A novel class of "GABAergic" agents: 1-aryl-3-(aminoalkylidene

Feb 1, 1989 - Note: In lieu of an abstract, this is the article's first page. Click to increase image size Free first page. View: PDF. Citing Articles...
0 downloads 0 Views 1MB Size
437

J . Med. Chem. 1989, 32, 437-444 acetonitrile-H20 (2:l) mixture.

Preparation of No-Carrier-Added "Tc Complexes of BPA-BAT. The necmier-added =TeBPA-BAT was prepared by a ligand-exchange reaction. A solution of 99mTeglucoheptonate (1-10 mCi, 0.1-1 mL), prepared by adding a desired amount of sodium [&Tc]pertechnetate into a freeze-dried kit (1 mg of glucoheptonate + 0.25 mg of stannous chloride), and 1mg of ligand was heated at 100 "C for 30 min. The product was extracted with chloroform (3 X 2 mL), and the combined extracts were dried over anhydrous sodium sulfate. The solution was then condensed with a stream of air and redissolved in saline (radiochemical yield >80%). No decomposition of the &Tc complex was observed on standing a t room temperature for 18 h. For separating the syn and the anti isomers, the racemic mixture was injected into HPLC and appropriate fractions were collected (Table I). After condensing, reextraction with ethyl acetate (3 X 1 mL), and condensing with a stream of nitrogen, the products were redissolved in saline (total radiochemical yield >50%,radiochemical and isomeric purity >98%). Partition Coefficients. The partition coefficient was measured by mixing the &Tc-BAT compound with 3 g each of 1-octanol and buffer (pH 7.0 or 7.4, 0.1 M phosphate) in a test tube. This test tube was vortexed for 3 min at room temperature and then centrifuged for 5 min. Two weighed samples (0.5 g each) from the 1-octanol and buffer layers were counted in a well counter. The partition coefficient was determined by calculating the ratio of counts per minute/gram of octanol to that of buffer. Samples from the octanol layer were repartitioned until consistent partition coefficient values were obtained. The measurement was repeated three times. Animal Distribution Studies. Male Sprague-Dawley rats (200-300 g) were injected intravenously (femural vein under ether anesthesia) with 0.2 mL of a saline solution containing the -Tc-BAT complex (0.5-20 ,uCi). At selected intervals following the injection, blood samples (1mL each) were collected by cardiac puncture, and the rats were sacrificed immediately thereafter by cardiectomy. The organs of interest were subsequently excised, weighed, and counted in a dual-channel automatic gamma counter ( B e c k " 5500). The % dose/organ values were determined by comparison of the tissue radioactivity with suitable dilutents of the injected dose. The % dose/gram values were computed from the % doselorgan values and the corresponding mean organ weights (mean organ weights: heart, 0.85 g; brain, 1.65 g; blood, 18 g; liver, 9 g; Kidneys, 1.9 g; lungs, 1.6 g). Finally, the braixblood ratio was calculated from the corresponding % dose/gram values. X-ray Crystallography. Crystal Data. wTc-BPA-BAT (syn): TCS~ON&ZH~,,dark brown, crystallized from 2:1 MeCN-water, monoclinic-b, P 2 , / n (no. 14), a = 15.241 ( 5 ) A, b = 15.658 (7) A, c = 11.385 (3) A, 0 = 109.91 (2)", from 24 reflections, T = -70 "C, V = 2554.6 A3, Z = 4, FW = 536.69, D c d d = 1.395 g/cm3, ,u(Mo) = 7.26 cm-'.

