| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TAP Pharmaceutical Products, Inc. (K.C., M.C.P., D.D.), Lake Forest, Illinois 60045; School of Nursing (C.W.), Georgetown University, Washington, D.C. 20007; Jenapharm GmbH & Co. (G.S.), 07745 Jena, Germany; and EnTec GmbH (W.E.), 07745 Jena, Germany
Correspondence: Address all correspondence and requests for reprints to: Kristof Chwalisz, M.D., Ph.D., TAP Pharmaceutical Products, Inc., 675 North Field Drive, Lake Forest, Illinois 60045. E-mail: Kristof.Chwalisz{at}TAP.com
 |
Abstract |
|---|
|
I. Introduction |
|---|
Progesterone is an important mitogen in breast epithelial cells (5). Mitotic activity in normal breast tissue peaks during the luteal phase (6). Synthetic progestins clearly increase mammographic breast density (7), an effect that is accompanied by an increase in the expression of proliferation markers (8). Furthermore, continuous administration of estrogen/progestin regimens, but not estrogen treatment alone, was associated with a slight, but significant, increase in breast cancer risk as reported by the Women’s Health Initiative study and other clinical studies in postmenopausal women (9, 10).
Progesterone mediates its physiological effects through interaction with the PR, expressed in multiple tissues as two isoforms, hPR-A and hPR-B. These isoforms are derived from the same gene by the action of two different promoters (11, 12). The full-length hPR-B and N-terminus truncated hPR-A have highly conserved DNA and ligand binding domains (LBDs). Both isoforms have a similar architecture composed of the ligand-dependent activation function (AF)-2 present in the carboxyl terminus, and AF-1, a transcription domain present in the amino terminus (13). The constitutive AF-1 can function independently of AF-2 or with AF-2 in a ligand-dependent fashion. The LBD participates in interaction of the inactive receptor with heat shock proteins and immunophilins as well as promoting receptor dimerization (13, 14). A third activation domain, the AF-3 domain, is located in the upstream sequence region of hPR-B (13, 14). AF-3 is composed of approximately 164 amino acids and is present only in the hPR-B isoform. Functional evaluation studies of the AF-3 domain suggest that AF-3 mediates hPR-B transactivational activity, potentially through inhibition of the inhibitory domain common to hPR-A and hPR-B (13, 14). Within in vitro systems (cell-free preparations and cell lines), hPR-B is a much stronger activator of gene transcription. PR-A is an active transcription factor but is weaker than PR-B. Under conditions where PR-A is inactive as a transcription factor, it has the ability to repress PR-B and other steroid receptors in vitro (15, 16). The DNA binding domain contains two asymmetric zinc fingers and two 
-helices perpendicular to one another that facilitate interaction with the hormone response element present in PR-target genes. The domains present in the hormone receptor undergo spatial modifications to accommodate the ligand. The conformational changes induced by the ligand to the LBD help to orchestrate the receptor agonist or antagonist responses mediated by the receptor on PR-responsive genes (17, 18, 19, 20).
The functions of PR isoforms in vivo recently have been characterized, based on studies in mice with selective ablation of PR-A and PR-B (21, 22). Selective ablation of PR-A produced a phenotype characterized by infertility, severe endometrial hyperplasia, anovulation, and ovarian abnormalities in the presence of normal mammary gland response to progesterone. In contrast, mice with selective ablation of PR-B were fertile, did not show altered responses to progesterone in the uterus, but exhibited severely disrupted pregnancy-induced mammary gland morphogenesis.
The expression and ratios of PR-A and PR-B isoforms vary between normal and malignant tissues. In breast cancer cells with inducible PR-A or PR-B, progesterone stimulated gene regulation in a PR isoform-specific manner. The majority of the genes were uniquely regulated by PR-B, with smaller subsets regulated by PR-A or both isoforms (11, 23). It has recently been demonstrated by the Horwitz group (23) that PR gene regulation in vitro is further differentiated by whether or not the receptor is bound by the ligand. A majority of the genes regulated by the unliganded receptor were regulated by PR-A. The majority of the genes were uniquely regulated by PR-B with smaller subsets regulated by PR-A or both isoforms (11). Overall, both in vitro and in vivo studies indicate that PR-A and PR-B can profoundly affect the biological responses to progesterone and synthetic PR ligands in different manners (21, 22, 24). Because the hPR-A/hPR-B ratio can vary in different physiological and pathological situations, the ultimate response to the ligand may be determined by cell concentration of the specific isoform (25, 26).
