| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Medicine (M.P., T.Fi., S.K., S.H., C.C.J., M.J.E.) and Medical and Molecular Genetics (D.L.K., D.L., T.Fo., M.J.E.), Indiana University School of Medicine, Indianapolis, Indiana 46202-5250
Address all correspondence and requests for reprints to: Munro Peacock, University Hospital and Out Patient Center, 550 North University Boulevard, Room 5595, Indianapolis, Indiana 46202-5250. E-mail: mpeacock{at}iupui.edu (Endocrinology); dkoller{at}iupui.edu (Genetics); tidings{at}iupui.edu (Endocrinology); subkrish{at}iupui.edu (Bioinformatics); dlai{at}iupui.edu (Bioinformatics); shui{at}iupui.edu (Biostatistics); cjohnsto{at}iupui.edu (Endocrinology); tforoud{at}iupui.edu (Genetics); or mecons{at}iupui.edu (Endocrinology).
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Methods: BMD at the spine and hip were measured in 323 pairs of brothers aged 18–61 yr (264 white pairs; 59 black pairs). Heritability was calculated and linkage analysis performed with spine and hip BMD phenotypes.
Results: Heritability estimates ranged from 0.61 to 0.87 and were not significantly different between white and black men. A 9-cM genome-wide scan followed by genotyping with more closely spaced markers identified suggestive QTL (logarithm of the odds > 2.2) for BMD on chromosomes 1q (spine), 2p (spine), 2q (hip), 14p (spine), 18 (hip), and 21 (hip). Comparison with published data in 774 pairs of premenopausal sisters suggested that the QTL on 1q (spine), 2q (hip), 14p (spine), and 21q (hip) were male specific, whereas those on 2p (spine) and 18 (hip) were not sex specific.
Conclusions: This study demonstrates that BMD in healthy men is highly heritable with similar estimates of the genetic contribution to BMD in both whites and blacks. Of the six QTL identified, three were specific for spine BMD and three were specific for hip BMD. When compared with published QTL for peak BMD in women from the same geographical region, four of the QTL appeared to be male specific. The occurrence of sex-specific genes in humans for BMD has potentially important implications for the pathogenesis and treatment of osteoporosis.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It is generally considered that peak BMD is a polygenic trait and that age-related osteoporosis is a complex disease, with both genetic and environmental determinants (7). In women, heritability of peak BMD is high (7). In men, however, there are few comparable studies. Thus, it has not been established in men whether the heritability of peak BMD is different from that estimated in premenopausal women or whether the heritability of BMD is similar in white and black men.
Studies in both recombinant inbred mice (8) and congenic mice (9) strongly suggest that many of the genes underlying bone mass in the mouse are sex specific. In addition, because studies in humans suggest that heritability of BMD is higher in mother-daughter and father-son pairs than across sexes (10, 11, 12), it is likely that studies in men and women will identify sex-specific quantitative trait loci (QTL) for BMD. Sex-specific genes may underlie the marked differences in skeletal size and structure that are observed between men and women. The presence of sex-specific genes may also imply that there are sex-specific variations in the response of bone to both environment and medications aimed at preventing and treating osteoporosis and other metabolic bone diseases.
The aims of this study were to establish the heritability of peak BMD in both white and black healthy men; perform a genome wide scan to identify QTL underlying BMD; and examine whether some of these loci may be sex specific by comparing the QTL identified in men with the previously published QTL from a sample of healthy premenopausal sister pairs drawn from the same geographical location (13).
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The sample of men comprised 264 pairs of healthy white and 59 pairs of healthy black full-brother pairs, aged 18–61 yr, recruited from 1999 to 2003 (Table 1
). Subjects were recruited by advertising. Exclusion criteria included a history of chronic disease, taking medications known to affect bone mass or metabolism, inability to have BMD measured because of obesity (136 kg), and abnormal blood tests assessed by safety biochemistry. No limit was placed on the difference in years between pairs. The studies were performed on the General Clinical Research Center at Indiana University. A blood sample was collected for DNA, health and lifestyle questionnaires administered, and anthropometrics obtained and BMD measured. In addition, blood was collected for DNA from one or both parents, where possible. Informed written consent, approved by Indiana University-Purdue University Indianapolis and Clarian Institutional Review Board, was obtained from all subjects. The sample of women comprised 774 pairs of healthy premenopausal white and black sister pairs from the same geographical region of the United States as the men (13). Recruitment methods and exclusion criteria were the same as for the men except that more than 10 yr of age between pairs was an exclusion criterion in sisters. Women on oral contraceptives were included, and use of these was examined as a covariate. In 19 of the families studied, there were both brother and sister pairs.
|
BMD was measured by dual-energy x-ray absorptiometry (DXA) (DPXL; Lunar Corp., Madison, WI) at lumbar spine (vertebrae 2–4) and hip (femoral neck, trochanter, and Ward’s area). Brothers were measured on the same machine, usually at the same visit. Image analysis was performed using software 4.6/4.7 (Lunar Corp.). Coefficient of variation measured in 115 pairs of sisters who had duplicate DXA measures made after they were repositioned on the machine was 1.0% for femoral neck, 1.75% for trochanter, and 2.4% for Ward’s area. Height and weight were measured with the subject nonfasting, gowned, and without shoes using a Harpenden stadiometer and a weighing scale (Scale-Tronix, White Plains, NY), which were regularly calibrated throughout the study.
