.18, P = 0.71
N.Ve/B.Bdr = −0.81, P = 0.03r = −0.74, P = 0.06r = −0.33, P = 0.48r = −0.92, P < 0.01r = −0.55, P = 0.20
N.Ve/Cg.Bd Sr = −0.36, P = 0.42 Sr = 0.10, P = 0.82 
N.Cr/B.BdSr = 0.59, P = 0.16Sr = 0.19, P = 0.69 Sr = 0.33, P = 0.47Sr = 0.79, P = 0.03
B.Ar/T.Arr = −0.13, P = 0.78r = −0.36, P = 0.42r = −0.28, P = 0.54r = −0.35, P = 0.44r = 0.10, P = 0.83
Ot.N/TVr = −0.08, P = 0.86r = 0.77, P = 0.04r = 0.42, P = 0.35r = −0.02, P = 0.96r = 0.04, P = 0.93
A-Ot.N/TVr = −0.56, P = 0.19r = 0.72, P = 0.07r = 0.16, P = 0.72r = −0.00, P = 1.00r = −0.19, P = 0.68
Thoroughbred racehorse age
Hy.Cg.Wir = 0.38, P = 0.28r = 0.50, P = 0.14r = 0.59, P = 0.07r = 0.62, P = 0.06r = 0.56, P = 0.10
Cl.Cg.Wir = 0.01, P = 0.98r = 0.25, P = 0.49r = 0.49, P = 0.15r = 0.38, P = 0.28r = 0.53, P = 0.11
No. of Tidemarksr = −0.33, P = 0.36r = 0.51, P = 0.13r = 0.33, P = 0.36r = −0.07, P = 0.85r = −0.02, P = 0.95
N.Ve/B.Bdr = −0.46, P = 0.19r = −0.30, P = 0.41r = −0.27, P = 0.45r = −0.33, P = 0.35r = −0.36, P = 0.30
N.Ve/Cg.Bd Sr = −0.28, P = 0.42Sr = −0.37, P = 0.30 Sr = −0.37, P = 0.30
N.Cr/B.BdSr = 0.70, P = 0.03Sr = 0.68, P = 0.03 Sr = 0.50, P = 0.15Sr = 0.13, P = 0.72
B.Ar/T.Arr = 0.10, P = 0.78r = −0.01, P = 0.98r = −0.33, P = 0.36r = −0.33, P = 0.35r = −0.35, P = 0.32
Ot.N/TVr = −0.22, P = 0.54r = −0.47, P = 0.17r = 0.22, P = 0.55r = 0.32, P = 0.37r = −0.22, P = 0.55
A-Ot.N/TVr = 0.39, P = 0.27r = −0.37, P = 0.29r = 0.10, P = 0.78r = 0.67, P = 0.04r = −0.12, P = 0.75
Thoroughbred racehorse cumulative racing activity
Hy.Cg.Wir = 0.22, P = 0.54r = 0.57, P = 0.08r = 0.55, P = 0.10r = 0.72, P = 0.02r = 0.60, P = 0.07
Cl.Cg.Wir = 0.25, P = 0.48r = 0.57, P = 0.09r = 0.65, P = 0.04r = 0.56, P = 0.09r = 0.67, P = 0.03
No. of Tidemarksr = −0.08, P = 0.83r = 0.70, P = 0.02r = 0.42, P = 0.22r = −0.31, P = 0.39r = 0.27, P = 0.45
N.Ve/B.Bdr = −0.59, P = 0.07r = −0.15, P = 0.66r = −0.22, P = 0.54r = −0.41, P = 0.24r = −0.34, P = 0.34
N.Ve/Cg.Bd Sr = −0.27, P = 0.45Sr = −0.46, P = 0.17 Sr = −0.18, P = 0.63
N.Cr/B.BdSr = 0.79, P < 0.01Sr = 0.59, P = 0.07 Sr = 0.72, P = 0.02Sr = 0.20, P = 0.58
B.Ar/T.Arr = 0.15, P = 0.68r = −0.13, P = 0.71r = −0.22, P = 0.54r = −0.27, P = 0.46r = −0.36, P = 0.31
Ot.N/TVr = −0.03, P = 0.93r = −0.35, P = 0.31r = 0.06, P = 0.87r = 0.08, P = 0.82r = −0.09, P = 0.80
A-Ot.N/TVr = 0.28, P = 0.44r = −0.28, P = 0.44r = −0.04, P = 0.92r = 0.36, P = 0.31r = −0.20, P = 0.59