The recognition of the important role of progesterone in reproduction led to the development of synthetic PR ligands with either agonist (progestins) or antagonist [PR antagonists (PAs)] properties. During the past 40 yr, numerous progestins have been synthesized. This class of compounds was crucial in the development of oral contraceptives, acknowledged as one of the most important medical breakthroughs of the last century. Progestins are widely used in the treatment of various gynecological disorders, including endometriosis and abnormal uterine bleeding, taking advantage of their antiproliferative effects on the endometrium. However, chronic use of progestins is associated with sometimes unacceptable side effects, including mood changes, depression, bloating, breakthrough bleeding, among others, that significantly limit their use (27).
The synthesis of mifepristone, the first glucocorticoid and PR antagonist (28), was a starting point of drug discovery and research programs in the area of PAs in several laboratories worldwide. Initially, this research was focused primarily on finding compounds with increased progesterone antagonistic potency and reduced antiglucocorticoid activity compared with mifepristone. Fertility control and treatment of breast cancer were viewed in the 1980s and early 1990s as the most promising indications for PAs. Since then, several 11ß-aryl substituted PAs with reduced antiglucocorticoid activity have been synthesized and characterized (29, 30, 31, 32). Steroidal PAs with a high degree of PR specificity, such as onapristone ("pure" PA) and ZK 137 316 ("almost pure" PA), were instrumental as pharmacological tools in defining the role of progesterone in various physiological processes, including ovulation, endometrial proliferation, implantation, uterine contractility, cervical ripening, and onset of labor (33, 34, 35, 36, 37, 38, 39). Although the PAs showed potential in the treatment of various gynecological disorders, none of these compounds has been developed for therapeutic indications other than termination of pregnancy and postcoital contraception because of various obstacles and undesirable effects on the endometrium (32, 40). Although low doses of mifepristone (2 or 5 mg) decreased endometrial proliferation (41), "unopposed" estrogenic effects including a high rate of endometrial hyperplasia (42, 43) were reported at higher mifepristone doses.
The selective PR modulators (SPRMs) have the potential to provide the beneficial effects of progestins and PAs while avoiding their drawbacks (Table 1
). The ideal SPRM would have antiproliferative effects on the endometrium and breast but would retain protective effects of ovarian estrogen on bones and cardiovascular system. It would not compromise ovarian estrogen production or induce irregular endometrial bleeding. In addition, it would exhibit neutral effects on the central nervous system and other systems. Furthermore, the ideal SPRM would suppress leiomyoma growth and the symptoms of endometriosis.
|
|
II. Terminology, Definitions, and Mechanism of Action of SPRMs |
|---|
More recently, a molecular definition of SERMs has been proposed based on their unique properties in cell-free systems (46). Because the molecular mechanisms of cell- and tissue-specific effects of steroid receptor ligands are similar within the family of nuclear receptors, the term selective receptor modulator (SRM) has recently been introduced (15). SRMs are now defined as a new class of nuclear receptor ligands that exhibit agonistic or antagonistic properties in a cell- and tissue context-dependent manner and should be distinguished from pure steroid receptor agonist and antagonists (15). These compounds change the conformation of the LBDs of nuclear receptors (47, 48) and influence their ability to interact with other proteins such as coactivators and corepressors. It has been postulated, therefore, that the relative balance of coactivators and corepressors within a given cell determines the relative agonist vs. antagonist activity of SRMs (15, 49, 50).
A. SPRM definition
The term SPRM (selective progesterone receptor modulator) has been proposed to describe the biological effects of PR ligands of 11ß-benzaldoxime-substituted estratrienes in keeping with the functional terminology adapted for SERMs (51, 52). Accordingly, SPRMs represent a class of PR ligands that exerts clinically relevant tissue-selective progesterone agonist, antagonist, or partial (mixed) agonist/antagonist effects on various progesterone target tissues in an in vivo situation depending on the biological action studied. This definition is consistent with the SRM definition proposed by Smith and O’Malley (15).
Animal and clinical studies discussed below clearly show that asoprisnil and other 11ß-benzaldoxime-substituted SPRMs exhibit partial agonist/antagonist effects in animals and humans and differ from PAs and progestins in a variety of models. In addition, asoprisnil demonstrates endometrial selectivity in nonhuman primates and humans. The PR specificity, i.e., the presence or absence of antiglucocorticoid activity, should not be included in the definition of a SPRM because some PAs do not significantly interact with the glucocorticoid receptor (GR).