T e B P A - B A T (anti): TcS2ON4CBH3,red, block, -0.30 x 0.40 0.35 mm, crystallized from acetonitrile-water, monoclinic-b, P2,/c (no. 14), a = 12.243 (7) A, b = 11.022 (2) A, c = 19.180 ( 5 ) A, p = 102.19 (3)", from 20 reflections, T = -70 "C, V = 2529.8 A3, Z = 4, FW = 537.70, Ddd = 1.411 g/cm3, ,u(Mo) = 7.33 cm-'. Data Collection and Treatment. For the syn isomer: Enraf-Nonius CAD4 diffractometer, graphite monochromator, Mo K a radiation, 6607 data collected, 2.6" < 28 3.0u(I). For the anti isomer: Enraf-Nonius CAD4 diffractometer, graphite monochromator, Mo KCYradiation, 6607 data collected, 2.2" < 28 M) and it had no apparent effect on GABA levels. Like other GABAergic compounds, F however, it antagonized cyclic GMP elevations in rat X h cerebellum induced by another glutamate decarboxylase inhibitor, isoniazid, and it enhanced the binding of [3H]E flunitrazepam in vivo.% This interesting profile prompted us to investigate more fully the structure-activity rela19 tionships in this class of compounds. H z Chemistry NR; R'z The l-aryl-3-(aminoalkylidene)oxindolesE and G were accessible by a variety of routes, which are summarized in Scheme I. Our initial synthesis involved the preparation of 1-aryloxindoles C via the method of Stoll6 et al.1° Thus, condensation of the commercially available diX phenylamines A with chloroacetyl chloride in refluxing benzene or toluene gave an essentially quantitative yield G of the a-chloroacetanilides, which could then be cyclized by heating in AlC13 to a melt at 180-200 OC for 10-15 min X = 3-C1, Y = H, 40%; 87, X = H, Y = 6-C1,7%; 88, X (pathway a). In the case of 3-chlorodiphenylamine, we = H, Y = 4-C1,5%),which were readily separated by silica obtained by this process all three possible isomers of C (86, gel chromatography. A more attractive and general route to 1-aryloxindoles, which avoided the formation of positional isomers, involved the preparation of arylindoles B (1) For recent reviews, see: (a) Enna, S. J. In Neuropharmacology of Central Nervous System and Behavioral Disorders; Acaby a modified Ullmann reaction of indole with substituted demic Press: New York, 1981; pp 507-537. (b)Lloyd, K. G.; aryl iodides or bromides in the presence of cuprous bromMorselli, P. L.; Deportere, H.; Fournier, B.; Zivkovic, B.; ide and sodium carbonate in refluxing N-methyl-2Scatton, B.; Broekkamp, C.; Worms, P.; Bartholini, G. Pharpyrrolidinone and subsequent conversion with N-chloromacol. Biochem. Behau. 1983,18, 957. (c) Bartholini, G. Med. succinimide and H3P04/acetic acid" to C. Several atRes. Reu. 1985, 5, 55. tempts to prepare C by a one-step condensation of sub(2) (a) Tallman, J. F.; Thomas, J. W.; Gallager, D. W. Nature 1978, 274, 383. (b) Costa, E. Arzneim.-Forsch. 1980,30, 858. stituted anilines with o-chlorophenylaceticacid catalyzed (3) Krogsgaard-Larsen,P.; Falch, E. Mol. Cell. Biochem. 1981,38, by copper oxide12 gave very poor results in our hands. 129. Treatment of the 1-aryloxindoles C with dimethylform(4) Hill, D. R.; Bowery, N. G. Nature 1981,290, 149. amide or dimethylacetamideacetals in refluxing CHC13 for (5) Meiners, B. A.; Salama, A. I. Eur. J. Pharmacol. 1982, 78,315. 1-6 h gave excellent yields of compounds E with Z = H (6) Johnston, G. A. R.; Krogsgaard-Larsen, P.; Stephanson, A. or CH3 (pathwayd). Use of 1equiv of the acetal was found Nature 1975,258, 627.

&

6

di

/

A

w

t

d3

(7) Lewis, J. R. J . Am. Med. Assoc. 1978, 240, 2190. (8) Metcalf, B. W. Biochem. Pharmacol. 1979, 28, 1705. (9) (a) Loscher, W. Neuropharmacology 1982, 21, 803.

(b) Loscher, W. Biochem. Pharmacol. 1979,28,1397. (c) Koe, B. K.; Minor, K. W.; Kondratas, E.; Lebel, L. A.; Koch, S. W. Drug. Dev. Res. 1986, 7, 255. (10) Stoll6, R. Chem. Ber. 1914, 47, 2120.

~~

(11) Lo, Y. S.; Walsh, D. A.; Welstead, W. J.; Mays, R. P.; Rose, E.

K.; Causey, D. H.; Duncan, R. L. J . Heterocycl. Chem. 1980, 17, 1663. (12) Bergmann, E. D.; Bilk-Sam& T. Bull. SOC.Chem. Fr. 1968, 1090.

