Portal de Periódicos da CAPES

Mice, Estrogen, and Postmenopausal Osteoporosis

1999; Oxford University Press; Volume: 14; Issue: 2 Linguagem: Inglês

10.1359/jbmr.1999.14.2.187

1523-4681

Russell T. Turner,

Bone Metabolism and Diseases

Samuels et al. have presented evidence that the well known osteosclerotic response of mice to high doses of estrogen is due to de novo bone formation. This finding raises fundamental questions regarding the mechanism of action of estrogen in regulating bone cell metabolism, as well as the appropriateness of the mouse model for studies related to the pathogenesis and treatment of postmenopausal bone loss. However, by considering the physiological effects of estrogen on bone, it will become clear that the effects described by Samuels and co-workers are unlikely to be due to a direct action of estrogen on bone via a conventional pathway. Indeed, the osteosclerotic mouse may be more accurately characterized as a model for complications resulting from a toxic response to estrogen excess. Unlike reproductive tissues, the major functions of the skeleton occur normally in sexually immature and hypogonadal individuals. These estrogen-independent functions include physical protection of vital organs, participation in maintenance of mineral homeostasis, and provision of the rigid scaffold essential for locomotion in vertebrates. However, epidemiological and experimental studies have identified estrogen deficiency as an important risk factor for osteoporosis.1 Two related specialized skeletal functions have been identified which are dependent upon estrogen: development of gender differences in bone mass and architecture, and adaptation of the female skeleton to store extra mineral in bone in preparation for the increased mineral demanded during reproduction. The distinction between sexual dimorphism and reproductive function is somewhat arbitrary because it is clear that some of the gender differences in bone architecture are directly related to reproduction. The sexual dimorphism of the skeleton is species-, bone-, and bone compartment–specific. Androgens, as well as estrogens, are important. The architecture of the adult skull, pelvis, and long bones exhibit many gender differences, the development of which can be reduced or prevented in laboratory animals by neonatal gonadectomy. Sex hormones influence peak bone mass and internal as well as external bone architecture. Because the effects of the hormone are not uniform throughout the skeleton no universal action of estrogen on bone cell metabolism is identifiable. This lack of uniformity suggests that the response of bone cells to estrogen is context dependent. The rat serves as a useful laboratory animal model for sexual dimorphism of the human skeleton. In growing rats, as in adolescent humans, the tibia increases in length and radius at a more rapid rate in males. This gender difference in growth leads to a greater peak cortical bone mass in males. The growth processes involved in determining bone mass include endochondral ossification (bone length), secondary intramembranous ossification (cross-sectional area), and endocortical modeling (medullary area). The essential role of sex hormones in mediating the gender differences in growth of the rat tibia are easily demonstrable. Gonadectomy largely extinguishes the gender differences in both radial and longitudinal bone growth; bone growth decreases in males and increases in females. Importantly, the gender differences can be reestablished in gonadectomized animals by administering androgens to males to increase growth and estrogens to females to reduce growth.1-3 The cellular mechanisms for estrogen action on the growth of the rat tibia have been extensively studied.1 Estrogen reduces proliferation of cartilage cells in the proliferative zone of the growth plate. The hormone may also decrease chondrocyte hypertrophy. The net result of these changes is a decrease in linear growth rate. On the per iosteal bone surface, estrogen inhibits preosteoblast proliferation and decreases osteoblast activity. On the endocortical bone surface, estrogen reduces osteoclast number and decreases bone resorption. In comparison with the rat, the gender difference in tibia size in the limited number of mouse strains that have been evaluated appears to be much less pronounced. Ovariectomy (OVX) has little influence on radial bone growth, longitudinal bone growth, and endocortical bone formation in growing mice.4 Estrogen treatment does not alter radial bone growth, suggesting, in comparison with domestic birds, rat, and humans, that the mouse periosteum is much less responsive to estrogen. The special adaptation of skeleton to the demands of