Design, Synthesis, and Structure-Activity Relationship of Tetrahydropyrido[4,3-d]pyrimidine Derivatives as Potent Smoothened Antagonists with in Vivo Activity

Article abstract of DOI:10.1021/acschemneuro.7b00153

Medulloblastoma is one of the most prevalent brain tumors in children. Aberrant hedgehog (Hh) pathway signaling is thought to be involved in the initiation and development of medulloblastoma. Vismodegib, the first FDA-approved cancer therapy based on inhibition of aberrant hedgehog signaling, targets smoothened (Smo), a G-protein coupled receptor (GPCR) central to the Hh pathway. Although vismodegib exhibits promising therapeutic efficacy in tumor treatment, concerns have been raised from its nonlinear pharmacokinetic (PK) profiles at high doses partly due to low aqueous solubility. Many patients experience adverse events such as muscle spasms and weight loss. In addition, drug resistance often arises among tumor cells during treatment with vismodegib. There is clearly an urgent need to explore novel Smo antagonists with improved potency and efficacy. Through a scaffold hopping strategy, we have identified a series of novel tetrahydropyrido[4,3-d]pyrimidine derivatives, which exhibited effective inhibition of Hh signaling. Among them, compound 24 is three times more potent than vismodegib in the NIH3T3-GRE-Luc reporter gene assay. Compound 24 has a lower melting point and much greater solubility compared with vismodegib, resulting in linear PK profiles when dosed orally at 10, 30, and 100 mg/kg in rats. Furthermore, compound 24 showed excellent PK profiles with a 72% oral bioavailability in beagle dogs. Compound 24 demonstrated overall favorable in vitro safety profiles with respect to CYP isoform and hERG inhibition. Finally, compound 24 led to significant regression of subcutaneous tumor generated by primary Ptch1-deficient medulloblastoma cells in SCID mouse. In conclusion, tetrahydropyrido[4,3-d]pyrimidine derivatives represent a novel set of Smo inhibitors that could potentially be utilized to treat medulloblastoma and other Hh pathway related malignancies.

Full text of DOI:10.1021/acschemneuro.7b00153

Subscriber access provided by University of Massachusetts Amherst Libraries
Article
Design, Synthesis, and Structure-Activity-Relationship
of Tetrahydropyrido[4,3-d]pyrimidine Derivatives as
Potent Smoothened Antagonists with in Vivo Activity
Wenfeng Lu, Yongqiang Liu, Haikuo Ma, Jiyue Zheng, Sheng Tian, Zhijian
Sun, Lusong Luo, Jiajun Li, Hongjian Zhang, Zeng-jie Yang, and Xiaohu Zhang
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.7b00153 • Publication Date (Web): 15 Jun 2017
Downloaded from http://pubs.acs.org on June 16, 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.
ACS Chemical Neuroscience 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 37
ACS Chemical Neuroscience
1
2
3
4
Design, Synthesis, and Structure-Activity-Relationship of Tetrahydropyrido[4,3-d]pyrimidine
5
6
7
Derivatives as Potent Smoothened Antagonists with in Vivo Activity
8
9
Wenfeng Lu,†,# Yongqiang Liu,‡,# Haikuo Ma,†,# Jiyue Zheng,† Sheng Tian,† Zhijian Sun,§ Lusong
Luo,*,§ Jiajun Li,† Hongjian Zhang,† ZengꢀJie Yang,*,†, ‡ and Xiaohu Zhang*,†
† Jiangsu Key Laboratory of Translational Research and Therapy for NeuroꢀPsychiatricꢀDiseases and
College of Pharmaceutical Sciences, Soochow University, Suzhou, Jiangsu 215021, P. R. China
‡ Cancer Biology Program, Fox Chase Cancer Center, Temple University Health System, Philadelphia,
Pennsylvania, USA.
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
§ BeiGene (Beijing) Co., Ltd., No. 30 Science Park Road, Zhongguancun Life Science Park, Beijing
102206, P. R. China
Abstract: Medulloblastoma is one of the most prevalent brain tumors in children. Aberrant hedgehog
(Hh) pathway signaling is thought to be involved in the initiation and development of medulloblastoma.
Vismodegib, the first FDAꢀapproved cancer therapy based on inhibition of aberrant hedgehog signaling,
targets smoothened (Smo), a Gꢀprotein coupled receptor (GPCR) central to the Hh pathway. Although
vismodegib exhibits promising therapeutic efficacy in tumor treatment, concerns have been raised
fromits nonꢀlinear pharmacokinetic (PK) profiles at high doses partly due to low aqueous solubility.
Many patients experience adverse events such as muscle spasms and weight loss. In addition, drug
resistance often arises among tumor cells during treatment with vismodegib. There is clearly an urgent
need to explore novel Smo antagonists with improved potency and efficacy. Through a scaffold
hopping strategy, we have identified a series of novel tetrahydropyrido[4,3ꢀd]pyrimidine derivatives,
which exhibited effective inhibition of Hh signaling. Among them, compound 24 is three times more
1
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 2 of 37
1
2
3
4
5
6
7
8
9
potent than vismodegib in the NIH3T3ꢀGREꢀLuc reporter gene assay. Compound 24 has lower melting
point and much greater solubility compared with vismodegib, resulting in linear PK profiles when
dosed orally at 10, 30, and 100 mg/kg in rats. Furthermore, compound 24 showed excellent PK profiles
with a 72% oral bioavailability in beagle dogs. Compound 24 demonstrated overall favorable in vitro
safety profiles with respect to CYP isoform and hERG inhibition. Finally, compound 24 led to
significant regression of subcutaneous tumor generated by primary Ptch1ꢀdeficient medulloblastoma
cells in SCID mouse. In conclusion, tetrahydropyrido[4,3ꢀd]pyrimidine derivatives represent a novel
set of Smo inhibitors that could potentially be utilized to treat medulloblastoma and other Hh pathway
related malignancies.
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
Keywords: medulloblastoma, hedgehog signaling pathway, Smoothened, antagonist, GPCR, scaffold
hopping, cancer therapy
2
ACS Paragon Plus Environment

