Structure-Based Optimization of Small-Molecule Inhibitors for the β-Catenin/B-Cell Lymphoma 9 Protein-Protein Interaction

Article abstract of DOI:10.1021/acs.jmedchem.8b00068

Structure-based optimization was conducted to improve the potency, selectivity, and cell-based activities of β-catenin/B-cell lymphoma 9 (BCL9) inhibitors based on the 4′-fluoro-N-phenyl-[1,1′-biphenyl]-3-carboxamide scaffold, which was designed to mimic the side chains of the hydrophobic α-helical hot spots at positions i, i + 3, and i + 7. Compound 29 was found to disrupt the β-catenin/BCL9 protein-protein interaction (PPI) with a K_i of 0.47 μM and >1900-fold selectivity for β-catenin/BCL9 over β-catenin/E-cadherin PPIs. The proposed binding mode of new inhibitors was consistent with the results of site-directed mutagenesis and structure-activity relationship studies. Cell-based studies indicated that 29 disrupted the β-catenin/BCL9 interaction without affecting the β-catenin/E-cadherin interaction, selectively suppressed transactivation of Wnt/β-catenin signaling, downregulated expression of Wnt target genes, and inhibited viability of Wnt/β-catenin-dependent cancer cells in dose-dependent manners. A comparison of the biochemical and cell-based assay results offered the directions for future inhibitor optimization.

Full text of DOI:10.1021/acs.jmedchem.8b00068

Subscriber access provided by Queen Mary, University of London
Structure-Based Optimization of Small-Molecule Inhibitors for
the #-Catenin/B-Cell Lymphoma 9 Protein–Protein Interaction
Min Zhang, Zhen Wang, Yongqiang Zhang, Wenxing Guo, and Haitao Ji
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00068 • Publication Date (Web): 22 Mar 2018
Downloaded from http://pubs.acs.org on March 22, 2018
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 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 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.
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 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
StructureꢀBased Optimization of SmallꢀMolecule
Inhibitors for the βꢀCatenin/BꢀCell Lymphoma 9
Protein–Protein Interaction
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
Min Zhang,† Zhen Wang,† Yongqiang Zhang, ‡ Wenxing Guo,§ Haitao Ji*
Drug Discovery Department, H. Lee Moffitt Cancer Center and Research Institute, Tampa,
Florida, USA.
Departments of Oncologic Sciences and Chemistry, University of South Florida, Tampa, Florida,
USA
The present addresses:
† These two authors contribute equally to this work.
‡
Shanghai Key Laboratory of New Drug Design, School of Pharmacy, State Key Laboratory of
Bioengineering Reactor, East China University of Science and Technology, Shanghai, 200237,
China.
§ The Graduate Program, Department of Chemistry, University of Utah, Salt Lake City, Utah,
84112, USA.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 70
1
2
3
4
5
* To whom correspondence should be addressed:
6
7
Haitao Ji, Ph.D.
8
9
Drug Discovery Department, H. Lee Moffitt Cancer Center and Research Institute, 12902
Magnolia Drive, MRC room 4047, Tampa, Florida 33612–9416, USA
Phone: 813ꢀ745ꢀ8070
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
Fax: 813ꢀ745ꢀ4506
Eꢀmail: Haitao.Ji@moffitt.org
2
ACS Paragon Plus Environment

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
Keywords: Structureꢀbased drug design, βꢀCatenin, Bꢀcell lymphoma 9, Protein–protein
5
6
interaction, Inhibitors, Selectivity, Wnt signaling, Synthesis.
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
Running title: Structureꢀbased optimization of βꢀcatenin/BCL9 inhibitors
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 70
1
2
3
4
ABSTRACT
5
6
7
8
Structureꢀbased optimization was conducted to improve the potency, selectivity, and cellꢀ
based activities of βꢀcatenin/Bꢀcell lymphoma 9 (BCL9) inhibitors based on the 4'ꢀfluoroꢀNꢀ
phenylꢀ[1,1'ꢀbiphenyl]ꢀ3ꢀcarboxamide scaffold, which was designed to mimic the side chains of
the hydrophobic αꢀhelical hot spots at positions i, i + 3, and i + 7. Compound 29 was found to
disrupt the βꢀcatenin/BCL9 protein–protein interaction (PPI) with a Ki of 0.47 ꢀM and >1900ꢀ
fold selectivity for βꢀcatenin/BCL9 over βꢀcatenin/Eꢀcadherin PPIs. The proposed binding mode
of new inhibitors was consistent with the results of siteꢀdirected mutagenesis and structureꢀ
activity relationship studies. Cellꢀbased studies indicated that 29 disrupted the βꢀcatenin/BCL9
interaction without affecting the βꢀcatenin/Eꢀcadherin interaction, selectively suppressed
transactivation of Wnt/βꢀcatenin signaling, downregulated expression of Wnt target genes, and
inhibited viability of Wnt/βꢀcatenin dependent cancer cells in doseꢀdependent manners. A
comparison of the biochemical and cellꢀbased assay results offered the directions for future
inhibitor optimization.
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
4
ACS Paragon Plus Environment

