New telmisartan-derived PPARγ agonists: Impact of the 3D-binding mode on the pharmacological profile

Article abstract of DOI:10.1016/j.ejmech.2016.08.027

In previous studies, the 4′-((2-propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,1′-biphenyl]-2-carboxylic acid was identified as pharmacophoric core for PPARγ activation. In this structure-activity relationship study the C2-alkyl chain was elongated and the 2-COOH group was changed to a carbamide/carbonitrile or shifted to the 3- or 4-position. Furthermore, the benzo[d]imidazole was exchanged by 2,3-dihydrobenzo[d]thiazole or 1H-indole. C2-propyl derivatives showed the profile of partial agonists, while elongation of the C2-chain to that of an n-heptyl group or a 4-COOH shift changed the pharmacological profile to that of a potent full agonist. This finding can be explained by binding to the LBD in different ligand conformations. Two anchoring points (Tyr473 and Arg288) exist in the LBD, which have to be contacted to achieve receptor activation. In a crystal violet chemosensitivity assay using COS-7?cells and LNCaP cells expressing PPARγ only the carbamide derivatives influenced the cell growth, independently on the presence of the PPARγ. Therefore, receptor mediated cytotoxicity can be excluded.

Full text of DOI:10.1016/j.ejmech.2016.08.027

European Journal of Medicinal Chemistry 124 (2016) 138e152
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
New telmisartan-derived PPARg agonists: Impact of the 3D-binding
mode on the pharmacological profile
Victoria Obermoser a, Margarethe E. Urban b, Manuela S. Murgueitio c, Gerhard Wolber c,
,
*
Ulrich Kintscher d, Ronald Gust a
a
€
Pharmaceutical Chemistry, Institute of Pharmacy, Universitat Innsbruck, Innrain 80-82, 6020, Innsbruck, Austria
b
c
€
€
Pharmaceutical Chemistry, Institute of Pharmacy, Freie Universitat Berlin, Konigin-Luise Str. 2þ4, 14195, Berlin, Germany
€
€
Computer-Aided Drug Design, Institute of Pharmacy, Freie Universitat Berlin, Konigin-Luise Str. 2þ4, 14195, Berlin, Germany
d
ꢀ
€
Institute of Pharmacology, Center for Cardiovascular Research, Charite Universitatsmedizin Berlin, Hessische Str. 3-4, 10115, Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
In previous studies, the 40-((2-propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-2-carboxylic
acid was identified as pharmacophoric core for PPARg activation. In this structure-activity relationship
Article history:
Received 23 June 2016
Received in revised form
12 August 2016
Accepted 13 August 2016
Available online 15 August 2016
study the C2-alkyl chain was elongated and the 2-COOH group was changed to a carbamide/carbonitrile
or shifted to the 3- or 4-position. Furthermore, the benzo[d]imidazole was exchanged by 2,3-
dihydrobenzo[d]thiazole or 1H-indole. C2-propyl derivatives showed the profile of partial agonists,
while elongation of the C2-chain to that of an n-heptyl group or a 4-COOH shift changed the pharma-
cological profile to that of a potent full agonist. This finding can be explained by binding to the LBD in
different ligand conformations. Two anchoring points (Tyr473 and Arg288) exist in the LBD, which have
to be contacted to achieve receptor activation. In a crystal violet chemosensitivity assay using COS-7 cells
Keywords:
PPAR
g
Transactivation
Cytotoxicity
and LNCaP cells expressing PPAR
g only the carbamide derivatives influenced the cell growth, indepen-
dently on the presence of the PPAR
g
. Therefore, receptor mediated cytotoxicity can be excluded.
Binding mode
Molecular modeling
© 2016 Elsevier Masson SAS. All rights reserved.
1. Introduction
elements (PPRE) in enhancer sites of target genes and regulates
gene transcription [3].
The peroxisome proliferator-activated receptor
g
(PPAR
g) be-
PPARg is one of the main regulators of adipogenesis, glucose and
longs to the hormone receptor superfamily of ligand-activated
transcription factors whose activity is regulated by direct binding
of hormones (steroids and iodothyronines), lipid metabolites, vi-
tamins and various xenobiotics [1].
lipid homeostasis and its activation has beneficial effects on serum
glucose levels [4].
Synthetic high-affinity ligands for PPARg are e.g. the thiazoli-
dinediones (TZDs), which act as so-called insulin-sensitizers and
were broadly used in the therapy of type II diabetes mellitus and
the metabolic syndrome. However, severe side effects like weight
gain, body fluid retention, congestive heart failure and loss of bone
Activation of PPARg leads to a conformational change, followed
by a heterodimerization with retinoic X receptors (RXRs), dissoci-
ation of corepressors and recruitment of coactivators [2]. Subse-
quently, the heterodimer binds to sequence-specific DNA response
mineral density go along with full activation of PPARg [5], which led
to withdrawal from the market or limited usage of all initially
approved TZDs [6e8].
Aside from these well-known and well-investigated metabolic
actions triggered by PPARg activation, target genes of PPARg are
also involved in cellular differentiation, development of various
organs, and carcinogenesis [9]. It was suggested that activation of
PPARg induces apoptosis in several malignant cell lines [10] and
promotes terminal differentiation of tumor cells [11]. In vitro
Abbreviations: Amax, maximum intrinsic activity; BPO, benzoyl peroxide;
DMEM, Dulbecco's modified Eagle medium; LBD, ligand binding domaine; mp,
melting point; NBS, N-bromosuccinimide; PPAR, peroxisome proliferator-activated
receptor; PPRE, peroxisome proliferator-activated receptor response element; RT,
room temperature; SAR, structure-activity relationship; SD, standard deviation; TG,
triglyceride; TZD, thiazolidinedione.
studies already demonstrated antiproliferative and proapoptotic
actions in prostate cancer for PPARg agonists out of the TZD class
* Corresponding author.
E-mail address: ronald.gust@uibk.ac.at (R. Gust).
http://dx.doi.org/10.1016/j.ejmech.2016.08.027
0223-5234/© 2016 Elsevier Masson SAS. All rights reserved.

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
139
[12]. Additionally, inhibition of tumorigenesis upon activation of
PPAR was also shown in vivo in many animal cancer models [13].
These findings are of high interest, because human prostate
cancer tissue expresses PPAR at high levels, while healthy human
prostate tissue expresses PPAR only at low levels [14,15]. This
different expression of PPAR can probably be used to develop
(2-COOH), 1-D (3-COOH), and 1-E (4-COOH), the respective
COOMe-substituted 4-(bromomethyl)-1,10-biphenyl (3a-3c) was
reacted with the 2-propyl-1H-benzo[d]imidazole 5a and NaH in
DMF giving the 1-([1,10-biphenyl]-4-ylmethyl)-2-propyl-1H-benzo
[d]imidazoles 4a-4c, which were subsequently ester cleaved under
basic conditions giving 1-A, 1-D, and 1-E (Scheme 1).
g
g
g
g
agonists suitable for the tumor-specific treatment of the prostate
cancer. For this purpose, however, the side effects not related to
tumor therapy have to be minimized.
The synthons 3a-3c are available in a palladium-catalyzed
Suzuki-Miyaura coupling of p-tolylboronic acid with either 2-, 3-,
or 4-bromobenzoic acid followed by esterification with methanol/
H2SO4 and halogenation of the methyl group with N-bromosucci-
nimide (NBS) and benzoylperoxid (BPO) in a Wohl-Ziegler reaction
(Scheme 1) [28e30].
In recent years, the concept of disconnecting the potency of
agonism of PPAR
selective modulation or partial activation of the PPAR
g
and efficacy has been developed in terms of
as some side
g
effects might be related to the classic agonistic activity of TZDs.
Full agonists replace helix H12 in a conformation promoting the
interaction with coactivators [16], while partial agonists activate
The C2-substituted benzo[d]imidazoles 5a-5d were obtained by
a modified Phillips-Ladenburg benzimidazole synthesis [31,32], in
which the ortho-phenylenediamine was reacted with the respec-
tive acylchloride to the N,N0-1,2-phenylenebisalkanoylamids, which
then built the benzo[d]imidazoles under elimination of a carboxylic
acid. In the next step, 5a-5d, respectively, were N1-alkylated with
40-(bromomethyl)-[1,10-biphenyl]-2-carbonitrile and NaH in DMF
(Scheme 2).
the PPARg via a mechanism that is independent from helix H12
[17]. In addition, it is conceivable not to seek for compounds with
higher potency and efficacy but to look for compounds with a dif-
ferential mechanism of action, e.g. alterations in the phosphoryla-
tion pattern of PPARg [18].
Telmisartan, an angiotensin type 1 (AT1) receptor antagonist
commonly used in antihypertensive therapy, was reported to act as
The
40-((2-alkyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carbamides 1-B, 2-B, 2-D, and 2-F and the respective
carboxylic acids 1-A, 2-A, 2-C, and 2-E were obtained in successive
base catalyzed hydrolyses starting from the 40-((2-alkyl-1H-benzo
[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-2-carbonitriles 1-C, 6b-
6d (Scheme 2). The carbonitrile and the carbamides represent very
stable educts, why the yields of the respective products are very
low, although very drastic conditions were used (KOH (30%),
methanol, reflux, 14d).
a partial agonist of PPARg in vitro. It demonstrated to have a specific
coregulator recruitment profile compared to the TZDs, which
translated to a different gene expression pattern [19e21]. In animal
studies, telmisartan proved to have a positive impact on plasma
glucose levels and insulin sensitivity by simultaneously preventing
from TZD-alike side effects [20,21]. In further studies, telmisartan
was found to have growth inhibitory effects on various cancer cell
lines of the urogenital tract via a PPARg-mediated pathway [22,23].
In the first step of the synthesis of the 2-propyl-2,3-
This interesting pharmacological profile induced us to select
telmisartan as lead structure. In structure-activity relationship
(SAR) studies we elucidated the 40-((2-propyl-1H-benzo[d]imida-
zol-1-yl)methyl)-[1,10-biphenyl]-2-carboxylic acid (1-A) as essen-
dihydrobenzo[d]thiazole
7
(Scheme 3), butyraldehyde was
condensed with 2-aminobenzenethiol to a Schiff base, which was
not isolated. After addition of methanol and stirring for 0.5 h at RT,
cyclization takes place and the product 7 can be isolated in high
yields. The N-alkylation was performed with 40-(bromomethyl)-
[1,10-biphenyl]-2-carbonitrile as described above.
tial scaffold to meet the minimum requirements for PPAR
g
activation with the same potency as telmisartan [24].
Substituents at position 5 were already introduced at the benzo
[d]imidazole and shifted to position 6, showing that both, type and
position of the substituent, have a strong influence on the com-
As C2 represents a stereo center in 8 and N1 becomes asym-
metric due to alkylation, a mixture of two diastereomeric forms in a
ratio of 5.5:4.5 were obtained (determined by 1H-NMR spectros-
copy). Fractional crystallization allowed only an enrichment of one
diastereomer (7:3) but no separation. Nevertheless, the signals in
the NMR-spectra could be assigned to both compounds (see
experimental part).
The diastereomeric mixture was then hydrolyzed to the
carbamide 3-B and carboxylic acid 3-A. No further attempts were
made to separate the isomers and the mixtures were submitted to
the in vitro testing.
pound's PPARg activity [25,26].
The C2-substituent influences potency and efficacy [27]. In this
study, we elucidate if the successive elongation of the C2-alkyl
residue in 1-A up to an n-heptyl chain leads to a change of the
pharmacological profile. A further aspect of this SAR study is the
significance of the COOH position at the biphenyl moiety in 1-A and
the consequence of an exchange by a carbonitrile or carboxamide
group. Of interest is also the relevance of the benzo[d]imidazole
core in 1-A, which we exchanged by a 2,3-dihydrobenzo[d]thiazole
or 1H-indole.
For the synthesis of the 1H-indole derivatives, an intramolecular
Wittig reaction was chosen, which includes three consecutive re-
actions [33,34]. 2-Aminobenzyl alcohol and triphenylphosphine
hydrobromide were dissolved in acetonitrile and heated to reflux.
The resulting 2-aminobenzyltriphenylphosphonium bromide 9
was isolated and reacted with butyryl chloride to the carbamide 10.
2-Butyramidobenzyltriphenylphosphonium bromide 10 was then
treated with potassium-tert-butylat in toluol. The resulting ylide
reacted with the carbonyl C-atom to a phosphor betaine and after
elimination of triphenylphosphine oxide, the 2-propyl-1H-indole
11 was build (Scheme 4). The subsequent alkylation with 40-(bro-
momethyl)-[1,10-biphenyl]-2-carbonitrile (/ 3-E) and hydrolysis
to the carbamide 3-D and carboxylic acid 3-C was the same as
described above.
The novel compounds were characterized in a PPAR
tiation assay using 3T3-L1 cells and in a PPAR transactivation assay
using transiently transfected COS-7 cells, both established models
to assess intracellular PPAR activation. Further, we performed a
crystal violet chemosensitivity assay in PPAR -negative COS-7 and
LNCaP prostate cancer cells expressing PPAR to evaluate a possible
cancer cell line specific PPAR -mediated cytotoxic effect. The
g differen-
g
g
g
g
g
binding mode of the new compounds has been analyzed in 3D
pharmacophore-driven docking experiments.
2. Results and discussion
2.1. Synthesis
All compounds were characterized by 1H-NMR spectroscopy
and mass spectrometry and final products, which are evaluated
in vitro for biological activity, additionally by elemental analysis.
For the synthesis of the 40-((2-propyl-1H-benzo[d]imidazol-1-
yl)methyl)-[1,10-biphenyl]-2/3/4-carboxylic acid derivatives 1-A