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Exercise-induced distal limb fractures are a common event in the Thoroughbred racehorse (Verheyen et al. 2006). The Thoroughbred is an extreme athlete, and large stresses are transmitted across the distal Mc-III joint surface during high speed running (Riggs et al. 1999b). Consequently, this model provides an opportunity to better understand joint adaptation in running athletes and how adaptive failure may lead to development of osteoarthritis and articular stress fracture. In this model, athletic activity in horses activates endochondral ossification of the joint surface and induces accumulation of site-specific microcracks at the joint surface (Muir et al. 2006). In the present work, we now show that site-specific microcrack accumulation is co-localized with development of an intra-cortical articular stress fracture, as well as development of sagittal plane linear defects in the subchondral plate of the condylar grooves.

Both hyaline and calcified cartilage were relatively thin in the condylar grooves when compared with the condyles in both groups of horses. Although differences between non-athletic horses and racing Thoroughbreds were not significant, calcified cartilage in particular was thinner in the Thoroughbred group. Calcified cartilage width generally increases with aging (Martinelli et al. 2002), in contrast to the decrease in Cl.Cg.Wi and increase in the number of tidemarks found with aging in humans (Lane & Bullough, 1980). Young 18-month-old Thoroughbreds have relatively thick calcified cartilage in the condylar grooves, with exercise having little effect on calcified cartilage adaptation at this age (Doube et al. 2007). With treadmill exercise, calcified cartilage thickness increases in the equine middle carpal joint (Murray et al. 1999). In our study, calcified cartilage width in the sagittal ridge and medial condyle was also positively correlated with cumulative athletic activity. However, racing Thoroughbreds had lower Cl.Cg.Wi values in the condylar groove regions when compared with the other regions of the joint. This suggests that low calcified cartilage width in the condylar grooves may be a key feature that promotes overload of the joint surface and development of articular microcracks. This feature was more apparent in the current study than in our earlier work (Muir et al. 2006).

Adaptation of subchondral plate bone with an increase in bone volume fraction occurs rapidly in response to training (Boyde & Firth, 2005), and subchondral plate bone volume fraction was increased in the racing Thoroughbred group. Although no dependent variable was highly correlated with horse age or cumulative racing activity over all joint regions, cartilage adaptation was more correlated than subchondral bone adaptation. This suggests that cartilage adaptation to cyclic load occurs slowly over time. Again, this finding contrasts previous work from our laboratory, which found that subchondral bone volume fraction was decreased with cumulative racing activity (Muir et al. 2006).

Although the number of tidemarks and the number density of vessels entering calcified cartilage were increased in the non-athletic horses, this was not associated with significantly thinner hyaline cartilage width. This suggests that habitual athleticism and high joint loads may have important effects on endochondral ossification of the joint surface over time. Taken together, these results suggest that athleticism may lead to loss of hyaline cartilage from the articular surface over time at high load sites, since in the non-athletic horses there was evidence of tidemark advancement, but not cartilage thinning. In humans, vascularization of calcified cartilage decreases with aging (Lane et al. 1977), and similar trends were identified in the present study for the lateral condyle and the medial condylar groove of non-athletic horses. Cartilage vascularization in the carpal bone of similarly aged horses had higher numbers of blood vessels entering calcified cartilage (Norrdin et al. 1999), suggesting that endochondral ossification may be variable between joints. Cartilage vascularization was often identified without an associated microcrack, but articular microcracks were typically co-localized with blood vessels, suggesting that vascularization of calcified cartilage promotes microcrack development (Muir et al. 2006). If cumulative athletic activity promotes loss of cartilage from the joint surface, then this will likely also promote adaptive failure, accumulation of microdamage in the joint surface and development of stress fracture.