Page 3 of 37
ACS Chemical Neuroscience
1
2
3
4
INTRODUCTION
5
6
7
8
9
Medulloblastoma is one of the most prevalent brain tumors in children. Aberrant hedgehog (Hh)
pathway signaling is thought to be involved in the initiation and development of medulloblastoma.1,2
The hedgehog (Hh) signaling pathway is a pivotal developmental pathway which is responsible for cell
growth, patterning, and migration during embryogenesis. Its function is, however, limited to tissue
maintenance and stem cell regulation in adults. Ordinarily, the secreted proteins Sonic hedgehog,
Indian hedgehog and Desert hedgehog bind to their cellular membrane receptor Patched (Ptch),
relieving the suppression of Ptch to Smoothened (Smo), a G protein coupled receptor. Activated Smo
then triggers a series of intracellular events which involve the gliomaꢀassociated oncogene transcription
factors’ (Gli1, Gli2, and Gli3) dissociation from suppressor of fused (SUFU) and translocation to the
nucleus to transcribe Hedgehog target genes.3,4 Over activation of Hh signaling has been implicated in
multiple human malignancies. Ligandꢀindependent activation of the Hh signaling pathway caused by
mutational inactivation of the inhibitory pathway components e.g. Ptch (lostꢀofꢀfunction of Ptch) or
activation of Smo (gainꢀofꢀfunction of Smo) leads to medulloblastoma and basal cell carcinoma.5,6
Ligandꢀdependent activation of Hh signaling is observed in colorectal, pancreatic, lung, prostate and
blood cancers.7ꢀ9 Therefore, mitigation of hyperactivation of Hh signaling represents an attractive
therapeutic strategy for cancer treatments.10ꢀ15
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
The first Hh signaling pathway inhibitor reported in the literature was cyclopamine (Figure 1), a
naturally occurring alkaloid later identified as a Smo inhibitor.16 IPIꢀ926 (Figure 1) was then developed
based on the framework of cyclopamine. IPIꢀ926 exhibited not only better chemical stability and
potency, but also much improved metabolic stability, pharmacokinetic profiles and other superior
pharmaceutical properties than those of cyclopamine.17ꢀ19 Infinity Pharmaceuticals initiated clinical
3
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 4 of 37
1
2
3
4
5
6
7
8
9
trials with IPIꢀ926. Unfortunately, all clinical development efforts were suspended following
disappointing results of a phase II study of IPIꢀ926 in patients with metastatic chondrosarcoma.20 More
encouragingly, NVPꢀLDE225, a synthetic small molecule Smo inhibitor, demonstrated promising
results in clinical development for advanced basal cell carcinoma.21,22 The mixed results underscored
the complex nature of Hh signaling and prompted debate over whether blockage of ligand dependent
Hh pathway activation could be effective in cancer therapy.14,15 Despite the uncertainty, many Smo
inhibitors have been developed in recent years (Figure 1).23ꢀ37 In January 2012, vismodegib (GDCꢀ0449,
Figure 1) was approved by FDA for patients with locally advanced or metastatic basal cell carcinoma
which was not suitable for operation.38 This approval highlights the firstꢀinꢀclass small molecule Smo
antagonist for the treatment of human cancer and validated decades of basic research in hedgehog
signaling pathway. However, vismodegib suffered from suboptimal physicochemical properties such as
low aqueous solubility, resulting in nonꢀlinear pharmacokinetic profiles at high doses. Adverse events
such as muscle spasms and weight loss occurred in many patients. In addition, vismodegib suffered
from resistance in patients.39ꢀ41 Accumulating evidence suggests that Smo antagonists with diverse
structures, even though binding to the same site as vismodegib, may be capable of blocking the mutant
Smo.20,25ꢀ27 This effect may be due to the extended interactions between Smo and the antagonists away
from the mutant site. Therefore, Smo inhibitors with diverse chemical structures and improved potency
and efficacy may be able to address the needs of patients who do not benefit from vismodegib. In July
2015, a second Smo inhibitor, sonidegib (NVPꢀLDE225, Figure 1), was approved by FDA for the
treatment of locally advanced basal cell carcinoma.42 This new approval highlights the continued
efforts and interests in research and development of hedgehog pathway inhibitors. Here we report our
work on a series of Smo antagonists based on a novel tetrahydropyrido[4,3ꢀd]pyrimidine template.
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
4
ACS Paragon Plus Environment

Page 5 of 37
ACS Chemical Neuroscience
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 1. Structures of advanced smoothened antagonists reported in the literature
RESULTS AND DISCUSSION
Compound Design Rationale. In our effort to pursue novel Smo inhibitors, we explored numerous
templates with the guidance of the recently solved Smo crystal structures43ꢀ47 and a pharmacological
model.34 Vismodegib (GDCꢀ0449)25 was obtained by optimization campaign originated from
highꢀthroughput screening hits. With the exception of the terminal methyl group, the skeleton of
vismodegib was constructed uniformly with sp2ꢀhybridized carbons, resulting in a very high melting
point (264 °C) and poor solubility (9.5 ꢁg/mL). In the process of optimizing vismodegib, the
orthoꢀchloro group was added to the right side ring to introduce tilt and reduce planarity of the aryl
amide, despite adding additional molecular weight and clogP, to improve its solubility.25 In clinical
development, vismodegib was hampered by its nonꢀlinear pharmacokinetic profiles at high doses, most
likely due to its high crystal packing energy and poor aqueous solubility. We have encountered a
similar problem in an adenosine antagonist program before.48,49 Molecular topology and planarity have
5
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 6 of 37
1
2
3
4
5
6
7
8
9
significant impact on physicochemical properties such as solubility is well documented in the
literature.50,51 In an attempt to improve the physicochemical properties of our designed compounds, we
proposed to introduce sp3ꢀhybridized saturation ring as a placeꢀholder, as featured in SEN79432 and
PFꢀ527485752, as a way to reduce planarity. We also proposed forming a ring on the tail part in
SEN794 and PFꢀ5274857 to decrease rotatable bonds, as a way of enhancing metabolic stability and
absorption.53 We applied this strategy to obtain a number of novel scaffolds exemplified by
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
tetrahydroimidazo[1,2ꢀa]pyrazine,54
tetrahydrothiazolo[5,4ꢀc]pyridine,55
and
tetrahydropyrido[4,3ꢀd]pyrimidine, as shown in Figure 2. We initially worked on
tetrahydroimidazo[1,2ꢀa]pyrazine scaffold based on its low clogP (2.4) and polar surface area (73). The
pKa (7.5) is also attractive because it may further improve the solubility of its derivatives, without
overtly changing their ionization stage. Unfortunately, the best compounds we could optimize from the
tetrahydroimidazo[1,2ꢀa]pyrazine scaffold were at least 5 times less potent compared with vismodegib
in our screening assay.54 Better results were achieved from optimization of the
tetrahydrothiazolo[5,4ꢀc]pyridine scaffold. We were able to obtain compounds as potent as vismodegib.
In addition, the leading compounds also showed good pharmacokinetic profiles and improved solubility.
Encouraged, we decided to also work on the tetrahydropyrido[4,3ꢀd]pyrimidine scaffold, because the
pyrimidine is a closer bioisostere to thiazole in size and the template has a much lower intrinsic clogP
(2.9, Figure 2). The crystal complex of vismodegib interacts with Smo receptor was released (PDB ID:
5L7I) recently by Byrne et al.43 Based on this work, we examined the interaction patterns between
compounds 12 and 16 (Table 1) generated from Template III (Figure 2) and Smo receptor by using SP
(standard precision) and XP (extra precision) scoring functions in Glide.56 The detailed docking
procedure has been reported in our previous study.57 According to Glide docking results, the predicted
6
ACS Paragon Plus Environment

