Polymorphisms in the CLCN7 Gene Modulate Bone Density in ...

0021-972X/06/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 91(3):995–1000 Copyright © 2006 by The Endocrine Society doi: 10.1210/jc.2005-2017

Polymorphisms in the CLCN7 Gene Modulate Bone Density in Postmenopausal Women and in Patients with Autosomal Dominant Osteopetrosis Type II U. Kornak, A. Ostertag, S. Branger, O. Benichou, and M.-C. de Vernejoul Max Planck Institute for Molecular Genetics and Institute for Medical Genetics (U.K.), Charite´ University Hospital, D-13353 Berlin, Germany; Institut National de la Sante´ et de la Recherche Me´dicale U 606 (A.O., S.B., O.B., M.-C.d.V.) and Fe´de´rationde Rhumatologie (M.-C.d.V.), Hoˆpital Lariboisie`re, 75010 Paris, France Context: Genetic factors are important determinants of bone mineral density (BMD). The fact that mutations in the ClC-7 chloride channel cause autosomal dominant osteopetrosis (ADOII) make the CLCN7 gene an attractive candidate for the regulation of bone density.

allele were investigated for association with the variability of the ADOII phenotype. Results: Analysis by multiple linear regression revealed a significant association between the ss genotype of the VNTR and higher Z-score values (P ⫽ 0.029). The haplotype 4, which comprises the long allele of the VNTR, was found to be significantly associated with lower femoral neck Z-score values (P ⫽ 0.011). Furthermore, we found an association of the ss genotype of the VNTR with lower levels of the bone resorption marker D-Pyr/Crea (P ⫽ 0.015), whereas haplotype 4 was associated with higher D-Pyr/Crea levels (P ⫽ 0.039). In the ADOII family, we could demonstrate that haplotype 3, which contains the s-allele of the VNTR, is associated with a slightly higher probability that mutation carriers develop osteopetrosis (P ⫽ 0.029). In both cases the association seems largely to be driven by the VNTR genotype but is further strengthened if surrounding polymorphisms are added to the analysis.

Objective: The objective of the study was to investigate the association between polymorphisms in the CLCN7 gene and BMD in postmenopausal women and with clinical variability in ADOII. Design: This was a genetic association study using five single-nucleotide polymorphisms and a variable number tandem repeat (VNTR) polymorphism in the CLCN7 gene. Participants: A total of 425 postmenopausal women aged 64 ⫾ 7 yr participated in the study. We also investigated an ADOII family with low penetrance comprising 18 mutation carriers. Main Outcome Measure(s): In our postmenopausal cohort, individual single-nucleotide polymorphism genotypes and haplotypes were analyzed for association with BMD at the lumbar spine and the femoral neck and with the bone resorption marker deoxypyridinoline (D-Pyr/Crea). The same polymorphisms on the nonmutated CLCN7

Conclusion: We observed a significant association of CLCN7 polymorphisms with the variance of BMD and bone resorption marker levels in postmenopausal women and with the variability of the ADOII phenotype. (J Clin Endocrinol Metab 91: 995–1000, 2006)

O

STEOPOROSIS IS A multifactorial disorder of growing importance due to the increasing life expectancy in developed countries. Bone mass is decreased during the ageing process because bone remodeling gets out of balance. In young healthy individuals, bone tissue that is resorbed by the osteoclast is replaced by the bone-forming osteoblast during the remodeling cycle. In the older individual, however, resorption overrides bone formation resulting in a reduced number of trabeculae and an expansion of the cortical bone (1). A large number of endogenous (PTH, vitamin D3, kidney function, etc.) and exogenous factors (physical activity, calcium intake, smoking, etc.) are known to influence the coordinated action of osteoblasts and osteoclasts in humans (reviewed in Ref. 1). Investigation of the variation of bone mineral density (BMD) in families and twin studies suggest

that the heritability of the peak bone mass is 60 – 80%. Many association studies investigating polymorphisms in candidate genes have revealed several genetic factors underlying variability of BMD and/or fracture rate (2). Osteoclasts produce a low pH microenvironment at the contact surface between the cell and the bone matrix in the so-called resorption lacuna. It has become clear recently that this process relies on the cooperation of a v-ATPase complex and the chloride channel ClC-7 (3). ClC-7 is a broadly expressed member of the CLC chloride channel family and resides mainly in late endosomes, lysosomes, and the ruffled membrane of osteoclasts (3). Different groups have identified CLCN7 mutations in autosomal recessive osteopetrosis and autosomal dominant osteopetrosis type II (ADOII) (4 – 6). The different phenotypes can be explained by a reduction of ClC-7 chloride channel function of a varying degree. Hence, in ADOII osteoclast function is only mildly impaired, leading to a rather benign increase in bone mass, whereas in ARO bone resorption seems to be completely abolished. The fact that the penetrance of the ADOII phenotype is incomplete (7, 8) indicates that there are modifying factors that determine whether mutation carriers become symptomatic. We hypothesized that polymorphisms in CLCN7 can lead

