Enantiomerically pure tetrahydroquinoline derivatives as in vivo potent antagonists of the glycine binding site associated to the NMDA receptor

Article abstract of DOI:10.1016/j.bmcl.2003.06.001

To identify neuroprotective agents after stroke, new substituted tetrahydroquinoline derivatives were designed as antagonists of the glycine binding site associated to the NMDA receptor, satisfying the key pharmacophoric requirements. In particular, the racemate 3c exhibited outstanding in vivo activity in the MCAo model in rats, when given iv both pre- and post-ischemia. Pure enantiomers 3c-(+) and 3c-(-) have been prepared following an original synthetic route. Despite the significant difference of activity observed in vitro, they shown similar neuroprotective profile in the MCAo model in rats.

Full text of DOI:10.1016/j.bmcl.2003.06.001

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3863–3866
Enantiomerically Pure Tetrahydroquinoline Derivatives as
In Vivo Potent Antagonists of the Glycine Binding Site
Associated to the NMDA Receptor
Romano Di Fabio,* Elvira Tranquillini, Barbara Bertani, Giuseppe Alvaro,
Fabrizio Micheli, Fabio Sabbatini, Maria Domenica Pizzi, Giorgio Pentassuglia,
Alessandra Pasquarello, Tommaso Messeri, Daniele Donati, Emiliangelo Ratti,
Roberto Arban, Giovanna Dal Forno, Angelo Reggianiy and Robert J. Barnaby
Medicines Research Centre, GlaxoSmithKline S.p.A, Via Fleming 4, 37135Verona, Italy
Received 23 May 2003; revised 27 May 2003; accepted 25 June 2003
Abstract—To identify neuroprotective agents after stroke, new substituted tetrahydroquinoline derivatives were designed as
antagonists of the glycine binding site associated to the NMDA receptor, satisfying the key pharmacophoric requirements. In par-
ticular, the racemate 3c exhibited outstandingin vivo activity in the MCAo model in rats, when given iv both pre- and post-ische-
mia. Pure enantiomers 3c-(+) and 3c-(À) have been prepared following an original synthetic route. Despite the significant
difference of activity observed in vitro, they shown similar neuroprotective profile in the MCAo model in rats.
# 2003 Elsevier Ltd. All rights reserved.
The control of the excitotoxic cascade triggered by the
activation of the NMDA receptor1 is a promisingstrat-
egy to discover effective neuroprotective agents able to
prevent cerebral tissues degeneration after stroke.2
compounds fulfillingthe necessary pharmacophore
requirements.15
In particular, after the preparation of 3a (R=H), the
first member of this series, the compound was char-
acterized both in terms of in vitro affinity to the glycine
bindingsite 19 affinity and in vivo in the NMDA induced
convulsions model in mice, after iv administration.20
High in vitro affinity was observed (pKi=8.18 vs 8.52
for compound 3a and 1, respectively) but poor in vivo
anticonvulsant activity (pIC50=1 mg/kg vs 0.06 mg/kg15
for compound 3a and 1, respectively). Therefore, the
substitution of the para position of the terminal aro-
matic moiety was explored in detail to maximize both
the in vitro affinity and to optimize the in vivo profile.
To this event, the pool of compounds shown in Table 1
was prepared (Scheme 1) startingfrom the free carboxyl
derivative 4, intermediate synthesized in large scale as
previously reported.8,17 This compound was transformed
in high yield into the corresponding N-phenyl amide
derivatives 6a–p in two steps, by sequential activation of
the carboxyl group via the corresponding 2-pyridyl thiol-
ester,18 to afford intermediate 5 in 67% yield after puri-
fication by flash chromatography, and reaction with
substituted anilines in toluene at reflux for 2 h. Finally,
The extensive exploration performed by former Glaxo-
Wellcome on non-competitive NMDA antagonists act-
3À5
ingat the lgycine bindingsite,
gave rise to the
identification of the glycine antagonists,6À15 of type 1
and 2 shown in Figure 1.
