3-Benzyl-1,3-oxazolidin-2-ones as mGluR2 positive allosteric modulators: Hit-to lead and lead optimization

Article abstract of DOI:10.1016/j.bmcl.2009.03.032

The discovery, synthesis and SAR of a novel series of 3-benzyl-1,3-oxazolidin-2-ones as positive allosteric modulators (PAMs) of mGluR2 is described. Expedient hit-to-lead work on a single HTS hit led to the identification of a ligand-efficient and struct

Full text of DOI:10.1016/j.bmcl.2009.03.032

Bioorganic & Medicinal Chemistry Letters 19 (2009) 2524–2529
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
3-Benzyl-1,3-oxazolidin-2-ones as mGluR2positive allosteric modulators: Hit-to
lead and lead optimization
b,
b
d
c
d
Allen J. Duplantier a, , Ivan Efremov , John Candler , Angela C. Doran , Alan H. Ganong , Jessica A. Haas ,
Ashley N. Hanks c, Kenneth G. Kraus a, John T. Lazzaro Jr. c, Jiemin Lu a, Noha Maklad a, Sheryl A. McCarthy c,
Theresa J. O’Sullivan b, Bruce N. Rogers a, Judith A. Siuciak c, , Douglas K. Spracklin d, Lei Zhang a
*
*
a CNS Chemistry, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
b Lead Generation Group, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
c Department of Neuroscience, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
d Department of Pharmacokinetics and Drug Metabolism, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery, synthesis and SAR of a novel series of 3-benzyl-1,3-oxazolidin-2-ones as positive allosteric
modulators (PAMs) of mGluR2 is described. Expedient hit-to-lead work on a single HTS hit led to the
identification of a ligand-efficient and structurally attractive series of mGluR2 PAMs. Human microsomal
clearance and suboptimal physicochemical properties of the initial lead were improved to give potent,
metabolically stable and orally available mGluR2 PAMs.
Received 6 January 2009
Revised 9 March 2009
Accepted 10 March 2009
Available online 14 March 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
mGluR2
Positive allosteric modulator
Carbamate
Schizophrenia
Metabotropic glutamate type 2 receptors (mGluR2) are G-pro-
tein coupled receptors1 that are localized presynaptically on gluta-
mate neurons and modulate neurotransmitter release.2 Known
mGluR2 agonists also have agonist activity at mGluR3 receptors
which are the closest mGluR family member with a high degree
of homology at the agonist binding site.1,3 Mixed mGluR2/3 ago-
nists have been shown to be effective in a variety of preclinical ani-
mal models of anxiety4,5 and schizophrenia,5–7 and recently,
LY2140023 was reportedly active in a phase IIb clinical trial against
the positive and negative symptoms of schizophrenia.8 Positive
allosteric modulators (PAMs) of mGluR2 have been reported by
others in the literature9 and most recently by our group.10 PAMs
may offer an advantage over agonists of mGluR2 as potentially
selective tools to gain a clear understanding of the mGluR2 sub-
type, and allow for chemotype diversity without the structural
restrictions imposed by the agonist binding pocket.11 Additional
advantages of PAMs may include less receptor desensitization,12
and upregulation of receptor-mediated effects only when the
receptor is stimulated by either naturally released or exogenous
agonist. This latter point, however, may complicate the interpreta-
tion of in vivo efficacy data of mGluR2 PAMs due to an inability to
compare the glutamate tone at the synapse of animal models with
that in patients. Herein, we describe the identification, SAR devel-
opment and optimization of a novel class of mGluR2 modulators
identified by high-throughput screening (HTS) of our compound
file.
Choosing an appropriate starting point from a variety of hits
identified by HTS campaigns is a critical stage in the drug discovery
process. Not surprisingly, the pharmaceutical industry at large has
paid significant attention to the development of a more efficient
and robust hit-to-lead process.13 Two critical stages in this process
are the hit assessment (identity and sample integrity confirmation
along with physicochemical properties assessment) and hit valida-
tion (re-assay and possibly resynthesis) steps.
Our approach to identification of mGluR2 PAM lead matter
involved an HTS campaign and some of the results from that effort
have recently been reported.10 The triage of hit structures drew our
attention to a unique chemotype represented by structure 1
(Fig. 1), an in-house compound that was structurally distinct from
known mGluR2 PAMs.9,10
Calculated properties of 1 (MW, logP, solubility, polar surface
area) coupled with synthetic enablement (Fig. 1) looked promising,
but the analytical evaluation of the original sample by 1H NMR and
* Corresponding authors. Tel.: +1 860 441 6009 (A.J.D.); tel.: +1 860 441 686 1373
(I.E.).
