Relation of common allelic variation at vitamin D receptor locus to bone mineral density and postmenopausal bone loss: cross sectional and longitudinal population study

LONGITUDINAL STUDY

In 1977, 315 women aged 45 to 54 living in the Copenhagen region were recruited from a questionnaire survey and a medical screening procedure for a two year follow up study on the prevention of postmenopausal bone loss. Women who were immobile, who had diseases known to affect bone metabolism, or who had at any time after the menopause received drugs known to affect bone metabolism (including hormone replacement therapy) were excluded from the study. Women who were aged 45-54 and had had a natural menopause six months to three years previously were included. Details of the selection process are given elsewhere.12

The 260 women who were alive and living in the area in 1995 were asked to participate in this study, and 149 agreed. Thirteen were excluded for the reasons described above, so 136 women were included in the longitudinal study. The mean (SD) number of years since the menopause in these 139 women at baseline was 2 (1). These women did not differ significantly from the 179 women who were not included in terms of weight, height, age, and years since the menopause at baseline. Total bone loss was calculated as a percentage for three periods. Early postmenopausal bone loss at the distal forearm over two years (1977-9) was calculated by simple linear regression of nine measurements of bone mineral content; late postmenopausal bone loss at the distal forearm, proximal femur, and lumbar spine was calculated over six years from one bone mineral density measurement in 1989 and one in 1995; long term postmenopausal bone loss over 18 years was calculated from the first measurement in 1977 and one measurement in 1995. Participants receiving hormone replacement therapy in 1977-9 and 1989-95 were excluded from the calculations of the given time period. Participants who had taken hormone replacement therapy for more than three years were excluded from calculations for 1977-95.

CROSS SECTIONAL STUDY

From 1992 to 1994, 61 116 short questionnaires were posted to randomly selected women living in the Copenhagen region in order to form a database from which to draw participants for future studies. In all, 30 657 questionnaires were returned to the centre. From these completed questionnaires 1008 premenopausal, perimenopausal, and postmenopausal women were selected to take part in clinical and epidemiological studies after excluding those who were immobile, who had diseases affecting bone metabolism, who had received hormone replacement therapy within the previous six months or for more than three years, or who had taken any other drug known to affect bone metabolism. Most of the women included in the postmenopausal group of this study were aged between 49 and 61 or between 65 and 73 years. One baseline measurement of bone mineral density in the distal forearm, proximal femur, and lumbar spine was taken in all 1008 participants. These measurements were taken from 1992 to 1994. In 1995 these women were asked for a blood sample for determination of vitamin D receptor genotype. In all, we obtained 489 samples. To this number was added 110 participants from the longitudinal study who fulfilled the inclusion criteria for the cross sectional study; we used their bone mineral densities from 1995.

These 599 women were divided into three different groups according to the time since the menopause. Fifty were premenopausal, 153 were perimenopausal (time since the menopause six months to three years), and 396 were postmenopausal (time since the menopause being more than three years).

In 1993 we also included a further 180 women from the same database in a clinical trial. The inclusion criteria were (a) a bone mineral density in the forearm of less than -1.5 of the standard deviation of the mean value in premenopausal women (corresponding to a bone mineral density 9% lower than that of an age matched reference population) and (b) that more than 10 years had elapsed since the menopause. The exclusion criteria were the same as described above. In 1995 we asked the 180 women for a blood sample to determine vitamin D receptor genotype; 125 gave their consent and formed the group with low bone mineral density in our study. In 1993 bone mineral density of the lower forearm was measured in all 125 participants. In 1993 we also measured bone mineral density in the lumbar spine and proximal femur in 124 of them (one participant did not turn up for appointment).

MEASUREMENTS

Vitamin D receptor genotype was determined by the polymerase chain reaction and analysis of restriction fragment length polymorphism. In brief, DNA was isolated from leucocytes from whole blood. A region of the receptor gene carrying the three reported polymorphic restriction sites BsmI, ApaI, and TaqI (situated in intron 8 and the 5' end of exon 9; 2100 base pairs)6 was amplified by the polymerase chain reaction. The specific primers for the vitamin D receptor were upstream primer positioned in exon 8 (5'- CAACCAAGACTACAAGTACCGCGTCAGTGA-3’) and downstream primer positioned in exon 9 (5'-CGAGCACAAGGGGCGTTAGC-3’). DNA (50 ng) was amplified on a Perkin Elmer 9600 Thermocycler (Forster City, CA) for 33 cycles at 94°C for 25 s, 63°C for 30 s, and 72°C for 4 minutes, with an additional extension of 5 minutes at 72°C in the final step. The products of the reaction were digested with the restriction endonucleases BsmI, ApaI, and TaqI. Resulting restriction fragments were separated by gel electrophoresis in a 1% agarose gel and the genotype directly determined by looking at the DNA fragment pattern under ultraviolet light. The technique was validated against the standard Southern blot technique before implementation, and samples were tested in random order and blinded to clinical information.

