There is considerable ongoing investment in the research and development of selective progesterone receptor (PR) modulators for the treatment of gynecological conditions such as endometriosis. Here, we provide the first report on the clinical evaluation of a nonsteroidal progesterone receptor antagonist 4-[3-cyclopropyl-1-(mesylmethyl)-5-methyl-1H-pyrazol-4-yl]oxy,-2,6-dimethylbenzonitrile (PF-02413873) in healthy female subjects. In in vitro assays, PF-02413873 behaved as a selective and fully competitive PR antagonist, blocking progesterone binding and PR nuclear translocation. The pharmacological mode of action of PF-02413873 seems to differ from the founding member of the class of steroidal PR antagonists, 11β-4-dimethylaminophenyl-17β-hydroxy-17α-propinyl-4,9-estradiene-3-one (RU-486; mifepristone). Exposure-effect data from studies in the cynomolgus macaque, however, demonstrated that PF-02413873 reduced endometrial functionalis thickness to a comparable degree to RU-486 and this effect was accompanied by a decrease in proliferation rate (as measured by bromodeoxyuridine incorporation) for both RU-486 and high-dose PF-02413873. These data were used to underwrite a clinical assessment of PF-02413873 in a randomized, double-blinded, third-party open, placebo-controlled, dose-escalation study in healthy female volunteers with dosing for 14 days. PF-02413873 blocked the follicular phase increase in endometrial thickness, the midcycle lutenizing hormone surge, and elevation in estradiol in a dose-dependent fashion compared with placebo. This is the first report of translational efficacy data with a nonsteroidal PR antagonist in cynomolgus macaque and human subjects.
The physiological effects of progesterone are mediated by two nuclear receptor transcription factors, progesterone receptor (PR)-A and PR-B, which are produced from a single gene and upon binding progesterone regulate the expression of specific gene networks in reproductive tissues. Both PR null mutation and selective disruption of the PR-A isoform in the mouse leads to a failure of ovulation caused by disabled follicular rupture in response to gonadotrophin stimulation (Mulac-Jericevic and Conneely, 2005). Although there is still much to be delineated in the mechanisms by which PR controls the function of reproductive tissue and the hypothalamic-pituitary-ovary axis, because alterations in PR function also seem to contribute to pathological conditions such as endometriosis and cancer, there continues to be considerable interest in agents that modulate PR activity (Turgeon et al., 2004; Marx, 2006; Spitz, 2009; Kobayashi et al., 2010). As a class, steroidal progesterone receptor antagonists, such as 11β-4-dimethylaminophenyl-17β-hydroxy-17α-propinyl-4,9-estradiene-3-one (RU-486)/mifepristone, have been approved clinically for use in pregnancy termination and emergency contraception (Spitz, 2009). Studies with RU-486 and other steroidal PR antagonists, such as onapristone, ZK-137316, and 11β-(4-acetylphenyl)-17β-hydroxy-17α-(1,1,2,2,2-pentafluoroethyl)estra-4,9-dien-3-one (ZK-230211), have revealed that these agents, as well as directly antagonizing progesterone function, can block ovarian function and arrest the effect of estrogen on the endometrium in women and nonhuman primates (Wolf et al., 1989; Slayden and Brenner, 1994; Slayden et al., 2001a, 2006; Baird et al., 2003; Brenner and Slayden, 2005; Chabbert-Buffet et al., 2005; Narvekar et al., 2006). The broader clinical utility of RU-486 as a new approach to the treatment of endometriosis, uterine fibroids, and dysfunctional uterine bleeding, for instance, is limited because of antagonism at the glucocorticoid receptor (GR) (Heikinheimo et al., 1987) and effects on corticotropin secretion. Consequently, there has been considerable medicinal chemistry investment focused on identifying alternative chemical equity that has greater selectivity for PR over GR as well as other nuclear hormone receptors (NHRs) (Attardi et al., 2002; Jones et al., 2005; Terefenko et al., 2005; Kern et al., 2007, 2009, 2010; Zhang et al., 2007a, 2008; Fensome et al., 2008; Dack et al., 2010).
The identification of drug-like, potent, and selective PR antagonists has been challenging. As well as being highly lipophilic, the ligand binding sites between homologous NHRs are highly conserved. Once bound, ligands can induce different NHR conformations, which, depending on cell type and context, can affect the specificity of the downstream signaling events (Dai et al., 2008; Kobayashi et al., 2010; Zhang et al., 2010), leading to complex expressions of pharmacology. The cornerstone of clinical translation is the development of a preclinical screen sequence that can be used to predict clinical outcome. We have previously reported on the in vitro pharmacological profile and exposure/effect relationship of a novel potent nonsteroidal PR antagonist, 2-[4-(4-cyano-phenoxy)-3,5-dicyclopropyl-1H-pyrazol-1-yl]-N-methylacetamide (PF-02367982) (de Giorgio-Miller et al., 2008) in the rabbit and the macaque. The development of this molecule was curtailed before reaching clinical testing and was superseded by a related compound, 4-[3-cyclopropyl-1-(mesylmethyl)-5-methyl-1H-pyrazol-4-yl]oxy,-2,6-dimethylbenzonitrile (PF-02413873). Here, we describe the in vitro pharmacological characteristics of PF-02413873, comparing them with those of RU-486, and report on the comparative in vivo effect of PF-02413873 and RU-486 in the cynomolgus macaque. These data were used to underwrite a clinical evaluation, the first reporting efficacy of a nonsteroidal PR antagonist on the inhibition of endometrial growth in healthy female subjects.
