Discovery of Benzamide Analogues as a Novel Class of 5-HT3 Receptor Agonists

Article abstract of DOI:10.1002/cmdc.201000444

A 5-HT₃ receptor agonist based on a benzamide scaffold was identified in a screening of a small commercial compound library, and an elaborate SAR study originating from this hit was performed. The design, synthesis, and functional characterisation of benzamide analogues at the 5-HT₃A receptor yielded substantial information concerning the analogues as 5-HT₃ receptor agonists. However, the potencies of the derived analogues were not significantly improved over that of the initial hit. The benzamide scaffold constitutes a novel type of 5-HT₃ receptor agonist, as it does not possess a positively charged functionality, which is essential for the binding of all orthosteric ligands to the receptor. Preliminary investigations suggest that the compounds may exert their effects on 5-HT₃ receptors by binding to an allosteric site in the receptor complex.

Full text of DOI:10.1002/cmdc.201000444

DOI: 10.1002/cmdc.201000444
Discovery of Benzamide Analogues as a Novel Class of
5-HT3 Receptor Agonists
Charlotte G. Jørgensen,*[a] Bente Frølund,[a] Jan Kehler,[b] and Anders A. Jensen*[a]
A 5-HT3 receptor agonist based on a benzamide scaffold was
identified in a screening of a small commercial compound li-
brary, and an elaborate SAR study originating from this hit was
performed. The design, synthesis, and functional characterisa-
tion of benzamide analogues at the 5-HT3A receptor yielded
substantial information concerning the analogues as 5-HT3 re-
ceptor agonists. However, the potencies of the derived ana-
logues were not significantly improved over that of the initial
hit. The benzamide scaffold constitutes a novel type of 5-HT3
receptor agonist, as it does not possess a positively charged
functionality, which is essential for the binding of all orthoster-
ic ligands to the receptor. Preliminary investigations suggest
that the compounds may exert their effects on 5-HT3 receptors
by binding to an allosteric site in the receptor complex.
Introduction
The neurotransmitter serotonin (5-hydroxytryptamine, 5-HT) is
found in virtually every major organ in the human body, where
it plays essential roles in numerous basic physiological func-
tions including developmental, cardiovascular, gastrointestinal
and endocrine functions, as well as in several central processes
involved in mood, libido, aggression, anxiety, cognition, sleep,
appetite sensation, and pain perception and processing.[1] Se-
rotonin mediates its physiological effects through seven classes
of membrane-bound receptors, six of which are composed of
7-transmembrane receptors.[1a] The seventh receptor class, the
5-HT3 receptors (5-HT3Rs), are ligand-gated ion channels that
belong to the superfamily of Cys-loop receptors, which also in-
cludes nicotinic acetylcholine (nACh), g-aminobutyric acid
type A (GABAA), and glycine receptors.[2]
this is the modulatory effects of benzodiazepines on GABAA
receptor signalling.[2d,3]
The medicinal chemistry development in the 5-HT3R field
has led to a wide range of high-affinity/potent orthosteric ago-
nists and competitive antagonists. Competitive 5-HT3R antago-
nists such as tropisetron, granisetron, and ondansetron, the so-
called ’-setrons’, are being administered in the clinic for the
nausea and emesis induced by chemotherapy and radiothera-
py against cancer and against postoperative nausea and eme-
sis.[2a,b] In contrast to the potent orthosteric ligands, the alloste-
ric ligands targeting 5-HT3Rs reported to date are all character-
ised by rather low potencies and by having activity on several
other Cys-loop receptors.
In the present study, we have discovered that compounds
based on the simple benzamide scaffold constitute a novel
class of 5-HT3R agonists. Based on a hit identified in a screen-
ing of a small compound library at the 5-HT3A receptor, we
have designed, synthesised, and functionally characterised a
series of benzamide analogues at 5-HT3Rs.
The Cys-loop receptors are pentameric complexes of subu-
nits, and the presence of several subunits in each receptor
family gives rise to a considerable degree of heterogeneity
when it comes to physiologically relevant receptor subtypes. In
the case of the human 5-HT3Rs, five different subunits have
been identified, termed 5-HT3A–5-HT3E.[1c,2a,c] Homomeric
5-HT3A receptors (consisting solely of 5-HT3A subunits) and
heteromeric complexes, wherein 5-HT3A co-assemble with
5-HT3B, 5-HT3C, 5-HT3D, or 5-HT3E, are formed from these
subunits. The two major physiological subtypes are the 5-HT3A
and 5-HT3AB receptors.[2a,c]
Results
Discovery of a novel 5-HT3R agonist
A compound library consisting of 3040 structurally diverse
compounds was screened for activity as agonists or potentia-
Signal transduction through the Cys-loop receptor is initiat-
ed by binding of the neurotransmitter to orthosteric sites situ-
ated at the interface between subunits in the extracellular part
of the receptor complex. However, the receptors are allosteric
proteins, and thus signalling can be initiated and modulated
by ligand binding to numerous other sites in the receptor
complex. Whereas the majority of Cys-loop receptor ligands
published to date target the orthosteric sites of the receptors,
several allosteric modulators acting through distinct sites to
potentiate or inhibit the response mediated by the orthosteric
agonist have been identified. The most prominent example of
[a] Dr. C. G. Jørgensen, Dr. B. Frølund, Dr. A. A. Jensen
Department of Medicinal Chemistry, University of Copenhagen
2 Universitetsparken, 2100 Copenhagen (Denmark)
Fax: (+45)35-33-60-40
E-mail: chj@farma.ku.dk
aaj@farma.ku.dk
[b] Dr. J. Kehler
H. Lundbeck A/S, 9 Ottiliavej, 2500 Valby (Denmark)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000444.
ChemMedChem 2011, 6, 725 – 736
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
725

