Calcified Tissue International

, Volume 84, Issue 3, pp 171–179

LRP5 Polymorphisms and Response to Risedronate Treatment in Osteoporotic Men

Authors

  • Marcin Kruk
    • Molecular Medicine Centre, Institute of Genetics and Molecular Medicine, Western General HospitalUniversity of Edinburgh
  • Stuart H. Ralston
    • Molecular Medicine Centre, Institute of Genetics and Molecular Medicine, Western General HospitalUniversity of Edinburgh
    • Molecular Medicine Centre, Institute of Genetics and Molecular Medicine, Western General HospitalUniversity of Edinburgh
Article

DOI: 10.1007/s00223-008-9207-5

Cite this article as:
Kruk, M., Ralston, S.H. & Albagha, O.M.E. Calcif Tissue Int (2009) 84: 171. doi:10.1007/s00223-008-9207-5

Abstract

Genetic factors are important in the pathogenesis of osteoporosis, but little is known about the genetic determinants of treatment response. Previous studies have shown that polymorphisms of the LRP5 gene are associated with bone mineral density (BMD), but the relationship between LRP5 polymorphisms and response to bisphosphonate treatment in osteoporosis has not been studied. In this study we investigated LRP5 polymorphisms in relation to treatment response in a group of 249 osteoporotic or osteopenic men who participated in a 24-month randomized double blind placebo-controlled trial of risedronate treatment. BMD and biochemical markers of bone turnover were measured at baseline and after 6, 12, and 24 months of follow-up. We analyzed two coding polymorphisms of LRP5, which have previously been associated with BMD, V667M (rs4988321) and A1330V (rs3736228), and found a significant association between the A1330V polymorphism and hip BMD at baseline. Subjects with the 1330 Val/Val genotype had 8.4% higher total-hip BMD compared with the other genotype groups (P = 0.009), and similar associations were observed at the femoral neck (P = 0.01) and trochanter (P = 0.002). There was no association between A1330V and spine BMD, however, or between the V667M polymorphism and BMD at any site. The difference in hip BMD between A1330V genotype groups remained significant throughout the study, but there was no evidence of a genotype–treatment interaction in either risedronate- or placebo-treated patients. In conclusion, the LRP5 A1330V polymorphism is associated with hip BMD in osteoporotic men, but allelic variations in LRP5 do not appear to be associated with response to bisphosphonate treatment.

Keywords

BoneLRP5OsteoporosisPolymorphismBone density

Osteoporosis is a common disease that is characterized by reduced bone mineral density (BMD), deteriorated bone microarchitecture, and an increased risk of fractures. Although the majority of individuals affected with osteoporosis are postmenopausal women, men also suffer from the disease, with >30% of all hip fractures occurring in men. Indeed, the 1-year mortality after a hip fracture is about twice as high in men compared with women [1, 2]. Genetic factors play an important role in osteoporosis and contribute 60–80% of the variance in BMD [35]. The low-density lipoprotein receptor–related protein 5 (LRP5) gene is an essential element of the Wnt signaling pathway and an established candidate for susceptibility to osteoporosis. Loss-of-function mutations cause the rare osteoporosis-pseudoglioma syndrome [68], whereas activating mutations are responsible for autosomal dominant inheritance of high bone mass [912]. Several common polymorphisms have been identified in LRP5 and extensively studied in relation to bone phenotypes such as BMD and fracture risk. A coding polymorphism located in exon 18 (A1330V, rs3736228) has been associated with BMD [1321] and increased fracture risk in both men and women [17, 22, 23]. The V667M (rs4988321) single-nucleotide polymorphism is another widely studied polymorphism located in exon 9 that has also been associated with various osteoporosis-related phenotypes [13, 15, 19, 24]. In many studies, the association between LRP5 polymorphisms and BMD has been particularly evident in men [17, 25], and interactions have been observed with physical activity. This has led to the suggestion that LRP5 alleles might modulate the skeletal response to mechanical loading [1820].

