I. INTRODUCTION

Top Previous Next References

Osteoporosis is defined as a condition characterized by reduced bone mass and disruption of bone architecture, resulting in increased bone fragility and increased fracture risk (294). These fractures, which particularly affect the hip, spine, and wrist, are a major cause of morbidity and mortality in elderly populations (65, 247, 248). Clinically, osteoporosis may be recognized by the presence of fragility fractures, but recently, diagnostic criteria based on bone mineral density measurements have been proposed (397), based on the well-documented inverse relationship between bone mineral density and fracture risk (70, 115, 160, 235, 390). According to this classification, osteoporosis is defined as a bone mineral density in the spine and/or proximal femur 2.5 or more standard deviations below normal peak bone mass. The term established osteoporosis is used when one or more fragility fractures have occurred.

The recognition, by Fuller Albright in 1948, of the central role of estrogen deficiency in the pathogenesis of postmenopausal osteoporosis (7) provided a major stimulus to research into this hitherto neglected condition and into the mechanisms by which estrogens affect bone. The advances that followed have been paralleled by a rapid growth in understanding of bone physiology and biochemistry; together, these have been responsible for significant improvements in the clinical management of patients with osteoporosis over the past two decades. In particular, Albright's fundamental observation provided the rationale for the use of estrogen replacement therapy in the prevention of postmenopausal osteoporosis and altered the widely held perception that osteoporosis was an inevitable and untreatable consequence of ageing.

Sex steroids play an essential role in the maintenance of bone health throughout life, and adverse effects of hormone deficiency can be seen in the young and old and in men and women. The mechanisms by which these effects are mediated remain incompletely understood and are the subject of enormous research effort. The potential therapeutic implications of progress in this field are, however, considerable and extend beyond osteoporosis. In this review relevant aspects of bone physiology and biochemistry are discussed, and current knowledge of the skeletal effects of sex steroids is reviewed.

	II. BONE COMPOSITION, STRUCTURE, AND FUNCTION

Top Previous Next References

The skeleton provides structural support for the body, protecting internal organs and housing the bone marrow. It also functions as a reservoir of calcium and phosphate ions and plays a major role in the homeostasis of these minerals. Bone consists of an extracellular matrix, the organic phase of which is composed of type I collagen, proteoglycans, and noncollagenous proteins including osteocalcin, bone sialoprotein, osteonectin, thrombospondin, and osteopontin. Bone matrix also contains growth factors and cytokines that have an important regulatory role in bone remodeling. The inorganic phase of bone matrix is composed mainly of calcium hydroxyapatite.

Approximately 80% of the skeleton is composed of cortical bone, which is found mainly in the shafts of long bones and surfaces of flat bones. It is composed of compact bone, which is laid down concentrically around central canals or Haversian systems, which contain blood vessels, lymphatic tissue, nerves, and connective tissue. Cancellous or trabecular bone is found mainly at the ends of long bones and in the inner parts of flat bones and consists of interconnecting plates and bars within which lies hematopoietic or fatty marrow. The surface-to-volume ratio of cancellous bone is much greater than that of cortical bone, and the potential for metabolic activity is correspondingly higher.

Osteoblasts are responsible for the formation and mineralization of bone. They are derived from pluripotent mesenchymal stem cells, which can also differentiate into chondrocytes, adipocytes, myoblasts, and fibroblasts (279, 280) (Fig. 1). The mechanisms by which commitment to the osteoblast phenotype is achieved are not fully established, but the core binding transcription factor Cbfa1 (also known as osteoblast stimulating factor 2 or Osf2) has recently been shown to be essential for osteoblast differentiation; thus loss of function mutant mice exhibit complete lack of ossification of cartilage (197, 273), and heterozygous loss of function causes cleidocranial dysplasia (255), a condition associated with patent fontanelles, abnormal dentition, short stature, and hypoplastic clavicles. In addition, a number of other factors are required for normal osteoblast differentiation including fibroblastic growth factors (FGFs), transforming growth factor- (TGF-), bone morphogenetic factors (BMPs), glucocorticoids, and 1,25-dihydroxyvitamin D [1,25(OH)₂D] (216).

View larger version (9K): [in this window] [in a new window]   Fig. 1. Possible differentiation pathways of the pluripotent mesenchymal stem cell.

In situ, osteoblasts actively involved in bone formation appear as monolayers of plump cuboidal cells in close juxtaposition to newly formed unmineralized bone (osteoid). Structural characteristics include a round nucleus at the base of the cell, a strongly basophilic cytoplasm, and a prominent Golgi complex (44). Cytoplasmic processes extend from the secretory side of the cell into the bone matrix and communicate with the osteocyte canalicular network. There are also gap junctions, composed of proteins called connexins, that connect the cytoplasm of adjacent cells (343, 410). Developing and mature osteoblasts express a number of products including type I collagen, alkaline phosphatase, osteopontin, and osteocalcin that may be used to identify the osteoblastic phenotype in vivo and in vitro.

Actively forming osteoblasts may subsequently undergo apoptosis or become bone-lining cells or osteocytes; both the latter are believed to represent further stages of maturation. Bone-lining cells are flat elongated cells with a spindle-shaped nucleus that lie along the endosteal membrane covering quiescent bone surfaces. Lining cells, together with the endosteal membrane, form a protective layer over the bone surface; their function is not well understood, but they may play a role in the activation of bone remodeling (32).

Osteocytes are small flattened cells within the bone matrix and are connected to one another and to osteoblastic cells on the bone surface by an extensive canalicular network that contains the bone extracellular fluid (1). The cytoplasmic projections within the canaliculi communicate via gap functions and enable osteocytes to respond to mechanical and biochemical stimuli (83, 308). Osteocytes are terminally differentiated and may ultimately undergo apoptosis or be phagocytosed during the process of osteoclastic resorption.

Osteocytes are believed to play a central role in the response to mechanical stimuli, sensing mechanical strains and initiating an appropriate modeling or remodeling response via a number of chemical messengers including glucose-6-phosphate dehydrogenase, nitric oxide, and insulin-like growth factors.

Osteoclasts are large, multinucleated bone-resorbing cells derived from hematopoietic precursors of the monocyte/macrophage lineage. They are formed by the fusion of mononuclear cells and are characterized by the presence of a ruffled border, which consists of a complex infolding of plasma membrane, and a prominent cytoskeleton. They are rich in lysosomal enzymes, including tartrate-resistant acid phosphatase (TRAP). During the process of bone resorption, hydrogen ions generated by carbonic anhydrase II are delivered across the plasma membrane by a proton pump to dissolve bone mineral. Subsequently, lysosomal enzymes including collagenase and cathepsins are released and degrade bone matrix. Attachment of osteoclasts to the bone surface is an essential prerequisite for resorption and is mediated by integrins, particularly av3, which bind matrix proteins containing the motif Arg-Gly-Asp (153); potential ligands include osteopontin, bone sialoprotein, thrombospondin, osteonectin, and type 1 collagen. Morphologically, attachment of the osteoclast to the bone surface is seen as an actin-containing ring (211) that surrounds completely the ruffled membrane.