Notably, some of these compounds exhibited out-
standingneuroprotective activity in animal model of
cerebral ischemia, after permanent occlusion of the
middle cerebral artery (MCA),16 without relevant side
effects in rodents. As part of a wide program aimed
toward the identification of new glycine antagonists, the
tetrahydroquinoline (THQ) derivatives of general struc-
ture 3, bearingthe exo a,b-unsaturated N-aryl amide
moiety at the C-4 position, were identified as a class of
*Correspondingauthor. Tel.: +39-045-921-8879; fax: +39-045-921-
8196; e-mail: rdf26781@gsk.com
yCurrent address: ScheringPlough, S. Raffaele Science Park, Milan,
Italy.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.06.001

3864
R. Di Fabio et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3863–3866
Scheme 1. Effect of the substituents in the para position of the terminal aromatic moiety; (a) 2,20dipyridyl disulphide, PPh3, THF, rt, 3 h, 67%; (b)
(i) ArNH2, toluene, 2 h, reflux, 78–85%; (ii) NaOH, EtOH/MeOH 95:5, rt, 30 min, 90%; (c) DEAD, PPh3, R*OH, THF, rt, 30 min, 0–100%; (d) (i)
LiOH, THF/H2O, 3:1, rt, 1 h; (ii) Me3SiONa, THF, rt, 1 h, 85–90% (two steps).
The most active compound of the series 3c was found to
be also the most potent in vivo in the NMDA induced
convulsions model in mice, after iv administration
(ED50=0.2 mg/kg).
Based on this preliminary profile, this compound was
characterized in the MCAo model in rats, both pre- and
post-ischemia. When 3c was given iv, 5 min prior occlu-
sion of the MCA (pre-ischemia), at the doses of 0.01,
0.1, 1 and 10 mg/kg, respectively, as shown in Table 2, a
significant reduction of the size of gross brain damage
Figure 1. New series of glycine antagonists.
was observed (brain damage volume 73.7Æ8.5 mm3
(n=10) for the vehicle treated group). At the same doses
the infarct volumes were 64.3Æ12.7 mm3 (n=8),
47.8Æ7.1 mm3 (n=8; p<0.05), 41.8Æ6.2 mm3 (n=8;
p<0.01) and 39.7Æ7.6 mm3 (n=8; p<0.01), respec-
tively [ED50=0.02 mg/kg (0.01-0.12)] for the group of
rats treated with 3c.
the hydrolysis of the ethyl ester with NaOH in EtOH/
MeOH 95:5, gave title racemates 3a–p quantitatively.21
As reported in Table 1, all the compounds prepared
exhibited high in vitro affinity, confirming the presence
in the north-east part of the receptor of a non-
hydrophobic pocket of limited size.15
Post-ischemia, when 3c was given both at 0.3 and 1 mg/
kg, iv, 6 h after MCA occlusion, a significant neuropro-
tective effect was also observed. In this case, the infarct
volume was, 36.6Æ5.9 mm3 (n=8; p<0.05) and
29.9Æ3.6 mm3 (n=8; p<0.01) at 0.3 and 1.0 mg/kg
dose, respectively (brain damage volume 73.6Æ6.5 mm3
(n=10) for the vehicle treated group). Considering the
outstandingneuroprotective activity observed, both
pre- and post-ischemia, a suitable synthetic strategy was
put in place to prepare sufficient amount of the single
enantiomers. In this event, as shown in Scheme 1, a
‘library’ of diastereoisomers was prepared reactingin
parallel the racemic carboxylic acid derivative 3c with a