Materials and Methods
PF-02413873 (Fig. 1) is an orally active nonsteroidal PR antagonist. The synthesis of PF-02413873 has been described elsewhere (Bradley et al., 2009). RU-486/mifepristone was purchased from Sigma-Aldrich (St. Louis, MO). The in vitro functional pharmacological properties of PF-02413873 for PR, androgen receptor (AR), GR, and mineralocorticoid receptor (MR) was determined as described previously (de Giorgio-Miller et al., 2008). CEREP (Poitiers, France) was used for supplemental wide receptor-profiling assay data and additional binding assays as described. Pamgene (Hertogenbosch, The Netherlands) was contracted to provide PR-LxxLL peptide binding data.
PR Nuclear Translocation Assay.
U2OS cells recombinantly expressing PR with a prolink tag and the nuclear restricted “enzyme acceptor” were purchased from DiscoveRx (Fremont, CA) and maintained in phenol red free minimal Eagle's medium supplemented with 10% (v/v) heat-inactivated fetal calf serum, 2 mM l-glutaMAX, 600 μg/ml Geneticin (G-418), and 250 μg/ml Hygromycin B. Cells were seeded in 384-well plates (10,000 cells, 20 μl/well) in phenol red free minimal Eagle's medium supplemented with 1% (v/v) charcoal/dextran-stripped fetal calf serum and 2 mM l-GlutaMAX and incubated overnight at 37°C. Diluent (5 μl; phosphate-buffered saline supplemented with 0.025% Pluronic acid F127 and 2.4% dimethyl sulfoxide) or serial dilutions of RU-486 (1 pM to 10 nM), or PF-02413873 (1 nM to 10 μM) were added to the cells alone or followed by serial dilutions of progesterone (5 μl; 1.58 pM to 158 nM) in diluent. Cells were incubated at 37°C for 3 h to enable nuclear translocation; after which DiscoveRx PathHunter reagents (15 μl) were added according to the manufacturer's instructions 60 min before reading on a luminescence counter (EnVison; PerkinElmer Life and Analytical Sciences, Waltham, MA).
Analysis of Plasma Samples for PF-02413873 and RU-486.
Plasma concentrations of RU-486 and PF-02413873 were determined in 50-μl aliquots of thawed macaque plasma using tert-butyl methylether extraction followed by high-performance liquid chromatography and on-line tandem mass spectrophotometric analysis. Plasma protein binding of PF-02413873 was determined by equilibrium dialysis, and unbound plasma concentrations were then calculated using values for “free fractions” of 0.046 and 0.031 for cynomolgus macaque and human, respectively. The free fraction value for RU-486 was 0.028 for cynomolgus macaque (de Giorgio-Miller et al., 2008).
Human plasma samples were analyzed for PF-02413873 concentrations at GVK Biosciences Ltd (Hyderabad, India) using a validated analytical specific liquid chromatography-tandem mass spectrophotometry assay. Specimens were stored at approximately −20°C until analysis and assayed within the 154 days of established stability data generated during validation. Calibration standard responses were linear over the range of 1 to 5 ng/ml, and the lower limit of quantification for PF-02413873 was 3 ng/ml. Assay precision, expressed as the between-day coefficient of variation (%) of the mean estimated concentrations of quality-control samples, was 6.2% for low (3 ng/ml), medium (249 ng/ml), and high (400 ng/ml) concentrations.
Evaluating the Effects of PF-02413873 on the Naturally Cycling Cynomolgus Macaque Endometrium.
