Isolation, Semisynthesis, and Molecular Modeling of Deoxypodophyllotoxin Analogs for an Anti-oral Cancer Agent

Full text of DOI:10.1002/bkcs.11979

Note
DOI: 10.1002/bkcs.11979
BULLETIN OF THE
E. Kim et al.
KOREAN CHEMICAL SOCIETY
Isolation, Semisynthesis, and Molecular Modeling of
Deoxypodophyllotoxin Analogs for an Anti-oral Cancer Agent
‡,
Eunae Kim,† Hyun Jung Kim,‡ Seung-Sik Cho,‡ Jung-Hyun Shim,‡, and Goo Yoon
*
*
†College of Pharmacy, Chosun University, Kwangju 61452, Republic of Korea
‡College of Pharmacy, Mokpo National University, Muan 58554, South Korea.
*E-mail: s1004jh@gmail.com; gyoon@mokpo.ac.kr
Received November 19, 2019, Accepted January 12, 2019
Keywords: Oral cancer, Deoxypodophyllotoxin, Molecular modeling, Semisynthesis
Oral cancer is one of the 10 most general cancers in the
world, often detected late and associated with a poor prog-
nosis, a lack of specific biomarkers, and expensive treat-
ments. Oral cancers are mostly oral squamous cell
carcinomas (OSCC) with sophisticated biological and clini-
cal behaviors.1 In general, the important risk factors of oral
cancer are tobacco, alcohol, ultraviolet (UV) light exposure,
and human papillomavirus infection.2,3 Despite advanced
cancer identification and therapy, there has been little
improvement in the 5-year survival rate of oral cancer
patients over the last few decades.4
the 40-methoxyl group with 30%-HBr-AcOH afforded 40-
demethylated DPT, compound 9.16
Compounds 10 and 11 were prepared from treatment of PT
(4) by a methanesulphonic acid/sodium iodide system in
CH2Cl2, followed by weak basic hydrolysis.17 Compounds 10
and 11 were then oxidized by pyridinium dichromate (PDC) in
CH2Cl2, leading to compounds 12 and 13. PDC was more
convenient than Dess–Martin periodinane.18 Compounds 14
and 15 were synthesized with PT (4) and trimethylsilyl cya-
nide in the presence of boron trifluoride diethyl etherate.19
Using trimethylsilyl cyanide in the presence of boron
trifluoride diethyl etherate as a nucleophilic agent was more
effective than any other materials, such as KCN, HCN, or
Hg(CN)2.19,20 The C4-configuration of semisynthetic com-
pounds was deduced from the reaction mechanism.21
We evaluated the antitumor activities of prepared com-
pounds (1–15) against oral cancer HN22 and HSC4 cell
lines by MTT assay. The results are presented in Table 1.
Almost all compounds except yatein (2) and nemerosin (3)
were more effective than positive control, 5-fluorouracil. In
general, the efficacy in the HN22 cell line tended to be bet-
ter than in the HSC4 cell line. DPT showed the strongest
inhibitory effect, with an IC50 value of 6.52 nM for HN22
cells and 7.26 nM for HSC4 cells. PT (4) and pic-
ropodophyllotoxin (PPT, 5) had moderate activity, whereas
yatein (2) and nemerosin (3), which had open C rings,
exhibited micromolar levels of efficacy. In particular,
nemerosin (3), with a double bond, had double the efficacy
of yatein (2). Therefore, cleavage of the C-ring seemed to
have the most important effect on efficacy. The difference
in efficacy between PT (4) and PPT (5) seemed to be due
to the difference in the absolute structure of the D-ring, but
the difference is only about twofold, so this structural dif-
ference did not have a significant effect. Catechol com-
pounds 6 and 7, which had opened A rings, had decreased
activity in the HN22 cell line like that of PT. Compounds
6 and 7 showed significant differences in efficacy between
the two cell lines. 6,7-Demethylene compounds, com-
pounds 6 and 7 are known to significantly decreased the
inhibition of tubulin polymerization.16 The efficacy of com-
pound 8, opened D ring, was similar to that of PT, and the
Therefore, we have been searching for lead compounds
from natural products as potential preclinical candidates for
oral cancer therapy for the last few years. Recently, we
reported that deoxypodophyllotoxin (DPT, 1), isolated
from Anthriscus sylvestris, showed a potent cytotoxic
effect on oral cancer cells.5 DPT is known to have many
bioactivities such as antiproliferative,5 antitumor,6 anti-
platelet aggregation,7 antiviral,8 insecticidal,9 anti-inflam-
matory, anti-allergic,10 and liver protective activities.11 The
main mechanism of action of DPT is the inhibition of
tubulin polymerization, which leads to cell cycle arrest at
the G2/M phase and apoptosis by caspase 3 and 7 activa-
tion.12 Unfortunately, although DPT has a strong in vitro
antitumor effect, it may be limited by in vivo insolubility
in water. The absence of a hydroxyl group at position 4 in
DPT blocks the synthesis of water-soluble derivatives,
which limits its clinical treatment.13 Therefore, DPT ana-
logs need to be developed to increase the water solubility
and maintain its efficacy. Herein, we report the isolation,
semisynthesis, and molecular modeling of DPT analogs for
new anti-oral cancer agents (Figure 1).