E-mail addresses: allen.j.duplantier@pfizer.com (A.J. Duplantier), ivan.efre-
mov@pfizer.com (I. Efremov).
Present address: The Biomarkers Consortium, Foundation for the NIH, Bethesda,
MD, USA.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.03.032

A. J. Duplantier et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2524–2529
2525
R1
N
O
OH
a, b, c
O
O
O
O
H
n = 0 or 1
( )n
4
5
R
O
NH2
Scheme 2. Reagents and conditions: (a) R1OH, Ph3P, DBAD, toluene, rt; (b)
aminoalcohol, NaBH(OAc)3, 1,2-dichloroethane, rt; (c) carbonyl diimidazole, 1,2-
dichloroethane, 80 °C, 2 h.
1
Figure 1. Presumed structure of the HTS hit
1
and potential synthetic
disconnections.
parallel synthesis. The key questions to answer were SAR respon-
siveness and an ability to increase potency while remaining in the
chemical space of CNS drugs.18 With this in mind, the key criteria
used for library design were systematic structural changes in the
scaffold, molecular weight, lipophilicity, polar surface area, and a
score from a proprietary in silico model assessing potential for
P-gp efflux. To methodically evaluate these properties from a struc-
tural perspective, our main parameters for the selection of library
precursors became: size, degree of saturation and polarity of ‘head’
substituents(R1 inScheme2);linkerlengthinR1; ringsizeofthe‘tail’
(carbamate) substituents; polarity and placement of additional sub-
stitution on the cyclic carbamate ring. Of particular note was an
attempt to introduce polarity in both head and tail parts of the scaf-
fold to balance the lipophilic nature of the hit structure 3. The library
was produced with 92% success rate and provided timely SAR infor-
mation used for decision making in this mGluR2 hit series. Some
observations are highlighted in Table 1.
The prepared set of analogs allowed for a confident assessment
of the lead potential for this series and highlighted key areas for
improvement. It was observed that the SAR was additive between
substituents.19 From a potency perspective, we found the SAR to be
very responsive. The optimal linker between ether oxygen and the
ring was a single-carbon chain, preferably branched (compounds 6,
7, 10 and 11, Table 1). Ring size of the carbamate substituent had
an interesting effect on activity — five-membered carbamate with
substitution next to oxygen was optimal with unsubstituted five-
and six-membered carbamates being inferior (compounds 3, 7
and 14). Substitution next to nitrogen in the five-membered carba-
mates was clearly detrimental and, while some substitution on the
six-membered carbamates was tolerated, it still resulted in loss of
potency. Compound 13 illustrates one possible avenue for polarity
modulation. It is worth noting that a corresponding N-methyl
HO
O
O
a, b
HN
O
O
O
O
N
OBu-t
O
N
O
2
3
Scheme 1. Proposed rationale for formation of compound 3 responsible for mGluR2
activity in the HTS hit. Reagents: (a) NaH, DMF; (b) (para-cyclohexyl-methyl-
oxy)benzyl chloride, DMF.
LC–MS revealed the presence of only trace amounts of 1. Indeed,
resynthesis and evaluation of the authentic compound 1 confirmed
that no desired pharmacological activity resided in this chemical
structure. Additional purification of the active batch followed by
a battery of analytical methods (1H and 13C NMR, HRMS, UV, IR),
allowed us to arrive at structure 3 as the active component, and
this structure was confirmed by X-ray crystallography (data not
shown). This result prompted us to investigate in detail the syn-
thetic route used for the original synthesis of compound 1. The
likely explanation for the formation of 3 appears to be an inadver-
tent use of an excess of base during deprotonation of the Boc-pro-
tected amino alcohol 2 (Scheme 1). The resulting dianion would
cause N-alkylation of 2 with para-(cyclohexyl methyloxy)benzyl
chloride rather than the intended O-alkylation. The subsequent
spontaneous cyclization of the intermediate would then lead to
the observed compound 3. We were pleased to see that upon reas-
say of the purified sample of 3, potent mGluR2 PAM activity
(EC50 = 117 nM) was observed.
