4(1H)-quinolone derivatives overcome acquired resistance to anti-microtubule agents by targeting the colchicine site of β-tubulin

Article abstract of DOI:10.1016/j.ejmech.2019.111584

Developing new therapeutic strategies to overcome drug resistance of cancer cells is an ongoing endeavor. From among 2 million chemicals, we identi?ed ethyl 4-oxo-2-phenyl-1,4-dihydroquinoline-6-carboxylate (AS1712) as a low-toxicity inhibitor of lung can

Full text of DOI:10.1016/j.ejmech.2019.111584

Journal Pre-proof
4(1H)-quinolone derivatives overcome acquired resistance to anti-microtubule agents
by targeting the colchicine site of β-tubulin
Ming-Shiu Lin, Tse-Ming Hong, Ting-Hung Chou, Shuenn-Chen Yang, Wei-Chia
Chung, Chia-Wei Weng, Mei-Ling Tsai, Ting-Jen Rachel Cheng, Jeremy J.W. Chen,
Te-Chang Lee, Chi-Huey Wong, Rong-Jie Chein, Pan-Chyr Yang
PII:
S0223-5234(19)30718-4
DOI:
https://doi.org/10.1016/j.ejmech.2019.111584
EJMECH 111584
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 14 June 2019
Revised Date: 30 July 2019
Accepted Date: 1 August 2019
Please cite this article as: M.-S. Lin, T.-M. Hong, T.-H. Chou, S.-C. Yang, W.-C. Chung, C.-W. Weng,
M.-L. Tsai, T.-J. Rachel Cheng, J.J.W. Chen, T.-C. Lee, C.-H. Wong, R.-J. Chein, P.-C. Yang, 4(1H)-
quinolone derivatives overcome acquired resistance to anti-microtubule agents by targeting the
colchicine site of β-tubulin, European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/
j.ejmech.2019.111584.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Masson SAS.

1
2
4(1H)-Quinolone Derivatives Overcome Acquired Resistance to Anti-microtubule Agents by
Targeting the Colchicine Site of β-Tubulin
3
4
Ming-Shiu Lin1, 2, Tse-Ming Hong3, Ting-Hung Chou4, Shuenn-Chen Yang2, Wei-Chia Chung2, Chia-
Wei Weng5, Mei-Ling Tsai4, Ting-Jen Rachel Cheng6, Jeremy J. W. Chen5, Te-Chang Lee1, 2, Chi-
5
6
Huey Wong6, Rong-Jie Chein4, *, and Pan-Chyr Yang1, 2, 7,
*
7
8
1Program in Molecular Medicine, National Yang-Ming University and Academia Sinica, Taipei,
Taiwan
9
10
11
12
13
2Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan
3Institute of Clinical Medicine, National Cheng Kung University, Tainan 701, Taiwan
4Institute of Chemistry, Academia Sinica, Taipei 115, Taiwan
5Institute of Biomedical Sciences, National Chung Hsing University, Taichung 402, Taiwan
6The Genomics Research Center, Academia Sinica, Taipei 115, Taiwan
14
15
7Department of Internal Medicine, College of Medicine, National Taiwan University, Taipei 100,
Taiwan
16
17
18
19
*Corresponding author
E-mail addresses: pcyang@ntu.edu.tw (P.C. Yang); rjchein@chem.sinica.edu.tw (R.J. Chein)
ABSTRACT: Developing new therapeutic strategies to overcome drug resistance of cancer cells is an
ongoing endeavor. From among 2 million chemicals, we identified ethyl 4-oxo-2-phenyl-1,4-
dihydroquinoline-6-carboxylate (AS1712) as a low-toxicity inhibitor of lung cancer cell proliferation
and xenograft tumor growth. We show that AS1712 is active against broad cancer cell lines and is
able to bind in the colchicine-binding pocket of β-tubulin, thereby inhibiting microtubule assembly
and, consequently, inducing mitotic arrest and apoptosis. Our cell-based structure-activity relationship
study identified a new lead compound, RJ-LC-15-8, which had a greater anti-proliferative potency for
H1975 cells than did AS1712, while maintaining a similar mechanism of action. Notably, AS1712
and RJ-LC-15-8 overcame P-glycoprotein efflux pump and β-tubulin alterations that lead to acquired
resistance against microtubule-targeting drugs of cancer cells. AS1712 and RJ-LC-15-8 may be lead
compounds that overcome acquired resistance to microtubule-targeting agents of cancer cells.
20
21
22
23
24
25
26
27
28
29
30
KEYWORDS: microtubule-targeting agents, acquired resistance, p-glycoprotein
1

1
Introduction
Cancer is the second most common cause of death worldwide, and lung cancer is the leading
2
3
4
5
6
cause of cancer-related deaths [1]. Despite advances in treatment during the past 20 years, the
prognosis for lung cancer patients remains poor [2]. Drug resistance limits the efficacy of anticancer
agents and is responsible for treatment failure in most cases [3-5]. Consequently, the development of
new drugs to overcome drug resistance and improve survival of cancer patients is urgently needed.
7
8
Microtubules are long, hollow cylinders composed of heterodimeric α/β-tubulin units. They are
an important component of the cell cytoskeleton and are involved in mitosis, motility, and organelle
transport [6-8]. Cancer cells depend on microtubules for mitosis and rapid division, which makes
microtubules a suitable target for anticancer agents [8-10]. Nucleation of microtubules occurs at the
centrosome, and, as they are highly dynamic structures that continuously assemble and disassemble,
microtubule-targeting agents (MTAs) can regulate their dynamics and, consequently, cause mitotic
arrest and subsequently cell apoptosis [10-12]. MTAs, i.e., taxanes, vinca alkaloids, colchicine, and
laulimalide, are classified according to their mode of action and tubulin-binding site [13]. The taxanes
(e.g., paclitaxel and ixabepilone) and the vinca alkaloids (e.g., vincristine and vinblastine) are the two
major types of MTAs that are widely used clinically to treat many different types of malignant
cancers, including those of the lung, breast, ovary, prostate, and head and neck and leukemia [14, 15].
However, cancers in most patients who have been treated for prolonged periods with MTA become
acquired resistant to the treatment protocol [3, 16].
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
The two major mechanisms of MTA-induced resistance involve mutations in β-tubulin and ATP-
binding cassette efflux transporters [17, 18]. A structurally altering mutation at a β-tubulin drug-
binding site or expression of a β-tubulin isotype with an altered conformation may interfere with the
interaction between an MTA and its binding site on β-tubulin, thereby causing treatment failure [19,
20]. Efflux transporters embedded in cell membranes may pump different types of anti-cancer drug
from the intracellular to the extracellular space, which decreases intracellular drug accumulation [21],
resulting in multidrug resistance (MDR) and making it more difficult to treat patients with cancer [22-
24]. P-glycoprotein (P-gp) was the first identified ATP-binding cassette efflux pump and has been
2

