WO1996022387A1 - Diagnostic method using estrogen receptor gene polymorphisms - Google Patents

Abstract

The present invention provides a method of assessing an individual's predisposition to low or high bone density, development of low or high bone turnover and/or responsiveness to therapy. The invention also provides a method of assessing an individual's predisposition to development of Type 1 or Type 2 diabetes. The method involves analysis of the estrogen receptor gene of the individual.

Description

DIAGNOSTIC METHOD USING ESTROGEN RECEPTOR GENE POLYMORPHISMS

The present invention relates to a method of assessing an individual's predisposition to low or high bone density, development of low or high bone turnover and/or responsiveness to therapy. The invention also relates to a method of assessing an individual's predisposition to development of Type 1 or Type 2 diabetes.

Osteoporosis is a debilitating bone disease that affects a high proportion of women and a lesser number of men. Due to the considerable public health problem associated with osteoporosis efforts have been focussed at identifying diagnostic and predictive markers associated with the disease. Since osteoporosis has a significant genetic component, the possibility exists for genetic prediction of susceptibility and an understanding of the underlying pathophysiology of disease. Prediction of those at risk of osteoporosis may reduce the incidence of this disease by focussing early attention at those in greatest need. Our aim was to define a genetic test that is capable of identifying those individuals at risk of osteoporosis and those protected from disease.

Osteoporosis is defined generally as low bone density associated with fracture however for genetic analysis, a bone density two standard deviations below young normal is a more useful criterion of osteoporosis since subsequent fracture is dependent on a traumatic event. Several humoral markers, notably osteocalcin, have been widely used as indicators of bone turnover. Genetic effects on osteocalcin levels have been reported in twin studies with a correlation between high osteocalcin levels and lower bone mineral density (Kelly et al., 1989). Furthermore, this genetic effect is explained in part by allelic variation in the vitamin D receptor gene which influences osteocalcin serum levels (Morrison et al. 1992) and bone mineral density (Morrison et al. 1994). In this study, genotypes of the estrogen receptor gene (ESR) are shown to be related to differences in bone density and inferred rates of loss of bone density. In addition, genotypes of the ESR are related to mean serum insulin levels, indicating that ESR genotypes could be useful in predicting risk of developing diabetes, both type 1 and type 2. A recent paper (2) described a male individual with a debilitating mutation of the estrogen receptor gene. This subject had low bone density and had inexplicable hyperinsulinemia. In retrospect, this case shows that complete abolition of the function of the ESR can result in both hyperinsulinemia and low bone density. The present invention demonstrates that common alleles of the ESR gene can be used to describe a frequent and milder form of disorder with lower mean bone density and/or greater rate of loss of bone density with age and/or higher serum insulin levels. Subjects homozygous for the presence of the Pvu II RFLP in the ESR gene exhibit mean traits compatible with a reduced functionality of the estrogen receptor gene and this results in the prediction that these subjects will respond differently to a wide range of process dependent on the estrogen receptor, whether ligand dependent or not. The discrimination of these genetic subtypes will be of utility in estrogen based therapies, both for bone disorders, diabetes and for reproductive problems and estrogen dependent breast cancer. These data demonstrate that the common alleles of the ESR are functionally distinct. In terms of diabetes therapy, this test indicates that a therapy based on manipulating estrogen receptor should raise insulin levels, providing an adjunct to injected insulin. During the time of developing type 1 diabetes when pancreatic islet cells are destroyed or become deficient in insulin secretion, treatment with an estrogen analogue could be expected to boost insulin secretion and have a sparing effect on islet cells and a limitation of the degree of patient crisis. In those patients with established type 1 diabetes, residual islet cells may be influenced to produce increased amounts of insulin under the influence of an estrogen receptor analogue, ameliorating the severity of disease and providing an oral therapy as an adjunct to injectable insulin.

Accordingly in a first aspect the present invention consists in a method of assessing an individual's predisposition to low or high bone density, development of high or low bone density and/or responsiveness to therapy comprising analysing allelic variation in relation to the estrogen receptor gene of the individual.

In a second aspect the present invention consists in a method of assessing an individual's predisposition to development of Type 1 or Type 2 diabetes and/or central abdominal fat and/or responsiveness to therapy comprising analysing allelic variation in relation to the estrogen receptor gene of the individual.

In a third aspect the present invention consists in a method of assessing an individual's predisposition to development of central adiposity and/or insulin resistance comprising analysing allelic variation in relation to the estrogen receptor gene of the individual.

In a fourth aspect the present invention consists in a method of treating Type 1 or Type 2 diabetes in an individual comprising administering to the individual an effective amount of an estrogen analogue.

In a preferred embodiment of the first aspect of the present the method further involves analysing allelic variation in relation to the vitamin D receptor gene and/or the retinoic acid receptor alpha gene of the individual. Further information regarding the analysis of these genes may be found in PCT/AU93/00394, and Australian Provisional patent application

Nos. PM 7015 and PM 7063. The disclosure of these applications is included herein by cross reference.

It is presently preferred that ESR genotypes are detected with the restriction endonuclease PvuII. However, as it will be appreciated by those skilled in the art, other restriction enzymes and other detection systems can be used to detect the same genetic variants, and that other genetic variants can have similar information content and be detected by other means. The Pvu2 RFLP may be in linkage disequilibrium with other sequence alteration, both known and unknown that mediate this effect. The invention can be practised in a number of ways in particular:

1. The use of Southern hybridisation blots to detect the estrogen receptor restriction fragment length polymorphism using a cDNA fragment as a probe. 2. The use of ESR genotypes detected by any means as a method of discriminating differences in risk of lower mean bone density, different rates of change of bone density, subsequent fracture risk and as a means of discriminating intrinsic differences in responsiveness to estrogen based therapies in health and disease.

3. The use of ESR genotypes, detected by any means, in combination with other genes, particularly the VDR gene and the RAR-alpha gene, in determining risk of low bone density, differing rates of loss of bone density and subsequent osteoporosis risk.

4. The use of ESR genotypes, detected by any means, in combination with other genes, particularly the VDR gene and the RAR-alpha gene, in determining response to therapy directed at reducing osteoporosis risk.

5. The use of ESR genotypes, detected by any means, in determining response to therapy directed at limiting type 1 and type 2 diabetes.

6. The use of ESR genotypes, detected by any means, which define differing levels of partial resistance to estrogens as a diagnostic tool in discriminating risk to a range of estrogen receptor dependent processes, such as infertility, breast cancer and reproductive cancer risk.

In order that the nature of the present invention may be more clearly understood preferred forms thereof will now be described with reference to the following examples.

EXAMPLES

Example 1

Interaction Between Estrogen Receptor and Vitamin D Receptor Genotypes

Contributes to Normal Variation in Bone Mineral Density.

