,July 196’7
h T I C O X V U L S A X T 4a,8a-?;APHTHALENEDICARBOXIMIDE
DERIVATIVES
GO7
The Anticonvulsant Activity of 4a,8a-Naphthalenedicarboximide and Its Derivatives’ EUGENE R. WAGNER AKD ALLAND. RUDZIK Chewiistry Research and Pharmacology Departments, H u m a n Health Research and Development Center, The Daw Chemacal Company, Zzonsville, Indiana 46077 Received February 16, 1967
1,4,3,8-Tetrahydro-ia,8a-naphthalenedicarboximide (I)has been found t o have anticonvulsant activity against perityletietetrazole-induced convulsions in mice and low toxicity. A structure-act,ivity study was made revealing that li-siibstitution destroyed t h i j activity as did saturation of the double bonds. Increasing t h e degree of ring ittisaturation as it1 4a,8a-naphthalenedicarboximide (VII) produced a compound with half of the activity of I . An hypothesis has been suggested to explain the structure-activity relationship in these compoiuids. This led to the syiithesis of 4,7-dihydro-3a,7a-itidandicarboximide(XIY), which also proved to be active.
earlier by Brigl and Herrmain6 Dicarboximide I Many substituted succiriimides have been studied was prepared from the corresponding anhydride IV by for anticonvulsant activity and several of these are refluxing in aqueous ammonia. The anhydride was clinically useful in the treatment of petit mal epilepsy. formed, along with diacid V and 3,G-dihydrophthalic These compounds manifest theniselves in pharmacological testing by preventing the clonic convulsions induced anhydride (VI),from a Diels-Alder addition of 2 moles of butadiene to 1 mole of acetylenedicarboxylic acid. by administration of pentylenetetrazole to mice. The most active succinimides possess a,a-dialkyl or a-alkyl0 COOH a-phenyl substitution with nitrogen unsubstituted or methylated. Alkyl-substituted compounds show little or no activity in the niaxinial electroshock test, which indicates compounds potentially effective against grand mal epilepsy; phenyl substitution, however, specifically IV V increases activity against electroshock, and a,a-dipheiiylsuccinimides are electroshock active and pentylTo determine a structure-activity relationship for erietetrazole inactive.2 Although some trisubstituted this unique tricyclic dicarboximide system, a number of N-substituted derivatives of I were prepared by succininiides are highly active in the pentylenetetraxole test, few tetrasubstituted derivatives have been t e ~ t e d . ~ treatment of the anhydride IV with the corresponding Surprisingly, however, a,a,/3,P-tetraniethylsuccinimide amine (see Table I). The N-methyldicarboximide showed markedly reduced activity, but it and the Shas been shown to be a pentylenetetrazole-like convulsant . 4 benzyl compound were still weakly active. However, We have found that 1,4,5,S-tetrahydro-4a,Sa-naph-a variety of other substituents destroyed the activity thalenedicarboximide (I), which is, in effect, a tetraaltogether (see Table 11). substituted succinimide, displayed significant activity Since N-substitution did not appear fruitful, attempts against pentylenetetrazole convulsions. It has an mere made to modify the carbocyclic system. The most obvious change involved saturation of the double ED50 of 22 mg/kg and relatively low toxicity (LDSo = bonds. The monounsaturated derivative, 1,4,5,6.7,s620 mg/kg) and thus possessed a high therapeutic ratio. Typical of alkyl-substituted succinimides, it was inhexahydro-4a,Sa-naphthalenedicarbo~imide(11), was artive against maximal electroshock. prepared by addition of butadiene to 3,4,5,6-tetrahydrophthalic anhydride. The resulting anhydride 0 Ri was converted to the diacid for isolation and treated 0 with urea in ethylene glycol at 200” to produce the imide 11. This proved to be corisiderably less active thaii I ; methylation with diazomethane formed the Smethyl derivative which was also inactive. Catalytic 0 0 0 hydrogenation of I1 as described by Brigl and HerrI, R = R i = R z = H I1 I11 IX, R = H ; R 1 = C H 3 ; R z = H mann6 produced I11 which was void of all biological X, R = R i = H ; RZ=CH3 activity in the pharmacodynamic screen. The next attempt was to move in the opposite direcThe first preparation of I was reported by Snat,zke tion, i.e., to introduce more double bonds into the rings. and Zanati,j although the corresponding hexahydro The goal was to prepare 4a,Sa-naphthalenedicararid octahydro derivatives I1 and I11 had been prepared boximide (VII), the parent compound. Although the
& @q$
(1) Presented before the Division of Medicinal Chemistry a t t h e 153rd National Meeting of the American Chemical Society, Miami Beach, Fla., April 196i. (2) A . Spinks and JT. S. Waring, Progr. M e d . Chem., 3, 261 (1963). (3) IV. J. Close and M. A. Spielman in “Medicinal Chemistry,” Val. V, iV. H. Hartunp, Ed., John Wiley and Sons, Inc., New York, N . T.,1961, Chapter 1. (4) G . Clien and B. Bohner, J . Pkarmucol. Ezptl. T h e m p . , 123, 212 (1958). ( 5 ) G . Snatzke and G . Zanati, Ann., 684, 62 (1966).
q@qp Q
(6) P. Brigl a n d
!!
