Synthesis and Evaluation of Heteroaryl-Substituted

Mar 8, 2006 - These findings show the importance of the blockade of aldosterone production as possible treatment for congestive heart failure and ...
1 downloads 0 Views 244KB Size
2222

J. Med. Chem. 2006, 49, 2222-2231

Synthesis and Evaluation of Heteroaryl-Substituted Dihydronaphthalenes and Indenes: Potent and Selective Inhibitors of Aldosterone Synthase (CYP11B2) for the Treatment of Congestive Heart Failure and Myocardial Fibrosis Marieke Voets,† Iris Antes,‡ Christiane Scherer,† Ursula Mu¨ller-Vieira,† Klaus Biemel,§ Sandrine Marchais-Oberwinkler,† and Rolf W. Hartmann*,† 8.2 Pharmaceutical and Medicinal Chemistry, Saarland UniVersity, P.O. Box 15 11 50, D-66041 Saarbruecken, Germany, Max Planck Institute for Informatics, Stuhlsatzenhausweg 85, D-66123 Saarbruecken, Germany, and Pharmacelsus CRO, Science Park 2, D-66123 Saarbruecken, Germany ReceiVed January 17, 2006

In this study, the synthesis and biological evaluation of heteroaryl-substituted dihydronaphthalenes and indenes (1-16) is described. The compounds were tested for activity by use of human CYP11B2 expressed in fission yeast and V79 MZh cells and for selectivity by use of human CYP11B1, CYP17, and CYP19. The most active inhibitor was the 6-methoxydihydronaphthalene 4 (IC50 ) 2 nM), showing a Ki value of 1.3 nM and a competitive type of inhibition. The 5-methoxyindene 3 was found to be the most selective CYP11B2 inhibitor (IC50 ) 4 nM; CYP11B1 IC50 ) 5684 nM), which also showed only marginal inhibition of human CYP3A4 and CYP2D6. Docking and molecular dynamics studies using our homology-modeled CYP11B2 structure were performed to understand some structure-activity relationships. Caco-2 cell experiments revealed highly cell-permeable compounds, and metabolic studies with 4 using rat liver microsomes showed sufficient stability. Introduction Aldosterone synthase (CYP11B2) is the key enzyme of mineralocorticoid biosynthesis. This cytochrome P450 enzyme catalyzes the conversion of its substrate, 11-deoxycorticosterone, to the mineralocorticoid aldosterone.1 It is responsible for the regulation of the salt and water household and thus for the regulation of blood pressure. Besides the classical adrenal biosynthetic pathway, extra-adrenal sites of aldosterone production have been identified in brain, blood vessels, and heart.2-4 Mineralocorticoid receptors have been detected in heart tissue, and it is known that aldosterone produced locally in the myocardium triggers myocardial fibrosis.5,6 Thus, in addition to the indirect effects of aldosterone on the failing heart from sodium retention, hypervolemia, and hypertension, aldosterone exerts also direct effects on the heart. These findings show the importance of the blockade of aldosterone production as possible treatment for congestive heart failure and myocardial fibrosis. The mineralocorticoid receptor antagonists spironolactone and eplerenone (Rales and Ephesus study) were able to reduce mortality in heart failure patients and in patients after myocardial infarction.7,8 The treatment with spironolactone, however, was accompanied by severe side effects because of its poor selectivity for the aldosterone receptor.7,9 Additionally, a correlation between the use of spironolactone and hyperkalemia-associated mortality was observed.10 Eplerenone turned out to be more selective, but the hyperkalemia problem can be expected to occur as well. In 1994, aldosterone synthase was already propagated as a novel pharmacological target by our group.11 More recently, a new therapeutic option for the treatment of congestive heart failure and myocardial fibrosis was suggested by us: inhibition of aldosterone production using nonsteroidal, selective CYP11B2 * To whom correspondence should be addressed: phone + 49 681 302 3424; fax + 49 681 302 4386; e-mail [email protected]. Homepage: http://www.uni-saarland.de/fak8/hartmann/index.htm. † Saarland University. ‡ Max Planck Institute for Informatics. § Pharmacelsus CRO.

