Design, synthesis, and pharmacological evaluation of phenoxy pyridyl derivatives as dual norepinephrine reuptake inhibitors and 5-HT1A partial agonists

Article abstract of DOI:10.1016/j.bmcl.2009.12.023

Preclinical studies suggest that compounds with dual norepinephrine reuptake inhibitor (NRI) and 5-HT_1A partial agonist properties may provide an important new therapeutic approach to ADHD, depression, and anxiety. Reported herein is the discovery of a novel chemical series with a favorable NRI and 5-HT_1A partial agonist pharmacological profile as well as excellent selectivity for the norepinephrine transporter over the dopamine transporter.

Full text of DOI:10.1016/j.bmcl.2009.12.023

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1114–1117
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis, and pharmacological evaluation of phenoxy pyridyl
derivatives as dual norepinephrine reuptake inhibitors and 5-HT1A
partial agonists
b
a
a
a
Amy B. Dounay a, , Nancy S. Barta , Brian M. Campbell , Corey Coleman , Elizabeth M. Collantes ,
Lynne Denny b, Satavisha Dutta b, David L. Gray a, Dongfeng Hou c, Rathna Iyer a, Samarendra N. Maiti c,
Daniel F. Ortwine b, Al Probert a, Nancy C. Stratman a, Rajendra Subedi c, Tammy Whisman a,
Wenjian Xu a, Kim Zoski a
*
a Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, United States
b Pfizer Global Research and Development, 2800 Plymouth Road, Ann Arbor, MI 48105, United States
c Naeja Pharmaceutical Research and Development, Edmonton, Alberta, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Preclinical studies suggest that compounds with dual norepinephrine reuptake inhibitor (NRI) and
5-HT1A partial agonist properties may provide an important new therapeutic approach to ADHD, depres-
sion, and anxiety. Reported herein is the discovery of a novel chemical series with a favorable NRI and
5-HT1A partial agonist pharmacological profile as well as excellent selectivity for the norepinephrine
transporter over the dopamine transporter.
Received 12 October 2009
Revised 2 December 2009
Accepted 4 December 2009
Available online 6 December 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Norepinephrine reuptake inhibitor
5-HT1A partial agonist
Drugs that increase monoamine neurotransmission are widely
used to treat a variety of neuropsychiatric disorders. For example,
stimulant medications such as methylphenidate and amphet-
amine, which increase norepinephrine and dopamine (DA) neuro-
transmission throughout the brain, are commonly prescribed to
treat ADHD; however, these medications also carry a burden of
abuse liability.1 Alternatively, norepinephrine reuptake inhibitors
(NRIs) such as atomoxetine, which more selectively increase nor-
epinephrine neurotransmission with less impact on dopamine,
are also used to treat ADHD. In addition, NRI agents are clinically
efficacious in the treatment of depression2 and anxiety.3 Preclinical
microdialysis studies suggest that 5-HT1A receptors play a crucial
role in regulating DA transmission in the prefrontal cortex upon
treatment with monoamine reuptake inhibitors.4 Additional stud-
ies indicate that adding 5-HT1A partial agonist properties to a nor-
epinephrine reuptake inhibitor may selectively enhance cortical
DA neurotransmission.5 Moreover, preclinical behavioral studies
measuring antidepressant efficacy and cognitive function demon-
strate that compounds possessing both NRI and 5-HT1A partial ago-
nist properties are more effective than NRI agents alone.5 Thus,
identification of compounds with dual NRI and 5-HT1A partial
agonist pharmacologies may provide an important new therapeu-
tic approach to ADHD, depression, and anxiety.
As part of our efforts to identify novel agents with dual NRI and
5-HT1A partial agonist pharmacology, we recently reported the dis-
covery of a diphenyl ether series exemplified by 1 (Fig. 1).6 This
compound showed excellent activity in our in vitro assays and
in vivo efficacy models. During our optimization of compounds
such as 1, we investigated alternate templates in which one or
more of the phenyl groups were replaced with a heterocyclic sys-
tem. Our goal was to maintain the favorable NRI/5-HT1A profile
of 1, while improving upon its overall monoamine transporter
selectivity profile. Additionally, we anticipated that introduction
of a heterocycle would lower log P, thereby reducing microsomal
H
N
H
N
H
N
R1
R4
N
R1
F
R2
R3
O
O
F
O
N
N
R5
R5
1
2
3
* Corresponding author.
