SEARCH

SEARCH BY CITATION

Keywords:

  • vertebra;
  • trabecular bone;
  • quantitative trait locus;
  • mouse;
  • genetics

Abstract

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

BMD, which reflects both cortical and cancellous bone, has been shown to be highly heritable; however, little is known about the specific genetic factors regulating trabecular bone. Genome-wide linkage analysis of vertebral trabecular bone traits in 914 adult female mice from the F2 intercross of C57BL/6J and C3H/HeJ inbred strains revealed a pattern of genetic regulation derived from 13 autosomes, with 5–13 QTLs associated with each of the traits. Ultimately, identification of genes that regulate trabecular bone traits may yield important information regarding mechanisms that regulate mechanical integrity of the skeleton.

Introduction: Both cortical and cancellous bone influence the mechanical integrity of the skeleton, with the relative contribution of each varying with skeletal site. Whereas areal BMD, which reflects both cortical and cancellous bone, has been shown to be highly heritable, little is known about the genetic determinants of trabecular bone density and architecture.

Materials and Methods: To identify heritable determinants of vertebral trabecular bone traits, we evaluated the fifth lumbar vertebra from 914 adult female mice from the F2 intercross of C57BL/6J (B6) and C3H/HeJ (C3H) progenitor strains. High-resolution μCT was used to assess total volume (TV), bone volume (BV), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), separation (Tb.Sp), and number (Tb.N) of the trabecular bone in the vertebral body in the progenitors (n = 8/strain) and female B6C3H-F2 progeny (n = 914). Genomic DNA from F2 progeny was screened for 118 PCR-based markers discriminating B6 and C3H alleles on all 19 autosomes.

Results and Conclusions: Despite having a slightly larger trabecular bone compartment, C3H progenitors had dramatically lower vertebral trabecular BV/TV (−53%) and Tb.N (−40%) and higher Tb.Sp (71%) compared with B6 progenitors (p < 0.001 for all). Genome-wide quantitative trait analysis revealed a pattern of genetic regulation derived from 13 autosomes, with 5–13 quantitative trait loci (QTLs) associated with each of the vertebral trabecular bone traits, exhibiting adjusted LOD scores ranging from 3.1 to 14.4. The variance explained in the F2 population by each of the individual QTL after adjusting for contributions from other QTLs ranged from 0.8% to 5.9%. Taken together, the QTLs explained 22–33% of the variance of the vertebral traits in the F2 population. In conclusion, we observed a complex pattern of genetic regulation for vertebral trabecular bone volume fraction and microarchitecture using the F2 intercross of the C57BL/6J and C3H/HeJ inbred mouse strains and identified a number of QTLs, some of which are distinct from those that were previously identified for total femoral and vertebral BMD. Identification of genes that regulate trabecular bone traits may ultimately yield important information regarding the mechanisms that regulate the acquisition and maintenance of mechanical integrity of the skeleton.


INTRODUCTION

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

OSTEOPOROSIS IS CHARACTERIZED by compromised bone strength leading to an increased risk of fracture.(1) Biomechanical principles dictate that a bone's strength is determined by the amount of material in the bone (i.e., mass), how this material is distributed (i.e., geometry and architecture), and the intrinsic properties of this material.(2) Thus, genetic and environmental factors that influence any of these characteristics will ultimately affect fracture risk.

Family history of fracture is a strong risk factor for future fractures, providing evidence that genetic factors may influence skeletal fragility.(3, 4) Indeed, twin, sib-pair, and intergenerational family studies conducted over the past several decades have established that ∼60-90% of the variance in BMD is attributable to genetic determinants.(5-7) Moreover, there is mounting evidence that genetic factors contribute not only to BMD, but also bone geometry(8, 9) and bone turnover,(10-13) and therefore may influence an individual's susceptibility for fracture through several mechanisms. These genetic epidemiology studies have provided strong motivation to identify specific chromosomal regions and/or genes that may contribute to skeletal characteristics that influence attainment and maintenance of bone strength and skeletal integrity.

Several recent clinical studies have yielded putative chromosomal regions that may harbor genes related to skeletal fragility.(5, 7, 14) However, reliance on human linkage studies alone to identify genes involved in osteoporosis may be inadequate because these studies frequently have low statistical power and are affected adversely by genetic and environmental heterogeneity of the study population.(5, 15) Moreover, to date, the majority of clinical studies have been limited to evaluation of composite phenotypes, such as areal BMD, a measurement that is partially influenced by bone size and does not distinguish between the cortical and trabecular bone compartments. Evaluation of more “proximal” phenotypes, such as bone formation rate, bone geometry, or bone density in distinct trabecular or cortical compartments, may enhance the chance of finding genes for osteoporosis by reducing the potential number of genes and exogenous factors involved and by clarifying genotype-phenotype associations.(5, 16) It has therefore been suggested that animal models may be useful for initial identification of candidate genes for subsequent in depth study in human populations.(17)

In this regard, inbred mouse strains have been successfully used to identify genes that play important roles in number of complex human diseases(18) and have recently been established as a valuable tool for studying the genetic regulation of skeletal phenotypes.(19, 20) Studies in mice offer a number of advantages, including the ability to generate large numbers of genetically identical subjects for study, controlled breeding with relatively short generation times, and standardized environmental conditions. Skeletal phenotypes in mice include those that are analogous to measurements made in human populations, such as BMD, bone size, and bone geometry. Thus, quantitative trait loci (QTL) for whole body and spine areal BMD,(21-25) total femoral and vertebral volumetric BMD,(26, 27) cortical thickness,(25, 28, 29) and skeletal size(17, 29-33) have been identified in mice. Another advantage afforded by studies in mice is the ability to assess phenotypes that require invasive or destructive testing techniques. Accordingly, QTLs for femoral strength have been identified.(34, 35) However, the phenotypes studied to date (in both mice and humans) reflect primarily cortical bone characteristics (i.e., femoral BMD, geometry, and strength measurements) or a composite of cortical and trabecular bone (i.e., areal BMD measurements). Thus, few studies, either in mice or humans, have evaluated whether there is evidence for genetic regulation of trabecular bone, an important contributor to bone strength at common fracture sites such as vertebral body and proximal femur.(36-39)

Therefore, our overall objective was to evaluate heritable determinants of vertebral trabecular bone volume fraction and microarchitecture using two inbred mouse strains, C57BL/6J (B6) and C3H/HeJ (C3H), of similar body size but with established differences in several skeletal characteristics.(40-49) We previously identified several chromosomal regions associated with total femoral and vertebral volumetric BMD (vBMD) in female progeny from a B6C3H F2 intercross(27) and recently verified one of these regions by generation of congenic sublines.(19) Thus, a second objective of our study was to test whether the genetic determinants of vertebral trabecular bone traits would be, in part, different from those previously identified for total femoral and vertebral vBMD.

MATERIALS AND METHODS

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

Animals

The study used two inbred strains of mice, C57BL/6J (B6) and C3H/HeJ (C3H), previously shown to differ widely in total femur vBMD.(40) Mice were produced and maintained in a research colony at the Jackson Laboratory under 14:10 h light:dark cycles. Females were group-housed in polycarbonate cages (51 in2, 3-5 mice/group). Water was acidified with HCl to achieve a pH of 2.8-3.2 (to prevent bacterial growth) and was freely available. All mice were fed autoclaved National Institutes of Health 31 (4% fat, 18% protein, with 1.20% Ca, 0.85% P; Purina Mills International, Richmond, IN, USA) ad libitum. Use of mice in this research project was reviewed and approved by the Institutional Animal Care and Use Committee of The Jackson Laboratory.