B. 11ß-Benzaldoxime-substituted SPRMs
A focused drug discovery program was initiated at EnTec GmbH and Jenapharm GmbH (Jena, Germany) to identify PR ligands with particularly pronounced partial PR-agonist activity. The goal of this program was to identify compounds that maintain the beneficial properties of both progestins and PAs but are devoid of their negative effects. We were particularly interested in compounds exhibiting strong endometrial antiproliferative effects and absence of labor-inducing effects (53). A large number of compounds with high PR affinity (greater than progesterone), reduced GR affinity (lower than mifepristone), and mixed progesterone agonist/antagonist activity in vivo were identified and tested in the class of 11ß-benzaldoxime-substituted estratrienes (J-compounds) (53). The selection of candidate compounds was performed primarily in animal models, including cycling guinea pigs (luteolysis inhibition test), pregnant guinea pigs (induction of labor), and rabbits (McPhail test), because cell-specific in vitro systems capable of distinguishing PR agonist and antagonist activities were not available at the time (53). The luteolysis inhibition test was particularly important in this respect because of its simplicity and sensitivity in detecting progesterone agonist and antagonist effects using multiple endpoints (53, 54).
Using animal models, these compounds were classified according to their ratio of progesterone agonist vs. antagonist activities. A mosaic of progesterone agonist and antagonist effects was found in these models. The most agonistic compounds of this series were asoprisnil (J867), J1042, asoprisnil ecamate (J956), and J912 (Fig. 1). It became evident that there was an inverse correlation between the amount of agonist activity in the rabbit uterus (McPhail test) and labor-inducing activity in pregnant guinea pigs (53, 54). In most cases, similar inverse correlations were found between the McPhail test and the luteolysis inhibition assay in guinea pigs (54). Asoprisnil (J867) was selected for further development based on its pronounced progesterone agonist and antiproliferative effects and the absence of labor-inducing activity.
|
A nonsteroidal benzimidazole-2-thione analog (compound 25) with high PR selectivity was recently reported (59). In the immature rat model, compound 25 provided a significant suppression of estrogen-induced endometrium hypertrophy as measured by luminal epithelial height. In contrast, compound 25 was inactive in the LH release assay in young ovariectomized rats. The differential activities observed in the in vivo progestagenic assays in rat models suggest that compound 25 can act as a SPRM.
A series of 5-benylidene-1,2-dihydrochromeno[3,4-f]quinolines were recently synthesized and tested in bioassays to evaluate their progestational activities and receptor- and tissue-selectivity profiles (60). These compounds exhibited high (more than 100-fold) PR selectivity over other steroid hormone receptors. One of these compounds, LG120920, demonstrated tissue selectivity toward uterus and vagina vs. breasts in a rodent model after oral administration.
It was recently reported that dexamethasone (Dex)-mesylate and Dex-oxetanone, each a derivative of the GR agonist Dex, exhibit mixed antagonist/agonist activities with both the PR-A and PR-B isoforms for different progesterone-responsive elements and in several cell lines (61). Interestingly, Dex-oxetanone and Dex-mesylate exhibited PR agonist activity in cell lines in which other partial PR agonists were inactive. This observation suggests that these compounds promote interaction with specific coregulators and exhibit PR isoform specific effects.
D. Molecular basis of tissue selectivity of SPRMs
Although tissue-selective effects of certain steroid receptor ligands have been known for a relatively long time, it was not until the late 1980s and early 1990s that a more complete understanding of this effect emerged. Two discoveries were crucial in this respect: 1) findings that nuclear receptor agonists and antagonists induce different conformational changes to the PR (47, 62), and 2) the discovery of nuclear receptor interacting coactivators and corepressors, proteins that interact with the ligand-receptor complex bound to the promoter of the target gene to induce or inhibit transcriptional activation of the gene (67, 68).
In 1991, O’Malley’s group challenged the view that PAs, such as mifepristone, simply compete with the natural ligand progesterone and "freeze" the PR in an "inactive" state by inhibiting the ability of progesterone to activate the PR (63). This study showed that mifepristone did in fact activate the receptor but induced a conformation of the PR that was incompatible with transcription. Subsequent collaborative studies between the McDonnell and O’Malley laboratories defined the molecular basis of this antagonism (47, 64). Specifically, it was demonstrated that the extreme carboxyl tail of the receptor may adopt two different positions: one when occupied by an agonist and the other when a PA (e.g., onapristone or mifepristone) was bound (47, 48, 65). Both of these conformational changes rendered the PR distinct from that of unliganded PR. Using similar techniques, a third class of PR ligands in the series of 16-substituted analogs of mifepristone (i.e., RTI 3021–020) was described, which induces changes in receptor conformation that are distinct from those induced by agonists or antagonists (66). These compounds exhibited partial agonist activity in vitro in a cell-specific manner.
