Correlation between brain/plasma ratios and efficacy in neuropathic pain models of selective metabotropic glutamate receptor 1 antagonists

Article abstract of DOI:10.1016/j.bmcl.2006.06.053

We have discovered a novel, potent, and selective triazafluorenone series of metabotropic glutamate receptor 1 (mGluR1) antagonists with efficacy in various rat pain models. Pharmacokinetic and pharmacodynamic profiles of these triazafluorenone analogs re

Full text of DOI:10.1016/j.bmcl.2006.06.053

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4936–4940
Correlation between brain/plasma ratios and efficacy
in neuropathic pain models of selective metabotropic
glutamate receptor 1 antagonists
Guo Zhu Zheng,* Pramila Bhatia, Teodozyj Kolasa, Meena Patel,
Odile F. El Kouhen, Renjie Chang, Marie E. Uchic, Loan Miller, Scott Baker,
Sonya G. Lehto, Prisca Honore, Jill M. Wetter, Kennan C. Marsh,
Robert B. Moreland, Jorge D. Brioni and Andrew O. Stewart
Neuroscience Research, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064-6115, USA
Received 3 May 2006; revised 14 June 2006; accepted 14 June 2006
Available online 30 June 2006
Abstract—We have discovered a novel, potent, and selective triazafluorenone series of metabotropic glutamate receptor 1 (mGluR1)
antagonists with efficacy in various rat pain models. Pharmacokinetic and pharmacodynamic profiles of these triazafluorenone ana-
logs revealed that brain/plasma ratios of these mGluR1 antagonists were important to achieve efficacy in neuropathic pain models.
This correlation could be used to guide our in vivo SAR (structure–activity relationship) modification. For example, compound 4a
has a brain/plasma ratio of 0.34, demonstrating only moderate efficacy in neuropathic pain models. On the other hand, antagonist
4b with a brain/plasma ratio of 2.70 was fully efficacious in neuropathic pain models.
Ó 2006 Elsevier Ltd. All rights reserved.
Glutamate is the most prominent excitatory neurotrans-
mitter in the central nervous system. Glutamate recep-
tors are divided into two major categories: ionotropic
and metabotropic. Ionotropic glutamate receptors are
responsible for fast neurotransmission. Metabotropic
glutamate receptors play a slower, modulatory role
through interactions with a variety of other intracellular
signaling systems. There are three different mGluR
groups, group I, II, and III. A total of eight distinct sub-
types, mGluR1 to mGluR8, based on their primary
sequence similarity, signal transduction linkages, and
pharmacological profile has been identified.1 Group I
mGluRs, mGluR1 and mGluR5, play key roles in the
central sensitization of pain, in addition to a variety of
functions with potential implications in neurological
and psychiatric disorders.2 Glutamate and other excit-
atory amino acids are released from nerve endings in
the periphery under inflammatory conditions. Gluta-
mate or group I mGluR agonists induce hyperalgesia
when administered peripherally.3 MGluRs modulate
pain transmission in the spinal cord, most likely via
sensitization of dorsal horn neurons to sustain
high intensity C-fiber input.1c Normalization of
glutamatergic neurotransmission in the spinal cord and
nociceptive afferents via inhibition of the group I
mGluRs is manifested in the attenuation of pain.4
We are interested in mGluR1 antagonists as a therapy for
the treatment of pain.5 The role of mGluR1 in the treat-
ment of pain has not been validated, due to lack of potent,
selective, and systemically active mGluR1 antagonists.
Earlier efforts to evaluate amino acid antagonists, com-
peting with the glutamate binding site, were not successful
due to poor selectivity, weak antagonism, and lack of
CNS availability.2 Literature reported, non-amino acid-
like mGluR1 antagonists, such as CPCCOEt [7-(hydroxy-
imino)cyclopropa[b]chromen-1a-carboxylate ethyl ester],6
BAY 36-7620,7 R214127,8 dicarboxypyrroles,9 and
JNJ16259685,10 showed limited efficacy in various pain
models, especially in neuropathic pain models.
Keywords: Metabotropic; Glutamate; Receptor; Antagonist; mGluR1;
Neuropathic pain.
Triazafluorenone derivatives (Fig. 1) were identified as
potent group I mGluR antagonists from high-through-
put screening (HTS). Our early SAR of this series has
*
Corresponding author. Tel.: +1 847 935 7513; fax: +1 847 938
0072; e-mail: guozhu.zheng@abbott.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.06.053

G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4936–4940
4937
1
N
N
N
3
N
N
N
N
9
NH2
N
N
a
b
N
O
O
O
O
F
O
N
N
N
7
S
5
O
S
S
N
6
S
5
6
7
hmGluR1 IC = 20 nM
50
Scheme 2. Reagents and conditions: (a) Me2NCH(OMe)2, EtOH,
reflux 16 h, 90%; (b) 4-methyl-3-fluoroaniline, toluene, p-TsOH (cat.),
reflux 16 h, 58%.
