SEARCH

SEARCH BY CITATION

Keywords:

  • maturation;
  • growth;
  • aging;
  • rabbit;
  • bone;
  • mechanical properties;
  • composition

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We characterized the composition and mechanical properties of cortical bone during maturation and growth and in adult life in the rabbit. We hypothesized that the collagen network develops earlier than the mineralized matrix. Growth was monitored, and the rabbits were euthanized at birth (newborn), and at 1, 3, 6, 9, and 18 months of age. The collagen network was assessed biochemically (collagen content, enzymatic and non-enzymatic cross-links) in specimens from the mid-diaphysis of the tibia and femur and biomechanically (tensile testing) from decalcified whole tibia specimens. The mineralized matrix was analyzed using pQCT and 3-point bend tests from intact femur specimens. The collagen content and the Young's modulus of the collagen matrix increased significantly until the rabbits were 3 months old, and thereafter remained stable. The amount of HP and LP collagen cross-links increased continuously from newborn to 18 months of age, whereas PEN cross-links increased after 6 months of age. Bone mineral density and the Young's modulus of the mineralized bone increased until the rabbits were at least 6 months old. We concluded that substantial changes take place during the normal process of development in both the biochemical and biomechanical properties of rabbit cortical bone. In cortical bone, the collagen network reaches its mature composition and mechanical strength prior to the mineralized matrix. © 2010 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 28:1626–1633, 2010

Collagen is the principal structural component of bone matrix, accounting for about 1/3 of mineralized bone tissues.1 During development, bone formation and growth rates are regulated by genetic and hormonal factors, and by external factors such as mechanical loading that interact to determine the final anatomy and strength of the skeleton.1–3 During this period, bone also changes in its composition, structure, and function. The organic matrix changes its composition and structural organization, and mechanical stiffness increases with increased bone mineral density (BMD).3, 4 The majority (>80%) of the mineral can be found within the collagen fibril network,5 but limited information is available about the factors that determine the normal deposition of mineral, for example, the timing and how this can determine the properties of the adult tissue.

The structure and function of maturing and growing bone were studied in several animal models.4, 6, 7 Most studies focused on bone development and growth in rabbits, investigating the development of the primary and secondary ossification centers or evaluating bone growth at the growth plate.8–10 Other studies showed that BMD increases and biomechanical properties change concurrently during maturation and growth.4, 6 However, the development of the collagen framework has been less extensively studied. During development, at least in the equine, the collagen matrix is formed rapidly, and the fibrils are restructured and re-orientated with time towards the main loading directions of the bone.4 In addition, changes also occur in the amounts of intra- and interfibrillar enzymatic (hydroxylysylpyridinoline, HP; lysylpyridinoline, LP) and non-enzymatic (pentosidine, PEN) cross-links.11–14 In experimental studies, the collagen framework in long bones of mice and rats undergoes significant structural modifications during growth and maturation.6, 15, 16 A previous study in mice found that the content and mechanical strength of the collagen matrix increased during early maturation,6 whereas during adult age, the collagen composition remained stable despite decreased stiffness of the collagen matrix.17 Also, in equine long bones, collagen orientation alters significantly during maturation and growth towards a more longitudinal fibril direction.18 Moreover, studies on equine subchondral bone showed that both the biochemical and structural modifications of the collagen network primarily develop during the first half year post-partum.4, 19 Comparison between studies is often difficult due to differences in species, anatomical location, type of bone (e.g., cortical, trabecular, or subchondral bone), and experimental conditions. Furthermore, how the properties of the collagen matrix, the mineral content, and the mechanical integrity of bone develop in relation to one another, and how they individually and together control the bone's mechanical resistance to fracture remains unknown.

