Calcified Tissue International

, Volume 73, Issue 3, pp 237–243

A CAG Repeat Polymorphism in the Androgen Receptor Gene is Associated with Reduced Bone Mass and Increased Risk of Osteoporotic Fractures

Authors

    • Department of Endocrinology and MetabolismAarhus University Hospital, Aarhus Amtssygehus, DK-8000 Aarhus
  • L. Stenkjær
    • Department of Endocrinology and MetabolismAarhus University Hospital, Aarhus Amtssygehus, DK-8000 Aarhus
  • M. Carstens
    • Department of Endocrinology and MetabolismAarhus University Hospital, Aarhus Amtssygehus, DK-8000 Aarhus
  • C. L. Tofteng
    • Osteoporosis Research Clinic, Department of EndocrinologyHvidovre Hospital, DK-2650 Hvidovre
  • E. F. Eriksen
    • Department of Endocrinology and MetabolismAarhus University Hospital, Aarhus Amtssygehus, DK-8000 Aarhus
Clinical Investigations

DOI: 10.1007/s00223-002-0019-8

Cite this article as:
Langdahl, B., Stenkjær, L., Carstens, M. et al. Calcif Tissue Int (2003) 73: 237. doi:10.1007/s00223-002-0019-8

Abstract

Osteoporosis is a common disease with a strong genetic component. Hypogonadism results in low bone mass and it increases significantly the risk of osteoporosis in both sexes. The estrogen and androgen receptor genes are therefore strong candidates for mediating the genetic influence on bone mass and risk of osteoporosis. A CAG repeat in the first exon of the androgen receptor (AR) is associated with reduced transcriptional activity of the AR. We therefore examined whether this CAG repeat polymorphism is associated with changes in bone mass and risk of osteoporotic fractures in 284 osteoporotic patients with vertebral fractures and 327 normal individuals. The number of CAG repeats varied between 13 and 30 in men and between 7 and 34 in women. The short and long alleles comprised 19.2 ± 2.5 and 19.0 ± 2.3 repeats (ns) and 22.7 ± 2.4 and 21.9 ± 2.4 (P < 0.01) in women with vertebral fractures and normal women, respectively. This difference was also reflected in the average number of CAG repeats: 21.0 ± 2.0 in osteoporotic women vs. 20.5 ± 2.0 in normal women (P < 0.05). 54.8% of women with osteoporotic fractures vs. 45.9% of normal women had average number of CAG repeats of 21 and more (χ2 = 3.11, P = 0.08). Logistic regression analyses revealed that both the average number of CAG repeats and the length of the long allele were significant predictors of osteoporotic fractures in women (P < 0.05 and P < 0.01, respectively). Men with vertebral fractures had 20.0 ± 2.8 CAG repeats compared with 20.7 ± 2.5 CAG repeats in normal men (ns). Linear regression analysis revealed that the length of the long allele was negatively correlated with BMD of the lumbar spine (P < 0.05) and femoral neck (P < 0.05) in women. In men, linear regression analyses demonstrated that BMD of the lumbar spine (P < 0.05), femoral neck (P < 0.05) and total hip (P < 0.05) was positively correlated with length of the CAG repeat polymorphism. In conclusion, we have demonstrated that the CAG repeat polymorphism in the first exon of the AR gene is associated with reduced bone mass and increased risk of osteoporotic fractures in women.

Keywords

Androgen receptorOsteoporosisBone massPolymorphismBone turnover

Osteoporosis is a common disease affecting one in three women and one in ten men above the age of 50 years. Twin and family studies have revealed that genetic factors are responsible for 50–85% of the inter-individual variation in bone mass [1, 2, 3]. However, the effect is most likely a combined effect of several genes, each with a modest effect [4]. The genes involved can be unraveled by two different approaches; linkage studies in families or candidate gene searches in case-control studies. If the genes involved in the pathogenesis of osteoporosis are many, they may be difficult to detect in linkage studies, whereas the candidate gene approach is more powerful [5].