Figure 1. Triazafluorenone analog identified from HTS.
been reported.5 The early leads of the series demonstrat-
ed full efficacy in various pain models, except neuro-
pathic pain models. While we thought that CNS
penetration, mainly brain concentration, was important
to predict efficacy of these mGluR1 antagonists in neu-
ropathic pain models, we were puzzled by the fact that
the brain concentration of these antagonists did not
correlate well with their efficacy in neuropathic pain
models. Apparently, brain distribution, not the total
concentration, played a more important role for these
compounds to demonstrate efficacy in neuropathic pain
models. With an ongoing effort to correlate local brain
concentrations of these compounds to the efficacy in
neuropathic pain models, we were also actively looking
for any other indications/parameters to guide our in vivo
SAR study, and to improve in vivo potency of these
compounds in neuropathic pain models. We describe
here that we had observed a correlation between brain/
plasma ratios of these triazafluorenone analogs and
their in vivo efficacy in neuropathic pain models.
3-fluoro-4-methylaniline in the presence of para-toluene-
sulfonic acid to form the final triazafluorenone product
7 (Scheme 2).
Triazafluorenone analog 13 was made from triflate
intermediate 12 (Scheme 3). Triflate 12 was generated
from intermediate 8 as described.13 Compound 9,
formed from 8 in the presence of N,N-dimethylformam-
ide dimethyl acetal, was treated with 4-ethylaniline to
produce tricyclic compound 10. The protecting group
PMB (para-methoxybenzyl) of 10 was then removed
with trifluoroacetic acid (TFA) to give 11.14 The hydroxy
group of 11 was then converted to a triflate group in the
presence of N,N-bis(trifluoromethanesulfonyl)phenyl-
amide.15 The final triazafluorenone analog 13 was
obtained by treating triflate 12 with corresponding
amine (Scheme 3).
The N-oxide analog 14 was obtained by treating 4b with
phthalic anhydride and hydrogen peroxide-urea com-
plex (58% yield) (Scheme 4).16
We have established several synthetic methods to access
these triazafluorenone analogs to accommodate varia-
tions at the different parts of the triazafluorenone phar-
macophore.5 The para-dimethylaminopyridine analogs
4a,b were synthesized from intermediate 2 in 2 steps.
Intermediate 2 was made with known procedures.5,11
The formamidine derivative 3 was obtained in quantita-
tive yield after condensation of intermediate 2 with N,N-
dimethylformamide dimethyl acetal. Formamidine 3
was then treated with primary amines, 4-ethylaniline
or 1-amino-1-homopiperidine, in the presence of para-
toluenesulfonic acid to form the desired triazafluore-
none products 4a and 4b, respectively (Scheme 1).
N
PMBO
PMBO
NH2
N
O
O
a
O
O
S
S
9
N
N
8
b
HO
PMBO
N
N
N
N
c
O
O
S
S
N
TfO
N
N
11
10
d
The meta-dimethylaminopyridine analog 7 was prepared
as described in Scheme 2. Intermediate 5 was made with
known procedures.5,12 The aminoester 5 was then
heated to reflux with N,N-dimethylformamide dimethyl
acetal to generate formamidine intermediate 6 quantita-
tively. The intermediate 6 was then reacted with
N
N
N
N
N
N
e
O
O
S
S
N
13
12
Scheme 3. Reagents and conditions: (a) Me2NCH(OMe)2, reflux 10 h,
100%; (b) p-ethylaniline, toluene, p-TsOH (cat.), microwave 160 °C,
1 h, 39%; (c) TFA, 0 °C, 1.2 h, 100%; (d) Tf2NPh, EtN(i-Pr)2, 80%;
(e) 2-(2-pyridyl)ethylmethyl-amine, 23 °C, 89%.
N
N
N
N
N
N
N
NH2
N
N
R1
a
b
N
O
O
O
O
O-
N+
N
N
Phthalic anhydride,
UHP, CH2Cl2
O
N
S
S
S
N
N
N
N
N
1
2
3
4a: R =4-ethylphenyl
58%
O
1
S
O
S
4b: R =1-homopiperidinyl
N
4b
14
Scheme 1. Reagents and conditions: (a) Me2NCH(OMe)2, reflux, 5 h
(99%); (b) R1NH2, toluene, p-TsOH, 4a (73%), 4b (45%).
