S. Vedi¹, D. W. Purdie², P. Ballard². S. Bord¹, A. C. Cooper² and J. E. Compston¹

¹Department of Medicine, University of Cambridge School of Clinical Medicine, Cambridge; and ²Centre for Metabolic Bone Disease, University of Hull, Hull, UK.

Dr J. E. Compston, Department of Medicine, Box 157, Addenbrooke’s Hospital, Cam¬bridge CB2 2QQ, UK. Tel: +44 (0) 1223 336867. Fax: +44 (0) 1223 336846. ; E-mail: jec1001@cam.ac.uk.

Abstract

Conventional hormone replacement therapy preserves bone mass predominantly by reducing bone turnover but does not exert significant anabolic skeletal effects. In contrast, high doses of estrogen have been shown to increase bone formation in animals and we have recently reported high bone mineral density values in women treated long-term with estradiol implant therapy. The aim of this study was to investigate the mechanisms by which high doses of estrogen may increase bone mass in postmenopausal women. Iliac crest biopsies were obtained from 12 women who had received long-term treatment with estradiol implants (at least 14 years), on demand, following hysterectomy and bilateral salpingo-oophorectomy. Indices of bone turn¬over, remodeling balance and cancellous bone structure were assessed by image analysis and compared with those of premenopausal women. Mean wall width was significantly higher in women treated with estradiol therapy than in premenopausal women (44.8 ± 4.8 vs 38.8 ± 2.8 pm; mean ± SD; p = 0.001) and eroded cavity area was significantly lower in the implant-treated women (3612 ± 956 vs 5418 ± 1404 μm²; p = 0.001). Bone formation rate at tissue level and activation frequency were lower in the women treated with implants, although the differences were not statistically significant. Indices of cancellous bone structure were generally similar between the two groups. These results provide the first direct evidence that high-dose estrogen therapy produces anabolic skeletal effects in postmeno¬pausal women and indicate that these are achieved by stimulation of osteoblastic activity.

Keywords: Anabolic; Bone remodeling; Estrogen; Postmenopausal women.

---

Introduction

The beneficial skeletal effects of menopausal estrogen replacement therapy are well documented [1-6]. Admin-istration of estrogen at or after the menopause prevents bone loss and reduces fracture rate in the spine, hip and radius - effects which are believed to be mediated predominantly by inhibition of osteoclastic bone resorption. The small increase in bone mineral density during the first 1-2 years of treatment that has been observed in many studies can be attributed to the infilling of the remodeling space that occurs as resorption is inhibited and existing cavities become filled in with new bone formed by osteoblasts.

Studies in animals have shown that estrogens not only inhibit osteoclastic activity but may also have stimula¬tory effects on osteoblasts, resulting in increased bone formation [7,8]. Direct evidence for a similar effect in humans is lacking although indirect evidence is provided by the observation that percutaneous estrogen implant therapy, which is associated with high serum estradiol levels, results in greater increases in bone mineral density than oral or transdermal estrogen administration, with which more physiologic concentrations of serum estradiol are achieved [9-12]. However, these data are derived mainly from cross-sectional studies in which testosterone implants were also given. More recently, high bone mineral density values were reported in a group of women receiving long-term treatment with percutaneous estradiol implants without co-administra¬tion of testosterone [13].

To explore further the skeletal changes induced by high doses of estrogen we have examined bone turnover, resorption cavity characteristics and cancellous structure in women receiving long-term percutaneous estradiol implant therapy following hysterectomy and bilateral salpingo-oophorectomy. Comparison of histomorpho- metric indices was made with a group of premenopausal women, based on the rationale that significant age- related bone loss had not occurred in the patient group prior to estrogen therapy and that any differences between the two groups would therefore reflect the effects of high-dose as opposed to physiologic estrogen replacement.

Patients and Methods

Patients and Controls

Twelve women, aged 52-67 years (mean 58 years), who had undergone total abdominal hysterectomy and bilateral salpingo-oophorectomy for a non-malignant indication and had received long-term estradiol implant therapy, agreed to take part in the study. These women were attending the estradiol implant clinic at the Department of Obstetrics and Gynaecology, Princess Royal, Hull, UK; all 12 consented to undergo iliac crest bone biopsy. Estradiol implants, 100 mg, had been inserted approximately 6-monthly, on demand, for at least 14 years, although for the last 2-3 years before this study the dose had been reduced to an average of 50 mg every 6 months. None of the women had received testosterone, bisphosphonates, fluoride, vitamin D or glucocorticoids, nor did any have a past or present history of illness associated with disturbances of bone or mineral metabolism.