It has long been known that osteoblastic or stromal cells are essential for osteoclastogenesis, and the identity of the factor concerned, termed "osteoclast differentiation factor" or ODF, has recently been reported as receptor activator of NFB ligand (RANKL), a new member of the tumor necrosis factor (TNF) ligand family, also termed TRANCE (TNF-related activation-induced cytokine) or osteoprotegerin ligand (OPGL) (413). The signaling receptor for RANKL is RANK, a type 1 transmembrane protein expressed by osteoclasts (9), whereas osteoprotegerin (OPG), a novel member of the TNF receptor superfamily, acts as a soluble decoy receptor that prevents RANKL from binding to and activating RANK on the osteoclast surface (198). The interaction of RANKL with RANK activates a cascade of intracellular events that involve activation of NFB and the protein kinase JNK, and interaction with TNF receptor-associated factors (TRAFs) (147). Macrophage-colony stimulating factor (M-CSF) production by osteoblastic/stromal cells is also essential for osteoclastogenesis (415), although unlike RANKL, it does not appear to have effects on osteoclast activity (364).

Osteoclast apoptosis is an important determinant of osteoclast activity. Like osteocytes, osteoclasts are terminally differentiated cells with a limited life span. The cytokines interleukin-1, TNF-, and M-CSF all reduce osteoclast apoptosis (348), thus prolonging the viability of these cells. In contrast, as discussed in section IVC, estrogen increases apoptosis of osteoclasts (158), an effect which is associated with increased production of TGF- and reduced expression of NFB-activated genes. Loss of function gene mutations associated with osteopetrosis, a group of disorders caused by osteoclast dysfunction, are shown in Table 1.

Like bone modeling, bone remodeling is a surface phenomenon. Remodeling serves to maintain the mechanical integrity of the adult skeleton and also provides a mechanism by which calcium and phosphate ions may be released from or conserved within the skeleton. It consists of the removal, by osteoclasts, of a quantum of bone followed by the formation by osteoblasts within the cavity so created of osteoid, which is subsequently mineralized. In normal adult bone, the processes of resorption and formation are coupled both in space and time; thus bone resorption always precedes formation (coupling), and in the young adult skeleton, the amounts of bone formed and resorbed are quantitatively similar (balance) (Fig. 2). The sites at which bone remodeling occurs are termed basic multicellular units (BMUs) or bone remodeling units. The life span of each remodeling unit in humans is believed to be between 2 and 8 mo, with most of this period being occupied by bone formation (287). In normal human adults, ~20% of the cancellous bone surface is undergoing remodeling at any given time.

View larger version (35K): [in this window] [in a new window]   Fig. 2. Schematic representation of bone remodeling. (From Compston JE. Bone morphology: quality, quantity and strength. In: Advances in Reproductive Endocrinology. Oestrogen Deficiency: Causes and Consequences, edited by Shaw RW. Carnforth, Lancs, UK: Parthenon, 1996, vol. 8, p. 63-84.)

The first stage in bone remodeling involves activation of the quiescent bone surface before resorption. Although the process of activation is not well understood, it is believed to involve retraction of lining cells and digestion of the endosteal membrane, the latter possibly occurring as a result of the production of collagenases by the lining cells (32). Osteoclast precursors are then attracted to the exposed mineralized bone surface and fuse to become functional osteoclasts that resorb bone. Exposure of the mineralized bone surface by this process of activation is thought to be an essential prerequisite for osteoclastic resorption. The presence of capillary sinusoids close to sites of bone remodeling suggests that circulating osteoclasts may pass through the vessel wall before bone resorption rather than being directly recruited from bone marrow (288). There is a close interdependence between angiogenesis and osteogenesis in developing bone (125, 151), a relationship which may also exist in adult bone.

The determinants of the sites at which bone remodeling is initiated have not been fully elucidated. However, it is likely that the location of activation and the subsequent remodeling process is critically dependent on mechanical factors, and sites of trabecular thinning may thus be favored. (60)

At the tissue and cell levels, there are two possible mechanisms of bone loss in osteoporosis (59) (Fig. 3). Quantitatively, the most important is an increase in the activation frequency (also termed high bone turnover) in which the number of remodeling units activated on the bone surface is increased; this results in a greater number of units undergoing bone resorption at any given time and is potentially reversible provided that bone remodeling is coupled and that remodeling balance is maintained. The second mechanism, which often coexists with increased bone turnover, is that of remodeling imbalance, in which the amount of bone formed within individual remodeling units is less than that resorbed due either to an increase in resorption, decrease in formation, or a combination of the two. This form of bone loss is irreversible once the remodeling cycle has been completed, at least in terms of that remodeling unit.

View larger version (40K): [in this window] [in a new window]   Fig. 3. Mechanisms of bone loss in osteoporosis. (From Compston JE. The skeletal effects of oestrogen depletion and replacement: histomorphometrical studies. In: Annual Review of the Management of Menopause, edited by Studd J. Carnforth, Lancs, UK: Parthenon, 2000, p. 287-296.)

These mechanisms of bone loss can be quantitatively assessed using histomorphometric techniques. The administration of two, time-spaced doses of a tetracycline compound before bone biopsy enables identification of actively forming bone surfaces (111) and calculation of bone turnover and activation frequency. The amounts of bone formed and resorbed within individual bone remodeling units can also be measured; the former is known as the wall width (72) and is a measure of osteoblast function. The erosion depth and other indices of resorption cavity size can be assessed after computerized or manual reconstruction of the eroded bone surface (55, 118).

The alterations in bone remodeling responsible for bone loss determine the accompanying changes in bone architecture, an important determinant of the mechanical strength of bone (62). In cancellous bone, either trabecular thinning or trabecular perforation and erosion may occur; these two processes are to some extent interdependent. Trabecular thinning is associated with better preservation of bone architecture than penetration and erosion of trabeculae, the latter having the greater adverse effects on bone strength. Increased activation frequency and/or increased resorption depth predispose to trabecular penetration and erosion, whereas low bone turnover states favor trabecular thinning.