Scheme 4.

4938
G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4936–4940
We have thus established several synthetic routes to
accommodate variations of these triazafluorenone ana-
logs for SAR studies, and to modify their ADME pro-
files. These triazafluorenone analogs are non-amino
acid-like and non-competitive mGluR1 antagonists
which bind at 7-TMD region of the receptor.6 Com-
pound 4a was a selective mGluR1 antagonist (mGluR1
IC50 = 3 0.9 nM). It was also active in mGluR5
(IC50 = 442 93 nM) with a ratio of mGluR5/mGluR1
around 147. Compound 4a was inactive in mGluR2,
mGluR4, and mGluR7 (IC50 > 10 lM). MGluR1 antag-
onist 4a was able to achieve full efficacy in various
animal pain models, including CFA (complete Freund’s
adjuvant-induced thermal hyperalgesia, ED50 = 15
lmol/kg, ip), Carrageenan (ED50 = 11 lmol/kg, ip),
formalin (ED50 = 19 lmol/kg, ip), skin incision
(ED50 = 47 lmol/kg, ip), and OA (osteoarthritic pain,
ED50 = 29 lmol/kg, ip). Pharmacokinetics analysis
revealed that compound 4a had moderate bioavailability
(Fip = 45%, Foral = 12%). There was no locomotor
side effect caused by this mGluR1 antagonist
(ED50 > 100 lmol/kg, ip). This compound was also
capable to penetrate BBB (blood–brain barrier) with a
brain/plasma ratio of 0.34 (30 lmol/kg, ip) (Table 1).
of neuropathic pain. We thought that total brain con-
centration might be able to correlate efficacy in neuro-
pathic pain models of these compounds, and could be
used to guide our in vivo SAR study. We quickly real-
ized that this was not case for these mGluR1 antago-
nists. Compound 7 had much lower total brain
concentration (total brain concentration: 125 41 ng/
g at 30 lmol/kg, ip; protein binding: 99.9%, free brain
concentration: 0.13 0.04 ng/g) than 4a (total brain
concentration: 1180 258 ng/g at 30 lmol/kg, ip;
97.8% protein binding, free brain concentration:
25.5 5.7 ng/g). Yet this compound 7 was more effica-
cious (Chung ED50 = 55 lmol/kg, ip) than the latter
antagonist 4a (Chung ED50 > 100 lmol/kg, ip) in neu-
ropathic pain models. Apparently, neither total brain
concentration, nor free brain concentration of these
compounds correlated well their in vivo efficacy in
neuropathic pain models. The assumption made from
these results was that the local distribution of mGluR1
antagonist in the brain might play a more crucial role
in the neuropathic pain models.
It is certainly a technical challenge to first identify the
actual location of the action site in the brain, and
then determine the local concentration of these
mGluR1 antagonists. We were looking for some type
of parameter(s) we could use to guide our in vivo
SAR study to improve efficacy in neuropathic pain
models. We did find that there was a correlation be-
tween brain/plasma ratio of these mGluR1 antagonists
and their efficacy in neuropathic pain models. We
used this correlation to guide our in vivo SAR study
and identified compound with full efficacy in neuro-
pathic pain models.
While we were encouraged by the efficacy demonstrated
by compound 4a in the above-mentioned pain models,
we were puzzled by its weak and partial efficacy in
neuropathic pain models, such as Chung (Spinal Nerve
(L5/L6) Ligation Model, ED50 > 100 lmol/kg, ip).17
The site of action for this target is not very clear. It
was known that mGluR1 has expression and distribu-
tion in both peripheral and CNS system.18 Compound
4a, with distribution in both peripheral and central
compartments, achieved full efficacy in various animal
pain models listed in Table 1, but not the Chung model
Compound 13, made according to Scheme 3 in an effort
to modify solubility and metabolic profiles of the dim-
ethylamino group in parent analog 4a, was a relatively
more polar compound with its 9-pyridylethylmethylami-
no group. This compound 13 had a lower brain/plasma
ratio (0.16) than its parent compound 4a (brain/plasma
ratio = 0.34), and was inactive in Chung model (11% at
100 lmol/kg, ip Table 2).
Table 1. Selectivity, efficacy, and side effects of triazafluorenone
mGluRl antagonist 4a
mGluRl
(nM)
mGluR2 mGluR4
(nM) (nM)
mGluR5 mGluR7
(nM) (nM)
In vitroa
In vivob
3 ( 0.9)
>10,000 >10,000
442 ( 93) >10,000
Incorporating a dimethylamino N-oxide group at the C9
position of the triazafluorenone pharmacophore gener-
ated a very polar mGluR1 antagonist 14. This N-oxide
analog was still quite potent and selective mGluR1
antagonist (IC50 = 33 17 nM). Compound 14 had a
very low brain/plasma ratio (0.05). Again, this com-
pound showed no efficacy in Chung model (12% at
100 lmol/kg, ip Table 2).