Although common genetic variants in the LRP5 gene predispose to osteoporosis, there is little information on whether or not LRP5 variants affect the response to treatment. This is an important issue clinically since polymorphisms in the vitamin D receptor and COLIA1 genes have previously been reported to influence response to bisphosphonate treatment [2629]. The mechanism by which LRP5 polymorphisms may affect response to bisphosphonate treatment is unclear; however, previous studies have shown that risedronate may modulate Wnt signaling in a mouse model of glucocorticoid-induced osteoporosis [30, 31]. In this study we therefore investigated the relationship between two coding polymorphisms of LRP5 and the response to risedronate therapy in a cohort of osteoporotic or osteopenic men who participated in a randomized double blind placebo-controlled trial.

Materials and Methods

Study Subjects

Subjects comprised a group of 249 men who participated in a randomized double blind placebo-controlled trial of risedronate treatment for osteoporosis and who consented to provide samples for genetic analysis (NIH clinical trial identifier NCT00619957, the results of this study are being prepared for publication elsewhere). The average age of participants (mean ± SD) was 59.7 ± 10.7 years and ranged 36–83 years. All participants had a BMD T score at the femoral neck of −2.0 or below (and a T score of <−1.0 at the spine) or a T score of −2.5 or below at the lumbar spine (and a T score of −1.0 or below at the femoral neck). Individuals with diseases known to affect bone metabolism (primary hyperparathyroidism, pituitary disease, hypogonadism, neoplasia, and thyrotoxicosis) revealed upon clinical examination and laboratory tests were excluded from the study. Subjects who were receiving treatment such as corticosteroid, testosterone, calcitriol, and bisphosphonate were also excluded from the study. Participants were enrolled from 25 centers located in Europe or the United States, and the majority (230/249, 92%) were of Caucasian origin. All participants gave written informed consent to be included in this study, which was approved by the appropriate ethics committees.

Treatment

Patients were randomized to receive 35 mg risedronate once a week or an identical placebo for 24 months and were given a daily supplement of calcium (equivalent to 1,000 mg elemental calcium) and vitamin D (400–500 IU). Treatment was given with 120 ml plain water at least 30 min before the first meal or drink of the day, and participants were instructed not to lie down for 30 min after taking the tablet.

Anthropometric and BMD Measurements

Physical examination of study subjects was performed at baseline and after 6, 12, and 24 months of treatment. Body weight was measured at all visits. Body height was measured at the 12- and 24-month follow-up visits.

BMD measurements were performed at the lumbar spine L1–L4 (LS), femoral neck (FN), trochanter (Troch), and total hip (TH) by dual-energy X-ray absorptiometry (DXA) using Hologic (Bedford, MA) and Lunar (Madison, WI) densitometers. The same machine was used for a given patient throughout the study. Bone densitometric measurements were performed at baseline and after 6, 12, and 24 months of treatment. Measurements were corrected on densitometer drift, if any, by referring to phantom scans, performed on the day of patient’s visit. The in vivo coefficients of variation (CV) for LS and FN measurements were, respectively, 1.1% and 1.5% for the Hologic instrument and 1.2% and 2.0% for the Lunar instrument. DXA measurements were performed and the necessary correction factors calculated in the SYNARC clinical research central laboratory (San Francisco, CA).

Bone Turnover Markers

Blood for standard hematology and biochemistry tests was taken at baseline and all follow-up visits. Specialized biochemical markers of bone turnover were measured on second voided morning urine samples and blood samples collected between 9:00 a.m. and noon. These included serum bone-specific alkaline phosphatase (bALP), serum type I collagen C-telopeptide (CTX-I), and urinary type I collagen N-telopeptide/creatinine (NTX/Creat). Serum bALP was measured using the Tandem-R Ostase kit (Beckman Coulter, Fullerton, CA). Serum CTX-I levels were measured by radioimmunoassay (Orion Diagnostica, Espoo, Finland), and urinary NTX was measured using NTX reagent kits (Vitros ECi; Ortho Clinical Diagnostics, Amersham, UK) and corrected for urinary creatinine excretion. The intra- and interassay CVs were, respectively, <6% and <10% for bALP, <7% and <%10 for serum CTX-I, <8% and <10% for urinary NTX.