Progeny for genetic analysis of vertebral bone density and microarchitecture were produced by mating B6 females (low total femoral vBMD) to C3H males (high total femoral vBMD) and intercrossing this B6C3F1 hybrid to produce F2 offspring.(27) A total of 1012 F2 females were raised, 986 were used for genetic assessment of total femoral vBMD by pQCT, and 914 of these provided L5 vertebrae suitable for assessment by μCT. The progenitors and F2 females were analyzed at 4 months of age, after peak total femur vBMD had been achieved.(40) Kidneys and spleens from each mouse were frozen in liquid nitrogen and stored at −60°C for later extraction of genomic DNA. The F2 males were not kept because of losses caused by aggressive behavior when group housed. At death, body weight was recorded, and the fifth lumbar vertebral body was excised and placed in 95% ethanol, as previously described.(26)

Assessment of vertebral morphometry by μCT

The intact fifth lumbar vertebrae were assessed using a desktop microtomographic imaging system (μCT20; Scanco Medical, Basserdorf, Switzerland) that employs a microfocus X-ray tube (10-μm focal spot) as a source.(50) The entire vertebra was scanned using a slice increment of 17 μm. CT images were reconstructed in 1024 × 1024 pixel matrices using a standard convolution-backprojection procedure with a Shepp and Logan filter. Images were stored in 3D arrays with an isotropic voxel size of 17 μm. The resulting grayscale images were segmented using a constrained 3D Gaussian filter (σ = 1.2, support = 1) to remove noise, and a fixed threshold (22% of maximal gray scale value) was used to extract the structure of mineralized tissue. A single operator outlined the trabecular bone region within the vertebral body for each CT slice, excluding both the cranial and caudal growth plate regions. Within this region, morphometric variables were computed from the binarized images using direct, three-dimensional techniques that do not rely on any prior assumptions about the underlying structure.(51, 52) For this region, we computed the total volume (TV, mm3), bone volume (BV, mm3), bone volume fraction (BV/TV, %), trabecular thickness (Tb.Th, μm), number (Tb.N, mm−1), and separation (Tb.Sp, μm). The structure model index (SMI) and connectivity density (Conn.D, mm−3) were also computed, although they were not subjected to quantitative trait analysis.(53, 54) SMI quantifies the plate-versus rod-like nature of the cancellous bone such that structures that are purely rod-like have an SMI of 3, whereas those that are purely plate-like have an SMI of 0.(54) The CV (SD/mean) in the progenitor strains was 8.8-11.6% for BV/TV of the cancellous bone in the vertebral body, 4.3% for Tb.Th, 13-16% for Tb.N, 14-16% for Tb.Sp, and 15-40% for ConnD. The reproducibility of the vertebral analysis was estimated by computing the mean CV for repeat measurements (six vertebrae were scanned three times each with repositioning between each scan). The CV for repeat measurements was 3.3% for BV/TV of vertebral cancellous bone (range, 1.7-2.4%), and 2.0% for Tb.Th, 5.4% for Tb.N, and 3.8% for Tb.Sp. After exclusion of specimens with broken pedicles and/or damaged vertebral bodies, complete μCT data were available for 914 specimens.

Genetic analyses

Preparation of genomic DNA and PCR reaction conditions were previously described by Beamer et al.(26, 27) Essentially, genotyping of individual mouse DNAs was accomplished by PCR using oligonucleotide primer pairs that amplify simple CA repeated sequences of anonymous genomic DNA that are of different length and that can uniquely discriminate between B6 and C3H genomes. Primer pairs identifying simple sequence length polymorphisms between B6 and C3H were selected from more than 6000+ available.(55) PCR products from B6, C3H, and (B6 × C3H) F1 hybrids were used as electrophoretic standards in every gel to identify the genotypes of F2 mice (i.e., homozygous B6 [b6/b6] or C3H [c3/c3] and heterozygous [b6/c3]). Genotype data were available for 986 of the F2 progeny.

The F2 progeny were tested for correlations of vertebral characteristics with segregation of 118 PCR-based simple sequence length polymorphic markers on the 19 autosomes. Four to nine polymorphic DNA markers, spaced at ∼13- to 14-cM intervals from centromere to telomere, were selected for each autosome. We have previously shown that the 15-cM genetic distance is capable of detecting major loci for quantitative traits in this experimental design, given the large F2 population size.(27, 56) Chromosome X alleles were not assessed because reciprocal F1 × F1 matings would be required to yield all possible allelic combinations necessary for genetic evaluation of the BMD phenotype in females.

Statistical analyses

Standard descriptive statistics were computed for all μCT variables. Characteristics of each progenitor strain were compared using unpaired Students t-tests, and differences were judged statistically significant when p < 0.05. Correlations among vertebral traits within the F2 population were estimated by Pearson correlation coefficients.

QTL selection strategy

We used a three stage strategy to identify and select the QTLs reported here. The first stage uses standard (single locus) genome scans to identify QTLs that display significant main effects. In the second stage, we carried out a search for significant pairs of loci. This involves an exhaustive search through all possible pairs of loci. Each pair is assessed for both additive and epistatic effects on the phenotype values. It is possible to identify significant pairs of loci with additive effects that were missed in the single locus genome scans, but the primary goal of the pairwise search is to identify epistatic QTLs. (We note in passing that no epistasis was detected in this study.) In the third stage, we collect all of the significant loci and any epistatic interactions identified in the genome scans and construct a multiple QTL model to describe the simultaneous effects of all QTLs on the phenotype. The method used here is analogous to multiple interval mapping,(57) but it was implemented using the algorithm described in Sen and Churchill.(58) We refined the multiple QTL model at this stage using a backward elimination approach. QTLs may be removed if they fail to achieve significance after the effects of other QTL have been taken into account. This is a conservative approach as only those QTLs that are detected in the genome scans and retained in the final model are reported. Further details of each stage in the analysis process are provided below.

We wish to make a note regarding the two distinct types of LOD scores reported in this study. Single locus genome scans are widely used and provide an established standard for detecting QTLs. The LOD scores obtained from these scans are unadjusted in the sense that the effects of any other QTL are not accounted for their calculation. When we consider the simultaneous effects of multiple QTLs, an adjusted LOD score is obtained. The adjusted LOD score takes into account the effects of other QTLs and may be higher or lower than the unadjusted LOD score. The difference is a matter of chance and is determined by the particular pattern of QTL segregation that occurred in the cross. The adjusted LOD score provides a more realistic assessment of QTL effects and has been advocated by others.(59) The LOD scores reported in Table 3 are adjusted LOD scores, and the genome scans (Fig. 3) used for initial detection of QTLs show unadjusted LOD scores.

thumbnail image

Figure FIG. 3.. Results of a genome-wide scan for the association between vertebral trabecular traits and molecular markers on each of the 19 autosomes: (A) total volume, (B) bone volume, (C) bone volume fraction, (D) trabecular thickness, and (E) trabecular number. The unadjusted, single-QTL LOD score is depicted on the y axis. The centromere is located at the left of the axis.

Download figure to PowerPoint

Table Table 3.. Adjusted LOD Scores, Percent Variance Explained in F2 Population (Shown in Parentheses), and Major Allele Effects With Designation for an Additive (Add) or Dominant (Dom) Inheritance Pattern*
Thumbnail image of

Genome scans

Genome scans were performed using the multiple imputation method of Sen and Churchill.(58) We used a 5-cM grid and 128 imputations to perform the interval mapping computations. The central idea behind imputation is as follows. We construct a grid of loci, called pseudomarkers, at 5-cM steps across the genome. Then, using the information in the observed marker genotypes, we impute genotype values at the pseudomarkers. To account for uncertainty in the imputed genotypes the process was repeated 128 times, called multiple imputations. The advantage of this approach is that, after the imputations have been performed, simple linear regression methods can be applied to each set of imputated genotypes. The results are combined across the multiple imputations to obtain an accurate approximation to the LOD score. We used this approach to compute both the single locus and the locus pair genome scans.