Hypogonadism results in low bone mass and it significantly increases the risk of osteoporosis in both sexes. The estrogen and the androgen receptor genes are therefore obvious candidates for mediating the genetic influence on bone mass and risk of osteoporosis. Previously, we and others have reported associations between polymorphisms in the estrogen receptor gene and reduced bone mass and increased risk of osteoporotic fractures [6, 7, 8]. The first exon of the androgen receptor (AR) harbors a CAG repeat polymorphism which causes a polyglutamine tract with length range of 8–33 repeats [9]. The CAG repeat polymorphism has been demonstrated to be correlated with transcriptional activity of the AR gene. Short repeat sequence was associated with an increase in the capacity of the receptor to activate transcription of reporter genes in vitro [10, 11, 12], whereas binding of androgen to the AR was unaffected [13]. Racial differences in the CAG sequence length have been demonstrated. African Americans have a higher prevalence of short CAG repeat sequences than other ethnic groups [9, 14].

In some studies the CAG repeat polymorphism has previously been found to be associated with male infertility [15, 16], adenomas of the prostate [17], and prostate cancer [18, 19, 20, 21], whereas other studies were not able to demonstrate any associations with infertility or diseases of the prostate gland [22, 23, 24, 25, 26, 27]. In women, the CAG repeat polymorphism has been demonstrated to be associated with endometrial, ovarian and breast cancers [28, 29, 30, 31] and hirsutism [32]. These results suggest that the CAG repeat polymorphism affects the responsivity of cells to androgens either directly or through effects on circulating levels of sex hormones. Furthermore, in some studies long CAG repeat sequences have been demonstrated to be associated with reduced bone mass in both women and men [33, 34], however, other studies revealed no effect of the CAG repeat polymorphism on bone mass or fracture risk [35, 36].

A (GGN)n repeat is also located in the first exon of the AR gene, this repeat encodes a polyglycine tract. The (GGN)n repeat is much less variable than the CAG repeat and has not consistently been associated with androgen-dependent diseases [26, 37].

We therefore investigated the effect of the CAG repeat polymorphism in the first exon of the AR gene on prevalence of osteoporotic fractures and bone mass in men and women.

Material and Methods

Study Subjects

The study was a case control study. The osteoporotic group consisted of 226 women and 58 men with primary spinal osteoporosis, referred to the Department of Endocrinology, Aarhus University Hospital. Primary spinal osteoporosis was defined by the presence of at least one nontraumatic fracture of the spine. The diagnosis was made after extensive examination for secondary causes. A trained radiologist examined X-rays of the thoracic and lumbar spine, and spinal fracture was defined as a 20% or more reduction of the anterior, central or posterior height of a vertebra. The normal control group comprised 255 normal women and 72 normal men without diseases or medications that could influence bone mass and turnover. The normal controls were recruited from the local community by invitations posted at places of work, senior citizens clubs, schools, educational institutions, hospitals, and at office of general practitioners. Characteristics of the osteoporotic patients and normal controls are presented in Table 1. For comparison of length of the CAG repeat sequence between osteoporotic patients and normal controls, we selected age-matched subgroups. The age-matched groups comprised 226 osteoporotic women (mean age: 64.3 ± 9.4 years), 193 normal women (mean age: 65.0 ± 8.6 years), 58 osteoporotic men (mean age: 56.0 ± 13.1 years) and 58 normal men (mean age: 56.5 ± 12.6 years). The study was approved by the local ethical committee and conducted according to Helsinki Declaration II.

Table 1

Characteristics of patients with vertebral fractures and normal controls

 

Osteoporosis

Normal

    
 

Women

Men

Women

Men

    

Number

226

58

255

72

    

Age (years)

64.3 ± 9.4

56.0 ± 13.1

57.9 ± 15.2

51.2 ± 15.8

    

Range

30–82

27–83

20–82

22–78

    

Pre/postmenopausal

18/204

81/174

    

Z-score (lumbar spine)

−1.48 ± 1.15

−2.42 ± 1.27

0.56 ± 1.28

−0.20 ± 1.32

    

Age and Z-score are given as means ± SD

Bone Mass Measurements

BMD of the lumbar spine and the following standard sites at the hip, femoral neck, trochanter, intertrochanteric region and Wards triangle were assessed using dual energy X-ray absorptiometry (DeXA) on a Hologic 1000® or a Norland® bone densitometer. Results obtained on the Norland® densitometer were corrected for the difference between the two densitometers using the correction formulas suggested by Genant et al. [38]. All BMD values are corrected for age and gender.

Biochemical Markers of Bone Turnover

Serum samples were collected in the morning after an 8-hour fasting period. All samples from osteoporotic patients were collected before start of anti-osteoporotic treatment.