0
0
VI1
VI11
R. Herrmann, Ber., 71B,2280 (1938)
1.0,
Ilexohart~itnlsleep time i r i miiiiites; experimental ,iiiimnle/coiiti.r)ls. ', A1:ixini:tl vlecLtloshock test : aiiimals prott?c:tc~d/:iiiiiii:il-; Sti,ychniiie lethality test : animals proterted/aiiimals tested. (1 I'eiit?.lcric.tetI.:izi)lr te4 : animals protecl.ed/aiiim:ils te-tcatl. tested. r Iiecerit work has shown t h a t this ciimpoiuid was activr agairiAt o s o t r e m o r i u e - i n ~ i ~trrmcir> ~ ~ e ~ l iii mice: E I h = 27.8 tng/kg. I T t r w ever', it was also fairly toxic: I,I)x = 121 mg/kg. '8
c*oriesponding nnhydritlt. VI11 hnd been obtained previousl). hy T'ogel, cf al.,7 t h c tlirarhoximidc XIS 1111liiroirii. Following F'ogel's procedure, I-TII p:ircd from IV in reasoiiablc yield, but it could Iiot h t h voiivertcd to the dicurboximide with aqueous ninnioriia :IS in ttic c:isc of IT. Thc :iiihytlride VI11 dissolvd i n ttic :~niinoni:tslowly :it room tcinpcrnturc hut :ic.iclific*:xl i ) I. Vrigel, \\
l l e c h r l , nnil \\
G r i r n m p , . I ~ ~ P I ('iirni ,.
76, 780 ( I ' l b I l
ASTICONVULSANT 4a,8a-SAPHTHALEh-EDICARBOXIMIDEDERIVATIVES
July 1967
raphy on silicic acid. It showed typical dicarboximide absorption bands in the infrared and its nmr spectrum displayed only vinyl proton peaks centered at 5.8 ppm similar to those reported by Vogel, et a l l 9and by Bloonifield and Smiley Irelan* for 4a,Sa-bridged naphthalenes. Biologically VI1 proved to be active against pentylenetetrazole, with an EDboof 42 mg/kg, or half as active as its precursor I. The S-methyl derivative, prepared with diazomethane, was not active. Two C-methyl-substituted derivatives of I were prepared by treating 3,G-dihydrophthalic anhydride (VI) with piperylene and isoprene. The resulting anhydrides were converted to the dicarboximides with ammonia. These products, 1-methyl- (IX) and 2-niethx 11,4,5,S-tetrahydro-4a1Sa naphthalenedicarboximide (X), were only weakly active against pentylenetetrazole. Epoxidation of one of the double bonds in I or of the single double bond in I1 with nz-chloroperbenzoic acid produced the epoxides X I and XI1 in which almost all of the activity against pentylenetetrazole convulsions was absent. A study of models indicated that a possible reason for this structure-activity relationship might involve steric hindrance around the dicarboximide group. Those compounds which showed very little activity also sterically blocked one or both sides of the heterocyclic ring. The t etrahydr0-4a,Sa-naphthalenedicarboximide I probably exists in the symmetrical exo-ezo configuration XIII, as indicated by the nmr splitting patterns of the methylene protons and as suggested by the work of Iiallos and Deslongchampslo on the configuration of the adduct of furan and acetylenedicarboxylica acid. I n this form, the four vinyl protons lie on a plane at the base of and perpendicular to the plane of the heterocyclic ring. I n the case of 11, 111,XI, and X I I , all of which were inactive, protons lie above this
609
ene prepared XVI. The cyano ester was cyclized to dicarboximide XIV by the method of Homing and Schock.12
dooR butadiene
CN
xv
CN
XVI
+ 0
XIV
As predicted. XIV was active against pentylenetetraeole convulsions with an ED50of 46 mg/kg. Also as expected, the S-methyl derivative was of much lower activity. Although it was useful in predicting activity of XIV, the hypothesis did not explain the lack of activity in a large number of other unhindered succinimides, nor the activity found in some trisubstituted conipounds which must be hindered at least as much as some of these.13 I n addition, IX and X were less active than one would have predicted since the steric hindrance of the methyl groups appears to he slight.'* Unfortunately none of the modifications of the 4a,Sanaphthalenedicarboximides could match the anticonvulsant activity of the tetrahydro derivative I. Compound I was also effective in protecting mice from a lethal dose of strychnine sulfate. The ED50 value of I against strychnine was found to be SG (79-94)mg/kg ip. Kone of the compounds in this series protected against the seizures induced by maximal electroshock, but several appeared to possess central nervous system depressant properties. Compound VI1 and the Sphenyl and S-cyclopropyl derivatives of I were found to potentiate the sleeping time induced by hexobarbital (a greater than twofold increase in sleeping time mas considered to be significant).