At this stage our attention shifted to the assessment of the
structure 3 as a starting point for lead identification. The calculated
and measured in vitro properties of 3 (cLogP, polar surface area,
MDCK,14 MDR15) were predictive of good brain penetration, and
its potency and low molecular weight suggested good ligand effi-
ciency.16,17 Among a variety of synthetic routes that could be used
for synthesis of this chemotype, we chose an alkylation/reductive
amination sequence followed by carbamate formation as repre-
sented in Scheme 2. The advantages of this route included amena-
bility toward parallel synthesis coupled with ready availability of
diverse precursors, thus enabling efficient SAR development.
While a virtual library comprising several thousands of products
could be enumerated using available starting materials, our goal was
to reach decision making on this chemotype in one iteration of
piperidine derivative was not active at 10 lM (data not shown).
As shown by the two diastereomeric isomers of 6 (compounds 8
and 17), greater desired activity resided within the R-configuration
of the stereogenic center on the carbamate ring.
Compound 17 (Fig. 2) displayed in vitro ADME properties predic-
tive of good brain penetration (MDCK AB >10 Â 10À6 cm/s and MDR
BA/AB <2.5),20 and this was confirmed in vivo in the mouse [brain/
plasma ratio = 3.0 (32 mg/kg, s.c.)]. In addition, 17 showed no signif-
icant CYP inhibition21 and dofetilide displacement assay data22 sug-
gested a low potential for hERG channel inhibition. However,
although 17 was a potent mGluR2 PAM (EC50 = 5 nM) with good
brain penetration, it also had high in vitro human liver microsomal
clearance (14.7 mL/min/kg), a high cLog P (5.12) and a melting point
(MP) below room temperature making it difficult to manufacture on
a kg scale. Our strategy to address these issues was to introduce an
ionizable center and/or increase the polarity, and to replace the alk-
ylether moiety with metabolically stable groups such as substituted
aryls while retaining low molecular weight (<400).
Amino analogs 20 and 21 were prepared by reductive amination
of aldehydes 18a and 18b, respectively, with (R)-(À)-1-amino-2-

2526
A. J. Duplantier et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2524–2529
Table 1
Functional activities and human microsomal clearance data of representative mGluR2 PAMs (general structure 5)
Compound
R1
carbamate
mGluR2 EC50 (nM)a
h-CLh (mL/min/kg)b
O
N
6c
7.9
16
O
O
O
O
N
7
8
9
3
15
22.1
75
9.9
13.6
10
O
N
O
N
O
O
O
N
117
9.1
O
N
10
11
203
236
<5.3
14
O
O
O
N
O
N
12
13
14
15
262
384
590
<5.3
14.5
<5.3
O
O
O
N
O
O
N
O
O
O
N
O
N
>10,000
>10,000
<5.3
O
N
O
16
NDd
a
EC50 values obtained from mGluR2 receptor transfected HEK cell line using FLIPR method in the presence of glutamate (EC10-20); means of at least three experiments; an
EC50 value >10,000 indicates that no curve was noted in the dose–response up to 10
lM.
b
Predicted hepatic clearance (CLh) from human liver microsomal stability assay.
c
Mixture of diastereomeric pairs.

A. J. Duplantier et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2524–2529
2527
O
mGluR2 EC50 = 5.0 nM
MW = 317.43
X
Y
z
cLogP = 5.12
c
MDCK AB = 12.3 x 10-6 cm/s
MDR BA/AB = 2.17
O
Y
z
X
HLM CL = 14.7 mL/min/kg
CYP1A2: 0% inhibition at 3 mM
CYP2C9: 27% inhibition at 3 mM
CYP2D6: 0% inhibition at 3 mM
CYP3A4: 0% inhibition at 3 mM
dofetilide: 3% inhibition at 10 mM
a, b
O
O
Y
z
N
O
O
O
d
N
24
N
O
H
O
23a-c
Ar
17
Y
z
22a: X = Br, Y = CH, Z = CH
22b: X = Cl, Y = N, Z = CH
22c: X = Br, Y = CH, Z = N
Figure 2. In vitro ADME profiling of a representative hit follow-up structure.
O
N
25-47
propanol followed by carbamate cyclization with carbonyl diimi-
dazole, subsequent deprotection of the dioxolane ring (to give
19a and b), and then reductive amination with either cyclohexyl
methylamine or piperidine (Scheme 3). The cyclic carbamate moi-
ety of intermediates 23a–c were prepared in a similar manner
from aldehydes 22a–c as described in Scheme 4. Pyridyl analog
24 was synthesized by arylation of 1-cyclohexylethanol with
23b. Suzuki coupling of the appropriate arylboronic acid and aryl-
halide (23a–c) provided biaryl analogs 25–47, as well as aldehyde
intermediates 48a and b (see Table 2 for definitions of Ar, Y and Z).
