1998 Oxford University Press
Human Molecular Genetics, 1998, Vol. 7, No. 13 2129–2133
Calcitonin receptor polymorphism is associated with a decreased fracture risk in post-menopausal women J. Taboulet, M. Frenkian, J. L. Frendo, N. Feingold1, A. Jullienne and M. C. de Vernejoul* INSERM U349, Hôpital Lariboisière, 6 rue Guy Patin, 75475 Paris cedex 10, France and 1INSERM U155, Université Paris 7, case 7041, 2 place Jussieu, 75251 Paris cedex 05, France Received July 16, 1998; Revised and Accepted September 29, 1998
High bone resorption by the osteoclast results in osteoporosis, a disease affecting 40% of women after the menopause. Calcitonin, used to treat osteoporosis, inhibits bone resorption via receptors located on the osteoclasts. Two alleles of the calcitonin receptor gene (CTR) exist: a base mutation T→C in the third intracellular C-terminal domain changes a proline (CCG) at position 447 to a leucine (CTG). We therefore studied the distribution of these alleles in a cohort of 215 postmenopausal Caucasian women suffering or not from osteoporotic fractures. The region of interest within the point mutation was amplified by PCR and screened for single strand conformation polymorphism. This work was followed by DNA sequencing of the fragments amplified. We found that bone mineral density (BMD) at the femoral neck was significantly higher in heterozygous subjects with the Rr genotype compared with the homozygous leucine (RR) and homozygous proline (rr) genotypes. Also, a decreased fracture risk was observed in heterozygote subjects. In conclusion, our results suggest that polymorphism of CTR could be associated with osteoporotic fractures and BMD in a population of post-menopausal women. CTR heterozygotes could produce both alleles of the receptor. The heterozygous advantage effect of Rr subjects could explain their protection against osteoporosis: higher bone density and decreased fracture risk. Establishing the genotype of the CTR gene in post-menopausal women could be of value in evaluating their risk of developing fractures. INTRODUCTION Osteoporosis is a severe disease considering its high prevalence and the economic burden of femoral neck fractures. This disease has a strong heritability component and is under polygenic
control (1–3). Indeed, a relationship between a mutation of the vitamin D receptor and low bone mass (4,5) has been reported, but the results were not reproducible (6,7). Also, polymorphism at the SP1 site of the collagen type I promoter is associated with osteoporosis (8). These two genes are implicated in osteoblastic bone formation. However, biochemical data show that agerelated bone loss and fracture development are associated with increased bone resorption (9,10). One of the candidate genes of osteoporosis is the calcitonin gene. Calcitonin is a hormone implicated in bone resorption and acts through specific receptors present in large numbers in the osteoclasts (11,12). This hormone decreases bone resorption and is therefore used to treat osteoporosis (13). The calcitonin receptor is a member of the seven transmembrane receptor family and has recently been sequenced and cloned in several species (14–16). Three isoforms (17–20) of the mRNA of the human calcitonin receptor, resulting from alternative splicing (21), have been described. We previously reported a point mutation polymorphism (T→C) in the 3′-region of the calcitonin receptor gene (CTR) which induced a Pro→Leu shift in the third intracellular domain of the protein (22). The aim of this study was to determine if this polymorphism of the calcitonin receptor gene was associated with osteoporotic fractures. We used single strand conformation polymorphism (SSCP) to study the distribution of the alleles of this receptor in 215 post-menopausal women. The women we studied had no osteoporotic fractures (n = 123) or presented one or more osteoporotic fractures (n = 92) of wrist or vertebrae. RESULTS Table 1 reports the clinical data for the two groups of postmenopausal women we studied. Patients with osteoporotic fractures were older, smaller, thinner and menopause onset was earlier. Their bone mineral density (BMD) was lower compared with patients without fractures: femoral neck (0.689 ± 0.10 versus 0.807 ± 0.12 g/cm2, P < 0.02); lumbar vertebrae (0.859 ± 0.14 versus 1.047 ± 0.16 g/cm2, P < 0.01).
