End-to-end annealing of microtubules in vitro.

Mixtures of pre-formed microtubules, polymerized from chicken erythrocyte and brain tubulin, rapidly anneal end-to-end in vitro in standard microtubule assembly buffer. The erythrocyte tubulin segments in annealed heteropolymers can be distinguished by an immunoelectron microscopic assay that uses an antibody specific for chicken erythrocyte beta-tubulin. An annealing process is consistent with the following observations: (a) Microtubule number decreases while the polymer mass remains constant. (b) As the total number of microtubules declines, the number of heteropolymers, and the number of segments contained in each heteropolymer, increases. (c) The size of the segments determined after annealing and antibody labeling is the same as the original microtubule polymers. (d) Points of discontinuity in the annealing heteropolymers can be observed directly by electron microscopy, and correspond to type-specific polymer domains. The junctions probably represent initial contact points during the annealing process. Microtubule annealing occurs rapidly in vitro and may be significant for determining properties of microtubule dynamics in vivo.

Immunofluorescence examination of beta tubulin expression and marginal band formation in developing chicken erythroblasts.

Chicken erythrocyte beta tubulin, a tubulin variant with unique biochemical and assembly properties, is found to be specifically contained in two chicken blood cell types--erythrocytes and thrombocytes. The beta tubulin variant is absent or present in low amounts in a variety of white blood cell types and other body tissues, as determined by immunofluorescence microscopy and a semi-quantitative immunoblotting procedure. During differentiation in the marrow the beta tubulin variant appears suddenly in mid-stage erythroblasts at the onset of hemoglobin synthesis, and forming marginal bands are seen in all subsequent polychromatophilic erythroblast stages. The developmental sequence of events in marginal band formation entails microtubule nucleation at the centrosome, followed by microtubule elongation, consolidation of loose parallel microtubules into a compact bundle, and microtubule association with the cell membrane.

Direct observation of microtubule treadmilling by electron microscopy.

Using an immunoelectron microscopic procedure, we directly observed the concurrent addition and loss of chicken brain tubulin subunits from the opposite ends of microtubules containing erythrocyte tubulin domains. The polarity of growth of the brain tubulin on the ends of erythrocyte microtubules was determined to be similar to growth off the ends of Chlamydomonas axonemes. The flux rate for brain tubulin subunits in vitro was low, approximately 0.9 micron/h. Tubulin subunit flux did not continue through the entire microtubule as expected, but ceased when erythrocyte tubulin domains became exposed, resulting in a metastable configuration that persisted for at least several hours. We attribute this to differences in the critical concentrations of erythrocyte and brain tubulin. The exchange of tubulin subunits into the walls of preformed microtubules other than at their ends was also determined to be insignificant, the exchange rate being less than the sensitivity of the assay, or less than 0.2%/h.

The relative contributions of polymer annealing and subunit exchange to microtubule dynamics in vitro.

Microtubules that are free of microtubule-associated protein undergo dynamic changes at steady state, becoming longer but fewer in number with time through a process which was previously assumed to be based entirely on mechanisms of subunit exchange at polymer ends. However, we recently demonstrated that brain and erythrocyte microtubules are capable of joining end-to-end and suggested that polymer annealing may also affect the dynamic behavior of microtubules in vitro (Rothwell, S. W., W. A. Grasser, and D. B. Murphy, 1986, J. Cell Biol. 102:619-627). In the present study, we first show that annealing is a general property of cytoplasmic microtubules and is not a specialized characteristic of erythrocyte microtubules by documenting annealing between tryosinolated and detyrosinolated brain microtubules. We then examine the contributions of polymer annealing and subunit exchange to microtubule dynamics by analyzing the composition and length of individual polymers in a mixture of brain and erythrocyte microtubules by immunoelectron microscopy. In concentrated preparations of short-length microtubules at polymer-mass steady state, annealing was observed to be the principal factor responsible for the increase in polymer length, whereas annealing and subunit exchange contributed about equally to the reduction in microtubule number.

Direct demonstration of actin filament annealing in vitro.