1
2
3
well characterized. MDR caused by P-gp overexpression is the most common resistance mechanism
known involving clinical cancer therapies [25-27], and, therefore, finding new agents that will
improve patient outcomes and circumvent drug resistance remains a major and unmet need.
4
5
In this study, we performed a two-step high-throughput screening of 2 million compounds to
search for novel molecules with therapeutic activity against lung cancer. Ethyl 4-oxo-2-phenyl-1,4-
dihydroquinoline-6-carboxylate (AS1712) was found to have anticancer activity in vitro and in vivo
and a superior therapeutic specificity for cancer instead of normal cells. We found that AS1712
inhibited the growth of various types of cancer cells, including those of the lung, prostate, ovary,
breast, colon, and head and neck and leukemia. By targeting the colchicine-binding site of β-tubulin,
AS1712 induced mitotic cell cycle arrest and apoptosis. Cell-based structure-activity relationship
study identified a new lead compound, RJ-LC-15-8, which had a greater anti-proliferative potency
and maintained a similar mechanism of action. Notably, the distinct tubulin-binding site of AS1712
and RJ-LC-15-8 circumvented resistance caused by P-gp overexpression.
6
7
8
9
10
11
12
13
14
Results
15
16
AS1712 Inhibited Lung Cancer Cell Proliferation In Vitro and In Vivo and Had Broad Anti-
Cancer Activity
17
18
19
20
21
22
23
24
25
Through high-throughput screening, a library of 2 million compounds was used to evaluate anti-
proliferation activities against EGFR tyrosine kinase inhibitor (TKI)-resistant lung cancer cells
(H1975) at 10 µM concentration. The coefficient of variation (CV%) and Z' values of the primary
screening were determined as 4.9% and 0.58, respectively. In the first screening, ~6,800 small
molecules exhibited >80% growth inhibition of H1975 cells. These compounds were further screened
against other eight lung cancer cells, including clinical isolates (CL25, CL83, CL97, CL100, CL141,
CL152), an ATCC cell line (PC-9) and its derivative resistant clone (PC9/IR). By filtered with the
criteria of IC50 ≤ 6 µM for all tested cells, only 232 small molecules were kept for the further study.
Ethyl 4-oxo-2-phenyl-1,4-dihydroquinoline-6-carboxylate (AS1712, Figure 1A) was one of these
3

1
2
molecules displaying superior cytotoxic e ects on lung cancer cells and a notable therapeutic window
between cancer cells and normal cells.
3
4
We first evaluated the cytotoxic effects of AS1712 on different non-small cell lung cancer cell
lines using an MTS assay and found that AS1712 induced cell death in different EGFR-status cell
lines, including A549 (wild-type EGFR), CL1-0 (wild-type EGFR), PC9 (EGFR exon 19 deletion),
and H1975 (T790M and L858R mutations in EGFR). The IC50 values for the affected cell lines ranged
from 16 to 33 nM (Figure 1B), which indicated that the cytotoxic effect of AS1712 was not related to
the EGFR status of lung cancer cells. Furthermore, AS1712 was relatively non-toxic against normal
bronchial epithelial (NBE) and human fibroblast (HFB) cells (IC50 > 10 µM; Figure 1B). The colony-
forming abilities of CL1-0 and H1975 cells were suppressed in a dose-dependent manner by AS1712
(Figure 1C). The aforementioned data suggested that AS1712 inhibited non-small cell lung cancer
proliferation in vitro.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
To evaluate the in vivo anti-tumor efficacy of AS1712, athymic BALB/c nude mice each bearing
an established subcutaneous H1975 tumor were intraperitoneally injected with DMSO (control) or
with 4 or 8 mg/kg AS1712, three times a week for 25 days (n = 6 per group). The tumor volume and
body mass of each mouse were monitored throughout the treatment period. Treatment with AS1712
markedly reduced H1975 xenograft tumor growth compared with that found for the control group
(average tumor size, 1749 ± 234.5 mm3 for DMSO treatment; 871.3 ± 106 mm3 for the 4 mg/kg and
587.2 ± 67 mm3 for 8 mg/kg treatments; all measured on day 25 and both p < 0.01; Figures 1D and
S1). The body masses and serum biochemical markers for liver and kidney functions, including ALT,
AST, BUN, and Cre, were non-toxic a ected on day 25 of the experiment (Figures S1D, E). Thus
AS1712 inhibited tumor growth with low toxicity in vivo.
4

1
2
3
Figure 1. AS1712 inhibited cancer cell proliferation in vitro and in vivo. (A) Chemical structure of
AS1712. (B) The cytotoxic effects of AS1712 on different non-small cell lung cancer and normal lung
cell lines. Cells were treated with AS1712 for 72 h, and cell viability was examined by the MTS assay.
Data are the mean of three determinations. (C) CL1-0 and H1975 cells were treated with AS1712
(12.5−200 nM) for 14 days. The colonies were fixed, stained with crystal violet, and then counted.
Data are the mean ± SD; *p < 0.05. (D) H1975 cells (2 × 106) were subcutaneously injected into
BALB/c nude mice. Mice were treated with DMSO or with 4 or 8 mg/kg of AS1712 intraperitoneally
three times a week for the next 25 days. Tumor volumes were measured before each injection. Data
are the mean ± SEM; *p < 0.05.
4
5
6
7
8
9
10
11
12
13
14
We also assessed the cytotoxicity of AS1712 against various types of cancer cells and found that
AS1712 decreased the viability of many different types of malignant cells, including those derived
from lung (A549, CL1-0, PC9, and H1975), breast (BT-549, Hs578t, MDA-MB-231, 37T, and 82T),
ovarian (SKOV-3, IGR-OV1, and Ovcar-3), head and neck (SAS, OECM1, and KB), colon (HCT116,
5

1
2
3
4
HT-29, Colon205, and SW620), and prostate (Du-145 and PC-3) cancers and from leukemia (Molt4
and CCRF-CEM). For the 23 cancer cell lines tested, their IC50 values ranged from 16 to 84 nM
(Table 1), suggesting that AS1712 or a derivative of AS1712 might be used for the treatment of
various types of cancer.
5
6
Table 1. Inhibitory Activities of AS1712 against the Proliferation of different Cancer Cell lines or
Normal cell lines.
Cell line
IC50 (nM)
Cell line
IC50 (nM)
Lung cancer
Head and Neck cancer
A549
CL1-0
PC9
20 ± 0.98
16 ± 2.48
17 ± 2.09
33 ± 4.86
SAS
OECM1
KB
78 ± 10.4
52 ± 1.94
37 ± 0.27
H1975
Colon cancer
HCT116
HT-29
50 ± 0.81
48 ± 0.62
57 ± 5.17
Breast cancer
BT-549
Hs578t
MDA-MB-
231
33 ± 1.51
41 ± 2.87
Colon205
55 ± 6.83
SW620
35 ± 1.18
37T
34 ± 2.99
35 ± 0.67
Prostate cancer
82T
Du-145
PC-3
47 ± 1.28
45 ± 1.77
Ovarian cancer
SKOV-3
IGR-OV1
Ovcar-3
50 ± 0.32
84 ± 2.07
28 ± 1.08
Leukemia
Molt4
38 ±1.03
45 ± 0.45
CCRF-CEM
Normal cell line
NBE
>10 µM
HFB
>10 µM
7
8
AS1712 Induced Apoptosis and Cell Cycle Arrest
9
To evaluate whether the cytotoxic effect of AS1712 correlated with apoptotic cell death, CL1-0,
10
H1975, and HFB cells were treated with AS1712 for 24 to 72ꢀh and then monitored for the activation
11
12
13
14
of major pro-apoptotic proteins. We found that AS1712 induced cleavage of poly(ADP-ribose)
polymerase (PARP) and caspase-3 in CL1-0 and H1975 cells in a time-dependent manner; PARP and
caspase-3 in the normal HFB cells were not affected by AS1712 (Figure S2A). We then looked for
apoptotic pathway proteins that might have been activated by AS1712 and found that AS1712
6