*To whom correspondence should be addressed. Tel: +33 1 49 95 63 58; Fax: +33 1 49 95 84 52; Email:
[email protected]
2130 Human Molecular Genetics, 1998, Vol. 7, No. 13 Table 1. Clinical data for the 215 post-menopausal women studied in the presence or absence of fractures – Fracture (n = 123) Age
65.06 ± 6.4
Weight
62.28 ± 10.04
Height
158.2 ± 5.5
+ Fracture (n = 92) 68.03 ± 7.0 59.7 ± 10.4 155.4 ± 6.5
49.8 ± 5.2
47.6 ± 6.8
BMD femoral neck (g/cm2)
0.807 ± 0.12
0.689 ± 0.10
BMD lumbar spine (g/cm2)
1.047 ± 0.16
0.859 ± 0.14
Age of menopause
Data are expressed as means ± SD. Patients with osteoporotic fractures were older, smaller, thinner and menopause onset was earlier. Their BMD was lower at the femoral neck and at the lumbar spine compared with patients without fractures.
Table 2. Observed and expected number of patients with the different genotypes in each group, with or without fractures RR
Rr
rr
Observed
53
64
6
Expected
59
52
12
Observed
52
32
8
Expected
50
36
6
– Fracture χ2 = 5.87, P < 0.05
+ Fracture χ2 = 0.45, P > 0.05
We have studied genomic DNA obtained from the 215 post-menopausal women by PCR–SSCP. Figure 1 displays the pattern of migration of some patients on the SSCP gels. We identified four bands of interest; the additional bands (1, 5 and 6) corresponded to another migration conformation, as confirmed by sequencing. The distribution of the genotypes is reported in Table 2. The distribution of the genotypes in both groups did not differ from the expected Hardy–Weinberg equilibrium for patients with fractures, but patients without fractures had an excess of heterozygotes (Table 2). The numbers of patients with or without osteoporotic fractures for the three genotypes are reported in Figure 2. Compared with the RR and rr homozygotes, the prevalence of Rr heterozygotes was lower in the group with osteoporotic fractures (χ2 = 6.63, P < 0.036, df = 2). As a consequence, the relative risk factor for vertebral fracture was 1.95 (95% confidence limits 1.16–3.3) for homozygotes. If women with wrist fractures were excluded from the analysis the difference between the genotypes was maintained (χ2 = 7.031, P < 0.029). Thus, CTR heterozygotes showed a distinct advantage over homozygotes as far as fractures were concerned. Low BMD is a strong risk factor for osteoporotic fractures. We therefore examined whether this parameter was different between the genotypes studied (Table 3). Age, weight, height and the age of menopause onset were not different. BMD at the femoral neck was significantly higher in patients with an Rr genotype than in patients with the RR genotype. The same trend, but not significant, for a higher value in Rr heterozygotes was observed for BMD at the lumbar spine. Covariance analysis with age, weight, height, age of menopause and duration of hormone
Figure 1. CTR polymorphism is at amino acid 447 where a T→C mutation changes a leucine to a proline. PCR amplification of genomic DNA was carried out using specific primers bordering the polymorphic 447 site. (A) Autoradiography of SSCP gel. Lanes a–f, DNA from individuals with the leucine allele (bands 1 and 3, additional bands 5 and 6); lanes i and j, DNA from patients with the proline allele (bands 2 and 4 and sometimes additional band 1); lanes g, h and k, DNA from patients with the two alleles; lane l, DNA from the TT cell line with the two alleles. ND, non-denatured DNA. (B) Partial sequence of PCR products amplified from patients homozygous for proline or leucine alleles.
Figure 2. Distribution of the genotypes of CTR in 215 post-menopausal women with (+ fracture) or without (– fracture) fractures. Analysis of the CTR genotype distribution in patients with osteoporotic fractures compared with patients without osteoporotic fractures showed that heterozygotes (Rr) have a distinct advantage over homozygotes (RR and rr) concerning fractures. – Fracture / + fracture χ2 = 6.63, P < 0.037.
2131 Human Genetics, 1998, 7, No. NucleicMolecular Acids Research, 1994, Vol. Vol. 22, No. 1 13 2131 receptor containing either leucine or proline were detected after reverse transcription. DISCUSSION
Figure 3. A polymorphic CTR site is related to BMD. BMD at the femoral neck and lumbar spine in 199 subjects according to their genotype. Results are expressed as the means ± SEM. Results were analyzed by covariance analysis to correct for the effects of age, weight, height, age of menopause and duration of hormone replacement therapy. p indicates the level of significance between the two means (Fisher’s LSD test). This analysis confirmed the existence of a significantly higher BMD value at the femoral neck in Rr versus RR and rr subjects and at the lumbar spine in Rr versus RR subjects.