Although it is known that athletic activity induces changes in all of the tissues of the bone end (Oettmeier et al. 1992), the specific reasons for development of pathological joint damage remain poorly understood. Development of arrays of macroscopic subchondral cracks in the condylar grooves is an important feature of this model and often leads to development of a parasagittal stress fracture (Riggs et al. 1999a; Radtke et al. 2003). These crack arrays are fractal in that the arrays appear similar at different orders of magnification (Stepnik et al. 2004) and arise from microcracks that originate in the calcified cartilage (Muir et al. 2006). We found regional variation in microcrack density was much greater in racing Thoroughbreds than in non-athletic horses, although microcracks were most evident in the condylar grooves in both groups. Consistent with our previous work (Muir et al. 2006), we found that microcracking of the joint surface was associated with sclerosis of underlying subchondral plate. These observations suggests that athleticism may lead to accumulation of joint damage and also suggest that direct healing (Boyde, 2003) or targeted remodeling (Burr, 2002) of joint microcracks is a slow process, and that the capacity for physiologic repair is poor. Microcrack accumulation in the lateral and medial condylar grooves was not significantly different, and yet condylar stress fracture is much more common in the lateral condylar groove (Rick et al. 1983). This suggests that once a critical defect has developed in the cortical shell of the distal end of Mc-III, damage in the lateral condylar groove is more likely to propagate into a stress fracture with habitual athleticism, although the specific biomechanical mechanism is unclear.

Microcracking of calcified cartilage can be a consequence of large transarticular loads (Vener et al. 1992). Joint microcracks typically had an oblique orientation relative to the joint surface, suggesting shear failure. The shear strength of bone is much lower than its compressive strength, particularly during habitual loading (Reilly & Burstein, 1975). Interestingly, condylar microcracks were orientated axially, whereas condylar groove microcracks were orientated abaxially at approximately 45º to the joint surface, suggesting that the linear parasagittal defects develop at the maximal tensile plane, where oblique microcracks with opposite orientations coalesce and promote fragmentation of the joint surface. This plane is located at the junction of the condyle and the condylar groove, the typical site for development of parasagittal macrocracks (Riggs et al. 1999a; Radtke et al. 2003; Muir et al. 2006).

We also found that the circular osteochondral lesions (Hornoff et al. 1981) adjacent to the axial margin of the condyle are intracortical articular stress fractures. Coalescence of microcracks orientated in an oblique plane to the joint surface, again at an angle of approximately 45º, led to development of a subchondral stress fracture. Often the overlying articular cartilage was still present, such that the fracture would not necessarily be evident by observation of the joint surface. With ongoing cyclic joint loading, collapse of the subchondral plate is likely a direct consequence of stress fracture development, leading to breakdown and loss of cartilage from the articular surface (Hornoff et al. 1981; Radtke et al. 2003; Muir et al. 2006). Development of this intracortical stress fracture is associated with the maximal tensile plane being located more abaxially on the condyle. Linear parasagittal subchondral stress fracture and circular condylar intracortical stress fracture can be found alone or both lesions can be found in the same joint (Radtke et al. 2003; Muir et al. 2006). This suggests that during habitual loading, the maximal tensile plane may move across the frontal plane of the joint. However, the specific biomechanical mechanism that might cause such shifts in the maximal tensile plane is unclear.

The role of subchondral plate adaptation in promoting cartilage degeneration in osteoarthritis is controversial (Burr, 2004). Our findings suggest that subchondral bone adaptation is likely an important physiological response to joint loading and is likely a key factor in the development of joint degeneration and osteoarthritis, even if adaptation is initially protective (Radin & Paul, 1971; Burr, 2004).

