Bazedoxifene Acetate: A Selective Estrogen Receptor Modulator with Improved Selectivity
http://www.100md.com
《内分泌学杂志》
Wyeth Research, Women’s Health Research Institute, Collegeville, Pennsylvania 19426
Abstract
We assessed the preclinical characteristics of a novel, stringently screened selective estrogen receptor modulator, bazedoxifene acetate, including its ability to bind to and activate estrogen receptors and promote increased bone mineral density and bone strength in rats, and the effects impacting the uterine endometrium, breast cancer cell proliferation, and central nervous system-associated vasomotor responses in an animal model. Bazedoxifene bound to estrogen receptor- with an IC50 of 26 nM, an affinity similar to that of raloxifene. Bazedoxifene did not stimulate proliferation of MCF-7 cells but did inhibit 17-estradiol-induced proliferation with an IC50 of 0.19 nM. In an immature rat uterine model, bazedoxifene (0.5 and 5.0 mg/kg) was associated with less increase in uterine wet weight than either ethinyl estradiol (10 μg/kg) or raloxifene (0.5 and 5.0 mg/kg). Histological analysis revealed that coadministration of bazedoxifene also appeared to reduce raloxifene-stimulated endometrial luminal epithelial cell and myometrial cell hypertrophy. In ovariectomized rats, bazedoxifene was associated with significant increases in bone mineral density at 6 wk, compared with control, and better compressive strength of bone samples from the L4 vertebrae, compared with samples from ovariectomized animals. In the morphine-addicted rat model of vasomotor activity, bone-sparing doses of bazedoxifene alone were not associated with 17-estradiol inhibition of increased vasomotor activity. Bazedoxifene acetate represents a promising new treatment for osteoporosis, with a potential for less uterine and vasomotor effects than selective estrogen receptor modulators currently used in clinical practice. Controlled clinical trial data will be needed to confirm these effects.
Introduction
ESTROGENS COMPRISE A RELATIVELY large class of compounds with substantial structural diversity with at least one common characteristic: they bind to the estrogen receptors (ERs) to transduce their activity. Over time different names have been assigned to members of this family to, more than anything, describe their functional role. The classical members include the steroidal estrogens, of which there are many, with estrone 17-estradiol, and estriol representing three of the most abundant and active in humans. There are a fair number of nonsteroidal estrogens, with diethylstilbestrol being the most often cited and used. The phytoestrogens are represented by a rather sizable number of compounds, again with a number of different structural backbones (1, 2, 3).
Many years ago, when it was hypothesized that estrogens may be implicated in breast cancer etiology, an effort to generate compounds that would antagonize an ER agonist like 17-estradiol was undertaken. Although not the first, tamoxifen, a triphenylethylene, proved to be a competitor for estrogen agonist binding to the ER and, indeed, did inhibit breast cancer cell proliferation (4, 5). Due to the presence of some undesirable side effects (6), efforts to improve on tamoxifen resulted in a group of compounds, which collectively became known simply as antiestrogens. Conceptually, a group of compounds now existed that would block physiologic responses to ER agonists like estrone, 17-estradiol, and estriol by competitively binding to the ER. If the antiestrogens were occupying the ER, no agonist activity would be imparted.
However, in the late 1980s, work with tamoxifen given to rats demonstrated that tamoxifen actually revealed mixed functional activity (7). That is, it acted as both an ER agonist and antagonist to varying degrees correlated to the physiologic end point under investigation. Not too long after this initial finding, other antiestrogens were also shown to display similar mixed functional activities, although not necessarily identical in character to one another. For example, raloxifene (RAL), originally created to function as a treatment for breast cancer (8), like tamoxifen, inhibited breast cancer cell proliferation yet, unlike tamoxifen, did not demonstrate the same level of uterine endometrial stimulation (9). Based on these data, the term selective ER modulator (SERM) was coined to describe these types of ligands (10), which display mixed functional estrogenic activities.
Irrespective of the descriptor used to categorize these compounds, they transduce their information via the ERs and all exhibit varying degrees of estrogen-like activity. If this group of diverse structures all binds to the ERs with relatively high affinity, how can the differences in gene activation and physiological outcomes be so different Before crystallization of the ligand binding domains of ER and ER were available, it was demonstrated by proteolytic digestion that the bound ligand affected the digestion pattern of the ERs, which was easily discerned on sodium dodecyl sulfate gels (11). These data supported the concept that the ER structural conformation varies according to whether it is bound and the effects were somewhat ligand specific. This was convincingly confirmed by x-ray crystallography of the ERs and cocrystallization with other peptides, specifically a small peptide fragment of glucocorticoid receptor interacting protein-1 or steroid receptor coactivator (SRC)-2 (12). Certain members of the p160 coactivator group like SRC-2, when bound to the ER, enhance the receptor’s transcriptional activity on certain promoters (13). When ER is liganded to 17-estradiol, SRC-2 interacts with a specific region of the ER exposed when helix 12 is positioned over the ligand binding pocket, allowing interaction to occur in a hydrophobic groove created by helices 4, 5, and 6. However, when tamoxifen or raloxifene are substituted for 17-estradiol, helix 12 is shifted to a position that blocks access of SRC-2 (and conceptually other p160 coactivators) to the appropriate site for them to function as coregulators. Simplistically, slight alterations in the receptor conformation alter its ability to interact efficiently with a coactivator, thus potentially reducing its effectiveness in the regulation of gene transcription. Of course, the conformation of the ER will affect not only potential protein-protein interactions but also biochemical modifications such as phosphorylation critical to ER function.
The selective activity associated with any estrogen, be it 17-estradiol or ICI182780 (referred to as a pure antagonist) depends on, first and foremost, the conformation it imparts to the ER on binding and the subsequent interactions/modifications that occur. The ER presented to a particular gene’s promoter will dictate the response. It appears that one gene may be more permissive when the ER is not in what would be considered its optimal conformational status to permit efficient transcriptional enhancement vs. another promoter, which will not permit such activity, plus everything in between. These responses appear to be cell type or tissue selective in addition to promoter selective (14, 15). This type of conceptual explanation for selective activity conveniently supports the notion that each estrogen, even though working through one of two receptors, can impact a tissue, cell, or gene with distinct activity (16).
Here we describe a new selective estrogen with activity distinct from other members of the SERM family. The compound is an indole-based ER ligand stringently selected to ensure an improved profile vs. its predecessors. The compound, bazedoxifene acetate (BZA), is shown to interact with the ERs, transactivate the ER, and positively affect the skeletal and lipid profile without stimulating the uterine endometrium, causing breast cancer cell proliferation or negatively impacting the central nervous system (CNS) in preclinical models.
Materials and Methods
All chemicals and reagents were reagent grade or better and obtained from various vendors. 17-ethinyl estradiol, 17-estradiol, RAL HCl, tamoxifen, 4-OH tamoxifen, and lasofoxifene were supplied by the Wyeth compound library (Princeton, NJ) that are either purchased or generated by Wyeth Medicinal Chemistry (Collegeville, PA) and undergo analytical confirmation before cataloging/use.
Animals
Sprague Dawley rats of varying age were obtained from Taconic Farms (Germantown, NY) for all animal experiments described. Animals were housed appropriately in Wyeth’s Association for Assessment and Accreditation of Laboratory Animal Care- and U.S. Department of Agriculture-sanctioned facility, and all procedures (Institutional Animal Care and Use Committee approved) were performed in accordance with all federal, state, and local guidelines.
Ligand binding
Interaction of BZA with human ER and ER was assessed with a solid phase competitive radioligand binding assay using [3H]-17- estradiol as previously described (17).
In vitro transcription assays
CHO-HER.14 (Chinese hamster ovarian cell line), HOB03CE6 (human osteoblast cell line), and D12 (hypothalamic neuronal cell line) and HepG2–313-89-A6 cells (hepatoma cell line) were used as the cell recipients of various transcription vectors containing estrogen responsive DNA sequence elements. The modified CHO cell line contains a stable integrated human estrogen receptor sequence as does the modified HepG2 cell line. The construction and details of their production can be retrieved from Harnish et al. (18)
CHO cells were plated at 50,000 cells/well in 12-well dishes and grown at 37 C/5% CO2 in phenol red-free MEM containing 10% charcoal-treated fetal bovine serum (FBS). The HOB03Ce6 cell line was maintained at 34 C in phenol red-free DMEM/F-12 containing 10% (vol/vol) heat-inactivated FBS, 1% (vol/vol) penicillin-streptomycin, and 2 mM GlutaMAX-1 (19). D12 cells were cultured at 37 C/5% CO2 in phenol red-free medium (DMEM/F-12, GlutaMAX, penicillin-streptomycin) containing 5% charcoal stripped FBS at approximately 2 x 106 cells per 150-mm dish. Twenty-four hours later, medium was changed to one containing only 2% stripped FBS with or without test compounds. HepG2 cells were maintained at 37 C in a 5% CO2 incubator in phenol red-free DMEM, 10% heat-inactivated FBS, 1% GlutaMAX, 1% MEM nonessential amino acids, 100 U/ml penicillin, and 100ug/ml streptomycin at 2.5 x 105 cells/well in 12-well dishes.
For transcription assays using an estrogen response element (ERE) combined with a weak promoter and luciferase reporter sequence, the following adenoviral construct was used. The adenovirus (Ad-5) contained two copies of the vitellogenin A2 ERE (5'-GGTCACAGTGACC-3') linked to –110 to +10 of the thymidine kinase promoter and a luciferase gene coding sequence. This construct was inserted in place of the E1a adenoviral gene. Cells are routinely infected with approximately 200 plaque-forming units/cell added directly to the cell culture medium for 2 h. After viral infection incubation time, the media are replaced with media containing test compounds or positive controls. Details of transfections and luciferase assay conditions can be found in Bodine et al. (19)
MCF-7 cell proliferation
Samples of the human breast tumor cell line, MCF-7, were obtained from the American Type Culture Collection (Manassas, VA) and maintained in DMEM/F-12 (50:50) with 10% FBS and 1 x GlutaMAX-1. They were passaged twice weekly at a concentration of 0.1 million cells/cm2 and generally used in this assay between passages 12 and 40.
For the proliferation assay, cells were plated at 20,000 cells/well in a 24-well plate in DMEM/F12 (50:50) (phenol red-free) with 10% charcoal/dextran-treated FBS and 1 x GlutaMAX-1. After overnight incubation, the medium was aspirated and treatments in DMEM/F12 (50:50) (phenol red-free) with 2% charcoal/dextran-treated FBS and 1 x GlutaMAX-1 were added to the wells. Each plate had a vehicle (baseline proliferation) and treatments. Treatments included 10 pM 17-estradiol determined to be the EC80 for 17-estradiol and 17-estradiol in combination with six concentrations of BZA. Treatments from d 1 were renewed on d 3 and d 6 by aspirating medium from wells and replacing with fresh medium and treatments. On d 7, cells were detached from the plate using trypsin-EDTA and counted using a Multisizer II (Coulter, Miami, FL).
Uterine evaluation
The effect of compounds on compounds on uterine weight was evaluated as previously described (20). Briefly, sexually immature Sprague Dawley rats were treated once daily for 3 d and euthanized approximately 24 h after administration of the last dose. The vehicle for sc dosing was 50% dimethylsulfoxide-50% 1x Dulbecco’s PBS, and the vehicle for oral administration was 2% Tween 80 and 0.5% methylcellulose. After death, uteri were excised and weighed after trimming associated fat and expressing any luminal fluid. A segment was placed in a 10% formalin solution for histological analysis and the remainder frozen on dry ice for isolation of RNA.