C. G. Jørgensen, A. A. Jensen, et al.
MED
tors at a 5-HT3A–HEK293 cell line in the FLIPRꢁ Membrane Po-
tential (FMP) Blue assay. The compounds were co-applied with
EC15 concentrations of serotonin onto the cells, and one of the
compounds, 1a (Table 1), significantly augmented 5-HT3A sig-
nalling. Compound 1a was subsequently found to be an ago-
nist at the receptor, as it elicited receptor signalling in the ab-
sence of serotonin. As the structure of 1a is quite different
from those of other published 5-HT3R agonists, we decided to
further investigate the properties and potential of this struc-
ture by the design, synthesis, and pharmacological characteri-
sation of a series of 1a analogues.
Table 1. Overview of designed and synthesised compounds.[a]
Compd R1
Compd R2
1a
1b
1c
1d
1e
1 f
1g
1h
1i
1j
1k
1l
1m
1n
1o
1p
1q
1r
4-OEt
4-H
2a
2b
2c
2d
2e
2 f
2g
2h
2i
2j
2k
2l
2m
2n
2o
2p
2q
2r
Me
Et
nPr
iPr
tBu
CH2-cPr
cPr
cPent
cHex
-(CH2)4-[b]
-(CH2)5-[b]
-(CH2)2O(CH2)2-[b]
(CH2)2OCH3
(CH2)3OCH3
(CH2)2OH
CH2CN
(CH2)2NHBoc
(CH2)2NH2
CH2COOCH3
CH2COOH
2,6-diMePh
Ph
4-OMe
4-OnPr
4-OPh
2-OMe
2-OEt
3-OMe
3,4-diOEt
3,4-OCH2O-
3,4-O(CH2)2O-
4-OH
Design strategy
The structural design of analogues was based on the initial hit
1a and a semisystematic investigation of different parts of the
molecule by dividing the lead structure into three structural
domains: 1) the aromatic ring system and substituents (variety
of alkoxy groups in addition to electron-donating and elec-
tron-withdrawing groups) primarily in the 4-position; 2) the
amine part of the amide (alkyl, alkyls with handles and aromat-
ic rings); and 3) the linker between the two above-mentioned
parts (the carbon length and the functional amide group)
(Figure 1 and Table 1).
3-OH
3,4-diOH
4-SMe
4-SEt
4-Me
4-Ph
4-NH2
4-NMe2
4-Cl
4-Br
4-I
1s
1t
2s
2t
2u
2v
2w
2x
1u
1v
1w
1x
1y
1z
1aa
1ab
1ac
1ad
1ae
1af
CH2Ph
(CH2)2Ph
4-F
4-SOMe
4-SOEt
4-CF3
4-CN
4-COOMe
4-COOH
4-NO2
4-NHBoc
Figure 1. Design principle based on the two original hits.
Moreover, a small selection of compounds having a N-
phenyl-2-naphthamide scaffold (Figure 1, see naphthamide)
was also designed, because another hit from the commercial li-
brary, compound 6b, had an interesting pharmacological pro-
file.
R=3-pentyl
Compd
Compd R1
R1
Compd YÀZ (R1 =OEt)
3a
3b
3c
4-OMe (n=1)
4-OMe (n=2)
4-OEt (n=2)
3d
3e
4-OMe
4-OEt
4a
4b
4c
4d
4e
-NHCO-
-COO-
-SO2NH-[c]
-CH2NH-
-NHCONH-
Synthesis of analogues
The choice of synthesis strategy was based on a parallel syn-
thetic approach. As such the syntheses of benzamides and het-
eroaromatic amides with different substituents were per-
formed using N,N’-carbonylimidazole (CDI) as the single acti-
vating reagent followed by evaporation of solvents and direct
purification using automated equipment (Scheme 1).
Compd HetAr or cHex
Compd Ar
5a
5b
5c
5d
5e
5 f
2-furanyl
3-furanyl
2-thienyl
3-thienyl
cHex
1-naphthyl
2-naphthyl
4-pyridinyl
6a
6b
6c
6d
6e
6 f
Ph
2,6-diMe-Ph
2,6-diEt-Ph
2,6-EtMe-Ph
2-Me-Ph
For the investigation of the substituents on the aromatic
ring system (Table 1, compounds 1a–1af), high yields (>75%)
were generally observed except in the case of the 4-nitro-
(1ae), 4-dimethylamine (1t), and 2-ethoxybenzoic acid (1g)
which gave less than moderate yields.
2-Et-Ph
5g
5h
6g
6h
2-OMe-Ph
2-OEt-Ph
A selection of the compounds was further derivatised to
give another subset of compounds potentially more soluble
than the first collection of compounds. As such tert-butoxycar-
bamate (1af), methyl ester (1ac), and ethers (1c, 1h, 1j) were
[a] Isolated yields of compounds are given in the Supporting Information.
[b] Compounds 2j–2l: pyrrolidine, piperidine, morpholine. [c] R1 =4-OMe.
726
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 725 – 736

Benzamide Analogues as 5-HT3 Agonists
Scheme 1. General synthetic principle using N,N’-carbonylimidazole (CDI) to
pre-activate the carboxylic acid in THF; further reaction with amine gave the
respective benzamide after purification (yields of 1a–1af: 37–92%; of 2a–
2x: 2%–quant.). The majority of yields were in the range of 80% to quanti-
tative).
submitted to acidic treatment, alkaline hydrolysis, or demethy-
lation with boron tribromide giving the free amine (1s), car-
boxylic acid (1ad), or phenol/catechol compounds (1l, 1m
and 1n). Thioethers (1o and 1p) were oxidised with m-CPBA
to give the sulfoxide (1y and 1z).
For the exploration of the amine moiety (Table 1, 2a–2x) ac-
tivated 4-ethoxybenzoic acid was preformed with CDI and
added to the relevant amines (Scheme 1). This selection of
compounds were synthesised in yields ranging from high to
quantitative yields (>85–100%) with the exception of amines
such as tert-butyl amine and 2-aminoethanol (products 2e and
2o).
The ethyleneamine (2r) and methylene carboxylic acid (2t)
were synthesised by nickel boride reduction followed by acidic
treatment of N-(cyanomethyl)-4-ethoxybenzamide (2p) and hy-
drolysis of methyl 2-(4-ethoxybenzamido)acetate (2s), respec-
tively. Compounds with elongation of the distance from
phenyl to amide (Table 1) were synthesised (CH2 (3a), CH=CH
(3d and 3e)) using the CDI approach with 2-phenylacetic and
cinnamic acid. In the case of phenylpropanamide (3b and 3c),
phenylacrylamides (3d and 3e) were reduced by hydrogena-
tion on Pd/C.
Figure 2. Zacopride, metoclopramide, and analogues (compounds 1a and
7a–7e).
Functional characterisation of the benzamide analogues at
5-HT3A in the FMP assay
The functional properties of all analogues in the series were
determined at the 5-HT3A–HEK293 cell line in the FLIPRꢁ Mem-
brane Potential (FMP) Blue assay. We and others have found
the pharmacological properties exhibited by 5-HT3Rs in this
assay and those in the other fluorescence-based assay used in
this study, the Ca2+/Fluo-4 assay, to be in good agreement
with those obtained from studies in conventional electro-
physiological set-ups.[5]
The majority of the analogues were determined to be ago-
nists at 5-HT3A, giving rise to concentration-dependent increas-
es in fluorescence in the FMP assay (Table 2 and Figure 3a).
Another group of analogues did not elicit a significant fluores-
cence response at 5-HT3A in the assay in the concentration
ranges that their respective solubility properties allowed us to
test them in. However, co-application of these analogues to-
gether with EC10–EC20 concentrations of serotonin resulted in
significant potentiation of the signal evoked by the neuro-
transmitter applied on its own (Table 3 and Figure 3a). This
group of compounds is termed ’potentiators’ in the following,
but the question of whether or not they are actually true po-
tentiators of 5-HT3R signalling is addressed in the Discussion
section below. The increase in fluorescence intensity evoked
by the agonists and by the ’potentiators’ (in the presence of
serotonin) could be eliminated completely by the presence of
standard 5-HT3R antagonist ondansetron in the assay (data not
shown). This strongly suggests that the responses observed in
the FMP assay are indeed mediated by 5-HT3A and cannot be
ascribed to inherent fluorescent properties or to nonspecific
cellular effects of the compounds. Finally, a third group of the
benzamide analogues were inactive at the 5-HT3A in the con-
centration ranges tested, both as agonists, ’potentiators’ and
antagonists (See footnote to Table 2).
With the sterically hindered compounds, 2,6-disubstituted
phenyls, (2u, 6b, 6c and 6d) generally low yields were ob-
served using CDI but sufficient amounts of products for phar-
macological investigations were obtained. The monosubstitut-
ed compounds such as 6e–6h gave from moderate to high
yields (61–97%).
The amide bond (Table 1) was transformed into isosteric
equivalents[4] by using borane reduction of amide 1a to ach-
ieve methyleneamino (4d), CDI activation of 2-ethylbutanoic
acid or 4-ethoxybenzoic acid giving the reversed amide (4a)
and ester (4b), sulfonamide formation (4c) from 4-ethoxyben-
zene-1-sulfonyl chloride, and urea (4e) synthesis from phenyl
4-ethoxyphenylcarbamate.
To compare the activity of the benzamide compounds with
those of the structurally related standard 5-HT3R antagonists
metoclopramide and zacopride, the relevant moieties were
used to prepare the mixed analogues (1a, 7a–7d) shown in
Figure 2.
Metoclopramide (7b) and zacopride hybrid compounds (1a,
7a, 7c, 7d) were prepared using the CDI reaction conditions
(Scheme 1), although in the case of metoclopramide 7b, DMF
was used instead of THF followed by traditional workup. The
majority of the analogues were isolated in moderate to high
yields (76–90%) although compound 7d and metoclopramide
7b gave low yield when THF was employed as the solvent.
As can be seen from Tables 2 and 3 and from the selected
examples of concentration–response curves shown in Figure 3,
the responses evoked by the vast majority of agonists and ’po-
tentiators’ characterised in the FMP assay did not reach satura-
ChemMedChem 2011, 6, 725 – 736
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
727