Page 5 of 70
Journal of Medicinal Chemistry
1
2
3
4
INTRODUCTION
5
6
7
8
The Wnt/βꢀcatenin signaling pathway plays a key role in directing cell proliferation,
differentiation, and survival.1,2 βꢀCatenin is the central mediator of this signaling pathway. The
lossꢀofꢀfunction mutations in the suppressor genes of the Wnt/βꢀcatenin signaling pathway, such
as adenomatous polyposis coli (APC) and Axin, and/or the activation mutations at the Nꢀterminal
phosphorylation sites of the βꢀcatenin gene (CTNNB1), stabilize βꢀcatenin into the
dephosphorylated state and result in accumulation of βꢀcatenin in the cytoplasm. βꢀCatenin is
then translocated into the cell nucleus, where it displaces the repressor protein Groucho/TLE,
binds with the Tꢀcell factor (Tcf)/lymphoid enhancerꢀbinding factor (Lef) family of transcription
factors, and recruits coꢀactivators, Bꢀcell lymphoma 9 (BCL9), Pygo, CREBꢀbinding protein
(CBP), etc. to activate transcription of specific Wnt/βꢀcatenin target genes, such as cyclin D1, cꢁ
myc, survivin, and LEF1. These Wnt target genes then cause the initiation and progression of
cancers3–6 and fibroses7–9. The other causes for the hyperactivation of Wnt/βꢀcatenin signaling
are the autocrine/paracrine activation of Wnt ligands, frizzled (Fzd), and dishevelled (Dvl), and
the epigenetic silencing of Wnt antagonist genes. Wnt/βꢀcatenin signaling is also hyperactive in
cancer stem cells to maintain the selfꢀrenewal of cancer stem cells,10–14 seed cancer metastasis,15
and drive the resistance to the current therapies16–19. On the other hand, tumor cell intrinsic
activation of βꢀcatenin signaling prevents the recruitment of CD103+ dendritic cells by
suppression of chemokine CCL4 expression and excludes CD8+ Tꢀcells from the tumor
microenvironment.20,21 The activation of βꢀcatenin signaling was also reported to promote
intratumoral regulator T cell (Treg) survival and infiltration, suppressing Tꢀcell immunity.22,23
Significant efforts have been made to discover inhibitors for the upstream effectors of the
Wnt/βꢀcatenin signaling pathway.24–31 However, the inhibition of the upstream Wnt effectors can
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
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 70
1
2
3
4
5
6
7
8
9
potentially cause the crossꢀregulatory effects on the noncanonical Wnt signaling pathways.
Second, these inhibitors cannot confer efficacy to the diseases that harbor the more downstream
APC and Axin lossꢀofꢀfunction mutations and the βꢀcatenin activation mutations. The formation
of the βꢀcateninꢀcontaining transcriptional complex in the cell nucleus is the penultimate step of
this pathway. The transcriptional overactivation of Wnt/βꢀcatenin signaling is dependent on the
formation of this complex, in which βꢀcatenin interacts with BCL9.32 BCL9 functions as a
scaffolding structure of the Wnt enhanceosome and brings βꢀcatenin to Tcf/Lef to transcribe
specific Wnt target genes that cause the diseases.33 The crystallographic analysis34,35 and the
biochemical studies36–38 reveal that the homology domain 2 (HD2) of BCL9/B9L adopts an αꢀ
helical structure to interact with the first armadillo repeat of βꢀcatenin. Mammalian genomes
encode two BCL9 paralogues, BCL9 and B9L (BCL9ꢀlike).39,40 The use of siRNAs39–42 and
shRNAs43 against BCL9/B9L markedly decreases βꢀcateninꢀdependent gene expression and
diminishes the growth of cancer cells in vitro and in vivo. The dominant negative constructs of
BCL9/B9L that lack the Cꢀterminal domains greatly inhibit the activity of Wnt/βꢀcatenin
signaling.39,41 Hence, the inhibitors that can disrupt the βꢀcatenin/BCL9 protein–protein
interaction (PPI) not only provide new chemical probes to understand the biology regulated by βꢀ
catenin signaling, but also offer the important starting points for drug discovery to treat advanced
cancers and fibroses.
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
Triazoleꢀstapled44 and olefinꢀstapled45 BCL9 L351–F374 αꢀhelical peptides have been
reported. The olefinꢀstapled BCL9 peptide was reported to target βꢀcatenin, disrupt the βꢀ
catenin/BCL9 PPI, and suppress transcription of Wnt target genes. This stapled peptide also
inhibited tumor cell growth, angiogenesis, and metastasis without overt damage to normal tissues
in mouse xenograft models for colorectal carcinoma and multiple myeloma.45 Compound
6
ACS Paragon Plus Environment