Histological preparation after fixation included dehydration, paraffin embedding, slide preparation, and staining/counterstaining with hematoxylin/eosin. Veterinary pathologists scored the following five parameters (1, 2, 3, 4): epithelial hypertrophy/hyperplasia, myometrial hypertrophy, luminal distention, stromal eosinophilia, and luminal epithelial apoptosis. Endometrial luminal epithelial cell height was measured with a micrometer on a Eclipse 800 photomicroscope (Nikon, Tokyo, Japan). Manual measurements spanning from the basal lamina to the apical surface were taken from five regions around the lumen, and an average epithelial cell height was determined.
Uterine RNA was isolated by polytron homogenization in Trizol followed by alcohol precipitation, air drying, and resuspension in buffer containing 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA (pH 8.0). Complement component 3 evaluation was performed by Northern blot analysis.
Vasomotor instability (hot flush)
Ovariectomized female (60 d) rats were obtained after surgery. The surgeries were performed minimally 7 d before initiation of any experiment. Vehicle and ethinyl estradiol (0.3 mg/kg) were included in each replicate. Bazedoxifene was administered orally in a saline, Tween-80, methylcellulose vehicle. A detailed description of methodology for evaluating vasomotor instability in rats has been published (21). Briefly, compound treatment (17-estradiol, ethinyl estradiol, or bazedoxifene) is initiated, and on the third day of treatment each animal receives a morphine pellet sc. This is followed by two more pellets on the fifth day of treatment. On the eighth day, a thermistor is taped to the animal’s tail to measure tail skin temperature for 15 min (to obtain baseline temperature) followed by a sc injection of naloxone (1 mg/kg). Tail skin temperature readings continue for 1 h after naloxone injection.
Vertebral bone densitometry
In mature (2.5–3 months, 250 g) ovariectomized rats, all treatments were initiated 3 d after ovariectomy and continued for 6 wk. Five weeks after the initiation of treatment and 1 wk before the termination of study, each rat was evaluated for bone mineral density (BMD). The BMDs of the proximal tibiae (PT) and fourth lumbar vertebrae (L4) were measured in anesthetized rats using a dual-energy x-ray absorptiometer (DXA) (Eclipse XR-26, Norland Corp., Ft. Atkins, WI). The DXA measurements for each rat were performed as follows: a preliminary scan was performed at a scan speed of 50 mm/sec with a scan resolution of 1.5 mm x 1.5 mm to determine the region of interest in PT and L4. Small subject software was employed at a scan speed of 10 mm/sec with resolution of 0.5 mm x 0.5 mm for final BMD measurements. A defined area of a 1.5 cm-wide region in the PT (starting from femur-tibia junction to 0.5 cm distal) or a 1.5 cm-wide area to cover the total length of L4 was selected for analysis. The BMDs for respective sites were computed by the software as a function of the attenuation of the dual beam (46.8 and 80 KeV) x-ray generated by the source underneath the subject and the detector traveling along the defined area above the subject. The data for BMD values were expressed in grams per square centimeter.
Peripheral quantitative computed tomography (pQCT)
Volumetric total (cortical and trabecular) and trabecular densities were evaluated in anesthetized rats using a XCT-960M (pQCT; Stratec Medizintechnik, Pforzheim, Germany). The right hind limb was placed in a polycarbonate tube affixed to a sliding platform that maintained it parallel to the aperture of the pQCT. The platform was adjusted so that the distal end of the femur and the proximal end of the tibia would be in the scanning field. A two-dimensional scout view was run for a length of 10 mm and line resolution of 0.2 mm. The pQCT scan was initiated 3.4 mm distal from the proximal end of tibia. The pQCT scan was 1 mm thick, had a voxel (three-dimensional pixel) size of 0.140 mm, and consisted of 145 projections through the slice. After the pQCT scan was completed, the image was displayed on the monitor. A region of interest in the tibia was analyzed for total density and the value was reported in milligrams per cubic centimeter. The outer 55% of the bone was mathematically removed away in a concentric spiral. The density of the remaining bone was reported as trabecular density in milligrams per cubic centimeter.
Tissue collection and serum cholesterol measurement
One week after BMD evaluation, serum cholesterol evaluation was performed. Total serum cholesterol was analyzed using a Hitachi 911 autoanalyzer (Roche Diagnostics, Indianapolis, IN). The uterus from each rat was removed, fat excised, luminal fluid blotted away, and weighed. The tibiae from both limbs and fifth lumbar vertebra were also dissected free of soft tissue, removed, and processed for histomorphometric analysis and compressive strength measurements, respectively.
Tibial histology
The proximal portion of the tibiae was fixed, dehydrated, and embedded undecalcified in a methyl methacrylate mixture. Longitudinal tissue sections (5–10 μM) were prepared on a Polycut S microtome (Reichert, Nussloch, Germany). Sections were stained with Goldner’s stain and coverslipped.
Cancellous (trabecular) bone content in the PT was quantified as two-dimensional bone mineral area (B.Ar) with an image analysis system (Videometric 150; Oncor, Gaithersburg, MD). The field of view from a microscope image of the slide was coupled to a microcomputer through a video camera (TMC-74 color RGB; Plunix, San Jose, CA). The mineralized tissue content of cancellous bone (B.Ar.) was quantified as the number of mineralized pixels relative to the total number of pixels in the field. Data are tabulated as percent mineralized bone area.
The regions of the tibiae selected for cancellous bone content analysis were designated as primary and secondary spongiosa. To standardize these areas for evaluation, the epiphyseal growth plate-metaphyseal junction was oriented parallel to the abscissa of the digitizing screen. Bone elements 1.06 mm (secondary spongiosa) and 0.02 mm (primary spongiosa) from the growth plate and equidistant from the flanking cortical elements were then quantified. The total area evaluated was 2.98 mm2 (1.42 mm wide and 2.1 mm long).
Biomechanical testing
The compressive strength of the fifth lumbar vertebra was determined using an Instron 4201 (Instron, Canton, MA) equipped with a 5000 newton (N) load cell as follows: all processes were removed using a diamond blade wafer saw (Buehler, Lake Bluff, IL). Two coplanar cuts, 4.9 mm apart, were made perpendicular to the cephalocaudal axis. This produces a uniform sample for compression. The sample was placed in a bone holding fixture that uses a piston moving parallel to the cephalocaudal axis to compress the sample. Data were analyzed by Instron Series IX software, which produced a load-deformation curve for each data set. Maximum load was calculated from the curve and expressed as newtons.
Statistical analysis for bone data
Treatment and vehicle groups were statistically compared using a software package (SAS Institute, Cary, NC) for one-way ANOVA with Dunnett’s test.
Results
The structure of bazedoxifene is shown in Fig. 1. Bazedoxifene is a new chemical entity with unique structural characteristics, compared with RAL and tamoxifen, two other compounds referred to as SERMs. It is believed the key for any one of these molecules to bind to the ERs is the presence of at least one phenolic group. For tamoxifen, hydroxylation at the 4-position is critical to its efficient interaction with ER. Arrows in the figure identify a few obvious structural differences between bazedoxifene and RAL. The design and synthesis of bazedoxifene are clearly described in a previous publication (22).
Ligand binding to ER and ER, transcriptional activity via ER
Bazedoxifene binds to ER and ER with high affinity. Its affinity for ER is about 10-fold lower than 17-estradiol or raloxifene (Table 1), but it is less ER selective than raloxifene.
In a simple transcription assay using a vitellogenin ERE upstream of a thymidine kinase promoter sequence, bazedoxifene does not activate ER in HepG2 cells, resulting in no detectable increase in luciferase activity vs. 17-estradiol, which has an EC50 of 0.06 nM in the same assay system (14). Similarly, RAL does not function as an ER agonist in this assay. In fact, bazedoxifene and RAL are potent antagonists in this assay whether the cellular background is a CHO (ovarian), HepG2 (hepatic), or GT1–7 (neuronal) cell (data not shown). When cotreated with 1.0 nM 17-estradiol, bazedoxifene has an IC50 of 22.0 nM in CHO cells, 4.97 nM in HepG2 cells, and 10.0 nM in GT1–7 cells. The results for raloxifene are quite similar (data not shown).
To determine whether a promoter-selective response could be associated with bazedoxifene, a hepatic lipase promoter luciferase construct was infected into HepG2 cells followed by treatment with 17-estradiol, tamoxifen, RAL, lasofoxifene, and bazedoxifene in dose response to determine their receptor agonist or antagonist activity (23). As seen in Table 2, 17-estradiol functions as an agonist on this promoter (EC50 = 26.0 nM), whereas for tamoxifen more than 10 μM is required and no detectable response to raloxifene or lasofoxifene could be measured. However, bazedoxifene did function as an agonist with an EC50 of 100.0 nM. Bazedoxifene is not as potent as 17-estradiol, but it does distinguish itself from these other SERMs, supporting the concept that subtle to moderate structural differentiation can impact the ability of any ligand to regulate a promoter’s activity.
MCF-7 cell proliferation
MCF-7 cell proliferation is accelerated dramatically by exposure to estrogen receptor agonists. The data in Fig. 2 were generated using an EC80 of 17-estradiol in this assay system, which is 10 pM, and maximum growth was set at 100% for comparative purposes. Not shown in the figure is bazedoxifene alone. Bazedoxifene does not stimulate MCF-7 cell proliferation, but when cells are cotreated with 17-estradiol, it demonstrates a dose-dependent inhibition of proliferation with an IC50 of 0.19 nM. Raloxifene behaves similarly with an IC50 of 1.0 nM.
Three-day immature rat uterine model
In this very sensitive model (24) for the evaluation of ER agonists, after 3 d of dosing, ethinyl estradiol (10 μg/kg) increased uterine wet weight to 406% of control (Table 3). When bazedoxifene was dosed at 0.5 and 5.0 mg/kg, there was an increase in wet weight (35%) at the lower dose and no significant difference at 5 mg/kg. In contrast to bazedoxifene, raloxifene at the same treatment dosages significantly increased weight wet 85.2 and 75.9% at the corresponding 0.5 and 5.0 mg/kg doses. One-to-one dosing of BZA and RAL (0.5 mg/kg) demonstrates a BZA inhibition of the RAL stimulation that converges to the BZA increase of 35%.
Whereas wet weight increases do indicate a uterine response, histological examination of the entire uterus reveals BZA does not affect luminal epithelial cell hypertrophy or hyperplasia, myometrial hypertrophy, or luminal distention. Photomicrographs of cross-sections of uteri from control, ethinyl estradiol, and BZA-treated rats reveals a greater than 3-fold increase in luminal cell height as a consequence of ethinyl estradiol treatment, whereas BZA treatment resulted in only a slight, insignificant increase in cell height (Fig. 3). Previous data have shown raloxifene increases luminal cell height (14, 25). The pathology review also confirmed previously reported results for raloxifene, which demonstrated increased luminal epithelial cell hypertrophy and myometrial hypertrophy. These changes for raloxifene were not the same magnitude as ethinyl estradiol but reflect definitive responses to treatment not seen with bazedoxifene (Table 4). Cotreatment of bazedoxifene with raloxifene completely abrogated the stimulation of the luminal epithelial cells and myometrium by raloxifene supporting an improved uterine profile for bazedoxifene (i.e. little or no uterine stimulation).
