J. Med. Chem. 1995,38,4367-4379
4367
Articles Synthesis and Structure-Activity Relationships of Substituted 1,4-Dihydroquinoxaline-2,3-diones: Antagonists of N-Methyl-D-aspartate (NMDA) Receptor Glycine Sites and Non-”DA Glutamate Receptors John F. W. Keana,*lt Sunil M. Kher,t Sui Xiong Cai,$ Christian M. Dinsmore,?Anne G. Glenn,+J. Guastella,’ Jin-Cheng Huang,$ Victor Ilyin,$JYixin Lu,t Pamela L. Mouser,t Richard M. Woodward,$ and Eckard Weber**$ Department of Chemistry, University of Oregon, Eugene, Oregon 97403, Acea Pharmaceuticals Inc., A Subsidiary of CoCensys, Inc., 213 Technology Drive, Irvine, California 92718, and Department of Pharmacology, University of California, Irvine, California 9271 7 Received May 15, 1995@
A series of mono-, di-, tri-, and tetrasubstituted 1,4-dihydroquinoxaline-2,3-diones ( a s ) were synthesized and evaluated as antagonists at N-methyl-D-aspartate (NMDAYglycine sites and a-amino-3-hydroxy-5-methylisoxazole-4-propionic acid-preferring non-NMDA receptors. Antagonist potencies were measured by electrical assays in Xenopus oocytes expressing rat whole brain poly(A)+RNA. Trisubstituted a s 17a (ACEA 10211, 1% (ACEA 1031), 24a, and 27, containing a nitro group in the 5 position and halogen in the 6 and 7 positions, displayed high potency (Kb 6-8 at the glycine site, moderate potency at non-NMDA receptors (Kb = 0.9-1.5 pM), and the highest (120-250-fold) selectivity in favor of glycine site antagonism over non-NMDA receptors. Tetrasubstituted &xs 17d,ewere more than 100-fold weaker glycine site antagonists than the corresponding trisubstituted QXs with F being better tolerated than C1 as a substituent at the 8 position. Di- and monosubstituted &xs showed progressively weaker antagonism compared to trisubstituted analogues. For example, removal of the 5-nitro group of 17a results in a -100-fold decrease in potency (10a,b,z),while removal of both halogens from 17a results in a -3000-fold decrease in potency (1Ov). In terms of steady-state inhibition, most QX substitution patterns favor antagonism at NMDNglycine sites over antagonism a t non-NMDA receptors. Among the QXs tested, only 17i was slightly selective for non-NMDA receptors.
-
Introduction N-Methyl-D-aspartate (NMDA) receptors are implicated in the pathology of numerous neurodegenerative disorders, including the acute brain damage which follows ischemic stroke and the more gradual loss of neurons associated with epilepsy, Alzheimer’s disease, and acquired immune deficiency syndrome (AIDS)related dementia.l In each case, neuronal damage is thought to be a consequence of the “excitotoxic” effects of glutamate, wherein excessive excitatory input causes pathological increases in intracellular Ca2+ and ultimately cell death.2 In terms of therapeutic intervention, there are at least four sites for antagonism of NMDA receptor^:^ (i) phencyclidine (PCP)-binding sites, located within the channel lumen and accessible in open-channel configur a t i o n ~ (ii) , ~ glutamate-binding sites, where antagonists compete with glutamate to inhibit channel activity,5 (iii) glycine coagonist sites, which must be occupied by glycine for glutamate to gate the channe1,‘j and (iv) polyamine inhibitory sites.7 Numerous studies have shown that PCP site ligands such as dizocilpine (MK-
* Authors to whom correspondence should be addressed. +
University of Oregon.
* Acea Pharmaceuticals Inc.
