Identification and characterization of a potential ischemia-selective N-methyl-D-aspartate (NMDA) receptor ion-channel blocker, CNS 5788

Article abstract of DOI:10.1016/S0960-894X(00)00695-8

The identification and characterization of a potentially ischemia-selective and orally-active sulfoxide based NMDA ion-channel blocker showing good neuroprotective activity, (R)-(+)-N-(2-chloro-5-methylthiophenyl)-N′-(3-methylsulfinylphenyl)- N′-methylguanidine (CNS 5788), is described.

Full text of DOI:10.1016/S0960-894X(00)00695-8

Bioorganic & Medicinal Chemistry Letters 11 (2001) 501±504
Identi®cation and Characterization of a Potential
Ischemia-Selective N-Methyl-D-aspartate (NMDA)
Receptor Ion-Channel Blocker, CNS 5788y
Seetharamaiyer Padmanabhan,* Michael E. Perlman, Lu Zhang, Deke Moore, Dan Zhou,
James B. Fischer, Graham J. Durant and Robert N. McBurney
Cambridge NeuroScience, Inc., 333 Providence Highway, Norwood, MA 02602, USA
Received 26 September 2000; accepted 6 December 2000
AbstractÐThe identi®cation and characterization of a potentially ischemia-selective and orally-active sulfoxide based NMDA ion-
channel blocker showing good neuroprotective activity, (R)-(+)-N-(2-chloro-5-methylthiophenyl)-N0-(3-methylsul®nylphenyl)-N0-
methylguanidine (CNS 5788), is described. # 2001 Elsevier Science Ltd. All rights reserved.
Many di€erent steps in the cascade of events leading to
neuronal cell death have been investigated as possible
targets for therapeutic intervention in the treatment of
stroke and traumatic brain injury. Amongst the most
e€ective mechanisms studied is by inhibition of the
e€ects of excessive concentrations of the excitatory
neurotransmittor glutamic acid, by the blockade of the
NMDA receptor directly at its associated ion channel.1
Many conventional NMDA ion-channel blockers
including dizolcipine2 (MK 801, 1), aptiganel3 (CNS
1102, 2), and N-(2-chloro-5-methylthiophenyl)-N0-
methyl-N0-(3-methylthiophenyl)guanidine (CNS 5161,
3) (Fig. 1) are neuroprotective in animal studies. To
date, it has been dicult to demonstrate the ecacy of
NMDA antagonists in stroke patients, in part because
of behavioral and cardiovascular side e€ects that appear
to limit upper dosage levels in clinical investigations.
containing drug molecules including the nonsteroidal
antiin¯ammatory drug molecule sulindac.6 In this
report we discuss the synthesis and biological evaluation
of compounds containing sulfoxide groups that are
derived from the potent NMDA ion-channel blocker 3
(CNS 5161), which is currently under clinical investiga-
tion for the treatment of chronic pain.
Synthesis of guanidine derivatives in the present study
was achieved through the condensation of the corre-
sponding cyanamides with the aniline hydrogen chloride
salts in suitable solvents such as toluene under re¯uxing
conditions7 (Scheme 1). In our laboratory, 2-chloro-5-
methylthioaniline (5) was prepared4 from commercially
available 2-chloro-5-methylthiobenzoic acid (4) using
diphenylphosphoryl azide in the presence of triethyl
amine in N,N-dimethylformamide as solvent in 56%
yield. For the preparation of 5 in multigram quantities,
this reaction was modi®ed,8 using t-butanol as solvent.
Subsequent hydrolysis of the resulting carbamate inter-
mediate (without isolation) using 10% hydrochloric
acid led to the isolation of 5 in 79% yield. Reaction of
the aniline derivative 5 with cyanogens bromide was car-
ried out in water to yield the desired 2-chloro-5-
methylthiophenylcyanamide4 (6) in 61% yield.
We hypothesized that a molecule that is relatively inert
under normal physiological conditions but that becomes
activated to generate a potent NMDA channel blocker
under hypoxic conditions, would be associated with
reduced side e€ects and improved therapeutic utility.