BMD, histology/histomorphometry, compressive strength
To determine the effect of BZA on the skeleton, a 6-wk ovariectomized rat model of osteopenia was used. Dose-response data for BZA demonstrate consistent, significant increased bone mass, compared with control animals via pQCT evaluation is achieved at 0.1 mg/kg·d with increases at 0.3, 1.0, and 3.0 mg/kg·d (Fig. 4). Several independent experiments evaluating BZA support the selection as 0.3 mg/k·d as the efficacious dosage in this model. Dose-response data with bazedoxifene and other SERMs like raloxifene demonstrate a steep dose response and a rapidly achieved maximal effect. Choosing the 0.3 mg/kg·d dose clearly impacts other responses measured with bazedoxifene such as uterine and vasomotor endpoints.
Histological assessment (Fig. 5 and Table 5) demonstrates the severe loss of trabecular bone in the primary spongiosia of ovariectomized rats. There is a decrease in B.Ar% from 16.66 + 1.07 to 14.11 + 0.79 in the ovariectomized animals. Bazedoxifene treatment for 6 wk protected the skeleton, maintaining bone mass and B.Ar% at 18.40 + 0.88, which was statistically different from the ovariectomized control. No statistically significant changes were seen for dynamic parameters of bone formation; however, bone volume per tissue volume (BV/TV) (percent) (shown in Table 5), osteoclast surface per bone surface (OsS/BS) (percent), and osteoclast number per bone surface (number per millimeter) (N.Oc/BS) were all statistically different from the vehicle control. Similar data were generated for RAL at its bone-sparing dose in the rat (3.0 mg/kg·d) (Table 5).
When trabecular core samples from the L4 vertebrae of the same groups of animals were tested for resistance to applied compressive force, the 0.3 mg/kg·d dose of bazedoxifene maintained compressive strength equivalent or better than the sham operated animals and was statistically significant (P 0.05), compared with the ovariectomized group (Table 5).
In addition to the bone data generated in this model, total cholesterol and wet uterine weight were also evaluated. A statistically significant decrease in total cholesterol was seen with bazedoxifene at 0.3 mg/kg·d and with raloxifene at 3.0 mg/kg·d. In contrast to the similarity in the lipid response for these two SERMs, the uterine wet weight change for bazedoxifene was not statistically different from the ovariectomized vehicle-treated animals, whereas that for raloxifene was 35% above that of bazedoxifene (191.6 vs. 149.3) and statistically different from the control group mean. All of these data are summarized in Table 5.
Vasomotor reactivity/thermoregulation
As women experience a reduction in ovarian output of estrogens, an apparent resetting of their thermoregulatory pathways occurs. The result is manifested as hot sweats at night and hot flushes throughout the entire day (26). The morphine-addicted rat model is an experimental paradigm designed to measure vasomotor changes via a thermoregulatory response initiated by a naloxone injection to the morphine-addicted rats (27). The acute response to naloxone is a rapid rise in tail skin temperature (with little or no core body temperature fluctuation). The increase in tail skin temperature is abated in rats pretreated with an ER agonist like 17-estradiol, ethinyl estradiol, or conjugated estrogens. In Fig. 6, 0.3 mg/kg ethinyl estradiol reduces the naloxone-induced rise in tail skin temperature to baseline. It is clear from these data that bazedoxifene alone does not act as an agonist in this model and, although not shown in this figure, neither does raloxifene at the same doses (0.1, 1.0, and 10 mg/kg). When bazedoxifene is coadministered with ethinyl estradiol, however, antagonism is detected at less than 1.0 mg/kg. This is also what occurs with raloxifene in this model (data not shown). At the bone-sparing efficacious dose of bazedoxifene (0.3 mg/kg·d), no antagonism is seen (data not shown), which may be an indicator for the potential lack of a vasomotor effect of bazedoxifene in postmenopausal women. This would be in contrast to raloxifene, which has a bone-sparing dose of 1–3 mg/kg in rats and a demonstrated increase in hot flushes in women taking the drug (28).
Discussion
The development of ER ligands has received a great deal of attention over the last several years with attempts to classify certain types as selective activators/modulators or SERMs (29, 30, 31). Classically, estrogens bind to their cognate receptors, which dimerize, bind to DNA, and regulate transcription of a variety of target genes. This oversimplified paradigm has been intensely evaluated and expanded to now demonstrate that several classes of proteins in addition to the liganded receptors are recruited to a transcriptional complex to enhance and regulate transcription. Interestingly, dependent on which ligand binds to the receptor, the transcriptional response can be modified to various extents (32). The apparent explanation for this phenomenon is the fact that each ligand that binds to the receptor places that receptor in a specific conformation. Crystallization of the ER ligand binding domain with a few different compounds supports this hypothesis (12). The conformation dictates which proteins can interact and what other biochemical modifications can occur. The optimal ligand would place the ER in the hypothetically most efficient conformation to enhance transcription. Certain genes will be more permissive than others, depending on what receptor complex is presented to them and this directly impacts transcriptional efficiency. The concept of selectivity is a derivative of this scenario.
Bazedoxifene represents a new chemical entity with demonstrable selective activity relating to the skeleton, CNS, lipid metabolism, uterus, and mammary gland. The development of SERMs, or at least the creation of the concept, was to have a compound effectively control menopausal symptoms and protect the skeleton like hormone therapy without the negatively associated side effects of breast cancer and thromboembolic events. Unfortunately, this has yet to be achieved; however, in preclinical evaluation, bazedoxifene appears to have improved the SERM profile beyond that achieved by raloxifene.
Perhaps the major issue relating to SERM activity is not so much what the molecule achieves as far as bone mass maintenance is concerned but more so as what the SERM does on targets that should not be touched either by agonist or antagonist activities. A key target for BZA was the uterine endometrium. At least three SERMs, levormeloxifene, idoxifene, and droloxifene, failed in clinical evaluation due to complications associated with uterine stimulation (33, 34). Based on uterine wet weight data generated in both immature and mature ovariectomized rats, one could have predicted that these compounds would stimulate the uterus. It is important to point out that this does not necessarily translate to endometrial hyperplasia. In fact, data generated in our laboratories have not demonstrated any SERM to stimulate a profound hyperplastic response, but clear hypertrophic endometrial and myometrial effects are detected coupled with increases in an epithelial-specific gene’s expression—complement component 3 (35). Comparison of bazedoxifene with RAL reveals that both affect increases in rat uterine wet weights; however, RAL increases wet weight to a greater degree than bazedoxifene, and the histological changes are overt with RAL, whereas bazedoxifene does not elicit any detectable modification in the endometrium or myometrium. In fact, the mild stimulation seen with RAL in rodents is manifested in humans, but the observed changes are not considered to be clinically relevant. That is, there is a slight increase in endometrial thickness and histologically observable glands, but there is no detectable hyperplasia (36). Hyperplasia is the key issue from a regulatory perspective when discussing endometrial safety in women followed by increased fluid production, which was the apparent problem with levormeloxifene and idoxifene. However, it is important to note the correlation between hyperplasia of the endometrial epithelium and endometrial cancer is not absolute and other uterine responses may be just as important (6).
Whereas the uterus has been a tissue target for SERMs, this has caused problems and is clearly a tissue that should not be stimulated. Perhaps the target with the greatest impact on a woman’s decision to take any estrogen or SERM is the effect these compounds may exert on the breast and their effect on breast cancer. The data presented address the potential impact bazedoxifene might have on existing breast tumor cells’ proliferation. Potent inhibition of 17-estradiol-stimulated breast cancer cell proliferation is probably a good predictor of the probable effect of a SERM on breast tumor growth. Similar data have been presented for other SERMs including RAL and tamoxifen (37), and RAL has demonstrated a reduction in the incidence of breast cancer in clinical trials (however they were not powered to specifically evaluate breast cancer rates) (38, 39).
These types of proliferation data do not address the influence of bazedoxifene or any other SERM on normal mammary gland differentiation. Normal mammary gland differentiation is regulated by 17-estradiol and progesterone. Estradiol primes the epithelium and progesterone regulates further ductal differentiation and epithelial proliferation (40, 41). Whether this process can be correlated with tumorigenesis is unknown, but preliminary bazedoxifene data in a rat mammary gland differentiation model show that it does not prime like estrone or 17-estradiol resulting in no detectable histological response including a total absence of proliferation (42).
The CNS is a target for estrogens and the role of estrogens in tempering/abating vasomotor instability associated with hot flushes is the primary motivation for women to use hormone replacement or estrogen therapies (43). Mechanistically, how an ER agonist like17-estradiol reduces the number and intensity of hot flushes is not known. It is most likely more than a simple temperature set point adjustment regulated by direct effects of the estrogen. The interaction of -adrenergic and serotonergic pathways is apparently also involved in temperature control (26). However, whether estrogens are directly regulating these pathways has yet to be discerned. Data generated comparing bazedoxifene and RAL in the rodent hot flush model would categorize both as antagonists at doses of 1.0 mg/kg and higher. The question is in a menopausal woman with a circulating estradiol in the range of less than 20 pg/ml, what impact would an ER antagonist have The rodent model suggests vasomotor responses to either of these SERMs is directly related to their bone efficacious dose, i.e. it is simply dose related. More RAL is needed to maintain bone mass, compared with bazedoxifene in rodent models (1–3 vs. 0.3 mg/kg, respectively); thus, at the bone-effective dose, RAL completely antagonizes the positive effect on tail skin temperature by 17-estradiol. There may be differences in blood-brain barrier penetrance, but this has not been proven. It does seem that an explanation for the increase in hot flushes reported in woman taking RAL (28) is simply dose related. Perhaps more interesting is the apparent fact that tamoxifen (28), lasofoxifene (44), and presumably all other SERMs studied to date have this antagonist effect on flushes, yet their agonist effects on other targets outside of the CNS are dissimilar. Clearly, elucidation of hot flush physiology will be necessary to address these questions.
The primary target for SERMs is the skeleton in which ER agonist activity is the goal. The preclinical data supporting bazedoxifene as an antiresorptive therapy for the prevention and treatment of postmenopausal osteoporosis are strong. Bazedoxifene reveals evidence of efficacy on maintaining bone mass at a dose as low as 0.1 mg/kg·d and consistently demonstrates maximal efficacy at approximately 0.3 mg/kg·d. Efficacy on skeletal parameters appears to be similar to what has been reported for the SERMs, RAL, and lasofoxifene (9, 45). In our laboratory, bazedoxifene consistently shows a positive effect on the maintenance of lumbar vertebral resistance to an applied compressive force, a surrogate for a reduced incidence of fracture. The histological quality of bone (proximal tibia) is maintained and correlates well with the increases in BMD and compressive force data, which one would align with a reduction in fracture risk. Bazedoxifene, like the other SERMs, does not result in a skeletal response like 17-estradiol or the bisphosphonate, alendronate, in an ovariectomized rat (46); however, as has been shown in the clinic for RAL, a modest increase in BMD is sufficient to result in a significant decrease in vertebral fractures not too different from the bisphosphonates (47).
Bazedoxifene, a tissue-selective estrogen or a SERM, has been selected based on its differentiation from other SERMs’ effects on two key target tissues: the uterus and CNS. Definitively, even though this group of compounds all function via activation or inhibition of ER function, their somewhat modest structural differences are adequate to result in physiological responses that are not superimposable. If one carefully examines the many effects that 17-estradiol imparts on its various targets (genes and tissues), an appropriate conclusion would be that it, too, is a SERM. In fact, it can be argued that all estrogens are SERMs, albeit with differing degrees of selectivity. Achieving the right mix is what results or would result in the perfect SERM. Clearly, bazedoxifene has not achieved this status, although it does distinguish itself from the others, especially with regard to its effects on the uterus. The question of whether a SERM can be developed that exhibits an optimal clinical profile on all relevant tissues is still unresolved. Our data, and data on other SERMs, suggest uterine and bone separation can be achieved at the expense of reducing vasomotor instability (hot flushes). Perhaps the solution will be combining compounds to gain the best from both. That combination could be the perfect SERM if a balance between the essential and less desirable characteristics of each is attainable.