We isolated DPT (1),5 yatein (2),14 and nemerosin (3)15
by cytotoxicity-guided fractionation from the roots of
A. sylvestris identified by comparison of their spectroscopic
data with previously reported values in the literature. The
semisynthesis
of
DPT
and
podophyllotoxin
(PT) derivatives is shown in Schemes 1 and 2. Treatment
of DPT with boron trichloride yielded catechol compound
6 and its 40-demethyl compound 7.16 Reduction of the lac-
tone ring with KOH led to compound 8. Demethylation of
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Wiley Online Library
1

Note
BULLETIN OF THE
ISSN (Print) 0253-2964 | (Online) 1229-5949
KOREAN CHEMICAL SOCIETY
H
H
OH
OH
H
H
4
H
Table 1. IC50 values of compounds in HN22 and HSC4 cell lines.
O
O
O
O
O
O
O
O
O
O
A
O
B
C
D O
O
O
O
Compound
HN22 (nM)
HSC4 (nM)
logSa
H
H
O
O
O
H
H
O
O
E
MeO
OMe
MeO
OMe
MeO
OMe
4'
OMe
MeO
OMe
MeO
OMe
OMe
DPT (1)
Yatein (2)
Nemerosin (3)
PT (4)
PPT (5)
6
7
8
6.52
7.26
−5.57
OMe
OMe
OMe
3.92 × 104
1.73 × 104
122
5.44 × 104
2.89 × 104
161
Deoxypodophyllotoxin (1)
Yatein (2)
Nemerosin (3)
Podophyllotoxin (4) Picropodophyllotoxin (5)
Figure 1. Chemical structures of compounds (1–3) isolated from
A. sylvestris, podophyllotoxin (4), and picropodophyllotoxin (5).
−4.75
−4.75
−4.68
−3.99
−4.58
−4.88
−4.06
−4.75
−4.76
−5.45
−4.71
−5.41
245
115
114
97
12
143
84.3
76.2
243
945
987
153.6
17.3
229.2
99.2
89.3
3.5 × 103
820
H
O
O
O
O
O
O
O
OH
OH
O
30%-HBr-AcOH,
ClCH2CH2Cl
O
10% KOH
H
O
O
9
10
11
12
13
14
15
MeO
OMe
MeO
OMe
MeO
OMe
OH
OMe
OMe
1
9
8
1) BCl3, CH2Cl2
2) BaCO3, H2O-acetone
2.5 × 103
398.2
92.7
HO
HO
HO
O
O
O
HO
151.7
1.7 × 103
+
O
5-Fluorouracilb
6.1 × 103
MeO
OMe
MeO
OMe
OMe
OH
a logS was developed by SILICOS-IT. The predicted values are the
decimal logarithm of the molar solubility in water.23
b Positive control.
