| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Arthritis and Metabolic Bone Disease Research Unit and Laboratory for Biomechanics (G.L.), B-3212 Leuven, Belgium
Address all correspondence and requests for reprints to: Dr. J. Dequeker, Arthritis and Metabolic Bone Disease Research Unit, UZ Pellenberg, Weligerveld 1, B-3212 Pellenberg, Belgium. E-mail: jan.dequeker{at}uz.kuleuven.ac.be
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although over the past 10 yr significant advances have been made in the biology of bone cells, especially because of the technical improvements of in vitro systems of bone cell cultures, it has also become clear that such in vitro systems oversimplify the in vivo situation. The study of animal models will therefore remain an important part of research to address the pathogenesis of bone disorders and/or the effects of newly developed drugs on bone metabolism. Moreover, before accepting any agent for clinical osteoporosis trials, the Food and Drug Administration (FDA) now requires its efficacy in at least one small (rodent) animal model and one larger animal with known intracortical bone remodeling (1).
The rat is by far the most commonly used animal in bone research because of a number of practical advantages: its low cost and ease of accomodation and care, its short lifespan that enables studies of age-related changes, its well defined genetic background, and its thoroughly documented mineral metabolism (1, 2, 3, 4, 5). Several rat models have been used, including the growing rat (6, 7), the aged rat (8, 9), the mature ovariectomized rat (2, 3), and the aged ovariectomized rat (10). As an alternative to the rat, other small animals, such as the guinea pig (11) or the mouse (12), have been proposed. The choice of an appropriate larger animal model is usually much more difficult because the suitability of each model depends largely upon the objectives of the study. There is no ideal model for the study of all aspects of a metabolic bone disorder such as osteoporosis; all have advantages and disadvantages that should be taken into account when selecting animal models for the study of a specific metabolic bone disorder (13, 14, 15). A nonlimited list of larger animals that are being used in bone research includes dogs (16, 17), sheep (15), (mini)pigs (18), poultry (19), dairy cows (20), and nonhuman primates (21, 22).
The suitability of an animal model in bone research is closely related to the degree of similarity between the results obtained in the model compared with the human situation. After all, the ultimate goal of the use of an animal model is to transfer the obtained results to clinical practice. Differential responses in animals compared with humans can be caused by many different factors, acting directly or indirectly on the bone metabolism. Usually, animal studies aim to evaluate the impact of a specific treatment or therapy during an experimental period on a predefined set of parameters relevant in bone research. However, little attention has focused on how the baseline values of these parameters in animal models compare with humans; yet this may largely influence whether or not the results of a study may also apply to the clinical situation. In this regard, the objective of the present study was to document differences in bone mineral density, mechanical competence, and biochemical composition in bone specimens from different species, including human samples.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Bone sample preparation
Cortical bone powder samples were prepared from femoral bone
specimens, according to previously established methods (7, 23), before
analysis of their biochemical composition. Briefly, the bones were
thoroughly cleaned from periosteum, associated soft tissue, and bone
marrow. The femoral shafts were defatted for 2 days in
trichloroethylene (renewed once daily), washed in distilled water, and
dried. Bone was pulverized with a beater mill cooled with liquid
nitrogen until bone powder with particle size between 40 and 160 µm
was retained. For density fractionation analysis (see below), bone
powder with a particle size smaller than 20 µm was prepared.
From the lumbar spine specimens, one (or two, if possible) cylindrical core(s) of trabecular bone (8.2 mm diameter) was taken in the axial direction of the specimens, using a core drill. From these samples, the end plates were cut plane-parallel, using a low-speed diamond saw, such that the height of the cylinders was exactly 10 mm. The obtained cylindrical cores with fixed dimensions were stored in PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, 1.4 mM KH2PO4, pH 7.3) at 4 C until bone mineral density measurements and mechanical testing were done. The remaining trabecular bone parts of the lumbar spine samples were collected immediately after the isolation of the cylindrical core(s) and treated as described above for the cortical femoral bone samples, and chemical composition was analyzed.
Bone mass measurement
Bone mineral content (BMC) and projected area (area) of the
trabecular fixed-size cylinders were measured by dual-energy x-ray
absorptiometry, using a quantitative digital radiography device
(QDR-1000/W, Hologic, Waltham, MA) that was calibrated with a Hologic
hydroxyapatite anthropometric spine phantom. The specimens were
scanned, with standard collimation of the x-ray beam, in ultrahigh
resolution mode and manufacturer-supplied software (Hologic, V4.47),
using 140/70 keV, average 3 mA, a scan step of 0.5 mm, and point
resolution of 0.5 mm. All specimens were scanned in two positions.
