Potential anxiolytic agents. Part 4: Novel orally-active N5-substituted pyrido[1,2-a]benzimidazoles with high GABA-A receptor affinity

Article abstract of DOI:10.1016/S0960-894X(02)00463-8

A series of pyrido[1,2-a]benzimidazoles (PBIs) with substitution on the N⁵-nitrogen has been synthesized and found to possess high affinity for the benzodiazepine (BZD) site on the GABA-A receptor. The compounds evaluated include those bearing a heteroalkyl group and heterocyclic rings. The most promising of these compounds is ethoxymethyl analogue 24, which has an IC₅₀ of 0.1 nM for the BZD site on the GABA-A receptor and has been advanced to human clinical trials.

Full text of DOI:10.1016/S0960-894X(02)00463-8

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2381–2386
Potential Anxiolytic Agents. Part 4: Novel Orally-Active
N5-Substituted Pyrido[1,2-a]benzimidazoles with
High GABA-A Receptor Affinity
Alfonzo D. Jordan,a,* Anil H. Vaidya,a Daniel I. Rosenthal,a Barry Dubinsky,a
Cheryl P. Kordik,a Pauline J. Sanfilippo,a Wu-Nan Wub and Allen B. Reitza
aDrug Discovery Division, Johnson & Johnson Pharmaceutical Research and Development,
Welsh and McKean Roads, PO Box 776, Spring House, PA 19477-0776, USA
bPreclinical Development, Johnson & Johnson Pharmaceutical Research and Development,
Welsh and McKean Roads, PO Box 776, Spring House, PA 19477-0776, USA
Received 15 March 2002; accepted 24 May 2002
Abstract—A series of pyrido[1,2-a]benzimidazoles (PBIs) with substitution on the N5-nitrogen has been synthesized and found to
possess high affinity for the benzodiazepine (BZD) site on the GABA-A receptor. The compounds evaluated include those bearing a
heteroalkyl group and heterocyclic rings. The most promising of these compounds is ethoxymethyl analogue 24, which has an IC50
of 0.1 nM for the BZD site on the GABA-A receptor and has been advanced to human clinical trials. # 2002 Elsevier Science Ltd.
All rights reserved.
Our continuing investigations into anxiolytics with an
improved margin of safety as compared to marketed
drugs have prompted us and others to explore the PBI
chemical series (viz. 1).1ꢀ7 During the course of these
studies, we found that substitution of N5 on the PBI
nucleus produced compounds with very high GABA-A
receptor affinities and favorable in vivo therapeutic
indices. Substitution at N5 was conducted originally in
an attempt to improve the physical properties of the
series such as aqueous solubility. The PBIs are generally
water insoluble which can complicate drug develop-
ment, even though their therapeutically effective dose is
expected to be quite low (ca. 1–10 mg/day). Therefore,
we focused our efforts on N5 substitution bearing polar
functionality (viz. 2) including those groups such as
amino or basic heterocycles which could form acid-
addition salts. In this paper we describe the synthesis and
structure–activity relationships of a series of N5-sub-
stituted PBI derivatives as potential anxiolytics.
Conversion of 3 to various amides 4 was accomplished
by heating an appropriately substituted aniline in a sui-
table solvent such as xylene or dimethylformamide.
Treatment of 4 with a strong base (NaH), 15-crown-5
and a reactive electrophile such as a halo alkyl ether in
dimethylformamide (DMF) promoted direct alkylation.
Alternatively, the sodium salt of 4 was prepared sepa-
rately using freshly formed EtONa and stored as a solid
and then added to the reaction mixture with the crown
ether. In addition, the alkylation reaction of 4 with di-
alkylaminoethyl halides proceeded under phase transfer
catalysis conditions utilizing aqueous sodium hydroxide
and benzyltrialkylammonium halide in chloroform. These
chloroethylamines demonstrated enhanced reactivity as
compared to the corresponding chloropropylamines
under the same reaction conditions, as the former reacts
presumably through an aziridinium intermediate.
Since only reactive electrophiles worked well in the
direct alkylation procedure, we developed Mitsunobu
reaction conditions for reactions of 4 with functiona-
lized alcohols, which proved to be fairly general except
for those substrates which bear an electron-withdrawing
group or a basic free NH group. Nonetheless, coupling of
4 with the appropriate alcohols was conducted with di-
ethyl (DEAD) or diisopropylazodicarboxylate (DIAD) or
1,10-(azodicarbonyl) dipiperidine (ADDP) and either tri-
phenyl or tributylphosphine in tetrahydrofuran or
Chemistry
N5-Substituted PBI analogues were prepared in two
steps from carboxylic acid ester 34 as shown in Scheme 1.
*Corresponding author. Tel.: +1-215-628-5638; fax: +1-215-628-
4985; e-mail: ajordan@prdus.jnj.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00463-8

2382
A. D. Jordan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2381–2386
ester 7 to afford 8. Incorporation of the ethoxymethyl
group onto N5 as described previously led to 9, which
produced 10 upon removal of the silyl group with
nBu4NF. The phenolic hydroxyl of 10 was further
reacted with electrophiles Mel and ClCO2Me to furnish
compounds 11 and 12.
Results and Discussion
The biological activity of the PBI derivatives is shown in
Tables 1–4, along with that of the reference benzodi-
azepine diazepam. Specifically, these compounds were
evaluated for their in vitro affinities (IC50 values) for the
BZD site on the GABA-A receptor by competition
experiments with the radioligand [3H]-flunitrazepam;11
compounds were tested at five concentrations in a tissue
preparation from rat cerebral cortex. In addition, the
GABA shift (G.S.), which is the ratio of the binding in
the absence and presence of GABA, was also deter-
mined. Our goal was to identify compounds with high
affinity for the GABA-A receptor and which have a
partial agonist profile (G.S.=1.2–1.8).4 These com-
pounds would be expected to have minimal side effects
such as daytime sedation and abuse liabilities which
are characteristic of full agonists such as diazepam
(G.S.>2.0). Our in vivo evaluation has consisted of a
series of tests including inhibition of pentylenetetrazole
(PTZ)-induced seizures in mice and experimentally
induced conflict in rats. In the rat conflict assay, the
data are presented as the dose at which efficacy is first
observed (minimum effective dose, MED) whereas in
the PTZ assay, the data are expressed as the dose that
Scheme 1. Reagents and conditions: (a) ArNH2, xylene, 140 ꢁC; (b)
NaH, 15-crown-5, RX, DMF or RCH2OH, DEAD, Ph3P, THF, rt or
RX, R4N+ Xꢀ, aq NaOH, CHCl3, rt.
dichloromethane to give the desired compound
directly.8ꢀ10 Certain ring modified compounds required
the synthesis of particular anilines prior to condensation
with 3. For example as shown in Scheme 2, phenol 5
was protected by silylation such as with reaction of t-
BuPh2SiCl, followed by reduction of the nitro group
with hydrogen gas at 55 psi and Pd/C in 3:1 mixture of
MeOH/EtOAc to give 6, and then condensation with
Scheme 2. Reagents and conditions: (a) tBuPh2SiCl, imidazole, DMF; (b) H2, 10% Pd/C, MeOH/EtOAc; (c) xylene, 140 ꢁC; (d) NaH, ClCH2OEt,
15-crown-5, DMF; (e) nBu4NF, THF; (f) Mel, K2CO3 18-crown-6, DMF; or CH3OCOCl, pyridine, CHCl3.

A. D. Jordan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2381–2386
Table 1. Biological data for ring B substituted PBI derivatives
2383
No.
R
GABAA IC50 (nM)
G.S.
PTZ (mouse) ED50 (mg/kg)
Anticonflict (rat) MED (mg/kg)
ip
po
ip
po
13
14
15
16
17
H
1.9
49.9
3.3
5.1
7.4
1.8
2.7
1.4
2.7
1.8
ꢂ1
>30
ꢂ3
30
10
1
10
<10
10
10
10
>30
3
>10
10
(Me)2N(CH2)2
(Me)2N(CH2)3
H2N(CH2)2
0.5
3
0.3
1
MeO(CH2)2NH(CH2)2
10
18
19
20
21
22
23
8.5
6.1
1.3
2.2
2.0
2.6
2.1
1.7
1
<30
30
>10
>10
10
—
—
1
0.9
0.3
1
10
>10
>10
30
1.6
10
—
35.0
5.6
ꢂ1
30
—
>1
>30
—
>10
24
25
26
EtOCH2
EtO(CH2)2
EtO(CH2)3
0.1
0.5
0.3
1.2
3.2
1.6
0.01
0.04
0.3
0.3
10
ꢂ10
—
0.4
1
1
ꢂ1
ꢂ1
27
2.3
2.2
ꢂ1
3
—
3
28
29
30
31
32
MeO(CH2)2OCH2
HO(CH2)2
HO(CH2)3
MeOCO2(CH2)2
MeCO2(CH2)2
Diazepam
0.8
1.2
1.1
0.3
0.2
4.9
2.2
2.7
1.8
2.2
1.2
2.2
0.003
ꢂ1
0.1
0.1
<30
ꢂ10
10
10
—
—
5
3
>10
10
3
3
>10
0.5
3
3
0.5
ꢂ1
ꢂ1
0.11
Table 2. Biological data for ring A substituted PBI derivatives
No.
R
GABAA IC50 (nM)
G.S.
PTZ (mouse) ED50 (mg/kg)
Anticonflict (rat) MED (mg/kg)
ip
po
ip
po
33
34
35
36
37
38
H
151.0
0.5
126.0
13.3
0.5
2.0
2.8
1.7
2.9
1.7
1.4
>1
<1
1
—
<1
<1
3
—
—
—
—
—
—
>10
>10
>10
>10
1
7-Cl
8,9-F2
6,8-F2
2
>30
—
<3
>30
6,7-F2
7,8-CH¼CH-CH¼CH—
2.2
>10

2384
A. D. Jordan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2381–2386
Table 3. Biological data for ring D modified PBI derivatives
No.
Ar
GABAA IC50 (nM)
G.S.
PTZ (mouse) ED50 (mg/kg)
Anticonflict (rat) MED (mg/kg)
ip
po
ip
po
39
40
10
11
12
41
42
43
44
0.6
1.2
1.4
0.4
4.3
1.0
1.0
0.7
0.7
1.6
2.0
2.0
2.2
3.2
1.7
2.1
2.3
1.6
1
3
>10
>10
ꢂ1
—
<3
0.3–3
0.3
0.1
1
0.1
1–3
0.03
1
—
0.1
—
—
—
—
—
1
3–30
<0.2
0.2
ꢂ1
ꢂ1
ꢂ1
ꢂ1
0.3
3
0.5
0.3
>10
10
Table 4. In vivo biological data for selected compounds
No.
PTZ (mouse)
ED50 (mg/kg)
Anticonflict (AC, rat)
MED (mg/kg)
EtOHa sleep (rat)
MED
Ratio
EtOH/AC
HSb ED50 (mouse)
Ratio
HS/PTZ
ip
po
ip
po
po
po
ip
po
27
30
10
po
24
28
0.01
0.003
<1
0.3
0.04
0.1
<3
0.5
10
ꢂ10
0.4
3
0.1
5
0.3
3
0.1
5
1
1
1
1
0.3
0.07
<1
0.1
675
300
3
40
Diazepam
0.1
5
6.3
13
aPotentiation of EtOH-induced sleep time.
bHorizontal screen.
antagonizes seizures induced in 50% of the mice (ED50).
The in vivo data upon intraperitoneal (ip) and oral
administration (po) are shown in the tables. The data in
Table 1 illustrate the replacement of the N5 hydrogen in
13 with various heteroatom groups and heterocyclic
rings, and allowing ring C to be unsubstituted, the 7-
fluoro substitution on ring A, and 2-fluoro substitution
on ring D. This pattern of fluoro substitution was
selected because it was found previously to impart high
levels of GABA-A receptor affinity.
In general, compounds with alkylamino substitution on
the N5 position (viz, 14–17) displayed lower affinity for
the GABA-A receptor than the direct unsubstituted

A. D. Jordan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2381–2386
2385
analogue 13 but were more orally active in the in vivo
assays. However, aromatic amine 18 and cyclic amine
19 were less potent in the binding assay and the PTZ in
vivo assay. The 2-pyridylethyl derivative 20 demon-
strated a 2-fold increase in potency (IC50 of 0.9 nM) at
the GABA-A receptor as compared to 13 and was
found to be more orally active in the pentylenetetrazole
and anticonflict biological tests. Imidazolylethyl com-
pound 21 showed comparable in vitro activity and
improved in vivo activity. Highly basic amidine 22
exhibited less activity in the binding assay with larger
G.S. value of 2.1. Similiarly, guanidine 23 revealed
diminished in vitro activity but a lower G.S. value of
1.7. Compounds 22 and 23 showed no improvement in
oral activity in the in vivo tests. Although formation of
the acid addition salts of compounds 14–21 did result in
an increase in their water solubility, the oral activities of
these compounds were not improved enough to warrant
further consideration.
(IC50=126.0 nM) than 24, whereas 6,8-, and 6,7-
difluorophenyl compounds 36 and 37 had 13.3 and 0.5
nM IC50s in vitro. In addition, 37 displayed 3-fold less
activity po in the anticonflict assay. These results sug-
gest that halogen substitution of the 6 and/or 7 posi-
tions are favorable for GABA-A receptor binding. In a
previous report, the A ring naphthyl derivative with 7,8
fusion was found to be the most potent of the three
possible benzo ring fused isomers.3 Incorporation of the
naphthyl ring as a phenyl replacement afforded com-
pound 38, which demonstrated weaker binding affinity
(IC50=2.2 nM) for the GABA-A receptor than 24 and
diminished in vivo activity in the rat anticonflict test.
Lastly, our SAR studies focused on modifications of
ring D (Table 3). The 2,6- and 2,4-difluorophenyl com-
pounds 39 and 40 showed slightly less in vitro activity,
but only 40 exhibited appreciable activity in the anti-
conflict assay. In vitro metabolism of 24 in rat and
human hepatic S9 fractions revealed that para hydro-
xylation on the D ring benzene ring was the major site
for metabolism (viz. 10). In fact, 10 was formed to the
extent of 60 and 30% upon incubation of 24 with rat
and human microsomes, respectively, after 1 h, and all
other metabolites only accounted for 10 and 20% of the
original drug sample. Therefore, 2-fluoro-4-hydroxy
phenyl compound 10 was prepared as shown earlier
(Scheme 2), and was found to be 14-fold less active in
vitro than 24, and 30-fold less active po in the rat con-
flict assay. The 2-fluoro-4-methoxy and 2-fluoro-4-
methoxycarbonyloxy compounds 11 and 12 were syn-
thesized as potential prodrugs of 10. The biological
activity of 11 was 3-fold better in the binding assay and
30-fold more active po in the rat conflict assay. Thio-
phene was utilized as a bioisosteric replacement for the
D ring phenyl. Specifically, N-(3-thienyl) and (2-thi-
enyl)carboxamides 41 and 42 were synthesized and each
was found to be significantly less active in vitro than 24
and less active po in in vivo efficacy tests. Incorporation
of a 5-methyl substituent (viz. 43) and a 3-chloro group
(viz. 44) on the 2-thienyl ring slightly enhanced binding
potency relative to 42 but nonetheless each analogue
was less active in the binding assay and po in the PTZ in
vivo assay than 24.
We then decided to incorporate oxygen-containing
groups at the N5 position of the PBI nucleus. We were
delighted to find that the ethoxymethyl derivative 24
displayed a superior IC50 of 0.1 nM to the GABA-A
receptor with an acceptable G.S.=1.2 as compared to
compound 13. In addition, compound 24 displayed
particularly excellent in vivo activity, including 0.04 mg/
kg MED po in the mouse PTZ and 0.4 mg/kg ED po in
the rat conflict test. Increasing the chain length by one
carbon produced ethoxyethyl compound 25, which was
5-fold less potent in vitro than 24 with an unexpectedly
large G.S. of 3.2. Compound 25 possessed 7-fold less
oral activity in the PTZ test and 3-fold less oral activity
in the conflict test. Ethoxypropyl homologue 26 dis-
played only 3-fold less potency in the binding assay but
7-fold less activity po in the mouse PTZ and 3-fold less
activity in the rat conflict tests. Cyclic ether compounds,
such as tetrahydrofuranyl methyl analogue 27 were
generally less active than acyclic ether 24. Diether 28,
with an IC50=0.8 nM and a rather high G.S. of 2.2, did
show 3-fold less po activity in the PTZ test and an 8-
fold reduction in the rat conflict test. Alcohols 29 and 30
showed an 8-fold reduction in in vitro activity and
exhibited diminished po in vivo activity. Methoxy-
carbonyloxy analogue 31 and acetate 32, designed as
prodrugs of 29, showed comparable in vivo potency of 3
mg/kg po in the Vogel conflict and PTZ tests, but were
generally less active than 24.
The in vivo biological data highlighting efficacy and side
effect testing for PBI compounds 24, 28, 40, and diaze-
pam are shown in Table 4. Several of these compounds
had pharmacological attributes that warranted further
investigation. In fact, compound 24 has excellent oral
efficacy, in both the PTZ (mice) and conflict assays (rats),
while possessing a beneficial separation from the side
effects of motor impairment in the horizontal screen (HS)
assay (mice; see Table 4). Compound 28, shows appreci-
able separation between efficacy in the PTZ assay and the
horizontal screen side effect test. Lastly, compound 40 was
found to display the least separation in the (HS/PTZ) side-
effect profile (Table 4). Among these compounds which
were evaluated in detail, there was no difference in the
relative degree of ethanol sleep time. Given that the
results for 24 are quite favorable relative to close ana-
logue 28 and diazepam, this compound is presently
being investigated more extensively.12
Based on the biological data discussed thus far, incor-
poration of the ethoxymethyl group on the N5 position
of the PBI nucleus was found to impart favorable
anxiolytic activity. The best compound for continued
structure–activity relationship (SAR) studies was 24.
The next region for SAR development in this series
centered on ring A and particularly halogen atom sub-
stitution, with the placement of halogen atoms at var-
ious positions on the phenyl ring (Table 2). Absence of a
fluoro group at position 7 (33) resulted in a ca. 1000-
fold decrease in vitro affinity versus compound 24. The
7-chlorophenyl analogue 34 exhibited in vitro potency
but a higher G.S. than 24. The 8,9-difluorophenyl com-
pound 35 was unexpectedly less active in vitro

2386
A. D. Jordan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2381–2386
Conclusions
References and Notes
We describe here a systematic series of modifications on
the PBI nucleus with the goal of improving biological
activity and physical properties. We examined substitu-
tions and modifications at the N-5 position and on rings
A, B, and D to produce potent, orally active anxiolytics
which modulate the BZD site of the GABA-A receptor
with a reasonably acceptable side effect profile. Ela-
boration of initial lead 13 by incorporation of various
heteroalkyl and heterocyclic alkyl groups at the N-5
position led to identification of the ethoxymethyl group
as preferred. Further investigations led to selection of
the 7-fluoro group as the best substituent on phenyl ring
A and the 2-fluoro group as the optimal group on
phenyl ring D. As a result, compound 24 was identified
during the course of these SAR studies as a potential
anxiolytic agent for human clinical evaluation.
1. Maryanoff, B. E.; Ho, W. H.; McComsey, D. F.; Reitz,
A. B.; Grous, P. P.; Nortey, S. O.; Shank, R. P.; Dubinsky, B.;
Taylor, R. J., Jr; Gardocki, J. F. J. Med. Chem. 1995, 38, 16.
2. Maryanoff, B. E.; McComsey, D. F.; Ho, W. H.; Shank,
R. P.; Dubinsky, B. Bioorg. Med. Chem. Lett. 1996, 6, 333.
3. Maryanoff, B. E.; Nortey, S. O.; McNally, J. J.; Sanfilippo,
P. J.; McComsey, D. F.; Dubinsky, B.; Shank, R. P.; Reitz,
A. B. Bioorg. Med. Chem. Lett. 1999, 9, 1547.
4. Reitz, A. B.; Jordan, A. D.; Sanfilippo, P. J.; Vavouyios-
Smith, A. US Patent 5,817,668, 6 October 1998.
5. Scott, M. K.; Demeter, D. A.; Nortey, S. O.; Dubinsky, B.;
Shank, A. P.; Reitz, A. B. Prog. Med. Chem. 1999, 36, 169.
6. Bokanov, A. I.; Evstratova, M. I.; Turchin, K. V.; Granik,
V. G.; Andreeva, N. I.; Asnina, V. V.; Golovina, S. M.;
Mashkovskii, M. D. Khim.-Farrn. Zh. 1997, 31, 27.
7. Xie, L.; Currie, K. S.; Albaugh, P.; Shaw, K.; Hutchison,
A. World Patent 99/40092, 1999.
8. Mitsunobu, O. Synthesis 1981, 1.
9. Hughes, D. Organic Reactions 1992, 42, 335.
10. Hughes, D. L. Org. Prep. Proced. Int. 1996, 28, 127.
11. Wood, P. L.; Loo, P.; Braunwalder, A.; Yokoyama, N.;
Cheney, D. L. J. Pharmacol. Exp. Ther. 1984, 231, 572. Wil-
liams, M.; Bennett, D. A.; Loo, P. S.; Braunwalder, A. F.;
Amrick, C. L.; Wilson, D. E.; Thompson, T. N.; Schmutz, M.;
Yokoyama, N.; Wasley, J. W. F. Ibid. 1989, 248, 89.
12. Dubinsky, B.; Carter, A. R.; Cheo-Isaacs, C. T.; Crooke,
J. J.; Deluca, S.; Devine, A.; Hochman, C.; Jordan, A. D.;
Reitz, A. B.; Rosenthal, D. I.; Shank, R. P.; Vaidya, A. H.
J. Pharmacol. Exp Ther. In press.
Acknowledgements
We thank the following colleagues for their advice and
assistance: Richard Shank, Bruce Maryanoff, Winston
Ho, Mark McDonnell, David McComsey, Malcolm
Scott, Samuel Nortey, James McNally, Louis Fitzpatrick,
and Anna Vavouyias-Smith.