inhibitors.12,13 One of the main advantages of this approach is that nonsteroidal inhibitors are expected to have less side effects on the endocrine system than the steroidal antagonists. Furthermore, the inhibition of aldosterone formation should be more advantageous than interfering at the receptor level while leaving the elevated aldosterone levels unaffected. However, it is crucial that the inhibitors do not affect 11β-hydroxylase (CYP11B1), the key enzyme of the glucocorticoid biosynthesis, to avoid unwanted effects. It is quite a challenge to develop inhibitors with selectivity toward CYP11B2, since CYP11B1 and CYP11B2 have a sequence homology of more than 93%.14 Previously, we evaluated amidinohydrazones as potential CYP11B2 inhibitors using a CYP11B2-expressing yeast cell line and the adrenocortical tumor cell line NCI-H295R.15 However, the compounds displayed almost no inhibitory activity toward CYP11B2. Recently, (imidazolyl- and pyridylmethylene)tetrahydronaphthalenes and -indanes I (Chart 1) were described as highly active and in some cases selective CYP11B2 inhibitors by our group.16,17 By keeping the pharmacophore and rigidifying the molecules, a new class of 3-pyridine-substituted naphthalenes II was discovered, which turned out to be potent and selective inhibitors of CYP11B2.18 The most interesting of these was the 6-methoxy-substituted naphthalene (IC50 CYP11B2 ) 6 nM, IC50 CYP11B1 ) 1577 nM, selectivity factor ) 263).18 Another way to make the pharmacophore rigid (compounds of type III) is examined in this paper. The synthesis of these compounds (Chart 2) and their biological activity toward the human CYP enzymes CYP11B2, CYP11B1, CYP17, and CYP19 are described. Molecular modeling and docking studies using our homology-modeled CYP11B2 structure16-18 were performed to understand some interesting structure-activity relationships. Cell permeation properties of selected compounds using Caco-2 cells and metabolic stability of compound 4 using rat hepatic microsomes were also investigated.

10.1021/jm060055x CCC: $33.50 © 2006 American Chemical Society Published on Web 03/08/2006

Potent SelectiVe Inhibitors of Aldosterone Synthase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 7 2223

Chart 1. Title Compounds III Derived from the Lead Compounds I and II

Chart 2. Title Compounds

Scheme 2a

a Conditions: (a) 3-pyridylacetonitrile, NaNH , DMF, 0-80 °C, 6.5 h; 2 (b) NaOH, EtOH, reflux, 24 h; (c) PPA, 110 °C, 20 min; (d) NaBH4, MeOH, 0 °C to RT, 1 h; (e) CH3COOH/H2SO4, 100 °C, 1 h; (f) R′MgX, toluene, reflux, 3 h; (g) HCl, 100 °C, 2 h.

Scheme 1a

a Conditions: (a)TBA Br , CH Cl /MeOH, RT, 0.5-1 h; (b) NaBH , 3 2 2 4 MeOH, 0 °C to RT, 1 h; (c) pTSA, toluene, reflux, 2 h; (d) 3- or 4-pyridylboronic acid, Na2CO3, Pd(PPh3)4, DME, 80 °C, 16 h.

Chemistry The pyridine-substituted indenes and dihydronaphthalenes 1, 2, 4, 5, and 11 were synthesized by the route shown in Scheme 1. The R-bromoketones 2iii, 4iii, and 5iii, synthesized by reacting the corresponding ketone 2iv, 4iv, or 5iv with tetrabutylammonium tribromide at room temperature, were reduced with NaBH4. The resulting alcohols 2ii, 4ii, and 5ii and the commercially available 2-bromo-1-indanol 1ii were refluxed in toluene with a catalytic amount of pTSA to yield the dehydrated compounds 1i, 2i, 4i, and 5i. The last step was the Suzuki crosscoupling with a 3- or 4-pyridylboronic acid in the presence of sodium carbonate and tetrakis(triphenylphosphine)palladium as a catalyst.19 The intermediates 2iii, 4iii, and 5iii and 1ii, 2ii, 4ii, and 5ii have one or two chiral centers. The resulting isomers were not separated and the racemic mixtures were used for the next reaction steps.