First, they were positioned with the cylinder axis perpendicular to the
scanning table (axial direction); second, their axis was parallel to
the table (lateral direction). Reproducibility, as coefficient of
variation (CV) from five measurements of the same bone specimen after
repositioning, was 0.38% (BMC) and 0.8% (area) for lateral
measurements, respectively, and 0.58% (BMC) and 0.45% (area) for
measurements in the axial direction through the cylinder. The BMC was
taken as the average value obtained for these two scanning positions.
The exact height of the core samples was recorded using a micrometer
with a readout up to 0.01 mm. From these data, the volumetric bone
mineral density (vBMD) was then calculated by dividing the BMC (mg)
through the volume of the core (calculated by multiplication of the
circular area measured in axial direction with the measured core
height).
Mechanical analysis
A compression test was made between the two parallel surfaces of
the wet cylindrical cores, as described earlier (24). Briefly, the
specimens were placed between flat platens, and a downward displacement
of the upper platen produced compression loading. The force was applied
by means of a hydraulic piston, moving downward at high speed (about
0.12 meters per second). The force was measured using a load cell
(Lebow model 3157, Philadelphia, PA) and the deformation of the sample
(displacement of the piston) with a linear extensometer (Philips
LO1314, Eindhoven, The Netherlands). The maximum force was determined
as the point at which the derivative of the curve force vs.
displacement is equal to zero. The maximum force was divided by the
area of cross-section of the samples, to yield the fracture stress.
Density fractionation of bone powder
The density fractionation by stepwise centrifugation method was
performed as described by Grynpas et al. (21), using the
cortical bone powder samples. Sieved bone powder, 200–300 mg (particle
size <20 µm), was added to a tube containing 35 ml of a 2.0 g/ml
density solution made up by a bromoform-toluene mixture and calibrated
with sink floats (Cargille Laboratories Inc, Cedar Grove, NJ). After
centrifugation of the solution for 30 min at 10,000 rpm (JA20, Beckman
Instruments, Fullerton, CA), the supernatant was separated from the
pellet and transferred to a new tube. The density of the supernatant
was modified to 1.9 g/ml by the addition of toluene and recentrifuged.
Under the same conditions, each precipitate obtained from solutions of
progressively decreasing density (at steps of 0.1 g/ml) was collected.
To obtain a range of mineral density greater than 2 g/ml, the
precipitate obtained from the initial 2 g/ml density solution was
resuspended in a solution of 2.3 g/ml. Successive centrifugation of
precipitates at progressively decreasing densities (steps of 0.1 g/ml)
provided the higher density fractions. The finally obtained series of
specific gravity fractions were centrifuged in 100% ethanol to remove
bromophenol and toluene residues, dried, and weighed. From weight
determinations of each fraction, the percent distribution of the
various bone density fractions in a sample of unfractionated bone
powder was calculated to generate mineralization profiles.
Ash weight and collagen analysis
Small amounts (
10 mg) of each bone powder sample were ashed
at 700 C for 6 h in a muffled furnace. The ashed segment was
weighed and expressed as percentage of dry bone weight. Hydroxyproline
concentration, as a marker for collagen content, was measured by the
method of Kivirikko et al. (25) and was used previously in
our laboratory for the analysis of bone samples (7, 26).
Analysis of the total amount of extractable proteins and
insulin-like growth factor-I (IGF-I)
Based on previous experience (7, 23, 26), the noncollagenous
proteins were extracted from the bone samples as described below.
Approximately 15 mg of each dry bone powder sample were extracted
subsequently with 1.5 ml of 0.5 M ammonium-EDTA (pH 8.0)
and 4 M guanidinium-HCl (pH 7.4). Each of the
extraction solutions contained a mixture of protease inhibitors (5
mM benzamidine, 10 mM 6-aminocaproic acid, 100
µM p-hydroxymercuribenzoic acid, pH 6.2) at 4
C. The extractions were carried out overnight in microcentrifuge tubes
by end-over-end rotation. After 18 h the solution was centrifuged
(12,000 rpm for 30 min), the supernatant containing noncollagenous
proteins was separated from the collagenous residue and desalted on a
Sephadex PD-10 column (Pharmacia, Uppsala, Sweden). Both extracts (with
EDTA and guanidinium-HCl, respectively) were analyzed separately; the
presented data represent the sum of the values found in both extracts.
The total protein concentration in the extracts was determined by
spectrophotometry (
= 280 nm) using BSA as standard. Before IGF-I
determination the desalted extracts were lyophilized in a Speed Vac
Concentrator (Heto, Allerod, Denmark) and dissolved in assay buffer.
