J. Med. Chem. 1993,36,101-110
101
Nonpeptide Angiotensin I1 Antagonists: N-Phenyl-1H-pyrroleDerivatives Are Angiotensin I1 Receptor Antagonists Philippe R. Bovy,'J David B. Reitz: Joseph T. Collins: Timothy S. Chamberlain,$Gillian M. Olins,t Valerie M. Corpus: Ellen G. McMahon; Maria A. Palomo; John P. Koepke,t Glenn J. Smits: Dean E. McGraw: and Jeffrey F. Gawt Cardiovascular Diseases Research, Searle & Company, and Monsanto Corporate Research, Monsanto Life Sciences Research Center, 700 Chesterfield Parkway North, St. Louis, Missouri 63198 Received May 8, 1992
A series of 5-[1-[4-[(4,5-disubstituted-1H-imidazol-l-yl)methyl]-substituted] -1H-pyrrol-2-yl]-lH-substitutedl-lH-pyrrol-2-yl]-1Htetrazoles and 541- 14-[(3,5-dibutyl-1H-1,2,4-triazol-l-yl)methyll tetrazoles were investigated as novel AT1-selective angiotensin I1receptor antagonists. Computerassisted modeling techniques were used to evaluate structural parameters in comparison to the related biphenyl system. New synthetic procedures have been developed to prepare the novel compounds. The best antagonists in this series had ICs0 values (rat uterine membrane receptor binding) in the lo+ M range and corresponding pA2 in isolated organ assay (rabbit aorta rings). Structure-activity relationships indicate some similarities with the finding in the biphenyl system. Substitution on the pyrrole ring modulates activity. Compound 5 antagonized angiotensin-induced blood pressure increase when administered to conscious rat at 30 mg/kg per os.
Introduction Angiotensin I1 (AII) receptor antagonists, renin inhibitors, and angiotensin I converting enzyme (ACE) inhibitors have all been used to demonstrate the involvement of the renin-angiotensin system in essential hypertension.1 The development during the last decades of orally active ACE inhibitors has lead to a new, increasingly utilized therapy for treating hypertension. Peptidic AI1 antagonists have been available for over 30 years as a means to block the renin-angiotensin system; however, their therapeutic use has been severely limited by their partial agonist activity, rapid clearance, and lack of oral bioavailabilitye2 Recently, a series of imidazole-based compounds has been identified as antagonists to the angiotensin I1 r e ~ e p t o r .In~ particular, ~~ biphenylimidazoleswere found to produce potent antihypertensive effects upon oral administration. Bioisostericreplacement of the carboxylic acid by a tetrazole improved both in vitro binding affinity and in vivo oral antihypertensive activity. The general structure-activity relationship of this series of biphenyl compounds has been recently reported.6 The biphenyl tetrazole known as DUP753 or MK954 (1) is currently under development by Du Pont Merck (phase I11clinical trial) for treatment of hypertension. This compound and related analogues have high AI1 receptor binding affinity, are competitiveAI1 antagonists devoid of agonist activity, and are orally active and antihypertensive in animal models. This seminal discovery initiated a flurry of activity in pharmaceuticalresearch. Reviews covering this area have recently appearedee By and large, the remarkable pharmacological properties associated with the biphenyl fragment reported by Duncia and Carini focused research efforts on replacement of the azole fragment while preserving the biphenyl moiety.' Towhom correapondenceshouldbe add& Dr. Phdippe R. Bovy, MomantoCorporateW a r c h , Monsanto Life Sciencea Research Center, 700 Cheaterfield Parkway North,St. Louie, MO 63198. + M o m t o Corporate Research. Searle & Co.