A number of approaches to the quantitative assessment of cancellous bone structure have been described. In histological sections of bone, trabecular width and spacing can be measured directly or calculated from area and perimeter measurements (289) and indirect assessment of connectivity made by the technique of strut analysis (119) or measurement of trabecular bone pattern factor (138) or marrow star volume (381). Finally, a number of techniques have been used to generate three-dimensional images of bone; these include reconstruction of serial sections; scanning and stereo microscopy; volumetric, high-resolution, and microcomputed tomography; and magnetic resonance imaging (124, 231). Such approaches enable direct assessment of connectivity and measurement of anisotropy, but their application in vivo is currently restricted by limited resolution, partial volume effects, and noise.

The regulation of bone remodeling involves a complex interplay between systemic hormones, mechanical stimuli, and locally produced cytokines, growth factors, and other mediators (Fig. 4). Much of our knowledge in this area is derived from in vitro experiments and may not always be relevant to the control of bone remodeling in vivo.

View larger version (34K): [in this window] [in a new window]   Fig. 4. Control of bone remodeling. (From Compston JE. Hormone replacement therapy for osteoporosis: clinical and pathophysiological aspects. Reprod Med Rev 3: 209-244, 1994.)

Mechanical stresses are a major determinant of bone modeling and remodeling, and it is generally believed that osteocytes are the major mechanosensory bone cell. Intermittent loading at physiological levels of strain results in rapid metabolic changes in osteocytes, one of the earliest manifestations of which is an increase in the production of glucose-6-phosphate dehydrogenase activity (293). The mechanisms by which osteocytes sense mechanical loading have not been fully established, but it is believed that the deformation resulting from strain stimulates the flow of interstitial fluid through the osteocyte canalicular network (299). Electrokinetic streaming potentials and/or fluid shear stress may then modulate production by the osteocyte of mediators such as prostaglandins and nitric oxide (264). These may then stimulate the production of other cytokines and growth factors, for example, insulin-like growth factor (IGF) (214).

Bone is a rich source of cytokines and growth factors (Fig. 5, Table 2) and also other mediators such as prostaglandins and nitric oxide. In addition, cells in the bone microenvironment play a major role in the regulation of bone remodeling, both as a source of bone cell precursors and by the production of bone active cytokines and growth factors. Table 2 lists the major cytokines and growth factors known to be implicated in bone metabolism. Those known to play an important role in mediating the effects of estrogen on bone are described in greater detail below.

View larger version (18K): [in this window] [in a new window]   Fig. 5. Effects of cytokines on osteoclast production and activity. TGF-, transforming growth factor-; IL, interleukin; TNF-, tumor necrosis factor-; M-CSF, macrophage-colony stimulating factor; GM-CSF, granulocyte/macrophage-colony stimulating factor.

Interleukin (IL)-1 and -1 are potent stimulators of bone resorption in vitro and in vivo (34, 129, 324). These effects are mediated both by an increase in the proliferation and differentiation of osteoclast precursors and also by increased osteoclastic activity (297, 354), the latter resulting at least in part from inhibitory effects on osteoclast apoptosis. Some of the effects of IL-1 on osteoclasts result from an increase in prostaglandin synthesis (34). IL-1 also has effects on osteoblasts, which are probably dependent on whether administration is continuous or intermittent (130, 335). In the former situation, inhibitory effects on bone formation are seen, whereas intermittent administration is associated with an increase in osteoblast proliferation and differentiation. The IL-1 receptor antagonist (IL-1ra) is a constitutively occurring inhibitor of IL-1 (139), inhibiting IL-1-induced stimulation of bone resorption both in vitro (330) and in vivo (136). TNF- and lymphotoxin (TNF-) are also potent stimulators of bone resorption (22, 173) and appear to act in a similar way to IL-1.

IL-6 also stimulates bone resorption, although by different mechanisms. Its production in bone is increased by other bone-resorbing cytokines and systemic hormones (for example, PTH) (101), and it also acts synergistically with these agents, increasing their bone resorptive effects (75). The effects of IL-6 in vivo may be modulated by the circulating levels of IL-6 soluble receptor (350).

Granulocyte/macrophage-colony stimulating factor (GM-CSF) acts on the early development of hematopoietic precursor cells, including osteoclasts (210). Unlike M-CSF, it is not essential for osteoclastogenesis, although it supports the differentiation of osteoclast precursors. GM-CSF has also been reported to increase the proliferation of osteoblastic cells in vitro (74) and in vivo (352), probably by an indirect action.

The TGF- superfamily includes the TGF- isoforms, the activins and inhibins, and BMPs (28). TGF- is present in a latent, biologically inert form in bone matrix, its active form being released in the process of bone resorption (298). It is a potent stimulator of bone formation (267), stimulating osteoblastic differentiation and the synthesis of bone matrix proteins and their receptors, while inhibiting the synthesis of proteases. Most data support inhibitory effects on osteoclastic bone resorption (29, 233) due to effects both on osteoclast formation and activity, the latter effect being mediated by stimulation of osteoclast apoptosis (157). Three main TGF- receptors exist (50): type I and type II, which are transmembrane serine/threonine kinases and function as signaling receptors (109), and type III, betaglycan, which is nonsignaling (389). It is believed that TGF- binds directly to the type II receptor, which is constitutively active, and that this complex is then recognized by the type I receptor to form a complex, with phosphorylation of the type I receptor by the type II receptor (401).

The BMPs are members of the TGF- superfamily. They possess osteoinductive properties, inducing differentiation of osteoblastic and chondroblastic precursor cells, and are similar to but not identical to TGF- in terms of their structure and activity (400). BMPs act as morphogens during embryogenesis, with the pattern of production of BMPs 2, 4, and 6 indicating a role in bone and cartilage formation. The regulation and precise functions of the BMPs remain to be elucidated, but estrogen-induced stimulation of the production of BMP-6 mRNA and protein has been demonstrated in human osteoblastic cell lines (311).

IGFs exist in two forms: IGF-I and -II. In the circulation, they form a large-molecular-weight complex with binding proteins (IGFBPs), and in the case of IGFBP3 and -5 complexes an acid-labile subunit (309). IGFs stimulate bone formation, their production by bone cells being regulated by a number of systemic hormones and locally produced factors (45). They increase proliferation of osteoblast precursors and enhance the synthesis and inhibit the degradation of type I collagen (145, 241). There are at least six IGFBPs (45, 212), all of which are expressed by bone cells in various in vitro systems (319). All IGFBPs bind IGFs with high affinity, preventing their interaction with the receptor. However, because of posttranslational modifications that result in changes in both structure and function, the IGFBPs may exert either stimulatory or inhibitory effects; thus, for example, IGFBP-1 and -3 have both stimulatory and inhibitory potential, IGFBP-2 and -4 are inhibitory, and IGFBP-5 is stimulatory (251). IGFBP-6 is inhibitory and exhibits a selective affinity for IGF-II over IGF-I. The complexity of the IGF axis is further increased by the action of IGFBP proteases, which affect the binding affinity of the binding proteins for IGFs and may themselves be regulated by IGFs (64, 85).