series of commercially available alcohols (R*OH)
belonging to the chiral pool. Several classic esterifica-
tion methods were attempted, usinga variety of car-
boxyl activatingagents (CDI; mixed anhydride; DCC;
EDC; Mitsunobu reaction). Amongthose, the Mitsu-
nobu reaction was proven to be the most effective
Table 1. In vitro affinity at the glycine binding site
Entry
R
pKia
1
H
8.52
8.18
7.76
8.52
8.18
8.18
8.17
8.18
8.15
7.72
8.38
8.32
7.78
8.15
7.73
3a
3b
3c
3d
3e
3f
3g
3h
3i
H
NH2
NHCOCH3
NHCOCH2OCH3
NHSO2CH3
NHSO2NH2
CH2CONH2
CH2CN
CH2COOH
CH2NHCONH2
CH2NHCOCH3
CH2NHCO-i-C3H7
CH¼CHCONH2
CH¼CHCN
3l
3m
3n
3o
3p
aInhibition of bindingof [ 3H] glycine. 19

R. Di Fabio et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3863–3866
3865
Table 2. Neuroprotective activity of racemate 3c in the MCAo model
in rats
Table 3. Neuroprotective activity of single enantiomers 3c-(+) and
3c-(À) post-ischemia in the MCAo model in rats (single dose given 6 h
after occlusion)
Type of study
Pre-ischaemia
Dose
(mg/kg, iv)
Infarct volume
(mm3ÆSEM)
Damage
reduction (%)
Compd
Dose
(mg/kg, iv)
Infarct volume
(mm3ÆSEM)b
Damage
reduction
0.01
0.1
1
64.3Æ12.7
47.8Æ7.1b
41.8Æ6.2c
39.7Æ7.6c
12.7
35.1
43.3
46.1
3c-(+)
0.3
1
55.6Æ9.5
26.3%
41.2%
44.4Æ6.3a
10
3c-(À)
0.3
1
58.7Æ10.1
42.9Æ7.7a
22.2%
43.2%
Vehicle
73.7Æ8.5
Post-ischaemiaa
0.3
1
36.6Æ5.9c
50.3
59.4
Vehicle
75.5Æ8.5
29.9Æ3.6c
ap<0.05 vs vehicle.
bp<0.01 vs vehicle.
Vehicle
73.6Æ6.5
aSingle dose given 6 h after occlusion.
bp<0.05 vs vehicle.
ferential stereospecific recognition of the enantiomers of
type R.23 The same kind of difference between the two
enantiomers was observed in vivo in the NMDA-
induced convulsion model in mice, being 3c-(+) sig-
nificantly more potent than 3c-(À) [ED50=0.31 mg/kg
and 1.52 for 3c-(+) and 3c-(À), respectively, vs 0.2 mg/
kgfor 3c].
cp<0.01 vs vehicle.
method. Accordingly, when 3c was treated with ‘diverse’
primary and secondary chiral alcohols in THF at room
temperature, in the presence of PPh3 and DEAD, the
correspondingester derivatives were obtained in vari-
able conversion (0–100% by HPLC analysis).
In the MCAo model in rat, when 3c-(+) was given iv,
6 h post-ischemia, the compound exhibited a sign-
ificant neuroprotective activity at 1 mg/kg (Table 3:
44.4Æ6.3 mm3 in treated animals; vehicle treated rats
75.5Æ8.5 mm3; p<0.05; n=8) but it was found inactive
at the lower dose (0.3 mg/kg, 55.6Æ9.5 mm3 in treated
animals; vehicle treated rats 75.5Æ8.5 mm3; n=8).
Total conversion was observed with (S)-(+)-1-inda-
nol, methyl-(S)-(+)-mandelate, methyl (R)-(À)-lactate,
tert-butyl (R)-(+)-lactate, methyl (R)-(À)-3-hydroxy-
2-methylpropionate and (R)-(À)-2-butanol; no reaction
occurred with (+)-menthol. The tert-butyl (R)-(+)-
lactate ester derivative 7c was chosen based on the
efficiency of the separation of the two diastereoisomers
both by tlc. (AcOEt/cyclohexanes 7:3, Rf=0.40 and
0.47 respectively) and HPLC (Supelcosil LC-CN, THF–
hexane 30:70, flow=0.8 mL/min, l=260 nm: retention
time=12.56 and 15.60 min, respectively).