Page 7 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
binding affinities of compound 12 was worse than, while compound 16 was better than that of
vismodegib against Smo receptor evaluated by the docking scores. For example, the docking scores of
compounds 12, 16, and vismodegib were ꢀ5.951, ꢀ9.491, and ꢀ7.828 by using SP scoring mode of Glide,
respectively. Similar results were observed in Glide XP scoring function studies. The computational
predictions indicated that promising Smo receptor antagonists may be obtained based on Template III.
In addition, we also found that the binding poses of compounds 12 and 16 predicted by Glide docking
were quite similar to that of the vismodegib. It demonstrated that the key interactions derived from the
vismodegibꢀSmo crystal complex were almost reserved by our scaffold optimizations. The docking
poses of compounds 12 and 16 from Glide SP docking were depicted in Figure 3 using DS2.5.58 Here
we report the results from the exploration of the tetrahydropyrido[4,3ꢀd]pyrimidine template.
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 2. Design rationale of new templates
7
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 8 of 37
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. The docking poses of compounds 12 (A) and 16 (B) derived from Glide SP docking in Smo
receptor.
Chemistry. The synthesis of compounds 12-16, 20, 21, 26, 27, and 32-34 was undertaken as outlined
in Scheme 1. The commercial available 2,5ꢀdichloropyridine 1 was treated with LDA and then
triisopropyl borate to afford boronic acid 2.59 The key intermediate 3 was obtained by Suzuki coupling
of 2 with 2ꢀbromoꢀ3,5ꢀdimethylꢀpyridine in the presence of Pd(dppf)Cl2 and K3PO4•3H2O. A twoꢀstep
sequence involving piperidone 4 treated with DMFꢀDMA in reflux followed by ring closure with
Sꢀmethylisothiourea sulfate and sodium acetate gave tetrahydropyrido[4,3ꢀd]pyrimidine 5. Oxidization
of intermediate 5 with mꢀCPBA provided sulfone 6. The intermediate 6 was substituted with
corresponding amines toafford intermediates 7a-k. Removal of the Boc of 7a-k with HCl afforded
amines 8a-k, which were reacted with 3 in the presence of CsF to give compounds 12-16, 20, 21, 26,
27, 32, and 33. Sulfonylation of 13 with methyl sulfonyl chloride gave compound 34.
Scheme 1.Synthesis of Analogues12-16, 20, 21, 26, 27,and 32-34
8
ACS Paragon Plus Environment

Page 9 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
Cl
Cl
N
N
Cl
b
a
N
HO
Cl
B
Cl
N
Cl
OH
1
2
3
7
8
9
Boc
Boc
Boc
N
N
N
N
d
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
N
e
c
O
S
N
N
S
O
O
5
6
4
Cl
N
Boc
f
N
N
g
HN
N
N
N
N
N
R
N
R
N
R
7a-k
8a-k
12-16, 20, 21, 26, 27, 32-34
7a
R = amino
7b R = methylamino
7c R = ethylamino
7d R = isopropylamino
7e R = cyclopropylamino
12
13
14
R = amino
R = methylamino
R = ethylamino
15 R = isopropylamino
16 R = cyclopropylamino
20 R = pyrrolidino
21 R = morpholino
R = 3,5-dimethylpiperazin-1-yl
R = (hydroxyethyl)methylamino
32 R = 1H-pyrazol-1-yl
33 R = anilino
7f
7g
7h
R = pyrrolidino
R = morpholino
R = 3,5-dimethylpiperazin-1-yl
h
26
27
7i R = (hydroxyethyl)methylamino
7j R = 1H-pyrazol-1-yl
7k R = anilino
34 R = N-methylmethanesulfonamido
aReagents and conditions: (a) LDA, THF, ꢀ78 °C, 1 h, then triisopropyl borate, THF, rt, 2 h; (b)
2ꢀbromoꢀ3,5ꢀdimethylꢀpyridine, Pd(dppf)Cl2, K3PO4•3H2O, dioxane/H2O, 110 °C, 12 h; (c) i)
DMFꢀDMA, 100 °C, 12 h; ii) Sꢀmethylisothiourea sulfate, EtONa, EtOH, 80 °C, 12 h; (d) mꢀCPBA,
CH2Cl2, 0 °C to rt, 12 h; (e) corresponding amines, tꢀBuOH, 90 °C,16 h for 7a-i; NaH, pyrazole, THF,
0 °C, 1 h for 7j; aniline, 100 °C, 48 h for 7k; (f) HCl/EtOAc, rt, 5 h; (j) 3, CsF, DMSO, 120 °C, 36 h;
(h) MsCl, TEA, CH2Cl2, 0 °C, 3 h.
Compounds 17-19, 22-25 and 28-31 were prepared as shown in Scheme 2. Removal of the Boc of
intermediate 5 with HCl afforded amine 9, which was reacted with 3 in the presence of CsF to give
intermediate 10. Intermediate 10 was oxidized with oxone, then displaced with corresponding amines
or alcohols to yield compounds 17-19, 22-25, and 28-31.
Scheme 2. Synthesis of Analogues17-19, 22-25, and 28-31a
9
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 10 of 37
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
aReagents and conditions: (a) HCl/EtOAc, rt, 5 h; (b) 3, CsF, DMSO, 120 °C, 36 h; (c) oxone, THF, rt,
12 h; (d) corresponding amines, tꢀBuOH, 80 °C, 12ꢀ72 h for 17ꢀ19, 22ꢀ25; EtONa, EtOH, 0 °C, 1 h for
28; NaH, iꢀPrOH, 0 °C, 2 h for 29; corresponding alcohols, NaH, THF, 0 °C, 3 h for 30 and 31.
The biheteroaryl halides 37a-i, 48a, and 48b were obtained through either Negishi coupling or Suzuki
coupling (Schemes 3 and 4). The 2ꢀchloride of intermediates 37a-i, 48a, and 48b were substituted with
amine 8e to give the compounds 38-46, 49, and 50.
Scheme 3. Synthesis of Analogues 38-46a
aReagents and conditions: (a) nꢀBuLi, THF, ꢀ60 °C, 2 h, then iodine, THF, ꢀ60 °C, 2 h; (b) Pd(PPh3)4,
THF, reflux, 12 h; (c) K3PO4•3H2O, Pd(dppf)Cl2, dioxane/H2O, 110 °C, 12 h for 37a and 37b; Na2CO3,
Pd(dppf)Cl2, DME/H2O, 100 °C, 16 h for 37c and 37h; K3PO4•3H2O, Pd(dppf)Cl2, THF/H2O, 60 °C,
16 h for 37d; 2 N K2CO3, Pd(PPh3)4, EtOH/toluene, 115 °C, 16 h for 37e and 37i; Na2CO3, Pd(PPh3)4,
10
ACS Paragon Plus Environment