therapy as co-factors confirmed the existence of a significantly higher BMD value at the femoral neck and lumbar spine in Rr versus RR. The women with the rr genotype also had a value lower than Rr but only at the femoral neck (Fig. 3). It is to be noted that in women aged >60 years presenting with vertebral fractures, measurement of BMD at the lumbar spine is subject to several artefacts in comparison with measurements at the femoral neck, which are more reliable (23). Indeed, in our population femoral neck BMD was negatively correlated with the number of fractures (r = –0.44, P < 0.0001). Table 3. Clinical data for the 215 post-menopausal women studied for each CTR genotype RR (n = 105) Age Weight
66 ± 7 61.8 ± 10.6
Rr (n = 96) 66 ± 7 61.2 ± 10.8
rr (n = 14) 70 ± 7 58.8 ± 7.8
157.1 ± 6.2
157.0 ± 5.8
48.6 ± 6.2
49.5 ± 5.9
49.2 ± 4.3
BMD femoral neck (g/cm2)
0.748 ± 0.118
0.780± 0.134
0.698 ± 0.131
BMD lumbar spine (g/cm2)
0.952 ± 0.176
0.997 ± 0.203
0.932 ± 0.161
Height Age of menopause
154.2 ± 5.5
Data are expressed as means ± SD. Age, weight, height and the age of menopause onset were not different. BMD at the femoral neck was significantly higher in patients with an Rr genotype than in patients with an rr genotype (ANOVA, P < 0.02), RR homozygotes had an intermediate value of bone density but the difference from the other genotypes was not significant. The BMD at the lumbar spine showed the same trend for Rr heterozygotes but fell short of significance.
We also determined if the two alleles were transcribed and expressed. Reverse transcription of RNA extracted from cell lines showed that T47D (RR) and HL60 (rr) transcribed the leucine and the proline allele, respectively. In the TT cell line (Rr), both alleles were transcribed, as PCR isoforms of the cDNA of the calcitonin
These results confirm the existence of two allelic forms of the human calcitonin receptor gene differing by a single base mutation changing a proline to a leucine in the peptidic sequence of hCTR (22). The distribution of calcitonin receptor alleles in a sample of the Japanese population (24) is quite different from that we observed in our Caucasian population. In Japan the proline homozygote (rr) is the most frequent genotype (70%), while in our sample proline homozygotes were rare (6.5%). The risk of vertebral fractures is higher in the Japanese population (25). It is interesting to note that the frequency of heterozygotes in the Japanese population is much lower (20%) than in the population we studied (41%) and could perhaps account for the increase in risk of vertebral fractures in this population (25). A recent work in a sample of post-menopausal Italian women without osteoporotic fractures show a similar distribution of the CTR genotypes, the rr genotype representing almost 20% of the population. No significant differences in BMD were observed between the three genotypes by ANOVA analysis (26). Our results clearly show that the subjects who are proline/ leucine heterozygotes have a higher bone density than the proline or leucine homozygotes. Moreover, these women have a strikingly lower incidence of vertebral fractures and fracture risk. The data on the genotypes according to fracture were compatible with an advantage of heterozygotes. This last hypothesis is supported by the low BMD in both homozygous groups compared with the heterozygotes. A further argument is the high conservation of the proline residue in all species studied with the exception of the human isoform. The absence of the proline residue could alter the secondary structure of the calcitonin receptor, more so as two other proline residues border this region. Progressive truncation of the C-terminal tail of the porcine receptor greatly decreased internalization of the ligand–receptor complex and reduced the magnitude of adenylate cyclase responses (27). This underlines the importance of the CTR C-terminal domain and suggests that the proline/leucine mutation could alter receptor biological activity. Finally, heterozygotes could produce both alleles of the receptor, resulting in an advantage as compared with homozygotes, especially as this genetic effect accounts for 13% of the total variance of BMD in covariance analysis. However, these results will need to be confirmed on a much larger series of patients as rr homozygotes are rare (n = 14). Several mRNA isoforms of hCTR (17–20) can be generated by alternative splicing (21) of one specific gene. The two main isoforms reported differ by the presence of a 16 amino acid insertion in the first intracytoplasmic domain. Both stimulate adenylate cyclase but only the truncated one is able to activate the inositol phosphate pathway (28). This diversity of mRNA receptor isoforms is independent of the allelic polymorphism we report here, since the same splicing events could arise on the two alleles, as assessed by sequences so far reported (20). The two mRNA isoforms of hCTR are expressed in the homozygous RR T47D cell line (18). HL60 cells, which are rr homozygotes, also express both isoforms. Therefore, there is no clear relationship between the occurrence of these mRNA isoforms and the polymorphism we describe.