reproduction reaches its highest development in birds. Prior to egg laying, the marrow cavities of certain bones are invaded by a complex network of cancellous bone, called medullary bone. This estrogen-dependent bone is limited to skeletal sites with hematopoietic marrow; treatment with the hormone reduces bone formation at the periosteum and promotes premature epiphyseal closure. This new bone, which serves no mechanical role, functions as a mineral reservoir for eggshell calcification. Medullary bone quickly disappears after the hen lays her final egg in a clutch. This seasonal bone provides an elegant adaptation to competing requirements of the avian skeleton, namely the ability to mobilize large amounts of mineral from the skeleton in the brief interval that an egg is located in the shell gland and the necessity to minimize body weight for flight.5 No other class of vertebrates, including those that lay eggs with a leathery shell (e.g., reptiles), form medullary bone. Nevertheless, special skeletal adaptations related to reproduction are observed in many species. For example, the turtle carapus provides a reservoir for mineral, which can be mobilized during reproduction. In mammals, increasing estrogen secretion during adolescence can lead to increased cancellous bone volume which can then be mobilized during pregnancy and lactation. Urist et al.6 described a process whereby estrogen increased cancellous bone volume in rodents by antagonizing resorption of the calcified growth plate cartilage and bone. This antiresorbing activity of the hormone leads to an increase in the retention of the primary spongiosa. Ongoing bone formation is unchanged or modestly reduced.7 However, because of the prolonged survival of the trabeculae, which is entirely due to decreased resorption, bone volume is increased. More recent studies investigating the cellular mechanism for estrogen action on cancellous bone volume in growing rats have demonstrated pronounced decreases in the production and overall number of chondroclasts.8 Estrogen-induced osteosclerosis in mice is often compared with avian medullary bone, which it superficially resembles. If the mechanisms of intramedullary bone formation by birds and mammals were similar, one might hypothesize that avian medullary bone represents a hypersensitivity of the skeleton to estrogen which evolved by natural selection. However, there are many observations that argue against this interpretation. Estrogen-induced endocortical bone formation is not a general phenomenon in vertebrates; it has only been described in mice. The endocrine requirements for bone formation differ between mice and birds. Androgens potentiate medullary bone formation in birds, whereas in mice, androgens are reported to antagonize the development of osteosclerosis.1, 7 There are also important differences in the histology, kinetics, and cellular mechanisms of estrogen-induced bone formation in mice and birds. The skeletal response of birds to estrogen is essentially immediate: a highly synchronized sequence of molecular and cellular events is rapidly initiated by moderate elevations of the hormone.9 Osteoblasts appear not only on endocortical bone surfaces but also deep within the marrow cavity. These islands of new bone formation rapidly collalese to form trabeculae which radiate from the endocortical surface at up to 1000 μ m/day.10 These new trabeculae differ from those in normal cancellous bone in mechanism of origination (intramembranous rather than endochondral), orientation (haphazard rather than organized), degree of mineralization (undermineralized), and matrix chemistry. In comparison, osteosclerosis in mice is delayed and once initiated occurs at a more leisurely pace. These remarkable changes in mice do not occur until the dose of estrogen surpasses physiological meaningful levels by more than 100-fold. Under these conditions, Samuels and colleagues report a continuous extension of the existing trabecular structure. Others have also described a gradual reduction of the volume of the marrow cavity through addition onto the endocortical bone surface.11 Estrogen treatment results in decreased hematopoiesis in both mice and birds, but the severe marrow cells toxicity seen in mice has not been reported in birds. The well established antiresorptive activity of estrogen was considered to be a major factor in the original descriptions of estrogen-induced osteosclerosis in growing mice.6, 11 This mechanism has been largely discounted by contemporary authors. However, the cartilage remnants (see Fig. 6 of the current paper by Samuels et al.) residing well within the marrow cavity of the estrogen-treated mice provides definitive proof that an inhibition