1
2
3
4
5
6
induced two intrinsic apoptosis pathway markers, cleavage of caspase-9 and expression of Bax, in a
dose-dependent manner (Figure 2A). Conversely, the extrinsic apoptosis pathway marker, caspase-8,
was not activated by AS1712 treatment (Figure 2A). We also found that a 24-h treatment with
AS1712 increased the cytosolic level of cytochrome c (Figure 2B), whereas the same dose of AS1712
did not induce apoptosis in HFB cells or cytochrome c release (Figures S2B and C). AS1712 therefore
induced activation of the intrinsic apoptosis pathway only in cancer cells.
7
8
Next, we explored whether the apoptotic effect of AS1712 was caused by aberrant cell cycle
progression. The distribution of CL1-0, H1975, and HFB cells among phases of the cell cycle was
examined by flow cytometry. AS1712 treatment increased the sub-G1 (apoptotic cell population) and
the G2/M-phase population of CL1-0 and H1975 cells in a dose-dependent manner (Figure 2C) but
not of HFB cells (Figure S2D). To clarify whether AS1712 induced G2 or mitotic cell cycle arrest, we
characterized the cell cycle at its checkpoints for CL1-0 and H1975 cells and found that AS1712
increased the mitotic-phase markers cyclin B1 and phospho-histone H3 (p-HH3) but not the G2-phase
marker cyclin A (Figure 2D). We also used high-content imaging to measure the mitosis index (the
population of mitotic cells) by staining for the specific mitotic phase marker p-HH3 and found that
AS1712 treatment caused CL1-0 and H1975 cells to accumulate in the mitotic phase in a dose-
dependent manner (Figure 2E). Furthermore, we directly evaluated the mitotic-spindle organization
using α-tubulin immunofluorescence staining and observed that arrangement of the mitotic spindles
was disrupted by AS1712 treatment (Figure 2F). The aforementioned data showed that AS1712
induced mitotic cell cycle arrest.
9
10
11
12
13
14
15
16
17
18
19
20
7

1
2
3
Figure 2. AS1712 induced apoptosis and cell cycle arrest in CL1-0 and H1975 lung cancer cells.
Cells were treated with the indicated concentrations of AS1712 for 24 h. Then, (A) pro-apoptotic
protein expression was detected by immunoblotting; (B) cytochrome c levels were assessed in the
cytosolic and mitochondrial fractions; (C) cells were stained with propidium iodide to determine their
DNA content by flow cytometry; (D) proteins were immunoblotted to examine their expression levels;
and (E) cells were subjected to high-content imaging by p-HH3 staining to calculate the population of
mitotic cells. Data are the mean ± SD; *p < 0.05. (F) Cells were treated with AS1712 for 6 h to
observe mitotic spindle organization by α-tubulin staining (green). Immunofluorescence was detected
by confocal laser microscopy. Nuclei were detected by DAPI staining (blue). Scale bar, 10 µm.
4
5
6
7
8
9
10
8

1
AS1712 Directly Inhibited Microtubule Polymerization
2
3
The observation that AS1712 treatment disrupted mitotic-spindle organization led us to
hypothesize that the treatment might interfere with dynamic microtubule association and/or
dissociation. Therefore, we performed an immunofluorescence assay to examine the status of
microtubule networks in AS1712-treated CL1-0, H1975, and HFB cells. AS1712 treatment of CL1-0
and H1975 cells resulted in depolymerization of their microtubules (Figure 3A), whereas such
treatment had only a minor effect on microtubules in HFB cells (Figure S2E). An in-cell microtubule
assembly assay was performed, after first separating assembled and disassembled microtubules into
particulate and soluble fractions, respectively, by centrifugation. Notably, AS1712 substantially
increased the amount of α-tubulin in the soluble fraction, a finding similar to that found for
nocodazole but not paclitaxel treatment (Figure 3B). We further confirmed that AS1712 induced
depolymerization of microtubules using an immunofluorescence-based microtubule polymerization
assay. H1975 cells were first treated with 40 nM AS1712 for 1 h, were then exposed to the cold for 30
min, and then were shifted to 37°C for various times. When treated with AS1712, the microtubules
failed to polymerize at 37°C (Figure 3C). Next, an in vitro tubulin polymerization assay was used
with purified tubulin to examine whether AS1712 directly interfered with microtubule assembly in a
cell-free system. As expected, the reference agents nocodazole and vincristine were found to inhibit
tubulin polymerization, whereas paclitaxel enhanced polymerization. AS1712 was also found to
directly inhibit tubulin polymerization (Figure 3D). We also carried out an in vitro microtubule
nucleation assay to test whether AS1712 directly inhibited the assembly of new microtubules at
purified centrosomes. Purified centrosomes and tubulins were incubated with AS1712 or DMSO.
AS1712 substantially inhibited the formation of microtubule asters from the centrosome (Figure 3E,
left panel). A quantitative analysis of microtubule fiber length also showed a dramatic difference
between AS1712 and DMSO treatments (Figure 3E, right panel). The aforementioned data showed
that AS1712 directly inhibited microtubule polymerization.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
9

1
2
3
4
5
6
7
8
Figure 3. AS1712 inhibited microtubule polymerization. (A) Cells were incubated with AS1712
for 6-h prior to examining the microtubule networks as α-tubulin staining (green). Nuclei were
detected by DAPI staining (blue). Scale bar, 50 µm. (B) Cells were incubated with DMSO, AS1712,
paclitaxel, or nocodazole for 6-h and then harvested and separated into soluble (tubulin monomers)
and insoluble (tubulin polymers) fractions. The amount of α-tubulin was detected by immunoblotting.
(C) Immunofluorescence-based microtubule polymerization assay H1975 cells were treated with 40
nM AS1712 for 1 h, after cold exposure for 30 min, and then were incubated at 37 °C for the
10

