Mapping Quantitative Trait Loci That Influence Femoral Cross-sectional Area in Mice


  • Robert F. Klein M.D.,

    Corresponding author
    1. Bone and Mineral Research Unit, Department of Medicine, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon, USA
    • Bone and Mineral Unit (CR113), Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97201-3098, USA
    Search for more papers by this author
  • Renn J. Turner,

    1. Bone and Mineral Research Unit, Department of Medicine, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon, USA
    Search for more papers by this author
  • Lisa D. Skinner,

    1. Department of Orthopedics, Oregon Health and Science University, Portland, Oregon, USA
    Search for more papers by this author
  • Kristina A. Vartanian,

    1. Bone and Mineral Research Unit, Department of Medicine, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon, USA
    Search for more papers by this author
  • Maqsood Serang,

    1. Department of Orthopedics, Oregon Health and Science University, Portland, Oregon, USA
    Search for more papers by this author
  • Amy S. Carlos,

    1. Bone and Mineral Research Unit, Department of Medicine, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon, USA
    Search for more papers by this author
  • Marie Shea,

    1. Department of Orthopedics, Oregon Health and Science University, Portland, Oregon, USA
    Search for more papers by this author
  • John K. Belknap,

    1. Department of Behavioral Neuroscience, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon, USA
    Search for more papers by this author
  • Eric S. Orwoll

    1. Bone and Mineral Research Unit, Department of Medicine, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon, USA
    Search for more papers by this author

  • The authors have no conflict of interest.


Size and shape are critical determinants of the mechanical properties of skeletal elements and can be anticipated to be highly heritable. Moreover, the genes responsible may be independent of those that regulate bone mineral density (BMD). To begin to identify the heritable determinants of skeletal geometry, we have examined femoral cross-sectional area (FCSA) in male and female mice from two inbred strains of mice with divergent FCSA (C57BL/6 [B6] and DBA/2 [D2]), a large genetically heterogeneous population (n = 964) of B6D2F2 mice and 18 BXD recombinant inbred (RI) strains derived from their F2 cross. Femora were harvested from 16-week-old mice and FCSA (bone and marrow space enclosed within the periosteum) was measured at the midshaft by digital image analysis. In all mouse populations examined, FCSA was positively correlated with body weight and weight-corrected FCSA (WC-FCSA) values were normally distributed in the BXD-RI and F2 populations, suggesting polygenic control of this trait. Genome-wide quantitative trait locus (QTL) analysis of the B6D2F2 population revealed regions on four different chromosomes that were very strongly linked to WC-FCSA (chromosomes 6, 8, 10, and X) in both genders. Evidence of gender-specific genetic influences on femoral geometry was also identified at three other chromosomal sites (chromosomes 2, 7, and 12). Supporting evidence for the WC-FCSA QTLs on chromosomes 2, 7, 8, 10, and 12 also was present in the RI strains. Interestingly, none of these WC-FCSA QTLs were identified in our previous QTL analysis of whole body BMD in the same B6D2F2 population. Thus, the genetic determinants of bone size appear to be largely, if not entirely, distinct from those that regulate BMD attainment. The identification of the genes responsible for geometric differences in bone development should reveal fundamentally important processes in the control of skeletal integrity.


Osteoporosis is the result of abnormalities in both the amount of bone and the architectural arrangement of bone tissue that lead to decreased skeletal strength and increased fracture risk. Osteoporotic fracture risk is determined by a complex interplay of factors, including a strong genetic component. For instance, the role of low bone density in determining fracture risk is well-established and the level of bone mineral density (BMD) achieved in early adulthood (peak BMD) is a highly heritable trait.(1–3) Although recent clinical reports show promise,(4–8) unraveling the genetic influences on bone mass will be difficult because of the genetic and cultural heterogeneity of patient populations. One approach to this problem is the use of animal models to pinpoint candidate genes for more focused human investigation. Recent studies have found significant differences in peak BMD of phenotypically normal inbred strains of mice.(9,10) Because these genetically distinct strains were raised in the same controlled environment, the observed interstrain differences indicate substantial genetic regulation of bone acquisition. Now, there is an ongoing effort using modern genetic analyses to map the chromosomal loci of genes responsible for the heritable differences in peak BMD in mice.(10–17)

