Identification and structure-activity relationships of ortho-biphenyl carboxamides as potent Smoothened antagonists inhibiting the Hedgehog signaling pathway

Article abstract of DOI:10.1016/j.bmcl.2008.11.096

Ortho-biphenyl carboxamides, originally prepared as inhibitors of microsomal triglyceride transfer protein (MTP) have been identified as novel inhibitors of the Hedgehog signaling pathway. Structure-activity relationship studies for this class of compound

Full text of DOI:10.1016/j.bmcl.2008.11.096

Bioorganic & Medicinal Chemistry Letters 19 (2009) 328–331
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification and structure–activity relationships of ortho-biphenyl
carboxamides as potent Smoothened antagonists inhibiting
the Hedgehog signaling pathway
*
Stefan Peukert , Rishi K. Jain, Adrian Geisser, Yingchuan Sun, Rui Zhang, Aaron Bourret, Adam Carlson,
Jennifer DaSilva, Arun Ramamurthy, Joseph F. Kelleher
Novartis Institutes for BioMedical Research, Inc., 250 Massachusetts Avenue, Cambridge, MA 02139, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Ortho-biphenyl carboxamides, originally prepared as inhibitors of microsomal triglyceride transfer pro-
tein (MTP) have been identified as novel inhibitors of the Hedgehog signaling pathway. Structure–activity
relationship studies for this class of compounds reduced MTP inhibitory activity and led to low nanomo-
lar Hedgehog inhibitors. Binding assays revealed that the compounds act as antagonists of Smoothened
and show cross-reactivity for both the human and mouse receptor.
Received 22 August 2008
Revised 24 November 2008
Accepted 24 November 2008
Available online 3 December 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
Hedgehog inhibitor
Smoothened antagonist
Antiproliferative agent
SAR
The Hedgehog (Hh) signal transduction pathway plays critical
roles in the development and homeostasis of many organs and tis-
sues.1 In the resting state, the 12-pass transmembrane protein
Patched (Ptch) inhibits activity of the membrane receptor Smooth-
ened (Smo) which resembles G-protein coupled receptors (GPCR)
in general topology. The Hh family protein ligands Sonic Hedgehog,
Desert Hh, and Indian Hh can all bind to Ptch, which releases the
repression of Smo activity. Active Smo signals via a cytosolic com-
plex of proteins leading to activation of the Gli family of transcrip-
tion factors. Active Gli1 signaling can cause cell proliferation and
differentiation. Genetic activation of the Hh pathway at or up-
stream of Smo is linked to tumorigenesis in several cancers such
as basal cell carcinoma and medulloblastoma.2
Treatment with small molecule Hh inhibitors such as HhAntag
and the natural product cyclopamine, both binding to Smo, in-
duced tumor remission in a genetic mouse model of medulloblas-
toma.3,4 We therefore became interested in identifying small
molecule inhibitors of the Hh pathway with cross-reactivity to
both the human and mouse Smo receptor to be of potential use
in a therapeutic application.5
boxamides such as LAB687 (Fig. 1) as micromolar inhibitors of
the Hh signaling pathway. In addition, binding assays showed
weak binding of this compound to the mouse and human Smo
receptors (Table 1). LAB687 was originally prepared as an antago-
nist of the microsomal triglyceride transfer protein (MTP) inhibit-
ing the secretion of apoB from Hep G2 cells with subnanomolar
activity (Table 1).6,7 Our initial aims were to identify compounds
with improved Hh and Smo potency over LAB687 and significantly
reduced ability to inhibit the secretion of apoB. Herein, we describe
the structure–activity relationships obtained from studying the ef-
fect of substituents on the biaryl unit (Fig. 1, red box), the stereo-
chemistry of the aminoindane moiety (yellow dot), the nature of
the linker (green box), and the western substituent (blue box).8
Synthesis: The preparation of the ortho-biphenyl carboxamides
4–26 follows the route described in Ksander et al. (Scheme 1).6
Ortho-biphenyl carboxylic acids 1 were converted into their acid
chlorides and coupled with either racemic or the enantiomerically
CF3
In this study we describe potent inhibitors of the Hh pathway
which act as antagonists of the Smo receptor. Screening of part
of our compound collection provided us with ortho-biphenyl car-
H
N
O
O
N
H
O
LAB687 (2a)
* Corresponding author. Tel.: +1 617 871 3644; fax: +1 617 871 4081.
E-mail addresses: stefan.peukert@novartis.com, stefanpeukert@yahoo.com
(S. Peukert).
Figure 1. Structure of LAB687, a Hedgehog inhibitor screening hit.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.11.096

S. Peukert et al. / Bioorg. Med. Chem. Lett. 19 (2009) 328–331
329
Table 1
In vitro activity in the Hedgehog cell assay, the Smoothened mouse (mu), and human (hu) receptor binding assay and the apoB secretion assay (IC50s, nM)9
R1
H
N
O
R2
*
X
N
H
R3
Compound
Chirality
R1
R2
R3
X
Hh assay
Smo mu binding
Smo hu binding
apoB assay
2a (LAB687)
2b
2
4
5
R
S
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
Cl
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
CH3
OCH3
H
H
H
COOMe
COOMe
COOMe
COPh
SO2Ph
CH2Ph
CH2-2-pyridyl
CH2-2-pyridyl
CH2-2-pyridyl
CH2-2-thiazolyl
CH2-2-thiazolyl
CH2-2-pyridyl
CH2-2-pyridyl
CH2-2-pyridyl
CH2-2-pyridyl
CH2-2-pyridyl
1200
2150
1330
600
1250
93
100
4890
100
4270
25
6800
>5000
200
2480
5090
2960
3420
7520
2700
0.9
4.3
rac
rac
rac
rac
rac
R
S
R
S
rac
rac
rac
rac
rac
6
7
110
8770
57
1190
10
8680
>5000
490
2710
1420
7a
7b
8a
8b
9
10
11
12
13
3.5
2350
0.6
120
770
30
140
H
H
H
H
790
4950
1000
H
1150
R1
R1
R1
R1
O
N
O
H
N
O
O
O
NH2
R2
R2
d
R2
a
b
R2
CO2H
c
*
X
*
*
H
N
H
N
H
N
H
X =
R3
4
COPh
SO2Ph
R3
1
R3
R3
5
2
3
6-16,19 &
21-26 CH2R
17,18 CHR'Ph
20
Ph
Scheme 1. Reagents and conditions: (a) oxalyl chloride, rt, 16 h; (b) rac, R or S (5-amino-indan-2-yl)-carbamic acid methyl ester, pyridine, CH2Cl2, 0 °C, 1 h; (c) TMSI, CH3CN,
rt, 16 h; (d) for 4: PhCOCl, DIPEA, CH2Cl2, 0 °C, 2 h; for 5: PhSO2Cl, DIPEA, CH2Cl2, 0 °C, 2 h; for 6–16 and 21–26: RCHO, sodium triacetoxy borohydride, CH2Cl2, rt, 16 h; for 17,
18: PhCOR0, Ti(iPrO)4, THF, 100–150 °C, 30 min, then NaBH4, EtOH, rt, 2 h; for 20: bromo-benzene, Pd2dba3, Xantphos, Cs2CO3, toluene, 130–140 °C, microwave synthesizer,
4 h.
pure S and R diaminoindane methylcarbamates7 to yield com-
pounds 2. Removal of the methylcarbamate with iodotetrameth-
ylsilane (TMSI) produced the ortho-biphenyl aminoindanes 3.
Reaction of 3 with either benzoyl chloride or phenylsulfonyl chlo-
ride provided the desired N-substituted 2-aminoindanes 4 and 5.
N-Alkyl derivatives 6–16, 19, and 21–26 were prepared by reduc-
tive amination of the free amines 3 with the appropriate aldehydes
in the presence of sodium triacetoxy borohydride as reducing
agent. On the other hand, N-alkyl derivatives 17 and 18 required
pre-formation of the imines followed by reduction with NaBH4.
Synthesis of N-phenyl compound 20 was best achieved using Buch-
wald–Hartwig amination of bromo-benzene with amine 3 and
microwave heating at 130–140 °C.
Results and discussion: Table 1 shows the structure–activity rela-
tionship (SAR) for representatives of the biphenyl-2-carboxylic
acid (2-amino-indan-5-yl)-amide series. A comparison of com-
pounds 4, 5, and 6 which differ only in the functional group con-
necting the aminoindane to the phenyl substituent suggests that
benzylamines are superior to carboxamides or sulfonamides in
terms of Hh inhibitory activity. Key for increasing Hh potency
and simultaneously decreasing MTP activity proved to be the ste-
reochemistry of the aminoindane moiety: A comparison of the
enantiomeric compound pairs 7a,b and 8a,b demonstrates that
the Hh activity of the N-benzyl amines resides predominantly
(P50:1) in the S-enantiomer whereas the R-enantiomers (7a, 8a)
are significantly more potent in inhibiting apoB secretion. Enantio-
meric purity was determined by chiral chromatography to be
>99.9% ee for both 8a and b practically excluding that the weak
Hh activity observed with the R-enantiomer 8a is caused by con-
tamination with small amounts of the S-enantiomer. This diverging
SAR for Hh and apoB activity is characteristic for the N-benzylam-
ines and not observed to the same degree for the other series, for
example, the carbamates (2a,b). We next explored the influence
of the biphenyl substituents R1, R2, and R3 on Hh potency. Moving
the methyl group from the ortho to the para position on the lower
phenyl ring (compounds 7 and 9) resulted in a drastic decrease of
activity. Other substituents such as the methoxy group in com-
pound 10 in the R3-position were not tolerated either. Replacing
the ortho methyl substituent with a hydrogen yielded compound
11 with a slight drop in activity compared to the corresponding
analogue 7. Finally, substituents R1 play a crucial role for activity,
too: The trifluoromethyl substituent is the best substituent identi-
fied so far; both chloro and hydrogen in this position (compounds
12 and 13) resulted in lower activity.

330
S. Peukert et al. / Bioorg. Med. Chem. Lett. 19 (2009) 328–331
For all compounds tested in Table 1, Hh inhibition and binding
pounds in Table 1 Hh pathway inhibitory activity and affinity to
the human and mouse Smo receptor track very well over the full
range of potency. The most potent inhibitors in Table 2 show a
ꢀ5-fold higher activity in the Hh cell assay and significantly higher
affinity to the Smo receptors than the natural product
cyclopamine.
In conclusion, we described the optimization based on SAR
studies of the micromolar hit LAB687 to afford low nanomolar
inhibitors (e.g., 21b, 23b) of the Hh signaling pathway which act
as antagonists for both the human and mouse Smo receptor. Dur-
ing this process the activity against the original target MTP of this
compound class was significantly reduced and the most potent Hh
inhibitors (e.g., 8b, 21b, 22b) have a ratio on-target/off-target of
ꢀ10:1. Therefore, these compounds have potential as novel lead
structures in the search for therapeutic agents for the treatment
of Hh dependent cancers.
to the mouse Smo receptor correlates well and suggests that the
observed Hh activity observed in the reporter-gene cell assay is
driven by Smo antagonism. Furthermore, the compounds show
good cross-reactivity between the human and mouse Smo receptor
with IC50 values for both receptors typically within a factor of two.
As the S-enantiomers of the N-benzyl amines provided the best
potency and selectivity, the SAR of this compound series was fur-
ther explored in more detail (Table 2). Replacing the phenyl substi-
tuent (6b) with either a 2-pyridyl or 4-pyridyl substituent (7b,
14b) resulted in a ꢀ5- to 7-fold drop in potency in the Hh assay
and a stronger drop in binding affinity to the mouse Smo receptor.
Substitutions on the phenyl ring were not well tolerated and the
activity of compounds 15b and 16b with para-substituents on
the phenyl group dropped significantly. The same was true by
replacing the pyridyl group with a bicyclic aromatic group such
as the 3-quinolinyl moiety in compound 19b. Substitution on the
methylene unit, including the methyl and trifluoromethyl substit-
uents in compounds 17b and 18b were well tolerated and resulted
in only slightly less potent Hh inhibitors compared to compound
6b. The methylene unit can be omitted attaching the phenyl ring
directly to the nitrogen (20b) producing a slight 2-fold reduction
in inhibitory activity. It was found that 5-membered heterocycles
at the benzylic position (8b, 21b, 22b) are among the best substit-
uents with IC50s in the low double-digit nanomolar range for Hh
pathway inhibition. An additional methyl substituent in the distal
position as in 23b and 25b maintained or even slightly increased
potency whereas a methyl group in the ortho-position resulted in
a compound (24b) with significantly less activity. Aryl substituents
on the methylene unit are vital for excellent potency: Replacement
of the aryl group with an alkyl group, as exemplified by the tetra-
hydrofuranyl substituent in 26b resulted in a compound with only
moderate micromolar inhibitory activity. As observed with com-
References and notes
1. McMahon, A. P.; Ingham, P. W.; Tabin, C. J. Curr. Top. Dev. Biol. 2003,
53, 1.
2. For recent reviews see: (a) Polkinghorn, W. R.; Tarbell, N. J. Nat. Clin. Pract. Oncol.
2007, 4, 295; (b) Rubin, L. L.; de Sauvage, F. J. Nat. Rev. Drug Disc. 2006, 5, 1026;
(c) di Magliano, M. P.; Hebrok, M. Nat. Rev. Cancer 2003, 3, 903.
3. Romer, J. T.; Kimura, H.; Magdaleno, S.; Sasai, K.; Fuller, C.; Baines, H., et al
Cancer Cell 2004, 6, 229.
4. Sanchez, P.; Ruiz i Altaba, A. Mech. Dev. 2005, 122, 223.
5. For recent publications on Hh inhibitors see: (a) Brunton, S. A.; Stibbard, J. H. A.;
Rubin, L. L.; Kruse, L. I.; Guicherit, O. M.; Boyd, E. A.; Price, S. J. Med. Chem. 2008,
51, 1108; (b) Remsberg, J. R.; Lou, H.; Tarasov, S. G.; Dean, M.; Tarasova, N. I. J.
Med. Chem. 2007, 50, 4534.
6. Ksander, G. M.; de Jesus, R.; Yuan, A.; Fink, C.; Moskal, M.; Carlson, E., et al J. Med.
Chem. 2001, 44, 4677.
7. Prashad, M.; Hu, B.; Repic, O.; Blacklock, T. J.; Acemoglu, M. Adv. Synth. Catal.
2001, 343, 461.
8. Jain, R. K.; Kelleher, J.; Peukert, S.; Sun, Y. PCT Patent Application
WO2007120827, 2007.
9. Hh cell assay: This assay measures the end stage of the Hh signaling pathway,
that is, the transcriptional modulation of Gli, using Luciferase as readout (Gli-
Luc assay). Test compounds were prepared for assay by serial dilution in DMSO
and then added to empty assay plates. TM3Hh12 cells (TM3 cells containing Hh-
responsive reporter gene construct pTA-8xGli-Luc, kind gift of Xu Wu, Genomics
Institute of the Novartis Foundation, La Jolla, CA) were cultured in F12 Ham’s/
DMEM (1:1) containing 5% horse serum, 2.5% fetal bovine serum (FBS), and
15 mM Hepes, pH 7.3. Cells were harvested by trypsin treatment, resuspended
in F12 Ham’s/DMEM (1:1) containing 5% FBS and 15 mM Hepes pH 7.3, added to
assay plates and incubated with test compounds for approximately 30 min at
37 °C in 5% CO2. 1 nM Hh-Ag 1.5 (Frank-Kamenetsky, M.; Zhang. X. M.; Bottega,
S.; Guicherit, O.; Wichterle, H.; Dudek, H. et al. J. Biol. 2002, 1, 10) was then
added to assay plates and incubated at 37 °C in the presence of 5% CO2. After
48 h, either Bright-Glo (Promega E2650) or MTS reagent (Promega G258B) was
added to the assay plates and luminescence or absorbance at 492 nm was
determined. IC50 values, defined as the inflection point of the logistic curve,
were determined by non-linear regression of the Gli-driven luciferase
luminescence or absorbance signal from MTS assay vs log10 (concentration)
of test antagonist using the R statistical software package.Smo binding assay:
Smo membranes were prepared from CHO-K1 cells which were stably
transfected with HA-tagged cDNA encoding human or mouse Smo. Test
compounds were prepared for assay by serial dilution in DMSO and then
Table 2
In vitro activity in the Hh cell assay, the Smo mouse (mu), and human (hu) receptor
binding assay and the apoB secretion assay (IC50s, nM)9
CF3
H
N
O
X
N
H
Compound
X
Hh
assay
Smo mu
binding
Smo hu
binding
apoB
assay
Cyclopamine
6b
7b
14b
15b
46
21
100
150
360
1200
8
57
190
93
280
14
120
240
120
CH2-phenyl
CH2-2-pyridyl
CH2-4-pyridyl
2350
CH2-(4-methoxy)-phenyl
H
N
added to empty assay plates. Smo membranes (10 lg total protein) were added
16b
450
360
340
to these assay plates and incubated with 1.5 nM [3H]-Hh-Ag 1.5 in binding
buffer (50 mM Tris–HCl, 10 mM EDTA, and 5 mM magnesium chloride, pH 7.2)
for 48 h at 37 °C. 96 well Unifilter GF/B filter plates (Perkin Elmer) were
prepared by 60 min incubation in binding buffer containing 0.5% w/v
polyethyleneimine (Acros) and 0.1% (w/v) bovine serum albumin (Jackson
Immuno Research) followed by three rinses with 2% beta-hydroxy propyl
cyclodextrin (HPCD, Acros) in distilled water. Assay plates were harvested into
filter plates using a Unifilter-96 cell harvester (Perkin Elmer). Loaded filter
plates were washed three times with 2% HPCD buffer to remove unbound [3H]-
Hh-Ag 1.5, dried in a 60 °C oven for 30 min and then cooled. The bottom of the
O
17b
18b
19b
20b
8b
CH(Me)-phenyl
CH(CF3)-phenyl
CH2-3-quinolinyl
Phenyl
CH2-2-thiazolyl
CH2-2-thienyl
CH2-2-furanyl
S
88
33
1290
41
25
10
29
57
350
82
10
9
32
55
210
130
30
180
1400
140
140
150
21b
22b
7
23
14
plate was sealed, 50 lL of Microscint-O (Perkin Elmer) was added to each well,
23b
17
4
7
170
10
N
S
N
the top of the plate was sealed and the plate was incubated 100 min to
overnight. The amount of bound [3H]-Hh-Ag 1.5 was determined by scintillation
counting on a TopCount (Perkin Elmer). The data were analyzed for saturation
binding using global fitting with Graphpad Prizm software.apoB secretion assay:
This assay monitors inhibition of the microsomal triglyceride transfer protein
(MTP) by measuring Apolipoprotein B (ApoB) production from HepG2 cells by
ELISA. HepG2 cells (ATCC HB-8065) were cultured in MEM media with Earle’s
24b
160
120
O
O
25b
26b
8
6
3700
16,700
salts and
L-glutamine containing 10 mM non-essential amino acids, 100 mM
sodium pyruvate, and 10% fetal bovine serum (complete media). Cells at 80%

S. Peukert et al. / Bioorg. Med. Chem. Lett. 19 (2009) 328–331
331
confluence were harvested with trypsin–EDTA 0.05%, resuspended in complete
media, and 20,000 cells per well were added to assay plates (Corning 3716) and
incubated with test compounds for 16 h at 37 °C in 5% CO2. Concentration of
ApoB secreted by the cells was determined using total human apolipoprotein B
(Apo B) ELISA Assay kit (Alerchek A70102) according to manufacturer’s
instructions, with the following modifications. The HepG2 cell supernatant
diluted sample was added to the ELISA microplate and incubated for 120 min at
room temperature. Washes, development, and termination of assay were
performed according to the manufacturer’s protocol and absorbance was
measured at 450 nm. IC50 values, defined as the inflection point of the logistic
curve, were determined by non-linear regression of the absorbance signal from
ApoB ELISA vs log10 (concentration) of test compound using the R statistical
software package.
was diluted 1:2 with the sample diluent provided in the kit. 100 lL of the