Genotyping

DNA was extracted from whole blood using standard methods, quantified, and diluted to a working concentration of 50 ng/μl. Both the A1330V and V667M polymorphisms were genotyped by DNA sequencing using standard automated methods on PCR-amplified fragments of genomic DNA. For V667M, the following primers were used to generate the PCR product: forward 5′-GGTGAGTCCTGAGCTC GGCACC, reverse 5′-GTCTGAAGCCTTTGAGGCAGG. For A1330V, the following primers were used: forward 5′-GCAGCGATGGAGGATGTGCGGGTG, reverse 5′-CAGAGCCCCTACTCCTGTGAGGCC. PCR was carried out in a volume of 25 μl containing 50 ng DNA using Taq DNA polymerase according to the manufacturer’s protocol (Qiagen, Crawley, UK). The following conditions were used: one cycle at 94°C for 3 min, 60°C for 60 s and 72°C for 90 s; 35 cycles at 94°C for 50 s, 60°C for 60 s, and 72°C for 90 s; followed by final extension at 72°C for 10 min. Genotypes were checked by two individuals; genotypes for 25 randomly selected samples were repeated, and no discrepancies were found.

Statistical Analysis

Statistical analyses were performed using SPSS v.14.0 software (SPSS, Inc., Woking, UK). Data are presented as the mean ± SD unless indicated otherwise. For BMD and bone turnover markers, the percentage change between measurements at baseline and follow-up visits were calculated and used in statistical analysis of response to treatment. The χ2 test was used to check for deviation from Hardy–Weinberg equilibrium. Linkage disequilibrium between the two polymorphisms was evaluated using D′ calculated from estimated haplotype frequencies using FAMHAP software (http://famhap.meb.uni-bonn.de/) [32]. Associations between genotype and continuous variables were studied using one-way analysis of variance (ANOVA) or Student’s t-test as appropriate. General linear model ANOVA (GLM-ANOVA) was used to adjust for confounding factors such as age, body weight, height, and baseline BMD. Genotype–treatment interaction analysis was performed in the whole study population using GLM-ANOVA by entering the treatment and genotype as interacting factors in the model. P < 0.05 was considered significant. The power of the study to detect differences in BMD of 0.5 SD units between genotype groups was 97% and 99% assuming a minor allele frequency of 9% (V667M) and 20% (A1330V), respectively.

Results

Characteristics of Study Population

Table 1 shows relevant demographic characteristics of the study population at baseline for the risedronate-treated and placebo groups individually and the whole study population. There was no significant difference between the placebo and treated subjects in terms of baseline values of any measured variable. We found no significant difference between men with different genotypes in relation to age, body weight, or body mass index (BMI) (data not shown). The genotype frequencies of the studied polymorphisms were in Hardy–Weinberg equilibrium, and the distribution did not differ in the two treatment groups (Table 1). After 24 months of treatment, BMD values for the risedronate-treated subjects increased by 5.8% ± 4.3 and 2.0% ± 3.4 from baseline at the LS and FN, respectively, compared to 0.8% ± 3.3 and 0.7% ± 3.4 in the placebo group (LS P < 0.001, FN P = 0.007) (Fig. 1). Additionally, values for biochemical markers of bone turnover were significantly lower in the treated group compared to the placebo group over the course of risedronate treatment (Fig. 1).
Table 1

Characteristics of the study population at baseline

Characteristic

All subjects

Placebo subjects

Treated subjects

Number

249

79

170

Age (years)

59.7 ± 10.7

59.7 ± 10.9

59.7 ± 10.7

Body height (cm)