We aimed to characterize the development of the collagen network, its cross-links, and its relationships to the mechanical properties in demineralized cortical bone of the rabbit. Simultaneously, these characteristics were evaluated in relation to quantified properties of the mineralized bone. Hence, we aimed to clarify the contributions of the mineral content and collagen matrix to the functional properties and to reveal the timing of the development of the collagen network and its relation to the mineralized matrix properties. We hypothesized that the development of the collagen network in bone precedes the progress of skeletal calcification and that the mechanical strength of the network reaches its mature levels prior to the mineralized matrix. The biochemical and mechanical development of the collagen and mineralized matrix in rabbit cortical bone was assessed from birth to mature adult age.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Experimental Conditions

Female New Zealand white rabbits were bred and cared for in the National Laboratory Animal Center, University of Kuopio, Finland. The animals were divided into six age groups at different stages of maturation: newborn (0 days), and 1, 3, 6, 9, and 18 months. Each group included 10–15 animals (Table 1). The rabbits were housed in cages with standard floor dimensions of 70 cm × 80 cm. Water and commercial food pellets were freely available (STANRAB up to the age of 11 weeks and RABMA thereafter; Special Diets Services, Witham, UK). Room temperature was maintained at 15°C, and the lights on/lights off cycle was 12/12 h. After euthanasia at predetermined experimental time points, the rabbits were weighed, and the femora and tibiae were dissected free from soft tissues while fresh and stored in airtight specimen tubes containing phosphate-buffered saline (PBS, pH 7.4) at −20°C. The distances between proximal and distal articulating surfaces were measured with a digital caliper. The study protocol was approved by the Animal Care and Use Committee of the University of Kuopio.

Table 1. Not All Age Groups and Bones Were Tested With All Methodsab
Age GroupAnalysis Methods
SizeBiochemistrypQCTTensile Testing3-Point-Bending
  • T, tibia; F, femur.

  • a

    The numbers of rabbits used in each age group and for each analysis method are given for the femur and tibia separately.

  • b

    Size includes weight and length measurements; Biochemistry includes biochemical analysis of collagen content and cross-links. pQCT includes measurements of bone mineral density. Tensile testing refers to mechanical testing of decalcified bones. 3-Point-bending refers to mechanical testing of intact (mineralized) bones.

NewbornT (15), F (15)T (15), F (14)F (14)T (14) 
1 monthT (15), F (15)T (15), F (15)F (15)T (13)F (15)
3 monthsT (12), F (12)T (12), F (12)F (12)T (10)F (12)
6 monthsT (12), F (12)T (12), F (12)F (12)T (11)F (12)
9 monthsT (15)T (13) T (13) 
18 monthsT (10), F (10)T (10), F (10)F (10)T (8)F (9)

Mechanical Tensile Testing of Decalcified Tibiae

The left tibias were decalcified with 0.5 M ethylene diaminotetraacetic acid (EDTA), pH 8, at 4°C for 2 months, and thereafter stored in cryogenic vials in PBS buffer at −20°C until tensile testing (Table 1). The EDTA solution was changed three times a week, and the release of calcium was followed with atomic absorption spectrophotometry. Prior to mechanical testing, specimens were thawed in a water bath (30°C) for 30 min. After thawing, the specimens were kept in an ice bath within the airtight specimen tubes containing PBS for at least 30 min to ensure stable test conditions for each specimen at room temperature (20°C). The newborn samples were tested with a linear servo motorized testing machine (Newport PM500-C Precision Motion Controller, Newport PM1A1798, 112N load cell, Honeywell Sensotec 31/1430-04, load cell accuracy ± 0.5 N). All other age groups were tested with a servohydraulic material testing device (Model 8874 Instron Co, Canton, MA, load cell accuracy ± 1 N). To ensure tight but gentle clamping of the specimens, a coarse, soft rubber material was used. The bone was centered between two clamps to ensure that the breaking point occurred at the mid-diaphysis. The decalcified tissue was subsequently exposed to tension. The bone cross-section was assumed to be ellipsoidal, and the area of the breaking point was calculated using the mean inner and outer semi-minor and semi-major axis distances from both breaking ends. The axis distances were measured from pictures of bone ends drawn using a stereomicroscope equipped with a drawing tube.20