The presence of atypical bone tissue in the subchondral plate of the condyles was a prominent feature of the racing Thoroughbred group, but not the non-athletic horses. The matrix changes that cause atypical basic fuchsin bulk-staining are likely to be load-related. We found that formation of this atypical tissue is co-localized with the development of joint microcracks and subchondral sclerosis. Microcracks typically develop within highly mineralized cortical bone (Wasserman et al. 2005). In addition to targeted remodeling of microcracks (Burr, 2002), Boyde (2003) has also suggested that direct healing of microcracks can occur in equine bone by infilling with densely mineralized matrix. Altered matrix mineralization is a likely explanation for atypical staining.

In the subchondral bone in both groups of horses, osteocyte density was decreased when compared with epiphyseal trabecular bone, even though bone volume fraction of compact bone is higher than trabecular bone; this suggests that loss of osteocytes from subchondral bone has occurred. Associated with this cell loss, we found a reduction in osteocyte connectivity and the presence of many osteocytes with atypical morphology. Atypical osteocyte density was increased in racing Thoroughbreds, suggesting the development of this osteocyte morphology is promoted by athletic habitual loading. In general, differences in osteocyte density between the subchondral and epiphyseal bone were most evident in the condylar grooves. Confocal microscopy revealed extensive disruption to the network of osteocytes in atypical subchondral bone tissue. In racing Thoroughbreds, osteocyte disruption was co-localized with both linear and diffuse damage, again suggesting that these findings are caused by cyclic mechanical overload of the joint surface. More than 99% of osteocytes in this type of bone had atypical morphology, suggesting that osteocyte loss is co-localized with microcrack formation and with matrix sclerosis.

A normal osteocyte population is a feature of healthy bone, with loss of osteocytes occurring in diseases such as osteoporosis (Mullender et al. 2005). However, low osteocyte density has not consistently been found at high strain sites in the equine skeleton predisposed to exercise-induced injury (Da Costa Gomez et al. 2005). The lack of significant difference in osteocyte density between groups in the present study, despite differences in microcrack density and bone volume fraction, suggests that osteocyte signaling is not the only mechanosensory pathway regulating bone adaptation and targeted remodeling in this model. Similarly, in contrast to other work (Vashishth et al. 2002), we did not find a significant relationship between osteocyte density and bone volume fraction. The innervation of bone also regulates functional adaptation to bone loading (Sample et al. 2008). However, the functional importance of the sensory innervation in the joint regions of horses (Nixon & Cummings, 1994) that accumulate microcracks with athletic habitual loading is unclear. As endochondral ossification of the joint develops, it is likely that the sensory innervation of cartilage is altered and such changes may be important for progression of osteoarthritis (Suri et al. 2007).

An important limitation of this study was that we only examined a relatively small number of bones. Differences in horse age, gender, and breed may have influenced our results. Morphometric data from horses towards the end of their natural lifespan would also help in understanding the influence of aging and habitual athleticism on joint adaptation. In future work, it would also be of interest to determine how habitual athleticism influences mineral to matrix ratio within the subchondral plate and whether direct healing of microcracks (Boyde, 2003) is evident in horses with joint microdamage.

In conclusion, our data suggest that articular stress fracture in Thoroughbred racehorses is a consequence of adaptive failure. Site-specific fractures affect both the condyles and the condylar grooves, and arise from microcrack arrays at the joint surface. Targeted remodeling is unlikely to be exclusively osteocyte-mediated because subchondral osteocyte densities were similar in both athletic and non-athletic horses, but stress fractures were only found in racing Thoroughbreds.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was supported by a grant from the AO VET Center, Switzerland. Anna Peterson was supported by a University of Wisconsin-Madison, School of Veterinary Medicine Summer Research Scholarship.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References