C. G. Jørgensen, A. A. Jensen, et al.
MED
Table 2. Functional properties of analogues as agonists at h5-HT3A–HEK293 cells in the FMP Blue assay.[a]
Agonists
EC50 [mm] (pEC50 ÆSEM)
Rmax ÆSEM [% of Rmax for 5-HT]
Serotonin
0.25 (6.60Æ0.03)
5.4 (5.26Æ0.06)
100
95Æ7
7d
Concentration of test compounds [mm]
Compd
10
30
100
300
1000
1a
1b
1c
1d
1 f
1g
1h
1i
NR
NR
NR
17Æ2
NR
NR
NR
NR
NR
NR
19Æ1
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
11 Æ2
NR
NR
15Æ3
NR
NR
37Æ2
NR
NR
NR
NR
NR
17Æ1
52Æ3
13Æ2
NR
17Æ1
NR
NR
NR
NR
NR
NR
34Æ3
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
30Æ2
NR
NR
NR
29Æ5
41Æ4
NR
39Æ4
NR
12Æ3
70Æ3
NR
74Æ7
29Æ4
45Æ7
91Æ4
12Æ3
NR
–
59Æ5
–
–
52Æ4
NR
13Æ2
17Æ2
21Æ2
23Æ2
95Æ3
121Æ5
43Æ2
13Æ2
97Æ7
34Æ4
22Æ2
NR
19Æ1
NR
NR
134Æ11
NR
23Æ2
NR
13Æ3
37Æ2
NR
25Æ2
NR
NR
NR
53Æ3
NR
NR
57Æ3
42Æ3
57Æ2
143Æ3
127Æ6
92Æ4
32Æ2
122Æ11
86Æ5
41Æ2
33Æ3
67Æ3
17Æ3
40Æ5
114 Æ7
13Æ2
82Æ4
29Æ4
43Æ6
77Æ2
13Æ4
–
99Æ5
–
1j
90Æ4
1k
1o
1p
1q
1t
1u
1v
1w
1x
1aa
1ab
1ac
2b
2c
2d
2e
2 f
–
–
–
–
–
–
–
–
–
–
–
–
–
–
45Æ3
63Æ4
–
–
–
–
–
–
–
2g
2k
2m
2r
17Æ2
12Æ2
12Æ1
–
2s
2u
3a
3b
3c
3d
3e
4a
4d
5a
5d
5e
5 f
9Æ3
39Æ5
19Æ2
20Æ3
97Æ6
103Æ7
47Æ2
129Æ11
82Æ5
15Æ2
47Æ3
30Æ5
31Æ2
–
–
–
–
–
–
–
–
–
NR
NR
12Æ4
9Æ1
NR
NR
NR
NR
NR
NR
NR
87Æ5
88Æ7
21Æ2
36Æ3
52Æ3
NR
19Æ1
NR
11Æ1
21Æ2
10Æ2
NR
25Æ2
NR
NR
NR
NR
NR
NR
–
–
5g
6h
7c
75Æ4
NR
NR
–
–
31Æ4
[a] The functional characterisation of the analogues in the assay was performed as described in the Experimental Section. The ability of the compounds to
potentiate 5-HT-induced 5-HT3A receptor signalling was testing by co-application of the compounds and an EC10–EC20 5-HT concentration onto the cells.
The responses obtained at various concentrations of the respective test compounds are given as % response of the maximal response elicited by 5-HT at
the receptor in the same plate on the same day. The data represent the mean ÆSEM of 3–4 independent experiments, except in a few cases where the
data are based on two independent experiments. The functional properties of the analogues that did not elicit an agonist response at the concentrations
tested but potentiated the signal when co-applied with EC10–EC20 concentrations of serotonin are presented in Table 3. Analogues 1e, 1l, 1m, 1n, 1r, 1y,
1z, 2a, 2j, 2n, 2o, 2p, 2t, 2v, 2w, 2x, 4c, 4e, 5c, 5h, 6a, 6e, 6 f, 6g, and 7a were inactive as agonists, potentiators or antagonists at the maximum
tested concentrations (100–300 mm). None of the compounds showed any response at a concentration of 3 mm; –: not tested due to solubility problems;
NR: no significant response were obtained at these compound concentrations compared to the control, that is, application of buffer (in the test for ago-
nism) or buffer supplemented with EC15–EC20 concentration of serotonin (in the test for allosteric potentiation).
tion levels in the concentration ranges tested. Hence, it was
not possible to determine the Rmax or EC50 values for the li-
gands. Instead, Tables 2 and 3 list the functional responses eli-
cited by different concentrations of the respective analogues
and are presented in percent relative to the Rmax value for sero-
tonin. The active analogues in the series were characterised by
728
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 725 – 736

Benzamide Analogues as 5-HT3 Agonists
Figure 3. Functional properties of selected benzamide analogues at the 5-HT3A–HEK293 cell line in the FMP Blue assay; experiments were performed as de-
scribed in the Experimental Section. a) Concentration–response curves for ten analogues tested as agonists (top panel) and for three analogues tested as ’po-
tentiators’ by co-application of the ligands with 0.15 mm serotonin (~EC15) (bottom panel). Responses are given as percent of the maximal response elicited
by serotonin in the assay, and concentration–response curves for serotonin are given for comparison. The figure depicts data from a single representative ex-
periment, and error bars are omitted for clarity. b) Concentration–response curves for serotonin alone or co-applied with four different concentrations of the
agonist 5a or the ’potentiator’ 6d. Responses are given as the changes in fluorescence intensity (DFU) elicited by the applied ligands in the assay. The figure
depicts data from a single experiment (a total of three independent experiments were performed for each of the two compounds with very similar results),
and data are given means ÆSD of duplicate determinations.
Table 3. Functional properties of analogues as ’potentiators’ at h5-HT3A–
logues such as 1d, 1o, 2u, 3d, and 3e, which all elicited sig-
nificant agonist responses at concentrations as low as 10 mm,
appeared to be the more potent agonists in the series. Inter-
estingly, several of the agonists (for example 1k, 1o, 1t, 1ac,
and 3d) elicited responses in the FMP assay significantly
higher than the Rmax value obtained for serotonin (Table 2 and
Figure 3a).
HEK293 cells in the FMP Blue assay.[a]
Concentration of test compounds [mm]
Compd
10
30
100
300
1000
1s
1ad
1ae
2h
2i
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
33Æ4
55Æ2
NR
31Æ5
NR
NR
59Æ5
–
68Æ4
41Æ2
30Æ3
34Æ3
55Æ3
83Æ3
31Æ2
54Æ3
29Æ2
33Æ3
–
–
–
70Æ4
–
–
–
–
–
–
–
–
–
To shed further light on the functional properties of agonists
and ’potentiators’, representative compounds from each class
were characterised in greater detail at the 5-HT3A–HEK293 cell
line in the FMP assay. The relationship between concentration
and response for serotonin was determined when the neuro-
transmitter was applied onto the cells alone or co-applied with
four different concentrations of 5a or 6d (Figure 3b). As ex-
pected, the agonist 5a gave rise to significant responses in the
absence of serotonin. Furthermore, co-application of serotonin
with increasing concentrations of 5a resulted in a left-shifted
concentration–response curve and an increased Rmax value for
serotonin (Figure 3b). Also in accordance with its profile in the
basic functional characterisation, the ’potentiator’ 6d did not
give rise to significant responses on its own in the FMP assay.
Instead, the compound increased the potency and maximal
2l
2q
4b
5b
6b
6c
39Æ3
45Æ3
34Æ1
–
–
6d
[a] Refer to Table 2 for footnote details.
fairly similar potencies, as they displayed activity at the recep-
tor in the mid- to high-micromolar range (Table 2 and Table 3).
However, 7d (the only analogue for which it was possible to
obtain a complete concentration–response curve) and ana-
ChemMedChem 2011, 6, 725 – 736
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
729