Acknowledgments
The list of individuals who have helped with this project are numerous. Thanks go to everyone at Wyeth in Women’s Health Research and Discovery Chemistry who participated in the development of bazedoxifene. Special thanks go to Drs. Doug Harnish and Istvan Merchanthaler for their work on gene transcription and the hot flush model, respectively, and Susan Jenkins for her work on MCF-7 cell proliferation.
Footnotes
Abbreviations: B.Ar, Bone mineral area; BMD, bone mineral density; BZA, bazedoxifene acetate; CHO, Chinese hamster ovarian cell line; CNS, central nervous system; DXA, dual-energy x-ray absorptiometry; ER, estrogen receptor; ERE, estrogen response element; FBS, fetal bovine serum; L4, fourth lumbar vertebrae; pQCT, peripheral quantitative computed tomography; PT, proximal tibiae; RAL, raloxifene; SERM, selective ER modulator; SRC, steroid receptor coactivator.
References
Lindsay R, Dempster DW, Jordan VC 1997 Estrogens and antiestrogens: basic and clinical aspects. Philadelphia: Lippincott-Raven
O’Connell MB 1995 Pharmacokinetic and pharmacologic variation between different estrogen products. J Clin Pharmacol 35(Suppl):18S–24S
Komm BS, Bodine PV 2001 Regulation of bone cell function by estrogens. In: Marcus R, Feldman D, Kelsey J, eds. Osteoporosis. 2nd ed. New York: Academic Press; 305–337
Cole MP, Jones CT, Todd ID 1971 A new anti-oestrogenic agent in late breast cancer. An early clinical appraisal of ICI46474. Br J Cancer 25:270–275
MacGregor JI, Jordan VC 1998 Basic guide to the mechanisms of antiestrogen action. Pharmacol Rev 50:151–196
Mourits MJE, de Vries EGE, Willemse PHB, ten Hoor KA, Hollema H, van der Zee AGJ 2001 Tamoxifen treatment and gynecologic side effects: a review. Obstet Gynecol 97:855–866
Gradishar WJ, Jordan VC 1997 Clinical potential of new antiestrogens. J Clin Oncol 15:840–852
Buzdar AU, Marcus C, Holmes F, Hug V, Hortobagyi G 1988 Phase II evaluation of Ly156758 in metastatic breast cancer. Oncology 45:344–345
Black LJ, Sato M, Rowley ER, Magee DE, Bekele A, Williams DC, Cullinan GJ, Bendele R, Kauffman RF, Bensch WR, Frolik CA, Termine JD, Bryant HU 1994 Raloxifene (LY139481 HCI) prevents bone loss and reduces serum cholesterol without causing uterine hypertrophy in ovariectomized rats. J Clin Invest 93:63–69
Sato M, Glasebrook AL, Bryant HU 1994 Raloxifene: a selective estrogen receptor modulator. J Bone Miner Metab 12(Suppl 2):S9–S20
McDonnell DP, Clemm DL, Hermann T, Goldman ME, Pike JW 1995 Analysis of estrogen receptor function in vitro reveals three distinct classes of antiestrogens. Mol Endocrinol 9:659–669
Shiau AK, Barstad D, Loria PM, Cheng L, Kushner PJ, Agard DA, Greene GL 1998 The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell 95:927–937
Glass CK, Rose DW, Rosenfeld MG 1997 Nuclear receptor coactivators. Curr Opin Cell Biol 9:222–232
Komm BS, Lyttle CR 2001 Developing a SERM: stringent preclinical selection criteria leading to an acceptable candidate (WAY-140424) for clinical evaluation. Ann NY Acad Sci 949:317–326
Van Den Bemd GJCM, Kuiper GGJM, Pols HAP, van Leeuwen JPTM 1999 Distinct effects on the conformation of estrogen receptor and by both the antiestrogens ICI 164,384 and ICI 182,780 leading to opposite effects on receptor stability. Biochem Biophys Res Commun 261:1–5
Lonard DM, Smith CL 2002 Molecular perspectives on selective estrogen receptor modulators (SERMs): progress in understanding their tissue-specific agonist and antagonist actions. Steroids 67:15–24
Harris HA, Bapat AR, Gonder DS, Frail DE 2002 The ligand binding profiles of estrogen receptors and are species dependent. Steroids 67:379–384
Harnish DC, Scicchitano MS, Adelman SJ, Lyttle CR, Karathanasis SK 2000 The role of CBP in estrogen receptor cross-talk with nuclear factor-B in HepG2 cells. Endocrinology 141:3403–3411
Bodine PV, Green J, Harris HA, Bhat RA, Stein GS, Lian JB, Komm BS 1997 Functional properties of a conditionally phenotypic, estrogen-responsive, human osteoblast cell line. J Cell Biochem 65:368–387
Harris HA, Katzenellenbogen JA, Katzenellenbogen BS 2002 Characterization of the biological roles of the estrogen receptors, ER and ER, in estrogen target tissues in vivo through the use of an ER-selective ligand. Endocrinology 143:4172–4177
Merchenthaler I, Funkhouser JM, Carver JM, Lundeen SG, Ghosh K, Winneker RC 1998 The effect of estrogens and antiestrogens in a rat model for hot flush. Maturitas 30:307–316
Miller CP, Collini MD, Tran BD, Harris HA, Kharode YP, Marzolf JT, Moran RA, Henderson RA, Bender RH, Unwalla RJ, Greenberger LM, Yardley JP, Abou-Gharbia MA, Lyttle CR, Komm BS 2001 Design, synthesis, and preclinical characterization of novel, highly selective indole estrogens. J Med Chem 44:1654–1657
Hadzopoulou-Cladaras M, Cardot P 1993 Identification of a cis-acting negative DNA element which modulates human hepatic triglyceride lipase gene expression. Biochemistry 32:9657–9667
Komm BS, Rusling DJ, Lyttle CR 1986 Estrogen regulation of protein synthesis in the immature rat uterus: the analysis of proteins released into the medium during in vitro incubation. Endocrinology 118:2411–2416
Ashby J, Odum J, Foster JR 1997 Activity of raloxifene in immature and ovariectomized rat uterotrophic assays. Regul Toxicol Pharmacol 25:226–231
Stearns V, Ullmer L, Lopez JF, Smith Y, Isaacs C, Hayes D 2002 Hot flushes. Lancet 360:1851–1861
Katovich MJ, Simpkins JW, Song IC, O’Meara J 1987 Effects of central application of naloxone on the skin temperature response in morphine-dependent rats. Brain Res Bull 19:81–85
Davies GC, Huster WJ, Lu Y, Plouffe Jr L, Lakshmanan M 1999 Adverse events reported by postmenopausal women in controlled trials with raloxifene. Obstet Gynecol 93:558–565
Katzenellenbogen BS, Katzenellenbogen JA 2002 Defining the "S" in SERMs. Science 295:2380–2381
Agnusdei D, Iori N 2000 Selective estrogen receptor modulators (SERMs): effects on multiple organ systems. Curr Med Chem 7:577–584
Burr ML 1987 Is asthma increasing J Epidemiol Community Health 41:185–189
Shang Y, Brown M 2002 Molecular determinants for the tissue specificity of SERMs. Science 295:2465–2468
Goldstein SR, Nanavati N 2002 Adverse events that are associated with the selective estrogen receptor modulator levormeloxifene in an aborted phase III osteoporosis treatment study. Am J Obstet Gynecol 187:521–527
Hendrix SL, McNeeley SG 2001 Effect of selective estrogen receptor modulators on reproductive tissues other than endometrium. Ann NY Acad Sci 949:243–250
Sundstrom SA, Komm BS, Ponce-de-Leon H, Yi Z, Teuscher C, Lyttle CR 1989 Estrogen regulation of tissue-specific expression of complement C3. J Biol Chem 264:16941–16947
Boss SM, Huster WJ, Neild JA, Glant MD, Eisenhut CC, Draper MW 1997 Effects of raloxifene hydrochloride on the endometrium of postmenopausal women. Am J Obstet Gynecol 177:1458–1464
O’Regan RM, Jordan VC 2001 Tamoxifen to raloxifene and beyond. Semin Oncol 28:260–273
Chlebowski RT, Col N, Winer EP, Collyar DE, Cummings SR, Vogel III VG, Burstein HJ, Eisen A, Lipkus I, Pfister DG 2002 American Society of Clinical Oncology technology assessment of pharmacologic interventions for breast cancer risk reduction including tamoxifen, raloxifene, and aromatase inhibition. J Clin Oncol 20:3328–3343
Cummings SR, Eckert S, Krueger KA, Grady D, Powles TJ, Cauley JA, Norton L, Nickelsen T, Bjarnason NH, Morrow M, Lippman ME, Black D, Glusman JE, Costa A, Jordan VC 1999 The effects of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. JAMA 281:2189–2197
Russo IH, Russo J 1998 Role of hormones in mammary cancer initiation and progression. J Mammary Gland Biol Neoplasia 3:49–61
Humphreys RC, Lydon J, O’Malley BW, Rosen JM 1997 Mammary gland development is mediated by both stromal and epithelial progesterone receptors. Mol Endocrinol 11:801–811
Komm BS, Kharode Y, Bodine PVN, Bex F 2003 Bazedoxifene+conjugated estrogens: a balanced combination to provide optimal "estrogenic" safety and efficacy. J Bone Miner Res 18(Suppl 2):SU385
Nelson HD 2004 Commonly used types of postmenopausal estrogen for treatment of hot flashes: scientific review. JAMA 291:1610–1620
Ettinger M, Schwartz E, Emkey R, Moffet AH, Bolognese M, Weiss SR, Lee A, Lasofoxifene, a next generation selective estrogen receptor modulator (SERM) in the prevention of bone loss in postmenopausal women. Program of the 86th Annual Meeting of the Endocrine Society, New Orleans, LA, 2004 (Abstract S35-2)
Ke HZ, Paralkar VM, Grasser WA, Crawford DT, Qi H, Simmons HA, Pirie CM, Chidsey-Frink KL, Owen TA, Smock SL, Chen HK, Jee WS, Cameron KO, Rosati RL, Brown TA, Dasilva-Jardine P, Thompson DD 1998 Effects of CP-336,156, a new, nonsteroidal estrogen agonist/antagonist, on bone, serum cholesterol, uterus and body composition in rat models. Endocrinology 139:2068–2076
Ma YL, Bryant HU, Zeng Q, Schmidt A, Hoover J, Cole HW, Yao W, Jee WS, Sato M 2003 New bone formation with teriparatide [human parathyroid hormone-(1–34)] is not retarded by long-term pretreatment with alendronate, estrogen, or raloxifene in ovariectomized rats. Endocrinology 144:2008–2015
Ettinger B, Black DM, Mitlak BH, Knickerbocker RK, Nickelsen T, Genant HK, Christiansen C, Delmas PD, Zanchetta JR, Stakkestad J, Glüer CC, Krueger K, Cohen FJ, Eckert S, Ensrud KE, Avioli LV, Lips P, Cummings SR 1999 Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene. Results from a 3-year randomized clinical trial. JAMA 282:637–645(Barry S. Komm, Yogendra P)
Abstract
We assessed the preclinical characteristics of a novel, stringently screened selective estrogen receptor modulator, bazedoxifene acetate, including its ability to bind to and activate estrogen receptors and promote increased bone mineral density and bone strength in rats, and the effects impacting the uterine endometrium, breast cancer cell proliferation, and central nervous system-associated vasomotor responses in an animal model. Bazedoxifene bound to estrogen receptor- with an IC50 of 26 nM, an affinity similar to that of raloxifene. Bazedoxifene did not stimulate proliferation of MCF-7 cells but did inhibit 17-estradiol-induced proliferation with an IC50 of 0.19 nM. In an immature rat uterine model, bazedoxifene (0.5 and 5.0 mg/kg) was associated with less increase in uterine wet weight than either ethinyl estradiol (10 μg/kg) or raloxifene (0.5 and 5.0 mg/kg). Histological analysis revealed that coadministration of bazedoxifene also appeared to reduce raloxifene-stimulated endometrial luminal epithelial cell and myometrial cell hypertrophy. In ovariectomized rats, bazedoxifene was associated with significant increases in bone mineral density at 6 wk, compared with control, and better compressive strength of bone samples from the L4 vertebrae, compared with samples from ovariectomized animals. In the morphine-addicted rat model of vasomotor activity, bone-sparing doses of bazedoxifene alone were not associated with 17-estradiol inhibition of increased vasomotor activity. Bazedoxifene acetate represents a promising new treatment for osteoporosis, with a potential for less uterine and vasomotor effects than selective estrogen receptor modulators currently used in clinical practice. Controlled clinical trial data will be needed to confirm these effects.