University of California. Permanent address: Institute of Cell Biophysics,Russian Academy of Sciences, Pushchino, Moscow Reg. 142292,Russia. Abstract published in Advance ACS Abstracts, October 1, 1995. 5
‘I
@
801)and glutamate site antagonists such as CGS 19755, LY 274614, and [3-((&)-2-carboxypiperazin-4-yl)prop-lyllphosphonic acid (CPP) have neuroprotective actions in animal models of stroke.* Unfortunately, the clinical potential of these classes of antagonists can be compromised by psychotomimetic side effe~ts.~JO For reasons that remain uncertain, the behavioral side effect profiles of glycine site antagonists are more encouraging.10-12 In addition, glycine site antagonists do not appear to cause neuronal vacuolization, a pathological phenomenon observed following treatment with potent PCP site ligands and competitive antag0ni~ts.l~ A wide variety of glycine site antagonists are now Among the more potent series are (i) the kynurenic acids, which include 7-chloro-5-iodo-kynurenic acid (IC50 in binding studies, -30 nM) (1),15 (ii) the 2-~arboxytetrahydroquinolines, which include (&)-trans2-c~bo~y-5,7-dichloro-4-(phenylureido)tetrahydroquinoline (IC50 8 nM) (2),16(iii) the tricyclic 1,4-dihydroquinoxaline-2,3-diones (QXs), which include (S)-g-bromo5-[(phenylcarbamoyl)methyll-6,7-dihydro-lH,5Hpyrido[l,2,3-delQX (Ki = 0.96 nM) (3),17 (iv) the benzazepines, which include 8-methyl-2,5-dihydro-2,5dioxo-3-hydroxy-lH-benzazepine(Kb 470 nM) (4),18 lHI-ones, and (v) the 3-substituted-4-hydroxyquinolin-2( which include 3-(3-phenoxyphenyl)-4-hydroxyquinolin2(Vn-one (IC50 2 nM) Blood-brain barrier penetration, one indication of in vivo bioavailability, varies drastically both between and
-
-
-
0022-2623/95/1838-4367$09.00/0 0 1995 American Chemical Society
4368 Journal of Medicinal Chemistry, 1995, Vol. 38, No. 22
Keana et al.
available, they were prepared from the requisite anilines 7 employing routine protection-deprotection sequences (Scheme 1). QXs 1Ox,y are known27and were prepared by chlorination and bromination, respectively, of commercially 1 2 available QX 6. 1Oz was prepared by chlorination of 1Oy (Scheme 2). A third approach (Scheme 3) for the synthesis of QXs (YCH2CoNHPh Q utilized selenium heterocycles to produce the 5-nitro QXs 16. Regiospecific nitration of selenium heterocycles of type 13b is known to furnish 14b.28Accordingly, the nitroaniline l l b was reduced to the corresponding H diamine 12b, which was then allowed to react with 3 4 selenium oxide in aqueous ethanol, giving (2,1,3)benzoselenadiazole 13b. Treating 13b with a mixture of nitric acid and sulfuric acid gave 14b, which was then converted to the corresponding 1,2-diamine 15b with hydriodic acid. Finally, 15b was condensed with oxalic acid to give 16b. A similar reaction sequence was followed using l l a , c to give 16a,c, respectively. Bro6, X = Y = H 5 minationZ9of 16b gave 27 (Table 1). All the compounds 6a, X = Y = N O z described in Scheme 3 were assigned the structures 6b, X = CN, Y = NO2 shown based on the lH NMR spin-spin splitting pattern and coupling constants observed for their aromatic within the different classes of antagonists. For example, protons. kynurenic acids