Experimental Section15 1,4,5,8-Tetrahydro-4a,8a-naphthalenedicarboxylicanhydride
(IV)was prepared by a modification of the procedure described XI1
XI
plane blocking one or both sides of the center ring. However, in I and VI1 the protons lie below this plane and the compounds were active.
H
XI11 This hypothesis led to the preparation of the corresponding unknown indandicarboxiniide XIV. By the steric hindrance argument XIV should be active since the five-membered ring would be completely below the hypothetical "plane." Activity in this compound mould also indicate that the presence of two double bonds was not required. Following literature procedures,'l 2-cyanocyclopentene-1-carboxylicester (XV) was obtained in good yield, and its adduct with butadi(9) E. Vogel, W.Maier, and J. Eimer, Thad., 655 (1966). (10) J. ICallos and P. Deslongchamps, Can. J . Chem , 44, 1239 (1966).
by Alder and Backendorf.16 A mixture of 250 g of acetylenedicarboxylic acid (2.19 moles), 750 ml of dioxane, and 550 ml of liquid butadiene was heated in a bomb at 170' for 20 hr. The reaction was cooled and the dioxane was removed in vacuo. The resulting brown oil, which weighed 489 g, crystallized on standing. Instead of distilling the residue, the prodnct was suspended in 2 1. of CCl, and the insoluble dicarboxylic acid T' was isolated by filtration. Some viscous polymer separated on dilution of the filtrate to 4 1. with CCL. A mixture of anhydrides IT and TI was obtained by concentration of the CCla solution and these ( l l ) ( a ) S. C . Sen-Gupta and A. J. Rhattacharyya, J . Indian Chem. Soc.. 17,183 (1940); (b) B. R . Baker, M.V. Querry, S. Bernstein, S. R. Safir, and Y. Subbarow, J . Ow.Chem., 12, 167 (1947); (8) G. Bialino, Farmaco (Pavia), Ed. Sei., 17,377 (1962); Chem. Ahstr., 58, 56848 (1963). (12) E. C. Hornina and R. U. Schock, Jr., J . A m . Chem. Soc., 70, 2945 (1948). (13) G. Chen, R. Portman, C . Ensor, and A . C. Bratton, ,Jr., J . Pharmacol. Ezptl. Therap., 103,54 (1951). (14) These unsymmetrical dicarboximides must occur as iacemic mixtures and it could he t h a t only one of the optical isomers is active. Howerer, stereo- or geometrical configuration may not be tao important since C. A . Miller, H . I. Schell, and L. 31. Long ( J . A m . Chem. Soc., 73,5608 (1951)) found no differences in t h e activity between cis and trans isomers in some substituted succinimides they studied. (15) All microanalyses mere performed by Midwest Microlab.. Inc.. Indianapolis, Snd. All melting points mere obtained on a Thomas-Hoover Unirnelt apparatus and are uncorrected. Infrared spectra were obtained on a Perkin-Elmer grating infrared spectrophotometer Model 33i. Nmr spectra were obtained using a Varian A-60 nmr spectrometer. (16) K. Alder and K. H . Haokenduxf, Ber.. 71B, 2199 (1938).
,July 1967
AXTICONVULSANT ARYLDIALKYLSUCCINIMIDEH
4,7-Dihydro-3a,7a-indandiearboximide (XIV).-A mixture of 100 g of 2-cyanocyclopentene-1-carboxylateand 250 ml of butadiene was heated a t liso in a bomb for 22 hr. On cooling, t h e contents of the bomb were poured into 8 1. of acetone. T h e cloud>- solution was filtered and taken to dryness in evmto t o leave 130 g of dark oil. Thih was poured into 2 1. of 95% ethanol and filtered again through Super-Cel. T h e clear yellow solution was taken to dryness arid the resulting oil distilled in vacuo. The first fraction weighing 59 g boiled over a range from belou100 to 113' ( 0 . 7 mm). The next fractioii weighing 15 g boiled from 11.; to 12.5' (0.6 m m ) mid contriiied the desired nitrile ester as indicated by iiifrared atid nnir anal This Iiitrile ester was cyclized by the procediire of Horning and Srhork.11 The mixtiire of 5 g of cyano ester, 25 ml of glacial acetic w i d , and 2.7 ml of roncentrated TTCl was heated on the steam b a t h for 2 hr. The rewltiiig soliitioii was cooled and poiired into 2-50 ml of water, and the oil which separated was extracted three times n-ith .jO-ml portions of CHC1,. The extracts were
