Structure-activity relationships of N-substituted piperazine amine reuptake inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.05.049

We report the structure-activity relationships of further analogues in a series of piperazine derivatives as dual inhibitors of serotonin and noradrenaline reuptake, that is, with additional substitution of the phenyl rings, or their replacement by hetero

Full text of DOI:10.1016/j.bmcl.2006.05.049

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4349–4353
Structure–activity relationships of N-substituted piperazine
amine reuptake inhibitors
M. Jonathan Fray,* Gerwyn Bish, Paul V. Fish, Alan Stobie,
Florian Wakenhut and Gavin A. Whitlock
Sandwich Chemistry, Pfizer Global Research and Development, Sandwich, Kent CT13 9NJ, UK
Received 11 April 2006; revised 16 May 2006; accepted 16 May 2006
Available online 5 June 2006
Abstract—We report the structure–activity relationships of further analogues in a series of piperazine derivatives as dual inhibitors
of serotonin and noradrenaline reuptake, that is, with additional substitution of the phenyl rings, or their replacement by hetero-
cycles. The enantiomers of compounds 1 and 2 were also profiled, and possessed drug-like physicochemical properties. In particular,
compound (À)-2 lacked potent inhibitory activity against any of the important cytochromes P450 and high selectivity over a wide
range of receptors, which is unusual for a compound that inhibits human amine transporters.
ꢀ 2006 Elsevier Ltd. All rights reserved.
In the preceding paper, we reported the discovery of a
series of N-substituted piperazines, for example 1 and
2, as inhibitors of serotonin (5-HT) and noradrenaline
(NA) reuptake, with selectivity over the dopamine
(DA) transporter.1 Such compounds have potential
utility in treating depression and stress urinary inconti-
nence.2
Thus, condensation of the appropriate aryl or heterocy-
clic aldehyde,5 N-Boc-piperazine and benzotriazole in
refluxing toluene with azeotropic removal of water affor-
ded intermediates 27, which were not isolated, but added
directly as a solution in toluene to a solution of the rel-
evant benzylic Grignard reagent (2 eq.) or zinc reagent.6
The resulting N-Boc piperazines were then deprotected
under standard conditions to give 3–22. Compounds
23 and 24 were prepared in the same way, using phen-
ethylmagnesium bromide as the Grignard component
for the former and phenylacetaldehyde instead of benz-
aldehyde for the latter. Replacement of N-Boc pipera-
zine with N-Boc homopiperazine afforded compounds
25 and 26. The structures of all the compounds made
are shown in Table 1 and Figure 1.
NH
N
1 R = Cl
2 R = OEt
R
We report here the structure–activity relationships of
further analogues in the series: additional substitution
of the phenyl rings, or their replacement by heterocycles
(compounds 3–22); the spatial requirements of the two
aromatic rings in relation to the piperazine (compounds
23, 24); and replacement of the piperazine ring by homo-
piperazine (compounds 25, 26). The enantiomers of
compounds 1 and 2 have also been separated and
profiled.
We also separated7 and profiled the enantiomers of com-
pounds 1 and 2, and, for compound 2, determined the
absolute configuration of the (À) isomer to be (R).8
Compounds 3–26 were assayed for their ability to inhib-
it the uptake of [3H]5-HT, [3H]NA and [3H]DA in
HEK293 cells expressing a single human amine trans-
porter.9 Reference data were measured for fluoxetine,
reboxetine, and duloxetine. Results for 1–22 are shown
in Table 1 and will be discussed first; those for 23–26
and the enantiomers of 1 and 2 are in Figures 1 and 2,
respectively.
Compounds 3–22 were conveniently prepared by the
methods previously reported (Scheme 1).1,3,4
Keywords: Serotonin; Noradrenaline; Reuptake inhibition; Neuro-
transmitters; Depression; Incontinence.
*
In the series of analogues 3–10, where the chlorophenyl
ring of 1 was replaced, most of the structural changes
Corresponding author. Tel.: +44 0 1304 649175; fax: +44 0 1304
656776; e-mail: jonathan.fray@pfizer.com
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.049

4350
M. J. Fray et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4349–4353
NH
NBoc
Ar
N
Ar
N
N
a
b,c
Ar
NBoc
+
CHO
HN
N
R
N
27
3-22
Scheme 1. Reagents and conditions: (a) benzotriazole, toluene, Dean–Stark apparatus, reflux 6–24 h; (b) RCH2MgHal, THF, À78–0 ꢁC, 30 min or
RCH2ZnHal, THF, 20 ꢁC, 3 h, 35–80%; (c) 50% TFA, CH2Cl2, 20 ꢁC, 80–95%.
Table 1. Inhibition of human amine reuptake by N-substituted piperazines
NH
Ar
N
R
Compound
IC50 (nM)
5-HT
5.4
13
Ar
R
NA
22
DA
1
Ph
Ph
2-Cl phenyl
1300
>4000
NT
2
2-OEt phenyl
2-Naphthyl
1-Naphthyl
2-Pyridyl
16
>400
>400
>400
>400
100
>400
120
82
3
Ph
Ph
17
12
4
NT
NT
5
Ph
Ph
230
38
6
4-Pyridyl
NT
NT
7
Ph
Ph
2,3-Cl2 phenyl
2,4-Cl2 phenyl
2,5-Cl2 phenyl
2,6-Cl2 phenyl
2-Cl phenyl
2-OEt phenyl
2-Cl phenyl
2-OEt phenyl
2-Cl phenyl
2-OEt phenyl
2-Cl phenyl
2-OEt phenyl
2-Cl phenyl
2-OEt phenyl
2-Cl phenyl
2-OEt phenyl
3.9
11
8.2
8
>4000
>4000
970
9
10
Ph
Ph
12
14
46
79
39
12
40
30
10
15
28
13
17
16
590
11
12
3-Pyridyl
3-Pyridyl
2-Thiazolyl
2-Thiazolyl
5-Thiazolyl
5-Thiazolyl
2-F phenyl
2-F phenyl
3-F phenyl
3-F phenyl
4-F phenyl
4-F phenyl
75
15
9200
>4000
NT
13
14
45
340
120
21
NT
NT
15
16
NT
NT
17
18
130
38
NT
19
20
25
22
2300
NT
21
22
94
28
1600
NT
Fluoxetine
Reboxetine
Duloxetine
5200
11
4400
>25,000
870
6.0
19
Note: IC50 values are a geometric mean of at least n = 4. A difference of <2-fold should not be considered significant. NT, not tested.
NH
N
NH
NH
NH
N
N
N
OEt
Cl
23
5-HT IC50
NA IC50 >400 nM NA IC50
24
26
5-HT IC50 = 300 nM
NA IC50 8.8 nM
25
=
23 nM 5-HT IC50
=
=
23 nM
48 nM
5-HT IC50 = 42 nM
NA IC50 = 76 nM
=
DA IC50 = 400 nM
Figure 1. Inhibition of human amine reuptake by compounds 24–26.

M. J. Fray et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4349–4353
4351
Figure 2. Inhibition of human amine reuptake by compounds 1, 2, and their enantiomers.
(benzene annelation, replacement with pyridyl, addi-
tional chlorine atom) had a dramatically deleterious ef-
fect on potency against the NA transporter, whereas
good potency against the 5-HT transporter was fre-
quently retained. These data confirm, as we had ob-
served in other analogues,1 that the NA transporter
cannot tolerate much steric bulk (as in naphthyl) or
polarity (as in pyridyl) in the substituent R. Only com-
pounds 7 and 9 possessed comparable potency to 1
against the 5-HT transporter, but were still 5- to 6-fold
less potent than 1 against the NA transporter. The next
part of Table 1 shows the results for compounds 11–22,
in which the 2-chlorophenyl or 2-ethoxyphenyl substitu-
ent was retained, and the other phenyl ring replaced by
heterocycles or substituted with a single fluorine atom.
Overall, these alterations generally led to a loss of poten-
cy against both the 5-HT and NA transporters. Com-
pound 1 had an IC50 ratio 5-HT/NA 1:4, whereas for
compound 2 this ratio is near unity. The nature of the
heterocyclic group affected the relative potencies, as dis-
cussed below. For example, the 3-pyridyl and 5-thiazolyl
analogues (11, 15) with a chlorophenyl substituent re-
tained a similar IC50 ratio to 1 (i.e. 5-HT/NA 1:4–6) at
about 3-fold lower potency. In the same way that 1
was 3-fold more potent vs. the 5-HT transporter than
2, 11, and 15 were 3–4 more potent than 12 and 16,
respectively, against the 5-HT transporter; however,
the corresponding ethoxy analogues (12, 16) showed en-
hanced potency (5- to 6-fold) for the NA transporter
compared with 11 and 15. Consequently, 12 and 16 dem-
onstrated 2- to 3-fold selectivity for the NA transporter
compared to 11 and 15, which were selective for the 5-
HT transporter. In the series of fluorophenyl analogues
17–22, the compounds were 2- to 6-fold less potent than
1 versus 5-HT; vs. the NA transporter, the ethoxy ana-
logues (18, 20, 22) were 62-fold less potent than 2,
whereas in the chloro analogues (17, 19, 21) the fluoro
substituent was clearly preferred in the 3-position of
the phenyl ring. None of the analogues’ profiles was
superior to the parent analogues 1 and 2, although all
the compounds that were screened against the DA trans-
porter had low potency.
progressed further, as, although 25 possessed a balanced
profile, it lacked potency, and 26 possessed significant
selectivity for the NA transporter.
Figure 2 shows the potency of the enantiomers of
compounds 1 and 2 compared with the racemates. The
enantiomers in both series possessed significant activity.
In the case of compound 1, the (À)-isomer was about
3-fold more potent than the (+)-enantiomer against both
5-HT and NA transporters, but around 2-fold less
potent versus the DA transporter. The differences in
potency for the ethoxy enantiomers are probably not
significant. It is interesting to note that the enantiomers
of duloxetine both have similar potencies for 5-HT and
NA reuptake inhibition.10
Further profiling was conducted on (+)-1, (À)-1, (+)-2
and (À)-2. All four compounds possess moderate lipo-
philicity (logD7.4 octanol = 1.8 and 1.5 for the chloro
and ethoxy analogues, respectively), and formed sta-
ble, high-melting, crystalline hydrochloride salts with
good aqueous solubility (1–5 mg/ml at pH 7.4). Com-
pounds ( )-1, (+)-2, and (À)-2 demonstrated rapid
flux rates across Caco2 cells (Table 2), without evi-
dence for any affinity for efflux transporter proteins,
suggesting that absorption in vivo should be complete.
The set of enantiomers was profiled against a panel of
cytochrome P450 (CYP) enzymes in comparison with
known inhibitors (Table 2). Although (+)-1 and (À)-
1 were weak inhibitors of CYPs 3A4, 2C9, 2C19,
and 1A2, they were both potent inhibitors of 2D6
(comparable potency to fluoxetine). In contrast, (+)-2
and (À)-2 were both relatively weak inhibitors of
2D6 (comparable potency to venlafaxine), as well as
the other CYPs. One interesting property of all the
enantiomers was that they demonstrated good meta-
bolic stability in human liver microsome preparations,
but very poor stability when incubated with rat or dog
liver microsomes (Table 2). The reason for these dif-
ferences is not understood, but may indicate the com-
pounds are metabolised by different enzymes in
different species.
Figure 1 shows the results for compounds 23–26. Both
compounds 23 and 24 were reasonably potent against
the 5-HT transporter (IC50 = 23 nM). The former was
very significantly less active versus NA (IC50 > 400 nM),
and the latter was moderately potent versus DA, so nei-
ther analogue was considered a very interesting lead.
Likewise, the homopiperazines 25 and 26 were not
Further data obtained with (À)-2 indicated it possessed
>100-fold selectivity for 5-HT/NA transporters over a
wide variety of G-protein coupled receptors, including
adrenergic, dopaminergic, muscarinic, nicotinic, and
opiate receptors. It did, however, demonstrate modest
binding affinity for sodium channels (site 2, IC50
0.41 lM), calcium channels (L-type, diltiazem site, IC50

4352
M. J. Fray et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4349–4353
Table 2. Caco flux rates, metabolic half-lives, and CYP inhibition IC50 data
Compound
Caco fluxa
HLMbt1/2 (min)
RLMc t1/2 (min)
DLMd t1/2 (min)
CYP inhibition IC50 (lM)
2D6
3A4
1A2
2C9
2C19
( )-1
(+)-1
(À)-1
(+)-2
(À)-2
20/17
NT
NT
120
>120
>120
76
NTe
NT
NT
NT
9
NT
0.6
0.5
10
NT
>30
>30
10
NT
>30
26
NT
23
NT
14
2
2
4
5
23
14
30/41
33/36
NT
>30
NT
>30
NT
>30
96
4
30
30
Notes to Table 2.
a Flux across Caco2 cells was measured at 25 lM concentration. Figures quoted correspond to the initial flux rates (·10À6 cmÀ1) for apical to
basolateral (first figure) and basolateral to apical (second figure) sides of the cells.
b HLM, human liver microsomes; the maximum t1/2 measurable was 120 min.
c RLM, rat liver microsomes.
d DLM, dog liver microsomes.
e NT, not tested.
3. Katritzky, A. R.; Yannakopoulou, K.; Lue, P.; Rasala,
D.; Urogdi, L. J. Chem. Soc. Perkin Trans. 1 1989, 225.
binding IC50 16 lM).
4. Katritzky, A. R.; Strah, S.; Belyakov, S. A. Tetrahedron
0.73 lM) and the hERG potassium channel (dofetilide
1998, 54, 7167.
We have described the structure–activity relationships
5. Dondoni, A.; Fantin, G.; Fogagnolo, M.; Medici, A.;
for a series of N-alkylated piperazine derivatives.
Pedrini, P. Synthesis 1987, 998, now available
Although none of the new analogues were superior to
commercially.
either compound 1 or 2, the profiles of the enantiomers
6. N-Boc piperazine and N-Boc homopiperazine were pur-
of 2 were very promising. It is notable that both enanti-
omers had very similar profiles and possessed good,
drug-like physicochemical properties. Compound (À)-2
lacked potent inhibitory activity against any of the
important CYPs450 and possessed high selectivity over
a wide range of receptors, which is unusual for a com-
pound that inhibits human amine transporters.
chased from Sigma–Aldrich. The organometallic reagents
used were either obtained commercially or generated from
the commercial benzyl/naphthylmethyl/picolyl chloride or
bromide immediately prior to use.
7. HPLC conditions: for (+)-1, (À)-1: Chiralcel OJ column,
eluting with hexane/ethanol/diethylamine 80:20:0.2, detec-
tion at 254 nm; for (+)-2 and (À)-2 Chiralcel OD column,
eluting with hexane/isopropanol/diethylamine 70:30:0.3,
detection at 254 nm.
8. Absolute configuration was determined by a single crystal
X-ray diffraction study of the edisylate salt, grown by slow
evaporation from methanol.
Acknowledgments
9. The assays were a modification of those described by
Blakely, R. D.; Clark, J. A.; Rudnick, G.; Amara, S. G.
Anal. Biochem. 1991, 194, 302, HEK293 cells expressing a
single human amine transporter protein (7500 cells/well in
Millipore 96-well filter bottom plates) were pre-incubated
at 25 ꢁC for 5 min with assay buffer containing vehicle
(DMSO in water) or test compound. Uptake of neuro-
transmitter into the cells was initiated by the addition of
tritiated 5-HT (50 nM), NA (200 nM) or DA (200 nM)
substrates, the samples were shaken in an incubator at
25 ꢁC for 5 min (5-HT, DA) or 15 min (NA). The assays
were stopped by an ice-cold buffer wash followed by
filtration. The filters were then dried before measuring the
amount of radioactivity taken up into the cells by
scintillation counting. Potency of test compounds was
quantified as IC50 values, i.e. concentration required to
inhibit the specific uptake of radiolabelled substrate into
the cells by 50% relative to maximum (vehicle only) over a
10-point dose response range. A minimum of four
measurements of the IC50 were made. Typically, the
IC50s did not fit a normal distribution. Thus, geometric
means were used to minimise the distorting effect of
outliers. From validation data sets on all three transporter
assays, we established that the 95% confidence intervals
were 1.3- (5-HT, DAT) to 1.5-fold (NA) of the IC50, i.e.
IC50 = 4 nM would have 95% confidence intervals of 1.4–
6.6 nM.
We acknowledge the contributions of Drs. Stephen Phil-
lips and Donald Newgreen and their teams (Discovery
Biology Department) for screening data, and Miles
Tackett, Arnaud Lemaitre, and Bhairavi Patel for com-
pound synthesis. We are also grateful to Paula Bryans
and Anne-Laure Boutet who performed the separation
of the enantiomers of 1 and 2, to Neil Feeder, who per-
formed the single crystal X-ray diffraction study, and to
Anthony Harrison, Nicola Lindsay, and Ian Gurrell of
the pharmacokinetics and metabolism department for
additional studies.
References and notes
1. Fray, M. J.; Bish, G.; Brown, A. D.; Fish, P. V.; Stobie,
A.; Wakenhut, F.; Whitlock, G. A.; Bioorg. Med. Chem.
Lett., preceding paper.
2. (a) Reith, M. E. A. Neurotransmitter Transporters:
Structure, Function and Regulation, 2nd ed., Humana
Press: Totowa, NJ, 2002; (b) Iversen, L. Mol. Psychi-
atry 2000, 5, 357; (c) Schatzberg, A. F. J. Clin.
Psychiatry 2000, 61(Suppl. 11), 9; (d) Barker, E. L.;
Blakely, R. D. In Psychopharmacology: The Fourth
Generation of Progress; Bloom, F. E., Kupfer, D. J.,
Eds.; Raven Press: Philadelphia, 1995; p 21; (e) Bump,
R. C.; Norton, P. A.; Zinner, N. R.; Yalcin, I. Obstet.
Gynecol. 2003, 102, 76.
10. (a) Boot, J. R.; Brace, G.; Delatour, C. L.; Dezutter, N.;
Fairhurst, J.; Findlay, J.; Gallagher, P. T.; Hoes, I.;
Mahadevan, S.; Mitchell, S. N.; Rathwell, R. E.; Richards,

M. J. Fray et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4349–4353
4353
S. J.; Simmonds, R. G.; Wallace, L.; Whatton, M. A.
Bioorg. Med. Chem. Lett. 2004, 14, 5395; (b) Bymaster, F.
P.; Beedle, E. E.; Findlay, J.; Gallagher, P. T.; Krushinski,
J. H.; Mitchell, S. N.; Robertson, D. W.; Thompson, D.
C.; Wallace, L.; Wong, D. T. Bioorg. Med. Chem. Lett.
2003, 13, 4477.