Transforming growth factor β (TGF-β) is an important regulator of bone metabolism, its effects being intertwined with those of estrogen and vitamin D. A T→C polymorphism in exon 1 of the TGF-β1 gene, which results in the substitution of proline for leucine, is associated with bone mineral density (BMD). However, it is not known whether this polymorphism affects the response to treatment with active vitamin D or to hormone replacement therapy (HRT) in individuals with osteoporosis. Changes in BMD at the lumbar spine (L2–L4 BMD) were compared among TGF-β1 genotypes in 363 postmenopausal Japanese women who were divided into three groups: an untreated, control group (n = 130), an active vitamin D treatment group (n = 117), and an HRT group (n = 116). TGF-β1 genotype was determined with an allele-specific polymerase chain reaction assay. In the control group, the rate of bone loss decreased according to the rank order of genotypes TT (homozygous for the T allele) > TC (heterozygous) > CC (homozygous for the C allele), with a significant difference detected between the CC and TT genotypes. The positive response of L2–L4 BMD to HRT increased according to the rank order of genotypes TT < TC < CC, although the differences among genotypes were not statistically significant. Individuals with the CC genotype responded to active vitamin D treatment with an annual increase in L2–L4 BMD of 1.6%, whereas those with the TT or TC genotypes similarly treated lost bone to a similar extent as did untreated subjects of the corresponding genotype. These results suggest that TGF-β1 genotype is associated with both the rate of bone loss and the response to active vitamin D treatment.
Bone mineral density (BMD), a major determinant of the risk of osteoporotic fracture, exhibits a strong genetic component, and several genes, including those encoding the vitamin D receptor (VDR), the estrogen receptor (ER), collagen Iα1, apolipoprotein E, parathyroid hormone, and transforming growth factor β1 (TGF-β1), have been implicated as genetic determinants of osteoporosis.(1–14) Although these gene polymorphisms have been associated with BMD, the underlying molecular mechanisms remain unclear. It is not known whether these polymorphisms affect the function of the gene product or whether they are merely linked to other genes that are responsible for BMD. VDR gene polymorphisms have been shown to be related to the response to treatment with active vitamin D and to fractional calcium absorption, suggesting that they affect the function of the VDR.(15,16) Regarding the estrogen receptor (ER), polymorphisms in ER-α gene do not appear to be associated with the response to hormone replacement therapy (HRT), although the involvement of ER-β gene polymorphisms remains to be elucidated.(17)
TGF-β is an important local regulator of bone metabolism, acting downstream of estrogen and cooperatively with vitamin D signaling pathways.(18–25) We have previously shown that a T→C polymorphism at nucleotide 29 of the TGF-β1 gene, which results in the replacement of leucine at position 10 in the signal sequence by a proline residue, is associated with BMD at the lumbar spine (L2–L4) in postmenopausal Japanese women, with the T allele representing an independent risk factor for the development of osteoporosis; individuals carrying the T allele were 15–33 times more susceptible to osteoporosis than were those homozygous for the C allele.(14) In addition, we have shown that BMD increases according to the rank order of genotypes TT < TC < CC in Japanese adolescents, suggesting that the TGF-β1 gene may be a determinant of peak bone mass.(26)
In the present study, we examined whether the T29→C polymorphism in the TGF-β1 gene modulates responses to treatment in individuals with postmenopausal osteoporosis. To address these issues, we investigated the relation between TGF-β1 genotype and changes in L2–L4 BMD in three groups of postmenopausal Japanese women: nontreated controls and those treated with either active vitamin D or HRT.
