Novel Nonsteroidal Progesterone Receptor (PR) Antagonists with a Phenanthridinone Skeleton

Abstract

The progesterone receptor (PR) plays an important role in various physiological systems, including female reproduction and the central nervous system, and PR antagonists are thought to be effective not only as contraceptive agents and abortifacients but also in the treatment of various diseases, including hormone-dependent cancers and endometriosis. Here, we identified phenanthridin-6(5H)-one derivatives as a new class of PR antagonists and investigated their structure–activity relationships. Among the synthesized compounds, 37, 40, and 46 exhibited very potent PR antagonistic activity with high selectivity for PR over other nuclear receptors. These compounds are structurally distinct from other nonsteroidal PR antagonists, including cyanoaryl derivatives, and should be useful for further studies of the clinical utility of PR antagonists.

Progesterone receptor (PR) is a member of the nuclear receptor superfamily of ligand-dependent transcription factors¹ and is expressed in female tissues including uterus, ovary, vagina, fallopian tubes and breast, and brain. It plays an important role in female reproduction, being associated with the establishment and maintenance of pregnancy and with alveolar development in the breast. Recently, there is increasing evidence suggesting that PR has a role in the central nervous system.^2,3 PR agonists, including the endogenous agonist progesterone (1, Figure 1), have been used extensively to treat gynecological disorders, as well as in female contraception and hormone replacement therapy. However, their steroidal skeleton brings with it the potential to cross-react with other steroid receptors, which can result in unwanted side effects, potentially limiting the benefits of these agents. To overcome these issues, nonsteroidal PR agonists such as tanaproget (2) have been developed⁴

Figure 1.

Chemical structures of PR agonists 1–2 and antagonists 3–10.

PR antagonists may be potentially useful for the treatment of various diseases including hormone-dependent cancers,⁵ uterine fibroids,⁶ endometriosis,⁷ and abortifacient. However, the steroidal PR antagonist mifepristone (3) is in only limited clinical use as an abortifacient at present. It is known that 3 demonstrates potent activities against other steroid receptors such as glucocorticoid receptor (GR). Therefore, development of selective nonsteroidal PR antagonists is necessary for the estimation of their clinical utility. Indeed, various nonsteroidal PR antagonists, including 4,⁸5,⁹ and 6,⁶ have been synthesized based on the structure of 2, with the aim of improving the selectivity for PR over other steroid receptors. Some of the present authors also developed 7–9.¹⁰⁻¹³ PR antagonists containing a cyanoaryl moiety, a common pharmacophore motif, found that quite small structural modifications cause agonist/antagonist activity switching.^10,8,9 A concern about such activity switching is that metabolites of an antagonist act as agonists. Although we recently reported nonsteroidal PR antagonist 10, which does not have a cyanoaryl moiety,¹⁴ there is still a need for a new, more robust pharmacophore motif of PR antagonists to avoid such activity switching.

Here, to obtain nonsteroidal PR antagonists without a cyanoaryl moiety, we focused on our previously developed liver X receptor (LXR) and retinoic acid receptor-related orphan receptor (ROR) ligands 12 and 13, which contain a phenanthridin-6(5H)-one skeleton as a cyclized carba-analog of T0901317 (11) (Figure S1).^15,16 Because hydroxycholesterols are endogenous ligands for both LXR and ROR, we hypothesized that if the phenanthridin-6(5H)-one scaffold acts as a steroid surrogate, phenanthridinone derivatives would also bind to other nuclear receptors that recognize endogenous steroidal ligands. Based on this idea, we successfully identified phenanthridinone derivatives with PR-antagonistic activity and examined their structure–activity relationships (SAR) and selectivity.

Compounds 12–24 were prepared as described in refs (15 and 16). PR-antagonistic activity was evaluated by assay of PR-regulated alkaline phosphatase activity in human breast cancer cell line T47D.¹⁷ Under our assay conditions, mifepristone (3) showed subnanomolar IC₅₀ value with a reproducibility (Table 1). We found that compound 14, a phenanthridin-6(5H)-one bearing an n-butyl group on the nitrogen atom, showed PR-antagonistic activity with the IC₅₀ value of 7.1 μM. Although the activity was not potent, we considered it promising, as this compound belongs to a different chemical class from other known PR antagonists. Thus, we focused on 14 as a lead compound and investigated the SAR of analogs we had previously synthesized.¹⁶ First, we examined SARs at the 2-position (Table 1). Introduction of alkyl groups (15–17) or a hydroxymethyl group (18) retained the PR-antagonistic activity. On the other hand, compound 19 bearing a hexafluoropropanol moiety showed the strongest activity, suggesting that trifluoromethyl groups at this position are favorable for potent activity.