IGF-I was determined by a RIA as previously outlined in detail (27).
Briefly, recombinant human IGF-I was used as standard, and a polyclonal
antiserum was raised in a guinea pig. The antiserum displayed less than
0.01% cross-reaction with insulin and less than 1% cross-reaction
with IGF-II. Inter- and intraassay coefficients of variation of the RIA
were 7.7 and 7.4%, respectively, with a detection limit of 1.3 ng
IGF-I/tube. The analytical recovery was 100 ± 2% (mean ±
SEM, n = 6). Interference with IGF-binding proteins
was evaluated by column chromatography (7). Different bone powder
extracts were spiked with radioactive labeled IGF-I and eluted on a
superose 12 column (Pharmacia). A single peak was observed, indicating
the lack of interference between IGF-I and its binding proteins in
these samples.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
.
Large variations were observed between the different species for all
examined parameters. Ash content was similar in human, dog, and pig
cortical bone, intermediate in cow, sheep, and chicken, and highest in
the rat. In five of the seven species, the collagen content, measured
by hydroxyproline, was inversely related to the degree of
mineralization. The femoral samples from pig and sheep, however, showed
lower collagen concentrations than expected, but this was compensated
for by a larger amount of extractable noncollagenous proteins compared
with the other species. Rat bone also showed a high concentration of
extractable proteins, whereas human and cow samples had the lowest
concentration. These extracts of the bone powder samples were analyzed
for IGF-I, a growth factor known to be involved in bone metabolism and
produced by osteoblasts, and showing a well conserved sequence across
species. IGF-I concentrations were highest in rat bone, intermediate in
human and dog bone, and very low in the other examined species.
|
). A shift
from the bone particles to higher density fractions was observed in the
species with a higher ash content in the unfractionated bone powder
samples. The order of the species from lowest to highest proportion of
high density bone particles was as follows: human, dog, pig, chicken,
cow, sheep, and rat. The chemical composition of the two main bone
particle fractions (density 2.0–2.1 and 2.1–2.2 g/ml, respectively)
was further analyzed, and the results are presented in Table 1. An increase of ash and a decrease of
collagen concentration were observed with increasing density in all
species. Within a specific density range, significant variations of ash
and collagen were observed between the species, similar to the
variations in the whole bone samples (Fig. 1
). This indicates that
species differences of bone composition in the whole bone powder
samples are found in all the bone particles of the sample and are not
solely the result of a relative shift of a part of the bone particles
to a higher/lower density. The total amount of extractable proteins was
in general higher in the lower mineralized density fraction (2.0–2.1
g/ml), but the concentration varied significantly among species (Table 1). Bone powder extracts from rats, however, showed a higher
concentration of extractable proteins in the higher mineralized density
fraction (2.1–2.2 g/ml). The IGF-I did not show large variations among
the different density fractions of a species; a similar interspecies
variation was found as in the unfractionated bone powder samples.
|
|
. In general, much less variation was
found among the examined species with regard to trabecular bone
composition compared with the compact bone samples (Fig. 1). The ash
concentrations in dog, pig, and sheep were very similar to the human
samples; only the ash content of the samples derived from the cow was
somewhat lower compared with the other samples. Inversely, collagen
concentration was higher in the cow samples compared with the other
species and was somewhat lower in pig and sheep. The cow showed the
largest contrast between cortical and trabecular bone samples: a
relative high ash content in cortical bone was combined with a relative
low ash content in the trabecular bone samples. In all species, the
amount of extractable proteins was significantly higher (on average, a
doubling of the amount) in the lumbar spine bone powder samples
compared with the cortical samples derived from the femur. The relative
differences between the species was similar, however, in the trabecular
and the cortical bone samples. Only the trabecular samples derived from
the cow showed a higher-than-expected amount of extractable proteins
based on the results from the cortical samples. Although IGF-I
contributes to the large pool of extractable proteins, large
differences between trabecular and cortical bone samples were observed
with regard to the IGF-I concentration.