172.0 ± 7.5

170.7 ± 7.7

172.5 ± 7.3

Body weight (kg)

74.1 ± 13.4

74.2 ± 14.6

74.0 ± 12.9

BMI (kg/m2)

25.0 ± 4.0

25.4 ± 4.0

24.9 ± 4.0

LS BMD (g/cm2)

0.804 ± 0.093

0.815 ± 0.085

0.799 ± 0.096

FN BMD (g/cm2)

0.674 ± 0.078

0.680 ± 0.074

0.671 ± 0.080

Troch BMD (g/cm2)

0.634 ± 0.091

0.636 ± 0.084

0.634 ± 0.094

TH BMD (g/cm2)

0.786 ± 0.094

0.788 ± 0.088

0.786 ± 0.098

bALP (μg/l)

12.6 ± 4.2

12.2 ± 3.7

12.7 ± 4.4

CTX-I (ng/ml)

0.41 ± 0.20

0.41 ± 0.19

0.41 ± 0.21

NTX-Creat (nmol/nmol Creat)

44.6 ± 24.0

43.9 ± 20.6

44.9 ± 25.3

LRP5 genotype frequency n (%)

    V667Ma

        V/V

208 (83.9)

62 (87.3)

140 (82.4)

        V/M

36 (14.5)

7 (9.9)

28 (16.5)

        M/M

4 (1.6)

2 (2.8)

2 (1.2)

    A1330Va

        A/A

163 (65.7)

47 (65.3)

112 (66.3)

        A/V

71 (28.6)

20 (27.8)

48 (28.4)

        V/V

14 (5.6)

5 (6.9)

9 (5.3)

aThe number of individuals with genotype information is different from the total number of subjects due to PCR failure of one sample from the treated group and eight samples from the placebo group

https://static-content.springer.com/image/art%3A10.1007%2Fs00223-008-9207-5/MediaObjects/223_2008_9207_Fig1_HTML.gif
Fig. 1

BMD (expressed as mean percentage change from baseline) and levels of biochemical markers of bone turnover throughout the course of risedronate treatment in the risedronate-treated group compared to the placebo group. Error bars represent SEM. *Significantly different from placebo group (P < 0.001), **P < 0.01

LRP5 Polymorphisms and Bone Phenotypes

The relationship between LRP5 genotypes, BMD, and biochemical markers of bone turnover at baseline is shown in Table 2. The A1330V polymorphism was associated with baseline BMD values at TH, FN, and Troch but not at LS. The mode of association appeared to be recessive in that patients homozygous for the 1330 Val allele (n = 14) had 8.4% higher BMD at TH compared with the other genotype groups (n = 232, P = 0.009). Corresponding values for FN and Troch were 7.5% (P = 0.015) and 11.7% (P = 0.002) (Table 2). Higher BMD values observed at baseline for the 1330 Val/Val genotype at the hip remained significant during the entire treatment period in risedronate-treated subjects. After 24 months of treatment, BMD values for patients with 1330 Val/Val genotype were 11.0%, 11.5%, and 14.7% higher than those with other genotype groups at FN (P = 0.006), TH (P = 0.004), and Troch (P = 0.003), respectively. Analysis of the V667M polymorphism in relation to baseline BMD showed no significant association with BMD at any measured skeletal site. There was no significant association between baseline levels of biochemical markers of bone turnover and the two studied LRP5 polymorphisms (Table 2).
Table 2

LRP5 polymorphisms in relation to BMD and bone turnover markers at baseline

Genotype

LS BMD (g/cm2)

FN BMD (g/cm2)

Troch BMD (g/cm2)

TH BMD (g/cm2)

bALP (μg/l)

CTX-I (ng/ml)

NTX (nmol/nmol Creat)

V667M

    V/V (n = 202)

0.807 ± 0.092

0.673 ± 0.075

0.634 ± 0.088

0.786 ± 0.093

12.5 ± 4.1

0.41 ± 0.20

44.4 ± 24.9

    V/M (n = 35) + M/M (n = 4)