An alternative synthetic route, based on the method by Bencze and Barsky,20 was used for the synthesis of the 3-pyridinesubstituted compounds 3 and 6-10 (Scheme 2). The alkylation of 3-pyridylacetonitrile with the brominated alkyl derivatives 3v, 6v, and 8v-10v yielded the nitriles 3iv, 6iv, and 8iv-10iv. Hydrolysis with NaOH resulted in the acids 3iii, 6iii, and 8iii10iii, which were cyclized to the corresponding indanones and tetralones 3ii, 6ii, and 8ii-10ii by treatment with PPA. The reduction of the ketones 3ii and 8ii-10ii with NaBH4 to the alcohols 3i and 8i-10i was followed by dehydration with sulfuric and acetic acids to give the desired indenes and dihydronaphthalenes 3 and 8-10. The Grignard reaction of the tetralone 6ii with an alkylmagnesium halogenide and subsequent dehydration of the resulting alcohols 6i and 7i with HCl yielded the 1-alkyl-substituted dihydronaphthalenes 6 and 7. The isomers of the intermediates 3iv, 6iv, and 8iv-10iv to 3i, 6i, and 8i10i were not separated but used as mixtures for the next reaction steps. The end products 8-10 were obtained as racemic mixtures. The synthesis of the 1-imidazole-substituted indenes and dihydronaphthalenes 12-14 is shown in Scheme 3.21 The R-bromoketones 2iii, 4iii, and 12iii were stirred with imidazole for 24 h to give the imidazole-substituted ketones 12ii-14ii. The reduction with NaBH4 to the corresponding alcohols 12i14i and the subsequent acid-catalyzed dehydration yielded the 1-imidazolyl compounds 12-14. The intermediates 12ii-14ii and 12i-14i were used as racemates or diastereomeric mictures, respectively. The (E)- and (Z)-3-styrylpyridines 15 and 16 were synthesized by a modified Wittig reaction (Scheme 4). The benzyl alcohol 15ii was transformed into the phosphonium salt 15i, which reacted with 3-pyridinecarbaldehyde in the presence of K2CO3 as a base and 18-crown-6 as phase-transfer catalyst.

2224

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 7

Voets et al.

Scheme 3a

a

Conditions: (a) imidazole, DMF, RT, 24 h; (b) NaBH4, MeOH, 0 °C to RT, 1 h; (c) CH3COOH, H2SO4, 100 °C, 1 h.

Scheme 4a

a Conditions: (a) PPh ‚HBr, benzene, reflux, 12 h; (b) 3-pyridinecar3 baldehyde, K2CO3, 18-crown-6, CH2Cl2, reflux, 12 h.