Direct electron microscopic examination confirms that short actin filaments rapidly anneal end-to-end in vitro, leading over time to an increase in filament length at steady state. During annealing of mixtures of native unlabeled filaments and glutaraldehyde-fixed filaments labeled with myosin subfragment-1, the structural polarity within heteropolymers is conserved absolutely. Annealing does not appear to require either ATP hydrolysis or the presence of exogenous actin monomers, suggesting that joining occurs through the direct association of filament ends. During recovery from sonication the initial rate of annealing is consistent with a second-order reaction involving the collision of two filament ends with an apparent annealing rate constant of 10(7) M-1s-1. This rapid phase lasts less than 10 s and is followed by a slow phase lasting minutes to hours. Annealing is calculated to contribute minimally to filament elongation during the initial stages of self-assembly. However, the rapid rate of annealing of sonicated fixed filaments observed in vitro suggests that it may be an efficient mechanism for repairing breaks in filaments and that annealing together with polymer-severing mechanisms may contribute significantly to the dynamics and function of actin filaments in vivo.

Echistatin is a potent inhibitor of bone resorption in culture.

The venom protein, s-echistatin, originally derived from the saw-scaled viper Echis carinatus, was found to be a potent inhibitor of bone resorption by isolated osteoclasts. This Arg24-Gly25-Asp26-(RGD)-containing protein inhibited the excavation of bone slices by rat osteoclasts (IC50 = 0.1 nM). It also inhibited the release of [3H]proline from labeled bone particles by chicken osteoclasts (IC50 = 100 nM). By comparison, the tetrapeptide Arg-Gly-Asp-Ser (RGDS) inhibited resorption by rat or chicken osteoclasts with an IC50 of 0.1 mM while ala24-echistatin was inactive. Video microscopy showed that rat osteoclast attachment to substrate was more sensitive to s-echistatin than was the attachment of mononuclear cells or chicken osteoclasts. The difference in sensitivity of rat and chicken osteoclasts to s-echistatin may be due to differences between receptors on rat and chicken osteoclasts for s-echistatin. Antibody localization of echistatin on these cells showed much greater echistatin binding to rat osteoclasts than to chicken osteoclasts. Laser scanning confocal microscopy after immunohistochemical staining showed that s-echistatin binds to osteoclasts, that s-echistatin receptors are most abundant at the osteoclast/glass interface, and that s-echistatin colocalizes with vinculin. Confocal interference reflection microscopy of osteoclasts incubated with s-echistatin, demonstrated colocalization of s-echistatin with the outer edges of clusters of grey contacts at the tips of some lamellipodia. Identification of the echistatin receptor as an integrin was confirmed by colocalization of echistatin fluorescence with staining for an alpha-like subunit. Attachment of bone particles labeled with [3H]proline to chicken osteoclasts confirmed that the mechanism of action of echistatin was to inhibit osteoclast binding to bone presumably by disrupting adhesion structures. These data demonstrate that osteoclasts bind to bone via an RGD-sequence as an obligatory step in bone resorption, that this RGD-binding integrin is at adhesion structures, and that it colocalizes with vinculin and has an alpha-like subunit.

Copolymerization of two distinct tubulin isotypes during microtubule assembly in vitro.

Cells contain multiple tubulin isotypes that are the products of different genes and posttranslational modifications. It has been proposed that tubulin isotypes become segregated into different classes of microtubules each adapted to specific activities and functions. To determine if mixtures of tubulin isotypes segregate into different classes of polymers in vitro, we used immunoelectron microscopy to examine the composition of microtubule copolymers that assembled from mixtures of purified tubulin subunits from chicken brain and erythrocytes, each of which has been shown to exhibit distinct assembly properties in vitro. We observed that (a) the two isotypes coassemble rapidly and efficiently despite the fact that each isotype exhibits its own unique biochemical and assembly properties; (b) at low monomer concentrations the ratio of tubulin isotypes changes along the lengths of elongating copolymers resulting in gradients in immuno-gold labeling; (c) two distinct classes of copolymers each containing a distinct ratio of isotypes assemble simultaneously in the same subunit mixture; and (d) subunits and polymers of different isotypes associate nearly equally well with each other, there being only a slight bias favoring interactions among subunits and polymers of the same isotype. The observations agree with previous studies on the homogeneous distribution of multiple isotypes within cells and suggest that if segregation of isotypes does occur in vivo, it is most likely directed by cell-specific microtubule-associated proteins (MAPs) or specialized intracellular conditions.

Effects of gestational age on prostaglandin EP receptor expression and functional involvement during in vitro contraction of the guinea pig uterus.