Abstract

Bone mineral density (BMD), a principal risk factor for osteoporotic fracture, is determined by both genetic and environmental factors. The present investigation examines the effect of estrogen receptor (ESR) gene polymorphisms and their interaction with the vitamin D receptor (VDR) gene alleles in the determination of BMD. One hundred and nineteen pairs of normal, healthy female twins, aged 41.8 + 14.1 years (mean + SD) were studied. BMD at the lumbar spine and femoral neck was measured by dual- energy x-ray absorptiometry. Polymorphisms of the ESR and VDR genes were determined by PCR on isolated blood leucocytes to detect the region of DNA susceptible to the Pvu-II and Bsm-I endonuclease sites, respectively. For ESR, the presence was coded as p and absence, P of the Pvu-π site; for VDR, the presence was coded as B and absence, b of the Bsm-I site. The distribution of ESR genotypes were as follows: PP 23.4%, Pp 42.6% and pp 48 34.0%. The distribution of VDR genotypes were: BB 22.7%, Bb 49.6% and bb 27.7%. There was no statistical association between ESR and VDR genotypic distribution (p = 0.52). Unadjusted lumbar spine BMD was highest in pp genotype (1.21 + 0.12 g/cm2; mean + SD), followed by Pp (1.16 + 0.14) and PP (1.14 + 0.19). After adjustment for age (or years since menopause) and weight, the effects of ESR genotypes was statistically significant (p = 0.034) and remain significant (p = 0.029) in the presence of effect of VDR genotypes. Further analyses indicated that effect of ESR genotypes was most pronounced in VDR BB and Bb groups, but not in the bb group. On the other hand, effect of VDR genotypes was greatest in ESR PP group compared to that of in Pp or pp group. No significant effect of ESR genotypes was observed in femoral neck BMD. These data indicate that normal functional variants of the ESR gene are associated with lumbar spine BMD and that the association between VDR genotypes and BMD may be mediated to a greater extent through the ESR genotypes. The complexity of genetic interaction requires further investigation at the molecular level. Low bone mineral density (BMD), particularly in late decades of life, is an important risk factor for fracture (Hui et al 1987, Nguyen et al 1993, Cummings et al 1993). Bone density in later life is the sum of peak bone mass achieved during premenopausal years and subsequent postmenopausal age-related bone loss. Although environmental and life style factors play a role in the determination of peak bone density, genetic factors appears to have a greater influence, accounting for between 75% to 80% of variance of bone mineral density (Smith et al 1973, Pocock et al 1987, Dequeker et al 1987). We have previously reported that polymorphisms of the vitamin D receptor (VDR) gene accounted for a significant proportion of the genetic variance (Morrison & Qi et al 1994). However, even after accounted for by specific environmental factors and VDR gene alleles, a sizeable proportion of variance of BMD remained unexplained. This indicates that there are other genes or environmental factors are involved in the regulation of BMD. The central role of estrogen deficiency in the pathogenesis of postmenopausal bone loss and osteoporosis is well established. Estrogen status and genetic factors have been suggested to play a functional role in achievement and maintenance of bone mass in premenopausal women (Armmento-Villareal 1992). Estrogen receptor is a major regulator of growth. Estrogen receptor status is a predictor of response to endocrine treatment of breast cancer (Vollenweider- Zerargui 1986). Recently, Smith et al (1994) reported an interesting case of a young adult man, who had a disruptive mutation in the estrogen receptor gene, such that resulting in estrogen resistance; he was tall stature with unfused epiphyses, and markedly low bone density. It is, thus, likely that common variation, rather than severe mutation, might exist in the ESR gene which contributes to low bone mineral density. More importantly, like VDR, ESR is a member of superfamily of nuclear receptors. Therefore, functional different alleles of ESR might be also expected to have impact on numerous estrogen receptor dependent processes. In this example, we examine the effect of ESR gene polymorphisms and their interaction with the VDR gene alleles in the determination of BMD in a sample of predominantly premenopausal twins.

Materials and Methods

Subjects

The majority of twins in this study were originally recruited for an investigation into genetic effects of bone density and bone turnover markers, in which details of study design and method of recruitment have been described elsewhere. Briefly, through the Australian Twin Registry, twins were identified and invited to participate in the study. After obtaining informed consent, each twin was interviewed separately by using a structured questionnaire to solicit demographic data, basic clinical and life¬ style history. All twins pairs had normal renal function as assessed by serum and/or creatinine clearance. Twins with a history of diseases or medication use which could affect bone mass and bone turnover were excluded from the study. Bone mineral density (BMD) at the lumbar spine and proximal femur were measured by a dual photon absorptiometry or dual photon energy X-ray absorptiometry (LUNAR Corp, Madison, WI). The intrasubject coefficient of variation for normal subjects at our institution was 1.2% for the lumbar spine and 1.9% for the femoral neck.

DNA analysis

Blood was collected into heparin treated tubes and leukocytes separated by sedimentation through physiological saline solution in a clinical centrifuge. Purified leukocytes were lysed in leukocyte lysis buffer (10 mM Tris-HCl, pH7.4, physiological saline and 0.5% w/v sodium dodecyl sulphate). Lysate was treated with proteinase K (Applied Biosciences, Palo Alto USA) at 50 μg/ml for 2 hour at 65 Celcius. DNA was extracted by repetitive phenol chloroform solvent extraction as described in Maniatis et al. (1982) and ethanol precipitated. DNA was redissolved in TE buffer

(lOmM Tris-HCl, ImM EDTA, pH 8.0) and quantitated by absorbance at 260 nM.

The ESR gene probe was a 2.1 kilobasepair cDNA (pOR-8) cloned in pGEM (Stratagene Inc.) described by Green et al (1986). Radioactive probes and Southern blotting were as previously described (Hill et al 1989). The

RFLP was a two allele Pvu-II variant bands at 1.6 and /or 0.6 kilobase and six invariant bands at 13, 7.5, 3.9, 3.5 and 1 kilobase were also identified by Southern blot (Hill et al 1989). Flanking primers (5'- CTGCCACCCTATCTGTATCTTTTCCTATTCTCC- 3' and 5'- TCTTTCTCTGCCACCCTGGCGTCGATTATCTGA-3') were then used in a

PCR amplification reaction consisting of the following steps: denaturation for 5 sec at 94oC; annealing at 64oC for 5 sec; and polymerase extension at 72oC for 90 sec. After amplification reaction, the PCR products were digested with Pvu-II and electrophoresed in a 1.5% agarose gel to type the sample according to Yaich et al 1992.

Statistical analysis

Statistical analysis of data was aimed at examining the association between ESR genotypes and BMD, controlling for specific environmental factors and the effects of VDR genotypes. Therefore, prior to the analysis of effects of genetic variability at the ESR gene locus, BMD was adjusted by linear regression for variation in age, sex, height, weight, menopausal status and VDR genotypes. Model-fitting analyses indicated that a model with age (or postmenopausal years) and weight was the most empirical. Studentised residuals [Myers, 1986] obtained from this model were then used as a adjusted measure of BMD. Analysis of variance was performed on the adjusted BMD to test the null hypothesis that phenotypic variation is unaffected by the ESR gene locus. Interaction between the ESR and VDR genotypes was also explored in this analysis of variance model. Scheffe's multiple comparison was used to make specific contrasts among genotypic means.

Results

Sixty-three MZ pairs (aged 41.8 + 14.1 years; mean + SD) and 56 DZ pairs (aged 40.7 + 13.2) were used in the analysis. There was no significant difference between zygosities in terms of mean and variance of age, weight, height, body mass index and years since menopause.

The distribution of ESR genotypes were as follows: PP 33 (23.4%), Pp 60 (42.6%) and pp 48 (34.0%), giving the allelic frequency of P being 0.447 + 0.029 (mean + SE). The distribution follows closely the Hardy- Weinberg equilibrium law (p = 0.61). The distribution of ESR and VDR genotypes were independent (p = 0.516): the allelic frequency of "P" among VDR Bsm-1 genotypes BB. Bb and bb was 0.48 + 0.062, 0.42 + 0.042 and 0.46 + 0.056, respectively (Table 1). There was significant difference between ESR genotypes with respect to age (p = 0.04), in which women with PP were about 7 years older than their counterparts with pp genotype, who in turn, were about 4 years younger than the heterozygous group (Pp). However, the three genotypic groups were comparable with respect to weight, height and menopausal status (Table 2). Unadjusted lumbar spine BMD was higher among subject with p allele: mean + SD for PP, Pp and pp genotypes were: 1.14 + 0.19 g cm2, 1.16 + 0.14 g cm2 and 1.21 + 0.12 g/cm2, respectively (p = 0.14; simple one- factor analysis of variance). The corresponding values for femoral neck BMD were 0.88 + 0.15 g/cm2, 0.90 + 0.11 g/cm2 and 0.91 + 0.12 g/cm2 respectively (p = 0.63).