The study population comprised 363 unrelated postmenopausal Japanese women (mean age ± SD, 62.2 years ± 10.8 years) who either visited outpatient clinics of or were admitted to one of the participating hospitals (National Chubu Hospital, National Hyogo-Chuo Hospital, Keio University Hospital, and the Research Institute and Practice for Involutional Diseases) between 1992 and 1998, and were followed up for measurement of BMD for at least 1 year (mean follow-up period ± SD, 2.8 years ± 1.7 years). Individuals with disorders known to cause abnormalities of bone metabolism, including diabetes mellitus; renal diseases; rheumatoid arthritis; and thyroid, parathyroid, or other endocrinological diseases, were excluded from the study. The study population was divided into three groups: the control group, the active vitamin D treatment group, and the HRT group. The control group consisted of 130 individuals who were followed up without treatment that affects bone metabolism; this group included individuals with osteoporosis or osteopenia as well as individuals with normal BMD. The active vitamin D treatment group comprised 117 individuals who were treated with alfacalcidol (1α-hydroxyvitamin D3, 0.5–1 μg daily) or calcitriol (1α,25-dihydroxyvitamin D3, 0.25–0.5 μg daily). The HRT group consisted of 116 individuals who were treated with conjugated equine estrogen (0.625 mg daily) either alone or in combination with medroxyprogesterone acetate in a cyclic (5–10 mg daily) or continuous (2.5 mg daily) regimen. Individuals in the active vitamin D treatment and HRT groups did not take other drugs that affect bone metabolism, nor did they switch to other drugs during the follow-up period. The study protocol was approved by the Committee on the Ethics of Human Research of National Chubu Hospital, Obu, Japan and the National Institute for Longevity Sciences, Obu, Japan and informed consent was obtained from each subject.
Measurement of BMD
L2–L4 BMD was measured before, during, and after the follow-up period by dual-energy X-ray absorptiometry with a DPX or DPX-L instrument (Lunar, Madison, WI, U.S.A.) or with an XR-26 or XR-36 machine (Norland Medical Systems, Fort Atkinson, WI, U.S.A.). The L2–L4 BMD values obtained with XR machines were adjusted with the use of a cross-calibration equation to be consistent with those obtained with the DPX machines.(27) The rate of change in L2–L4 BMD was expressed as the percentage change per year between BMD values before and those after treatment. BMD was expressed as percentages of the young adult (20–44 years) reference mean (YARM) and the age-matched reference mean (AMRM) on the basis of data for Japanese women provided by the Japanese Society for Bone and Mineral Research.(28) The diagnosis of osteoporosis was based on the criteria recommended by the Japanese Society for Bone and Mineral Research, which are either an L2–L4 BMD of <80% of the YARM in the presence of nontraumatic vertebral fracture or an L2–L4 BMD of <70% of the YARM in the absence of nontraumatic vertebral fracture.(28) Vertebral fractures were diagnosed from lateral X-ray films of the thoracic and lumbar spine on the basis of the criteria described by Riggs et al.(29)
Genotyping the TGF-β1 gene
Venous blood was collected into tubes containing EDTA-disodium salt (50 mM), and genomic DNA was isolated. The genotype for the TGF-β1 gene was determined with an allele-specific polymerase chain reaction assay as previously described.(14) The assay accurately detected the T29→C polymorphism in exon 1 of the TGF-β1 gene, which was confirmed by direct DNA sequencing with a fluorescence-based automated DNA sequencer (Prism 310; Applied Biosystems, Foster City, CA, U.S.A.).
Data were compared among TGF-β1 genotypes by oneway analysis of variance and Scheffe's multiple range test. Qualitative data were compared by the chi-square test. Allele frequencies were estimated by the gene counting method, and the chi-square test was used to identify significant departure from Hardy-Weinberg equilibrium. We performed multivariable regression analysis to adjust factors, with the annual change in L2–L4 BMD as a dependent variable and independent variables including age, height, body weight, time since menopause, L2–L4 BMD, and follow-up period. A p value of <0.05 was considered statistically significant.
The baseline characteristics of the three groups of study subjects are summarized in Table 1. Time since menopause was shorter and, consequently, subjects were younger in the HRT group, compared with the other two groups. At baseline, L2–L4 BMD was highest in the control group and lowest in the active vitamin D treatment group. The three groups were followed up for measurement of L2–L4 BMD for more than 1 year, and the annual changes in this parameter were analyzed according to TGF-β1 genotype. The genotype distributions within all three groups as well as in the total study population were in Hardy-Weinberg equilibrium, suggesting that the subjects shared a homogeneous genetic background. The distributions of TGF-β1 genotype did not differ significantly between control group and active vitamin D treatment or HRT group. Age, height, body weight, body mass index, time since menopause, follow-up period, and frequencies of ovariectomy or of hip or vertebral fractures did not differ significantly among the TGF-β1 genotypes in each group (data not shown).