Prostaglandin E(2) (PGE(2)) exerts diverse biological effects through four G-protein-coupled cell surface receptor subtypes, EP1-4. This study's objective was to characterize EP1-4 receptor mRNA expression within pregnant guinea pig myometrium during early implantation stage (gestation day [GD] 6) and late stage gestation (GD 50) and evaluate in vitro contractile activity of receptor subtype selective agonists. Using RT-PCR, qualitative gene expression patterns of EP2, EP3, and EP4 mRNA were detected in the myometrium and remained unchanged between the gestational ages. EP1 mRNA remained undetected in pregnant tissue. In vitro contractile activity was evaluated in GD 6 and GD 50 myometrium using vehicle and EP agonists PGE(2), 17-phenyl trinor PGE(2), sulprostone, misoprostol, and CP-533,536. All spasmogens in pregnant myometrium were EP1/EP3 selective agonists, though likely acting via EP3 receptors in this test model. CP-533,536--a highly selective EP2 receptor agonist--and the vehicle failed to induce myometrial contraction at both gestational ages.

Presence of messenger ribonucleic acid encoding osteocalcin, a marker of bone turnover, in bone marrow megakaryocytes and peripheral blood platelets.

Osteocalcin (Oc), an abundant gamma-carboxylated protein (mol wt, 5800) of bone matrix, is used as a serum marker of bone turnover because it is considered to be uniquely synthesized by the osteoblast. Our finding of Oc messenger RNA (mRNA) in rat tibial diaphyseal marrow led us to investigate the cellular origins of Oc mRNA in peripheral blood and bone marrow. In anticoagulated blood, Oc mRNA was detected in total RNA prepared from buffy coat cells (BCC). Fractionation of rat and human blood showed that platelets contain Oc mRNA identical to that found in bone cells. In rat bone marrow, Oc mRNA is highly enriched in the platelet-producing megakaryocyte population. Depending upon the RIA used, immunoreactive Oc was either undetectable or present at very low levels in platelets and megakaryocytes, suggesting that synthesis of Oc by these cells may be under strong translational regulation. In addition, Oc levels were higher in serum vs. plasma obtained from the same blood, suggesting that Oc may be released by platelets during blood clotting. Interestingly, the magnitude of this difference was greater in female rats. Injection of 1,25-dihydroxyvitamin D3 dose-dependently increased plasma Oc, but did not cause correlative changes in steady state levels of Oc mRNA in BCC. During rat growth, plasma Oc was maximal, whereas Oc mRNA levels in BBC were low. This relationship was reversed during aging. A correlation between Oc mRNA levels in BCC and rat age suggests a developmental regulation of Oc mRNA levels in platelets. These data indicate that Oc mRNA is not restricted to cells on mineralizing surfaces, but is also found in megakaryocytes and peripheral blood platelets, which possibly contribute to the Oc levels in blood and the regulation of bone turnover.

Expression of the parathyroid hormone-related protein gene in the avian oviduct: potential role as a local modulator of vascular smooth muscle tension and shell gland motility during the egg-laying cycle.

The phylogenetic conservation of the primary structure of PTH-related protein (PTHrP) supports an important, yet undetermined, role(s) for this molecule in the biology of birds and mammals. As an initial step toward understanding the function of PTHrP in birds, we investigated the expression of PTHrP mRNA in tissues of the egg-laying hen. This analysis revealed that PTHrP mRNA is expressed at various levels in lung, brain, heart, and tissues of the digestive tract, including the proventriculus (secretory stomach), gizzard, and small intestine. In the oviduct tissues of adult birds, PTHrP mRNA was detected in the isthmus (membrane-secreting) and shell gland (calcium-secreting) portions, but not in magnum (albumin secreting) tissue. During oviduct development, high levels of PTHrP mRNA present in the oviducts of the 12-week-old bird suggest a role for PTHrP in oviduct development. Interestingly, as the oviduct matures, relatively high levels of PTHrP mRNA segregate with the distal tissues that ultimately differentiate into the isthmus and shell gland (uterus). To address a possible role for PTHrP in the differentiated function of the shell gland, we followed the expression of PTHrP in the shell gland at different times in the laying cycle and found levels of PTHrP to transiently increase as the egg moves through the oviduct, gradually returning to basal levels in the 15-h calcification period. We localized the cycle-associated fluctuations in PTHrP mRNA levels to the shell gland serosa and smooth muscle layer. Immunoreactive PTHrP was localized to the serosal membrane as well as the smooth muscle layer of serosal arterioles, suggesting that PTHrP may modulate vascular smooth muscle activity. In support of this hypothesis, synthetic chicken PTHrP (1-34)NH2 was found to relax the resting tension of isolated shell gland blood vessels in a dose-dependent manner. Together, these data indicate that the expression of the PTHrP gene in the avian oviduct is both temporally and spatially regulated during the egg-laying cycle and that PTHrP may function as an autocrine/paracrine modulator of shell gland smooth muscle activity of both ductal and vascular origins. The vasorelaxant property of N-terminal fragments of PTHrP supports a role for this molecule in the temporal increase in blood flow to the shell gland during egg calcification.