140
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
Scheme 1. Synthesis of 40-((2-propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-2/3/4-carboxylic acid derivatives. i) KOH, Pd(OAc)2, H2O, RT; ii) H2SO4, methanol,
iii) NBS, BPO, CCl4, T, 4 h; iv) NaH, DMF abs., RT; v) NaOH, methanol, T, 12 h.
DT, 5 h;
D
D
Scheme 2. Synthesis of 1-([1,10-biphenyl]- 4-ylmethyl)- 2-alkyl-1H-benzo[d]imidazole derivatives. i)
DT, 3 h; ii) NaH, DMF abs., RT; iii) KOH (30%) methanol, DT, 14d.
2.2. In vitro SAR studies
involved in the LBD binding via H-bridges. Therefore, we studied
the influence of an exchange of the carboxyl group by an carbamide
or carbonitrile and a shift of the COOH group from position 2 to the
3/4-position on potency and pharmacological profile of the
respective compound.
As depicted in Fig. 1a and b telmisartan, 1-A (2-COOH) and 1-C
(2-CN) show identical concentration-activity curves representing
partial agonists with a maximum activation of about 50% (telmi-
All compounds were characterized in terms of PPARg activation
in a cell-based dual luciferase transactivation assay (Fig. 1).
The full PPAR agonist pioglitazone was used as internal refer-
ence and its maximum PPAR activation was set as 100%. Table 1
g
g
lists the EC50 values and the intrinsic activity (percentage of pio-
glitazone's activation) at the highest concentration used.
One point of interest is the 2-COOH group at the biphenyl
moiety, which can be exchanged by a bioisosteric tetrazole without
changing the activity [24]. It is well known that the carboxylate is
sartan: EC50 ¼ 4.71
Amax ¼ 49.6%; 1-C: EC50 ¼ 5.01
carbamide 1-B shows reduced partial agonism (Amax ¼ 21.1%) but
m
M, Amax ¼ 50.9%; 1-A: EC50 ¼ 4.17
mM,
m
M, Amax ¼ 51.3%). Interestingly, the

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
141
Scheme 3. Synthesis of 3-([1,10-biphenyl]-4-ylmethyl)-2-propyl-2,3-dihydrobenzo[d]thiazole derivatives. i) RT, 0.5 h; ii) methanol abs., RT, 0.5 h; iii) NaH, DMF, RT; iv) KOH (30%)
methanol, T, 14d.
D
Scheme 4. Synthesis of 1-([1,10-biphenyl]-4-ylmethyl)-2-propyl-1H-indole derivatives. i) acetonitrile,
RT; iv) KOH (30%) methanol, T, 14d.
DT 6 h; ii) CH2Cl2, DT, 3.5 h; iii) t-BuOK, toluol, DT, 15 min; iii) NaH, DMF abs.,
D
slightly increased potency (EC50 ¼ 3.15
m
M).
EC50 ¼ 0.21
with an elongated alkyl chain are more potent agonists than tel-
misartan (EC50 4.71 M) and comparable to pioglitazone
(EC50 ¼ 0.32 M). Parallely, the pharmacological profile changed
m
M) < 2-E (heptyl; EC50 ¼ 0.17
mM). All compounds
The shift of the carboxylic acid from the ortho- to the meta-
position does not change the pharmacological profile (1-D:
¼
m
EC50
(EC50 ¼ 1.88
¼
6.24
m
M, Amax
¼
44.7%). However, both potency
m
m
M) and maximum activity (Amax ¼ 84.1%) increased
from that of a partial agonist to that of a full agonist: 1-A
(Amax ¼ 49.6%) ¼ 2-A (Amax ¼ 55.4%) < 2-C (Amax ¼ 75.0%) < 2-E
(Amax ¼ 82.5%) (Fig. 1d-f).
by shifting the carboxylic acid to the para-position (Fig. 1c). Com-
pound 1-E represents now a very potent full agonist.
The alkyl chain at position 2 at the benzo[d]imidazole core in-
fluences both, the pharmacological profile and the potency. We
could already show that the potency reaches a maximum if the
benzo[d]imidazole bears a propyl or butyl chain. Furthermore, the
efficacy can be increased if a benzyl or a phenethyl moiety is
introduced at C2 [27]. In case of a 4-chlorobenzyl residue the
pharmacological profile changed from a partial agonist to that of a
full agonist. In continuation of this study, we elucidate if the suc-
cessive elongation of the C2-alkyl residue in 1-A up to an n-heptyl
chain leads to similar effects.
The influence of a longer alkyl chain at C2 on the activity of
carbamide derivatives is not as significant. 1-B (propyl;
EC50 ¼ 3.15
Amax ¼ 16.2%) and 2-D (hexyl; EC50 ¼ 3.06
nearly the same results in the transactivation assay. Only the heptyl
m
M, Amax ¼ 21.1%), 2-B (pentyl; EC50 ¼ 3.90
mM,
m
M, Amax ¼ 37.8%) show
derivative 2-F possesses a higher potency (EC50 ¼ 0.84
mM) with a
comparable maximum activation of Amax ¼ 32.8%.
In the third class of compounds (3-A to 3-E), the necessity of the
benzo[d]imidazole for biological activity of the derivatives was
studied. If the core in 1-A is exchanged by a 2,3-dihydrobenzo[d]
thiazole, the potency strongly increased: 3-A shows an
The potency strongly increases in the range: 1-A (propyl;
EC50 ¼ 4.17
m
M) < 2-A (pentyl; EC50 ¼ 0.40
mM) < 2-C (hexyl;
EC50 ¼ 1.05 M with an intrinsic activity of Amax ¼ 57.4%. In the case
m

142
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
Fig. 1. PPAR
g
activation measured in a luciferase transactivation assay. COS-7 cells were transiently transfected with the pGal4-hPPARgDEF and pGal5-Tk-pGL3 followed by
stimulation with pioglitazone ( ), or telmisartan ( ), or the synthesized compounds a) essential scaffold 1-A ( ) b) 1-B ( ), 1-C ( ); c) 1-D ( ), 1-E ( ); d) 2-A ( ), 2-B ( ); e) 2-C
;
;
;
;
(
), 2-D ( ); f) 2-E ( ), 2-F ( ); g) 3-A ( ), 3-B ( ); h) 3-C ( ), 3-D ( ), 3-E ( ) in a concentration-dependent manner. Firefly-luciferase was measured after 42 h and normalized
; ; ; ;
with the activity of the co-transfected pCMV-Renilla. Data points represent mean SD of ꢀ3 independent experiments with threefold determination.
of 3-B (carbamide derivative) only a marginal transactivation
M) and
its curve progression does not allow a statement about the phar-
As depicted in Fig. 1h, the data of the 1H-indole derivatives 3-C,
(Amax ¼ 33.1%) is determined at higher concentrations (20
m
3-D and 3-E give sigmoid concentration-activation curves. The
maximum intrinsic activation is in each case lower than 50%: 3-C
(Amax ¼ 40.0%) > 3-E (Amax ¼ 21.7%) > 3-D (Amax ¼ 11.0%). From
macological profile (Fig. 1g).