Significance thresholds for the genome scans were established by permutation analysis and take into account the multiple testing that is implicit in a genome-wide search. For single locus scans, all loci that cleared the 5% genome-wide threshold (3.0 based on 1000 permutations) were selected. For the pairwise scans, all locus pairs that cleared the 5% genome-wide threshold (8.7 based on 1000 imputations) were subjected to secondary tests.(60) We did the permutation test for the locus pair scans on a 5-cM grid with 16 imputations (instead of 128 used for the locus pair scans). This is a conservative approach because it leads to a higher permutation test LOD threshold than we would obtain using 128 imputations. We first tested the interaction component of the two locus model using a 0.0001 nominal significance level. This corresponds to a 5% genome-wide adjusted threshold. The stringency required to account for multiple testing in the search for epistatic interactions may have contributed to their apparent absence in this study. To determine if both loci in a pair were acting additively, we conducted tests for each of the loci in any pair after accounting for the effect of the other locus. We used a nominal p value of 0.002, corresponding to the single locus genome-wide threshold for these tests.

Multiple QTL modeling

All loci and locus pairs selected in the previous step were analyzed simultaneously in a multiple-QTL model.(59) Each QTL locus was represented by the nearest typed marker to the selected locus. Then each term in the model was dropped, and the change in the LOD score was noted. This corresponds to a type III ANOVA sums of squares, also known as adjusted sums of squares. If the nominal p value of the change in the LOD score was greater than 0.002 for the main effects and greater than 0.0001 for the interactions, that term was dropped. This backward elimination procedure(61) was continued until no additional terms could be dropped.

Note that because of the known limitations of using ratios in statistical analyses,(62) the genetic analyses for BV/TV were performed using BV-adjusted-for-TV (by linear regression). Also note that the LOD scores obtained from the 1D genome scans are obtained by fitting a one-QTL model to the data at a particular locus and comparing it to a null model. The LOD scores in Table 3 are adjusted LOD scores obtained from the type III ANOVA analysis in the backward elimination step above. They reflect the contribution of a locus adjusting for all other terms in the final multiple-QTL model. In general, the adjusted LOD scores are smaller than the (unadjusted) LOD scores from the 1D genome scans. This is usually because markers on different chromosomes are, in practice, correlated. Occasionally, when there are two linked QTLs in repulsion, the adjusted LOD score may be larger than the unadjusted LOD score for each. Similar considerations apply to unadjusted LOD scores from 2D genome scans.(60)

During this study, it was discovered that a paracentric chromosomal inversion within chromosome 6, from ∼30 to 52 cM, is fixed in the Jackson Laboratory's C3H/HeJ strain that we used for the F2 cross (http://www.jax.org/jaxmice). This inversion suppresses meiotic recombination, yielding potentially artifactual evidence of genetic linkage of phenotypes within this region. Thus, because of this inversion, QTL mapping data on chromosome 6 may not be reliable. We have chosen to take a conservative approach and have excluded results from QTL mapping on chromosome 6 for this publication.

RESULTS

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

Description of phenotypes in progenitor strains

As expected, the vertebral trabecular bone traits varied significantly in the progenitor strains (Table 1; Fig. 1). The total volume of the trabecular compartment in the vertebral body was 21% higher in the C3H mice (p < 0.001). Despite this, the BV and BV/TV were 40-50% lower in the C3H mice compared with the B6 mice (p < 0.001). Trabecular thickness was similar in both progenitor strains. Thus, the observed decrease in BV/TV was attributable primarily to fewer trabeculae in the vertebral bodies of C3H mice. These decreases in BV/TV and architecture were associated with a nearly 5-fold reduction in connectivity density and a greater than 3-fold increase in the structure model index in the C3H compared with the B6 mice (p < 0.001 for both, Table 1).

thumbnail image

Figure FIG. 1.. μCT images of representative fifth lumbar vertebral body of the B6 (left) and C3H (right) progenitors showing the volume of interest that was defined for the vertebral trabecular region (solid red line).

Download figure to PowerPoint

Table Table 1.. Description of Traits for the Trabecular Bone of the Vertebral Body in B6 and C3H Progenitor Strains (Mean ± SD)
Thumbnail image of

Quantitative trait data in F2 population and heritability estimates

Evaluating associations within the vertebral trabecular bone parameters (Table 2), the strongest correlation was seen between Tb.Sp and Tb.N (r = −0.96). BV/TV was also strongly positively correlated to Tb.Th (r = 0.81), Tb.N (r = 0.81), and connectivity density (r = 0.72), and strongly negatively correlated to Tb.Sp (r = −0.77) and SMI (r = −0.73, Table 2). Trabecular TV and BV/TV were only very weakly associated with each other, again indicating that a larger trabecular volume is not necessarily predictive of increased BV fraction. Body weight was weakly correlated with all of the vertebral trabecular traits (r = 0.13-0.22, p < 0.0001), and therefore the values were not adjusted for body weight in the QTL analyses.

Table Table 2.. Correlation Matrix for Vertebral Trabecular Bone Characteristics
Thumbnail image of

The values for trabecular TV, BV, BV/TV, Tb.Th, Tb.Sp, and Tb.N in the F2 progeny approximated a normal distribution, indicative of traits that are affected by a number of genes (Figs. 2A-2E). Arrows denoting the location of mean values for the C3H and B6 progenitor strains are superimposed on the F2 distributions for these traits. For vertebral trabecular BV/TV (Fig. 2C), the mean values for the progenitors are located toward opposite ends of the F2 distribution, suggesting that the C3H progenitor carries alleles that predominantly yield a low BV/TV, whereas the B6 progenitor alleles predominantly yield a high BV/TV. In contrast, the C3H and B6 progenitors had equal Tb.Th (Fig. 2D), yet there was a normal distribution of Tb.Th values in the F2 population. This phenomenon indicates that both of the progenitor strains carry alleles that can both positively and negatively influence trabecular thickness. The normal distribution of Tb.Th in the F2 population is evidence of the independent segregation of these alleles.

thumbnail image

Figure FIG. 2.. Distribution of vertebral (A) trabecular total volume, (B) bone volume, (C) bone volume fraction, (D) trabecular thickness, and (E) trabecular number in the B6C3H F2 progeny at 4 months of age. The mean values for the progenitor strains are designated by the arrows, and the solid line depicts a normal Gaussian distribution.

Download figure to PowerPoint

Identification, mapping, and inheritance patterns of QTL for vertebral trabecular BV/TV

Genome-wide quantitative trait analyses was conducted for vertebral trabecular TV, BV, BV adjusted for TV (i.e., BV/TV), Tb.Th, and Tb.N. Trabecular BV/TV is strongly correlated to most of the other vertebral trabecular traits (Table 2) and is likely to be a strong predictor of vertebral strength. Thus, as an example of the genetic analyses, results for vertebral trabecular BV/TV will be presented in detail, whereas the major findings for the remainder of the vertebral trabecular bone traits are summarized in Table 3 and Fig. 3 and are discussed later.

Genome-wide quantitative trait analysis revealed 13 QTLs for trabecular BV/TV, with adjusted LOD scores ranging from 3.5 to 13.1 (Table 3). Results for the whole genome scan for vertebral trabecular BV/TV are depicted graphically in Fig. 3A, which plots the LOD score from a 1D genome scan versus each autosome. Genome-wide corrected data are shown, with the critical level for significance in a 1D scan (p < 0.01) indicated. QTLs were located on chromosomes 1, 4, 8, 9, 10, 12, 13, 14, and 17, with two QTLs on chromosomes 1 and 9 and three on chromosome 12 (Table 3). After adjusting for contributions from the other QTL, each of the individual QTL explained between 0.8% and 3.0% of the variance in vertebral trabecular BV/TV in the F2 population (Table 3). The total variance explained for BV given TV (BV/TV) was 29% from the total regression model, after adjusting for contributions of the QTL for TV.