Subsequent discoveries of coactivators (67) and corepressors (68) determined the differences in the molecular mechanism of action of PR agonists and antagonists and thus provided a molecular explanation for tissue-selective action of SPRMs and other SRMs (15). Coregulators (coactivators and corepressors) are nuclear proteins that form multiple complexes with nuclear receptors and modulate their transcriptional activity (69, 70). It has been postulated that coactivators enhance transcriptional activity of nuclear receptors, whereas corepressors elicit inhibitory effects on nuclear receptors. Pure steroid receptor antagonists change the conformation of the steroid receptor such that it conversely favors interaction with corepressors or inhibits interactions with coactivators. In contrast, pure agonists promote the interaction of the nuclear receptor with coactivators. Finally, partial agonists induce an intermediate state of interaction between nuclear receptors and coactivators (15, 49, 71). This intermediate conformation allows nuclear receptors to interact to some degree with both coactivators and corepressors.
The family of p160 nuclear hormone receptor coactivators consists of three classes of steroid receptor coactivators (SRCs). The conformational changes induced by agonist ligands to the hormone receptor are associated with exposure of an amino acid interactive surface needed for binding of these coregulatory proteins to the nuclear DNA bound receptor dimmer (17). The coactivators have a binding site defined as the nuclear receptor box motifs, which consists of three conserved LXXLL amino acid regions. The nuclear receptor box motif region is responsible for binding of the coactivator to the LBD of the receptor (17, 72). The recruitment of coactivators to the DNA-receptor-ligand complex may promote additional recruitment of other coactivators to the complex to facilitate basal activation of the transcriptional machinery (17). Several studies have suggested that coactivators serve as interactive adaptors between the N-terminus AF-1 and the C-terminus AF-2 (73). The role of corepressors in steroid hormone antagonistic activity may be explained by failure of the ligand to induce conformational changes to the receptor that provides an active AF-2. Several studies have also suggested a direct interaction between the corepressor protein and the receptor (71, 74).
The relative balance of coactivator and corepressor expression within a given target cell determines the relative agonist vs. antagonist activity of SRMs (15). It is known that the availability of both coactivators and corepressors is dependent on tissue and hormonal milieu. Thus, cell type-specific and promoter-specific differences in coregulator recruitment determine the tissue selectivity of SRMs. To date, more than 50 coactivators have been cloned and characterized, and the list is still growing (15).
The same principles have been demonstrated for the PR. Conformational changes induced by PAs such as mifepristone, onapristone, or CDB-2914 permit the receptor to interact with the potent transcriptional repressors silencing mediator for retinoid and thyroid hormone receptors and nuclear receptor corepressor (50, 71). The Kate Horwitz group (71) showed for the first time that the direction of transcription by antagonist-occupied steroid receptors could be controlled by the ratio of coactivators to corepressors recruited to the transcription complex by promoter-bound receptors. It seems that all compounds investigated to date that induce antagonist (mifepristone-like) conformation of PR protein promote the recruitment of corepressors resulting in transactivation inhibition of target genes in vitro (49, 50, 66). Antagonist-activated PR is unable to interact with the coactivator proteins required to induce transcriptional activity on target genes. Because the corepressors and coactivators seem to be expressed in a tissue-specific manner, the pharmacodynamic effects of a given SPRM may be determined by tissue expression of both coactivators and corepressors (49, 69).
The classification of asoprisnil and other 11ß-benzaldoxime-substituted estratrienes as SPRMs was based primarily on their pharmacodynamic properties in vivo, i.e., partial agonist/antagonist effects and endometrial/uterine selectivity. Recently, we initiated molecular studies with these compounds that are ongoing. Preliminary results demonstrate that asoprisnil, in the absence of progesterone, induces minimal transactivation of the transiently transfected mouse mammary tumor virus-reporter gene and the endogenously expressed Sgk (serum and glucocorticoid responsive kinase) gene in T47D breast cancer cells expressing both PR isoforms. Asoprisnil induced recruitment of PR to the promoters of both genes and promoted recruitment of the coactivator SRC-1, a member of the p160 family of coactivators. In the presence of progesterone, asoprisnil inhibits the progesterone-induced transcription of these genes (D. Edwards, personal communication).