Control values were obtained from 12 premenopausal women, aged 23-40 years (mean 31.3 years), with endometriosis, prior to treatment with gonadotropin-releasing hormone analogs [14]. None of these women had a past or present history of any illness associated with bone disease and none had taken drugs known to affect bone or mineral metabolism.

Bone Histomorphometry

Trans-iliac bone biopsies were obtained using an 8-mm internal diameter trephine under local anesthetic and mild sedation after the administration of two time¬spaced oral doses of a tetracycline (demeclocycline). Informed written consent was obtained from all women and the study was approved by the Hull and East Riding research ethics committee. Biopsies were embedded in LR White medium resin (London Resin Co.). Eight micrometer undecalcified sections were stained by the von Kossa technique or with 1% toluidine blue.

Histomorphometric assessment was made using a Digicad digitizing tablet and cursor with an LED point light source (Kontron) and an Olympus BHS-BH2 binocular transmitted light microscope with a BH2-DA drawing attachment (Olympus Optical UK, London). All histomorphometric data are described according to ASBMR nomenclature [15]. All measurements were made in masked fashion by the same observer (S.V.) with the exception of tetracycline-based indices in the premenopausal women; because of the lack of sufficient remaining bone tissue, fresh histologic sections could not be obtained from these biopsies and it was therefore necessary to use the values previously obtained by another observer.

Primary Measurements

Bone area/tissue area (B.Ar/ T.Ar), osteoid perimeter/bone perimeter (O.Pm/B.Pm) and osteoid seam width (O.Wi) were measured on von- Kossa-stained sections on a minimum of 25 fields from three to six sections. Osteoid width was measured at four approximately equidistant points, or eight points on seams longer than 600 μm in length. A minimum of 20 seams per biopsy was measured on the same sections used for O.Pm. All seams with a width of 3 μm or more were measured.

The mean width of completed bone remodeling units (W.Wi) was measured on toluidine-blue-stained sections viewed under polarized light at x 156 magnification. A minimum of 25 BMUs was measured on each biopsy from between three and eight sections. Tetracycline labeling was viewed by fluorescence microscopy on a minimum of six 15 μm unstained sections at x 156 magnification.

Mineralizing perimeter (Md.Pm) was calculated as follows:

• Md.Pm/B.Pm (%) = dL.Pm + (0.5 x sL.Pm)/B.Pm

where dL.Pm is the double-labeled perimeter and sL.Pm is the single-labeled perimeter.

The mean distance between double labels was measured directly at x312 magnification using the digitizing tablet and cursor. Measurements were made at approximately four equidistant points along the double labels. A minimum of 20 labels was measured for each biopsy on a minimum of six sections.

Mineral apposition rate was calculated as :

• MAR (μm/day) = L.Wi/LP

where L.Wi is the inter-label width and LP is the labeling period (12 days).

Derived Indices

Adjusted apposition rate (Aj.AR), mineralization lag time (Mlt) and osteoid maturation period (Omt) were calculated as follows:

• Aj.AR (μm/day) = MAR x (M.Pm/O.Pm%)
• Mlt (days) = O.Wi/Aj.AR
• Omt (days) = O.Wi/MAR

The tissue-based bone formation rate (BFR/B.Pm) was calculated as follows:

• BFR/B.Pm (μm²/qm/day) = MAR x (M.Pm/B.Pm%)

Activation frequency (Acf) was calculated as:

• Acf (yr^-1) = (BFR/B.Pm)/W.Wi

Measurement of Resorption Cavity Characteristics

The method described by Garrahan et al. [16] was adapted for use with the digitizing tablet and light cursor for measurement. Cavities were identified on toluidine blue stained sections viewed under polarized light at x 156 magnification and measured at x 375 or x 750 magnification depending on the size of the cavity. Criteria for identification of resorption cavities included interruption of lamellae at an angle to the bone perimeter, absence of osteoid tissue and depth of greater than 3 μm. A minimum of 20 cavities was assessed for each biopsy. The following indices were obtained:

• Mean eroded depth (E.De, μm)
• Maximum eroded depth (E.De.Max, μm)
• Eroded cavity area (E.Ar, μm² )
• Mean reconstructed cavity length (μm)
• Cement line length (μm)

Strut Analysis

This method has been described in detail previously [17]. The bone section is viewed with a CCD camera mounted on a light box, allowing the whole section to appear within a single field of view (magnification x 9). Images of the whole bone section are captured on an 386DX IBM-compatible AT-based computer system containing a Virtuoso frame store (Primagraphic, UK). All analysis software was written in the ‘C’ language using a library of image-processing subroutines (Foster Findlay Associates, Newcastle, UK). The stored images are converted to binary images that can be interactively edited to remove minor specimen preparation artifacts. The right and left corticomedullary junctions are defined automatically using a procedure that has been described previously [18]. The upper and lower boundaries of the section are defined interactively by the operator. The upper and lower boundaries and right and left corticomedullary delineations define the ‘active’ regions upon which all measurements are performed.

Binary images are skeletonized to give a symmetric axis of the original bone profile. The computer automatically identifies trabecular junctions, or nodes (Nd) and the ends of trabeculae, or termini (Tm). Individual trabeculae, or struts, are defined topologi¬cally: the indices thus generated include the ratio of nodes to termini (Nd/Tm) and the terminus-to-terminus strut length (Tm.Tm), node-to-loop strut length (Nd.Lp) and node-to-terminus strut length (Nd.Tm), each of which is expressed as a percentage of the total strut length.

Statistical analysis was performed using an unpaired Student’s t-test after log transformation of the data. Results are expressed as the mean ± SD.

Results

Details of the patient group are shown in Table 1. The time since hysterectomy and bilateral salpingo-oophor- ectomy was 17-23 years (mean 20 years) and the age at surgical menopause ranged between 33 and 48 years (mean 37.6 years). The mean spinal bone mineral density was 1.42 ± 0.24 g/cm² and the mean T-score was +1.73 ± 2.0; corresponding figures for femoral neck bone mineral density were 1.13 ± 0.17 g/cm² and +1.24 ±1.42.

Table 1. Details of the patient group.

Histomorphometric indices are shown in Tables 2-4. There were no significant differences in cancellous bone area, osteoid perimeter, mineralizing perimeter or osteoid seam width between the women receiving estradiol implants and the premenopausal controls (Table 2). The mineral apposition rate was significantly lower in the implant-treated women than in the controls (0.63 ± 0.10 vs 0.89 ± 0.39 μm/day; p = 0.0004) and there was a trend towards lower bone formation rate and activation frequency in these women, although the differences were not statistically significant (0.030 ± 0.023 vs 0.048 ± 0.040 μm²/pm/day and 0.25 ± 0.21 vs 0.47 ± 0.42 /yr^-1; Table 2).

Table 2. Comparison of indices of bone formation in postmenopausal women on long-term high-dose estrogen implant therapy versus control group.

Indices related to remodeling balance are shown in Table 3. In women receiving estradiol implants the mean value for wall width was significantly higher than in the control group of premenopausal women (44.8 ± 4.8 vs 38.8 ± 3.8 μm; p = 0.001). There were no significant differences in the mean or maximum eroded depth between the two groups. However, the eroded cavity area was significantly lower in the implant-treated women (3612 ± 956 vs 5418 ± 1404 μm²; p = 0.001), as were the reconstructed surface length and cement line length (175 ± 25.3 vs 259 ± 26.6 μm; p < 0.0001) and 234 ± 29.0 vs 284 ± 30.9 μm; p = 0.0004).

Table 3. Comparison of indices of remodeling balance in postmenopausal women on long-term high-dose estrogen implant therapy versus control group.

The results of strut analysis are shown in Table 4. No significant differences were observed between the two groups in the node-to-terminus ratio or in the terminus- to-terminus, node-to-loop or node-to-terminus strut lengths. The node-to-node strut length was significantly lower in the women receiving implants than in the premenopausal controls (22.5 ± 6.8 vs 29.4 ± 6.0%; p = 0.018).