Table 1. SAR at the 2-Position and the Nitrogen Atom^a.

compounds	R1	R2	PR IC50 (nM)
Mifepristone (3)			0.073 ± 0.0047 (N = 19)
7			123
9			95.3
10			107
14	H	n-Bu	7100
15	Me	n-Bu	7900
16	Et	n-Bu	7800
17	t-Bu	n-Bu	8300
18	CH2OH	n-Bu	8600
19	(CF3)2COH	n-Bu	1200
12	(CF3)2COH	H	780
20	(CF3)2COH	Me	340
21	(CF3)2COH	Et	1400
22	(CF3)2COH	n-Pr	3800
23	(CF3)2COH	n-Hex	2500
24	(CF3)2COH	n-C9H19	10000

^aAlkaline phosphatase assay. Mean IC₅₀ values with standard error of mean (SEM) from 1 or N times independent experiments.

Next, we evaluated substituent effects on the nitrogen atom. Hydrogen analog 12 and methyl analog 20 increased PR-antagonistic activity, affording submicromolar IC₅₀ (Table 1). Introduction of a longer alkyl group (22–24) resulted in a decrease of antagonistic activity. Thus, we concluded that hydrogen or a methyl group is a suitable substituent on the nitrogen atom.

Next, the effect of introduction of a methoxy group at every position (methoxy scanning) was investigated (25–31; Table 2) because many types of nuclear receptor ligands possess a methoxy group(s).¹⁸ 3-Methoxy and 4-methoxy analogs 26 and 27 showed more than 3-fold and 2-fold enhanced activity compared with the unsubstituted analog 19, respectively. In contrast, 9-methoxy analog 30 showed 5-fold weaker activity. We believe that introduction of a substituent into every position is favorable in order to investigate substituent effects. It is important to note that phenanthridin-6(5H)-one analogs bearing substituent(s) at any position can be quite easily synthesized, and this represents a considerable advantage compared to the steroid skeleton. Based on the SARs shown in Tables 1–3, we proceeded to synthesize new compounds, aiming to obtain more potent antagonists.

Table 2. SAR at the 4-Position and Other Substitution Effects^a.

compounds	R1	R2	PR IC50 (nM)
25	n-Bu	1-MeO	4000
26	n-Bu	3-MeO	410
27	n-Bu	4-MeO	560
28	n-Bu	7-MeO	12000
29	n-Bu	8-MeO	1100
30	n-Bu	9-MeO	5800
31	n-Bu	10-MeO	2700
32	H	4-F	130
33	H	7-F	730
34	H	8-F	310
35	H	9-F	1000
36	H	4-MeO	270 ± 20 (N = 2)
37	H	4-Me	150 ± 17 (N = 5)
38	H	4-Et	210 ± 3.7 (N = 2)
39	H	4-n-Pr	360 ± 15 (N = 2)
40	H	4-Me, 8-F	46 ± 4.4 (N = 4)
45	Me	4-Me	180 ± 8.9 (N = 3)
46	Me	4-Me, 8-F	27 ± 16 (N = 2)

^aAlkaline phosphatase assay. Mean IC₅₀ values with standard error of mean (SEM) from 1 or N times independent experiments.

Table 3. Binding Affinity of Representative PR Antagonists^a.

compounds	PR IC50 (nM)
Mifepristone	5.2 ± 0.53
37	1500 ± 540
40	340 ± 150
46	150 ± 110

^a4 nM [1,2,6,7-³H]progesterone was used.