DNA and marker genotyping
DNA was isolated using standard techniques and stored in a DNA repository. A genome scan was performed at the Center for Inherited Disease Research using automated fluorescent microsatellite analysis. PCR products were sized on an ABI 3700 sequencer (PE Applied Biosystems, Foster City, CA). The marker set used was a modification of the CHLC version 9 marker set with 402 markers, at average spacing of 9 cM and average heterozygosity of 0.76 (www.cidr.jhmi.edu). A total of 251,652 genotypes were produced from 626 samples. The error rate, based on 22,513 paired genotypes from duplicate samples, was 0.05%. The percentage genotypic data missing was 3.53. The marker genotype data were used to verify full sibling relationships among subjects using the computer programs RELATIVE (14) and RELPAIR (15), and half-sibling pairs (n = 13) were eliminated from further analysis.
Genetic analysis
Stepwise regression analysis was employed on BMD using height, weight, pack-years of smoking, and age to identify significant covariates with BMD. Weight and age, which were significant predictors of BMD (P < 0.10) in regression model fitting, were used as covariates in the genetic modeling. Race-specific regression residuals, representing covariate-adjusted BMD values, were computed and used in all analysis.
Heritability of hip and spine BMD were calculated from the sibling pair data using the variance components modeling as implemented in the computer program SOLAR (16). Differences in heritability estimates between whites and blacks were tested by transforming the calculated heritability values into normal deviates using Fisher’s z transformation (17). The statistical comparison was then made using a standard z test.
Multipoint quantitative linkage analysis was performed for the BMD phenotypes using the maximum likelihood variance estimation method as implemented in the computer package Mapmaker/SIBS (18). Chromosomal positions, marker order, and map positions were obtained from the Marshfield electronic database (www.marshmed.org/genetics/). Logarithm of the odds (LOD) scores were computed at 1 cM intervals along each autosome and the X chromosome. In families with more than two brothers, all possible sibling pairs were used in the analysis. Observed allele frequencies in the individuals genotyped for the genome screen were used. To confirm the robustness of linkage, analyses were also performed using only independent sibling pairs and implementing the more conservative Haseman-Elston regression approach (19).
Follow-up genotyping
To further examine the linked chromosomal region identified in the genome wide scan of this sample of 323 brother pairs, denser microsatellite genotyping was performed in chromosomal regions producing suggestive linkage (LOD > 2.2). Microsatellite markers were selected based on size, heterozygosity, and position, with approximately 5 cM marker coverage of BMD linkage regions. The Marshfield electronic database (http://research.marshfieldclinic.org/genetics/) was used to determine chromosomal position, map position, and order of the genetic marker. Primer sequences were found on The Genome Database (http://www.gdb.org), and PCR was performed using multiplex PCR kits (QIAGEN, Valencia, CA), according to the kit microsatellite amplification protocol. Nine to eleven markers were amplified in each multiplex reaction, and products were sized on an ABI 3100 genetic analyzer (Applied Biosystems). Data were analyzed using Genotyper software (version 3.6; Applied Biosystems) and read independently by two observers. Linkage analysis was performed as described above.
Evaluation of sex-specific QTL
Premenopausal sister pairs (n = 774) from the same geographical region in the United States (Indiana) were studied under an identical protocol (13). A genome wide scan was performed by the Center for Inherited Disease Research using an identical panel of microsatellite markers. QTL for peak BMD were identified using the spine (20) and hip (13) BMD. Genotypic and phenotypic data from the brother and sister sample were combined to perform a joint linkage analysis in chromosomal regions demonstrating linkage in the brother sample. To correct for BMD differences between men and women and blacks and whites, race-specific regression residuals were computed separately for each sex. These sex- and race-specific residuals, representing the degree to which an individual’s BMD differed from the mean for her/his sex, were employed as the phenotype in the combined male/female linkage analysis.