Reductive amination of 48a and b with dimethylamine afforded
amino analogs 49a and b, respectively (Scheme 5).
Efforts to address the high CL and low MP entailed the introduc-
tion of a basic amine into the cycloalkyl moiety of 17. To that end,
amino compounds 20 and 21 reduced cLogP and in vitro metabolic
clearance, and provided a salt handle. However, the introduction of
polar functionality into this part of the molecule proved to be detri-
mental to mGluR2 activity (see Table 2). Substitution of the central
phenyl ring of compound 17 with a pyridyl group (24) had a less dra-
matic effect on potency (EC50 = 88 nM), but did not improve CL.
As an alternative strategy to reduce CL and increase the MP, we
replaced the alkylether moiety of 17 with phenyl23 (25). Not
surprisingly, 25 was also metabolically labile, presumably due to
oxidation around the phenyl rings, so we attempted to block metab-
olism by systematically substituting F, Cl and OMe groups around
the terminal phenyl ring (compounds 26–34, Table 2). The interest-
ing SAR that emerged from this exercise was the discovery that
O
Scheme 4. Reagents and conditions: (a) (R)-(À)-1-amino-2-propanol, Na(OAc)3BH,
CH2Cl2, 18 h, rt; (b) carbonyl diimidazole, THF, 80 °C, 4 h; (c) 1-cyclohexylethanol,
potassium tert-butoxide, THF, 120 °C (microwave), 10 min, 20%; (d) ArB(OH)2,
Pd(PPh3)4, Na2CO3, EtOH, 120 °C (microwave), 5 min.
ortho-substitution (26, 29 and 32) improved potency and para-sub-
stitution (28 and 31) lowered clearance. The high CL observed with
the4-methoxy analog34was likelydue to metabolic demethylation,
as the corresponding ethoxy (36) and trifluoromethoxy (37) analogs
were stable in the presence of human liver microsomes. The most
interesting compounds identified were the di-substituted analogs
38–40, where the substituent at the 4-position blocked metabolism
and the second substituent appeared to improve potency. Com-
pounds 38–40 were obtained in solid form, but their low MPs
(<90 °C) and low aqueous solubility were difficult to address due
to the lack of an ionizable center from which to prepare alternative
salt forms. To that end, attempts to introduce heterocyclic (41–47)
and amino (49a and b) functionality led to metabolically stable ana-
logs with improved solubility and lower cLogP (<4), but once again
with reduced mGluR2 potency.
Selected compounds were profiled further for selectivity against
a panel of CNS receptors (data not shown), and compound 40 was
found to have moderate 5-HT2A activity (Ki = 323 nM).
Interestingly, only biaryl analogs with fluoro-substitution had
N
HN
O
R
O
R
O
H
O
20
N
a, b, c
d
O
O
O
N
O
N
HN
18a: R = H
18b: R = Me
19a, b
O
21
N
O
Scheme 3. Reagents and conditions: (a) (R)-(À)-1-amino-2-propanol, Na(OAc)3BH, CH2Cl2, 18 h, rt; (b) carbonyl diimidazole, THF, 80 °C, 4 h; (c) sulfuric acid (cat.), acetone,
reflux, 32%; (d) Na(OAc)3BH, CH2Cl2, 18 h, rt.