Page 7 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
screening identified a small molecule, carnosic acid, that disrupted the βꢀcatenin/BCL9 PPI and
inhibited βꢀcatenin dependent transcription.35,46 In the previous study, we designed 4'ꢀfluoroꢀNꢀ
phenylꢀ[1,1'ꢀbiphenyl]ꢀ3ꢀcarboxamide as a generalizable scaffold that itself can directly mimic
the side chains of αꢀhelical hot spots at positions i, i + 3, and i + 7.47 The further derivatization
generated 1 in Figure 1. The biochemical AlphaScreen assay indicated that 1 disrupted the βꢀ
catenin/BCL9 PPI with an inhibition constant (Ki) of 2.1 ± 0.41 ꢀM and showed 125ꢀfold
selectivity for βꢀcatenin/BCL9 over βꢀcatenin/Eꢀcadherin PPIs.47 The proposed binding mode of
1 was evaluated with the siteꢀdirected mutagenesis studies using βꢀcatenin D145A, E155A,
D145A/E155A, L159S, and L156S/L178S and the structureꢀactivity relationship (SAR) studies
of 29 derivatives. The cellular bioavailability studies indicated that 1 can easily be taken up by
cancer cells.48 This compound also dissociated the βꢀcatenin/BCL9 PPI in cells, selectively
suppressed transactivation of Wnt/βꢀcatenin signaling, downregulated expression of Wnt target
genes, and inhibited growth of Wnt/βꢀcateninꢀdependent cancer cells. To further improve the
potency and selectivity of this series of inhibitors, herein, we report the structureꢀbased
optimization, synthesis, and biological characterization of new derivatives.
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
1. Inhibitor Design and Synthesis
(a) Explore the hydrophobic interaction. As described in the previous studies, the βꢀ
catenin/BCL9 PPI interface has two hot regions.34,36–38,41,47,49 In hot region 1, residues D162,
E163, and D164 of human βꢀcatenin form an acidic knob to interact with H358 and R359 of
human BCL9.34,36–38,41 This PPI interface is shallow and dominated by electrostatic interactions.
In hot region 2, residues L366 (i), I369 (i + 3), and L373 (i + 7) of human BCL9 interact with a
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 70
1
2
3
4
5
6
7
8
9
surface pocket that is lined with L159, V167, A171, M174, L178, L148, A149, A152, and L156
of human βꢀcatenin, as shown in Supplementary Figure S1.34,41,44,47 This PPI interface is
dominated by the hydrophobic interaction. The model of the interaction of 1 with βꢀcatenin
reported in the previous study47 was used as the starting point for inhibitor optimization. The first
step of inhibitor design based on 1 was to increase its hydrophobic interactions with βꢀcatenin.
Compounds 2–8 in Figure 1 were designed and synthesized. The synthetic route for 2 and 3 is
shown in Scheme 1. The synthetic route for 4–8 is shown in Supplementary Scheme S1. The
AlphaScreen assay50 was used to evaluate the inhibitory activities of 2–8 in in vitro biochemical
assays using fullꢀlength βꢀcatenin and the BCL9 peptide (G350–P375). The results were
expressed as the Ki in Figure 1. Compounds 2 and 3 displayed the similar biochemical activities
as 1, indicating the limited gain of the potency by increasing the hydrophobic interaction.
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. The chemical structures and the AlphaScreen Ki values of 1–11. Each set of data was
expressed as mean ± standard deviation (n = 3).
8
ACS Paragon Plus Environment

Page 9 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The 3, 4ꢀdifluorophenyl ring of 1 was also replaced with the aliphatic ring to explore the
hydrophobic interaction. Compounds 9–11 were designed and synthesized. The synthetic route
of 9–11 is shown in Supplementary Scheme S2. The AlphaScreen assay results in Figure 1
indicated that these three compounds were not as potent as 1.
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
(b) Search for the favorable polar interactions. Although the hot region 2 is primarily
hydrophobic, it does have the helical dipole and the polar functional groups. One strategy we
used was the replacement of the CH group in the aromatic rings of 1 with the N atom.51 This
method introduces the polarity to the compounds and reduces the lipophilicity. Compounds 12–
16 were designed and synthesized. The synthetic routes for 12 and 13 are shown in
Supplementary Scheme S3. The synthetic routes for 14–16 are shown in Supplementary Scheme
S4. The biochemical assay results of 12–16 in Figure 2 indicated none of these compounds were
as potent as 1.
Figure 2. The chemical structures and the AlphaScreen Ki values of 12–16. Each set of data was
expressed as mean ± standard deviation (n = 3).
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 70
1
2
3
4
5
6
7
8
9
(c) Search for the favorable H-bonding and salt bridge interactions. The crystal structures of
βꢀcatenin in complex with the BCL9 HD2 domain shows βꢀcatenin Lys181 is adjacent to the hot
region 2.34,35 Another strategy for inhibitor optimization was targeting this residue. Compounds
17–23 were designed to form Hꢀbonding interactions with βꢀcatenin Lys181. The synthetic route
for 17–23 is shown in Supplementary Scheme S5. Carboxylic acid and bioisosteres were
introduced to 1 to form salt bridge interactions with Lys181. Compounds 24–30 were designed
and synthesized. The synthetic route for 24–26 is shown in Supplementary Scheme S6. The
synthetic routes for 27–30 are shown in Scheme 2. The AlphaScreen assay results of 17–30 are
shown in Figure 3. The biochemical inhibitory activities of 27 and 30 were comparable with that
of 1. Compound 29 exhibited the Ki of 0.47 ± 0.14 ꢀM for disruption of the βꢀcatenin/BCL9 PPI.
This Ki value is very close to the dissociation constant (KD) of 0.47 ꢀM for the βꢀcatenin/BCL9
interaction.34
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
10
ACS Paragon Plus Environment

