Non-competitive inhibitors of metabotropic glutamate receptor 5 (mGluR5)

Article abstract of DOI:10.1016/j.bmcl.2005.03.089

Based on a pharmacophore alignment on known non-competitive mGluR5 inhibitors applying 4SCan technology, a new lead series was identified and further structurally investigated. K_i's as low as around 100 nM were achieved.

Full text of DOI:10.1016/j.bmcl.2005.03.089

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2876–2880
Non-competitive inhibitors of metabotropic glutamate
receptor 5 (mGluR5)
Stefan Tasler,a,* Jurgen Kraus,a Stefano Pegoraro,a Andrea Aschenbrenner,a
¨
Elena Poggesi,b Rodolfo Testa,b Gianni Mottab and Amedeo Leonardib,*
a4SC AG, Am Klopferspitz 19a, 82152 Martinsried, Germany
bRecordati Industria Chimica E Farmaceutica S. p. A., Via Matteo Civitali 1, 20148 Milan, Italy
Received 12 November 2004; revised 23 March 2005; accepted 23 March 2005
Available online 5 May 2005
Abstract—Based on a pharmacophore alignment on known non-competitive mGluR5 inhibitors applying 4SCanÒ technology, a
new lead series was identified and further structurally investigated. KiÕs as low as around 100 nM were achieved.
Ó 2005 Elsevier Ltd. All rights reserved.
Activation of metabotropic glutamate receptors
(mGluRs) modulates neuroplasticity and neuronal excit-
ability, rendering these receptors intriguing therapeutic
targets for certain neurological and psychiatric disor-
ders, such as ischemia, schizophrenia, or chronic neuro-
degenerative diseases. Interference with this process has
a potential for analgesic or anxiolytic treatment.1–5
mGluRs belong to type C superfamily of G-protein-cou-
pled receptors and are further organized into groups
according to sequence homology, signal transduction
mechanism, and agonist selectivity.3,6 Group I recep-
tors, that is, mGluR1 and mGluR5, are mainly
excitatory and preferentially activate phosphoinositide-
specific phospholipase C. Receptors of groups II and
III (mGluRs 2, 3, and 4, 6–8, respectively), are nega-
tively coupled to adenylate cyclase.1–3,5,7 Antagonists
for group I or agonists for group II/III mGluRs possess
potential as neuroprotective agents. Especially for
mGluR5, an important role in nociception was proven,
making this receptor highly interesting for the treatment
of neuropathic/chronic pain.1,3,4
tions of the receptor, resulting in G-protein activation.
Non-competitive inhibitors bind to the transmembrane
domain and seem to interrupt the transfer of conforma-
tional changes from the amino-terminal domain through
the transmembrane region toward G-protein activa-
tion.3,4,7–10 In contrast to the glutamate binding site,
which is naturally highly conserved throughout all
groups of mGluRs, it was anticipated to achieve higher
subtype selectivity addressing the transmembrane do-
main with a non-competitive inhibitor rather than with
a competitive one for the active site. As crystal structure
data was only available for a truncated extracellular do-
main, including the active site,8 a usually preferable in
silico docking approach could not be applied to the dis-
covery of new structural types of non-competitive inhibi-
tors. With two known non-competitive inhibitors for
mGluR5 at hand, MMPEP (1) from Novartis and 2
from NPS Pharmaceuticals (Fig. 1), displaying IC50 val-
ues of 10 and 74 nM at mGluR5, respectively,11,12
a
pharmacophore alignment on these compounds was per-
13
formed using 4SCanÒ with a virtual library of around
3.3 Mio commercially available compounds. This proce-
dure resulted in the selection of 188 compounds for an
initial and further 123 for a follow-up screen in a bind-
ing assay on mGluR5,14 which directly led to the identi-
fication of a lead structure (Fig. 1), with representatives
possessing KiÕs below 200 nM and an excellent selectivity
profile: Compounds A3/B5 and B3 (Tables 1 and 2) were
virtually inactive at mGluR1, all mGluRs of group II
and III, 5-HT1A, l opioid and dopamine D2 receptors,
with KiÕs abov e 10lM.
Binding of glutamate to the extracellular binding site of
mGluRs stabilizes an active conformation within a com-
plex equilibrium of different active/inactive conforma-
Keywords: Metabotropic glutamate receptor; Molecular modeling;
Non-competitive antagonist; Neuroplasticity.
*
Corresponding authors. Tel.: +49 89 7007630; fax: +49 89 70076329
(S.T.); fax: +39 02 48709017 (A.L.); e-mail addresses: stefan.tasler@
4sc.com; leonardi.a@recordati.it
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.03.089