2528
A. J. Duplantier et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2524–2529
Table 2
Functional activities, in vitro human microsomal clearance data and cLogP values of representative mGluR2 PAMs
cLogPd
a
b
Entry
Y
Z
Ar
mGluR2 EC50 (nM)
h-CLh (mL/min/kg)
20
21
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
49a
49b
—
—
N
—
—
—
—
—
Phenyl
>10,000
>10,000
88
142
35
280
84
28
158
76
41
275
53
26
97
183
30
32
70
4790
7550
4380
380
595
3400
1800
3060
6360
8.1
4.11
3.58
4.52
4.13
4.27
4.27
4.27
4.59
4.84
4.84
3.49
4.05
4.05
4.02
4.58
5.16
4.98
5.44
4.42
2.63
2.63
1.68
4.19
3.72
3.15
2.94
3.96
3.66
<5.3
15.4
>18.7
>18.7
18.1
5.7
>18.7
NDc
8.4
>18.7
>18.7
18.6
>18.7
<5.3
<5.3
8.3
<5.3
7.3
13.2
<5.3
ND
<5.3
<5.3
<5.3
<5.3
ND
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
N
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
N
2-Fluorophenyl
3-Fluorophenyl
4-Fluorophenyl
2-Chlorophenyl
3-Chlorophenyl
4-Chlorophenyl
2-Methoxyphenyl
3-Methoxyphenyl
4-Methoxyphenyl
2-Ethoxyphenyl
4-Ethoxyphenyl
4-Trifluoro-methoxyphenyl
3-Chloro-4-fluorophenyl
2-Fluoro-4-trifluoro-methoxyphenyl
2,4-Difluorophenyl
4-Pyridyl
3-Pyridyl
Pyrimidin-5-yl
2-Fluoro-4-trifluoro-methoxyphenyl
3-Chloro-4-fluorophenyl
2,4-Difluorophenyl
2,4-Difluorophenyl
—
N
N
CH
—
—
—
—
—
9.7
a
EC50 values obtained from rat mGluR2 receptor transfected HEK cell line using FLIPR method in the presence of glutamate (EC10–20); means of at least three experiments.
Predicted hepatic clearance from human liver microsomal stability assay.
ND, not determined.
b
c
d
In silico calculation of logP.
R1
Table 3
Human 5-HT2A binding activity
Compounds
h 5-HT2A (Ki, nM)
Compounds
h 5-HT2A (Ki, nM)
a
R2
17
38
39
>3200
2100
>3100
40
44
46
323
>3100
>3100
49a: R1 = CH2NMe2, R2 = H
49b: R1 = H, R2 = CH2NMe2
O
O
containing a pyridyl nitrogen in place of phenyl (44 and 46) had
much weaker 5-HT2A binding activity (Table 3).
N
48a: R1 = CHO, R2 = H
48b: R1 = H, R2 = CHO
Compounds 38, 40 and 44, as well as the mGluR2 agonist,
LY3547404, were evaluated in an anti-psychosis model, testing a
compound’s ability to attenuate methamphetamine-induced
hyperlocomotor activity in mice.24 The mGluR2 agonist showed
significant activity at 0.01 mg/kg, s.c., whereas the mGluR2 PAMs,
38 and 40, required a minimal effective dose (MED) of 10 mg/kg,
s.c. At the MED dose, compounds 38 and 40 had free plasma expo-
sures of 33 and 211 nM, respectively, and brain/plasma ratios >2
(see Table 4). To determine the amount of free drug in the brain
Scheme 5. Reagents and conditions: (a) Dimethylamine, Na(OAc)3BH, CH2Cl2, 18 h,
rt.
sub-micromolar activity against 5-HT2A, as compounds containing
any larger group such as chloro (38) or trifluoromethoxy (39), or
Table 4
Methamphetamine induced locomotor activity in micea
c
d
Compounds
MEDb (mg/kg, sc)
[Brain] (ng/g)
[Plasma] (ng/mL)
B/P
Fup; Fub
—
[Plasma]free; [brain]free (nM)
LY354740
0.01
10
10
—
—
—
—
38
40
44
2750
2230
10,090
750
1070
3550
3.7
2.1
2.8
0.014; 0.004
0.06; 0.009
0.06; 0.010
33; 34
211; 66
575; 267
>32
a
Concentrations reported at MED except for 44 (17.8 mg/kg).
Minimal effective dose; compounds were formulated in 5:5:90 DMSO/cremophor/saline; animals sacrificed at 2 h post dose except for 40 (30 min post dose); brain and
b
plasma levels analyzed.
c
Fraction unbound in mouse plasma and brain.
[Plasma]free = [plasma] Â Fup; [brain]free = [brain] Â Fub.
d

A. J. Duplantier et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2524–2529
2529
Table 5
Pharmacokinetic data in rats.
Schmidt for 5-HT2A data, Elaine Tseng for exposure data for com-
pound 40, Dianne Wong for generating B/P data for compound
17, and our ADME technology group for microsomal clearance
data.
Compounds
CLp (mL/min/kg)a
102
18.1
Vdss (L/kg)
T1/2 (h)
%Fb
38
44
3.2
4.7
0.8
3.8
64
100
References and notes
a
1 mg/kg iv dose.