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
143
Table 1
compounds caused a distinctly reduced TG accumulation compared
to the COOH derivatives especially at 1 M. The latter (2-A, 2-C and
2-E), which proved to be partial or strong partial agonists in the
PPAR transactivation assay, induced differentiation at both tested
concentrations (1 M and 10 M) to approximately the same extent
as measured for the full agonist pioglitazone. It should be noted
that all derivatives are more potent than telmisartan.
The general trends were also determined in the third class of
compounds: 2-COOH (3-C) > 2-CONH2 (3-D) > 2-CN (3-E); 2-COOH
EC50 values and maximum instrinsic activity [Amax] determined in the luciferase
transactivation assay using the hPPAR
m
g
-LBD.
Compound
EC50
[m
M]a,c
Amax [%max]b,c
g
m
m
Pioglitazone
Telmisartan
1-A
1-B
1-C
1-D
1-E
2-A
2-B
2-C
2-D
2-E
2-F
3-A
3-B
3-C
3-D
0.32 0.08
4.71 0.26
4.17 0.28
3.15 0.26
5.01 0.22
6.24 0.17
1.88 0.10
0.40 0.03
3.90 0.07
0.21 0.02
3.06 0.09
0.17 0.02
0.84 0.06
1.05 0.07
n.d.
100
50.9 2.1
49.6 2.3
21.1 2.4
51.3 3.6
44.7 3.4
84.1 3.8
55.4 3.0
16.2 1.0
75.0 0.5
37.8 0.7
82.5 3.0
32.8 1.2
57.4 1.1
33.1 5.0
40.0 1.7
11.0 1.7
21.7 2.1
derivatives (3-A, 3-C) reached at 10 mM nearly the effects of pio-
glitazone (>3-fold TG accumulation). These results further imply
that the activity in the adipocyte differentiation assay is not
dependent on the used heterocyclic core. Furthermore, there is no
correlation with the transactivation assay.
As multiple studies established an association between full
PPARg agonism and adipocyte differentiation in vitro and weight
gain in vivo, compound 1-E is certainly outstanding in this SAR
study due to its non-adipogenic effect although it represents a
potent full agonist.
1.86 0.09
n.d.
8.03 0.25
3-E
During the last years an increasing number of studies were
a
EC50 values were calculated by means of concentration-PPAR
g
activation-curves
published showing the antitumor potency of PPAR
antagonists in vitro and in vivo. The cytotoxicities are discussed to
be PPAR -dependent or -independent. Therefore, we studied the
growth inhibitory effects of the class 1e3 agonists in the androgen-
dependent LNCaP prostate cancer cell line expressing PPAR [35] as
g agonists and
(Fig. 1).
b
The maximum activation [Amax] indicates the highest reached response trig-
gered by the compounds and is shown in relation to pioglitazone (100%).
g
c
Data values represent mean SD of three independent experiments.
g
well as in healthy COS-7 cells, which were used as transfection
vector in the luciferase transactivation assay.
The test concentration (50 mM) was chosen according to previ-
ously published studies with telmisartan in different cell lines of
the urogenital system [23].
The results depicted in Fig. 3 indicate a growth inhibitory effect
of about 55.8% in LNCaP cells for pioglitazone, while it was inactive
in COS-7 cells. On the other hand telmisartan did not influence the
growth of both cell lines.
Among the class 1 compounds a clear structure-activity rela-
tionship was observed. The cell growth was reduced in the series 1-
A (2-COOH: inactive) < 1-B (CONH2: 33.0%) < 1-C (2-CN: 61.9%) The
shift of the carboxylic acid did not influence the effects of the
compounds (1-D, 1-E) on the growth of both cell lines.
these curves the EC50 values can be calculated showing that 3-C
(EC50 ¼ 1.86
m
M) is a more potent partial agonist than 1-A. The
intrinsic activity and the potency of 3-E (EC50 ¼ 8.03
mM) is strongly
decreased compared to 1-A and from the curve of 3-D an EC50 value
cannot be calculated beyond doubt.
In the next step, we studied the relevance of the structural
modifications in an adipocyte differentiation assay. In 3T3-L1 cells,
the PPARg activation strongly correlates with the degree of differ-
entiation, which was analyzed by Oil-Red O staining. Telmisartan
and pioglitazone were used as positive controls and DMSO as
vehicle. Results are shown in Fig. 2. Pioglitazone as full agonist
induced a triglyceride accumulation (TG accumulation) up to 4-fold
compared to the vehicle. The partial agonist telmisartan was nearly
inactive at a concentration of 1
mM and showed an about 2-fold TG
Derivatives out of class 2 exerted effects on the cells dependent
on the length of the C2-alkyl chain and the 2-substituent. While all
2-COOH bearing compounds were inactive, 2-B, 2-D, and 2-F
showed pronounced antiproliferative effects with a slight prefer-
ence for COS-7 cells. 2-D was the most active derivative with a
reduced cell growth of 78.5% in LNCaP cells and 92.8% in COS-
7 cells. The dihydrobenzo[d]thiazole derivatives demonstrated a
comparable pharmacological profile (3-A inactive; 3-B inhibition of
accumulation at 10 M (Fig. 2). Compounds 1-A, 1-B, and 1-C with
COOH, CONH2 and CN substituents at 2-position were comparably
active, whereas the shift of the carboxylic acid from the ortho- (1-A
m
(10
mM): 3.2-fold) to the meta- (1-D (10
mM): 2.5-fold) and finally to
the para-position (1-E (10
mM): 1.2-fold) reduced strongly the TG-
accumulation.
In the second class of compounds, the 2-carbamide-substituted
Fig. 2. Compound screening in an adipocyte differentiation assay. 3T3-L1 cells were differentiated for 9 days with the indicated compounds at 1 mM ( ) and 10 mM (-). Telmisartan
(T) and pioglitazone (P) served as positive controls and DMSO (V) as vehicle. Cells were stained with Oil-Red O and the dye was subsequently extracted with 80% isopropanol.
Intracellular trigylceride accumulation was determined by measuring absorbance of extraction solution at 515 nm. Values were set in ratio to the vehicle value. Data points
represent mean SD of three independent experiments.

144
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
Fig. 3. Cytotoxicity screening of the respective compounds in an in vitro chemosensitivity assay. LNCaP ( ) and COS-7 cells (-) were treated with the compounds; telmisartan (T) or
pioglitazone (P) at a concentration of 50 M. Cell biomass was determinded after 72 h via crystal violet staining and subsequent measurement of absorbance at 595 nm. Cell biomass
m
is represented as % of cell biomass of vehicle (V) treated control.
79.0% in LNCaP cells and 69.2% in COS-7 cells). Interestingly, the
influence of 1H-indole derivatives against LNCaP cells was higher
than against COS-7 cells. While only 3-E reduced the growth of
COS-7 cells (by 36.8%) 3-C, 3-D and 3-E were cytotoxic on the LNCaP
cell line (growth reduction 27.4e54.1%).
propyl moiety fills an extension of the arm I, which is formed by
pushing out the side chain of His323 in comparison to its position in
other crystal structures [25]. Hydrophobic contacts are formed with
Tyr473, Ile472, Tyr327 and Ile326. Additionally,
p-stacking in-
teractions are formed by the distal benzo[d]imidazole ring and
Phe363 in arm I. The biphenyl moiety is located in arms II and III
and forms hydrophobic contacts with Met364, Ile341, Val339 and
Ile326. The carboxylate is located in direction to Arg288 but not
close enough to form H-bonds to residues of helix 2.
Based on these results it seems to be very unlikely that the
active compounds possess a PPAR
the 1H-indole derivatives and pioglitazone were more active on
PPAR expressing LNCaP cells than on COS-7 cells.
g-mediated mode of action. Only
g
Next, the synthesized compounds were docked into the same
crystal structure in order to identify a potential binding mode.
Binding poses were selected according to their similarity to the
telmisartan binding mode putting special emphasis on the hydro-
phobic contacts formed with the arm I extension resulting from the
movement of His323. This resulted in the predicted binding modes
discussed below.
2.3. Docking studies to predict PPARg binding
In order to investigate the binding mode of the new compound
series to PPAR , molecular modeling studies were performed. The
high-resolution structure of PPAR co-crystalized with telmisartan
g
g
and the coactivator peptide from steroid receptor coactivator-1
(PDB code 3VN2 [36]) was selected to investigate potential bind-
ing modes of the compounds by computational docking
experiments.
Binding conformations and rationalization of structure-activity
relationship derived from docking experiments
Compounds 1-A, 1-B, 1-C, and 1-D show an orientation of the
annelated ring system comparable to telmisartan. The propyl
moiety in 1-A is located in the extension of arm I created by the
His323 movement and forms hydrophobic contacts with Val322,
Ile472, Leu469, Val293 and Ile326 (Fig. 5A). The position of the
benzo[d]imidazole is stabilized by an H-bond formed by the N30-
nitrogen and the hydroxyl group of Tyr473. The biphenyl moiety is
located in arms II and III and in hydrophobic contacts with Leu330,
Ile341, Ile326 and Met364. The carboxylate in 1-A is bound to
Arg288 by an H-bridge and a negative charge interaction (Fig. 5A).
In the case of 1-B the carboxylate group is exchanged by a
First, we performed a 3D pharmacophore analysis to identify
key interactions formed by telmisartan (Fig. 4): the highly hydro-
phobic receptor binding site is characterized by a Y-shaped form
that can be subdivided into the arms I, II and III [37e39]. In the
aforementioned crystal structure, telmisartan adopts a U-shaped
conformation around Cys285 [36], which adjusts our previous hy-
pothesis on telmisartan binding [25]: the central benzo[d]imid-
azole ring occupies arm I, and its N30 nitrogen forms a hydrogen
bond with the hydroxyl group of Tyr473 that is part of helix 12,
representing the only observed H-bond formed by the ligand. The
Fig. 4. Binding mode of telmisartan to PPAR
depicted in grey. B) Telmisartan in the PPAR binding site. Key interactions are marked in color code: yellow spheres e hydrophobic contacts, blue ring e
bond acceptor. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
g
. A) PPAR
g
co-crystallized with telmisartan. The receptor is shown in blue, the ligand in stick mode. Its van der Waals surface is
g
p
-stacking, red arrow e H-