Effects of CP-336,156, a new, nonsteroidal estrogen agonist/antagonist, on bone, serum cholesterol, uterus and body composition in rat models.

We have discovered a new, nonsteroidal, potent estrogen agonist/antagonist, CP-336,156. CP-336,156 binds selectively and with high affinity to the human estrogen receptor-alpha with a half-inhibition concentration of 1.5 nM, which is similar to that seen with estradiol (4.8 nM). When given orally to immature (3-week-old) female Sprague-Dawley rats for 3 days at doses of 0.1, 1.0, 10, or 100 microg/kg x day, unlike 17alpha-ethynyl estradiol, CP-336,156 had no effect on uterine wet or dry weight. Similarly, no uterine hypertrophy was observed in aged (17-month-old) female rats treated (p.o.) with CP-336,156 at 10 or 100 microg/kg x day for 28 days. We also found that CP-336,156 decreased total serum cholesterol and fat body mass and had no effect on lean body mass in these aged female rats. In 5-month-old ovariectomized (OVX) Sprague-Dawley female rats, CP-336,156 completely prevented OVX-induced increases in body weight gain, total serum cholesterol, and serum osteocalcin at doses between 10 and 1000 microg/kg x day after 4 weeks. At these doses, CP-336,156 completely prevented OVX-induced bone loss and inhibited the increased bone turnover associated with estrogen deficiency in lumbar vertebrae, proximal tibiae, and distal femora. Similar to estrogen, CP-336,156 induced apoptosis and p53 expression with a concomitant decrease in the number of tartrate-resistant acid phosphatase-positive multinuclear cells in rat bone marrow cell cultures in vitro, suggesting that the induction of apoptosis may be a mechanism for the estrogenic activities of CP-336,156 in bone. In summary, CP-336,156 is a new, orally active, nonsteroidal, potent estrogen agonist/antagonist that has similar effects in bone as estradiol but without the uterine-stimulating effects associated with estradiol in rats.

Prostaglandin E2 receptor subtype EP-2 is not involved in the induction of non-pregnant guinea pig uterine contractions associated with terminal pregnancy.

Prostaglandin E2 (PGE2) exerts its biological effects through 4 different receptor subtypes, EP-1, EP-2, EP-3, and EP-4. Recently we have demonstrated the importance of the prostaglandin E2 receptor subtype EP-2 in the healing of bone defects and fractures. This discovery led to the identification of CP-533,536, an EP-2 selective agonist, a promising therapeutic alternative for the enhancement of bone healing and the treatment of fractures (J Bone Miner Res 18 (2003) 2033). PGE2 has a myriad of effects throughout the body including the induction of uterine contractions, which results in termination of pregnancies. Our objective in this study was to determine the role of the EP-2 receptor and specifically that of CP-533,536, an EP-2 specific agonist, to induce uterine contractions and terminate pregnancy in guinea pigs, an animal model of human pregnancy. Preliminary experiments confirmed earlier reports that the guinea pig uterus was more sensitive than that of the rat. The guinea pig uterus contains the four PGE2 receptor subtypes, and ex vivo treatment of the uterus with PGE2 as expected causes profound uterine contractions. However, using receptor selective prostaglandin agonists including CP-533,536 we showed that the EP-1 and 3 receptors not the EP-2 receptor is responsible for the induction of uterine contractions of PGE2. Further, CP-533,536 did not antagonize the ability of PGE2 to induce uterine contractions in this model.

Myostatin expression in age and denervation-induced skeletal muscle atrophy.

