Article pubs.acs.org/jmc
Novel Pyridyl- or Isoquinolinyl-Substituted Indolines and Indoles as Potent and Selective Aldosterone Synthase Inhibitors Lina Yin,†,‡, ⊥ Qingzhong Hu,†, ⊥ Juliette Emmerich,† Michael Man-Chu Lo,∥ Edward Metzger,∥ Amjad Ali,∥ and Rolf W. Hartmann*,†,§ †
Pharmaceutical and Medicinal Chemistry, Saarland University, Campus C2.3, D-66123 Saarbrücken, Germany ElexoPharm GmbH, Campus A1, D-66123 Saarbrücken, Germany § Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Campus C2.3, D-66123 Saarbrücken, Germany ∥ Discovery Chemistry, Merck Research Laboratories, 126 East Lincoln Avenue, Rahway, New Jersey 07065, United States ‡
S Supporting Information *
ABSTRACT: Pathologically, high levels of aldosterone are associated with severe cardiovascular diseases such as congestive heart failure, hypertension, and myocardial fibrosis. The inhibition of aldosterone synthase (CYP11B2) to reduce aldosterone levels has been proposed as a promising treatment for diseases related to CYP11B2 because it is the crucial enzyme in the biosynthesis of aldosterone. A series of novel pyridyl- or isoquinolinyl-substituted indolines and indoles was designed via a ligand-based approach. The synthesized compounds were tested and found to be strong CYP11B2 inhibitors. The most potent ones showed IC50 values of less than 3 nM, being similarly potent as fadrozole and LCI699. Among them, compounds 14 and 23 showed good selectivity over the highly homologous CYP11B1, with selectivity factors (SF = IC50 CYP11B1/IC50 CYP11B2) around 170; thus, they are superior to fadrozole and LCI699 (SFs < 15). These potent CYP11B2 inhibitors exhibited no inhibition (IC50 > 50 μM) of a panel of hepatic CYP enzymes including CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A4 and the crucial steroidogenic enzymes, CYP17 and CYP19. Because of these advantageous profiles, compounds 14 and 23 are considered to be candidates for further in vivo evaluation.
■
escape” phenomenon observed for ACE inhibitors4 and the overcoming of the MR blockage as a result of aldosterone accumulation caused by long-term application of MR antagonists.5 Recently, it has been demonstrated that CYP11B2 inhibitors are effective in decreasing plasma aldosterone levels in rats and humans.6 Experiments in heart failure models in rats indicated that the administration of the potent CYP11B2 inhibitor FAD286 (R-enantiomer of fadrozole, a nonselective aromatase inhibitor, Chart 1) improved cardiac hemodynamics and cardiac function, and the effects were more significant than those achieved by spironolactone.7 Application of the CYP11B2 inhibitor LCI699 (Chart 1, IC 50 = 0.2 nM) in patients with primary aldosteronism or hypertension reduced blood pressure and increased potassium and renin plasma concentrations.8 Over the past decade, several classes of CYP11B2 inhibitors have been designed by our group9 and others.10 The inhibitory mechanism exploited for the design of these compounds makes use of an sp2 hybrid N, which reversibly coordinates to the
INTRODUCTION Aldosterone is a major mineralocorticoid that regulates electrolyte and volume homeostasis after binding to the mineralocorticoid receptor (MR). Normally, its secretion is controlled by the cooperation of the renin−angiotensin− aldosterone system, adrenocorticotropic hormone, and potassium concentration. However, in some pathological conditions, aldosterone levels can be abnormally elevated. High aldosterone levels are associated with severe cardiovascular diseases such as congestive heart failure, myocardial fibrosis, and certain forms of hypertension,1 which can lead to sudden death. Aldosterone is synthesized from 11-deoxycorticosterone via three consecutive steps catalyzed by aldosterone synthase (CYP11B2), which is a mitochondial cytochrome P450 (CYP) enzyme located in the outer layer of the adrenal cortex.2 Reduction of high aldosterone levels through the inhibition of CYP11B2 has been proposed, therefore, as a novel strategy for the treatment of aldosterone-related heart diseases.3 Such a therapy is believed to be superior to treatments with the currently used angiotensin-converting enzyme (ACE) inhibitors (like enalapril, Chart 1) and MR antagonists (like spironolactone, Chart 1). This is due to the “aldosterone © XXXX American Chemical Society
Received: January 27, 2014
A
dx.doi.org/10.1021/jm500140c | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Chart 1. Structures of ACE Inhibitor Enalapril, MR Antagonist Spironolatone, Aromatase Inhibitor Fadrozole (R-enantiomer: FAD 286), and CYP11B2 Inhibitor LCI699
Chart 2. Inhibitor Design Conception