In 1977-89 bone mineral content and bone mineral density of the forearm were measured by single photon absorptiometry using iodine-125 (Bone Mineral Analyzer 1100; M(empty set)lsgaard Medical, H(empty set)rsholm, Denmark)13 and in 1995 by single energy x ray absorptiometry (DTX100; Osteometer, R(empty set)dovre, Denmark).14 The correlation between the two techniques in measuring bone mineral content at the distal forearm is good (R = 0.99), with measurements lying along the line of equality.15 Furthermore, the same external standard and calibration system has been used by this centre since 1977, thus making the results obtained at the different times comparable.

The bone mineral densities of the lumbar spine and of the proximal femur were measured by dual energy x ray absorptiometry. In the longitudinal study the machine used was a QDR 1000 and in the cross sectional study a QDR 2000 (both from Hologic, Waltham, MA).16

STATISTICAL ANALYSIS

Simple and multiple linear regression analyses were carried out to determine the relation between bone mineral density, age, weight, height, years since the menopause, and vitamin D receptor genotype. Analysis of variance using the general linear model of the SAS Institute17 was used for comparison between the genotypes, with a significance level of 5%. Differences in genotype frequencies between the four groups in the cross sectional study were tested by likelihood ratio χ² statistics.

Bone mineral density was expressed as a percentage of expected values; it was calculated for each group as the ratio of the observed bone mineral density to the value expected within the study group as a function (determined by linear regression) of years since the menopause. For the premenopausal women the percentage was calculated as the ratio of the observed bone mineral density to the mean value in the premenopausal group.

VITAMIN D RECEPTOR GENOTYPES AND BONE MINERAL DENSITY

Many studies have shown osteoporosis to be hereditable.4 5 8 Given the many and complex factors controlling bone resorption and bone formation, the hereditability of osteoporosis is likely to be multifactorial. However, Morrison et al found that common allelic variation in a single gene coding for the vitamin D receptor accounted for up to 75% of the total genetic effect.6 Others have since confirmed the relation of the vitamin D receptor gene with bone mineraldensity,7 11 18 although Hustmeyer et al found no relation between them.8 Preliminary data from a large population study in Rotterdam show an association of the baT haplotype with low bone mineral density,19 contrasting with the alleles reported by others.7 11 18

Some groups have shown the relation between vitamin D receptor genotypes and bone mineral density to be more pronounced in premenopausal women18 and to decline with age.20 We could not confirm this. Our premenopausal group was, however, too small to rule out a minor relation between genotype and peak bone mass, but the relation between genotypes and bone mineral density in our larger perimenopausal group was not significant. More importantly, the relation was clearly not significant in our postmenopausal group at the age when fractures begin to appear.

Like Looney et al,21 we did not find a higher prevalence of the BB genotype in women with a low bone mineral density. In the forearm bone mineral density was 9% below that of normal age matched women; these women with a low bone mineral density corresponded to the spinal wedge fracture group (mild osteoporosis) of Overgaard et al.22

VITAMIN D RECEPTOR GENOTYPE AND RATE OF BONE LOSS

The first work on vitamin D receptor gene polymorphism showed a noticeable influence of receptor genotype on bone turnover, measured as circulating osteocalcin concentrations.23 Subjects homozygous for BB had rates twice as high as those with Bb or bb genotypes.23 As biochemical markers of bone turnover may predict the rate of bone loss,24 we investigated whether receptor genotype affects the rate of bone loss. We followed up 136 women for 18 years with repeated bone mineral density measurements. We found no significant association of receptor genotype (BsmI, ApaI, or TaqI) with early postmenopausal bone loss over two years at the lower forearm, which is in contrast to the findings of Krall et al.11 However, in their study of early postmenopausal bone loss Keen et al found no significant relation of vitamin D receptor genotype (TaqI) with rate of bone loss at the lumbar spine or femoral neck,10 thus supporting our findings.

Late postmenopausal bone loss was determined over six years from measurements at the lower forearm, hip and lumbar spine. Again there was no significant relation between receptor genotype and rate of bone loss. Finally, when long term postmenopausal bone loss at the lower forearm was considered the differences between the genotypes were clearly not significant. Our cross sectional study also supports the lack of association between vitamin D receptor genotype and rate of bone loss (fig 1).

Our selection process for the cross sectional study has not led to any bias in bone mineral density or vitamin D receptor genotype. The genotype frequencies in our study are similar to the frequencies found in white women.8 19 23 The cross sectional study does, however, have a slightly unequal age distribution in the postmenopausal group, with a relative lack of women between the ages of 61 and 65, but this is unlikely to have any bearing on the conclusions of this study.

CONCLUSION

If the vitamin D receptor genotype is related to bone mineral density in premenopausal women it is probably not clinically relevant since this relation seems to be lost at the menopause, whereas fractures occur mainly after the menopause. We therefore conclude, that receptor genotype is not related to bone density in postmenopausal women. Vitamin D receptor genotype was also not related to early (age 51-53), late (age 63 to 69), or long term postmenopausal bone loss (age 51 to 69).