After fixation to zero offset load, the specimens were stretched at a predefined rate (mm/s) until failure. The specimen length, l (i.e., the free length between the clamps), and the pulling speed, v, were chosen to ensure that the measurements in the different age groups were comparable. The specimen length ranged between 2 and 12 mm, and the strain rate varied from 0.1 to 0.6 mm/s for bones of newborn and 18-month-old animals, respectively (Table 2). These values were set so that each specimen would break in the range of 5–10 s.16 Time, force, and actuator displacement were recorded during the test. From these data, the force and deformation at failure were obtained, and the ultimate strain at failure was calculated. The energy was calculated by integrating the force–deformation curve from 0 to the breaking point, and the Young's modulus was defined as the slope of the linear part of the stress–strain curve, using approximately 20–60% of the loading curve.

Table 2. The Parameters Used During Tensile Testing Depended on the Bone Dimension and the Animal's Agea
 L0 (mm)υ (mm/s)
  • a

    L0, specimen length (distance between the clamps); v, the displacement rate.

Newborn20.1
1 month60.3
3 months120.6
6 months120.6
9 months120.6
18 months120.6

Biochemical Analysis

High-pressure liquid chromatography (HPLC) was used for biochemical assessment of collagen content and collagen cross-linking. Mid-diaphyseal samples were cut from the left tibia after mechanical testing (Table 1). After the 3-point-bending tests, the femur samples were decalcified with 0.5 M EDTA, using the same protocol as for the tibia (above). After washing to remove excessive EDTA, samples were weighed, freeze-dried, and hydrolyzed in 6 M HCl. The yield of hydroxyproline in the hydrolysis was estimated using an external standard. The amount of collagen was calculated by assuming 300 mol of hydroxyproline per 1 mol of collagen. The amounts of the enzymatic collagen cross-links, that is, HP, LP, and non-enzymatic collagen cross-link PEN were measured by using reversed-phase HPLC analysis,13, 21, 22 with the help of HP, LP, and PEN standards that were kind gifts of Dr. Jeroen de Groot, TNO, Leiden, The Netherlands. The quantities are expressed as nmol/mg (collagen), pmol/mg (HP, LP), or fmol/mg (PEN) wet weight, and as mmol (HP, LP) or µmol (PEN) per mol collagen.

Peripheral Quantitative Computed Tomography (pQCT) Measurements

BMD (g/cm3) of the left femurs (Table 1) was measured with pQCT (Stratec XCT 960A, v. 5.20, Norland Stratec Medizintechnik GmbH, Birkenfeld, Germany). At the midshaft, one cross-sectional slice (1.25 mm) was scanned with 0.092 × 0.092 × 1.25 mm3 voxel size. The location was determined from the scout view of the pQCT device. Mean volumetric cortical BMD and cross-sectional cortical area (CSA) were measured with attenuation threshold values of 0.690 and 0.464 mg/cm3 to enhance the accuracy of both measurements.23, 24 The measurement was repeated, and the average was used for the analyses.

Mechanical Testing of Mineralized Bone

Following the pQCT measurements, the left femurs were subjected to 3-point bend tests (Table 1).6, 25 Testing was not performed on the newborn samples due to sample size limitations on the testing device. Prior to testing, specimens were refrigerated for 12 h and thereafter at room temperature for 1 h to stabilize the test conditions. The servohydraulic testing device (Instron model 8874) was used. The span length was set individually to 50% of the total length of the bone to ensure comparable testing for each specimen and age group. The press head and the two support points were rounded (10 mm diameter) to avoid shear load and cutting. The bone was centered on the supports and positioned horizontally with the anterior surface upwards. The force was directed vertically to the mid-shaft. After preconditioning to a load of 5 N using a speed 0.2 mm/min, each bone was tested to failure with a constant speed of 1.0 mm/min. Time, force, and displacement were recorded. Energy was defined by integrating the force–displacement curve from 0 to the maximal load. The stress was defined as equation image, where F is the maximum (breaking) force, L the support span length, c the radius of the specimen at the loading site, and I the axial moment of inertia obtained from the pQCT measurements.26 The elastic region of the load–displacement curve was used to define the displacement for the calculations of Young's modulus.25–27