C. G. Jørgensen, A. A. Jensen, et al.
MED
response of serotonin in a concentration-dependent manner
(Figure 3b).
Functional characterisation of selected benzamide ana-
logues at 5-HT3A and 5-HT3AB in the Ca2+/Fluo-4 assay
To verify the functional properties displayed by the analogues
at 5-HT3A in the FMP assay and to investigate their pharmacol-
ogy at different 5-HT3R subtypes, a selection of eight ana-
logues from the compound series were characterised at the
two major physiological 5-HT3Rs, the 5-HT3A and 5-HT3AB sub-
types, in the Ca2+/Fluo-4 assay. In the Ca2+/Fluo-4 assay, the
responses of several of these analogues reached saturation
levels, suggesting
a general left-shifted concentration–re-
sponse relationship for the analogues in this assay relative to
the FMP assay. This fact enabled the determination of Rmax and
EC50 values for five of the eight selected analogues (Table 4
and Figure 4).
In general, a good concordance exists between the function-
al properties observed for the selected analogues at 5-HT3A in
the two fluorescence-based assays. Compounds 1o, 1p, 3d,
and 3e were the more potent analogues, whereas 1h and 1t
displayed somewhat weaker agonist potencies. Furthermore,
1p, 3d and in particular 1o and 1t, displayed significantly
higher Rmax values than serotonin at both 5-HT3A and 5-HT3AB
in the assay (Table 4 and Figure 4). Thus, the apparent super-
agonism displayed by 1o and 1t in the FMP assay (Table 2 and
Figure 3) was reproduced in the Ca2+/Fluo-4 assay. Finally, nei-
ther of the two ’potentiators’ in the selection, 6d nor 5b, dis-
played significant agonist activity in the concentration ranges
tested (Figure 4 and Table 4). When co-applied with EC10–EC20
concentrations of serotonin, 6d was able to enhance the
signal evoked by the neurotransmitter on its own, whereas 5b
was not (Table 4). As for subtype selectivity, the eight selected
analogues displayed similar potencies and efficacies at the
5-HT3A and 5-HT3AB subtypes (Table 4).
Figure 4. Functional properties of serotonin and eight benzamide analogues
as agonists at the human a) 5-HT3A and b) 5-HT3AB receptors stably ex-
pressed in HEK293 cells in the Ca2+/Fluo-4 assay. The assay was performed
as described in the Experimental Section. Responses are given as the
change in fluorescence intensity (DFU) elicited by the ligands in the assay.
The figure depicts data from a single experiment (representative of the total
of three individual experiments performed for each compound), and data
are given means ÆSD of duplicate determinations.
cell lines stably expressing the three major neuronal nAChR
subtypes in the FMP (a4b2 and a3b4 nAChRs) and Ca2+/Fluo-4
(a7 nAChR) assays. None of the analogues 1o, 1p, 1t, 3d, 3e,
5b, or 6d displayed significant activity at any of the three re-
ceptors when tested as agonists, as potentiators, or as antago-
nists in concentrations up to 300 mm, and 1h was inactive at
all three receptors in concentrations up to 1000 mm (shown as
an example for the a7 nAChR in Figure 5).
Functional characterisation of selected benzamide ana-
logues at nAChRs
The degree of selectivity of the analogues for the 5-HT3Rs over
other Cys-loop receptors was investigated by functional char-
acterisation of eight analogues from the series at mammalian
Table 4. Functional properties of selected analogues at h5-HT3A– and h5-HT3AB–HEK293 cell lines in the Ca2+/Fluo-4 assay.[a]
h5-HT3A
nH ÆSEM
h5-HT3AB
nH ÆSEM
Agonists[b]
pEC50 ÆSEM
Rmax ÆSEM [%]
pEC50 ÆSEM
Rmax ÆSEM [%]
Serotonin
1o
6.62Æ0.02
4.97Æ0.09
4.92Æ0.04
4.34Æ0.06
5.06Æ0.07
5.11Æ0.06
3.1Æ0.3
3.1Æ0.4
2.0Æ0.4
2.1Æ0.3
3.0Æ0.4
1.9Æ0.3
100
6.75Æ0.07
5.11Æ0.09
5.16Æ0.07
4.59Æ0.05
5.27Æ0.06
5.15Æ0.09
2.9Æ0.2
3.2Æ0.4
2.3Æ0.2
1.9Æ0.3
2.7Æ0.2
2.3Æ0.4
100
207Æ11
163Æ3
223Æ6
183Æ12
106Æ9
240Æ12
190Æ14
232Æ10
175Æ6
99Æ6
1p
1t
3d
3e
[a] The experiments were performed as described in the Experimental Section, and the data is based on 3–4 individual experiments. [b] Analogue 5b was
inactive as an agonist, as a ’potentiator’ and as antagonist; compound 1h displayed weak agonist response at concentrations >100 mm, and compound
6d potentiated serotonin-mediated signalling at concentrations ꢀ30 mm.
730
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 725 – 736

Benzamide Analogues as 5-HT3 Agonists
pride, and tropisetron displayed Ki values of 550 nm (pKi Æ
SEM=6.26Æ0.07, n=3), 2.7 nm (pKi ÆSEM=8.57Æ0.07, n=
3), and 0.54 nm (pKi ÆSEM=9.27Æ0.07, n=3), respectively
(Figure 6). The Ki values for the three reference antagonists are
Figure 6. Binding profiles of standard competitive 5-HT3R antagonists tropi-
setron, zacopride, metoclopramide, and 12 benzamide analogues using
membranes from 5-HT3A–HEK293 cells in the [3H]GR65630 binding assay.
The binding assay was performed as described in the Experimental Section.
The figure depicts data from a single representative experiment, and error
bars are omitted for clarity.
in good agreement with previously reported binding affinities
to 5-HT3Rs in neuroblastoma cell lines.[7]
Figure 5. Functional properties of seven benzamide analogues at the human
a7 nAChR stably expressed in GH3 cells in the Ca2+/Fluo-4 assay. The assay
was performed as described in the Experimental Section. Responses are
given as the change in fluorescence intensity (DFU) elicited by the ligands
in the assay. a) Test for agonist activity at the a7 nAChR. b) Test for potentia-
tion of the response elicited by 2 mm ACh (~EC20) at the a7 nAChR. The
figure depicts data from a single experiment (representative of the total of
2–3 individual experiments performed for each compound).
Discussion
In the present study, we have discovered a novel class of
5-HT3R agonists having a benzamide scaffold. The original hit
compound 1a identified in a screening of a commercial com-
pound library at the 5-HT3A receptor has been subjected to ex-
tensive medicinal chemistry optimisation, and the functional
properties of a total of 102 analogues have been characterised
at 5-HT3A. Following an overview of the structure–activity rela-
tionships displayed by the analogues, we will discuss the infor-
mation concerning the functional properties and the mode of
action of the benzamide analogues obtained in this study.
Characterisation of the binding affinities of selected benz-
amide analogues to the orthosteric site of 5-HT3A
Twelve analogues from the series (1h, 1k, 1p, 1t, 1ac, 3d, 3e,
5a, 5b, 6b, 7c, and 7d) were characterised in a competition
binding assay to membranes from 5-HT3A–HEK293 cells using
the orthosteric radioligand [3H]GR65630. In a previous study,
the KD value for this radioligand at 5-HT3A has been deter-
mined to be 180 pm using this assay.[6] To facilitate the puta-
tive competition of the benzamide analogues with
[3H]GR65630 for the orthosteric site, we used a tracer concen-
tration of 50 pm, considerably below the KD value, in these
experiments.
Structure–activity relationships of the benzamide analogues
We cannot claim that our extensive medicinal chemistry explo-
rations into the original hit from the screening and the benz-
amide scaffold resulted in analogues with substantially im-
proved 5-HT3R activities (Tables 2 and 3). Furthermore, the lim-
ited solubility of the compounds has made it impossible to de-
termine accurate EC50 and Rmax values for most of the ana-
logues. Hence, the following discussion of the structure–
activity relationships of the analogues will mainly focus on ob-
served trends rather than on pronounced differences between
the analogues.
None of the selected analogues significantly inhibited the
binding of [3H]GR65630 to the receptor in the concentration
ranges tested, with the exception of 7d which displaced radio-
ligand binding in a concentration-dependent manner with a Ki
value of 9.2 mm (pKi ÆSEM=5.04Æ0.09, n=3). The standard
competitive 5-HT3R antagonists metoclopramide (7b), zaco-
ChemMedChem 2011, 6, 725 – 736
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
731