The agonist profile of asoprisnil in the absence of progesterone may be explained by conformational changes to the carboxy-terminal transcriptional activation domain of the PR. The conformational change induced by this molecule would expose the LXXLL motifs of the LBD of PR for recruitment of coactivator proteins. The ligand-receptor coactivator complex would then enhance the communication with basal transcriptional apparatus for activation of the target gene (17). In the presence of progesterone, the antagonist effect induced by asoprisnil may be explained by its binding to the PR followed by heterodimerization with progesterone bound to PR. The heterodimers are less effective at binding to the response element of the target genes, which would be translated into silencing of gene transcription (13). Figure 2 presents the hypothetical mechanism of action of SPRMs.
|
|
III. Reproductive Pharmacology of Asoprisnil and Structurally Related SPRMs |
|---|
B. PR-mediated effects in animal models
PR-mediated responses were studied in different animal models, including rabbits, guinea pigs, and nonhuman primates (52, 54, 75). In all of these models, partial progesterone agonist and antagonist effects were clearly evident.
1. PR-mediated effects in the rabbit uterus.
The classical McPhail test of endometrial transformation in estrogen-primed rabbits allows the assessment of the progesterone agonist and antagonist activity of a PR ligand (76). Progesterone and synthetic progestins stimulate both proliferation and differentiation of the rabbit uterine epithelium. The treatment-induced endometrial changes are histologically evaluated and semiquantitatively assessed using the McPhail scores on a scale of 0–4, with 4 being a maximal postovulatory response. Asoprisnil, J912, and several other 11ß-benzaldoxime-substituted SPRMs exhibited partial progesterone agonist and antagonistic effects, depending on whether progesterone was absent or present (Figs. 3, A and B, and 4) (52, 54). The PA mifepristone was used as one of the reference compounds and showed the most pronounced antagonist properties but no agonist activity in rabbits, even at high doses. In contrast, 11ß-benzaldoxime-substituted SPRMs clearly showed partial agonist effects in the absence of progesterone. Asoprisnil elevated McPhail scores in a dose-dependent manner; however, the agonist effects never reached the maximum response of progesterone. Such a "plateau response" is characteristic of a partial agonist. Similar observations were made with J1042 and J912 (54). In the antagonistic mode of the test (presence of progesterone), asoprisnil and other 11ß-benzaldoxime-substituted SPRMs showed relatively weak partial antagonist effects, and none of the compounds reached the inhibition level achieved with mifepristone, irrespective of the dose.
|
3. Animal models of pregnancy.
The effects of 11ß-benzaldoxime-substituted SPRMs on pregnancy were dependent upon the species investigated and stage of pregnancy. Although these compounds exhibited some inhibitory effects on implantation in rats, most likely by disrupting the delicate estrogen/progesterone balance and thus the decidualization process, they were marginally effective, or even ineffective, in influencing more advanced stages of pregnancy in guinea pigs (52).
The guinea pig (which is not a rodent) is probably the only relevant small animal model of human pregnancy and parturition (34, 38, 78). In guinea pigs, as in primates, the placenta produces progesterone during advanced pregnancy, and labor and parturition are brought about by the increase in uterine contractility and cervical ripening that occur in the presence of high progesterone levels. Importantly, pure PAs are very effective in inducing labor and parturition in guinea pigs, pointing to the possibility of "functional" progesterone withdrawal in this species, i.e., via PR-related mechanisms (39). This is in contrast to rodent models, in which resorption of the conceptus is frequently observed after treatment with antifertile agents, and parturition is the result of progesterone withdrawal. Asoprisnil, unlike the PAs mifepristone and onapristone (39), was marginally active in midpregnancy and completely ineffective in inducing preterm parturition in guinea pigs over a wide range of doses (52) (Fig. 5). Importantly, the tested doses (1 mg/animal and 10 mg/animal) were much higher than those producing an arrest of proliferation in the genital tract in nonpregnant guinea pigs (0.3 mg/animal; our unpublished data).
|
|
IV. Pharmacodynamic Effects of 11ß-Benzaldoxime-Substituted SPRMs in Nonhuman Primates |
|---|
shows representative histological section of the endometrium from control animals and an animal treated with asoprisnil for 90 d (79). Overall, these studies showed that 11ß-benzaldoxime-substituted SPRMs suppressed endometrial proliferation and induced amenorrhea in nonhuman primates due to an unknown mechanism. In monkeys, the antiproliferative effects were specific to the endometrium because there was no evidence of functional antiestrogenic effects in the vagina and oviducts (75). These experiments also provided the first evidence of the endometrial selectivity of asoprisnil and other 11ß-benzaldoxime-substituted SPRMs in a nonhuman primate model. Based on these studies, we hypothesized that these compounds may control both endometrial bleeding and proliferation via a vascular effect specific to the endometrium (75).
|
) (51). This observation indicates that asoprisnil, unlike progestins, has the potential to suppress mammary gland proliferation.