Statistical Analysis

Age-dependent changes in the biochemical contents and mechanical properties were analyzed using the Mann–Whitney U-test between the age groups. Also, Pearson's correlation (R) analyses were used to explore relationships between mechanical and biochemical properties, and relationships between collagen and mineralized matrix properties. Statistical analysis was performed with SPSS statistical software (v. 15, SPSS, Inc., Chicago, IL).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Growth

The growth of the animals was evident up to 6 months (Fig. 1). Weight, bone length, and bone cross-sectional area all increased significantly (p < 0.001) between each age group until the rabbits were 6 months old, thereafter remaining unchanged.

thumbnail image

Figure 1. Growth was monitored via the rabbit's weight (a), length of the femur and tibia (b), and cross-sectional area of the femur and tibia (c). Results are expressed as mean ± SD. All three parameters indicated significant increases between each time point until 6 months (**p < 0.001).

Download figure to PowerPoint

Biochemical Analysis

The amount of collagen in the bone increased significantly between each time point until 3 months (Fig. 2a). The amount of enzymatic HP and LP cross-links, however, increased slowly but steadily to 18 months (Fig. 2b,c). Only a few significant differences were found between consecutive age groups. The amount of non-enzymatic cross-link PEN decreased during early time points (between newborn and 1–3 months), later increasing (Fig. 2d). Differences between the enzymatic and non-enzymatic cross-links were further demonstrated in relation to the amount of collagen (Fig. 2e,f). Also, the amounts of collagen, HP, and LP were significantly higher (p < 0.05) in the tibia than in the femur.

thumbnail image

Figure 2. Biochemical determination of the collagen content (a) and the collagen cross-links (b–d) in the mid-diaphysis of the femur and tibia. Results are expressed as mean ± SD. Collagen content increased significantly until 3 months (a). The enzymatic cross-links HP and LP increased slowly throughout life (b,c). The non-enzymatic cross-link PEN only showed a significant increase between the last two time-points (d). The relative amount of HP (e) and PEN (f) cross-links are shown (*p < 0.05, **p < 0.001).

Download figure to PowerPoint

pQCT

BMD and cortical thickness increased significantly (p < 0.001) until 6 months (Fig. 3a,b).

thumbnail image

Figure 3. Cortical BMD (a) and cortical thickness (b) increased significantly between each time point until the age of 6 months. Results are expressed as mean ± SD (**p < 0.001).

Download figure to PowerPoint

Tensile Testing of Collagen Network

The mechanical characteristics of the collagen network of the tibia mainly changed during the first 3 months (Fig. 4). Maximal force, energy, and Young's modulus all increased, while ultimate strain decreased significantly between birth and 1 month and between 1 and 3 months (p < 0.05 and p < 0.001). Maximal force continued to increase to 6 months (p < 0.05) while energy showed an increasing trend (Fig. 4a,b). Maximal force and energy decreased significantly (p < 0.05) between 6 and 18 months (Fig. 4a,b). All other properties remained constant from 3 to 18 months.

thumbnail image

Figure 4. Tensile testing of decalcified tibia. The force (a), energy (b), ultimate strain (c), and modulus (d) increased significantly between each time point until 3 months. The force further increased between 3 and 6 months of age. The force and energy showed a significant decrease between 6 and 18 months (a,b). Results are expressed as mean ± SD (*p < 0.05, **p < 0.001).

