Optimization of a novel class of benzimidazole-based farnesoid X receptor (FXR) agonists to improve physicochemical and ADME properties

Article abstract of DOI:10.1016/j.bmcl.2010.12.123

Structure-guided lead optimization of recently described benzimidazolyl acetamides addressed the key liabilities of the previous lead compound 1. These efforts culminated in the discovery of 4-{(S)-2-[2-(4-chloro-phenyl)-5,6- difluoro-benzoimidazol-1-yl]-2-cyclohexyl-acetylamino}-3-fluoro-benzoic acid 7g, a highly potent and selective FXR agonist with excellent physicochemical and ADME properties and potent lipid lowering activity after oral administration to LDL receptor deficient mice.

Full text of DOI:10.1016/j.bmcl.2010.12.123

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1134–1140
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Optimization of a novel class of benzimidazole-based farnesoid X receptor
(FXR) agonists to improve physicochemical and ADME properties
Hans G. F. Richter a, , G. M. Benson a, K. H. Bleicher a, D. Blum a, E. Chaput a, N. Clemann a, S. Feng b,
⇑
C. Gardes a, U. Grether a, P. Hartman a, B. Kuhn a, R. E. Martin a, J.-M. Plancher a, M. G. Rudolph a,
F. Schuler a, S. Taylor a
a F. Hoffmann-La Roche Ltd, Pharmaceutical Research, Grenzacherstrasse, CH-4070 Basel, Switzerland
b Roche R&D Center (China) Ltd, 720 Cai Lun Road, Building 5, Pudong, Shanghai 201203, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Structure-guided lead optimization of recently described benzimidazolyl acetamides addressed the key
liabilities of the previous lead compound 1. These efforts culminated in the discovery of 4-{(S)-2-[2-(4-
chloro-phenyl)-5,6-difluoro-benzoimidazol-1-yl]-2-cyclohexyl-acetylamino}-3-fluoro-benzoic acid 7g, a
highly potent and selective FXR agonist with excellent physicochemical and ADME properties and potent
lipid lowering activity after oral administration to LDL receptor deficient mice.
Received 9 November 2010
Revised 22 December 2010
Accepted 23 December 2010
Available online 31 December 2010
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Farnesoid X receptor agonist
FXR
Nuclear receptor
Lead optimization
ADME
X-ray structure
The farnesoid X receptor (FXR, NR1H4) is a ligand-activated
transcription factor and member of the nuclear hormone receptor
superfamily. Cloned in 19951 and de-orphanized in 1999,2 FXR
has been identified as a key sensor for bile acids, with chenode-
oxycholic acid (CDCA, Fig. 1) being the most effective endogenous
activator.
Consistent with its function in regulating bile acid synthesis and
profile, FXR is mainly expressed in organs that are involved in the
enterohepatic circulation of bile acids that is, liver and intestine as
well as kidneys and adrenal glands.3 Upon activation of FXR by
bile acids it represses the expression of CYP7A1 and CYP8B1, the
rate-limiting enzyme of bile acid synthesis from cholesterol and
of bile acid hydrophilicity, respectively, by increasing the expres-
sion of the atypical nuclear receptor small heterodimer partner
(SHP). SHP then decreases the pro-transcriptional regulation of
these genes by a third nuclear receptor liver receptor homologue
1 (LRH-1). FXR also controls enterohepatic circulation of bile acids
by regulating the expression of bile acid transporters and binding
proteins such as BSEP, IBABP, and NTCP.4
lipids resulting in a reduction of plasma cholesterol and triglycer-
ides. Besides its central role in the maintenance of bile acid homeo-
stasis, FXR also regulates genes affecting triglyceride metabolism
(e.g., apoCII, apoCIII, SREBP1-1c)5 leading to reduced plasma
triglyceride levels. FXR is also involved in glucose homeostasis. In
various mouse models6 FXR agonists were shown to decrease plas-
ma glucose levels through the modulation of different target genes
involved in the regulation of hepatic gluconeogenesis (e.g., PEPCK
or G6Pase) and glycogen synthesis (e.g., phosphorylation of
GSK3b), respectively. In addition, FXR was shown to improve insu-
lin sensitivity, although the underlying molecular mechanisms and
the contribution of FXR-independent pathways still need to be
determined. Consistent with these mechanisms and with positive
results from pre-clinical animal studies, FXR modulators have a
great therapeutic potential in the treatment of dyslipidemia,
atherosclerosis and diabetes. Further studies also suggest a possi-
ble role for FXR agonists in treating cholestasis,7 gallstone disease,8
nonalcoholic steatohepatitis (NASH)9 and diabetic nephropathy,10
although the final proof in humans is still to be established.