Conversion of a sulfoxide derivative to an active azir-
idinium species within a hypoxic solid tumor provides
good analogy for a sulfoxide prodrug for ischemic con-
ditions.5 There are numerous precedents5 for sulfoxide-
Methylation of 3-methylthioaniline was carried out fol-
lowing the reported procedure.4 Oxidation of N-methyl-3-
methylthioaniline (7) to N-methyl-3-methylsul®nylaniline
(8), was studied under di€erent reaction conditions using
di€erent reagents, such as sodium periodate, Oxone1,
hydrogen peroxide (alone and in the presence of catalysts)
yPresented at the 219th ACS meeting held at San Francisco, CA on
March 25, 2000. Abstract of Papers, Medicinal Chemistry Division, 095.
*Corresponding author. Tel.: +1-781-255-6700; fax: +1-781-255-6712;
e-mail: padman@cambneuro.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00695-8

502
S. Padmanabhan et al. / Bioorg. Med. Chem. Lett. 11 (2001) 501±504
Oxidation of 2-chloro-5-methylthiophenylcyanamide (6)
was carried out using sodium periodate in aqueous
methanol at room temperature to the corresponding
sulfoxide derivative 11 in 65% yield. Condensation of
11 with N-methyl-3-methylthioaniline hydrochloride
(12) at 135±140 ꢁC led to the isolation of 13 in 32%
yield.
Synthesis of the disulfoxide 14 was achieved by sodium
periodate oxidation of 3 (CNS 5161) in aqueous
methanol in 73% yield.
All the three sulfoxide derivatives (10, 13, 14) exhibited
reduced in vitro anity {[3H]-MK-801 binding assay}12
for the NMDA ion-channel binding site and markedly
reduced sodium channel blocking activity (14C-guanidi-
nium ¯ux assay)13 compared with parent sul®de, 3. The
sulfoxide 10 (CNS 5655), retained the highest level of
NMDA binding anity (ꢀ80-fold reduction) compared
to the other sulfoxides 13 and 14 (Table 1).
Figure 1.
Sulfoxides 10, 13 and 14 were studied in various animal
models14 and 10 showed the preferred pharmacological
pro®le, including activity in the rat middle cerebral
artery occlusion model15 (MCAO model) for stroke
(32% protection at 4.5 mg/kg) and reduced side e€ects,
based on the Irwin general behavioral screen16 and the
rotarod assay for motor e€ects17 (Table 2).
Both the enantiomers of 10 were obtained through pre-
parative chiral HPLC separation (Fig. 2) of the racemic
compound as the free base.18 Further evaluation of the
enantiomers revealed that the in vitro potencies of the
enantiomers do not appear to be signi®cantly di€erent
from the racemate 10 (Table 1). However, based on the
rat MCAO data, the (+)-enantiomer, 15 (CNS 5788)
appears to be at least as potent as 10 (32% protection at
4.5 mg/kg) whereas the (À)-enantiomer, did not show
signi®cant neuroprotection at this dose.
Scheme 1.
to reduce the formation of sulfone derivative and to
avoid the coloration of ®nal product. After considerable
experimentation, sulfoxide (8) was prepared in ꢀ90%
yield through oxidation9 using hydrogen peroxide (5%)
in chloroform in the presence of tri¯uoroacetone.10
Reaction of N-methyl-3-methylsul®nylaniline (8) with
hydrogen chloride in methanol gave the corresponding
hydrochloride (9). Condensation of the cyanamide 6
and anilinehydrochloride 9 in re¯uxing toluene a€orded
10 in ꢀ20% yield initially. Following reinvestigation11
of the synthesis of 10, the yield was improved from 17%
to 50% using a mixture of solvents (dichloromethane/
toluene 1:1) for the condensation step.
The absolute con®guration of the active enantiomer 15
was determined as R by X-ray crystallography19 and
this enantiomer retains the favorable pharmacological
pro®le of the racemate in exhibiting good neuroprotec-
tion at doses not associated with the unwanted CNS
side e€ects as revealed by the Irwin test and rotarod
assays.
In pharmacokinetics experiments for monitoring the
conversion of racemic sulfoxide 10 (CNS 5655) to the
parent sul®de 3 (CNS 5161) in vivo, a plasma HPLC
Table 1. In vitro properties of parent sul®de 3 and derived sulfoxides
Compound
NMDA 3H-MK-801
Binding Ki (nM)
Neuronal Guan.