C. G. Jørgensen, A. A. Jensen, et al.
MED
The benzamide analogue can be divided into three structur-
al domains: the aromatic ring system, the amine moiety, and
the ’linker’ between them. The SAR observations made for all
three are discussed below. Furthermore, the findings for a sub-
group of analogues clarifying the structural similarities be-
tween the benzamide analogues and standard 5-HT3R antago-
nists such as zacopride and metoclopramide are discussed.
ative to 1a. Interestingly, the N-methylated analogue 2a is
completely inactive in the concentration ranges tested, which
suggests a need for some bulk of the N-substituent for recep-
tor binding. The fact that introduction of hydrophilic N-sub-
stituents in 2m, 2n, 2o, 2p, 2r, 2s, and 2t result in inactive or
almost inactive analogues suggests that this part of the mole-
cule docks into a hydrophobic cavity in the receptor.
The aromatic ring system
Linker
A comparison of the agonist properties of 1b with those of its
monomethoxy-substituted analogues demonstrates that al-
though the introduction of a methoxy group in the 3- or 4-po-
sitions of the phenyl ring (1h and 1c) increases potency of the
benzamide analogue approximately tenfold, the 2-methoxy an-
alogue 1 f is roughly equipotent with the unsubstituted ana-
logue. The same trend is observed when comparing the 4-
ethoxy and 2-ethoxy analogues 1a and 1g with 1b. Therefore
the presence of an electron donor in the 3- or 4-positions ap-
pears to be beneficial for the 5-HT3R activity. Interestingly, in
contrast to the increased 5-HT3R activity arising from methoxy
and ethoxy substitutions at the 3- and 4-positions, analogues
with hydroxy groups at these positions (1l and 1m) are in-
active (Table 2 and Table 3).
In a third subgroup of analogues we investigated the impor-
tance of the amide moiety and the distance between this
group and the phenyl ring for the 5-HT3R activity of the benz-
amide analogue. Substitution of the carbonyl group in 4d with
a methylene group retains agonist activity at the 5-HT3R, albeit
somewhat lower than that of 1a. This is quite surprising, as
the conversion of the amide to an amine not only introduces a
positive charge in the structure but also changes its planar
structure. The homologues 3a, 3b, and 3c all displayed activi-
ty lower than that of 1a and 1c. In contrast, the corresponding
unsaturated analogues 3d and 3e are amongst the most
potent agonists in the series. This suggests that the planar
structure and the decreased loss in entropy arising from this
conformational restraint is important for the activity of the
homologues, and 3d and 3e certainly constitute interesting
leads for future medicinal chemistry optimisation efforts.
As expected from the increased activity of the 4-ethoxy and
4-methoxy analogues 1a and 1c, the corresponding thioether
analogues 1o and 1p display significantly higher potencies
than 1b, 1o so much so that it belongs to the more potent
analogues in the series. Similarly the 4-OH analogue 1l and
the 4-NH2 analogue 1s displayed weak activities, and thus the
relatively high agonist potency and efficacy of the derived 4-
N(CH3)2 analogue 1t is interesting. Finally, analogues with large
substituents such as phenoxy (1e) and phenyl (1r) groups in
the 4-position are completely inactive, perhaps indicative of a
steric clash between the 4-substituent and residues lining the
binding pocket for the benzamide analogue (Tables 2 and 3).
However, replacement of the phenyl with 1-naphthyl and 2-
naphthyl rings (5 f and 5g) results in analogues with decreased
or similar agonist activities as 1b, respectively.
To investigate the importance of the phenyl group for the
pharmacological activity of the analogues, it was substituted
with other ring systems. The 2-thiophen, 3-furan, and 3-thio-
phen analogues (5c, 5b, and 5d) were without activity, where-
as the 2-furan analogue 5a is slightly more potent than 1b.
Surprisingly, some receptor activity is retained in the more flex-
ible cyclohexane analogue 5e.
Overall SAR conclusions
The observed trends for the functional properties of the benz-
amide analogues outlined above form the basis for an under-
standing of their structure–activity relationships. A few outliers
in the form of analogues displaying unexpected functionalities
exist, most notably 4d and 5e, both of which retain activity,
albeit decreased, at the 5-HT3R despite substantial changes
being introduced into the benzamide scaffold. Nevertheless,
the bulk of the data seems to be coherent and reasonable
from a SAR perspective. O-Alkyl or S-alkyl substituents in the 3-
and 4-positions of the phenyl ring and a relatively bulky N-
alkyl amide group substituent in the benzamide analogue
have been identified as contributors to 5-HT3R activity of the
compounds, and interestingly it is possible to extend the
length of the linker between the phenyl and amide groups
and retain activity.
Structural similarity between the benzamide analogues and
orthosteric 5-HT3R antagonists
The amine group
The competitive 5-HT3R antagonists metoclopramide (7b) and
zacopride (7e) both contain a benzamide group. To investigate
the relationship between the benzamide analogues in this
study and these standard antagonists, the metoclopramide an-
alogue 7c and the zacopride analogue 7d, which both contain
a benzamide group with a substitution pattern identical to
that in 1a, were synthesised and characterised functionally.
Furthermore, the functional properties of a 1a analogue (7a)
with a substitution pattern identical to those of the benzamide
In another series of analogues the contributions of the amine
moiety system to the 5-HT3R activity was investigated by vary-
ing this functionality using the 4-ethoxy-substituted benza-
mide scaffold as a basis. Here, 1a, the original hit identified in
the screening with its N-(3-pentyl)-amide substituent is found
to be superior to the derived analogues (Tables 2 and 3). Ana-
logues mono-N-substituted with small aliphatic groups (2b,
2c, 2d, 2e, 2 f, 2g, 2h and 2i) all display decreased activity rel-
732
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 725 – 736