	III. LIFETIME CHANGES IN BONE MASS: EFFECTS OF SEX STEROIDS

Top Previous Next References

Bone mass increases throughout childhood and adolescence (30, 31, 126); in prepubertal children, there is a close relationship between bone mass and body height, but this becomes less evident during puberty. In girls the rate of increase in bone mass decreases rapidly after the menarche, whereas gains in bone mass in boys persist up to 17 yr of age (30, 353) and are closely linked to pubertal stage and androgen status (200). Although by the age of 17 or 18 in both sexes the vast majority of peak bone mass has already been achieved, small increases in bone mass during the third decade of life have been demonstrated in several studies (31, 116, 290, 310); however, this finding has not been consistently reported (159, 236, 266). Peak bone mass is attained in the third decade of life and maintained until the fifth decade, when age-related bone loss commences both in men and women, thereafter persisting throughout life (140, 174, 239, 240, 313, 314, 318) (Fig. 6).

View larger version (12K): [in this window] [in a new window]   Fig. 6. Lifetime changes in bone mass. (From Compston JE. Osteoporosis, corticosteroids and inflammatory bowel disease. Aliment Pharmacol Ther 9: 237-250, 1995.)

The onset of age-related bone loss has not been well defined. In cross-sectional studies, bone loss has been documented in healthy premenopausal women at the spine, proximal femur, and forearm (14), and this finding has also been confirmed in prospective studies (13, 54, 337, 340). In women there is an acceleration in the rate of bone loss at the time of the menopause, the duration of which has not been well characterized but is probably between 5 and 10 yr (16, 88, 123, 161, 265). In men, relatively few data are available, but bone loss is generally believed to begin during the fifth decade of life; thereafter, both in women and men, bone loss continues throughout life (140, 174, 239, 240, 313, 314, 318).

Genetic factors are important determinants of peak bone mass, and up to 60-80% of its variance is genetically determined (51, 78, 182). The basis of this effect has not been fully defined, and a number of genetic polymorphisms are likely to be involved. A polymorphism in the regulatory region of the collagen 1AI gene at a recognition site for the transcription factor Sp1 has been demonstrated to correlate with bone mineral density and fracture in several populations (131, 366); there are many other potential candidates including the vitamin D receptor gene, estrogen receptor gene, and genes for many cytokines and growth factors (306). Other determinants of peak bone mass include nutrition, calcium intake, physical activity, and hormonal status.

Sex steroids play an important role in bone growth and the attainment of peak bone mass. They are responsible for the sexual dimorphism of the skeleton, which emerges during adolescence (369); the male skeleton is characterized by larger bone size (even when corrected for body height and weight) with both a larger diameter and greater cortical thickness in the long bones. Volumetric bone mineral density is, however, very similar in young adult men and women (183), but the larger bone size in men confers significant biomechanical advantages and, in part, explains the lower incidence of fragility fractures compared with women. Estrogen is essential for normal closure of the growth plates in both sexes; thus estrogen resistance and aromatase deficiency in men are associated with delayed bone age and tall stature despite normal or high circulating concentrations of testosterone (252, 336).

Hypogonadism has adverse effects on the attainment of peak bone mass both in men and women. Late menarche has been associated with reduced bone mineral density (321, 340) and premenopausal amenorrhea resulting from anorexia nervosa (24, 315), excessive exercise (84, 234), and hyperprolactinemia (23), and a variety of other disorders (73) also result in low bone density. Reduced spinal bone mineral density has been reported in women with asymptomatic disturbances of ovulation (i.e., without amenorrhea) (305), although this finding has not been universal (79, 388), and premature menopause, whether natural or induced, is a major risk factor for osteoporosis (12). Low bone mineral density values have also been reported in Turner's syndrome, predominantly reflecting the smaller bone size associated with this condition (260, 263, 322), which is believed to be due to resistance to growth hormone (374).

The role of androgens in growth of the male skeleton during puberty is supported by several observations. Androgen deficiency due to hypogonadotropic hypogonadism is associated with low bone mineral density (103), while administration of testosterone before epiphyseal closure leads to increases in bone mass (102) and testosterone administration to prepubertal boys results in increased bone calcium accretion (238). The timing of puberty may also be important, with some studies indicating that late puberty is associated with reduced bone mineral density and peak bone mass later in life (21, 104); in these subjects, increases in bone mineral density were reported in response to testosterone therapy. Notwithstanding these observations, however, the effects of estrogen resistance and aromatase deficiency on skeletal mass (253, 336) indicate that estrogens also play an important role in skeletal development in males during adolescence; furthermore, it is uncertain to what extent the skeletal effects of androgens are mediated by local metabolism to estrogens. Finally, there is evidence that androgens also have effects on the attainment of peak bone mass in women (42, 43, 71), conditions of androgen excess in women being associated with higher bone mineral density (42, 81).

Estrogen deficiency is a major pathogenetic factor in the bone loss associated with the menopause and the subsequent development, in some women, of postmenopausal osteoporosis. Estrogen replacement at or after menopause, whether natural or induced, prevents menopausal bone loss and characteristically results in an increase in bone mineral density during the first 12-18 mo of treatment (52, 96, 218, 259, 346). This increase, which is typically between 3 and 5% but may be as much as 10% (53, 219), is attributed to the simultaneous reduction in activation frequency and formation of new bone within existing resorption cavities when an antiresorptive agent is administered in high turnover states. There is evidence, almost exclusively from observational studies, that estrogen replacement is associated with a reduction in fracture risk at the hip, spine, and wrist (162, 187, 249, 261, 285, 396); however, such studies are biased by the better health status of women who choose to take estrogens as opposed to those who do not and are thus likely to overestimate any benefit (58).