First Published Online December 20, 2005 Abbreviations: ADOII, Autosomal dominant osteopetrosis type II; BMD, bone mineral density; D-Pyr/Crea, deoxypyridinoline; FN, femoral neck; HRT, hormone replacement therapy; l, long; LS, lumbar spine; s, short; SNP, single-nucleotide polymorphism; VNTR, variable number tandem repeat. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

995

996

J Clin Endocrinol Metab, March 2006, 91(3):995–1000

Kornak et al. • CLCN7 Polymorphisms and Bone Density

to a mild alteration of chloride channel function, which could either produce a beneficial higher peak bone mass or aggravate the phenotype in osteopetrosis. To test this hypothesis, we investigated a panel of six different polymorphisms in CLCN7 in a large family with ADOII and postmenopausal women. Our results demonstrate a weak but significant association with BMD in both groups.

with other families affected with ADOII, the phenotype was exceptional because of low radiological penetrance: 47% of the mutation carriers presented with segmentary osteosclerosis and 11% with a high bone density measured by dual-energy x-ray absorptiometry without segmentary osteosclerosis on x-rays. Nine affected carriers, nine nonaffected carriers, and 14 noncarriers from this family were available for genotyping.

Patients and Methods BMD association study

rs3751884 and rs12926089 were detected by restriction digestion of PCR products by SmaI and MslI, respectively. rs2235580, rs2235579, and rs2294542 were analyzed using the Snapshot method (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. Primers used for amplification by a standard PCR protocol and probes for Snapshot primer extensions are given in Table 2. Length of the DNA fragments was determined by electrophoresis on 2.5% agarose gels and fragment length analysis using an ABI 3100 capillary sequencer (Applied Biosystems). Linkage disequilibrium was analyzed by the Haploview software to determine haplotype blocks (http://www.broad.mit.edu/mpg/haploview). Individual haplotypes were calculated using the FAMHAP software (10).

For the BMD association study, 425 Caucasian women were recruited and investigated at a single clinical center (Hoˆpital Lariboisie`re, Paris, France). Two hundred eight-two women were selected from a population of 500 postmenopausal women already being monitored to evaluate determinant factors for bone loss (Viggos study). We included women between 50 and 80 yr of age who agreed with DNA sampling. Individuals who had been treated with corticosteroids for more than a month, bisphosphonates, or fluoride were excluded. Thirty percent of the included individuals had a T score less than ⫺2.5 at the femoral neck (FN) or lumbar spine (LS). One hundred forty-three postmenopausal women between 50 to 80 yr of age were referred to our outpatient clinic for osteoporosis. These patients underwent clinical, biochemical, and radiological assessment to exclude secondary causes of osteoporosis. All of these included individuals had a T score less than ⫺2.5 at the FN or LS. All 425 women were interviewed to evaluate risk factors for osteoporosis (personal and familial history of fractures, physical exercise, nutritional habits) and had an x-ray of the spine. Vertebral crush fractures were defined by a decrease in height of 20% or greater. This population is described in Table 1. BMD was measured at the FN and the LS (L2-L4) using a DPX-L instrument (Lunar Corp., Madison, WI). The instrument was calibrated on a daily basis. T-score and Z-score values were based on a French reference population obtained from several centers (provided by Lunar France). Urinary excretion of free deoxypyridinoline was measured by an automated chemiluminescent immunoassay (Pyrilinks-D Immulite; Diagnostic Products Corp., Los Angeles, CA). To avoid intraindividual variability, samplings were all performed on first morning void, at the same time in the morning (⫾ 1 h) on fasting subjects. Informed consent for genotyping of osteoporosis candidate genes was obtained from all participants, and the study was approved by the Ethics Committee of the Hoˆpital Lariboisie`re in Paris.