The in-life phase of the study was performed at Covance (Münster, Germany) in cynomolgus macaques (weight range 3.7–5.7 kg and ages 5–6 years) previously used to assess the effects of PF-02367982 (de Giorgio-Miller et al., 2008). The study was approved by local and Covance ethics review boards and conducted in accordance with local animal husbandry procedures and legislation. A chronic dosing study was supported with a preliminary single-dose oral pharmacokinetics study from which doses for the pivotal efficacy study were extrapolated. For this, 20 sexually mature female cynomolgus macaques underwent daily menstrual cycle inspection by morning examination of external genitalia and vaginal smears. Menstrual bleeding was checked by inserting a cotton bud into the vagina. Animals were allowed to complete an observation cycle before any surgical intervention or drug administration. All animals completed a normal menstrual cycle within 35 days. The first day of the next observation cycle was deemed the first day of menstruation and the first day of dosing. Animals were dosed with vehicle [2% (w/v) hydroxyprolyl cellulose, 0.1% (v/v) Tween 80, and 0.1% (w/v) sodium lauryl sulfate in water], PF-02413873 (2.5 and 10 mg/kg b.i.d), and RU-486 (20 mg/kg q.i.d.) in a vehicle suspension by oral gavage (n = 5 per dose group). The doses were selected after a pilot single dose (3 mg/kg) PK study and by projection from a previous study with RU-486 and the related nonsteroidal progesterone receptor antagonist PF-02367982 (de Giorgio-Miller et al., 2008) with the objective of targeting 0.5 and 2 nmol · h/ml PF-02413873 exposures. Animals were dosed for 10 days and 1 h before euthanasia, and approximately 4 h after the final drug dose, the animals were treated with bromodeoxyuridine (BrdU; 100 mg/animal i.v.). Drug exposure levels were determined by pharmacokinetic sampling on days 1, 5, and 10 of dosing.
Multiple sections of vagina, cervix, and uterus were embedded in paraffin before sectioning at 4 μm and staining with hematoxylin and eosin using the Leica (Wetzlar, Germany) 5030 auto-stainer. A representative section of endometrium, which had the full thickness of endometrium and where possible longitudinal sections of endometrial glands, was chosen for evaluation. Sections were assessed for degree of reduction in endometrial thickness relative to negative control. A virtual 0.63× image was used to measure the endometrial thickness. All measurements were performed using Image Pro-Plus (Media Cybernetics, Inc., Bethesda, MD). The border of endometrium and myometrium was manually delineated using the measurements line drawing tool within Image Pro-Plus. Between 6 and 15 representative distances (from luminal epithelial surface to the endometrial/myometrial border) were captured per sample. Whenever possible, all measurements followed the gland direction. The basalis and functionalis zones were manually delineated using the measurements line drawing tool within Image Pro-Plus, and representative thickness measurements were drawn along glands. Between 6 and 10 representative measurements were taken for each sample.
Immunohistochemical Assessment of BrdU and Androgen Receptor Expression.
The paraffin blocks containing the optimal endometrial morphology were chosen and stained for BrdU and AR by using a Ventana XT (Ventana Medical Systems, Tucson, AZ). A primary rat anti-BrdU antibody (1/100) was purchased from Abcam Inc. (Cambridge, MA) and developed with a rabbit anti-rat biotin (1/200) from Vector Laboratories (Burlingame, CA). The DabMap kit used was manufactured by Ventana Medical Systems following protocol number 136. Sections were stained for the androgen receptor using a rabbit anti-human AR (1/100) primary antibody from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) and donkey anti-rabbit biotin (1/200) secondary from Jackson ImmunoResearch Laboratories Inc. (Suffolk, UK). Slides were dehydrated through graded alcohols into xylene and coverslipped using a Leica CV5030. All slides were cleaned, and digital images were produced using a Nanozoomer digital slide scanner (Hamamatsu Corporation, Bridgewater, NJ) at 20× magnification. BrdU-positive nuclei were counted manually in a blinded fashion, and the proportion of positive-to-negative nuclei was recorded for between five and nine fields (equivalent to approximately 300–700 nuclei). For AR expression, Image Pro-Plus was used to identify total areas of positive- and negative-stained nuclei per field of view. The optimum red-green-blue threshold level was chosen per stain and remained constant for each sample. The mean intensity of positive and negative nuclei was also recorded per sample. This was used to calculate the available (dynamic) range of intensities for each sample. A corrected intensity (the average intensity of positive area) was calculated, and from this, an androgen receptor index was determined (= corrected intensity × positive area/total area), where corrected intensity is the average intensity of positive area (on increasing scale of 0–100) determined by the following: ((255 − mean brown staining) − (255 − mean blue staining))/(dynamic range/100), where dynamic range = 210 − (255 − mean blue staining).
Nuclei area was determined from the pixel size of the region of interest, multiplying by the relevant objective magnification.
All statistical analysis was performed using Prism 5 (GraphPad Software Inc., San Diego, CA). Statistical significance (p ≤ 0.05) was determined using two-way analysis of variance with a Bonferroni adjustment for multiple comparisons.
Clinical Evaluation of PF-02413873 Given Chronically to Women of Child-Bearing Potential.