Even in postmenopausal women, the small amounts of estrogen produced endogenously are determinants both of bone mineral density and fracture risk. In a large population-based study it was demonstrated that women aged 65 yr or older with serum estradiol levels between 10 and 25 pg/ml had significantly higher bone mineral density in the hip, spine, calcaneus, and proximal radius than those with estradiol levels below 5 pg/ml (97). Furthermore, women with undetectable serum estradiol levels had a significantly increased risk of hip and vertebral fractures compared with those with levels above 5 pg/ml, and this risk was further increased in the presence of high serum concentrations of sex hormone binding globulin (68). These interesting and unexpected data challenge the perception that endogenous estrogen production in postmenopausal women does not have physiological skeletal effects and emphasize the potential functional significance of relatively low concentrations of the hormone in the bone microenvironment. In this respect, the presence in human osteoblastic cells of 17-hydroxysteroid dehydrogenases (17-HSDs), which interconvert estradiol, and the relatively inactive estrone (and testosterone) may be relevant, providing a mechanism for the local regulation of intracellular ligand supply for estrogen receptors (82). Four isoforms of this enzyme have been cloned (6, 122, 228, 409), with 17-HSD I and III being mainly involved in the reduction of estrone to estradiol and testosterone to dihydrotestosterone and 17-HSD II and IV in the oxidation of estradiol to estrone.

The relationship between the age-related decline in serum testosterone levels and reduction in bone mineral density in men is less well documented, and although some studies have demonstrated such a correlation (106, 257), this finding has not been universal (244). However, hypogonadism is believed to be an important pathogenetic factor in male osteoporosis (272, 341); in the majority of such cases, there are no overt clinical manifestations of hypogonadism, the diagnosis being established by the presence of low free serum testosterone levels. Klinefelter's syndrome is associated with low bone mineral density (107, 152), and castration in adult men is followed by rapid bone loss with evidence of increased bone turnover (345), similar changes being described after the administration of gonadotrophin-releasing hormone analogs (127). The extent to which conversion of androgens to estrogen in bone is responsible for the effects of androgens in adult men is unclear; some studies have reported closer correlations between bone mineral density and estrogen than androgen status (134, 199). Furthermore, prevention by estrogens of bone loss associated with cyproterone acetate in trans-sexual men has been reported (220), and there is indirect evidence that the beneficial effects of testosterone on bone mineral density in eugonadal men with osteoporosis may be partly mediated by conversion to estrogens (10).

	IV. SKELETAL EFFECTS OF ESTROGEN: MECHANISMS OF ACTION

Top Previous Next References

Estrogen has a diverse range of actions involving growth, differentiation, and function in many target tissues. The mechanisms by which these actions are achieved have not been fully established, but it is thought that many of the effects of estrogen are mediated by a genomic pathway involving ligand/receptor interaction. The importance of nongenomic mechanisms, in which the ligand interacts with plasma membrane receptors, is increasingly recognized in the mediation of rapid responses to estrogen (39, 393) and in the ROS osteoblastic cell line rapid activation of mitogen-activated protein kinase by estrogen has recently been reported (89). In addition, there is evidence for nongenomic effects of estrogen on osteoclasts, rapid tyrosine phosphorylation of several proteins, including src, being reported in avian osteoclasts after administration of 17-estradiol (38).

Estrogen receptors (ERs) belong to a family of steroid hormone receptors that include receptors for glucocorticoids, androgens, progestins, and mineralocorticoids (135) and can be considered as ligand-regulated transcription factors. ERs consist of several domains, defined according to their function (Fig. 6). The AF-1 and AF-2 sites (activation functions 1 and 2) activate gene transcription, with the AF-1 being constitutively active and responsible for promotor-specific activation, independent of the presence of ligand, whereas AF-2 is ligand specific (20, 392). The C region contains the highly conserved DNA-binding domain with two zinc fingers that are essential for DNA binding (208). The classical estrogen response element (ERE) consists of an inverted hexanucleotide repeat (A/GGGTCA) separated by three nucleotides. The hormone binding domain is in the COOH terminus of the molecule and is responsible for specific ligand recognition and binding. The E region, and possibly also the C region, contains a 90-kDa heat shock protein function (229).

At least two main ER subtypes exist, namely, ER and ER. ER was originally cloned from the uterus (133) and, more recently, ER was cloned, initially from a rat prostate cDNA library (90, 204, 254, 358). The ER shows close structural homology with the ER molecule, especially in the DNA binding domain and, to a lesser extent, in the ligand binding domain (Fig. 7). The binding affinities of estradiol and other ligands including SERMs and phytoestrogens for the two ER subtypes are very similar (203). Several isoforms of the ER and at least two of ER, created by alternative splicing or alternative initiation of translation, have been demonstrated (mainly at mRNA level); one of these does not bind estrogen and may act as a dominant negative inhibitor of ER-mediated activity (207).

View larger version (18K): [in this window] [in a new window]   Fig. 7. Structure of estrogen receptors (ER)  and . The percentage figures indicate the degree of structural homology for each domain between the two receptor subtypes; these are similar in the rat, mouse, and human.

Mice with loss of function mutations of the ER gene (ERKO) show only minor skeletal abnormalities with reduced longitudinal bone growth, particularly in females, and modest reductions in bone mineral density which are, in contrast, more prominent in males (66, 286). These changes differ from those observed in human males with ER resistance (336) or aromatase deficiency (253), in which longitudinal growth is increased. In the ER knock-out model (BERKO), increased cortical bone mineral content and periosteal diameter have been reported in females, but the males exhibit a normal skeletal phenotype (383). No effect on ovariectomy induced bone loss was demonstrated in these mice; this observation, together with the normal trabecular bone mineral density in the intact females, indicates that ER does not mediate the protective skeletal effects of estrogen in this species. To date, therefore, the knock-out models do not indicate a major role for either of the two known ER subtypes in mediating estrogen-induced effects on the skeleton, possibly reflecting the presence of other, as yet unidentified ERs.

The tissue distribution of the ERs is overlapping but not identical, and at least in some tissues where both receptor subtypes exist, they are cell specific, possibly indicating different functions (202). In keeping with the diverse actions of estrogen, ERs are widely distributed and are found in the central nervous system, heart, blood vessels, mammary gland, uterus, testis, epididymus, bladder, ovary, kidney, intestine, prostate, and bone (90, 203, 204, 206, 254, 292, 358). However, it should be recognized that current knowledge of the tissue distribution of the two receptor subtypes is based mainly on localization of mRNA rather than protein.

The presence of ER (presumably ER) on rat and human osteoblastic cells was first reported in 1988 (91, 276) and subsequently extended to osteoclasts (295) and osteocytes (35). However, the relative proportion and distribution of the two receptor subtypes in bone remains to be established. ER mRNA has been reported on rat osteoblastic cells (268) and also in a human osteoblast cell line, SV-HFO (11). Recently, Vidal et al. (382) reported the presence of ER mRNA in human osteoblast cell lines and cultures and have also demonstrated the presence of ER protein in these cells, both in vitro and in vivo. Furthermore, ER protein was identified in osteocytes, where the staining was nuclear, and in osteoclasts, in which staining was predominantly cytoplasmic. Interestingly, these workers noted the presence of nuclear and cytoplasmic staining for ER in some bone marrow cells, an observation consistent with the recent report of ER expression in megakaryocytes in human bone marrow (33).