The resulting E- and Z-isomers could be easily separated by column chromatography. Biological Results Inhibition of Human Adrenal Corticoids Producing CYP11B1 and CYP11B2 In Vitro (Table 1). For the determination of the inhibitory activity of our compounds toward human CYP11B2, our screening assay was used.12 Human CYP11B2 expressing fission yeast was incubated with [14C]deoxycorticosterone as substrate and the inhibitor at a concentration of 500 nM. The product formation was monitored by HPTLC using a phosphoimager. Most of the 3-pyridine-substituted derivatives showed a higher inhibitory activity than the reference fadrozole (68%), an aromatase (CYP19) inhibitor that is in use for the treatment of breast cancer and is known to reduce aldosterone and cortisol levels in adrenal slices22 and in vivo.23 Only the 7-methoxy compound 5 exhibited a moderate activity (51%), and the racemic 4-ethyl-substituted 3,4-dihydronaphthalene 10 had little activity (20%). In general, the 3-pyridine-substituted indenes showed a slightly lower activity than the 3,4-dihydronaphthalenes. The shift of the methoxy substituent from 6- into 7-position of the 3,4-dihydronaphthalene core reduced activity, resulting in a moderate inhibitor 5. Introduction of a methyl substituent in 1-, 3-, or 4-position of the 3,4-dihydronaphthalene moiety did not change potency (6, 8, 9), but a larger ethyl substituent in the 4-position resulted in a loss of activity (10). The 1-imidazolylnaphthalenes 12-14 and the ring open compounds 15 and 16 showed moderate activity (25-54%). The 4-pyridine-substituted dihydronaphthalene 11 had almost no activity (15%). The most potent compounds, showing more than 60% inhibition in Schizosaccharomyces pombe, and a few less potent compounds were tested for activity and selectivity in V79 MZh cells expressing either CYP11B1 or CYP11B2.12,24 [14C]Deoxycorticosterone was used as substrate, and the products were monitored as in the yeast assay. In Table 1 the IC50 values are presented. All the compounds with more than 60% inhibition in the yeast assay exhibited very high activity toward CYP11B2 with IC50 values in the range of 2-30 nM. They were also highly selective by showing only moderate inhibition of CYP11B1 (IC50 ) 503-5684 nM). Compounds 1, 2, and 4 were over 180-fold more selective for CYP11B2 in contrast to the reference fadrozole, which displayed a selectivity factor of 10. The 5-methoxy-substituted indene 3 was the most selective compound with an exceptional selectivity factor of 1421.

Selected compounds with less than 60% inhibition in the yeast assay (10, 11, 13, 14, and 16) also showed only moderate or no activity at all in the CYP11B2 expressing V79 cells (IC50 values in the range between 176 and 2834 nM). An interesting finding is that the Z-isomer 16 exhibited some selectivity for CYP11B1 as compared to CYP11B2 with a 2.5-fold lower IC50 for the 11β-hydroxylase. By use of V79 cells, selected compounds (1-4) were tested for their type of CYP11B2 inhibition. They turned out to be competitive inhibitors (in Figure 1, compound 3 is shown as an example) and revealed Ki values of 8.4, 4.6, 2.6, and 1.3 nM, respectively (KM value of deoxycorticosterone ) 185 nM). Inhibition of Human CYP17, CYP19, and Hepatic CYPs in Vitro (Table 1). The selectivity of the compounds toward other steroidogenic CYP enzymes (CYP19, producing estrogens, and CYP17, forming androgens) was investigated. The IC50 values of the compounds for CYP19 were determined in vitro by use of human placental microsomes and [1-3H]androstenedione as substrate as described by Thompson and Siiteri25 with our modification.26 All the compounds showed low or no inhibitory activity with IC50 values in the range between 814 and >36 000 nM, thus being much less active than the reference fadrozole (IC50 ) 30 nM). The percent inhibition values of the compounds toward CYP17 were determined in vitro with progesterone as substrate and the 50000g sediment of Escherichia coli recombinantly expressing human CYP17.27,28 The indenes 1 and 3 and the 3and 4-methyl-3,4-dihydronaphthalenes 8 and 9 displayed a weaker inhibition at a concentration of 2.5 µM than the reference ketoconazole (40%), which is an antimycotic and an unspecific inhibitor of several CYPs. The 3,4-dihydronaphthalenes 2, 4, and 7 showed similar or slightly higher inhibition (37-62%). The most potent CYP17 inhibitor was the 1-methyl-substituted compound 6, exhibiting an IC50 value of 1085 nM. Thus, almost all of the tested compounds are very weak CYP17 inhibitors, having moderate to very low percent inhibition values that correspond to IC50 values of approximately 2500 nM or higher. The most selective compound 3 with respect to CYP11B1 inhibition was further tested toward two crucial human hepatic CYP enzymes: CYP3A4, being responsible for 75% of the drug metabolism, and CYP2D6, for which a genetic polymorphism is described. Only marginal inhibitory effects could be observed (CYP3A4, ketoconazole IC50 ) 50 nM, 3, 7.0% ( 1.0% inhibition at 50 nM; CYP2D6, quinidine IC50 ) 20 nM, 3, 8.5% ( 2.2% inhibition at 20 nM) Computational Results Docking and protein modeling studies were performed to verify our homology-modeled CYP11B2 structure16-18 and to explain interesting structure-activity relationships of two imidazolyl and one pyridyl compound (Table 2). Previously, compound 23 was found to be a highly active inhibitor of CYP11B2.18 In this study, we observed that the corresponding dihydronaphthalene compound 13 shows a much lower activity.