To determine how allelic variation at each of the QTL affected vertebral trabecular BV/TV, we computed the mean BV/TV for F2 mice after dividing them into groups that had allele combinations of b6/b6, b6/c3, or c3/c3 for each QTL. Figure 4 shows plots of these allele groups for the major QTLs (chromosomes 1, 4, 8, and 13). At these major loci, C3H alleles, which were previously associated with high cortical bone mass,(27, 40) had both positive and negative effects on vertebral trabecular BV/TV. In particular, C3H alleles were associated with increased BV/TV for the QTL on chromosomes 1 (D1Mit14) and 4 (D4Mit187) and decreased BV/TV from the QTLs on chromosomes 8 (D8Mit75), 9 (D9Mit196), and 13 (D13Mit245). Of these major QTLs, seven show additive inheritance, with heterozygous mice having significantly different BV/TV than mice with either b6/b6 or c3/c3 alleles (the difference between b6/b6 and c3/c3 mice ranging from 15% to 22%). Six of the major QTLs for BV/TV show dominant inheritance patterns; in both cases, B6 displays the dominant allele. Although not shown graphically, this type of analysis was performed for all QTL (Table 3). At the remainder of the QTLs, C3H alleles were associated with both increased (chromosomes 10 and 12 [D12Mit79] and 14) and decreased trabecular BV/TV (chromosomes 1 [D1Mit282], 12 [D12Mit215 and D12Mit114], and 17). The inheritance pattern for the QTLs on chromosomes 1(D1Mit14), 4, 8, 12, and 17 was additive. In comparison, the inheritance pattern for the QTLs on chromosomes 1 (D1Mit282), 9, 10, 13, and 14 was dominant (Table 3). We did not identify any major gene-gene interactions for BV/TV or other trabecular bone traits.

thumbnail image

Figure FIG. 4.. Effects of C3H (c3) vs. B6 (b6) alleles at the major QTLs affecting vertebral trabecular BV/TV (chromosomes 1, 4, 8, and 13). Significant differences (p < 0.01) are indicated by the lowercase letters, where “a” indicates that the mean value for mice with c3/c3 alleles is significantly different than mice with b6/b6 alleles; “b” indicates the mean value for mice with c3/b6 value is significantly different than those with b6/b6 alleles; and “c” indicates the mean value of mice with c3/c3 alleles differs from mice with c3/b6 alleles. The number of mice in each group are shown in parentheses. Error bars represent SEM.

Download figure to PowerPoint

Identification, mapping, and inheritance patterns of QTL for other vertebral trabecular bone traits

Results from the genome-wide scan and regression analyses for the remainder of the vertebral trabecular bone traits, including trabecular TV, BV, Tb.Th, and Tb.N are shown in Fig. 3 and Table 3. Figure 4 shows the results from the 1D genome scan (unadjusted LOD scores), whereas Table 3 shows the best MIT marker for the QTL region, adjusted LOD score, percent variation in the F2 population explained by each QTL, whether the trait is influenced positively or negatively by B6 and C3H alleles, and the mode of inheritance. Overall, these scans revealed a pattern of genetic regulation derived from 13 autosomes. Five to 10 QTLs were associated with each trait, with adjusted LOD scores ranging from 3.1 to 14.4.

Five QTLs, on chromosomes 1, 9, 10, 11, and 15, were identified for the total volume of the trabecular region in the vertebral body (Table 3, Fig. 3). Three of the five QTLs for trabecular TV were shared by Tb.N (chromosomes 1, 9, and 10), suggesting that the size of the trabecular region and the total number of trabecular struts may have similar genetic regulation. C3H alleles were associated with decreased trabecular TV for the QTLs on chromosomes 1 and 10 but were associated with increased trabecular TV for the QTLs on chromosomes 9, 11, and 15. The regression model including all the QTLs explained 27% of the variance in the distribution of trabecular TV in the F2 population, with each individual QTL explaining between 1.5% and 3.1% of the variance after adjusting for the contributions of the other QTLs. The QTLs for trabecular TV displayed both dominant and additive inheritance.

For BV of the trabecular region, we identified six QTLs (chromosomes 4, 8, 9, 12, 13, and 17) with adjusted LOD scores ranging from 4.2 to 9.5 (Fig. 3). All of these were shared QTLs with BV/TV (Table 3). None of the QTLs identified for trabecular BV were coincident with those for trabecular TV. The regression model containing these six QTL explained 33% of the variation in trabecular BV in the F2 population. The QTL for BV exhibited both dominant and additive inheritance.

Regarding the microarchitectural traits, we identified 9 QTLs for Tb.Th and 10 for Tb.N (Table 3; Fig. 3). The QTLs identified for Tb.Sp were identical to those for Tb.N (data not shown). In nearly all instances, the QTLs identified for Tb.Th and Tb.N overlapped with those identified for trabecular BV/TV (Table 3). Interestingly, Tb.Th and Tb.N themselves shared only four common QTLs: one each on chromosomes 4 and 8 and two on chromosome 12. QTL identified for Tb.Th but not Tb.N included those on chromosomes 7, 9 (D9Mit297), 14, 17, and 18. QTL identified for Tb.N but not Tb.Th included those on chromosomes 1, 9 (D9Mit95 and D9Mit196), 12 (D12Mit215), and 13. The strongest QTL for both Tb.Th and Tb.N was on chromosome 4, by itself explaining 5.9% and 5.5%, respectively, of the variation of these traits in the F2 population. Altogether, the regression models for Tb.Th and Tb.N explained 22% and 32%, respectively, of the variance of these traits in the F2 population. Inheritance patterns for the Tb.Th and Tb.N QTLs were both dominant and additive (Table 3).

DISCUSSION

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

Previous studies have shown that genetic factors contribute to many skeletal traits. Here we show that, in B6C3H-F2 mice, vertebral trabecular bone traits have strong heritable components. Genome-wide scans identified several QTLs that regulate these vertebral trabecular bone traits, some of which overlap and some of which are distinct from those QTL that were previously identified for femoral and vertebral vBMD in the same sample.(27) Overall, the genome-wide scans revealed a pattern of genetic regulation derived from 13 autosomes, with 5-13 QTLs associated with each of the vertebral trabecular bone traits. This is the first study that we are aware of that identifies QTLs associated with trabecular bone characteristics and thus may represent an important contribution toward understanding the genetic regulation of skeletal fragility.

In previous studies with this B6C3F2 population, we have shown strong heritable determinants for femoral and vertebral vBMD, as assessed by pQCT,(19, 27) and for serum insulin-like growth factor (IGF)-1 levels.(56, 64) Here we confirm and extend these observations with evidence that most trabecular bone traits are also heritable. We previously reported that C3H mice had greater vBMD of the vertebra than did B6 mice at 4, 8, and 12 months of age.(27) This observation differs from findings here, which indicate that C3H mice, although having slightly larger vertebrae, have much lower trabecular BV/TV in the vertebral body than do B6 mice. This difference is explained by noting that vBMD by pQCT reflects both the cortical and trabecular bone compartments, and in fact, is likely dominated by contributions from cortical bone. C3H mice have greater cortical bone mass and mineralization than B6 mice,(47, 49) thereby explaining the different observations for vertebral vBMD and vertebral trabecular BV/TV.

The QTLs for vertebral trabecular traits were distributed widely across the genome, indicating that genetic regulation of these features is complex (Table 3). The size of the trabecular region (i.e., TV) and bone volume (i.e., BV) in the trabecular region had distinct QTLs (not shared by the reciprocal phenotype), suggesting that vertebral size and trabecular mass may be differentially regulated by genetic factors. Eight of 13 QTLs for BV/TV were shared with either BV or TV. Only one QTL identified for trabecular BV/TV was not shared with either Tb.Th or Tb.N.