of the rate of replacement of calcified growth plate cartilage by bone contributes to the cancellous osteosclerosis in estrogen-treated mice. This mechanism does not contribute to medullary bone formation in adult birds. Medullary bone formation in birds is prevented by simultaneous administration of 5-flurouracil, a potent inhibitor of cell proliferation.10 Estrogen receptors have been identified in the putative target cells, and the response to estrogen is prevented by simultaneous administration of the antiestrogen tamoxifen.12 The target cells and role of estrogen in mediating the induction of osteosclerosis in mice has not been determined, but the levels of estrogen required to induce bone formation greatly exceeds those necessary to promote full development of the reproductive tract, suggesting that the response is unrelated to the hormonal activity of estrogen. Administration of −f1500 μg/kg of 17β-estradiol was insufficient to induce endocortical bone formation in mice.4 This high dose contrasts to the 5 μg/kg dose which is sufficient to saturate the estrogen receptor ligand binding sites in estrogen target tissues.13 Thus, there is no evidence that estrogen-induced osteosclerosis in mice is relevent to any physiological action of the hormone. The many differences suggest that the mechanism for estrogen-induced osteosclerosis in mice may differ fundamentally from medullary bone formation in birds. Estrogen also results in a form of osteosclerosis in growing rats, but the mechanism of its formation differs, at least in part, from that described in mice.6, 7 Osteosclerosis in rats is limited to cancellous bone and the response occurs at much lower doses of the hormone than in the mouse, including physiological doses. Unlike mice and birds, the osteosclerotic response in rats decreases with age as the growth rate decreases. The resulting increase in cancellous bone volume in estrogen-treated rats is due to the same mechanisms that result in the gender differences in cancellous bone volume, namely decreased resorption of calcified cartilage during vascular invasion of the growth plate, decreased resorption of cartilage and bone matrix in the primary spongiosa, and decreased resorption of bone matrix in the secondary spongiosa.8 The net result of these three processes is an increase in cancellous bone volume. There is compelling evidence for the antiresorption-based mechanisms described in the previous paragraph: estrogen treatment in rats prolongs retention of cartilage remnants in trabeculae; this phenomenon is accompanied by decreases in the number of chondroclasts and osteoclasts; osteosclerosis is not observed in rats at skeletal sites with fused epiphysis; and estrogen-induced osteosclerosis is strikingly similar to osteopetrosis, a condition resulting from defective bone resorption. Additionally, administration to growing rats of potent inhibitors of bone formation such as bisphosphonates and glucocorticoids results in similar increases in bone volume. The putative anabolic action of estrogen on bone formation in other mammalian species cited by Samuels and co-workers is not representative of the literature. The vast number of peer-reviewed published studies in laboratory animals as well as in humans have failed to identify any evidence for an increase in the rate of bone formation following estrogen treatment, regardless of the duration of treatment, dose of estrogen, or endogenous circulating levels of the hormone. Attempts to confirm reports of estrogen-induced stimulation of bone formation in the rat were unsuccessful.14 If physiological levels of estrogen support normal bone formation, as has been proposed by some investigators,15 then an acute decrease in circulating levels of the hormone should result in transient decreases in bone formation. Several laboratories have performed careful time-course studies following OVX and have not detected this predicted decrease in formation by dynamic bone histomorphometry.16-18 Similarly, no decrease in serum osteocalcin, a marker of total bone formation, was detected.19 Also, according to this model, treatment with estrogen should result in a transient rise in bone formation. However, no rise in osteoblast perimeter, fluorochrome label perimeter, calculated bone formation rate, and serum osteocalcin is reported in estrogen-treated rats.14, 18 Other studies have reported that estrogen results in rapid ultrastructural changes consistent with decreased osteoblast activity.20 This interpretation of a direct inhibitory effect of estrogen on boneformation is supported by measurement of decreased collagen synthesis14, 21 and decreased expression of steady-state mRNA levels for bone matrix proteins.22 The