Table 4. Indices of cancellous bone connectivity in postmenopausal women on long-term high-dose estrogen implant therapy versus control group.

Discussion

These results provide histologic evidence that estrogen implant therapy produces anabolic skeletal effects in postmenopausal women and indicate that these are achieved by stimulation of osteoblastic activity, resulting in increased bone formation at the cellular level. The mean bone mineral density at the spine and femur in women treated with percutaneous estrogen was higher than that seen in normal premenopausal women, suggesting not only that peak bone mass had been preserved but also that estrogen replacement had further increased bone mass above this level. This contention is supported by the significantly greater thickness of completed cancellous bone structural units in the women receiving high-dose implant therapy when compared with premenopausal women. Iliac crest cancellous bone area was not significantly higher in the implant-treated women, probably reflecting the large measurement variance associated with its assessment and the relatively small number of subjects studied.

For ethical reasons it is difficult to obtain bone biopsies from normal premenopausal women and, in the present study, comparison of histomorphometric indices was made with a group of premenopausal women with endometriosis who had taken part in a previous study of the effects of gonadotropin-releasing hormone agonists [14]. Although it has been suggested, on the basis of measurements of volumetric bone density in the distal radius, that bone mass may be reduced in women with endometriosis [19], in a subsequent study of 85 women with endometriosis Lane et al. [20] reported normal spinal bone mineral density. In our group of premeno¬pausal women, serum estradiol concentrations were normal (mean 374 pmol/l; range 232-516) and spinal bone mineral density was also normal, providing justification for their use as controls in this study.

Several other lines of evidence support the contention that estrogens, in high doses, may stimulate bone formation in the human skeleton. The use of oral contraceptives in premenopausal women has been reported in some studies to be associated with increased bone mineral density, although this finding has not been universal [21-24]. Low bone mass has been reported in a young man with estrogen resistance due to a mutation in the estrogen receptor gene [25] and in a young male with aromatase deficiency [26]; furthermore, in a young man with the latter condition, estrogen replacement was associated with a 19.8% increase in spinal bone mass after 30 months of treatment [27]. Finally, estrogen deficiency in adolescents is associated with failure to attain peak bone mass; in a study of adolescent and young adult females, an integrated estimate of endogen¬ous and exogenous estrogen exposure was significantly correlated with bone mineral density of the spine and wrist [28].

The measurement of wall width in cancellous bone provides an index of osteoblast cellular activity [29]; because of the long life span of the formative phase of the bone remodeling cycle and relatively low bone turnover in adult human bone, assessment of wall width after relatively short treatment periods does not accurately reflect the effects of intervention since only a small proportion of completed remodeling units will have been formed during the treatment period. Prospec¬tive studies of women treated with conventional hormone replacement therapy have shown no change [30] or a small decrease [31] in wall width after 1 or 2 years treatment respectively. In women treated with percutaneous estrogen implants (75 mg 6-monthly), Holland et al. [32] also reported no significant change in wall width after 1 year. In the present study the long treatment period in the women receiving estrogen implants insured that the majority of structural units identified were formed subsequent to the commencement of treatment.

The changes observed in indices of resorption cavity size in implant-treated women are consistent with those reported earlier in a prospective study of women treated with conventional hormone replacement therapy, in which a consistent trend toward a reduction in resorption cavity size was seen [31]. In the present study, the cavity area was significantly smaller in implant-treated women than in premenopausal controls - a difference that resulted from a smaller erosion cavity length rather than depth. The effects of estrogen on osteoclast activity are incompletely understood but inhibition of osteoclasto- genesis [33] and stimulation of osteoclast apoptosis have been reported [34]. Although the relative contributions of osteoclast number and activity to erosion cavity depth and length have not been defined, it is conceivable that a reduction in the former may preferentially affect the surface extent of individual resorption cavities rather than their depth, which is likely to be determined predominantly by the activity of individual osteoclasts. This proposition is supported by the study of Delaisse et al. [35] on the effects of cysteine proteinase inhibitors on the resorptive activity of chick osteoclasts.