Serum cross-linked carboxyterminal telopeptide of type I collagen (s-ICTP) was measured by an equilibrium radioimmunoassay; the intra-assay CV was 5% and the inter-assay CV was 6% [39]. S-osteocalcin was determined by radioimmunoassay using rabbit antiserum against bovine bone-gla-protein [40]. The intra-assay CV was 5% and the inter-assay CV was 10%.

CAG Repeat Polymorphism

DNA was isolated from whole blood leukocytes as described by Kunkel et al. [41]. The CAG-repeat polymorphism was examined by a PCR-based method as previously described [42]. In short, PCR was performed in a final volume of 25 ul containing 100 ng genomic DNA, 150 ng of each primer and AmpliTaqGold polymerase (Perkin Elmer®) using standard conditions on a Perkin Elmer Termocycler 2400®: 32 cycles of 94°C for 1 min., 58°C for 1 min and 72°C for 1 min. Prior to the first cycle, initial denaturation was performed at 94°C for 10 min and the last cycle was followed by an extension step of 8 min at 72°C. Downstream primer = 5′TCCAGAATCTGTTCCAGAGCGTGC and upstream primer 5′GCTGTGAAGGTTGCTGTTCCTCAT. The PCR product was analyzed on a 4.25% PAGE gel using the GeneScan® software package (Perkin Elmer®).

Statistical Analysis

Differences in length of CAG repeat sequence between osteoporotic patients and age-matched normal controls were tested using Students t-test for unpaired data. The effects of the CAG repeat polymorphism on fractures and BMD were examined by logistic and linear regression analyses. Differences in BMD and levels of biochemical markers between individuals with a number of CAG repeats below and above average were tested using ANOVA and Students t-test for unpaired data. The level of significance was set at 0.05.

Results

The androgen receptor gene is located on the X chromosome and men therefore have only one copy of this gene. The number of CAG repeats varied between 13 and 30 in men and between 7 and 34 in women. The cumulative frequency distributions of the CAG repeat polymorphism in osteoporotic and normal men are shown in Figure 1. Men with vertebral fractures had 20.0 ± 2.8 CAG repeats compared with 20.7 ± 2.5 CAG repeats in normal men (ns). Men with a number of CAG repeats below average (≤20) did not have increased risk of osteoporotic fractures (χ2 = 0.90, ns) and logistic regression analysis revealed that number of CAG repeats was not a predictor of osteoporotic fractures in men.

https://static-content.springer.com/image/art%3A10.1007%2Fs00223-002-0019-8/MediaObjects/fig1.gif
Figure 1

The cumulative frequency distribution of the CAG repeat polymorphism in age-matched osteoporotic and normal men.

The cumulative frequency distributions of the short and long alleles and the average length of the two alleles of the CAG repeat polymorphism in osteoporotic and normal women are shown Figure 2. The short allele comprised 19.2 ± 2.5 and 19.0 ± 2.3 repeats in women with vertebral fractures and normal women, respectively (ns). The long allele comprised 22.7 ± 2.4 repeats in osteoporotic women compared with 21.9 ± 2.4 in normal women (P < 0.01). This difference was also reflected in the mean number of CAG repeats on the two alleles: 21.0 ± 2.0 in osteoporotic women compared with 20.5 ± 2.0 in normal women (P < 0.05). A total of 54.8% of women with osteoporotic fractures vs. 45.9% of normal women had a number of CAG repeats above average (≥21) (χ2 = 3.11, P = 0.08). Logistic regression analyses revealed that both the average number of CAG repeats and the length of the long allele were significant predictors of osteoporotic fractures in women (P < 0.05 and P < 0.01, respectively). An increase of 5 CAG repeats in the average number of CAG repeats or in repeats in the long allele increased the risk of osteoporotic fractures by 3.3% and 7.1%, respectively.

https://static-content.springer.com/image/art%3A10.1007%2Fs00223-002-0019-8/MediaObjects/fig2.gif
Figure 2

The cumulative frequency distributions of the short and long alleles and the average length of the two alleles of the CAG repeat polymorphism in age-matched osteoporotic and normal women.

BMD at all measured sites was lower in women with a mean number of CAG repeats above average, however, the differences were not significant (Table 2). If the CAG repeat alleles are divided into two groups, short alleles (CAG ≤ 20) and long alleles (CAG ≥ 21), women can be either homozygous for the short allele (SS) or the long allele (LL) or heterozygous (SL). Women with the SS genotype had higher BMD at all sites investigated (Fig. 3), however, the differences were only significant at the lumbar spine. Linear regression analysis revealed that the length of the long allele was negatively correlated with BMD of the lumbar spine (P < 0.05) and femoral neck (P < 0.05).