1
2
3
4
5
6
indicated times. Scale bar, 50 µm. (D) Purified porcine tubulin and GTP were incubated at 37 °C with
5 µM AS1712, 5 µM paclitaxel, 5 µM vincristine, 5 µM nocodazole, or with DMSO control.
Microtubule polymerization was assessed every min for 1 h by monitoring the A340. (E) Purified
centrosomes were first incubated with AS1712 and then with tubulin. Microtubule asters that formed
were stained with monoclonal anti-α-tubulin (green). Fiber lengths were measured (n = 20 asters per
group) Data are the mean ± SD. Scale bar, 50 µm.
7
Scheme 1. Synthesis of AS1712 analogs 2a-2m
8
9
10
11
A Cell-Based Structure-Activity Relationship (SAR) Study of AS1712 Led to the Identification
of the More Potent Anti-Cancer Compound RJ-LC-15-8
12
13
14
15
To find a more potent AS1712-based compound with increased anti-cancer activity, we carried
out an SAR study using H1975 cells and AS1712 analogs, which possessed the 4(1H)-quinolone
scaffold. As shown in Scheme 1A and Table 2, AS1712 was first hydrolyzed to the acid 2a, which
had an IC50 > 5 µM. Next 2a was transformed into the ethyl amide 2b, which had an IC50 = 3.7 µM.
11

1
2
Both analogs had substantially less activity against H1975 cells then did AS1712. We then focused on
structural optimization of the phenyl group at the C-2 position of the 4(1H)-quinolone scaffold. The
effects of substituents at the meta or para positions of the aryl ring were explored using 2c−2j, which
were readily synthesized by reactions involving ethyl 3-aryl-3-oxypropanoates and ethyl 4-
aminobenzoates, and cyclization of the resulting β-enamino esters 1 in hot diphenyl ether (Scheme
1B). Addition of an electron-donating group (2c, IC50 = 2.2 µM), -withdrawing groups (2d, IC50 = 2.1
µM and 2e, IC50 = 4.5 µM), or a hydroxyl group (2f, IC50 = 475 nM) at the para position of the phenyl
group resulted in substantial decreases in inhibitory activity against H1975 cells. Interestingly, while
exploring the effects of meta substituents, we found that 2h possessing a meta-fluorophenyl moiety
had increased inhibitory activity (IC50 = 24 nM) against H1975 cells, whereas 2i and 2j, which bore a
trifluoromethyl and a hydroxyl group respectively at the meta position of the phenyl ring showed
substantially decreased potencies (IC50 = 780 nM and 1.95 µM, respectively) compared with that of
AS1712. The introduction of a meta-methoxy group (2g) on the tail aryl ring slightly reduced activity
(IC50 = 44 nM). The effects of substituents at the 6-position of the 4(1H)-quinolone scaffold were also
explored by replacing the ethyl group with less or more sterically bulky groups (2k, IC50 = 207 nM
and 2l, IC50 = 48 nM, respectively) or by removal of the oxo group (2m, Scheme 1C, IC50 = 810 nM).
The results showed that the original ethyl carbonate is the optimal substituent at the C-6 position.
Overall, the data gained from the SAR study suggested that 2h (denoted RJ-LC-15-8) possessed a
greater anti-cellular proliferation potency (IC50 = 24 nM) against H1975 cells than did AS1712 and,
therefore, deserves to be further developed into a drug for the treatment of cancers with MDR.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
12

1
2
Table 2. Cell-based SAR study of AS1712 for the inhibition of H1975 cells
Compound
R1
R2
IC50
33 ± 4.86 nM
>5 µM
AS1712
H
COOEt
H
H
COOH
CONHEt
COOEt
COOEt
COOEt
2a
2b
2c
2d
2e
3.7
2.2
±
±
0.73µM
0.20µM
4’-OMe
4’-F
2.1 ± 0.21 µM
4’-Br
4.5
±
0.90µM
31.6 nM
4’-OH
3’-OMe
3’-F
COOEt
COOEt
COOEt
COOEt
COOEt
COOMe
COOiPr
CH2OEt
2f
475
±
2g
44 ± 1.39 nM
24 ± 1.12 nM
2h(RJ-LC-15-8)
3’-CF3
3’-OH
H
2i
2j
780
1.95 ± 0.20 µM
207 39.1nM
48 ± 5.1 nM
810 0.12 nM
±
12.1 nM
2k
2l
±
H
H
2m
±
3
4
5
RJ-LC-15-8 Inhibited Tumor Growth Using a Mechanism Similar to AS1712 In Vitro and In
Vivo
6
7
To verify that the mechanism of action of RJ-LC-15-8 is similar to AS1712, we performed
additional experiments with H1975 cells. Immunoblotting showed that RJ-LC-15-8 induced cleavage
of PARP and increased the expression of the mitotic phase markers cyclinB1 and p-HH3 (Figure
S3A). In situ immunofluorescence of α-tubulin staining showed that RJ-LC-15-8 treatment inhibited
microtubule polymerization (Figure S3B). The in vitro tubulin polymerization assay also showed
direct inhibition by RJ-LC-15-8 of microtubule polymerization (Figure S3C). Next, we evaluated the
anti-tumor ability of RJ-LC-15-8 using the H1975 xenograft model, and the data showed that, in
8
9
10
11
12
13

1
2
comparison with the control group, RJ-LC-15-8 inhibited tumor growth (average tumor size, 2153 ±
126.3 mm3 for the DMSO control treatment and 1333 ± 121.9 mm3 for the 4 mg/kg and 972.5 ± 52.3
mm3 for the 8 mg/kg RJ-LC-15-8 treatments on day 21; both p < 0.01; Figure S3D). At the end of the
experiment, we found that RJ-LC-15-8 treatment had not affected body mass (Figures S3E). We also
evaluated the cytotoxicity of RJ-LC-15-8 against triple-negative breast cancer (TNBC). We found RJ-
LC-15-8 was more effective than AS1712 in different TNBC cell lines, with IC50 values ranging from
21 to 35 nM (Figures S3F). The MDA-MB-231 xenograft model showed RJ-LC-15-8 could inhibit
TNBC tumor growth in comparison with the control group (average tumor size, 218 ± 20.9 mm3 for
the DMSO control treatment and 105 ± 9.4 mm3 for the 4 mg/kg and 73.6 ± 8.2 mm3 for the 8 mg/kg
RJ-LC-15-8 treatments on day 35; both p < 0.05; Figure S3G). RJ-LC-15-8 treatment did not affect
body mass in MDA-MB-231 xenograft model (Figures S3H).
3
4
5
6
7
8
9
10
11
12
Scheme 2. Conjugate magnetic beads on AS1712.
13
14
15
AS1712 Binds the Colchicine-Binding Pocket of β-Tubulin
16
17
18
Inhibition by AS1712 of microtubule polymerization led us to hypothesize that AS1712 directly
targeted tubulin. To provide evidence for this hypothesis, we performed a pull-down assay using
magnetic beads conjugated with AS1712. First, the alkyne 4 was synthesized via a three-step
14