Accurate assessment of remodeling balance is difficult, largely because of the problems associated with measurement of completed erosion depth. The method used for the latter in the present study assesses the size of all cavities present and thus underestimates completed erosion depth by an unknown amount. Nevertheless, values for mean and maximum erosion depth were similar to those obtained in premenopausal women and this, in combination with the substantially greater values for mean wall width, provides strong evidence in favor of a positive remodeling balance.

Strut analysis of cancellous bone generally revealed similar values in the implant-treated women and premenopausal controls, indicating that connectivity had been preserved in the former group. In women with postmenopausal osteoporosis, conventional hormone replacement therapy has also been shown to preserve existing cancellous bone architecture [36]; similar effects would be predicted in the implant-treated women in the present study since the anabolic effect of estrogen was mediated via increased mean wall width rather than by de novo bone formation.

Because of the lack of availability of fresh histologic sections in the control group of premenopausal women, it was not possible for tetracycline-based measurements to be performed by the same observer. It is therefore possible that the significant differences in mineral apposition rate and its derived measurements, bone formation rate at tissue level and activation frequency, from the implant-treated women might have arisen from inter-observer variation and these data thus have to interpreted with some caution. This would also explain the apparent contradiction in the observations that mineral apposition rate was lower but wall width higher in the implant-treated women when compared with values in premenopausal women; alternatively, it is possible that mineral apposition rate had previously been higher but decreased when the dose of estradiol was reduced. Nevertheless, the values for bone formation rate and activation frequency in the implant-treated women were very similar to those measured, by the same observer, in women with postmenopausal osteoporosis after 2 years of treatment with conventional hormone replacement therapy [31], suggesting that bone turnover was relatively suppressed in the women treated with estradiol implants.

The mechanisms by which estrogens produce anabolic skeletal effects remain to be identified. Estrogen receptors have been identified on osteoblasts and osteocytes [37-39], raising the possibility that increased bone formation may result from direct stimulation of these cells. However, the estrogen receptor alpha does not appear to be required for anabolic skeletal effects of estrogen in mice [40], suggesting involvement of the estrogen receptor beta [41-43] or, alternatively, mechan¬isms not mediated by the estrogen receptor. In vitro, estrogen has been reported to stimulate the production of transforming growth factor beta, insulin-like growth factors and type 1 collagen [44,45], all of which are associated with increased bone formation.

Nearly all drugs currently used in the prevention and treatment of osteoporosis act predominantly by inhibi¬tion of bone turnover and resorption, and thus identification of agents that produce substantial gains in bone mass is of considerable interest. Although high doses of estrogens may be associated with increased short-term and long-term side-effects, there might be a role for a finite period of such treatment in perimeno- pausal women with severely reduced bone mass. The anabolic skeletal effects of estrogen may also be relevant to selective estrogen receptor modulators, which it may be possible to administer in higher doses because of their lack of unwanted estrogenic effects outside the skeleton [46]. Finally, greater understanding of the mechanisms by which estrogen stimulates bone formation may ultimately lead to the development of novel therapeutic interventions.

Acknowledgements

We are grateful to the Wellcome Trust and to Organon Laboratories, UK for financial support. Kumiko Yamaguchi made the tetracycline-based measurements in the premenopausal women.

---

References

Hutchinson A, Polansky SM, Feinstein AR. Postmenopausal estrogens protect against fractures of the hip and distal radius. Lancet 1979;II:705-9. 

Lindsay R, Hart DM, Forrest C, Baird C. Prevention of spinal osteoporosis in oophorectomised women. Lancet 1980;II:1151-3. 

Christiansen C, Christensen MS, McNair PL, Hagen C, Stocklund KE, Transb0l I. Prevention of early menopausal bone loss: conducted 2-year study in 315 normal females. Eur J Clin Invest 1980;10:273-9. 

Weiss NS, Ure CL, Ballard JH, Williams AR, Daling JR. Decreased risk of fractures of the hip and lower forearm with postmenopausal use of estrogens. N Engl J Med 1980;303:1195- 8. 

Ettinger B, Genant HK, Cann CE. Long-term estrogen replace¬ment therapy prevents bone loss and fractures. Ann Intern Med 1985;102:319-24. 