A significant relation between TGF-β1 genotype and changes in L2–L4 BMD was observed (Fig. 1). In the control group, the rate of bone loss decreased according to the rank order of genotypes TT > TC > CC, with a significant difference apparent between individuals with the CC and TT genotypes (Fig. 1A). In the active vitamin D treatment group, women with the TT or TC genotypes lost bone to a similar extent as did untreated women with the corresponding genotypes, whereas individuals with the CC genotype responded to active vitamin D treatment with an annual increase in L2–L4 BMD of 1.6% (Fig. 1B). In the HRT group, L2–L4 BMD increased irrespectively of genotype; although the annual gain increased according to the rank order of genotypes TT < TC < CC, there were no significant differences among genotypes (Fig. 1C).
Multivariable regression analysis, with adjustment for age, height, body weight, time since menopause, L2–L4 BMD, and follow-up period, revealed that the annual increase in BMD in individuals with the CC genotype who were treated with active vitamin D was significantly greater than that in controls with the same genotype, whereas no significant differences were observed between controls and active vitamin D–treated individuals with either the TC or TT genotypes (Table 2). For all genotypes, the annual gain in BMD was significantly greater for individuals in the HRT group than for controls (Table 2).
Table Table 1.. Baseline Characteristics of the Study Subjects
Active vitamin D
Data are means ± SD.
*p < 0.01 versus control and active vitamin D treatment groups; †p < 0.05 versus control; ‡p < 0.01 versus control and HRT; §p < 0.01 versus control; ‖p < 0.05 versus control and active vitamin D; ¶p < 0.01 versus control, p < 0.05 versus HRT.
No. of subjects
65.4 ± 8.4
67.7 ± 9.1
53.2 ± 9.1*
150.6 ± 5.3
148.5 ± 6.9†
154.5 ± 5.0*
Body weight (kg)
50.3 ± 7.6
47.3 ± 8.1‡
50.3 ± 5.4
Body mass index (kg/m2)
22.2 ± 2.9
21.3 ± 3.8
21.1 ± 2.2†
Time since menopause (years)
15.8 ± 8.9
19.5 ± 9.4§
7.4 ± 6.7*
Femoral neck fracture (%)
Vertebral fracture (%)
L2–L4 BMD (g/cm2)
0.868 ± 0.182
0.759 ± 0.159‡
0.830 ± 0.154
L2–L4 BMD (%YARM)
73.3 ± 15.2
64.8 ± 13.9¶
69.8 ± 12.9
L2–L4 BMD (%AMRM)
95.0 ± 17.5
86.6 ± 18.7§
79.6 ± 13.7*
Follow-up period (years)
2.3 ± 1.3
2.6 ± 1.5
3.7 ± 1.9*
TGF-β1 is produced by osteoblasts, stored in substantial amounts in the bone matrix, and an important regulator of both skeletal development and homeostasis of bone metabolism.(30) TGF-β is released during bone resorption and activated by the acidic microenvironment.(31) Once activated, TGF-β stimulates the proliferation and differentiation of, as well as matrix production by, osteoblasts and induces bone formation in vivo.(30,32) We have now shown that the T29→C polymorphism in the TGF-β1 gene is associated with the rate of bone loss in postmenopausal Japanese women. This observation, together with our previous demonstration that this polymorphism is related to BMD in high school students, suggests that TGF-β1 genotype affects both the peak bone mass attained during adolescence and the rate of bone loss later in life, at least in the Japanese population.(26)
The Leu10→Pro polymorphism of TGF-β1 is located in the 29-residue signal peptide, which is thought to target newly synthesized protein to the endoplasmic reticulum.(33,34) Our previous observation that the serum concentration of TGF-β1 increases with the number of C alleles suggests that the polymorphism may affect intracellular trafficking or export efficiency of TGF-β1.(14) The observation that TGF-β1–deficient mice exhibit reduced bone mass lends further support to the concept that bone mass is regulated by TGF-β1 in a dose-dependent manner.(35)
We have now shown that TGF-β1 genotype affects the therapeutic response to active vitamin D in postmenopausal women with osteoporosis. Multiple regression analysis, with adjustment for age, height, body weight, time since menopause, L2–L4 BMD, and follow-up period, revealed that the annual increase in BMD in individuals with the CC genotype who were treated with active vitamin D was significantly greater than that in controls with the same genotype, whereas no significant differences were observed between controls and active vitamin D-treated individuals with the T allele. Calcitriol and its prodrug alfacalcidol have been used to treat osteoporosis in various countries including Japan.(36) Several randomized, prospective clinical trials performed in the United States have shown that calcitriol is effective in reducing both bone loss and the frequency of new vertebral fractures.(37–39) However, the effect of active vitamin D on bone mass is relatively small compared with those of HRT and bisphosphonates, and the associated risk of developing hypercalcemia and hypercalciuria or urinary stones presents a major problem in its clinical use, possibly explaining why it is not approved for the treatment of osteoporosis in the United States. Our observation that individuals with the CC genotype respond to active vitamin D treatment with an annual gain in BMD of 1.6%, which was similar to the increase observed in the HRT group, suggests that the modest effect of active vitamin D on bone mass in previous clinical studies is attributable to the inclusion of nonresponders with the TT or TC genotypes, who constitute ∼80% of the population. Thus, our observation suggests that TGF-β1 genotype is a genetic factor that affects the response to active vitamin D treatment.