S. Tasler et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2876–2880
2877
Table 1. Keto-ester derivatives based on structure A (Scheme 1) and
their inhibitory activity against mGluR5
CN
N
N
N
O
Compound
Structure A
Ki [nM] (inhib.@ 10 lM)
O
1
2
A1
— (18%)
O
R’
A2
A3
A4
A5
9475
172
401
171
O
O
O
N
( )n
R
O
I
tether
II
center
Figure 1. Known non-competitive inhibitors 1 and 2 for mGluR5 and
the identified new lead structure.
Based on this structure, a synthetic route was estab-
lished, allowing for a simple derivatization of portions
I and II (Fig. 1) of the molecule in a highly convergent
manner. Starting either with 4-aminobenzoic acid (3a)
or its sodium salt 3b enables the initial attachment of
the anhydride- (I) or phenacyl-portion (II), respectively,
and subsequent systematic SAR exploration on the
remaining side of the molecule (Scheme 1). Whereas carb-
oxylate 3b reacted directly with corresponding phena-
cylbromides 6a15 in DMF to give keto-ester 7, KF as
weak base was required for this step to take place using
a carboxylic acid 5.16 Imide formation with differently
substituted anhydrides 4 was achieved either neatly
(melting point of the anhydride <150 °C), or under
microwave irradiation. For substrates with melting
points between 150 and 190 °C, a smooth reaction pro-
ceeding was achieved by addition of a few drops DMF
to the neat reaction mixture to give a highly concen-
trated solution/melt at 160 °C. For substrates with a
mono-cyclic six-membered or a heteroaromatic anhy-
dride, conversion usually stopped at the stage of the cor-
responding amide, lacking the final condensation to the
imide unit. In those cases, EDC in DMF was necessary
for imide closure, additional catalytic DMAP was added
to conversions proceeding too sluggishly.17
A6
A7
A8
— (32%)
303
18,022
O
A9
— (39%)
1604
O
O
A10
A11
A12
A13
A14
O
O
— (21%)
2037
O
O
3786
NO2
For both portions I and II of the substrate, around 30
derivatives were synthesized. meta-Connectivity at the
central ring of these keto-esters proved to be detrimental
to the activity. As becomes evident from Table 1 for part
II, activity is clearly improving going from a plane phe-
nyl ring (A1) via mono-substitution (A2 and A9) toward
3,4-disubstitution (A3, A10). Mono-substitution in
either 3- or 4-position (Cl, Br, NO2, CN, CF3, and
piperidine tested in addition to A2 and A9) resulted in
low activities in general, with two exceptions being pyr-
rolidine- and morpholine-substituted derivatives A15
and A16, which showed activity in the low micromolar
range. Connecting the 3- and 4-substituents by ring clo-
sure, resulting in a more rigid framework, retained activ-
ity in case of an additional six-membered ring (A5, A12
compared to A3, A10), but led to a drop in activity with
five-membered rings (slightly for A4, drastically for
A11). Comparing the upper and lower halves of Table
1, the more lipophilic derivatives clearly afforded lower
O
1683
NO2
N
A15
A16
1088
3359
O
N
KiÕs. In order to explore the spatial nature of this hydro-
phobic pocket, portion II was further extended either
with sterically demanding substituents (A6, A8, and with
portion II being adamantyl) or flat groups (A7, or por-
tion II being a p-biphenyl). Only rather flat structures

2878
S. Tasler et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2876–2880
Table 2. Keto-ester derivatives based on structure B (Scheme 1) and
their inhibitory activity against mGluR5
Table 2 (continued)
Compound
Structure B
Ki [nM] (inhib.@ 10 lM)
Compound
Structure B
Ki [nM] (inhib.@ 10 lM)
O
N
O
B14
B15
4795
H
B1
N
190
O
O
O
H
N
O
6426
H
N
O
B2
B3
150
113
O
H
O
N
O
B16
B17
3267
1836
O
N
O
N
O
O
O
N
O
N
O
O
B4
B5
98
Br
Br
O
B18
B19
— (41%)
— (34%)
O
N
O
172
N
H
O
O
OH
O
N
B6
B7
216
seem to fit nicely, with naphthalene (A7) or tetrahydro-
naphthalene (A5) already filling the space available.
More steric bulk led to a dramatic decrease of activity.
A different disubstitution pattern (e.g., 2,5-dimethyl)
was not tolerated at all. The same applied to the incor-
poration of heteroaromatics in portion II of the
molecule.
O
O
N
1012
O
O
N
As to part I of the molecule, hexahydro- and tetrahydro-
phthalimide portions were most favorable (Table 2).
The imide portion and the second cycle were usually
cis-anellated, the only exception being compound B2
(racemate). Adding a further substituent to the hexa-
hydrophthalic moiety resulted in a slight increase in
activity (B3, B4), but structurally, an additional stereo-
center was introduced, whose relative configuration
was investigated for B3. Here, an almost 1:1 ratio for
exo- versus endo-methyl was determined by spin analysis
B8
— (28%)
— (33%)
— (10%)
506
O
O
N
B9
O
N
O
N
1
of its H NMR spectrum. Installing a double bond at
B10
B11
that part eliminated the problem of testing mixtures of
diastereomers and kept the biological activity almost
within the same range (B5, B6). Increasing steric de-
mand (norbornene unit in B7), or switching to phthalic
portions (as e.g., in B8) or analogous heteroaromatics
(like e.g., in B9) was detrimental to activity, as was an
endocyclic central double bond as in B10 and its six-ring
homologue (28% inhibition at 10 lM).
O
O
N
O
O
N
B12
B13
411
Mono-cyclic imides, however, were tolerated, leading to
only a slight decrease in activity of about factor 3, as
long as five-membered imides were incorporated with-
out an endocyclic double bond (cf. B11, B12 vs B14,
B15). The latter seems to be a general conclusion when
including the poor results obtained for B10 (and its
O
O
N
— (38%)