On the other hand, BMD was found to be related to demographic characteristics such as age, weight, height and menopausal status. Model- fitting analyses (Table 3) and stepwise regression analysis suggested that the model incorporating effects of post-menopausal years and weight is the most parsimonious one: the two variables collectively accounted for 26% and 31% of variation of lumbar spine and femoral neck BMD, respectively. At the lumbar spine, the effect of ESR genotypes was statistically significant, either in the presence or absence of VDR genotypes; however, a model with interaction between ESR and VDR genotypes was more empirical, accounting for 46% of variation of lumbar spine BMD. At the femoral neck, there was no statistically significant effect ESR genotypes, either in main-effect or interaction effect model, however, there was significant effect of VDR genotypes (Table 3). Analysis of the interaction model indicated that the effect of ESR genotypes was not constant across VDR genotypes. For example, among subjects with BB and Bb genotypes, the ESR P allele was associated with a decrease in BMD by 0.026 g/cm2 and 0.024 g/cm2, respectively, such that spinal BMD in women with BBPP genotype was 13% lower than those of BBpp genotype (p = 0.03). On the other hand, among subjects with bb genotype, higher BMD ( + 0.006 g/cm2) was associated with the ESR allele P, such that subjects with PP has 5.7% higher BMD than those with pp (p = 0.20). Similar trends were observed at the femoral neck, however, the extent of effect was small and statistically insignificant (Table 4).

Discussion

The present analysis shows that ESR genotypes, as detected by frequent RFLPs, are associated with different mean BMD traits, either in the presence or absence of the effect of VDR genotypes. The magnitude of the ESR genotypic effect in this cohort was, however, less than that of the VDR gene locus. This finding is, in fact, agreeable to a recent study in a Japanese population, in which higher BMD was associated with the p allele (Hosoi et al 1995).

It was widely accepted that the central role of estrogen deficiency in the pathogenesis of postmenopausal bone loss and osteoporosis by stimulating osteoclastic bone resorption. Estradiol can act directly on osteoblasts by receptor mediated mechanism and thereby modulate the extracellular matrix and other proteins involved in the maintenance of skeletal mineralization and remodeling (Erisksen et al 1988). ESR is a member of superfamily of nuclear receptors for hydrophobic ligands including the steroid hormones, thyroid hormone, vitamin D3 and retinoids. As a class, these receptors are transcription factors that are regulated allosterically by ligand binding. In this manner, it is expected that natural variants of a receptor gene, especially the estrogen receptor, could be responsible for differences in physiological parameters. ESR gene, like VDR gene, is therefore a another candidate for a prime regulator of peak bone mass in adulthood.

The present analysis further indicates an interaction effects between ESR and VDR genotypes in the determination of BMD at the lumbar spine. The interaction suggests that polymorphisms of the ESR gene may modulate the expression of VDR gene effect in different populations. Indeed, the difference between VDR BB and bb genotypes was greatest (28%) among subjects with ESR PP genotype, but was somewhat lower among subjects with Pp(12%) or pp (7%). On the other hand, the most pronounced effect of ESR genotypes was observed in subjects with VDR BB genotype, but not significant in those with Bb or bb genotypes.

Since the distribution of ESR genotypes was independent to that of VDR genotypes, the observed interaction between the two gene loci can not be attributed to random association. A simple model of functionally different gene products resulting in altered binding capacity is a plausible explanation for the observed phenomenon, but is also possible that they are caused by more complex mechanisms. Indeed, Bellido et al (1993) has obtained evidence showing that ESR is up-regulated by bone-active steroid systemic hormone calcitriol. Keeting et al (1992) found that the presence of calcitriol is required for the full expression ESR in human osteoblast-like cell line.

Nevertheless, the molecular basis of this functional allelic variation is unknown. Extrapolating from other steroid receptors, such as the glucocorticoid, androgen, and vitamin D, hormone resistance has clearly been linked to deletions and point mutations of respective genes (Harmon et al 1989, Dieken et al 1990, Brown et al 1988, Ritchie et al 1989). Since the steroid hormone receptors are closely related in their domain structure and function as ligand-inducible transcriptional regulators, one would expect the type of ESR defect associated for estrogen resistance to be analogous to those described for the glucocorticoid receptor, androgen receptor and vitamin D receptor. Pvu-II polymorphic site locate in the first intron (Yaich et al 1992), 400bp upstream from exon 2, with a point mutation (T to C) in the recognition sequence CAGCTG responsible for the PP allele. The location of the RFLP in the intron makes it unlikely that the polymorphism is correlated with ESR expression. However, it could not be ruled out the possibility that the polymorphism is in linkage disequilibrium with another ESR polymorphic site which does affect ESR function and another unidentified gene is responsible for this effect.

The non-significant effect of ESR genotypes on femoral neck BMD is of considerable interest. In the present twin data, heritability (proportion of variance of a phenotype "explained" by genetic factors, denoted by H2) for lumbar spine was estimated to be 0.92 which is considerably higher than that for femoral neck BMD (H2 = 0.74). This is consistent with a number of other twin studies (Smith et al 1973, Christian et al 1989) which showed that genetic effects in the lumbar spine is more pronounced than in the femoral neck since lumbar spine rich in trabecular bone which appear more sensitive than cortical bone to the associated sex hormonal changes.

In summary, the present study identifies that normal functional variants of the estrogen receptor gene are associated with lumbar spine BMD.

Table 1: Frequency distribution of ESR and VDR genotypes

VDR genotype

ESR genotype Ml BB Bb bb

PP 8 17 8 33 (25.0) (24.3) (20.1) (23.4)

Pp 15 25 20 60 (46.9) (35.7) (51.2) (42.6)

PP 9 28 11 48 (28.1) (40.0) (28.7) (34.0)

All 32 70 39 141 (22.7) (49.6) (27.7) (100)

Chi-square test of independence p = 0.516.

Numbers in brackets are percent of total for each ESR genotype (across column). Table 2. Demographic characteristics by genotype.

Characteristics P P PP P value ¹

N 33 60 39

Age (yrs) 46.9 ±11.8 43.8 ±11.1 40.2 ± 12.5 0.04

Weight (kg) 63.1 ±8.6 65.1 ±12.8 62.2 ± 10.6 0.39

Height (cm) 162.0 ±6.3 162.7 ±6.3 162.1 ±8.0 0.82

BMI kg/m2) 24.0 ±2.9 24.6 ±4.6 23.7 ±4.0 0.56

Number of postmenopause 13 12 10

Years 10.6 ±5.5 7.3 ±6.1 10.1 ±9.9 0.48 postmenopause.

(vrs)

Numerical values are mean ± SD

1 P values were calculated based on one-way analysis of variance, in which

ESR genotypes were treated as independent factor.

Table 3. Summary of model-fitting analyses.

Lumbar spine Femoral neck

Model

R2 MSE1 R2 MSE1

1. Age + weight 0.21 0.0169 0.31 0.0109

2. Age + weight + height 0.21 0.0171 0.31 0.0109

3. Age + BMI 0.17 0.0179 0.26 0.0117

4. YMP + weight 0.26 0.0159 0.31 0.0109

5. Age + YMP + weight 0.26 0.0160 0.33 0.0105 Model 4 + ESR 0.29 0.0158* 0.33 0.0106 Model 4 + ESR + VDR 0.42 0.0131** 0.37 0.0104

Model 4 + ESR + VDR + ESR x VDR 0.46 0.0121' 0.36 0.0103

1 MSE: Mean square error

* Effects of ESR is statistically significant at p = 0.048;

** Effects of ESR is statistically significant at p = 0.034, of VDR: p<

0.0001;

*** Effects of interaction between ESR and VDR is significant at p = 0.029 level.

Table 4. Interaction between VDR and ESR genotypes: adjusted lumbar spine and femoral neck BMD.