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
145
Fig. 5. Impact of the substitution of the benzo[d]imidazole through an 1H-indole. A) 3D view of the binding mode of compound 1-A. One H-bond between the benzo[d]imidazole
moiety and Tyr473 is depicted as a red arrow, the charge interaction with Arg288 as a red star. Hydrophobic contacts are shown as yellow spheres. B) Binding mode predicted for
compound 3-C is analogous to the one for 1-A. However, the stabilizing H-bond to Tyr473 is missing. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)
carbamide moiety, thus the negative charge interaction between
the ligand and Arg288 is missing. Furthermore, the charge distri-
bution in the carbamide seems to cause a repulsion to Arg288
which could be the reason for the decrease in intrinsic activity
(lowering of Amax). Compound 1-C bearing a 2-carbonitrile group
possesses the same pharmacological profile as 1-A. The stabiliza-
tion through the H-bond from Tyr473 to the N30-nitrogen of the
heterocycle seems, therefore, to be sufficient to stabilize the bind-
ing in the LBD and maintain the intrinsic activity.
Alternatively, agonists can also activate the receptor through
binding of the 2-carboxylate group to Arg288. The 1H-indole de-
rivative 3-C is located nearly identically to 1-A in the LBD (Fig. 5B)
without having contact to Tyr473. This missing link leads only to a
marginal reduction of the intrinsic activity (Amax: 49.6% / 40.0%).
However, if both key interactions are missing, e.g. by the exchange
of the carboxylate by a carbamide (3-D), the binding to Arg288 is
intrinsic activity (Amax ¼ 84.1%). The pharmacological profile is now
that of a full agonist.
Interestingly, the elongation of the C2-propyl chain in 1-A to
pentyl (2-A), hexyl (2-C), or heptyl (2-E) forces the benzo[d]imid-
azole in a binding mode comparable to 1-E and in opposite direc-
tion as in telmisartan or 1-A. The long alkyl chains do not fit the
cavity extension close to His323 (Fig. 7). The biphenyl moiety of the
compounds is oriented in arms II and III and the binding is stabi-
lized in the case of the carboxylated compounds by the charge
interaction to Arg288.
The intrinsic activity (49.6% for propyl (1-A), 55.4% for pentyl (2-
A), 75.0% for hexyl (2-C) and 82.5% for heptyl (2-E)) as well as the
potency (EC50 (1-A) ¼ 4.19
m
M; EC50 (2-A) ¼ 0.40
mM; EC50 (2-
M) of the 2-carboxylate contain-
C) ¼ 0.21 M; EC50 (2-E) ¼ 0.17
m
m
ing compounds increase with longer alkyl groups. 2-E is a very
potent full agonist. Its long heptyl chain reaches into arm II (Fig. 7B)
and forms multiple hydrophobic contacts with residues as Ile281,
Leu356 and Leu353 what further stabilizes the activated confor-
mation of the receptor.
impossible leading to
a nearly complete loss of activity
(Amax ¼ 11.0%).
Surprisingly, the shift of the 2-carboxylate to the 4-position (1-
A / 1-E) causes the ligand to flip and the propyl substituent is now
located in the binding site of arm I (Fig. 6A). In this case the benzene
part of the benzo[d]imidazole ring is embedded in the aperture
close to His323. The biphenyl moiety is slightly shifted and the
carboxylate forms two H-bonds to the backbone nitrogens of
Ser342 and Glu343, which presumably causes an increased stabi-
The lack of the Arg288 interaction in the compounds with a
carbamide substituent (2-B, 2-D and 2-F) causes a decrease in
intrinsic activity and potency, showing that it is essential for the
orientation and stabilization of these ligands and functions as an
anchoring point in the binding site.
The exchange of the benzo[d]imidazole core through a 2,3-
dihydrobenzo[d]thiazole ring causes a moderate increase in activ-
ity for 3-A compared to 1-A. In the predicted binding mode, the 2,3-
lization of helix 3 and the
orientation leads to an increase in potency (EC50 ¼ 1.88
b
-sheet region in our model. This
M) and
m
Fig. 6. Binding modes of compounds 1-E and 3-A that share the flipped ring orientation. A) Predicted binding mode for compound 1-E forming two H-bonds to residues Ser342 and
Glu343 (depicted as red arrows) that stabilize the binding mode. B) 3D view of the binding pose of compound 3-A. The ligand is anchored by the charge interaction with Arg288.
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

146
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
Fig. 7. Overview on the binding mode of the compounds with varying alkyl-chain length. A) Superimposed compounds 2-A to 2-F that show the flipped orientation of the benzo[d]
imidazole in the binding pocket B) Binding mode of compound 2-E. The long heptanyl chain forms multiple hydrophobic contacts with residues as Ile281, Leu356 and Leu353.
dihydrobenzo[d]thiazole compound is comparable oriented as
compounds 2-A to 2-F (Fig. 6B). The flexibility of the ring system
makes it possible for the benzene moiety of the heterocycle to be
deeply embedded into the cavity region close to His323, which
could explain the enhanced activity. Interestingly, also in this series
the exchange of the carboxylate by a carbamide (3-B) causes a loss
of activity and potency showing again that the charge interaction
between the carboxylate and Arg288 is necessary to stabilize this
binding pose.
with Ser342 and Glu343 further anchoring points for full agonists
were identified.
4. Experimental section
4.1. General material and methods
All reagents and solvents were purchased from Acros Organics,
Sigma-Aldrich, Alfa Aesar or Merck. All reactions were monitored
by TLC, performed on silica gel plates 60 F254 (Merck, Darmstadt/
Germany). Visualization on TLC was achieved by UV light. Column
chromatography was performed with Merck silica gel 60H, grain
size <0.063 mm, 230 mesh ASTM (Darmstadt/Germany). Melting
points were determined on a B 545 Büchi (Flawil/Schweiz) capillary
melting point apparatus. 1H-NMR spectra were obtained with an
Avance DPX-400 spectrometer (Bruker, Karlsruhe/Germany) at
400 MHz using TMS as an internal standard. 13C-NMR spectra were
measured with a Gemini 2000 NMR spectrometer (Varian, Palo
Alto/USA) at 50 MHz using TMS as an internal standard. Elemental
analyses were carried out at the Microlaboratory of the Freie Uni-
3. Conclusions
Extending our studies on PPARg agonists, new derivatives of the
lead structure 1-A have been synthesized with the following
modifications: elongation of the alkyl chain at position 2 of the
benzo[d]imidazole, exchange or shift of the carboxylic acid residue
at the biphenyl moiety and exchange of the central benzo[d]
imidazole.
In pharmacological in vitro studies most of the new compounds
demonstrated potent PPAR
and 8.03 M) with distinct pharmacological profiles ranging from
weak partial agonists to full agonists. Interestingly, data from the
PPAR activation assay correlate only partially with those from the
g activation (EC50 values between 0.17
m
€
versitat Berlin with a Vario EL elemental analyzer (Elementar,
Hanau/Germany). Mass spectra were obtained with an Orbitrap
Elite ESI mass spectrometer (Thermo Fisher, Waltham/USA). The
analytical results are within 0.4% of theoretical values, indicating
that the purity of all tested compounds was higher than 95%. EIMS
spectra were obtained with a CH-7A-Varian MAT, 70 eV (Mel-
bourne/Australia). Absorbance and luminescence in biological as-
says were measured with an Enspire Multimode Plate Reader
(Perkin-Elmer, Life Sciences, Brunn/Austria).
g
functional preadipocyte differentiation assay. For example, the full
agonist 1-E induced TG accumulation only to a very low extent. Also
in the cytotoxicity assay using PPAR
no thorough correlation with the results from the PPAR
activation assay was observed indicating that occurring cytotoxic
effect are likely to be independent from PPAR
In 3D pharmacophore-driven docking experiments, we
demonstrated that in the LBD of the PPAR two anchoring points
g
positive prostate cancer cells
g
trans-
g
.
g
(Tyr473 and Arg288) through which an agonist can activate the
receptor exists. Either the N30-nitrogen of the 1-([1,10-biphenyl]-4-
ylmethyl)-1H-benzo[d]imidazole core (H-bridge with the Tyr473)
or an H-bridge and a negative charge interaction from the 2-
carboxylate to Arg288 must be present to achieve the pharmaco-
logical profile of a partial agonist (binding mode 1).
If the carboxylate is shifted to the 4-position or the C2-alkyl
chain is elongated, the orientation of the benzo[d]imidazole is
turned by 180ꢁ and the compounds have the profile of full agonists
(binding mode 2). It is very likely, that the 1-([1,10-biphenyl]-4-
ylmethyl)-1H-benzo[d]imidazole derivatives can bind to the LBD
in both conformations. However, the longer the C2-chain the
higher is the proportion of binding mode 2 which goes along with
lower EC50 values and higher maximum activation. Furthermore,
4.2. Chemistry
4.2.1. General method for the preparation of 40-((2-propyl-1H-
benzo[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-carboxylic acids
To a stirred solution of the respective methyl 40-((2-propyl-1H-
benzo[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-carboxylate (0.10 g,
0.26 mmol) in 10 ml of abs. methanol, 0.21 ml of NaOH (10%)
(0.52 mmol) were added and heated to reflux for 12 h. After cooling
to RT, the mixture was acidified with HCl and the product was
extracted with methylene chloride. After drying over Na2SO4, the
solvent was removed in vacuo and the crude product was purified
by column chromatography on silica gel with methylene chloride/
methanol (9:1) and recrystallization from diethyl ether and
methanol.