2132 Human Molecular Genetics, 1998, Vol. 7, No. 13 Calcitonin’s physiological role (29) in age-related bone loss and osteoporosis is controversial (30). However, it is the only protein which binds specifically to the osteoclast membrane and has a well-demonstrated direct anti-resorptive activity. Our results add an important point to the debate because a mutation of the receptor for this hormone, in a region of likely functional significance, is associated with fracture and low bone mass. Furthermore, the association of a mutation in the receptor for this hormone with osteoporosis fits with the concept that bone loss with age and fractures are associated with high resorption. The higher bone density and decreased fracture risk in patients heterozygous for the CTR polymorphism could also be due to interaction of the receptor gene with other genes implicated in osteoporosis. Finally, our results suggest that the calcitonin receptor gene is one of the candidate genes for osteoporosis.
MATERIALS AND METHODS Clinical subjects We studied 215 post-menopausal French women of Caucasian origin at our clinic. All the subjects gave their informed consent for the study. The women were either volunteers, part of an epidemiological cohort (n = 143) or referred for vertebral fractures (n = 72). All the women completed an osteoporosisoriented questionnaire (including a dietary questionnaire) and had a spine X-ray. Vertebral crush fractures were assessed according to the Genant criteria (31). The bone density was measured in 199 patients at the lumbar spine and the femoral neck with a Lunar DPX-L (Madison, WI). Among the volunteers of the epidemiological cohort, 20 had presented with an osteoporotic fracture (wrist or vertebrae). We therefore had a group of 123 women (mean age 65 ± 6, 50–81) without osteoporotic fractures. Of the sample 31% had received hormone replacement therapy (HRT) for 2.4 ± 4.8 years. According to the WHO definition of osteoporosis based on bone density, 4.1% of them had osteoporosis, 52.1% osteopenia and 43.8% were normal. Among the 92 women with osteoporotic fractures, 75 had presented with crushed vertebrae and 17 with wrist fractures. The average number of crushed vertebrae was 3.3 ± 1.5. The weight and height, and age of menopause were lower in women with fractures compared with women without fractures.
Tissue culture
Biochemistry DNA extraction. DNA extraction from the patients and the TT, HL 60 and T47D cell lines was carried out using Trizol solution (Life Technologies). RNA extraction. Total RNA was extracted from TT and HL60 cells using the guanidine isothiocyanate/chloroform method (32). Reverse transcription of RNA. An aliquot of 1 µg total RNA extracted from the cell lines was reversed transcribed (Superscript II; Life Technologies) using an oligo(dT) primer for 10 min at 23C and extension was carried out at 42C for 50 min. PCR amplification of the cDNA. Aliquots of 200 ng cDNA were first denaturated for 5 min at 95C. The amplification reaction was carried out using Taq polymerase (Gibco BRL) for 30 cycles: 94C, 30 s; 60C, 30 s; 72C, 30 s. The reaction was terminated by 5 min elongation at 72C. PCR–SSCP analysis of the calcitonin receptor gene. We searched for the proline/leucine mutation by PCR–SSCP on DNA extracted from cell lines or from whole blood. PCR amplification was performed using two oligonucleotides primers (Genosys) (forward, 5′-ATTCAGTGGAACCAGCTTG-3′; reverse 5′-GATGGCTCAGTGATCACGAT-3′) located on each side of the base pair mutation. [33P]dCTP (2 µCi, 2000 Ci/mmol; Dupont de Nemours–NEN) was included. Aliquots of 200 ng genomic DNA were first denaturated for 5 min at 95C. The amplification reaction was carried out using Taq polymerase (Gibco BRL) for 30 cycles: 94C, 1 min; 60C, 1 min; 72C, 2 min. The reaction was terminated by 5 min elongation at 72C. The denaturated amplification products were separated overnight by gel electrophoresis (6% polyacrylamide, 10% glycerol, non-denaturing gel) at room temperature. DNA sequencing of PCR products. After vacuum drying of the gel, autoradiography was performed for 24–48 h. The bands revealed in the SSCP gels were extracted, re-amplified by PCR using the same oligonucleotides primers, purified and sequenced (Double Strand Sequencing kit; Life Technologies). Samples from 100 patients picked at random were also subjected to PCR using the same primers we used for SSCP and digested overnight with AluI (24). Fragments were separated by agarose gel electrophoresis and stained with ethidium bromide. Identical genotypes were observed with both methods. Statistical analysis
TT cell line (medullary thyroid carcinoma cell line). TT cells were grown in RPMI 1640 medium with 10 mM HEPES, 6 mM glutamine supplemented with 15% fetal calf serum (FCS). Medium was changed every 2 days until confluence. HL 60 (promyelocytic cell line) and T47D (breast carcinoma cell line) cell lines. The cells were cultured in RPMI 1640 in the presence of 2 mM glutamine, penicillin/streptomycin supplemented with 10% FCS. Cells were seeded at a concentration of 250 000 cells/ml. Medium was changed every 2 days. Cells were grown at 37C in 5% CO2.