Flux IC50 (mM)
Cardiac Guan.
Flux IC50 (mM)
3 (CNS 5161)
10 (CNS 5655)
13
14
15 (CNS5788)
(À)-Enantiomer
1.94Æ0.37
179Æ23
1.75Æ0.41
83.1Æ0.50
>25
2.25Æ0.25
25.98Æ3.55
29.5Æ2.1
81.0Æ37.4
23.7Æ2
834Æ114
6587Æ2150
233Æ29
>25
>25
>25
321Æ5
26.0Æ2.05

S. Padmanabhan et al. / Bioorg. Med. Chem. Lett. 11 (2001) 501±504
Table 2. In vivo properties and comparison of CNS 5161 (3) and CNS 5655 (10)
503
Assay system
Measure
3 (CNS 5161)
10 (CNS 5655)
Comparison of potency
CNS 5655 versus CNS 5161
Rat pup hypoxia-ischemia
neuroprotecion model
% Reduction in lesion
volume@the dose
(mg/Kg) tested (ip)
74Æ14%@4 (11)a
100Æ0%@20 (13)a
60Æ22%@10 (15)a
3- to 4-fold less potent
Rat MCAO Neuroprotection
model
% Reduction in total
lesion volume at the
dose tested
33Æ13%@1.5 (6)a
32Æ8%@4.5 (12)a
ꢀ3-fold less potent
(combination iv and ip)
Rat rotarod motor
coordination test
ED50 To cause rats to
fall from the rod (iv)
ED50=0.14 mg/kg
0.25 mg/kg
ED50=8.8 mg/kg
4 mg/kg
63-fold less potent
16-fold less potent
Irwin general behavior
screen
Lowest dose at which
de®nite behavioral
e€ects seen (esp.
excitation) (iv)
Audiogenic seizure mouse
model (DBA/2)
ED50 To block the
audiogenic seizures
ED50=0.9 mg/kg
ED50=12 mg/kg
13-fold less potent
Oral activity
DBA/2 testÐCNS 5655
Irwin testÐCNS 5161
ꢀ25% for CNS 5655
<5% for CNS 5161
aNumbers in parentheses refer to the number of animals tested.
assay was developed for simultaneous determinations of
both compounds in plasma in the range of 10±1000 ng/
ml. Rats (three animals/time pt.) were dosed with 10
mg/kg racemate CNS 5655 (10) using a 5 min iv infu-
sion as a formulation of the drug 3 mg/ml in 40 mM
acetate and 0.3 M mannitol and three control animals
were dosed with vehicle alone.
tion. The half-life of CNS 5788 was about 32 min, the
same as the 31 min, half-life found for CNS 5655. The
CNS 5161 appeared to peak at 30 min and to then
decline with a half-life of 73 min, compared to 55 min
after CNS 5655 administration. The overall percentage
conversion of CNS 5788 to CNS 5161 was ꢀ6%, com-
pared to the 5% seen previously with CNS 5655
administration. Figure 3 shows the plasma levels versus
time curves for 15 (CNS 5788) and 10 (CNS 5655).
The results indicated that after an initial build-up of
sul®de, CNS 5161 (3) (Cmax at 30 min), data for both
sulfoxide and sul®de can be ®t well with single expo-
nential decays, which yield half-lives of 31 and 55 min,
respectively. Based on area under curve (AUC) values
of 93,000 for 10 and 4700 for 3, about 5% of putative
sulfoxide prodrug is reduced to the sul®de. The plasma
levels of sul®de, achieved after the 10 mg/kg iv dose of
CNS 5655 (10) are in the range of the plasma levels
achieved by neuroprotective doses of CNS 5161 (3) in
MCAO experiments (Fig. 3).
Conversion of a methylthio group in 3 (pKa=9.04, log
D7.4=2.5) to the sulfoxide 10 or 15 (pKa=8.63, log
D7.4=0.83) lowers the pKa by 0.41 units (versus 0.5
predicted from electronic substituent e€ects) and the
lipophilicity (log D7.4) by 1.67 units. The di€erence in
log D7.4 may be re¯ected in the improved oral activity
(25±33%) of 10 (CNS 5655) and 15 (CNS 5788) in the
DBA/2 mouse anti-seizure test relative to the parent
sul®de, 3 (Table 2).