heme iron. Such a strategy was successful in the past for aromatase (CYP19) inhibitors,11 17α-hydroxylase-17,20-lyase (CYP17) inhibitors,12 and 11β-hydroxylase (CYP11B1) inhibitors.13 In the case of CYP11B2 inhibitors, extremely potent compounds with IC50 values in a subnanomolar range were obtained.14 To avoid severe side effects, the compounds should not inhibit other steroidogenic CYP enzymes because they are crucial for the biosynthesis of other steroid hormones, for example, CYP17 for androgens, CYP19 for estrogens, and CYP11B1 for glucocorticoids. However, because of the identical catalytic mechanism and the homology throughout the steroidogenic CYP enzyme family, it is a challenge to develop selective steroidogenic CYP enzyme inhibitors, especially between CYP11B1 and CYP11B2, which share more than 93% homology. Nevertheless, highly selective inhibition can be achieved, as demonstrated by some CYP1715 and CYP11B216 inhibitors. These results showed that the key to high selectivity often resides in subtle structural differences caused by small substituents,15 as has been observed in dual CYP enzyme inhibitors.17 In this study, a series of novel CYP11B2 inhibitors was designed via a ligand-based approach. Reference compound 114a is a potent CYP11B2 inhibitor (IC50 = 28 nM) with a dihydroquinolinone core (Chart 2). The fusion of an additional five-membered ring between the lactam N atom and the adjacent benzene moiety strongly elevated inhibitory potency (reference compound 2, IC50 = 1.1 nM).16a To look for another class of potent inhibitors and further explore the enzyme binding pocket, the five-membered ring was sustained, whereas the lactam ring was cleaved, leading to an indoline core (Chart 2). Because the carbonyl oxygen of the lactam was identified as a hydrogen-bond acceptor interacting with the enzyme,14a the amido moiety was preserved. Various R1 substituents with different bulkiness and electrostatic potential were subsequently introduced into the amido moiety to probe the binding pocket, leading to compounds 3−12. Because the heme-complexing heterocycle was a determinant factor of inhibitory potency, modifications on this ring were also performed, resulting in compounds 13−24. The replacement of the indoline core by an indole moiety led to compounds 25−29. These compounds were subsequently tested for CYP11B2 inhibition and also for
selectivity over CYP11B1, CYP17, and CYP19 for safety evaluation.
■
RESULTS AND DISCUSSION Chemistry. The synthetic routes are depicted in Schemes 1−3. Two consecutive key steps were used for the synthesis of compounds 3−16 and 24−29, which include a Suzuki coupling reaction with pyridin-3-yl or isoquinolin-4-yl boronic acid to introduce the N-containing heterocycle and an acylation in order to furnish various substituents on the indoline or indole core (Schemes 1 and 3). For indoline derivatives 3−16 and 24, the amino group was first protected with Boc to ensure high yields in the Suzuki coupling. Before acylation, the protecting group was cleaved using trifluoroacetic (Scheme 1). With regard to the compounds with different substituents on the pyridyl moiety (17−23), boride 17a was employed as a common building block to achieve last-stage diversification via Suzuki coupling with various substituted bromo-3-pyridines (Scheme 2). Intermediate 17a was prepared via acetylation of 5-bromoindoline and subsequent borylation using bis(pinacolato)diboron and Pd(dppf)2Cl2 as a catalyst. A simple stoichiometric reduction of ketone 19 with sodium borohydride was performed to obtain alcohol 21 as a racemate. Interestingly, for compounds 8 and 9, in which the indoline core was furnished with a substituted phenylcarbonyl moiety, the signal of one aromatic proton was missing from the 1H NMR when it was measured at room temperature. However, after heating to 373 K, the signal appeared at around 7.7 ppm (see the Supporting Information for details). This observation might be caused by the bulky phenylcarbonyl substituents, which block the rotation of amid bond at room temperature and thus conceal one signal of the aromatic protons. The structures of compounds 8 and 9 were further confirmed by 13C NMR, MS, and X-ray diffraction (see the Supporting Information). Biological Results. Inhibition of CYP11B2 and CYP11B1. The synthesized compounds were tested for their inhibitory potencies using V79 MZh cells expressing either human B
dx.doi.org/10.1021/jm500140c | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Scheme 1a
a Reagents and conditions: (i) Boc2O, NaHCO3, THF, H2O, 0 °C to rt; (ii) method A: pyridin-3-ylboronic acid or isoquinolin-4-ylboronic acid, Ph(PPh3)4, Na2CO3, DME, H2O, 90 °C; (iii) method B: CF3COOH, CH2Cl2, 0 °C to rt; (iv) method C: RCOCl, pyridine, THF, 0 °C to rt.