To formally test the hypothesis that QTL identified in the brothers’ genome screen were sex specific, the traditional Haseman-Elston regression test for linkage (19) was extended. This test uses phenotypic data, such as BMD, and marker genotype from each sibling pair. In each sibling pair, the difference in BMD is calculated and then squared. The genotypic marker information is used to compute the number of alleles identical by descent (IBD) that is shared by the sibling pair. Alleles are IBD if siblings inherit the same allele from the same parent. Because each sibling has inherited two alleles at a marker locus for each autosomal chromosome, a sibling pair can share 0, 1, or 2 alleles IBD.
The Haseman-Elston regression test is based on the principle that if the marker being tested is in close physical proximity to a gene influencing BMD, then siblings with similar BMD would be expected to share marker alleles IBD near the BMD locus. Conversely, siblings with dissimilar BMD would be expected to share few marker alleles IBD near a BMD locus. The following original formulation of the Haseman-Elston linkage test employs a regression approach to examine the relationship between the difference in BMD between sibling pairs and the extent of marker IBD allele sharing:
 | (1) |
We have modified the Haseman-Elston model to employ this approach to test for sex-specific effects of a QTL. The key change to the regression formula is the addition of a variable pair type, which indicates whether the sibling pair is a male/male or female/female pair. The model now becomes:
 | (2) |
The test for sex-specific QTL was performed only at QTL in which linkage was detected in brother pairs. Multipoint estimates for IBD were obtained for each sibling pair at the QTL to be tested using the Mapmaker/SIBS software package (18). The analysis of covariance model, equation 2
described above, is fitted using the generalized linear models procedure in the SAS software package (version 9, 2004; SAS Institute, Cary, NC). A Bonferroni adjustment for multiple comparisons was used to set the significance level of the P values for the test of sex specificity, taking the number of independent chromosome regions tested into account.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. There was no difference in mean height, weight, and age between American white and black men. BMD at spine and hip were higher (P < 0.001) in blacks than whites. The sample of white brothers arose from 185 families and the black brothers from 40 families (Table 2). A blood sample to extract DNA was collected from 136 parents of the brother pairs. No phenotypic information was obtained from these parents. The genotypic data from the parents were used in the linkage analysis to improve the ability to accurately estimate marker allele sharing among brother pairs.
|
). There were no significant differences in heritability estimates between white and black men (all P > 0.30).
|
). Follow-up genotyping using more closely spaced markers in the QTL regions increased the LOD score in five of the regions. For spine BMD, the maximum LOD score on chromosome 1q increased from 3.1 to 3.9, chromosome 2p increased from 3.1 to 3.7, and chromosome 14p increased from 4.6 to 4.8. Evidence of linkage for hip BMD increased on chromosome 18p from 2.4 to 3.0 (femoral neck), chromosome 21p increased from 2.6 to 2.8 (femoral neck), from 2.8 to 3.1 (Ward’s area), and 2.2 to 2.8 (trochanter). Follow-up genotyping of markers at the QTL on chromosome 2q did not appreciably change the LOD scores for femoral neck and Ward’s area. Evidence of linkage on chromosome 1q, 2p, and 14p was with spine BMD and chromosome 2q, 18q, and 21 was with hip BMD. Similar evidence of linkage was observed in these regions when employing independent sibling pairs and the Haseman-Elston regression approach (19). The linkage findings were primarily due to variation in bone mineral content. LOD scores for bone area itself were uniformly low (0.6 or less) in all chromosomal regions of interest.
|
and Fig. 2). On the other hand, QTL on chromosomes 2p (spine, P = 0.02), and 18 (hip, P = 0.09) were not male specific (Table 4).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This genome wide scan in men identified six QTL whose maximum LOD score was greater than 2.2, suggesting linkage with BMD. Furthermore, genotyping using more closely spaced markers continued to support linkage these six QTL with an increase in LOD score in five chromosomal regions and no change in the LOD score in one region. These findings support the concept that BMD is a polygenic trait (7), although there is also evidence from segregation analysis that BMD may be largely controlled by a major gene (22, 23).
On chromosomes 1q, 2p, and 14, the maximum LOD scores were greater than 3.6, the level considered as significant evidence of linkage (21). Three QTL were specific for spine BMD and three QTL were specific for hip BMD. Considering the high correlation between BMD at hip and spine in healthy subjects (7), it is perhaps surprising that so many of the QTL are specific to only the hip or spine. However, independent QTL for hip and spine have also been found in studies of premenopausal women (13, 20). This may imply that the high correlation between skeletal sites for BMD in humans is determined more by universal environmental factors acting throughout life, such as exercise and diet, than by common genes.