1
2
procedure from ethyl 3-(3-hydroxyphenyl)-3-oxopropanoatealkyne (Scheme 2A) and then coupled to
magnetic beads equipped with polyethylene glycol12-azido linkers to generate “mag-beads-AS1712”
(Scheme 2B). After incubating H1975 lysates separately with mag-beads-AS1712 or mag-beads-
control overnight at 4 °C, immunoblotting showed that the mag-beads-AS1712 had pulled down only
β-tubulin (Figure 4A). Next, we performed a surface plasmon resonance study to measure the value of
the dissociation constant (KD) for the tubulin-AS1712 complex. Different concentrations of AS1712
were injected over a tubulin-immobilized sensor chip surface, and the formation of complexes was
monitored. The number of resonance units, which reflected the amount of complex formed, increased
in a dose-dependent manner (Figure 4B). The KD value for the complex was 13.8 µM as determined
by fitting the data to a steady-state model. The data indicated that AS1712 directly interacted with
tubulin.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
We further characterized the AS1712-binding site on tubulin using a limited proteolysis assay.
Two different tubulin-binding sites for microtubule-destabilizing agents are known: the vinca site and
the colchicine site. When a drug binds to tubulin, it induces conformational changes that can be
probed by a limited trypsin-digestion assay [28] (Figure 4C). The colchicine-site agent nocodazole
caused β-tubulin to unfold and produced an increased amount of βcol after a limited trypsin digestion.
In contrast, the vinca-site agent vincristine stabilized the α- and β-tubulin folds and decreased the
amount of the βcol, αN, and αC fragments produced in comparison with the control digest. The tryptic
pattern for the tubulin sample exposed to AS1712 was similar to that of nocodazole (Figure 4C). A
competition assay was performed to confirm that AS1712 binds to the colchicine site. The binding
site competition between the test compound and colchicine will decrease the intrinsic fluorescence of
colchicine-tubulin complex by reducing the amount of colchicine bound [29]. The presence of
AS1712 decreased the colchicine fluorescence in a dose-dependent manner as did nocodazole,
whereas vincristine did not (Figure 4D).
25
26
27
We performed a molecular modeling study to explore the binding mode of AS1712 in the
colchicine pocket of a heterodimeric tubulin crystal structure [30]. The orientation of colchicine,
which was the original ligand in the crystal structure, was used as a reference. The binding pose of
15

1
2
3
4
5
6
7
8
AS1712 was generated by molecular docking into the colchicine binding pocket could be
superimposed onto the colchicine structure with an RMSD value of 0.383 Å (Figure 4E). The docking
energy score for the AS1712 complex (–32.8706 kcal/mol) was higher than that for the colchicine
complex (–4.47969 kcal/mol). Various types of bonds were found between AS1712 and the α/β-
tubulin dimer within a distance of 7 Å (Figure 4F), including hydrogen bonds between AS1712 and
Ala250 and Lys254 of β-tubulin and Gln11 of α-tubulin, as well as pi-bonds with Leu248, Lys254,
Met259, Ala316, and Lys352 of the β- tubulin (Figure 4G). These results suggested that AS1712
binds to β-tubulin at its colchicine-binding pocket.
16

1
2
3
4
5
Figure 4. AS1712 binds to the colchicine-binding pocket of β-tubulin. (A) H1975 cell lysates
were incubated with mag-beads-control or -AS1712 overnight and precipitated proteins were
detected by immunoblotting. (B) Binding affinity of AS1712 to tubulin as measured by surface
plasmon resonance. (C) Tubulin (1 mg/mL) was incubated with 100 µM of nocodazole, vincristine,
17

1
2
AS1712, or DMSO control for 30 min at 30 °C. Then, the samples were subjected to trypsin digestion
for 10 min on ice. The positions of the tryptic cleavage products, αN, αC, and β‑col, are shown to the
3
4
5
6
7
right of the gel. Data are the mean ± SD. (D) Colchicine competition assay. AS1712 and nocodazole
decreased the fluorescence of the colchicine-tubulin complex. Data are the mean ± SD. (E)
Superimposition of the docked AS1712-tubulin (orange) and colchicine-tubulin complexes (purple).
(F) Interactions between the tubulin heterodimer and AS1712. (G) Schematic of the AS1712 and
tubulin-residue interactions.
8
9
AS1712 and RJ-LC-15-8 Overcame P-Glycoprotein−Mediated Multidrug Resistance
10
11
12
13
14
15
16
17
18
19
20
21
22
A major mechanism of MTA-induced MDR is P-gp overexpression, which is also the most
common resistance mechanism that is clinically relevant to cancer therapy. Therefore, we tested the
cytotoxic effects of AS1712 and RJ-LC-15-8 on the KBtax, KBvin, and CEM/VBL cancer cell lines,
which overexpress P-gp. The KBtax and KBvin cell lines were developed by exposure of the human
oral cancer cell line KB to paclitaxel and vincristine, respectively [31]. CEM/VBL cells were
developed by exposure of the human T-lymphoblastic leukemia cell line CCRF-CEM to vinblastine
[31]. The resistance factor (RF) values for the KBtax, KBvin, and CEM/VBL cells against paclitaxel
were 10, 70, and 214, respectively, and the values against vincristine were 32, 163, and 440,
respectively (Table 3). These results confirmed that these cell lines were resistant to paclitaxel and
vincristine. By contrast, AS1712 and RJ-LC-15-8 were cytotoxic to these cell lines, and their effects
on these cells were equal to or greater than those on the parent cells (Table 3). Notably, ixabepilone, a
second-line MTA used for advanced breast cancer that no longer responds to available
chemotherapies, did not effectively inhibit the viability of these multidrug resistant cell lines (Table 3).
23
24
25
18