Several of the QTLs identified in this study were previously identified for total vertebral vBMD and total femoral vBMD in this B6C3H F2 population(27) and for whole body BMD(21, 23) and femoral cortical thickness index(28) in studies from different mouse strains (Table 4), indicating that these QTLs are broadly implicated in the regulation of several skeletal traits. However, a number of QTLs were not previously identified and therefore seem to be unique to vertebral trabecular bone volume fraction and microarchitecture, including regions on chromosomes 9 (D9Mit297, D9Mit95), 10, and 12 (D12Mit114, D12Mit79). The presence of these unique QTLs suggests that genetic regulation of skeletal traits may exhibit specificity for a skeletal site and/or bone compartment. The notion of skeletal site specificity is supported in part by the observations that, whereas C3H mice have lower trabecular BV/TV than B6 in the vertebra, they have higher trabecular BV/TV than B6 in the metaphyseal regions of the long bones.(44, 65, 66) Thus, it should be noted that the QTL identified in this study were for vertebral trabecular bone traits and that they may differ from those for metaphyseal trabecular bone traits.

Table Table 4.. Comparison of QTL for Vertebral Trabecular Bone Density and Architecture With Those Previously Identified for Volumetric and Areal BMD in Mouse Genome-Wide Scans
Thumbnail image of

Notable is the lack of overlap with QTLs previously identified for spine BMD in mice generated by reciprocal crosses between SAMP6 and either SAMR1 or AKR/J strains.(22) Interestingly, however, three of the four QTLs previously identified for serum IGF-1 (on chromosomes 10, 11, and 15)(56) overlap with QTLs for the total volume of the trabecular bone region, perhaps indicating some common genetic or biological pathway for regulating bone size or the size of the trabecular compartment.

The total variance for vertebral trabecular traits in the B6C3-F2 population explained by the QTLs ranged from 22% to 33%. Whereas this is potentially a large amount of explained variability, it is much less than the previously published heritability estimates for skeletal traits in mice. There are several possible reasons for this apparent discrepancy. First, it should be recognized that the broad-sense heritability predictions may overestimate the true genetic effects, because they are based on the assumption that the estimate of environmental variation based on the parental strains is applicable to the F2 population, and this may not always be the case. Second, the detected QTL only represent the major loci segregating in this cross, and there are likely to be additional chromosomal regions that effect vertebral trabecular bone traits, which cannot be reliably detected. However, if these regions with low LOD scores included genes influencing our trabecular bone traits, they each would likely provide little additional explanation for the variance in the F2 population, but collectively may explain a larger proportion of variation in these complex traits. Other sources, such as undetected gene-gene interactions may also contribute to trait variability. Moreover, the X chromosome was not studied in this cross and may contain important genes contributing to trabecular bone mass and architecture. Taking these explanations together, we expect that the proportion of variance explained by the QTLs to be smaller than heritability estimates.

Defining the exact location of a QTL for any quantitative trait is difficult, even in a very large F2 intercross population. Initial genome-wide scans generally identify QTL regions ranging in size from 10 to 30 cM. Using the 1.5 LOD method proposed by Dupuis and Siegmund,(67) the size of the regions containing a QTL for vertebral trabecular bone ranged from ∼9 to 66 cM in genetic distance. It is estimated that, based on finding ∼30,000-35,000 genes in the mouse genome,(68) there are roughly 20-30 genes/cM. Thus, each of our QTL regions is predicted to contain ∼250-2000 genes. It is therefore likely that each QTL contains multiple genes that regulate trabecular bone traits. This large number of potential genes precludes discussion of candidate genes at this point. Clearly, additional studies using crosses among different strains and nested congenic and/or recombinant inbred strains will be required to further define the chromosomal regions regulating trabecular bone traits.(20)

The long-term goal of these studies is to identify human genes that contribute to skeletal fragility through their effects on trabecular density and architecture. In this regard, several of the regions we identified in the mouse are homologous with regions identified in human genome-wide scans for various skeletal traits, including QTLs for spine BMD (Table 5). In particular, the QTL previously identified for spine BMD on human 1q21-23, 6p11-12, and 14q31-34(69, 71, 74) are syntenic with the QTLs on mouse chromosomes 1 (82 cM), 17 (4 cM), and 12 (53 cM), respectively. However, given the lack of statistical power in many human genome-wide linkage studies and that none of the human studies has evaluated a pure trabecular phenotype, it is not surprising that many of our QTLs have not yet been identified in human studies. Moreover, it may be that human genes do not carry suitable variants for these traits.

Table Table 5.. Homologous Linkage Relationships Between Mouse Chromosomal Regions With QTLs for Vertebral Trabecular Traits and Human Chromosomal Regions and Listing of QTLs Identified in Genome-Wide Scans in Humans
Thumbnail image of

There are several potential limitations associated with this study. We used bone traits derived from μCT imaging for our genetic analyses. The specific traits were chosen to reflect aspects of the size of the trabecular region, the amount of bone in the trabecular region, and microarchitecture. It is likely that these traits are the key determinants of the mechanical behavior of the vertebrae. However, we did not have a direct assessment of vertebral strength in these mice, and ultimately, compression testing of the vertebra would be useful to define relationships between the μCT-derived variables and mechanical behavior. In addition, the vertebral strength data could be used to confirm the QTLs identified in this study and/or to identify new QTLs. Second, these QTLs were identified in two inbred strains (C3H and B6) that are polymorphic at only about 50% of the Mit markers. Therefore, there may be additional QTLs that we were unable to identify using these progenitor strains. Genetic evaluation of crosses with different mouse strains may reveal additional QTLs for vertebral bone traits. Third, because of the well-established complications for evaluating recombination events caused by inverted chromosomal regions, we chose the conservative approach of not reporting QTL data for the central region of chromosome 6 (i.e., 30-52 cM). However, given the caveat that the mapping data may be unreliable, it is of interest to note that we detected significant QTLs (adjusted LOD scores of 3.1-3.3) for total trabecular volume, BV/TV, and trabecular number with markers in the region of the newly identified inversion region of chromosome 6. Our decision to exclude these QTLs from the results was based on genetic mapping considerations alone, because additional knowledge from the B6.C3H-6T congenic strain (which carries the same C3H chromosome 6 segment) confirms that there is a QTL regulating serum IGF-1 within this region and that this QTL either directly or indirectly (through modulating serum IGF-1) influences vertebral trabecular BV fraction and microarchitecture.(64)

A fourth limitation was that we only studied female mice. Therefore, we cannot ascertain whether there is gender-specific regulation of vertebral trabecular bone traits, as was shown for whole body BMD and femoral cross-sectional geometry.(17, 24) Moreover, evaluation of genetic contributions to vertebral bone strength in male mice may be important, because vertebral fractures are less common in men than women.

As a final limitation, it should be noted that we measured the vertebral traits at a single age. The 4-month time point was initially chosen as the age of peak total femoral and vertebral volumetric bone density, assessed by pQCT. Beamer et al.(27) reported that total femur vBMD derived from pQCT peaked at 4 months of age and was maintained until 12 months of age, whereas total vertebral vBMD peaked at age 4 months but declined thereafter. These vBMD measurements of the total femur and total vertebrae are likely to be influenced primarily by the cortical bone compartment. In contrast, the vertebral traits we assessed included indices that were purely trabecular bone. Halloran et al.(76) showed that trabecular BV/TV of the proximal tibia in male C57BL/6J mice peaks at 8-12 weeks of age and declines thereafter. Thus, it is likely that the trabecular traits in this study reflected a combination of what was achieved at the time of peak trabecular BV/TV acquisition, as well as the subsequent degree of age-related bone loss and possible environmental adaptations. However, it should be noted that the female F2 mice in this study are a mix of B6 and C3H, and both sex- and strain-related differences in age-related trabecular bone loss have yet to be established. Moreover, the age-related changes in vertebral trabecular bone in mice have not yet been studied, and they may differ from those seen in the metaphysis of long bones. Future studies in congenic mouse strains carrying the QTLs identified for vertebral trabecular traits are needed to determine the age- and sex-specific effects of these QTLs.