Results from preliminary pharmacokinetics experiments
with the acetate salt of CNS 5788 (15) were strikingly
similar to those found with CNS 5655 (10) administra-
A comparison of the biological activities of the sulfoxide-
based channel blocker 10/15 with the parent sul®de 3 in
various animal models is given in Table 2.
Figure 3. Drug levels of CNS 5161 after iv dosing of rats with CNS
5655 (10) and CNS 5788 (15).
Figure 2. Preparative chiral HPLC chromatogram of 10.

504
S. Padmanabhan et al. / Bioorg. Med. Chem. Lett. 11 (2001) 501±504
In conclusion, a sulfoxide based NMDA ion channel
blocker, 15 (CNS 5788), that shows good neuroprotection,
reduced side e€ects and improved oral activity has been
identi®ed and CNS 5788 may have potential for the
treatment of stroke and other ischemia related disorders.
The interesting properties of CNS 5788 (15) led to the
studies on a chiral synthesis, which has been commu-
nicated separately.20
min at room temperature. The reaction mixture was stirred
overnight (ꢀ36 h) and the disappearance of the starting material
was observed on TLC (hexanes/EtOAc, 4:1). The reaction
mixture was diluted with water (20 mL) and repeatedly
extracted with chloroform (3Â20 mL). Combined organic
extracts were washed with sodium thiosulfate solution (10%,
25 mL), dried and concentrated to a€ord the reaction mixture.
The residue was chromatographed on silica gel using gradually
a mixtures of hexanes/ethyl acetate initially 4:1, and later 3:2.
These elutions remove any traces of the unreacted starting
material (if any) and most of the sulfone formed. Finally elution
with ethyl acetate and later ethyl acetate containing 5%
methanol eluted the desired sulfoxide 8 and the fractions were
combined, concentrated under reduced pressure to yield the
sulfoxide (8) as an oil 10 g (90%); TLC CHCl3/MeOH (19:1); Rf
0.48; 1H NMR (CDCl3) d 2.69 (s, 3H, S(O)Me), 2.86 (s, 3H,
NMe), 6.67 (dd, 1H), 6.81 (dd, 1H), 6.92 (d, 1H), 7.27 (m, 1H).
10. Lupattelli, P.; Ruzzioconi, R.; Scafato, P.; Degl'Innocenti,
A.; Paolobelli, A. B. Synth. Commun. 1997, 27, 441.
11. A solution of 3-methylsul®nyl-N-methylaniline hydro-
chloride (9, 2.8 g), 2-chloro-5-methylthiophenylcyanamide (6,
1.8 g) in dichloromethane (25 mL) and toluene (25 mL). The
reaction mixture was slowly heated in an oil bath to 115±
120 ꢁC with a distillation setup for collecting the solvent dis-
tilled during the reaction. After the complete distillation of
dichloromethane the reaction mixture was maintained at 115±
120 ꢁC for 1 h, cooled to room temperature. The residue, after
the solvent removal, was puri®ed by silica gel column chro-
matography using, initially, chloroform and gradually to
chloroform containing 5% methanol as eluents. The almost
colorless product, CNS 5655 10, obtained upon concentration
of fractions, was dried under high vacuum. 1.90 g (53%); mp
156±160 ꢁC; TLC CHCl3/MeOH 4:1; Rf 0.33; 1H NMR
(CD3OD) d 2.46 (s, 3H, SMe), 2.82 (s, 3H, S(O)Me), 3.53 (s,
3H, NMe), 7.19 (s, 1H), 7.21 (d, 1H), 7.40 (dd, 1H), 7.65 (m,
1H), 7.70 (m, 2H), 7.81 (d, 1H).
Acknowledgements
The authors wish to acknowledge Ruth Lavin, Langyu
Wu, C. Chen, S. Wang, W. F. Holt, C. Kirk, K. Burke-
Howie, T. Wolcott and D. Daly for their valuable con-
tributions to the work.