A Novel Class of "GABAergic" Agents

to be optimal, since excess acetal resulted in poorer yields and the formation of byproducts. The compounds in which the a substituent formed a ring with one of the N substituents (e.g., the pyrrolidinylidene derivatives) were accessible by reaction of the aryloxindoles C with either lactam acetals or with a Vilsmeir-Haack complex formed from the lactam and POC13 (pathway d). In the latter case, preparation of the Vilsmeier complex with POCl:, a t 0 "C, followed by treatment with an oxindole in a suitable solvent, gave after several hours a t 50-80 "C the desired compounds E, along with varying amounts of 2-chloroindoles as byproducts. The Vilsmeier complex method was also successful with N-methylformanilide to give 49, and to a lesser extent with NJV-dimethylbenzamide to give 60. The compounds in Table IV with R = H (72,77, and 78) were prepared by heating the 1-aryloxindole C with an excess of the 0-methyl lactim in toluene':, (pathway d). The congeners of E with varying amine components were available in most cases by treating E (NR,R, = NMeJ with an excess (2-5 equiv) of the desired amine either neat or in ethanol solution (pathway g). In some instances, this displacement of the dimethylamino group could not be effected even under more forceful conditions (i.e., refluxing ethanolic solution), but compounds such as 40,46, and 61 were accessible via the intermediate hydroxy- or alkoxyalkylidenes D (pathway c, e) according to the methods of Winn and Kyncl.14 We also prepared a large number of l-aryl-3-(aminoalky1idene)oxindolesby converting oxindoles H to 3-(aminoalky1idene)oxindoles F (pathway h), followed by an Ullmann arylation step (pathway f). While this provided an efficient route to achieve variability in the pendant aryl ring, the yields to step f were good only when Z = H. Thus, oxindole, or substituted oxindoles (prepared from reduction of substituted isatins or oxidation of substituted indoles), were converted with dimethylformamide or dimethylacetamide acetals in refluxing CHC13to the 3-(aminoalkylidene)oxindoles, (i.e., Z = H or CH3, 99-105, Table V). These intermediates were allowed to react with NaH in dry DMF to give the insoluble sodium salts, which were then treated with a variety of substituted aryl bromides or iodides in the presence of Cu2Br, at elevated temperatures for 24-72 h to give the compounds E (3,4, 6-36) in yields ranging up to 72%. Although the ethylidene homologues such as 52 could also be prepared in this manner, yields were often extremely poor (8% for 52), possibly due to competitive deprotonation of the a-methyl group; pathways a,d or b,d were clearly preferable for the preparation of the ethylidene compounds. Similarly, some pyrrolidinylidene oxindoles F (e.g., 106 and 107) were prepared by condensation of substituted oxindoles with lactam acetals (pathway h), and subsequent Ullmann arylation (pathway f) gave compounds E (64-68,79-81). Again, the yields in this arylation step were generally poor, but adequate to obtain test samples.

Pharmacology Antagonism of MPA-Induced Convulsions.15 Charles River male mice, Swiss CD strain, 17-21 g, were fasted for 18 h prior to testing. Compounds were administered subcutaneously at 100 mg/kg in a vehicle consisting of 5% ethanol, 5% emulphor 620, and 90% saline. Vehicle alone served as a control treatment. If active, the com(13) Glushkov, R. G.; Volokova, V. A; Magidson, 0. Y. Khim.-Farm. Zh.1967, 1(9),25; Chem. Abstr. 1968, 68, 105143f. (14) Winn, W.; Kync!, J. J. U.S. Patent 4,145,422. (15) Adcock, T.; Taberner, P. V. Biochem. Pharmacol. 1978, 27, 246.