*
0022-2623/93/1836-0101$04.00/0
Relatively few non-biphenyl-containing compounds endowed with AI1 antagonist activity have been reported aside from the original compounds reported by Takedas and the initial work by Du Pontes Noteworthy is a family of 4-(carboxybenzyl)imidazole-5-acrylicacid9 and their tetrazolyl equivalents.10 Typical of this c h is SKF 108666 which was developed from the early Takeda lead to enhance affinity by mimicking critical binding elementa of the natural peptidic agonist in its binding conformationell More recently, N-alkanoic acid derivatives were reported to have antagonist effect as well.12 In addition, phenyl-5-benzo[blfuran, phenyl-6-benzo[bl thiophene, and phenyl-lH-indolesurrogates13J4for the biphenyl have also been recently disclosed for which antagonist activity was claimed. Replacement of the terminal aromatic ring of the biphenyl with a 3-carboxy2-furanyl moiety reduced binding affinity by a factor of approximately 20. The presence of 2',6'-dimethoxy substituents on the biphenyl moiety was found to significantly (10-fold)decrease affinity for the receptor with respect to the unsubstituted analogue.16 A recent report revealed that naphthalene and tetrahydronaphthaleneare mediocre substitutes for the biphenyl spacer.16 We became interested in the hypothesis that N-phenylpyrrole could be a suitable replacement for the biphenyl group. The validity of the hypothesis was evaluated by molecular modeling methods, and several model compounds were prepared by new synthetic procedures. Several combinations of the two recognized pharmacophoric groups (the acidic function and the azole) were attempted. A modeling study was undertaken to estimate how 1-phenylpyrrolespacers (Table I, models I11and IV) would compare to the biphenyl spacers (Table I, models I and 11)in terms of distances between comparableatom. Computational methods have been employed to calculate the potential hypersurfaces of the anionic model structures I-IV at a STO 3G basis with respect to the two rotatable bonds labeled "tor a" and "tor b". The structures were fully optimized (except for tor a and tor b) at each point on the surface. The distances between the para carbon atom of the phenyl ring and the acidic group have 0 1993 American Chemical Society
Booy e t 01.
102 Journal of Medicinal Chemistry, 1993, Vol. 36, No.1 Table I. Comparison of the Carboxylic Acid and Tetrazole Derivatives in the Biphenyl and N-Phenylpyrrole Seriesa
Scheme 11. C02CH3
16
dlb d2c
I
I1
I11
IV
5.40 5.50
5.23 5.54
5.39 5.49
5.34 5.71
0 The four structures were anions for which the geometry was fully optimized a t the STO 3G basis ueing the ab initio electronic structure code, CADPAC version 4 (R.D. Amos and J. Rice, Cambridge University). b Distances expressed in angstroms between the para carbon of the phenyl ring and the carbon of the acidic fragment (independent of conformation due to axial symmetry). Distances expressed in angstroms between the para carbon of the phenyl ring and the center of gravity of the atoms of the acidic fragment (independent of conformation due to axial symmetry).
18: X = CCH20H, R3 = CI 20: X = CH20H, X = CI 21: X = N , R3 = nBu
CN
14
Scheme I*
17
CN
23
CHO
13
5N 14
C\02CH3
16
0 (a)p-Toluidine; acetic acid; DMF, POC13. (b) NHzOH; (AcO)nO. (c) KOH, ethylene glycol; MeOH, HC1.
been measured on the models of the optimized structures and fall within a 0.23-A range. This result indicates that the pyrrole derivatives envisaged for synthesiswould allow similar relative spatial positioning of the acidic and azole pharmacophoric groups. In derivatives I1 and IV, the bulkier tetrazole rings force the biaryl system out of planarity slightly more than the carboxylic group in derivatives I and 111. The energy difference between the more stable conformations for the carboxylate and the tetrazole derivative is more pronounced in the biphenyl system than in the N-phenylpyrrole system. The potential energy surfaces indicate that the carboxylic acid minima correspond to sharp w e b 4-6 kcal/ mol below a mean valley that extend alongthe tor a minima and encompassesthe entire range of tor b values. On the other hand, the tetrazoles have less well defined minima and can rotate more freely (lower energy increase)around their minimal energy conformations. However, in all fragments, torsional angles of 90° encounter a very high energy barrier. Solvation, although not included in this study, would lead to the same general conclusions as the molecules have similar electronic structures. Consequently, we felt that, from a structural point of view, the N-phenylpyrrolefragment provided an excellent substitute for the biphenyl fragment as a spacer. Thus, we set forth to investigate a family of substituted Ntolylpyrroles 2-11 for potential AI1 antagonist activity. Chemistry Scheme I describes a general synthetic pathway to the l-p-tolyl-2-cyano-l~-pyrrolederivative (14) based on a described procedure.17 Treatment of 4-aminotoluidine with 2,6-diethoxytetrahydrofuranin glacial acetic acid at reflux gave good yield of the desired pyrrole. Formylation
26: X = R3 = CONH2 24:X=N,R3=nBu a (a) NBS, CC4, AIBN. (b) 19,22, or 2k NaH or tBuOK, DMF. (c) NaOH, H20. (d) MesSnN3, HCl(g), EtOH or Si02.