VDR genotypes

ESR genotypes

BB Bb bb

Lumbar spine BMD

Mean ±SD

PP 1.01 ±0.04 1.18 ±0.03 1.29 ±0.04

Pp 1.08 + 0.03 1.14 + 0.02 1.21 ±0.03

PP 1.14 ±0.04 1.22 ±0.02 1.22 ±0.03

P value

PP vs Pp 0.16 0.18 0.07

PP vs pp 0.03 0.36 0.20

Pp vs pp 0.29 0.01 0.64

Femoral neck BMD

Mean ± SD

PP 0.83 ±0.04 0.90 ±0.03 0.98 ±0.04

Pp 0.87 ±0.03 0.90 ± 0.02 0.92 ± 0.02

PP 0.89 ±0.03 0.91 ±0.02 0.92 ±0.03

P values

PP vs Pp 0.38 0.79 0.09

PP vs pp 0.24 0.82 0.24

Pp vs pp 0.65 0.57 0.66 DNA analysis, Southern blot, PCR (polymerase chain reaction) and RFLP analysis using endonuclease digestion. Blood was collected into heparin treated tubes and leukocytes separated by sedimentation through physiological saline solution in a clinical centrifuge. Purified leukocytes were lysed in leukocyte lysis buffer (10 mM Tris-HCl, pH7.4, physiological saline and 0.5% w/v sodium dodecyl sulphate). Lysate was treated with proteinase K (Applied Biosciences, Palo Alto USA) at 50 μg/ml for 2 hour at 65 Celcius. DNA was extracted by repetitive phenol chloroform solvent extraction as described in Maniatis et al. (1982) and ethanol precipitated.

DNA was redissolved in TE buffer (lOmM Tris-HCl, lmM EDTA, pH 8.0) and quantitated by absorbance at 260 nM. Other methods of DNA preparation (Kawasaki, 1990) are compatible with the PCR procedure.

ESR genotypes were detected with the restriction endonuclease PvuII, using previously reported RFLPs in the ESR gene. As will be appreciated by those skilled in the art other restriction enzymes and other detection systems may detect the same genetic variants, and other genetic variants may have similar information content and be detected by other means. The Pvu2 RFLP may be in linkage disequilibrium with other sequence alteration which may mediate this effect.

ESR genotypes were analysed with respect to bone density measured by dual X-ray absorptiometry at both the lumbar region of the spine (LS BMD) and the femoral neck region (FN BMD) in 235 female subjects. Fasting insulin measured in serum was analysed in a subset of these subjects. These data showed that subjects homozygous for the presence of a Pvu II restriction endonuclease cleavage site detected by the ESR cDNA probe, have higher mean serum insulin levels, and a faster rate of decline in bone density after the menopause.

Statistical analysis. The multiple regression technique was used testing various models of fit of genetic data to clinical data. A biologically plausible model was used in which the effect of the ESR genotype varied with the age of the subject, a so called interaction term was devised where ESR genotype (arbitrarily coded as 1, 2 and 3) was multiplied by a subjects age. If ESR genotypes have different slopes of the age- bone density relationship, then both ESR and ESRxAGE variables will be able to co-exist with Age in an equation.

Bone density declines with age, and any effect of a gene on bone density may vary with age. An interaction term between ESR genotype and age is included in the analysis.

The result: that ESR and ESR-AGE terms co-exist in the same equation indicates that ESR has a strong affect on rate of loss of BMD with time as well as the mean BMD.

Multiple Regression Yl.FN BMD 4X Variables

Count: R: R-squared: Adj. R- RMS squared: Residual:

235 .66 .436 .426 .122

Analysis of Variance

Source: DF: Sum Mean F-test: Squares: Square:

REGRESSI 4 2.644 .661 44.372 ON

RESIDUAL 230 3.427 .015 p = .0001

TOTAL 234 6.071

No Residual Statistics Computed

Note: 170 cases deleted with missing values Multiple Regression Y1:FN BMD 4X Variables

Beta Coefficient Table

Variable: Coeffici Std. Std. t-Value: Probabil ent: Err.: Coeff. : ity:

INTERCE 1.058 PT

ESR -.088 .032 -.417 2.758 .0063 genotype

AGE -.011 .002 -.9 6.502 .0001

KG .005 .001 .353 7.107 .0001

Age-ESR .002 1.971 1.652 2.892 .0042

Multiple Regression Yl.FN R 4X Variables

Confidence Intervals and Partial F Table

Variable: 95% 95% 90% 90% Partial Lower: Upper: Lower: Upper: F:

INTERCE PT

ESR -.151 -.025 -.141 -.035 7.606

AGE -.014 -.007 -.013 -.008 42.272

KG .004 .006 .004 .006 50.51

Age-ESR -3.881 3.885 -3.253 3.257 8.362 The results confirm the contribution of ESR genotypes and ESR effects on the relationship between bone density and age. The interaction term Age time ESR genotype (Age-ESR) is significant in the presence of both age and the ESR term, meaning that the combined term adds more to the equation. This is very strong evidence that the rate of change of bone density after the menopause is related to the ESR genotype. More than 40% of the variance in Femoral neck bone density is explained with this simple equation. These data demonstrate conclusively that ESR genotypes are related to FN BMD and that a strong age-genotype interaction effect exists. The direction of the effect is such that the presence of the site genotype ESR (pp) is correlated with a faster rate of decline. A similar result held for LS BMD.

Effect of ESR on bone density. The effect of another major gene, the vitamin D receptor gene, with impact on bone density was reduced by considering only VDR gene heterozygotes (Bb). The effect of age was limited by only considering premenopausal females. ESR alleles had a significant effect on bone density when analysed in this manner.

ESR combined with other genes.

We have previously described that VDR gene alleles and RAR-alpha gene alleles are related to bone density. The following analysis shows that ESR genotypes can be used in conjunction with VDR and RAR-alpha genotypes to form a three gene genetic test for genetic susceptibility to different bone densities, different rates of loss of bone with time and therefore different risk status for osteoporotic fracture.

Analysis. 166 subjects were genotyped for the three markers (VDR using Bsm-1), RAR-alpha using Pst-1 and ESR using the Pvu II. Multiple regression analysis with age and weight and ESR-Age interaction terms showed that all genotypes can co-exist in the same equation and by extrapolation that all genotypes contribute to the prediction of risk. The subsequent three gene combined test is therefore an improvement over the state of the current art Multiple Regression Yl:LS BMD 5X Variables

Count: R: R-squared: Adj. R- RMS squared: Residual:

166 .575 .33 .309 .127

Analysis of Variance Table

Source: DF: Sum Mean F-test: Squares: Square:

REGRESSI 5 1.265 .253 15.768 ON

RESIDUAL 160 2.568 .016 p = .0001

TOTAL 165 3.833

Residual Information Table

SS[e(i)-e(l-l)]: e > 0: e < 0: DW test:

5.522 88 78 2.151

Multiple Regression Yl.LS BMD 5X Variables Beta Coefficient Table

Variable: Coeffici Std. Std. t-Value: Probabil ent: Err.: Coeff.: ity:

INTERCEPT 1.543

ESR -.114 .051 -.583 2.223 .0276

AGE -.011 .003 -.812 4.13 .0001

VDR .088 .014 .413 6.273 .0001 Bsm-1

RAR- -.041 1.975 1.655 2.865 .0047 alpha

ESR-AGE .003 .001 .695 2.389 .018

Multiple Regression Y1:LS BMD 5X Variables

Confidence Intervals and Partial F Table

Variable: 95% 95% 90% 90% Partial F: Lower: Upper: Lower: Upper:

INTERCEPT

ESR -.215 -.013 -.199 -.029 4.942

AGE -.016 -.005 -.015 -.006 17.053

VDR Bsm-1 .061 .116 .065 .112 39.351

RAR-alpha -3.943 3.86 -3.309 3.226 8.209

ESR-AGE 4.582E-4 .005 .001 .004 5.709

This analysis demonstrates that the ESR genotype contributes to the utility of the previously configured genetic tests, in that each marker contributes to the analysis and therefore to the prediction of differing bone density states. SERUM INSULIN.