In conclusion, we observed a complex pattern of genetic regulation for vertebral trabecular bone density and architecture using the F2 intercross of the C57BL/6J and C3H/HeJ inbred mouse strains, identifying a number of QTL, some of which are distinct from those that were previously identified for total femoral and vertebral vBMD. Our findings have significant implications because the majority of osteoporotic fractures occur in the spine or metaphyseal regions, sites that are rich in trabecular bone. Trabecular bone characteristics are important determinants of bone strength at these skeletal sites. Specifically, it has been estimated that in the vertebral body, the portion of compressive load carried by the trabecular centrum, ranges from 20% to 50% at the mid-transverse plane, and may be up to 70% of the load at the endplate.(38, 60, 61) Vertebral trabecular bone properties are strongly associated with vertebral strength(38, 63) and seem to dominate the damage behavior of human cadaveric vertebral bodies.(77) In addition, trabecular bone properties may be important contributors to the strength of the proximal femur, where the percent of load carried by trabecular (versus cortical) bone is predicted to be 70% at the subcapital region, 50% at the mid-neck, and 20% at the intertrochanteric region.(39) The importance of trabecular bone features in the etiology of osteoporotic fractures is further supported by increasing evidence for differences in trabecular bone architecture and mechanical behavior between individuals with fractures and those who are fracture free.(78-83) Altogether, identification of genes that regulate trabecular bone density and architecture may ultimately yield important information with respect to mechanisms that regulate the acquisition and maintenance of mechanical integrity of the skeleton.