Introduction
ESTROGENS COMPRISE A RELATIVELY large class of compounds with substantial structural diversity with at least one common characteristic: they bind to the estrogen receptors (ERs) to transduce their activity. Over time different names have been assigned to members of this family to, more than anything, describe their functional role. The classical members include the steroidal estrogens, of which there are many, with estrone 17-estradiol, and estriol representing three of the most abundant and active in humans. There are a fair number of nonsteroidal estrogens, with diethylstilbestrol being the most often cited and used. The phytoestrogens are represented by a rather sizable number of compounds, again with a number of different structural backbones (1, 2, 3).
Many years ago, when it was hypothesized that estrogens may be implicated in breast cancer etiology, an effort to generate compounds that would antagonize an ER agonist like 17-estradiol was undertaken. Although not the first, tamoxifen, a triphenylethylene, proved to be a competitor for estrogen agonist binding to the ER and, indeed, did inhibit breast cancer cell proliferation (4, 5). Due to the presence of some undesirable side effects (6), efforts to improve on tamoxifen resulted in a group of compounds, which collectively became known simply as antiestrogens. Conceptually, a group of compounds now existed that would block physiologic responses to ER agonists like estrone, 17-estradiol, and estriol by competitively binding to the ER. If the antiestrogens were occupying the ER, no agonist activity would be imparted.
However, in the late 1980s, work with tamoxifen given to rats demonstrated that tamoxifen actually revealed mixed functional activity (7). That is, it acted as both an ER agonist and antagonist to varying degrees correlated to the physiologic end point under investigation. Not too long after this initial finding, other antiestrogens were also shown to display similar mixed functional activities, although not necessarily identical in character to one another. For example, raloxifene (RAL), originally created to function as a treatment for breast cancer (8), like tamoxifen, inhibited breast cancer cell proliferation yet, unlike tamoxifen, did not demonstrate the same level of uterine endometrial stimulation (9). Based on these data, the term selective ER modulator (SERM) was coined to describe these types of ligands (10), which display mixed functional estrogenic activities.
Irrespective of the descriptor used to categorize these compounds, they transduce their information via the ERs and all exhibit varying degrees of estrogen-like activity. If this group of diverse structures all binds to the ERs with relatively high affinity, how can the differences in gene activation and physiological outcomes be so different Before crystallization of the ligand binding domains of ER and ER were available, it was demonstrated by proteolytic digestion that the bound ligand affected the digestion pattern of the ERs, which was easily discerned on sodium dodecyl sulfate gels (11). These data supported the concept that the ER structural conformation varies according to whether it is bound and the effects were somewhat ligand specific. This was convincingly confirmed by x-ray crystallography of the ERs and cocrystallization with other peptides, specifically a small peptide fragment of glucocorticoid receptor interacting protein-1 or steroid receptor coactivator (SRC)-2 (12). Certain members of the p160 coactivator group like SRC-2, when bound to the ER, enhance the receptor’s transcriptional activity on certain promoters (13). When ER is liganded to 17-estradiol, SRC-2 interacts with a specific region of the ER exposed when helix 12 is positioned over the ligand binding pocket, allowing interaction to occur in a hydrophobic groove created by helices 4, 5, and 6. However, when tamoxifen or raloxifene are substituted for 17-estradiol, helix 12 is shifted to a position that blocks access of SRC-2 (and conceptually other p160 coactivators) to the appropriate site for them to function as coregulators. Simplistically, slight alterations in the receptor conformation alter its ability to interact efficiently with a coactivator, thus potentially reducing its effectiveness in the regulation of gene transcription. Of course, the conformation of the ER will affect not only potential protein-protein interactions but also biochemical modifications such as phosphorylation critical to ER function.
The selective activity associated with any estrogen, be it 17-estradiol or ICI182780 (referred to as a pure antagonist) depends on, first and foremost, the conformation it imparts to the ER on binding and the subsequent interactions/modifications that occur. The ER presented to a particular gene’s promoter will dictate the response. It appears that one gene may be more permissive when the ER is not in what would be considered its optimal conformational status to permit efficient transcriptional enhancement vs. another promoter, which will not permit such activity, plus everything in between. These responses appear to be cell type or tissue selective in addition to promoter selective (14, 15). This type of conceptual explanation for selective activity conveniently supports the notion that each estrogen, even though working through one of two receptors, can impact a tissue, cell, or gene with distinct activity (16).
Here we describe a new selective estrogen with activity distinct from other members of the SERM family. The compound is an indole-based ER ligand stringently selected to ensure an improved profile vs. its predecessors. The compound, bazedoxifene acetate (BZA), is shown to interact with the ERs, transactivate the ER, and positively affect the skeletal and lipid profile without stimulating the uterine endometrium, causing breast cancer cell proliferation or negatively impacting the central nervous system (CNS) in preclinical models.
Materials and Methods
All chemicals and reagents were reagent grade or better and obtained from various vendors. 17-ethinyl estradiol, 17-estradiol, RAL HCl, tamoxifen, 4-OH tamoxifen, and lasofoxifene were supplied by the Wyeth compound library (Princeton, NJ) that are either purchased or generated by Wyeth Medicinal Chemistry (Collegeville, PA) and undergo analytical confirmation before cataloging/use.
Animals
Sprague Dawley rats of varying age were obtained from Taconic Farms (Germantown, NY) for all animal experiments described. Animals were housed appropriately in Wyeth’s Association for Assessment and Accreditation of Laboratory Animal Care- and U.S. Department of Agriculture-sanctioned facility, and all procedures (Institutional Animal Care and Use Committee approved) were performed in accordance with all federal, state, and local guidelines.
Ligand binding
Interaction of BZA with human ER and ER was assessed with a solid phase competitive radioligand binding assay using [3H]-17- estradiol as previously described (17).
In vitro transcription assays
CHO-HER.14 (Chinese hamster ovarian cell line), HOB03CE6 (human osteoblast cell line), and D12 (hypothalamic neuronal cell line) and HepG2–313-89-A6 cells (hepatoma cell line) were used as the cell recipients of various transcription vectors containing estrogen responsive DNA sequence elements. The modified CHO cell line contains a stable integrated human estrogen receptor sequence as does the modified HepG2 cell line. The construction and details of their production can be retrieved from Harnish et al. (18)
CHO cells were plated at 50,000 cells/well in 12-well dishes and grown at 37 C/5% CO2 in phenol red-free MEM containing 10% charcoal-treated fetal bovine serum (FBS). The HOB03Ce6 cell line was maintained at 34 C in phenol red-free DMEM/F-12 containing 10% (vol/vol) heat-inactivated FBS, 1% (vol/vol) penicillin-streptomycin, and 2 mM GlutaMAX-1 (19). D12 cells were cultured at 37 C/5% CO2 in phenol red-free medium (DMEM/F-12, GlutaMAX, penicillin-streptomycin) containing 5% charcoal stripped FBS at approximately 2 x 106 cells per 150-mm dish. Twenty-four hours later, medium was changed to one containing only 2% stripped FBS with or without test compounds. HepG2 cells were maintained at 37 C in a 5% CO2 incubator in phenol red-free DMEM, 10% heat-inactivated FBS, 1% GlutaMAX, 1% MEM nonessential amino acids, 100 U/ml penicillin, and 100ug/ml streptomycin at 2.5 x 105 cells/well in 12-well dishes.
For transcription assays using an estrogen response element (ERE) combined with a weak promoter and luciferase reporter sequence, the following adenoviral construct was used. The adenovirus (Ad-5) contained two copies of the vitellogenin A2 ERE (5'-GGTCACAGTGACC-3') linked to –110 to +10 of the thymidine kinase promoter and a luciferase gene coding sequence. This construct was inserted in place of the E1a adenoviral gene. Cells are routinely infected with approximately 200 plaque-forming units/cell added directly to the cell culture medium for 2 h. After viral infection incubation time, the media are replaced with media containing test compounds or positive controls. Details of transfections and luciferase assay conditions can be found in Bodine et al. (19)
MCF-7 cell proliferation
Samples of the human breast tumor cell line, MCF-7, were obtained from the American Type Culture Collection (Manassas, VA) and maintained in DMEM/F-12 (50:50) with 10% FBS and 1 x GlutaMAX-1. They were passaged twice weekly at a concentration of 0.1 million cells/cm2 and generally used in this assay between passages 12 and 40.
For the proliferation assay, cells were plated at 20,000 cells/well in a 24-well plate in DMEM/F12 (50:50) (phenol red-free) with 10% charcoal/dextran-treated FBS and 1 x GlutaMAX-1. After overnight incubation, the medium was aspirated and treatments in DMEM/F12 (50:50) (phenol red-free) with 2% charcoal/dextran-treated FBS and 1 x GlutaMAX-1 were added to the wells. Each plate had a vehicle (baseline proliferation) and treatments. Treatments included 10 pM 17-estradiol determined to be the EC80 for 17-estradiol and 17-estradiol in combination with six concentrations of BZA. Treatments from d 1 were renewed on d 3 and d 6 by aspirating medium from wells and replacing with fresh medium and treatments. On d 7, cells were detached from the plate using trypsin-EDTA and counted using a Multisizer II (Coulter, Miami, FL).
Uterine evaluation
The effect of compounds on compounds on uterine weight was evaluated as previously described (20). Briefly, sexually immature Sprague Dawley rats were treated once daily for 3 d and euthanized approximately 24 h after administration of the last dose. The vehicle for sc dosing was 50% dimethylsulfoxide-50% 1x Dulbecco’s PBS, and the vehicle for oral administration was 2% Tween 80 and 0.5% methylcellulose. After death, uteri were excised and weighed after trimming associated fat and expressing any luminal fluid. A segment was placed in a 10% formalin solution for histological analysis and the remainder frozen on dry ice for isolation of RNA.