In men, two suggestive QTL were identified in a study of 29 two- and three-generation Mexican-American families using a genome wide scan (24). On chromosome 2p, a QTL was identified linked to hip BMD. This QTL is close to the QTL we found on chromosome 2p. However, in our study, evidence of linkage to this region was identified with spine BMD, which was not measured in the Mexican-American study. We did not detect evidence of linkage with any of the three hip BMD phenotypes that were measured in our sample of brother pairs. The second QTL identified in the study of Mexican-American men was linked to hip BMD on chromosome 13q. We found no equivalent QTL at that position. The discrepancies in the linkage results between the studies probably relate to differences in racial composition and age ranges. Our sample of men had a mean age of 34 yr and a range of 18–61 yr, whereas the Mexican-American men were older, with an average age of 42 yr, and had a wider age range, from 18 to 96 yr. Thus, the BMD in Mexican-American men may also be influenced by bone loss.
Criteria for identifying sex-specific QTL have not been established. In one study of inbred mice, the investigators defined sex-specific effects as 3 orders of magnitude differences in the P values (8). In congenic mouse studies, sex-specific effects were determined by ANOVA (9). In our study we used a relatively novel approach based on the Haseman-Elston test of linkage for sibling pair data. We extended this model to include a term representing an interaction between the sex of the sibling pair members and genetic sharing at marker loci. The significance of this interaction term has been examined for all chromosomal regions, demonstrating evidence of linkage in our sample of brother pairs. Significance criteria for the test of sex specificity were obtained using a Bonferroni correction, taking the number of chromosomal regions tested into account. Using these criteria, we concluded that four of the six QTL identified in our study were male specific. In this study because the sample size in our brother pairs is only about a third of that in our sister pairs, the test for sex-specific effect on QTL was performed only at QTL positions in which linkage was detected by linkage analysis of brother pairs. Thus, we are not in a position at this time to test for sex-specific QTL in our sample of sisters.
The proportion of sex-specific QTL is perhaps surprising because it is generally assumed that most of the genes underlying a complex trait are the same in both sexes. However, studies in humans suggest that heritability of BMD is higher in mother-daughter and father-son pairs than across sexes (10, 11, 12). Importantly, studies in mice indicate that there are more sex-specific QTL than sex-shared QTL. In studies of recombinant strains of mice, it was found that nine QTL for total body bone mass were in males and seven in females (8). However, only two of the QTL were shared between the two sexes, and five QTL were specific to female mice and six QTL were male specific. In the studies with congenic mice, five QTL were identified that influenced femoral strength. Of these, two imparted sex-specific effects on femoral structure and one on BMD and bone strength. To investigate whether any of the QTL we identified in men were syntenic with those in mice (25), we compared our QTL with the published QTL in mice (8, 9). The male-specific QTL on chromosome 18 for hip BMD in our sample of brothers was syntenic with the male-specific QTL reported by Orwoll and his colleagues (8).
QTL for hip BMD previously reported in women on chromosomes 14q and15q (13) were not detected in our sample of men. Similarly, the QTL on chromosome 1q for spine BMD in premenopausal women (20) does not appear to be the same QTL as the one we identified in men in this study because the linked regions only partially overlap.
The study has several limitations. The sample size in the brothers is relatively small, and the current findings need to be corroborated in both a larger sample and other populations. Furthermore, with a larger sample size of men, the QTL identified in women can also be tested for sex specificity. Most linkage studies have used BMD measured by DXA as the phenotype of interest because it is readily available in large populations, it accounts for about 70% of bone strength, and it predicts risk of fracture. However, DXA provides areal density of combined cortical and trabecular bone. Thus, it may be a surrogate for the phenotype that is truly under genetic control. Thus, corroborative genetic studies are needed using phenotypes of volumetric density of cortical and trabecular bone and bone structure.
In conclusion, we found that heritability of peak BMD in men is high and that in a sample of 323 brother pairs, six suggestive QTL underlying BMD have been identified. Using sister pairs from the same region of the United States who had an identical genome wide scan, it appears that four of these QTL may be male specific.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online March 1, 2005
Abbreviations: BMD, Bone mineral density; DXA, dual-energy x-ray absorptiometry; IBD, identical by descent; LOD, logarithm of the odds; QTL, quantitative trait loci.
Received November 1, 2004.
Accepted February 18, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C.-L. Cheung, P. C. Sham, V. Chan, A. D. Paterson, K. D. K. Luk, and A. W. C. Kung Identification of LTBP2 on Chromosome 14q as a Novel Candidate Gene for Bone Mineral Density Variation and Fracture Risk Association J. Clin. Endocrinol. Metab., November 1, 2008; 93(11): 4448 - 4455. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Edderkaoui, D. J. Baylink, W. G. Beamer, J. E. Wergedal, R. Porte, A. Chaudhuri, and S. Mohan Identification of mouse Duffy Antigen Receptor for Chemokines (Darc) as a BMD QTL gene Genome Res., May 1, 2007; 17(5): 577 - 585. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