|
. Both the BMC and vBMD values were
approximately 50% lower in the human samples compared with the samples
from the dog and pig. The BMC and vBMD in dog and pig were in turn much
lower in comparison with cow and sheep. From the raw data of the
mechanical tests, the fracture stress was derived. Fracture stress was
lowest in the samples from human and pig, intermediate in dog and cow
samples, and highest in sheep samples. The relation between fracture
stress and mineral mass (BMC) and density (vBMD), respectively, in the
trabecular bone samples derived from different species is visualized in
Fig. 4. This figure illustrates that each
of the examined species forms a separate entity with no or only limited
overlaps with the data from other species. It suggests, furthermore,
that the relative contribution of BMC and/or vBMD measurements to the
bone quality, measured as fracture stress, is likely to be
species-dependent.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Biochemical analysis of the trabecular and cortical bone of the included species reveals several interesting points. First, significant interspecies differences are present for each of the examined bone composition parameters, both in cortical and trabecular bone. An early study by Biltz and Pellegrino (31) reported on the biochemical analysis of a limited number of cortical bone samples, derived from a variety of different species. The data on ash and hydroxyproline content in these samples were comparable with the data of the present study. The composition of human and rat cortical bone has also been investigated by Mbuyi and Dequeker (32), who compared the collagen content and the content of some nonspecific markers for noncollagenous proteins, such as sialic acid, uronic acid, and hexoses. Higher collagen and lower EDTA-extractable protein contents were found in human compared with rat bone. This was confirmed by the results of the present study, which extends the biochemical analyses of the latter two studies to the analysis of several other species currently used in bone research, to the simultaneous analysis of trabecular bone, and to the analysis of more specific parameters. However, the physiological and phylogenetic implications of the species variations in bone composition remain to be determined. Second, the density fractionation analyses of the cortical bone samples indicate that both a shift of a portion of the microscopic bone particles toward another density and composition, as well as compositional differences within a specific density range, may contribute to the interspecies differences observed in the samples of whole bone. These results support the hypothesis that the material properties of the bone particles are highly species-dependent and might contribute to interspecies differences in bone quality characteristics. It is well recognized that, in addition to bone mass and bone architecture, bone composition contributes to bone quality properties (34). Third, it is clear that bone composition in some species more closely resembles human bone composition than others. In particular, it should be noted that, of all species examined, the rat differs most, whereas the bone composition of the dog most resembles that of human bone. This observation supports the requirement of the FDA that evaluation of drugs should be performed, in addition to the rat, in at least one larger animal. Our results support the recommendation by Rodgers et al. (28) to use the relatively inexpensive and well characterized rat model for preliminary screenings of new pharmacological agents or therapeutic modalities, followed by verification in other species, before undertaking clinical trials in human patients.
Results of bone density measurements and mechanical testing also show significant interspecies differences. Human bone specimens constitute a clearly separated entity compared with all other species, both with regard to BMC and vBMD analyses. It is, however, speculative to ascribe the higher fracture incidence in humans compared with other species to the fact that humans show the lowest BMC, vBMD, and fracture stress values. Based on these analyses, the pig and dog most resemble the human situation. A positive intraspecies correlation between bone density and fracture stress was observed. In general, this correlation was also found at the interspecies level. However, comparison of individual species does not necessarily follow this correlation. Fracture stress in dogs is higher compared with pigs, although similar BMC and vBMD values are observed. Similarly, sheep show higher fracture stress compared with cows, but bone density measurement data are comparable. This suggests that the relative contribution of BMC and/or vBMD data to the bone quality, measured as fracture stress, is species-dependent. Whereas BMC and BMD measurements in humans have been shown to predict up to 70% of bone strength (35), it is clear from our data that such a relationship cannot simply be transferred to the situation in animal models or vice versa.
The present analysis has several limitations. The number of animals per species was very limited; yet the intraspecies variation was much lower compared with the interspecies variation. The age and sex of the animals may also influence the results. To minimize the bias of these confounders, we used sexually mature female animals as is frequently done in bone studies. Other intraspecies differences, e.g. animal strain, are difficult to correct for, unless multiple breeds per species are included in the study. This was, however, beyond the objectives of this study. Finally, a limiting factor of our study might be found in the choice of the examined anatomical sites. Although bone samples from the same anatomical sites (femur and lumbar spine) were analyzed for all species examined, it cannot be excluded that at least part of the observed differences may be attributed to differences in load-bearing of the lumbar spine and the femur. From among these species, humans are the only ones to walk in an upright position, and it could be expected that, by evolution, this leads to differences in architectural and mechanical bone properties. Indeed, the human lumbar spine is clearly different from all other examined species, in that its bone density and fracture stress are much lower compared with the other species. It would be of interest to document whether similar characteristics of the lumbar spine are also observed in other upright walking species, such as nonhuman primates.
It seems unlikely that the large interspecies differences reported here can all be attributed to the above-mentioned confounding factors alone. The fact that all the analyses described in this study have been performed in a single laboratory using the same analytical procedures for all samples strengthens the present interspecies comparison. Nevertheless, it is impossible to exclude the impact of all these confounders on the results from the real genetic interspecies differences. Moreover, the small number of animals per species does not allow statistical analyses. The presented data should therefore be interpreted as relative indications of qualitative differences only, and not as absolute reference values for further studies.