E-mail address: Amy.Dounay@pfizer.com (A.B. Dounay).
Figure 1. Exploration of the 2-phenoxypyridyl template.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.023

A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1114–1117
1115
Boc
N
Our medicinal chemistry strategy for the development of series
2 and 3 was designed with the following objectives: (a) optimiza-
tion of the NET inhibitor potency; (b) optimization of selectivity for
NET over the other monoamine transporters, with a primary focus
on NET/DAT and a secondary focus on NET/SERT; (c) optimization
of the 5-HT1A binding potency and partial agonist profile. Our pre-
vious studies on compounds such as 1 suggested that a detailed
survey of the SAR of phenyl ring substitutions in series 2 could
be a fruitful approach toward achieving our objectives in this
series.
Boc
N
Boc
N
a
b
B
O
O
O
OTf
4
5
6
Scheme 1. Reagents and conditions: (a) LDA, PhNTf2, THF, ꢀ78 °C to rt; (b)
bis(neopentylglycolato) diboron, Pd(dppf)Cl2, K3PO4, toluene/dioxane, 110 °C, 71%
yield (two steps).
Medicinal chemistry studies in this series commenced with
compound 2a, which bears an unsubstituted phenyl group. This
compound is a very weak inhibitor of NET, but still maintains mod-
erate 5-HT1A binding activity. These results are consistent with our
previous findings in series 1, which suggested that the presence of
a substituent at the 2-phenyl position (R1) is important for potency
at NET.6 Addition of a fluoro group at the R1 position provides a
modest improvement in potency at NET (2b), but a slightly larger
substituent imparts a more significant enhancement of NET po-
tency, with chloro, methoxy, and trifluoromethoxy groups all being
well tolerated by both NET and 5-HT1A (2c–f). Compounds 2g–h
demonstrate that NET and 5-HT1A affinities are not significantly af-
fected by the 2,3-dihalo substitution pattern (R1, R2 = Cl, F); how-
ever, placement of a chloro group at the 3-position significantly
enhanced potency at SERT. We noted that the 5-HT1A affinity is
not significantly affected by the addition of substituents at the
4-phenyl position (R3), whereas the transporters are particularly
sensitive to modifications at this position. Thus, compounds 2i–j
(R3 = Me, Cl) are significantly less potent at NET and more potent
at SERT than their respective parent compounds (2b, 2c); compar-
ison with 2k confirms that the sensitivity of substitution at the R3
position is primarily a steric effect. Substitution at the R4 position
with a fluorine atom (2l) results in a slightly weaker 5-HT1A bind-
ing, but does not significantly affect potency at the transport-
ers.Lipophilic efficiency (LipE) has been recently described as a
parameter that incorporates both potency and lipophilicity (e.g.,
c log P, c log D, or measured log D).16,17 For assessment of series 1
and 2, we calculated lipophilic efficiency using the following
equation:
clearance and also potentially enhancing safety margins in the ser-
ies.7 Following a preliminary survey of heterocyclic replacements
for the phenyl groups in 1, we focused our medicinal chemistry ef-
forts on the novel phenoxy pyridyl series 2 and 3.
The synthesis of compounds 2a–p was achieved in straightfor-
ward fashion beginning with commercially available N-Boc-pipe-
ridinone (4, Scheme 1). Treatment of
4 with LDA and N-
phenyltriflimide provided the vinyl triflate 5. Subsequent reaction
of 5 with bis(neopentylglycolato)diboron in the presence of cata-
lytic Pd(dppf)Cl2 yielded vinyl boronate 6 as the key Suzuki cou-
pling partner required for preparation of piperidine analogs 2a–p.