and 2-carboxytetrahydroquinolinesare Nitration (KNo3,HzS04 or mo3, TFA) of QXs 10alargely inactive in rodents following systemic adminish, 6, and 10n,x furnished 17a-h,i,n,x, respectively tration, whereas some 3-phenyl-4-hydroxyquinolin- (Scheme 4). KNOOFA was the reagent combination 2(1H)-ones have anticonvulsant effects a t oral doses as of choice when 10s were substituted with two or more low as 1 mg/kg.lg fluorine atoms on the aromatic ring. 17x was further QXs 6a,b were initially reported as selective antagonitrated (KN03,HzS04) to 18x. Nitro QXs 17a-c were nists a t a-amino-3-hydroxy-5-methylisoxazole-4-propi- reduced with tin(I1) chloride dihydrate to the correonic acid (AMPA)-preferring non-NMDA receptors.20 sponding amino QXs 18a-c. Further, 18a was acetylSubsequent characterization revealed that the pharmaated with acetyl chloride to give 19. cological profile of these compounds also includes modNitration of unsymmetrically 6,7-disubstituted QXs erately potent antagonism a t NMDA receptor glycine was not satisfactory in many instances using convensites.21 Herein, we report the chemistry and pharmational nitration procedures described in Scheme 4. cology of a series of QXs with novel substitution patterns While nitration of 1Of,g gave the single nitro isomers, on the benzene ring. The steady-state inhibitory potenother nitrations produced mixtures of two isomers that cies of these molecules were measured a t rat brain were difficult to separate. Hence it was necessary to NMDA and non-NMDA receptors expressed in Xenopus develop a regiospecific procedure to prepare 5-nitro-6,7oocytes. Several trisubstituted compounds were found disubstituted QXs 24a-d (Scheme 5). to be potent and selective glycine site antagonists. One The approach is based on the observation that treatof these is ACEA 1021 (17a), a neuroprotective drug ment of 3,4-dihydroquinoxalin-2(lH)-ones23 with excurrently undergoing clinical trials for stroke. A precess fuming nitric acid (10-20 equiv) in TFA results liminary report describing some of these results has both in nitration exclusively a t the 5 position and in appeared previously.22 oxidation to the QXs 24.30Accordingly, fluorobenzenes 20 were converted into N-phenyl glycinates 22. ReducChemistry tion of 22 and concomitant cyclization gave quinoxalinSynthesis. Table 1 lists the compounds that have 2(1H)-ones 23. Finally, treatment of 23 with excess been synthesized together with supporting data. QX 6 HN03 in TFA gave analytically pure QXs 24. fuming was commercially available, while QXs 10 (Scheme 1) QXs 26a,b (Scheme 6) were prepared by reduction of were prepared by condensation of diethyl oxalate23or, the nitro QXs 17c’,n to amino QXs 25a,b respectively, better, oxalic acidz4 with the corresponding 1,2-difollowed by diazotization and subsequent treatment aminobenzenes 9. When 9s were not commercially with copper(1) chloride in hydrochloric acid. available, they were prepared from 8 by one of several Structural Considerations. Structures of the synmethods. Treatment of the appropriate 2-nitroanilines thetic intermediates and tested ligands were assigned 8 with tin(I1) chloride dihydrate in refluxing ethanol or by applying