Page 11 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
dioxane/H2O, 130 °C, 16 h for 37f; (d) 8e, CsF, DMSO, 120 °C, 36 h.
Scheme 4. Synthesis of Analogues 49 and 50a
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
aReagents and conditions: (a) bis(pinacolato)diboron, AcOK, Pd(dppf)Cl2, dioxane, 110 °C, 6 h, then
2ꢀbromoꢀ3,5ꢀdimethylpyridine or 2ꢀbromoꢀ3ꢀmethylpyridine, K3PO4•3H2O, dioxane/H2O, 110 °C, 8 h;
(b) 8e, LiCl, K2CO3, DMSO, 140 °C, 36 h.
Compound 55 was prepared according to Scheme 5. Treatment of 2ꢀbromoꢀ3,5ꢀdimethylꢀpyridine 51
with nꢀBuLi and then N,Nꢀdimethylpropionamide to provide ketone 52. Reaction of 52 with
DMFꢀDMA and followed by cyclization with Sꢀmethylisothiourea sulfate provided thioether 53.
Oxidation of the thioether group of 53 with oxone afforded sulfone 54, which was substituted with 8e
to give final compound 55.
Scheme 5. Synthesis of Compound 55a
O
N
N
Br
O
S
a
c
d
b
N
N
S
N
N
N
O
N
N
52
51
53
54
N
N
N
N
N
N
H
55
aReagents and conditions: (a) nꢀBuLi, THF, ꢀ78 °C, 10 min, then N,Nꢀdimethylpropionamide, THF,
ꢀ78 °C, 1 h; (b) (i) DMFꢀDMA, 80 °C, 12 h; (ii) Sꢀmethylisothiourea sulfate, EtONa, EtOH, reflux, 12
h; (c) oxone, THF, 0 °C to rt, 16 h; (d) 8e, DIPEA, tꢀBuOH, reflux, 24 h.
Compounds 63 and 64 were prepared as outlined in Scheme 6. Aniline 58 was prepared from the
11
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 12 of 37
1
2
3
4
5
6
7
8
9
commercial available 56 in two steps via diazotization followed by reduction. Aniline 58 was converted
to borate ester 59 and then coupled with 2ꢀbromoꢀ3,5ꢀdimethylꢀpyridine or 2ꢀbromopyridine to give
intermediates 60a and 60b, respectively. Iodine was introduced by diazotizationꢀiodination to give 61a
and 61b. Buchwald coupling of amine 9 with 61a or 61b were carried outin the presence of Pd(dba)2
and XPhos to afford 62a and 62b, respectively. The thioether group of 62a and 62b was oxidized using
sodium perborate, and then displaced with cyclopropylamine to give compounds 63 and 64,
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
Scheme 6. Synthesis of Analogues 63 and 64a
aReagents and conditions: (a) NaNO2, HCl, H2O, 0 °C, 0.5 h, then KI, H2O, 0 °C to rt, 0.5 h; (b) Fe,
NH4Cl, EtOH/H2O, 75 °C, 3 h; (c) bis(pinacolato)diboron, Pd(dppf)Cl2, AcOK, DMF, 120 °C, 2.5 h; (d)
51 or 2ꢀbromopyridine, K3PO4•3H2O, Pd(PPh3)4, dioxane/H2O, 110 °C, 12 h; (e) NaNO2, HCl, H2O,
0 °C, 1 h, then KI, H2O, 0 °C to rt, 1 h; (f) 9, Pd(dba)2, XPhos, Cs2CO3, dioxane, 110 °C, 12 h; (g) (i)
NaBO3•4H2O, MeCN, rt, 16 h; (ii) cyclopropylamine, tꢀBuOH, 90 °C, 48 h.
Evaluation of Pharmacological Activity and Structure Activity Relationship. We have developed
two assays to characterize the synthesized compounds. The detailed experimental procedures were
reported before.54,55,60 Briefly, the primary screening assay was a cell based NIH3T3ꢀGREꢀLuc reporter
12
ACS Paragon Plus Environment

Page 13 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
gene assay with 10 nM SAG as the Hh pathway agonist.23,61,62 This assay is suitable to identify Smo
antagonists and Gli inhibitors but is not applicable to identify compounds interacting with targets
upstream of Smo, such as hedgehog protein inhibitors63, or hedgehog acyltransferase inhibitors64. In
order to confirm that the synthetic compounds were interacting with Smo, we developed a competitive
displacement assay based on fluorescence signal. In this assay, the synthesized compounds were
evaluated for their ability to displace BODIPYꢀcyclopamine in the U2OS cells overꢀexpressing human
Smo.16,65,66
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
The structureꢀactivityꢀrelationship is summarized in Tables 1 and 2.
Table 1. SAR of the rightꢀhand side structural elements
NꢀGꢀL
NꢀGꢀL
IC50 (nM)
± SEMa
SMOꢀBCB
IC50 (nM)
± SEM b
SMOꢀBCB
IC50 (nM)
± SEM b
IC50
(nM) ±
SEMa
Compd
R1
Compd
R1
12
13
>1000
24
25
7.5 ± 2.4
110 ± 24
24 ± 2.5
71 ± 1.0
330 ± 97
14
27 ± 11
26
14 ± 2.7
NH
15
16
17
28 ± 11
15 ± 10
58 ± 27
27
28
29
46 ± 23
150 ± 52
70 ± 28
41 ± 12
41 ± 9.0
O
13
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 14 of 37
1
2
3
4
5
18
19 ± 5.3
30
63 ± 21
6
7
8
9
19
20
21
76 ± 42
36 ± 11
66 ± 17
31
32
33
54 ± 23
98 ± 50
56 ± 53
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
22
23
240 ± 40
56 ± 29
34
100 ± 44
21 ± 12
Vismodegib c
92 ± 13
a Inhibition of luminescence signaling in NIH3T3ꢀGREꢀLuc reporter gene assay (NꢀGꢀL) with 10 nM
SAG as the Hh pathway agonist. Data are expressed as geometric mean values of at least two runs ±
the standard error measurement (SEM).
b Inhibition of BODIPYꢀcyclopamine fluorescence signaling in the competitive displacement
experiment using U2OS cells overꢀexpress human SMO. Data are expressed from a single IC50
determination.
c Vismodegib was run as standard in each assay. Data are expressed as geometric mean values of six
runs ± the standard error measurement (SEM).
Table2. SAR of the leftꢀhand side structural elements
NꢀGꢀL IC50
(nM) ± SEMa
Compd
R
X
Y
N
Z
CH
Cl
350 ± 43
38
14
ACS Paragon Plus Environment