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
Figure 3. The chemical structures and the AlphaScreen Ki values of 17–30. Each set of data was
expressed as mean ± standard deviation (n = 3).
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
(d) Introduce conformational constraints. Two compounds, 31 and 32, were designed and
synthesized. The synthetic routes for 31 and 32 are shown in Scheme 3. The biochemical assay
results of these two compounds are shown in Figure 4. Compound 32 exhibited the similar
biochemical inhibitory activity as 1.
Figure 4. The chemical structures and the AlphaScreen Ki values of 31 and 32. Each set of data
was expressed as mean ± standard deviation (n = 3).
2. Biochemical Characterizations.
(a) Inhibitor selectivity between β-catenin/BCL9 and β-catenin/E-cadherin PPIs. βꢀCatenin
is at least associated with two functions in cells. One is that the nuclear βꢀcatenin interacts with
Tcf/Lef, BCL9/B9L, CBP/p300, etc. to culminate Wnt/βꢀcatenin signaling. The second is that the
cytosolic βꢀcatenin interacts with Eꢀcadherin to act as a structural component of adherens
junctions. The crystallographic analyses of βꢀcatenin in complexes with BCL934,35 and Eꢀ
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 70
1
2
3
4
5
6
7
8
9
cadherin52 indicated that the interface of βꢀcatenin to interact with BCL9 was also used to bind
region V of Eꢀcadherin. Compounds 29 and 30 were assessed for their selectivity between βꢀ
catenin/BCL9 and βꢀcatenin/Eꢀcadherin PPIs using the AlphaScreen selectivity assay.50 As
shown in Table 1, these two compounds exhibited high selectivity for βꢀcatenin/BCL9 over βꢀ
catenin/Eꢀcadherin PPIs with 29 of >1900ꢀfold.
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 1. The selectivities of 29 and 30 between βꢀcatenin/BCL9 over βꢀcatenin/Eꢀcadherin
interactions. Each set of data was expressed as mean ± standard deviation (n = 3).
Ki ± SD (ꢀM)
βꢀcatenin/Eꢀcadherin
Selectivity
BCL9/cadherin
>1900
Compounds
βꢀcatenin/BCL9
0.47 ± 0.14
1.6 ± 0.23
>915
>915
29
30
>570
(b) Site-directed mutagenesis studies. In the previous studies, compound 1 was reported to
bind with βꢀcatenin but not the BCL9 HD2 domain in isothermal titration calorimetry (ITC)
studies.47 The crystallographic studies indicated that D145, E155, and K181 were not involved in
the βꢀcatenin/BCL9 binding.34,35 Our AlphaScreen and fluorescence polarization (FP)
experiments indicated that βꢀcatenin D145A,47 E155A,47 D145A/E155A,47 and K181A mutants
had no effect on the βꢀcatenin/BCL9 PPI, as shown in Figure 5A. This result allowed us to use
the AlphaScreen competitive inhibition assays to evaluate the effects of these mutations on the Ki
of 3, 27, and 29. As shown in Figure 5B, the Ki values of 3 for wild type βꢀcatenin/BCL9, βꢀ
catenin D145A/BCL9, and βꢀcatenin D145A and E155A/BCL9 interactions were 1.3 ± 0.80, 4.2
± 1.1 and 12 ± 3.8 ꢀM, respectively, indicating that the carboxylic side chains of both D145 and
12
ACS Paragon Plus Environment

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
E155 of βꢀcatenin were important for the inhibitory potency of 3. However, βꢀcatenin K181 was
not targeted by 3, because the Ki value of 3 for the βꢀcatenin K181A/BCL9 interaction was 1.7 ±
0.26 ꢀM. This result is consistent with the proposed binding mode of 3 with βꢀcatenin in Figure
5C (The stick models of Figure 5C is shown in Supplementary Figure S2A). Compounds 27 and
29 were designed to form salt bridge interactions with βꢀcatenin K181A. The AlphaScreen Ki of
these two compounds for wild type βꢀcatenin/BCL9, βꢀcatenin D145A/BCL9, βꢀcatenin D145A
and E155A/BCL9, and βꢀcatenin K181A/BCL9 PPIs are shown in Figure 5B. βꢀCatenin D145,
E155, and K181 are important for 27 and 29 to bind with βꢀcatenin and disrupt the βꢀ
catenin/BCL9 PPI. This result is consistent with the proposed binding mode in Figure 5D (The
stick models of Figure 5D is shown in Supplementary Figure S2B).
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. (A) The apparent KD values of wildꢀtype and mutant βꢀcatenin proteins with wildꢀtype
BCL9 HD2 domain. (B) AlphaScreen competitive inhibition assay of 3, 27, and 29 to disrupt
wildꢀtype βꢀcatenin/wildꢀtype BCL9 and mutant βꢀcatenin/wildꢀtype BCL9 PPIs. Each set of
data was expressed as mean ± standard deviation (n = 3). (C) AutoDock docking result of 3 with
βꢀcatenin (PDB ID, 2GL7). (D) AutoDock docking result of 29 with βꢀcatenin (PDB ID, 2GL7).
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 70
1
2
3
4
5
6
3. Cell-Based Studies.
7
8
9
(a) Inhibition of viability of Wnt/β-catenin-dependent cancer cells. The MTS assay was
conducted to determine the effects of 1–32 on inhibition of cancer cell growth. Three cancer cell
lines with hyperactive Wnt/βꢀcatenin signaling, SW480, HCT116, and MDAꢀMBꢀ231, and one
cancer cell line with the normal Wnt/βꢀcatenin signaling pathway, A549, were used in this study.
The tested compounds were incubated with cells for 72 h before the addition of the MTS reagent
(3ꢀ(4,5ꢀdimethylthiazolꢀ2ꢀyl)ꢀ5ꢀ(3ꢀcarboxymethoxyphenyl)ꢀ2ꢀ(4ꢀsulfophenyl)ꢀ2Hꢀtetrazolium,
inner salt) and the electron coupling reagent, phenazine methosulfate (PMS), for detection.53,54
The half maximal inhibitory concentrations (IC50) of these compounds are shown in Table 2. All
of the compounds inhibited viability of cancer cells. Among them, compounds 4, 8, 14, 15, 27,
29, and 31 are selective for Wnt/βꢀcatenin hyperactive cancer cells over the cancer cells that have
normal Wnt signaling.
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
(b) Lactate dehydrogenase (LDH) cytotoxicity assay. The cytotoxic compounds may cause
cell death by damaging of the cell membrane through the nonspecific manner. The LDH release
assay55,56 with the relatively short inhibitor incubation time57 (4 h in this study) was conducted to
determine the cytotoxicity of the tested compounds. The result is shown in Supplementary Figure
S3. Compounds 1–18 except 12 exhibited nonspecific cytotoxicity at high concentrations, such
as 50 and 100 ꢀM. Compounds 19–31 did not exhibit obvious cytotoxicity in the LDH assay
even at the very high concentration of 800 ꢀM.
Table 2. The MTS assay to monitor the inhibitory activities of 1–32 on viability of cancer cells,
and the Wntꢀresponsive TOPFlash luciferase reporter assay to examine effect on Wnt/βꢀcatenin
14
ACS Paragon Plus Environment