Myostatin is hypothesized to regulate skeletal muscle mass and to be a potential target for therapeutic intervention in sarcopenia. To clarify whether myostatin is invariably associated with sarcopenia, this study examined the levels of expression of myostatin mRNA and protein in Sprague Dawley rats during aging- and denervation-induced sarcopenia. The level of myostatin mRNA in the gastrocnemius decreased progressively with age being 9, 34 and 56% lower at 6, 12 and 27 months, respectively, compared with mRNA levels at 1.5 months. In contrast, two low molecular mass isoforms of myostatin protein identified by Western blotting increased progressively with age. With denervation, myostatin mRNA was 31% higher on day 1 but by 14 days after sciatic neurectomy when the muscle had atrophied 50%, myostatin expression decreased 34% relative to the sham operated limb. Western analysis of the denervated gastrocnemius showed that myostatin protein levels varied in parallel with mRNA. These disparate patterns of expression of myostatin during age- and denervation-induced atrophy suggest that the regulation of myostatin is complex and variable depending on whether the atrophy is slowly or rapidly progressive.

Regulation of osteoclast differentiation by statins.

HMG-CoA reductase inhibitor (statin) treatment is frontline therapy for lowering plasma cholesterol levels in patients with hyperlipidemia. In a few case studies, analysis of clinical data has revealed a decreased risk of fracture in patients on statin therapy. However, this reduction in the incidence of fracture is not always observed nor is it supported by an increase in bone density, which further complicates our understanding of the role of statins in bone metabolism. Thus, the precise role of statins in bone metabolism remains poorly understood. In this study, we examined the effect of statin treatment on osteoclastogenesis. Treatment with lovastatin resulted in a significant, dose-dependent decrease in the numbers of differentiated osteoclasts and decreased cholesterol biosynthesis activity with an EC(50) similar to that observed in freshly isolated rat or cultured human liver cells. Studies assessing the role of mevalonate metabolites in the development of the osteoclasts demonstrated that geranylgeraniol, but not squalene or farnesol was important for the development and differentiation of osteoclasts, implicating protein geranylgeranylation rather than protein farnesylation as a key factor in the osteoclast differentiation process. In conclusion, our data indicate that lovastatin inhibits osteoclast development through inhibition of geranylgeranylation of key prenylated proteins and that the bone effects of statins are at least partially due to their effects on osteoclast numbers.

BMP-6 exerts its osteoinductive effect through activation of IGF-I and EGF pathways.

We have recently shown that human recombinant BMP-6 (rhBMP-6), given systematically, can restore bone in animal models of osteoporosis. To further elucidate the underlying mechanisms of new bone formation following systemic application of BMPs, we conducted gene expression profiling experiments using bone samples of oophrectomised mice treated with BMP-6. Gene set enrichment analysis revealed enrichment of insulin-like growth factor-I and epidermal growth factor related pathways in animals treated with BMP-6. Significant upregulation of IGF-I and EGF expression in bones of BMP-6 treated mice was confirmed by quantitative PCR. To develop an in vitro model for evaluation of the effects of BMP-6 on cells of human origin, we cultured primary human osteoblasts. Treatment with rhBMP-6 accelerated cell differentiation as indicated by the formation of mineralised nodules by day 18 of culture versus 28-30 days in vehicle treated cultures. In addition, alkaline phosphatase gene expression and activity were dramatically increased upon BMP-6 treatment. Expression of IGF-I and EGF was upregulated in human osteoblast cells treated with BMP-6. These results collectively indicate that BMP-6 exerts its osteoinductive effect, at least in part, through IGF-I and EGF pathways, which can be observed both in a murine model of osteopenia and in human osteoblasts.

Role of EP2 and EP4 receptor-selective agonists of prostaglandin E(2) in acute and chronic kidney failure.

We tested the efficacy of three selective agonists of prostaglandin E(2) (PGE(2)) receptor, EP2 (CP-536,745-01), EP2/4 (CP-043,305-02), and EP4 (CP-044,519-02), in two models of acute and chronic kidney failure. In the nephrotoxic mercury chloride (HgCl(2)) rat model of acute kidney failure systemically administered EP4 agonist reduced the serum creatinine values and increased the survival rate. Although the EP2 or the EP2/4 agonist did not change the serum creatinine values, the EP2 receptor agonist increased the survival rate. Histological evaluation of kidneys from EP4-treated rats indicated less proximal tubular necrosis and less apoptotic cells. In a rat model of chronic renal failure, the three receptor agonists decreased the serum creatinine and increased the glomerular filtration rate at 9 weeks following therapy. Kidneys treated with the EP4 agonist had less glomerular sclerosis, better preservation of proximal and distal tubules and blood vessels, increased convoluted epithelium proliferation and less apoptotic cells. Nephrectomy had no influence on the expression of the EP4 receptor, whereas EP2 receptor expression was reduced by 50% and then corrected following treatment with EP2 and EP2/4 receptor agonists. These findings suggest that PGE(2) has an important role in acute kidney failure via the EP4 receptor, whereas in chronic kidney failure both EP2 and EP4 receptors are equally important in preserving the progression of chronic kidney failure. Thus, agonism of EP2 and EP4 receptors may provide a basis for treating acute and chronic kidney failure.