Journal of Medicinal Chemistry, 1989, Vol. 32, No. 2 439

pound was further evaluated at lower points on a 0.5 X log,, dosage scale to determine an EDm value. Solution concentrations were varied a t different doses to provide a constant injection volume of 10 mL/kg. Groups of five mice were treated with each test compound and 1h later with MPA, 32 mg/kg, intraperitoneally. The animals were then observed continuously for 10 min. In untreated mice this MPA challenge caused clonic convulsions within 4 min of treatment. Protection against MPA convulsions in a given mouse was said to occur if no convulsions occurred during the 10-min test period. Enhancement of [3H]Flunitrazepam Binding in Vivo. [3H]Flunitrazepam ( [3H]FNP)binding to mouse brain was measured as described previouslyQcJ6by the method of Chang and Snyder.17 Groups of five mice were injected intraperitoneally with 320 pmol/ kg of the test compound or vehicle 1 h prior to an intravenous injection of 200 pCi/kg [3H]FNP. Twenty minutes after the [:,H]FNP injection, the mice were sacrificed by cervical dislocation, and the brains were removed and immediately frozen. Each brain was weighed quickly and homogenized in 40 volumes (w/v) ice-cold 50 mM Tris-HC1 pH 7.7 buffer using a Brinkmann Polytron. Triplicate 1.0-mL samples were filtered through Whatman GF/B glass fiber filters under vacuum and washed with two 5-mL aliquots of the ice-cold buffer. The bound [3H]FNP was measured by adding the filters to vials containing 10 mL of Aquasol-2 and counting the radioactivity. Bound [3H]FNP for drug-treated mice was calculated as percentage of bound [:,H]FNP for vehicle-treated mice. Antagonism of Cyclic GMP Accumulation after Isoniazid. Experiments to determine the effect of drugs on cyclic GMP accumulation after isoniazid were conducted as described by Koe and Lebel.18 Sprague-Dawley CD male rats received ip drug or vehicle 90 min before administration of isoniazid, 450 mg/kg sc (n = 7). Thirty minutes later the rats were killed by focused microwave irradiation of the brain. The cerebellum was dissected and assayed for cyclic GMP by radioimmunoassay. Results and Discussion The SAR results are summarized in Tables I-IV, and key oxindoles are compared with reference compounds in Table VI. Analogues of 2 with modification of substituents in the pendant phenyl ring and in the aromatic portion of the oxindole ring are listed in Table I. The 3-chloro derivative 2 was clearly more potent in antagonizing MPA than were the 2- or 4-chloro congeners 3 and 4. Re9lacement of chlorine by fluorine, methyl, or ethyl gave MPA-active agents 5, 7, and 10; 7 was again more potent than the 4-methyl congener 9, although the 2methyl congener 8 had surprisingly good MPA activity. Both the 3- or 4-methoxy derivatives (13 and 14) were weakly MPA active. Replacement of the chlorine in 2 with CF3gave good MPA activity, but substitution by a variety of other electron-withdrawing substituents led to inactive compounds (17-20). Potency enhancement was observed by combining the 4-methoxy group with a 3-chloro, 3fluoro, or 3-methyl substituent (26,27,29) but surprisingly not by combining the 3- and 4-methoxy groups (31). Addition of chlorine in any of the positions of the oxindole nucleus of 2 led to inactive compounds (32-36). The influence of substituting the dimethylamino group of 2 with other amines is shown in Table 11. The MPA (16) Koe, B. K. Drug. Deu. Res. 1983, 3, 421. (17) Chang, R. S. L.; Snyder, S. H. Eur. J . Pharmacol. 1978, 48, 213. (18) Koe, B. K.; Lebel, L. A. Eur. J. Pharmacol. 1983, 90,97.

Surges et al.

440 Journal of Medicinal Chemistry, 1989, Vol. 32, No. 2

Table I. 3-[(Dimethylamino)methylene]oxindoles

yield, no. 1 2 3 4

X

Y

method D1 D1 G G D1 G G G G Gh G G G G G G G G G G G G G G G

~ p , C 135-g 116-8 169-71 176.5-9 87-88.5 118-21 105-8 105-7 135-7 112-4 131-3 200-3 106-8.5 136-8 99-102h 103-6 113-6

MPA antagonism: EDSO >100 26.1

recrystal solvent" Et20 EtiO Et20.P Et20 Et20,PE Et20 Et20.P

[3H]FNP bound: % isoniazid control f antagonism,e S.E. ED50 150 f 6 138 f 6 1-3.2 131 f 6 3.2-10 117 f 4 191 f 11