At this stage, the substituent on the nitrogen atom was fixed to hydrogen, and we focused on fluoro derivatives to investigate the effect of the electron withdrawing substituent. Synthesis of the novel analogs 32–40 is illustrated in Scheme 1. Briefly, ortho-substituted anilines 41 were treated with hexafluoroacetone trihydrate to give 1,1,1,3,3,3-hexafluoro-2-hydroxypropyl derivatives 42. Anilines 42 were condensed with o-bromo-/iodobenzoic acids to give amides 43. After protection of the amide and hydroxyl group with SEM, palladium-catalyzed intramolecular cyclization afforded 44. Compounds 32–40 were obtained by deprotection of the SEM groups. The 4-fluoro analog 32 and 8-fluoro analog 34 were 6-fold and 3-fold more potent than the unsubstituted analog 12, respectively (Table 2). These preferences of fluorine-substitution position are roughly consistent with the results of the methoxy scanning, that is, 4- and 8-substituted analogs showed enhanced activity whereas 9-substituted analogs showed weaker activity.

Scheme 1. Synthesis of the Novel Analogs 32–40.

Reagents and conditions: (a) CF₃COCF₃·1.5H₂O, p-TsOH·H₂O, toluene, reflux, 8–49%; (b) 2-iodobenzoic acids or 2-bromobenzoic acids, EDC, DMAP, DMF, 100 °C, 19–91%; (c) SEMCl, NaH, DMF, 0 °C to rt; (d) Pd(OAc)₂, PCy₃·HBF₄, Cs₂CO₃, DMA, 130 °C, 9–44% in 2 steps; (e) TBAF, THF, reflux, 30–80%.

We next focused on the 4-position because 4-fluoro analog 32 showed the most potent activity among the methoxy- and fluorine-substituted compounds, and 4-substituted analogs are easier to synthesize than 3-substituted analogs; in the course of synthesis of the 3-substitued analog, the 1-substitued analog is also obtained. 4-Methoxy analog 36 was more potent than the unsubstituted analog 12, in accordance with the results for the n-butyl analog 19 and 27 (Table 1). 4-Alkyl analogs 37–39 showed greater activity than the unsubstituted analog 12. In particular, 4-methyl analog 37 had around 0.1 μM IC₅₀ value (Table 2).

Next, we planned to introduce a fluorine atom at the 8-position of 37, because the 8-fluoro analog 34 showed enhanced antagonistic activity. Indeed, the 4-methyl and 8-fluoro-substituted analog 40 showed even more potent activity than 37. Finally, based on the SARs shown in Table 2, we investigated N-methyl analogs 45–46 (Scheme S1). The antagonistic activity of 4-methyl analog 45 was stronger than that of unsubstituted 20 as expected but weaker than that of N–H analog 37. On the other hand, the N,4-dimethyl and 8-fluoro analog 46 showed very potent antagonistic activity (Table 2), being more potent than the cyanoaryl,¹⁰⁻¹³ and noncyanoaryl¹⁴ chemical classes of nonsteroidal PR antagonists in alkaline phosphatase assay (Table S1). In addition, these representative compounds did not show PR agonistic activity at 10 μM. This result suggests that the phenanthridinone skeleton might be a more robust pharmacophore motif of PR antagonists to avoid such activity switching. Therefore, we selected 37, 40, and 46 as representative compounds and investigated their biological activities in more detail.

The alkaline phosphatase inhibitory activity assay utilized for SAR analyses may be affected not only by inhibition of PR-mediated transcription (antagonistic activity) but also by direct inhibition of the enzymatic activity. Therefore, to validate the PR-antagonistic activity of representative compounds, we measured the expression of another PR-regulated protein, the β1 subunit of Na/K-ATPase,¹⁹ in T47D cells by means of Western blotting. Potent PR antagonists 37, 40, and 46 decreased progesterone-induced β1-Na/K ATPase expression, whereas the weak antagonist 24 had little effect on the β1-Na/K ATPase level at the evaluated concentration (Figure 2). The activity of 1 μM 40 and 46 was equal to or greater than 0.3 μM mifepristone, indicating that these nonsteroidal antagonists are very potent, at least under these assay conditions.

Figure 2.

Western blot analysis of β1-Na/K-ATPase in T47D cells in the presence of test compounds.

We next investigated whether the representative compounds bind directly to PR, using human PR ligand-binding domain (LBD) and ³H-labeled progesterone (1) (Table 3). Compounds 37, 40, and 46 all exhibited strong binding affinity. These results, together with the results of alkaline phosphatase assay, Western blotting of β1-Na/K ATPase, and PR binding assay, indicate that compounds 37, 40, and 46 bind directly to the progesterone-binding pocket of PR and inhibit expression of PR-regulated genes; thus, they function as PR antagonists.

Some of the analogs shown in Tables 1 and 2 were identified as RORs inverse agonists or LXRs antagonists. And PR, GR, and androgen receptor (AR) all belong to the class I steroid receptor family. Thus, selectivity of PR antagonists toward AR and GR is an important concern, and selectivity of PR antagonists over AR and GR has been investigated⁸ (3: PR IC₅₀: 0.3 nM, AR IC₅₀: 5 nM, GR IC₅₀: 0.8 nM; 4: PR IC₅₀: 16 nM, AR IC₅₀: 1300 nM, GR IC₅₀: 1800 nM). Here, the activities of the representative PR antagonists toward other nuclear receptors, including RORα/β/γ inverse agonistic activities, and LXRα/β, GR and AR antagonistic activities, were evaluated by means of reporter gene assays.²⁰ Among these compounds, 37 showed more than 100-fold selectivity for PR over all the NRs evaluated (Table 4). Compounds 40 and 46 also showed more than 300-fold selectivity over these NRs except for AR. For AR, 40 and 46 showed over 30- and 10-fold selectivity, respectively. Compounds 37 demonstrated medium Caco-2 permeability, sufficient human liver microsomal stability, and acceptable LogP value (Table S2).

Table 4. Selectivity of Representative Compounds for PR over other NRs.

compounds	PRa,b	RORαa,c	RORβa,c	RORγa,c	LXRαa,c	LXRβa,c	ARa,c	GRαa,c
37	0.15	NAd	>20	>20	>20	>20	>20	>20
40	0.046	NAd	>20	NAd	>20	NAd	1.8	>20
46	0.027	NAd	>20	NAd	>20	>20	0.3	>20

^aIC₅₀ (μM).

^bAlkaline phosphatase assay.

^cReporter gene assay.

^dNo activity at 20 μM. T0901317 (0.3 μM), T0901317 (0.1 μM), dihydrotestosterone (0.3 nM), and dexamethasone (1 nM) were used as agonists for LXRα, LXRβ, AR, and GR, respectively.

In summary, we have identified a series of PR antagonists with a phenanthridin-6(5H)-one skeleton and investigated their SARs by means of alkaline phosphatase assay using T47D cells. Among them, 37 exhibited potent PR-antagonistic activity with the IC₅₀ value of around 0.1 μM and showed more than 100-fold selectivity over other NRs examined. Compounds 46 and 40 showed very potent PR-antagonistic activity, with IC₅₀ values of 0.01 μM order. These compounds showed over 100-fold selectivity for PR over RORs, LXRs, and GR and about more than 10-fold selectivity over AR. The PR-antagonistic activity of these representative compounds was validated by means of Western blot analysis of a PR-regulated protein (β1-Na/K-ATPase) and PR-binding assay. Compounds 37, 40, and 46 are PR antagonists belonging to a novel chemical class distinct from other PR antagonists, and their activity is similar to or stronger than those of reported PR antagonists, including cyanoaryl compounds. The phenanthridinone skeleton might be a more robust pharmacophore motif of PR antagonists to avoid activity switching. Thus, these compounds should open up new possibilities for exploitation of the pharmaceutical potential of PR antagonists.

Glossary

Abbreviations

PR

progesterone receptor

GR

glucocorticoid receptor

AR

androgen receptor

LXR

liver X receptor

ROR

retinoic acid receptor-related orphan receptor

Ts

p-toluenesulfonyl

EDC

N-ethyl-N′-(3-(dimethylamino)propyl)carbodiimide

DMAP

4-(N,N-dimethylamino)pyridine

DMF

dimethylformamide

SEM

2-(trimethylsilyl)ethoxymethyl

Cy

cyclohexyl

DMA

dimethylacetamide

TBAF

tetrabutylammonium fluoride

THF

tetrahydrofuran.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.8b00058.

• Structures of phenanthridinones and their lead compound, synthesis of 45 and 46, tables of PR antagonistic activity and physicochemical and pharmaceutical potential of PR antagonists, experimental section, and purity of the synthesized compounds (PDF)

The authors declare no competing financial interest.