Intermediate filaments in the cytoskeletons of fish chromatophores.

When fish pigment cells (melanophores, erythrophores) are lysed by a modified Kleinschmidt method on a buffer-air interface and examined by electron microscopy, large numbers of intermediate filaments are observed. The intermediate filament networks are distinct from actin and tubulin, and entrap the pigment as determined by stereo viewing of freeze-dried rotary-shadowed specimens. During lysis, under conditions that do not preserve actin filaments or microtubules, the area covered by dispersed pigment granules reaches a maximum size and remains stable for many minutes, suggesting that intermediate filaments are responsible for holding the pigment in position and preventing further cytoplasmic dispersion. These observations demonstrate that fish pigment cells contain large numbers of intermediate filaments and suggest that they may be important for coordinating pigment granule movement.

Parathyroid hormone-related protein in the rat urinary bladder: a smooth muscle relaxant produced locally in response to mechanical stretch.

Parathyroid hormone-related protein (PTHrP) gene expression in the pregnant rat uterus has been shown to be dependent on occupancy of the uterus by the fetus. To further test the hypothesis that the synthesis of PTHrP in smooth muscle tissue is regulated by mechanical stretch, we conducted experiments using the rat urinary bladder as a model of an expansible hollow organ. The results indicate that PTHrP mRNA levels do change in response to the stretch of the bladder wall. Under normal conditions PTHrP mRNA levels in the bladder correlated with the urine volume-namely, the extent of bladder distension. When bladders were maintained empty in vivo, PTHrP mRNA levels decreased gradually. Conversely, when bladders were distended by the accumulation of urine, levels of PTHrP mRNA increased dramatically with time. When distension was limited to one-half of the bladder, the increase in PTHrP mRNA was observed only in the distended portion. Histochemical studies performed on distended bladder tissue indicated the presence of PTHrP immunoreactivity in smooth muscle cells. Isolated organ bath studies were used to examine the possible physiological role of PTHrP in smooth muscle tonicity. In vitro responsiveness of bladder muscle strips to exogenous PTHrP was dependent on the in vivo condition of the bladder. In muscle strips obtained from bladders kept empty in vivo, PTHrP-(1-34)-NH2 relaxed carbachol-induced contraction in a dose-dependent manner but failed to relax the contraction in muscle strips from distended bladders that had high endogenous PTHrP expression. These results and the previous findings in the rat uterus suggest a physiological role of PTHrP in bladder smooth muscle function.

Common mechanism for the estrogen agonist and antagonist activities of droloxifene.

The incidence of postmenopausal osteoporosis is increasing as the population ages. Even though estrogen replacement therapy has proven beneficial in reducing the number of skeletal fractures, the known risks and associated side-effects of estrogen replacement therapy make compliance poor. Recent research has focused on the development of tissue specific estrogen agonist/antagonists such as droloxifene which can prevent estrogen deficiency-induced bone loss without causing uterine hypertrophy. Furthermore, droloxifene acts as a full estrogen antagonist on breast tissue and is being evaluated for treatment of advanced breast cancer. In this report we propose a common mechanism of action for droloxifene that underlies its estrogen agonist and antagonist effects in different tissues. Droloxifene and estrogen, which have identical effects on bone in vivo, both induced p53 expression and apoptosis in cells of in vitro rat bone marrow cultures resulting in a decrease in the number of bone-resorbing osteoclasts. Droloxifene is growth inhibitory in MCF-7 human breast cancer cells and therefore acts as an antagonist, whereas estrogen is mitogenic to these cells and acts as an agonist. Droloxifene, but not estrogen, induced p53 expression and apoptosis in MCF-7 cells. These results indicate that the induction of apoptosis by droloxifene may be the common mechanism for both its estrogen agonist effects in bone and its antagonist effects in breast tissue.