Table 2

BMD of the lumbar spine, femoral neck and total hip in osteoporotic patients, normal controls, men and women with mean number of CAG repeats below (Short) or above (Long) average (21 repeats)

  

No

Lumbar spine

Femoral neck

Total hip

     

Women

Short

235

0.904 ± 0.171

0.725 ± 0.116

0.835 ± 0.146

     
 

Long

238

0.878 ± 0.178

0.710 ± 0.120

0.824 ± 0.149

     

T-test

  

0.11

0.19

0.43

     

Men

Short

72

0.773 ± 0.199

0.648 ± 0.132

0.737 ± 0.164

     
 

Long

56

0.836 ± 0.166

0.671 ± 0.109

0.818 ± 0.159

     

T-test

  

0.06

0.29

0.02

     

Osteoporosis

Short

134

0.740 ± 0.148

0.629 ± 0.097

0.700 ± 0.125

     
 

Long

143

0.758 ± 0.122

0.640 ± 0.105

0.722 ± 0.124

     

T-test

  

0.29

0.35

0.19

     

Normal

Short

173

0.976 ± 0.142

0.767 ± 0.109

0.898 ± 0.117

     
 

Long

151

0.977 ± 0.151

0.762 ± 0.100

0.900 ± 0.119

     

T-test

  

0.97

0.68

0.84

     

All

Short

307

0.873 ± 0.186

0.707 ± 0.124

0.815 ± 0.155

     
 

Long

294

0.870 ± 0.176

0.703 ± 0.119

0.823 ± 0.150

     

T-test

  

0.84

0.70

0.57

     

BMD is given as mean ± SD (g/cm2)

https://static-content.springer.com/image/art%3A10.1007%2Fs00223-002-0019-8/MediaObjects/fig3.gif
Figure 3

BMD of the lumbar spine, femoral neck and total hip in women according to CAG repeat length genotypes. S = number of CAG repeats ≤20, L = CAG repeats ≥21. *P < 0.05 versus SS genotype; **P < 0.01 versus SS genotype.

In men, BMD at all sites was higher in men with long CAG repeat alleles. BMD (lumbar spine) was 0.773 ± 0.199 g/cm2 and 0.836 ± 0.166 g/cm2 in men with short and long alleles, respectively (P = 0.06). BMD (total hip) was 0.865 ± 0.201 g/cm2 in men with short alleles and 0.965 ± 0.191 g/cm2 in men with long alleles (P < 0.05). Linear regression analyses demonstrated that BMD of the lumbar spine (P < 0.05), femoral neck (P < 0.05) and total hip (P < 0.05) was positively correlated with length of the CAG repeat polymorphism. No significant differences in BMD could be demonstrated between individuals with short and long CAG repeat alleles when patients with osteoporotic fractures and normal controls were examined separately.

No significant differences in biochemical markers of bone turnover could be demonstrated between individuals with short and those with long CAG repeat alleles (Table 3).

Table 3

Biochemical markers of bone turnover in osteoporotic patients, normal controls, men and women with mean number of CAG repeats below (Short) or above (Long) average (21 repeats)

  

No

S-ICTP

BGP

    

Women

Short

107

3.29 ± 1.18

16.7 ± 7.4

    
 

Long

104

3.31 ± 1.11

17.4 ± 7.7

    

T-test

  

0.90

0.53

    

Men

Short

41

2.86 ± 0.90

14.5 ± 4.7

    
 

Long

33

2.66 ± 0.59

14.2 ± 3.8

    

T-test

  

0.28

0.75

    

Osteoporosis

Short

74

3.99 ± 1.43

19.4 ± 8.8

    
 

Long

84

3.88 ± 1.25

18.2 ± 8.1

    

T-test

  

0.75

0.55

    

Normal

Short

103

2.94 ± 0.89

15.1 ± 5.8

    
 

Long

91

2.87 ± 0.80

16.0 ± 6.6

    

T-test

  

0.60

0.32

    

All

Short

148

3.18 ± 1.13

16.1 ± 6.9

    
 

Long

137

3.15 ± 1.05

16.6 ± 7.1

    

T-test

  

0.84

0.58

    

Biochemical markers are given as mean ± SD (μg/L)