It is also known that a family history of fracture approximately doubles the risk of hip fracture in older women.(18) This increased risk is independent of the additional risk associated with low bone mass, suggesting that other skeletal characteristics may both contribute to fracture risk and have a genetic basis. From a mechanical perspective, the strength of a bone is derived from material properties that govern its ability to withstand loading stress and its structural geometry that determines the magnitudes of those stresses within the bone tissue itself. Elements of femoral neck geometry such as hip axis length contribute independently to the risk of hip fracture(19–21) and osteoporotic fractures are more likely in subjects with smaller bone size.(22–24) Studies of stress fractures in young, healthy adults indicated that diaphyseal dimensions (even after correction for body weight) were significantly smaller in fracture cases.(25–27) Bone morphology is strongly driven by mechanical stimuli associated with increasing body mass.(28) However, gender, physical activity, and dietary and genetic factors also influence skeletal structure.(29)

Although studies using inbred strains of mice indicate genetic differences in BMD, phenotypes that characterize skeletal geometry have not been well defined. For a long bone, one of the most important geometric properties influencing its ability to resist fracture is its cross-sectional area.(30) For the same amount of bone tissue, long bones with larger total cross-sectional area will possess increased bending strength but lower BMD compared with bones with smaller cross-sectional area. The purpose of this study was to investigate the genetic determinants of femoral cross-sectional area (FCSA) in mice. The C57BL/6 (B6) and DBA/2 (D2) inbred mouse strains exhibit substantial differences in femoral geometry,(14) despite similar lean body mass(31) and activity levels.(32) Thus, we felt this experimental animal system would be optimal for identifying genetic influences impacting directly on skeletal processes to determine bone size. Therefore, we examined male and female mice from the B6 and D2 parental strains, a large genetically heterogeneous population of B6D2F2 mice and 18 BXD recombinant inbred (RI) strains derived from their F2 cross.

Using this murine model, we sought to examine the heritability of femoral geometry and identify chromosomal regions linked to FCSA. Previously, we have identified five chromosomal regions that influence whole body BMD in mice derived from the same B6 and D2 progenitors.(33,34) Although areal measures of BMD are clearly linked to mechanical strength, it is apparent that areal BMD is a complex phenotype. Bone size may contribute to the relationship between areal BMD and bone strength. Therefore, another objective of these studies was to examine the hypothesis that the genetic determinants of BMD and FCSA may be shared, at least in part.



All mice used in these experiments were bred under identical conditions at the Portland Veterans Affairs (VA) Veterinary Medical Unit from breeding stock originally obtained from the The Jackson Laboratory (Bar Harbor, ME, USA) no more than three generations before this work. At the time of weaning, the mice were group-housed (two to five animals per cage) and maintained with ad libitum water and laboratory rodent chow (Diet 5001: 23% protein, 10% fat, 0.95% calcium, and 0.67% phosphorus; PMI Feeds, Inc., St. Louis, MO, USA) in a 12 h light/dark cycle (6:00 a.m. to 6:00 p.m.) at 21 ± 2°C. Mice from the two progenitor strains B6 and D2 and 18 of the BXD RI strains were available in sufficient number (n ≥ 10 of each gender) for testing. B6D2F1 mice were bred locally from The Jackson Laboratory parental lines (B6 females X D2 males; B6D2) and intercrossed to generate a total of 994 B6D2F2 mice. Of this F2 population, 786 mice (393 female and 393 male) were all of the progeny from consecutive B6D2 F1 X F1 litters and the remaining 208 mice (all female) were obtained at the same time from additional B6D2 F1 X F1 litters from which the male progeny had been assigned to another researcher. Assuming a minimum heritability for a given quantitative trait locus (QTL) at 4% and a power of 0.9, the sample size to attain significance (p = 5 × 10−5) using selective genotyping (vide infra) was estimated to be ∼1000 as described by Soller et al.(35) and Lander and Botstein.(36) All procedures were approved by the VA Institutional Animal Care and Use Committee and performed in accordance with National Institutes of Health (NIH) guidelines for the care and use of animals in research.