Model (rat)
ED50
(lmol/kg, ip)
CFA
15
Carrageenan
Formalin
Skin incision
OA
11
19
47
29
Chung
>100
It seemed clear that mGluR1 antagonists with high
brain/plasma ratio might be advantageous in terms of
their in vivo efficacy in neuropathic pain models. Com-
pound 7, as discussed earlier, had very low total brain
concentration compared to the parent compound 4a.
Yet this compound 7 was more efficacious (72% at
100 lmol/kg, ip) than 4a (41% at 100 lmol/kg, ip).
Compound 7 did have a higher brain/plasma ratio of
0.51, compared to compound 4a (brain/plasma
ratio = 0.34).
Model (rat)
ED50
(lmol/kg, ip)
Side effectsb Locomotor
Rotorod
>100
>300
a 1321N1 cells expressing human mGluRs, mean of multiple results
with standard error of mean.
b Tests performed 30 min after intraperitoneal administration of
compound in rats (6 rats per group). Vehicle was 10% DMSO/PEG
(5 mL/kg).

G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4936–4940
4939
Table 2. Correlation between brain/plasma ratios and efficacy in neuropathic pain model of triazafluorenone mGluRl antagonists
Structure
mGluR1a
IC50 (nM)
mGluR5a
IC50 (nM)
clogPb
ratioc
(brain/plasma)
Chungd % effect
at 100 lmol/kg
-
O
+
N
N
N
N
33 ( 17)
>10,000
1.26
0.05 ( 0.02)
12%
S
O
N
14
N
N
N
N
52 ( 16)
>10,000
4.84
0.16 ( 0.06)
11%
O
S
N
13
N
N
N
N
5 ( 3)
1330 ( 490)
4.24
0.34 ( 0.08)
41%e
O
S
4a
N
N
N
O
21 ( 15)
>10,000
>10,000
3.85
2.80
0.51 ( 0.1)
72%f
F
N
S
7
N
N
N
N
1 ( 05)
2.70 ( 0.4)
100%f
O
S
N
4b
a 1321N1 cells expressing either rat mGluRl or mGluR5, mean of multiple results with standard error of mean.
b Calculated from SPARK ACD program.
c Determined from the corresponding in vivo experiments at 30 llmol/kg ip (see Footnote d). Mean of multiple results with standard error of mean.
d Tests performed 30 min after intraperitoneal administration of compound in rats (6 rats per group). Vehicle was 10% DMSO/PEG (5 mL/kg).
e P < 0.05 versus vehicle group.
f P < 0.01 versus vehicle group.
Compound 4b, a potent and selective mGluR1 antago-
nist (IC50 = 1 0.5 nM), was able to penetrate BBB
better than compound 7, based on their brain/plasma
ratios (2.70 and 0.51, respectively). With the highest
brain/plasma ratio of 2.70 (total brain concentra-
tion = 851 ng/g at 30 lmol/kg, ip) among this series of
mGluR1 antagonists, compound 4b now achieved full
efficacy at 100 lmol/kg (ED50 = 22 lmol/kg, ip) in the
Chung model of neuropathic pain.
brain/plasma ratio (2.70), it demonstrated full efficacy
in neuropathic pain model.
References and notes
1. (a) Nakanishi, S.; Masu, M. Annu. Rev. Biophys. Biomol.
Struct. 1994, 23, 319; (b) Bockaert, J.; Pin, J.-P. EMBO J.
1999, 18, 1723; (c) Hermans, E.; Challiss, R. A. J.
Biochem. J. 2001, 359, 465.
Combined with clogP values of these compounds
(Table 2), the use of brain/plasma ratio to guide our
SAR study is now even more practical.
2. Schoepp, D. D.; Jane, D. E.; Monn, J. A. Neuropharma-
cology 1999, 38, 1431.
3. (a) Fisher, K.; Codere, T. J. Neuroreport 1996, 7, 2743; (b)
Fisher, K.; Coderre, T. J. Neuroreport 1998, 9, 1169.
4. Walker, K.; Bowes, M.; Panesar, M.; Davis, A.; Gentry,
C.; Kesingland, A.; Gasparini, F.; Spooren, F.; Stoehr, N.;
Pagano, A.; Flor, P. J.; Vranesic, I.; Lingenhoehl, K.;
Johnson, E. C.; Varney, M.; Urban, L.; Kuhn, R.