well-established ortho-, para-, or metaethyl acetate furnished 9,25as did agitating solutions directing effects of substituents already present on the of 8 in methanol with P d C under hydrogen gas. A aromatic ring and are consistent with lH NMR spectra different procedure was used in the case of the symand elemental composition by HRMS. In the case of metrically substituted 2,6-dinitroaniline 8u. Selective the F-containing QXs, structures were confirmed by the reduction to 9u was achieved by treatment with a magnitude of the F-H coupling constant ( J ) in the ‘H refluxing solution of ammonium sulfide in ethanol and NMR spectra.31 Structures of the corresponding C1 o r water.26 When nitroanilines 8 were not commercially
Substituted 1,4-Dihydroquinoxaline-2,3-diones
Journal of Medicinal Chemistry, 1995, Vol. 38, No. 22 4369
Table 1. Physical Data, NMR Spectroscopic Data, and Methods of Preparation for Substituted QXs
compd no. 6 10a lob 1oc 10d 10e 1Of log 10h 1Oi lOj
R5 H H H C1 C1 F H H H Br Br
R6 R7 Rg H H H C1 C1 H Br Br H H CF3 H C1 C1 H C 1 F H Br CF3 H C1 CFB H F F H C1 C1 H H Br H
mp("C) >300 >400d 335e 346-348 355 >250 >360 > 360 >360 332-335 356-358
starting material; yield methods" (%I comm available 9a; H 74 49 7b; B, E, H 36 8c; G 23 8d; D, G 62 7e; B, D, H 46 7f; A, D, H 13 7g; A, D,H 89 8h; E, H 24 8i; D, G 16 8j; D, G
formulab
lH NMRc
7.23 (2H), 12.02 (2H) 7.37 (2H), 11.96 (2H) 7.30, 7.53, 11.9 (b, 2H) 7.25, 11.62, 12.12 6.93 (d. J = 9.6). 12.14. 12.17 (b) 7.54, 7:74, 12.05, 12.30 7.27, 7.48, 12.14 (2H) 7.05 (d, J = 9.31, 11.94 (2H) 7.30, 11.19 (b), 12.14 (b) 7.21 (d, J = 2.11, 7.53 (d,J = 2.11, 11.1(b), 12.1 (b) 7.28, 7.56, 11.7 (b, 2H) 62 10k 332-334 9k;G Br H CF3 H 7.51 (d, J = 1.81, 7.56 (d,J = 1.51, 22 101 307-309 71; A, D, H CF3 H Br H 11.34, 12.22 7.65, 7.69, 11.58, 12.31 12 CioH4FsN202f H CF3 H 308-3109 7m; A, D, H 10m CF3 7.37 (d, J = 1.81, 7.47 (d, J = 1.8), 31 CgH4ClF3N20J 10n CF3 H C1 H 302-304 7n; A, D, H 11.35, 12.22 7.18 (dd, J = 8.7, 2.71, 7.36 (d, J = 8.7, 9 CgH4F4N202f CF3 H F H 300-302 70; A, D, H 100 2.7), 11.27, 12.24 7.05(d, J = 1.8),7.32(d,J = 1.8), 41 326-328h 9p; D, G C1 H C1 H 1OP 11.5 (b), 12.1 (b) 46 C~HzF4Nn02f 12.33 (b. 2H) F F F F 330-331' 8q;D, G 1oq 9 C ; H ~ F ~ N ~ O ~ 6.91(mi, 12.02, 12.19 >360 7r; B, D, H 10r F F F H 7.08 (d, J = 11.4), 7.38 (d,J = 5.1), 20 CgH4F4N202f 10s H F CF3 H 330-333 79; A, D, H 12.02, 12.23 52 CgH4ClN304.0.5HzO 7.40, 7.90, 11.19, 12.34 315-317j 8t; F, H lot NO2 H C1 H 6.93(d,J=10.2),7.08(d,J=7.2) 45 CgH4ClFN202f H C 1 F H 344-348 9u;G 1ou 7.26(t,J=8.1),7.46(d,J=7.8),7.87 77 279-283k 9v; H 1ov NO2 H H H (dd, J = 8.4, 0.9), 11.69, 12.31 7.05 (dd, J = 10.8, 7.2), 11.6 (b, 2H) 8 322-326 8h; P, D, G low C l F F H 7.07 (m, 3H), 11.94 (2H) 29 H C 1 H H 1OX >360m 6; P 7.04 (d, J = 91, 7.24 (m, 2H), 11.96 53 H B r H H >35On 6; P 1oY (b, 2H) 7.24, 7.37, 11.8, 12.2 42 102 H Br C1 H >360 1oy; P 7.64,8.11, 11.41, 12.43 22 329-332" l l a ; I CF3 H 16a NO2 H 7.22 (d, J = 9.01, 7.36 (d, J = 8.71, 23 H llb;I 16b 321 NO2 C1 H 12.14, 12.22 7.15(d,J=8.7),7.47(d,J=8.7),12.12 36 >350 llc; I 16c NO2 Br H H (b), 12.22 7.38,12.28, 12.37 (b) 90 342-344 loa; J 17a NO2 C1 C1 H 7.48, 12.22, 12.26 77 352-354 lob: J 17b NO2 Br Br H 7.86, 12,.03,12.43 (b) 14 loci J 17c H CF3 NO2 305 7.48, 12.21, 12.42 1Oc; J 70 17c' 345 NO2 CF3 H 11.9 (b), 12.5 (b) 10d; J 65 C1 C1 C1 324 17d 12.10 (b), 12.50 7 C1 F 310 10e; K F 17e 74 7.61, 12.34, 12.60 (b) 17f 333-335 lOf;J Br CFB H 80 7.62. 