Commercially available 3-bromo-2-chloropyridine (7) also served
as a key starting material in the synthetic route toward 2a–p
(Scheme 2). The addition of phenols 8 to 7 to form the 2-phenoxy
pyridyl intermediate 9 was initially explored using K2CO3 in DMF
at 150 °C. Although these reaction conditions were generally useful
for the preparation of analogs in this series, sluggish reactions and
low conversions were observed with some sterically hindered (e.g.,
2,6-disubstituted) or electron-deficient phenols. In these cases,
alternate reaction conditions employing Cs2CO3 in DMSO at
120 °C provided superior results (68–96% yield). The Suzuki cou-
pling of boronate ester 6 and 3-bromo-2-phenoxypyridine 9 pro-
ceeded in good yields using Pd(dppf)Cl2 and CsF in DMF at
100 °C. Finally, hydrogenation of the alkene and removal of the
Boc group proceeded smoothly under standard conditions to pro-
vide analogs 2a–p. Our synthetic route toward piperazine analogs
3a–b also took advantage of the key 3-bromo-2-phenoxy interme-
diate 9 (Scheme 2). In this case, palladium-mediated amination of
9 with N-Boc-piperazine was accomplished using Pd2dba3, Dave-
Phos, and potassium t-butoxide in toluene at 60 °C.8
LipE ¼ ꢀlog Ki ꢀ clog D12
ð1Þ
It is noteworthy that numerous compounds in series 2 achieve
high NET and 5-HT1A binding affinities while reducing c log P by 1–
2 units compared to compound 1 (Table 1). Thus, many compounds
in series 2 show significant improvement in LipE compared to 1. It
is also of note that compounds in phenoxy pyridyl series 2, unlike
their counterparts in phenoxy phenyl series 1, are generally devoid
of activity at DAT. The structural basis for selectivity over DAT in
this series is not well understood. Nevertheless, the modifications
in the substitution pattern on the phenyl ring that cause significant
changes in NET and SERT affinities do not impact DAT affinity.
In the course of optimizing for NET potency in the phenoxy pyr-
idyl series 2, analogs were computationally investigated by analyz-
ing their fit to the pharmacophore model previously built for the
norepinephrine transporter.18 This model consists of (1) a basic
nitrogen which interacts with a receptor site point, likely an acidic
hydrogen bond donating group; (2) a primary aromatic pocket
flanked by receptor walls or excluded volumes; and (3) a second
pocket that is presumed to be at least partially solvent-exposed,
as the SAR in this area is relatively insensitive to small structural
changes.
H
N
R1
R4
R1
R1
R4
Br
O
R2
R3
HO
R5
R2
R3
O
R2
R3
a or b
c-e
N
N
R5
R5
R4
8
9
2
f, e
H
N
N
R1
R4
R2
R3
O
N
R5
3
Mapping of series 2 analogs onto the model (Fig. 2a) indicates
that the piperidine nitrogen maintains a good interaction with
the receptor site point, the primary aromatic site is occupied by
the phenyl group while the pyridine is directed into an area of
Scheme 2. Reagents and conditions: (a) 2-chloro-3-bromopyridine (7), K2CO3,
DMF, 150 °C, 8–94%; (b)Cs2CO3, DMSO, 120 °C, 68–96%; (c) 6, Pd(dppf)Cl2, CsF, DMF,
100 °C, 33–85%; (d) Pd/C, H2, EtOH, 68–98%; (e) AcCl/MeOH or HCl/MeOH, rt, 49–
91%; (f) Pd2dba3, Dave-Phos, t-BuOK/toluene, 60 °C, N-Boc-piperazine, 60–82%.

1116
A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1114–1117
Table 1
Binding affinities at 5-HT1A receptor, NET, DAT, and SERT
Compound
R1
R2
R3
R4
R5
5-HT1A Kia (nM)
NET Kib (nM)
DAT Kib (nM)
SERT Kib (nM)
c log Pc
LipEd (5-HT1A
)
LipEd (NET)
1
4.6
101
13
16
92
60
71
37
173
5.5
37
29
17
24
11
7.2
10
4.5
3.7
32
278
4310
>6180
5150
6410
>6170
>6650
310
>6650
>6140
>5500
2590
1260
1890
728
57
610
292
203
2.0
4.1
2.5
2.5
3.0
2.1
3.0
3.4
3.6
2.6
3.0
3.7
2.6
3.2
2.2
3.0
2.4
3.0
3.2
3.2
6.6
7.1
8.0
7.5
8.4
6.9
6.2
6.5
6.7
7.9
6.5
7.3
7.3
8.9
8.1
8.1
7.6
7.2
7.7
5.8
5.7
6.3
6.8
7.9
6.4
5.9
6.7
5.7
5.7
5.5
6.6
6.2
7.9
7.6
6.8
7.2
5.0
5.5
2a
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
2m
2n
2o
2p
3a
3b
H
F
H
H
H
H
H
H
Cl
F
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
Cl
F
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
F
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OMe
Cl
F
Me
H
2440
536
74
292
330
140
25
1810
921
413
143
247
276
33
Cl
OMe
Me
CF3
Cl
F
F
Cl
F
Cl
F
F
F
F
109
36
2.7
30
>6220
6410
251
887
194
1850
102
>4710
>4710
>6180
>6630
>6370
>6320
>6320
181
27
747
663
Cl
F
Cl
a
b
c
Serotonin1A RBSHA binding assay; Ki values are the mean of at least two experiments carried out in duplicate.9
Monoamine transporter binding scintillation proximity assay (SPA); Ki values are the mean of at least two experiments carried out in duplicate.10
Calculated logarithm of octanol/water partition coefficient.11
LipE = ꢀlog Ki ꢀ c log D.12
d
the steric restriction in the model. Phenyl is open to substitution in
the R1 position but R3 is limited to small substitution, as bigger
groups could project onto the sterically sensitive area and result
in decrease in binding affinity. Hence, the introduction of a methyl
and chlorine substituents at R3 resulted in compounds with de-
creased NET activity (e.g., 2i–j) as these substituents come close
to the receptor wall and cause unfavorable steric interactions. On
the contrary, the presence of a substituent at R1 seems to be impor-
tant for modulating NET activity. Hence 2c–d, which are substi-
tuted with chlorine or methoxy groups, respectively, are well
accommodated in the active region and account for the good NET
activity generally associated with the presence of small substitu-
ents at this position. It is to be noted that while a methyl or chlo-
rine at the 3-position reduced NET affinity, potency at SERT is
enhanced. This can be explained by the spatial information pro-
vided by the SERT pharmacophore model.18 As shown in Figure
2b, these R3-groups are well accommodated in the primary aro-
matic pocket for SERT. Moreover, previous findings indicated that
the presence of an electronegative functional group in this aro-
matic region is favorable for SERT binding affinity.19
In addition to our studies on series 2 compounds, we also inves-
tigated series 3 compounds in which the piperidine is replaced by a
piperazine. Compounds 3a and 3b (Table 2) exemplify our findings
in this series. Although the piperazine system provides a slight
improvement in 5-HT1A binding affinity, this structural modifica-
tion also causes a dramatic decrease in NET binding affinity. These
data are consistent with our pharmacophore model, in which a ba-
sic amine is a key element required for potency against NET. The
reduced basicity of the piperazines (calculated pKa = 7.9)20 in com-
parison with the piperidine (calculated pKa = 10.2) likely results in
weakening of the key interaction between the amine moiety and
the transporter.
Ultimately, the best overall pharmacological profile in series 2 is
achieved with a 2,6-disubstitution pattern in which one of the sub-
stituents is either F or Cl (2m–p). This subset of compounds
showed the greatest improvement in LipE in comparison to lead
compound 1. For example, compound 2n, in which the c log P is
reduced by one unit compared to 1, achieves LipE of 8.1 and 7.6
for 5-HT1A and NET, respectively. Compound 2n was profiled in
our panel of functional assays and compared favorably to com-
pound 1 with respect to 5-HT1A and NET functional activities and
Figure 2. Mapping of compound 2j onto the (a) NET and (b) SERT pharmacophores
showing sterically disfavored areas (excluded volumes shown in orange and yellow
for NET and SERT, respectively). A receptor interaction site point placed 2.8 Å
(optimum hydrogen bonding distance) from the basic amine is shown in gray.
the model that is currently not well defined. The observed SAR
trends with various substitutions on the phenyl are consistent with

A. B. Dounay et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1114–1117
1117
Table 2
Functional activity at 5-HT1A receptor, NET, DAT, and SERT
Compound 5-HT1A
EC50
5-HT1A
IAa
%
NET EC50b (nM) DAT EC50b (nM) SERT EC50b (nM) DAT EC50/NET EC50 SERT EC50/NET EC50 CL int, appc
(ll/min/mg)
1
2n
341
89
84
37
28
9
498
6880
915
409
18
764
33
45
9.7
<8.0
a
b
c
Serotonin1A GTPc
S functional assay; EC50 and intrinsic activity (IA) values are the mean of at least two experiments carried out in duplicate.13
Monoamine functional assays; EC50 and intrinsic activity (IA) values are the mean of at least two experiments carried out in duplicate.14
Human liver microsome clearance.15
3. Dannon, P. N.; Iancu, I.; Grunhaus, L. Hum. Psychopharmacol. 2002, 17, 329.
4. Weikop, P.; Kehr, J.; Scheel-Krüger, J. J. Psychopharmacol. 2007, 21, 795.
5. Campbell, B. M., Pfizer Global Research and Development, unpublished results.
6. Gray, D. L.; Xu, W.; Campbell, B. M.; Dounay, A. B.; Barta, N. S.; Boroski, S.;
Denny, L.; Evans, L.; Stratman, N.; Probert, A. Bioorg. Med. Chem. Lett. 2009, 19,
6604.
7. Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo, G. A.; Devraj, R. V.;
Ellsworth, E.; Fobian, Y. M.; Gibbs, M. E.; Gilles, R. W.; Greene, N.; Huang, E.;
Krieger-Burke, T.; Loesel, J.; Wager, T.; Whiteley, L.; Zhang, Y. Bioorg. Med.
Chem. Lett. 2008, 18, 4872.
monoamine transporter selectivity (Table 2). This compound was
further profiled using ex vivo receptor occupancy studies.21 Use
of this technique to study compound 2n demonstrated that it is
brain penetrant and binds to the target receptors in rats following
a 10 mg/kg subcutaneous injection. Ex vivo NET and 5-HT1A recep-
tor occupancies were determined to be 75.3 1.7% (n = 4) and
37.4 5.1% (n = 4), respectively.22 These values are consistent with
NET and 5-HT1A agonist occupancy values required to drive in vivo
functional activity.21,23–25
In summary, our efforts on the synthesis and pharmacological
evaluation of phenoxy pyridyl series 2 have led to the discovery
of novel compounds with a favorable NRI and 5-HT1A partial ago-
nist pharmacological profile as well as excellent selectivity for
the norepinephrine transporter over the dopamine transporter.
SAR studies demonstrate a number of key features of this series:
(a) The NET and SERT potencies are sensitive to electronic and ste-
ric modifications on the phenyl ring, and affinities toward these
transporters can be tuned by careful choice of substituents; (b)
5-HT1A partial agonist properties are relatively insensitive to sub-
stituent modifications on the phenyl ring; (c) Owing to the pres-
ence of the pyridyl nitrogen, compounds in this series are devoid
of activity at DAT; (d) Incorporation of the pyridyl nitrogen lowers
8. Old, D. W.; Wolfe, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1998, 120, 9722.
9. Assay conducted as described in Graham, J. M.; Coughenour, L. L.; Barr, B. M.;
Rock, D. L.; Nikam, S. S. Bioorg. Med. Chem. Lett. 2008, 18, 489.
10. The affinity of test compounds for binding to human NET, DAT, and SERT were
assessed by measuring inhibition of binding to [3H]nisoxetine, [3H]WIN
35,428, and [3H]citalopram, respectively, using a scintillation proximity assay
similar to that described in published protocol Bymaster, F. P.; Beedle, E. E.;
Findlay, J.; Gallagher, P. T.; Krushinski, J. H.; Mitchell, S.; Robertson, D. W.;
Thompson, D. C.; Wallace, L.; Wong, D. T. Bioorg. Med. Chem. Lett. 2003, 13,
4477. Transporters were obtained from membranes of HEK 293 cell lines stably
transfected with the human NET, DAT, or SERT. All binding assays were
conducted at room temperature in the presence of 180 mM NaCl with 30 mM
HEPES, pH 7.4 buffer with the test system consisting of 30
([3H]nisoxetine 12–20 nM, [3H]WIN 35,428 12–20 nM, or [3H]citalopram (2–
3 nM), 0.5 of test compound, vehicle control, or nonspecific binding
component (10 M desipramine for NET, 10 M nomifensine for DAT, and
10 M fluoxetine for SERT), and 30 l of cell membrane coupled to SPA beads
(12 g/well membrane protein; 0.5 mg/well WGA PVT SPA Beads (GE
ll of radioligand
ll
l
l
l
l
l
Healthcare BioSciences Corp., Piscataway, NJ)).
11. Values for c log P calculated using the BIOBYTE (www.biobyte.com) program
c log P, version 4.3.
12. Log D at pH 7.4 calculated using ACD version 9.3.
13. Assay conducted as described in published protocol Newman-Tancredi, A.;
Assie, M.-B.; Martel, J.-C.; Cosi, C.; Slot, L. B.; Palmier, C.; Rauly-Lestienne, I.;
Colpaert, F.; Vacher, B.; Cussac, D. Br. J. Pharmacol. 2007, 151, 237.
14. Monoamine functional assay conducted as described in Ref. 6.
15. Assay method adapted from published protocols (a) Riley, R. J.; McGinnity, D.
F.; Austin, R. P. Drug Metab. Dispos. 2005, 33, 1304; (b) Obach, R. S. Drug Metab.
Dispos. 1999, 27, 1350.
c
log P, and thus affords improved LipE, while also reducing micro-
somal clearance in the series. Compound 2n, a leading compound
in the series achieves excellent potency at NET and 5-HT1A, a
5-HT1A partial agonist profile, and selectivity for NET over both
DAT and SERT.
Acknowledgments
16. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
17. For a recent report on the use of LipE for lead series assessment, see Ryckmans,
T.; Edwards, M. P.; Horne, V. A.; Correia, A. M.; Owen, D. R.; Thompson, L. R.;
Tran, I.; Tutt, M. F.; Young, T. Bioorg. Med. Chem. Lett. 2009, 19, 4406.
18. Collantes, E. M.; Ortwine, D. F. Abstracts of Papers, 238th National Meeting of
the American Chemical Society: Washington, DC, Aug 16–20, 2009; COMP-229.
19. Butler, S. G.; Meegan, M. J. Curr. Med. Chem. 2008, 15, 1737.
20. Calculated pKa values obtained using ACD version 9.3.
We would like to thank Michael Stier for coordinating chemis-
try outsourcing efforts and David Favor for assistance with analog
preparation.
Supplementary data
21. Grimwood, S.; Hartig, P. R. Pharmacol. Ther. 2009, 122, 281.
22. Ex vivo receptor occupancy data were collected at 1 h post-dose following sc
administration. Drug plasma concentration = 777 144 ng/ml (n = 4). The data
are listed as mean SEM. Assay protocol is described in Ref. 6.
23. Takano, A.; Gulyás, B.; Varrone, A.; Maguire, R. P.; Halldin, C. Eur. J. Nucl. Med.
Mol. Imaging 2009, 36, 1308.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.12.023.
References and notes
24. Passchier, J.; van Waarde, A.; Pieterman, R. M.; Willemsen, A. T. M.; Vaalburg,
W. Psychopharmacology 2001, 155, 193.
25. Nakayama, T.; Suhara, T.; Okubo, Y.; Ichimiya, T.; Yasuno, F.; Maeda, J.; Takano,
A.; Saijo, T.; Suzuki, K. Psychopharmacology 2002, 165, 37.
1. Volkow, N. D.; Swanson, J. M. Am. J. Psychiatry 2003, 160, 1909.
2. (a) Chuluunkhuu, G.; Nakahara, N.; Yanagisawa, S.; Kamae, I. Kobe J. Med. Sci.
2008, 54, E147; (b) Schatzberg, A. F. J. Clin. Psychiatry 2000, 61, 31.