Synthesis and biological evaluation of 2-(4-fluorophenoxy)-2-phenyl-ethyl piperazines as serotonin-selective reuptake inhibitors with a potentially improved adverse reaction profile

Article abstract of DOI:10.1016/j.bmc.2003.12.021

Three new 2-(4-fluorophenoxy)-2-phenyl-ethyl piperazines, 1-(3-chlorophenyl)-4-[2-(4-fluorophenoxy)-2-phenylethyl]-piperazine 7, 1-[2-(4-fluorophenoxy)-2-phenylethyl]-4-(2-methoxyphenyl)-piperazine 8, and 1-[2-(4-fluorophenoxy)-2-phenylethyl]-4-(3-trifluoromethylphenyl)-piperazine 9, modeled after the potent antidepressant fluoxetine and coupled with several functionalized piperazines, have been prepared by chemical synthesis as selective serotonin reuptake inhibitors (SSRIs) with a potentially improved adverse reaction profile. Typical SSRIs, although very effective in the treatment of depression, still face the troublesome side effect of sexual dysfunction. A number of pharmacological agents-notably, drugs in the piperazine class-have been used to reverse SSRI-induced sexual dysfunction, and evidence for developing an improved SSRI by coupling a fluoxetine congener with the pharmacophore of a reversal agent holds promise. Preliminary data indicates that the hydrochloride (HCl) salts 10, 11, and 12 each exhibit single-site binding at the site of the serotonin reuptake transporter (SERT). However, each of the three compounds are much less potent than typical SSRIs, showing micromolar (μM) affinity for the SERT with IC₅₀ values of 1.45 μM, 3.27 μM, and 9.56 μM, respectively. Further biological evaluation of compounds 10, 11, and 12 is needed before definitive conclusions can be made with regard to each compound's potential for use as an SSRI-type candidate which is devoid of sexual side effects. Nevertheless, the initial findings are quite encouraging, thus lending credence to the idea of hybridizing an SSRI congener with that of the pharmacophore of an agent known to reverse or treat SSRI-induced sexual dysfunction.

Full text of DOI:10.1016/j.bmc.2003.12.021

Bioorganic & Medicinal Chemistry 12 (2004) 1483–1491
Synthesis and biological evaluation of 2-(4-fluorophenoxy)-2-
phenyl-ethyl piperazines as serotonin-selective reuptake inhibitors
with a potentially improved adverse reaction profile
James M. Dorsey,a Maria G. Miranda,a Nicholas V. Cozzib and Kevin G. Pinneya,*
aDepartment of Chemistry and Biochemistry and The Center for Drug Discovery, Baylor University, Waco, TX 76798-7348, USA
bDepartment of Pharmacology and Toxicology, Brody School of Medicine, East Carolina University, Greenville, NC 27834, USA
Received 3 October 2003; accepted 18 December 2003
Abstract—Three new 2-(4-fluorophenoxy)-2-phenyl-ethyl piperazines, 1-(3-chlorophenyl)-4-[2-(4-fluorophenoxy)-2-phenylethyl]-
piperazine 7, 1-[2-(4-fluorophenoxy)-2-phenylethyl]-4-(2-methoxyphenyl)-piperazine 8, and 1-[2-(4-fluorophenoxy)-2-phenylethyl]-
4-(3-trifluoromethylphenyl)-piperazine 9, modeled after the potent antidepressant fluoxetine and coupled with several functiona-
lized piperazines, have been prepared by chemical synthesis as selective serotonin reuptake inhibitors (SSRIs) with a potentially
improved adverse reaction profile. Typical SSRIs, although very effective in the treatment of depression, still face the troublesome
side effect of sexual dysfunction. A number of pharmacological agents-notably, drugs in the piperazine class-have been used to
reverse SSRI-induced sexual dysfunction, and evidence for developing an improved SSRI by coupling a fluoxetine congener with
the pharmacophore of a reversal agent holds promise. Preliminary data indicates that the hydrochloride (HCl) salts 10, 11, and 12
each exhibit single-site binding at the site of the serotonin reuptake transporter (SERT). However, each of the three compounds are
much less potent than typical SSRIs, showing micromolar (mM) affinity for the SERT with IC50 values of 1.45 mM, 3.27 mM, and
9.56 mM, respectively. Further biological evaluation of compounds 10, 11, and 12 is needed before definitive conclusions can be
made with regard to each compound’s potential for use as an SSRI-type candidate which is devoid of sexual side effects. Never-
theless, the initial findings are quite encouraging, thus lending credence to the idea of hybridizing an SSRI congener with that of the
pharmacophore of an agent known to reverse or treat SSRI-induced sexual dysfunction.
# 2004 Elsevier Ltd. All rights reserved.
1. Introduction
the piperazine class-have been used to reverse SSRI-
induced sexual dysfunction, and evidence for developing
an improved SSRI by coupling a fluoxetine congener
with the pharmacophore of a reversal agent holds
promise.
In the late-1980s, the serotonin-selective reuptake inhi-
bitor (SSRI) fluoxetine became the mainstay of treat-
ment for clinical depression-replacing the more toxic
tricyclic antidepressants (TCAs).1 SSRIs have a more
favorable adverse reaction profile in comparison to the
TCAs and are much easier to tolerate. The structures of
some of the typically prescribed SSRIs are depicted in
Figure 1. However, SSRIs are notinnocuous drugs. In
spite of the progress in the treatment of depression with
the advent of the SSRIs, some troublesome side effects
still remain-most notably, sexual dysfunction. As pre-
scribing of SSRIs increased, the side effect of sexual
dysfunction emerged, and clinicians searched for meth-
ods to treat or reverse SSRI-induced sexual dysfunction.
A number of pharmacological agents-notably, drugs in
Itis widely believed ht ata number of monoamine neu-
rotransmitters (Fig. 2) have a role in the etiology of
depression and, consequently, its treatment. The most
notable of these neurotransmitters include nor-
epinephrine (NE) and serotonin (5-HT).2 Itis well
known that inhibiting the reuptake of NE or 5-HT (or
both) has elicited a favorable response in patients being
treated for depression. Without question, however,
inhibition of 5-HT reuptake has been the most widely
studied, and treatment of depression has focused on
inhibiting the reuptake mechanism of the neuro-
transmitter.3 Indeed, it was this premise which led to the
hypothesis that altering 5-HT function could lead to an
effective mechanism in the treatment of depression.4
* Corresponding author. Tel.: +1-254-710-4117; fax: +1-254-710-
4272; e-mail: kevin_pinney@baylor.edu
0968-0896/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2003.12.021

1484
J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
The basis for the role of serotonin in the etiology of
depression was hypothesized by Lapin and Oxenkrug in
1969.5 Before that time, most hypotheses relied on the
premise that biogenic amines including NE, 5-HT and
dopamine (DA) were being produced in insufficient
quantities in the limbic system of afflicted patients.
Adding to the confusion is the fact that medications
which alter the action or release of each of these neuro-
transmitters have elicited favorable responses in clini-
cally depressed patients. Examples like bupropion and
venlafaxine (Fig. 3), both administered as racemates,
have mechanisms of action which are not confined to
one neurotransmitter system.6
have a 50- to 100-fold or greater selectivity for the inhi-
bition of serotonin uptake in vitro when compared to
their ability to inhibit NE or DA uptake.8
Sexual dysfunction is a common problem in affective
disorders such as major depression. It is important to
note, however, that the drugs used in the treatment of
depression can cause sexual dysfunction as well. Anti-
depressants such as the TCAs, monoamine oxidase
inhibitors, and the SSRIs have all been reported to
cause sexual side effects.9 The inhibitory effects of
the SSRIs on sexual function are well documented, and the
frequency of occurrence varies widely. Some sources list
ranges varying from 1% up to 75%10 while the majority
of sources estimate the occurrences in the 20–40%
range.11 By switching to a less serotonergic drug, it is
surmised that the non-selective effects of 5-HT can be
reduced within the synapse of the 5-HT neuron.
Although the connection between serotonergic activity
and sexual function is not straightforward, it is believed
that the increased central serotonergic activity of SSRIs
(at the receptor subtypes 5-HT1A, 5-HT1B, 5-HT2C, 5-
HT2A and 5-HT3) plays a role in the emergence of sex-
ual dysfunction.10 It follows, therefore, that reducing
the activity of SSRIs at these receptors could allow the
suppression or inhibition of the sexual response to be
corrected. Strategies which have been tried in treating or
reversing sexual dysfunction include: titration of an
antidepressant dose, scheduling of a drug holiday dur-
ing treatment, and administering a second drug (along
with an SSRI, for example) to counteract the SSRI’s
sexual side effects.11
Research in the treatment of depression remains focused
on 5-HT because, to date, a good number of selective 5-
HT reuptake blockers tested in clinical trials have been
effective in treating major depression. The general
mechanism of action is hypothesized as enhanced 5-HT
neurotransmission due to the increased availability of 5-
HT in the synapse of these neurons.7 SSRIs generally
The drugs which have been administered along with a
SSRI include a broad range of medicinal agents (struc-
tures depicted in Fig. 4). Drugs such as: amantadine
(normally used in the treatment of Parkinsonism),
cyproheptadine, buspirone, trazodone, nefazodone,
methylphenidate, mianserin, yohimbine, and even silde-
nafil citrate. Bupropion has also been of benefit and is
depicted in Figure 3.
Figure 1. Representative serotonin-selective reuptake inhibitors.
The success of each of these agents varies widely, but
the best studied agents include bupropion and buspir-
one.13 Buspirone, which has a pyrimidinyl-piperazine
group, is generally used in the treatment of generalized
anxiety disorder and has been used with some success in
treating patients with SSRI-induced sexual dysfunc-
tion.12 It is interesting to note that other piperazine-
containing groups (specifically, functionalized phenylpi-
perazines) have likewise been used with success in
treating sexual dysfunction. These agents include: silde-
nafil citrate, trazodone, and nefazodone.
Figure 2. Structures of endogenous monoamine neurotransmitters.
Based on the information that buspirone, trazodone,
and nefazodone have all been used successfully in treat-
ing SSRI-induced sexual dysfunction, it is proposed that
coupling the piperazinyl moiety to a structural homo-
logue of an SSRI may have utility in synthesizing an
SSRI candidate with an adverse reaction profile devoid
of sexual side effects. The strategy, therefore, is to
covalently couple a structural homologue of the SSRI
antidepressant, fluoxetine, with that of the piperazinyl-
Figure 3. Antidepressant compounds not confined to one neuro-
transmitter system.

J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
1485
containing portion of each of the following compounds:
buspirone, trazodone, and nefazodone (Fig. 4).
Itwas decided htata racemic mixutre of 2-amino-1-
phenylethanol 1 would serve as a good source of start-
ing material. The primary amine in compound 1 was
protected with di-tert-butyl-dicarbonate (t-boc anhy-
dride) 2, to afford protected alcohol 3. This was subse-
quently converted to 4-fluorophenyl ether 4 using
Mitsunobu conditions, a combination of triphenylphos-
phine (PPh3) and diethylazodicarboxylate (DEAD). The
protected ether was deprotected with 4N HCl in 1,4-
dioxane, giving free amine 5. Bromide 6 was afforded by
treating free amine 5 with a combination of TiBr4 and
tert-butyl nitrite (Scheme 1).
2. Results and discussion
In the attempt to create a novel serotonin reuptake
inhibitor with an improved side effect profile, three new
2-(4-fluorophenoxy)-2-phenyl-ethyl piperazines, fluox-
etine structural homologues with three different piper-
azine side chains, have been synthesized and biologically
evaluated for their potential as antidepressants.
Figure 4. Drugs administered with an SSRI in a effort to reverse/treat sexual dysfunction.
Scheme 1. Synthesis of 2-(4-fluorophenoxy)-2-phenyl-ethyl bromide. Reagents and conditions: (i) DMF, 15 min at 50 ꢀC, 8 h atrt78%; (ii) DMF,
4-fluorophenol, PPh3, DEAD dropwise at0 ꢀC, 8 h atrt, 43%; (iii) 4 N HCl in Dioxane, 1.5 h atrt, 64%; (iv) DMF, TiBr 4, t-butyl nitrite, 1 h at rt,
33%.

1486
J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
Scheme 2. Synthesis of 2-(4-fluorophenoxy)-2-phenyl-ethyl piperazines and their corresponding hydrochloride salts. Reagents and conditions:
(i) K2CO3, DMF, reflux 16–18 h, 18–39%; (ii) 2 M HCl in anh. ether, 6–68%.
Bromide intermediate 6, synthesized as described in
Scheme 1, was subsequently used as starting material in
the next set of reactions. 1-(3-chlorophenyl)-piperazine
and m-trifluoromethylphenyl-piperazine were commer-
cially available in their free base form, while 1-(2-meth-
oxyphenyl)-piperazine was available as the hydrochloride
salt. In this latter case, the free base form of the com-
pound was generated using a strong base (e.g.; NaOH)
in aqueous solution followed by extraction with ethyl
ether. Compounds 7, 8, and 9 were obtained by a
nucleophilic substitution reaction between the bromide
in compound 6 and the secondary amino group of each
piperazine-containing compound. The corresponding
hydrochloride salts (HCl), 10, 11, and 12, were obtained
by treatment of the free base form of each with 2M HCl
in diethyl ether. These compounds were used as such in
IC50 testing of 5-HT reuptake activity (Scheme 2).
from each other at the p<0.05 level of significance
(one-way ANOVA followed by a Student–Newman-
Keuls t-test). Compound 12 was the least potent com-
pound in the series with an IC50 value of 9.56Æ2.4 mM;
this was only one-sixth the potency of compound 10 and
three-fold less potent than compound 11 (p<0.05). Itis
apparent that the new compounds are significantly less
potent than fluoxetine, which exhibits nanomolar (nM)
affinity for the SERT,17 butitis notpossible athtis
stage to decide whether the decreased potency is due to
electrostatic, stereochemical, or steric considerations or
some combination of these. For example, the phenylpi-
perazine moiety of the new compounds adds consider-
able bulk to the side-chain amine of fluoxetine and also
changes its electronic character with the addition of new
lone pair electrons and an electron-rich aromatic ring. It
is possible that the electron-withdrawing, hydrophobic
trifluoromethyl group present on the phenyl ring of 12 is
responsible for the attenuated serotonin uptake inhibi-
tion activity of this compound when compared to com-
pounds 10 and 11, but the methoxy substitution on
compound 11 also represents a positional isomer and
this may influence binding through a steric mechanism.
Interestingly, tests conducted by Hyttel in 1982 indicated
that trazodone, mianserin, and bupropion had IC50
values for serotonin uptake of 0.58 mM, 1.2 mM, and
19.0 mM, respectively.13 This finding is then encouraging
given that trazodone, mianserin, and bupropion have all
The specific uptake of [3H]5-HT by human platelets was
typically >90% specific as defined by 10 mM fluoxetine.
Dose-response curves for inhibition of [3H]5-HT uptake
by the test compounds are presented in Figures 5, 6 and
7. The inhibition curves all had slope coefficients that
did not differ from unity, indicating that the compounds
all interacted with a single site on the SERT. Compounds
10 and 11 were the most potent inhibitors of [3H]5-HT
accumulation, with IC50 values of 1.45Æ0.08 mM and
3.27Æ0.23 mM, respectively. These values did not differ
Figure 5. Inhibition of [3H]5-HT uptake into human platelets by
compound 10.
Figure 6. Inhibition of [3H]5-HT uptake into human platelets by
compound 11.

J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
1487
1
(300–400 mesh) obtained from Merck EM Science. H
and 13C NMR were recorded in deuterated chloroform
(CDCl3) using an AMX 360 MHz (90 MHz for 13C)
and/or DPX Avance 300 MHz (75 MHz for 13C) Bruker
NMR spectrometers. Peaks are listed as broad (b),
singlet (s), doublet (d), triplet (t), quartet (q), or multi-
plet (m), with the coupling constant (J) expressed in
hertz (Hz). High resolution mass spectra were recorded
using a VG/Fisons High Resolution GC/Mass Spectro-
meter using perfluorokerosene (PFK) as the calibration
standard. All elemental analyses were conducted by
Atlantic Microlab, Inc. in Norcross, Georgia. Melting
points were determined in open capillary tubes using a
Thomas-Hoover electronically heated melting point
apparatus and are uncorrected.
4.1.1. (2-Hydroxy-2-phenylethyl)-carbamic acid tert-butyl
ester (3). To a well-stirred solution of 2-amino-1-phe-
nylethanol 1 (5.05 g,ꢀ36.8 mmol) in anhydrous DMF (50
mL) under N2 at25 C was added 1.5 molar equivalents
of di-tert-butyl dicarbonate (t-boc anhydride) (13.24 g,
60.7 mmol). After dissolution of the t-boc anhydride,
the mixture was brought to 50 ꢀC for approximately 15
min and then allowed to cool to 25 ꢀC. After stirring
overnight (8–12 h), ethyl acetate (EtOAc) was added to
the reaction mixture. Water was then added, and the
organic layer was washed three times with water and
subsequently dried over anhydrous magnesium sulfate
(MgSO4). The mixture was then filtered and the solvent
removed under reduced pressure. The resulting residue
was suspended in a minimal amountof EOt Ac and
heated to 55–60 ꢀC until completely dissolved. The mix-
ture was allowed to cool overnight, remaining undis-
turbed, and afforded the desired product 3 (6.73 g, 28.4
Figure 7. Inhibition of [3H]5-HT uptake into human platelets by
compound 12.
been used with success in reversing or treating SSRI-
induced sexual dysfunction.
3. Conclusions
We have prepared three new fluoxetine analogues, 2-(4-
piperazines,
fluorophenoxy)-2-phenyl-ethyl
which
demonstrate single-site binding at the site of the ser-
otonin reuptake transporter (SERT). However, each of
the three compounds are much less potent than typical
SSRIs, such as parentcompound fluoxeint e, which
exhibits nanomolar (nM) affinity for the SERT. Com-
pounds 10, 11, and 12, on the other hand, have micro-
molar (mM) affinity for the SERT with IC50 values of
1.45 mM, 3.27 mM, and 9.56 mM, respectively. It is sur-
mised that the electron-withdrawing, hydrophobic tri-
fluoromethyl group present on the phenyl ring of 12 is
responsible for the reduced serotonin reuptake activity
of this compound when compared with that of com-
pounds 10 and 11. Further biological evaluation of
compounds 10, 11, and 12 is needed before definitive
conclusions can be made with regard to each com-
pound’s potential for use as an SSRI-type candidate
which is devoid of sexual side effects.
1
mmol, 78%) as translucent, slightly yellow crystals. H
NMR (CDCl3): d 7.31 (5H, m), 4.97 (1H, bs), 4.82 (1H,
dd, J=3.9, 3.7 Hz), 3.47 (1H, ddd, J=14.0, 6.5, 3.2 Hz),
3.27 (1H, dd, J=7.8, 5.4 Hz), 3.22 (1H, dd, J=8.6, 6.1
Hz), 1.71 (1H, bs), 1.44 (9H, s). 13C- NMR (CDCl3): d
157.0 (faint), 142.2, 128.9, 128.9, 128.2, 126.3, 126.3,
80.3, 74.4, 48.8, 28.8, 28.8, 28.8. Melting point: 123–
124 ꢀC.
4.1.2. [2-(4-Fluorophenoxy)-2-phenylethyl]-carbamic acid
tert-butyl ester (4). To a well-stirred solution of 3 (6.73
g, 28.4mmol) in anhydrous THF (40 mL) under N2 at
25 ꢀC was added 2 molar equivalents (equiv) of 4-fluoro-
phenol (6.55 g, 58.4 mmol). After complete dissolution
of the 4-fluorophenol, 1.5 equiv of triphenylphosphine
(PPh3) (11.68 g, 44.5 mmol) was added to the mixture
and allowed to completely dissolve. The reaction vessel
was then cooled in an ice bath to 0 ꢀC. After 20 min, 1.5
equiv of diethylazodicarboxylate (DEAD) (7.96 g, 45.7
mmol) was slowly added dropwise to the stirring mix-
ture. The reaction vessel was subsequently allowed to
return to room temperature (25 ꢀC). After stirring over-
night (8–12 h), the triphenylphosphine salts formed
during the reaction were removed by filtration and the
solventevaporated under reduced pressure. The residue
was adsorbed onto silica gel and purified by flash col-
umn chromatography (silica gel, 5:95 EtOAc/hexanes)
which afforded the desired intermediate 4 (4.04 g, 12.2
4. Experimental
4.1. General methods
Chemicals used in synthesis were obtained from the
Aldrich Chemical Company, Fisher Scientific and
ACROS Chemicals. Tetrahydrofuran (THF), dried with
potassium and benzophenone, and methylene chloride,
dried with calcium hydride, were freshly distilled imme-
diately prior to use. Dimethylformamide (DMF) was
purchased as an anhydrous commercial productin a
SureSeal1 container from the Aldrich Chemical Com-
pany. Silica gel plates for analytical thin layer chroma-
tography (TLC) were obtained from EM Science.
Column chromatography was performed using silica gel
1
mmol, 43%) as a white solid. H NMR (CDCl3): d 7.34

1488
J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
(1H, d, J=5.4 Hz), 7.33 (1H, d, J=8.0 Hz), 7.31 (1H, d,
J=4.8 Hz), 7.30 (1H, d, J=7.1 Hz), 7.27 (1H, d, J=4.2
Hz), 6.89 (1H, d, J=10.6 Hz), 6.85 (1H, d, J=9.0 Hz),
6.78 (1H, d, J=9.2 Hz), 6.76 (1H, d, J=9.2 Hz), 5.16
(1H, dd, J=8.5, 3.4 Hz), 5.02 (1H, bs), 3.63 (1H, ddd,
J=14.4, 8.0, 3.7 Hz), 3.39 (1H, ddd, J=13.3, 8.3, 4.6
Hz), 1.43 (9H, s). 13C NMR (CDCl3): d 158.7, 156.0,
155.8, 153.8, 138.5, 128.7, 128.1, 128.1, 126.1, 126.1,
117.0, 116.9, 115.9, 115.6, 80.0, 68.1, 47.2, 28.4, 28.4,
28.4. 19F-NMR (CDCl3): d À123.8. Melting point: 98–
102 ꢀC.
four 30 mL portions of ethyl ether. The combined
organic layers were subsequently dried over anhydrous
MgSO4. The solventwas hten removed under reduced
pressure. Analytical thin-layer chromatography (TLC)
indicated that the desired product had an Rf value of
approx. 0.85 when developed in 30:70 EtOAc/hexanes.
Purification was performed using flash column chroma-
tography (silica gel, 100% hexanes) to isolate the
desired bromide product 6 (0.74 g, 2.6 mmol, 33%) as a
yellow oil. 1H NMR (CDCl3): d 7.37 (5H, m), 6.89 (1H,
d, J=8.0 Hz), 6.87 (1H, d, J=8.0 Hz), 6.83 (1H, d,
J=4.5 Hz), 6.81 (1H, d, J=4.6 Hz), 5.22 (1H, dd,
J=8.4, 4.0 Hz), 3.72 (1H, dd, J=10.9, 8.4 Hz), 3.61
(1H, dd, J=10.9, 4.0 Hz). 13C NMR (CDCl3): d 159.6,
156.5, 154.3, 154.3, 138.9, 129.3, 129.2, 126.8, 118.0, 117.9,
116.4, 116.1, 81.7, 36.4. 19F NMR (CDCl3): d À123.2.
4.1.3. 2-(4-Fluorophenoxy)-2-phenyl-ethylamine (5). A 30
mL volume of 4 N HCl in dioxane was prepared by
adding dropwise 10 mL of 12 N (i.e.; concentrated)
hydrochloric acid to 20 mL of well-stirred amount of
1,4-dioxane. The t-boc protected 4-fluorophenyl ether
intermediate 4 (4.04 g, 12.2 mmol) was dissolved in a
minimal amountof 1,4-dioxane. The 4 N HCl was then
added dropwise to the reaction vessel at 25 ꢀC over
approximately a 10 min period. After 90 min, a satu-
rated solution of sodium bicarbonate (NaHCO3) was
added to the reaction mixture. The aqueous layer was
washed with three separate 30 mL portions of CH2Cl2.
The combined organic layers were then dried over
anhydrous MgSO4 and filtered. Purification was per-
formed by flash column chromatography (on silica gel)
starting with 100% EtOAc until any remaining starting
material had eluted; then the polarity of the eluent was
increased using 30:70 methanol/EtOAc. Analytical thin-
layer chromatography showed that the desired product
had an Rf value of approximately 0.4 when developed in
30:70 methanol/EtOAc. Fractions containing the
desired productwere combined and hte solventwas
removed under reduced pressure. The resultant residue
after solvent evaporation was re-suspended in CH2Cl2
and filtered. The solvent from the filtrate was evapo-
rated under reduced pressure and afforded the desired
free amine 5 (1.80 g, 7.8 mmol, 64%) as a slightly yellow
4.1.5. 1-(3-Chlorophenyl)-4-[2-(4-fluorophenoxy)-2-phenyl-
ethyl]-piperazine (7). To a well-stirred solution of
approx. 1.1 equiv of 1-(3-chlorophenyl)-piperazine (0.26
g, 1.3 mmol) in 20 mL DMF at25 ꢀC under N2 and fit-
ted with a reflux condenser was added a minimum of
1.25 equiv of anhydrous K2CO3 (0.27 g, 1.9 mmol). The
starting material 6 (0.24 g, 0.83 mmol) was dissolved in
a small amountof DMF (approx. 5 mL) and added ot
the reaction mixture. After stirring for 10 min, the
reaction mixture was then brought to reflux temperature
and allowed to reflux under an inert atmosphere for 16
h. Upon completion of the reaction, water was added
and the reaction mixture was extracted with four 30 mL
portions of CH2Cl2. The organic layers were combined
and dried over anhydrous MgSO4. The solventwas
removed under reduced pressure and the residue was
adsorbed onto silica gel and purified via flash column
chromatography. Analytical thin layer chromatography
(TLC) indicated the desired product had an Rf value of
approx. 0.75 when developed in 30:70 EtOAc/hexanes.
Purification via flash column chromatography (on silica
gel) was performed with 500 mL of hexanes until any
remaining starting material had completely eluted; then
the polarity of the eluent mixture was increased to 10:90
EtOAc/hexanes to afford the desired product 7 (0.24 g,
0.57 mmol, 69%) as an opalescentoil. 1H NMR
(CDCl3): d 7.31 (5H, m), 7.15 (1H, dd, J=8.1, 8.0 Hz),
6.82 (7H, m), 5.27 (1H, dd, J=8.4, 3.2 Hz), 3.19 (4H,
dd, J=5.1, 5.0 Hz), 3.04 (1H, dd, J=13.8, 8.5 Hz), 2.82
(2H, ddd, J=9.4, 8.1, 5.0 Hz), 2.73 (3H, ddd, J=6.9,
6.9, 3.1 Hz). 13C NMR (CDCl3): d 154.3, 152.7, 147
(faint), 140.6, 135.3, 130.4, 129.1, 129.1, 128.3, 126.5,
126.5, 119.7, 117.6, 117.5, 116.3, 116.1, 116.0, 114.3,
80.1, 65.7, 53.8, 53.8, 49.2, 49.2. 19F NMR (CDCl3): d-
À124.0. HRMS (EI) M+ calculated for C24H24ClFN2O:
410.91, found 410.16. Elemental analysis (%) calculated
for C24H24ClFN2O: C, 70.15; H, 5.89; N, 6.82; F, 4.62;
Cl, 8.63. Found: C, 70.43; H, 6.02; N, 6.75; F, n/a; Cl,
n/a.
1
oil. H NMR (CDCl3): d 7.28 (5H, m), 6.86 (1H, d,
J=9.3 Hz), 6.83 (1H, d, J=8.0 Hz), 6.79 (1H, d, J=9.3
Hz), 6.78 (1H, d, J=9.4 Hz), 5.06 (1H, dd, J=7.1, 4.4
Hz), 3.14 (1H, d, J=6.8 Hz), 3.12 (1H, d, J=7.7 Hz),
2.35 (2H, bs). 13C NMR (CDCl3): d 158.6, 156.0, 154.0,
139.0, 128.7, 128.0, 126.1, 126.1, 117.0, 116.9, 115.8,
115.6, 82.3, 49.0. 19F-NMR (CDCl3): d À124.0.
4.1.4. 2-(4-Fluorophenoxy)-2-phenyl-ethyl bromide (6).
To a rapidly stirred solution of 1.1 equiv of TiBr4 (3.15
ꢀ
g, 8.67 mmol) in 40 mL of DMF at25 C under N2 was
added dropwise 1.1 equiv of t-butyl nitrite (1.2 mL of
90% t-butyl nitrite, d=0.86, 8.6 mmol) dissolved in
approx. 5 mL of DMF. Upon addition of the t-butyl
nitrite, a color change from reddish orange to orange-
yellow was observed. The primary amine starting mate-
rial 5 (1.80 g, 7.8 mmol) was dissolved in approximately
5 mL of DMF and added dropwise to the reaction flask
over a 20-min period. Gas evolution was observed dur-
ing addition of the amine starting material and was
complete within approximately 5 min following its
complete addition. After complete gas evolution, the
reaction mixture was added to 150 mL of 20% aqueous
hydrochloric acid (20% aq HCl) and extracted with
4.1.6. 1-[2-(4-Fluorophenoxy)-2-phenylethyl]-4-(2-methoxy-
phenyl)-piperazine (8). a. Formation of the free base
form of 1-(2-methoxyphenyl)-piperazine.
The free base form of 1-(2-methoxyphenyl)-piperazine
was formed by dissolving a liberal amountof hte

J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
1489
corresponding hydrochloride saltin deionized waetr
and adding an excess of 2 M aqueous NaOH. After 10
min, the aqueous mixture was extracted with four 30
mL portions of ethyl ether. The combined organic lay-
ers were dried over anhydrous MgSO4 and the solvent
was removed under reduced pressure. The secondary
amine (free base form) of 1-(2-methoxyphenyl)-piper-
azine was isolated as an off-white solid and confirmed
min, the reaction mixture was then brought to reflux
temperature and allowed to reflux under an inert atmo-
sphere for 16 h. Upon completion of the reaction, water
was added and the reaction mixture was extracted with
four 30 mL portions of CH2Cl2. The organic layers were
combined and dried over anhydrous MgSO4. The sol-
ventwas removed under reduced pressure and ht e resi-
due was adsorbed onto silica gel and purified via flash
column chromatography. Analytical thin layer chroma-
tography (TLC) indicated the desired product had an Rf
value of approx. 0.75–0.8 when developed in 30:70
EtOAc/hexanes. Purification via flash column chroma-
tography (on silica gel) was performed with 500 mL of
hexanes until any remaining starting material had com-
pletely eluted; then the polarity of the eluent mixture
was increased to 10:90 EtOAc/hexanes to afford the
desired product 9 (0.11 g, 0.25 mmol, 18%) as an opa-
lescentoil. 1H NMR (CDCl3): d 7.32 (6H, m), 7.09 (1H,
d, J=7.6 Hz), 7.07 (1H, d, J=7.8 Hz), 7.05 (1H, d,
J=7.0 Hz), 6.88 (1H, d, J=9.3 Hz), 6.85 (1H, d, J=8.2
Hz), 6.80 (1H, d, J=9.3 Hz), 6.79 (1H, d, J=9.3 Hz),
5.28 (1H, dd, J=8.4, 3.3 Hz), 3.24 (4H, dd, J=5.0, 5.0
Hz), 3.05 (1H, dd, J=13.9, 8.5 Hz), 2.84 (2H, ddd,
J=10.8, 5.2, 5.2 Hz), 2.74 (3H, m). 13C NMR (CDCl3):
d 152 (faint), 151 (faint), 145 (faint), 140 (faint), 129.5,
129.5, 128.7, 128.7, 127.9, 126.1, 126.1, 118.6, 117.2,
117.1, 115.8, 115.8, 115.6, 112, 112, 79.7, 65.2, 53.4,
53.4, 48.7, 48.7. 19F NMR (CDCl3): d À63.1, À123.5.
HRMS (EI) M+ calculated for C25H24F4N2O: 444.46,
found 444.18. Elemental analysis (%) calculated for
C25H24F4N2O: C, 67.56; H, 5.44; F, 17.10; N, 6.30.
Found: C, 68.20; H, 5.82; F, 15.22; N, 5.90.
1
via NMR spectroscopy. H NMR (CDCl3): d 7.07 (1H,
ddd, J=9.1, 8.0, 4.5 Hz), 6.94 (1H, d, J=8.6 Hz), 6.93
(1H, d, J=4.1 Hz), 6.89 (1H, d, J=7.7 Hz), 3.87 (3H,
s), 3.38 (4H, d, J=15.3 Hz), 3.38 (2H, d, J=4.0 Hz),
3.14 (1H, bs), 2.85 (1H, bs), 1.83 (1H, s). 13C NMR
(CDCl3): d 152.2, 141.7, 122.7, 120.9, 118.1, 111.1, 55.2,
52.0, 52.0, 46.3, 46.3 Melting point: 180–183 ꢀC (dec.).
b. Formation the coupled product 8
To a well-stirred solution of approx. 1.6 equiv of 1-(2-
methoxyphenyl)-piperazine (0.34 g, 1.78 mmol) in 20
mL DMF at25 ꢀC under N2 and fitted with a reflux
condenser was added a minimum of 1.25 equiv of anhy-
drous K2CO3 (0.34 g, 2.45 mmol). The starting material
6 (0.245 g, 0.83 mmol) was dissolved in a small amount
of DMF (approx. 5 mL) and added to the reaction
mixture. After stirring for 10 min, the reaction mixture
was then brought to reflux temperature and allowed to
reflux under an inertamt osphere for 16 h. Upon com-
pletion of the reaction, water was added and the reac-
tion mixture was extracted with four 30 mL portions
of CH2Cl2. The organic layers were combined and
dried over anhydrous MgSO4. The solventwas removed
under reduced pressure and the residue was adsorbed
onto silica gel and purified via flash column chromato-
graphy. Analytical thin layer chromatography (TLC)
indicated the desired product had an Rf value of approx.
0.7 when developed in 30:70 EtOAc/hexanes. The flash
column chromatography (on silica gel) was performed
with 500 mL of hexanes until any remaining starting
material had completely eluted; then the polarity of the
eluent mixture was increased to 10:90 EtOAc/hexanes to
afford the desired product 8 (0.114 g, 0.28 mmol, 34%)
as an opalescentoil. 1H NMR (CDCl3): d 7.32 (5H, m),
6.9 (8H, m), 5.30 (1H, dd, J=8.5, 3.1 Hz), 3.86 (3H, s),
3.1 (5H, dd, J=13.7, 8.5 Hz), 2.8 (5H, m). 13C NMR
(CDCl3): d 152.6, 152.6, 141.7, 140.9, 129.1, 129.1,
129.1, 128.2, 126.5, 126.5, 123.3, 121.4, 118.6, 117.7,
117.6, 116.2, 116.0, 111.5, 79.9, 66.0, 55.7, 55.7, 54.2,
51.1, 51.1. 19F NMR (CDCl3): d-124.2. HRMS (EI) M+
calculated for C25H27FN2O2: 406.49; found 406.21.
Elemental analysis (%) calculated for C25H27FN2O2: C,
73.87; H, 6.69; F, 4.67; N, 6.89. Found: C, 73.84; H,
6.95; F, n/a; N, 6.92.
4.1.8. Formation of the hydrochloride salt (10) from piper-
azinyl-containing base (7). To a well-stirred solution of
7 (0.235 g, 0.57 mmol) in 30mL of anhydrous THF was
added 10 equiv of 2M HCl in anhydrous diethyl ether (3
mL, 6.0 mmol) at25 ꢀC under N2. The mixture was
allowed to stir for approx. 12 to 16 h. Upon completion
of the reaction, the reaction mixture was filtered and the
filtered residue was washed with freshly distilled
CH2Cl2. The solute was then allowed to dry and was
subsequently re-dissolved in approx. 5 mL of anhydrous
EtOH. The resulting solution was refluxed briefly for 5–
10 min and the solvent then was removed under reduced
pressure. The remaining solid was placed under vacuum
until all residual solvent was removed affording the
desired hydrochloride salt 10 as a greenish crystalline
solid (0.17 g, 0.39mmol, 68%). Confirmation of the HCl
salt 10 was made by treating a small amount of the final
product with strong base, extracting with ethyl ether,
removing the organic solvent under reduced pressure,
and confirming via NMR using CDCl3. NMR data
corresponded with those results previously reported for
the free base form 7. TLC was also performed (30:70
EtOAc/hexanes) and the Rf value of approx. 0.75 mat-
ched that of 7. Reverse phase TLC was performed on
the HCl salt 10 and Rf values of 0.3 in 100% CH3CN
and 0.8 in 2:98H2O/CH3CN were recorded, respectively.
4.1.7. 1-[2-(4-Fluorophenoxy)-2-phenylethyl]-4-(3-trifluoro-
methylphenyl)-piperazine (9). To a well-stirred solution
of approx. 1.1 equiv of m-trifluoromethylphenyl-piper-
azine (0.44 g, 1.9 mmol) in 20 mL DMF at25 ꢀC under
N2 and fitted with a reflux condenser was added a
minimum of 1.25 equiv of anhydrous K2CO3 (0.33 g, 2.4
mmol). The starting material 6 (0.41 g, 1.4 mmol) was
dissolved in a small amountof DMF (approx. 5 mL)
and added to the reaction mixture. After stirring for 10
4.1.9. Formation of the hydrochloride salt (11) from piper-
azinyl-containing base (8). To a well-stirred solution of
8 (0.11 4 g, 0.28 mmol) in 30 mL of anhydrous THF was

1490
J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
added 10 eq. of 2 M HCl in anhydrous diethyl ether (1.6
mL, 3.2 mmol) at25 ꢀC under N2. The mixture was
allowed to stir for approx. 12 to 16 h. Upon completion
of the reaction, the reaction mixture was filtered and the
filtered residue was washed with freshly distilled
CH2Cl2. The solute was then allowed to dry and was
subsequently re-dissolved in approx. 5 mL of anhydrous
EtOH. The resulting solution was refluxed briefly for 5–
10 min and the solvent was then removed under reduced
pressure. The remaining solid was placed under vacuum
until all residual solvent was removed affording the
desired hydrochloride salt 11 as a white crystalline solid
(0.0112 g, 0.025 mmol, 9%), albeitin low yield. Con-
firmation of the HCl salt 11 was made by treating a
small amount of the final product with strong base,
extracting with ethyl ether, removing the organic sol-
ventunder reduced pressure, and confirming via NMR
using CDCl3. NMR data corresponded with those
results previously reported for the free base form 8.
TLC was also performed (30:70 EtOAc/hexanes) and
the Rf value of approx. 0.7 matched that of 8. Reverse
phase TLC was performed on the HCl salt 11 and Rf
values of 0.25 in 100% CH3CN and 0.75 in 2:98H2O/
CH3CN were recorded, respectively.
KH2PO4, 2.4 mM CaCl2, 5.2 mM d-glucose, 25.0 mM
HEPES, 0.1 mM sodium ascorbate, 0.1 mM pargyline.
The buffer was adjusted to pH 7.4 with 5 M NaOH. The
ability of platelets to accumulate [3H]serotonin was
measured in the absence and presence of test drugs as
follows: a 490 mL aliquot of the platelet suspension was
added to glass tubes containing 5 mL test drug (various
concentrations, dissolved in DMSO), 5 mL DMSO (for
total determinations), or 5 mL fluoxetine hydrochloride
dissolved in DMSO (for nonspecific determinations;
final concentration, 10 mM). The assay tubes were pre-
incubated in a 37 ꢀC shaking water bath for 5 min. The
tubes were then returned to the ice bath and chilled for
10 min. [3H]Serotonin was added (5 mL of stock solu-
tion; final concentration, 10 nM), giving a total incu-
bation volume of 500 mL. All tubes were returned to the
37 ꢀC shaking water bath for 5 min to initiate neuro-
transmitter uptake. Uptake was terminated by chilling
the test tubes in the ice bath. After adding 3 mL ice-cold
KRH, each assay tube was immediately vacuum filtered
through glass fiber filters (Whatman GF/B) pretreated
with 0.1% polyethyleneimine. Filters were washed with
2Â3 mL ice-cold KRH, allowed to dry briefly under
vacuum, then placed in liquid scintillation vials. Scintil-
lation cocktail (8 mL) was added and the vials were
sealed, vortexed, and allowed to stand overnight.
Radioactivity was measured using liquid scintillation
spectroscopy (Packard Tri-Carb 1600 CA). Specific
uptake was defined as uptake at 37 ꢀC minus uptake in
the presence of 10 mM fluoxetine. Under these condi-
tions, specific [3H]serotonin uptake was typically 90%
of total uptake. The IC50 value for each test drug was
determined from displacement curves using at least
6 drug concentrations, each run in triplicate. Data were
transformed from dpm to% specific uptake and fitted
to a four-parameter logistic curve using commercial
computer software, from which the IC50 values are
obtained.
4.1.10. Formation of the hydrochloride salt (12) from piper-
azinyl-containing base (9). To a well-stirred solution of 9
(0.11 g, 0.25 mmol) in 30 mL of anhydrous THF was
added 10 equiv of 2 M HCl in anhydrous diethyl ether
ꢀ
(1.4 mL, 2.8 mmol) at25 C under N2. The mixture was
allowed to stir for approx. 12 to 16 h. Upon completion
of the reaction, the reaction mixture was filtered and the
filtered residue was washed with freshly distilled
CH2Cl2. The solute was then allowed to dry and was
subsequently re-dissolved in approx. 5 mL of anhydrous
EtOH. The resulting solution was refluxed briefly for 5–
10 min and the solvent then was removed under reduced
pressure. The remaining solid was placed under vacuum
until all residual solvent was removed affording the
desired hydrochloride salt 12 (0.0496 g, 0.103 mmol,
42%). Confirmation of the HCl salt 12 was made by
treating a small amount of the final product with strong
base, extracting with ethyl ether, removing the organic
solventunder reduced pressure, and confirming via
NMR using CDCl3. NMR data corresponded with
those previously reported for the free base form 9. TLC
was also performed (30:70 EtOAc/hexanes) and the Rf
value of approx. 0.75–0.8 matched that of 9. Reverse
phase TLC was performed on the HCl salt 12 and Rf
values of 0.35 in 100% CH3CN and 0.85 in 2:98H2O/
CH3CN were recorded, respectively.
Acknowledgements
K.G.P. thanks The Welch Foundation (Grant No. AA-
1278), Baylor University, and Baylor University Research
Committee for financial support of this project.
References and notes
1. Richelson, E. Mayo Clin. Proc. 1994, 69, 1069.
2. Hyttel, J.; Larsen, J. J. Acta Pharmacologica et Tox-
icologica 1985, 56 (S1), 146.
3. Schatz, P. F. Journal of Chemical Education 1997, 74 (7), 746.
4. Price, L. H.; Charney, D. S.; Delgado, P. L.; Goodman,
W. K.; Krystal, J. H.; Woods, S. W.; Heninger, G. R. J.
Clin. Psychiatry 1990, 51 (Suppl. 4), 44.
5. Asberg, M.; Eriksson, B.; Martensson, B.; Traskman-
Bendz, L.; Wagner, A. J. Clin. Psychiatry 1986,
47 (Suppl. 4), 23.
6. Frazer, A. Depression 1994, 2, 1.
7. Blier, P.; de Montigny, C.; Chaput, Y. J. Clin. Psychiatry
1990, 51 (Supp. 4), 14.
8. Lemberger, L.; Fuller, R. W.; Zerbe, R. L. Clinical Neu-
ropharmacology 1985, 8 (4), 299.
4.2. [3H] 5-HT uptake assay
Outdated human platelets were obtained from the blood
bank at Pitt County Memorial Hospital, Greenville,
NC. Platelets from 5–10 donors were pooled, 10%
dimethylsulfoxide was added, and aliquots were stored
frozen at À80 ꢀC until use. For assays, 5 mL of platelets
were thawed and suspended in 20 mL ice-cold Krebs-
Ringer-HEPES (KRH) buffer as previously described.14
The KRH buffer contained the following ingredients:
124.0 mM NaCl, 2.9 mM KCl, 1.3 mM MgSO4, 1.2 mM

J. M. Dorsey et al. / Bioorg. Med. Chem. 12 (2004) 1483–1491
1491
9. Herran, A.; Palacios-Araus, L.; Vazquez-Barquero, J. L.;
Diez-Manrique, J. F. J. Clin. Psychopharmacol. 1997,
17 (1), 67.
11. StDennis, C. U.S. Pharmacist 1999, 1, 65.
12. Gonzalez-Heydrich, J.; Peroutka, S. J. J. Clin. Psychiatry
1990, 51 (Suppl. 4), 5.
10. Modell, J. G.; Katholi, C. R.; Modell, J. A.; DePalma,
R. L. Clinical Pharmacology and Therapeutics 1997, 61 (4),
476.
13. Hyttel, J. Prog. Neuro-Psychopharmacol. & Biol. Psychiat.
1982, 6, 281.
14. Cozzi, N. V.; Foley, K. F. Biotechniques 2002, 32, 486.