Samples from ADOII family The mutation 797fsX925 in family K affected with ADOII and the phenotypical findings had been reported before (6, 9). In brief, compared TABLE 1. Demographic, densitometric, and biochemical marker data of 425 postmenopausal study participants Age (yr) Weight (kg) Height (cm) BMI (kg/m2) Age at menopause (yr) Duration of menopause (yr) HRT users (%) Duration of HRT use (yr) LS BMD (g/cm2) Z-score FN BMD (g/cm2) Z-score Fractures Vertebral (% of subjects) Colles (% of subjects) FN (% of subjects) D-Pyr/creat (nmol/mmol)

64.0 (7.1) 60.3 (9.3) 158.0 (6.0) 24.1 (3.5) 49.2 (5.7) 15.2 (8.6) 43.3 8.2 (5.9) 0.964 (0.183) 0.302 (1.577) 0.743 (0.121) 0.029 (1.035) 22.8 8.5 2.1 5.537 (2.353)

Genotyping

Statistical methods Results were expressed as mean ⫾ sd. The means were compared by ANOVA for the quantitative data. All tests were two sided, and the significance level was adjusted for multiple testing according to the recommendations given by Sankoh et al. (11). Taking into account the average correlation of the six polymorphisms (R2 ⫽ 0.54), we determined the level for significance as P ⬍ 0.032 by the Armitage and Parmar method. A modified Bonferroni correction as proposed by Holm was used to determine the significance level for the association of haplotypes with Z-score or BMD. Using a multiple regression model, we evaluated the Z-score [adjusted for hormone replacement therapy (HRT) use and duration, as a continuous response data] as a function of the genotype (included as binary variable) and three environment variables (weight, height, and duration of menopause as quantitative variables). The model included the four factors and their combined effects (first-order interaction terms). To select the best-fitting and most parsimonious subsets, we conducted a stepwise model selection. Diagnostic tests were performed to detect nonnormality and outliers. Statistical analysis was performed using S-PLUS 6.1 software program for Windows (S-PLUS 6 for Windows, Insightful Corp., Seattle, WA.).

Results

Because the LD structure of the CLCN7 genomic region as revealed by HapMap data (http://www.hapmap.org) did not allow to identify any tagging single-nucleotide polymorphisms (SNPs), we chose six equally spaced sufficiently informative sequence variants (Table 2). Only SNP rs12926089 leads to an amino acid change (V418M). Whereas five polymorphisms are SNPs, one is a 50-bp variable number tandem repeat (VNTR) residing in intron 8. In the general population, the number of the repeat units ranges between two and nine; the most frequent alleles contain three or five repeat units. We defined alleles harboring two or three repeat units as short (s) and those with four to nine repeats as long (l) to simplify the statistical analysis. A similar approach was taken by Becherini et al. (12) in the analysis of a TA polymorphism in the ESR1 gene. All polymorphisms did not deviate significantly from the Hardy-Weinberg equilibrium (P ⬎ 0.05). We were not able to detect any significant association of the genotype of any individual polymorphism with Z-score

Kornak et al. • CLCN7 Polymorphisms and Bone Density

J Clin Endocrinol Metab, March 2006, 91(3):995–1000

997

TABLE 2. Analyzed sequence polymorphisms in the CLCN7 gene Name

Position

Localization

Polymorphism

rs3751884 rs2235580 Intron 8 VNTR rs2235579 rs12926089 rs2294542

1464851 1455163 1447617 1443880 1442858 1435898

Exon 1 Intron 1 Intron 8 Exon 14 Exon 15 3⬘ UTR

A/G C/T 50-bp repeat A/T C/T A/G

AA change

none none V418M

Frequency (%)

50 50 30 48 10 40

The position within the gene and the frequency of the rare allele are given. UTR, Untranslated region.