Benzamide Analogues as 5-HT3 Agonists
groups in zacopride (7e) and metoclopramide (7b) were char-
acterised. Whereas both 7a and 7c displayed negligible 5-HT3R
activities, 7d was the most potent agonist in the study
(Table 2). It is interesting to note that the difference in the sub-
stitution patterns on the phenyl rings of zacopride (7b) and
7d has dramatic implications for the functional properties of
the ligands, the former being a potent antagonist and the
latter a relatively potent agonist. On the other hand, caution
should be taken when comparing the functional profiles of 7d
with those of the other benzamide analogues in the series. As
discussed below, the agonism exhibited by 7d may arise from
an entirely different mode of action at the receptor than that
underlying the agonist responses elicited by the other benz-
amide analogues in the series.
portant to verify this intriguing functional property in conven-
tional electrophysiological setups.
Selectivity profile of the benzamide analogues
Judging from the functional properties exhibited by a selection
of the analogues at 5-HT3A and 5-HT3AB (Table 4 and Figure 4),
the benzamide compounds do not appear to differentiate be-
tween the two subtypes. This is hardly surprising considering
that relatively few ligands are reported to possess subtype se-
lectivity to date. To investigate whether the observed activity
of the benzamide analogue was specific for the 5-HT3Rs or if
the ligands act promiscuously at Cys-loop receptors, a selec-
tion of them were also characterised at the a7, a4b2, and
a3b4 nAChRs. The nAChRs are the Cys-loop receptor family
members that display the closest amino acid sequence homol-
ogy to the 5-HT3Rs, and a considerable number of compounds
acting at 5-HT3Rs have been reported to also possess activities
at the nAChRs, in particular the a7 subtype.[8] Hence, the lack
of activity of the benzamide analogues as agonists, potentia-
tors, or antagonists at these nAChRs suggests that they may
be selective for 5-HT3Rs over other Cys-loop receptors
(Figure 4). Furthermore, in our research group, the compound
library containing the hit compound 1a has been screened
against about ten other mammalian cell lines expressing other
neurotransmitter receptors and transporters (including GABAA
receptors, glycine receptors, and excitatory amino acid trans-
porters), without 1a showing any activity at these targets
(data not shown). Although this does suggest some specificity
of the benzamide analogues for the 5-HT3Rs over other mem-
brane-bound receptors, it does seem very unlikely that these li-
gands are completely selective, considering their relatively
simple structures and their relatively modest activities at the
5-HT3Rs.
Functional properties of the benzamide analogues
In this section, some of the interesting properties exhibited by
the benzamide analogues, such as the ’potentiation’ displayed
by some and the superagonism displayed by others, their
selectivity profile, and their mode of action are discussed.
The ’potentiators’
Several of the benzamide analogues were inactive as agonists
at 5-HT3A at the concentrations tested but significantly aug-
mented the signal elicited by EC10–EC20 concentrations of sero-
tonin (Figure 3 and Table 3). The difference between agonists
and ’potentiators’ was further investigated for 5a and 6d,
where the former is clearly able to activate the receptor on its
own, the latter is inactive on its own and yet causes a left-shift
in the concentration–response curve and increased maximal
response of serotonin (Figure 3b). A possible explanation of
the ’potentiator’ profile is that the compounds actually are
true allosteric potentiators having no intrinsic activity them-
selves. Alternatively, the observed enhancement may simply
be the cumulated response caused by serotonin and a low-po-
tency agonist, which may not be able to elicit a significant re-
sponse applied alone in the FMP assay. On the basis of pres-
ently available data, we are not able to reject either of these
two explanations.
Molecular basis for the 5-HT3R activity of the benzamide
analogue
The benzamide structures presented herein represent an inter-
esting 5-HT3R agonist scaffold, as the scaffold does not contain
a positive charge at physiological pH. A substantial number of
structural biology, mutagenesis, and medicinal chemistry stud-
ies of the interactions between 5-HT3Rs and other Cys-loop re-
ceptors and their orthosteric ligands have demonstrated the
essential role of a positive charge in the ligands for proper
binding to the receptors.[2a,e,9] The structural similarity between
the benzamide analogues in this study and competitive 5-
HT3Rs antagonists metoclopramide and zacopride on the one
hand could suggest that the analogues could target the
domain of the orthosteric site occupied by the benzamide
moiety in metoclopramide and zacopride. However, the ab-
sence of the positive charge in the benzamide analogue
makes it questionable whether it would be able to establish
sufficiently strong interactions with residues around the or-
thosteric site to establish binding to this site. This is supported
by the inability of the analogues 1h, 1k, 1p, 1t, 1ac, 3d, 3e,
5a, and 6b to displace [3H]GR65630 binding to the receptor
The superagonists
Interestingly, the 5-HT3R signalling elicited by a considerable
number of the agonists in the series was characterised by sub-
stantially higher maximal responses than that caused by sero-
tonin (Tables 2 and 4). The superagonism of these analogues
was observed in two different fluorescence-based assays using
two different 5-HT3A–HEK293 cell lines and a 5-HT3AB–HEK293
cell line, and these responses could be eliminated completely
with 5-HT3R-specific antagonists. Hence, we believe that this
observation is a true reflection of the functional properties of
the compounds at the receptors rather than an assay-related
artefact. However, as the observed responses in these assays
represent indirect measures of receptor signalling, it will be im-
ChemMedChem 2011, 6, 725 – 736
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
733