Table Table 2.. Multivariable Regression Analysis of Annual Changes in L2–L4 BMD Comparing Control and Treatment Groups for Individuals of Each TGF-β1 Genotype
Data are p values for comparisons with controls of the same genotype, calculated with adjustment for age, height, body weight, time since menopause, L2–L4 BMD, and follow-up period.
Active vitamin D
Polymorphisms in the VDR gene (Bsm I, Apa I, and Taq I) have been associated with the outcome of active vitamin D therapy in Japanese women, suggesting that the effect of the TGF-β1 polymorphism might be related to that of VDR genotype.(15) When we examined the relation between VDR polymorphisms and changes in L2–L4 BMD in 78 individuals randomly selected from the active vitamin D treatment group, the annual gain in BMD in response to active vitamin D therapy increased according to the rank order of genotypes AABBtt < AaBbTt < AabbTT < aabbTT (uppercase letters denote the absence and lowercase letters denote the presence of the site for the restriction enzymes Apa I [A/a], Bsm I [B/b], and Taq I [T/t]; data not shown), similar to the results of the previous study.(15) However, we did not detect a significant relation between the genotype distribution for TGF-β1 and that for VDR (Bsm I, Apa I, and Taq I; data not shown), indicating that the effect of TGF-β1 genotype is likely independent of that of the VDR genotype.
The molecular mechanism by which TGF-β1 genotype affects the response to active vitamin D therapy remains to be elucidated. Evidence suggests the existence of substantial cross-talk between TGF-β signaling and signal transduction through the nuclear VDR in the skeletal system during both development and bone remodeling. 1α,25-Dihydroxyvitamin D3 stimulates the production and release of TGF-β from osteoblasts, and vitamin D deficiency results in reduced TGF-β concentrations in cortical bone.(22,23) In addition, the amount of TGF-β stored in bone has been shown to decrease with age.(40) These observations suggest that the age-related decrease in the serum concentration of 1α,25-dihydroxyvitamin D3 may result in a similar reduction in the skeletal concentration of TGF-β and a consequent increase in susceptibility to involutional osteoporosis. Yanagisawa et al. recently showed that Smad3, a downstream signaling molecule for TGF-β, directly interacts with the VDR and functions as a coactivator of transcription of VDR target genes.(25) Thus, TGF-β and Smad proteins may cooperate with active vitamin D and the VDR in bone remodeling, and the TGF-β1 polymorphism, by affecting the intracellular activity of Smad proteins, may have a substantial impact on the biological function of the VDR system.
HRT, the treatment of choice for postmenopausal osteoporosis, resulted in an overall annual increase in L2–L4 BMD of 1.2% in the present study subjects.(41) In multivariable regression analysis, with adjustment for age, height, body weight, time since menopause, L2–L4 BMD, and follow-up period, the annual gain in BMD was significantly greater for individuals in the HRT group than for controls with any genotype. Also, the annual gain in BMD in response to HRT increased according to the rank order of TGF-β1 genotypes TT < TC < CC; however, the differences among genotypes were not statistically significant, possibly because of the relatively small population size. A differential response to HRT according to TGF-β1 genotype is supported by the observations that estrogen stimulates the production of TGF-β1 by osteoblastic cells and that TGF-β1 acts as a local mediator of the skeletal effects of estrogen.(18–21)
In conclusion, the present study suggests the importance of the T29→C polymorphism of the TGF-β1 gene as a genetic factor affecting the rate of bone loss as well as of treatment outcome, especially with regard to active vitamin D therapy, in postmenopausal osteoporosis.
This work was supported in part by a Research Grant for Longevity Sciences (Y.Y.) and by a Health Sciences Research Grant for Research on Human Genome and Gene Therapy (K.I.) from the Ministry of Health and Welfare of Japan.