Fasting serum insulin is not normally distributed, so analysis was based on both log normal transformed data and untransformed raw data (Figure 1). ANOVA analysis was used to detect differences in serum insulin levels between different ESR genotypes. Students t-test was used for pairwise comparisons of extremes of genotype. Multiple regression was used for multiple variables.

Statistical analysis.

Multiple regression: Fasting serum insulin (logarithmically transformed) was related to weight (kilograms) and to ESR genotype in multiple regression.

Multiple Regression Yl:ln (1+x) of Insulin 2X Variables

Count: R: R-squared: Adj. R- RMS squared: Residual:

68 .36 .13 .103 .427

Analysis of Variance

Source: DF: Sum Mean F-test: Squares: Square:

REGRESSION 2 1.769 .884 4.847

RESIDUAL 65 11.86 .182 p = .0109

TOTAL 67 13.629

No Residual Statistics Computed

Note: 337 cases deleted with missing values Multiple Regression Yl:IN(l+x) of Insulin 2X Variables

Beta Coefficient Table

Variable: Coeffici Std. Std. t-Value: Probabil ent: Err.: Coeff.: ity:

INTERCEPT 1.006

ESR .144 .071 .237 2.028 .0487

KG .011 .005 .243 2.08 .0415

Multiple Regression Yl:IN(l+x) of av ins 2X Variables

Confidence Intervals and Partial F Table

Variable: 95% 95% 90% 90% Partial Lower: Upper: Lower: Upper: F:

INTERCEPT

ESR .002 .286 .026 .263 4.113

KG 4.297E-4 .021 .002 .019 4.327

Excess fat is known to be a strong correlate of serum insulin levels. Fat was measured in the subjects as spinal fat and as central abdominal fat using dual energy X-ray absorptiometry. Spinal fat was taken as that visible in the lumbar spine bone density window of the Dexa output. The ESR genotype remained significant in the analysis in the presence of the spinal fat variable, indicating that the ESR genotype provides additional information above that derived from the central obesity parameter. An increased proportion of the variance of the serum insulin was explained by this analysis.

Multiple Regression Yl .Insulin 2X Variables

Count: R: R-squared: Adj. R- RMS squared: Residual:

68 .608 .367 .347 3.373

Analysis of Variance

Source: DF: Sum Mean F-test: Squares: Square:

REGRESSION 2 428.587 214.294 18.635

RESIDUAL 65 739.551 11.378 p = .0001

TOTAL 67 1168.138

No Residual Statistics Computed

Note: 337 cases deleted with missing values

Multiple Regression Yl:Insulin 2X Variables

Beta Coefficient Table

Variable: Coeffici Std. Std. t-Value: Probability: ent: Err.: Coeff.:

INTERCEPT -.954 spine fat .284 .051 .548 5.534 .0001

ESR 1.517 .556 .269 2.725 .0082

Multiple Regression Yl.Insulin 2X Variables

Confidence Intervals and Partial F Table

Variable: 95% 95% 90% 90% Partial Lower: Upper: Lower: Upper: F:

INTERCEPT spine fat .182 .387 .199 .37 30.624

ESR .405 2.628 .588 2.445 7.427

A similar result was found using central abdominal fat with a weaker p value for ESR of 0.06. This is explained by the second example which shoes that ESR is a determinator of central abdominal fat itself. This analysis shows that ESR provides additional information above that derived from the relationship of fat to serum insulin.

Example: The effect of estrogen receptor genotypes on central abdominal fat:

1. INTRODUCTION TO THE ANALYSIS . A step wise regression approach was undertaken using a range of variables derived from within-pair differences in traits associated with fat metabolism and diabetes susceptibility.

The stepwise regression model ranks the variables in terms of within- pair differences in explaining the within pair difference in the target variable, in this case the within pair difference in central abdominal fat.

2. THE VARIABLES AND HYPOTHESIS .

The subjects of this study were a different twin cohort to the previous study, derived from middle aged female population. Serum parameters were measured and whole body dexa analysis using a Hologic QDR machine was used to determine central abdominal fat content. Central abdominal fat (CAF) is highly related to fasting insulin levels and the physiological phenomenon of insulin resistance.

The hypothesis is that the particular ESR genotypes have a difference by state, which means that a particular genotype will predispose to a direction of effect on the population mean of a trait. The genotypes are hypothesized to have codominant effects, so that heterozygotes will be intermediate in trait mean in comparison to homozygotes. Within a twin pair, the within-pair difference in genotype should relate to the with-pair difference in any particular trait that is influenced by the genotype in question. Since our hypothesis is that a direction of genotype effect exists we use the within pair differences which have sign, rather than absolute values or squared values that become positive and lose directional information. Step-wise regression enters variables in order of their ability to explain the variance in the variable in question (in this case the within-pair difference in central fat). Variables not in the equation. Variable Within-pair difference in:

ΔTOTAL FAT total body fat

ΔESR estrogen receptor genotype detected by

Pvu-π

ΔMENO STATUS menopausal status.

ΔWT weight (kilograms)

ΔHT Height (cm)

ΔBMI Body mass index (weight /square of height in meters)

ΔLBM Lean body mass derived form Dexa scan.

ΔSHBG Sex hormone binding globulin.

ΔEST Circulating estradiol concentrations.

ΔDHEAS Circulating Dyhydroxyepiandrostendione concentrations.

ΔCURRENT ERT Current estrogen replacement therapy.

ΔSMOKING Current smoking.

3. RESULTS.

Summary Information for Statistical Analysis: Step wise regression.

F to Enter 4

F to Remove 3.996

Number of Steps 2

Variables Entered 2

Variables Forced none

Step l a. ANALYSIS OF VARIANCE TABLE

R: R-squared: Adj. R-squared: RMS Residual:

0.851 0.724 0.713 356.442

Source DF: Sum Squares: Mean Square: F-test:

REGRESSION 1 8349299.438 8349299.438 65.716

RESIDUAL 25 3176281.08 127051.243

TOTAL 26 11525580.519

VARIABLES IN EQUATION

INTERCEPT -111.906

Variable: Coefficient: Std. Err.: Std. Coeff.: F to Remove

ΔTOTAL FAT 0.07 0.009 0.851 65.716

c. VARIABLES NOT IN EQUATION

Variable: Par. Corr: F to Enter:

ΔMENO STATUS -0.083 0.167

ΔWT -0.06 0.086

ΔHT -0.25 1.596

ΔBMI 0.179 0.791

ΔLBM -0.086 0.177

ΔSHBG -0.148 0.534

ΔEST -0.153 0.572

ΔDHEAS 0.105 0.266

ΔCURRENT ERT 0.246 1.549

ΔSMOKING 0.056 0.075

ΔESR¹ -0.408 4.786 STEP 2

a. ANALYSIS OF VARIANCE TABLE

R: R-squared: Adj.R-squared: RMS Residual:

0.878 0.77 0.751 332.179

Source DF: Sum Squares: Mean Square: F-test:

REGRESSION 2 8877358.138 4438679.069 40.226

RESIDUAL 24 2648222.38 110342.599

TOTAL 26 11525580.519

b. VARIABLES IN EQUATION

Variable: Coefficient: Std. Err.: Std. Coeff.: F to Remove:

INTERCEPT -138.21

ΔTOTAL FAT 0.073 0.008 0.891 80.138

ΔESR' -286.265 130.857 -0.218 4.786

c. VARIABLES NOT IN EQUAΗON

Variable: Par. Corr: F to Enter:

ΔMENO STATUS -0.084 0.163

ΔWT 0.054 0.068

ΔHT -0.227 1.254

ΔBMI 0.251 1.542

ΔLBM 0.033 0.025

ΔSHBG -0.092 0.197

ΔEST -0.161 0.614

ΔDHEAS 0.313 2.491

ΔCURRENT ERT 0.281 1.979

ΔSMOKING -0.113 0.297 Discussion:

The only two variables related to the within pair difference in central abdominal fat measured by quantitative dexa analysis was the difference in total fat, which is to be expected since abdominal fat is a component of total fat, and the estrogen receptor genotypic difference. The contribution to the sum of the squares (adjusted for total fat) contributed by the estrogen receptor was (using adjusted r-squared values) 75.1%-71.3= 3.8%. This indicates that the estrogen receptor genotypic status can explain a significant proportion of the variance in a highly environmentally influenced trait, central abdominal fat. In addition, this analysis ranks estrogen receptor genotypic status above all other available variables that might conceivably be related to central abdominal fat, except the total fat measure. In the previous example, we saw that estrogen receptor is related to fasting serum insulin levels even in the presence of spinal fat data. Spinal fat is derived from a para-spinal dexa instrument window and is an approximate measure of central abdominal fat. This analysis shows that the estrogen receptor genotype is a controller of central abdominal fat partitioning when total fat is taken into account. The data show that estrogen receptor can have two inter-related effects on physiological mechanisms which influence type 1 and type 2 diabetes susceptibility: namely that ESR regulates fasting insulin levels independent of spinal fat and the ESR regulates central abdominal fat independent of total fat. The mapping of this locus in type 1 diabetes (2), the linkage of ESR to fasting insulin levels in normal healthy subjects and hyperinsulinemia in the ESR mutant patient all point to the ESR as a susceptibility locus for both type 1 and type 2 diabetes (11). The data presented here should stimulate research into the role of ESR in diabetes and should alert investigators to the potential of estrogen analogues as agents for controlling insulin resistance. Recently, peptide hormone signalling by IGF-1 and EGF has been shown to result in phosphorylation of the estrogen receptor and activation by non- ligand mediated mechanisms (12). Diabetes is a highly prevalent disease which is yielding to genetic analysis. The multifactorial aetiology of type 1 diabetes (insulin dependent diabetes or IDDM) involves susceptibility genes and autoimmune reactions leading to pancreatic insufficiency of insulin secretion. Type 2 diabetes (non-insulin dependent diabetes or NIDDM) is also caused by environmental effects and genetic predispositions but has a later age of onset and is characterised by tissue resistance to insulin action. Environmental influences in type 2 diabetes include the accumulation of central abdominal fat as opposed to peripheral or subcutaneous fat (1). Regional fat distribution is strongly related to fasting insulin levels

(7) we used dual energy X-ray absorptiometry to measure centrally located fat content as spinal fat (a subset of total central fat). Fasting insulin was strongly related to spinal fat content (n=90, R2= 0.265, p=0.0001). When subjects were categorised according to ESR genotypes, significantly different mean insulin levels were observed in each genotype (p=0.04 by ANOVA). The within pair difference in spinal fat (ΔSPINAL FAT) in MZ twins can be attributed to individual specific environmental differences and measurement error. The within-pair difference in insulin (ΔINS) was strongly related to ΔSPINAL FAT in MZ twins (R2=0.44, p=0.0003) and less so in DZ twins (R2=0.14, p=0.1). The difference between correlations in MZ (who share all their genes) and DZ twins (who share only half their genes), suggests a genetic effect on insulin independent of central fat. In multiple regression ignoring zygosity, both ESR genotypes (coded; 1, 2, 3) and central fat (p=0.01 and p=0.001, respectively) were significantly and independently related to insulin levels. In forward stepwise regression, only ESR genotype and central fat remained in the equation (age, BMI and kg eliminated). ESR genotypes explained 6.8% of the variance in fat corrected insulin levels (p=0.01). In lean subjects (BMI <24kg/m2) both fat and ESR were significant in regression (n=46, R2=0.29, p=0.0003 for fat) and the strength of ESR effect improved, explaining 15% of the residual variance (p=0.008), indicating that the relationship is not dependent on individuals with excess BMI. As DZ twin pairs (n=20) can be treated as equal age siblings, we used the sib-pair variance method (8) and found a significant regression (R=-0.54, p=0.01) between the number of ESR alleles identical by state (ΔESR) and the within pair variance (ΔIN²). In DZ twins, ΔINS was related to both ΔSPINAL FAT and ΔESR in multiple regression (R2=0.28, p=0.06). Natural logarithmic transformation as (In 1+ insulin) normalised insulin values and improved all significance scores (analysis not shown).

The observation that simple RFLP in the ESR can explain 6.8% of the variance in central fat-corrected fasting insulin in normal subjects suggests that ESR is a major regulator of this quantitative trait. The result is most likely due to genetic differences in common ESR alleles rather than an unknown linked gene. The hyperinsulinemia and acanthosis nigricans in the estrogen receptor null mutation patient (2) suggests that the estrogen receptor itself is responsible for this effect rather than a linked gene. While the literature on ESR effects on insulin and diabetes is small, ESR is found in adipocytes and could potentially regulate insulin receptor (9). Finally, estrogen treatment prevents type 1 diabetes in a beta-cell expressing H-ras transgenic mouse model (10).

Despite the differences in environmental agents involved, genetic phenomena involved in type 1 and type 2 diabetes may have unexpected coincidences. Davies et al. (11) conducted a genome wide screen for susceptibility loci linked to type 1 diabetes using affected sib pairs. A notable feature was that several diabetes susceptibility genes reside in the vicinity of genes with powerful hormonal actions, in particular: estrogen receptor (ESR), tumour necrosis factor alpha (TNF-a), insulin (INS), and fibroblast growth factor (FGF-3). The susceptibility gene mapping close to ESR at 6q25 was assigned IDDM5. We suggest that IDDM5 is the estrogen receptor gene. Subsequent published data has replicated the mapping of IDDM5 with a peak at the ESR locus (American Journal of Human Genetics 57:911-919,1995), however no claim has been made that IDDM5 is the estrogen receptor and in no paper by these international experts has it been claimed that the ESR is a likely candidate for a gene related to type 1 or type 2 diabetes.

In a second twin study described above we have demonstrated that the estrogen receptor gene locus regulates the within-pair difference in central abdominal fat (a potent risk factor for type 2 diabetes, hypertension and heart disease) independent of the total fat content of the subjects. The data demonstrates that the ESR controls fat partitioning, permitting the detection of an effect on central abdominal fat in the presence of total fat data. These data verify that the ESR is involved in the regulation of central abdominal fat as well as the set point of fasting insulin levels. The effect of the ESR on the distribution of central fat may be derived from cross talk between the estrogen receptor and the peroxisome proliferator associated receptor-gamma (PPAR-gamma) or similar receptors which are essential regulators of adipocyte differentiation in fatty tissue. The ESR and the PPAR recognise similar DNA sequences in the promoters of target genes and can therefore be expected to interact in regulation. An example of this interaction has been demonstrated where an estrogen responsive promoter was activated by PPAR (13). Adipocytes contain estrogen receptor (14). An example exists in the literature showing that postmenopausal hormone replacement therapy reduces central abdominal fat content as measured by Dexa (15). An important distinction in that work in terms of prior art was that the hormone was a mixture of estrogen and progesterone and it was impossible to distinguish at that time which hormone was responsible for the effect.

The mapping of this locus in type 1 diabetes, the linkage of ESR to fasting insulin levels in normal healthy subjects and hyperinsulinemia in the estrogen receptor mutant patient all point to the ESR as a susceptibility locus for both type 1 and type 2 diabetes. The data presented here should stimulate research into the role of estrogen receptor in diabetes and should alert investigators to the potential of estrogen analogues as agents for controlling insulin resistance.