Femoral geometry

All mice were studied at 4 months of age when the acquisition of adult bone mass is complete.(37) Mice were killed by CO2 inhalation and weighed to the nearest 0.1 g. The left femurs were harvested immediately. The femoral length was measured with vernier calipers, and the midpoint on the anterior surface was marked with indelible ink. Then, the femur was sectioned at midshaft using an Isomet low-speed saw. Next, the distal section was embedded in clay with the condyles down; the cut surface was exposed. The cut section then was imaged with a high-resolution video camera and the data were digitized and entered into a digital image analysis program (Optimas, Inc., Seattle, WA, USA) to determine FCSA (bone plus marrow area enclosed within the periosteal surface). Measurements of FCSA with this optical method strongly correlated (r = 0.94) with those obtained with X-ray microtomographic scanning (model 1074; SkyScan, Aartselaar, Belgium).

Polymerase chain reaction genotyping

Genomic DNA was isolated from individual mouse spleens using a salting-out method.(38) Mice were genotyped with microsatellite markers chosen from the Massachusetts Institute of Technology database ( and the Mouse Genome Database ( All polymerase chain reaction (PCR) primers were purchased from Research Genetics (Huntsville, AL, USA). Amplification was performed on a Perkin-Elmer 9700 thermocycler (Perkin-Elmer Cetus, Branchburg, NJ, USA). Reaction mixture was 25 μl of total volume, consisting of ∼150 ng of genomic DNA, 264 nM of both forward and reverse primers, 0.2 mM of each deoxynucleoside triphosphate (dNTP), 1 U of Taq polymerase (Perkin-Elmer Cetus), 2.5 μl of GeneAmp 10× PCR buffer containing 100 mM of Tris-HCl (pH 8.3), 15 mM of MgCl2, 500 mM of KCl, and 0.01% (wt/vol) gelatin. Thermal cycling included two 5-minute denaturation steps at 95°C and then at 80°C; 40 cycles of 30 s at 94°C, 30 s at 53°C, and 30 s at 72°C; and a final extension step for 10 minutes at 72°C. PCR products were separated on 4% agarose gels and visualized with ethidium bromide staining.

Data analysis

Phenotype values for all strains are presented as the mean ± SEM. ANOVA was used to detect significant strain differences and provide an estimate of heritability. In the B6D2F2 experiment, only the extreme scoring of 15% at each tail of the trait distribution for each gender were genotyped as a cost-saving measure. The effect of genotype on phenotype was assessed at each individual marker within the male and female B6D2F2 data sets by one-way ANOVA. The B6D2F2 genotyping data were also subjected to analysis using the Map Manager QT (β-version 28) computer program ( to determine the position of the peak logarithm of the likelihood for linkage (LOD) and 1 LOD support intervals.(39) For the chromosome X linkage analysis, the backcross (rather than intercross) option in Map Manager QT was used to deal with the fact that the female F2 mice can only be B6B6 or B6D2 for chromosome X markers and the males can only be B6Y or D2Y. A gender difference was ascribed when the LOD score for a phenotype-genotype correlation differed by three or more between the two genders, which is equivalent to about p < 0.001. Individual p values for the BXD RI QTL evaluations were determined using Student's t-test (one-tailed). StatView statistical software for the Macintosh and Systat for the DOS environment(40) were used to perform all other statistics. This same B6D2F2 population was also used to map whole body BMD.(33,34) Although derived from the same F1 parents, the mice genotyped previously for BMD were largely different from the mice genotyped in this experiment for FCSA. Therefore, we did not adjust our significance thresholds to account for increased type I errors caused by multiple traits because the mice genotyped were largely independent samples.

The criterion for statistical significance was that recommended by Lander and Kruglyak(41) for an F2 population (p = 5 × 10−5), which presumes a near infinite number of mice and markers in a full genome search, and the permutation test(42) implemented in the Map Manager QT program, which estimates an appropriate significance threshold based on the finite numbers of markers and mice actually used in our study. The latter was p = 2.6 × 10−4, a much more relaxed criterion. The difference between the two criteria is caused by the fact that our search did not fully cover the genome and that we used finite numbers of mice and markers. The permutation test has the advantage that is not sensitive to departures from a normal trait distribution, and it is specific to the finite nature of the experiments reported in this study.