corrected only for the use and duration of HRT (data not shown). The lowest P value was observed for the association of the intron 8 VNTR length with FN Z-score (P ⫽ 0.15) (Table 3). Multiple linear regression analysis of the Z-score was performed using weight, height, duration of menopause, and CLCN7 polymorphisms as covariables. Using this method, the intron 8 genotype could be shown to have an impact on FN Z-score (P ⫽ 0.028), whereas no association of the intron 8 VNTR with LS Z-score could be detected. The linear regression model was highly significant (P ⬍ 0.0001) and predicted that 1% of the observed Z-score variance can be explained by the intron 8 genotype (see Table 5). Furthermore, individuals with the ss genotype of the intron 8 VNTR had significantly lower urine levels of the bone resorption marker deoxypyridinoline (D-Pyr/Crea) than the remaining individuals (P ⫽ 0.015) (Fig. 1). Analysis of the LD structure revealed that rs2235580, intron 8 VNTR, rs2235579, and rs12926089 are in linkage disequilibrium (rs2235580-rs12926089, R2 ⫽ 0.92, ␹2 ⫽ 50.2). Four haplotypes with a frequency of more than 5% were deduced that explained 84% of the genotypes found in our postmenopausal patient sample. None of these haplotypes showed a significant association with Z-score without any further correction (Table 4). In the multiple regression analysis, however, haplotype 4 (T-l-A-C) could be shown to have a significant association with FN Z-score (P ⫽ 0.011) and accounted for 1.3% of its variance (Table 5). Additionally, individuals with haplotype 4 on one or both alleles could be shown to have higher D-Pyr/Crea levels (P ⫽ 0.039). This corresponds well to the observed tendency that the intron 8 ss genotype entails higher FN bone density, whereas haplotype 4, which comprises the l-allele of the intron 8 VNTR, is a risk factor for decreased FN bone density. As previously reported, the heterozygous CLCN7 mutation 797fsX925 in family K gives rise to a rather mild ADOII phenotype, presumably because the C-terminally elongated channel protein is still functional (6). Nine mutation carriers from this family clearly showed radiographic signs of os-

teopetrosis (either sandwich vertebrae or diffuse osteosclerosis), whereas the other nine carries were unaffected (9). Eighteen mutation carriers and 14 noncarrier DNAs were available for genotyping. Because of the existing linkage disequilibrium among markers rs2235580, intron 8 VNTR, rs2235579, and rs12926089, haplotype analysis was performed. Of the 13 individuals carrying haplotype 3 (T-s-A-C), nine were affected, whereas all of the haplotype 2 (C-l-T-C) carriers were unaffected (P ⫽ 0.029) (Fig. 2). One nonaffected person showed haplotype 5 (C-l-T-T), which comprises the V418M polymorphism. No difference existed between the frequency of the haplotypes of nonmutated alleles from unaffected carriers and noncarriers (haplotype 3: four of nine vs. 11 of 28; P ⫽ 0.7). This suggests that the polymorphisms on the nonmutated allele might partially contribute to the susceptibility for the disease. As seen in the osteoporosis cohort, the haplotype comprising the intron 8 s-allele seems to be associated with an increase in BMD, although the contribution of the individual tested markers cannot be disentangled due to their mutual correlation. Discussion

We here report that a VNTR polymorphism in intron 8 of the CLCN7 gene is associated with FN Z-score and with lower levels of the bone resorption marker deoxypyridino-

TABLE 3. Association of intron 8 VNTR genotype with Z-score adjusted for HRT use Intron 8

LS Z-score (g/cm2) FN Z-score (g/cm2)

ss (n ⫽ 70)

sl and ll (n ⫽ 353)

P value

0.402 (1.714)

0.257 (1.551)

0.49

0.184 (1.117)

⫺0.014 (1.016)

0.15

Association of the intron 8 genotype with Z-score of the FN and the LS as assessed by ANOVA analysis (quantitative data). The postmenopausal individuals were subdivided in two groups according to their intron 8 genotype.

FIG. 1. D-Pyr/Crea urine levels in relation to the intron 8 VNTR genotype (left) or haplotype 4 (right). Urines were collected on a fasting urine sample in the morning. We compared subjects with genotype ss of intron 8 and the other genotypes by ANOVA, *, P ⫽ 0.015; and subjects without haplotype 4 with all other genotypes, *, P ⫽ 0.039. The proportion of subjects treated with HRT was not significantly different between both subgroups either for intron 8 or haplotype 4.