formulab H H 47 Ci7H~Nz0 87 Ci7Hi,CiN20 3-C1 H >loo 12 2x1 H C17H15C1N20 10 C17H15C1N20*1/4H20 >loo 4-C1 H 39.9 40 C17H15FN20m1/4H20 H 5 3-F 24 >loo H 6 3-Br C17H15BrN20 14.1 184 f 128 38 H 7 3-CH3 C18H18N20 18 10 H 8 2-CH3 C18Hl8N20 56 PhH-PE C18H18N20 23 H 9 4-CH; 14.1 H 39 10 3-Et Et20.P C19H2oNzO >loo 34 H 11 3-Ph C23H20N20 Et20 >100 83 CHZC12.H C17H16N202 H 12 3-OH 217 f 9 100 51 H 17 %NO2 EA-H C17H15N303 >loo 18 120-2 H 18 3-CHO Et20 C18H16N202 192 f 9 >loo 22 140-3 H 19 3-CN EA*H C18H15N30 >loo 24 180.5-3 Tol*PE H 20 3-CONMez C20H21N302~'/4H20 >loo 32 192-5 H 21 4-NMez EA C19H21N30 56 32 144-7 H 22 4-SCH3 EA-P C18H18" 147 f 4 -10 >100 19 161-3 H 23 2,5-C& Et20 C17H14C12N20 0.1-0.3 144 f 5 >100 28 149-51 H 24 3,4-C12 Et20 C17H14C12"20 153 f 6 >100 72 147-9.5 H 25 3,5-C12 C17H14C12N20 244 f 4 7.2 G' 38 148-51 H 26 3-C1, 4-OCH3 PhH*P C18H17C1N202 194 f 8 6.6 G 10 113-4 H 27 3-F, 4-OCH3 Et20.P Cl8H17FN202 149 f 11 EtzOmPhH C18H17ClN202 27 161-3 >loo G H 28 2-OCH3, 5-C1 9.1 18 172-4 G' PhHePE CigHzoN202 H 29 3-CH3, 4-OCH3 15.3 33 147-50 Gk 30 343, 4-OMe, 5-Me H Et20.P C19H19C"2 138f 6 100 6 164-7 H G 31 3,4-(OMe)2 Et20 C19H20N203 >loo 5 CHC13.EtzO C17H14C12N20 G 169-70 4-C1 32 3-C1 24 165-8 >loo G 5-C1 33 3-C1 Cl7Hl4C12N20 >loo 2 128-30 G 6-C1 34 3-C1 CH C17H14C12N20 169 f 2 31 172-5 >100 G 7-C1 35 3-C1 MeOH.Et20 C17H14ClzNz0 >100 9 128-30 G 5-OMe 36 3-C1 MeOH.Et20 C18H17C1Nz02 18 148-51 >100 PhH-P C,,Hi 37 H 6-C1 D1 ., XINqO " a Recrystallization solvents: P (pentane); P E (petroleum ether), PhH (benzene); H (hexane), To1 (toluene), EA (ethyl acetate), IPE (isopropyl ether), CH (cyclohexane). *Analysis for C, H, and N was within f0.4% of the theoretical values. "Dm, mg/kg sc. dCompounds were administered a t 320 pmol/kg ip, 60 min before [3H]FNP,pCi/kg iv, and 80 min before killing, unless otherwise noted. P C0.05 or less vs contemporary controls for all entries except as indicated by NS (nonsignificant). eED50,mg/kg ip, for antagonism of isoniazid induced elevation of cyclic GMP in mouse brain. fBelgian Patent 849, 626; mp 130-2 "C. 8 At 178 pmol/kg, ip. "Prepared from 3-ethyliodobenzene (Homer, L.; Subramanian, P. Ann. Chem. 1968, 714,91; bp 51-52 "C (0.05 mmHg)). 'Prepared from 2-chloro-4-iodoanisole (Matheson, D.; McCombie, H. J. Chem. SOC.1931, 1103; mp 83 "C). 'Prepared from 4-bromo-2-methylanisole (Nasipuri, D.; Roy, D. N. Chem. Abstr. 1960, 59, 8803e mp 68 "C). kPrepared from 4-bromo-2-chloro-6-methylanisole (bp 122-124 "C (0.9 mmHg)) which was prepared in 81% by methylating 4-bromo-6-chloro-o-cresol (Aldrich) with dimethyl sulfate in acetone. %

."

activity of 2 was shared only by the primary amine analogue 39 and by the diethylamino analogue 40. The inactivity of the monomethyl amine derivative 38 was especially surprising. In this regard it is interesting to note that all alkylideneoxindoles of this series are mixtures of rapidly equilibrating 2 and E isomers as measured by NMR.lg In derivatives with a proton on nitrogen, such as 38 and 39, the 2 isomer predominates in solvents such as chloroform by stabilization via a hydrogen bond to the oxindole carbonyl. However, a preferred 2 configuration cannot explain the inactivity of 38, because a similar preference is shown by the MPA active analogue 39.