7
6
Scheme 1. Synthetic pathway for DPT derivatives.
epipodophyllotoxin (14) were less effective than epipodo-
phyllotoxin (11) and 4-cyano-4-deoxy-epipodophyllotoxin
(15). However, in the case of 40-demethylated podophyllo-
toxone (12) and podophyllotoxone (13), 40-demethylated
podophyllotoxone (12) was more effective than podophyllo-
toxone (13). Podophyllotoxone (13) is known as an anti-
prostate agent and inhibitor of the tubulin polymerization.24
4-Cyano-4-deoxy-40-demethylated epipodophyllotoxin (14)
was reported to exhibit an anti-cancer effect against L1210
(mouse lymphocytic leukemia) and KB (human nasopharyn-
geal cancer) cell lines.19 Although 4-cyano-4-deoxy-epipodo-
phyllotoxin (15) showed efficacy similar to PT and compound
8, its water solubility was similar to that of DPT, which
reduced the possibility of another structural transformation.
From these results, the structural modification of the hydroxyl
group at position 40 of similar compound 9 could be expected
to yield a new anti-oral cancer agent with increased water solu-
bility and an improved pharmacokinetic profile. In addition,
the carboxylic acid due to D-ring opening like compound
8 and the hydroxyl group at position 40 of the like 40-
demethylated podophyllotoxone (12) also offer potential as
prodrugs. Cleavage of the C ring was confirmed to have one of
the biggest roles in the deterioration of efficacy. DPT deriva-
tives are known for their various anticancer activities, but their
role in oral cancer treatment is uncertain. We confirmed that
DPT analogs were effective in treating oral cancer and the find-
ings provide a starting point for further optimization of DPT
derivatives as anti-oral agents.
O
H
H
O
O
O
O
PDC
CH2Cl2
MeO
OMe
OH
OR
OH
H
H
O
O
12 R = H
O
O
O
13 R = Me
O
NaI, CH3SO3H
CH2Cl2
H
H
O
O
CN
H
MeO
OMe
BF3-Et2O
Me3SiCN
CH2Cl2
MeO
OMe
O
O
OR
10 R = H
OMe
O
4
H
O
11 R = Me
MeO
OMe
OR
14 R = H
15 R = Me
Scheme 2. Synthetic pathway for PT derivatives.
carboxylic acid group of the opened D ring showed the possi-
bility of another structural transformation. The efficacy of 40-
demethylated DPT, compound 9, was about half that of DPT,
but 10 times more potent than that of PT. It showed the
increased water solubility and the best potency besides DPT.
However, an in vivo experiment in BDF1 mice bearing
murine Lewis lung carcinoma cells demonstrated a loss of anti-
tumor activity. Compound 9 needed esterification of the pheno-
lic hydroxyl group with carbamic, carbonic, and amino
acids to increase antitumor activity.22 Compound 11, epi-
podophyllotoxin, showed better efficacy than that of PT, and
40-demethylated epipodophyllotoxin, compound 10, exhibited
a worse effect than that of PT. Like the relationship between
DPT (1) and 40-demethylated DPT (9), the 40-demethylated
epipodophyllotoxin (10) and 4-cyano-4-deoxy-40-demethylated
To understand the mechanism of action, we attempted
the docking simulations and the molecular dynamics simu-
lation to predict the tubulin–ligand complex. Known as the
mechanism of action of DPT, the occupied colchicine
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
2

BULLETIN OF THE
Note
KOREAN CHEMICAL SOCIETY
domain prevented microtubule assembly by maintaining a
curved αβ heterodimer structure.24 However, the complex
DPT structure was not demonstrated in the X-ray experi-
ment. According to the computational result, DPT was
located at the αβ interface and was bound to the β-subunit
(Figure 2(a)). The T5 loop of the α-subunit (αT5) came
closer to the T7 loop of the β-subunit (βT7) and looked like
a lid on the binding pocket. An interaction between αT5
and βT7 could induce the curved tubulin. DPT was totally
surrounded by H7 and H8 helices and the strand S8 and S9
of the β-subunit driven by the hydrophobic interactions. In
more detail, the A, B, C, and D rings of DPT were deeply
supported by the Leu255 of helix H8. The hydrophobic
E-ring of DPT was contacted by the hydrophobic side chains:
Cys241 of helix H7, Ala316 and Val318 of strand S8, and
Ala354 of strand S9. Compared to the tubulin-PT complex