References and Notes
1. Meldrum, B. Cerebrovasc. Brain Metab. Rev. 1990, 2, 27.
2. Wong, E. H.; Kemp, J. A.; Priestley, T.; Knight, A. R.;
Woodru€, G. N.; Iverson, L. L. Proc. Natl. Acad. Sci. U.S.A.
1986, 83, 7104.
3. Reddy, N. L.; Hu, L.-Y.; Cotter, R. E.; Fischer, J. B.;
Wong, W. J.; McBurney, R. N.; Holmes, L. D.; Wong, S. T.;
Prasad, R.; Keana, J. F. W. J. Med. Chem. 1994, 37, 260.
4. Hu, L.-Y.; Guo, J.; Magar, S. S.; Fischer, J. B.; Burke-
Howie, K. J.; Durant, G. J. J. Med. Chem. 1997, 40, 4281.
5. Kwon, C. H.; Blanco, D. R.; Baturay, N. J. Med. Chem.
1992, 35, 2139.
6. Druggan, D. E.; Hooke, K. F.; Noll, R. M.; Hucker, H. B.;
Van Arman, C. G. Biochem. Pharmacol. 1978, 27, 2311.
7. Kampf, A. Chem. Ber. 1904, 37, 1681.
8. To a suspension of 2-chloro-5-methylthiobenzoic acid (4,
20 g) in t-butanol (75 mL) triethyl amine (21 mL) was added
followed by the dropwise addition of diphenylphosphoryl
azide (24 mL) at room temperature. The reaction mixture was
slowly heated to re¯ux for 6 h, cooled to room temperature,
and solvent was removed under reduced pressure. The crude
reaction mixture was dissolved in tetrahydrofuran (50 mL)
followed by the addition of aq HCl (1:1, 50 mL) and was
re¯uxed for 6 h. The reaction mixture was concentrated under
reduced pressure and treated with NaOH (25%) to bring the
pH to 12 with cooling in ice water bath. The product was
repeatedly extracted into ethyl acetate (4Â50 mL) and the
organic layer was washed with water (20 mL). Concentration
of the organic layer led to the isolation of 2-chloro-5-methyl-
thioaniline, 5 (13.6 g, 79%); TLC: hexanes/EtOAc (4:1); Rf 0.58;
1H NMR (CDCl3) d 2.45 (s, 3H), 4.1 (br., 2H, NH2), 6.55 (m,
1H), 6.65 (m, 1H), 7.15 (m, 1H).
12. Keana, J. F. W.; McBurney, R. N.; Scherz, M. W.;
Fischer, J. B.; Hamilton, P. N.; Smith, S. M.; Server, A. C.;
Finkbeiner, S.; Stevens, C. F.; Jahi, C.; Weber, E. Proc. Natl.
Acad. Sci. U.S.A. 1989, 86, 5631.
13. Reith, M. E. A. Eur. J. Pharmacol. Mol. Pharmacol. Sec.
1990, 188, 33.
14. Compounds 13 (24% @ 20 mg/kg) and 14 (0% @ 20 mg/
kg) did not show any signi®cant protection in the hypoxia-
ischemia neuroprotection model.
15. Zhou, D.; Wang, S.; Hu, L.; Durant, G. J.; Fischer, J. B.;
Knapp, A.; Holt, W. F.; Gamzu, E. R. Soc. Neurosci. Abstr.
1997, 23, 2433.
16. Irwin, S. Psychopharmacologia 1968, 13, 222.
17. Dunham, N. W.; Miya, T. S. J. Am. Pharm. Ass. 1957, 46,
208.
18. Resolution of racemate through chiral separation was
carried out by Chiral Technologies, Exton, PA.
9. To a solution of N-methyl-3-methylthioaniline (7, 10 g) and
tri¯uoroacetone (1 mL) in chloroform (50 mL) hydrogen peroxide
(41 mL, 5%) solution was added dropwise over a period of 15
19. Burkhart, B. M.; Duax, W. L. Unpublished results.
20. Padmanabhan, S.; Lavin, R. C.; Durant, G. J. Tetra-
hedron: Asymmetry 2000, 11, 3455.