Shown in Table 1 are body weight and FCSA values for groups of skeletally mature, 16-week-old B6 and D2 mice. As expected, male mice exhibited heavier body weights than female mice, but there was no progenitor strain difference in body weight. However, as has been established in a previous report,(14) FCSA was ∼50% greater in B6 mice compared with D2 mice and a gender difference in was also observed in both progenitor inbred strains.

Table Table 1.. Comparison of Body Weight and FCSA in Male and Female DBA/2 (D2) and C57BL/6 (B6) Progenitor Strains
original image

Next, we examined FCSA in an F2 population derived from the B6 and D2 progenitor strains (B6D2F2). Again, male mice exhibited greater body weight than female mice (33.1 ± 0.2 g vs. 26.5 ± 0.1 g; p < 0.0001) and larger FCSA (1.769 ± 0.213 mm2 vs. 1.522 ± 0.141 mm2; p = 6.5 × 10−85). Similar to that observed in the progenitor strains (data not shown), a relationship between body weight and FCSA was observed in both male and female F2 mice (male mice, r2 = 0.13 and p = 2.5 × 10−13; female mice, r2 = 0.24 and p = 4.2 × 10−36). Mathematically varying the body weight value had little impact on the correlation and thus for simplicity we used unadjusted weight values to correct the individual F2 FCSA values for the impact of weight (weight-corrected FCSA [WC-FCSA]). In both genders, the distributions of WC-FCSA were continuous, inferring polygenic control of this skeletal trait in both male and female mice (Fig. 1). Tests of skewness and kurtosis showed slight but significant departures from normality. To deal with this, we used the permutation test to establish significance thresholds, which are not sensitive to departures from normality.(42) As noted in the following paragraphs, this had no effect on the significance of any QTLs reported here.

Figure FIG. 1.

Distribution of WC-FCSA in female and male populations of B6D2F2 mice. The mean WC-FCSA ± SD values for female (n = 574) and male (n = 390) F2 mice were 1.522 ± 0.141 mm2 and 1.769 ± 0.213 mm2, respectively (p = 6.5 × 10−85). The animals from each tail of the frequency distribution that were used for selective genotyping are indicated by hatched bars.

To identify the putative location of relevant FCSA-related QTLs in male and female mice, we carried out a whole genome scan of the two B6D2F2 populations with 115 selected microsatellite markers distributed across the 19 autosomes and the X chromosome at an average spacing of 15 cM. Thirty percent of each F2 population (15% from each tail of the frequency distribution; Fig. 1) was genotyped. Interval mapping of data from the two F2 populations identified four loci residing within chromosomes 6, 8, 10, and X in which associations between genotype and WC-FCSA were present in both genders (Table 2). In each instance, a large difference in mean WC-FCSA values was apparent between mice homozygous for the B6 allele versus those homozygous for the D2 allele, as was the approximate intermediate value of heterozygous mice (no dominance). Evidence for gender specificity at three other chromosomal loci was also found (chromosomes 2, 7, and 12; Table 2). Analysis of the male F2 population revealed a significant association between genotype and phenotype at D2Mit81 that was not present in female F2 mice, whereas significant associations between genotype and phenotype at D7Mit297 and D12Mit219 were detected in female F2 mice that were absent in the male population. At all seven QTLs, the B6 allele of a gene (or genes) in the identified chromosomal regions contributed to increased FCSA.

Table Table 2.. Candidate QTL Sites for WC-FCSA in Male and Female B6D2F2 Mice
original image

To determine the peak logarithm of the likelihood for linkage (or LOD score) and 1 LOD support intervals for the QTLs on chromosomes 6, 8, 10, and X, the combined (male and female) B6D2F2 data set was analyzed with the Map Manager QT software program(39) (Fig. 2). For the chromosome 2, 7, and 12 QTLs, interval mapping was performed on the male and female B6D2F2 data sets separately (Fig. 3). In each instance, the unconstrained model for QTL effects, which entailed no a priori assumptions regarding mode of inheritance, is shown. These linkage studies indicated that the levels of statistical significance required to confirm linkage proposed by Lander and Kruglyak(41) of LOD score ≥4.3 were attained for all seven FCSA-related QTLs. Currently, guidelines for the determination of gender divergence have not to be established by the genetics community. However, using our predefined criterion for gender difference (LOD score difference >3), the chromosomes 2, 7, and 12 QTLs plainly exhibited gender specificity (Fig. 3).