Neuropharmacology 2001, 40, 1.
5. Zheng, G. Z.; Bhatia, P.; Daanen, J.; Kolasa, T.; Patel,
M.; Latshaw, S.; El Kouhen, O. F.; Chang, R.; Uchic, M.
E.; Miller, L.; Nakane, M.; Lehto, S. G.; Honore, M. P.;
In summary, we have observed a correlation between
brain/plasma ratios and in vivo efficacy in neuropathic
pain model of our triazafluorenone mGluR1 antago-
nists. Potent and selective mGluR1 antagonists with
low brain/plasma ratios demonstrated little or no efficacy
in neuropathic pain model. When an mGluR1 antago-
nist, such as compound 4b, achieved a favorable high

4940
G. Z. Zheng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4936–4940
Moreland, R. B.; Jorge, D.; Brioni, J. D.; Stewart, A. O.
J. Med. Chem. 2005, 48, 7374.
6. Litschig, S.; Gasparini, F.; Rueegg, D.; Stoehr, N.; Flor,
P. J.; Vranesic, I.; Prezeau, L.; Pin, J. P.; Thomsen, C.;
Kuhn, R. Mol. Pharmacol. 1999, 55, 453.
7. Carroll, F. Y.; Stolle, A.; Beart, P. M.; Voerste, A.; Brabet,
I.; Mauler, F.; Joly, C.; Antonicek, H.; Bockaert, J.; Muller,
T.; Pin, J. P.; Prezeau, L. Mol. Pharmacol. 2001, 59, 965.
8. Lavreysen, H.; Janssen, C.; Bischoff, F.; Langlois, X.;
Leysen, J. E.; Lesage, A. S. Mol. Pharmacol. 2003, 63, 1082.
9. Micheli, F.; Fabio, R. D.; Cavanni, P.; Rimland, J. M.;
Capelli, A. M.; Chiamulera, C.; Corsi, M.; Corti, C.;
Donati, D.; Feriani, A.; Ferraguti, F.; Maffeis, M.; Missio,
A.; Ratti, E.; Paio, A.; Pachera, R.; Quartaroli, M.;
Reggiani, A.; Sabbatini, F. M.; Trist, D. G.; Ugolini, A.;
Vitulli, G. Bioorg. Med. Chem. 2003, 11, 171.
Med. Chem. 1986, 29, 1046; (c) Mittelbach, M.; Kastner,
G.; Junek, H. Arch. Pharm. 1985, 318, 481; (d) Mekhei-
mer, R. A.; Ahmed, E. K.; Ei-Fahham, H. A.; Kamel, L.
H. Synthesis 2001, 1, 97.
12. (a) Bertini, V.; Lucchesini, F.; Pocci, M.; Munno, A. D.
Heterocycles 1995, 41, 675; (b) Hutchins, R. O. J.
Heterocycl. Chem. 1985, 22, 401.
13. (a) Nicholl, L.; Tarsio, P. J.; Point, S.; Blohm, H. US
Patent 2,824,121, 1958; (b) Cywin, C. L.; Chen, Z.
Emeigh, J.; Fleck, R. W.; Hao, M. H.; Hickey, E.; Liu,
W.; Marshall, D. R.; Morwick, T.; Nemoto, P.; Sorcek,
R. J.; Sun, S.; Wu, J. P. WO Patent 03/103661 A1,
2003.
14. Smith, A. B.; Rano, T. A.; Chida, N.; Sulikowski, G. A.;
Wood, J. J. Am. Chem. Soc. 1992, 114, 8008.
15. Hendrickson, J. B.; Bair, K. W. J. Org. Chem. 1977, 42,
3875.
16. Kaczmarek, L.; Balicki, R.; Nantka-Namirski, P. Chem.
Ber. 1992, 125, 1965.
10. Lavreysen, H.; Wouters, R.; Bischoff, F.; Pereira, S. N.;
Langlois, X.; Blokland, S.; Somers, M.; Dillen, L.; Lesage,
A. S. J. Neuropharmacology 2004, 47, 961.
11. (a) Fisher, H. GE Patent 834104, 1949; (b) Gate, E. N.;
Threadgill, M. D.; Stevens, M. F. G.; Chubb, D.; Vickers,
J. M.; Langdon, S. P.; Hickman, J. A.; Gescher, A. J.
17. Kim, S. H.; Chung, J. M. Pain 1992, 50, 355.
18. Lavreysen, H.; Pereira, S. N.; Leysen, J. E.; Langlois, X.;
Lesage, A. S. J. Neuropharmacology 2004, 46, 609.