1
2
Table 3. The cytotoxicity of AS1712 and RJ-LC-15-8 against KB, CCRF-CEM cells and their
multidrug resistant sublines
CCRF-
CEM
KB
KBtax
KBvin
CEM/VBL
Compound
RFa
RF
RF
IC50 (nM)
37 ± 0.27
IC50 (nM)
30 ± 0.65
IC50 (nM)
37 ± 0.77
IC50 (nM)
45 ± 0.45
IC50 (nM)
44 ± 0.98
AS1712
RJ-LC-15-8
Paclitaxel
0.81
0.88
10
1
1
0.97
0.8
34 ± 0.39
10 ± 0.32
5 ± 0.31
8 ± 0.37
30 ± 1.35
101 ± 5.52
49 ± 2.96
35 ± 0.71
35 ± 0.88
708 ± 46.9
817 ± 47.4
193 ± 24.2
20 ± 1.71
4 ± 0.02
3 ± 0.04
3 ± 0.06
16 ± 0.52
857 ± 9.89
1320 ± 20.6
150 ± 6.59
70
163
24
214
440
50
Vincristine
32
Ixabepilone
4.3
3
4
aRF, resistance factor (IC50 in resistant cell line/IC50 in parent cell line).
5
6
To assess the activity of RJ-LC-15-8 against tumor viability in vivo, BALB/c mice were grafted
with KBtax cells and then intraperitoneally injected with DMSO, paclitaxel, (15 mg/kg), or RJ-LC-
15-8 (8 mg/kg) three times a week (n = 5 per group). After 21 days of treatment, the RJ-LC-15-8
group showed substantially decreased tumor growth as compared with the paclitaxel and DMSO
groups (average tumor size, 1109 ± 128.3 mm3 for DMSO; 868.7 ± 121.4 mm3 for paclitaxel; and
528.4 ± 101.6 mm3 for RJ-LC-15-8; Figures 5A and S4A). TUNEL staining of the tumor tissues
showed that the RJ-LC-15-8 treatment induced apoptosis (Figure 5B). Body mass and serum
biochemical markers for liver and kidney functions were not negatively affected at the end of the
experiment (Figures S4B and C). Thus RJ-LC-15-8 inhibited the growth of an MDR-type cancer
tumor with low toxicity.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Because RJ-LC-15-8 displayed inhibitory activity against the KBtax cells, we sought to
determine whether its effects were related to P-gp, either through a direct inhibition or because it is
not a P-gp substrate. Therefore, we examined the effects of RJ-LC-15-8 on P-gp efflux activity and its
expression. A rhodamine efflux assay showed that RJ-LC-15-8 did not directly inhibit pump activity
in KBtax cells but that verapamil, a known P-gp inhibitor, did (Figure 5C). Our immunoblotting study
showed that RJ-LC-15-8 did not affect P-gp expression in KBtax cells, whereas it induced cleavage of
the apoptosis markers PARP and caspase-9 and increased the expression of the mitotic phase markers
cyclin B1 and p-HH3 (Figure 5D). In situ immunofluorescence staining of α-tubulin showed that RJ-
19

1
2
3
4
LC-15-8 treatment inhibited microtubule polymerization and disrupted mitotic spindle organization
(Figure 5E). RJ-LC-15-8 induced apoptosis in multidrug resistant cells without interfering directly
with P-gp expression OR activity and, more broadly, the mechanism of cytotoxicity of RJ-LC-15-8
was not related to P-gp.
5
6
7
8
9
Figure 5. RJ-LC-15-8 inhibited KBtax tumor growth and induced apoptosis without interfering
with P-gp. (A) KBtax (1×107) cells were subcutaneously injected into BALB/c nude mice. Mice were
treated with DMSO, paclitaxel (15 mg/kg), or RJ-LC-15-8 (8 mg/kg) intraperitoneally three times a
week for 21 days. The tumor volume was measured before each treatment. Data are shown as the
20

1
2
3
4
5
6
mean ± SEM. (B) Tumor slices were histologically stained with hematoxylin and eosin (H&E), and
the cell apoptosis status was determined by TUNEL staining. (C) Rhodamine efflux assay for P-gp
pumping activity. (D) Samples of KBtax cells were treated with one of the indicated concentrations of
RJ-LC-15-8 for 24 h and then harvested to assess expression of the indicated proteins. (E) The
microtubule networks and mitotic spindle organization in KBtax cells after a 6-h treatment with 30
nM RJ-LC-15-8.
7
Discussion
8
9
Cancer is the second leading cause of death worldwide and afflicts ~40% of the global population.
Lung, breast, colorectal, and prostate cancers are the most common types and account for ~46% of all
cancer deaths, with lung cancer being responsible for the highest number of deaths [1]. MTAs are
widely used to treat advanced cancers, but prolonged treatment with MTAs reduces their clinical
efficacy owing to the development of drug resistance by the cancers [32-34]. Thus, the development
of new agents that improve drug efficacy and overcome drug resistance is an ongoing effort. For the
study reported herein, we showed that AS1712 has the potential to treat various types of cancers and
the ability to overcome MDR caused by MTAs with good efficacy and a good therapeutic window
(Figure 6).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Cancer cells rely on microtubules to advance through mitosis and rapidly divide, which has made
microtubules an important therapeutic target in cancer treatments for >50 years [35]. When MTAs
alter microtubule dynamics, they cause mitotic arrest and trigger mitochondria-mediated intrinsic
apoptosis [29, 36, 37]. We demonstrated that AS1712 inhibited microtubule polymerization and
nucleation and disrupted mitotic spindle organization. AS1712 also induced mitotic arrest and then
triggered release of cytochrome c and cleavage of caspase-9 to activate intrinsic apoptosis in several
cancer cell lines (Figure 6). AS1712 also had good in vivo activity against tumor growth with no
noticeable toxicity. The anti-cancer activities of AS1712 are similar to those of MTAs currently used
clinically, although AS1712 has a different tubulin-binding site.
21

1
2
3
4
5
6
7
8
9
There are at least four binding sites on tubulin with a stabilizing or destabilizing function [38].
The colchicine-binding site is located on β-tubulin at its interface with α-tubulin [39]. Colchicine-
binding site inhibitors (CBSIs) destabilize microtubule assembly and exhibit anti-angiogenesis and
vascular-disruption activities, which are not found for inhibitors of the other functional sites [40, 41].
Unlike taxanes and vinca alkaloids, CBSIs can counteract the effects of overexpression of β-tubulin
isoforms, and MDR mechanisms have little effect on CBSI activity [35, 42, 43]. Our results showed
that AS1712 directly interacts with the colchicine-binding pocket of β-tubulin, indicating that AS1712
should be classified as a CBSI and suggested that, because of its binding site, AS1712 may not be
susceptible to resistance induced by conformational changes in β-tubulin caused by mutations.
10
11
12
13
14
15
Currently, no CBSIs are approved for cancer therapy because colchicine and its derivatives are
toxic to humans [38, 44]. However, AS1712 showed a >100-fold therapeutic window for the cell lines
analyzed here and did not inhibit microtubule polymerization, cell-cycle arrest, or apoptosis in HFB
cells at the highest treatment dosage that was used for the cancer cells. AS1712 also showed no
notable toxicity in the animal model. Thus AS1712 may be a CBSI with low toxicity and has the
potential to act as a lead compound for further development of anti-cancer agents.
16
17
18
19
20
21
22
23
24
25
26
27
MDR is a major cause of treatment failure in various cancer therapies [27]. Overexpression of P-
gp is the main MDR mechanism involved in MTA-induced resistance because most MTAs, including
paclitaxel, vincristine, vinblastine, and docetaxel, are exported from cells by P-gp [17, 45, 46]. P-gp
pumps many different types of cytotoxic drugs out of cancer cells, thereby decreasing intracellular
drug concentrations and resulting in treatment failure [46]. AS1712 inhibited the proliferation of
KBtax, KBvin, and CEM/VBL cells that overexpressed P-gp with IC50 values comparable to those of
their respective parental control cell lines. Our results suggested that AS1712 either directly inhibited
P-gp activity or did not function as a P-gp substrate, thereby against P-gp−overexpressing cancer cells.
Our rhodamine efflux assay and immunoblotting study ruled out the possibility that AS1712 directly
inhibits P-gp pump activity or P-gp expression, which supports the hypothesis that AS1712 is not a P-
gp substrate. Although other P-gp inhibitors, such as verapamil, cyclosporine, and valspodar, can
reverse the MDR phenotype of cancer cells, clinical trials involving these inhibitors showed toxic side
22