Figure FIG. 2.

LOD plots of gender-independent WC-FCSA QTLs on chromosomes 6, 8, 10, and X as determined by an interval mapping approach (Map Manager QT). In each case, the LOD curves exceeded the Lander and Kruglyak(41) significance threshold of 4.3 (df = 2) for F2 data. For each plot, the results for the “additive” and “free” QTL models were virtually identical; therefore, only the free model is shown in each case. The chromosomal positions of the individual markers(51) used for mapping are indicated at the top of each graph. The 1 LOD CIs for each QTL are represented by the black horizontal bar located just above the abscissa of each graph.

Figure FIG. 3.

LOD plots of gender-dependent WC-FCSA QTLs on chromosomes 2, 7, and 12 as determined by an interval mapping approach (Map Manager QT). For each plot, the results for the “additive” and “free” QTL models were virtually identical; therefore, only the free model is shown in each case. The chromosomal positions of the individual markers(51) used for mapping are indicated at the top of each graph. The 1 LOD CIs for each QTL are represented by the black horizontal bar located just above the abscissa of each graph. In female mice, the LOD curves exceeded the Lander and Kruglyak(41) significance threshold of 4.3 (df = 2) for the chromosome 7 and 12 QTLs and for chromosome 2 in male mice.

These results definitely indicated that FCSA is a heritable trait in laboratory mice. To better quantify the degree of heritability of this skeletal phenotype, we extended our studies to a panel of recombinant inbred strains (BXD RI) derived from the same B6 and D2 strains. As expected from the prior analyses of these progenitor strains and their F2 population, a highly significant relationship between body weight and FCSA also existed in the BXD RI population (male mice, r2 = 0.15 and p = 7.3 × 10−8; female mice, r2 = 0.12 and p = 1.6 × 10−6). Therefore, the individual BXD RI FCSA values were corrected for body weight by regression residuals (WC-FCSA). Figure 4 shows the strain mean WC-FCSA values plotted for each gender. WC-FCSA varied by 63% across the male BXD RI panel, ranging from 1.384 ± 0.163 mm2 in BXD-24 to 2.259 ± 0.161 mm2 in BXD-1. WC-FCSA varied by 58% across the female BXD RI panel, ranging from 1.155 ± 0.053 mm2 in BXD-24 to 1.820 ± 0.077 mm2 in BXD-1. Consistent with our findings in the F2 populations, the BXD RI strain distribution for WC-FCSA was also gender specific. Two-way ANOVA of RI strain results from male and female mice combined revealed a strain-by-gender interaction (F1,17,320 = 11.4; p < 0.0001). The CV for the individual strain mean WC-FCSA values (a measure of the environmental variance) ranged from 3% to 9% for female mice and 3% to 13% for male mice. We also estimated the split-half reliability coefficient based on the correlation of odd with even numbered subjects for strain means, which indexes the reliability or consistency of measurement of the phenotype (genotypic differences). The correlation between the two halves of the data using the Spearman-Brown formula was 0.992 (p < 0.0001) for male RI mice and 0.989 (p < 0.0001) for female RI mice, thus showing that WC-FCSA is a very reliable phenotype in these inbred strains of laboratory mice. Heritability (h2) or the proportion of the total variation due to additive genetic sources can be estimated from the adjusted r2 from a one-way ANOVA by strain. The estimated narrow sense heritability for WC-FCSA in the male RI series was 0.75 (F17,175 = 31.34, p < 0.0001) and 0.70 (F17,179 = 25.04, p < 0.0001) in the female RI series, thus showing a substantial degree of genetic control of this trait in both genders. Of note, the estimated narrow sense heritabilities for FCSA (without adjusting for the influence of body weight) essentially were unchanged at 0.76 (F17,175 = 33.12; p < 0.0001) and 0.69 (F17,179 = 24.21; p < 0.0001) for male and female BXD-RI mice, respectively.

Figure FIG. 4.