ER protein has also been demonstrated in the growth plates of rodents and rabbits, where it is localized in the proliferative and early hypertrophic zone (184). The observation in rats that loss of expression at sexual maturity is associated with failure of epiphyseal closure is consistent with the well-documented role of estrogen in this process.

In target cells, 17-estradiol diffuses through the plasma membrane and binds to the ER (Fig. 8). On binding, heat shock proteins dissociate, and the receptor undergoes a conformational change and dimerization (164, 229). The receptor/ligand complex then binds to response elements within the promotor area of target genes, resulting in transcriptional activation and modulation of gene expression. In addition, ERs can regulate the transcription of genes that lack classical EREs in their promotor region by modulating the activity of other transcription factors such as AP-1, NFB, and Sp1 (120, 302, 342). The conformational change that occurs in the ligand-binding domain of the receptor enables the AF-2 function of the ER to interact with coactivators and corepressors in a ligand-dependent manner; in the case of 17-estradiol, this results in the formation of a transcriptionally competent complex and the initiation of gene transcription (154, 165, 195).

View larger version (23K): [in this window] [in a new window]   Fig. 8. Estrogen signaling pathways. The ligand 17-estradiol is transported to the nucleus where it forms a complex with the estrogen receptor (ER). This subsequently undergoes dimerization and conformational change resulting in the formation of a transcriptionally competent complex that binds to response elements in target estrogen-sensitive genes. In addition to the classical ERE and the AP-1 site shown in the diagram, other transcription factors such as NFB and Sp1 can interact with the ER and modulate gene transcription.

A number of estrogen-induced effects on gene expression in osteoblasts have been described (275). These include induction of TIEG, a TGF--inducible gene that inhibits DNA synthesis (351), IGF-I (93, 94), and TGF- (274, 276). Increased BMP-6 mRNA expression has also been reported in response to estrogen in a fetal osteoblastic cell line (311). Reports on the effects of estrogen on DNA synthesis and proliferation and bone matrix protein production have produced conflicting results, possibly as a result of differences in the in vitro systems investigated and, in particular, the stage of differentiation of osteoblasts in these systems (275). Thus, in osteoblastic cells, for which estrogen acts as a mitogen, increased expression of alkaline phosphatase and type I collagen has been reported (230, 416), whereas in cells that show no proliferative response to estrogen, stimulation of type I collagen and osteocalcin expression have been demonstrated with no increase in alkaline phosphatase (181). Third, in systems in which estrogen has antiproliferative effects, stimulation of alkaline phosphatase expression has been reported, with suppression of osteocalcin and variable effects on type I collagen expression (317). Estrogen also increases expression of the receptors for 1,25(OH)₂D (95), growth hormone (163), and progesterone (334); modulates PTH responsiveness in osteoblastic cells (93, 112); and increases expression of IGFBP-4, as well as reducing its proteolytic breakdown (180).

The report by Pensler et al. (295) that ERs were present on osteoclasts has since been confirmed by a number of groups in bone from humans (155, 277), chicks (276), mice (150, 250), and rabbits (232). Levels of the ER on osteoclasts are generally low and, as discussed below, the antiresorptive effects of estrogen may largely be mediated by modulation of cytokine production by cells in the bone microenvironment rather than by direct effects on osteoclasts. However, estrogen-induced reduction in the expression of mRNAs and secretion of several lysosomal enzymes, including cathepsin L, -glucuronidase, and cathepsin K have been reported in osteoclasts in vitro (201, 278).

The bone-preserving action of estrogen is mediated predominantly if not solely through effects on osteoclast number and activity, the latter encompassing both resorptive activity per se and the life span of the cell. Studies in ovariectomized rodents have demonstrated an increase in the proliferation and differentiation of osteoclast precursors (168, 169), increased numbers of stromal/osteoblastic cells (170, 190), and reduced osteoclast apoptosis (158). These effects are, in turn, believed to be largely mediated via cytokines involved in the regulation of osteoclastogenesis and osteoclastic activity. Studies in postmenopausal women have demonstrated increased production of IL-1, GM-CSF, and TNF- by monocytes in the bone microenvironment after natural or surgical menopause, these changes being abrogated by the administration of exogenous estrogen (281, 282, 307). In support of these observations, treatment with TNF binding protein prevents bone loss in ovariectomized rats but has no effect in estrogen-replete animals (189, 194). The increase in IL-1 activity associated with estrogen deficiency is a result not only of increased IL-1 synthesis but also of decreased production of IL-1ra (283); thus treatment of ovariectomized rats with IL-1ra decreases bone loss (191) by blocking the proliferation and differentiation of osteoclast precursors (188). Mice that are unable to synthesize or respond to either IL-1 (8) or TNF- (301) do not exhibit the bone loss seen in normal animals after ovariectomy, and simultaneous inhibition of IL-1 and TNF activity is required completely to prevent bone loss after ovariectomy in normal mature rats (189). However, these animals have a normal bone phenotype with no evidence of abnormal remodeling activity when sex hormone status is normal (167). These observations emphasize the interdependent nature of cytokine regulation; IL-1, IL-6, and TNF- not only induce their own synthesis but also have synergistic autocrine effects, TNF- and IL-1 acting to increase production of TNF and IL-6, and PTH synergizing with TNF to stimulate IL-6 production (80, 108, 167, 291).

Estrogen also inhibits the production of IL-6 by blocking the activity of the transcription factors NFB and CCAAT/enhancer binding protein  that are required for activation of the IL-6 promotor (114, 209, 303, 342). In vivo studies in ovariectomized mice have demonstrated increased production of IL-6 from bone marrow cells (168) and increased expression of the IL-6 soluble receptor IL-6R, through which the effects of IL-6 are mediated, may also contribute (217). Transgenic mice overexpressing IL-6 do not exhibit osteopenia or increased osteoclastogenesis (193, 349, 399), and IL-6-deficient mice exhibit a normal bone phenotype, although they are protected from ovariectomy-induced bone loss (301). The role of IL-6 in the pathogenesis of menopausal bone loss in women remains to be fully established.