changes in cancellous bone mass following estrogen treatment argue against a stimulatory effect of the hormone on bone formation. Estrogen increases cancellous bone volume in rapidly growing bones, but other antiresorbing agents have identical effects.23 In contrast to anabolic agents, such as parathyroid hormone, estrogen has minimal effects on bone volume following cessation of growth. Taken as a whole there is no compelling evidence that estrogen-induced increases in the rate of bone formation contribute to bone mass. The apparent exception, estrogen-induced osteosclerosis in mice, may be more accurately viewed as a nonspecific, possibly toxic response to estrogen excess. The view of Samuels and colleagues that high-dose estrogen is nontoxic is not supported by the literature. Indeed, the doses used to induce osteosclerosis is commonly used as a model for marrow cell toxicity. In the dog, for example, elevating estrogen levels using doses that are lower than the doses given to mice to induce osteosclerosis lead to anemia, hemorrhage, and death.24 Estrogen toxicity is species specific and rodents are remarkedly insensitive to estrogen toxicity, possibly due to their ability to transfer hematopoiesis to spleen and liver. Nevertheless, the subcutaneous doses used in the mice to induce osteosclerosis (30 mg/kg) approach the LD50 of estrogen administered intraperitoneally and are far from benign.24 A small sample of the side effects reported in the mouse include: bone marrow hypocellularity and stem cell myelotoxicity, reduction in lymphocyte production, induction of CD5 + B cells to produce autoantibodies, reduced T-cell development in the thymus, inhibition of bone marrow hematopoiesis, and induction of liver hematopoiesis. These changes indicate that high-dose estrogen is highly toxic to bone marrow of the mouse. Since marrow ablation is known to result in endocortical and cancellous bone formation, it is possible that estrogen-induced osteosclerosis occurs secondarily to bone marrow toxicity.25 This possibility is supported by the knowledge that osteosclerosis is associated in a variety of species, including humans, with disturbances of the hematopoietic system.26 The results of Samuels et al. and others lead to questions regarding the relevance of the mouse as an animal model for the pathogenesis of postmenopausal bone loss. On the one hand, both the availability of numerous inbred mouse strains with differences in peak bone mass and susceptibility to bone loss, and the extraordinary power of transgenic technology make the mouse a highly desirable animal model. On the other hand, the possible remarkable difference in the respective responses of the mouse and human skeleton to estrogen raises questions as to the usefulness of the OVX mouse as a model for postmenopausal bone loss. There have been limited studies in OVX mice, but they have generally reported cancellous osteopenia with the bone loss resulting from a net increase in bone resorption. This is consistent with the mechanism of bone loss in estrogen-depleted women. Importantly, near physiological estrogen replacement prevents the bone loss without initiating intramedullary bone formation. These important findings suggest that judiciously used, the mouse is a valuable laboratory animal model for investigation of the pathogenesis of ovarian hormone-induced bone loss as well as the physiological actions of estrogen. Doses of estrogen that greatly exceed the threshold level necessary to saturate estrogen receptor ligand binding sites are required to induce bone formation in the mouse. The resulting osteosclerotic response is not related to any known physiological function of the hormone and is closely associated with bone marrow toxicity. Estrogen-induced endocortical bone formation in mice has not been described in other mammalian species and is only superficially similar to medullary bone formation in birds, suggesting fundamental differences in the respective mechanisms leading to bone formation in birds and mice. De novo formation of cancellous bone is also unique to mice among mammalian species. Thus, it is highly debatable whether the osteosclerotic mouse model will provide any additional insight into either the pathogenesis of osteoporosis or the mechanism of action of estrogen on bone cells. However, the precise mechanism that leads to the osteosclerotic response is of great interest because it is likely that the proposed nonspecific toxic effects of estrogen excess lead to release of highly potent marrow cell–derived factors which promote osteoblast differentiation. National Institutes of Health grant AR 41418 has supported studies on the mechanism of action of estrogen on the skeleton.