BXD RI strain means (±SEM) for WC-FCSA in male and female BXD RI mice. Femoral areal measurements were performed on animals as described in the Materials and Methods section. Eighteen BXD RI strains were studied. Ten female mice and 9–10 male mice from each RI strain were examined for a total of 354 mice (180 female and 176 male mice). Missing strains either did not exist because they became extinct during the inbreeding process (e.g., BXD-3) or were poor breeders (e.g., BXD-25) so that insufficient numbers were available for testing.

Although the number of RI strains (i.e., genotypes) available for these studies was insufficient for formal QTL analysis, we did explore the BXD RI strain set for supporting evidence of phenotype-genotype associations at each of the seven genomic regions suggested by the F2 analysis. As is shown in Table 3, statistical support for the FCSA QTLs on chromosomes 2, 7, 8, 10, and 12 was also present in the RI strains. Moreover, the gender specificity of the three QTLs on chromosomes 2, 7, and 12 paralleled that observed in the F2 analyses. No support for the putative chromosome 6 and X QTLs was derived from the RI analyses, but in the latter case only 3 of the 18 strains examined possessed D2 genomic material at the region of interest. The preponderance of B6 genomic material on chromosome X was a consequence of the particular mating strategy used in generating the BXD RI strain set.(43)

Table Table 3.. Mean WC-FCSA Values (Mean ± SEM) for BXD-RI Strains With the Two Possible Genotypes (Homozygous B6 or Homozygous D2) for a Microsatellite Marker Within Each QTL Region for Chromosomes 2, 7, 8, 10, and 12
original image

Previously, we identified five chromosomal regions that influence whole body BMD in mice derived from the same B6 and D2 progenitors.(33,34) However, we observed little or no correlation between WC-BMD and FCSA values in either the B6D2F2 (male mice, r2 = 0.03; p = 3.1 × 10−4; female mice, r2 = 0.03; p = 6.0 × 10−5) or BXD RI (male mice, r2 = 0.00; p = 0.83; female mice, r2 = 0.01; p = 0.33) populations. Consequently, it was not surprising that no overlap was observed between the chromosomal regions associated with peak whole body BMD (distal chromosome 1, midchromosome 2, proximal chromosome 7, and midchromosome 11) and those associated with FCSA (proximal chromosome 2, middistal chromosome 6, midchromosome 7, middistal chromosome 8, distal chromosome 10, proximal chromosome 12, and midchromosome X).


Several skeletal characteristics, including BMD, other material properties, and structure, are important determinants of fracture resistance. Clearly, BMD is a heritable trait, and bone size and certain bone material properties have also been suggested to be influenced by genetic factors. In these studies we have used complementary approaches to confirm the strong heritability of WC-FCSA, to identify chromosomal regions associated with WC-FCSA, and to show that the genetic contributions to WC-FCSA are to some extent gender specific.

There is considerable individual variability in skeletal dimensions. Although some of this variability may be environmental, previous reports of differences in bone size between inbred strains of mice show the importance of genetic factors as well.(9,10,12,14) Here, we confirm and extend that information with the finding that the narrow sense heritability of FCSA, after correction for body mass, is ≈70% in both male and female BXD RI mice. The large genetic contribution to the determination of bone size has important implications. First, size has an important effect on the biomechanical properties of bone. The diameter of bone is exponentially related to strength and so small increases in size may have a major influence on breaking strength. If the magnitude of the genetic component of bone size in the mouse is mirrored in humans, the tendency of osteoporotic fractures to be familial may be in part the result of the heritability of bone size. Second, because size has an important effect on areal measures of BMD,(44,45) the heritability of bone density measures could be to some extent the effect of the genetic determinants of size. This illustrates the complex nature of BMD as a phenotype and the need to examine its components individually.(46)