The study was managed by Pfizer Global Research and Development (the sponsor) and conducted by investigators contracted by and under the direction of the sponsor. The final protocol, its amendments, and informed consent documentation were reviewed and approved by the Institutional Review Board and/or Independent Ethics Committee at each of the investigational centers participating in the study. This study was conducted in compliance with the ethical principles originating in or derived from the Declaration of Helsinki and in compliance with all International Conference on Harmonization Good Clinical Practice Guidelines. In addition, all local regulatory requirements were followed; in particular, those affording greater protection to the safety of study participants. The study was conducted at Pfizer Clinical Research Units in Brussels, Belgium, and New Haven, CT and the Vince and Associates Clinical Research Centre, Overland Park, KS. Medical clinical monitoring was conducted by the sponsor or its designated representatives. Study drug was packaged, labeled, and shipped by the study sponsor. Written informed consent was obtained from all participants in this study before screening. This was a double-blind, third-party open, randomized, placebo-controlled, parallel group, oral dose-escalation study in healthy women of child-bearing potential, aged 18 to 35 with a regular menstrual cycle, with at least one functioning ovary and a normal endometrium. Subjects had to be willing to have their cycle synchronized with the combined pill for at least 2 weeks (and up to 8 weeks) until 7 days before randomization and, if nonabstinent, have a male partner who was willing to use adequate contraception (Fig. 2). Subjects were excluded if they had an abnormal papanicolaou smear, chlamydia, or gonorrhea, or if on screening transvaginal ultrasound any of the following were detected: uterine fibroids >3 cm in diameter, which distorted the uterine cavity, uterine polyp, hydrosalpinx or tubo-ovarian mass, or benign or malignant ovarian mass (including functional ovarian cysts >4 cm).
Doses were selected on the basis of a previous preliminary single-dose safety and toleration dose-escalation study. A sample size of 10 subjects per cohort (4:1 active/placebo) for four cohorts was planned to allow detection of a difference of 5 mm in endometrial thickness with >80% power at a two-sided significance level of 5%. The calculation was based on assumptions that endometrial thickness is normally distributed, the between-subject S.D. estimated to be 3.2 mm based on published data (Premkumar et al., 2007), and the data would be pooled across the cohorts for the statistical analysis. Cohorts were run sequentially, and within each cohort subjects were randomized to treatment. Each dose cohort was intended to be dosed for 14 days (hereafter referred to as the treatment period) and then observed for another 14 days (hereafter referred to as the observation period). Initially for each cohort, 10 subjects were randomized to treatment, and subjects withdrawing for reasons unrelated to safety and toleration were replaced. The dose cohorts were as follows: cohort 1, eight subjects received 20 mg of PF-02413873 and three subjects received placebo; cohort 2, nine subjects received 100 mg of PF-02413873 and two subjects received placebo; cohort 3, eight subjects received 500 mg of PF-02413873 and two subjects received placebo, dosing was stopped on day 11 because of the occurrence of rash-related adverse events in four subjects on active treatment; however, subjects continued in the study and other procedures were completed as planned; cohort 4, eight subjects received 250 mg of PF-02413873 and two subjects received placebo.
Four subjects (one on placebo, one each on doses of 100 mg, 250, and 500 mg of PF-02413873) were discontinued from the study and as such had incomplete data. Endometrial thickness and all other data were collected for all other subjects.
PF-02413873 or placebo was administered as an extemporaneous preparation with a standard meal. The safety/tolerability and steady-state pharmacokinetic data of PF-02413873 (up to 24 h after the last dose) were reviewed after each cohort. Each dose escalation was based on safety (including determination of potential prolongation of corrected QT interval), tolerability, and pharmacokinetic data from the previous dose.
Blood samples were collected for confirmation of PF-02413873 exposure and the determination of plasma hormone concentrations during the treatment and observation periods (days 0–24). The clinical laboratory sample analyses were performed at the local laboratories. All pharmacokinetic samples were sent to SGS Cephac Europe (Paris, France) for urine analysis and GVK Biosciences Ltd for plasma analysis.
Real-time ultrasonography was conducted with a 5.0- to 7.5-MHz vaginal transducer predose (usually day −1) and on day 13 (notional). In brief, the vaginal probe was covered with a coupling gel and inserted into a condom, which was coated with gel and inserted into the vaginal fornix, with the subject in the lithotomy position. During the ultrasound examination particular attention was paid to the uterus and ovaries. After obtaining a proper longitudinal view of the uterus, the uterus was observed for a few minutes during which the endometrium was continuously focused on, and any wave-like contractions in the subendometrial layer were observed. A still image of the uterus was captured between contractions, and the endometrium was arbitrarily divided into three roughly equal portions for the purposes of measurement: an upper third of the endometrial stripe at the uterine fundus, middle third between upper and lower portions, and a lower third that was adjacent to the cervix. In each of the three regions of the uterus, the following measurements (to the nearest 0.1 mm) were obtained in the relaxed phase between any subendometrial or myometrial contractions: 1) antero-posterior thickness of any pool of intrauterine fluid (measurement A); 2) the endometrial thickness (obtained by placing electronic calipers at the anterior and posterior uterine walls at the margins of the basal layers of the endometrium delineated by the highly echogenic interface between endometrium and inner myometrium) (measurement B); the true endometrial thickness was measured as B − A; and 3) the longitudinal, transverse, and antero-posterior length of the uterus.