Download figure to PowerPoint

3-Point Bending of Intact Femur

Maximal force, energy, Young's modulus, and moment of inertia of intact bone all increased significantly (p < 0.01 and p < 0.001) up to 6 months (Fig. 5a–d). Most values remained constant from 6 to 18 months, except for energy, which decreased significantly (p < 0.01) during this period (Fig. 5b).

thumbnail image

Figure 5. 3-Point-bending of intact femur. Ultimate force (a), energy (b), modulus (c), and moment of inertia (d) increased significantly between each time point until 6 months. Also, energy decreased between 6 and 18 months. Results are expressed as mean ± SD (*p < 0.05, **p < 0.001).

Download figure to PowerPoint

Structural Correlations With Biomechanical Properties

Mineral content correlated clearly with weight, exhibiting an exponential function (R2 = 0.97, p < 0.01) (Fig. 6a). When mineralized and decalcified matrices were compared, the energy needed to break the mineralized femur and to tear the decalcified collagen network of the tibia correlated significantly (R2 = 0.63, p < 0.01) (Fig. 6b). Collagen content explained the modulus of the decalcified bone matrix (R2 = 0.62, p < 0.01) (Fig. 6c), whereas the amount of enzymatic cross-links, HP, (R2 = 0.44, p < 0.01) or LP (R2 = 0.42, p < 0.01) correlated with ultimate strain (Fig. 6d).

thumbnail image

Figure 6. BMD was correlated with the animal's weight (a). Energy obtained from testing the collagen matrix and the intact bone, respectively, showed high correlation (b). The collagen content explained the modulus of the decalcified tissue (c) and the amount of enzymatic cross-links (HP) correlated with the ultimate strain of the decalcified matrix (d) (*p < 0.05, **p < 0.01).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We found significant changes in the biochemical and mechanical properties of bone during maturation and growth of female rabbits. Some changes occurred over the lifetime, such as the increase in BMD and non-enzymatic PEN cross-links, but the most significant changes took place before the bone tissue reached maturity. Collagen content increased up to 3 months of age. Most mechanical characteristics of the collagen network also reached their mature properties at about 3 months. However, the force needed to break the collagen network increased further until 6 months, most likely due to continued bone growth. This coincided with a continued increase in the numbers of collagen cross-links, indicating that the cross-links augment the stability and mechanical strength of the collagen network. Conversely, mineralized bone tissue demonstrated continued mineral deposition, increase in cortical thickness, and enhanced bone mechanical properties up to at least 6 months of age (Fig. 7). Knott and Bailey12 proposed that the mineralization status of bone depends on the maturation of the collagen network and that post-translational modifications of the network play a key role in mineralization. Wassen et al.5 demonstrated that the collagen network structure directs the mineralization process. Others have proposed that mineralization prevents further cross-linking.28, 29 These findings are consistent with our results, where the mineralization followed the development of the collagen network and continued to develop even after the collagen content reached its mature levels. Our study clearly indicates that the compositional and mechanical properties of bone change most during the early dynamic and rapid period of growth, that is, the first 3–6 months of life in the rabbit. Weight gain and increase of physical activity are most likely stimuli that determine this process. Withholding physical activity during early life may affect the development process, and hence, the biochemical composition of the bone matrix.6, 19

thumbnail image

Figure 7. Comparison between the development of the collagen content and the mineral density in the mid-diaphysis of the femur showed that the collagen content reached its maximal levels at 3 months of age, whereas mineral content continued to increase at least until 6 months. Results are expressed as mean ± SD (*p < 0.05, **p < 0.001).

Download figure to PowerPoint

We investigated bone properties in both the femur and tibia. Timing of the development of certain properties can be different in these two long bones. For example, studies of the growth plates in rabbits showed that the bone growth rate in the tibia is higher.8 Furthermore, articular cartilage in the rabbit tibia reaches its mature composition and structure before femoral cartilage.30 However, we measured both BMD and collagen content at the same location (mid-diaphysis of the femur). The comparison was based on different, but standard techniques. Mineral content measured by pQCT correlates closely with mineral content based on non-destructive neutron activation analysis and flame atomic absorption spectrometry,31 and with ash weight.32 Nonetheless, comparison between these techniques must be done with care. Therefore, only a comparison of the trends with age was conducted, showing that collagen content reached its maximal levels before the mineralized matrix (Fig. 7).