Effects of estrogen on stromal/osteoblastic cells, which support osteoclastogenesis, have been reported. Thus estrogen deficiency is associated with an increase in this cell population (170), and increased synthesis of M-CSF and osteopontin has been reported in vitro and in ovariectomized animals (100, 190, 411). Recently, it has also been shown that estrogen increases levels of OPG mRNA and protein in osteoblastic cells (148). In addition, estrogen plays an important role in the regulation of osteoclast activity. The cytokines IL-1, IL-6, TNF-, and M-CSF have all been shown to inhibit apoptosis in osteoclasts (156, 172), whereas TGF-, the production of which is decreased in estrogen deficiency states, stimulates apoptosis (158). Estrogen may also directly stimulate apoptosis by decreasing expression of NFB-activated genes that normally suppress apoptosis (171). Interestingly, the reverse effect has been reported for osteocytes, acute estrogen withdrawal in humans being associated with increased apoptosis of osteocytes (357).

Evidence for a role of nitric oxide in bone loss associated with estrogen deficiency is provided by the observation that nitroglycerine, a nitric oxide donor, alleviates bone loss induced by ovariectomy in rats and that in the presence of N^G-nitro-L-arginine methyl ester, an inhibitor of nitric oxide synthase (NOS), estrogen was ineffective in reversing bone loss (398). This is consistent with earlier studies in the guinea pig demonstrating estrogen-induced regulation of the constitutive NOS enzymes, epithelial NOS and neuronal NOS (394), and with the inhibitory effect of high nitric oxide concentrations on osteoclastogenesis and osteoclastic activity (although there is some evidence that lower concentrations of NO have a stimulatory effect on bone resorption) (98). Interestingly, functional ERs have been demonstrated in bone endothelial cells in vitro (36), supporting a role for estrogens in angiogenesis and hence, potentially, access of osteoclasts to remodeling bone surfaces (288).

The role of estrogen in the regulation of osteoclast activity is thus mediated via effects on osteoclast number and activity. The former action is determined both by direct cytokine-induced effects on osteoclast proliferation and differentiation and by modulation of the stromal/osteoblastic cell population that supports osteoclastogenesis. Changes in osteoclast activity are probably mediated predominantly through effects on apoptosis.

Ovariectomy leads to the development of rapid cancellous bone loss in some species, particularly the rat, with an increase in osteoclast and osteoblast number and also an increase in osteoclast size (408). In young rats, much of the apparent cancellous bone loss occurs as a result of increased resorption of calcified cartilage by chondroclasts (405). Bone formation rates are increased, consistent with high bone turnover, and these changes persist for at least 1 yr after ovariectomy (407). Studies of cancellous bone architecture in the ovariectomized rat have demonstrated that bone loss is accompanied by osteoclastic perforation and erosion of trabecular plates without trabecular thinning (77), indicating that both the number and activity of osteoclasts are increased in estrogen-deficient states. In cortical bone, increased bone resorption results in an increase in the volume of the medullary canal in the tibiae (175); however, there is also an increase in bone formation at the periosteal surface that may exceed endocortical resorption in rapidly growing rats (362). Osteoclast numbers are increased at the endocortical surface. These changes, both in cancellous and cortical bone, can be prevented by administration of estrogen (362, 406).

It is important to emphasize that sexually mature rodents should be used for these models to avoid confounding effects of estrogen deficiency on longitudinal growth (192). Other animals that have been studied as models of estrogen deficiency-induced bone loss include mice, ferrets, dog, sheep, swine, and monkeys. These species vary in their skeletal responsiveness to estrogen depletion and are less well established than the rat model (121, 192).

Histomorphometric data on the skeletal changes associated with menopausal bone loss are sparse and restricted to cross-sectional studies in relatively small numbers of women. Some of these studies have provided evidence for an increase in bone turnover during the menopause, both in cortical and cancellous (37, 86, 377), although this finding has not been universal (246). These somewhat conflicting data contrast with results obtained from kinetic and biochemical measurements of bone turnover, which have invariably demonstrated an increase in bone turnover during menopause (143, 365). Furthermore, estrogen replacement therapy is associated with a return to premenopausal values of biochemical markers of bone resorption and formation. The failure of histomorphometric studies to demonstrate unequivocally an increase in bone turnover in association with menopause is likely to be attributable to several factors including the small numbers studied, lack of prospective data, and the large measurement variance associated with bone histomorphometry.

A consistent finding in untreated postmenopausal women has been a reduction in wall width, indicating reduced bone formation at the cellular level and hence a reduction in osteoblast activity (221, 377). The age at which this reduction occurs is uncertain. Thus Lips et al. (221) reported an age-related reduction in mean wall width in 22 men and 14 women aged between 18 and 82 yr, whereas in another study, the age-related reduction in women and men appeared to begin after the age of 50 yr (377). However, the cross-sectional design of both these studies makes it difficult to determine accurately the age of onset of change. Whether this change is specifically related to estrogen deficiency is uncertain; similar changes occur in men, and conventional estrogen replacement at menopause has not been demonstrated to reverse this change. In women, an age-related decrease in wall width has also been reported in cortical bone in some, but not all, studies (37, 110, 166). Measurement of resorption depth has demonstrated a small decrease or no change in postmenopausal women, suggesting that the negative remodeling balance is primarily due to reduced bone formation (67, 92). However, studies of acute estrogen deficiency in premenopausal women, induced by administration of gonadotrophin releasing hormone analogs, suggest that there may be a transient increase in resorption depth (63). In these women, rapid and significant disruption of cancellous bone architecture was observed after 6-mo therapy; these changes are unlikely to be due solely to increased bone turnover and would be consistent with an early and transient increase in osteoclastic activity, resulting in increased cavity depth and trabecular penetration and erosion. Furthermore, in cortical bone, an increase in resorption depth within Haversian systems was demonstrated in these patients (17).

The greater age-related disruption of cancellous bone architecture in women than in men (60, 245) also supports the contention that estrogen deficiency is associated with increased erosion depth. Studies of cancellous bone structure in women have clearly demonstrated a reduction in trabecular continuity and loss of whole trabeculae after menopause. Whether there is significant trabecular thinning is less certain; some studies have reported significant or nonsignificant decreases in trabecular width, whereas others have found no change (2, 25, 61, 386, 395). The increase in trabecular separation that has consistently been demonstrated in postmenopausal women may thus mainly reflect loss of whole trabeculae rather than trabecular thinning. It is also possible that there is preferential erosion of thin trabeculae so that the contribution of trabecular thinning to bone loss is underestimated.

There have been relatively few bone histomorphometric studies of the effects of hormone replacement therapy. Evidence that hormone replacement reduces bone turnover was first reported by Riggs et al. (312) in a prospective study of 17 women with established osteoporosis. Iliac crest bone biopsies were obtained before and either 2.5-4 mo (short-term) or 26-42 mo (long-term) after estrogen replacement. After 2.5-4 mo, there was a significant reduction in bone-resorbing but not bone-forming surfaces, both of these being evaluated by microradiography; in contrast, after 26-42 mo, there was a significant reduction in both resorbing and forming surfaces. These data thus indicate that estrogen replacement reduces bone turnover, a suppressive effect on bone resorption being followed by a later decrease in bone formation.