Conclusion. The present invention relates to a method of predicting genetic risk of having 1. low bone density 2. different serum fasting insulin 3. different central abdominal fat and 4. different rates of post menopausal bone loss and therefore osteoporosis risk using genetic analysis of the estrogen receptor gene. This gene can be combined with the two others previously demonstrated to regulate bone density (the vitamin D receptor gene and the retinoic acid receptor alpha gene) to provide a genetic test capable of discriminating risk. The genetic system permits the definition of groups of people with differing risks of subsequent bone disease, in particular post¬ menopausal osteoporosis. The relationship between ESR genotypes and serum insulin indicates that ESR genotypes play a role in insulin biology and therefore by inference to the development of insulin related disorders such as type 1 and type 2 diabetes. The genetic system provides a means of discriminating relative risk of bone fracture in later years such that particular genotypes have a higher risk of fracture, as a result of the relationship between bone density and fracture. The use of alleles of the ESR in the prediction of bone density and their relationship to serum insulin provide a finer means of discriminating bone density characteristics associated with genetic resistance and susceptibility to osteoporosis as well as other pathophysiological processes associated with the estrogen endocrine system, including in the broadest possible interpretation physiological processes influenced by the estrogen receptor.

In summary, the present inventor has developed a method which can provide useful information regarding the prognosis of an individual, in particular:- a. assessing ESR genotype provides a discrimination of different mean bone density set points, pertaining to various skeletal sites: with examples of the femoral neck , lumbar spine, Ward's triangle and trochanteric region.

b. identifying functionally different alleles of the ESR gene provides a means of discriminating differences in future bone loss and thus provides a prognostic indicator for future disease risk as demonstrated by differences in the inferred rate of post menopausal bone loss detectable in different ESR genotypes. c. functionally different ESR variants detects differences in mean fasting insulin levels representing a difference in set points of serum insulin metabolism which it is realised relates strongly to metabolic disturbances associated and causative of type 2 diabetes. It is recognised that functional different alleles of the ESR could relate to susceptibility to type 1 diabetes.

d. functionally different alleles of the ESR detects genetic groups with further differences in serum osteocalcin levels and hence different states of bone turnover reflecting differences in the physiological handling of osteocalcin.

e. functionally different alleles of the ESR detects identifies particular genetic groups with differences in the rate of loss of bone density after the menopause inferred from cross sectional data. ESR genotyping provides a means of identifying subjects at risk of higher bone loss at the menopause. ESR genotype provides an adjunct to the decision regarding therapy.

f. identification of functionally different alleles of the ESR provides a new test and as combined with other clinical markers, particularly osteocalcin levels and other genes related to bone density in particular the vitamin D receptor (VDR) and the retinoic acid receptor alpha (RAR-a), is able to discriminate ultra high bone density subjects who have low probability of suffering osteoporosis, as defined by low bone density and fracture.

g. ESR gene allele testing is capable of detecting subjects at higher relative risk of osteoporotic bone fracture due to higher rates of loss of bone density at all ages and from other causes.

h. within subjects who have already suffered bone fracture, functionally different alleles of the ESR identifies, still discriminates those with lower bone density and therefore a higher risk of subsequent additional fracture. i. alleles of the ESR gene provide a test which discriminates different levels of ESR gene functionality which can be related to the total syndrome of partial estrogen resistance, including low bone density, elevated insulin, relative central fat distribution, diabetes risk and problematic fertility.

j. alleles of the ESR gene can be used in conjunction with measures of spinal and central fat as a means of discriminating genetic and morphological effects on serum insulin as it pertains to the pre-diabetic and diabetic subject.

k. alleles of the ESR can be used to determine relative risk of developing central abdominal fat distribution.

j. ESR can be used as a target gene for the development of specific pharmaceuticals to treat osteoporosis, diabetes and central abdominal obesity.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

References

Arjmandi, B.H.; Salih, M.A.; Herbert, D.C.; Sims, S.H.and Kalu, D.N.

Evidence for estrogen receptor-linked calcium transport in the intestine. Bone and Min 1993; 21: 63-74.

Armamento-Villareal R, Villareal DT, Avioli LV, Civitelli R. Estrogen status and heredity are major determinants of premenopausal bone mass. J Clin

Invest 1992; 90:2464-2471.

Brommage, R.; Binacua, C. and Carrie, A. An-associated increase in intestinal calcium absorption during the rat estrous cycle is blunted by ovarectomy. Reproduction 1993; 49: 544-548.

Baski, S.N.; Kenny, A.D. Acute effect of estradiol on the renal vitamin D hydroxylase in Japanes. Quail; Biochem.pharmacol 1978; 27: 2765-2768.

Bellido T, Girasole G, Passeri G, Yu XP, Mocharla H, Jilka RL, Notides A, Manolagas SC. Demonstration of estrogen and vitamin D receptors in bone marrow-derived stromal cells: up-regulation of the estrogen receptor by 1,25-

Dihydroxyvitamin-D3. Edocrinology 1993; 133:553-562.

Brown TR, Lubahn DB, Wilson EM, Joseph DR, French FS, Migeon CJ.

Deletion of the steroid-binding domain of the human androgen receptor gene in one family with complete androgen insensitivity sydrome: evidence foe further genetic heterogeneity in this syndrome. Pro Nati Acad Sci. U.S.A.

1988; 85: 8.51-8.55.

Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K et al.

Bone density at various sites for prediction of hip fractures. Lancet 1993; 341:72-75.

Christian JC, Yu PL, Slemenda CW, Johnston CC. Heritability of bone mass: a longitudinal study in ageing male twins. Am J Hum Genet 1989 44:429-433.

Chan, S.D.H.; Chiu, D.H.and Atkins, D. Oophorectomy lead to a selective decrease in l,25-(OH)2 Vitamin D receptors in rat jejunal villous cells. Clin Science 1984; 66:745-748.

Dequeker J, Nijs J, Verstraeten A, Geusan P, Gevers G. Genetic determinants of bone mineral content at the spine and radius: a twin study. Bone 1987

8:207-209. Dieken ES, Meese EU and Miesfeld RL. nt' glucocorticoid receptor transcripts lack sequences encoding the amino-terminal transcriptional modulatory domain. Mol Cell Bio 1990; 10: 4574-4581.

Eriksen, E.F.; Colvarrd, D.S.; Berg, N.J.; Graham.M.L.; Mann, K.G.; Spelsherg, T.C. ; Riggs, B.L. Evidence of estrogen receptors in normal human osteoblast- like cells. Science 1988; 241: 84-87.

Evans RA, Marel GM, Lancaster EK, Kos S, Evans M, Wong SYP. Bone mass is low in relatives of osteoporotic patients. Ann Int Med 1988 109:870-873. Gennari, C; Agnusdei, D.; Nardi, P.; Civitelli. R. Estrogen preserves a normal intestinal responsiveness to 1,25-didroxy-vitamin D 3 in oophorectomized women. J Clin Endocrinol metab 1990; 71(5): 1288-1293. Green, G.L.; Gilna, P.; Waterfield, M.; Baker, A.; Hore, Y.; Shine, J. Sequence and expression of human estrogen receptor complementary DNA. Science 1986; 231: 1150-1154. Gallgher, J.C. etd. The pathogenesis of osteoporosis. J Bone Miner Res 1990; 9(3):215-227.

Harmon JM, Elsasser MS, Eisen LP, Urda LA, Ashraf J, and Thompson EB, Glucocorticoid receptoe expression in receptorless mutats isolated from the human leukemic cell line CEM-CT. Mol Endocrinol 1989; 3: 734-743. Heanney, R.P.; Recker, R.R.; Sanille, P.D. Menopausal changes in calcium balance perfomance. J Lab Clin Med, 1978; 92:953-963 Hill, S.M.; Fuqua, S.A.W.; Chamness, G.C.; Green, G.L.; McGuire, W.L. Estrogen receptor expression in human breast cancer associated with an estrogen receptor gene restriction fragment length polymorphism. Cancer 1989; 49: 145-148.

Horowitz, M.C. Cytokines and estrogen in bone: anti-osteoporotic effects. Science 1993; 260: 626-627.

Hosoi T, Inoue S, Kobayashi S, Sano M, Nakamura T, Ouchi Y, Shiraki M, Orimo H. Association of bone mineral density with polymorphisms of estrogen receptor gene in postmenopausal women. J Bone Miner Res 1995; 10: Abstract No. 125.