F based on oral dose of 10 mg/kg 38 and 3 mg/kg 44.
b
1. Pin, J.-P.; Galvez, T.; Prezeau, L. Pharmacol. Ther. 2003, 98, 325.
2. Poisik, O.; Raju, D. V.; Verreault, M.; Rodriguez, A.; Abeniyi, O. A.; Conn, P. J.;
Smith, Y. Neuropharmacology 2005, 49, 57.
3. Kew, J. N. C.; Kemp, J. A. Psychopharmacology 2005, 179, 4.
4. Grillon, C.; Cordova, J.; Levine, L. R. Psychopharmacology 2003, 168, 446.
5. Rorick-Kehn, L. M.; Johnson, B. G.; Knitowski, K. M.; Salhoff, C. R.; Witkin, J. M.;
Perry, K. W.; Griffey, K. I.; Tizzano, J. P.; Monn, J. A.; McKinzie, D. L.; Schoepp, D.
D. Psychopharmacology 2007, 193, 121.
6. Marek, G. J. Curr. Opin. Pharmacol. 2004, 4, 18.
7. Linden, A.-M.; Schoepp, D. D. Drug Discovery Today: Therapeutic Strategies 2006,
3(4), 507.
8. Patil, S. T.; Zhang, L.; Martenyi, F.; Lowe, S. L.; Jackson, K. A.; Andreev, B. V.;
Avedisova, A. S.; Bardenstein, L. M.; Gurovich, I. Y.; Morozova, M. A.; Mosolov, S.
N.; Neznanov, N. G.; Reznik, A. M.; Smulevich, A. B.; Tochilov, V. A.; Johnson, B.
G.; Monn, J. A.; Schoepp, D. D. Nat. Med. 2007, 13, 1102.
9. For a review, see: Rudd, M. T.; McCauley, J. A. Curr. Top. Med. Chem. 2005, 5, 869.
10. Zhang, L.; Rogers, B. N.; Duplantier, A. J.; McHardy, S. F.; Efremov, I.; Berke, H.;
Qian, W.; Zhang, A. Q.; Maklad, N.; Candler, J.; Doran, A. C.; Lazzaro, J. T., Jr.;
Ganong, A. H. Bioorg. Med. Chem. Lett. 2008, 18, 5493.
([brain]free), the unbound fractions in mouse brain (Fub) were mea-
sured. The [brain]free of 38 (2 h post dose) and 40 (0.5 h post dose)
at the MED were calculated to be 34 and 66 nM, respectively, sim-
ilar to their in vitro EC50 values (30 and 70 nM, respectively). How-
ever, for compound 40 it is unclear how much of its activity in the
locomotor assay is due to 5-HT2A, or whether 5-HT2A activity could
act synergistically with mGluR2 in this model.25 Compound 44 was
not active at doses up to 32 mg/kg, likely reflecting the larger hur-
dle of covering a less potent EC50 of 380 nM. Finally, compounds 38
and 44 were both shown to have good oral bioavailability in rats
(Table 5); and 44 had low plasma CL and a moderate half-life
(T1/2 = 3.8 h). Compound 38, however, appeared to have high clear-
ance (CL = 102 mL/min/kg, undefined explanation), suggestive of
11. Recasens, M.; Guiramand, J.; Aimar, R.; Abdulkarium, A.; Barbanel, G. Curr. Drug
Targets 2007, 8, 651.
extrahepatic metabolism, and
a moderate volume (4.7 L/kg),
12. Gjoni, T.; Urwyler, S. Neuropharmacology 2008, 55, 1293.
13. (a) Goodnow, R. A., Jr.; Gillespie, P. Prog. Med. Chem. 2007, 45, 1; (b) Wunberg,
T.; Hendrix, M.; Hillisch, A.; Lobell, M.; Meier, H.; Schmeck, C.; Wild, H.; Hinzen,
B. Drug Discovery Today 2006, 11, 175; (c) Keseru, G.; Makara, G. Drug Discovery
Today 2006, 11, 741; (d) Alanine, A.; Nettekoven, M.; Roberts, E.; Thomas, A. W.
Comb. Chem. High Throughput Screen. 2003, 6, 51.
resulting in a short half-life (0.8 h). It is noteworthy that in the
presence of rat liver microsomes, compound 38 had a relatively
high in vitro CL of 51 mL/min/kg.