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
147
4.2.1.1. 40-((2-Propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carboxylic acid (1-A).
Colorless crystals; yield 0.09 g (0.24 mmol, 92%); mp 247 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d
12.97 (s, 1H, COOH), 7.99 (d,
3J ¼ 8.4 Hz, 2H, H2 þ H6), 7.76 (d, 3J ¼ 8.4 Hz, 2H, H3 þ H5), 7.70 (d,
3J ¼ 8.3 Hz, 2H, H20 þ H60), 7.6 (m, 1H, H400), 7.47 (m, 1H, H700), 7.20
(d, 3J ¼ 8.3 Hz, 2H, H30 þ H50), 7.16 (m, 2H, H500 þ H600), 5.56 (s, 2H,
CH2 800), 2.84 (t, 3J ¼ 7.4 Hz, 2H, CH2 1000), 1.79 (tq, 3J ¼ 7.4 Hz,
3J ¼ 7.3 Hz, 2H, CH2 2000), 0.96 (t, 3J ¼ 7.3 Hz, 3H, CH3 3000). 13C-NMR
(200 MHz, DMSO-d6) d 167.65,155.60,144.23,142.95, 138.67,137.87,
135.84, 130.48, 130.23, 127.87, 127.68, 127.24, 122.27, 121.92, 119.06,
110.69, 46.16, 29.06, 20.89, 14.35. HR-MS: calcd for C24H22N2O2
[M þ H]þ: 371.1773; found: 371.1747. MS: (EI, 200 ꢁC) m/z 370.3 (Mþ
61%); Anal. calcd for C24H22N2O2: C 77.81, H 5.99, N 7.56; found: C
77.78, H 6.01, N 7.56.
4.2.1.4. 40-((2-Propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carbonitrile (1-C). The 1H-benzo[d]imidazole 5a
(2.50 mmol) was dissolved in 10 ml of abs. DMF. Sodium hydride
(60% dispersion in mineral oil) (0.2 g, 5 mmol) was added with
cooling. After the gas formation stopped, 40-bromomethyl-[1,10-
biphenyl]-2-carbonitrile dissolved in 2 ml of abs. DMF was added
dropwise and the solution was stirred for 1 h under cooling in an ice
bath and further 2 h at RT. Water was added and after the solution
was acidified with HCl, the product was extracted three times with
methylene chloride. The combined organic layers were dried over
Na2SO4 and evaporated to dryness in vacuo. The crude product was
purified by column chromatography on silica gel with methylene
chloride/methanol (95:5) and recrystallized from diethyl ether/
methanol.
Colorless crystals; yield 0.08 g (0.22 mmol, 85%); mp 153 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 12.72 (s, 1H, COOH), 7.70 (dd,
1H, H3), 7.59 (m,1H, H400), 7.54 (ddd, 1H, H5), 7.50 (m, 1H, H700), 7.43
(ddd, 1H, H4), 7.33 (dd, 1H, H6), 7.28 (d, 3J ¼ 8.1 Hz, 2H, H20 þ H60),
7.17 (m, 2H, H500 þ H600), 7.12 (d, 3J ¼ 8.1 Hz, 2H, H30 þ H50), 5.53 (s,
2H, CH2 800), 2.84 (t, 3J ¼ 7.4 Hz, 2H, CH2 1000), 1.79 (tq, 3J ¼ 7.4 Hz, 2H,
CH2 2000), 0.96 (t, 3J ¼ 7.4 Hz, 3H, CH3 3000). MS: (EI, 170 ꢁC) m/z 370.4
(M þ 100%); Anal. calcd for C24H22N2O2: C 77.81, H 5.99, N 7.56;
found: C 78.05, H 6.34, N 7.56.
4.2.1.2. 40-((2-Propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-3-carboxylic acid (1-D).
Colorless crystals; yield 0.09 g (0.24 mmol, 92%); mp 234 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 13.07 (s, 1H, COOH), 8.14 (dd,
Grey-colored wax-like paste; yield 0.62 g (1.76 mmol, 70%).
1H, H2), 7.89 (m, 2H, H4 þ H6), 7.66 (d, 3J ¼ 8.3 Hz, 2H, H20 þ H60),
1H-NMR (400 MHz/DMSO-d6):
d 7.93 (d, 1H, H3), 7.76 (ddd, 1H,
H400), 7.47 (m, 1H, H77.20 (d, 2H,
H5), 7.61 (m,1H, H400), 7.57 (m, 2H, H4 þ H6), 7.55 (d, 3J ¼ 8.3 Hz, 2H,
H20 þ H60), 7.50 (m,1H, H700), 7.23 (d, 3J ¼ 8.3 Hz, 2H, H30 þ H50), 7.17
(m, 2H, H500 þ H600), 5.59 (s, 2H, CH2 800), 2.85 (t, 3J ¼ 7.5 Hz, 2H, CH2
1000), 1.78 (tq, 3J ¼ 7.5 Hz, 3J ¼ 7.4 Hz, 2H, CH2 2000), 0.96 (t, 3J ¼ 7.4 Hz,
3H, CH3 3000). MS: (EI, 50 ꢁC) m/z 351.3 (Mþ 67%); Anal. calcd for
7.58 (m, 2H, H5
þ
3J ¼ 8.3 Hz,H30 þ H50), 7.16 (m, 2H, H500 þ H600), 5.55 (s, 2H, CH2 800),
3
2.84 (t, J ¼ 7.4 Hz, 2H, CH2 1000), 1.78 (tq, 3J ¼ 7.3 Hz, 2H, CH2 2000),
0.96 (t, 3J ¼ 7.3 Hz, 3H, CH3 3000). MS: (EI, 200 ꢁC) m/z 370.3
(M þ 71%); Anal. calcd for C24H22N2O2: C 77.81, H 5.99, N 7.56;
found: C 77.79, H 6.00, N 7.57.
C
24H21N3: C 82.02, H 6.02, N 11.96; found: C 82.10, H 6.20, N 11.76.
4.2.2. General method for the preparation of C2-substituted 40-
((1H-benzo[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-2-carbamides
and 40-((1H-benzo[d]imidazol-1-yl)methyl)-[1,10-biphenyl]-2-
carboxylic acids
4.2.1.3.. 40-((2-Propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-4-carboxylic acid (1-E).
The respective 40-((2-alkyl-1H-benzo[d]imidazol-1-yl)methyl)-
[1,10-biphenyl]-2-carbonitrile (1.00 mmol) was dissolved in 15 ml of
methanol and combined with 5.61 ml of KOH (30%) (30 mmol). The
mixture was heated to reflux for 14d. After cooling to RT, HCl was
added until the solution was acidified. Subsequently, the crude
product was isolated by extraction with methylene chloride. After
drying over Na2SO4, the solvent was removed in vacuo and the
products were separated by column chromatography on silica gel
with methylene chloride/methanol (95:5) and recrystallized from
diethyl ether/methanol.

148
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
4.2.2.1. 40-((2-Propyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carbamide (1-B).
H400), 7.50 (m, 1H, H700), 7.44 (ddd, 1H, H4), 7.41 (d, 1H, H3), 7.39
(ddd, 1H, H5), 7.37 (d, 3J ¼ 8.3 Hz, 2H, H20 þ H60), 7.31 (dd, 1H, H6),
7.25 (s, 1H, NH), 7.16 (m, 2H, H500 þ H600), 7.11 (d, 3J ¼ 8.3 Hz, 2H,
H30 þ H50), 5.52 (s, 2H, CH2 800), 2.85 (t, 2H, CH2 1000), 1.76 (tt, 2H, CH2
2000), 1.31 (m, 4H, CH2 3000 þ CH2 4000), 0.84 (t, 3H, CH3 5000). MS: (EI,
100 ꢁC) m/z 397.4 (Mþ 58%); Anal. calcd for C26H27N3O: C 78.56, H
6.85, N 10.57; found: C 78.57, H 6.61, N 10.32.
4.2.2.4. 40-((2-Hexyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carboxylic acid (2-C).
Colorless crystals; yield 0.04 g (0.11 mmol, 11%); mp 170 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 7.62 (s, 1H, NH), 7.59 (m, 1H,
H400), 7.49 (m, 1H, H700), 7.44 (ddd, 1H, H4), 7.40 (dd, 1H, H3) 7.39
3
(ddd, 1H, H5) 7.37 (d, J ¼ 8.3 Hz, 2H, H20 þ H60), 7.32 (d, 1H, H6),
7.24 (s, 1H, NH), 7.16 (m, 2H, H500 þ H600), 7.11 (d, 3J ¼ 8.3 Hz, 2H,
H30 þ H50), 5.52 (s, 2H, CH2 800), 2.84 (t, 3J ¼ 7.5 Hz, 2H, CH2 1000), 1.79
(tq, 3J ¼ 7.5 Hz, 3J ¼ 7.4 Hz, 2H, CH2 2000), 0.97 (t, 3J ¼ 7.4 Hz, 3H, CH3
3000). MS: (EI, 200 ꢁC) m/z 369.4 (Mþ 79%); Anal. calcd for C24H23N3O:
C 78.02, H 6.27, N 11.37; found C 78.22, H 6.26, N 11.39.
4.2.2.2. 40-((2-Pentyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carboxylic acid (2-A).
Beige-colored powder; yield 0.08 g (0.19 mmol, 21%); mp 194 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 12.75 (s, 1H, COOH), 7.70 (dd,
1H, H3), 7.59 (m, 1H, H400), 7.53 (ddd, 1H, H5), 7.50 (m, 1H, H700), 7.43
(ddd, 1H, H4), 7.31 (dd, 1H, H6), 7.28 (d, 3J ¼ 8.3 Hz, 2H, H20 þ H60),
7.16 (m, 2H, H500 þ H600), 7.12 (d, 3J ¼ 8.3 Hz, 2H, H30 þ H50), 5.53 (s,
2H, CH2 800), 2.85 (t, 3J ¼ 7.5 Hz, 2H, CH2 1000), 1.74 (tt, 3J ¼ 7.5 Hz, 2H,
CH2 2000), 1.35 (m, 2H, CH2 3000), 1.24 (m, 4H, CH2 4000 þ CH2 5000), 0.83
(t, 3H, CH3). 13C-NMR (200 MHz, DMSO-d6)
d 170.11, 155.76, 142.91,
140.91, 140.59, 136.61, 135.92, 132.91, 131.32, 130.93, 129.62, 129.21,
127.84, 126.79, 122.23, 121.86, 119.00, 110.69, 46.25, 31.55, 28.94,
27.44, 27.15, 22.53, 14.48. HR-MS: calcd for C27H28N2O2 [MþH]þ:
413.2273; found: 413.2219. MS: (EI, 150 ꢁC) m/z 412.2 (Mþ 55%);
Anal. calcd for C27H28N2O2: C 78.61, H 6.84, N 6.79; found: C 78.70,
H 7.02, N 6.91.
Beige-colored powder; yield 0.08 g (0.20 mmol, 22%); mp
193 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 12.73 (s, 1H, COOH), 7.70 (ddd,
4.2.2.5. 40-((2-Hexyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carbamide (2-D).
1H, H3), 7.59 (m, 1H, H400), 7.54 (ddd, 1H, H5), 7.51 (m,01H, H700), 7.43
(ddd,1H, H4), 7.32 (ddd,1H, H6), 7.28 (d, 2H, H20 þ H6 ), 7.17 (m, 2H,
H500 þ H600), 7.12 (d, 2H, H30 þ H50), 5.53 (s, 2H, CH2 800), 2.85 (t, 2H,
CH2 1000), 1.74 (tt, 2H, CH2 2000), 1.30 (m, 4H, CH2 3000 þ CH2 4000), 0.84 (t,
3H, CH3 5000). HR-MS: calcd for C26H26N2O2 [M þ H]þ: 399.2073;
found: 399.2059. MS: (EI, 200 ꢁC) m/z 398.4 (Mþ 54%); Anal. calcd
for C26H26N2O2: C 78.36, H 6.58, N 7.03; found: C 78.06, H 6.87, N
7.16.
4.2.2.3. 40-((2-Pentyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carbamide (2-B).
Beige-colored powder; yield 0.04 g (0.10 mmol, 11%); mp 154 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 7.63 (s, 1H, NH), 7.59 (m, 1H,
H400), 7.50 (m, 1H, H700), 7.44 (ddd, 1H, H4), 7.41 (d, 1H, H3), 7.39
(ddd, 1H, H5), 7.37 (d, 3J ¼ 8.3 Hz, 2H, H20 þ H60), 7.31 (dd, 1H, H6),
7.37 (s, 1H, NH), 7.16 (m, 2H, H500 þ H600), 7.11 (d, 3J ¼ 8.3 Hz, 2H,
H30 þ H50), 5.52 (s, 2H, CH2 800), 2.86 (t, 3J ¼ 7.4 Hz, 2H, CH2 1000), 1.76
(tt, 3J ¼ 7.4 Hz, 2H, CH2 2000), 1.35 (m, 2H, CH2 3000), 1.25 (m, 4H, CH2
4000 þ CH2 5000), 0.84 (t, 3H, CH3 6000). MS: (EI, 200 ꢁC) m/z 411.3 (Mþ
46%); Anal. calcd for C27H29N3O: C 78.80, H 7.10, N 10.21; found: C
78.97, H 7.35, N 10.18.
Beige-colored powder; yield 0.04 g (0.10 mmol, 11%); mp 152 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 7.63 (s, 1H, NH), 7.59 (m, 1H,