12.35 342-345 log; J C1 CF; H 17g 7.33; 11.86, 12.25 68 10h; K F H 288-290 F 17h 78 H >300 6; J 7.21 (d. J = 8.7). 7.92 (m. 2H). 12.14 17i NO2 H (b, lH), 12.33'(b, 1H) 91 7.53, 11.8 (b), 12.46 10n;J 278-280 17n NO2 C1 H 37 7.25, 7.8, 12.18, 12.30 lox; J >370 17x C1 NO2 H 5.94 (2H),6.60, 11.32, 11.87 62 17a; L >360 c1 c1 H 18a 5.84 (2H),6.73, 11.26, 11.81 59 17 b; L 324-326 Br Br H 18b 5.92 (2H), 7.24, 11.32, 11.44 93 17c;L 18c H CF3 NH2 >360 35 7.75, 10.82, 11.53 17x; J 298-300 18x NO2 C1 H 2.06 (3H). 7.24,9.62, 11.66. 12.10 88 18a; M 320-322 c1 c1 H 19 7.35, 12.25 (2H) ' 26 Br C1 H 338-343 20a; N 24a 11 7.18 (d, J = 9.01, 12.21, 12.32 Br F 24b H 323-327 20 b; N 7.46 (d, J = 6.31, 12.01 (b), 12.18 5 F Br H 24c 316-321 20c; N 7.37 (d, J = 6.91, 12.02 (b), 12.22 8 24d F C1 H 308-310 20d; N 7.49, 11.85, 12.22 52 26a 2250 17c'; L, 0 C1 CF3 H 7.52, 11.05 (b), 12.27 17n; L, 0 69 26b c1 c1 H 305-307 7.48, 12.22, 12.32 (b) 16b: P 44 27 C1 Br H 2350 a For details of synthesis, refer to experimentals. Analyses for C, H, and N are within f0.4% of the theoretical values, unless otherwise noted. Unless otherwise noted, values listed are for one-proton singlets. Abbreviations: b = broad singlet, d = doublet, dd = double 0 e Lit.23mp ~ 3 6 "C. 0 f HRMS f doublet, m = multiplet, and t = triplet. Splitting values are reported in hertz (Hz). Lit.23mp ~ 3 6 "C. 0.002, purity by HPLC, >95.0%. 8 Lit.44mp 333 "C. Lit.45mp 320 "C. Lit.46mp 300 "C. j Lit.44mp 320-323 "C. Lit.44mp 284-285 "C. Contaminated by 11% dechloro QX. See ref 27. " Lit.44mp 342-344 "C. Contaminated by 5% debromo QX. p See ref 23. I
,
4370 Journal of Medicinal Chemistry, 1995, Vol. 38, No. 22
Keana et al.
Scheme 1"
fkNH2 R5
R7
R8
X
aorborc
E 78,, XX = HNO2
doreorf
9,X=NH2
c
H H CI CI F H f, H H Br Br Br 1, CF3 m. H n, CF3 0, CF3 D. . . CI g , F r , F S, H t, NO2 u, H V, NO2 w , CI a, b, C, d, e,
=
1
10
R6
R7
Re
CI Br H CI CI Br CI F CI H H H CFI H H H F F F H CI H F
CI Br c F3 CI
H H H H H H H H H H H H CFI H H H F H H H H H H
c F3
Br H CI F CI F F CF3 CI F H F
Reagents: (a) i. AcpO, ii. HN03, HzS04, iii. HC1; (b) i. AcnO, CHC13, ii. KNO3, HzSO4; (c) i. (CF3CO)pO, ii. HN03, HzS04, iii. KzC03, MeOH; (d) SnCl2.2H20, EtOH or EtOAc; (e) PcUC, EtOH; (0 (NH&3, aqueous EtOH; (g) (COOEt12; (h) (COOH)2*2HzO,2 N HC1; (i) same as 10m, Table 1.
Scheme 2"
R6@Nxo R5
aorb
~
H
N
R7
O
4 1l a , W = NOz; Y = CF3 llb,W=NOp;Y=CI, l l c , W = NO2; Y = Br
13a, Y = CF3; X = 2 = H 13b, Y = CI; X = 2 = H 13c, Y = Br; X = 2 = H
l z a , W = NH2; Y = CF3 1 Zb, W = NH2; Y = CI 12c, W = NHz; Y = Br
14a, X = H; Y = CF3; 2 = NOz 14b, X = NO2; Y = CI; Z = H 14c, X = NOz;Y = Br; 2 = H
J
4
X
H
16a, X = H; Y = CF3; 2 = NO2 16b, X = N02; Y = CI; 2 = H 16c, X = NO2; Y = Br: 2 =H
15a, X = H; Y = CF3; 2 = NOz 15b, X 5 NO2: Y = CI; 2 = H 15c, X = NOz; Y = Br; 2 = H
Reagents: (a)SnCl2.2H20 in EtOH; (b) SeO2, aqueous EtOH; ( c ) HN03, HzS04; (d) 48% HI; (e) (COOH)y2H20,2 N HC1.