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
transactivation. Each set of data was expressed as mean ± standard deviation (n = 3). The data
for 1 was reported in the previous study.47
7
8
9
MTs IC50 ± SD (ꢀM)
Wnt/βꢀcatenin hyperactive
TOPFlash IC50 ± SD (ꢀM)
βꢀcateninꢀactivated
normal Wnt
A549
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
No.
1
SW480
HCT116
3.1 ± 0.83
5.6 ± 2.2
7.5 ± 2.0
5.7 ± 1.9
1.4 ± 0.6
MDAꢀMBꢀ231
2.3 ± 0.25
2.1 ± 1.1
2.8 ± 1.7
3.1 ± 1.5
1.4 ± 0.7
HEK239
4.5 ± 1.3
10 ± 3.4
6.2 ± 1.9
SW480
3.5 ± 0.32
1.8 ± 0.94
2.0 ± 0.80
1.2 ± 0.62
2.2 ± 1.3
1.6 ± 0.71
2.6 ± 0.67
0.90 ± 0.51
6.6 ± 2.7
3.7 ± 2.0
3.6 ± 2.7
8.8 ± 0.62
3.7 ± 1.4
9.4 ± 3.3
2.9 ± 0.33
12 ± 3.5
1.2 ± 0.27
5.2 ± 1.3
2
3
23
± 5.0
4
4.5 ± 2.0
5
6
7
5.6 ± 2.0
3.8 ± 1.5
6.9 ± 3.2
5.6 ± 2.7
4.1 ± 1.8
10 ± 2.4
3.8 ± 1.5
8.9 ± 2.7
4.8 ± 2.4
2.0 ± 0.29
8
9
3.4 ± 1.6
2.3 ± 1.0
8.7 ± 2.3
10
11
10 ± 3.2
10 ± 3.3
21
15
± 8.0
± 4.1
12
13
14
15
16
3.0 ± 1.4
1.7 ± 0.99
9.9 ± 3.7
8.7 ± 3.5
7.8 ± 3.3
3.6 ± 1.2
3.1 ± 1.3
26 ± 7.4
29 ± 7.9
22 ± 4.5
10 ± 1.2
5.9 ± 1.0
7.7 ± 1.2
4.3 ± 0.74
6.1 ± 2.9
>25
17
18
19
20
21
22
23
>25
190
± 15
160
220
160
95
± 19
± 58
>200
>200
± 9.1
± 8.9
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 70
1
2
3
4
5
6
7
8
9
>400
>400
>400
140
86
>400
368 ± 30
>400
24
25
26
27
28
29
30
>400
>400
391 ± 36
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
± 11
190 ± 7.8
92 ± 16
410 ± 42
± 13
41
± 6.2
± 21
55 ± 7.4
240 ± 29
73 ± 8.9
>400
25 ± 4.1
50 ± 5.2
25 ± 2.7
43 ± 3.4
150
430 ± 30
3.6 ± 1.9
9.6 ± 2.5
2.2 ± 1.0
15 ± 6.1
3.1 ± 1.2
23 ± 4.4
7.5 ± 2.9
26 ± 3.1
14 ± 1.5
12 ± 1.5
7.9 ± 1.4
31
32
1.7 ± 0.90
(c) Inhibition of β-catenin/BCL9/Tcf transcriptional activity. The TOPFlash luciferase
reporter assay (in which the luciferase reporter has three wildꢀtype Tcf4 binding sites) was
performed with 1–4, 8, 10, 11, 14, 15, 17, 18, 21, and 29–32. Their IC50 values are shown in
Table 2. Among them, compounds 1–4, 8, 10, 11, 14, 15, and 29–32 inhibited the TOPFlash
luciferase activities. However, all of these compounds, except 29 and 30, inhibited the FOPFlash
luciferase reporter (with three mutant Tcf4 binding sites) activities at the higher concentrations.
The results of 1–4, 8, 11, 14, 15, and 29–32 are shown in Figure 6. These results indicated that
only 29 and 30 are specific for the Wnt/βꢀcatenin signaling pathway. Compounds 1–4, 8, 11, 14,
and 15 inhibited FOPFlash luciferase reporter activity because these compounds underwent the
primary necrosis by damaging the cell membrane at high concentrations.
16
ACS Paragon Plus Environment

Page 17 of 70
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 6. Wntꢀresponsive FOPFlash luciferase reporter assay results of inhibitors 1–4, 8, 11, 14,
15, and 29–32 in βꢀcatenin activated HEK293 cells.
(d) Inhibition of expression of Wnt/β-catenin target genes. Axin2 is a specific target gene of
the Wnt/βꢀcatenin signaling pathway.58 Cyclin D1, LEF1, and cꢁmyc are upregulated in cancer
cells with hyperactive Wnt/βꢀcatenin signaling to promote tumorigenesis. Quantitative realꢀtime
PCR (qPCR) studies were conducted for 3, 8, 29, and 31 (Figure 7). Similar to 1,47 compounds 3
and 29 did not inhibit houseꢀkeeper gene HPRT, but doseꢀdependently inhibited Wnt target
genes, Axin2, cyclin D1, and LEF1.
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 70
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 7. qPCR studies to determine the changes of mRNA expression of AXIN2, LEF1, and
cyclic D1 in SW480 cells in response to different concentrations of 3 (A), 8 (B), 29 (C), and 31
(D). Houseꢀkeeper gene HPRT was used as the reference. Each set of data was expressed as
mean ± standard deviation (n = 3).
As shown in Figure 8A,B, Western blot experiments indicated that the protein expression
levels of cyclin D1 and cꢀmyc were significantly decreased after treatment of 3 and 29,
respectively. Both compounds can inhibit the level of the active form of βꢀcatenin (ABC), which
is nuclear βꢀcatenin, in dose dependent manners. These two compounds had no effect on the
level of Eꢀcadherinꢀbound βꢀcatenin (i.e., total βꢀcatenin), indicating these two compounds did
not affect the upstream nodes of Wnt pathway that regulate degradation of cytosolic βꢀcatenin.
18
ACS Paragon Plus Environment