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
149
4.2.2.6. 40-((2-Heptyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carboxylic acid (2-E).
by column chromatography on silica gel with diethyl ether/meth-
anol (95:5) and recrystallized from diethyl ether/methanol.
4.2.3.1. 40-((2-Propyl-2,3-dihydro-1,3-benzo[d]thiazol-3-yl)-methyl)-
[1,10-biphenyl]-2-carboxylic acid (3-A).
Colorless crystals; yield 0.07 g (0.16 mmol, 19%); mp 172 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 12.75 (s, 1H, COOH), 7.66 (d,1H,
H3), 7.59 (m, 1H, H400), 7.50 (m, 2H, H5þH700), 7.41 (ddd, 1H, H4),
Brown solid; yield (diastereomer 1 þ 2): 0.06 g (0.15 mmol,16%);
3
mp 83 ꢁC.
7.30 (d, 1H, H6), 7.29 (d, J ¼ 8.1 Hz, 2H, H20 þ H60), 7.16 (m, 2H,
1H-NMR (400 MHz/DMSO-d6):
d (diastereomer 1) 12.79 (s, 1H,
H500 þ H600), 7.11 (d, 3J ¼ 8.1 Hz, 2H, H30 þ H50), 5.52 (s, 2H, CH2 800),
2.85 (t, 3J ¼ 7.5 Hz, 2H, CH2 1000), 1.73 (tt, 3J ¼ 7.5 Hz, 2H, CH2 2000), 1.29
(m, 8H, CH2 3000 þ CH2 4000 þ CH2 5000 þ CH2 6000), 0.83 (t, 3H, CH3
COOH), 7.68 (d, 1H, H3), 7.56 (ddd, 1H, H4), 7.44 (ddd, 1H, H5), 7.38
(d, 1H, H6), 7.35 (d, 3J ¼ 8.2 Hz, 2H, H20 þ H60), 7.28 (d, 3J ¼ 8.2 Hz,
2H, H30 þ H50), 6.90 (dd, 1H, H700), 6.86 (ddd, 1H, H500), 6.70 (dd, 1H,
H400), 6.49 (ddd, 1H, H600), 6.14 (s, broad, 1H, CH2 800), 4.11 (d, 1H, CH2
800), 3.53 (m, 1H, H200), 1.54e1.30 (m, 4H, CH2 1000 þ CH2 2000), 0.83 (t,
7000).13C NMR (200 MHz, DMSO-d6)
d 170.38, 155.74, 142.90, 140.67,
136.49, 135.90, 130.96, 130.81, 129.49, 129.20, 127.75, 126.74,122.22,
121.85, 118.99, 110.68, 46.25, 31.71, 29.23, 29.00, 27.48, 27.14, 22.61,
14.47. HR-MS: calcd for C28H30N2O2 [MþH]þ: 427.2373; found:
427.2372. MS: (EI, 225 ꢁC) m/z 426.3 (Mþ 47%); Anal. calcd for
3H, CH3 3000). 1H-NMR (400 MHz/DMSO-d6):
d (diastereomer 2)
12.79 (s, 1H, COOH), 7.68 (d, 1H, H3), 7.54 (ddd, 1H, H4), 7.43 (ddd,
1H, H5), 7.38 (d, 1H, H6), 7.28 (d, 2H, H20 þ H60), 7.23 (d, 3J ¼ 8.3 Hz,
2H, H30 þ H50), 6.90 (dd, 1H, H700), 6.86 (ddd, 1H, H500), 6.73 (dd, 1H,
H400), 6.54 (ddd, 1H, H600), 6.01 (s, broad, 1H, CH2 800), 4.45 (d, 1H,
CH2 800), 3.69 (m, 1H, H200), 1.54e1.30 (m, 4H, CH2 1000 þ CH2 2000), 0.83
(t, 3H, CH3 3000). HR-MS: calcd for C24H23NO2S [MþH]þ: 390.1473;
found: 390.1511. MS: (EI, 175 ꢁC) m/z 389.3 (Mþ 95%).
C
28H30N2O2 (x 0.5H2O): C 77.21, H 7.17, N 6.43; found: C 77.57, H
7.34, N 6.53.
4.2.2.7. 40-((2-Heptyl-1H-benzo[d]imidazol-1-yl)methyl)-[1,10-
biphenyl]-2-carbamide (2-F).
4.2.3.2. 40-((2-Propyl-2,3-dihydro-1,3-benzo[d]thiazol-3-yl)-
methyl)-[1,10-biphenyl]-2-carbamide (3-B).
Colorless crystals; yield 0.05 g (0.12 mmol, 14%); mp 153 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 7.62 (s, 1H, NH), 7.59 (m, 1H,
H400), 7.50 (m, 1H, H700), 7.44 (ddd, 1H, H4), 7.41 (d, 1H, H3), 7.40
(ddd, 1H, H5), 7.37 (d, 3J ¼ 8.3 Hz, 2H, H20 þ H60), 7.31 (d, 1H, H6),
7.23 (s, 1H, NH), 7.16 (m, 2H, H500 þ H600), 7.11 (d, 3J ¼ 8.3 Hz, 2H,
H30 þ H50), 5.51 (s, 2H, CH2 800), 2.85 (t, 3J ¼ 7.5 Hz, 2H, CH2 1000), 1.75
(tt, 3J ¼ 7.5 Hz, 2H, CH2 2000), 1.29 (m, 8H, CH2 3000 þ CH2 4000 þ CH2
5000 þ CH2 6000), 0.84 (t, 3H, CH3 7000). HR-MS: calcd for C28H31N3O
[MþH]þ: 426.2573; found: 426.2533. MS: (EI, 150 ꢁC) m/z 425.4
(Mþ 40%).
Colorless crystals; yield 0.03 g (0.08 mmol, 9%); mp 106 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d (diastereomer 1) 7.64 (s, 1H,
NH), 7.49e7.37 (m, 4H, H3þH4þH5þH6), 7.36 (m, 4H,
H20 þ H30 þ H50 þ H60), 7.29 (s, 1H, NH), 6.92 (dd, 1H, H700), 6.86
(ddd, 1H, H500), 6.70 (dd, 1H, H400), 6.49 (ddd, 1H, H600), 6.10 (d, 1H,
CH2 800), 4.10 (d, 1H, CH2 800), 3.53 (m, 1H, H200), 1.55e1.29 (m, 4H,
CH2 1000 þ CH2 2000), 0.84 (t, 3H, CH3 3000). 1H-NMR (400 MHz/DMSO-
d6):
d (diastereomer 2) 7.64 (s, 1H, NH), 7.49e7.37 (m, 4H,
H3þH4þH5þH6), 7.36 (m, 4H, H20 þ H30 þ H50 þ H60), 7.29 (s, 1H,
NH), 6.92 (dd, 1H, H700), 6.86 (ddd, 1H, H500), 6.73 (dd, 1H, H400), 6.54
(ddd, 1H, H600), 6.00 (s, broad, 1H, CH2 800), 4.43 (d, 1H, CH2 800), 3.68
(m, 1H, H200), 1.55e1.29 (m, 4H, CH2 1000 þ CH2 2000), 0.84 (t, 3H, CH3
3000). MS: (EI, 150 ꢁC) m/z 388.4 (Mþ 66%).
4.2.3. Hydrolysis reaction of 40-((2-propyl-2,3-dihydro-1,3-benzo
[d]thiazol-3-yl)methyl)-[1,10-biphenyl]-2-carbonitrile
40-[(2-Propyl-2,3-dihydro-1,3-benzo[d]thiazol-3-yl)methyl]-
[1,10-biphenyl]-2-carbonitrile 8 (0.35 g, 0.94 mmol) was dissolved
in 15 ml of methanol and combined with 5.30 ml of KOH (30%)
(28.3 mmol). The mixture was heated to reflux for 14d. After
cooling to RT, HCl was added until the solution was acidified.
Subsequently, the crude product was isolated by extraction with
methylene chloride. After drying over Na2SO4, the solvent was
removed in vacuo and the products (3-A and 3-B) were separated
4.2.4. 40-((2-Propyl-1H-indol-1-yl)methyl)-[1,10-biphenyl]-2-
carbonitrile (3-E)
2-Propyl-1H-indole 11 (0.20 g, 5.02 mmol) in 10 ml of abs. DMF
was carefully reacted with NaH (60% dispersion in mineral oil)
(0.18 g, 4.46 mmol) while cooling in an ice bath. After the gas

150
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
formation stopped, 40-bromomethyl-[1,10-biphenyl]-2-carbonitrile
(0.76 g, 2.76 mmol) dissolved in 2 ml of abs. DMF was added
dropwise and the solution was stirred for 1 h under cooling in an ice
bath and further 2 h at RT. Water was added and after the solution
was acidified with HCl, the product was extracted three times with
methylene chloride. The combined organic layers were dried over
Na2SO4 and evaporated to dryness in vacuo. The crude product was
purified by column chromatography on silica gel with ethyl acetate/
n-hexane (1:3).
370.1773; found: 370.1797. MS: (EI, 150 ꢁC) m/z 369.4 (Mþ 92%);
Anal. calcd for C25H23NO2 (x 0.5H2O): C 79.34, H 6.39, N 3.70;
found: C 79.04, H 6.36, N 3.92.
4.2.5.2. 40-((2-Propyl-1H-indol-1-yl)methyl)-[1,10-biphenyl]-2-
carbamid (3-D).
White solid, yield 0.04 g (0.11 mmol, 11%); mp 196 ꢁC.
1H-NMR (400 MHz/DMSO-d6):
d 7.62 (s, 1H, NH), 7.49 (dd, 1H,
H700), 7.44 (ddd, 1H, H4), 7.38 (d, 1H, H400), 7.38 (ddd, 1H, H5), 7.35
(d, 1H, H3), 7.31 (d, 1H, H6), 7.33 (d, 3J ¼ 8.2 Hz, 2H, H20 þ H60), 7.25
(s, 1H, NH), 7.03 (ddd, 1H, H500), 7.00 (ddd, 1H, H600), 6.96 (d,
3J ¼ 8.2 Hz, 2H, H30 þ H50), 6.32 (s, 1H, H300), 5.45 (s, 2H, CH2 800),
2.70 (t, 3J ¼ 7.6 Hz, 2H, CH2 1000), 1.70 (tq, 3J ¼ 7.6 Hz, 3J ¼ 7.4 Hz, 2H,
CH2 2000), 0.98 (t, 3J ¼ 7.4 Hz, 3H, CH3 3000). MS: (EI, 150 ꢁC) m/z 368.4
(Mþ 73%); Anal. calcd for C25H24N2O: C 81.49, H 6.57, N 7.60; found:
C 81.69, H 6.53, N 7.54.
Yellow oil; yield 0.50 g (1.43 mmol, 57%).
1H-NMR (400 MHz/DMSO-d6):
d 7.92 (dd, 1H, H3), 7.75 (ddd, 1H,
H5), 7.57 (d, 1H, H6), 7.55 (ddd, 1H, H4), 7.50 (m, 1H, H700), 7.50 (d,
4
2H, H20 þ H60), 7.38 (dd, J ¼ 0.7 Hz, 1H, H400), 7.09 (d, 2H,
H30 þ H50), 7.04 (ddd, 1H, H500), 6.99 (ddd, 1H, H600), 6.11 (d,
4J ¼ 0.7 Hz, 1H, H300), 5.51 (s, 2H, CH2 800), 2.7 (t, 3J ¼ 7.5 Hz, 2H, CH2
1000), 1.68 (tq, 3J ¼ 7.5 Hz, 3J ¼ 7.4 Hz, 2H, CH2 2000), 0.97 (t, 3J ¼ 7.4 Hz,
3H, CH3 3000). MS: (EI, 75 ꢁC) m/z 350.5 (Mþ 88%); Anal. calcd for
4.3. Biology
C
25H22N2: C 85.68, H 6.33, N 7.99; found: C 85.70, H 6.39, N 8.07.
4.3.1. General cell culture methods
4.2.5. Hydrolysis reaction of 40-((2-propyl-1H-indol-1-yl)methyl)-
[1,10-biphenyl]-2-carbonitrile
The mouse embryonic 3T3-L1 cell line, the monkey kidney-
derived COS-7 cell line and the human prostate adenoma cell line
LNCaP were obtained from the American Type Culture Collection
(ATCC). The 3T3-L1 and the COS-7 cells were maintained as
monolayer cultures in Dulbecco's modified Eagle medium (DMEM)
with phenol red, 4.5 g/l glucose, 584 mg/l L-glutamine (GE
Healthcare), supplemented with fetal calf serum (FCS, 10%; Bio-
chrom). LNCaP cells were cultivated as monolayer culture in poly-
D-lysine treated culture vessels in RPMI 1640 without phenol red
(GE Healthcare), supplemented with FCS (10%) and L-glutamine
(584 mg/l; GE Healthcare). The cells were cultivated in a humidified
atmosphere (5% CO2/95% air) at 37 ꢁC and passaged once (LNCaP) or
twice (COS-7, 3T3-L1) a week.
40-((2-Propyl-1H-indol-1-yl)methyl)-[1,10-biphenyl]-2-
carbonitrile (0.35 g, 1.00 mmol) 3-E was dissolved in 15 ml of
methanol and combined with 5.61 ml of KOH (30%) (30 mmol). The
mixture was heated to reflux for 24d. After cooling to RT, HCl was
added until the solution was acidified. Subsequently, the crude
product was isolated by extraction with methylene chloride. After
drying over Na2SO4, the solvent was removed in vacuo and the
products (3-C and 3-D) were separated by column chromatography
on silica gel with methylene chloride/methanol (95:5) and recrys-
tallized from diethyl ether/methanol.
4.2.5.1. 40-((2-Propyl-1H-indol-1-yl)methyl)-[1,10-biphenyl]-2-
carboxylic acid (3-C).
4.3.2. PPARg transactivation assay
Transient transfection (Lipofectamin 2000; Life Technologies)
and dual-luciferase reporter assay (Promega) were performed ac-
cording to the manufacturer's protocol. COS-7 cells were seeded in
96-well microtiter plates at a density of 104 cells/well in completed
DMEM and incubated at 37 ꢁC and 5% CO2 for 24 h prior to trans-
fection. After changing to serum- and antibiotic-free medium, cells
were transiently transfected with pGal5-TK-Luc (30 ng), pGal4-
PPARg-LBD (3 ng) and pRenilla-CMV (1 ng) in OptiMEM (25 ml;
Gibco). After 4 h, the selected compounds, pioglitazone or vehicle
(DMSO) were added at indicated concentrations, and luciferase
activity was measured after 42 h.
Brown oil, yield 0.06 g (0.16 mmol, 16%).
1H-NMR (400 MHz/DMSO-d6):
d
12.75 (s, 1H, COOH), 7.67 (dd,
4.3.3. 3T3-L1 differentiation assay
1H, H3), 7.52 (dd, 1H, H5), 7.49 (d, 1H, H700), 7.42 (ddd, 1H, H4), 7.37
3T3-L1 preadipocytes were differentiated by a modified proto-
col previously described [19]. 3T3-L1 cells were seeded in 6-well
plates and in a postconfluent state treated with completed DMEM
3
(d, 1H, H400), 7.32 (dd, 1H, H6), 7.24 (d, J ¼ 8.1 Hz, 2H, H20 þ H60),
7.03 (ddd, 1H, H500), 7.00 (m, 1H, H600), 6.98 (d, 3J ¼ 8.1 Hz, 2H,
H30 þ H50), 6.32 (d, 1H, H300), 5.46 (s, 2H, CH2 800), 2.69 (t, 3J ¼ 7.5 Hz,
containing insulin (0.17
days, medium was changed to completed DMEM containing insulin
(0.17 M) and incubated for further 3 days. Subsequently, cells were
mM) and dexamethasone (1 mM). After 3
3
2H, CH2 1000), 1.69 (tq, J ¼ 7.5 Hz, 3J ¼ 7.4 Hz, 2H, CH2 2000), 0.97 (t,
3J ¼ 7.4 Hz, 3H, CH3 3000). HR-MS: calcd for C25H23NO2 [MþH]þ:
m

V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
151
incubated with solely completed medium for 3 more days. During
important compounds described in this article.
the whole incubation period, the medium contained pioglitazone,
telmisartan, vehicle DMSO, or the compounds at two different
References
concentrations (1 mM and 10 mM), respectively. At day 9 of differ-
[1] M. Lehrke, M.A. Lazar, The many faces of PPARg, Cell 123 (2005) 993e999.
entiation, cells were washed with phosphate buffered saline (PBS)
and intracellular trigycerides were stained with Oil-Red O for 1 h.
Intracellularly bound dye was extracted with isopropanol 80% and
the absorbance of the extraction solution was measured at 515 nm.
[2] N.J. McKenna, B.W. O'Malley, Combinatorial control of gene expression by
nuclear receptors and coregulators, Cell 108 (2002) 465e474.
[3] N. Viswakarma, Y. Jia, L. Bai, A. Vluggens, J. Borensztajn, J. Xu, J.K. Reddy,
Coactivators in PPAR-regulated gene expression, PPAR Res. 2010 (2010).
Article ID 250126.
[4] P. Tontonoz, B.M. Spiegelman, Fat and beyond: The diverse biology of PPARg,
4.3.4. In vitro chemosensitivity assay
Annu. Rev. Biochem. 77 (2008) 289e312.
[5] T.A. Cock, S.M. Houten, J. Auwerx, Peroxisome proliferator-activated receptor
: too much of a good thing causes harm, EMBO Rep. 5 (2004) 142e147.
[6] E.J. Murphy, T.J. Davern, A.O. Shakil, L. Shick, U. Masharani, H. Chow, C. Freise,
W.M. Lee, N.M. Bass, Troglitazone-induced fulminant hepatic failure. Acute
liver failure study group, Dig. Dis. Sci. 3 (2000) 549e553.
[7] A.M. Gallagher, L. Smeeth, S. Seabroke, H.G.M. Leufkens, T.P. van Staa, Risk of
death and cardiovascular outcomes with thiazolidinediones: a study with the
general practice research database and secondary care data, PLoS One 6
(2011) e28157.
[8] J.D. Lewis, A. Ferrara, T. Peng, M. Hedderson, W.B. Bilker, C.P. Quesenberry,
D.J. Vaughn, L. Nessel, J. Selby, B.L. Strom, Risk of bladder cancer among dia-
betic patients treated with pioglitazone: interim report of a longitudinal
cohort study, Diabetes Care 34 (2011) 916e922.
The in vitro testing of the new compounds for their growth
inhibitory activity on prostate adenocarcinoma cells and healthy
COS-7 cells was carried out according to a modified protocol pre-
viously described [40,41]. LNCaP prostate adenocarcinoma cells
(5 ꢂ 104 cells/well) and COS-7 cells (3 ꢂ 104 cells/well) were seeded
in their exponential growth phase in complete DMEM in 96-well
microtiter plates, which were in case of LNCaP cells previously
coated with poly-D-lysine. Plates were kept at 37 ꢁC for 24 h prior to
addition of complete medium containing the appropriate concen-
g
tration (final concentration 50
mM) of pioglitazone, telmisartan,
[9] M. Ahmadian, J.M. Suh, N. Hah, C. Liddle, A.R. Atkins, M. Downes, R.M. Evans,
vehicle DMSO, or compounds, respectively. After an incubation
time of 72 h, medium was aspirated, cells were washed with PBS
and fixed with a solution of 1% (v/v) glutaric acid in PBS. Cell
biomass was determined via staining of the chromatin of adherent
cells with crystal violet and subsequent measurement of absor-
bance at 590 nm. Cell viability is expressed as percentage of cell
viability of vehicle-treated control which was set at 100%. Results
are the mean SD of at least two independent experiments. The
IC50 values were calculated with Prism 7 (GraphPad, San Diego, CA)
using nonlinear regression and the log of the inhibitor versus var-
iable slope response equation. The bottom constraint was set to 0%.
PPARg signaling and metabolism: the good, the bad and the future, Nat. Med.
99 (2013) 557e566.
[10] H.P. Koeffler, Peroxisome proliferator-activated receptor
Cancer. Res. 9 (2003) 1e9.
g and cancers, Clin.
[11] D. Shao, S.M. Rangwala, S.T. Bailey, S.L. Krakow, M.J. Reginato, M.A. Lazar,
Interdomain communication regulating ligand binding by PPARg, Nature 396
(1998) 377e380.
[12] G. Venkatachalam, A.P. Kumar, K.R. Sakharkar, S. Thangavel, M.-V. Clement,
M.K. Sakharkar, PPAR disease gene network and identification of therapeutic
targets for prostate cancer, J. Drug Target. 19 (2011) 781e796.
[13] S. Sikka, L. Chen, G. Sethi, A.P. Kumar, Targeting PPAR signaling cascade for
g
g
the prevention and treatment of prostate cancer, PPAR Res. 2012 (2012).
Article ID 968040.
[14] P.S. Nelson, N. Clegg, H. Arnold, C. Ferguson, M. Bonham, J. White, L. Hood,
B. Lin, The program of androgen-responsive genes in neoplastic prostate
epithelium, Proc. Natl. Acad. Sci. 99 (2002) 11890e11895.
4.4. Docking
[15] T. Kubota, K. Koshizuka, E.A. Williamson, H. Asou, J.W. Said, S. Holden,
I. Miyoshi, H. Phillip Koeffler, Ligand for peroxisome proliferator-activated
Docking studies with the synthesized PPARg modulators were
receptor
g (troglitazone) has potent antitumor effect against human pros-
performed using GOLD Suite v5.2.2 [42]. Default settings were
applied and the GoldScore [43] selected as scoring function. Before
docking the protein crystal structure of PPARg co-crystalized with
tate cancer both in vitro and in vivo, Cancer Res. 58 (1998) 3344e3352.
[16] D.J. Kojetin, T.P. Burris, Small molecule modulation of nuclear receptor
conformational dynamics: implications for function and drug discovery, Mol.
Pharmacol. 83 (2013) 1e8.
telmisartan (PDB code: 3VN2 [44]) was prepared using MOE
2014.09 [45]. All waters with exception of those located in the
binding site (HOH645, HOH625, HOH613 and HOH632) were
removed. Next, hydrogens were added using the protonate 3D
application with standard parameters. All compounds were docked
[17] J.B. Bruning, M.J. Chalmers, S. Prasad, S.A. Busby, T.M. Kamenecka, Y. He,
K.W. Nettles, P.R. Griffin, Partial agonists activate PPAR
g using a helix 12 in-
dependent mechanism, Structure 15 (2007) 1258e1271.
[18] J.H. Choi, A.S. Banks, J.L. Estall, S. Kajimura, P. Bostrom, D. Laznik, J.L. Ruas,
M.J. Chalmers, T.M. Kamenecka, M. Bluher, P.R. Griffin, B.M. Spiegelman, Anti-
diabetic drugs inhibit obesity-linked phosphorylation of PPAR
g by Cdk5,
Nature 466 (2010) 451e456.
into the PPARg binding site and ten poses per molecule retrieved.
[19] M. Schupp, J. Janke, R. Clasen, T. Unger, U. Kintscher, Angiotensin type 1 re-
All ligand poses were minimized in the binding site using the
MMFF94 force field. The resulting poses were then further
inspected using LigandScout 3.12 and selected using a telmisartan-
derived 3D pharmacophore as customized scoring function
[46e48].
ceptor blockers induce peroxisome proliferatoreactivated receptor
Circulation 109 (2004) 2054e2057.
[20] M. Schupp, M. Clemenz, R. Gineste, H. Witt, J. Janke, S. Helleboid, N. Hennuyer,
P. Ruiz, T. Unger, B. Staels, U. Kintscher, Molecular characterization of new
g activity,
selective peroxisome proliferatoreactivated receptor
g modulators with
angiotensin receptor blocking activity, Diabetes 54 (2005) 3442e3452.
[21] S.C. Benson, H.A. Pershadsingh, C.I. Ho, A. Chittiboyina, P. Desai, M. Pravenec,
N. Qi, J. Wang, M.A. Avery, T.W. Kurtz, Identification of telmisartan as a unique
Acknowledgements
angiotensin II receptor antagonist with selective PPAR
g
emodulating activity,
Hypertension 43 (2004) 993e1002.
[22] M. Matsuyama, R. Yoshimura, Peroxisome proliferator-activated receptor
g
is a
We thank Alexandra Knox (Department of Pharmaceutical
Chemistry, University of Innsbruck) for recording the 13C-NMR
spectra and Monika Cziferszky (Department of Pharmaceutical
Chemistry, University of Innsbruck) for carrying out the high res-
olution mass analyses. This work was supported by Research Grant
potent target for prevention and treatment in human prostate and testicular
cancer, PPAR Res. 2008 (2008).
[23] H. Ishiguro, Y. Ishiguro, Y. Kubota, H. Uemura, Regulation of prostate cancer
cell growth and PSA expression by angiotensin II receptor blocker with
peroxisome proliferator-activated receptor
g ligand like action, Prostate 67
(2007) 924e932.
Gu
285/7-2
and
Ki
712/3-2
from
the
Deutsche
[24] M. Goebel, M. Clemenz, B. Staels, T. Unger, U. Kintscher, R. Gust, Character-
ization of new PPAR agonists: analysis of telmisartan's structural compo-
g
Forschungsgemeinschaft.
nents, ChemMedChem 4 (2009) 445e456.
[25] M. Goebel, G. Wolber, P. Markt, B. Staels, T. Unger, U. Kintscher, R. Gust,
Appendix A. Supplementary data
Characterization of new PPAR
g agonists: benzimidazole derivatives d
importance of positions 5 and 6, and computational studies on the binding
mode, Bioorg. Med. Chem. 18 (2010) 5885e5895.
[26] L. Herbst, M. Goebel, S. Bandholtz, R. Gust, U. Kintscher, Characterization of
Supplementary data associated with this article can be found in
the online version, at http://dx.doi.org/10.1016/j.ejmech.2016.08.
027. These data include MOL files and InChiKeys of the most
telmisartan-derived PPAR
g agonists: importance of moiety shift from position
6 to 5 on potency, efficacy and cofactor recruitment, ChemMedChem 7 (2012)

152
V. Obermoser et al. / European Journal of Medicinal Chemistry 124 (2016) 138e152
ꢀ
1935e1942.
[27] M. Goebel, B. Staels, T. Unger, U. Kintscher, R. Gust, Characterization of new
PPAR agonists: benzimidazole derivatives e the importance of position 2,
[38] L. Guasch, E. Sala, A. Castell-Auví, L. Cedo, K.R. Liedl, G. Wolber,
ꢀ
M. Muehlbacher, M. Mulero, M. Pinent, A. Ardevol, C. Valls, G. Pujadas,
ꢀ
g
S. Garcia-Vallve, Identification of PPAR
g
partial agonists of natural origin (I):
ChemMedChem 4 (2009) 1136e1142.
development of a virtual screening procedure and in vitro validation, PLoS
One 7 (2012) e50816.
[39] R.B. El-Houri, J. Mortier, M.S. Murgueitio, G. Wolber, L.P. Christensen, Identi-
[28] D.N. Korolev, N.A. Bumagin, An improved protocol for ligandless Suzu-
kieMiyaura coupling in water, Tetrahedron Lett. 47 (2006) 4225e4229.
[29] C. Djerassi, Brominations with N-bromosuccinimide and related compounds.
The Wohl-Ziegler reaction, Chem. Rev. 43 (1948) 271e317.
[30] K.C. Santhosh, G.C. Paul, E. De Clercq, C. Pannecouque, M. Witvrouw,
T.L. Loftus, J.A. Turpin, R.W. Buckheit, M. Cushman, Correlation of anti-HIV
activity with anion spacing in a series of cosalane analogues with extended
polycarboxylate pharmacophores, J. Med. Chem. 44 (2001) 703e714.
[31] M.A. Phillips, The formation of 2-substituted benzimidazoles, J. Chem. Soc. 13
(1928) 2393e2399.
[32] D.V. Ramana, E. Kantharaj, Mass spectrometer as a probe in the synthesis of 2-
substituted benzimidazoles, Tetrahedron Lett. 50 (1994) 2485e2496.
[33] L. Capuano, A. Ahlhelm, H. Hartmann, New syntheses of 2-acylbenzofurans, 2-
acylindoles, 2-indolylcarboxylates and 2-quinolones by intramolecular Wittig
reaction, Chem. Ber. 119 (1986) 2069e2074.
fication of PPARg agonists from natural sources using different in silico ap-
proaches, Planta Med. 81 (2015) 488e494.
€
€
[40] H. Reile, H. Birnbock, G. Bernhardt, T. Spruss, H. Schonenberger, Computerized
determination of growth kinetic curves and doubling times from cells in
microculture, Anal. Biochem. 187 (1990) 262e267.
€
€
[41] G. Bernhardt, H. Reile, H. Birnbock, T. Spruss, H. Schonenberger, Standardized
kinetic microassay to quantify differential chemosensitivity on the basis of
proliferative activity, J. Cancer Res. Clin. Oncol. 118 (1992) 35e43.
[42] GOLD Suite v5.2.2, in, Cambridge Crystallographic Data Centre, Cambridge,
UK, 2013.
[43] M.L. Verdonk, J.C. Cole, M.J. Hartshorn, C.W. Murray, R.D. Taylor, Improved
protein-ligand docking using GOLD, Proteins 52 (2003) 609e623.
[44] Y. Amano, T. Yamaguchi, K. Ohno, T. Niimi, M. Orita, H. Sakashita, M. Takeuchi,
[34] S. Taira, H. Danjo, T. Imamoto, A facile synthesis of 2-substituted indoles from
(2-aminobenzyl)triphenylphosphonium salts, Tetrahedron Lett. 43 (2002)
2885e2888.
Structural basis for telmisartan-mediated partial activation of PPARg,
Hypertens. Res. 35 (2012) 715e719.
[45] Molecular Operating Environment (MOE), 2014.09, in, Chemical Computing
Group Inc., 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A
2R7, 2015.
[46] T. Seidel, G. Ibis, F. Bendix, G. Wolber, Strategies for 3D pharmacophore-based
virtual screening, Drug Discov. Today Technol. 7 (2010) e221ee228.
[47] G. Wolber, A.A. Dornhofer, T. Langer, Efficient overlay of small organic mol-
ecules using 3D pharmacophores, J. Comp. Aided Mol. Des. 20 (2006)
773e788.
[35] J.O. Nwankwo, M.E.C. Robbins, Peroxisome proliferator-activated receptor
g
expression in human malignant and normal brain, breast and prostate-
derived cells, Prostagl. Leukot. Essent. Fat. Acids 64 (2001) 241e245.
[36] Y. Amano, T. Yamaguchi, K. Ohno, T. Niimi, M. Orita, H. Sakashita, M. Takeuchi,
Structural basis for telmisartan-mediated partial activation of PPARg,
Hypertens. Res. 35 (2012) 715e719.
[37] P. Markt, R.K. Petersen, E.N. Flindt, K. Kristiansen, J. Kirchmair, G. Spitzer,
S. Distinto, D. Schuster, G. Wolber, C. Laggner, T. Langer, Discovery of novel
[48] G. Wolber, T. Langer, LigandScout: 3-d pharmacophores derived from protein-
bound ligands and their use as virtual screening filters, J. Chem. Inf. Model. 45
(2005) 160e169.
PPAR ligands by
a virtual screening approach based on pharmacophore
modeling, 3D shape, and electrostatic similarity screening, J. Med. Chem. 51
(2008) 6303e6317.