1
2
3
4
5
6
7
8
effects and failed to improve clinical outcomes [47, 48]. With its inability to act as a P-gp substrate,
AS1712 provides a different treatment strategy for MDR cancers. Ixabepilone is a second-line MTA
used for treatment of advanced breast cancers that no longer respond to currently available
chemotherapies [49, 50]; however, we found that some MDR cell lines resist treatment by ixabepilone,
which thereby restricts its clinical use. In contrast, AS1712 has the potential to treat various types of
cancers, including those of the lung, breast, ovary, prostate, and head and neck and leukemia that were
first clinically subjected to taxanes or vinca alkaloids [14, 15]. Our results suggest that AS1712 or its
derivative RJ-LC-15-8 may counteract multidrug resistance in various types of cancers.
9
10
Figure 6. Proposed Mechanism of Action of AS1712 on anti-cancer effects.
11
12
Conclusions
13
14
15
16
17
For the study reported herein, we showed that AS1712 directly targets the colchicine-binding site
of β-tubulin to disassemble microtubules, which subsequently induces mitotic arrest and activates
apoptosis. AS1712 can overcome the MTAs-induced acquired resistance, including β-tubulin
alterations and P-gp mediated multidrug resistance. We suggest that AS1712 and its derivative RJ-
LC-15-8 may have great potential for cancer therapy.
18
19
23

1
2
3
Experimental section
4
5
6
7
Chemical library. We used the chemical library “GRC-2M” from Genomics Research Center at
Academia Sinica. The 2-million compound library contains pure natural products, approved drugs,
known bioactive inhibitors, commercial synthetic molecules, and proprietary collections. Compounds
were prepared at 1 mM in 1536-well propylene plates.
8
9
Antibodies and Reagents. Antibodies against PARP, caspase-3, caspase-8, caspase-9, Bax, and
MDR-1 were purchased from Cell Signaling. Antibodies against cyclin B1, cyclin A, GADPH, and
MOCT-1 were obtained from Santa Cruz. Antibodies against α-tubulin, β-tubulin, and β-actin were
from Sigma Aldrich. The antibody preparation against cytochrome c was from Abcam. The antibody
preparation against p-HH3 (S10) was from Merck Millipore. Horseradish peroxidase−labeled anti-
mouse and anti-rabbit antibodies were purchased from Santa Cruz. Fluorescein
isothiocyanate−labeled anti-mouse antibody was from Thermo Fisher Scientific. Roswell Park
Memorial Institute (RPMI) 1640 medium, Dulbecco’s Modified Eagle medium (DMEM), and fetal
bovine serum were purchased from GIBCO Life Technologies. The in situ cell death detection
(TUNEL) kit was purchased from Roche. The tubulin polymerization assay kit was obtained from
Cytoskeleton. The BCA protein assay kit was purchased from Pierce. The Series S Sensor Chip SA
and HBS-EP+ running buffer were obtained from GE Healthcare.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Cell Lines and Cell Culture. We had isolated CL1-0 cells from a 64-year-old male patient with
a poorly differentiated adenocarcinoma [51]. PC9 cells were obtained from Dr. C.H. Yang (Graduate
Institute of Oncology, Cancer Research Center, National Taiwan University). Molt4 leukemia cells
were obtained from Dr. Tang K. Tang (Institute of Biomedical Sciences, Academia Sinica). KB,
KBtax, KBvin, CCRF-CEM, and CEM/VBL cells were obtained from Dr. T.C. Lee (Institute of
Biomedical Sciences, Academia Sinica). The KB cell line was originally derived from an oral
epidermal carcinoma and has been shown to be contaminated with HeLa cells [31]. Hs578T, MDA-
24

1
2
MB-231, 37T, and 82T cells were obtained from Dr. Y.S. Lu (Department of Internal Medicine,
College of Medicine, National Taiwan University). 37T and 82T cells were isolated from breast
cancer patients and cultured in 2% IH medium. SKOV-3, IGR-OV1, Ovcar-3, HCT116, HT-29,
Colon205, SW620, Du-145, and PC-3 cells were obtained from Dr. S.L. Yu (Department of Clinical
Laboratory Sciences and Medical Biotechnology, National Taiwan University). The NBE cell line
was provided by Dr. Reen Wu (Department of Anatomy, Physiology and Cell Biology, University of
California Davis) and cultured in bronchial epithelial basal medium. SAS and OECM1 cells were
obtained from Dr. C.Y. Chen (Graduate Institute of Health Industry Technology and Research Center
for Industry of Human Ecology, Chang Gung University of Science and Technology). A549, H1975,
BT-549, and HFB cells were purchased from American Type Culture Collection (Rockville, MD,
USA). In general, cells were cultured in RPMI 1640 medium or DMEM supplemented with 10% heat-
inactivated fetal bovine serum and 1% penicillin-streptomycin in a humidified incubator under an
atmosphere of 5% CO2 and 95% air.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Cell Cytotoxicity Assay. Cells were seeded into the wells of 96-well plates (2000 cells/well).
After an overnight culture, the cells were treated with different concentrations of a compound for 72 h.
The CellTiter 96 AQueous one-solution cell proliferation assay (MTS) reagent was used to determine
cytotoxicity. After treatment with a compound, 20 µL of the MTS reagent was added to each well and
the plates were incubated at 37 °C for 1−4 h. Next, the absorbance at 490 nm of the culture medium in
each well was determined using a VersaMax microplate reader (Molecular Devices).
20
21
22
23
24
Colony Formation Assay. Cells were seeded into the wells of six-well plates (1000 cells/well).
After overnight culture, the cells were treated with various concentrations of AS1712 for 2 weeks.
Then, the colonies in each well were fixed with 3.7% paraformaldehyde, stained with 0.04% crystal
violet, and dissolved in 1 ml DMSO. The absorbance at 590 nm of each DMSO solution was read
using the VersaMax microplate reader.
25
26
Cell Cycle Analysis. After treatment, cells were trypsinized and fixed in ice-cold 70% ethanol
overnight. Fixed cells were washed with PBS and stained with 0.5 mL of a propidium iodide (10
25