Discussion

All human female cells undergo inactivation of one of their X chromosomes. Normally, inactivation is considered to be random. However, skewed inactivation occurs and seems to be an inherited character linked to specific loci of the X chromosome [43, 44]. Skewed inactivation is defined as inactivation of the same allele in more than 80% of cells examined [45]. Calvo et al. [46] have examined inactivation of the androgen receptor gene in 54 women with hirsutism and 13 normal controls. They found that skewed inactivation of the AR is rare, but more importantly, the frequency of skewed inactivation was similar among patients and controls and the shorter and longer CAG repeat sequence was subjected to skewed inactivation with similar frequency. Based on these results we made the calculations of the effect of the CAG repeat polymorphism on fracture risk, bone mass and biochemical markers of bone turnover in women, assuming random inactivation of the X chromosome with respect to the AR gene.

We found that number of CAG repeats in the long allele and the average number of CAG repeats in the AR were higher in osteoporotic women and that BMD was negatively correlated with average number of repeats and length of the long allele. The effect of this polymorphism in pre- and perimenopausal women has previously been examined by Sowers et al. [33]. Numbers of CAG repeats were smaller in their population than in ours, however, they found the same inverse association between BMD and number of CAG repeats as we did in this study. However, in a Finnish study of perimenopausal healthy women, no association between this polymorphism and bone mass could be demonstrated [36]. Based on the in vitro study by Kazemi-Esfarjani et al. [11] it has been estimated that a decrement of six CAG repeats causes a 12% increase in transcriptional activation. Although the magnitude of the effect of the AR polyglutamine length and transcriptional activation function in vitro may appear relatively modest, these differences may, over a lifetime, have a substantial impact on bone mass. Furthermore, Westberg et al. [47] demonstrated that fewer CAG repeats are associated with higher serum levels of androgens. Both these two mechanisms might therefore be contributing to the positive effect on bone mass and protection against fractures of low numbers of CAG repeats in elderly women.

In this study we could not demonstrate any effect of number of CAG repeats on the risk of osteoporotic fractures in men. However, we found that CAG repeats above average were associated with increased bone mass at both the lumbar spine and the hip. Zitsmann et al. [34] have previously examined the effect of this polymorphism on phalangeal QUS in younger men (<50 years) and found that a high number of CAG repeats was associated with lower QUS values and higher levels of biochemical markers of bone turnover. However, in a cohort of community-dwelling elderly men (≥71 years) Van Pottelbergh et al. [35] found no associations between this polymorphism and BMD or biochemical markers of bone turnover. In studies examining the effects of length of the CAG repeat polymorphism on classic androgen-sensitive diseases such as prostate cancer and male infertility, it has been demonstrated that reduced number of repeats is associated with increased androgenic effect of the receptor [15, 16, 17, 18, 19, 20, 21]. Based on these results it could be expected that low number of repeats would be associated with increased bone mass. However, it has been suggested that androgens are not mediating their effect on bone solely through the androgen receptor; androgens are converted to estrogen and influence bone through the estrogen receptor [48, 49]. Krithivas et al. [50] examined the effect of the CAG repeat polymorphism on serum levels of testosterone in 882 men between 40 and 70 years of age. At the first examination, when the men were 53 years old, no effect of the polymorphisms on testosterone levels could be detected, however, eight years later, it could be demonstrated that men with a low number of repeats had a more rapid age-related decline in serum levels of testosterone compared with men with a high number of repeats. This finding could not be confirmed in the smaller study by van Pottelbergh et al. [35]. A reduced level of testosterone in elderly men could explain the reduced bone mass in men with a low number of repeats. A reduction in serum levels of androgens as response to few CAG repeats has also been suggested as an explanation for the association found between early onset rheumatoid arthritis in men and short length of the CAG repeat sequence [51]. However, we are aware that the power of the results obtained in men in this study is not strong because of the limited number of men included.

In conclusion, we have demonstrated that the CAG repeat polymorphism in the first exon of the AR gene is associated with reduced bone mass and increased risk of osteoporotic fractures in women. Further studies will be needed to clarify whether this polymorphism also affects bone mass and fracture risk in men.

Acknowledgements

The authors thank The Novo Nordisk Fonden, The Institute of Experimental Clinical Research at University of Aarhus, The Fonden til Lægevidenskabens Fremme, and The Eli Lilly Research Foundation.

Copyright information

© Springer-Verlag 2003