Since the deorphanization of FXR several potent steroidal
and non-steroidal FXR agonists have been reported ( Fig. 1).
GW4064,11 the first high-affinity non-steroidal ligand, and later
Fexaramine,12 served as useful tool compounds to investigate the
FXR pharmacology and to elucidate the topology and active confor-
mation of the FXR binding site. However, sub-optimal in vitro and
The combined effects of FXR agonism on bile acid composition
and pool size is a decrease in the intestinal absorption of dietary
⇑
Corresponding author. Tel.: +41 61 688 1330; fax: +41 61 688 6459.
E-mail address: hans.richter@roche.com (H.G.F. Richter).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.12.123

H. G. F. Richter et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1134–1140
1135
O
N
O
N
O
OH
Cl
O
O
Cl
N
O
Cl
Fexaramine
GW4064
O
O
N
F
OH
H
F
H
N
H
O
O
H
H
HO
OH
H
R
R
FXR-450 (WAY-362450), R = CH3
Reference compound, R = H
CDCA (R = H)
6-ECDCA (R = Et; INT-747)
Figure 1. Representative FXR agonists.
in vivo characteristics, such as poor rat pharmacokinetic properties
or instability under UV light and the presence of a potentially toxic
stilbene moiety in the case of GW4064, precluded further develop-
ment of these compounds.13 Researchers from Wyeth and Exelixis
published the synthesis of the non-steroidal agonist FXR-450 that
recently entered clinical trials.14 A follow-up paper described the
approaches applied for improving the physicochemical properties
of FXR-450.15 The most advanced FXR agonist to date is the
semi-synthetic bile acid derivative 6-ECDCA (a mixed FXR/GPBAR1
agonist)16 which recently completed Phase II clinical trials in
patients with primary biliary cirrhosis.17
We recently reported on the synthesis and structure–activity
relationship (SAR) profiling of a novel class of potent and selective
benzimidazole-based FXR agonists as exemplified by our lead
structure 1 (Fig. 2).18 Although 1, a potent, partial agonist at the
FXR receptor is endowed with oral activity in an in vivo model of
dyslipidemia, it was found to have poor physicochemical proper-
ties such as high lipophilicity and poor aqueous solubility. Surpris-
ingly, the compound also inhibited the hERG potassium channel
FXR with 6-ECDCA revealed that the carboxylate of 6-ECDCA is
located in this region and interacts through charge-assisted hydro-
gen bonds with the corresponding rat Arg328.19 We therefore
initiated a structure-guided design program in which we investi-
gated a variety of polarity-conferring substituents at different
distances from and orientation to the amide group, targeting in
particular an electrostatic interaction with Arg335.
Two strategies were followed for the synthesis of the novel
benzimidazoles. For both, single enantiomers were obtained
by chiral chromatography. The assignment of the absolute
stereochemistry was based on several co-crystal structures (incl.
crystallizations with racemates in which only the S enantiomers
bound) as well as small molecule X-ray crystallography of
in vitro with an IC50 of 1.6 lM.
Our lead optimization efforts, therefore, concentrated on the
introduction of polarity to reduce the high lipophilicity and in-
crease solubility and bioavailability (mouse F = 1.7%) and to over-
come the hERG liability of 1.
In accordance with the hydrophobic nature of the ligand bind-
ing site of human FXR, the original SAR18 indicated little scope
for the introducing polarity into these novel agonists. However,
following further analysis of structural information generated
in-house, we identified a more polar, as yet unexplored region in
FXR consisting of Gln267, Asn297, His298, Arg335 and three water
molecules in the proximity of the amide-bound cyclohexyl group
(Fig. 3). Comparison with a published co-crystal structures of rat
F
N
N
Cl
Cl
H
H
N
Figure 3. X-ray structure of the 5,6-di-H analogue18 of benzimidazolyl acetamide 1
bound to human FXR (PDB: 3OKI) illustrating the targeted polar binding site region.
The front of the FXR binding site is removed for clarity. Protein–ligand hydrogen
bonds to Ser336 and Tyr373 are depicted as red dashed lines. The structure is
overlaid with the 6-ECDCA/rat FXR complex (magenta, PDB: 1ot7, protein not
shown).
O
1
Figure 2. Lead compound from the benzimidazolyl acetamide class.

