Pharmacophore-based discovery, synthesis, and biological evaluation of 4-phenyl-1-arylalkyl piperidines as dopamine transporter inhibitors

Article abstract of DOI:10.1016/S0960-894X(00)00703-4

Pharmacophore-based discovery, synthesis, and structure-activity relationship (SAR) of a series of 4-phenyl-1-arylalkyl piperidines are disclosed. These compounds have been evaluated for their ability to inhibit reuptake of dopamine (DA) into striatal nerve endings (synaptosomes). The lead compound 5 and the most potent analogue 43 were found to have significant functional antagonism.

Full text of DOI:10.1016/S0960-894X(00)00703-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 495±500
Pharmacophore-Based Discovery, Synthesis, and Biological
Evaluation of 4-Phenyl-1-arylalkyl Piperidines as Dopamine
Transporter Inhibitors
Sukumar Sakamuri,a,b Istvan J. Enyedy,a,c,d Alan P. Kozikowski,a,b Wahiduz A. Zaman,e
Kenneth M. Johnsone and Shaomeng Wanga,c,d,
*
aDrug Discovery Program, Georgetown University Medical Center, 3900 Reservoir Road, NW, Washington, DC 20007-2197, USA
bDepartment of Neurology, Georgetown University Medical Center, 3900 Reservoir Road, NW, Washington, DC 20007-2197, USA
cDepartment of Oncology, Georgetown University Medical Center, 3900 Reservoir Road, NW, Washington, DC 20007-2197, USA
dDepartment of Neuroscience, Georgetown University Medical Center, 3900 Reservoir Road, NW, Washington, DC 20007-2197, USA
eDepartment of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, TX 77555, USA
Received 22 September 2000; accepted 5 December 2000
AbstractÐPharmacophore-based discovery, synthesis, and structure±activity relationship (SAR) of a series of 4-phenyl-1-arylalkyl
piperidines are disclosed. These compounds have been evaluated for their ability to inhibit reuptake of dopamine (DA) into striatal
nerve endings (synaptosomes). The lead compound 5 and the most potent analogue 43 were found to have signi®cant functional
antagonism. # 2001 Elsevier Science Ltd. All rights reserved.
Because of its powerful reinforcing properties, (R)-
cocaine (1) has great abuse potential. The level of cocaine
abuse has reached epidemic proportions worldwide in
recent years, and immediate therapies are needed for its
treatment.1 Cocaine reinforcing and stimulant properties
have been associated with its ability to bind to mono-
amine transporter systems, particularly the dopamine
transporter (DAT). Several studies have shown that
compounds with DAT inhibitory activity may serve as
potential therapeutic agents for cocaine abuse treat-
ment.1c While no speci®c pharmacotherapy is currently
available in clinic, two potent DAT inhibitors are now
in clinical trials for the treatment of cocaine abuse.2
Over the last 20 years, extensive chemical and pharma-
cological studies have been performed on several classes
of DAT inhibitors, including tropanes (mainly com-
pounds of the WIN series, 2), benzotropanes, piper-
azines (also called GBR series, 4) and most recently
piperidines (3).3
agonists',4 and have used for this purpose a novel 3D-
database pharmacophore searching approach. Novel
DAT inhibitors are then evaluated as potential cocaine
antagonists in an in vitro functional antagonism assay.
In this functional antagonism assay, the IC50 value of
cocaine in the presence of approximate IC10 to IC50
concentrations of candidate antagonist compounds was
then compared to the IC50 value of cocaine alone. Sig-
ni®cant di€erences in IC50 values were compared to
theoretical IC50 values expected from models of `same site'
antagonism.5 IC50 values greater than those expected for
`same site' antagonism were taken as evidence of func-
tional antagonism. This test was performed under pre-
incubation conditions to allow slowly equilibrating
compounds to reach equilibrium. Further, any artifactual
di€erences in IC50 (Ki) due to di€erences in tempera-
ture, tissue preparation, etc. were negated in this assay
as binding of cocaine and the putative antagonists to
both the cocaine binding site and the transporter
occurred under identical conditions. This insures that a
right-shift in the cocaine inhibition curve beyond what
is expected for two drugs acting at the same site is a true
measure of functional antagonism.
We are interested in the discovery of novel DAT inhibitors
that can be used as either cocaine antagonists or `partial
We have recently reported the discovery of hydroxy-
piperidines as a novel class of DAT inhibitors with
*Corresponding author. Tel.: +1-202-687-2028; fax: +1-202-687-0617;
e-mail: wangs@giccs.georgetown.edu
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00703-4

496
S. Sakamuri et al. / Bioorg. Med. Chem. Lett. 11 (2001) 495±500
signi®cant functional antagonism against cocaine.4
Furthermore, using the same pharmacophore model,
several additional classes of novel DAT inhibitors were
discovered.6 Compound 5 represents one such novel
lead compound. Compound 5 was found to have a Ki
value of 1.56 mMin the inhibition of DA reuptake
(Table 1), a relatively weak DAT inhibitor. However,
when 5 was tested in our functional antagonism assay as
described previously,4 it exhibits a signi®cant functional
antagonism. For example, in the presence of 5 (500 or
1000 nM), the experimental IC50 values of cocaine in the
inhibition of DA reuptake were signi®cantly increased
(Table 2). These values are signi®cantly greater than the
theoretical values calculated using the same binding-site
model.5 These data thus suggested that 5 has signi®cant
functional antagonism against cocaine. Therefore,
despite its weak potency, 5 may represent an interesting
lead for further chemical modi®cations.
using the QUANTA/CHARMm program.7 Upon the
completion of conformation generation and energy
minimization, the most stable conformation was identi-
®ed (the global minimum in vacuum). It is noted, how-
ever, that the lowest energy conformation may not be
the bioactive conformation. For this reason, other low
energy conformations, typically within 5 kcal/mol of the
global minimum were identi®ed. It was found that 5 and
WIN 35065 have a very good overlap with their low
energy conformations with respect to the two binding
elements de®ned in the pharmacophore model (i.e. the
nitrogen atom and the phenyl ring at the 4-position of
the piperidine 5 overlap very well with the correspond-
ing structural elements of WIN 35065). The N-sub-
stituent of piperidine 5 has two di€erent orientations,
and the energy di€erence is less than 1 kcal/mol between
these two conformations. One conformation has rea-
sonable overlap with the third binding element, the
methyl ester in the WIN compound (Fig. 2a). Thus, our
molecular modeling suggested that despite the structural
di€erences between 5 and WIN 35065, they may interact
with DAT in a similar binding mode, although it is
important to keep in mind that these calculations were
done in vacuo.
Since the pharmacophore model (Fig. 1) used to identify
the lead compound was developed based upon cocaine
and WIN compounds, we carried out molecular model-
ing studies to investigate the overlap between the lead
compound 5 and WIN 35065. Conformational analysis
(generation and energy minimization) was performed
Compound 5 has a weak potency and its potency has to
be improved signi®cantly in order to have any ther-
apeutic value. For this purpose, we have carried out
SAR studies on this class of compounds. Compound 5
may be divided into three basic structural elements, the
piperidine ring, the 4-phenyl substituent (head), and the
phenyl ring (tail) tethered by a three-carbon linker.
Accordingly, we made chemical modi®cations to the
head, the tail, and the linker. The resulting compounds
were evaluated for their ability to inhibit DA reuptake.
The most potent compound was tested for its functional
antagonism. It is of note that some N-arylalkyl-sub-
stituted tropanes with DAT inhibitory activity,8 and
piperidinylbutyrophenones with potent neuroleptic
activity have been reported.9a The SAR studies and the
evaluation of the in vitro functional antagonism against
cocaine of the most potent compound are the central
focus of this report.
Figure 1. A 3D pharmacophore model derived from cocaine and WIN
35065 compounds used in 3D database pharmacophore search.

S. Sakamuri et al. / Bioorg. Med. Chem. Lett. 11 (2001) 495±500
497
The synthesis of compounds, outlined in Scheme 1, was
accomplished by standard procedures.9 4-Piperidone
hydrochloride was converted to its N-benzyl derivative
7 using benzyl bromide and potassium carbonate in
DMF at 70 ꢀC in 70% yield. 1-Benzyl-4-piperidone (7)
was reacted with suitably substituted aryl Grignard
reagents in THF at 0 ꢀC to give compounds 8, 9, or 10
in good yields. One-step removal of the 4-hydroxyl and
N-benzyl groups under hydrogenolysis conditions
always gave mixtures. The 4-hydroxyl group was there-
fore ®rst removed in a two-step sequence to give com-
pounds 14 and 15 in quantitative yield. Compound 16
was prepared by debenzylation of compound 8 in 92%
yield. To avoid the problem of reduction of the chlorine
substituent under palladium catalyzed hydrogenation
conditions in compound 13, we have chosen Wilk-
inson's catalyst for the double bond reduction and 1-
chloroethyl chloroformate for N-debenzylation.10
Accordingly, compound 18 was prepared in good yield
by reduction of the double bond in compound 13 using
Wilkinson's catalyst, followed by the N-debenzylation
using 1-chloroethyl chloroformate. Compounds 14, 15, 16,
and 18 were alkylated with various arylalkyl halides in
DMF using potassium carbonate at 60 ꢀC to give com-
pounds 32±43. The alkylation yields vary from 53 to 87%
depending on the arylalkyl halide used (Scheme 3 and
Table 1). The preparation of arylalkyl halides 21, 22, 25, 28,
and 31 is shown in Scheme 2; the remaining arylalkyl
halides were purchased from commercial suppliers.
The synthesized compounds were tested for their ability
to inhibit the reuptake of DA, and the data are sum-
marized in Table 1.4 Compound 32 di€ers from the lead
compound 5 only by a 4-¯uoro substituent in the `head'
phenyl group and is slightly less potent than 5. To
investigate the e€ect of the length of the linker, com-
pound 33 (one carbon shorter than 32) and 34 (one
carbon longer than 32) were synthesized and tested.
Both compounds 33 and 34 are less potent than 32 but
the e€ect is moderate. To investigate the e€ect of the
carbonyl group in the linker, the carbonyl group in 32
Figure 2. Superposition of two low energy conformations of lead
compound 5 (yellow), and compound 43 (green) on WIN 35065 com-
pound (gray).
Scheme 1. Reagents and conditions: (a) K2CO3, PhCH2Br, DMF, 70 ꢀC, 70%; (b) (p-X)PhMgBr, THF, 0 ꢀC; (c) P2O5, toulene, re¯ux; (d) H2, Pd/C,
EtOH; (e) (Ph3P)3RhCl, H2, toulene, 70 ꢀC, 91%; (f) (1) 1-chloroethyl chloroformate, 1,2-dichloroethane, re¯ux, 2 h; (2) MeOH, re¯ux, 82% over
two steps.

498
S. Sakamuri et al. / Bioorg. Med. Chem. Lett. 11 (2001) 495±500
Scheme 2. Reagents and conditions: (a) K2CO3, DMF, 1,2-dibromoethane, 70 ꢀC; (b) 1 MBH ꢁ THF, THF, 0 ꢀC to re¯ux, 90%; (c) PBr3, ether,
3
0 ꢀC; (d) SOCl2, MeOH, 0 ꢀC; (e) 1 MDIBAL-H, CH 2Cl2, À78 ꢀC, 85%; (f) Pd/C, H2, EtOH, 100%; (g) Ph3PˆCHCO2Et, CH2Cl2, room tem-
perature, 64%; (h) DIBAL-H, CH2Cl2, 0 ꢀC, 81%; (i) Ph3P, CBr4, CH2Cl2, room temperature, 76%.
was replaced by a methylene group, and the resulting
compound 35 was found to be slightly more potent than
32. To further investigate the e€ect of the length of the
linker, compounds 36±38 were synthesized and tested.
Compound 36 with a one-carbon linker is at least 5-fold
less potent than 35, suggesting the importance of the
length of the linker. While compound 37 with a two-
carbon linker has a potency comparable to that of 35,
compound 38 with a four-carbon linker is slightly less
potent than 35. These data suggest that the optimal
length of the linker is either two- or three-carbon atoms,
although the e€ect is only marginal except for com-
pound 36. To investigate the electronic e€ect of the lin-
ker, the carbonyl group in the linker was replaced by an
oxygen atom. The resulting compound 39 was 3-fold
more potent than 32. To investigate the e€ect of the 4-
¯uoro substituent in the `head' phenyl ring, compound
40 without the 4-¯uoro substituent was synthesized and
found to be 3-fold less potent than 39. An additional
hydroxyl group at the 4-position of the piperidine ring
has a detrimental e€ect since compound 41 is 7-fold less
potent than compound 40 without the hydroxyl group.
It is known from the literature that a para chloro sub-
stituent in the phenyl ring of the WIN series is optimal
for activity.11 To investigate the e€ect of the sub-
stituents on both the `head' and the `tail' phenyl rings,
compounds 42 and 43 were synthesized and tested.12
Compound 42 with a 4-chloro substituent in the `tail'
phenyl ring is slightly more potent than compound 39
with a 4-¯uoro substituent. However, compound 43
with a 4-chloro substituent in both the `head' and the
`tail' phenyl rings is 9-fold more potent than compound
39 and 7-fold more potent than compound 42. These
data suggest that a substituent on the `head' phenyl ring
signi®cantly improves the potency of the compounds
relative to the substituent on the `tail' phenyl ring. 4-
Chloro substituted compound 43 is a reasonably potent
DAT inhibitor with a Ki value of 90nM, and is 17-fold
more potent than the lead compound 5.
The most potent compound 43 was then tested for its
functional antagonism against cocaine. It was found
that in the presence of 20 and 50 nMof 43, the IC50
values of cocaine in inhibition of DA reuptake were
increased and these values are signi®cantly greater than
the calculated theoretical values, if one assumes that
cocaine and 43 bind to the same binding site at the DAT
(Table 2). Therefore, compound 43 is not only a rea-
sonably potent DAT inhibitor but also possesses sig-
ni®cant functional antagonism. It is of interest that
modi®cation of the structure had no e€ect on antag-
onism per se. That is, although there was about a 17-
fold di€erence between the potencies of 5 and 43,
there is a similar degree of right-shift in the cocaine
inhibition curve when these compounds are tested at
similar concentrations relative to their own Ki. The
precise mechanism of the functional antagonism is not
clear.
In conclusion, a lead compound (5) with moderate
activity but signi®cant functional antagonism was dis-
covered using a pharmacophore-based 3D-database
searching approach. Through chemical modi®cations,
we identi®ed a new analogue (compound 43), which is
17-fold more potent than the lead compound in the
inhibition of DA reuptake with a Ki value of 90 nM.
Similar to the lead compound, this more potent analogue
43 was found to have signi®cant functional antagonism.
Further pharmacological and behavioral studies are
under way to investigate the mechanism of functional
antagonism of 43, and its therapeutic potential for the
treatment of cocaine abuse.

S. Sakamuri et al. / Bioorg. Med. Chem. Lett. 11 (2001) 495±500
499
Scheme 3.
Table 1. Chemical structures of DAT inhibitors and their potency as inhibitors of DA reuptake
R2-Br or R2-Cl
R1
X
Yield (%)
Compd
[3H]DA uptake Ki (mM)
H
H
H
F
Ð
60
Lead (5)
32
1.56 Æ 0.06
2.67 Æ 0.39
H
H
F
F
68
73
33
34
4.33 Æ 0.28
3.17 Æ 0.19
H
H
H
H
F
F
F
F
53
82
59
66
35
36
37
38
1.86 Æ 0.01
>1 0
1.36 Æ 0.02
2.54 Æ 0.02
H
H
OH
F
H
H
68
63
54
39
40
41
0.84 Æ 0.08
2.72 Æ 0.03
17.55 Æ 1.53
H
H
F
Cl
87
71
42
43
0.63 Æ 0.06
0.09 Æ 0.01
Table 2. Experimental and theoretical IC50 values of cocaine in the presence of lead compound 5 and compound 43 for [3H]-dopamine uptake by
striatal membrane preparations
[3H]-DA uptake
Experimental conditions
Experimental IC50 (mean Æ SEM)
Theoretical IC50 (mean Æ SEM)
Cocaine
297 Æ 22 nM
Cocaine + lead (5) (500 nM)
Cocaine + lead (5) (1000 nM)
Cocaine + compound 43 (20 nM)
Cocaine + compound 43 (50 nM)
538 Æ 25 nM383
686 Æ 4 nM471
490 Æ 74 nM354
696 Æ 32 nM441
Æ 13 nM
Æ 13 nM
Æ 17 nM
Æ 22 nM
Acknowledgements
References and Notes
The authors are indebted to the National Institute on
Drug Abuse (NIDA), National Institutes of Health for
their ®nancial support of this work (DA-11545). The
lead compound used in the initial screening was pro-
vided by the Chemistry Branch, Developmental Ther-
apeutics Program, National Cancer Institute, NIH.
1. (a) Ritz, M. C.; Lamb, R. J.; Goldberg, S. R.; Kuhar, M. J.
Science 1987, 237, 1219. (b) Carroll, F. I.; Lewin, A. H.; Bis-
was, J. Pharm. News 1994, 1, 11. (c) Carroll, F. I.; Howell,
L. L.; Kuhar, M. J. Med. Chem. 1999, 42, 2721.
2. NIDA Research Report. Cocaine Abuse and Addiction:
NIH Publication No. 99-4342, May 1999.

500
S. Sakamuri et al. / Bioorg. Med. Chem. Lett. 11 (2001) 495±500
3. (a) Smith, M. P.; Hoepping, A.; Johnson, K. M.; Trzcinska,
M.; Kozikowski, A. P. Drug Discovery Today 1999, 7, 322. (b)
Singh, S. Chem. Rev. 2000, 100, 925. (c) Zhao, L.; Johnson, K. M.;
Zhang, M.; Flippen-Anderson, J.; Kozikowski, A. P. J. Med.
Chem. 2000, 43, 3283. (d) Sakamuri, S.; George, C.; Flippen-
Anderson, J.; Kozikowski, A. P. Tetrahedron Lett. 2000, 41, 2055.
4. Wang, S.; Sakamuri, S.; Enyedy, I. J.; Kozikowski, A. P.;
Deschaux, O.; Bandyopadhyay, B. C.; Tella, S. R.; Zaman,
W. A.; Johnson, K. M. J. Med. Chem. 2000, 43, 351.
5. (a) Slusher, B. S.; Ti€any, C. W.; Olkowski, J. L.; Jackson,
P. F. Drug Alcohol Depend. 1997, 48, 43. (b) Reith, M. E. A.;
Selmeci, F. Naunyn-Schmiedeberg's Arch. Pharmacol. 1992,
345, 309.
6. (a) Sakamuri, S.; Enyedy, I. J.; Kozikowski, A. P.; Wang,
S. Tetrahedron Lett. 2000, 41, 9949. (b) Wang, S.; Sakamuri,
S.; Enyedy, I. J.; Kozikowski, A. P.; Flippen-Anderson, J.;
Farkas, T.; Zaman, W. A.; Johnson, K. M. J. Med. Chem.,
submitted for publication.
7. (a) QUANTA, a molecular modeling system, is supplied by
Molecular Simulations Inc., 9685 Scranton Road, San Diego,
CA 92121±3752, USA (b) Brooks, B. R.; Bruccoleri, R. E.;
Olafson, B. D.; States, D. J.; Swaminathan, S.; Jaun, B.; Kar-
plus, M. J. Comput. Chem. 1983, 4, 187.
8. Efange, S. M. N.; Kamath, A. P.; Khare, A. B.; Kung, M.;
Mach, R. H.; Parsons, S. M. J. Med. Chem. 1997, 40, 3905.
9. (a) Iorio, M. A.; Reymer, T. P.; Frigeni, V. J. Med. Chem.
1987, 30, 1906. (b) Beilstein, Handbuch der Organischen Che-
mie, Band xx, p. 291.
10. Olofson, R. A.; Martz, J. T.; Senet, J. P.; Piteau, M.;
Malfroof, T. A. J. Org. Chem. 1984, 49, 2081.
11. (a) Carroll, F. I.; Gao, Y.; Rahaman, M. A.; Abraham, P.;
Parham, K.; Lewin, A. H.; Boja, J. W.; Kuhar, M. J. Med.
Chem. 1991, 34, 2719. (b) Carroll, F. I.; Mascarella, S. W.;
Kuzemko, M. A.; Gao, Y.; Abraham, P.; Lewin, A. H.; Boja,
J. W.; Kuhar, M. J. Med. Chem. 1994, 37, 2865.
1
12. Compound 43: H NMR (300 MHz, CDCl3) d 1.69±1.82
(4H, m), 2.17±2.26 (2H, m), 2.43±2.53 (1H, m), 2.83 (2H, t,
J=5.9 Hz), 3.10 (2H, d, J=11.7 Hz), 4.10 (2H, t, J=6.1 Hz),
6.84 (2H, dd, J=2.2 Hz, 6.9 Hz), 7.14 (2H, dd, J=1.7 Hz,
8.3 Hz), 7.21±7.28 (4H, m); 13C NMR (CDCl3) d 33.2, 41.8,
54.6, 57.4, 66.3, 115.8, 125.6, 128.1, 128.5, 129.3, 131.6, 144.6,
157.3; MS m/z (%) 348 (MÀ1, 4), 210 (64), 208 (100).