This result was explained in terms of unbalanced PK/
PD profile in rat of the two enantiomers, beingthe most
potent enantiomer 3c-(+) was also the less abundant
compound in plasma (mean AUC0Àinf=3.44, 19.8 and
11.5 mgh/mL for 3c-(+), 3c-(À) and 3c, respectively). In
fact, when the less in vitro active enantiomer 3c-(À) was
evaluated in the MCAo model, as in the case of 3c-(+),
despite the significant difference of in vitro potency, a
similar neuroprotective profile was observed (Table 2:
1 mg/kg, 42.9Æ7.7 mm3 in treated animals; vehicle trea-
ted rats 75.5Æ8.5 mm3; p<0.05; n=8).
After large scale preparation of diastereoisomers 7c-(+)
and 7c-(À) (performed both by flash chromatography or
preparative HPLC), the next issue was the identification
of mild hydrolysis conditions able to prevent both the
racemization of the a-amino acid type stereogenic centre
and to minimize the formation of lactic acid duringthe
final removal the ester group. To that end, when the
reaction was performed in the presence of NaOH in
EtOH/MeOH 95:5, at room temperature for 3 h, a par-
tial conversion of the ester into the correspondingcar-
boxyl acid was observed, due to the competitive
hydrolysis of the tert-butyl lactate ester and formation of
significant amount of lactic acid. Conversely, when the
basic hydrolysis was attempted with LiOH in THF/H2O
3:1 at room temperature for 30min, the reaction was che-
moselective and the single enantiomers 3c-(+) and 3c-(À)
were isolated in high yield after acid work up, and then
In conclusion, enantiomerically pure THQ derivatives 3c-
(+) and 3c-(À) have been identified as in vitro and in vivo
potent glycine antagonists. Based on their outstanding
pharmacological profile observed in animal models of
cerebral ischemia, both enatiomers can be considered as
promisingand effective neuroprotective agents.
Acknowledgements
+
quantitatively transformed into the correspondingNa
salts in the presence of (CH3)3SiONa in THF.22
The authors would like to thank Dr. C. Marchioro and
Dr. M. Hamdan and their co-workers for the NMR and
MS support, respectively. Thanks are also due to Dr. S.
Bellini for the HPLC analysis and [a]D data.
The two enantiomers were then evaluated in the bind-
ingassay in vitro. A significant difference of potency
was observed [pKi=8.79 and 7.30, for 3c-(+) and 3c-
(À), respectively vs 8.52 for the racemate 3c].
References and Notes
Based on these results, it can be stated that the receptor
binding site of the glycine antagonists shows a pre-
1. Collingridge, G. L.; Watkins, J. C. The NMDA Receptor,
2nd Ed.; IRL: Oxford, UK, 1994.

3866
R. Di Fabio et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3863–3866
2. For a review on different neuroprotective approaches
potentially useful to control the progression of the brain
damage after stroke, see: Zivin, J. A. Drugs 1997, 54, 83.
3. Di Fabio, R.; Gaviraghi, G.; Reggiani, A. La Chim. l’In-
dust. 1996, 78, 283.
4. Leeson, P. D.; Iversen, L. L. J. Med. Chem. 1994, 37, 4053.
5. Iversen, L. L.; Kemp, J. A.; Leeson, P. D. TiPS 1993, 14, 20.
6. Di Fabio, R.; Araldi, G.; Baraldi, D.; Cugola, A.; Donati,
D.; Gastaldi, P.; Giacobbe, S. A.; Micheli, F.; Pentassuglia, G.
Farmaco 2001, 56, 791.
Donati, D.; Feriani, A.; Gastaldi, P.; Gaviraghi, G.; Hewkin,
C. T.; Micheli, F.; Missio, A.; Mugnaini, M.; Pecunioso, A.;
Quaglia, A. M.; Ratti, E.; Rossi, L.; Tedesco, G.; Trist, D. G.;
Reggiani, A. J. Med. Chem. 1997, 40, 841.