Scheme 4" lOa-lOh, 6, Ion, l o x
aorb
H
A,
N H
"@Nx:
R7
RB
55
17 R5 R6 18a, NH2 CI 18b, NHz Br lac, CI H
R7
i8] -
CI Br CF3 NHZ
C
R6
R7
CI CI Br Br H CF3 NO2 CF3 CI CI F CI Br CF3 CI CF3 F H H CI CF3 H CI
Re H H NO2 H CI F H
R5 R6 R7 Br compounds were assigned by analogy to the fluorine 19, NHAc CI CI derivatives. H The structures of 6,7-dichloro-1,4-dihydro-5-nitroquiH noxaline-2,3-dione (17a),7-chloro-5-(trifluoromethyl)H R5 R6 R7 R8 H 1,4-dihydro-6-nitroquinoxaline-2,3-dione (17111,and a lax, NO2 NO2 CI H cH 5-chloro-7-(trifluoromethyl)-l,4-dihydro-6-nitroquinoxaReagents: (a) m03, H2S04; (b) HN03, TFA; (c) SnClp.2HpO line-2,3-dione (17c') were established by single-crystal in EtOH or EtOAc; (d) AcC1. X-ray crystallographic determinations (see supporting information). The X-ray studies indicated that in all AMPA-preferring subtypes. The AMPA receptors were three QXs, the nitro group is twisted out of the plane of activated by application of kainic acid, which elicits nonthe aromatic ring. desensitizing responses and hence larger steady-state currents than AMPA. The membrane currents were Biology inward at -70 mV and had a monophasic time course Potencies of QXs a t mammalian NMDA and AMPA(not i l l ~ s t r a t e d ) . ~ ~ - ~ ~ preferring non-NMDA receptors were assessed by elecApparent agonist affinities (ECjos) were estimated trical assays in Xenopus oocytes expressing rat whole from concentration-response curves (Figure 1). For brain poly(A)+RNA. NMDA receptors were selectively NMDA receptor ligands, glycine receptor affinity was activated by coapplication of NMDA and glycine.6 measured at a fixed concentration of 100 pM NMDA, Membrane current responses elicited by NMDMglycine and NMDA receptor affinity was measured using 10 pM were inward a t a holding potential of -70 mV, showing glycine. Current ranges and mean maximum responses an initial spike of current and a subsequent more slowly in concentration-response experiments for glycine, developing peak. For all pharmacological assays the NMDA, and kainic acid were 367-740 nA (527 f 59 spike of current, due to secondary activation of Ca2+gated C1- channels,32was ignored and response amplinA, n = 6), 420-790 nA (569 f 99 nA, n = 41, and 10402925 nA (1932 f 668 nA,n = 31, respectively. Levels tudes were measured a t the peak of the second phase of receptor expression were similar for assays of an(e.g., Figure 1, arrow). Non-NMDA receptors expressed in oocytes by rat brain mRNA are predominantly tagonist potency.
Journal of Medicinal Chemistry, 1995, Vol. 38,No. 22 4371
Substituted 1,4-Dihydroquinoxaline-2,3-diones
Scheme 5" R6
a, Br b, Br
c, F 20, X = F; Y = H
a&
d, F
R7
CI F
Br CI
-fJ
21,X=F;Y=NOp
I C
t
No2 H
H ..
R6n;aod
R7
2 min
R7 R6k(Nxo N O
H
H
23
24
\
Reagents: (a) mo3, HzS04; (b) NHzCHzCOO-Na+, DMF, HzO; (c) SnCly2Hz0, EtOH; (d) fuming HN03, TFA.
Scheme 6u 17C', 17n
a
H z N & R5 N ~ oH R7
0
' N
- ci&Nxo R7
0.6 0.4
H
R5
b
-- 0.8
0
' N
H
H
0.2 0.01
0.1
R5
a, CI
100
[antagonist]rM
26
25
10
1
R7
CF3 b, CF3 CI
a Reagents: (a) SnCly2Hz0, EtOH; (b) concentrated HC1, NaNOZ, CuC1.
IC50 values of QXs at NMDA receptor glycine sites and non-NMDA receptors were estimated from partial (three- or four-point) concentration-inhibition curves (e.g., Figure 2). Oocytes were pretreated with antagonist for -30 s prior to receptor activation, and inhibition was fully washed out following 1-5 min of wash. Antagonist concentrations were selected such that levels of inhibition ranged between -20% and 80%, i.e., within the pseudolinear portion of semilog plots and spanning the IC50 value. Fixed agonist concentrations were 1pM glycine (-80% saturating)/100 pM NMDA for NMDA receptors and 20 pM kainate (-10% saturating) for nonNMDA receptors. For the majority of compounds, slope values for concentration-inhibition curves ranged between -1.3 and -0.9. Compounds showing potentially atypical slopes were 17x, 19, lOm, and IBx, slope >-0.9 in NMDA assays, 6 and 16b, slope >-0.9in non-NMDA assays, lOi,d,r, 16c, and 26b, slope