Lufkin EG, Wahner HW, O’Fallon WM, et al. Treatment of postmenopausal osteoporosis with transdermal estrogen. Ann Intern Med 1992;117:1-9. 

Edwards MW, Bain SD, Bailey MC, Lantry MM, Howard GA. 17ß-Estradiol stimulation of endosteal bone formation in the ovariectomised mouse: an animal model for the evaluation of bone-targeted estrogens. Bone 1992;13:29-34. 

Chow J, Tobias JH, Colston KW, Chambers TJ . Estrogen maintains trabecular bone volume in rats not only by suppression of resorption but also by stimulation of bone formation. J Clin Invest 1992;89:74-8. 

Savvas M, Studd JWW, Fogelman I, Dooley M, Montgomery J, Murby B. Skeletal effects of oral oestrogen compared with subcutaneous oestrogen and testosterone in postmenopausal women. BMJ 1988;297:331-3. 

Studd JWW, Savvas M, Fogelman I, Garnett T, Watson NR, Cooper D. The relationship between plasma estradiol and the increase in bone density in women following treatment with subcutaneous hormone implants. Am J Obstet Gynecol 1990;163:1474-9. 

Garnett T, Studd J, Watson N, Savvas M. A cross-sectional study of the effects of long-term percutaneous hormone replacement therapy on bone density. Obstet Gynaecol 1991;78:1002-7. 

Ryde SJS, Bowen-Simpkins K, Bowen-Simpkins P, Evans WD, Morgan WD, Compston JE. The effect of oestradiol implants on regional and total bone mass: a three year longitudinal study. Clin Endocrinol 1994;40:33-8. 

Wahab M, Ballard P, Purdie DW, Cooper A, Willson JC. The effect of long-term oestradiol implantation on bone mineral density in postmenopausal women who have undergone hyster¬ectomy and bilateral oophorectomy. Br J Obstet Gynaecol 1997;104:728-31. 

Compston JE, Yamaguchi K, Croucher PI, Garrahan NJ, Lindsay PC, Shaw RW. The effects of gonadotrophin-releasing hormone agonists on iliac crest cancellous bone structure in women with endometriosis. Bone 1995;16:261-7. 

Parfitt AM, Drezner MK, Glorieux FH, et al. Bone histomorpho- metry: standardisation of nomenclature, symbols and units. J Bone Miner Res 1987;2:595-610. 

Garrahan NJ, Croucher PI, Compston JE. A computerised technique for the quantitative assessment of resorption cavities in trabecular bone. Bone 1990;11:241-6. 

Garrahan NJ, Mellish RWE, Compston JE. A new method for the two-dimensional analysis of bone structure in human iliac crest biopsies. J Microsc 1986;142:341-9. 

Croucher PI, Garrahan NJ, Compston JE. Assessment of cancellous bone structure: comparison of strut analysis, trabecular bone pattern factor, and marrow space star volume. J Bone Miner Res 1996;11:955-61. 

Comite F, Delman M, Hutchinson-Williams K, DeCherney AH, Jensen P. Reduced bone mass in reproductive-aged women with endometriosis. J Clin Endocrinol Metab 1989;69:837-42. 

Lane N, Baptista J, Snow-Harter C. Bone mineral density of the lumbar spine in endometriosis subjects compared to an age- similar control population. J Clin Endocrinol Metab 1991; 72:510-4. 

Lindsay R, Tohme J, Kanders B. The effect of oral contraceptive use on vertebral bone mass in pre- and post-menopausal women. Contraception 1986;34:333-40. 

Stevenson JC, Lees B, Devenport M, Cust MP, Ganger KE. Determinants of bone density in normal women: risk factors for future osteoporosis? BMJ 1989;298:924-8. 

Mazess RB, Barden HS. Bone density in premenopausal women: effects of age, dietary intake, physical activity, smoking and birth-control pills. Am J Clin Nutr 1991;53:132-42. 

Murphy S, Khaw K-T, Compston JE. Lack of relationship between hip and spine bone mineral density and oral contra¬ceptive use. Eur J Clin Invest 1993;23:108-11. 

Smith EC, Boyd J, Frank GR, et al. Estrogen resistance caused by a mutation in the estrogen-receptor gene in a man. N Engl J Med 1994;331:1056-61. 