Page 15 of 37
ACS Chemical Neuroscience
1
2
3
4
5
CH
N
Cl
67 ± 16
39
6
7
8
9
CH
CH
N
N
Cl
Cl
400 ± 70
40
41
550 ± 140
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
N
N
CH
CH
N
N
Cl
Cl
>1000
42
43
N
330 ± 99
N
CH
CH
CH
N
N
N
Cl
Cl
Cl
190 ± 52
650 ± 340
96 ± 16
44
45
46
CH
CH
N
N
N
N
H
H
74 ± 19
54 ± 6
49
50
55
CH3
450 ± 150
11 ± 8.5
CH
CH
CH
CH
Cl
Cl
63
64
110 ± 16
21 ± 12
Vismodegibb
a
Inhibition of luminescence signaling in NIH3T3ꢀGREꢀLuc reporter gene assay (NꢀGꢀL) with 10 nM
SAG as the Hh pathway agonist. Data are expressed as geometric mean values of at least two runs ±
the standard error measurement (SEM).
b
Vismodegib was run as standard in each assay. Data are expressed as geometric mean values of six
15
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 16 of 37
1
2
3
4
runs ± the standard error measurement (SEM).
5
6
7
8
9
The simplest compound of this series of analogues was inactive (compound 12, >1000 nM). Addition
of a methyl group on the rightꢀhand side terminal nitrogen effectively brought down activity to 330 nM.
The trend continued as ethyl, isopropyl, tꢀbutyl quickly brought activity down to low nanoꢀmolar range
(compounds 14, 15, 18; 27, 28, and 19 nM, respectively). Compound 18 was as potent as vismodegib
(21 nM) in the same assay. Cyclization of the alkyl groups was well tolerated, as demonstrated by
compounds 16, 17, and 19 (15, 58, and 76 nM, respectively). Compound 16 was slightly more active
compared with vismodegib. Secondary amines, either with cyclic structures (compounds 20ꢀ25) or
acyclic structure (compound 27), were well tolerated. Among them, compound 24 was three times
more potent than vismodegib. We also incorporated ether linkage (compounds 28ꢀ31) and aromatic
elements (compounds 32 and 33) to the terminal nitrogen, without significantly improving potency.
Finally, sulfonamide 34 (100 nM) was synthesized, but it was five times less potent compared with
vismodegib. The overall SAR trend largely followed the tetrahydrothiazolo[5,4ꢀc]pyridine series.55
Having established the SAR on the rightꢀhand side region of this scaffold, we turned our focus on the
optimization of the leftꢀhand side biꢀaryl region of this template. To do so, cyclopropyl amine was
arbitrarily fixed on the rightꢀhand side (Table 2). Removal of the orthoꢀmethyl group on the leftꢀhand
terminal 2ꢀpyridine ring led to significant potency loss (compound 38, 350 nM). This was in contrast to
the moderate potency loss when the paraꢀmethyl group was removed (compound 39, 67 nM). Potency
was almost unchanged when both methyl groups were removed (compound 40, 400 nM). This
observation indicated that the biꢀaryl group might need to be tilted to ensure a better fit or interaction
with the smoothened receptor. This is in consistence with the crystal structure and molecular modeling.
Replacement of the paraꢀmethyl with a trifluoromethyl slightly reduced potency. Introduction of an
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
16
ACS Paragon Plus Environment

Page 17 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
additional nitrogen to the terminal pyridine ring led to very different results: while 2,5ꢀpyrazine was
tolerated, 2,6ꢀpyrimidine led to totally inactive compound (compounds 43, 42; 330, >1000 nM,
respectively). As a bioisostere of 2ꢀpyridine, 2ꢀthiazole was active (compound 44, 190 nM). We were
unable to improve potency either by moving the nitrogen around the ring (compound 45, 650 nM), or
by adding an extra ring (compound 46, 96 nM) to the scaffold. Overall, the initial 4,6ꢀdimethyl
2–pyridine remained the best structural element after the SAR study. We then focused our optimization
on the internal ring of the leftꢀhand side biꢀaryl element. Removal of the chlorine led to moderate
reduction of potency (compound 49, 74 nM). Replacement of the chloro pyridine with methyl
pyrimidine led to over twenty folds potency loss (compound 55, 450 nM), while introduction of chloro
phenyl ring improved the potency slightly (compound 63, 11 nM). This improvement was maintained
when the terminal aryl ring was replaced by 2ꢀpyridine, as compound 64 was four times more potent
compared with compound 40 (110 nM vs. 400 nM). Although compound 63 was twice as potent as
vismodegib, given the unfavorably high clogP (clogP of 63, 4.8 vs. clogP of 16, 4.1), the chloro phenyl
element was deꢀprioritized. In summary, through a scaffold hopping strategy, we obtained a novel
tetrahydropyrido[4,3ꢀd]pyrimidine template which provided compounds with hedgehog inhibition
activity more potent than vismodegib (e.g. compounds 24 and 63) evaluated by the reporter gene assay.
We were able to identify the preferred structural pieces on the rightꢀhand side and on the leftꢀhand side
after the initial SAR study. Compounds 16 and 24 were identified as early leads. Although compound
16 was as potent as vismodegib in our primary screening assay, it exhibited a number of potential
problems that warranted to be monitored closely: it showed moderate CYP inhibition at10 ꢁM
concentration (Table 4). Furthermore, the cyclopropyl amine structural element had been linked to
reactive metabolites in numerous cases which might lead to toxicity.67,68 Conversely, compound 24
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
17
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 18 of 37
1
2
3
4
5
6
showed no significant inhibition to multiple CYP isoforms, suggesting a more favorable drugꢀdrug
interaction profile.
7
8
9
The reporter gene assay can effectively identify active compounds downstream of smoothened and is
suitable for high through screening. However, as compounds with common cell toxicity can also inhibit
the reporter gene and therefore cause false positive results, we used a Wnt pathway reporter gene assay
as a counter screening assay to minimize the possibility for false positive results.69,70 Nonspecific cell
toxicity was confirmed not to be the source of hedgehog inhibition activity (data not shown). Finally,
we used a fluorescenceꢀbased competitive displacement assay to confirm that the new compounds were
indeed Smo antagonists. Compounds 16 and 24 effectively displaced BODIPYꢀcyclopamine in the
U2OS cells overꢀexpressing human Smo (41 and 24 nM, respectively. vismodegib was less active in
this assay at 92 nM, Table 1), validating Smo as their molecular target.
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
Evaluation of Physicochemical Properties. By designing the novel scaffold, we aimed to maintain
the excellent potency inherent from the known templates while fostering improved physicochemical
properties (e.g. solubility). As a way to achieve this goal, we introduced sp3ꢀhybridized carbons, thus
reducing planarity and disrupting crystal packing, in hope to lower the melting point and improve
solubility. The solubility and melting point of compounds 16 and 24 from the
tetrohydrothiazolopyridine scaffold along with vismodegib were determined in the same assay
conditions and the results were summarized in Table 3 (the data reported in the literature was also
included for comparison). We selected these two compounds to compare with vismodegib because
these two compounds were equally or more potent compared with vismodegib (Table 1). In addition,
these compounds had similar clogP to vismodegib (Table 3). Under the current experimental conditions,
vismodegib demonstrated a solubility of 4.1 ꢁg/mL at pH 6.5, while compounds 16 and 24
18
ACS Paragon Plus Environment