Bone mineral content correlated with the weights of the animals (Fig. 6a). Previous studies showed linear correlations between these parameters,6, 33 but did not include newborn animals. The energy needed to tear the decalcified collagen network of the tibia was about two times higher than that measured from the 3-point bend tests of intact bone (Fig. 6b). This confirms previous studies suggesting that the collagen network is responsible for storing energy in the bone and that the mineralized matrix decreases the capability to store energy.3, 34 We found a relationship between the collagen content and the modulus of the decalcified tibia bone matrix (Fig. 6c). Moreover, the amount of enzymatic cross-links, HP and LP, correlated with ultimate strain (Fig. 6d). This supports earlier findings in mice, which indicated that the collagen network's ultimate displacement was associated with the amount of LP cross-links.17 However, we did not determine other cross-links in connective tissues, such as pyrrolic cross-links, MODIC, GODIC, DOGDIC, or glucosepane.11, 35, 36

A limitation of our study is that due to the highly varying size of the specimens, all variables could not be kept constant between age groups during mechanical testing (Fig. 1). However, the parameters were set to achieve similar relative rates and span length for each age group and specimen. We believe this is also the strength of our study, since we know of no previous study in which mechanical and compositional properties were measured over the entire life span of the rabbit. Furthermore, mechanical testing was conducted at room temperature. The mechanical properties of bone are influenced by temperature; for accurate measurements, specimens should be tested at 37°C. However, this is not always practical. Previous studies showed that testing at room temperature increased the modulus of bone by about 2–4% compared to testing at 37°C.26 Thus, the error caused by testing at room temperature is not large, and any temperature dependence would be predicted to influence each experimental group equally.

Previous studies in mice described comparable compositional and mechanical developmental curves.6, 17 However, those studies did not include newborns. Studies on equine bone showed that the biochemical and structural modifications of the collagen network primarily take place during the first 6 months post-partum.4, 18, 19 Accounting for variation in timing of skeletal maturity, our results generally agree with these previous findings. However, due to the choice of experimental time points in the previous studies, the time points when the collagen and mineral content reached their respective mature levels were not established.

All rabbits used in our study were females. Evidence exists that gender affects cortical bone growth rate and adaptation to exercise.37 However, we cannot draw any conclusions on gender effects. The rabbits in our study become skeletally mature between 6–9 months of age, in line with a recent report that demonstrated negligible bone growth in rabbits after 6–8 months, while articular cartilage showed mature structure by 3–4 months.38 It was evident that the bone's capacity to store energy had decreased at 18 months (Figs. 4c and 5c). This may be due to decreased viscoelasticity of the collagen matrix with age,34 which may be related to an increase in PEN concentration.12, 14, 17 Due to the non-enzymatic character, PEN cross-links reflect the relative tissue age and matrix turnover. The decrease of PEN concentration during the early time points, expressed per collagen, was most likely due to the rapid increase in collagen content during early age. With age PEN cross-links increase as synthesis and turnover of collagen are reduced. When the animals get older, the cross-links increase because collagen turnover is low. This is supported by our findings (Fig. 2d) and in earlier studies in mice.6, 17

We conclude that profound biochemical and biomechanical changes take place during the early process of maturation and growth in both the mineralized and collagen matrix of rabbit cortical bone. Our study demonstrates that collagen content and biomechanical characteristics of the collagen network reached a mature state at or after 3 months of age in the rabbit, whereas the mineralized matrix reached its mature levels at or after 6 months of age.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

None of the authors have any conflicts of interest. Financial support was received from the Ministry of Education, Finland (25/627/2006 and 61/627/2005), Academy of Finland (200970, 113112, and 110595), and the European Commission (BONEQUAL—219980).

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES