Structure-activity relationship of trihexyphenidyl analogs with respect to the dopamine transporter in the on going search for a cocaine inhibitor

Article abstract of DOI:10.1016/j.ejmech.2005.04.016

A series of trihexyphenidyl (THP) analogs were used to search for a derivative that could serve as a cocaine inhibitor. A compound that blocks binding of the cocaine analog carboxyfluorotropane (CFT), allows dopamine uptake and exhibits low side effects could serve as a good candidate for that purpose. All analogs were tested for the extent to which they inhibit CFT binding, dopamine uptake and n-methyl scopolamine (NMS) binding. Several structure-function relationships emerged. Methylation/halogenation of THP's benzene ring enhanced the compound's ability to block CFT binding in comparison to its ability to block dopamine uptake (5a-e). Replacement of the cyclohexyl ring with a benzene ring tended to create compounds that had lower affinities to the dopamine transporter (7b compared to THP, 7d compared to 5h, 7c compared to 8c) and modification of THP's piperidine ring tended to enhance affinity to the dopamine transporter (5f-h, 8a, 8c). One analog (5f) that showed little muscarinic activity indicating that it would probably have few side effects was investigated for its effects as an in vivo cocaine inhibitor. However, it showed few antagonistic effects in vivo. Nevertheless, this work greatly elucidates the structure-function relationships required for potential cocaine inhibitors and so lays out promising directions for future research.

Full text of DOI:10.1016/j.ejmech.2005.04.016

European Journal of Medicinal Chemistry 40 (2005) 1013–1021
www.elsevier.com/locate/ejmech
Original article
Structure–activity relationship of trihexyphenidyl analogs with respect to
the dopamine transporter in the on going search for a cocaine inhibitor
D.E. Dar a,*, M. Thiruvazhi b, P. Abraham b, S. Kitayama a, T.A. Kopajtic a, A. Gamliel c,
B.S. Slusher d, F.I. Carroll b, G.R. Uhl a,e
a Molecular Neurobiology Branch, Intramural Research Program, National Institute on Drug Abuse, National Institutes of Health, P.O. Box 5180,
Baltimore, MD 21224, USA
b Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC 277091, USA
c Quark Biotech Inc., Weizmann Science Park, P.O. Box 4071, Nes Ziona 70400, Israel
d Department of Research, Guilford Pharmaceuticals, 6611 Tributary Street, Baltimore, MD 21224, USA
e The Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
Received 13 April 2005; received in revised form 16 April 2005; accepted 18 April 2005
Available online 11 July 2005
Abstract
A series of trihexyphenidyl (THP) analogs were used to search for a derivative that could serve as a cocaine inhibitor. A compound that
blocks binding of the cocaine analog carboxyfluorotropane (CFT), allows dopamine uptake and exhibits low side effects could serve as a good
candidate for that purpose. All analogs were tested for the extent to which they inhibit CFT binding, dopamine uptake and n-methyl scopo-
lamine (NMS) binding. Several structure–function relationships emerged. Methylation/halogenation of THP’s benzene ring enhanced the
compound’s ability to block CFT binding in comparison to its ability to block dopamine uptake (5a–e). Replacement of the cyclohexyl ring
with a benzene ring tended to create compounds that had lower affinities to the dopamine transporter (7b compared to THP, 7d compared to
5h, 7c compared to 8c) and modification of THP’s piperidine ring tended to enhance affinity to the dopamine transporter (5f–h, 8a, 8c). One
analog (5f) that showed little muscarinic activity indicating that it would probably have few side effects was investigated for its effects as an
in vivo cocaine inhibitor. However, it showed few antagonistic effects in vivo. Nevertheless, this work greatly elucidates the structure–function
relationships required for potential cocaine inhibitors and so lays out promising directions for future research.
© 2005 Elsevier SAS. All rights reserved.
Keywords: Trihexyphenidyl; Dopamine; Cocaine; n-methylscopolamine; Transporter; Uptake; Binding
1. Introduction
the dopamine (DAT), serotonin (SERT) and norepinephrine
(NET) transporters. The good fit between DAT potencies and
behavioral reward described in their work, coupled with data
from lesion [2] and dopamine release studies [3–5], sug-
gested that DAT was the principal site for cocaine induced
reward [1].
Studies using site directed mutagenesis [6–9], molecular
modeling [10], binding, [11,12] and thermodynamic ap-
proaches [13] have all suggested that there may be separate
or overlapping, but not identical, cocaine and dopamine rec-
ognition sites on the DAT. Although the exact relationship of
each site with the DAT is not clearly understood, these stud-
ies indicate on the possibility to construct a small drug mol-
ecule that could impede cocaine recognition but still allow
normal dopamine uptake.
Cocaine abuse is a major continuing concern in the West-
ern world and there is currently no cocaine antagonist or
cocaine substitute available. Ritz et al. [1] examined the cor-
relations between psychostimulant potencies in tests of behav-
ioral reward and their potencies in blocking each of the
monoamine transporters at which cocaine displays activity:
Abbreviations: CFT, carboxyfluorotropane; CHO, Chinese hamster ovary;
DAT, dopamine transporter; NET, norepinephrine transporter; NMS,
n-methylscopolamine; SERT, serotonin transporter; THP, trihexyphenidyl.
* Corresponding author. Present address: Weizmann Institute of Science,
Rehovot 76100, Israel. Tel.: +972 8 934 4418; fax: +972 8 934 4131.
E-mail address: dalit.dar@weizmann.ac.il (D.E. Dar).
0223-5234/$ - see front matter © 2005 Elsevier SAS. All rights reserved.
doi:10.1016/j.ejmech.2005.04.016

1014
D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
In order to characterize the cocaine binding site on the DAT,
extensive investigations were conducted to examine the struc-
ture–function relationships of a series of cocaine- or carboxy-
fluorotropane (CFT)-based compounds [14–18]. Those stud-
ies focused on compounds that bind better to the DAT than to
the SERT or NET [19–21]. They identified, for example, sev-
eral 3-phenyltropane cocaine analogues that demonstrated a
slower onset and longer duration of action than cocaine
[22,23] and were proposed for cocaine replacement therapy
similar to methadone substitution for heroin addiction
[24–27]. They did not, however, report on compound/s that
can block cocaine recognition without impairing the DAT’s
normal dopamine uptake function.
The term cocaine selective inhibitor is used here to describe
a compound that block the cocaine site on the DAT better
than it block the dopamine site. To find a structure capable of
at least partially blocking the binding of cocaine, while allow-
ing for dopamine uptake, the activities of almost 100 non-
tropane compounds containing heteroatoms and aromatic fea-
tures compatible with possible interactions predicted from
initial DAT models were screened (reviewed in [27]). Initial
screening identified the muscarinic anticholinergic drug THP
as providing some selective blockage of cocaine analog rec-
ognition, while being less potent in terms of dopamine uptake
inhibition [28].
We now report data on the synthesis and structure–
function relationships of a series of novel THP congeners.
The pharmacological characteristics of these analogs in rat
brain membranes and in cells transfected with the dopamine
transporter under two different sets of conditions were com-
pared and a similar trend was revealed in all systems. Three
main types of modifications of THP’s basic structure (Fig. 1)
were performed: (1) methylation/halogenation of the ben-
zene ring; (2) replacement of THP’s cyclohexyl ring with a
benzene ring; and (3) insertions/deletions on THP’s piperi-
dine ring. The structure–function activities of THP conge-
ners are described and discussed in detail. The side effects
profile of a compound that retains some selective ability to
block cocaine while manifesting a much lower potency than
THP with respect to blocking muscarinic cholinergic recep-
tors was investigated using a 61-site NovaScreen assay. The
implications of the current study for future development of
small molecules as drugs targeting the dopamine transporter
site is discussed.
Scheme 1
.
priate acetamides (2) were treated with the corresponding
cyclohexyl substituted phenyl ketone (3) to provide the cor-
responding propionamides (4). The commercially unavail-
able acetamides (2) were prepared by acetylation of the appro-
priate amines (1). On reduction with borane dimethylsulfide
complex, the propionamides gave the target aminoalcohols
(5). An alternate route was employed to prepare derivatives 7
and 8a, 8c (Scheme 2). The substituted 3-aminopropionic acid
ethyl esters (6) were prepared by the action of ethyl acrylate
on the appropriate amines (1), which, on treatment with phe-
nyl magnesium bromide, provided the diphenyl amine alco-
hols (7). Careful reduction of one of the phenyl groups in a
hydrogen atmosphere over platinum oxide (IV) gave the tar-
get propanolamines (8a and 8c). The various substituents (R1,
R2, X and Y) on each product are identified in Table 1.
2. Chemistry
THP derivatives were prepared by two general methods.
As illustrated in Scheme 1, the lithium derivatives of appro-
Fig. 1
. The structure of THP. A–C refer to the phenylpropane, cyclohexyl
and piperidine rings of THP accordingly.
Scheme 2
.

D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
1015
Table 1
Physical properties of THP analogs
Congener
5a
5b
5c
5d
5e
5f
5g
5h
5i
7a
7b
7c
X (para)
Y (meta)
R3
R1 R2
Salts
Melting point (°C)
> 260
CH3
H
H
C6H11
C6H11
C6H11
C6H11
C6H11
C6H11
C6H11
C6H11
C6H11
C6H5
C6H5
C6H5
C6H5
C6H11
C6H11
–CH2(CH2)3CH2–
As above
HCl
CH3
H
As above
As above
As above
As above
As above
As above
As above
As above
As above
As above
2 HCl
243–245
> 260
Cl
H
As above
Cl
Cl
H
As above
242–243
255–256
187–189
190–191
222–224
217–219
202–204
189–195
235–237
196–197
182–184
235–237
Cl
H
As above
CH2CH2C(OH)(C6H5)CH2CH2
CH2CH2CH(C6H5)CH2CH2
CH2CH2CH(CH3)CH2CH2
CH2CH2CH2(C6H4)CH2CH2
CH2CH3 CH2CH3
CH2CH2CH2CH2CH2
CH2CH2N(CH3)CH2CH2
CH2CH2CH(CH3)CH2CH2
CH2CH3 CH2CH3
CH2CH2N(CH3)CH2CH2
H
H
H
H
H
H
H
H
H
H
H
H
7d
8a
8c
H
H
HCl
H
H
HCl
H
H
2 HCl
3. Results
ited increased selectivity, inhibiting [3H]CFT binding more
than [3H]dopamine uptake. They, thus produced higher val-
ues for the ratios of Ki ([3H]dopamine uptake) to Ki ([3H]CFT
binding) in DAT-expressing cells. Para substituents were gen-
erally more selective than meta substituents (Table 2, condi-
tion sets I and II; 5a compared to 5b; 5c compared to 5d).
Chloride-substituted compounds had higher potencies than
methyl-substituted compounds with respect to inhibiting both
[3H]dopamine uptake and [3H]CFT binding (Table 2, condi-
tion sets I and II; 5c compared to 5a; 5d compared to 5b). In
DAT-expressing cells using condition set I (Table 2), though
not in expressing cells using condition set II (Table 2) or rat
striatum (Table 2, condition set III), [3H]dopamine uptake
was inhibited nearly equipotently by the dicholoro-substituted
5e and the monochloro-substituted 5d compounds. The 3′4′
dichloro substitution increased potency with respect to
[3H]CFT binding inhibition by about three times, and im-
proved selectivity 22-fold in DAT-expressing cells (Table 2,
condition set I). In striatal tissue, halogenation of the phenyl
ring also lead to more potent inhibition of [3H]dopamine
uptake and [3H]CFT binding than methylation at the same
position (Table 2, condition set III; 5c compared to 5a; 5d
compared to 5b). The meta substituent was more potent than
the para substituents (Table 2, condition set III; 5b compared
to 5a; 5d compared to 5c).
3.1. Potencies of THP and its analogs in dopamine
transporter uptake and binding
Initial results from screening assays demonstrated that THP
could block the CFT binding with a significantly greater
potency than it exhibited in inhibiting dopamine uptake. Com-
pounds that were closely related structural analogs of THP
were then synthesized and tested (Table 1). Many of the pre-
vious studies that tested the activity of compounds at the DAT
used homogenates of rat brain membranes [29,30,32]. In this
study, each compound was tested for its activity at the DAT
under three sets of conditions. In condition set I, Chinese ham-
ster ovary (CHO) cells stably transfected with the human DAT
were used. An identical buffer composition was used in the
[3H]dopamine uptake and [3H]CFT binding experiments,
which were repeated three to five times each. Condition set II
was similar to condition set I, with minor modifications, and
the experiment was conducted only once. In condition set III,
homogenated rat brains were used. Different buffers were used
in the [3H]dopamine uptake and [3H]CFT binding experi-
ments, which were repeated two to three times each. Condi-
tion set I is, therefore, considered the most conservative.
Although the specific conditions of each assay influenced the
results, they did not usually change the overall tendency. The
potencies with which the congeners influenced dopamine
uptake and cocaine analog binding differed, with the ratios
of Ki for [3H]dopamine uptake to Ki of [3H]CFT binding rang-
ing from 2 to 3 up to more than 20 in DAT-expressing cells.
Table 2 shows Ki values for these THP analogs’ inhibition of
[3H]dopamine uptake and [3H]CFT binding. Inhibition curves
for some of the compounds are presented in Fig. 2.
Replacement of THP’s cyclohexyl ring (Fig. 1) with a phe-
nyl ring usually resulted in reduced affinity for both
[3H]dopamine uptake and [3H]CFT binding in DAT-
expressing cells. This was true when comparing data from 7b
to THP (Table 2, condition sets I and II), data from 7d to 5h
(Table 2, condition set I) and data from 7c to 8c (Table 2,
condition set I), although it did not hold for comparisons of
data from 7a to 8a (Table 2, condition sets I–III). Substitut-
ing an aromatic ring for the cyclohexyl ring also reduced affin-
Compounds with methyl- or halogen substituents placed
at meta or para positions on the phenyl ring (Fig. 1) exhib-

1016
D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
Table 2
Affinities of THP and related analogs for the human and rat dopamine transporter
Ki (µM) hDATI
Ki (µM) hDATII
[3H]dopamine [3H]CFT Ratio [3H]dopamine
35,428
Ki (µM) rDATIII
[3H]CFT 35,428 Ratio
Test Compound [3H]dopamine
[3H]CFT 35,428 Ratio
THP
5a
5b
5c
5d
5e
5f
5g
5h
5i
7a
7b
7c
7d
8a
8c
79.8 4.7
54.6 2.1
18.5 3.2
16.2 0.3
10.2 2.2
11.6 2.4
8.9 1.1
26.9 3.1
1.2
2.96
10.85
4.33
10.15
6.76
21.94
4
142
39.5
46
21
6.3
16
1.4
6.8 26.3 4.6
6.3 13.1
2.9 6.8 0.5
10.4 3.8
12.2 1.1
6.5 0.9
4.6 0.3
2.3 0.2
1.5 0.2
0.8 0.5
0.9 0.04
0.63 0.03
9.4 0.8
6.4 2.6
2.1
2
5
1
4.3 0.8
1.6 0.5
1.5 0.45
0.53 0.12
2.2 0.2
3.8 1.1
13.8 3.6
48 21.5
11.3 1.9
40.8 7.8
20.9 2.4
23.9 4.2
10.7 2.3
2.4 0.1
1.5
1.65
1.8
0.875
1.7
1.3
1.25
0.7
1.2
3.1
1.3
2.6
1.1
1
14.4
12
1
2.8
1.3
0.8
1.7
4.3 2.7 0.4
3.2 0.7 0.1
8.1 1.5 0.4
4.8 0.8 0.2
5.3 11.8 0.8
4.2
6.5
8
28.5 6.8
40.1 6.1
384 33.4
32.3 3.3
139 12.3
55.8 4.8
77.2 2.6
42.1 7.9
8.9 0.7
7.5
2.89
8
105
> 300
37
20
> 300
12.5
50
ND
4.4 0.5
14 0.8
2.86
3.39
2.67
3.23
3.93
3.7
3
12
2
> 100
44
ND
33.9 2.2
11.5 1.4
9.8 1.4
6.9 1.6
23.5
15
1.9 13.3 2.2
17.9
105
54
7
5
16
3.4 14.2 2.8
5.9 3.8 0.3
13.1
2
11
1.9
3.8 0.7
The affinities of the above compounds for the cocaine/CFT site and the dopamine site on the dopamine transporter were determined by inhibition of [3H]CFT
binding and [3H]dopamine uptake to CHO cells stably transfected with the human dopamine transporter (condition sets I and II) and on rat striatum (condition
set III). Data are presented as the mean S.E.M. values for three to five experiments (condition I); one to two experiments (condition II), two to three experi-
ments (condition III); ND = not determined.
ity for [3H]dopamine uptake in rat striatum (Table 2, condi-
tion set III; 7d compared to 5h; 7c compared to 8c).
Five out of the six THP analogs with modifications on the
nitrogen-containing (piperidine or piperazine) ring (Fig. 1)
demonstrated significantly higher affinities for both
[3H]dopamine uptake and [3H]CFT binding sites than did THP
itself in DAT-expressing cell systems (Fig. 2 and Table 2, con-
dition sets I and II). Ranking the molecules in decreasing order
Fig. 2.
Representative curves for the inhibition of [3H]dopamine uptake, [3H]CFT binding and [3H]NMS binding by THP and several of its related analogs in
CHO cells stably transfected with the human dopamine transporter ([3H]dopamine uptake and [3H]CFT binding) or in rat brain membranes ([3H]NMS binding).
(A) THP, (B) Congener 5e, (C) Congener 5f, (D) Congener 5g (E) Congener 7b, (F) Congener 8a (see Table 1 and the Appendix for descriptions). Competition
studies were conducted with various concentrations of test compound and a fixed concentration of radioligand as indicated in the methods. Data for the graphs
were taken from the experiments presented in Table 2 (condition set I) and Table 3.

D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
1017
of potency yields: 4-methylpiperazine (8c) ≈ 4-hydroxy-4-
phenylpiperadine (5f) > 4-phenylpiperadine (5g) > 4-methyl-
piperadine (5h) ≈ the diethylamino hydroxyphenylcyclohexy-
lpropane derivative (8a). Binding and uptake assays performed
under the most conservative condition set revealed ratios of
Ki ([3H]dopamine uptake) to Ki ([3H]CFT binding) for that
group of compounds ranging between 2.9 and 7.5. Assays in
rat striatum also revealed that substituting the piperidine
ring with a hydroxyphenyl (5f), 4-phenylpiperadine (5g),
4-methylpiperidine (5h), 3′-diethylamine (8a) or 4-methyl
piperazine (8c) increased a compound’s potency in inhibit-
ing [3H]dopamine uptake compared to the parent compounds
(Table 2, condition set III).
and the hydroxyphenylpiperidine (5f; IC50 = 11.2 µM) sub-
stitutions were especially effective. Each of the other tested
analogs displayed a lower affinity than THP for muscarinic
receptors. Potencies varied from 5 nM (Table 3: 5b or 5d) to
250 µM (Table 3: 8c).
3.3. Potencies of a lead THP analog in inhibiting binding
of a number of radioligands
Congener 5f appeared to provide an interesting potential
lead compound for in vivo testing, as it had reduced musca-
rinic activity and it displayed moderately selective potencies
in blocking cocaine analog recognition. To further assess the
possible in vivo actions of this structure, 5f was employed in
single concentration (10 µM) tests for inhibition of radioli-
gand binding to 61 diverse brain sites (NovaScreen, Ocean-
icx Biosystems Inc.). No significant activity was revealed in
53 tests. At that concentration, 5f inhibited the binding of
[3H]LSD, [3H]naloxone, [3H]DTG, [3H]nitrendipine,
[3H]batrachotoxin A and [125I]neurokinin A by more than
50%. Follow up assays revealed that 5f displaced bound
sigma-1 receptor ligand [3H]SKF10047 potently
(Ki = 106 nM, n = 1) but that its potency to displace the opi-
ate receptor ligand [3H]naloxone was much lower
(Ki = 1.4 µM, n = 1).
Congener 5i inhibited [3H]dopamine uptake only mod-
estly in DAT-expressing cell systems, but inhibited [3H]CFT
binding in a complex fashion that made determining mean-
ingful Ki values difficult (Table 2, condition set I). In rat stri-
atal synaptosomes, 5i inhibited [3H]dopamine uptake with
potencies substantially higher than those displayed in the
DAT-expressing cell lines (Table 2, condition set III).
3.2. Potencies of THP analogs in inhibiting binding
of a muscarinic cholinergic ligand
With the exception of 5f, THP and its congeners remained
substantially more potent in inhibiting [3H]NMS binding to
muscarinic cholinergic receptors than they were in inhibiting
[3H]CFT binding or [3H]dopamine uptake through the DAT
(Table 3). The THP analog series of compounds displayed
various potencies at displacing [3H]NMS bound to mem-
branes prepared from rat frontal cortex (Fig. 2 and Table 3).
Addition of a second aromatic ring to THP (Fig. 1) substan-
tially reduced affinities for muscarinic receptors (7a com-
pared to 8a; 7b compared to THP; 7c compared to 8c; 7d
compared to 5h). The phenylpiperidine (5g; IC50 = 4.2 µM)
3.4. Potencies of 5f in behavioral assessments of baseline
and cocaine induced locomotion and reward
C57B1/6J male mice were treated with 1 or 4 mg/kg of 5f,
60 min before they were injected with 30 mg/kg of cocaine.
Their locomotive activity was measured for 60 min after
cocaine injection. Pretreatment with 1 and 4 mg/kg of 5f
reduced cocaine induced locomotive activity by 11% and
14%, respectively (n = 1, data not shown). In condition place
preference (CPP) tests, mice injected with 1 or 2.5 mg/kg of
5f displayed no significant alteration in the time spent in the
chamber paired with this agent (n = 1, data not shown). Mice
injected with cocaine alone (no pretreatment) spent signifi-
cantly more time on the cocaine paired side. Mice pretreated
with 5f before cocaine conditioning spent almost exactly the
same time in each chamber as mice pretreated with saline.
Table 3
Affinities of THP and related analogs for the muscarinic receptor
Test compound
[3H]NMS IC50 (nM)
3.1 0.09
14.1 2.3
5.9 0.56
19.5 4.1
5.5 0.76
10.4 2.3
11,186.5 817
4,160 196
68.3 4.5
312 68
THP
5a
5b
5c
5d
5e
5f
5g
5h
5i
4. Discussion
This work has implications for considerations of the
structure/activity properties of the dopamine transporter and
muscarinic cholinergic receptors. It demonstrates the possi-
bility, and the difficulty, of designing and selecting com-
pounds capable of serving as a new class of DAT blockers.
As mentioned earlier, the vast majority of studies investi-
gating fine interactions of ligands with the DAT have focused
on cocaine analogs [19,33–37]. Structure requirements elu-
cidated for those tropane congeners suggested that 2b-carboxy
esterification, basic nitrogens capable of protonization at
7a
7b
7c
7d
8a
8c
155 17.5
19.7 1.4
768 94
351 36
21.2
2
254 14
Affinities of THP analogs to the muscarinic cholinergic receptor were deter-
mines by inhibition of [3H]NMS binding to rat frontal cortex membranes.
Values for each compound are mean S.E.M. calculated from three to six
independent experiments.

1018
D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
physiological pH, R-configuration of the tropane and a
3b-aromatic ring at C-3 are required to varying extents
[19,29,30,38]. THP’s basic structure is substantially differ-
ent from that of cocaine, though they both possess an aro-
matic ring and a nitrogen.
A recent study has demonstrated that THP attenuated CPP
induced by methamphetamine, but not that induced by cocaine
[43]. In the same study, THP failed to attenuate locomotive
activity induced by any of the stimulants. It is not clear what
mechanism caused the observed attenuation in CPP. It would
be interesting to test the in vivo effect of compounds with
greater selectivity for each of these drugs. We were not able
to test all the described analogs for inhibition of locomotion
or other behavioral tests in the current study. In preliminary
behavioral studies, congener 5f demonstrated only a small
degree of antagonism towards cocaine activity in vivo. This
could be because of this compound’s relatively low selectiv-
ity, its poor ability to reach the brain or because it is rapidly
metabolized. Another possibility is that the dopamine uptake
and cocaine binding properties of the DAT can not be phar-
macologically separated. Extensive DAT mutagenesis stud-
ies, however, indicated that alanine or glycine substitutions
of tryptophane [8], proline [9] and phenylalanine [7] can selec-
tively reduce cocaine analog or dopamine affinities. It is con-
ceivable that a compound capable of acting on a certain DAT
site without affecting other sites will eventually be designed.
An agent capable of selectively blocking a specific site on
the DAT could be beneficial for treating cocaine overdoses,
on the one hand, and for preventing neurotoxicities, on the
other. Neurotoxicity may be caused by toxins such as MPP+,
which induces Parkinsonism [44]. DAT mutagenesis data sug-
gested that the residues important for MPP+ uptake are not
identical to those important for dopamine uptake by the same
transporter [45]. Conceivably, small molecules capable of
interfering with the uptake of an MPP+-like substrate with-
out significantly interfering with dopamine uptake may pro-
vide a novel approach to antagonizing neurotoxicities.
In conclusion, we present here results from a structure–
function study of a series of novel THP analogs. All the ana-
logs tested demonstrated a better affinity for the cocaine bind-
ing site than the dopamine uptake site. We have more
specifically defined the characteristics necessary for a drug
to be potent at the DAT. One of these analogs, the 5f conge-
ner, had few inihibitory effects in behavioral tests. Unfortu-
nately, other analogs were not tested. As more compounds
exhibiting selective DAT inhibition are developed, further
structure/activity studies could help to enrich the generalized
structure–activity requirements tentatively proposed for selec-
tive cocaine antagonism.
The structure–activity relationships of the THP analogs
described in detail in this report demonstrated that modifica-
tions of THP’s phenylpropane group (5a–5e), cyclohexyl
group (7a–7d) and piperidine group (5f–5i, 8a, 8c) each pro-
duced typical pharmacological characteristics. As indicated
in Section 3, methylation/halogenation of the phenylpropane
ring created compounds with better Ki dopamine uptake/Ki
CFT binding selectivity, ranging between 4.33 and 22
(Table 2, condition set I). In agreement with previous studies
using cocaine [38] and aryloxantropane analogs [39], chlo-
ride substituents (5c–5e) on the aromatic ring increased affin-
ity for the DAT and followed the same order of Cl2 > Cl. In
addition, chloride and methyl substitution on the phenyl ring
dramatically enhanced selectivity (Table 2, condition set I).
Congener 5e, which has two chloride substitutions on its phe-
nylpropane ring, displayed the highest affinity for the
cocaine/CFT site and the highest selectivity for that site among
the THP congeners (Table 2, condition set I; Fig. 2). A com-
pound with this kind of selectivity that does not interfere with
other receptors might serve as a cocaine inhibitor.
Replacing the cyclohexyl ring with an aromatic ring (7a–
7d) tended to reduce the compound’s affinity for the dopam-
ine transporter as measured by both [3H]dopamine inhibition
and by [3H]CFT displacement (Table 2, condition set I). This
may be a result of changing the octanol/water partition of the
molecule, thus creating a more lipophilic compound with
reduced ability for docking at its ligand site. Substitutions on
the piperidine ring (5f–5h, 8a, 8c) also influenced both
[3H]CFT binding and [3H]dopamine uptake and resulted in
compounds that demonstrated higher affinities for those sites.
This tendency was robust and appeared in the three sets of
conditions tested (Table 2). It may be related to the availabil-
ity of the nitrogen atom for protonization, which may be
required for binding to the transporter.
The cholinergic properties of THP congeners also dis-
played several structure–function profiles. Compounds that
had 4-phenyl (5g) or 4-hydroxyphenyl (5f) on the piperidine
ring showed a markedly reduced affinity to the muscarinic
receptor (Table 3).
It was anticipated that congener 5f would exhibit low side
effects due to its reduced affinity for the muscarinic cholin-
ergic receptor (Table 3) and it was, therefore, selected for
further measurement. The fact that 5f exhibited binding affin-
ity to the Sigma-1 receptor is not surprising, as THP and other
M1 selective antagonists demonstrated potent inhibitory activ-
ity at that site [40]. Recent studies demonstrated several func-
tional links between cocaine activity and the Sigma-1 recep-
tor [41,42]. If THP and its congeners act as antagonists at
that site, it should be interesting to relate their potency at the
Sigma-1 receptor to their ability to reduce cocaine’s reward-
ing effects.
5. Experimental
5.1. Chemistry: synthesis of trihexyphenidyl analogs
5.1.1. Synthesis of substituted phenylcyclohexyl ketones
(3a–e)
Cyclohexyl n-methyl-n-methoxy carboxamide was pre-
pared from cyclohexyl carboxylic acid chloride and
n-methoxy-n-methyl amine. To a solution of this carboxam-
ide (1.71 g., 10 mmol) in ether (20 ml) at –78 °C was added

D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
1019
the appropriately substituted phenylmagnesium bromide
5.1.5. Preparation of 1-cyclohexyl-1-phenylpropanols (8a
and 8c)
A solution of the diphenyl propanols (7a and 7c) in acetic
(30 ml, 1.0 M in ether solution). The reaction mixture was
stirred overnight while allowing the temperature to reach
23–25 °C.After addition of ammonium chloride solution, the
ether fraction yielded pure samples of ketones 3a (100%) and
3b (100%). The 3c ketone was purified (yield, 71%) on a
silica gel column (2% ethyl acetate in hexane). Ketones 3d
(yield, 31%) and 3e (yield, 32%) were prepared in a similar
manner by the action of cyclohexylmagnesium chloride on
the appropriate benzaldehyde. The resulting benzyl alcohols
were oxidized with pyridinium dichromate solution to give
the corresponding ketones.All ketones gave characteristic IR
and NMR spectra and were used without further purification
in the next step.
acid (10 ml/mmol) was hydrogenated over platinum oxide
(IV) catalyst (20 mg/mmol) at atmospheric pressure over-
night. The catalyst was removed and the filtrate was rendered
basic with dilute ammonium hydroxide solution and was
extracted with ether (20 ml/mmol). The ether solution was
dried using Na2SO4, evaporated and the residue was purified
by silica gel column chromatography, and eluted with a sol-
vent mixture of ethyl acetate (5–10%) in hexane. The target
compounds (8a and 8c; Scheme 2) were characterized as their
hydrogen chlorides. The physical properties are listed in the
Appendix (see Section 5.1.6).
5.1.2. General procedure for preparation
of 3-piperidino-1-substituted propanols (5a–i)
5.1.6. Appendix
5a 3-Piperidino-1-cyclohexyl-1-(p-tolyl)-1-propanol
hydrochloride. C21H33NO·HCl.
5b 3-Piperidino-1-cyclohexyl-1-(m-tolyl)-1-propanol
hydrochloride. C21H33NO·HCl.
5c 3-Piperidino-1-cyclohexyl-1-(4-chlorophenyl)-1-pro-
panol hydrochloride. C20H30NOCl·HCl.
5d 3-Piperidino-1-cyclohexyl-1-(3-chlorophenyl)-1-
propanol hydrochloride. C20H30NOCl·HCl.
5e 3-Piperidino-1-cyclohexyl-1-(3,4-dichlorophenyl)-1-
propanol hydrochloride. C20H30NOCl2·HCl.
5f 3-(4-Hydroxy-4-phenyl)piperidino-1-cyclohexyl-1-
pheny-l-propanol hydrochloride. C26H35NO2·HCl.
5g 3-(4-Phenyl)piperidino-1-cyclohexyl-1-phenyl-1-pro-
panol hydrochloride. C26H35NO·HCl.0.5 H2O.
5h 3-(4-Methyl)piperidino-1-cyclohexyl-1-phenyl-1-pro-
panol hydrochloride. C21H32NO·HCl.
5i 3-(1,2,3,4-Tetrahydroisoquinolino)-1-cyclohexyl-1-
phenyl-propan hydrochloride. C24H30NO·HCl.
7a 3-Diethylamino-1,1-diphenyl-1-propanol hydrochlo-
ride. C19H25NO HCl·0.25H2O.
7b 3-Piperidino-1,1-diphenyl-1-propanol hydrochloride.
C20H25NO HCl.
To a freshly prepared and stirred solution of
LiN[CH(CH3)2]2 (1.1 mmol) in THF (5 ml) at 0 °C was added
a solution of the appropriate n-acetylpiperidine (2) (1 mmol
in 4 ml THF). After stirring for 15 min the reaction mixture
was cooled to –78 °C and the appropriate ketone (3) was
added. The reaction mixture was allowed to warm up to
–15 °C over a period of 1.5 h and was stirred for 1 h at that
temperature. The reaction was quenched with a solution of
acetic acid in ether (15%, 10 ml followed by 10 ml water).
The aqueous fraction was extracted with ethylacetate. The
combined organic fraction was washed with brine and evapo-
rated. The crude amides (4a–i) were purified by column chro-
matography on silica gel (20% ethyl acetate/hexane). The pure
amides obtained (yield, 44–82%) were used in the next step.
BH3/(CH3)2S (2.0 M in THF ml/mmol of amide) was added
to a solution of the appropriate amide 4 (4 ml/mmol in THF).
After stirring at 0 °C for 1 h and at room temperature for 2 h,
methanol was added carefully to remove excess borane. The
residue obtained on evaporation of the solvent was converted
to a hydrogen chloride salt and crystallized from methanol/
ether (Scheme 1). The compound’s physical properties are
listed in the Appendix (see Section 5.1.6).
7c 3-(4-Methyl)piperazeno-1,1-diphenyl-1-propanol hy-
drochloride. C20H26NO·2HCl.
7d 3-(4-Methyl)piperidino-1,1-diphenyl-1-propanol hy-
drochloride. C21H27NO·HCl.
8a 3-Diethylamino-1-cyclohexyl-1-phenyl-1-propanol
hydrochloride. C19H31NO·HCl·0.25H2O.
8c 3-(4-Methyl)piperazino-1-cyclohexyl-1-phenyl-1-
propanol hydrochloride. C20H32N2O·2HCl.
5.1.3. Synthesis of ethyl-3-dialkylamino propionates (6a–d)
An equimolar mixture of ethyl acrylate and the appropri-
ate secondary amine (1) were heated to 100 °C for 4–8 h. The
resulting propionates were purified by distillation and were
used in the next step without further purification (Scheme 2).
5.1.4. Preparation of diphenylpropanols (7a–d)
To a stirred solution of each of 6a–c (in ether, 5 ml/mmol)
at 0 °C was added phenylmagnesium bromide in ether
(3 eq./1 eq. of 6) and the mixture was heated to reflux for 3 h.
After quenching the reaction with water at 0 °C the ether layer
was washed with brine and dried. The residue obtained on
evaporation was purified on a silica gel column. Eluting with
a solvent mixture of ethyl acetate and hexane (1:20) gave
diphenyl analogs (7a–d; Scheme 2) characterized as hydro-
gen chloride salts. Their physical properties are listed in the
Appendix (Section 5.1.6).
5.2. In-vitro assays
5.2.1. [3H]CFT binding and [3H]dopamine uptake
5.2.1.1. Condition set I. CHO cells stably expressing the
human DAT in pcDNAI were selected in G418. DAT-
expressing subclones were expanded and maintained on
F12 medium supplemented with 10% fetal bovine serum, 1%
penicillin–streptomycin and 200 µg/ml of G418. Cells were

1020
D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
distributed to 24 well plates 2 days prior to harvest. DAT-
expressing cells were subjected to radioligand CFT binding
and dopamine uptake assays after washing with Krebs–
Ringer–Hepes (KRH) buffer (in mM: NaCl 125, NaHepes
25, KCl 4.8, MgSO4 1.2, glucose 5.6, CaCl2 1.3, KH2PO4
1.2, pH 7.4). For binding, cells were incubated with 2 nM of
[3H]CFT (84.5 Ci/mmol; NEN) and the unlabeled com-
pounds for 2 h at 4 °C. For uptake assays, cells were incu-
bated with 1 µM [3H]dopamine (NEN stock diluted to ca.
23 mCi/mmol) and unlabeled compounds for 5 min at 37 °C
in KRH supplemented with 10 µM ascorbate. Three washes
with KRH terminated both binding and uptake assays. Cells
were dissolved in 0.25 ml 1% SDS, and their radioactivity
was assessed. 30 µM of (–)cocaine was added to parallel incu-
bations to provide estimates of non-specific binding and
uptake.
radioactivity was estimated. Parallel incubations containing
1 µM atropine provided estimates of non-specific binding.
5.2.3. NovaScreen
NovaScreen (Hanover, Maryland) uses radioligands selec-
tive for 61 neurotransmitters, ion channels, transporters, sec-
ond messengers, steroids, prostaglandins, growth factors, hor-
mones and enzyme receptor sites. The efficacy of compound
5f (10 µM) in competing against these radoligands for bind-
ing sites on whole rat brain membrane preparations was
assessed in duplicate.
5.3. In-vivo assays
5.3.1. Locomotive activity
C57B1/6J male mice obtained from Jackson Laboratories
(Bar Harbor, ME, USA), group housed with food and water
available ad lib and maintained on a 12 hourly day–night cycle
were housed in our facilities for a minimum of 2 weeks prior
to testing. Mice were placed in 46 × 25 × 19 (l × w × h) clear
plastic cages and locomotive activity was measured using
Columbus Instruments (Columbus, OH, USA) Opto-Varimex
activity monitors in sound-attenuating chambers under dim
ambient light conditions. Total distance traveled was mea-
sured for 60 min for mice injected with 1 or 4 mg/kg of 5f.
Mice were then injected with 30 mg/kg of cocaine HCl and
their activity was assessed for an additional 60 min.
5.2.1.2. Condition set II. Conditions were identical to those
above except that cells were distributed to 96-well plates,
3–4 days prior to the experiment, pre-incubated with unla-
beled test compounds for 2 h in cell culture media at 37 °C,
washed three times and then incubated with 5 nM [3H]CFT
or 100 nM [3H]dopamine and the tested compounds for
10 min at 25 °C. Binding and uptake were terminated by five
washes with KRH and radioactivity was assessed. 200 µM
(–)cocaine provided estimates of non-specific binding and
uptake.
5.3.2. Conditioned place preference
5.2.1.3. Condition set III. Striata were dissected from the
brains of 200–250 g male Sprague–Dawley rats. Binding stud-
ies were conducted as previously described [29]. Striata were
homogenized using a polytron, and membranes washed twice
resuspended in 0.32 M sucrose phosphate buffer (pH 7.4) and
recentrifuged. Membranes prepared from 1 mg of striatum
(wet weight) were incubated with 0.5 nM [3H]CFT in 0.5 ml
sucrose phosphate buffer and unlabeled test compound at 4 °C
for 2 h. Uptake studies were conducted as previously de-
scribed [30]. Striatal synaptosomes were prepared, pre-
incubated at 30 °C for 10 min with 1 µM pargyline and unla-
beled test compounds and incubated for 3 min with 0.5 nM
[3H]dopamine in 1 ml of buffer (in mM: NaCl 126, KCl 4.8,
CaCl2 1.3, NaH2PO4 16, MgSO4 1.4, dextrose 11 and ascor-
bic acid 1.1). Binding and uptake reactions were quenched
by rapid filtration onto GF/B filters presoaked in 0.05% poly-
ethyleneimine. Non-specific binding was determined using
30 or 100 µM of (–)cocaine for [3H]CFT binding and
[3H]dopamine uptake experiments, respectively.
CPP was assessed using two compartment chamber with
one 18x18x18 compartment with a wire mesh floor and the
other with concorb bedding. During two days of twice daily
conditioning sessions, C57B1/6J male mice were restricted
for 20 min after injection with cocaine or saline, with or with-
out pretreatment with congener 5f to one side of the appara-
tus. A single CPP assessment session followed the last con-
ditioning session by 24 h. In this session, mice had access to
both compartments and the proportion of the 20 min session
spent on each side was recorded using Optivarimax animal
activity monitor (Columbus Instruments; Columbus, OH,
USA).
Acknowledgments
The authors acknowledge financial support from NIDA-
IRP.
5.2.2. [3H]n-methyl scopolamine binding
Displacement of [3H]NMS from membranes prepared from
rat frontal cortex was determined as described previously [31].
Tissue was homogenized, washed twice in 50 mM Tris–HCl
and then membranes were prepared from 0.2 mg of tissue
incubated with 0.06 nM of [3H]NMS (NEN, 84 Ci/mmol)
and various unlabeled test compounds in 2 ml buffer at 22 °C
for 90 min. Reactions were terminated by rapid filtration, and
References
[1] M.C. Ritz, R.J. Lamb, S.R. Goldberg, M.J. Kuhar, Science 237 (1987)
1219–1223.
[2] G.M. Sizemore, C. Co, T.R. Koves, T.J. Martin, J.E. Smith, Psychop-
harmacol. 171 (2004) 413–420.
[3] M.D. Schechter, J.W. Boja, Pharmacol. Biochem. Behav. 30 (1988)
1089–1092.

D.E. Dar et al. / European Journal of Medicinal Chemistry 40 (2005) 1013–1021
1021
[4] D. French, J.M. Witkin, Pharmacol. Biochem. Behav. 46 (1993) 989–
993.
[5] E.J. Bilsky, M.J. Montegut, M.L. Nichols, L.D. Reid, Pharmacol.
Biochem. Behav. 59 (1998) 215–220.
[26] K.L. Preston, R.R. Griffiths, E.J. Cone, W.D. Darwin, C.W. Gor-
odesky, Drug Alcohol Depend. 18 (1986) 195–202.
[27] F.I. Carroll, A.H. Lewin, M.J. Kuhar, J.W. Boja, NIDA Res. Monogr.
134 (1993) 229–237.
[28] S. Kitayama, K. Morita, T. Dohi, J.B. Wang, S.C. Davis, G.R. Uhl,
Ann. N. Y. Acad. Sci. 801 (1996) 388–393.
[29] F.I. Carroll, Y. Gao, M.A. Rahman, P. Abraham, K. Parham,
A.H. Lewin, J.W. Boja, M.J. Kuhar, J. Med. Chem. 34 (1991) 2719–
2725.
[30] J.W. Boja, R.M. McNeill, A.H. Lewin, P. Abraham, F.I. Carroll,
M.J. Kuhar, Neuroreport 3 (1992) 984–986.
[31] S.B. Freedman, E.A. Harley, L.L. Iversen, Br. J. Pharmacol. 93 (1988)
437–445.
[32] B.E. Blough, C.R. Holmquist, P. Abraham, M.J. Kuhar, F.I. Carroll,
Bioorg. Med. Chem. Lett. 10 (2000) 2445–2447.
[33] F.I. Carroll, A.H. Lewin, M.J. Kuhar, in: M.E.A. Reith (Ed.), Neu-
rotransmitter Transporters: Structure, Function, and Regulation,
Humana, Totowa, NJ, 1997, pp. 263–295.
[34] B.A. Bennet, C.H. Wichems, C.K. Hollingworth, H.M.L. Davies,
C. Thornley, T. Sexton, S.R. Childers, J. Pharmacol. Exp. Ther. 272
(1995) 1176–1186.
[35] P.C. Meltzer, A.Y. Liang, B.K. Madras, J. Med. Chem. 37 (1994)
2001–2010.
[36] H.M.L. Davies, E. Saikali, T. Sexton, S.R. Childers, Eur. J.
Pharmacol.-Mol. Pharmacol. 244 (1993) 93–97.
[37] M.M. Goodman, M.-P. Kung, G.W. Kabalka, H.F. Kung, R. Switzer, J.
Med. Chem. 37 (1994) 1535–1542.
[38] X.-S. He, L.P. Raymon, M.V. Mattson,A.H. Eldefrawi, B.R. de Costa,
J. Med. Chem. 36 (1993) 4075–4081.
[39] F.I. Carroll, P. Abraham, A.H. Lewin, K.A. Parham, J.W. Boja,
M.J. Kuhar, J. Med. Chem. 35 (1992) 2497–2500.
[40] R.L. Hudkins, D.L. DeHaven-Hodkins, Life Sci. 49 (1991) 1229–
1235.
[41] D. Zhang, L. Zhang,Y. Tang, Q. Zhang, D. Lou, F.R. Sharp, J. Zhang,
M. Xu, Neuropsychopharmacol (May 2005) [Epub ahead of print].
[42] T. Maurice, P. Romieu, Pharmacopsychiatry 37 (Suppl. 3) (2004)
S198–S207.
[6] S. Kitayama, S. Shimada, H. Xu, L. Markham, D.M. Donovan,
G.R. Uhl, Proc. Natl. Acad. Sci. USA 89 (1992) 7782–7785.
[7] Z. Lin, W. Wang, T. Kopajtic, R.S. Revay, G.R. Uhl, Mol. Pharmacol.
56 (1999) 434–447.
[8] Z. Lin, W. Wang, G.R. Uhl, Mol. Pharmacol. 58 (2000) 1581–1592.
[9] Z. Lin, M. Itokawa, G.R. Uhl, FASEB J. 14 (2000) 715–728.
[10] O. Edvarson, S.G. Dahl, Brain Res. Mol. Brain Res. 27 (1994) 265–
274.
[11] D.O. Calligaro, M.E. Eldefrawi, Membr. Biochem. 7 (1988) 87–106.
[12] T. Maurice, G. Barbanel, J.M. Kamenka, J. Vignon, Neuropharmacol.
30 (1991) 591–598.
[13] J.J. Bonnet, S. Benmansour, J. Costentin, E.M. Parker, L.X. Cubeddu,
J. Pharmacol. Exp. Ther. 253 (1990) 1206–1214.
[14] F.I. Carroll, L.L. Howell, M.J. Kuhar, J. Med. Chem. 42 (1999)
2721–2736.
[15] R. Spanagel, F. Weiss, Trends Neurosci. 22 (1999) 521–527.
[16] A.H. Newman, Expert Opin. Ther. Patents 10 (2000) 1095–1122.
[17] F.I. Carroll, A.H. Lewin, S.W. Mascarella, in: M.E.A. Reith (Ed.),
Neurotransmitter Transporters: Structure, Function and Regulation,
second ed, Humana Press, Totowas, NJ, 2001, pp. 381–432.
[18] F.I. Carroll, N. Pawlush, M.J. Kuhar, G.T. Pollard, J.L. Howard, J.
Med. Chem. 47 (2004) 296–302.
[19] F.I. Carroll, A.H. Lewin, J.W. Boja, M.J. Kuhar, J. Med. Chem. 35
(1992) 969–981.
[20] F.I. Carroll, P. Kotian, A. Dehghani, J.L. Gray, M.A. Kuzemko,
K.A. Parham, P.Abraham,A.H. Lewin, J.W. Boja, M.J. Kuhar, J. Med.
Chem. 38 (1995) 379–388.
[21] R.B. Rothman, B. Lewis, C. Dersch, H. Xu, L. Radesca, B.R. de
Costa, K.C. Rice, R.B. Kilburn, H. Akunne, A. Pert, Synapse 14
(1993) 34–39.
[22] C.D. Cook, F.I. Carroll, P.M. Beardsley, Psychopharmacol. (Berlin)
159 (2001) 58–63.
[23] S.I. Dworkin, P. Lambert, G.M. Sizemore, F.I. Carroll, M.J. Kuhar,
Synapse 30 (1998) 49–55.
[43] K. Shimosato, S. Watanabe, S. Kitayama, Naunyn-Schmiedeberg’s
Arch. Pharmacol. 364 (364) (2001) 74–80.
[24] V.P. Dole, M.E.A. Nyswander, J. Am. Med. Assoc. 193 (1965) 646–
650.
[44] M. Gerlach, P. Riedere, H. Przuntek, M.B.Youdim, Eur. J. Pharmacol.
208 (1991) 273–286.
[25] D.R. Jasinski, K.L. Preston, Drug Alcohol Depend. 17 (1986) 301–
310.
[45] S. Kitayama, J.B. Wang, G.R. Uhl GR, Synapse 15, Dopamine trans-
porter mutants selectively enhance MPP+ transport 15 (1993) 58–62.