Page 19 of 70
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 8. (A and B) Western blot analysis to monitor the change of protein expression of cyclin
D1, cꢀmyc, the active form of βꢀcatenin (ABC), and the total βꢀcatenin in response to different
concentrations of 3 (A) and 29 (B) in SW480 cells. βꢀTubulin was used as the internal reference.
The quantitative data of Western blot analyses is shown in Supplementary Table S1 (C) CoꢀIP
experiments to evaluate the disruption of the βꢀcatenin/BCL9 PPI by 29 in HCT116 cell lysate.
(D) CoꢀIP experiments to evaluate the disruption of the βꢀcatenin/BCL9 PPI by 29 and the
selectivity of 29 for βꢀcatenin/BCL9 over βꢀcatenin/Eꢀcadherin PPIs in HCT116 cells. IP,
immunoprecipitation; IB, immunoblotting; input, 10% of cell lysate. Each experiment was
performed in triplicate. The quantitative data of coꢀIP experiments is shown in Supplementary
Table S2.
(e) Disruption of the β-catenin/BCL9 PPI in cells and cell-based inhibitor selectivity studies.
The coꢀimmunoprecipitation (coꢀIP) experiments were conducted to assess the effect of 29 for
disruption of the interaction between fullꢀlength βꢀcatenin and fullꢀlength BCL9 using HCT116
cell lysates. As shown in Figure 8C, compound 29 can doseꢀdependently disrupt the fullꢀlength
βꢀcatenin/fullꢀlength BCL9 PPI after 4 h incubation with HCT116 cell lysates. The coꢀIP
experiments were also performed with HCT116 cells to evaluate the effect of 29 on the
disruption of the βꢀcatenin/BCL9 PPI and the selectivity over the βꢀcatenin/Eꢀcadherin PPI in the
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 70
1
2
3
4
5
6
7
8
cellular context. As shown in Figure 8D, compound 29 inhibited the βꢀcatenin/BCL9 PPI in a
doseꢀdependent manner. A parallel experiment indicated that 29 did not affect the βꢀcatenin/Eꢀ
cadherin PPI at the concentrations that were sufficient to disrupt the βꢀcatenin/BCL9 PPI.
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
4. Chemistry.
The synthetic route for 2 and 3 is shown in Scheme 1. The amide bond coupling reaction
between the previously reported 3359 and 3447 generated key intermediate 35, which then
underwent the Suzuki reaction with (6ꢀfluoronaphthalenꢀ2ꢀyl) boronic acid or (1ꢀbenzothiophenꢀ
6ꢀyl) boronic acid to afford 36 in good yield. The Boc deprotection of 36 with 4 M HCl offered
the final products 2 and 3.
Scheme 1.
20
ACS Paragon Plus Environment

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The synthetic routes for 27–30 are shown in Scheme 2. Compound 33 was subjected to the
Curtius rearrangement to generate 37. The deprotection of the Cbz protecting group gave 38. The
amide bond coupling reaction of 38 with another previously reported key intermediate 3947, and
then the Suzuki reaction with 3ꢀcyanophenylboronic acid and 4ꢀcyanophenylboronic acid
generated 41a,b, respectively. The cycloaddition reaction of nitriles 41 with nꢁBu3SnN3 and then
the deprotection of the Boc protecting group with 4 M HCl offered 27 and 28. The Suzuki
reaction between the previously reported 4347 and 3ꢀcyanophenylboronic acid afforded 44. The
hydrolysis of the methyl ester of 44 and the Curtis rearrangement reactions furnished 46. The
deprotection of the Cbz group and the amide bond coupling reaction with 39 gave 48. The
cycloaddition reaction of 48 with nꢀBu3SnN3 resulted in 49, which offered the final product 29
after the deprotection of the Boc group. The Suzuki reaction between 43 and (1ꢀbenzothiophenꢀ
6ꢀyl) boronic acid generated 50. The deprotection of the methyl ester of 50 and the amide
coupling reaction with 46 afforded 52, which then underwent the cycloaddition reaction to
furnish 53. The deprotection of the Boc group of 53 generated final product 30.
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.
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 70
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
22
ACS Paragon Plus Environment

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
The synthetic routes for 31 and 32 are shown in Scheme 3. The esterification of 4ꢀbromoꢀ3ꢀ
hydroxybenzoic acid and then the Mitsunobu reaction with (R)ꢀNꢀBocꢀ3ꢀpyrrolidinol offered 55.
The Suzuki reaction of 55 with 3, 4ꢀdifluorophenylboronic acid and the hydrolysis of the methyl
ester of 56 afforded 57. The Curtius rearrangement of 57 and the deprotection of the Cbz group
generated 59. The methyl ester group of the previously synthesized intermediate 6047 was
reduced to a hydroxymethyl group by DIBALꢀH. The alcohol 61 was then oxidized to aldehyde
62 by PCC. The expoxidation of 62 by trimethylsulfoxonium iodide offered 63. The palladiumꢀ
catalyzed sequential reaction60 using the previously reported aniline 6447, aldehyde 62, and
acrylic acid produced 65. The other regioisomer 65a was not observed in this reaction. This
difference might be correlated with the steric hindrance of the 3, 4ꢀdifluorophenyl ring. The
deprotection of the Boc protecting group afforded the final product 31. The FeCl3 promoted
tandem reaction of aniline 59 with styrene oxide 63 furnished 66.61 The deprotection of the Boc
protecting group provided the final product 32.
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
Scheme 3.
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 70
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
DISCUSSION
Wnt/βꢀcatenin signaling is frequently observed hyperactive in cancers. The most effective
target to develop anticancer inhibitors for the Wnt/βꢀcatenin signaling pathway is the βꢀcateninꢀ
containing transcriptional complex, because this complex mediates the signaling at the
downstream node in the pathway. However, despite numerous efforts, this target has proven to
be challenging.62 To date, only one series of drugꢀlike inhibitors targets this transcriptional
24
ACS Paragon Plus Environment