Histological preparation after fixation included dehydration, paraffin embedding, slide preparation, and staining/counterstaining with hematoxylin/eosin. Veterinary pathologists scored the following five parameters (1, 2, 3, 4): epithelial hypertrophy/hyperplasia, myometrial hypertrophy, luminal distention, stromal eosinophilia, and luminal epithelial apoptosis. Endometrial luminal epithelial cell height was measured with a micrometer on a Eclipse 800 photomicroscope (Nikon, Tokyo, Japan). Manual measurements spanning from the basal lamina to the apical surface were taken from five regions around the lumen, and an average epithelial cell height was determined.
Uterine RNA was isolated by polytron homogenization in Trizol followed by alcohol precipitation, air drying, and resuspension in buffer containing 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA (pH 8.0). Complement component 3 evaluation was performed by Northern blot analysis.
Vasomotor instability (hot flush)
Ovariectomized female (60 d) rats were obtained after surgery. The surgeries were performed minimally 7 d before initiation of any experiment. Vehicle and ethinyl estradiol (0.3 mg/kg) were included in each replicate. Bazedoxifene was administered orally in a saline, Tween-80, methylcellulose vehicle. A detailed description of methodology for evaluating vasomotor instability in rats has been published (21). Briefly, compound treatment (17-estradiol, ethinyl estradiol, or bazedoxifene) is initiated, and on the third day of treatment each animal receives a morphine pellet sc. This is followed by two more pellets on the fifth day of treatment. On the eighth day, a thermistor is taped to the animal’s tail to measure tail skin temperature for 15 min (to obtain baseline temperature) followed by a sc injection of naloxone (1 mg/kg). Tail skin temperature readings continue for 1 h after naloxone injection.
Vertebral bone densitometry
In mature (2.5–3 months, 250 g) ovariectomized rats, all treatments were initiated 3 d after ovariectomy and continued for 6 wk. Five weeks after the initiation of treatment and 1 wk before the termination of study, each rat was evaluated for bone mineral density (BMD). The BMDs of the proximal tibiae (PT) and fourth lumbar vertebrae (L4) were measured in anesthetized rats using a dual-energy x-ray absorptiometer (DXA) (Eclipse XR-26, Norland Corp., Ft. Atkins, WI). The DXA measurements for each rat were performed as follows: a preliminary scan was performed at a scan speed of 50 mm/sec with a scan resolution of 1.5 mm x 1.5 mm to determine the region of interest in PT and L4. Small subject software was employed at a scan speed of 10 mm/sec with resolution of 0.5 mm x 0.5 mm for final BMD measurements. A defined area of a 1.5 cm-wide region in the PT (starting from femur-tibia junction to 0.5 cm distal) or a 1.5 cm-wide area to cover the total length of L4 was selected for analysis. The BMDs for respective sites were computed by the software as a function of the attenuation of the dual beam (46.8 and 80 KeV) x-ray generated by the source underneath the subject and the detector traveling along the defined area above the subject. The data for BMD values were expressed in grams per square centimeter.
Peripheral quantitative computed tomography (pQCT)
Volumetric total (cortical and trabecular) and trabecular densities were evaluated in anesthetized rats using a XCT-960M (pQCT; Stratec Medizintechnik, Pforzheim, Germany). The right hind limb was placed in a polycarbonate tube affixed to a sliding platform that maintained it parallel to the aperture of the pQCT. The platform was adjusted so that the distal end of the femur and the proximal end of the tibia would be in the scanning field. A two-dimensional scout view was run for a length of 10 mm and line resolution of 0.2 mm. The pQCT scan was initiated 3.4 mm distal from the proximal end of tibia. The pQCT scan was 1 mm thick, had a voxel (three-dimensional pixel) size of 0.140 mm, and consisted of 145 projections through the slice. After the pQCT scan was completed, the image was displayed on the monitor. A region of interest in the tibia was analyzed for total density and the value was reported in milligrams per cubic centimeter. The outer 55% of the bone was mathematically removed away in a concentric spiral. The density of the remaining bone was reported as trabecular density in milligrams per cubic centimeter.
Tissue collection and serum cholesterol measurement
One week after BMD evaluation, serum cholesterol evaluation was performed. Total serum cholesterol was analyzed using a Hitachi 911 autoanalyzer (Roche Diagnostics, Indianapolis, IN). The uterus from each rat was removed, fat excised, luminal fluid blotted away, and weighed. The tibiae from both limbs and fifth lumbar vertebra were also dissected free of soft tissue, removed, and processed for histomorphometric analysis and compressive strength measurements, respectively.
Tibial histology
The proximal portion of the tibiae was fixed, dehydrated, and embedded undecalcified in a methyl methacrylate mixture. Longitudinal tissue sections (5–10 μM) were prepared on a Polycut S microtome (Reichert, Nussloch, Germany). Sections were stained with Goldner’s stain and coverslipped.
Cancellous (trabecular) bone content in the PT was quantified as two-dimensional bone mineral area (B.Ar) with an image analysis system (Videometric 150; Oncor, Gaithersburg, MD). The field of view from a microscope image of the slide was coupled to a microcomputer through a video camera (TMC-74 color RGB; Plunix, San Jose, CA). The mineralized tissue content of cancellous bone (B.Ar.) was quantified as the number of mineralized pixels relative to the total number of pixels in the field. Data are tabulated as percent mineralized bone area.
The regions of the tibiae selected for cancellous bone content analysis were designated as primary and secondary spongiosa. To standardize these areas for evaluation, the epiphyseal growth plate-metaphyseal junction was oriented parallel to the abscissa of the digitizing screen. Bone elements 1.06 mm (secondary spongiosa) and 0.02 mm (primary spongiosa) from the growth plate and equidistant from the flanking cortical elements were then quantified. The total area evaluated was 2.98 mm2 (1.42 mm wide and 2.1 mm long).
Biomechanical testing
The compressive strength of the fifth lumbar vertebra was determined using an Instron 4201 (Instron, Canton, MA) equipped with a 5000 newton (N) load cell as follows: all processes were removed using a diamond blade wafer saw (Buehler, Lake Bluff, IL). Two coplanar cuts, 4.9 mm apart, were made perpendicular to the cephalocaudal axis. This produces a uniform sample for compression. The sample was placed in a bone holding fixture that uses a piston moving parallel to the cephalocaudal axis to compress the sample. Data were analyzed by Instron Series IX software, which produced a load-deformation curve for each data set. Maximum load was calculated from the curve and expressed as newtons.
Statistical analysis for bone data
Treatment and vehicle groups were statistically compared using a software package (SAS Institute, Cary, NC) for one-way ANOVA with Dunnett’s test.
Results
The structure of bazedoxifene is shown in Fig. 1. Bazedoxifene is a new chemical entity with unique structural characteristics, compared with RAL and tamoxifen, two other compounds referred to as SERMs. It is believed the key for any one of these molecules to bind to the ERs is the presence of at least one phenolic group. For tamoxifen, hydroxylation at the 4-position is critical to its efficient interaction with ER. Arrows in the figure identify a few obvious structural differences between bazedoxifene and RAL. The design and synthesis of bazedoxifene are clearly described in a previous publication (22).
Ligand binding to ER and ER, transcriptional activity via ER
Bazedoxifene binds to ER and ER with high affinity. Its affinity for ER is about 10-fold lower than 17-estradiol or raloxifene (Table 1), but it is less ER selective than raloxifene.
In a simple transcription assay using a vitellogenin ERE upstream of a thymidine kinase promoter sequence, bazedoxifene does not activate ER in HepG2 cells, resulting in no detectable increase in luciferase activity vs. 17-estradiol, which has an EC50 of 0.06 nM in the same assay system (14). Similarly, RAL does not function as an ER agonist in this assay. In fact, bazedoxifene and RAL are potent antagonists in this assay whether the cellular background is a CHO (ovarian), HepG2 (hepatic), or GT1–7 (neuronal) cell (data not shown). When cotreated with 1.0 nM 17-estradiol, bazedoxifene has an IC50 of 22.0 nM in CHO cells, 4.97 nM in HepG2 cells, and 10.0 nM in GT1–7 cells. The results for raloxifene are quite similar (data not shown).
To determine whether a promoter-selective response could be associated with bazedoxifene, a hepatic lipase promoter luciferase construct was infected into HepG2 cells followed by treatment with 17-estradiol, tamoxifen, RAL, lasofoxifene, and bazedoxifene in dose response to determine their receptor agonist or antagonist activity (23). As seen in Table 2, 17-estradiol functions as an agonist on this promoter (EC50 = 26.0 nM), whereas for tamoxifen more than 10 μM is required and no detectable response to raloxifene or lasofoxifene could be measured. However, bazedoxifene did function as an agonist with an EC50 of 100.0 nM. Bazedoxifene is not as potent as 17-estradiol, but it does distinguish itself from these other SERMs, supporting the concept that subtle to moderate structural differentiation can impact the ability of any ligand to regulate a promoter’s activity.
MCF-7 cell proliferation
MCF-7 cell proliferation is accelerated dramatically by exposure to estrogen receptor agonists. The data in Fig. 2 were generated using an EC80 of 17-estradiol in this assay system, which is 10 pM, and maximum growth was set at 100% for comparative purposes. Not shown in the figure is bazedoxifene alone. Bazedoxifene does not stimulate MCF-7 cell proliferation, but when cells are cotreated with 17-estradiol, it demonstrates a dose-dependent inhibition of proliferation with an IC50 of 0.19 nM. Raloxifene behaves similarly with an IC50 of 1.0 nM.
Three-day immature rat uterine model
In this very sensitive model (24) for the evaluation of ER agonists, after 3 d of dosing, ethinyl estradiol (10 μg/kg) increased uterine wet weight to 406% of control (Table 3). When bazedoxifene was dosed at 0.5 and 5.0 mg/kg, there was an increase in wet weight (35%) at the lower dose and no significant difference at 5 mg/kg. In contrast to bazedoxifene, raloxifene at the same treatment dosages significantly increased weight wet 85.2 and 75.9% at the corresponding 0.5 and 5.0 mg/kg doses. One-to-one dosing of BZA and RAL (0.5 mg/kg) demonstrates a BZA inhibition of the RAL stimulation that converges to the BZA increase of 35%.
Whereas wet weight increases do indicate a uterine response, histological examination of the entire uterus reveals BZA does not affect luminal epithelial cell hypertrophy or hyperplasia, myometrial hypertrophy, or luminal distention. Photomicrographs of cross-sections of uteri from control, ethinyl estradiol, and BZA-treated rats reveals a greater than 3-fold increase in luminal cell height as a consequence of ethinyl estradiol treatment, whereas BZA treatment resulted in only a slight, insignificant increase in cell height (Fig. 3). Previous data have shown raloxifene increases luminal cell height (14, 25). The pathology review also confirmed previously reported results for raloxifene, which demonstrated increased luminal epithelial cell hypertrophy and myometrial hypertrophy. These changes for raloxifene were not the same magnitude as ethinyl estradiol but reflect definitive responses to treatment not seen with bazedoxifene (Table 4). Cotreatment of bazedoxifene with raloxifene completely abrogated the stimulation of the luminal epithelial cells and myometrium by raloxifene supporting an improved uterine profile for bazedoxifene (i.e. little or no uterine stimulation).