Results are expressed as means ± SD and were analyzed using variance analysis. Bone density measurements were corrected for age, weight, height, age of menopause and duration of HRT by covariance analysis. Frequency of alleles and fractures were analyzed using the χ2 distribution. ACKNOWLEDGEMENTS This worked was supported by a GENSET fellowship to M.F. and by an individual grant from the French Ministry of Research and Education to J.L.F.
2133 Human Genetics, 1998, 7, No. NucleicMolecular Acids Research, 1994, Vol. Vol. 22, No. 1 13 2133 ABBREVIATIONS BMD, bone mineral density; CTR, calcitonin receptor; HRT, hormone replacement therapy; RT–PCR, reverse transcription– polymerase chain reaction; SSCP, single strand conformation polymorphism. REFERENCES 1. Pocock, N., Eisman, J.A., Hopper, J.L., Yeates, M.G., Sambrook, P.N. and Eberl, S. (1987) Genetic determinants of bone mass in adults: a twin study. J. Clin. Invest., 80, 706–710. 2. Slemenda, C.W., Christian, J.C., Williams, C.J., Norton, J.A. and Johnston, C.J. (1991) Genetic determinants of bone mass in adult women: a reevaluation of the twin model and the potential importance of gene interaction on heritability estimates. J. Bone Mineral Res., 6, 561–567. 3. Seeman, E., Tsalamandris, C., Formica, C., Hopper, J.L. and Mckay, J. (1994) Reduced femoral neck bone density in the daughters of women with hip fractures: the role of low peak bone density in the pathogenesis of osteoporosis. J. Bone Mineral Res., 9, 733–739. 4. Morrison, N.A. et al. (1994) Prediction of bone density from vitamin D receptor alleles [see comments]. Nature, 367, 284–287. 5. Yanagi, H. et al. (1996) Vitamin D receptor gene polymorphisms are associated with osteoporosis in Japanese women [letter; comment]. J. Clin. Endocrinol. Metab., 81, 4179–4181. 6. Garnero, P., Borel, O., Sornay, R.E., Arlot, M.E. and Delmas, P.D. (1996) Vitamin D receptor gene polymorphisms are not related to bone turnover, rate of bone loss and bone mass in postmenopausal women: the OFELY Study. J. Bone Mineral Res., 11, 827–834. 7. Houston, L.A., Grant, S.F., Reid, D.M. and Ralston, S.H. (1996) Vitamin D receptor polymorphism, bone mineral density and osteoporotic vertebral fracture: studies in a UK population. Bone, 18, 249–252. 8. Grant, S.F. et al. (1996) Reduced bone density and osteoporosis associated with a polymorphic Sp1 binding site in the collagen type I alpha 1 gene. Nature Genet., 14, 203–205. 9. Seibel, M.J. et al. (1993) Urinary hydroxypyridinium crosslinks of collagen as markers of bone resorption and estrogen efficacy in postmenopausal osteoporosis. J. Bone Mineral Res., 8, 881–889. 10. Garnero, P. et al. (1996) Markers of bone resorption predict hip fracture in elderly women: the EPIDOS Prospective Study. J. Bone Mineral Res., 11, 1531–1538. 11. Chambers, T.J. and Magnus, C.J. (1982) Calcitonin alters behaviour of isolated osteoclasts. J. Pathol., 136, 27–39. 12. Nicholson, G.C., Moseley, J.M., Sexton, P.M., Mendelsohn, F.A. and Martin, T.J. (1986) Abundant calcitonin receptors in isolated rat osteoclasts. Biochemical and autoradiographic characterization. J. Clin. Invest., 78, 355–360. 13. Reginster, J.Y. (1993) Calcitonin for prevention and treatment of osteoporosis [review]. Am. J. Med., 95, 44S–47S. 14. Lin, H.Y. et al. (1991) Expression cloning of an adenylate cyclase-coupled calcitonin receptor. Science, 254, 1022–1024.