of the l-p-tolylpyrrole was carried out according to the Vilsmeyer-Haack reaction. The l-p-tolyl-2-formyl-1Hpyrrole (13) derivative was then transformed into the corresponding oxime by treatment with hydroxylamine hydrochloride which furnished the l-(rl-methylphenyl)2-cyano-lH-pyrrole (14) when treated with acetic anhydride. We found that the nitrile 14provides an easyaccese to both the carboxylic derivatives 16 and the tetrazole derivatives. Scheme I1 describes the sequence of reactions which led to the preparation of the targeted molecules 2-6. We found that radical bromination of the benzylic position of the ester 16 was readily achieved with NBS in carbon tetrachloride in the presence of AIBN or dibenzoyl peroxide. Typically, in this reaction, traces of the a,adibromo derivative were generated. Methyl l-(a-bromot o l y l ) - ~ - p ~ o l e - 2 - ~ b o x y l(17) a t ewas used to alkylate the azoles 19 and 28. The alkylation was performed in dimethylformamideon the azole’s anions generated with NaH or tBuOK. In the case of 2-n-butyl-4(6)-chloro-6(4)(hydroxymethy1)imidazole (19), two regioisomers were produced which were separated by silica gel chromatography. Hydrolysis of the ester with dilute aqueous base completed the sequence leading to compounds 2-4. The radical brominationprocedure on 1-(4methylphenyl)-lHpyrrole-2-carbonitrile (14) led to 1-[4-(bromomethyl)phenyll-W-pyrrole-2-carbonitrile(231, which was then used to alkylate the anions of azoles 22 and 26 to provide respectively the nitriles 24 and 26. The tetrazole derivatives were obtained by a l,&dipolar cycloaddition with trimethyltin azide as described by Sisido et al.l* The N-(trimethylstanny1)tetrazoles can be converted to the free tetrazole by anhydrous hydrogen chloride in an ethereal or alcoholicsolution. We have also observed that the passage through a silica gel column was usually an effective procedure to cleave the trimethylstannyl group.
Nonpeptide Angiotensin 11 Antagonists
Journal of Medicinal Chemistry, 1993, Vol. 36, No. 1 103
Scheme 1111
Scheme IV* H
O
C
~
O
CHaO CI
23 27
H
3
C
e
N
L
L C ~ H 3~c + & H o CHNOH y
-
35 H3C+NTcN
36
37
cF3
14
29
n-BU
F.
14
32
7
(a) p-Toluidine, AcOH, refl, 2 h. (b) NHzOH, NazCOa aq. (c) (AcO)zO,refl, 3 h. (d) NBS, CCh, AIBN. (e) tBuOK, 22, DMF,rt. (f) MeaSnN3, HCl(g), EtOH or Si02. 0
As described in the Discussion, we were interested in introducing electron-withdrawing substituents on the pyrrole ring. These Substituents were introduced at various stages of the synthetic scheme. Chlorination in positions 3 and 5 of 1-[4-(bromomethyl)phenyl]-lHpyrrole-2-carbonitrile(23) was obtained by treatment with 2 equiv of N-chlorosuccinimide (Scheme 111). 144Methylphenyl)-lH-pyrrole-2-carbonitrile(14) was trifluoromethylated photochemically with trifluoromethyl iodide in the presence of m e r ~ u r y . ~ Fluorination 9*~ of the same nitrile intermediate was achieved in 25% yield by the use of xenon difluoride in dichloromethane.20 On the other hand, the nitro group on the 3 position of the pyrrole was obtained from direct nitration of the nitrile 24, in conditions previously described21for related compounds. In all cases, the position of the substituents on the pyrrole ring was deduced from the analysis of the proton and/or carbon NMR spectra and from analogy with other cases previously reported in the 1iterature.z2 Scheme IV describes the synthesis of the 3-tetrazole derivative 7. The construction of l-p-tolyl-3-formyl-Wpyrrole (38) was conducted in one efficient step from p-toluidine and 3-formyl-2,5-dimethoxytetrahydrofuran. All subsequent synthetic steps were adapted from the sequence described in Schemes I and 11. Biology
Newly synthesized compounds were evaluated for specific binding to rat uterine AI1 receptors (ICw) and for antagonismof AII-induced contraction of rabbit aorta rings (pAz).% In the binding assay, the ability of compounds to prevent lZI-angiotensin I1 binding to a rat uterine membrane preparationu was determined, and the calculated IC@ values are listed in Tables I1 and 111. Control experiments indicated an ICs0 value for AI1 of 2.2 nM. All