BMD, histology/histomorphometry, compressive strength
To determine the effect of BZA on the skeleton, a 6-wk ovariectomized rat model of osteopenia was used. Dose-response data for BZA demonstrate consistent, significant increased bone mass, compared with control animals via pQCT evaluation is achieved at 0.1 mg/kg·d with increases at 0.3, 1.0, and 3.0 mg/kg·d (Fig. 4). Several independent experiments evaluating BZA support the selection as 0.3 mg/k·d as the efficacious dosage in this model. Dose-response data with bazedoxifene and other SERMs like raloxifene demonstrate a steep dose response and a rapidly achieved maximal effect. Choosing the 0.3 mg/kg·d dose clearly impacts other responses measured with bazedoxifene such as uterine and vasomotor endpoints.
Histological assessment (Fig. 5 and Table 5) demonstrates the severe loss of trabecular bone in the primary spongiosia of ovariectomized rats. There is a decrease in B.Ar% from 16.66 + 1.07 to 14.11 + 0.79 in the ovariectomized animals. Bazedoxifene treatment for 6 wk protected the skeleton, maintaining bone mass and B.Ar% at 18.40 + 0.88, which was statistically different from the ovariectomized control. No statistically significant changes were seen for dynamic parameters of bone formation; however, bone volume per tissue volume (BV/TV) (percent) (shown in Table 5), osteoclast surface per bone surface (OsS/BS) (percent), and osteoclast number per bone surface (number per millimeter) (N.Oc/BS) were all statistically different from the vehicle control. Similar data were generated for RAL at its bone-sparing dose in the rat (3.0 mg/kg·d) (Table 5).
When trabecular core samples from the L4 vertebrae of the same groups of animals were tested for resistance to applied compressive force, the 0.3 mg/kg·d dose of bazedoxifene maintained compressive strength equivalent or better than the sham operated animals and was statistically significant (P 0.05), compared with the ovariectomized group (Table 5).
In addition to the bone data generated in this model, total cholesterol and wet uterine weight were also evaluated. A statistically significant decrease in total cholesterol was seen with bazedoxifene at 0.3 mg/kg·d and with raloxifene at 3.0 mg/kg·d. In contrast to the similarity in the lipid response for these two SERMs, the uterine wet weight change for bazedoxifene was not statistically different from the ovariectomized vehicle-treated animals, whereas that for raloxifene was 35% above that of bazedoxifene (191.6 vs. 149.3) and statistically different from the control group mean. All of these data are summarized in Table 5.
Vasomotor reactivity/thermoregulation
As women experience a reduction in ovarian output of estrogens, an apparent resetting of their thermoregulatory pathways occurs. The result is manifested as hot sweats at night and hot flushes throughout the entire day (26). The morphine-addicted rat model is an experimental paradigm designed to measure vasomotor changes via a thermoregulatory response initiated by a naloxone injection to the morphine-addicted rats (27). The acute response to naloxone is a rapid rise in tail skin temperature (with little or no core body temperature fluctuation). The increase in tail skin temperature is abated in rats pretreated with an ER agonist like 17-estradiol, ethinyl estradiol, or conjugated estrogens. In Fig. 6, 0.3 mg/kg ethinyl estradiol reduces the naloxone-induced rise in tail skin temperature to baseline. It is clear from these data that bazedoxifene alone does not act as an agonist in this model and, although not shown in this figure, neither does raloxifene at the same doses (0.1, 1.0, and 10 mg/kg). When bazedoxifene is coadministered with ethinyl estradiol, however, antagonism is detected at less than 1.0 mg/kg. This is also what occurs with raloxifene in this model (data not shown). At the bone-sparing efficacious dose of bazedoxifene (0.3 mg/kg·d), no antagonism is seen (data not shown), which may be an indicator for the potential lack of a vasomotor effect of bazedoxifene in postmenopausal women. This would be in contrast to raloxifene, which has a bone-sparing dose of 1–3 mg/kg in rats and a demonstrated increase in hot flushes in women taking the drug (28).
Discussion
The development of ER ligands has received a great deal of attention over the last several years with attempts to classify certain types as selective activators/modulators or SERMs (29, 30, 31). Classically, estrogens bind to their cognate receptors, which dimerize, bind to DNA, and regulate transcription of a variety of target genes. This oversimplified paradigm has been intensely evaluated and expanded to now demonstrate that several classes of proteins in addition to the liganded receptors are recruited to a transcriptional complex to enhance and regulate transcription. Interestingly, dependent on which ligand binds to the receptor, the transcriptional response can be modified to various extents (32). The apparent explanation for this phenomenon is the fact that each ligand that binds to the receptor places that receptor in a specific conformation. Crystallization of the ER ligand binding domain with a few different compounds supports this hypothesis (12). The conformation dictates which proteins can interact and what other biochemical modifications can occur. The optimal ligand would place the ER in the hypothetically most efficient conformation to enhance transcription. Certain genes will be more permissive than others, depending on what receptor complex is presented to them and this directly impacts transcriptional efficiency. The concept of selectivity is a derivative of this scenario.
Bazedoxifene represents a new chemical entity with demonstrable selective activity relating to the skeleton, CNS, lipid metabolism, uterus, and mammary gland. The development of SERMs, or at least the creation of the concept, was to have a compound effectively control menopausal symptoms and protect the skeleton like hormone therapy without the negatively associated side effects of breast cancer and thromboembolic events. Unfortunately, this has yet to be achieved; however, in preclinical evaluation, bazedoxifene appears to have improved the SERM profile beyond that achieved by raloxifene.
Perhaps the major issue relating to SERM activity is not so much what the molecule achieves as far as bone mass maintenance is concerned but more so as what the SERM does on targets that should not be touched either by agonist or antagonist activities. A key target for BZA was the uterine endometrium. At least three SERMs, levormeloxifene, idoxifene, and droloxifene, failed in clinical evaluation due to complications associated with uterine stimulation (33, 34). Based on uterine wet weight data generated in both immature and mature ovariectomized rats, one could have predicted that these compounds would stimulate the uterus. It is important to point out that this does not necessarily translate to endometrial hyperplasia. In fact, data generated in our laboratories have not demonstrated any SERM to stimulate a profound hyperplastic response, but clear hypertrophic endometrial and myometrial effects are detected coupled with increases in an epithelial-specific gene’s expression—complement component 3 (35). Comparison of bazedoxifene with RAL reveals that both affect increases in rat uterine wet weights; however, RAL increases wet weight to a greater degree than bazedoxifene, and the histological changes are overt with RAL, whereas bazedoxifene does not elicit any detectable modification in the endometrium or myometrium. In fact, the mild stimulation seen with RAL in rodents is manifested in humans, but the observed changes are not considered to be clinically relevant. That is, there is a slight increase in endometrial thickness and histologically observable glands, but there is no detectable hyperplasia (36). Hyperplasia is the key issue from a regulatory perspective when discussing endometrial safety in women followed by increased fluid production, which was the apparent problem with levormeloxifene and idoxifene. However, it is important to note the correlation between hyperplasia of the endometrial epithelium and endometrial cancer is not absolute and other uterine responses may be just as important (6).
Whereas the uterus has been a tissue target for SERMs, this has caused problems and is clearly a tissue that should not be stimulated. Perhaps the target with the greatest impact on a woman’s decision to take any estrogen or SERM is the effect these compounds may exert on the breast and their effect on breast cancer. The data presented address the potential impact bazedoxifene might have on existing breast tumor cells’ proliferation. Potent inhibition of 17-estradiol-stimulated breast cancer cell proliferation is probably a good predictor of the probable effect of a SERM on breast tumor growth. Similar data have been presented for other SERMs including RAL and tamoxifen (37), and RAL has demonstrated a reduction in the incidence of breast cancer in clinical trials (however they were not powered to specifically evaluate breast cancer rates) (38, 39).
These types of proliferation data do not address the influence of bazedoxifene or any other SERM on normal mammary gland differentiation. Normal mammary gland differentiation is regulated by 17-estradiol and progesterone. Estradiol primes the epithelium and progesterone regulates further ductal differentiation and epithelial proliferation (40, 41). Whether this process can be correlated with tumorigenesis is unknown, but preliminary bazedoxifene data in a rat mammary gland differentiation model show that it does not prime like estrone or 17-estradiol resulting in no detectable histological response including a total absence of proliferation (42).
The CNS is a target for estrogens and the role of estrogens in tempering/abating vasomotor instability associated with hot flushes is the primary motivation for women to use hormone replacement or estrogen therapies (43). Mechanistically, how an ER agonist like17-estradiol reduces the number and intensity of hot flushes is not known. It is most likely more than a simple temperature set point adjustment regulated by direct effects of the estrogen. The interaction of -adrenergic and serotonergic pathways is apparently also involved in temperature control (26). However, whether estrogens are directly regulating these pathways has yet to be discerned. Data generated comparing bazedoxifene and RAL in the rodent hot flush model would categorize both as antagonists at doses of 1.0 mg/kg and higher. The question is in a menopausal woman with a circulating estradiol in the range of less than 20 pg/ml, what impact would an ER antagonist have The rodent model suggests vasomotor responses to either of these SERMs is directly related to their bone efficacious dose, i.e. it is simply dose related. More RAL is needed to maintain bone mass, compared with bazedoxifene in rodent models (1–3 vs. 0.3 mg/kg, respectively); thus, at the bone-effective dose, RAL completely antagonizes the positive effect on tail skin temperature by 17-estradiol. There may be differences in blood-brain barrier penetrance, but this has not been proven. It does seem that an explanation for the increase in hot flushes reported in woman taking RAL (28) is simply dose related. Perhaps more interesting is the apparent fact that tamoxifen (28), lasofoxifene (44), and presumably all other SERMs studied to date have this antagonist effect on flushes, yet their agonist effects on other targets outside of the CNS are dissimilar. Clearly, elucidation of hot flush physiology will be necessary to address these questions.
The primary target for SERMs is the skeleton in which ER agonist activity is the goal. The preclinical data supporting bazedoxifene as an antiresorptive therapy for the prevention and treatment of postmenopausal osteoporosis are strong. Bazedoxifene reveals evidence of efficacy on maintaining bone mass at a dose as low as 0.1 mg/kg·d and consistently demonstrates maximal efficacy at approximately 0.3 mg/kg·d. Efficacy on skeletal parameters appears to be similar to what has been reported for the SERMs, RAL, and lasofoxifene (9, 45). In our laboratory, bazedoxifene consistently shows a positive effect on the maintenance of lumbar vertebral resistance to an applied compressive force, a surrogate for a reduced incidence of fracture. The histological quality of bone (proximal tibia) is maintained and correlates well with the increases in BMD and compressive force data, which one would align with a reduction in fracture risk. Bazedoxifene, like the other SERMs, does not result in a skeletal response like 17-estradiol or the bisphosphonate, alendronate, in an ovariectomized rat (46); however, as has been shown in the clinic for RAL, a modest increase in BMD is sufficient to result in a significant decrease in vertebral fractures not too different from the bisphosphonates (47).
Bazedoxifene, a tissue-selective estrogen or a SERM, has been selected based on its differentiation from other SERMs’ effects on two key target tissues: the uterus and CNS. Definitively, even though this group of compounds all function via activation or inhibition of ER function, their somewhat modest structural differences are adequate to result in physiological responses that are not superimposable. If one carefully examines the many effects that 17-estradiol imparts on its various targets (genes and tissues), an appropriate conclusion would be that it, too, is a SERM. In fact, it can be argued that all estrogens are SERMs, albeit with differing degrees of selectivity. Achieving the right mix is what results or would result in the perfect SERM. Clearly, bazedoxifene has not achieved this status, although it does distinguish itself from the others, especially with regard to its effects on the uterus. The question of whether a SERM can be developed that exhibits an optimal clinical profile on all relevant tissues is still unresolved. Our data, and data on other SERMs, suggest uterine and bone separation can be achieved at the expense of reducing vasomotor instability (hot flushes). Perhaps the solution will be combining compounds to gain the best from both. That combination could be the perfect SERM if a balance between the essential and less desirable characteristics of each is attainable.