S. Tasler et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2876–2880
2879
O
O
O
Br
O
R'
OH
O
R'
O
O
6
O
N
O
N
4a
O
ii
5
O
O
A
O
i
OR
H2N
iii
O
O
i or iv;
(v)
3a: R = H
3b: R = Na
O
O
O
O
Br
O
O
O
N
(
)
n
H2N
(
)
B
7
n
R
O
6a
O
R
4
O
˚
Scheme 1. Strategy for synthesizing a library of keto-esters. Reagents and conditions: (i) DMF, molecular sieves 4 A, microwave, 160 °C, 90 min; (ii)
KF, DMF, 65 °C, 3 h; (iii) DMF, rt, 2 h; (iv) neat, 160 °C, 1 h; (v) EDC, optional: DMAP, DMF, rt, 20 h.
bridge
O
N
C
O
2-3 atoms
4 atoms
5-6 atoms
O
O
O
O
O
O
H
N
O
O
O
N
H
N
H
N
N
H
N
H
N
N
O
O
O
O
O
O
O
H
C1
C4
C7
C10
C13
C16
O
O
O
O
H
N
H
N
N
N
N
N
N
H
O
N
H
H
H
O
O
C2
C5
C8
C11
C14
C17
O
O
O
H
N
H
N
H
N
N
O
N
H
N
H
N
O
O
C3
C6
C9
C12
C15
Figure 2. Replacement of the keto-ester bridge by several alternatives, activity on mGluR5 in general below 35% inhibition at 10 lM.
six-ring homologue) and phthalic derivatives B8 and B9.
Six-membered imides still displayed some activity in the
low micromolar range at best (B16, B17 vs B18). Two
carbonyl functionalities adjacent to the nitrogen proved
to be essential (B11 vs B13), as did the imide structure in
general as compared to an amide (B5 vs B19).
to be the saturated five-membered imides B3 and B11
with t1/2 around 1–2 h.19
Addressing the instability at the ester bridge, a variety of
different tether units were synthesized (Fig. 2). Having
started with the obvious, a replacement of the ester unit
by an amide (C4, C5), almost no activity was detectable.
Upon attaining a few equally poor results, also with
shortened (C1–C3) or lengthened (C16, C17) tethers,
semi-empirical calculations of partial charge distribu-
tions were performed for the parent keto-ester A3/B5,
several compounds with new bridging units already
tested (C3–C7, C9, C10, C12) and for numerous pro-
posed structures as one possibility to gain some insight
into reasons for the loss in activity when changing just
an atom within the tether and to prioritize new syntheses.
By comparison of the charge distribution maxima/min-
ima throughout the whole molecule to that of keto-ester
A3/B5, compound C11 displayed high similarities and
was therefore most promising. However, not even this
For a proof of concept, the effect of these keto-esters in a
cell assay had to be investigated, in which the intracellu-
lar Ca2+ release was monitored, stimulated by the full
agonist quisqualic acid.18 And indeed, keto-esters
showed non-competitive antagonistic activity. However,
these compounds displayed some instability in human
plasma, with the keto-ester bridge and the imide portion
being vulnerable to hydrolysis. Inconveniently, stability
in portion I of the compounds increased for those struc-
tures leading to a decrease in activity, with lactam B13
being most stable. Its half-life due to hydrolysis of only
the ester bridge was determined to be 8–17 h in depen-
dence on human plasma samples. Most unstable proved