Page 19 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
demonstrated a much improved solubility of 2770 ꢁg/mL and 90 ꢁg/mL, respectively, at pH 6.5.
Furthermore, the melting point of compounds 16 and 24 was determined to be 60 °C lower than that of
vismodegib (Table 3), suggesting that the added sp3ꢀhybridized carbons indeed lowered the crystal
lattice energy.
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. Solubility and melting point of leading compounds
Solubility (ꢁg/mL)c
Compd
16
clogPa
4.1
Melting point (°C)b
pH = 1.0
pH = 6.5
154
187
250
264d
>70000
2770
3.1
>6000
3150 ± 90
>3000d
90
4.1 ± 0.5
9.5d
24
Vismodegib
3.6
4.0d
a
b
c
Calculated by Molinspiration. All compounds were tested as crystalline HCl salts. All compounds
were tested as crystalline HCl salts. Data are measured three times ± the standard error measurement
(SEM). d Reported in the literature.25
Evaluation of in Vitro Safety. Compounds 16 and 24 were evaluated in a CYP screening assay in
order to assess their drugꢀdrug interaction potentials and liver safety. Compound 16 exhibited moderate
inhibition of CYP3A4, 1A2, 2C9, and 2C19 at 10 ꢁM. In contrast, compound 24 showed minimal
inhibition of all CYP isoforms tested at the same concentration, suggesting low drugꢀdrug interaction
liability. This may partly due to the relatively lower clogP value of compound 24. Compound 24 was
further profiled by a standard patch clamp (express) experiment for its inhibition of the human
etherꢀaꢀgoꢀgo related gene (hERG) potassium channel. Compound 24 exhibited very low inhibition
(IC50 = 38 ꢁM) of hERG, indicating minimal cardiotoxicity liability associated with blockade of this
19
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 20 of 37
1
2
3
4
key potassium channel. Compound 24 was found to be highly permeable (33× 10ꢀ6 cm/s, Cacoꢀ2), and
5
6
7
not likely to be a substrate for efflux (Table 4).
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
Table 4. In vitro safety data of 16 and 24
CYP Inhibition (%) b
Cacoꢀ2
hERG
Compd
clogP a
Papp(AꢀB)
(10ꢀ6 cm/s)
efflux ratio
(ꢁM)
3A4 2D6 1A2 2C9 2C19
(BꢀA)/(AꢀB)
4.1
3.1
4.0
60
8.0
2.7
27
21
24
76
49
81
69
30
51
70
33
44
16
24
38
33
0.53
Vismodegib
a Calculated by Molinspiration. b All compounds were tested at 10 ꢁM concentration.
Evaluation of Pharmacokinetic Properties and in Vivo Safety. Encouraged by the overall
satisfactory potency and in vitro safety data,we decided to assess the in vivo pharmacokinetic profile of
compound 24. When dosed orally in male SpragueꢀDawley (SD) rats, compound 24 exhibited a Tmax of
0.5 h and a Cmax of 2180 ng/mL. Compound 24 displayed a moderate clearance (18.6 mL/min/kg), a
good steady state volume of 1.7 L/kg, and a halfꢀlife of 2.3 h. The oral exposure of compound 24
yielded an AUC of 5541 ng·h/mL, leading to an estimated oral bioavailability of 62% (Table 5).
Further, compound 24 demonstrated a proportional PK profile when the oral doses were increased to 30
and 100 mg/kg (Table 5). The enhanced oral exposure of compound 24 at 30 and 100 mg/kg validated
the initial goal to improve physicochemical property as a way to improve drug exposure at higher doses.
Compound 24 can cross the blood brain barrier with an exposure of 831 ng/g at one hour after oral
dosing of 10 mg/kg, resulting in a brain/plasma ratio of 0.4. Given the tumor inhibition efficacy models
are mostly carried out in mouse, the PK of compound 24 was also determined in ICR mouse. After oral
dosing (10 mg/kg), compound 24 exhibited a Tmax of 0.25 h and a Cmax of 4132 ng/mL. Compound 24
20
ACS Paragon Plus Environment