Page 25 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
complex, ICGꢀ001 and its secondꢀgeneration prodrug PRIꢀ724.63 PRIꢀ724 is in clinical trial
phase II.64 ICGꢀ001 and PRIꢀ724 bind with transcription coactivator CBP to disrupt the βꢀ
catenin/CBP interaction.65 In the structure of the βꢀcateninꢀcontaining transcriptional complex,
the βꢀcatenin/BCL9 PPI represents an unique target for inhibitor design because: (1) BCL9/B9L
acts as an essential adaptor protein to recruit other cofactors and load βꢀcatenin to the Wnt
enhancesome, facilitating the access of βꢀcatenin to Tcf/Lef.33 The βꢀcatenin/BCL9 PPI is
required for transcription of Wnt responsive genes32,39,40,66–69 and tumorigensis41–43,70–72; (2) this
PPI has a moderately strong binding affinity (KD = 0.47 ꢀM); and (3) the βꢀcatenin interface for
binding with BCL9 is relatively small and has little overlap with those for the other βꢀcatenin
partners. Along with the efforts of the other groups to discover βꢀcatenin/BCL9 inhibitors,35,44–46
we reported the design and synthesis of 1 as a new smallꢀmolecule inhibitor for this PPI.47
In process of inhibitor optimization, we found that the gain of potency was limited by
exploration of the hydrophobic interaction based on 1. Also, the more hydrophobic compounds
cause cytotoxicity by damaging the cell membrane, as evident in the LDH release assay. The
substitution of the CH group of the aromatic rings of 1 with a N atom made the positive impact
on the molecular and physiological properties of 12–14, which displayed the lower cytotoxicity
in the LDH release assay. However, none of them were more potent than 1. Compound 1 is a
lipophilic base. The previous bioinformatic analyses indicated that the lipophilic bases were
prone to causing target promiscuity.73–76 Hence, in process of improving inhibitor potency and
selectivity for the Wnt/βꢀcatenin signaling pathway, we also aimed to introduce the polar groups;
compound 29 was obtained. This compound exhibits a Ki of 0.47 ꢀM for disruption of the βꢀ
catenin/BCL9 PPI and >1900ꢀfold selectivity for βꢀcatenin/BCL9 over βꢀcatenin/Eꢀcadherin PPIs
in biochemical AlphaScreen assays. This compound doseꢀdependently inhibited TOPFlash
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
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 70
1
2
3
4
5
6
7
8
luciferase reporter activity with an IC50 of 25 ꢀM, did not exhibit cytotoxicity in the LDH release
assay, and had no effect on FOPFlash luciferase reporter activity up to the concentration of 200
ꢀM.
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
The Ki of 29 for disruption of the βꢀcatenin/BCL9 PPI was 0.47 ꢀM. The TOPFlash
luciferase reporter (with three Tcf binding sites) activity of 29 was 25 ꢀM. This difference could
be caused by the zwitterion nature of this compound. The intracellular concentration of 29 was
determined by the method described in the previous study.48 The result was shown in the
Supplementary Note. The cellꢀbound concentration of 29 at 37 °C was determined to be 6.6
nmol/million HCT116 cells for the 6 h incubation in 5 mL of DMEM media with 10% FBS,
when the input concentration was 50 ꢀM of 29. The 4 °C incubation experiment indicated a
significant amount (77%) of cellꢀbound 29 bound with the cell membrane through the
nonspecific manner. The cellꢀbound concentration of 29 was 5.1 nmol/million HCT116 cells for
6 h incubation at 4 °C in 5 mL of DMEM media with 10% FBS. This result is different from that
of 1.48 The cellꢀbound concentration of 1 at 37 °C was determined to be 6.7 nmol/million MDAꢀ
MBꢀ231 cells for the 6 h incubation in 5 mL of DMEM media with 10% FBS, when the input
concentration was 2 ꢀM of 1. At 4 °C, the cellꢀbound concentration of 1 was 1 nmol/million
MDAꢀMBꢀ231 cells for 6 h incubation at 4 °C in 5 mL of DMEM media with 10% FBS.
Therefore, only 15% of 1 bound with the cell membrane by the nonspecific manner.
Four cancer cells were examined in cellꢀbased studies. SW480 and HCT116 cells have
hyperactive Wnt signaling. SW480 cells harbor the deletion of APC,77 whereas HCT116 cells
harbor the heterozygous threeꢀbase deletion of CTNNB178 that blocks βꢀcatenin phosphorylation
and ubiquitination. Although the expression levels of βꢀcatenin in MDAꢀMBꢀ231 cells are
low,79–81 MDAꢀMBꢀ231 cells have high expression levels of upstream Wnt effectors, lowꢀ
26
ACS Paragon Plus Environment