|
|
V. Metabolism and Pharmacokinetics of Asoprisnil |
|---|
Because J912 exhibits stronger antagonist activity and weaker agonist activity than the parent compound, it is likely that the overall effect in vivo depends on the balance between asoprisnil and J912.
|
VI. Pharmacodynamic Effects of Asoprisnil in Healthy Women |
|---|
The endometrial biopsies obtained from women treated with asoprisnil were consistent with the partial (mixed) progesterone agonist/antagonist activity of asoprisnil and showed asynchronous differentiation of endometrial epithelium and stroma. These appearances, which seem specific to 11ß-benzaldoxime-substituted SPRMs, were characterized by weak secretory effects on endometrial glands with no or infrequent mitotic figures and variable effects on endometrial stroma ranging from stromal compaction to focal predecidual changes. This study also revealed the formation of unusual, thick-walled arterioles in the endometrium of women exposed to asoprisnil (Fig. 8). Thick-walled vessels were consistently found in endometrial biopsies taken from subjects exposed to asoprisnil for 3 months and longer (our unpublished data). These vessels clearly differed from those typically observed in the endometrium from women treated with continuous progestins, which show patchy appearance of abnormally small and abnormally large, thin-walled, fragile vessels in the superficial endometrial areas (83, 84, 85).
|
|
VII. SPRMs in the Treatment of Uterine Leiomyomata |
|---|
A. Rationale
Although the initial steps in the pathogenesis of uterine fibroids are most likely due to chromosomal aberrations and/or the effects of specific genes (90), their development is highly dependent on ovarian steroid hormones. Traditionally, estrogen has been considered the major mitogenic factor in the uterus. However, there is growing evidence from biochemical, histological, clinical, and pharmacological studies indicating that progesterone and PR play a key role in uterine fibroid growth and development. Several investigators have shown an increased concentration of both PR isoforms (PR-A and PR-B) in leiomyoma tissue compared with adjacent myometrium (91, 92, 93). Furthermore, there was an increase in mitotic activity in fibroid tissue relative to the adjacent myometrial tissue during the luteal phase (94) and after treatment with medroxyprogesterone acetate (95). Increased expression of the proliferation marker Ki-67 in the leiomyoma compared with the normal myometrium has also been described, and its up-regulation was linked to progesterone (91). Epidermal growth factor (EGF) mRNA was increased in leiomyomata only during the secretory phase of the cycle in leiomyomata, suggesting that progesterone, not estrogen, controls the expression of this important growth factor (96). In addition, studies in vitro show that progesterone suppresses apoptosis and stimulates proliferation of leiomyoma cells (97, 98, 99, 100, 101). The apoptosis-inhibiting protein Bcl-2 was up-regulated in leiomyomata relative to the adjacent myometrium, and its expression was most abundant during the secretory phase of the menstrual cycle (100). Progesterone markedly increased Bcl-2 protein expression in primary leiomyoma cell cultures. In these cells, both E2 and progesterone increased the expression of proliferating cell nuclear antigen; whereas in cultures of normal myometrial cells only E2 had a similar effect. This study also showed that leiomyoma cells contained immunoreactive EGF and that progesterone treatment resulted in an increase in EGF expression in the cells, whereas E2 up-regulated EGF receptor (97).
Clinical studies with both progestins and the PA mifepristone also indicate that progesterone may be the more important hormone in fibroid growth compared with estrogen. The synthetic progestins, medroxyprogesterone acetate and norethindrone, when used as add-back therapy in combination with GnRH agonists (GnRH-a), attenuate or reverse the inhibitory effects of GnRH-a on leiomyoma size (102, 103), whereas mifepristone was shown, in small uncontrolled studies, to reduce leiomyoma volume (43, 104). Interestingly, the mifepristone effects were accompanied by a reduction in uterine blood flow (105), suggesting that progesterone plays an important role in the regulation of uterine perfusion.