Acknowledgments
The list of individuals who have helped with this project are numerous. Thanks go to everyone at Wyeth in Women’s Health Research and Discovery Chemistry who participated in the development of bazedoxifene. Special thanks go to Drs. Doug Harnish and Istvan Merchanthaler for their work on gene transcription and the hot flush model, respectively, and Susan Jenkins for her work on MCF-7 cell proliferation.
Footnotes
Abbreviations: B.Ar, Bone mineral area; BMD, bone mineral density; BZA, bazedoxifene acetate; CHO, Chinese hamster ovarian cell line; CNS, central nervous system; DXA, dual-energy x-ray absorptiometry; ER, estrogen receptor; ERE, estrogen response element; FBS, fetal bovine serum; L4, fourth lumbar vertebrae; pQCT, peripheral quantitative computed tomography; PT, proximal tibiae; RAL, raloxifene; SERM, selective ER modulator; SRC, steroid receptor coactivator.
References
Lindsay R, Dempster DW, Jordan VC 1997 Estrogens and antiestrogens: basic and clinical aspects. Philadelphia: Lippincott-Raven
O’Connell MB 1995 Pharmacokinetic and pharmacologic variation between different estrogen products. J Clin Pharmacol 35(Suppl):18S–24S
Komm BS, Bodine PV 2001 Regulation of bone cell function by estrogens. In: Marcus R, Feldman D, Kelsey J, eds. Osteoporosis. 2nd ed. New York: Academic Press; 305–337
Cole MP, Jones CT, Todd ID 1971 A new anti-oestrogenic agent in late breast cancer. An early clinical appraisal of ICI46474. Br J Cancer 25:270–275
MacGregor JI, Jordan VC 1998 Basic guide to the mechanisms of antiestrogen action. Pharmacol Rev 50:151–196
Mourits MJE, de Vries EGE, Willemse PHB, ten Hoor KA, Hollema H, van der Zee AGJ 2001 Tamoxifen treatment and gynecologic side effects: a review. Obstet Gynecol 97:855–866
Gradishar WJ, Jordan VC 1997 Clinical potential of new antiestrogens. J Clin Oncol 15:840–852
Buzdar AU, Marcus C, Holmes F, Hug V, Hortobagyi G 1988 Phase II evaluation of Ly156758 in metastatic breast cancer. Oncology 45:344–345
Black LJ, Sato M, Rowley ER, Magee DE, Bekele A, Williams DC, Cullinan GJ, Bendele R, Kauffman RF, Bensch WR, Frolik CA, Termine JD, Bryant HU 1994 Raloxifene (LY139481 HCI) prevents bone loss and reduces serum cholesterol without causing uterine hypertrophy in ovariectomized rats. J Clin Invest 93:63–69
Sato M, Glasebrook AL, Bryant HU 1994 Raloxifene: a selective estrogen receptor modulator. J Bone Miner Metab 12(Suppl 2):S9–S20
McDonnell DP, Clemm DL, Hermann T, Goldman ME, Pike JW 1995 Analysis of estrogen receptor function in vitro reveals three distinct classes of antiestrogens. Mol Endocrinol 9:659–669
Shiau AK, Barstad D, Loria PM, Cheng L, Kushner PJ, Agard DA, Greene GL 1998 The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell 95:927–937
Glass CK, Rose DW, Rosenfeld MG 1997 Nuclear receptor coactivators. Curr Opin Cell Biol 9:222–232
Komm BS, Lyttle CR 2001 Developing a SERM: stringent preclinical selection criteria leading to an acceptable candidate (WAY-140424) for clinical evaluation. Ann NY Acad Sci 949:317–326
Van Den Bemd GJCM, Kuiper GGJM, Pols HAP, van Leeuwen JPTM 1999 Distinct effects on the conformation of estrogen receptor and by both the antiestrogens ICI 164,384 and ICI 182,780 leading to opposite effects on receptor stability. Biochem Biophys Res Commun 261:1–5
Lonard DM, Smith CL 2002 Molecular perspectives on selective estrogen receptor modulators (SERMs): progress in understanding their tissue-specific agonist and antagonist actions. Steroids 67:15–24
Harris HA, Bapat AR, Gonder DS, Frail DE 2002 The ligand binding profiles of estrogen receptors and are species dependent. Steroids 67:379–384
Harnish DC, Scicchitano MS, Adelman SJ, Lyttle CR, Karathanasis SK 2000 The role of CBP in estrogen receptor cross-talk with nuclear factor-B in HepG2 cells. Endocrinology 141:3403–3411
Bodine PV, Green J, Harris HA, Bhat RA, Stein GS, Lian JB, Komm BS 1997 Functional properties of a conditionally phenotypic, estrogen-responsive, human osteoblast cell line. J Cell Biochem 65:368–387
Harris HA, Katzenellenbogen JA, Katzenellenbogen BS 2002 Characterization of the biological roles of the estrogen receptors, ER and ER, in estrogen target tissues in vivo through the use of an ER-selective ligand. Endocrinology 143:4172–4177
Merchenthaler I, Funkhouser JM, Carver JM, Lundeen SG, Ghosh K, Winneker RC 1998 The effect of estrogens and antiestrogens in a rat model for hot flush. Maturitas 30:307–316
Miller CP, Collini MD, Tran BD, Harris HA, Kharode YP, Marzolf JT, Moran RA, Henderson RA, Bender RH, Unwalla RJ, Greenberger LM, Yardley JP, Abou-Gharbia MA, Lyttle CR, Komm BS 2001 Design, synthesis, and preclinical characterization of novel, highly selective indole estrogens. J Med Chem 44:1654–1657
Hadzopoulou-Cladaras M, Cardot P 1993 Identification of a cis-acting negative DNA element which modulates human hepatic triglyceride lipase gene expression. Biochemistry 32:9657–9667
Komm BS, Rusling DJ, Lyttle CR 1986 Estrogen regulation of protein synthesis in the immature rat uterus: the analysis of proteins released into the medium during in vitro incubation. Endocrinology 118:2411–2416
Ashby J, Odum J, Foster JR 1997 Activity of raloxifene in immature and ovariectomized rat uterotrophic assays. Regul Toxicol Pharmacol 25:226–231
Stearns V, Ullmer L, Lopez JF, Smith Y, Isaacs C, Hayes D 2002 Hot flushes. Lancet 360:1851–1861
Katovich MJ, Simpkins JW, Song IC, O’Meara J 1987 Effects of central application of naloxone on the skin temperature response in morphine-dependent rats. Brain Res Bull 19:81–85
Davies GC, Huster WJ, Lu Y, Plouffe Jr L, Lakshmanan M 1999 Adverse events reported by postmenopausal women in controlled trials with raloxifene. Obstet Gynecol 93:558–565
Katzenellenbogen BS, Katzenellenbogen JA 2002 Defining the "S" in SERMs. Science 295:2380–2381
Agnusdei D, Iori N 2000 Selective estrogen receptor modulators (SERMs): effects on multiple organ systems. Curr Med Chem 7:577–584
Burr ML 1987 Is asthma increasing J Epidemiol Community Health 41:185–189
Shang Y, Brown M 2002 Molecular determinants for the tissue specificity of SERMs. Science 295:2465–2468
Goldstein SR, Nanavati N 2002 Adverse events that are associated with the selective estrogen receptor modulator levormeloxifene in an aborted phase III osteoporosis treatment study. Am J Obstet Gynecol 187:521–527
Hendrix SL, McNeeley SG 2001 Effect of selective estrogen receptor modulators on reproductive tissues other than endometrium. Ann NY Acad Sci 949:243–250
Sundstrom SA, Komm BS, Ponce-de-Leon H, Yi Z, Teuscher C, Lyttle CR 1989 Estrogen regulation of tissue-specific expression of complement C3. J Biol Chem 264:16941–16947
Boss SM, Huster WJ, Neild JA, Glant MD, Eisenhut CC, Draper MW 1997 Effects of raloxifene hydrochloride on the endometrium of postmenopausal women. Am J Obstet Gynecol 177:1458–1464
O’Regan RM, Jordan VC 2001 Tamoxifen to raloxifene and beyond. Semin Oncol 28:260–273
Chlebowski RT, Col N, Winer EP, Collyar DE, Cummings SR, Vogel III VG, Burstein HJ, Eisen A, Lipkus I, Pfister DG 2002 American Society of Clinical Oncology technology assessment of pharmacologic interventions for breast cancer risk reduction including tamoxifen, raloxifene, and aromatase inhibition. J Clin Oncol 20:3328–3343
Cummings SR, Eckert S, Krueger KA, Grady D, Powles TJ, Cauley JA, Norton L, Nickelsen T, Bjarnason NH, Morrow M, Lippman ME, Black D, Glusman JE, Costa A, Jordan VC 1999 The effects of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. JAMA 281:2189–2197
Russo IH, Russo J 1998 Role of hormones in mammary cancer initiation and progression. J Mammary Gland Biol Neoplasia 3:49–61
Humphreys RC, Lydon J, O’Malley BW, Rosen JM 1997 Mammary gland development is mediated by both stromal and epithelial progesterone receptors. Mol Endocrinol 11:801–811
Komm BS, Kharode Y, Bodine PVN, Bex F 2003 Bazedoxifene+conjugated estrogens: a balanced combination to provide optimal "estrogenic" safety and efficacy. J Bone Miner Res 18(Suppl 2):SU385
Nelson HD 2004 Commonly used types of postmenopausal estrogen for treatment of hot flashes: scientific review. JAMA 291:1610–1620
Ettinger M, Schwartz E, Emkey R, Moffet AH, Bolognese M, Weiss SR, Lee A, Lasofoxifene, a next generation selective estrogen receptor modulator (SERM) in the prevention of bone loss in postmenopausal women. Program of the 86th Annual Meeting of the Endocrine Society, New Orleans, LA, 2004 (Abstract S35-2)
Ke HZ, Paralkar VM, Grasser WA, Crawford DT, Qi H, Simmons HA, Pirie CM, Chidsey-Frink KL, Owen TA, Smock SL, Chen HK, Jee WS, Cameron KO, Rosati RL, Brown TA, Dasilva-Jardine P, Thompson DD 1998 Effects of CP-336,156, a new, nonsteroidal estrogen agonist/antagonist, on bone, serum cholesterol, uterus and body composition in rat models. Endocrinology 139:2068–2076
Ma YL, Bryant HU, Zeng Q, Schmidt A, Hoover J, Cole HW, Yao W, Jee WS, Sato M 2003 New bone formation with teriparatide [human parathyroid hormone-(1–34)] is not retarded by long-term pretreatment with alendronate, estrogen, or raloxifene in ovariectomized rats. Endocrinology 144:2008–2015
Ettinger B, Black DM, Mitlak BH, Knickerbocker RK, Nickelsen T, Genant HK, Christiansen C, Delmas PD, Zanchetta JR, Stakkestad J, Glüer CC, Krueger K, Cohen FJ, Eckert S, Ensrud KE, Avioli LV, Lips P, Cummings SR 1999 Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene. Results from a 3-year randomized clinical trial. JAMA 282:637–645(Barry S. Komm, Yogendra P)