1
2
µg/mL)/RNase (20 µg/mL) solution for 30 minutes. The cellular DNA content was analyzed with a
FACSCanto flow cytometer (Becton Dickinson).
3
4
Immunoprecipitation and Immunoblotting. The cells were harvested and lysed in RIPA buffer
(20 mM Tris-HCl, 150 mM NaCl, 1 mM Na2EDTA, 1% sodium deoxycholate, 1% IGEPAL CA-630,
0.1% SDS, pH 7.5) with protease inhibitors (Roche). Protein concentration was measured by the BCA
assay. For immunoprecipitation, protein extracts were incubated with mag-beads-AS1712 or mag-
beads-control at 4 °C overnight and then washed five times with PBS before immunoblotting. For
immunoblotting, equivalent amounts of denatured proteins were separated through SDS-PAGE gels
and then transferred to PVDF membranes (Millipore), which were each blocked with 5% nonfat milk
in PBS and 0.1% Tween-20 (Sigma Aldrich) for 1 h. Then each membrane was incubated with a
primary antibody overnight at 4 °C, followed by incubation with a horseradish peroxidase−conjugated
secondary antibody for 1 h at room temperature. The signal was visualized by enhanced
chemiluminescence detection, and images were acquired using a BioSpectrum Imaging System
(UVP).
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Immunofluorescence Staining. Cells or microtubule asters were fixed with 3.7%
paraformaldehyde, permeabilized with 0.1% Triton X-100, stained with anti-α-tubulin (1:1000
dilution) at 4 °C overnight, and then incubated with a fluorescein isothiocyanate−conjugated
secondary antibody (1:500) at 37 °C for 1 h. Cells were then mounted with Prolong Gold Antifade
Reagent and stained with DAPI (Thermo Fisher Scientific), after which images were acquired with an
LSM 700 laser scanning confocal microscope (Carl Zeiss).
21
22
23
24
25
26
In Vivo Microtubule Assembly Assay. Cell samples were treated with a compound for 6 h, and
the cells were then harvested and were lysed in 20 mM Tris-HCl, 1 mM MgCl2, 2 mM EGTA, 0.5%
IGEPAL CA-630, 2 mM PMSF, 200 U/mL aprotinin, and a protease inhibitor cocktail (pH 6.8). After
centrifugation of each lysate at 16,000 rpm for 15 min at room temperature, its supernatant containing
soluble monomeric tubulin and the particulate fraction containing polymerized tubulin were
separately subjected to immunoblotting.
26

1
2
3
4
5
In Vitro Tubulin Polymerization Assay. Reagents from a Tubulin Polymerization Assay kit
(Cytoskeleton) were used according to the manufacturer’s instructions. Purified porcine tubulin (4
mg/mL; >99% purity) in G-PEM buffer (80 mM PIPES, 2 mM MgCl2, 0.5 mM EGTA, 1 mM GTP,
15% glycerol, pH 6.9) was mixed with each test compound in the wells of a 96-well plate. Then, the
A340 value was recorded each minute for 1 h at 37 °C with the VersaMax microplate reader.
6
7
In Vitro Microtubule Nucleation Assay. Purified centrosomes were incubated with or without
AS1712 and with porcine tubulin (Cytoskeleton) in 80 mM PIPES, 1 mM MgCl2, 1 mM EGTA, and 1
mM GTP, pH 6.8 for 8 min at 37 °C. Microtubule asters were fixed in 1% glutaraldehyde, sedimented
onto acid-treated coverslips, and finally subjected to immunofluorescence with anti-α-tubulin. The
fiber lengths of the microtubule asters were measured using MetaMorph image analysis software (n =
20/group).
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Binding Affinity Assay. Surface plasmon resonance technology was used to analyze the binding
affinity of AS1712 for tubulin. A Series S Sensor Chip SA was conditioned with three consecutive 1-
min injections of 1 M NaCl in 50 mM NaOH, and then 50 µg/mL biotin-tubulin (Cytoskeleton) was
immobilized onto the sensor chip surface in a Biacore T200 system (GE Healthcare) to attain 3,000
resonance units. One flow cell on the chip was left free to serve as the negative control. Biocytin
(Sigma Aldrich) was used to block any remaining unmodified streptavidin sites. Different
concentrations of AS1712 in 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.05% surfactant P20,
pH 7.4 were separately injected over the sensor chip surface to measure the association and
dissociation of AS1712 and tubulin at 25 °C. The value of KD was calculated using the steady-state
fitting mode in Biacore T200 Evaluation Software.
22
23
24
25
26
27
Limited Proteolysis Assay. Tubulin (1 mg/mL) was incubated individually with a test compound
for 30 min at 30 °C in 0.1 M morpholino ethanesulfonic acid, 1 mM MgCl2, 1 mM EGTA, pH 6.9,
and then 25 µg/mL trypsin (TPCK-treated, Sigma Aldrich; 1:40 [w/w] trypsin/tubulin) was added to
the solution for a 10-min digestion at 4 °C. Samples were electrophoresed through a 15% SDS-PAGE
gel and stained with Coomassie brilliant blue R250. The images were scanned and quantified using a
BioSpectrum Imaging System (UVP).
27

1
2
3
4
5
6
Colchicine Competition Assay. Tubulin (3 µM) was mixed with colchicine (3 µM) and then
incubated with the various concentrations of AS1712, nocodazole, or vincristine in PEM buffer (80
mM PIPES, 2 mM MgCl2, 0.5 mM EGTA, pH 6.9) at 37 °C for a total volume of 100 µl per well in
black 96-well plates for 1 h. Fluorescence intensities (excitation and emission wavelengths at 365 and
435 nm, respectively) were recorded using an Infinite M200 plate reader (Tecan). Fluorescence values
were normalized to the DMSO control.
7
8
Molecular Modeling. The DMA-colchicine-bound tubulin heterodimer crystal structure (PDB-
ID: 1SA0) [30] was used for the docking study. Molecular docking was conducted using Discovery
Studio (Accelrys Inc.) to assess the probably binding mode of AS1712 in the colchicine-binding site
of tubulin. The initial crystal structure was prepared using the Prepare Protein protocol in Discovery
Studio that protonates the structure and inserts missing loop regions. The docking was performed
using the CDOCKER [52] docking protocol in Discovery Studio based on the grid-based molecular
docking method and the CHARMm force field for binding-energy minimization. Random ligand
conformations were generated from each initial ligand structure via random rotations and high-
temperature molecular dynamics. The random conformations were refined by grid-based simulated
annealing and minimization. The pose was each chosen based on their highest docking score for
further interaction analyses. The post-docking analyses were performed using the View Interactions
tool in Discovery Studio to identify the ligand-protein hydrogen bonds, bumps, and Pi interactions.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Rhodamine Efflux Assay. KBtax cells (2.5 × 105/test) were collected and held on ice for 1 h in
cold medium with 5 mM rhodamine. Then, the cells were resuspended in medium warmed to 37 °C
with one of the indicated compounds and incubated for 1 h in a water bath at the same temperature for
efflux determination. The fluorescence intensities were quantified by FACSCanto flow cytometry.
23
24
25
26
Xenograft Tumor Model. Using the H1975, MDA-MB-231, and KBtax xenograft models,
tumors were established after inoculation into the right flanks of 6-week-old BALB/c nude mice
(NARLabs, Taipei, Taiwan). After establishment of the tumors, vehicle and test compounds, which
had been dissolved in DMSO, were each injected intraperitoneally into a separate group of animals
28