A more detailed histomorphometric analysis of the effects of hormone replacement therapy on bone remodeling was later reported in a study of postmenopausal women with established osteoporosis (344). Bone formation rate at tissue level and activation frequency, both indices of bone turnover, were significantly decreased at 1 yr to ~50% of the pretreatment value, but no significant changes were observed in resorption depth or wall width, suggesting that remodeling balance was unchanged. However, because of the long life span of the bone remodeling unit in humans and, in particular, the time required for formation to be completed, a period of at least 2 yr is required to demonstrate changes in wall width induced either by disease or treatment. In contrast, because the resorptive component of the remodeling cycle is relatively rapid, changes may be seen over a much shorter period of time. Similar changes in bone turnover were reported in osteoporotic postmenopausal women after a 1-yr treatment with transdermal estrogen (226). Activation frequency and bone formation rate were both significantly lower in the posttreatment biopsies, bone turnover being suppressed to well below pretreatment values. A reduction in activation frequency, but not bone formation rate, was also reported in a study of postmenopausal women with low bone mineral density after treatment for 1 yr with percutaneous estradiol therapy (149). Finally, in a 2-yr prospective treatment study in postmenopausal women with osteopenia or osteoporosis, a significant reduction in bone turnover was observed; in addition, there was a trend toward decreased resorption cavity size after treatment, consistent with suppression of osteoclastic activity by hormone replacement therapy and a small reduction in wall width, possibly reflecting compensatory changes in response to the reduction in resorption cavity size (379). In this cohort, there was no significant change in cancellous bone structure during the study period, indicating that hormone replacement therapy preserves existing bone microstructure but does not reverse previously induced structural disruption (378).

These studies thus provide strong evidence that hormone replacement therapy, whether given as estrogen alone or combined with a progestin, preserves bone mass predominantly by reducing bone turnover. The relative contribution to this action of effects on the process of activation per se and those on osteoclast number and activity have not been established; a role for the latter mechanism is supported by the well-documented effects of estrogen on osteoclast proliferation, differentiation, and activity demonstrated in vitro. The effects of estrogen administration on remodeling balance remain to be fully defined, but there is at present no evidence that, when given in conventional doses, estrogens increase bone formation at the cellular level. It is therefore possible that the age-related decrease in wall width may be an estrogen-independent phenomenon. Conversely, there is some evidence that estrogen replacement reduces resorption cavity size and hence improves this component of remodeling imbalance.

Evidence from animal studies indicates that high doses of estrogens have anabolic skeletal effects (87, 356), but until recently, it was unknown whether similar effects occur in the human skeleton. Percutaneous estrogen implant therapy has been reported to be associated with higher bone mineral density levels than oral or transdermal hormone replacement, an observation that may be related to the higher serum estradiol concentrations associated with parenteral treatment (117, 323, 328, 347). Many of these studies, however, were cross-sectional and involved the coadministration of testosterone implants, thus providing only indirect evidence for an anabolic skeletal effect of estrogen.

Recently, Wahab et al. (385) reported high bone mineral density values in a cohort of women who had received long-term high-dose estradiol implant therapy, without testosterone. A histomorphometric assessment of iliac crest bone from a subgroup of this cohort was performed, and the values obtained compared with those of healthy premenopausal women (375), based on the rationale that significant age-related bone loss had not occurred in the patient group before estradiol replacement and that any differences between the two groups would therefore reflect effects of high dose as opposed to physiological estrogen replacement. The results of this study demonstrated a significantly higher wall width in the implant-treated group (Fig. 9), providing direct histological evidence that high-dose estrogens produce anabolic skeletal effects in postmenopausal women and indicating that these are achieved by stimulation of osteoblastic activity, resulting in increased bone formation at cellular level and hence a more positive remodeling balance.

View larger version (50K): [in this window] [in a new window]   Fig. 9. Wall width in women treated with high-dose, long-term estradiol and normal premenopausal women. High-dose estradiol therapy was associated with a significantly higher wall width than that found in normal premenopausal women, reflecting increased bone formation at the cellular level due to increased osteoblastic activity. Data are shown as means ± SD. (From Compston JE. The skeletal effects of oestrogen depletion and replacement: histomorphometrical studies. In: Annual Review of the Management of Menopause, edited by Studd J. Carnforth, Lancs, UK: Parthenon, 2000, p. 287-296.)

These findings have recently been confirmed in a prospective study of women undergoing treatment with estradiol implant therapy (185). In this study, not only was a significant increase in wall width observed, but changes indicative of increased connectivity of cancellous bone structure were also demonstrated. This raises the interesting possibility that the anabolic skeletal effects associated with high-dose estrogen therapy in postmenopausal women may result not only from improvement in remodeling balance but also de novo bone formation; the latter mechanism has been described in mice (326), but further studies are required to investigate its potential contribution to the observed changes in the human skeleton.

	V. EFFECTS OF PROGESTERONE ON BONE

Top Previous Next References

Relatively little is known about the effects of progestins on bone metabolism. Normal human osteoblastic cells express progesterone receptors (196), and stimulation of the proliferation and differentiation of these cells has been reported in response to relatively high doses of progesterone (46). In the ovariectomised rat model, progesterone was reported to have similar effects to estrogen in one study (15) but antagonistic actions in another (360).

Menopausal estrogen therapy in women with an intact uterus is combined with a progestin to prevent increase in endometrial cancer risk associated with the use of unopposed estrogen. Some of the progestrogens used in these formulations, particularly 19-nortestosterone derivatives, may independently have beneficial effects on bone mass, although the evidence in this area is conflicting (3, 5, 316, 331). Thus preservation of bone mineral density in postmenopausal women treated with norethisterone was demonstrated in metacarpal cortical bone (3), but Hart et al. (142) reported that norgestrel therapy was associated with significant bone loss at this site in a similar cohort. In a study of the effects of medroxyprogesterone in early postmenopausal women, Gallagher et al. (114a) demonstrated preservation of total body bone mineral density (reflecting predominantly cortical bone) but significant losses at the spine, forearm, and metacarpal cortex. Consistent with these findings, Adachi et al. (5) were unable to demonstrate any beneficial effect of medroxyprogesterone on bone mineral density in the lumbar spine or proximal femur in postmenopausal women taking estrogen replacement therapy. However, increases in bone mineral density have been