1136
H. G. F. Richter et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1134–1140
Table 1
Binding and functional activity at human FXR for selected benzimidazolyl acetamides
R1
N
N
R3
R2
H
H
N
R4
O
a
Compound
R1, R2
R3
R4
IC50
(l
M)
EC50 (l
M) (efficacy)b
GW4064 (see Fig. 1)
Reference compound (see Fig. 1)
4-Cl-Ph
0.112
0.021
0.013
0.91 (99%)
0.085 (39%)
0.1 (42%)
1
F, Cl
C6H11
7a
7b
7c
7d
7e
7f
7g
7h
7i
7j
7k
7l
7m
7n
7o
7p
7q
7r
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, H
H, H
F, F
F, F
F, F
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
6-Cl-pyridin-3-yl
2,6-(MeO)2-pyridin-3-yl
2-MeO-6-Cl-pyridin-3-yl
C6H11
0.010
0.041
0.61
0.05
0.32c
1.3c
0.037
0.009
0.02
0.008
0.74
0.004
1.4c
0.05 (31%)
0.3 (40%)
2.7 (26%)
2.5 (27%)
trans-4-OH–C6H10
trans-4-CO2H-C6H10
4-CO2H-Ph
2-CO2H-Ph
3-CO2H-Ph
>40c
10.7 (14%)c
0.3 (30%)
0.3 (54%)
0.5 (51%)
0.05 (59%)
10.4 (32%)
1.4 (48%)
4.9 (14%)c
2.0 (18%)c
0.036 (17%)
0.6 (21%)
10.0 (16%)c
5.0 (27%)
9.9 (26%)
10.4 (10%)c
2-F-4-CO2H-Ph
2-Cl-4-CO2H-Ph
2-Me-4-CO2H-Ph
2-CF3-4-CO2H-Ph
2-MeO-4-CO2H-Ph
2,6-F2-4-CO2H-Ph
4-CO2H-CH2-Ph
4-CO2H-CH2CH2-Ph
2-F-4-CO2H-CH2CH2-Ph
2-F-4-CO2H-Cc-Pr-O-Ph
trans-4-CO2H-C6H10-CH2
trans-4-CO2H-CH2-C6H10
trans-4-CO2H-CH2O-C6H10
4-Me-5-CO2H-thiazole
4-CF3-5-CO2H-thiazole
4-CO2H-2-Py
4-CO2H-3-Py
2-F-4-(1H-tetrazol-2-yl)-Ph
2-F-4-CO2H-Ph
2-CF3-4-CO2H-Ph
trans-4-CO2H-C6H10
trans-4-CO2H-C6H10
trans-4-CO2H-C6H10
0.06c
0.013
0.001
3.9c
0.61
0.83
0.09c
0.20c
0.20c
3.1c
0.008
0.04
0.01
1.7
0.21
0.29
7s
7t
7u
7v
7w
7x
7y
7z
7aa
7ab
7ac
>40c
10.6 (20%)c
>40c
0.3 (21%)
1.1 (25%)
0.09 (35%)
9.7 (18%)
1.7 (19%)
1.1 (21%)
a
SPA scintillation proximity assay (SPA). Binding affinity was assessed in a radioligand displacement assay according to the method of Nichols, J. S. et al.22 Values are means
of at least two experiments.
b
Transactivation assay. Transactivation activity was measured using a luciferase transcriptional reporter-gene assay according to the method of Jingami.23 Efficacies are
relative to GW4064 set at 100%.
c
Value for racemic mixture.
compounds 7d and 7g that proved S configuration at the chiral
center.20
converting the cyano group in 6 using standard procedures.
Analogue 7z (R1 = R2 = H) was prepared using the same protocol,
starting from commercially available 2-(4-chlorophenyl)benz-
imidazole.
The first synthetic route was based on a four-component Ugi
reaction treating Boc-protected phenylene diamines with carbox-
ylic acids, aldehydes and isonitriles as previously reported (1, 7a,
7y, 7aa, 7ab, 7ac, Table 1).18 Although this method gave us rapid
access to the target molecules, the limited number of available iso-
nitriles of interest triggered the development of an additional syn-
thetic route as depicted in Scheme 1. This methodology was used
preferably for the synthesis of ‘symmetrical’ benzimidazoles
(R1 = R2 = F or H). Benzimidazole 2 was synthesized by condensa-
tion of commercially available 4,5-difluoro-benzene-1,2-diamine
The structure–activity relationship profile for the newly synthe-
sized compounds is summarized in Table 1. Based on our previous
SAR we chose the equipotent 6-fluoro analogue of 1 (7a) to explore
the amide exit vector. The binding affinity of the newly synthe-
sized compounds was evaluated using a scintillation proximity
assay (SPA).22 Functional activity was determined in a Gal-4 trans-
activation assay23 using the human FXR ligand binding domain
(maximal efficacy relative to maximal efficacy of GW4064 in this
assay). As with analogues of 1,18 the R enantiomers were shown
to be significantly less potent agonists or else inactive (at concen-
trations up to 40 lM) versus the S enantiomers. Consequently, we
first tested the racemate activity and, if this proved interesting,
single enantiomers were prepared and tested.