Endometrial thickness was measured on a midline sagital image of the uterus and was a summation of the antero-posterior width of both the anterior and posterior endometrial layers, exclusive of possible intracavitary content. For each region, true endometrial thickness was analyzed by using analysis of covariance with baseline (day −1) as a covariate. From the analysis of covariance, the difference in treatment means and the 95% confidence interval (C.I.) for the difference were presented.
The daily estradiol and LH plasma concentrations in individual subjects dosed with PF-02413873 or placebo were used to confirm ovulation. The presence of an LH peak (defined as >1.5 × baseline LH) after a rise in estradiol to >200 pg/ml (visual inspection by an expert reviewer) was taken to signify ovulation. Where there was no LH peak present, the value recorded on the mean peak day was used, where mean peak day was the mean day of LH peak for those subjects with a peak.
In Vitro Pharmacological Profile of PF-02413873.
PF-02413873 was identified from a drug discovery campaign for agents that could block PR signaling function. PF-02413873 was assessed for its ability to block an in vitro native progesterone response in a human T47D mammary carcinoma cell-based functional reporter gene assay (de Giorgio-Miller et al., 2008) and block binding to PR. PF-02413873 blocked radioligand binding to PR in a CEREP MCF-7 cytosol binding assay with a Ki value of 2.6 nM (Table 1). PF-02413873 showed potent PR antagonist activity with a derived Ki value of 9.7 nM (95% C.I. 7.3 −13.0 nM; n = 9; Table 1) in the T47D native functional assay. Although PF-02413873 did not seem to elicit a PR agonist response at concentrations below 3 μM, at concentrations higher than this PF-02413873 seemed to induce a partial PR agonist response, an effect that has been observed with other nonsteroidal PR-As (Zhang et al., 2007b). To explore this further, Schild experiments were conducted with Lew and Angus nonlinear regression analysis (Lew and Angus, 1995) where dextral displacement of a progesterone concentration response curve was observed for both the steroidal antiprogestin RU-486 and PF-02413873 (Fig. 3) with pKB values of 10.5 (95% C.I. 10.4–10.5; n = 3) and 8.0 (95% C.I. 7.9–8.0; n = 3), respectively. In contrast to RU-486, at high doses of PF-02413873 there was an apparent increase in the basal response, indicative of the agonism of PF-02413873 observed at high concentrations in the absence of progesterone. Similar data were generated in a recombinant β-lactamase reporter cell line where the effects PF-02413873 on the progesterone-induced response on PR are coupled to a mouse mammary tumor virus promoter (de Giorgio-Miller et al., 2008), suggesting the observed pharmacological response was not caused by the T47D cell line (data not shown). When the effect of RU-486 and PF-02413873 was assessed in an enzyme complementation assay that measures PR nuclear translocation, RU-486 induced nuclear translocation consistent with its functional antagonist potency (Fig. 4A) and did not alter the progesterone concentration response (Fig. 4B). In contrast, PF-02413873 did not induce nuclear translocation at antagonist concentrations, but seemed to block progesterone-induced nuclear translocation (Fig. 4C). Only at high concentrations (>3 μM) was there an apparent facilitation of nuclear translocation (Fig. 4A), consistent with the observed agonism that PF-02413873 seems to induce in other systems (Fig. 3).
We explored the nature of the PF-02413873 interaction in more detail by evaluating the interaction between a recombinantly expressed glutathione transferase-tagged PR ligand binding domain and suite of immobilized peptides covering the LxxLL motifs of all the major NHR coactivators/corepressors in response to increasing concentrations of progesterone, PF-02413873, and RU-486. Representative examples of the ligand-induced interactions between PR and LxxLL peptides from NCoA-1 and NCoR-1 are depicted in Fig. 5. In the system, progesterone induced interactions between PR and the coactivator NCoA-1 peptide and reduced interactions with a peptide raised against the LxxLL motif on the corepressor NCoR-1 peptide in a concentration-dependent manner. This profile was broadly mimicked in the PF-02413873 concentration effect response, although both the magnitude and potency of the effect were considerably reduced compared with progesterone (Fig. 5). In contrast, RU-486 facilitated PR recruitment to the NCoR-1 peptide and reduced PR affinity for the NCoA-1 peptide in a concentration-dependent fashion.
PF-02413873 selectivity was evaluated in recombinant functional reporter assays expressing the homologous nuclear hormone receptors, GR, AR, and MR as well as against a panel of broad receptors, enzymes, and ion channels available at CEREP (Table 1). In these systems, PF-02413873 demonstrated more than 30-fold selectivity for PR.
In Vivo Effect of PF-02413873 in Naturally Cycling Cynomolgus Macaques.