Hui SL, Slemenda CW, Johnton CC. Age and bone mass as predictors of fracture in a prospective studies. J. Clin. Invest.1987; 81:1804-1809. Hustmyer FG, Peacock M, Hui S, Johnston CC, Christian J. Bone mineral density in relation to polymorphism at the vitamin D receptor gene locus. J Clin Invest 1994 94:2130-2134.

Keeting PE, Scott RE, Colvard DS, Anderson MA, Oursler MJ, Spelsberg TC, Riggs BL. Development and characterisation of a rapidly proliferating, well- differentiated cell line derived from normal adult human osteoblast-like cells transfected with SV40 large T antigen. J Bone Miner Res 1992; 7:127-136 Kelly PJ, Hopper JL, Macaskill GT, Pocock NA, Sambrook PN, Eisman JA. Genetic factors in bone turnover. J Clin Endocrinol Metab 1991 72:808-813. Kelly PJ, Nguyen TV, Hopper JL, Pocock NA, Sambrook PN, Eisman JA. Changes in axial bone density with age: a twin study. J Bone Min Res 1993 8:11-17.

Melhus H, Kindmark A, Amer S, Wilen B, Lindh E, Ljunghall S. Vitamin D receptor genotypes in osteoporosis. Lancet 1994; 344:949-950. Morrison NA, Yeoman R, Kelly PJ, Eisman JA. Contribution of trans-acting factor alleles to normal vitamin D receptor gene polymorphisms and circulating osteocalcin. Proc Nat Acad Sci USA 1992 89:6665-6669. Morrison NA, Qi JC, Tokita A, Kelly PJ, Croft L, Nguyen TV, Sambrook PN, Eisman JA. Prediction of bone density by vitamin D receptor alleles. Nature 1994; 367:284-287.

Miyaura C, Kusano K, Masuzawa T, Chaki O, Onoe M, Sasaki T, Tatsuya T, Koishihara Y, Ohsugi Y and Suda T. Endogenous bone-resorbing factors in estrogen deficiency: Cooperative effects of IL-1 and IL-6. J Bone Miner Res 1995; 10; 1365-1373. Nguyen T, Sambrook P, Kelly P, Jones G, Lord S, Freund J, Eisman J.

Prediction of osteoporotic fractures by postural instability and bone density. Br Med J 1993; 307:1111-1115.

Pocock NA, Eisman JA, Hopper JL, Yeates M, Sambrook PN, Eberl S. Genetic determinants of bone mass in adults. J Clin Invest 1987 80:706-710 Ritchie HH, Hughes MR, Thompson ET, Malloy PJ, Hochberg Z, Feldman D, Pike JW, and O' malley BW. An ochre mutation in the vitamin D receptor gene causes herediary 1,25-dihydroxyvitamin D3-resistant richets in three families. Pro Natl Acad Sci U.S.A. 1989; 86: 9783-9787. Riggs, B.L.; Gallagher, J.C; De Luca, H.F.; Zinmeister, A.R. Studies on the impaired calcium absorption in postmenopausal osteoporosis. In vitamin D, chemical, Biohemical and clinical endocrinology of calcium metablism. P.903.Ed, Norman, A.W.; Shaefer, K.; Herrath, D.V.; Grigoleit, H.G.; Walter, deGruyter; Berlin, Yew York (1982).

Riggs BL, Nguyen TV, Melton LJ in, Morrison NA, O'Fallon WM, Kelly PJ, Egan KS, Sambrook PN, Muhs JM, Eisman JA. The contribution of vitamin D receptor gene alleles to bone mineral density in normal and osteoporotic women. J Bone Miner Res 1995; In press. Smith DM, Nance WE, Kang KW, Christian JC, Johnston CC. Genetic factors in determining bone mass. J Clin Invest 1973 52:2800-2808. Seeman E. Hopper JL, Bach LA et al. Reduced bone mass in daughters of women with osteoporosis. N Engl J Med 1989 320:554-558. Spector TD, Keen RW, Arden NK, Morrison NA, Major PJ, Nguyen TV. Kelly PJ, Baker JR, Sambrook PN, Lanchbury JS, Eisman JA. Influence of vitamin D receptor genotype on bone density in postmenopausal women: a UK twin study. Br Med J, In press.

Stock, J.L.; Coderre, J.A.; Mallette, L.E. Effects of a short course of estrogen on mineral metabolism in postmenopausal women. J.Clin Endocroml Melab 1985; 61: 595-600.

Smith. E.P.; Boyd, J.; Frank, G.R.; Takahashi, H.; Cohen, R.M.; Specker, B.; Williams, T.C.; Lubahn, D.B.; Korach, K.S. Estrogen resistance caused by a mutation in the estrogen-receptor gene in a man. N Engl J Med 1994; 331(16): 1088-1089. Vollenweider-Zerargui L, Barrelet L, Wong Y, Lemarchand-Bereau T, Gomez F. The predictive value of estrogen and progesterone receptors' concentration on the use clinical behaviour of breast cancer in women. Cancer 1986; 57:1171-1180.

1. Turner R.C. Hattersley A.T., Shaw J.T.E. & Levy J.C. Diabetes. 44, 1-10 (1995).

2. Smith E.P. et al. N Engl J Med. 331,1088-1089 (1994).

3. Korach KS. Science. 266, 1524-1527 (1994).

4. Ignar-Trowbridge D.M. et al. Mol Endocrinol. 7, 992-998 (1993). 5. Newton C.J. et al. J Steroid Biochem & Mol Biol. 48, 481-486 (1994).

6. Yaich L. et al. Cancer Res. 52, 77-83 (1992).

7. Kohrt W.M. et al. Diabetes. 42, 273-281 (1993).

8. Haseman JK & Elston, R.C. Behavl Genet. 2, 3-19 (1972). 9. Pedersen S.B., Borglum J.D., Moller-Pedersen T. & Richelsen B. Mol Cell Endocrinol. 85, 13-19 (1992).

10. Efrat S. Endocrinology. 128, 897-901 (1991).

11. Davies J.L. et al. Nature 371, 130-136 (1994).

12. Kato S., Endoh H., Masuhiro Y., Kitamoto T., Uchiyama S., Sasaki H, Masushige S., Gotoh Y., Nishida E.. Kawashima H., Metzger D., Chambon P.

Nature. 270: 1491-1494 (1995).

13. Keller H., Givel F., Perroud M., Whali W. Molecular Endocrinology. 9:794-804 (1995).

14. Pedersen SB., Borglum JD., Eriksen EF., Richelsen B. Biochimica et Biophysica Acta. 1093:80-86 (1991).

15. Haarbo J., Marslew U., Gotfredsen A., Christiansen C. Postmenopausal hormone replacement therapy prevents central distribution of body fat after menopause. Metabolism: Clinical & Experimental. 40:1323-1326 (1991).

Claims

CLAIMS :-

1. A method of assessing an individual's predisposition to low or high bone density, development of high or low bone density and/or responsiveness to therapy comprising analysing allelic variation in relation to the estrogen receptor gene of the individual.

2. A method of assessing an individual's predisposition to development of Type 1 or Type 2 diabetes and/or responsiveness to therapy comprising analysing allelic variation in relation to the estrogen receptor gene of the individual.

3. A method of assessing an individual's predisposition to development of central adiposity and/or insulin resistance comprising analysing allelic variation in relation to the estrogen receptor gene of the individual.

4. A method as claimed in claim 1 in which the method further involves analysing allelic variation in relation to the vitamin D receptor gene and/or the retinoic acid receptor alpha gene of the individual.

5. A method as claimed in any of claims 1 to 4 in which the ESR genotypes are detected with the restriction endonuclease PvuII.

6. A method of treating Type 1 or Type 2 diabetes in an individual comprising administering to the individual an effective amount of an estrogen analogue.