Introduction of a hydroxy group in the 4-position of the amide-
bound cyclohexyl ring in 7a (7b) led to fourfold drop in binding
affinity and potency while the efficacy was maintained. Replacing
the hydroxy function in 7b with a carboxylate group (7c) led to a
and 4-chloro-benzoic acid. Alkylation of 2 with
a-bromo-cyclo-
hexyl-acetic acid ester 3, prepared in a three-step, one-pot proce-
dure from commercially available cyclohexyl-acetic acid,21 gave
compound 4. Cleavage of the ethyl ester in 4 provided the acid 5.
Amide coupling of 5 with different amines and anilines yielded
amides 6. Methyl or ethyl ester groups on substituent R3 in inter-
mediates 6 (either prepared via Ugi reaction or as described above)
were cleaved under basic conditions to provide the racemic acids 7.
Analogue 7x carrying a tetrazole-substituted R3 was prepared by

H. G. F. Richter et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1134–1140
1137
F
F
NH2
NH2
HO
O
HO
b
Cl
a
O
Br
O
O
3
F
F
F
F
F
F
N
N
N
N
N
c
d
Cl
Cl
Cl
N
H
O
O
2
4
5
O
OH
e or f or g
R3 =
*
*
*
F
F
F
F
N
N
N
N
h or i
Cl
Cl
F
F
F
HO
N
O
HO
N
O
NH
O
N
O
R3
O
R3
7g
(S enantiomer)
7p
7x
(S enantiomer)
7
6
NH
NH
(S enantiomer)
R3 with carboxylic acid
or tetrazole group
R3 with carboxylic acid ester
or cyano group
Scheme 1. Synthesis of novel benzimidazolyl acetamides. Reagents and conditions: (a) polyphosphoric acid, 160 °C (77%); (b) (i) SOCl2, reflux; (ii) Br2, reflux; (iii) EtOH, 0 °C
to rt (99% crude over three-steps, one-pot); (c) 3 (3.75 equiv), Cs2CO3 (3.75 equiv), DMF, 100 °C (79%); (d) LiOH (3 equiv), dioxane, water, reflux (99%); (e) SOCl2, reflux; then
amine, pyridine (7i, 7j, 7w, 62–84%) or DMAP (7d, 7g, 7h, 7k, 7l, 7p, 7s, 7v, 7x, 47–97%) or Huenig’s base (7b (66%), 7q (67%)), CH2Cl2, 0 °C to rt; (f) HATU, Huenig’s base,
CH2Cl2, rt (7c, 7e, 7f, 7m, 7n, 7t, 7u, 17–82%); (g) 2-chloro-1-methylpyridium iodide, NEt3, CH2Cl2, 40 °C (7r, 86%); (h) LiOH, dioxane, water, reflux (67–99%) or NaOH,
acetonitrile, water, microwave, 100 °C (7c, 99%); (i) NaN3, HNEt3Cl, o-xylene, 145 °C (7x, 29%).
more pronounced loss in binding affinity and potency. Binding
affinity was improved by exchanging the cyclohexyl with a phenyl
ring (7d). These early results indicated that the chosen exit vector
and receptor sub-pocket tolerates the introduction of polar groups.
Moving the carboxylate group in 7d to the 2- or 3-position (7e, 7f)
led to a loss in binding activity and an even more pronounced
reduction of potency, with derivative 7e failing to trigger a func-
tional response. Acyclic aliphatic carboxylic acid residues
(CH2)nCO2H, n = 1–5 were significantly less active (data not
shown). Substantially higher potency could be achieved by intro-
ducing fluoro, chloro, methyl or trifluoromethyl substituents in
the 2-position of the 4-carboxy-substituted phenyl ring (7g–7j).
For 7j these effects are most pronounced with gains of binding
affinity and potency of 7- and 49-fold, respectively.
the binding affinity (7m), whereas the longer ethylene spacer re-
stored the binding affinity (7n). The activity of 7n could again be
improved through the introduction of a fluorine substituent in
the 2-position (7o). By combining the 2-fluoro substituent with a
Analysis of the structure of human FXR complexed with 7g
(Fig. 4) suggested additional beneficial contacts of the 2-substitu-
ent of the phenyl ring with two Met residues in a small hydropho-
bic sub-pocket as the main reason for the stronger interaction. It is
also noteworthy that 2-substitution induces a change of the
relative orientation of the phenyl and amide planes from in-plane
(2-H, X-ray structure not shown) to orthogonal (2-des-H), resulting
in different ring interactions. The perpendicular arrangement is
also the preferred conformation of the amide-bound cyclohexyl
ring (Fig. 3). The X-ray structure of 7g24 confirms that polar inter-
actions are made between the newly introduced carboxylate group
and its protein environment. Larger 2-substituents, for example,
2-methoxy (7k) again diminished binding affinity and potency,
likely because the sub-pocket is too small. In contrast, introduction
of a second fluoro substituent at the 2’ position of 7g (7l) was
tolerated.
Figure 4. X-ray structure of benzimidazolyl acetamide 7g bound to the ligand
binding domain of human FXR (PDB: 3OMM. Resolution: 2.1 Å). Water-mediated
and direct protein–ligand hydrogen bonds to Gln267, Arg335, Ser336 and Tyr373
(red, dashed lines) and hydrophobic interactions of the 2-F substituent to Met294
and Met332 (blue, dashed lines) are shown.
We found that the introduction of a methylene spacer between
the phenyl ring and the carboxylate group significantly reduced

1138
H. G. F. Richter et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1134–1140
4-cyclopropoxy side-chain on the phenyl ring (7p) we identified
the analogue with highest binding affinity in this series although
this structural variation had little impact on potency as compared
to 7g. We also investigated the effects of additional methylene
groups in the series of cyclohexyl analogues. Whereas the incorpo-
ration of a methylene (or a CH2O) group between the cyclohexyl
ring and the carboxylate group had only little effect on the binding
affinity (7r, 7s vs 7c), moving the methylene spacer between the
amide nitrogen and the cyclohexyl ring led to a drop of binding
affinity and potency (7q). From a number of five-membered
heterocycles investigated only compounds 7t and 7u showed
sub-micromolar binding affinity, although functional activity was
poor or even absent for 7t and 7u, respectively. Pyridine carboxylic
acids (7v, 7w) were also less active than the corresponding phenyl
derivative 7d. As a next step we investigated tetrazoles as bioisos-
teric replacements for the carboxylate group. As exemplified by 7x,
this change led to a fivefold improved binding affinity and compa-
rable potency as for 7g. In an attempt to reduce overall molecular
weight and lipophilicity we also investigated the 6-des-fluoro and
5,6-des-fluoro substitution pattern on the benzimidazole core. As
shown for 7y and 7z, this variation had little effect on the binding
affinity and only a trend towards lower potency was observed
compared to the 5,6-di-fluoro counterparts 7g and 7j, respectively.
We finally investigated a number of pyridine substituents as less
lipophilic replacements for the 4-chloro-phenyl substituent. Con-
sistent with our previously generated SAR, unsubstituted pyridine
substituents diminished the activity (data not shown). The 3-aza-
analogue of 7c (7aa) also showed weaker binding affinity and a
significant drop in potency. However, appropriately substituted
pyridines such as 7ab and 7ac, especially with a 2-methoxy substi-
tuent on the pyridine ring, showed a two to threefold increase in
binding affinity with a trend to higher potency as compared to
7c, underlining the broad scope for structural variation within this
compound series also at this position.
possible that differences in binding to cellular proteins and/or
recruitment of co-activators or co-repressors to the artificial
FXR/GAL4 complex within the nucleus could explain why in vitro
activity and efficacy did not better predict in vivo efficacy.
Following the SAR assessment several compounds were se-
lected for profiling of their physicochemical properties, in vitro
metabolic stability, inhibition of three major human CYP isoforms,
and potential to inhibit the potassium hERG channel. The results
are summarized in Table 2. Whereas the introduction of a hydroxy
group in 7a (7b) slightly improved the solubility, lipophilicity
stayed in a high range. As expected, analogues with a carboxylate
or tetrazole moiety had significantly better aqueous solubility
and much lower lipophilicity (log D 1.91–3.06).
The permeability of compounds 7d and 7g in an artificial per-
meation assay were evaluated and both samples showed a high
permeability over the full pH range of 3.0–8.0. Based on similarity
of physicochemical properties, permeability is expected to be high
for the other compounds as well.
All of the compounds showed an improved CYP450 profile ver-
sus 7a except for tetrazole 7x, which was found to have a compa-
rable profile to 7a but inferior profile to its carboxylate counterpart
7h, indicating low drug–drug interaction potential. Metabolic sta-
bility in mouse and human microsomes was greatly improved for
all the tested compounds versus 7a.
Finally, we tested the compounds for their hERG potassium
channel inhibitory activity. Compared to 7a, compounds 7b and
7c showed a fivefold reduced hERG activity, although moderate
activity was retained (IC50 7.3
Compounds 7d, 7g, 7p, and 7z however showed substantially
attenuated hERG affinity with IC50 values greater than 20 M.
lM and 7.6 lM, respectively).
l
Interestingly, we found that the introduction of a carboxyl group
onto the phenyl ring was not sufficient to circumvent the hERG
activity of the tested derivatives. By replacing the 2-fluoro substi-
tuent in 7g by a chloro (7h) or a methyl group (7i) the compounds
picked up hERG activity again with 7i being the most potent com-
pound in this series with an IC50 of 0.7 lM.
Selectivity of the compounds for FXR was evaluated using a
panel of reporter-gene transcriptional assays for other nuclear
Comparison of in vitro binding, potency and efficacy data with
in vivo efficacy (plasma triglyceride and LDL cholesterol lowering
in LDL receptor deficient mice) led us to identify binding IC50s as
the best predictor for in vivo activity provided that in vitro efficacy
was P10% in the transactivation assay (results not shown). It is
receptors including PPAR-a, -d, -c, LXR-a, -b and RXR-a. No
Table 2
Solubility, lipophilicity, CYP450 inhibition, microsomal clearance and hERG inhibition data for selected benzimidazolyl acetamides
R1
N
N
R3
R2
H
H
N
R4
O
Solubilitya
Compound
R1, R2
R3
R4
log Db
CYP IC50
2D6, 2C9
(l
M) 3A4,
Clint (h/m)c ((
mg protein)
lL/min)/
hERG
IC50
d
(lg/mL)
(l
M)
1
7a
7b
7c
7d
7g
7h
7i
7p
7x
7z
7ab
7ac
F, Cl
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
F, F
H, H
F, F
F, F
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
2,6-(MeO)2-pyridin-3-yl
2-MeO-6-Cl-pyridin-3-yl
C6H11
C6H11
trans-4-OH-C6H10
trans-4-CO2H-C6H10
4-CO2H-Ph
<1
<1
17
395
88
115
51
>4
>4
>4
4.5, >50, >50
7.6, 32, 6.7
33, >50, 40
>50, >50, >50
37, >50, 32
39, >50, 14
23, >50, 18
>50, >50, 19
>50, >50, 29
7.7, 31, 7.1
49, 36, 28
137/909
221/1654
30/20
5/6
8/19
9/13
31/26
12/8
0/0
12/11
12/21
1/17
1.6
1.4
7.3
7.6
>20
>20
7.8
2.49
3.06
2.66
2.98
2.83
2.32
2.96
2.34
1.91
2.05
2-F-4-CO2H-Ph
2-Cl-4-CO2H-Ph
2-Me-4-CO2H-Ph
2-F-4-CO2H-Cc-Pr-O-Ph
2-F-4-(1H-tetrazol-2-yl)-Ph
2-CF3-4-CO2H-Ph
trans-4-CO2H-C6H10
trans-4-CO2H-C6H10
32
0.7
280
490
175
465
440
>20
>10
>20
>10
NDe
>50, >50, >50
NDe
4/0
a
b
c
Aqueous solubility at pH 6.5 in 0.05 M phosphate buffer.
pH 7.4.
Intrinsic clearance in human (h) and mouse (m) liver microsomes.
Inhibition of hERG potassium channel determined by whole-cell patch-clamp experiments in a transfected CHO cell line.
ND = not determined.
d
e

H. G. F. Richter et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1134–1140
1139
Table 3
Pharmacokinetic parameters of compound 7g after intravenous and oral administration in male C57BL/6 mice and male Wistar ratsa
Species
Dose (mg/kg)
Cl ((mL/min)/kg)
10.0
Cmax/dose (ng/mL)
Vss (L/kg)
AUC/dose (ng ⁄ h/mL)
543.3
t1/2 (h)
F (%)
Male C57BL/6 mouseb
4 (iv)
10 (po)
2 (iv)
205.4
1.24
1.61
33%
59%
Male Wistar ratc
16.3
269.6
0.62
606.5
1.16
5 (po)
a
7.5% gelatin in 0.62% aqueous NaCl and 30% N-methylpyrrolidinone in aqueous100 mM TRIS buffer pH 8.5 were used as vehicles for oral and intravenous administration,
respectively.
b
n = 2 per time point.
n = 3 per time point.
c
cross-reactivity was observed with these receptors up to a concen-
tration of 10 M (data not shown).
drug design allowed us to address the key liabilities of the previous
lead compound 1 through the identification of an exit vector which
permitted the introduction of polar residues. Several compounds
l
The pharmacokinetic parameters of 7g were determined in
mice and rats and the data are summarized in Table 3. After oral
dosing in mice the compound was characterized by a low clearance
(Cl) of 10 mL/min/kg, moderate half-life (t1/2) of 1.6 h, volume of
distribution (Vss) of 1.24 L/kg and moderate oral bioavailability
(F) of 33% which resulted in an overall good exposure (AUC) of
5433 ng ⁄ h/mL. In rats, 7g had a slightly higher clearance of
16 mL/min/kg, a moderate half-life of 1.2 h, low volume of distri-
bution of 0.62 L/kg and a good bioavailability of 59%, leading to a
good exposure of 3033 ng ⁄ h/mL.
with
a negatively charged functionality displayed excellent
physicochemical, ADME and in vitro safety properties culminating
in the discovery of 4-{(S)-2-[2-(4-chloro-phenyl)-5,6-difluoro-
benzoimidazol-1-yl]-2-cyclohexyl-acetylamino}-3-fluoro-benzoic
acid (7g) with favorable rodent pharmacokinetic properties and
potent plasma lipid lowering effects in LDLRÀ/À mice after oral
administration.
Acknowledgement
Based on the high in vitro potency of 7g for human FXR (Table
1) and murine FXR (IC50 0.29 lM, EC50 0.87 lM (38%)), selectivity
The work of Martin Ritter and Kersten Klar is gratefully
acknowledged.
and favorable pharmacokinetic properties we evaluated its effects
on plasma lipid profiles in high-fat diet fed LDL receptor deficient
(LDLRÀ/À) mice.25 LDLRÀ/À mice are reported to respond to FXR
treatment with decreases in plasma cholesterol and triglycerides
as a consequence of reduced intestinal absorption and decreased
synthesis.26 After five days of treatment (10 mg/kg po, QD), 7g
caused a statistically significant decrease in plasma total choles-
terol (TC, À41%), low density lipoprotein cholesterol (LDL-C,
À33%) and triglycerides (TG, À59%). These reductions were compa-
rable to the effects seen with the reference compound (ethyl ester
analogue of FXR-450, Scheme 1) dosed at 30 mg/kg/d (Fig. 5). PK
monitoring of 7g revealed a plasma exposure of 885 ng/mL
References and notes
1. (a) Forman, B. M. et al Cell 1995, 81, 687; (b) Seol, W.; Choi, H.-S.; Moore, D. D.
Mol. Endocrinol. 1995, 9, 72.
2. (a) Makishima, M.; Okamoto, A. Y.; Repa, J. J.; Tu, H.; Learned, R. M.; Luk, A.;
Hull, M. V.; Lustig, K. D.; Mangelsdorf, D. J.; Shanz, B. Science 1999, 284, 1362;
(b) Parks, D. J. et al Science 1999, 284, 1365; (c) Wang, H.; Chen, J.; Hollister, K.;
Sowers, L. C.; Forman, B. M. Mol. Cell 1999, 3, 543.
3. Zhang, Y.; Kast-Woelbern, H. R.; Edwards, P. A. J. Biol. Chem. 2003, 278, 104.
4. Lee, F. Y.; Lee, H.; Hubbert, M. L.; Edwards, P. A.; Zhang, Y. Trends Biochem. Sci.
2006, 31, 572.
5. Zhang, Y.; Edwards, P. A. FEBS Lett. 2008, 582, 10.
(1.63
pharmacokinetic properties of 7g versus 1 (plasma exposure after
2 h of 0.95
M at a dose of 30 mg/kg).18
In summary, we developed a novel class of potent and selective
non-steroidal FXR agonists based on a benzimidazole acetamide
scaffold. Our lead optimization efforts guided by structure-based
l
M) after 2 h. This result further confirmed the improved
6. (a) Zhang, Y.; Lee, F. Y.; Barrera, G.; Lee, H.; Vales, C.; Gonzalez, F. J.; Willson, T.
M.; Edwards, P. A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 1006; (b) van Dijk, T. H.;
Grefhorst, A., et al J. Biol. Chem. 2006, 281, 11039; (c) Ma, K.; Saha, P. K.; Chan,
L.; Moore, D. D. J. Clin. Invest. 2006, 116, 1102.
7. (a) Stedman, C.; Liddle, C.; Coulter, S.; Sonoda, J.; Alvarez, J. G.; Evans, R. M.;
Downes, M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 11323; (b) Jones, S. A. et al J.
Clin. Invest. 2003, 112, 1678; (c) Fiorucci, S. et al J. Pharmacol. Exp. Ther. 2005,
313, 604.
l
8. (a) Doggrell, S. A. Curr. Opin. Invest. Drugs 2006, 7, 344; (b) Moschetta, A.;
Bookout, A. L.; Mangelsdorf, D. J. Nat. Med. 2004, 10, 1235.
9. (a) Cariou, B. Diabetes Metab. 2008, 34, 685; (b) Kong, B.; Luyendyk, J. P.; Tawfik,
O.; Guo, G. L. J. Pharmacol. Exp. Ther. 2009, 328, 116.
10. Levi, M. et al Diabetes 2007, 56, 2485.
11. Maloney, P. R. et al J. Med. Chem. 2000, 43, 2971.
12. Evans, R. M. et al Mol. Cell 2003, 11, 1079.
13. Navas, F., III et al Bioorg. Med. Chem. Lett. 2009, 19, 4733.
14. Flatt, B. et al J. Med. Chem. 2009, 52, 904.
140
120
100
15. Lundquist, J. T., IV; Harnish, D. C., et al J. Med. Chem. 2010, 53, 1774.
16. Pellicciari, R.; Fiorucci, S.; Camaioni, E.; Clerici, C.; Costantino, G.; Maloney, P.
R.; Morelli, A.; Parks, D. J.; Willson, T. M. J. Med. Chem. 2002, 45, 3569.
17. Press release Intercept Pharmaceuticals; http://www.interceptpharma.com.
18. Bleicher, K. H. et al Bioorg. Med. Chem. Lett. 2011, 21, 191.
19. Mi, L.-Z.; Devarakonda, S.; Harp, J. M.; Han, Q.; Pellicciari, R.; Willson, T. M.;
Khorasanizadeh, S.; Rastinejad, F. Mol. Cell 2003, 11, 1093.
20. Crystallographic data for structures 7d and 7g have been deposited at the
Cambridge Crystallographic Data Centre under the deposition numbers CCDC
804960 (7d) and 804961 (7g), respectively. Copies of these data can be
obtained free of charge via the Internet at www.ccdc.cam.ac.uk/conts/
retrieving.html (or from the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB21EZ, UK; fax: +44 1223 336 033; or
deposit@ccdc.cam.ac.uk).
**
80
60
40
20
0
**
**
**
**
**
**
Total cholesterol
LDL-C
Triglycerides
7g (10 mg/kg)
Control
Reference (30 mg/kg)
Figure 5. Effect of 7g versus reference compound on plasma total cholesterol,
LDL-C and triglycerides in LDLRÀ/À mice fed a high-fat diet for 15 days after 5 days
oral treatment: (⁄⁄) P <0.01 versus vehicle control, one way ANOVA followed by
Dunnett‘s post hoc test.
21. Schwenk, E.; Papa, D. J. Am. Chem. Soc. 1948, 70, 3626.
22. (a) Nichols, J. S.; Parks, D. J.; Consler, T. G.; Blanchard, S. G. Anal. Biochem. 1998,
257, 112; (b) Nichols, J. S.; Parks, D. J.; Consler, T. G.; Blanchard, S. G. Anal.
Biochem. 1998, 263, 126 (corrigendum).

1140
H. G. F. Richter et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1134–1140
23. Kanaya, E.; Shiraki, T.; Jingami, H. Biochem. J. 2004, 382, 913.
24. Coordinates and diffraction data of the FXR co-crystal structures of 7d, 7g, 7i
and 7l have been deposited with the Protein Data Bank (www.pdb.org) under
accession codes 3OOF, 3OMM, 3OLF and 3OOK, respectively.
25. Repa, J. J.; Turley, S. D.; Quan, G.; Dietschy, J. M. J. Lipid Res. 2005, 46, 779.
26. (a) Zhang, Y.; Yin, L.; Anderson, J.; Ma, H.; Gonzalez, F. J.; Willson, T. M.;
Edwards, P. A. J. Biol. Chem. 2010, 285, 3035; (b) Mencarelli, A.; Fiorucci, S. J.
Cell. Mol. Med. 2010, 14, 79.