Page 21 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
displayed a moderate clearance (20 mL/min/kg), a good steady state volume of 1.0 L/kg, and a halfꢀlife
of 1.0 h. The oral exposure of compound 24 was at an AUC of 4183 ng·h/mL, resulting in an estimated
bioavailability of 49% (Table 5). Based on the encouraging rodent PK data, compound 24 was further
tested in Beagle dogs. When dosed orally at 5 mg/kg, compound 24 exhibited a Tmax of 1 h and a Cmax
of 1300 ng/mL. Compound 24 displayed a moderate clearance (8.1 mL/min/kg), a good steady state
volume of 1.8 L/kg, and a halfꢀlife of 5.4 h. The oral exposure of compound 24 was excellent with an
AUC of 7656 ng·h/mL, resulting in an estimated bioavailability of 72% in beagle dogs (Table 5).
Pharmacokinetic evaluation in rodents and dogs indicated that compound 24 had the potential to
mitigate the solubilityꢀlimited PK seen in vismodegib.
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. Pharmacokinetic data of 24
Dose
AUC0ꢀ24h
CL
Vd,ss
Cmax
Tmax
(h)
t1/2
(h)
Compd species route
F%
49
(mg/kg) (ng·h/mL)
(mL/min/kg) (L/kg)
(ng/mL)
24
Mouse
Rat
i.v.
p.o.
i.v.
2
10
2
1701
4183
20
1.0
1.7
1.2
1.0
4132
0.25
1780
18.6
p.o.
10
30
100
1
5541
2180
8150
0.5
1
2.3
2.4
2.6
2.2
5.4
62
41900
251000
2062
18400
1
Dog
i.v.
8.1
1.8
p.o.
5
7656
1300
1
72
Based on the aforementioned data, compound 24 was evaluated in a dose escalating toxicity study
in mouse (Figure 4). The compound was well tolerated at all doses (100, 250, and 500 mg/kg,
21
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 22 of 37
1
2
3
4
5
6
7
8
9
daily oral gavage) and all animal exhibited normal behavior and consumed normal amount of food
in the 14ꢀday observation period. Slight increase of liver weight was observed at 250 and 500 mg/kg,
and no adverse effects were observed in other organs. The exposure of 24 after 2 h at 100 mg/kg was
22200 ng/mL and 32300 ng/mL on day 1 and day 14, respectively, indicating no significant CYP
induction or compound accumulation had occurred.
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 4. The mean body weight of male (n = 5) and female (n = 5) ICR mouse after receiving 100,
250, and 500 mg/kg compound 24by daily oral gavage in the 14ꢀday observation period.
Evaluation of In Vivo Efficacy. We next tested the capacity of compound 24 in inhibiting tumor
growth by utilizing a subcutaneous xenograft model of Ptch1+/ꢀ mouse medulloblastoma.71,72 Tumor
cells were freshly isolated from medulloblastoma formed in Ptch1+/ꢀ mice, in which one allele of
Ptch1 gene is mutant. 15ꢀ20% of Ptch1+/ꢀ mice develop medulloblastoma in their cerebella at
22
ACS Paragon Plus Environment

Page 23 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
around 30 weeks of age. The procedure of isolating tumor cells was previously described.73 In
brief, tumor tissue was enzymatically digested in a papain solution for 30 minutes to obtain a
singleꢀcell suspension, which was then centrifuged through a 35% and 65% Percoll gradient. Cells
from the 35% to 65% interface were harvested and suspended in Dulbecco's PBS plus 0.5% BSA.
The above procedure yields a cell suspension enriched with tumor cells (>95%). SCID mice with
established tumors (tumors volume ≥ 200 mm3) were randomly divided into three groups and treated
with vehicle, 50 mg/kg of compound 24 or vismodegib daily by oral gavage. As shown in Figure 5A,
although tumor grew steadily as treated with vehicle, both compound 24 and vismodegib significantly
suppressed tumor growth, with 94% and 97% growth inhibition, respectively. Meanwhile, no obvious
change was observed in the behavior and body weight of mice during compound 24 treatment (Figure
5B), suggesting that compound 24 exhibited dramatic antiꢀtumor effects without visible signs of gross
toxicity.
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
To determine whether tumor growth inhibition by compound 24 was due to repressed Hh pathway
activity, we harvested tumor tissue after drug treatment and evaluated Hh pathway activation based on
Gli1 mRNA expression examined by qꢀPCR. Compared with vehicle treatment, 24 and vismodegib
resulted in 94% and 86% inhibition of Gli1 mRNA expression, respectively, indicating that 24 has
comparable inhibitory effects on Hh pathway activation with vismodegib (Figure 5C).
23
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 24 of 37
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 5. Comparison of preclinical efficacy of compound 24 and vismodegib (p.o., 50 mg/kg) on
Ptch1+/ꢀ medulloblastoma allograft in SCID mice
To further confirm the antiꢀtumor effects of compound 24, we tested compound 24 in
medulloblastoma of Math1ꢀcre:SmoM2 mice74,75, a mouse model harboring constitutively active
Smo (SmoM2) allele, which was conditionally activated by Cre recombinase in granule cell
precursors. The Math1ꢀcre:SmoM2 mice have a heavy tumor burden in cerebellum, the mouse
was very sick and hunched up within one month. Treatment with compound 24 for 2 weeks
significantly reduced the tumor volume in cerebellum, and apparently improved the physical
conditions of the mouse (Figure 6). The results above indicate that compound 24 has significant
antiꢀtumor effects without signs of toxicity.
24
ACS Paragon Plus Environment

Page 25 of 37
ACS Chemical Neuroscience
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 6. Treatment of Math1cre/SmoM2 medulloblastoma mouse with 100 mg/kg of compound
24 by oral administration for 2 weeks (start treatment when the mice are 2 weeks old, upper panel)
and control (lower panel).
CONCLUSIONS
Through
a
scaffold hopping strategy, we have identified
a
series of novel
tetrahydropyrido[4,3ꢀd]pyrimidine derivatives, which exhibited effective inhibition of Hh signaling.
Among them, compound 24 is three times more potent than vismodegib in the NIH3T3ꢀGREꢀLuc
reporter gene assay. Compound 24 has lower melting point and much greater solubility compared with
vismodegib, resulting in linear PK profiles when dosed orally at 10, 30, and 100 mg/kg in rats. In
addition, compound 24 showed excellent PK profiles with a 72% oral bioavailability in beagle dogs.
Compound 24 demonstrated overall favorable in vitro safety profiles with respect to CYP isoform and
hERG inhibition. Finally, compound 24 led to significant regression of subcutaneous tumor generated
by primary Ptch1ꢀdeficient medulloblastoma cells in SCID mouse. In conclusion,
tetrahydropyrido[4,3ꢀd]pyrimidine derivatives represent a novel set of Smo inhibitors that could
potentially be utilized to treat medulloblastoma and other Hh pathway related malignancies.
EXPERIMENTAL METHODS
25
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 26 of 37
1
2
3
4
5
6
Chemistry. Full experimental details, characterization of intermediates and final compounds can be
found in the Supporting Information.
7
8
9
In vitro biological assays. NIH3T3ꢀGREꢀLuc reporter gene assay and BODIPYꢀCyclopamine binding
assay were performed aspreviously described.52 The detailed experimental procedures can be found in
the Supporting Information.
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
Solubility. Solubility at pH = 6.5 and PH = 1 were measured as previouslydescribed.52 The detailed
experimental procedures can be found in the Supporting Information.
Permeability. Cacoꢀ2 permeability test was performed according to the methodology described in the
literature.76
CYP inhibitory potency. CYP inhibitory potency was evaluated as previously reported.66 The detailed
experimental procedures can be found in the Supporting Information.
hERG assays.77 Compound 24 was evaluated for block of the hERG K channel using CHO cells stably
expressing the hERG gene and the QPatch platform (Sophion, Ballerup, Denmark). K tail currents were
measured at −50 mV following a 500 ms depolarization to +20 mV from a holding voltage of −80 mV.
The external solution contained 4 mM K+, 1 mM Mg2+,and 2 mM Ca2+. Compound effects were
quantified 4 min after application to the cells. Pulses were elicited every 20 s.
In vivo mouse PK in-life phase. Male ICR mice received either a single intravenous (bolus) injection
or single oral administration (by gavage) of compound 24 (HCl salt). Doses of 2 mg/kg (i.v.) and 10
mg/kg (p.o.) were given as solutions in saline. Blood samples were collected by heart puncture from n
= 3 animals per route of administration after 0.08 (i.v. only), 0.25, 0.5, 1, 2, 4, 6 (p.o. only), 8 h, and 24
h and were further processed to obtain plasma.
In vivo rat PK in-life phase. Male SpragueꢀDawley rats (body weight, 220ꢀ250 g) received either a
26
ACS Paragon Plus Environment