B. Clinical studies with asoprisnil
The preliminary results of the phase II study with asoprisnil in subjects with uterine fibroids have been reported in abstract form (106, 107). In this multicenter, double-blind, placebo-controlled study, asoprisnil (5, 10, and 25 mg) was administered orally once daily for 12 wk. Uterine bleeding was assessed using a daily bleeding diary, monthly bleeding scores, and various hematological parameters. The volumes of the dominant fibroid and the uterus were measured by ultrasound. Asoprisnil significantly suppressed both the duration and intensity of uterine bleeding in a dose-dependent manner without inducing unscheduled bleeding. It also increased hemoglobin concentrations compared with placebo. In addition, asoprisnil treatment dose-dependently induced amenorrhea during the entire treatment period. Daily bleeding diaries showed a dramatic dose-dependent suppression of bleeding reflected in a significant reduction in average monthly spotting and bleeding scores. Similar effects were observed in women with menorrhagia at baseline. Asoprisnil also reduced the uterine volume and the volume of the largest leiomyoma in a dose-dependent manner, and at 10 and 25 mg significantly suppressed pressure symptoms (bloating and pelvic pressure), compared with placebo by wk 12. Asoprisnil did not decrease ovarian estrogen production in subjects with leiomyomata during 3 months of treatment. Consistent with the absence of antiglucocorticoid activity in humans at clinically relevant doses, there were no increases in serum concentrations of cortisol or dehydroepiandrosterone sulfate with asoprisnil (108). Asoprisnil was well tolerated during the 12 wk of treatment and the follow-up period. The most frequently reported adverse events were headache and abdominal pain. The adverse events were evenly distributed among all groups, including placebo. Large, phase III studies in subjects with menorrhagia associated with uterine fibroids are ongoing.
|
VIII. SPRMs in the Treatment of Endometriosis |
|---|
The major goal of the current medical treatment of women with endometriosis is to create an acyclic, hypoestrogenic environment either by blocking ovarian estrogen secretion (GnRH-a, or GnRH-antagonists), or by locally inhibiting estrogenic stimulation of the ectopic endometrium (progestins, androgenic progestins). Current medical therapies for women with endometriosis have severe side effects, such as the hypoestrogenic state induced by GnRH-a, breakthrough bleeding and mood changes from chronically administered progestins, and acne, hirsutism, and voice changes associated with androgens. Hence, a major objective of new therapies for women with endometriosis is to have fewer side effects while maintaining treatment effectiveness.
A. Rationale
SPRMs hold the potential of greater efficacy and flexibility than traditional treatments for endometriosis based on 1) selective inhibition of endometrial proliferation without systemic effects of estrogen deprivation, 2) reversible suppression of endometrial bleeding via a direct effect on endometrial blood vessels, and 3) the potential to suppress endometrial prostaglandin production in a tissue-specific manner (51).
Because the primary objective of any therapy for endometriosis is amelioration of pain, the effects of asoprisnil on endometrial production of prostaglandins are of particular importance for this indication. The endometrium-selective suppression of uterine prostaglandins by asoprisnil and structurally related SPRMs has been demonstrated in the guinea pig model (54, 117). In this species, prostaglandins of endometrial origin, produced in a pulsatile manner during the late luteal phase, are responsible for luteolysis; progesterone and PR tightly regulate these processes. Asoprisnil and some other 11ß-benzaldoxime-substituted SPRMs partially suppressed uterine prostaglandin F2 production and down-regulated uterine COX-2 expression in guinea pigs, without producing "unopposed" estrogenic effects in the genital tract. In humans, the corpus luteum is regulated differently, by an ovarian mechanism. However, the mechanisms for the synthesis of prostaglandins by the endometrium during the menstrual cycle might be similar in humans and guinea pigs, as suggested by studies in nonhuman primates (118) and one human study with mifepristone (119). The effects of asoprisnil on uterine prostaglandins and COX-expression are currently being evaluated in clinical studies.
B. Clinical studies with asoprisnil
Two randomized, placebo-controlled, dose-finding phase II studies of asoprisnil (0.5, 1.5, 5, 10, and 25 mg) have been conducted in subjects with pain from endometriosis. One of these studies has been reported in abstract form (120). Subjects with a laparoscopic diagnosis of endometriosis, exhibiting moderate or severe pain at baseline were treated with asoprisnil (5, 10, and 25 mg) or placebo for 12 wk. The effect of asoprisnil on daily pain was measured in subject diaries using a 4-point grading scale for three categories of pain (nonmenstrual pelvic pain, dysmenorrhea, and dyspareunia). An additional assessment of pain was made during monthly visits using a modified Biberoglu and Behrman 4-point pain scale. All three asoprisnil doses significantly reduced the average daily combined nonmenstrual pelvic pain/dysmenorrhea scores at all treatment months compared with placebo. All effective doses of asoprisnil showed similar effects on pain; however, the effect on bleeding pattern was dose-dependent. A separate study with an identical design using lower asoprisnil doses (0.5, 1.5, and 5 mg) showed that 5 mg is the minimum effective dose for pain relief in subjects with endometriosis (our unpublished data). Both studies also confirmed favorable safety and tolerability profiles of asoprisnil during short-term treatment. Adverse events were evenly distributed among treatment and placebo groups and were generally mild and self-limiting. No serious, drug-related adverse events were reported during treatment or follow-up period.
|
IX. Outlook and Concluding Remarks |
|---|
Asoprisnil, the first SPRM to reach an advanced stage of clinical development, was shown to induce a reversible amenorrhea, shrink uterine fibroids, and suppress endometriosis-associated pain without estrogen deprivation.