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
density lipoproteinꢀrelated proteins 5 and 6 (LRP 5/6),81,82 frizzled (Fzd),83,84 and dishevelled.79
MDAꢀMBꢀ231 cells are susceptible to the activation by Wnt ligands to promote cancer cell
growth in vitro and in vivo.83,85 βꢀCatenin siRNA knocks down βꢀcatenin expression in MDAꢀ
MBꢀ231 cells and blocks cancer cell migration and invasion.80,86 These three cancer cell lines
were chosen based on their dependence on βꢀcatenin for growth and survival. On the other hand,
A549 cells do not contain active Wnt/βꢀcatenin signaling.87,88
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 this study, the tetrazolium MTS assay was used to examine the inhibitory effects of new
compounds on cancer cell viability. The MTS assay, however, cannot distinguish the effects on
cell death and cell growth inhibition. LDH release was used as the preferred biomarker to
examine the effect on cell death because LDH is a soluble cytoplasmic enzyme that is present in
all examined cells of this study. It is released to the extracellular space when the cytoplasmic
membrane is damaged. Through the combinational use of the MTS (compounds were incubated
with cancer cells for 72 h) and LDH release (compounds were incubated with cancer cells for 4
h) assays, the effects of these compounds on cytotoxicity and on inhibition of the viability of
cancer cells can be differentiated.
Compounds 2–18 displayed high activities in MTS assays. However, they also exhibited
cytotoxicity in the LDH release assay. They are not suitable for the further inhibitor
optimization. In contrast, compounds 24–30 inhibited growth of cancer cells in the MTS assays
but had no effect on LDH release activity, indicating that the future inhibitor optimization should
take into consideration decreasing compound hydrophobicity, and compound 29 can be used as
the new starting point. Compound 29 also exhibits selectivity on inhibiting the viability of Wntꢀ
activated cancer cells over the cancer cells that have normal Wnt signaling. It is interesting that
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 70
1
2
3
4
5
6
the IC50 of 29 for the inhibition of TOPFlash luciferase reporter activity is lower than that for the
MTS assay.
7
8
9
Similar to carnosic acid35 and 147, compounds 3 and 29 also selectively decrease the level of
active form of βꢀcatenin (i.e., nuclear βꢀcatenin) without affecting the level of total βꢀcatenin
(i.e., Eꢀcadherinꢀbound βꢀcatenin). With the result that 29 disrupts the βꢀcatenin/BCL9
interaction in coꢀIP experiments (Figures 8C,D), there are at least two possible mechanisms why
29 promotes the selective degradation of nuclear βꢀcatenin. One is that 29 disrupts the interaction
of βꢀcatenin with BCL9, and free βꢀcatenin is susceptible to the H1 (the first αꢀhelix of βꢀcatenin
armadillo repeats)ꢀdependent proteasomal degradation35. The second mechanism is that 29 binds
with βꢀcatenin after the disruption of the PPI, destabilizes βꢀcatenin tertiary structure, and
promotes βꢀcatenin degradation. The further mechanistic studies are required to address this
question.
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
Although 29 is more potent and selective than 1 in biochemical assays, it should be noted
that 29 still has multiple carboaromatic rings, which are often considered to be the potential risk
of the high serum albumin binding, and the inhibition of cytochrome P450 and human etherꢀaꢀ
goꢀgoꢀrelated gene product (hERG), potentially lowering the developability of the compound.89–
92
In addition to improving inhibitor potency and maintaining inhibitor selectivity, the further
optimization will be centered on the replacement of the aromatic rings with the aliphatic
alternatives and the increase of the threeꢀdimensional features of the designed compounds.
CONCLUSION
The Wnt/βꢀcatenin signaling pathway is frequently implicated in tumorigenesis of many
cancers. The pathway is not only involved in the proliferation of cancer cells, but more
28
ACS Paragon Plus Environment

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
importantly responsible for cancer progression, metastasis, and drug resistance. The activation of
βꢀcatenin signaling also blocks recruitment of cytotoxic T cells, promotes Treg cell survival and
infiltration, and excludes anticancer immunity. On the other hand, it has been proven difficult to
discover inhibitors for the downstream effectors of the Wnt/βꢀcatenin signaling pathway. The βꢀ
catenin/BCL9 PPI in the transcriptional complex is an interesting target for chemical probe
development and drug discovery. In this study, structureꢀbased design and synthesis was
conducted to optimize the inhibitors based on the 4'ꢀfluoroꢀNꢀphenylꢀ[1,1'ꢀbiphenyl]ꢀ3ꢀ
carboxamide scaffold. Compound 29 was found to disrupt the βꢀcatenin/BCL9 PPI with a Ki of
0.47 ꢀM and exhibited high selectivity (>1900ꢀfold) for βꢀcatenin/BCL9 over βꢀcatenin/Eꢀ
cadherin PPIs in the biochemical AlphaScreen assays. The siteꢀdirected mutagenesis studies
indicated that βꢀcatenin D145, E155, and K181 are important for the inhibitors to disrupt the βꢀ
catenin/BCL9 PPI. The cellꢀbased studies indicated that 29 can selectively inhibit growth of
cancer cells that have hyperactive Wnt/βꢀcatenin signaling and did not show cytotoxicity in the
LDH release assays. The coꢀIP experiments indicated that this compound disrupted the fullꢀ
length βꢀcatenin/BCL9 PPI while leaving the βꢀcatenin/Eꢀcadherin PPIs unaffected. The
TOPFlash/FOPFlash luciferase reporter assays indicated that 29 can selectively suppress
transactivation of Wnt/βꢀcatenin signaling. This compound also downregulated expression of
Wnt target genes in doseꢀdependent manners at both mRNA and protein levels, and only
inhibited the downstream nodes of the Wnt/βꢀcatenin signaling pathway. This study offered the
directions for future inhibitor optimization.
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
29
ACS Paragon Plus Environment