new compounds produced a biphasic displacement curve, indicating the presence of high-affinity (80%)and lowaffmity (20%) binding sites with ICs0 values ranging between 30 nM and 2 pM for the high-affmity sites and between 10 and 710 pM for the low-affinitysites. Recent studies using selective AI1 receptor ligands have revealed that there are two AI1 receptor subtypes in various target t i s s ~ e s .Specifically, ~ ~ ~ ~ ~ peptides27 and non-peptides28 have been identified that bind selectively to the highaffinity (AT1) and to the low-affmity AI1 receptor population (AT& Analysis of the structure-activity relations among these compounds reveals that the AT1 receptors are involved in vascular contractile activity and blood pressure regulation while the function of AT2 receptors remains largely unknown. The binding assay was followed with a functional test for the ability of a compound to antagonizethe angiotensin 11-induced contraction of isolated rabbit aorta rings.m From this assay, pA2 values were derived and are listed in Tables I1 and 111. In this assay, the compounds bated shifted the AI1 concentration-response curve to the right in a concentration-dependent and parallel manner. Even in the presence of 0.1 pM of the compounds testad, the maximal response to AI1 was attainable, indicating the fdyreversible and competitivenature of the antagonism. No agonist effect was observed for the compoundstested. The in vivo activity was determined by assessing the inhibition of the pressor response to a 50 nglkg per min iv infusion of angiotensin I1 in conscioua rate. Figure 1displays the dose-response curve for inhibition of AIIinduced pressor response of compound 8. A dose of 30 mg/kg administered id blocked the pressor response for at least 3 h. Discussion Figure 2 depicta several key compounds from this work and from the literature to help evaluate the effect of replacement of the terminal phenyl ring of the biphenyl by a pyrrole ring. In the carboxylic acids series, the introduction of the N-phenylpyrrole fragment gave compounds which are
104 Journal of Medicinal Chemistry, 1993, Vol. 36,No.1
Bouy e t al.
Table 11. Comparison of Angiotensin I1 Antagonist Activity (Measured by Binding Affinities and pA2) of the New N-Toluylpyrrole Derivatives 2-6, Containing Various Heterocycles and Acidic Functions, to DUP 753, 1, and the Corresponding Carboxylic Acid 40
1 (DUP 763) 40 ( E D 7711 (1MI))g 2
3 4 5 6
c1 CHzOH n-CdH9 n-CdHg CONHz
CCH2OH cc1 N N CCONHz
CO2H CO2H CO2H CNiH CNiH
3tv
>1m
170 870 1100 270 33 290
3300 59000 140000 159000 260000 3 m
8.lf 7.1 5.9 6.1 6.3
8.0 5.9
183.5-184.5 168.0-170.0 180.0-181.5 181.0-182.5 107.5-108.6 125.5-127.0 216.0-221.0
a New
compounds were identified by a combination of their spectroscopic data and confirmed by exact mass measurement and/or analysis for C, H,N. b Concentration of the test compound inhibiting specific binding of AI1 by 50% was derived from analysis of plots of the percentage of specific binding VE the log concentration of the compound (at least 10 concentrations ranging over 5 orders of magnitude were used). Data are derived from one experiment performed with doses in triplicate. The IC60 values for each receptor subclass separately were obtained by the methodology described in ref 23. pA2 values were obtained as described in the Experimental Section. e The low pA2 value for this compound may be due to a low solubility in the conditions of this assay. 'Literature values are 19 nM for the affinity to rat adrenal cortex receptors and 8.48 for rabbit aorta pA2 (ref 3). 8 See refs 15 and 24. Table 111. Influence of Substitution of the Pyrrole Ring on Angiotensin I1 Antagonist Activity As Measured by Binding Affinities and PA2
no.' 5 7 8 9 10 11
R3
H H 3,5-C1 5-CF3 5-F
Q 2-CNdH 3-CN.jH 2-CNdH 2-CN4H 2-CN4H 2-CNJi
AT1 33 2000 52
AT2 26oooO 2 2 m 81oooO 450000 23oooO 93oooO
88 72
480
pAzd
m n "C 125.5-12