Page 27 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
single intravenous (bolus) injection or a single oral administration (by gavage) of the compound (HCl
salts). Doses of 2 mg/kg (per compound, i.v.), 10 mg/kg for compound 16 (p.o.) and 10/30/100 mg/kg
for compound 24 (p.o.) were given as solutions in saline. Consecutive blood samples were collected via
retroꢀorbital bleeding from n = 3 animals per route of administration after 0.08 (i.v. only), 0.25 (p.o.
only), 0.5, 1, 2, 4, 8 h, and 24 h and were further processed to obtain plasma.
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
In vivo dog PK in-life phase. Female Beagle dogs received either a single intravenous (bolus)
injection or an oral administration (by gavage) of compound 24 (HCl salt). Doses of 1 and 5 mg/kg
were given intravenously and per orally as a solution in saline. Consecutive blood samples were taken
by puncture of the vena cephalica from n = 3 animals per route of administration after 0.08 (i.v. only),
0.25, 0.5, 1, 2, 4, 8, and 24 h and were further processed to obtain plasma.
Bioanalysis. The concentrations of compound in plasma were analyzed by a LCꢀMS/MS system as
previously reported.55
14-Day repeat-dose mouse toxicity study. ICR mice were randomly assigned to 4 groups by using a
simple randomization procedure based on body weight and sex. 5 male and 5 female animals were
assigned to each group. Animals in group 1 to 4 were administered (p.o.) with compound 24 (HCl salt)
at 0, 100, 250, and 500 mg/kg/day as solutions in saline (10 mL/kg/day) for consecutive 14 days, once
daily. Cage side observation was given to all animals twice daily. A detailed physical examination was
conducted for all animals once daily during the dosing periods. Body weights were measured and
recorded once prior to group assignment, and once on Day 1, Day 4, Day 7, Day10, and Day13 during
dosing period. Food consumption was measured on Day 1, Day 4, Day 7, Day10, and Day13. On the
necropsy days of the dosing period, the survival animals were transferred to pathology department for
necropsy, they were euthanized by carbon dioxide inhalation and the organs (brain, adrenal gland, heart,
27
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 28 of 37
1
2
3
4
kidney, liver, spleen, testis, and thymus) were weighed.
5
6
7
8
9
Subcutaneous tumor allograft. Experiments were performed in female CB17 SCID mice between 8
and 12 weeks of age. CB17 SCID mice were bred in the Fox Chase Cancer Center Laboratory Animal
Facility (LAF). All animals were maintained in the LAF at Fox Chase Cancer Center and all
experiments were performed in accordance with procedures approved by the Fox Chase Cancer Center
Animal Care and Use Committee. Tumor cells from Ptch1+/ꢀ mouse medulloblastoma were resuspended
in reduced growth factor matrigel (BD 356231) with iceꢀcold PBS in 80:20 ratio, 2 x 106 cells in 50 ꢁL
matrigel were injected into the right flank area of the animals. The mice were randomly divided into 3
groups consisting of 6 mice each when the tumor volume reach around 200 mm3, the mice were
administrated with vehicle control (MCT solution: 0.5% Methyl Cellulose with 0.2% Tween 80 in
distilled water), 50 mg/kg Vismodegib (dissolved in MCT solution) and 50 mg/kg compound 24
(dissolved in distilled water) daily by oral gavage for two weeks. The tumor volume were calculated
using the formula V = (W2 × L)/2, where V is tumor volume, W is tumor width, L is tumor length.
q-PCR for Gli1 mRNA expression. Tumor tissues were homogenized and total RNA was isolated
using Trizol reagent (Sigma) following the manufacturer's instructions. cDNA was synthesized using
oligo(dT) and Superscript II reverse transcriptase (Invitrogen). Quantitative PCR reactions were
performed in triplicate using iQ SYBR Green Supermix (BioꢀRad) and the BioꢀRad iQ5 Multicolor
RealꢀTime PCR Detection System. Gli1 expression was normalized to Actin. Primer sequences are
available upon request.
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
ASSOCIATED CONTENT
Supporting Information
28
ACS Paragon Plus Environment

Page 29 of 37
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
In vitro biological assays S2, Solubility test S3, Evaluation of CYP inhibitory potency S3, Results of
14ꢀDay repeatꢀdose mouse toxicity study S4, Chemistry S5, Spectral Data and spectra. The Supporting
Information is available free of charge on the ACS Publications website.
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
AUTHOR INFORMATION
Corresponding Author
*(X.Z.) Eꢀmail: xiaohuzhang@suda.edu.cn; Tel.: +86 512 65880380; fax: +86 512 65880380.
*(L.L.) Eꢀmail: lusong.luo@beigene.com
.
*(Z.Y.) Eꢀmail: ZengꢀJie.Yang@fccc.edu.
Author Contributions
#W.L., Y.L., and H.M. contributed equallyto this paper.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was supported by National Natural Science Foundation of China (Grant No. 81473090,
81473278,
81572724,
21502133,81273372),
Natural
Science
Foundation of
Jiangsu
Province (BK20140313), BM2013003, and PAPD (A Project Funded by the Priority Academic
Program Development of Jiangsu Higher Education Institutions).
ABBREVIATIONS USED
AUC, area under the plasma timeꢀconcentration curve; CL, clearance; CYP, cytochrome P450;
mꢀCPBA,
mꢀchloroperoxybenzoic
acid;
DIPEA,
N,N-diisopropylethylamine;
DMF,
N,Nꢀdimethylformamide; DMFꢀDMA, N,Nꢀdimethylformamide dimethyl acetal; DMSO, dimethyl
29
ACS Paragon Plus Environment