2880
S. Tasler et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2876–2880
8. Kunishima, N.; Shimada, Y.; Tsuji, Y.; Sato, T.; Yamam-
oto, M.; Kumasaka, T.; Nakanishi, S.; Jingami, H.;
Morikawa, K. Nature 2000, 407, 971.
prediction allowed for an identification of a compound
with an activity above 35% inhibition at 10 lM.
9. Parmentier, M.-L.; Prezeau, L.; Bockaert, J.; Pin, J. P.
Trends Pharmacol. Sci. 2002, 23, 268.
10. Pin, J. P.; De Colle, C.; Bessis, A.-S.; Acher, F. Eur. J.
Pharmacol. 1999, 375, 277.
11. Gasparini, F.; Andres, H.; Flor, P. J.; Heinrich, M.;
In summary, a highly potent class of new mGluR5
inhibitors was identified using 4SCanÒ in silico technol-
ogy, with activities down to KiÕs of around 100 nM.
Numerous derivatives synthesized allowed for a conclu-
sive SAR study, offering a few possibilities for substitu-
tion at either portion I or II of the molecule, retaining
adequate activity but providing different influences on
physicochemical properties. A chemical stabilization of
these substrates failed so far but is still under
investigation.
Inderbitzin, W.; Lingenho¨hl, K.; Muller, H.; Munk, V. C.;
¨
Omilusik, K.; Stierlin, C.; Stoehr, N.; Vranesic, I.; Kuhn,
R. Bioorg. Med. Chem. Lett. 2002, 12, 407.
12. Van Wagenen, B. C.; Stormann, T. M.; Moe, S. T.;
Sheehan, S. M.; McLeod, D. A.; Smith, D. L.; Isaac, M.
B.; Slassi, A. WO Patent 01/12627 A1, 2001.
13. For a pharmacophore alignment, compounds of a virtual
library are partitioned into fragments, an anchor fragment
is selected and aligned on the parent structure, for
example, MMPEP (1), and adjusted in a way that similar
electrostatic and hydrophobic interaction possibilities
become superimposed (unlike a structural alignment, in
which similar parts of the molecules are superimposed).
Next fragment is then connected to the first one and again
adjusted for maximal interaction identities based on the
position chosen for the first fragment. Several permuta-
tions of starting fragments are evaluated, resulting in a
pharmacophore alignment of maximal similarities in
interaction possibilities. For details, see: Seifert, M. H.
J.; Wolf, K.; Vitt, D. Biosilico 2003, 1, 143.
14. Affinity of compounds at mGlu5 receptors was evaluated
utilizing rat brain membranes and the ligand [3H]MMPEP
according to: Mutel, V.; Ellis, G. J.; Adam, G.; Chaboz,
S.; Nilly, A.; Messer, J.; Blueuel, Z.; Metzler, V.;
Malherbe, P.; Schlaeger, E.-J.; Roughley, B. S.; Faull, R.
L. M.; Richards, G. J. Neurochem. 2000, 75, 2590.
15. If not commercially available, synthesized according to:
Organikum, 21st ed.; Schwetlick, K., Ed.; Wiley-VCH:
Weilheim, 2001, pp 613–614.
Acknowledgements
The authors want to thank Cathal Meere and Oliver
Muller for their synthetic contributions, Dr. Babett
¨
Krauss for her sophisticated NMR input, Dr. Thomas
Herz for performing semi-empirical calculations, Karin
Tentschert and Ingo No¨renberg for investigations on
plasma stability, and Carla Destefani for the skillful
technical assistance in binding experiments. Discussions
with and support by Dr. Daniel Vitt were highly
appreciated.
References and notes
1. Neugebauer, V. Expert Rev. Neurother. 2001, 1, 89.
2. Watkins, J. C. Biochem. Soc. Trans. 2000, 28, 297.
3. Bra¨uner-Osborne, H.; Egebjerg, J.; Nielsen, E. O.; Mad-
sen, U.; Krogsgaard-Larsen, P. J. Med. Chem. 2000, 43,
2609.
16. Clark, J. H.; Miller, J. M. Tetrahedron Lett. 1977, 599.
17. Products showed >95% purity as determined by LC–MS
and NMR spectroscopy.
4. Pellicciari, R.; Costantino, G. Curr. Opin. Chem. Biol.
1999, 3, 433.
5. Schoepp, D. D.; Jane, D. E.; Monn, J. A. Neuropharma-
cology 1999, 38, 1431.
6. Conn, P. J.; Pin, J. P. Annu. Rev. Pharmacol. Toxicol.
1997, 37, 205.
7. Hermans, E.; Challiss, R. A. J. Biochem. J. 2001, 359, 465.
18. Functional effects of the compounds were evaluated in a
FLIPR assay on CHO cells transfected with mGlu5
receptors.
19. All compounds tested were completely stable in physio-
logical buffer (PBS) for at least 24 h.