998

J Clin Endocrinol Metab, March 2006, 91(3):995–1000

Kornak et al. • CLCN7 Polymorphisms and Bone Density

TABLE 4. Association of the four most common CLCN7 haplotypes with Z-score of the FN and LS adjusted for HRT use in postmenopausal women Genotype

Haplotype 1/1 1/⫺ ⫺/⫺ Haplotype 2/2 2/⫺ ⫺/⫺ Haplotype 3/3 3/⫺ ⫺/⫺ Haplotype 4/4 4/⫺ ⫺/⫺

FN

n

LS

Mean (SD)

P value

Mean (SD)

P value

1 59 363

⫺0.749 (NA) 0.000 (1.267) 0.035 (0.997)

0.73

0.188 (NA) 0.152 (1.638) 0.332 (1.572)

0.72

53 182 188

⫺0.119 (0.912) ⫺0.001 (0.968) 0.098 (1.130)

0.37

0.168 (1.615) 0.347 (1.500) 0.306 (1.647)

0.77

54 158 211

0.192 (1.114) 0.015 (1.013) ⫺0.005 (1.035)

0.46

0.349 (1.575) 0.380 (1.579) 0.241 (1.584)

0.70

8 94 321

0.300 (1.603) 0.003 (1.039) 0.029 (1.023)

0.74

0.013 (2.110) 0.211 (1.643) 0.342 (1.549)

0.68

1 (C-l-A-C)

2 (C-l-T-C)

3 (T-s-A-C)

4 (T-l-A-C)

Data are presented by mean (SD). P value tests the difference between each haplotype assessed by ANOVA (quantitative data).

line (D-Pyr/Crea) in postmenopausal women. This association is stronger if haplotype 4 consisting of three SNPs surrounding the intronic VNTR is analyzed. Approximately 1.3% of the observed FN Z-score variance can be explained by the presence of haplotype 4. Furthermore, CLCN7 polymorphisms influence the probability that a carrier of a mild CLCN7 mutation becomes affected with ADOII. The putative chloride channel ClC-7 has been shown to be of crucial importance for the function of osteoclasts (3). A 50% reduction of ClC-7 expression, however, is not sufficient to considerably alter bone density in heterozygous human and murine carriers of ClC-7 loss-of-function mutations. In ADOII, however, heterozygous dominant mutations further lower the dosage of functional ClC-7 channel protein dimers, thereby reaching the threshold at which ClC-7 becomes rate limiting for the resorption process (13). At this threshold any further disturbance of ClC-7 expression, e.g. by polymorphic sequence changes, could result in changes in BMD. Individuals showing the intron 8 ss genotype have a slightly higher FN Z-score, whereas haplotype 4, which contains the intron 8 l-allele, is associated with lower bone density. These findings suggest that the association of CLCN7 with BMD is largely driven by the intron 8 genotype. The analysis of haplotype 4, however, is justified because it decreases the P value so that the association remains significant

even after a very conservative Bonferroni correction for multiple testing. Additionally, the fraction of the FN Z-score variance explained by CLCN7 sequence changes is higher for haplotype 4 than for intron 8 alone. Furthermore, we were able to assess only a dominant effect of haplotype 4 due to the limited number of homozygous individuals in our sample. It can be speculated that the observed effect on BMD might be even stronger if a sufficiently high number of homozygous haplotype 4 carriers is investigated. Surprisingly, no significant association of CLCN7 polymorphisms or haplotypes with LS Z-score could be detected. This is in contrast with the increased vertebral bone density observed in ADOII (14). In postmenopausal osteoporosis bone loss is related to an increase in bone resorption, which is reflected by elevated levels of bone resorption markers (15). In line with our finding that intron 8 and haplotype 4 genotypes are associated with Z-score values, we could show that there was an association between the intron 8 VNTR length and haplotype 4 with a marker of bone resorption (D-Pyr/Crea). It seems therefore likely that the higher BMD seen in individuals with the ss genotype or without haplotype 4, respectively, stems from a reduced bone resorption activity, which is compatible with the physiological role of ClC-7. Due to the globally reduced osteoclast activity, ADOII

TABLE 5. Multiple linear regression model for Z-score (BMD adjusted for age and HRT use) including weight, height, duration of menopause and the genotype as covariates FN

Intron 8 Haplotype Haplotype Haplotype Haplotype

1 2 3 4

LS

R2a

P valueb

Genetic fraction (%)

R2a

P valueb

Genetic fraction (%)

0.245 0.226 0.226 0.226 0.246

⬍0.05 NSd NSd NSd ⬍0.05

1 0 0 0 1.3

0.122 0.117 0.117 0.133 0.117

NSd NSd NSd NSc NSd

0 0 0 nd 0

R2 is the coefficient for the correlation of the respective genotype. Genetic fraction denotes the fraction of the observed Z-score variance that can be attributed to the genotype. a R2 of multiple linear regression model including covariates as weight, height, duration of menopause, and genetic variable. b P value of genetic variable. c Genetic variable was included in the final model but not significantly by itself. d Genetic variable was not included in the final model.