15. Gorn, A.H. et al. (1992) Cloning, characterization and expression of a human calcitonin receptor from an ovarian carcinoma cell line. J. Clin. Invest., 90, 1726–1735. 16. Shyu, J.F., Inoue, D., Baron, R. and Horne, W.C. (1996) The deletion of 14 amino acids in the seventh transmembrane domain of a naturally occurring calcitonin receptor isoform alters ligand binding and selectively abolishes coupling to phospholipase C. J. Biol. Chem., 271, 31127–31134. 17. Frendo, J.L. et al. (1994) An isoform of the human calcitonin receptor is expressed in TT cells and in medullary carcinoma of the thyroid. FEBS Lett., 342, 214–216. 18. Kuestner, R.E. et al. (1994) Cloning and characterization of an abundant subtype of the human calcitonin receptor. Mol. Pharmacol., 46, 246–255. 19. Albrandt, K. et al. (1995) Molecular cloning and functional expression of a third isoform of the human calcitonin receptor and partial characterization of the calcitonin receptor gene. Endocrinology, 136, 5377–5384. 20. Gorn, A.H. et al. (1995) Expression of two human skeletal calcitonin receptor isoforms cloned from a giant cell tumor of bone. The first intracellular domain modulates ligand binding and signal transduction. J. Clin. Invest., 95, 2680–2691. 21. Nussenzveig, D.R., Mathew, S. and Gershengorn, M.C. (1995) Alternative splicing of a 48-nucleotide exon generates two isoforms of the human calcitonin receptor. Endocrinology, 136, 2047–2051. 22. Taboulet, J. et al. (1996) Evidence for two allelic forms of calcitonin receptor gene: distribution in normal and osteoporotic women. J. Bone Mineral Res., 11 (suppl. 1), S204. 23. Genant, H.K. et al. (1996) Noninvasive assessment of bone mineral and structure: state of the art. J. Bone Mineral Res., 11, 707–730. 24. Nakamura, M. et al. (1997) Allelic variants of human calcitonin receptor in the Japanese population. Hum. Genet., 99, 38–41. 25. Ross, P.D. et al. (1995) Vertebral fracture prevalence in women in Hiroshima compared to Caucasians or Japanese in the US. Int. J. Epidemiol., 24, 1171–1177. 26. Masi, L. et al. (1998) Allelic variants of the Human Calcitonin Receptor: distribution and association with bone mass in postmenopausal italian women. Biochem. Biophys. Res. Commun., 245, 622–626. 27. Findlay, D.M. et al. (1994) Truncation of the porcine calcitonin receptor cytoplasmic tail inhibits internalization and signal transduction but increases receptor affinity. Mol. Endocrinol., 8, 1691–1700. 28. Nussenzveig, D.R., Thaw, C.N. and Gershengorn, M.C. (1994) Inhibition of inositol phosphate second messenger formation by intracellular loop one of a human calcitonin receptor. Expression and mutational analysis of synthetic receptor genes. J. Biol. Chem., 269, 28123–28129. 29. Boden, S.D. and Kaplan, F.S. (1990) Calcium homeostasis. Orthop. Clin. North Am., 21, 31–42. 30. Body, J.J. (1993) Calcitonin: from the determination of circulating levels in various physiological and pathological conditions to the demonstration of lymphocyte receptors [review]. Hormone Res., 39, 166–170. 31. Genant, H.K., Wu, C.Y., Van Ruijk,C. and Nevitt, M.C. (1993) Vertebral fracture assessement using a semi quantitative technique. J. Bone Mineral Res., 8, 1137–1148. 32. Chomczynski, P.A.S. (1987) Single step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Anal. Biochem., 162, 156–159.