Potent SelectiVe Inhibitors of Aldosterone Synthase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 7 2225

Table 1. Inhibition of Human Adrenal CYP11B1 and CYP11B2, Human CYP17 and Human CYP19 In Vitro

% inhibitiona compd

n

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ketoconazole fadrozole

0 1 0 1 1

R

R′

H H 5-OMe 6-OMe 7-OMe Me Et H H H

0 1 1

R′′

H H 6-OMe E-isomer Z-isomer

H H 3-Me 4-Me 4-Et

IC50 valueb (nM)

IC50 valuec (nM)

% inhibitiond

humane CYP11B2

V79 11B1f CYP11B1

V79 11B2g CYP11B2

humanh CYP19

humani CYP17

90 95 66 98 51 94 81 84 82 20 15 33 34 35 25 54 36 68

2391 1729 5684 578 nd 1268 2117 503 1291 1615 2529 nd 639 763 nd 288 224 10

13 7 4 2 45 7 30 5 13 176 2834 448 334 411 nd 735 81 1

>36 000 4073 >36 000 814 nd 6045 1507 1787 3551 nd nd nd nd nd nd nd 40 30

2 37 7 62 nd 69k 57 27 23 nd nd nd nd nd nd nd nd 5

selectivityj factor 184 247 1421 275 166 71 97 99 9 1 2 2 0.4 10

Mean value of four determinations, standard deviation less than 10%. b Mean value of four determinations, standard deviation less than 20%. nd ) not determined. c Mean value of four determinations, standard deviation less than 5%. d Mean value of four determinations, standard deviation less than 10%. nd ) not determined. e S. pombe expressing human CYP11B2; substrate deoxycorticosterone, 100 nM; inhibitor, 500 nM. f Hamster fibroblasts expressing human CYP11B1; substrate deoxycorticosterone, 100 nM. g Hamster fibroblasts expressing human CYP11B2; substrate deoxycorticosterone, 100 nM. h Human placental CYP19; 1 mg/mL protein; substrate androstenedione, 500 nM. i E.coli expressing human CYP17; 5 mg/mL protein; substrate progesterone, 2.5 µM; inhibitor, 2.5 µM. j IC50 CYP11B1/IC50 CYP11B2. k IC50 CYP17 ) 1085 nM. a

Table 2. IC50 Values for CYP11B2 Inhibition of Selected Compounds

Figure 1. Lineweaver-Burk plot of 3 indicating competitive type of inhibition: ([) no inhibitor; (2) 3, 3 nM; (9) 3, 15 nM.

However, its corresponding pyridyl compound 2 is again highly active. To explain this interesting phenomenon, compounds 2, 13, and 23 were docked into the CYP11B2 model and the structural and energetic aspects of the protein-inhibitor interactions of the energetically most favorable docked complexes were analyzed. Subsequently the dynamics of the complexes were studied through molecular dynamics simulations of the docked complexes. In Figure 2, the best scoring docking poses for the three compounds are shown. The ligand-protein interactions are indicated by green lines. Figure 2b shows the least active compound 13. In addition to the Fe-N interaction with the heme group, it forms aromatic-hydrophobic interactions between its phenyl and imidazolyl ring and the binding pocket amino acids Leu227/Phe231 and Leu382/Thr318. Due to the imidazolyl/