Kornak et al. • CLCN7 Polymorphisms and Bone Density

FIG. 2. A, Pedigree showing all CLCN7 mutation carriers from family K. The haplotypes of the nonmutation-carrying allele are given inside the symbols. The distribution of haplotypes in affected (f) and nonaffected (䡺) individuals was significantly different (P ⫽ 0.029) as assessed by Fisher’s exact test. B, Representative x-rays of three members of family K. II-1 displays typical sandwich vertebrae, whereas a diffuse sclerosis is seen in II-2. The LS of mutation carrier III-6 appears completely normal.

normally has a segmental radiological appearance (7, 8). In contrast, a diffuse bone sclerosis is seen in autosomal dominant osteopetrosis type I, which results from mutations in the LRP5 gene that regulates osteoblast differentiation (16). Similar to the findings in LRP5, the ADOII family described here demonstrates that the phenotype of patients with heterozygous mutations in CLCN7 can consist of high bone density only (Fig. 2) (9). Furthermore, polymorphisms in LRP5 turned out to be associated with higher BMD in a population sample (17). It has been reported by other groups that the penetrance of ADOII can be incomplete (4). In our ADOII family only 50% of the mutation carriers are affected. To understand the variability of the phenotype in this family, we asked whether the nonmutation-carrying CLCN7 allele could be associated with the expression of the disorder. The frequency of the SNPs rs2235580, rs2235579, and rs12926089 and the s-allele of the intron 8 VNTR were significantly different in affected and nonaffected carriers. There are, however, three unaffected individuals with a mutated allele and a haplotype 3 allele showing that the nonmutated CLCN7 allele might be only one factor among others influencing penetrance. Chu et al. (7) found suggestive linkage to a modifier locus on chromosome 9, but due to the low number of tested individuals, we could not test for further modifying loci. In an earlier work, Frattini et al. (5) did not find any

J Clin Endocrinol Metab, March 2006, 91(3):995–1000

999

association of expression of the ADOII phenotype and intron 8 repeat length alone by comparing carrier parents with more severely affected siblings. We cannot exclude that in our ADOII family the haplotype 2 allele might contain another so-far-unknown sequence variation that limits ClC-7 function. Because this family is of the same ethnic background as the postmenopausal women investigated in our association study, this could be true for these individuals as well. V418M (rs12926089) is the nonsynonymous polymorphism with the highest frequency (10%) in the CLCN7 gene. Although V418M was described elsewhere to be associated with BMD in the general population, we could not detect any significant effect, which could be due to the size of our sample (18). Our study had a power of 30% to detect a significant difference between subgroups. Furthermore, Campos-Xavier et al. (8) as well as Chu et al. (7) found an association of V418M with the severity of the ADOII phenotype in several families. We could not reproduce this finding because only one of the 18 mutation carriers of our ADOII family harbored this polymorphism (Fig. 2). This person was not affected. On the other hand, our study and the data of Pettersson et al. (18) demonstrate that the intron 8 VNTR and V418M are in linkage disequilibrium so that the contributions of both polymorphisms can hardly be separated from each other. The mechanism by which the intronic VNTR might influence the function of the CLCN7 gene product (ClC-7) is elusive. It is remarkable that each repeat unit contains the complete splice acceptor consensus sequence of the intron 8/exon 9 border. We could not detect any alternative splice products by RT-PCR (data not shown). Furthermore, this VNTR is a predicted CpG island that could have regulatory function. No repetitive elements can be found in intron 8 of the murine ortholog, thus precluding an analysis in the mouse. Further work will be needed to specify whether the VNTR exerts any direct influence or whether another polymorphism within the haplotype block might be causative. Several studies have investigated the association between polymorphisms within candidate genes implicated in osteoclast function and BMD. The gene TNFRSF1B encodes the receptor activator of nuclear factor-␬B receptor that is expressed by osteoclast precursors and governs their differentiation (19). In a population study using a multiple linear regression model, polymorphisms in TNFRSF1B accounted for a similar level of variance of BMD as the CLCN7 polymorphism we report here (20). The gene TCIRG1 encodes the ␣3-subunit of the proton pump protein complex that is functionally linked to ClC-7 in the osteoclast acid secretion mechanism. Two studies (21, 22) investigated whether TCIRG1 is associated with osteoporosis and came to conflicting conclusions. Thus, a growing amount of data indicates that genes essential for osteoclast function are worthwhile candidates for studying the polygenic regulation of bone mass. Association studies on such genes allow for an evaluation of candidate genes that might represent new drug targets. In line with that, a recent report (23) indicates that a putative ClC-7 inhibitor has a beneficial effect on BMD in a rat osteoporosis model. Although the contribution of CLCN7 polymorphisms to BMD variance in postmenopausal women identified in our

