3,4-Dihydro-2H-benzoxazinones are 5-HT1A receptor antagonists with potent 5-HT reuptake inhibitory activity

Article abstract of DOI:10.1016/j.bmcl.2004.11.030

Starting from a high throughput screening hit, a series of 3,4-dihydro-2H-benzoxazinones has been identified with both high affinity for the 5-HT_1A receptor and potent 5-HT reuptake inhibitory activity. The 5-(2-methyl)quinolinyloxy derivative

Full text of DOI:10.1016/j.bmcl.2004.11.030

Bioorganic & Medicinal Chemistry Letters 15 (2005) 737–741
3,4-Dihydro-2H-benzoxazinones are 5-HT1A receptor
antagonists with potent 5-HT reuptake inhibitory activity
Peter J. Atkinson,a Steven M. Bromidge,a Mark S. Duxon,b Laramie M. Gaster,c
Michael S. Hadley,a Beverley Hammond,b Christopher N. Johnson,b,*
Derek N. Middlemiss,a Stephanie E. North,d Gary W. Price,a Harshad K. Rami,b
Graham J. Riley,a Claire M. Scott,a Tracey E. Shaw,a Kathryn R. Starr,a Geoffrey Stemp,b
Kevin M. Thewlis,a David R. Thomas,a Mervyn Thompson,b Antonio K. K. Vonga
and Jeannette M. Watsona
aPsychiatry Center of Excellence for Drug Discovery, GlaxoSmithKline, New Frontiers Science Park, Third Avenue,
Harlow, Essex CM19 5AW, UK
bNeurology and GI Center of Excellence for Drug Discovery, GlaxoSmithKline, New Frontiers Science Park,
Third Avenue, Harlow, Essex CM19 5AW, UK
cDiscovery Research, GlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow, Essex CM19 5AW, UK
dWorldwide Bioanalysis DMPK, GlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow, Essex CM19 5AW, UK
Received 13 September 2004; revised 2 November 2004; accepted 4 November 2004
Available online 23 December 2004
Abstract—Starting from a high throughput screening hit, a series of 3,4-dihydro-2H-benzoxazinones has been identified with both
high affinity for the 5-HT1A receptor and potent 5-HT reuptake inhibitory activity. The 5-(2-methyl)quinolinyloxy derivative com-
bined high 5-HT1A/1B/1D receptor affinities with low intrinsic activity and potent inhibition of the 5-HT reuptake site (pKi 8.2). This
compound also had good oral bioavailability and brain penetration in the rat.
Ó 2004 Elsevier Ltd. All rights reserved.
Abnormalities in brain serotonin (5-HT) neurotransmis-
sion are thought to be the underlying cause of mood dis-
orders, such as depression and anxiety.1 Synaptic 5-HT
levels are under the control of 5-HT transporters (SerT),
located presynaptically, and 5-HT autoreceptors that
reside on cell bodies (5-HT1A receptor subtype) and on
nerve terminals (5-HT1B and/or 5-HT1D receptor sub-
type).2 Despite the success of selective serotonin reup-
take inhibitors (SSRIs) in the treatment of depression
and anxiety, one potential downside is a long latency
to therapeutic onset.3 The lack of clinical efficacy follow-
ing acute administration of SSRIs has been hypothesised
to be due to indirect 5-HT1A autoreceptor activation
resulting in an inhibitory effect on 5-HT cell firing and
net synaptic 5-HT levels. Consequently, SSRIs only
effectively elevate 5-HT levels following desensitisation
of 5-HT1A autoreceptors after chronic dosing.4 There-
fore, concomitant blockade of 5-HT1A receptors should
theoretically enhance 5-HT levels above that of an SSRI
alone.5 This has been confirmed in microdialysis studies
which demonstrated that co-administration of the selec-
tive 5-HT1A antagonist WAY-100635 with an SSRI re-
sulted in an immediate increase in central 5-HT levels.6
Furthermore, in the rat High-Light Social Interaction
model of anxiety, the onset of anxiolytic activity of par-
oxetine was accelerated from 21 to 7 days by co-admin-
istration with WAY-100635.7 Therefore, a combined
SSRI/5-HT1A receptor antagonist should have a reduced
latency to onset of therapeutic effects relative to SSRIs.5
This hypothesis is also supported by the recent meta-
analysis of clinical trials data which suggests that the
5-HT1A autoreceptor antagonist ( )-pindolol hastens
the therapeutic response to SSRIs.8
In order to identify compounds with such a dual profile,
high throughput screening was carried out using a [35S]-
GTPcS binding assay with CHO cells expressing human
*
Corresponding author. Tel.: +44 1279 627735; fax: +44 1279
622790; e-mail: christopher_n_johnson@gsk.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.11.030

738
P. J. Atkinson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 737–741
(h) 5-HT1A receptors. Compounds with good antagonist
potency were then tested for blockade of [3H]-5-HT
reuptake into rat cortical synaptosomes. From this ap-
proach, the diastereomeric mixture of 3,4-dihydro-2H-
benzoxazinones 19 was identified as a lead. This letter
describes the optimisation of benzoxazinones related
to 1.
ated with 4-(3-bromopropyloxy)indole10 to give 6, albeit
in low yield. The remaining compounds were prepared
according to Scheme 2. Phenol 3 underwent Mitsonobu
reaction with N-Boc protected cyclic aminoalcohols 10
and 20, then reductive cyclisation and Boc removal as
described above gave benzoxazinones 11 and 21, respec-
tively. Improved yields for the final stage were obtained
either by using the appropriate aryloxyethyl bromide11
in the presence of excess diisopropylethylamine as base
(Method A), or by employing reductive alkylation con-
ditions using the appropriate aryloxyacetaldehyde12
(Method B). The piperazine analogue 26 was prepared
by reductive amination of aldehyde 2413 with 1-Boc-
piperazine, followed by deprotection to give 25, then
subsequent alkylation according to Method A. Demeth-
ylation of 27 was followed by pyridine ring reduction,
nitration, Boc protection and alkylation with methyl
bromoacetate, to give nitroester 28. Reductive cyclisa-
tion as previously described gave key intermediate 29
from which methylene-linked benzoxazinones 30 and
31 were prepared according to Method A. All final com-
pounds were purified by chromatography and isolated
as their hydrochloride salts.
Me
H
O
N
O
O
N
H
OH
O
N
H
1
Novel compounds 6, 12–19, 22, 23, 26, 30 and 31 were
prepared according to Schemes 1 and 2. Alkylation of
nitrophenol 2 with methyl bromoacetate was followed
by debenzoylation using sodium methoxide in methanol
to give 3 (Scheme 1), then Mitsonobu reaction with N-
Boc-ethanolamine 7 gave 4. Nitro group reduction and
concomitant cyclisation was followed by acid-mediated
removal of the N-Boc group to give 5, which was alkyl-
NO2
OH
NO2
O
NO2
O
PhCO2
HO
Boc
O
(i), (ii)
90%
(iii)
N
H
Boc
OH
CO2Me
O
N
CO2Me
H
H
2
3
7
4
85%
(iv), (v) 75%
O
H
N
O
N
O
(vi)
20%
O
Br
O
N
H
H2N
O
O
8
6 (Table 1)
5
N
H
N
H
Scheme 1. Reagents and conditions: (i) BrCH2CO2Me, K2CO3, acetone, reflux, 18 h; (ii) NaOMe, MeOH, 20 °C, 1 h; (iii) 7, Bu3P, 1,10-
(azodicarbonyl)dipiperidine, 20 °C, 6 h; (iv) Pd–C, cyclohexene, EtOH, reflux, 2 h; (v) CF3CO2H, CH2Cl2, 20 °C, 18 h; (vi) 8, NaHCO3, EtOH,
reflux, 20 h.
H
N
H
N
O
O
O
O
OH
(i) - (iii)
65%
(iv)
(Table 2)
N
HN
HN
HN
30 - 70%
ArO
O
O
.HCl
14, 15, 19 (Method A)
10
11
12, 13, 16 - 18 (Method B)
( )n
( )n
H
( )n
H
N
H
HN
N
(iv)
(i) - (iii)
OH
O
O
O
N
O
O
O
ArO
H
H
.HCl
35 - 40%
60 - 65%
O
O
21
20
22, 23 (Table 3)
H
N
H
N
H
N
OHC
O
N
N
N
(v), (iii)
(iv)
HN
ArO
O
O
O
90%
45%
.2HCl
25
26 (Table 3)
24
(vi) - (x)
H
N
H
N
NO2
O
O
O
(iv)
(ii), (iii)
70%
N
N
HN
N
CO2Me
50 - 85%
ArO
Boc
O
OMe 75%
.HCl
O
28
27
30, 31 (Table 3)
29
Scheme 2. Reagents and conditions: (i) 3, Ph3P, DIAD, THF, 20 °C, 18 h; (ii) H2, Pd–C, 3 bar, 50 °C, 18 h; (iii) HCl, Et2O, 2-propanol, reflux, 18 h;
(iv) ArOCH2CH2Br, i-Pr2NEt, 2-propanol, reflux, 48 h (Method A), OR ArOCH2CHO, NaBH(OAc)3, CH2Cl2, 20 °C, 18 h (Method B); (v) 1-Boc-
piperazine, NaBH(OAc)3, 20 °C, 18 h; (vi) 48% aq HBr, reflux, 8 h; (vii) H2, PtO2, H2SO4, MeOH, 20 °C, 18 h; (viii) HNO3, AcOH, 20 °C, 2 h; (ix)
(Boc)2O, NEt3, H2O, THF, 20 °C, 18 h; (x) BrCH2CO2Me, K2CO3, acetone, reflux, 18 h.

P. J. Atkinson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 737–741
739
Table 1. 5-HT1A receptor binding affinity (pKi), 5-HT1A intrinsic
activity (IA), SerT potency (pKi) and b2 adrenergic receptor binding
affinity (pKi): individual diastereoisomers of 1 (1a–d) and 6a
Table 3. 5-HT1A receptor binding affinity (pKi), 5-HT1A intrinsic
activity (IA) and SerT potency (pKi): 5-quinolinyloxy derivativesa
R
H
R2
N
O
O
(CH2)2-LINKER
N
H
O
N
O
O
N
H
O
R1
O
Compoundb
R
H
Linker
5-HT1A IA SerT
N
H
O
19
22
23
7.9
8.9
8.8
0.1 7.5
0.3 7.3
0.3 7.1
c
b2
Compoundb
R1
R2
5-HT1A
IA
SerT
N
N
1
OH
OH
OH
OH
OH
H
Me
Me
Me
Me
Me
H
9.1
7.9
8.7
8.1
9.3
8.8
0.3
0.4
0.3
0.3
0.2
0.6
7.3
6.9
7.1
7.0
6.5
7.1
9.2
7.7
8.6
7.7
8.8
6.0
H
H
O
O
1a
1b
1c
1d
6
N
CH2
CH2
CH2
N
a All pKi values represent the mean of at least three experiments, each
within 0.3 of the mean. IA values represent the mean of three
experiments, each within 0.1 of the mean.
b All new compounds gave satisfactory analytical and/or mass spectral
data.15
c Displacement of [125I]-iodocyanopindolol from human cloned b2
receptors expressed in CHO cells.
26
30
31
H
8.0
8.9
9.5
0.2 7.5
0.2 8.2
0.2 8.2
N
N
N
H
Me
a All pKi values represent the mean of at least three experiments, each
within 0.3 of the mean. IA values represent the mean of three
experiments, each within 0.1 of the mean.
Compounds 1, 6, 12–19, 22, 23, 26, 30 and 31 were evalu-
ated using displacement of [3H]-8-OH-DPAT binding
from h5-HT1A receptors expressed in CHO cells, whilst
potency for the serotonin transporter (SerT) was as-
sessed by measuring the inhibition of reuptake of [3H]-
5-HT into rat cortical synaptosomes, with data ex-
pressed as pKi values. Functional activity was measured
using [35S]-GTPcS binding in HEK293 cells expressing
h5-HT1A receptors, with intrinsic activity (IA) expressed
relative to the 5-HT response (5-HT = 1). Data are
shown in Tables 1–3.
b All new compounds gave satisfactory analytical and/or mass spectral
data.15
SerT potency (Table 1), but also possessed high affinity
for b2 adrenergic receptors (pKi 9.2). Separation of the
four constituent diastereomers (1a–d) was carried out
using chiral HPLC.14 Encouragingly, 5-HT1A and SerT
activity did not reside in separate diastereomers, but
b2 activity unfortunately ran parallel with 5-HT1A activ-
ity. The adrenergic activity is not surprising given the
presence of an aryloxypropanolamine moiety in the mole-
cules (cf. pindolol). Therefore, the simplified analogue 6
in which the hydroxy group is removed from the linker
was prepared and found to retain both 5-HT1A and SerT
activity with greatly reduced b2 affinity.
Diastereomeric mixture 1 had high affinity in the 5-
HT1A receptor binding assay together with moderate
Table 2. 5-HT1A receptor binding affinity (pKi), 5-HT1A intrinsic
activity (IA) and SerT potency (pKi): variation of LHS aryloxy moietya
H
O
N
O
Although 6 had improved selectivity, this was at the cost
of increased 5-HT1A IA, and reduced SerT potency (tar-
get pKi P 8). The effect of increased conformational
constraint of the central basic linker was investigated,
and the piperidinyloxy analogue 12 proved to be a key
breakthrough (Table 2), exhibiting 5-HT1A and SerT
pKi > 8 together with low 5-HT1A IA. Although 12
showed >100-fold selectivity over b2 adrenoceptors, it
had <100-fold selectivity over a range of other mono-
amine receptors and poor in vitro metabolic stability
in human and rat liver microsomes.16 The SAR of the
left-hand aryloxy moiety was next investigated, and
from ca. 90 variations it was found that bicyclic aryl
groups had the most promising profiles (13–19, Table
2). However, very few achieved pKi P 8 at both
5-HT1A and SerT, and most had similar or higher
5-HT1A IA than 12. 2-Cyanoindolyl 13 had pKi > 8 at
both target sites, but had high b2 adrenoceptor binding
N
ArO
O
Compoundb Ar
5-HT1A IA
SerT
12
13
14
4-Indolyl
8.4
8.8
8.7
0.2
0.3
0.9
8.3
8.1
7.6
4-Indolyl(2-cyano)
1,2-Dihydrobenzo-
[b]furanyl(2,2-dimethyl)
1-Naphthyl
15
16
17
18
19
8.6
8.1
7.0
8.2
7.9
0.4
0.8
7.0
7.3
1-Isoquinolinyl
4-Quinolinyl
8-Quinolinyl
ND 7.0
0.5
0.1
7.4
7.5
5-Quinolinyl
ND = not determined.
a All pKi values represent the mean of at least three experiments, each
within 0.3 of the mean. IA values represent the mean of three
experiments, each within 0.1 of the mean.
b All new compounds gave satisfactory analytical and/or mass spectral
data.15

740
P. J. Atkinson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 737–741
affinity (pKi 7.3). Dihydrobenzofuran 14 had excellent
5-HT1A affinity but very high IA. The encouraging
profile of 1-naphthyl 15 prompted investigation of
isoquinolinyl and quinolinyl derivatives 16–19. 5-Quin-
olinyloxy 19 was an important lead, having the lowest
5-HT1A IA. Consequently, replacement of the piperidin-
yloxy linker moiety by a range of alternative basic link-
ers was then investigated in the 5-quinolinyloxy series
(Table 3).
investigation of the role of enhanced 5-HT transmission
in the treatment of depression and anxiety.
References and notes
1. (a) Hirschfeld, R. M. A. J. Clin. Psychiat. 2000, 61(Suppl.
6), 4–6; (b) Feighner, J. P. J. Clin. Psychiat. 1999,
60(Suppl. 4), 4–11.
2. Middlemiss, D. N.; Price, G. W.; Watson, J. M. Curr.
Opin. Pharmacol. 2002, 2, 18–22.
3. Blier, P.; Bergeron, R. J. Clin. Psychiat. 1998, 59(Suppl.
5), 16–25.
4. (a) Kreiss, D. S.; Lucki, I. JPET 1995, 274, 866–876;
(b) Blier, P. E.; De Montigny, C.; Chaput, Y. J.
Clin. Psychopharmacol. 1987, 7(6 Suppl.), 24S–
35S.
5. Artigas, F. Trends Pharmacol. Sci. 1993, 14, 263.
6. (a) Dawson, L. A.; Nguyen, H. Q. Neuropharmacology
2000, 39, 1044–1052; (b) Romero, L.; Hervas, I.; Artigas,
F. Neurosci. Lett. 1996, 219, 123–126.
7. Duxon, M. S.; Starr, K. R.; Upton, N. Br. J. Pharmacol.
2000, 130, 1713–1719.
8. Ballestesteros, J.; Callado, L. F. J. Affect. Disorders 2004,
79, 137–147.
9. Duckworth, D. M., Jennings, L. J. A., European Patent
171702; Chem. Abstr. 1986, 105, 97484.
10. Prepared by treatment of 4-hydroxyindole with sodium
hydride in N,N-dimethylformamide and subsequent reac-
tion with 1,3-dibromopropane at 20 °C.
11. Prepared from the corresponding phenol by treatment
with excess 1,2-dibromoethane and potassium carbonate
in 2-butanone at reflux.
12. Intermediates for 12 and 13 were prepared according to
the method described in: Sasai, H.; Yamada, Y. M. A.;
Suzuki, T.; Shibasaki, M. Tetrahedron 1994, 50, 12313–
12318. Intermediates for 16 and 17 were prepared from 1-
chloroisoquinoline and 4-chloroquinoline, respectively, by
reaction with the sodium salt of 2-hydroxyacetaldehyde
dimethylacetal in N,N-dimethylformamide and subse-
quent acid hydrolysis. For 18, (8-quinolinyloxy)-acet-
aldehyde was prepared by oxidative cleavage of 8-allyl-
oxyquinoline with sodium periodate in the presence of
osmium tetroxide.
Molecular modelling suggested that the 3-pyrrolidin-
ylmethyl 22 and 3-piperidinylmethyl 23 would overlap
with 4-piperidinyl 19. Both compounds had improved
5-HT1A affinity compared to 19 but with slightly higher
IA. 4-Piperazinylmethyl 26 had a similar in vitro profile
to 19, but 4-piperidinylmethyl 30 represented a signifi-
cant advance, having pKi 8.9 at 5-HT1A with low IA
and, importantly, pKi > 8 at the reuptake site. Encour-
agingly, microsomal metabolic stability of 30 was
improved relative to 12 and, in a rat steady-state CNS
penetration assay,17 30 showed good brain penetration
(Brain:Blood = 0.8:1). However, the compound was rap-
idly cleared (CLb 115 mL/min/kg). In an attempt to
rationalise this observation, the in vitro metabolic stabi-
lity of 30 was determined in rat and human liver S9 frac-
tion.16 Intrinsic clearance was high (rat CLi 6 mL/min/g,
human CLi 42 mL/min/g), suggesting the possible
involvement of aldehyde oxidase (AO) in the metabo-
lism of 30. It has been previously shown that AO can
oxidise quinolinyl groups at C-2,16 so the corresponding
5-(2-methyl)quinolinyl derivative 31 (SB-649915) was
prepared in which this potential site of metabolism is
blocked. Gratifyingly, this compound was apparently
stable to aldehyde oxidase (rat and human S9 CLi
<0.6 mL/min/g) and had significantly reduced clearance
in a rat PK study (CLb 42 mL/min/kg). The oral bio-
availability of 31 was estimated at 45% and Brain:Blood
ratio at steady-state was 0.4:1. In addition 31 had excel-
lent 5-HT1A/SerT activity (pKi 9.5 and 8.2, respectively)
and low 5-HT1A IA (0.2).
Compound 31 was further profiled against a range of
monoamine receptors and transporters (5-HT, norad-
renergic and dopaminergic) and was found to have sig-
nificant affinity for 5-HT1B and 5-HT1D receptors
(pKi 8.1 and 8.7, respectively).18 Selectivity was other-
wise excellent, with pKi 6 6.3 at all other receptors
and transporters tested.
13. Belliotti, T. R.; Wustrow, D. J.; Brink, W. A.; Zoski, K.
T.; Shih, Y.-H.; Whetzel, S. Z.; Georgic, L. M.; Corbin, A.
E.; Akunne, H. C.; Heffner, T. G.; Pugsley, T. A.; Wise, L.
D. J. Med. Chem. 1999, 42, 5181–5187.
14. HPLC method: Chiralpak AD, 50/50/0.1 hexane/ethanol/
triethylamine.
15. 1H NMR spectra were recorded at 250 MHz in CDCl3 as
solvent. Compound 31, mp 207–208 °C (dihydrochloride
salt); dH 1.25–1.38 (2H, m), 1.49 (1H, m), 1.65 (2H, m),
2.14 (2H, m), 2.45 (2H, d, J = 7 Hz), 2.72 (3H, s), 2.94
(2H, t, J = 6 Hz), 3.05 (2H, m), 4.27 (2H, t, J = 6 Hz), 4.58
(2H, s), 6.56 (1H, d, J = 2 Hz), 6.73 (1H, dd, J = 7, 2 Hz),
6.78 (1H, d, J = 8 Hz), 7.86 (1H, d, J = 7 Hz), 7.23 (1H, d,
J = 8 Hz), 7.50–7.64 (2H, m), 8.42 (1H, d, J = 8 Hz), 8.75
(1H, br s) (free base).
16. Austin, N. E.; Baldwin, S. J.; Cutler, L.; Deeks, N.; Kelly,
P. J.; Nash, M.; Shardlow, C. E.; Stemp, G.; Thewlis, K.;
Ayrton, A.; Jeffrey, P. Xenobiotica 2001, 31, 677–686, S9
clearance values were determined in the absence of co-
factors.
Starting from screening hit 1, which had high 5-HT1A
receptor affinity, moderate SerT potency and poor selecti-
vity versus b2 adrenergic receptors, compound 31 (SB-
649915) has been identified as a high affinity 5-HT1A
antagonist combined with potent 5-HT reuptake inhibi-
tory activity. Compound 31 has low 5-HT1A intrin-
sic activity in the [35S]-GTPcS binding assay using
HEK293 cells expressing the h5-HT1A receptor and
has been shown to have no intrinsic activity as measured
by GTPcS autoradiography19 in a human dorsal raphe
nucleus preparation. Compound 31 has an excellent
selectivity profile, together with good oral bioavailabi-
lity and brain penetration in the rat, and therefore rep-
resents a useful compound for the further in vivo
17. CNS penetration at steady-state in rat. Compounds
were dissolved in 2% (v/v) DMSO in 5% (w/v) dextrose
aq and administered at a constant infusion rate over 12 h
at a target dose rate of 0.3 mg free base/kg/h. Blood

P. J. Atkinson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 737–741
741
samples were removed during the latter part of the
infusion to confirm steady-state blood concentrations.
Blood and brain samples were analysed by LC/MS/MS.
18. Displacement of [3H]-5-HT from human cloned
5-HT1B and 5-HT1D receptors expressed in CHO cells.
19. Watson, J.; Roberts, C.; Scott, C.; Kendall, I.; Collin, L.;
Day, N. C.; Harries, M. H.; Soffin, E.; Davies, C. H.;
Randall, A. D.; Heightman, T.; Gaster, L.; Wyman, P.;
Parker, C.; Price, G. W.; Middlemiss, D. N. Br. J.
Pharmacol. 2001, 133, 797–806.