Acknowledgements

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

This research was supported by National Institutes of Health Grants AR43618, CA34196, and AR46530.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Anonymous 2001 Osteoporosis prevention, diagnosis, and therapy. JAMA 285:785795.
  • 2
    Bouxsein ML 2001 Biomechanics of age-related fractures. In: MarcusR, FeldmanD, KelseyJL (eds.) Osteoporosis, 2nd ed. Academic Press, San Diego, CA, USA, pp. 509534.
  • 3
    Fox KM, Cummings SR, Powell-Threets K, Stone K 1998 Family history and risk of osteoporotic fracture. Study of Osteoporotic Fractures Research Group. Osteoporos Int 8:557562.
  • 4
    Keen RW, Hart DJ, Arden NK, Doyle DV, Spector TD 1999 Family history of appendicular fracture and risk of osteoporosis: A population-based study. Osteoporos Int 10:161166.
  • 5
    Eisman JA 1999 Genetics of osteoporosis. Endocr Rev 20:788804.
  • 6
    Stewart TL, Ralston SH 2000 Role of genetic factors in the pathogenesis of osteoporosis. J Endocrinol 166:235245.
  • 7
    Peacock M, Turner CH, Econs MJ, Foroud T 2002 Genetics of osteoporosis. Endocr Rev 23:303326.
  • 8
    Koller DL, Liu G, Econs MJ, Hui SL, Morin PA, Joslyn G, Rodriguez LA, Conneally PM, Christian JC, Johnston CC Jr, Foroud T, Peacock M 2001 Genome screen for quantitative trait loci underlying normal variation in femoral structure. J Bone Miner Res 16:985991.
  • 9
    Livshits G, Yakovenko K, Kobyliansky E 2003 Quantitative genetic study of radiographic hand bone size and geometry. Bone 32:191198.
  • 10
    Kelly PJ, Hopper JL, Macaskill GT, Pocock NA, Sambrook PN, Eisman JA 1991 Genetic factors in bone turnover. J Clin Endocrinol Metab 72:808813.
  • 11
    Tokita A, Kelly PJ, Nguyen TV, Qi JC, Morrison NA, Risteli L, Risteli J, Sambrook PN, Eisman JA 1994 Genetic influences on type I collagen synthesis and degradation: Further evidence for genetic regulation of bone turnover. J Clin Endocrinol Metab 78:14611466.
  • 12
    Garnero P, Arden NK, Griffiths G, Delmas PD, Spector TD 1996 Genetic influence on bone turnover in postmenopausal twins. J Clin Endocrinol Metab 81:140146.
  • 13
    Harris M, Nguyen TV, Howard GM, Kelly PJ, Eisman JA 1998 Genetic and environmental correlations between bone formation and bone mineral density: A twin study. Bone 22:141145.
  • 14
    Uitterlinden A, Van Leeuwen J, Pols H 2001 Genetics and genomics of osteoporosis. In: MarcusR, FeldmanD, KelseyJ (eds.) Osteoporosis, Vol. 2. Academic Press, San Diego, CA, USA, pp. 639668.
  • 15
    Nguyen TV, Blangero J, Eisman JA 2000 Genetic epidemiological approaches to the search for osteoporosis genes. J Bone Miner Res 15:392401.
  • 16
    Lee C 2002 Irresistable force meets immovable object: SNP mapping of complex diseases. Trends Genet 18:7669.
  • 17
    Klein RF, Turner RJ, Skinner LD, Vartanian KA, Serang M, Carlos AS, Shea M, Belknap JK, Orwoll ES 2002 Mapping quantitative trait loci that influence femoral cross-sectional area in mice. J Bone Miner Res 17:17521760.
  • 18
    Korstanje R, Paigen B 2002 From QTL to gene: The harvest begins. Nat Genet 31:235236.
  • 19
    Shultz KL, Donahue LR, Bouxsein ML, Baylink DJ, Rosen CJ, Beamer WG 2003 Congenic strains of mice for verification and genetic decomposition of quantitative trait loci for femoral bone mineral density. J Bone Miner Res 18:175185.
  • 20
    Shmookler Reis RJ 2003 From QTL mapping to genes: The long and winding road. J Bone Miner Res 18:186189.
  • 21
    Klein RF, Mitchell SR, Phillips TJ, Belknap JK, Orwoll ES 1998 Quantitative trait loci affecting peak bone mineral density in mice. J Bone Miner Res 13:16481656.
  • 22
    Benes H, Weinstein RS, Zheng W, Thaden JJ, Jilka RL, Manolagas SC, Shmookler Reis RJ 2000 Chromosomal mapping of osteopenia-associated quantitative trait loci using closely related mouse strains. J Bone Miner Res 15:626633.
  • 23
    Klein OF, Carlos AS, Vartanian KA, Chambers VK, Turner EJ, Phillips TJ, Belknap JK, Orwoll ES 2001 Confirmation and fine mapping of chromosomal regions influencing peak bone mass in mice. J Bone Miner Res 16:19531961.
  • 24
    Orwoll ES, Belknap JK, Klein RF 2001 Gender specificity in the genetic determinants of peak bone mass. J Bone Miner Res 16:19621971.
  • 25
    Shimizu M, Higuchi K, Kasai S, Tsuboyama T, Matsushita M, Mori M, Shimizu Y, Nakamura T, Hosokawa M 2001 Chromosome 13 locus, Pbd2, regulates bone density in mice. J Bone Miner Res 16:19721982.
  • 26
    Beamer WG, Shultz KL, Churchill GA, Frankel WN, Baylink DJ, Rosen CJ, Donahue LR 1999 Quantitative trait loci for bone density in C57BL/6J and CAST/EiJ inbred mice. Mamm Genome 10:10431049.
  • 27
    Beamer WG, Shultz KL, Donahue LR, Churchill GA, Sen S, Wergedal JR, Baylink DJ, Rosen CJ 2001 Quantitative trait loci for femoral and lumbar vertebral bone mineral density in C57BL/6J and C3H/HeJ inbred strains of mice. J Bone Miner Res 16:11951206.
  • 28
    Shimizu M, Higuchi K, Bennett B, Xia C, Tsuboyama T, Kasai S, Chiba T, Fujisawa H, Kogishi K, Kitado H, Kimoto M, Takeda N, Matsushita M, Okumura H, Serikawa T, Nakamura T, Johnson TE, Hosokawa M 1999 Identification of peak bone mass QTL in a spontaneously osteoporotic mouse strain. Mamm Genome 10:8187.
  • 29
    Volkman S, Galecki A, Burke D, Paczas M, Moalli M, Miller R, Goldstein S 2002 Identification of five genetic markers in mice that influence femoral size and shape. J Bone Miner Res 17:S321.
  • 30
    Drake TA, Schadt E, Hannani K, Kabo JM, Krass K, Colinayo V, Greaser LE III, Goldin J, Lusis AJ 2001 Genetic loci determining bone density in mice with diet-induced atherosclerosis. Physiol Genomics 5:205215.
  • 31
    Corva PM, Medrano JF 2001 Quantitative trait loci (QTLs) mapping for growth traits in the mouse: A review. Genet Sel Evol 33:105132.
  • 32
    Corva PM, Horvat S, Medrano JF 2001 Quantitative trait loci affecting growth in high growth (hg) mice. Mamm Genome 12:284290.
  • 33
    Christians JK, Bingham VK, Oliver FK, Heath TT, Keightley PD 2003 Characterization of a QTL affecting skeletal size in mice. Mamm Genome 14:175183.
  • 34
    Li X, Masinde G, Gu W, Wergedal J, Mohan S, Baylink DJ 2002 Genetic dissection of femur breaking strength in a large population (MRL/MpJ × SJL/J) of F2 Mice: Single QTL effects, epistasis, and pleiotropy. Genomics 79:734740.
  • 35
    Koller D, Schriefer J, Sun Q, Shultz K, Donahue L, Rosen C, Foroud T, Beamer W, Turner C 2003 Genetic effects for femoral biomechanics, structure and density in C57BL/6J and C3H/HeJ inbred mouse strains. J Bone Miner Res 18:17581765.
  • 36
    Rockoff SD, Sweet E, Bleustein J 1969 The relative contribution of trabecular and cortical bone to the strength of human lumbar vertebrae. Calcif Tissue Res 3:163175.
  • 37
    Mosekilde L, Mosekilde L 1990 Sex differences in age-related changes in vertebral body size, density and biomechanical competence in normal individuals. Bone 11:6773.
  • 38
    Silva M, Keaveny T, Hayes W 1997 Load sharing between the shell and centrum in the lumbar vertebral body. Spine 22:140150.
  • 39
    Lotz J, Cheal E, Hayes W 1995 Stress distributions within the proximal femur during gait and falls: Implications for osteoporotic fracture. Osteoporos Int 5:252261.
  • 40
    Beamer WG, Donahue LR, Rosen CJ, Baylink DJ 1996 Genetic variability in adult bone density among inbred strains of mice. Bone 18:397403.
  • 41
    Rosen CJ, Dimai HP, Vereault D, Donahue LR, Beamer WG, Farley J, Linkhart S, Linkhart T, Mohan S, Baylink DJ 1997 Circulating and skeletal insulin-like growth factor-I (IGF-I) concentrations in two inbred strains of mice with different bone mineral densities. Bone 21:217223.
  • 42
    Dimai HP, Linkhart TA, Linkhart SG, Donahue LR, Beamer WG, Rosen CJ, Farley JR, Baylink DJ 1998 Alkaline phosphatase levels and osteoprogenitor cell numbers suggest bone formation may contribute to peak bone density differences between two inbred strains of mice. Bone 22:211216.
  • 43
    Akhter MP, Cullen DM, Pedersen EA, Kimmel DB, Recker RR 1998 Bone response to in vivo mechanical loading in two breeds of mice. Calcif Tissue Int 63:442449.
  • 44
    Sheng MH, Baylink DJ, Beamer WG, Donahue LR, Rosen CJ, Lau KH, Wergedal JE 1999 Histomorphometric studies show that bone formation and bone mineral apposition rates are greater in C3H/HeJ (high-density) than C57BL/6J (low-density) mice during growth. Bone 25:421429.
  • 45
    Kodama Y, Dimai HP, Wergedal J, Sheng M, Malpe R, Kutilek S, Beamer W, Donahue LR, Rosen C, Baylink DJ, Farley J 1999 Cortical tibial bone volume in two strains of mice: Effects of sciatic neurectomy and genetic regulation of bone response to mechanical loading. Bone 25:183190.
  • 46