1000

J Clin Endocrinol Metab, March 2006, 91(3):995–1000

study is in the same range as that of other tested candidate genes, it is well known that association results can vary greatly, depending on the investigated sample (24). Further replication of our findings in other populations is therefore necessary to finally demonstrate the degree of relevance of CLCN7 for the complex pathogenesis of osteoporosis.

Kornak et al. • CLCN7 Polymorphisms and Bone Density

10. 11. 12.

Acknowledgments We thank Wim van Hul for providing genotype information, Klaus Rohde for help with the statistics, Nico Ruf for performing QASEP analyses, and Bjo¨rn Fischer for his great technical assistance. Received September 9, 2005. Accepted December 8, 2005. Address all correspondence and requests for reprints to: Marie-Christine de Vernejoul, Institut National de la Sante´ et de la Recherche Me´dicale U 606, Hoˆpital Lariboisie`re, 2, rue Ambroise Pare´, 75010 Paris, France. E-mail: [email protected] (U.K.); [email protected] (S.B.); [email protected] (A.O.); [email protected] (O.B.); [email protected] (M.-C.d.V.). U.K. had a Poste Vert position from the Institut National de la Sante´ et de la Recherche Me´dicale to realize part of this work. M.-C.d.V. received a grant Direction de la Recherche Clinique at AP/HP (AOR 01049).

13.

14.

15.

16.

17.

References 1. Rizzoli R, Bonjour JP, Ferrari SL 2001 Osteoporosis, genetics and hormones. J Mol Endocrinol 26:79 –94 2. Ralston SH 2003 Genetic determinants of susceptibility to osteoporosis. Curr Opin Pharmacol 3:286 –290 3. Kornak U, Kasper D, Bosl MR, Kaiser E, Schweizer M, Schulz A, Friedrich W, Delling G, Jentsch TJ 2001 Loss of the ClC-7 chloride channel leads to osteopetrosis in mice and man. Cell 104:205–215 4. Waguespack SG, Koller DL, White KE, Fishburn T, Carn G, Buckwalter KA, Johnson M, Kocisko M, Evans WE, Foroud T, Econs MJ 2003 Chloride channel 7 (ClCN7) gene mutations and autosomal dominant osteopetrosis, type II. J Bone Miner Res 18:1513–1518 5. Frattini A, Pangrazio A, Susani L, Sobacchi C, Mirolo M, Abinun M, Andolina M, Flanagan A, Horwitz EM, Mihci E, Notarangelo LD, Ramenghi U, Teti A, Van Hove J, Vujic D, Young T, Albertini A, Orchard PJ, Vezzoni P, Villa A 2003 Chloride channel ClCN7 mutations are responsible for severe recessive, dominant, and intermediate osteopetrosis. J Bone Miner Res 18: 1740 –1747 6. Cleiren E, Benichou O, Van Hul E, Gram J, Bollerslev J, Singer FR, Beaverson K, Aledo A, Whyte MP, Yoneyama T, deVernejoul MC, Van Hul W 2001 Albers-Schonberg disease (autosomal dominant osteopetrosis, type II) results from mutations in the ClCN7 chloride channel gene. Hum Mol Genet 10: 2861–2867 7. Chu K, Koller DL, Snyder R, Fishburn T, Lai D, Waguespack SG, Foroud T, Econs MJ 2005 Analysis of variation in expression of autosomal dominant osteopetrosis type 2: searching for modifier genes. Bone 37:655– 661 8. Campos-Xavier AB, Casanova JL, Doumaz Y, Feingold J, Munnich A, Cormier-Daire V 2005 Intrafamilial phenotypic variability of osteopetrosis due to chloride channel 7 (CLCN7) mutations. Am J Med Genet A 133:216 –218 9. Benichou OD, Van Hul E Van Hul W, de Vernejoul MC 2001 Exclusion of

18.

19.

20.

21.

22.

23.