PF-02413873 doses of PF-02413873 (2.5 and 10 mg/kg p.o. b.i.d.) and RU-486 (20 mg/kg p.o. q.i.d.) were selected based on previous experience (de Giorgio-Miller et al., 2008) and pilot PK data (Table 2). The effects of PF-02413873 and RU-486 on estrogen-induced endometrial growth were confirmed by histology on samples of uterus obtained from cynomolgus macaques dosed with vehicle, RU-486 (20 mg/kg q.i.d.), and PF-02413873 (2.5 and 10 mg/kg b.i.d.) for 10 days from the start of the menstrual cycle. Both RU-486 and PF-02413873 induced a statistically significant reduction in endometrial thickness compared with vehicle control animals (Figs. 6 and 7). The cynomolgus and human PR primary amino acid sequence are highly homologous (data not shown), and the PF-02413873 exposures achieved <10× the multiple of the human Ki value (Table 2). The mean percentage change in endometrial thickness in animals dosed with PF-02413873 (2.5 and 10 mg/kg p.o. b.i.d.) or the steroidal progesterone receptor antagonist RU-486 (20 mg/kg p.o. q.i.d.) was −43, −56, and −34%, respectively, compared with a vehicle control (Fig. 7). The effect on endometrial thickness seemed to be caused by specific changes in the thickness of the functionalis compartment, because there were no observed statistically significant changes in basalis thickness. To correlate the change in endometrial thickness observed in cynomolgus macaques treated with RU-486 and PF-02413873 with changes in proliferation rate, samples of endometrium were stained for BrdU accumulation. Positive nuclei in luminal epithelium as well as functionalis glandular epithelium and stromal cell compartments were counted for each group. RU-486 induced a specific statistically significant reduction in BrdU accumulation in the functionalis stromal compartment by 88% compared with the vehicle. With the caveat of the single time point of evaluation, the effects of PF-02413873 were in contrast to the effects of RU-486 and only the highest dose of PF-02413873 tested (10 mg/kg b.i.d.) induced a statistically significant reduction in BrdU incorporation of 43% compared with the vehicle control group (Fig. 7B).
The effects of both RU-486 and PF-02413873 on the anticipated increase in endometrial AR expression (Slayden et al., 2001b; Brenner et al., 2002, 2003; Slayden and Brenner, 2004) were similarly determined by semiquantitative immunohistochemistry. AR-positive nuclei in the functionalis stroma were counted for each group. RU-486 induced an approximate 2-fold increase in AR expression compared with vehicle. In contrast, despite similar effects of PF-02413873 on endometrial thickness as RU-486, PF-02413873 was without apparent effect on endometrial AR expression compared with the vehicle control for all doses tested (Table 3).
Effect of PF-02413873 on Endometrial Thickness in Women.
Using a similar methodology to that used in the macaque, the effects of PF-02413873 on endometrial growth during the early proliferative phase were examined in healthy women of child-bearing potential. This was a double-blind, third-party open, randomized, placebo-controlled, parallel group study evaluating escalating multiple doses of PF-02413873 (20, 100, 250, and 500 mg) or placebo given once a day for 14 days. Dosing in the 500-mg cohort was stopped after the day 10 dosing because of the occurrence of rash in four subjects on active treatment; however, subjects continued in the study and other procedures were completed as planned. A single incidence of muculopapular rash was also reported in the 250-mg dose cohort.
Plasma concentrations of PF-02413873 were quantifiable after all administered doses (Table 4). After single- and multiple-dose oral administration, maximum concentrations of PF-02413873 were achieved between 3 and 5 h on days 1 and 14. The mean Cmax of the highest dose tested was approximately 10 times the human Ki value (Table 4). The estimated t½ ranged between 34 and 48 h across the dose range, and plasma concentration-time profiles exhibited at least a biexponential decline over time (data not shown). PF-02413873 showed a slightly less than dose-proportional increase of Cmax, after both single and multiple dosing. For a 5-fold increase in dose (20–100 mg) there was a 4-fold increase in Cmax, and for a 25-fold increase in dose (20–500 mg) there was a 15-fold increase in Cmax. PF-02413873 accumulation was less than 2-fold with q.i.d. oral dosing (Table 4).
The main efficacy endpoints were a change in endometrial thickness compared with the placebo control group measured by transvaginal ultrasound and change in ovarian hormone and LH responses over the period of dosing. Three different zones of the endometrium were assessed by ultrasound. At the highest dose tested PF-02413873 induced a reduction in the thickness of the uterine fundus and middle third regions (Fig. 8; Table 5), and these differences were statistically significant at the 5% level. This effect was more marked for the uterine fundus data where the difference was 5.4 mm (95% C.I., 2.5, 8.4). The differences in endometrial thickness observed at the lower doses of PF-02413873 (20, 100, and 250 mg) were smaller compared with placebo (Table 5). PF-02413873 had no inhibitory effect on the lower-third region compared with subjects receiving placebo (Fig. 8C).