Scheme 2a
Reagents and conditions: (i) method C: CH3COCl, pyridine, THF, 0 °C to rt; (ii) bis(pinacolato)diboron, Pd(dppf)2Cl2, KOAc, dioxane, 105 °C; (iii) method A: R-substituted 3-bromopyridine, Ph(PPh3)4, Na2CO3, DME, H2O, 90 °C; (iv) NaBH4, MeOH, 0 °C.
a
Scheme 3a
a Reagents and conditions: (i) method A: pyridin-3-yl boronic acid or isoquinolin-4-yl boronic acid, Ph(PPh3)4, Na2CO3, DME, H2O, 90 °C; (ii) method D: Ac2O, Et3N, DMAP, CH2Cl2, 60 °C.
CYP11B1 or CYP11B2.18 The results are presented with fadrozole and LCI699 as references in Tables 1 and 2. Various acyl moieties were introduced into the indoline core (Table 1) to probe the enzyme binding pocket, and steric hindrance was consequently identified as a crucial determinant for CYP11B2 inhibition. Bearing a methyl group, compound 3 exhibited a strong inhibition of CYP11B2 with an IC50 value of 60 nM and a selectivity factor (SF = IC50 CYP11B1/IC50 CYP11B2) of 47 over CYP11B1. In contrast, ethyl analogue 4 showed a significant decrease in both potency and selectivity (IC50 = 228 nM and SF = 21). With a further increase in the bulkiness of the substituent, the inhibitory potency of the corresponding compounds dropped dramatically. Cyclopropyl substitution (3) led to a 20-fold decrease in potency (IC50 = 1285 nM) compared to that of methyl analogue 3. Similarly, compounds
with isopropyl (6) and 2-chloroethyl (7) showed even weaker inhibition, with IC50 values over 5000 nM. A steric clash for bulky substituents at the indoline core was further demonstrated by the finding that ethyl analogue 4 was around 200-fold less potent than the corresponding ring-closed parent compound 2 (IC50 = 1.1 nM, Table 1). Moreover, introduction of aromatic rings (4-F Ph, 4-OMe Ph, and 3-OMe Ph in compounds 8−10, respectively, as well as 2-thienyl in compound 11) led to only moderate inhibitors (IC50 values ranging from 600 to 3000 nM) with moderate to low selectivity (SF < 50). A benzyl substitution (12) even resulted in a preference for CYP11B1 inhibition. To further validate the finding that the size of the substituents at the amido moiety strongly influences CYP11B2 inhibition and also to enhance inhibitory potency, C
dx.doi.org/10.1021/jm500140c | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
wide range of substituents were introduced at the 4- or 5position, including electron-withdrawing (F, CN, and Ac), electron-donating (Me, MeO, 1-OH-Et, and Ph), and bulky (Ph) groups as well as hydrogen-bond donor (1-OH-Et) and acceptors (F, MeO, and 1-OH-Et) (Table 2). Positions adjacent to the N, however, were left unsubstituted to avoid steric hindrance for the coordination. These modifications sustained or improved both potency and selectivity in comparison to that of nonsubstituted analogue 3 (IC50 = 60 nM), except for compound 17 (5-CN), which is 4-fold less active, with an IC50 value of 426 nM. This is probably because the CN group reduces the electron density at the sp2 hybrid N and thus weakens the coordination to the heme iron. Intriguingly, although having strong electron-withdrawing groups similar to that of CN, 5-F (18) led to a similar inhibition (IC50 = 61 nM) as that ofcompound 3, whereas 5-Ac (19) resulted in an IC50 value of 36 