Suppression of menstruation irrespective of the effects on luteinization and progesterone withdrawal is a surprising finding that has not been previously described with any known pharmacological agent. This observation indicates that endometrium is its preferred target. The exact mechanism of the antiproliferative effects of asoprisnil on the endometrium is still under evaluation. The morphological data generated to date suggest that asoprisnil may directly or indirectly target the endometrial and uterine vasculature in a tissue-specific manner. Unusual, thick-walled endometrial arterioles are consistently observed in the endometrium of subjects treated with asoprisnil. These vessels clearly differ from those observed in many women using long-acting progestins that stimulate the formation of thin-walled microvessels responsible for the troublesome symptoms of breakthrough bleeding and spotting (121). Hence, the asoprisnil-induced morphological changes in endometrial vessels and perivascular stroma might be, at least in part, responsible for amenorrhea.
Similarly, the exact mechanism of uterine fibroid volume reduction induced by asoprisnil remains to be determined. We hypothesize that this effect may be due to a reduction in uterine blood flow. Clinical studies addressing this potential mechanism are currently under way. In addition, SPRMs may directly suppress proliferation of leiomyomata; preliminary results of studies with asoprisnil in primary leiomyoma cell cultures support this hypothesis (T. Maruo, personal communication).
The effects of SPRMs on the breast are of particular interest. Mammary gland development is predominantly postnatal and is controlled by a complex interplay between reproductive hormones E2, progesterone, and prolactin and local growth factors (5). There is growing evidence that progesterone is an important mitogen in the epithelial breast cells, which is in contrast to its inhibitory effects on the endometrial epithelial cells (122). PAs suppress mammary gland proliferation and inhibit growth of PR-positive mammary gland tumors in experimental settings, whereas progestins have an opposite effect in these models (123). Overall, both clinical and preclinical studies suggest that progesterone acts as a primary mitogen in the postmenopausal breast where E2 levels are much lower, whereas estrogen is the key proliferative factor in the premenopausal breast (5). Our toxicological studies in monkeys suggest that asoprisnil has a potential to inhibit breast proliferation (Fig. 6
). Additional studies in rats to further characterize the effects of asoprisnil on breast proliferation are ongoing.
Asoprisnil and other 11ß-benzaldoxime substituted SPRMs were discovered and selected for further development based on studies in various animal models. These models were instrumental in characterizing their partial progesterone agonist and antagonist activities, antiproliferative effects, and inhibitory effects on endometrial prostaglandins of these compounds. They also provided the first evidence of tissue-selective effects. More recent molecular studies suggest that asoprisnil-liganded PR promotes the recruitment of the coactivator SRC-1 (D. Edwards, personal communication) that could explain the partial agonist effects observed in animal models and humans.
Cloning and characterization of coregulator molecules provide a key to understanding the pharmacology of tissue-specific responses of hormones and SPRMs (15). These advances have the potential to explain the biological effects of SPRMs and to guide future drug discovery in this area.
Early short-term clinical studies with asoprisnil (phase I and II) demonstrated its favorable safety and tolerability profile. Long-term safety, however, still needs to be established in large, phase III studies that are still ongoing.
|
Top
Abstract
I. Introduction
, Progesterone; •, asoprisnil; 
, J912; 
, mifepristone. Data are presented as mean McPhail scores. The animals (n = 5–6/group) were primed with 5 µg E2 sc for 6 d, followed by treatment with the test compounds for 7 d in the absence (agonist activity) or presence (antagonist activity) of progesterone (1 mg/animal). A, Agonist activity. Progesterone produced a maximum agonistic effect reaching a plateau at 0.3 mg. Asoprisnil showed pronounced partial agonist effects with maximum effects below those of progesterone. J912 produced partial agonist effects that were weaker than those of asoprisnil. Mifepristone did not exert any agonist activity in this model. B, Antagonistic activity. Asoprisnil showed partial antagonist activity at doses above 3 mg/animal. Note that the antagonistic activity did not increase at higher doses. Mifepristone led to a dose-dependent inhibition of progesterone effect with a complete inhibition at more than 3.0 mg/animal. J912 produced partial antagonist effects that were more pronounced than those of asoprisnil. The maximum inhibition of a progesterone effect, as seen with mifepristone, was not observed with J912. [Reproduced from D. DeManno et al.: Steroids 68:1019–1032, 2003 (