The effects of PF-02413873 on plasma LH and estradiol levels were determined by sampling during the dosing period and washout phase for all PF-02413873-dosed subjects and compared with placebo. The data were compiled and centered around the midcycle LH surge (Fig. 9). The peak plasma concentrations of LH and FSH (data not shown) were lower in subjects treated with PF-02413873 doses compared with placebo (Fig. 9A). This reduction in peak plasma LH concentrations was most marked in the PF-02413873 highest-dose (500 mg) group compared with placebo (Fig. 9B). Peak plasma estradiol concentrations (Fig. 9C) followed a similar pattern as the plasma LH profiles, a clear suppression in the individuals treated with the highest dose of PF-02413873. Ovulation was defined as the presence of an LH peak (defined as >1.5× increase over baseline LH) after a rise in estradiol to >200 pg/ml. Inhibition of ovulation response rate is summarized in Table 6. There was an apparent dose-dependent inhibition of ovulation. Inhibition of ovulation was highest (85%) for the PF-02413873 500-mg dose and persisted after cessation of dosing (Table 6). Both suppression of endometrial growth and ovarian function were achieved with PF-02413873 plasma exposures that were reasonable multiples of the primary pharmacology (Table 4).
The role of progesterone beyond its function as a pregnancy hormone has evolved only in recent years. Studies in knockout mice and by pharmacological modulation in nonhuman primates have revealed a direct role for the progesterone receptor in ovarian function and endometrial growth (Conneely et al., 2001; Slayden et al., 2001a; Brenner et al., 2010), suppressing the proliferative effects of estrogen on the endometrium (Wolf et al., 1989; Hodgen et al., 1994; Slayden and Brenner, 2004; Slayden et al., 2006; Brenner et al., 2010).
As a consequence, there has been considerable interest in the class of PR antagonists, typified by agents such as RU-486, as alternative contraceptives and for the treatment of gynecological conditions such as endometriosis and uterine fibroids (Spitz, 2009). The development of selective and safe steroidal PR antagonists has been challenging, both because of reported hepatoxicity and potential dose-limiting antiglucocorticoid effects. More recently, histological evaluations of subjects dosed for more than 3 months on steroidal PR antagonists seem to induce a characteristic cystic histological change in the endometrium that may be difficult to distinguish from endometrial hyperplasia without specialist evaluation (Williams et al., 2007; Mutter et al., 2008; Ioffe et al., 2009). Building on our previous experience (de Giorgio-Miller et al., 2008), we have identified and characterized a novel and selective nonsteroidal PR antagonist, PF-02413873.
PF-02413873 exhibited potent PR antagonism in human in vitro functional and binding assays and more than 30-fold selectivity over closely related members of the nuclear hormone receptor family as well as other enzymes, receptors, and ion channels (Table 1). We characterized the nature of the PF-02413873 interaction with PR in more detail by Schild analysis and assessment on nuclear translocation and LxxLL peptide binding compared with RU-486. RU-486 behaved in these assays as anticipated from previous literature reports, facilitating PR translocation and antagonizing PR function though recruitment of corepressors (Bocquel et al., 1993; Madauss et al., 2007; Afhüppe et al., 2010). In contrast, PF-02413873 blocked PR nuclear translocation at concentrations that blocked PR function in the T47D functional assay. At suprapharmacological concentrations, however, it induced a nuclear translocation and partial agonist activation of the receptor, observations that were matched with a recruitment of coactivator at these concentrations. When taken together, these data suggest that PF-02413873 has mixed antagonist-agonist pharmacology. At concentrations that block PR function, PF-02413873 behaves as a neutral antagonist, inhibiting progesterone binding and nuclear translocation in a competitive manner. At suprapharmacological concentrations, PF-02413873 may induce a conformational change in the PR complex to facilitate nuclear translocation and partial agonism. These pharmacological properties of PF-02413873 and other nonsteroidal examples we have profiled (data not shown) seem to contrast with the class of steroidal antiprogestins exemplified by RU-486.