The compounds with the a-substituted side chain listed in Table I11 were prepared in an effort to decrease, by steric hindrance, the hydrolysis of the aminoalkylideneoxindoles to the corresponding (biologically inactive) 3acyloxindoles observed in acidic media. For instance, compound 2 was hydrolyzed under acidic conditions (50% methanol-water, pH 2) with a half-life of 35 min.20 To our surprise, while the introduction of the a-methyl group doubled the potency with some (e.g. 5 1 vs 2,54 vs 26) but not all (e.g. 52 vs 7, 55 vs 27, 57 vs 39) compounds, it decreased dramatically the acid stability, contrary to our expectations. For example, 51 was hydrolyzed a t pH 2 in

(19) We are grateful to Dr. E. B. Whipple of Pfizer Central Research for these NMR determinations.

(20) We are grateful to Dr. T. A. Hagen of Pfizer Central Research

for these stability studies.

Journal of Medicinal Chemistry, 1989, Vol. 32, No. 2 441

A Novel Class of "GABAergic" Agents

Table 11. Amino Variations at the 3-Position

[3H]FNP bound,d % control f S.E. 138 f 6 119 f 3 134 f 4

MPA antagonism: ED, 26.1 >100

isoniazid mp, recrystal antagonism: no. W method % OC solvent" formulab EDw 2 NMe2 D1 87 116-8 Et20 C17H&lN20 1-3.2 38 NHMe F2 73 151.5-3 EtOH C1eH13ClN20 39 NHZ F2 93 22.1 Et0 C15H11ClN2O*'/,H20 141-3 40 NEtz F1 59 74-5.5 CigH1gClN20 18 Et20.P 41 NHPr F2 88 144-6 ClBH17C1N20 >56 f 42 NPrz F2 54 68-71 C21H23ClN20 >100 Et20 43 pyrrolidino F2 78 108-12 ClgH17ClNzO >100 Et20 153-5 44 piperidino F2 62 >56 131 f 5 C2OH&lN20 EA-H 45 morpholino F2 80 130-2.5 ClgH17ClN202 >100 121 f 3 Et2O.P 164-6 46 imidazo F1 11 >100 139 f 8 Cl8Hl2C" To1 47 NHCHzPh F2 90 97-9 CzzH17ClNzO >100 EtOH F2 3 118-20 EA*H C23H1gClN20*'/2HzO >lo0 135 f 5 48 N(Me)CH2Ph 49 N(Me)Ph E 19 127-9 Et20.P C22H17ClNzO >100 133 f 6 50 OEt C 17 132-4 EA-H C17HikClN02 >100 "See note a , in Table I. bSee note b, in Table I. cSee note c, in Table I. dSee note d , in Table I. eSee note e, in Table I. 'Precipitated from reaction. yield,

e

Table 111. Branched 3-(Aminoalkylidene)oxindoles

X

Y 2 3-C1 H H 51 3 4 1 H 52 3-Me H 53 4-OMe 54 341, 4-OMe H 55 3-F, 4-OMe H 56 H 6-C1 H 57 3-C1 58 H 6-C1 6-C1 59 H H 60 3-C1 H 61 3 4 1

no.

W NMez NMe, NMez NMe2 NMe, NMez NMez NH2 NH2 pyrrolidino NMez NH2

2 H Me Me Me Me Me Me Me Me Me Ph CF3

meth- yield, ~fl-p, od % C D1 87 116-8 D1 36 113-6 G 8 96-9 D1 66 176-7.5 D1 74 138-41 D1 48 187-9 5 118-21 D1 F2 70 200-2 29 170-2.5 F2 F2 28 128-31 E 4 124-6 F1 84 197-200

recrystal solventa Et20 Et20 PE Et20 Et20 Et20 Et20.PE EtOH EtOH Et20.P Et2O.P

formulab

MPA antagonism,c ED*n 26.1 13.3 17.8 >100 2.8 >100 35.9 I I

>loo >56 >56

>loo

[3H]FNP isoniazid bound: % control i antagonism: S.E. ED, 138 f 6 1-3.2 136 f d 1

-

157 f 98 103 f gNS 213 f 9

>56

h

"See note a, in Table I. bSee note b, in Table I. CSeenote c, in Table I. dSee note d , in Table I. eSee note e , in Table I. 'At 178 pmollkg ip. # A t 100 pmol/kg ip. hPrecipitated from reaction.

50% methanol-water with a half-life of only 2 min.20 Therefore, the observed potency increase must be due to enhanced intrinsic activity, but the relative acid instability made these derivatives less attractive for further development. The inactivity of 55 against MPA convulsions was surprising but can probably by explained by poor absorption after sc administration; 55 showed an ED, of 32-100 mg/ kg after oral administration. The inactivity of 57 was probably due to extreme insolubility of the compound in vehicle, leading to poor absorption. An interesting class of compounds was ultimately obtained by forming a ring between the nitrogen and the a position of the side chain. Excellent MPA activity was found in the N-methylpyrrolidinylidene compounds 66,69, and 70 (Table IV). Of these, 66 appeared to be particularly

interesting since it did not exhibit the twitches and myoclonic jerks seen with 69 and 70 a t higher doses. At low pH compound 66 was converted to the ring-opened 3-(4aminobutanoyl)oxindole,which can readily recyclize to 66, thereby resulting in greatly enhanced overall acid stability for 66 relative to 51 or 2. Another interesting compound in this table was 63. This direct analogue of 2 surprisingly had no MPA activity when administered sc. However, 63 was equipotent to 66 when given intraperitoneally (ED5,, 3.2 mg/kg) in the MPA test, and it was also quite potent after ip administration in reducing the isoniazid-induced cyclic GMP elevation and in enhancing [3H]FNP binding. Therefore, absorption after sc administration may be inadequate for certain members of this series, and the SAR results in the standard MPA test reflect a combination of

-

442 Journal of Medicinal Chemistry, 1989, Vol. 32, No. 2

Sarges et al.

Table IV. Cyclic Derivatives

x

yield,

n

65 66 67 68

H 3-C1 4-C1 3-OMe 4-OMe 3-CN 3-CONMe,

Y H H H H H H H

69 70 71 72 73 74 75 76

3-C1, 4-OMe 3-F, 4-OMe 3-C1 3-C1 3-C1 3-C1 3-C1 3-C1

H H H H H H H H

1 Me 1 Me

no. 62 63

64

1 1 1 1 1 1 1

R1 Me Me Me Me Me Me Me

G G

50 86 48 20 58 16 17

yp, C 165-7.5 139-41 174-6 127-9 129-31 172-4 170-2

recrystal solventa formulab CH2C12.IPE CigH18N20 CH2C12eIPE C1gH17ClNZO C1gH17ClN20 CHC13.Et20 CmHzoNzOz CH2C12.IPE CmHzoNzOz CHC13.EtO CmH17N30 CHC13.Et20 C22H23N302'

E E E D2 E E E G

42 12 7 44 38 33 53 29

126-7.5 68-71 162-5 158-61 105-7 128-30 72-5 115-7

EbO Et20.P EA-PE EA-H EA*H EbO Et0.P Et20

method D1 E G G

D1

%

MPA antagonism: ED50 >lo0 >lo0

>56 32-100 3.2-5.6 >lo0 >IO0

[3H]FNP isoniazid bound,d % control f antagonism: S.E. ED50 143 f 7 192 f 19 1-3.2 127 f 9 214 f 9 211 f 7 0.3-7 121 f 6 164 f 15

'/4H20

CzoH19C1N202 CzoHlgFNzOz 2 Me CNHigClNzO 3 H C2oH&lN20 1 Et CzoHlgClN20 1 Ph CuH19ClN20 1 CH2-Ph C25HzlClNzO 1 CHZ-Ph-C1, C25HmC12N20. '/zHzO 77 3-F, 4-OMe H 1 H D2 38 169-71 EA ClgH17FN202 78 H H 3 H D2 46 144-6 C2oHzoNzO PhHqPE 79 H 4-C1 1 Me E 20 78-80 Et20.H CigH17ClN20 80 H 6-C1 1 Me E 12 177-9 EbO C1gH17ClNZO 81 3-C1 6-C1 1 Me G 43 103-6 Et20CH ClgH&12N20 See note a, in Table I. *See note b, in Table I. See note c, in Table I. See note d, in Table I. ip.

3.6 5.6 >IO0 >lo0 >lo0 >lo0 >lo0 >lo0

220 f 7f 225 f 12f 135 f 2 138 f 9 164 f 6 125 f 5 149 f 9f 131 f 4

-56 >lo0 >lo0 190 f 7 >lo0 158 f 8 >lo0 189 f 10 e See note e, in Table I. 'At 100 wmol/kg

Table V. 3-(l-Aminoalkylidene)oxindoles NRlR2

Y

no.

Y

z

R1

a-i;.

R2

method

yield, %

mp, "C

formulaa

H H CH3 CH3 D1 94 193-6b ClZH12N20 204-7 C12H14N20 CH3 D1 87 H CH3 CH3 4-C1 CllHllClN~OJ/8H20 DlC 82 220-3 CH3 H CH3 5-C1 H CH3 CH3 Dld 67 204-6 CllHllC" 6-C1 224-6 C11H11C1N20"/4H2Oe H CH3 CH3 D1 100 7-C1 H CH3 CH3 D If 70 230-2 CllHllC1N20 5-OCH3 H CH3 CH3 Dlf 66 222-5 Ci~Hi4Nz02"/zHzO D1 47 269-72' C13H14N20 H -(CH2)3CH3 D1 50 227-9 C13H&1N20h 6-C1 -(CH2)3CH3 Osee note b, in Table I. bReference 30. CPreparedaccording to the method of Koelsch, C. F. J. Am. Chem. SOC. 1944,66, 2019. dFrom 5-chlorooxindole (Aldrich Chem. Co.). e m / e : 224, 222. f U S . 3,882,236, from the oxindoles via the isatins. gchandramohan, M. R. Znd. J . Chem. 1974, 12, 940; mp 267-8 "C. hm/e: 250, 248. 99 100 101 102 103 104 105 106 107

varying intrinsic activity and varying bioavailability. Conclusion These results suggest that many members of this oxindole series exert GABAergic activity. All the oxindoles of this series which were tested significantly enhance [3H]FN P binding in vivo, with the exception of the extremely insoluble compound 57. Such activity has been reported for GABA agonists and GABAergic agents, e.g. progabide16 and tracazolate?,* but not to the degree observed with the best members in this oxindole series (potency rank order: 69 > 66 > 63 > tracazolate > progabide). Antagonism of isoniazid parallels enhancement of [3H]FNPbinding in this series within the subset of compounds for which both tests

were run. It is not clear whether isoniazid activity is an expression of GABAergic activity or secondary to a benzodiazepine-like effect, especially since benzodiazepines (e.g., diazepam) and tracazolate are also active in this test (Table VI). MPA activity in this series does not always parallel [3H]FNP binding activity or isoniazid activity, probably reflecting poor bioavailability for certain compounds after sc dosing, as we have shown in the case of 63, which gave good MPA activity after ip administration. Nevertheless, several compounds, such as 66, 69, and 70 showed an excellent combination of intrinsic activity, as measured by [3H]FNP binding, and of MPA antagonism after sc administration. Whether or not the GABAergic activity of these oxindoles is related to binding at the

Journal of Medicinal Chemistry, 1989, Vol. 32, No. 2 443

A Novel Class of "GABAergic" Agents

Table VI. Biological Activity of Key Oxindoles and of Reference Compounds

PI

X Y

X

63 CI 66 69

no.

mouse LD,,' mg/kg sc iP >lo00 >320 >320 >lo00 32-100 2320 4.2 3.2-10 >lo00 -1000 >loo0 >loo0

MPA antagonism: ED,, mg/kg sc ip >loo -3.2 3.2-5.6 -3.2 3.6 0.7

H CI

Y H OCH3 OCH,

isoniazid antagonism:

[3H]FNPbound: % control f S.E. at the following dose (pmol/kg ip): 1000

320 192 f 19 211 f 7

100 132 f 7 192 f 9 220 f 7

32

10

ED,, 3.2

mg/kg'ip 1-3.2 0.3-1

63 105 f qNS 105 f qNS 125 f 4 103 f 3NS 105 f gNS 66 159 f 5 128 f 8 115 f gNS 69 124 f 6' 1.5 muscimol 178 f 15 93 f 3NS 89 f gNS loo progabide 212 128 f d Na valproate tracazolate >loo0 >loo 127 f 7 -3 39f2 66f2