nM, which is 2-fold stronger. This interesting observation might be a compromise between the weakened coordination and the hydrogen bonds formed by F or Ac with Glu310. In contrast, all electrondonating groups that were introduced promoted CYP11B2 inhibition. The OMe group (20) significantly increased inhibitory potency to 16 nM, whereas the 1-OH-Et moiety (21) rendered only a slight improvement (IC50 = 55 nM). It is interesting that the bulky 5-phenyl group (22) elevated CYP11B2 inhibition to 3 nM, which indicates the occupation of an additional hydrophobic pocket near the heme that has been exploited by fadrazole (PDB ID: 4FDH).19 The same explanation could also be valid for compound 13 with an isoquinoline moiety. Similarly, methyl substitution at the 4position of the pyridyl in compound 3 boosted the CYP11B2 inhibition from 60 to 2.2 nM (compound 23). This could be explained by a conformation induced by this Me group being located in ortho to the aryl−aryl bond, which is favorable for the interaction with the active site of the enzyme. Such an ortho effect might also contribute to the high potency of compound 13, in which the fused benzene nucleus twists the isoquinolinyl moiety out of the indoline plane. More importantly, compound 23 exhibited excellent selectivity over CYP11B1 (SF = 166), which is much better than that of the references, fadrozole and LCI699 (SFs of 7.9 and 15, respectively). Because an Et group at the amido moiety is tolerated and, in the case of compound 14, high selectivity is obtained (SF = 184), an Et analogue of compound 23 was synthesized by combining the Et amido moiety and the 4-Me substitution of the pyridyl into one molecule, leading to compound 24. Although 24 is also very potent (IC50 = 14 nM), the selectivity is much worse compared to that of Me compound 23 (SF of 21 vs 166). Furthermore, replacement of the indoline core by an indole moiety elevated the CYP11B2 inhibition of the compound with a nonsubstituted pyridyl by nearly 3-fold to 23 nM (25, Table 2) compared to that of its parent compound, 3. In contrast, for analogues comprising isoquinolinyl and 4-methyl-pyridin-3-yl, this exchange only led to similar potency as compared to that of the indoline compounds (for comparison, see compound pairs 13 and 27, 23 and 28, and 24 and 29). Similar to that observed in the indoline series, the employment of Et instead of a Me group at the amido moiety decreased CYP11B2 inhibition by 3to 8-fold (for comparison, see compound pairs 25 and 26 as well as 28 and 29). Although these indole analogues were demonstrated to be very potent CYP11B2 inhibitors with IC50 values less than 25 nM, with the exception of compound 26
Table 1. Inhibition of CYP11B1 and CYP11B2 by Compounds 3−16
percent inhibition at 500 nMc
IC50 (nM)d
compd
R
11B2a
11B1b
11B2a
11B1b
SFe
3 4 5 6 7 8 9 10 11 12 13 14 15 16 2 Fadrozole LCI699
Me Et c-propyl i-propyl ClCH2CH2 4-F Ph 4-MeO Ph 3-MeO Ph 2-thienyl Bn Me Et c-propyl i-propyl
87 66 28 7.3 7.0 15 26 40 14 2.2 100 100 98 90
14 0.8 26 11 6.0 2.1 1.6 28f 4.9 26 92 39 29 11
60 228 1285 >5000 >5000 2833 1433 607 3070 >5000 0.7 2.8 30 152 1.1 0.8 0.2
2848 4737 1423 4100 >5000 >5000 >5000 2300 >5000 1523 52 516 3723 6093 715 6.3 2.9
47 21 1