amino acid side-chain interactions, a distortion of the Fe-N binding geometry can be observed. The corresponding highly active naphthalene compound 23, shown in Figure 2c, obviously forms phenyl-CH3 interactions with both aromatic rings of the naphthalene and Leu382, Thr318, Leu227, and Phe231. However, the imidazole moiety is forming only the Fe-N interaction, and no distortion of the Fe-N binding geometry by interactions with binding pocket residues was found. In Figure 2a, the docked complex of compound 2 is presented. Its binding pattern is the same as for compound 13. However, due to the larger size of the pyridine ring and the different geometry, the dihydronaphthalene moiety of the inhibitor is placed differently in the pocket, leading to interactions with other amino acids (Val129 and Ala313 instead of Leu227, Phe231, and Leu382). Its Fe-N interaction is not distorted as it is with compound 13. The main difference between the binding patterns of compounds 2/23 and compound 13 is the distortion of the Fe-N

2226

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 7

Voets et al.

Figure 2. Structure of the CYP11B2 binding pocket with the docked inhibitors 2 (a), 13 (b), and 23 (c). The structures shown correspond to the docking poses with the lowest energy score, and all amino acids of the binding pocket that interact with the inhibitors are given. The inhibitors are presented in yellow, nitrogen atoms are colored blue, and oxygen atoms are in red. The green lines correspond to the protein-inhibitor interactions found by FlexX. Table 3. Caco-2 Cell Permeation of Highly Active and Selective CYP11B2 Inhibitors by a Multiple Dosing Approach

Figure 3. Superimposition of the pyridine-substituted compounds 2, 4, 6, 8, and 9 (gray) and the imidazole-substituted compounds 12-14 (yellow).

interaction in the latter case, which indicates that this might be the main reason for the lower activity of compound 13. The same tendency as observed for the IC50 values (Table 2) can be seen for the energy scores of the docking poses shown in Figure 2, which are -50.137 for compound 2, -42.739 for compound 13, and -48.116 for compound 23. However, these energies cannot solely explain the large difference in the IC50 values of compounds 2/23 and compound 13. Obviously, not static interactions alone but also the dynamics of the system must be considered. Consequently, molecular dynamics simulations were performed for the docked complexes of compounds 2 and 13. Identical simulations had been carried out for compound 23 in a previous study,18 and a correlation between the stability of the docked poses during molecular dynamics simulations and the IC50 values of 23 and other inhibitors was observed. These results were confirmed in the present work: During the 1 ns simulations, compound 13 moved away (∼4 Å) from the heme cofactor, whereas compound 2 stayed close to its docked position, as did compound 23 in the previous work.18 This separates the three complexes in two classes: compound 2 and 23 form a strong and dynamically stable Feinhibitor interaction, whereas compound 13 does not. To further explain the differences in the orientation of the pyridyl and imidazolyl compounds in the binding pocket, superimposition of these two classes of compounds was performed. From Figure 3 it is clear that when the lone pairs of the heterocyclic nitrogens are pointing in the same direction, the corresponding dihydronaphthalene cores are located with a different angle toward the plane of the heme. Caco-2 Cell Permeability Screening and Metabolic Stability. Selected compounds (1-4 and 6-9) that showed high CYP11B2 inhibitory activity and selectivity toward other CYP enzymes were further screened for cell permeability properties on Caco-2 monolayers. The Caco-2 cells are a human epithelial colon carcinoma cell line and a suitable in vitro model for the evaluation of intestinal absorption of drugs.29 To increase the

compd

Papp (×10-6 cm/s)a,b multiple dosing ( SD

1 2 3 4 6 7 8 9 atenolol ketoprofene testosterone erythromycin

6.6 ( 0.5 14.6 ( 0.6 12.5 ( 0.2 13.5 ( 0.6 3.4 ( 0.1 21.6 ( 1.6 14.7 ( 0.6 16.3 ( 3.0 0.1 ( 0.03 25.7 ( 0.5 9.4 ( 0.2