0.788 ± 0.100

0.676 ± 0.093

0.636 ± 0.104

0.789 ± 0.105

12.8 ± 4.5

0.41 ± 0.18

44.8 ± 18.4

    P valuea

0.23

0.82

0.89

0.88

0.65

0.98

0.91

A1330V

    A/A (n = 159)

0.809 ± 0.087

0.671 ± 0.071

0.627 ± 0.083

0.781 ± 0.089

12.5 ± 4.3

0.42 ± 0.21

43.9 ± 25.8

    A/V (n = 68)

0.800 ± 0.099

0.671 ± 0.085

0.637 ± 0.095

0.788 ± 0.098

13.0 ± 4.2

0.41 ± 0.19

47.6 ± 19.9

    V/V (n = 14)

0.791 ± 0.092

0.721 ± 0.100

0.704 ± 0.131

0.849 ± 0.122

10.8 ± 3.1

0.31 ± 0.12

37.6 ± 18.0

    P valueb

0.65

0.05

0.007

0.028

0.25

0.25

0.31

    A/V + V/V (n = 82)

0.799 ± 0.097

0.680 ± 0.089

0.649 ± 0.103

0.798 ± 0.103

12.7 ± 4.1

0.40 ± 0.18

45.9 ± 19.7

    P valuea

0.39

0.42

0.07

0.16

0.74

0.55

0.55

    A/A + A/V (n = 227)

0.806 ± 0.091

0.671 ± 0.076

0.630 ± 0.087

0.783 ± 0.092

12.6 ± 4.2

0.42 ± 0.20

45.0 ± 24.2

    P valuec

0.52

0.015

0.002

0.009

0.15

0.10

0.26

aDominant model

bGenotype test

cRecessive model. Due to the small number of subjects with the M/M genotype (n = 4), only the dominant model was investigated for the V667M polymorphism. All analyses were performed using GML-ANOVA, adjusting for age, body weight, and height

LRP5 Polymorphisms and Response to Treatment

Analysis of treatment response in relation to LRP5 polymorphisms showed no significant differences in percentage changes of BMD from baseline between genotypes defined by the A1330V (Fig. 2) or V667M polymorphism (data not shown) in the placebo- or risedronate-treated subjects. Levels of bone turnover markers in relation to A1330V LRP5 genotypes during the treatment period are shown in Fig. 3. We observed a trend for reduced bALP and CTX-I after 6 months of risedronate treatment in patients with at least one copy of the 1330Ala allele compared to patients with 1330 Val/Val. However, the association was of borderline significance, and a similar trend was observed in the placebo group (Fig. 3). We found no significant interaction between the two studied LRP5 polymorphisms and treatment in relation to biochemical markers of bone turnover when data were analyzed in the whole study population (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00223-008-9207-5/MediaObjects/223_2008_9207_Fig2_HTML.gif
Fig. 2

LRP5 A1330V polymorphism in relation to changes in BMD values from baseline throughout the course of risedronate or placebo treatment. Analysis was performed using GML-ANOVA, adjusting for age, body weight, height, and baseline BMD values. None of the associations was statistically significant. Error bars represent SEM

https://static-content.springer.com/image/art%3A10.1007%2Fs00223-008-9207-5/MediaObjects/223_2008_9207_Fig3_HTML.gif
Fig. 3

LRP5 A1330V polymorphism in relation to biochemical markers of bone turnover throughout the course of risedronate or placebo treatment. Values are presented as mean percentage changes from baseline. Analysis was performed using GLM-ANOVA, adjusting for age and baseline values. Error bars represent SEM

The V667M and A1330V polymorphisms were in strong linkage disequilibrium (D′ = 0.71), and haplotype analysis showed no significant association between haplotypes defined by these two polymorphism and BMD, biochemical markers of bone turnover, or response to risedronate treatment (data not shown).