    Chen C, Kalu DN 1999 Strain differences in bone density and calcium metabolism between C3H/HeJ and C57BL/6J mice. Bone 25:413420.
  • 47
    Turner CH, Hsieh YF, Muller R, Bouxsein ML, Baylink DJ, Rosen CJ, Grynpas MD, Donahue LR, Beamer WG 2000 Genetic regulation of cortical and trabecular bone strength and microstructure in inbred strains of mice. J Bone Miner Res 15:11261131.
  • 48
    Richman C, Kutilek S, Miyakoshi N, Srivastava AK, Beamer WG, Donahue LR, Rosen CJ, Wergedal JE, Baylink DJ, Mohan S 2001 Postnatal and pubertal skeletal changes contribute predominantly to the differences in peak bone density between C3H/HeJ and C57BL/6J mice. J Bone Miner Res 16:386397.
  • 49
    Jepsen KJ, Akkus OJ, Majeska RJ, Nadeau JH 2003 Hierarchical relationship between bone traits and mechanical properties in inbred mice. Mamm Genome 14:97104.
  • 50
    Ruegsegger P, Koller B, Müller R 1996 A microtomographic system for the nondestructive evaluation of bone architecture. Calcif Tissue Int 58:2429.
  • 51
    Hildebrand T, Ruegsegger P 1997 A new method for the model independent assessment of thickness in three-dimensional images. J Microsc 185:6775.
  • 52
    Hildebrand T, Laib A, Muller R, Dequeker J, Ruegsegger P 1999 Direct three-dimensional morphometric analysis of human cancellous bone: Microstructural data from spine, femur, iliac crest, and calcaneus. J Bone Miner Res 14:11671174.
  • 53
    Odgaard A 1997 Three-dimensional methods for quantification of cancellous bone architecture. Bone 20:315328.
  • 54
    Hildebrand T, Rüesegger P 1997 Quantification of bone microarchitecture with the structure model index. Comp Methods Biomech Biomed Eng 1:523.
  • 55
    Dietrich WF, Miller J, Steen R, Merchant MA, Damron-Boles D, Husain Z, Dredge R, Daly MJ, Ingalls KA, O'Connor TJ 1996 A comprehensive genetic map of the mouse genome. Nature 380:149152.
  • 56
    Rosen CJ, Churchill GA, Donahue LR, Shultz KL, Burgess JK, Powell DR, Ackert C, Beamer WG 2000 Mapping quantitative trait loci for serum insulin-like growth factor-1 levels in mice. Bone 27:521528.
  • 57
    Kao C, Zneg Z-B, Teasdale R 1999 Multiple interval mapping for quantitative trait loci. Genetics 152:12031216.
  • 58
    Sen S, Churchill GA 2001 A statistical framework for quantitative trait mapping. Genetics 159:371387.
  • 59
    Jansen R 1993 Interval mapping of multiple quantitative trait loci. Genetics 135:205211.
  • 60
    Mosekilde L 1998 The effect of modelling and remodelling on human vertebral body architecture. Technol Health Care 6:287297.
  • 61
    Homminga J, Weinans H, Gowin W, Felsenberg D, Huiskes R 2001 Osteoporosis changes the amount of vertebral trabecular bone at risk of fracture but not the vertebral load distribution. Spine 26:15551561.
  • 62
    Tanner J 1949 Fallacy of per-weight and per-surface are standards, and their relation to spurious correlation. J Appl Physiol 2:115.
  • 63
    Thomsen JS, Ebbesen EN, Mosekilde L 2002 Predicting human vertebral bone strength by vertebral static histomorphometry. Bone 30:502508.
  • 64
    Bouxsein ML, Rosen CJ, Turner CH, Ackert CL, Shultz KL, Donahue LR, Churchill G, Adamo ML, Powell DR, Turner RT, Muller R, Beamer WG 2002 Generation of a new congenic mouse strain to test the relationships among serum insulin-like growth factor I, bone mineral density, and skeletal morphology in vivo. J Bone Miner Res 17:570579.
  • 65
    Akhter MP, Iwaniec UT, Covey MA, Cullen DM, Kimmel DB, Recker RR 2000 Genetic variations in bone density, histomorphometry, and strength in mice. Calcif Tissue Int 67:337344.
  • 66
    Bouxsein ML, Solanki P, Müller R, Turner CH, Shultz KL, Ackert-Bicknell C, Donahue LR, Rosen CJ, Beamer W 2002 B6. C3H congenic mouse strains exhibit site-specific alterations in femoral and vertebral trabecular bone mass and microarchitecture. J Bone Miner Res 17:S176.
  • 67
    Dupuis J, Siegmund D 1999 Statistical methods for mapping quantitative trait loci from a dense set of markers. Genetics 151:373386.
  • 68
    Waterston RH, Lindblad-Toh K, Birney E, Rogers J, Abril JF, Agarwal P, Agarwala R, Ainscough R, Alexandersson M, An P, Antonarakis SE, Attwood J, Baertsch R, Bailey J, Barlow K, Beck S, Berry E, Birren B, Bloom T, Bork P, Botcherby M, Bray N, Brent MR, Brown DG, Brown SD, Bult C, Burton J, Butler J, Campbell RD, Carninci P, Cawley S, Chiaromonte F, Chinwalla AT, Church DM, Clamp M, Clee C, Collins FS, Cook LL, Copley RR, Coulson A, Couronne O, Cuff J, Curwen V, Cutts T, Daly M, David R, Davies J, Delehaunty KD, Deri J, Dermitzakis ET, Dewey C, Dickens NJ, Diekhans M, Dodge S, Dubchak I, Dunn DM, Eddy SR, Elnitski L, Emes RD, Eswara P, Eyras E, Felsenfeld A, Fewell GA, Flicek P, Foley K, Frankel WN, Fulton LA, Fulton RS, Furey TS, Gage D, Gibbs RA, Glusman G, Gnerre S, Goldman N, Goodstadt L, Grafham D, Graves TA, Green ED, Gregory S, Guigo R, Guyer M, Hardison RC, Haussler D, Hayashizaki Y, Hillier LW, Hinrichs A, Hlavina W, Holzer T, Hsu F, Hua A, Hubbard T, Hunt A, Jackson I, Jaffe DB, Johnson LS, Jones M, Jones TA, Joy A, Kamal M, Karlsson EK, et al. 2002 Initial sequencing and comparative analysis of the mouse genome. Nature 420:520562.
  • 69
    Koller DL, Econs MJ, Morin PA, Christian JC, Hui SL, Parry P, Curran ME, Rodriguez LA, Conneally PM, Joslyn G, Peacock M, Johnston CC, Foroud T 2000 Genome screen for QTLs contributing to normal variation in bone mineral density and osteoporosis. J Clin Endocrinol Metab 85:31163120.
  • 70
    Devoto M, Specchia C, Li HH, Caminis J, Tenenhouse A, Rodriguez H, Spotila LD 2001 Variance component linkage analysis indicates a QTL for femoral neck bone mineral density on chromosome 1p36. Hum Mol Genet 10:24472452.
  • 71
    Wilson SG, Reed PW, Bansal A, Chiano M, Lindersson M, Langdown M, Prince RL, Thompson D, Thompson E, Bailey M, Kleyn PW, Sambrook P, Shi MM, Spector TD 2003 Comparison of genome screens for two independent cohorts provides replication of suggestive linkage of bone mineral density to 3p21 and 1p36. Am J Hum Genet 72:144155.
  • 72
    Devoto M, Shimoya K, Caminis J, Ott J, Tenenhouse A, Whyte MP, Sereda L, Hall S, Considine E, Williams CJ, Tromp G, Kuivaniemi H, Ala-Kokko L, Prockop DJ, Spotila LD 1998 First-stage autosomal genome screen in extended pedigrees suggests genes predisposing to low bone mineral density on chromosomes 1p, 2p and 4q. Eur J Hum Genet 6:151157.
  • 73
    Deng HW, Xu FH, Huang QY, Shen H, Deng H, Conway T, Liu YJ, Liu YZ, Li JL, Zhang HT, Davies KM, Recker RR 2002 A whole-genome linkage scan suggests several genomic regions potentially containing quantitative trait Loci for osteoporosis. J Clin Endocrinol Metab 87:51515159.
  • 74
    Karasik D, Myers RH, Cupples LA, Hannan MT, Gagnon DR, Herbert A, Kiel DP 2002 Genome screen for quantitative trait loci contributing to normal variation in bone mineral density: The Framingham Study. J Bone Miner Res 17:17181727.
  • 75
    Ferrari S 2004 Genomics and osteoporosis: What are the implications? IBMS Bone Key 2010.1138/ibmske: 20040121.
  • 76
    Halloran BP, Ferguson VL, Simske SJ, Burghardt A, Venton LL, Majumdar S 2002 Changes in bone structure and mass with advancing age in the male C57BL/6J mouse. J Bone Miner Res 17:10441050.
  • 77
    Kopperdahl DL, Pearlman JL, Keaveny TM 2000 Biomechanical consequences of an isolated overload on the human vertebral body. J Orthop Res 18:685690.
  • 78
    Ciarelli TE, Fyhrie DP, Schaffler MB, Goldstein SA 2000 Variations in three-dimensional cancellous bone architecture of the proximal femur in female hip fractures and in controls. J Bone Miner Res 15:3240.
  • 79
    Legrand E, Chappard D, Pascaretti C, Duquenne M, Krebs S, Rohmer V, Basle MF, Audran M 2000 Trabecular bone microarchitecture, bone mineral density, and vertebral fractures in male osteoporosis. J Bone Miner Res 15:1319.
  • 80
    Hordon LD, Raisi M, Aaron JE, Paxton SK, Beneton M, Kanis JA 2000 Trabecular architecture in women and men of similar bone mass with and without vertebral fracture: I. two-dimensional histology. Bone 27:271276.
  • 81
    Aaron JE, Shore PA, Shore RC, Beneton M, Kanis JA 2000 Trabecular architecture in women and men of similar bone mass with and without vertebral fracture: II. three-dimensional histology. Bone 27:277282.
  • 82
    Mann V, Hobson EE, Li B, Stewart TL, Grant SF, Robins SP, Aspden RM, Ralston SH 2001 A COL1A1 Sp1 binding site polymorphism predisposes to osteoporotic fracture by affecting bone density and quality. J Clin Invest 107:899907.
  • 83
    Homminga J, McCreadie BR, Ciarelli TE, Weinans H, Goldstein SA, Huiskes R 2002 Cancellous bone mechanical properties from normals and patients with hip fractures differ on the structure level, not on the bone hard tissue level. Bone 30:759764.