C. G. Jørgensen, A. A. Jensen, et al.
MED
and purified by column chromatography (hexane/ethyl acetate in
the majority of cases).
(Figure 6). These analogues are amongst the most potent 5-
HT3R ligands in the series as determined by the functional
assays, and they have been tested in the binding assay in con-
centrations well above those needed to elicit their functional
response. Although we cannot completely exclude the possi-
bility that the lack of detectable binding affinity could be at-
tributed to specific assay conditions, this observation suggests
that the benzamide analogue may target a site other than that
targeted by [3H]GR65630 in the 5-HT3R, or at least that the
binding of the two ligands to the receptor is not mutually ex-
clusive. Conversely, the zacopride analogue 7d appears to
target the orthosteric site as it, similar to metoclopramide (7b)
and zacopride (7e), competes for binding with [3H]GR65630.
Thus, it seems that the fusion of 1a with the azabicyclo-
[2.2.2]oct-3-yl ring system has converted it into an orthosteric
ligand, and that the ’pure’ benzamide analogue mediates its
agonism through binding to an allosteric site at the receptor.
Although several allosteric modulators of Cys-loop receptors
have been reported, allosteric ligands with intrinsic activity as
agonists are very rare. Thus, it is important to stress that future
investigations are needed to fully elucidate the complex mode
of action of the benzamide analogues at the 5-HT3R.
Procedure P2: Synthesis of amides deviating on the alkyl moiety. A
standard solution of activated acid was prepared from 4-ethoxy-
benzoic acid (1.99 g, 12 mmol) in THF (12 mL) and CDI (1.94 g,
12 mmol). Appropriate amounts (~1 mmol) of activated acid was
taken out and transferred to the vessel containing amine
(~0.67 mmol, but weight was accurately measured). The following
reaction conditions and purification were performed according to
procedure P1.
4-Ethoxy-N-(pentan-3-yl)benzamide (1a): Preparation according
to procedure P1 using acid (172 mg), amine (61 mg) and CDI
(168 mg). Column chromatography (Combiflash, heptane/EtOAc)
gave a white solid (139 mg, yield of 84%). Preparation according
to procedure P2 using an activated acid solution (~1.12 mL corre-
sponding to 1 mmol) and amine (56 mg). Combiflash (heptane/
EtOAc) gave a white solid (132 mg, yield of 87%). Rf =0.34 (hep-
tane/EtOAc 2:1); 1H NMR (Bruker, CDCl3): d=0.95 (t, J=7 Hz, 2ꢂ
CH3, 6H), 1.43 (t, J=7 Hz, CH3, 3H), 1.50–1.70 (m, 2ꢂCH2, 4H),
3.97–4.02 (m, CH, 1H), 4.08 (q, J=7 Hz, OCH2, 2H), 5.68 (brd, NH,
1H), 6.92 (d, J=9 Hz, 2ꢂCH, 2H), 7.72 ppm (d, J=9 Hz, 2ꢂCH,
2H); 13C NMR (Gemini, CDCl3): d=10.6, 15.0, 27.8, 52.5, 63.8, 114.2,
127.2, 128.6, 161.4, 166.8 ppm; GCMS: 65, 93, 121, 149, 165, 206,
235 [C14H21NO2+]; HRMS: found: 236.1645, calcd: 236.1645
[C14H21NO2 +H+]+; Anal. found: C 71.33, H 9.13, N 5.91, calcd for
C14H21NO2: C 71.46, H 8.99, N 5.95%.
Conclusions
In conclusion, the benzamide analogues in this study consti-
tute a novel class of 5-HT3R agonists. The benzamide scaffold
is unique from other 5-HT3R agonists (and orthosteric ligands
in general), in that it does not contain a positively charged
functionality at physiological pH, and preliminary investiga-
tions suggest that it may exert its effects via an allosteric site
at the receptor complex. In ongoing projects in our lab we are
studying the functional properties of these ligands and the
molecular basis for their activity at the receptor in greater
detail.
Pharmacology
Materials. Culture media, serum, antibiotics, and buffers for cell cul-
ture and Fluo-4/AM were obtained from Invitrogen (Paisley, UK).
Serotonin and probenecid were purchased from Sigma (St. Louis,
MO, USA), and zacopride was obtained from Tocris Cookson (Bris-
tol, UK). The FLIPRꢁ Membrane Potential (FMP) Blue assay dye and
the Fluo-4/AM dye were purchased from Molecular Devices (Craw-
ley, UK) and Molecular Probes (Eugene, OR, USA), respectively.
[3H]GR65630 and Opti-FluorTM scintillation fluid were obtained from
PerkinElmer (Boston, MA, USA). The compound library consisting of
3040 different compounds was purchased from Chembridge Cor-
poration (San Diego, CA, USA). The stable h5-HT3A–HEK293 cell
line used for the FMP assay and in the [3H]GR65630 binding experi-
ments was a kind gift from Dr. J. Egebjerg (H. Lundbeck A/S, Valby,
Denmark), and the stable h5-HT3A– and h5-HT3AB–HEK293 cell
lines used in the studies in the Ca2+/Fluo-4 assay were kind gifts
from C. Rojas (Johns Hopkins University School of Medicine, Balti-
more, MD, USA).[10] The ma4b2-HEK293T, ra3b4–HEK293 and ha7–
GH3 cell lines were kind gifts from Dr. J. A. Stitzel (University of Col-
orado, Boulder, CO, USA), Drs. Y. Xiao and K. Kellar (Georgetown
University School of Medicine, Washington DC, USA) and Dr. D.
Feuerbach (Novartis, Basel, Switzerland), respectively.[11]
Experimental Section
Chemistry
Please refer to Supporting Information for experimental details of
all isolated compounds of which the purity (>95%) was estab-
lished through NMR, GCMS, and HRMS, in addition to elemental
analysis of 1/10 of the compounds (1a, 1u, 1ab, 2c, 2d, 4c, 4e,
5g, 7a, 7b, 7c and 7d).
General procedures for amide synthesis
Procedure P1: Synthesis of amides deviating on the phenyl moiety.
N,N’-Carbonyldiimidazole (CDI) (~1 mmol) was added to a suspen-
sion of benzoic acid (~1 mmol) in THF (1–2 mL) in 7 mL glass ves-
sels. The mixture was left shaking using an IKA heating/shaker
block for 5–10 min whereupon gas evolution subsided and the ac-
tivated acid became soluble in the majority of reactions. The
amine (~0.67 mmol but weight was accurately measured) was
added. The reaction vessel was sealed and left shaking overnight
at 608C. After cooling the reaction mixture, THF was evaporated
using a Genevac concentrator, followed by immobilisation on SiO2,
Cell culture. The ma4b2–HEK293T, ra3b4–HEK293 and ha7–GH3
and h5–HT3A-HEK293 cells (from Dr. Egebjerg) were grown in Dul-
becco’s Modified Eagle Medium supplemented with 100 UmLÀ1
penicillin, 100 mgmLÀ1 streptomycin, 10% fetal bovine serum and
with 0.1 mgmLÀ1 zeocin and 0.5 mgmLÀ1 hygromycin (ma4b2),
0.1 mgmLÀ1 G-418 (ha7) or 1 mgmLÀ1 G-418 (ra3b4 and h5-HT3A).
The h5-HT3A– and h5-HT3AB–HEK293 cell lines (from Dr. Rojas)
were grown in RPMI 1640 containing 100 UmLÀ1 penicillin,
100 mgmLÀ1 streptomycin and 10% fetal bovine serum and supple-
734
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 725 – 736