We sought to determine whether the action of PF-02413873 could antagonize PR in vivo by studying the effects in pregnant rats and rabbits (data not shown) as well as on endometrial thickness in the intact macaque. In pregnant rats and rabbits, PF-02413873 induced the dose-dependent and complete resorption of fetuses (data not shown), supporting its credentials as a PR antagonist. When macaques were dosed with vehicle, RU-486, or PF-02413873 from the first day of menstruation, both RU-486 and PF-02413873 induced a statistically significant reduction in the thickness of the endometrial functionalis layer compared with vehicle control animals. In contrast, the basalis thickness was unaffected, an observation that is consistent with other reports (Wolf et al., 1989; Slayden and Brenner, 1994; Slayden et al., 1998; Greb et al., 1999; Brenner et al., 2010). The effects on endometrial thickness were also coupled to a reduction in proliferation rate at the highest PF-02413873 dose tested compared with the vehicle control group. Although this may, in part, be caused by the doses of PF-02413873 only achieving a modest multiple of Ki compared with that achieved the RU-486 (Table 2), it can not be ruled out that PF-02413873 achieved its effects on the endometrium by a different mechanism to RU-486. To this latter point, one of the anticipated pharmacological consequences of PR antagonism is an up-regulation of AR expression in the endometrium, especially when assessed during luteal phase sampling (Slayden et al., 2001b; Brenner et al., 2002, 2003; Narvekar et al., 2004; Slayden and Brenner, 2004; Heikinheimo et al., 2007; de Giorgio-Miller et al., 2008). The consequences of PR blockade on AR expression during follicular phase dosing were different for RU-486- and PF-02413873-treated animals because only RU-486 had a statistically significant effect on AR expression compared with control animals. This observation when coupled with the data captured in Figs. 4 and 5 suggest that the mode of action of PF-02413873 and RU-486 in the macaque seem to be different.
We explored the pharmacological effects of PF-02413873 by undertaking a clinical evaluation in healthy female subjects. PF-02413873 had been previously assessed in a preliminary single-dose escalation to determine PK, safety, and toleration (Bungay et al., 2011). PF-02413873 was considered safe and well tolerated up to the maximal dose tested (3 g) and behaved with a terminal half-life of approximately 40 h (Bungay et al., 2011). These data were used to underwrite a 14-day, multiple-dose, double-blind, third-party open, randomized, and placebo-controlled study to determine the effects of PF-02413873 on endometrial growth and ovarian function. During this multiple dose study, idiosyncratic maculopapular rash was observed in both the 250-mg (n = 1) and 500-mg (n = 4) cohorts, which resulted in early curtailment of dosing in the 500-mg dose cohort. The rash resolved on cessation of dosing. The mechanism driving the rash is still under investigation and, although the possibility that this is a PR-mediated effect can not be excluded, because there are reports of rash with progestins, its incidence as well as that reported for steroidal PR antagonists is low. Fortunately, there were sufficient data collected from all dose cohorts to determine the effect of PF-02413873 on endometrial thickness and plasma LH/estradiol levels (Figs. 8 and 9). In a manner consistent with the data generated in the macaque, PF-02413873 induced a dose-dependent and statistically significant reduction in endometrial thickness at the highest PF-02413873 dose tested. The PF-02413873 plasma exposures at this dose were similar to those that reduced endometrial thickness in the macaque (compare Tables 2 and 4). In addition, on the basis of the plasma LH and estradiol profiles as well as the apparent breakthrough in the normal endocrine profile during the postdosing observation phase (Fig. 9), the menstrual cycle seemed to be blocked for the duration of exposure with the 500-mg dose. Taken together, these data strongly underwrite the translational confidence in the screen sequence leading to the identification of PF-02413873.
In conclusion, we have pharmacologically characterized a novel, selective nonsteroidal PR antagonist, PF-02413873. We have demonstrated that PF-02413873 suppresses endometrial growth in the macaque and human and established some preliminary evidence to suggest that the mechanism by which it achieves this seems to be different compared with conventional steroidal PR antagonists. Whether by retaining unligated PR in the cytosol or some other mechanism, the inhibition of PR function by PF-02413873 deserves further experimental investigation because this mode of action may give rise to an entirely different profile of effect on the endometrium than that characterized by the class of steroidal PR antagonists (Mutter et al., 2008; Ioffe et al., 2009).
Participated in research design: Howe, Mount, Brown, Bungay, Gibson, Hawcock, Richard, Tweedy, and Pullen.
Conducted experiments: Howe, Bess, Brown, Bungay, Hawcock, Ramsey, and Pullen.
Contributed new reagents or analytic tools: Gibson.
Performed data analysis: Howe, Bess, Brown, Bungay, Gibson, Hawcock, Jones, Walley, McLeod, Apfeldorfer, Ramsey, and Pullen.
Wrote or contributed to the writing of the manuscript: Howe, Mount, Bungay, Gibson, Hawcock, Richard, Walley, and Pullen.
We thank Baerbel Wittke for clinical pharmacology analysis support at Pfizer Global Research and Development; Dr. A. Fuchs at Covance for expert guidance and managing cynomolgus macaque studies; Dr. Herco van Liere (Pamgene) for characterizing the nature of the ligand-induced PR/LxxLL peptide interaction; and the volunteers who participated in the clinical study.
This work was supported by Pfizer Global Research and Development.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- progesterone receptor
- nuclear hormone receptor
- glucocorticoid receptor
- androgen receptor
- mineralocorticoid receptor
- lutenizing hormone
- confidence interval
- nuclear receptor corepressor 1
- nuclear receptor coactivator 1.
- Received May 11, 2011.
- Accepted August 4, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics