Discovery of Novel Macrocyclic Hedgehog Pathway Inhibitors Acting by Suppressing the Gli-Mediated Transcription

Article abstract of DOI:10.1021/acs.jmedchem.7b01185

A systemic medicinal chemistry campaign was conducted based on a literature hit compound 5 bearing the 4,5-dihydro-2H-benzo[b][1,5]oxazocin-6(3H)-one core through cyclization of two side substituents of the bicyclic skeleton combined with N-atom walking or ring walking and the central ring expansion or extraction approaches, leading to several series of structurally unique tricyclic compounds. Among these, compound 29a was identified as the most potent against the Hedgehog (Hh) signaling pathway showing an IC₅₀ value of 23 nM. Mechanism studies indicated that compound 29a inhibited the Hh signaling pathway by suppressing the expression of the transcriptional factors Gli rather than by interrupting the binding of Gli with DNA. We further observed that 29a was equally potent against both Smo wild type and the two major resistant mutants (Smo D473H and Smo W535L). It potently inhibited the proliferation of medulloblastoma cells and showed significant tumor growth inhibition in the ptch±p53-/- medulloblastoma allograft mice model. Though more studies are needed to clarify the precise interaction pattern of 29a with Gli, its promising in vitro and in vivo properties encourage further profiling as a new-generation Hh signaling inhibitor to treat tumors primarily or secondarily resistant to current Smo inhibitors.

Full text of DOI:10.1021/acs.jmedchem.7b01185

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Discovery of Novel Macrocyclic Hedgehog Pathway Inhibitors
Acting by Suppressing the Gli-mediated Transcription
Gang Liu, Wenjing Huang, Juan Wang, Xiaohua Liu, Jun
Yang, Yu Zhang, Yong Geng, Wenfu Tan, and Ao Zhang
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01185 • Publication Date (Web): 05 Sep 2017
Downloaded from http://pubs.acs.org on September 6, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
Discovery of Novel Macrocyclic Hedgehog Pathway Inhibitors Acting by
Suppressing the Gli-mediated Transcription
7
8
9
Gang Liu,†,∫,§ Wenjing Huang,‡,§ Juan Wang,‡,§ Xiaohua Liu,† Jun Yang,‡ Yu Zhang,‡ Yong Geng,†,∫
Wenfu Tan, ,‡ and Ao Zhang*,†,ǁ,∫
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†CAS Key Laboratory of Receptor Research, and the State Key Laboratory of Drug Research,
Shanghai Institute of Materia Medica (SIMM), Shanghai 201203, China
‡Department of Pharmacology, School of Pharmacy, Fudan University, Shanghai 201203, China
ǁSchool of Life Science and Technology, ShanghaiTech University, Shanghai 201210, China
∫University of Chinese Academy of Sciences, Beijing 100049, China
§These authors contributed equally to this work.
*To whom correspondence should be addressed. For A.Z.: phone: +86ꢀ21ꢀ50806035; fax:
86ꢀ21ꢀ50806035; Eꢀmail: aozhang@simm.ac.cn; For W.T.: phone: +86ꢀ21ꢀ51980039; fax:
86ꢀ21ꢀ51980039; Eꢀmail: wftan@fudan.edu.cn.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 75
1
2
3
4
Abstract.
5
6
7
8
A systemic medicinal chemistry campaign was conducted based on a literature hit compound 5
bearing the 4,5ꢀdihydroꢀ2Hꢀbenzo[b][1,5]oxazocinꢀ6(3H)ꢀone core through cyclization of two side
substituents of the bicyclic skeleton combined with Nꢀatom walking or ringꢀwalking and the central
ring expansion or extraction approaches, leading to several series of structurally unique tricyclic
compounds. Among these, compound 29a was identified as the most potent against the hedgehog
(Hh) signaling pathway showing an IC50 value of 23 nM. Mechanism studies indicated that
compound 29a inhibited the Hh signaling pathway by suppressing the expression of the
transcriptional factors Gli, rather than by interrupting the binding of Gli with DNA. We further
observed that 29a was equally potent against both Smo wild type and the two major resistant mutants
(Smo D473H and Smo W535L). It potently inhibited the proliferation of medulloblastoma cells and
showed significant tumor growth inhibition in the ptch+/ꢀ;p53ꢀ/ꢀ medulloblastoma allograft mice
model. Though more studies are needed to clarify the precise interaction pattern of 29a with Gli, its
promising in vitro and in vivo properties encourage further profiling as a newꢀgeneration Hh
signaling inhibitor to treat tumors primarily or secondarily resistant to current Smo inhibitors.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 3 of 75
Journal of Medicinal Chemistry
1
2
3
4
INTRODUCTION
5
6
7
8
The Hedgehog (Hh) signaling pathway is a major regulator governing cell proliferation and
differentiation, tissue polarity and patterning. It is initiated by binding of its 12ꢀpass transmembrane
receptor, patched (Ptch) with one of the three ligands, sonic hedgehog (Shh), indian hedgehog (Ihh),
and desert hedgehog (Dhh). The membrane protein Smoothened (Smo), structurally similar to
Gꢀprotein coupled receptors (GPCRs), is then released and accumulated in cilia. The ciliary
accumulated Smo can promote activation of transcriptional factor Gli (Gli1, Gli2, and Gli3) by
inhibition of suppressor of fused (Sufu) and protein kinase A (PKA), ultimately leading to target
genes transcription.1,2 It has been long recognized that aberrant activation of the Hh pathway after
birth drives initiation and maintenance of numerous types of cancers, such as basal cellular
carcinoma (BCC) and medulloblastoma.2,3 Molecular mechanisms behind the aberrant activation of
the Hh pathway in cancers include irregular elevation of its ligands in stromal cells or tumor cells,
Ptchꢀinactivating mutations, Smoꢀactivating mutations, Sufuꢀinactivating mutations, as well as
genetic changes in Gli1 or Gli2.2,4
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Addiction of various types of tumors to the hyperactive Hh pathway activity has attracted
intensive effort to develop targeted Hh pathway inhibitors for therapeutic use.5ꢀ7 Smo, a central
effector of the Hh pathway, has emerged as a predominant drug target in the development of small
molecular Hh pathway inhibitors.8ꢀ10 All the current drug candidates in clinical trials for combating
Hhꢀdriven cancers act directly on Smo,10 among which, vismodegib11,12 (2, GDCꢀ0449) from
Genentech, and sonidegib13,14 (3, NVPꢀLDE225) from Novartis were successively approved for
clinical use of locally advanced or metastatic BCC in 2012 and 2015, respectively (Figure 1).
However, tumors harboring aberrant Hh activity by genetic alterations downstream from Ptch as
mentioned above, are conceivably insensitive to these Smo inhibitors, a phenomenon proved by both
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 75
1
2
3
4
5
6
7
8
9
bench and bedside investigations.15ꢀ18 Moreover, like other molecular targeted antiꢀcancer drugs,
acquired resistance is rapidly developed in approximately 20% patients after initial dramatic
response to 2.19,20 This subpopulation of patients was further demonstrated to be refractory to 3 as
well.18 Studies in mice and humans have delineated the molecular basis underlying how tumors
evolved resistance to Smo inhibitors, including additional mutations in Smo, lossꢀofꢀfunction
mutation of Sufu, and amplification of Gli2.21ꢀ25 Hence, the primary and acquired resistance to
current Smo inhibitors highlights the critical need to develop novel nextꢀgeneration Hh inhibitors,
either with distinct chemical structures or with alternative mechanisms of action, particularly those
acting downstream of Smo.26
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Recently, emerging evidence has showed that targeting Hh pathway at the level of its final
effector Gli represents one promising strategy to combat primary and acquired resistance to current
ACS Paragon Plus Environment

Page 5 of 75
Journal of Medicinal Chemistry
1
2
3
4
Smo inhibitors.23,27ꢀ31 Several types of small molecule Gli inhibitors have been developed, such as
5
6
7
the Gli antagonist
ꢀ
2,2'ꢀ((2ꢀ(pyridinꢀ4ꢀyl)dihydropyrimidineꢀ1,3(2H,4H)ꢀdiyl)bis(methylene)
8
9
bis(N,Nꢀ dimethylaniline) (GANT61),32 arsenic trioxide (As2O3),28,33 the bromodomainꢀcontaining
protein 4 (BRD4) inhibitor ꢀ (6S)ꢀ4ꢀ(4ꢀchlorophenyl)ꢀ2,3,9ꢀtrimethylꢀ6Hꢀthieno[3,2ꢀf][1,2,4]
triazolo[4,3ꢀa] [1,4]diazepineꢀ6ꢀacetic acid 1,1ꢀdimethylethyl ester ((+)ꢀJQ1),27 and among others34,35
Unfortunately, these Gli inhibitors generally exhibited low efficacy against the Hh pathway activity
and need further structural optimization. Recently, Schreiber and coꢀworkers36,37 performed a
cellꢀbased highꢀthroughput screening and identified compound 5 (BRD9526)37 bearing the
4,5ꢀdihydroꢀ 2Hꢀbenzo[b][1,5]oxazocinꢀ6(3H)ꢀone bicyclic core 4 that showed an EC50 value of 60
nM in the Hhꢀinduced differentiation of C3H10T1/2 cells and partially lowered the Gli1 expression
at concentrations of 10 ꢁM. Although the cellular potency is weak and its mechanism whether acting
at the level of Gli or downstream of SuFu is not clear, the structural novelty, distinctly different from
the wellꢀknown Hh inhibitors 1-3, encourages us to conduct a systemic structural modification of 5
to generate more potent newꢀgeneration Hh pathway inhibitors. Since the Schreiber group has
demonstrated that, among the eight stereoisomers relevant to the three stereogenic carbons, only the
two C2,C3ꢀcis isomers showed high potency,37 we decide to design several series of compounds
(formula IꢀIII, Figure 1) by cyclization of C2 and C3 of 5 to form an additional piperidinyl ring
sitting on the eightꢀmembered 4,5ꢀdihydroꢀ2Hꢀbenzo[b][1,5]oxazocinꢀ6(3H)ꢀone center (I, n = 0,
Figure 1) or on the nineꢀmembered 2,3,4,5ꢀtetrahydrobenzo[h][1,5]oxazoninꢀ6(7H)ꢀone center (I, n =
1) in cisꢀconfiguration. Compound series II can be viewed from series I through an Nꢀatom walking
process (from 2ꢀ to 3ꢀposition on the ring as shown in red arrow), whereas series III can be viewed
from I through a piperidinyl ringꢀwalking process combined with a centerꢀring contraction from the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 75
1
2
3
4
5
6
7
8
9
eightꢀmembered 1,5ꢀoxazocanꢀ6ꢀone to the sevenꢀmembered 3,4ꢀdihydrobenzo[f][1,4]
oxazepinꢀ5(2H)ꢀone center. Compared to the prototypic bicyclic compound 5, all the new
compounds contain a unique fused tricyclic framework with reduced stereogenic carbons. Herein, we
report our synthesis and pharmacological evaluation of these new compounds as Hh pathway
inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Chemistry. The synthesis of tricyclic benzo[b]pyrido[4,3ꢀg][1,5]oxazocinꢀ7(12aH)ꢀones is shown
in Scheme 1. Starting from 1ꢀbenzylꢀ3ꢀoxopiperidineꢀ4ꢀcarboxylate 6, the cisꢀconfigured
hydroxypiperidine 7 was prepared by following the literature procedure.38 Hydrolysis of 7 followed
by
amidation
with
various
amines
in
the
presence
of
1ꢀ[bis(dimethylamino)methylene]ꢀ1Hꢀ1,2,3ꢀtriazolo[4,5ꢀb]pyridinium 3ꢀoxid hexafluorophosphate
(HATU), 1ꢀhydroxyꢀ7ꢀazabenzotriazole (HOAT) and diisopropyl ethylamine (DIPEA) provided
amides 8a-e in 57ꢀ78% overall yields. Without purification, protection39 of 8a-e via treatment with
tertꢀbutyldimethylsilyl trifluoromethanesulfonate (TBSOTf) afforded the TBSꢀethers 9a-e in 85ꢀ93%
yields. Reduction37 of amides 9a-e with borane dimethylsulfide (BH3.DMS) in THF followed by
coupling with 2ꢀfluoroꢀ3ꢀnitrobenzoic acid under the standard condensation conditions (HATU,
HOAT, DIPEA) delivered amides 11a-e, which were subsequently cyclized37 by treating with CsF to
furnish the key tricyclic intermediates ꢀ (4a,12a)ꢀcisꢀ2,3,4,4a,5,6ꢀhexahydroꢀ1Hꢀbenzo[b]pyrido
[4,3ꢀg][1,5]oxazocinꢀ7(12aH)ꢀones 12a-e in 68ꢀ91% yields. Reduction of 12a-e with Fe powder in
ethyl alcohol and subsequent acylation of the resulting aniline intermediate with
cyclopropanecarbonyl chloride produced the cyclopropylꢀsubstituted amides 14a-e in 76ꢀ92%
overall yields. Debenzylation of 14a-e with Pd(OH)2/C under hydrogen atmosphere gave the key
intermediates 15a-e in 43ꢀ86% yields. Meanwhile, the Nꢀcyclohexanecarbonylꢀsubstituted analog 17
ACS Paragon Plus Environment

Page 7 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
was obtained from 13b in 50% overall yield by following similar reaction procedures as that for
preparation of 15a-e.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2,6,11ꢀTrisubstituted benzo[b]pyrido[4,3ꢀg][1,5]oxazocinꢀ7(12aH)ꢀones 19-22 were synthesized
as described in Scheme 2. Treatment of 15a-b with 4ꢀchlorobenzeneꢀ1ꢀsulfonyl chloride led to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 75
1
2
3
4
5
6
7
8
9
arylsulfonamides 18a-b in 83ꢀ92% yields, which were further converted to final products 19a,b and
19a', respectively, by removal of the OꢀPMB (PMB = paraꢀmethoxybenzyl) group with
2,3ꢀdichloroꢀ5,6ꢀdicyanoꢀ1,4ꢀbenzoquinone (DDQ).37 Compounds 19a and 19a' were generated as a
pair of C4a,C12aꢀcis diastereomeric isomers with opposite optical rotations in about 1:1 ratio, due to
the existence of the Rꢀconfigurated N6ꢀsubstituent. Both isomers were isolated by HPLC and their
absolute configurations were not determined. Similarly, the intermediate 17 was converted to
arylsulfonamides 19c through a similar procedure as that for 19a,b. In the meantime, N6ꢀsubstituted
analogues 20-22 were prepared as well in 77ꢀ89% yields by treating intermediates 15c-e with
4ꢀchlorobenzeneꢀ1ꢀsulfonyl chloride under the aforementioned conditions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of compounds 19-22.a
O
O
O
S
OH
1
N
Cl
b
19a, (-)-isomer
NH
3
from 18a
a
12a
4a
O
15a-b
48%
83-92%
OH
7
N
19a'
, (+)-isomer
O
R
from 18b
OH
19b
41%
OPMB
OPMB
18a,
R
R =
18b
,
O
O
H
O
S
N
O
N
Cl
NH
O
NH
b
a
O
18c
44%
87%
N
N
O
17
OPMB
O
19c
OH
O
O
O
S
N
Cl
NH
a
O
15c-e
77-89%
N
R =
N
N
R
O
O
20
21
22
aReagents and conditions: (a) 4-chlorobenzene-1-sulfonyl chloride, 2,6-lutidine, CH2Cl2, 0 oC, 3
h; (b) DDQ, pH = 7, buffer/CH2Cl2, rt, 12 h.
ACS Paragon Plus Environment

Page 9 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
As shown in Scheme 3, treatment of compound 15e with diverse alkylꢀ or arylꢀsulfonyl
chlorides provided sulfonamides 23a-h in 73ꢀ96% yields. Meanwhile, treating 15e with acyl
7
8
9
chlorides or isocyanates afforded substituted benzamides 23i-k in 74ꢀ93% yields.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Synthesis of compounds 23a-k.a
O
H
N
NH
O
O
O
O
O
O
O
O
O
S
S
S
N
R =
S
N
N
a
b
c
O
O
O
O
O
O
15e
OMe
OMe
S
S
Cl
NHAc
Cl
a
73-96%
d
e
f
i
O
O
O
O
R
O
O
N
S
S
S
N
Cl
N
NH
11
2
Cl
O
h
g
F
CF3
O
7
O
N
N
N
O
H
H
O
F
j
k
23a-k
aReagents and conditions: (a) for 23a-h, sulfonyl chlorides, 2,6-lutidine, CH2Cl2, 0 oC, 3 h,
73-96%; for 23i, acyl chlorides, 2,6-lutidine, CH2Cl2, 0 oC, 3 h, 93%; for 23j-k, isocynates,
CH2Cl2, rt, 3 h, 74-75%.
The synthesis of the 2,3,4,4a,5,6ꢀhexahydroꢀ1Hꢀbenzo[b]pyrido[3,4ꢀg][1,5]oxazocinꢀ
7(12aH)ꢀones 29a-g is illustrated in Scheme 4. The cisꢀconfigurated hydroxypiperidine 25 was
prepared according to the literature procedure.38 Hydrolysis of 25 followed by amination with
(4ꢀmethoxyphenyl)methanamine and protection of the hydroxyl with TBSOTf delivered amide 26 in
49% overall yield. The N3,N6,C11ꢀtrisubstituted (4a,12a)ꢀcis 2,3,4,4a,5,6ꢀhexahydroꢀ1Hꢀ
benzo[b]pyrido[3,4ꢀg][1,5]oxazocinꢀ7(12aH)ꢀone 27 was obtained through BH3.DMS reduction of
the amido moiety, condensation with 2ꢀfluoroꢀ3ꢀnitrobenzoic acid and CsFꢀassisted cyclization by
following similar reaction procedures as that for preparation of 12a-e. Reduction of 27 with Fe
powder and subsequent amidation with cyclopropanecarbonyl chloride produced the intermediate 28
in 79% overall yield. Removal of the tertꢀbutoxycarbonyl (Boc) group with TFA followed by treating
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 75
1
2
3
4
5
6
7
8
9
with various arylsulfonyl chlorides afforded target compounds 29a-g as racemates in 66ꢀ82% yields.
The formation of the tricyclic skeleton was confirmed by NMR correlations between Hꢀ12a (δH =
4.60 ppm) and Cꢀ12a (δC = 69.44 ppm) as well as the correlation between Hꢀ4a (δH = 1.15 – 1.10
ppm) and Cꢀ4a (δC = 15.98 ppm) in the heteronuclear single quantum coherence spectrum (HSQC) of
compound 29a (See SI).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 11 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In addition, the racemic 28 was optically resolved by preparative chiralꢀHPLC to provide 28-1
and 28-2 as the (+)ꢀ and (ꢀ)ꢀisomers, respectively. They were subsequently converted to the optically
pure compounds 29a-1 ((+)ꢀisomer) and 29a-2 ((ꢀ)ꢀisomer), respectively (chiralꢀHPLC report, see
SI). To determine the absolute configuration of 29a-1 and 29a-2, we prepared camphorsulfonamide
29h-1 by treating (+)ꢀenantiomer 28-1 with D(+)ꢀ10ꢀcamphorsulfonyl chloride. The absolute
configuration of 29h-1 was successfully secured by singleꢀcrystal Xꢀray diffraction analysis as
(4aR,12aS)ꢀisomer (CCDC: 1564315). Accordingly, compounds 29a-1 and 29a-2 were deduced to be
(4aR,12aS)ꢀ and (4aS,12aR)ꢀenantiomer, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 75
1
2
3
4
5
6
7
8
9
As shown in Scheme 5, compounds 32a-c bearing the (4a,13a)ꢀcisꢀ1,2,3,4,4a,5,6,13aꢀ
octahydrobenzo[h]pyrido[4,3ꢀb][1,5]oxazoninꢀ7(8H)ꢀone scaffold were prepared using compound 26
as the starting material. Reduction of 26 with BH3.DMS37 and condensation with
2ꢀ(2ꢀfluoroꢀ5ꢀnitrophenyl)acetic acid followed by cyclization of the resulting intermediate with the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
assistance
of
TBAF
furnished
1,2,3,4,4a,5,6,13aꢀoctahydrobenzo[h]pyrido[4,3ꢀb][1,5]
oxazoninꢀ7(8H)ꢀone 30 in 19% overall yield. Subsequent reduction of the nitro group with Fe
powder followed by acylation of the resulting aniline with cyclopropanecarbonyl chloride gave
compound 31 in 72% overall yield. Removal of the NꢀBoc group followed by treating with various
arylsulfonyl chlorides yielded target compounds 32a-c in 71ꢀ83% yields. Similarly, the isomeric
(4a,13a)ꢀcisꢀ1,2,3,4,4a,5,6,13aꢀoctahydrobenzo[h]ꢀpyrido[3,4ꢀb][1,5]oxazoninꢀ7(8H)ꢀones
35a,b
were prepared by following a set of similar reaction procedures starting from compound 9e as that
for preparation of compounds 32a-c.
As shown in Scheme 6, compound 37 bearing the 3,4,12,12aꢀtetrahydroꢀ1Hꢀbenzo[f]
pyrazino[2,1ꢀc][1,4]oxazepinꢀ6(2H)ꢀone scaffold was prepared by condensation40 of tertꢀbutyl
3ꢀ(hydroxymethyl)piperazineꢀ1ꢀcarboxylate (36) with 2ꢀfluoroꢀ3ꢀnitrobenzoic acid under
ACS Paragon Plus Environment

Page 13 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HATU/Et3N followed by NaHꢀpromoted cyclization41 in 52% overall yield. Reduction of the nitro
group with Fe powder followed by amidation with cyclopropanecarbonyl chloride provided amide 38
in 69% overall yield. Removal of the NꢀBoc group with TFA followed by sulfonamidation with
diverse arylsulfonyl chlorides delivered target compounds 39a-c in 82ꢀ87% yields.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Structure-Activity Relationship (SAR) Study. All the new compounds were first evaluated for
their ability to inhibit the Hh signaling pathway by dual luciferase reporter assays with light II cells,
which were NIHꢀ3T3 cells stably transfected with a Gliꢀresponsive firefly luciferase reporter and
Renillaꢀluciferase expression vector.42 IC50 values were tested only for compounds showing
inhibitory rates greater than 50% at the concentration of 1 µM. Compounds 19a, 19a', 29a-1, and
29a-2 are single stereoisomer, whereas the rest of tested compounds were racemic mixtures.
As shown in Table 1, we first evaluated benzo[b]pyrido[4,3ꢀg][1,5]oxazocinꢀ7(12aH)ꢀones
19-22 bearing N2ꢀ(paraꢀchlorophenyl) sulfonyl and C11ꢀcyclopropanecarboxamido moieties with
diverse N6ꢀsubstituents. Unfortunately, the two diastereoisomers 19a and 19a' bearing the same
Rꢀconfigured Nꢀ(1ꢀhydroxypropanꢀ2ꢀyl) moiety as the N6ꢀsubstituent as that in 5 showed rather weak
inhibitory effects at the concentration of 1 µM against the Hh signaling pathway. The 2ꢀhydroxyethyl
substituted analog 19b also exhibited negligible activity. Replacement of the C11ꢀsubstituent with
cyclohexanecarboxamido moiety afforded compound 19c showing low inhibition as well. Again,
rather weak inhibitory effect was observed from 21 bearing pyridineꢀ3ꢀmethyl as the N6ꢀsubstituent.
To our delight, the Nꢀmethylpiperidinꢀ4ꢀyl substituted analog 20 and N6ꢀ(paraꢀmethoxybenzyl)
substituted analog 22 showed appreciable inhibition against the Hh signaling pathway with an IC50
value of 344 nM and 304 nM, respectively. Both compounds are slightly more potent than the
prototypic compound 5 (IC50 = 383 nM) in our test system.
Table 1. Hh pathway inhibition of compounds by dual luciferase reporter assays in light II cells.a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 75
1
2
3
4
5
6
7
8
O
O
O
S
Cl
1
N
NH
3
12a
4a
11
O
7
N
9
R
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Gliꢀluc reporter
Gliꢀluc reporter
R
R
1 ꢁM (%)
IC50 (nM)
1 ꢁM (%)
IC50 (nM)
OH
34 ± 16
>1000
61 ± 23
28 ± 8
344 ± 6
19a
20
OH
40 ± 12
25 ± 12
>1000
>1000
>1000
N
19a'
21
97 ± 8
304 ± 59
O
19b
22
67 ± 22
93 ± 6
383 ± 74
39 ± 2.0
5
2
32 ± 9
>1000
19c
aThe IC50 values are shown as the mean ± SD (nM) from three separate experiments.
Since compound 22 showed an inhibitory potency compatible to that of 5, we then retained
paraꢀmethoxybenzyl moiety as the N6ꢀsubstituent together with cyclopropanecarboxiamo moiety as
the C11ꢀsubstituent, and screened various sulfonyl or acyl groups as the N2ꢀsubstituent. As shown in
Table 2, methanesulfonyl substituted compound 23a showed much reduced inhibitory effect,
compared to compound 22. Likewise, both cyclohexanesulfonyl and 4ꢀmethylpiperazineꢀ1ꢀsulfonyl
ACS Paragon Plus Environment

Page 15 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
analogs 23b and 23c showed inappreciable potency. However, moderate to good potency was found
from arylsulfonyl substituted compounds 23dꢀf. Compounds 23d and 23f bearing two substituents
showed lower potency, whereas the (4ꢀacylamino)phenylsulfonyl analog 23e showed an enhanced
potency with an IC50 value of 42 nM, which is 7ꢀ and 9ꢀfold more potent than 22 and 5, respectively.
Compound 23e is also nearly equally potent to the clinically prescribed drug 2 (39 nM, Table 1). The
result indicated that our tricyclic framework well maintains the interaction with the pathway as that
of the bicyclic framework 5, and an Nꢀarylsulfonyl substituent on the piperidine ring seems optimal.
Thus analysis is further validated from N2ꢀheteroarylsulfonyl substituted 23g and 23h that exhibited
good potency as well, especially 1ꢀmethylꢀ1Hꢀpyrazoleꢀ3ꢀsulfonamide 23h having a high potency of
94 nM, which is 4ꢀfold more potent than both 22 and 5.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Hh pathway inhibition of compounds by dual luciferase reporter assays in light II cells.a
Gliꢀluc reporter
Gliꢀluc reporter
R
R
1 ꢁM (%)
IC50 (nM)
1 ꢁM (%)
IC50 (nM)
45 ± 16
42 ± 0.9
>1000
113 ± 3
74 ± 8
167 ± 88
23a
23g
23h
>1000
>1000
94 ± 25
>1000
23b
23c
45 ± 6
22 ± 13
23i
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 75
1
2
3
4
5
6
102 ± 5
766 ± 134
100 ± 5
292 ± 104
23j
7
8
9
23d
23e
23f
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
105 ± 5
54 ± 16
42 ± 7.0
100 ± 5
26 ± 6.0
23k
>1000
ꢀ
ꢀ
-
aThe IC50 values are shown as the mean ± SD (nM) from three separate experiments.
Completely no inhibitory effect was observed from 4ꢀchlorobenzoyl substituted analog 23i,
indicating both the oxygen atoms in the sulfonyl moiety of either 23e or 23h other than the only
oxygen atom in the acyl 23i involved in the compoundꢀtarget interaction model. Further, compounds
23j
and
23k
bearing
N2ꢀ(4ꢀ(trifluoromethyl)phenyl)carbamic
moiety
and
N2ꢀ(2,4ꢀdifluorophenyl)carbamic moiety, respectively, displayed good potency, especially compound
23k showing an IC50 value of 26 nM that is more than 10ꢀfold more potent than the prototypic
compound 5 and even 1.5ꢀfold more potent than the clinically prescribed Hh inhibitor 2. This result
highlights the critical contribution of the piperidine Nꢀsubstituent in our tricyclic skeleton and both
arylsulfonyl and arylacyl bearing appropriate substituents are well tolerant.
Compounds 29a-g can be viewed as isomers of 22e, 23d-e, 23g-h and 23j-k, respectively, with
the N2ꢀatom walking to N3 position. By taking advantage of the result in Table 1 and Table 2, we
directly elected the piperidine Nꢀsubstituents showing good potency to the isomeric tricyclic
framework. Gratifyingly, this small series of compounds retained good to high potency against the
Hh signaling pathway, especially compounds 29a and 29b showing IC50 values of 23 and 33 nM,
respectively (Table 3). Other Nꢀsubstituted analogues (29c-d, 29f-g) showed much less potency,
except the 5ꢀchlorothiopheneꢀ2sulfonyl 29e which retained high potency of 46 nM. Since racemic
29a showed the highest potency, we then evaluated its two enantiomeric isomers. The result shows
ACS Paragon Plus Environment

Page 17 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
that the stereogenic character plays a critical role on the Hh activity. The (+)ꢀenantiomer 29a-1 with
4aR,12aSꢀconfiguration retains the high potency (12 nM), whereas the (ꢀ)ꢀenantiomer 29a-2 with
4aS,12aRꢀconfiguration is nearly 17ꢀfold less potent (200 nM). These results may be roughly
rationalized based on the Xꢀray structure of compound 29h-1, where the phenyl and piperidinyl rings
are bended toward each other to form an appropriate dihedral angle, which can be adjusted by both
the stereogenic configuration of the piperidine ring (upside or downside of the central ring) and the
location of the piperidine Nꢀatom, together with its substituent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Hh pathway inhibition of compounds by dual luciferase reporter assays in light II cells.a
Gliꢀluc reporter
Gliꢀluc reporter
R
R
1 ꢁM (%)
IC50 (nM)
1 ꢁM (%)
IC50 (nM)
97 ± 9
117 ± 15
92 ± 0.8
23 ± 9
87 ± 1
108 ± 9
68 ± 9
145 ± 47
29d
29e
29a
29a-1, (+)ꢀisomer
12 ± 1
46 ± 4
>1000
29a-2, (ꢀ)ꢀisomer
200 ± 44
29f
107 ± 15
82 ± 18
33 ± 17
87 ± 9
540 ± 79
29b
29g
773 ± 268
_
_
_
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 75
1
2
3
4
29c
5
6
7
aThe IC50 values are shown as the mean ± SD (nM) from three separate experiments.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compounds 32a-c represent a subseries of ringꢀexpanded analogs bearing a nineꢀmembered
1,5ꢀoxazonanꢀ6ꢀone structural motif along with the cyclopropanecarboxamido moiety moving
downside to C10. Although bearing similar substituents at N3ꢀ, N6ꢀ and C10 as that in 29a, 29d and
29e, these new compounds showed negligible potency against the Hh pathway. Similarly,
compounds 35a,b are the ringꢀenlarged analogs of 22 and 23h, and both showed low inhibitory
effects with IC50 values of 502 nM and >1ꢁM, respectively (Table 4). We envisioned that in addition
to the location of C10ꢀsubstituent, the larger central ring significantly altered the dihedral angle
between the phenyl and piperidnyl of the tricyclic skeleton and subsequently impeded the interaction
patterns.
Table 4. Hh pathway inhibition of compounds by dual luciferase reporter assays in light II cells.a
Gliꢀluc reporter
Gliꢀluc reporter
R
R
IC50 (nM)
1 ꢁM (%)
IC50 (nM)
1 ꢁM (%)
36 ± 23
>1000
60 ± 11
502 ± 92
>1000
32a
32b
35a
35b
ꢀ15 ± 21
>1000
16 ± 3
ACS Paragon Plus Environment

Page 19 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
_
_
_
65 ± 5
1620 ± 43
32c
8
9
aThe IC50 values are shown as the mean ± SD (nM) from three separate experiments.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 5. Hh pathway inhibition of compounds by dual luciferase reporter assays in light II cells.a
Gliꢀluc reporter
Gliꢀluc reporter
R
R
1 ꢁM (%)
IC50 (nM)
1 ꢁM (%)
IC50 (nM)
35 ± 26
>1000
80 ± 11
482 ± 27
39a
39b
39c
58 ± 7
>1000
_
_
_
aThe IC50 values are shown as the mean ± SD (nM) from three separate experiments.
To further determine the effect of the topology of the tricyclic framework on the Hh pathway
activity, compounds 39a-c bearing the sevenꢀmembered 3,4ꢀdihydrobenzo[f][1,4]oxazepinꢀ
5(2H)ꢀone center were designed and tested (Table 5). As mentioned earlier, these compounds can be
viewed as the central ringꢀcontracting and the piperidineꢀring moving (on the central ring) analogs of
19a-22. Disappointingly, these compounds showed weak inhibitory effects at the concentration of 1
ꢁM, and only compound 39c bearing a N2ꢀ(5ꢀchloroꢀthiophenꢀ2ꢀyl)ꢀsulfonyl substituent showed a
moderate IC50 value of 482 nM.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 75
1
2
3
4
Table 6. Cytotoxicity and hERG Inhibition of Potent Compounds
5
6
7
IC50
8
9
Compound
Gliꢀluc reporter
assay
squamous carcinoma
KB cell
hERG
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
23 ± 9
> 20 µM
> 20 µM
29a
Overall Safety Profile of Compound 29a. From the SAR analysis above, compound 29a,
bearing the tricyclic center ꢀ (4a,12a)ꢀcis 1,2,3,4,4a,5,6,12aꢀoctahydroꢀ7Hꢀbenzo[b]pyrido[3,4ꢀg][1,5]
oxazocinꢀ7ꢀone turns out to be the most potent against the Hh signaling pathway showing an IC50
values of 23 nM. It is up to 17ꢀfold more potent than the prototypic compound 5 and 1.7ꢀfold more
potent than the clinically prescribed Hh pathway inhibitor drug 2. Although the (+)ꢀisomer 29a-1 is
slightly more potent than both the racemic and the (ꢀ)ꢀisomer, it was difficult to prepare in large scale.
Therefore, to quickly gain the developmental value of the tricyclic skeleton as the newꢀgeneration
Hh inhibitor, racemic 29a was selected for further evaluation. First, to exclude the chemical structure
relevant cytotoxicity and cardiac toxicity, we tested the inhibitory effects of 29a against the
proliferation of squamous carcinoma KB cells and against the hERG channel.42 As shown in Table 6,
29a showed no significant effects against the proliferation of KB cell (IC50 > 20 ꢁM). This
compound also has no cardiac toxicity potential (IC50 > 20 ꢁM), indicating its overall safety suitable
for further profiling.
Compound 29a Inhibited the Hh Pathway Activity. As shown in Figure 2A, compound 29a
was found to inhibit the Gliꢀluciferase activity stimulated by ShhN conditioned medium (ShhN CM)
in a doseꢀdependent manner as revealed by the normalized Gliꢀluciferase activity (the relative firely
and renilla luciferase values are separately shown in supplementary Figure S2). Further, we
examined the effects of 29a on the mRNA expression of Gli1 and ptch1, two Gli transcriptional
target genes which frequently serve as readout of the Hh signaling pathway activity.31 As shown in
ACS Paragon Plus Environment

Page 21 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figures 2B and 2C, exposure of compound 29a to NIHꢀ3T3 cells or C3H10T1/2 cells was found to
significantly inhibit the expression of Gli1 mRNA and ptch1 stimulated by ShhN CM, confirming
that 29a is indeed a potent inhibitor of the Hh signaling pathway. It is of note that compared to the
higher potency of 29a than 2 in the luciferase reporter assay, the inhibitory effect of 29a in the Gli
transcriptional experiment is somewhat lower than that of 2. The potency discrepancy likely reflected
that the inhibitory activity of 29a in the latter experiment plateaus with about 30% Gli reporter
activity remaining, whereas compound 2 can inhibit Hh pathway activity completely.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C
Control CM
ShhN CM
1.8
2
1.8
1.6
1.4
1.2
1
ptch1
Gli1
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0.8
0.6
0.4
0.2
0
0
-
-
-
-
+
-
2 (100 nM)
-
-
-
-
-
+
2 (100 nM)
-
+
29a (300 nM)
+
29a (300 nM)
-
Figure 2. Inhibition of 29a on the Hh activity. A, dose curve of 29a on the Gliꢀluciferase activity of light II cells after
treated with multiple concentrations of 29a. B-C, representative results of 29a on the mRNA expression of Gli target
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 75
1
2
3
genes Gli1 and ptch1 in NIHꢀ3T3 cells (B) and C3H10T1/2 cells (C). After various treatments as indicated for 36 h, cells
4
5
6
7
were collected for RTꢀqPCR analysis of the Gli1 and ptch1 mRNA expression. The results were expressed as the mean ±
SD.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 29a Selectively Inhibited the Hh Signaling Pathway. To exclude the possibility that
29a inhibited the Hh pathway activity due to its nonꢀspecific offꢀtarget effects, we used the dual
luciferase reporter assay to examine its effect on two other unrelated transcriptional factors NFꢀκB in
response to TNFꢀα, and Tꢀcell factor (TCF)/lymphoid enhancer factor (LEF) stimulated by
prostaglandin E2 (PGE2).42,43 As shown in Figure 3A, at the concentration of 1 ꢁM, a concentration
completely inhibiting the Hh pathway activity, 29a had no inhibitory effect on the NFꢀκB
transcriptional activity, whereas its specific inhibitor, (E)ꢀ3ꢀtosylacrylonitrile (BAY11ꢀ7082)
significantly suppressed the NFꢀκB transcriptional activity. Moreover, compound 29a at the
concentration of 1 ꢁM exhibited only minor inhibitory effect on the TCF/LEF activity in response to
PGE2 (Figure 3B), whereas its specific inhibitor (Nꢀ[2ꢀ((pꢀbromocinnamyl)ꢀamino)ethyl]ꢀ
5ꢀisoquinolinesulfonamide dihydrochloride salt (H89) exhibited significant inhibition. Hence, these
data suggested that 29a selectively inhibited the Hh signaling pathway. The target specificity of 29a
was further investigated by a kinase selectivity profiling against a panel of 468 kinases (including
403 nonꢀmutated kinases) at 1.0 ꢁM using the DiscoveRx screening platform, and 29a showed no
significant inhibition against all these kinases (Figure 3C).
ACS Paragon Plus Environment

Page 23 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Selective inhibition of 29a on the Hh signaling pathway. A, NFꢀκB activity in HEK293 cells after various
treatments as indicated for 24 h were measured by dual luciferase reporter analysis. B, LS174T cells after various
treatments as indicated for 6 h were collected for dual luciferase reporter analysis of TCF/LEF activity. The results were
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 75
1
2
3
expressed as the mean ± SD. C, KinomeScan kinase selectivity profile for 29a at a concentration of 1.0 ꢁM against a
4
5
diverse panel of 468 kinases by DiscoveRx.
6
7
8
9
Compound 29a Inhibited the Hh Signaling Pathway Activity Provoked by Gli1 and Gli2,
Respectively. Having well characterized the inhibitory effect of 29a on Hh signaling pathway, we
continued to map its molecular target. Since Sufu negatively controls Hh signaling pathway by
binding and sequestering the Gli transcriptional factor in cytoplasm,44 we then limited the Sufu
expression in NIHꢀ3T3 cells by lentivirus harboring Sufu short hairpin RNA (shRNA) (Figure 4A),
and determined whether 29a acts downstream or upstream of Sufu. The reverse transcription and
quantitative polymerase chain reaction (RTꢀqPCR) analysis showed that limiting the Sufu expression
abundantly increased the mRNA expression of Gli1 and ptch1 (Figure 4B), whereas treatment with
29a significantly inhibited the Gli1 and ptch1 mRNA expression (Figure 4B), a phenomenon similar
to that of BRD4 inhibitor (+)ꢀJQ1 that inhibits the Hh pathway by epigenetically targeting the
transcriptional factor Gli.27 Meanwhile, the Smo inhibitor 2 failed to reduce the mRNA expression of
Gli1 and ptch1 (Figure 4B). To exclude the possible offꢀtarget effect, we tested 29a against BRD4
and no significant inhibition was observed at the concentration of 10 ꢁM (data not shown). These
data clearly suggested that 29a inhibited the Hh signaling pathway by acting downstream of Sufu
rather than Smo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 25 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
shRNA control
Sufu shRNA
0.6
B
1.2
Gli1
ptch1
A
1
0.5
0.4
0.3
0.2
0.1
0
shRNA control
Sufu shRNA
0.8
0.6
0.4
0.2
0
Sufu
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
GAPDH
-
-
-
-
-
-
-
+
-
-
-
-
-
-
2 (100 nM)
-
-
+
-
2 (100 nM)
+
-
-
(+)-JQ1 (1 µM)
+
-
-
(+)-JQ1 (1 µM)
29a (300 nM)
+
29a (300 nM)
-
-
+
-
-
C
D
Agonist
IC50 (nM)
23
ShhN CM
Exogenous expression
of Gli1
29.9
Exogenous expression
of Gli2
24.9
Log [29a] nM
Figure 4. Inhibitory effects of 29a on the Hh signaling pathway. A, western blot analysis of the expression of Sufu in
NIHꢀ3T3 cells after infected with lentivirus harboring shRNA control or Sufu shRNA; B, effect of 29a on the expression
of Gli target genes Gli1 and ptch1 mRNA expression in the NIHꢀ3T3 cells infected with Sufu shRNA; C, effect of 29a
on the Gli luciferase activity initiated by ectopic expression of Gli1 and Gli2 plasmids, respectively. The results were
expressed as the mean ± SD; D, IC50 values of 29a inhibiting the Gliꢀluciferase activity provoked by ShhN CM, ectopic
expression of Gli1 and Gli2, respectively.
Having characterized 29a acting downstream of Sufu, we then set out to determine whether 29a
can inhibit the Hh pathway activity initiated by Gli, the last node of the Hh signaling pathway. As
Gli1 and Gli2 are mainly responsible for stimulating the Hh pathway activity, we analyzed the effect
of 29a on the Hh pathway activity provoked by forced expression of Gli1 and Gli2 in light II cells,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 75
1
2
3
4
5
6
7
8
9
respectively. As shown in Figure 4C, 29a significantly suppressed the Gliꢀluciferase activity
stimulated by either ectopic expression of Gli1 or Gli2 in a doseꢀdependent manner, whereas
Smoꢀtargeting inhibitor 2 had no effects (Figure 4C; the relative firely and renilla luciferase values
are separately shown in supplementary Figure S3A, B). Meanwhile, the IC50 values of 29a to inhibit
the Gliꢀluciferase activity provoked by artificially expressed Gli1 or Gli2 were similar (23ꢀ30 nM) to
those of inhibiting the Gliꢀluciferase activity in response to ShhN CM (Figure 4D). Taken together,
these data clearly demonstrated that 29a potently inhibited the Hh pathway activity provoked by Gli.
Compound 29a had No Influence on the DNA Sequence Binding Ability of Gli, but
Decreased the Protein Expression of Gli1 and Gli2. As we have showed that 29a inhibited the Hh
pathway activity provoked by ectopic expression of Gli1 or Gli2, we decide to further clarify its
molecular mechanism. First, we tested the effect of 29a on the binding ability of transcriptional
factor Gli1 to the DNA sequence of its transcriptional target genes, one of the most critical steps for
initiating the transcriptional activity of Gli. In the electrophoretic mobility shift assay (EMSA) using
biotinꢀlabeled Gliꢀbinding sequence in the promoter of Gli target gene twist1,45 compound 29a
exhibited no inhibitory effect on the DNA binding ability of Gli1, at the concentration of 1 uM that
completely suppressed the Hh activity in response to ShhN CM (Figure 5A; line 6 compared to line
2). However, different concentrations of nonꢀlabeled Gli binding sequence completely inhibited the
Gli1 binding to the DNA sequence (Figure 5A; lanes 4 and 5 compared to lane 2, respectively).
These results suggested that the inhibitory effect of 29a on the Hh pathway activity was not caused
by interruption of the binding of Gli to the DNA sequence of its target genes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 27 of 75
Journal of Medicinal Chemistry
1
2
3
4
1
2
3
4
5
6
A
line Nuclear protein
probe
5
6
7
1
Bio-Gli BS
-
8
9
2
3
4
5
NIH-3T3-Gli1
293T
Bio-Gli BS
Bio-Gli BS
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NIH-3T3-Gli1
NIH-3T3-Gli1
Bio-Gli1+50×Cold-Gli BS
Bio-Gli1+100×Cold-Gli BS
Bio-Gli1 BS+ 29a (1 µM)
NSB
6
NIH-3T3-Gli1
Arrow: DNA binding proteins
NSB: Non-Specific Binding
P: Free Biotin Labeled Probe
BS: binding sequence
P
1, 29a: 0 µM
2, 29a: 0.05 µM
1.5
3, 29a: 0.5 µM
4, 29a: 1.0 µM
1.5
B
29a (µM)
control
0.05
0.5
1.0
Gli1
Gli2
1
0.5
0
1
***
***
**
***
***
***
0.5
0
GAPDH
1 2 3 4
1 2 3 4
Gli1 -Flag
29a (nM)
1.2
C
GFP
1, 29a: 0 nM
1
0.8
0.6
0.4
0.2
0
*
0
50 500 1000
**
***
2, 29a: 50 nM
3, 29a: 500 nM
4, 29a: 1000 nM
Flag
GAPDH
1 2 3 4
Gli2 -HA
1.2
0.8
GFP
D
29a (nM)
1, 29a: 0 nM
2, 29a: 50 nM
*
0
50 500 1000
***
***
HA
GAPDH
3, 29a: 500 nM
4, 29a: 1000 nM
0.4
0
1 2 3 4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 75
1
2
3
4
5
6
7
8
9
Figure 5. Effects of 29a on the DNA sequence binding ability of Gli, and the protein expression of Gli1 and Gli2. A,
effect of 29a on the DNA binding ability of Gli1 examined by EMSA; BꢀD, Representative immunoblots and
semiꢀquantitative analysis of the effect of 29a on the expression of endogenous Gli1 and Gli2 proteins (B) and
artificially expressed Gli1 (C) and Gli2 (D) in the NIHꢀ3T3 cells. Statistical analysis using Student’s Tꢀtest (*, P < 0.05;
**, P < 0.01; ***, P < 0.001). Cells after subjected to 29a at indicated concentrations for 24 h were collected for western
blot analysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Further, we examined whether 29a influences the expression of Gli proteins. As shown in Figure
5B, exposure of NIHꢀ3T3 cells to different concentrations of 29a doseꢀdependently suppressed the
endogenous expression of Gli1 and Gli2. It should be noted that the endogenous Gli1 expression in
NIHꢀ3T3 cells might be caused by basal activity of both canonical and nonꢀcanonical Hh pathway
activity as the cells were cultured5,46 with 10% FBS (fetal bovine serum) when exposed to 29a.
Moreover, using the antibodies targeting the Tag of respective plasmids, we found that 29a also
reduced the ectopic expression of Gli1 (Figure 5C) and Gli2 (Figure 5D) in NIHꢀ3T3 cells.
Collectively, these observations indicated that 29a inhibited the Hh pathway activity by a mechanism
not involving disruption of DNA binding, but likely involving reduction of the Gli proteins
expression, including Gli1 and Gli2.
Compound 29a Significantly Suppressed the Hh Pathway Activity Provoked by Smo
Mutants Which Cause Resistance to Current Smo Inhibitors. Smo mutations in its ligand
binding pocket or in pivotal residues that ensuring Smo receptor autoinhibition are the predominant
molecular mechanisms, responsible for the resistance to current Smo inhibitors.23 Having profiled
that compound 29a inhibited the Hh signaling pathway by reducing the expression of its final
effectors Glis, we envisioned that compound 29a may inhibit the Hh pathway activity provoked by
Smo mutants, especially Smo D473H and Smo W535L, two key Smo resistant mutations observed
clinically.23 To this aim, we separately transfected three distinct Smo plasmids, including Smo wild
type and the two Smo mutants D473H and W535L into light II cells and analyzed the effect of
ACS Paragon Plus Environment

Page 29 of 75
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compound 29a on the Hh pathway activity provoked by these ectopic plasmids. As shown in Figure
6AꢀC (the relative firely and renilla luciferase values are separately shown in supplementary Figure
S4AꢀC), compound 29a robustly inhibited the Gliꢀluciferase activity provoked by all these three
distinct Smo plasmids in a doseꢀdependent manner. Meanwhile, the IC50 values revealed that
compound 29a exhibited statistically similar potency in inhibiting the Gliꢀluciferase activity
provoked by Smo wild type (35.72 nM), Smo D473H (37.46 nM), and Smo W535L (39.81 nM),
respectively (Figure 6D). Hence, these data supported the potential value of compound 29a to
combat the resistance to current Smo inhibitors caused by distinct Smo mutations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Suppression of 29a on the Hh pathway activity provoked by Smo wild type and mutants. A, effect of 29a
on the Gli luciferase activity initiated by ectopic expression of wild type Smo plasmid. B, effect of 29a on the Gli
luciferase activity initiated by ectopic expression of Smo D473H plasmid. C, effect of 29a on the Gli luciferase activity
initiated by ectopic expression of Smo W535L plasmid. D, IC50 values of 29a inhibiting the Gliꢀluciferase activity
provoked by Smo wild type, Smo D473H, and Smo W535L plasmids, respectively. Light II cells transfected with Smo
wild type, SmoW539L or SmoD473H plasmids were exposed to 29a for 36 h, and then were harvested for dual luciferase
ACS Paragon Plus Environment