Benzamide Analogues as 5-HT3 Agonists
mented with 0.5 mgmLÀ1 G-418 (h5-HT3A) or 0.5 mgmLÀ1 G-418
were harvested at 80–90% confluency and scraped into the assay
buffer [50 mm Tris-HCl (pH 7.2)], homogenised using a polytron for
10 sec, and centrifuged for 20 min at 50000 g. The resulting pellet
were homogenised in assay buffer (30 mL) and centrifuged again.
Then the cell pellet were resuspended in the assay buffer, and the
membranes were incubated with 50 pm [3H]GR65630 and various
concentrations of the test compounds. The total reaction volume
was 800 mL, and nonspecific binding was determined in reactions
with 10 mm tropisetron. The assay mixtures were incubated for
2.5 h at room temperature while shaking. GF/C filters were pre-
soaked for 1 h in a 0.2% polyethyleneimine solution, and binding
was terminated by filtration through these filters using a 48-well
cell harvester and washing with ice-cold isotonic NaCl solution (3ꢂ
4 mL). Following this, the filters were dried, Opti-FluorTM (3 mL)
were added, and the amount of bound radioactivity was deter-
mined in a scintillation counter. The fraction of specifically bound
radioligand was <10% of the total amount of radioligand. The
binding experiments were performed in duplicate three times for
each compound.
and 3 mgmLÀ1 blasticidin (h5-HT3AB).
The FLIPRꢀ Membrane Potential Blue assay. The screening of the
compound library at the h5-HT3A–HEK293 cell line and the subse-
quent functional characterisation of the compounds at the ra3b4–,
ma4b2– and h5-HT3A–HEK293 cell lines were performed in the
FLIPRꢁ Membrane Potential (FMP) Blue assay essentially as previ-
ously described.[12] Cells were split into poly-d-lysine-coated black
96-wells plates with clear bottom (BD Biosciences, Bedford, MA,
USA); 16–24 h later the medium was aspirated, and the cells were
washed with Krebs buffer (100 mL) [140 mm NaCl, 4.7 mm KCl,
2.5 mm CaCl2, 1.2 mm MgCl2, 11 mm HEPES, 10 mm d-glucose,
pH 7.4]. Krebs buffer (50 mL) was added to the wells (in the antago-
nist experiments, various concentrations of the compound were
dissolved in the buffer) and then additional Krebs buffer (50 mL)
supplemented with 1 mgmLÀ1 FMP Blue assay dye was added to
each well. The plate was incubated at 378C in a humidified 5%
CO2 incubator for 30 min and assayed in a NOVOstarTM plate reader
(BMG Labtechnologies, Offenburg, Germany) measuring emission
[in fluorescence units (FU)] at 560 nm caused by excitation at
530 nm before and up to 1 min after addition of agonist solution
(33 mL). Unless specified otherwise in the figures, serotonin was
used as reference agonist for h5-HT3A, whereas epibatidine was
used for the two nAChRs. In the screening of the compound library
at h5-HT3A, the compounds were co-applied with EC15 concentra-
tions of serotonin onto the cells. Experiments were performed in
duplicate at least three times for each compound tested. For the
testing of the compounds as potentiators, they were co-applied to
the cells with EC10–EC20 concentrations of the agonist for the re-
spective receptors. For the testing of the compounds as antago-
nists, an EC70–EC95 concentration of the agonist was used.
Data analysis. Concentration–response curves for agonists and ’po-
tentiators’ and concentration–inhibition curves for antagonists in
the functional assays were constructed based in the difference in
the fluorescence units (DFU) between the maximal fluorescence re-
cording made before and after addition of agonist obtained for dif-
ferent concentrations of the respective ligands. These curves and
the concentration–inhibition curves were generated by nonweight-
ed least-squares fits using the program KaleidaGraph 3.6 (Synergy
Software).
Acknowledgements
The Ca2+/Fluo-4 assay. The functional characterisation of com-
pounds at the ha7-GH3, h5-HT3A- and h5-HT3AB cell lines were
performed in the Ca2+/Fluo-4 assay essentially as previously de-
scribed.[12] Cell lines were split into poly-d-lysine-coated black 96-
well plates with clear bottom. Following 16–24 h incubation in the
case of the 5-HT3R cell lines or a 64–72 h incubation of the ha7-
GH3 cells, the culture medium was aspirated and the cells were in-
cubated in assay buffer (50 mL) [Hank’s Buffered Saline Solution
containing 20 mm HEPES, 1 mm CaCl2, 1 mm MgCl2 and 2.5 mm
probenecid, pH 7.4] supplemented with 6 mm Fluo-4/AM at 378C
for 1 h. Then the buffer was aspirated, the cells were washed once
with assay buffer (100 mL), and then assay buffer (100 mL) was
added to the cells. In the experiments with the h5-HT3A and h5-
HT3AB cell lines this final assay buffer was supplemented with
10 mm CaCl2, and the assay buffer used for the ha7-GH3 cells con-
tained 100 mm genistein. The 96-well plate was assayed in a NO-
VOstarTM microplate reader measuring emission [in fluorescence
units (FU)] at 520 nm caused by excitation at 485 nm before and
up to 60 sec after addition of agonist solution (33 mL) in assay
buffer. Serotonin was used as agonist for the h5-HT3A and h5-
HT3AB receptors, whereas ACh was used for the a7 nAChR. The ex-
periments were performed in duplicate at least three times for
each compound tested. For the testing of the compounds as po-
tentiators, they were co-applied to the cells with EC10-EC20 concen-
trations of the agonist for the respective receptors. For the testing
of the compounds as antagonists, an EC70–EC95 concentration of
the agonist was used.
The authors thank the Lundbeck Foundation (C.G.J. and A.A.J.),
the Carlsberg Foundation (C.G.J.), the Danish Medical Research
Council (A.A.J.), the Aase og Ejner Danielsens Foundation (A.A.J.),
and the Danish Chemical Society (C.G.J., travelling grants) for fi-
nancial support. Drs. Egebjerg, Rojas, Stitzel, Xiao, Kellar, and
Feuerbach are thanked for their generous gifts of the various cell
lines.
Keywords: agonists · allosterism · benzamides · ion channels ·
serotonin type 3 receptors
[1] a) D. E. Nichols, C. D. Nichols, Chem. Rev. 2008, 108, 1614–1641; b) M.
Berger, J. A. Gray, B. L. Roth, Annu. Rev. Med. 2009, 60, 355–366; c) J.
Walstab, G. Rappold, B. Niesler, Pharmacol. Ther. 2010, 128, 146–169.
[2] a) N. M. Barnes, T. G. Hales, S. C. R. Lummis, J. A. Peters, Neuropharmacol-
ogy 2009, 56, 273–284; b) A. J. Thompson, S. C. R. Lummis, Expert Opin.
Ther. Targets 2007, 11, 527–540; c) A. A. Jensen, P. A. Davies, H. Brꢃuner-
Osborne, K. Krzywkowski, Trends Pharmacol. Sci. 2008, 29, 437–444;
d) W. Sieghart, Adv. Pharmacol. 2006, 54, 231–263; e) A. Taly, P. J. Cor-
ringer, D. Guedin, P. Lestage, J. P. Changeux, Nat. Rev. Drug Discovery
2009, 8, 733–750; f) J. W. Lynch, Physiol. Rev. 2004, 84, 1051–1095.
[3] D. Bertrand, M. Gopalakrishnan, Biochem. Pharmacol. 2007, 74, 1155–
1163.
[4] G. A. Patani, E. J. LaVoie, Chem. Rev. 1996, 96, 3147–3176.
[5] a) A. E. Dubin, R. Huvar, M. R. D’Andrea, J. Pyati, J. Y. Zhu, K. C. Joy, S. J.
Wilson, J. E. Galindo, C. A. Glass, L. Luo, M. R. Jackson, T. W. Lovenberg,
M. G. Erlander, J. Biol. Chem. 1999, 274, 30799–30810; b) K. L. Price, S. C.
Lummis, J. Neurosci. Methods 2005, 149, 172–177; c) K. Krzywkowski,
A. A. Jensen, C. N. Connolly, H. Brauner-Osborne, Pharmacogenet. Ge-
nomics 2007, 17, 255–266.
[3H]GR65630 binding. Competition binding studies of selected com-
pounds using [3H]GR65630 and the stable h5-HT3A–HEK293 cell
line were performed essentially as described previously.[6] The cells
ChemMedChem 2011, 6, 725 – 736
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
735

C. G. Jørgensen, A. A. Jensen, et al.
MED
[6] A. A. Jensen, D. P. Zlotos, T. Liljefors, J. Med. Chem. 2007, 50, 4616–
4629.
2004, 41, 907–917; c) C. D. Dellisanti, Y. Yao, J. C. Stroud, Z. Z. Wang, L.
Chen, Nat. Neurosci. 2007, 10, 953–962.
[10] C. Rojas, M. Stathis, A. G. Thomas, E. B. Massuda, J. Alt, J. Zhang, E. Ru-
benstein, S. Sebastiani, S. Cantoreggi, S. H. Snyder, B. Slusher, Anesth.
Analg. 2008, 107, 469–478.
[11] a) M. S. Karadsheh, M. S. Shah, X. Tang, R. L. Macdonald, J. A. Stitzel, J.
Neurochem. 2004, 91, 1138–1150; b) Y. Xiao, E. L. Meyer, J. M. Thomp-
son, A. Surin, J. Wroblewski, K. J. Kellar, Mol. Pharmacol. 1998, 54, 322–
333; c) D. Feuerbach, K. Lingenhohl, P. Dobbins, J. Mosbacher, N. Cor-
bett, J. Nozulak, D. Hoyer, Neuropharmacology 2005, 48, 215–227.
[12] A. A. Jensen, M. L. Bergmann, T. Sander, T. Balle, J. Biol. Chem. 2010, 285,
10141–10153.
[7] N. A. Sharif, E. H. Wong, D. N. Loury, E. Stefanich, A. D. Michel, R. M.
Eglen, R. L. Whiting, Br. J. Pharmacol. 1991, 102, 919–925.
[8] a) J. E. Macor, D. Gurley, T. Lanthorn, J. Loch, R. A. Mack, G. Mullen, O.
Tran, N. Wright, J. C. Gordon, Bioorg. Med. Chem. Lett. 2001, 11, 319–
321; b) F. G. Boess, J. De Vry, C. Erb, T. Flessner, M. Hendrix, J. Luithle, C.
Methfessel, B. Riedl, K. Schnizler, F. J. van der Staay, M. van Kampen,
W. B. Wiese, G. Koenig, J. Pharmacol. Exp. Ther. 2007, 321, 716–725;
c) D. A. Gurley, T. H. Lanthorn, Neurosci. Lett. 1998, 247, 107–110;
d) L. M. Broad, C. Felthouse, R. Zwart, G. I. McPhie, K. H. Pearson, P. J.
Craig, L. Wallace, R. J. Broadmore, J. R. Boot, M. Keenan, S. R. Baker, E.
Sher, Eur. J. Pharmacol. 2002, 452, 137–144.
[9] a) D. L. Beene, G. S. Brandt, W. Zhong, N. Zacharias, H. A. Lester, D. A.
Dougherty, Biochemistry 2002, 41, 10262–10269; b) P. H. Celie, S. E. van
Rossum-Fikkert, W. J. van Dijk, K. Brejc, A. B. Smit, T. K. Sixma, Neuron
Received: October 12, 2010
Revised: November 24, 2010
Published online on January 18, 2011
736
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 725 – 736