24.

the chromosomal 1p21 region in a large pedigree with a phenotypic variant of type II autosomal dominant osteopetrosis. Joint Bone Spine 68:327–333 Becker T, Knapp M 2004 Maximum-likelihood estimation of haplotype frequencies in nuclear families. Genet Epidemiol 27:21–32 Sankoh AJ, Huque MF, Dubey SD 1997 Some comments on frequently used multiple endpoint adjustment methods in clinical trials. Stat Med 16:2529 –2542 Becherini L, Gennari L, Masi L, Mansani R, Massart F, Morelli A, Falchetti A, Gonnelli S, Fiorelli G, Tanini A, Brandi ML 2000 Evidence of a linkage disequilibrium between polymorphisms in the human estrogen receptor ␣ gene and their relationship to bone mass variation in postmenopausal Italian women. Hum Mol Genet 9:2043–2050 Henriksen K, Gram J, Schaller S, Dahl BH, Dziegiel MH, Bollerslev J, Karsdal MA 2004 Characterization of osteoclasts from patients harboring a G215R mutation in ClC-7 causing autosomal dominant osteopetrosis type II. Am J Pathol 164:1537–1545 Benichou OD, Laredo JD, de Vernejoul MC 2000 Type II autosomal dominant osteopetrosis (Albers-Schonberg disease): clinical and radiological manifestations in 42 patients. Bone 26:87–93 Eastell R, Colwell A, Hampton L, Reeve J 1997 Biochemical markers of bone resorption compared with estimates of bone resorption from radiotracer kinetic studies in osteoporosis. J Bone Miner Res 12:59 – 65 Van Wesenbeeck L, Cleiren E, Gram J, Beals RK, Benichou O, Scopelliti D, Key L, Renton T, Bartels C, Gong Y, Warman ML, De Vernejoul MC, Bollerslev J, Van Hul W 2003 Six novel missense mutations in the LDL receptorrelated protein 5 (LRP5) gene in different conditions with an increased bone density. Am J Hum Genet 72:763–771 Ferrari SL, Deutsch S, Choudhury U, Chevalley T, Bonjour JP, Dermitzakis ET, Rizzoli R, Antonarakis SE 2004 Polymorphisms in the low-density lipoprotein receptor-related protein 5 (LRP5) gene are associated with variation in vertebral bone mass, vertebral bone size, and stature in whites. Am J Hum Genet 74:866 – 875 Pettersson U, Albagha OM, Mirolo M, Taranta A, Frattini A, McGuigan FE, Vezzoni P, Teti A, Van Hul W, Reid DM, Villa A, Ralston SH 2005 Polymorphisms of the CLCN7 gene are associated with BMD in women. J Bone Miner Res 20:1960 –1967 Suda T, Takahashi N, Udagawa N, Jimi E, Gillespie MT, Martin TJ 1999 Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. Endocr Rev 20:345–357 Albagha OM, Tasker PN, McGuigan FE, Reid DM, Ralston SH 2002 Linkage disequilibrium between polymorphisms in the human TNFRSF1B gene and their association with bone mass in perimenopausal women. Hum Mol Genet 11:2289 –2295 Carn G, Koller DL, Peacock M, Evans WE, Conneally PM, Johnston Jr CC, Foroud T, Econs MJ 2002 Sibling pair linkage and association studies between peak bone mineral density and the gene locus for the osteoclast-specific subunit (OC116) of the vacuolar proton pump on chromosome 11p12–13. J Clin Endocrinol Metab 87:3819 –3824 Sobacchi C, Vezzoni P, Reid DM, McGuigan FE, Frattini A, Mirolo M, Albhaga OM, Musio A, Villa A, Ralston SH 2004 Association between a polymorphism affecting an AP1 binding site in the promoter of the TCIRG1 gene and bone mass in women. Calcif Tissue Int 74:35– 41 Schaller S, Henriksen K, Sveigaard C, Heegaard AM, Helix N, Stahlhut M, Ovejero MC, Johansen JV, Solberg H, Andersen TL, Hougaard D, Berryman M, Shiodt CB, Sorensen BH, Lichtenberg J, Christophersen P, Foged NT, Delaisse JM, Engsig MT, Karsdal MA 2004 The chloride channel inhibitor n53736 prevents bone resorption in ovariectomized rats without changing bone formation. J Bone Miner Res 19:1144 –1153 Ioannidis JP, Ntzani EE, Trikalinos TA, Contopoulos-Ioannidis DG 2001 Replication validity of genetic association studies. Nat Genet 29:306 –309

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.