Discussion

Common polymorphisms in the LRP5 gene have recently emerged as an important determinant of BMD and other osteoporosis-related phenotypes in several populations [1322, 24, 33]. Indeed, in a recent genomewide association study, polymorphisms at the LRP5 locus were significantly associated with BMD [23, 34]. However, our knowledge of the relationship between candidate gene polymorphisms and response to treatment of osteoporosis is very limited. There are no such studies published for the LRP5 gene and very few for other osteoporosis candidate genes [2629, 35]. In this study we therefore investigated the relationship between common coding LRP5 polymorphisms in relation to BMD, biochemical markers of bone turnover, and response to risedronate treatment in osteoporotic or osteopenic men. In agreement with previous studies, we found a significant association between the A1330V polymorphism of LRP5 and BMD in the men who took part in this study. The association was significant at all regions of the hip but not significant at the spine. In contrast, we observed no association between the V667M polymorphisms of LRP5 and BMD at either site, although the study population was relatively small and we could not exclude a modest association. Although the A1330V polymorphism was associated with reduced BMD, we found no evidence of an interaction between genotype and treatment response since homozygotes for the low BMD–associated allele responded to risedronate equally as well as the other genotype groups. This is an important finding clinically since it demonstrates that individuals who are genetically predisposed to osteoporosis by carrying an unfavorable variant of LRP5 still respond normally to risedronate, which is commonly used in the treatment of osteoporosis. Although we observed a weak association between the A1330V polymorphism and markers of bone turnover throughout the course of treatment, this difference was of borderline significance and was observed only at 6 months and a similar trend was found in the placebo group. This could have been a false-positive result since no evidence of genotype–treatment interaction was found when data were analyzed in the whole study group.

It is of interest that in this study, patients with the 1330 Val/Val genotype had higher BMD at all hip sites compared to patients with other genotype groups. This finding is in contrast to many previous studies in which the 1330 Val allele was associated with reduced BMD [1719, 21, 23, 24]. However, in agreement with our findings, Kiel et al. [20] found that the 1330 Val allele was associated with increased BMD values in men with reduced physical activity levels, whereas the same allele was associated with decreased BMD in physically active men. Moreover, Brixen et al. [19] found that the LRP5 A1330V polymorphism is associated with BMD in physically active men only in whom the 1330 Val allele was associated with reduced BMD. These findings combined with those observed from mice expressing a mutant Lrp5 gene [36, 37] highlighted the importance of physical activity and mechanical loading in modulating the effect of LRP5 on BMD determination. Although we did not have data on physical activity levels in our subjects, it is well established that reduced physical activity is an important risk factor for low BMD and osteoporosis [38], which were the criteria for enrolling our subjects. The reasons for these differences between studies might be explained by differences in physical activity between cohorts or the possibility that A1330V is not the functional polymorphism causing the association but, rather, is in linkage disequilibrium, with a causal polymorphism elsewhere in the gene.

The mechanism by which physical activity could modify the effect of the A1330V polymorphism on bone remodeling is unclear. However, since previous studies have shown that this polymorphism influences LRP5-mediated wnt signaling [20], one possible mechanism is that metabolic signals induced by mechanical loading could differentially act on LRP5 variants to influence LRP5-mediated wnt signaling. In support of this notion, previous studies have shown that allelic variations at the interleukin–6 locus modify cortical bone resorption in response to physical exercise [39].

In conclusion, our data showed that the LRP5 A1330V polymorphism is associated with hip BMD in osteoporotic men but found no evidence to suggest that allelic variation in the LRP5 gene regulates the response to risedronate therapy. However, further studies using larger clinical trials are required to confirm these findings.

Acknowledgments

The study was supported by a research grant from Procter & Gamble Pharmaceuticals. O. M. E. A. is funded by the Arthritis Research Campaign UK (15389 and 16303). S. H. R. acted as a consultant for Novartis, Procter & Gamble, and Sanofi Aventis pharmaceuticals.

Copyright information

© Springer Science+Business Media, LLC 2009