Morishima A, Grumbach MM, Simpson ER, Fisher C, Qin K. Aromatase deficiency in male and female siblings caused by a novel mutation and the physiological role of estrogens. J Clin Endocrinol Metab 1995;80:3689-98. 

Morishima A, Grumach MM, Bilezikian JP. Estrogen markedly increases bone mass in an estrogen deficient young man with aromatase deficiency. J Bone Miner Res 1997;12(Suppl 1):S126. 

Dhuper S, Warren MP, Brooks-Gunn J, Fox R. Effects of hormonal status on bone density in adolescent girls. J Clin Endocrinol Metab 1990;71:1083-8. 

Lips P, Courpron P, Meunier PJ. Mean wall thickness of trabecular bone packets in the human iliac crest: changes with age. Calcif Tissue Int 1987;26:13-7. 

Steiniche T, Hasling C, Charles P, Eriksen EF, Mosekilde L, Melsen F. A randomised study of the effects of estrogen/gestagen or high dose oral calcium on trabecular bone remodelling in postmenopausal osteoporosis. Bone 1989;10:313-20. 

Vedi S, Compston JE. The effects of long-term hormone replacement therapy on bone remodelling in postmenopausal women. Bone 1996;19:535-9. 

Holland EFN, Chow JWM, Studd JWW, Leather AT, Chambers TJ. Histomorphometric changes in the skeleton of postmenopau¬sal women with low bone mineral density treated with percutaneous implants. Obstet Gynecol 1994;83:387-91. 

Jilka RL, Hangoc G, Girasole G, et al. Increased osteoclast development after estrogen loss. Science 1992;257:88-91. 

Hughes DE, Dai A, Tiffee JC, Li HH, Mundy GR, Boyce BF. Estrogen promotes apoptosis of murine osteoclasts mediated by TGF-B. Nature Med 1996;2:1132-6. 

Delaisse JM, Boyde A, Maconnachie E, et al. The effects of inhibitors of cysteine proteinases and collagenase on the resorptive activity of isolated osteoclasts. Bone 1987;8:305-13. 

Vedi S, Croucher PI, Garrahan NJ, Compston JE. Effects of hormone replacement therapy on cancellous bone microstructure in postmenopausal women. Bone 1996;19:69-72. 

Komm BS, Terpening CM, Benz DJ, et al. Estrogenic binding, receptor mRNA and biologic response in osteoblast-like osteosarcoma cells. Science 1988;241:81-3. 

Eriksen EF, Colvard DS, Berg NJ, et al. Evidence of estrogen receptors in normal human osteoblast-like cells. Science 1988;241:84-6. 

Braidman IP, Davenport LK, Carter DH, Selby PL, Mawer EB, Freemont AJ. Preliminary in situ identification of estrogen target cells in bone. J Bone Miner Res 1995;10:74-80. 

Pan LC, Ke HZ, Simmons HA, et al. Estrogen receptor alpha knockout mice lose trabecular and cortical bone following ovariectomy. J Bone Miner Res 1997;12(Suppl 1):S134. 

Onoe Y, Miyaura C, Ohta H, Nozawa S, Suda T. Expression of estrogen receptor beta in rat bone. Endocrinology 1997;38:4509- 12. 

Kuiper GGJM, Gustafsson J-A. The novel estrogen receptor-B subtype: potential role in the cell- and promotor-specific actions of estrogens and anti-estrogens. FEBS Lett 1997;410:87-90. 

Kuiper GGJM, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson J-A. Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc Natl Acad Sci USA 1996;93:5925-30. 

Ernst M, Heath JK, Rodan GA. Estradiol effects on proliferation, messenger ribonucleic acid for collagen and insulin-like growth factor-1, and parathyroid hormone-stimulating adenylate cyclase activity in osteoblastic cells from calvariae and long bones. Endocrinology 1998;125:825-33. 

Yang NN, Bryant HU, Hardikar S, et al. Estrogen and raloxifene stimulate transforming growth factor-B3 gene expression in rat bone: a potential mechanism for estrogen- or raloxifene-mediated bone maintenance. Endocrinology 1996;137:2075-84. 

Compston JE. Designer oestrogens: fact or fantasy? Lancet 1997;350:676-7.