In summary, we have identified a promising new series of
mGluR2 PAMs, characterized by high brain penetration and good
ligand efficiency. This work highlighted the importance of a thor-
ough hit assessment process as part of the hit-to-lead paradigm.
Judicious library design allowed to reach a decision making point
on this series in one iteration of synthesis. Modification of the alk-
ylether moiety of 17 led to a series of biaryl analogs with lower hu-
man microsomal CL and improved physical properties (38–40).
Efforts to introduce polar functionality and/or an ionizable center
were successful in that they improved solubility, reduced cLogP,
provided a salt handle and lowered human microsomal CL, but
these compounds had generally less potent mGluR2 PAM activity.
Compound 38 was found to be active in an in vivo methamphet-
amine-induced hyperlocomotor model at a free brain exposure
nearly equal to its in vitro mGluR2 PAM EC50 of 30 nM, but the total
14. MS-based quantification of the basal/apical transfer rate of a test compound at
2 lM across contiguous monolayers from MDCK (Madin Darby Canine Kidney)
cells.
15. Ratio from the MS-based quantification of apical/basal and basal/apical
transfer rates of a test compound at 2
from MDR1-transfected MDCK cells.
lM across contiguous monolayers
16. Hopkins, A. L.; Groom, C. R.; Alexander, A. Drug Discovery Today 2004, 9, 430.
17. Ligand efficiency is calculated based on binding affinity. Since the mGluR2
assay used is
a functional assay and the EC50 value depends on the
concentration of glutamate present in the assay, the precise assessment of
the ligand efficiency based on this assay was not possible. At the same time,
comparison of LE profiles of multiple chemotypes using the same mGluR2
assay provided
a valuable perspective when assessing lead generation
potential of different chemical series.
18. Hitchcock, S. A. Curr. Opin. Chem. Biol. 2008, 12, 318.
19. (a) Patel, Y.; Gillet, V. J.; Howe, T.; Pastor, J.; Oyarzabal, J.; Willett, P. J. Med.
Chem. 2008, 51, 7552; (b) Dill, K. A. J. Biol. Chem. 1997, 272, 701.
20. Feng, B.; Mills, J. B.; Davidson, R. E.; Mireles, R. J.; Janiszewski, J. S.; Troutman,
M. D.; de Morais, S. M. Drug Metab. Dispos. 2008, 36, 268.
plasma exposure required to achieve this was high (2.3 lM) due to
21. Fluorescence-based measurement of CYP inhibition by incubating the
high protein binding. We conclude that future endeavors will have
to address the high lipophilicity seemingly required for mGluR2
PAM binding in the transmembrane domain versus the reduced
lipophilicity necessary for lower protein binding.
recombinant CYP (Cytochrome P450) with
corresponding fluorescent probe at 3 M.
22. Scintillation proximity assay measuring inhibition of 3H-dofetilide binding to
membranes from hERG-expressing HEK cells at 10 M.
a test compound and the
l
l
23. In efforts to obtain analogs in solid form, we compared the MPs of small
alkoxyphenyl ethers (oils) to benzylphenyl ether (ꢀ40 °C), phenyl ether
(ꢀ28 °C) and biphenyl (ꢀ70 °C).
Acknowledgments
24. (a) Hirota, S.; Kawashima, N.; Chaki, S.; Okuyama, S. CNS Drug Rev. 2003, 9(4),
375; (b) Satow, A.; Maehara, S.; Ise, S.; Hikichi, H.; Fukushima, M.; Suzuki, G.;
Kimura, T.; Tanaka, T.; Ito, S.; Kawamoto, H.; Ohta, H. J. Pharmacol. Exp. Ther.
2008, 326, 577.
The authors thank Bruce Posner and Csilla Csiki Jorgensen for
the HTS work, Jon Bordner for the single crystal X-ray structure
elucidation of compound 3, Joel Schachter and Diane Lang for con-
structing cells and Professor Shigetada Nakanishi of Kyoto Univer-
sity for providing cDNA. We also thank Rebecca O’Connor and Anne
25. Gonzalez-Maeso, J.; Ang, R. L.; Yuen, T.; Chan, P.; Weisstaub, N. V.; Lopez-
Gimenez, J. F.; Zhou, M.; Okawa, Y.; Callado, L. F.; Milligan, G.; Gingrich, J. A.;
Filizola, M.; Meana, J. J.; Sealfon, S. C. Nat. Lett. 2008, 1.