QTL analysis of a B6D2F2 population revealed seven chromosomal sites (chromosomes 2, 6, 7, 8, 10, 12, and X) that were strongly correlated with WC-FCSA. Importantly, these chromosomal regions were distinct from those we previously reported to be related to whole body BMD,(33,34) suggesting that these two important phenotypes are, to a large extent, influenced by independent genetic mechanisms. This is the first report of QTLs for femoral cross-sectional area, and this result should provide direction in the effort to identify genes that influence bone size. It can be expected that other structural phenotypes will also be influenced by genetic factors, potentially via other genes acting on those specific traits. Indeed, in a study of 16-month-old female B6D2F2 mice, Drake et al.(17) identified significant linkage between femoral length and loci on proximal chromosome 3 and between femoral diaphyseal width and loci on midchromosome 7. Given the important contribution of cross-sectional width to cross-sectional areal measures, it is interesting to note that the WC-FCSA chromosome 7 QTL we have identified in female B6D2F2 mice overlaps the femoral diaphyseal width QTL. However, the genome scan for femoral diaphyseal width did not identify linkage with other FCSA QTLs (chromosomes 6, 8, 10, 12, and X), perhaps because of important experimental differences in skeletal phenotype (linear vs. areal measure) or environment (animal age and diet) between the two studies.

Based on the Map Manager QT software analysis, we estimate that the QTLs on chromosomes 2, 6, 7, 8, 10, 12, and X accounted for as much as 19, 7, 10, 10, 6, 10, and 18%, respectively, of the variance in WC-FCSA in the B6D2F2 populations. Taking into consideration the fact that only the extreme ends of our F2 population were genotyped, we recognize that these are likely to be overestimates but together could explain much of the genetic variance predicted from the RI analyses.

Three of the seven QTLs associated with bone size showed a marked gender dependence. A region of chromosome 2 was related to WC-FCSA in male mice only, whereas in female mice WC-FCSA was associated with an area on chromosomes 7 and 12. Moreover, a region on chromosome X was associated with bone size in both genders. Previously, we reported an effect of gender on genetic determinants of BMD,(34) and a similar effect of gender on FCSA suggests that gender exerts a widespread influence on the genetics of skeletal phenotypes. This result is consistent with the obvious difference in bone size between male and female mice. Gender differences in size have been suggested to be the result of several influences, including (among others) the differential actions of sex steroids on periosteal bone growth; differences in muscle strength, physical activity, and biomechanical stress; or distinct effects of growth factor action. The presence of a gender-based genetic effect may be mediated via influences on these hormonal or biomechanical mechanisms, but may also be the result of independent pathways. Clearly, these effects of gender must be considered when evaluating the relationship of bone phenotypes and genetics.

Our findings involve the cross-sectional area of the femoral shaft. Although we predict that bone morphology at other anatomical sites is also influenced by genetic factors, the chromosomal regions may be somewhat distinct from those noted here. In addition, the QTLs identified in this model (derived from the B6 and D2 parental strains) may differ from those in mice derived from other genetic stocks. In the case of bone density, QTLs derived from one genetic model have considerable (but not identical) overlap with those found in other models,(11,12,33,34,47,48) and similar results can be expected with structural phenotypes. Similarly, chromosomal regions associated with BMD in mouse models have also been identified in syntenic regions of the human genome, and the QTLs associated with geometry in the mouse should also inform parallel human studies.

Comparative gene mapping in humans and rodents has revealed evidence for substantial conservation of gene order during mammalian evolution. Based on the considerable linkage conservation (estimated to be ∼ 80%) between mouse and human genomes, findings in this animal model system should aid in the identification of specific candidate genes for study in humans. It is interesting to note that the chromosome 2, 6, 8, 10, and 12 QTLs identified here contain genomic regions thought to be homologous with 9q, 17q, 4q, 19p, and 7q regions of the human genome, respectively. These five regions were among the seven areas recently reported by Koller and colleagues to be associated with femoral structural phenotypes in an autosomal genome screen of 309 white sister pairs.(49) The mouse and human X chromosomes bear striking linkage homology, but unfortunately Koller's genome screen did not extend to the X chromosome. As of yet, linkage studies have identified relatively broad chromosomal regions that contain QTLs regulating bone size. Additional research, for instance using congenic strains bearing chromosomal regions of interest,(50) should allow these regions to be narrowed, thus making more feasible the identification and testing of candidate genes.

In summary, FCSA is highly heritable in mice, and we have identified chromosomal regions strongly related to femoral size. Four regions are related to cross-sectional area in both male and female mice, and three additional chromosomal segments appear to exert their influence in a gender-specific manner. Cross-sectional area is an important determinant of bone strength and thus genetic factors may affect fracture risk via an influence on bone size.


This work was supported by funds from the NIH (AA 10760 and AR 44659) and the Medical Research Service of the VA.