16. Tamura, A.; Graham, D. I.; McCullogh, J.; Teasdale,
G. M. J. Cerebr. Flow Metab. 1981, 1, 53.
17. For an alternative synthesis of THQ derivatives see: Car-
ling, R. W.; Leeson, P. D.; Moseley, A. M.; Baker, R.; Foster,
A. C.; Grimwood, S.; Kemp, J. A.; Marshall, G. R. J. Med.
Chem. 1992, 35, 1942.
7. Gaviraghi, G.; Di Fabio, R., Donati, D.; Feriani, A.; Ratti,
E., Reggiani, A.; Trist, D. G. In Medicinal Chemistry into the
Millennium, Campbell M. M.; Blagbrough, I. S., Eds.; Royal
Society of Chemistry, Cambridge, 2001; p 85–92
8. Di Fabio, R.; Alvaro, G.; Bertani, B.; Giacobbe, S. Can. J.
Chem. 2000, 78, 809.
9. Donati, D.; Di Fabio, R. Pharm. Acta Helv. 2000, 74, 239.
10. Di Fabio, R.; Conti, N.; De Magistris, E.; Feriani, A.;
Provera, S.; Sabbatini, F. M.; Reggiani, A.; Rovatti, L. J.
Med. Chem. 1999, 42, 3486.
11. Micheli, F.; Di Fabio, R.; Capelli, A. M.; Cugola, A.;
Curcuruto, O.; Feriani, A.; Gastaldi, P.; Gaviraghi, G.;
Marchioro, C.; Orlandi, A.; Pozzan, A.; Quaglia, A. M.;
Reggiani, A.; van Amsterdam, F. Arch. Pharm. Pharm. Med.
Chem. 1999, 332, 73.
18. Di Fabio, R.; Summa, V.; Rossi, T. Tetrahedron 1993, 49,
2239, and references therein.
19. Affinity to glycine binding site was measured by inhibi-
tion of the bindingof [ 3H] glycine to crude synaptic mem-
branes prepared from adult rats cerebral cortex: Kishimoto,
H.; Simon, J. R.; Aprison, M. H. J. Neurochem. 1981, 37,
1015.
20. Chiamulera, C.; Costa, S.; Reggiani, A. Psycopharmacol-
ogy 1990, 112, 551. For a detailed description of the experi-
mental procedures of the in vitro and in vivo biological tests,
see ref 10.
21. Compounds 3a–p: C, H and N analyses were within 0.4%
of the theoretical values; mp were over 250 ꢀC.
22. 3c-(+): [a]D=+16ꢀ (c 0.25, DMSO); 3c-(À): [a]D=À16ꢀ
(c 0.25, DMSO).
12. Hewkin, C. T.; Di Fabio, R.; Conti, N.; Cugola, A.; Gas-
taldi, P.; Micheli, F.; Quaglia, A. M. Arch. Pharm. Med.
Chem. 1999, 332, 55.
13. Di Fabio, R.; Cugola, A.; Donati, D.; Feriani, A.; Gavir-
aghi, G.; Ratti, E.; Trist, D. G.; Reggiani, A. Drugs Future
1998, 23, 61.
14. Giacobbe, S. A.; Baraldi, D.; Di Fabio, R. Bioorg. Med.
Chem. Lett. 1998, 8, 1689.
15. Di Fabio, R.; Capelli, A. M.; Conti, N.; Cugola, A.;
23. After the completion of this study, the enantiomer 3c-(+)
was re-synthesized (R. Di Fabio and G. Alvaro, unpublished
results) followingthe same kind of asymmetric synthesis
recently published for the preparation of some analogue THQ
derivatives: Di Fabio, R.; Alvaro, G.; Bertani, B.; Donati, D.;
Giacobbe, S.; Marchioro, C.; Palma, C.; Lynn, S. M. JOC,
2002, 67, 7319. By this approach, it was clearly proven that the
absolute configuration of the C-2 stereogenic center of the
enantiomer 3c-(+) was of type R.