In summary, this study shows marked interspecies differences with regard to bone composition, bone density, and bone mechanical competence. None of these animal models is similar to the human situation for all examined parameters. Some animals, however, more closely resemble humans than others. In particular, with regard to bone density and quality parameters of the lumbar spine, humans appear to be very different from the other species examined. Based on a combination of all the parameters examined in the animals of our study, we conclude that the characteristics of human bone are best approximated by the properties of dog bone. The large biochemical differences in bone composition in the rat and human indicate that bone research data derived from this most frequently used animal model should be transferred to the clinical situation with utmost care. It would be speculative to predict whether the human bone response to a treatment will also be best resembled by the dog model. Moreover, other factors (e.g. physiology, endocrinology) that were not evaluated in this study also influence bone response. It is obvious that the interspecies variations that are demonstrated in this study may affect other clinical and experimental bone parameters and should therefore be taken into account when selecting an animal model for bone research.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received August 6, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-vitamin
D3 on bone strength and composition in growing rats with
and without corticosteroid treatment. Calcif Tissue Int 55:443–450[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  A. R. Pagano, K. Yasuda, K. R. Roneker, T. D. Crenshaw, and X. G. Lei Supplemental Escherichia coli Phytase and Strontium Enhance Bone Strength of Young Pigs Fed a Phosphorus-Adequate Diet J. Nutr., July 1, 2007; 137(7): 1795 - 1801. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Barker, T. Takahashi, A. Toms, P. Gregson, and J. H. Kuiper Reconstruction of femoral defects in revision hip surgery: RISK OF FRACTURE AND STEM MIGRATION AFTER IMPACTION BONE GRAFTING J Bone Joint Surg Br, June 1, 2006; 88-B(6): 832 - 836. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. G. R. Garcia, F. L. Mombiela, C. J. de la Fuente, M. G. Aranguez, D. V. Escribano, and J. V. Martin The Influence of the Size and Condition of the Reamers on Bone Temperature During Intramedullary Reaming J. Bone Joint Surg. Am., May 1, 2004; 86(5): 994 - 999. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Ahmad, T. R. Gardner, M. Groh, J. Arnouk, and W. N. Levine Mechanical Properties of Soft Tissue Femoral Fixation Devices for Anterior Cruciate Ligament Reconstruction Am. J. Sports Med., April 1, 2004; 32(3): 635 - 640. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. T. Nurmi, H. Sievanen, P. Kannus, M. Jarvinen, and T. L. N. Jarvinen Porcine Tibia Is a Poor Substitute for Human Cadaver Tibia for Evaluating Interference Screw Fixation Am. J. Sports Med., April 1, 2004; 32(3): 765 - 771. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Verma, M. A. Noerdlinger, N. Hallab, C. A. Bush-Joseph, and B. R. Bach Jr. Effects of Graft Rotation on Initial Biomechanical Failure Characteristics of Bone-Patellar Tendon-Bone Constructs Am. J. Sports Med., September 1, 2003; 31(5): 708 - 713. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Seeherman, R. Li, and J. Wozney A Review of Preclinical Program Development for Evaluating Injectable Carriers for Osteogenic Factors J. Bone Joint Surg. Am., August 1, 2003; 85(90003): 96 - 108. [Full Text] [PDF]
|
|
|
|
|
|
W. Singhatat, K. W. Lawhorn, S. M. Howell, and M. L. Hull How Four Weeks of Implantation Affect the Strength and Stiffness of a Tendon Graft in a Bone Tunnel: A Study of Two Fixation Devices in an Extraarticular Model in Ovine Am. J. Sports Med., July 1, 2002; 30(4): 506 - 513. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. T. Nurmi, T. L. N. Jarvinen, P. Kannus, H. Sievanen, J. Toukosalo, and M. Jarvinen Compaction Versus Extraction Drilling for Fixation of the Hamstring Tendon Graft in Anterior Cruciate Ligament Reconstruction Am. J. Sports Med., March 1, 2002; 30(2): 167 - 173. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J Wilkin, D. Devendra, J. Dequeker, and F. P Luyten For and against: Bone densitometry is not a good predictor of hip fracture For Against BMJ, October 6, 2001; 323(7316): 795 - 799. [Full Text] [PDF]
|
|
|
|
|
|
K. P. Black, M. M. Saunders, K. C. Stube, M. J.R. Moulton, and C. R. Jacobs Effects of Interference Fit Screw Length on Tibial Tunnel Fixation for Anterior Cruciate Ligament Reconstruction Am. J. Sports Med., November 1, 2000; 28(6): 846 - 849. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