seen on X-rays,25 it was similar to the binding interaction of
the DPT-tubulin complex. It would not be established by the
structure–activity relationship. However, based on an analy-
sis of the side chain complex contacts, additional interactions
of the binding site can be considered. When a substituent of
DPT was expended with a functional group, it tried to find
accessible amino acids to make a hydrogen bond, an ionic
bond, or van der Waals interaction. Notably, Asn258 of helix
H8 could be an important key in forming the extra-hydrogen
bonding such as proton acceptor or donor. It could be related
to the structure–activity relationship of PT with 4α-hydroxyl
group and compound 11 with 4β-hydroxyl group. The con-
figuration of C-4β was shown to positively affect anticancer
efficacy. To form extra-hydrogen bonding with Asn258, it
should be extended at the C-4β position of DPT and two spe-
cific derivatives could be proposed, 4β-carboxyl-4-deoxy-
podophyllotoxin (model 1), and 4β-carboxyl-40-hydroxyl
(demethylated)-deoxypodophyllotoxin (model 2) (Figure 2
(b)). Although the synthesis of the model compounds21 and
the anticancer effect of model 2 against some cancer cell
lines19 were already reported, anti-oral cancer effect and the
mechanism of action such as destabilizer of tubulin did not yet
have investigated. When the designed compound model 1 was
modeled by the docking simulation, it formed a hydrogen
bond with the side chain of Asn258 (Supporting Information
Figure S1). Specifically, the β direction of the C-ring was in
significant contact with helix H8. Furthermore, when the
activity and predicted water solubility of DPT (1) was com-
pared with compound 9, the substituted 40-hydroxyl group of
the E-ring improved the water solubility and slightly was
affected for the antitumor activity (Table 1). Then, the two-
position substituted compound (model 2) adding a C-4β car-
boxyl group and a 40-hydroxyl group on the E-ring increased
pharmacokinetic optimization. The predicted log S value was
about −3.98, indicating that it could be more soluble than the
synthesized compounds.
In conclusion, bioassay-guided fractionation of A. sylvestris
led to compounds 1–3 with antitumor activity against oral can-
cer. DPT and PT derivatives were semisynthesized and molec-
ular modeling was performed for the pharmacological and
pharmacokinetic advanced DPT analog. The ring opening
playing the most important role in the efficacy was identified
as the cleavage of the C ring. The structure of compounds 8,
9, and 12 showed potential for the pharmacological and phar-
macokinetic optimization. From these studies, we generated
two advanced DPT analog structures with better pharmacolog-
ical and/or pharmacokinetic potential. The synthesis and eval-
uation of these compounds, as well as mechanistic studies as
anti-oral cancer agents, are currently underway.
Figure 2. (a) The predicted binding pose of DPT in soluble tubu-
lin and (b) 3-dimensional structures of the ligands, such as DPT,
PT, model 1 (4β-carboxyl-4-deoxypodophyllotoxin), and model
2 (4β-carboxyl-40-hydroxyl(demethylated)-deoxypodophyllotoxin).
The ligand was located at the αβ interface of tubulin. The binding
domain of DPT was identical to the colchicine domain. The T7
loop of the β-subunit surrounded the tetra-ring of DPT together
with helix H8 (Leu255), and strands S9. The helix H7, and the
strand S8 and S9 held the tri-methoxy benzene of DTP, such as
the hydrophobic interactions; Cys241 of H7, Ala316 and Leu318
of S8, and Ala354 of S9 (sphere). In addition, the T5 loop of the
α-subunit was flexible but seemed like a lid. Asn258 of H8 could
be additional hydrogen bond sites as the proton donor and the pro-
ton acceptor, simultaneously. To form the hydrogen bond, we
proposed the models 1 and 2 compounds with the carboxyl sub-
stituent at C-4. Adding the hydroxyl group, model 2 has better
solubility than does model 1.
Acknowledgments. This study was supported by research
fund from Mokpo National University, 2017.
Supporting Information. Additional supporting informa-
tion may be found online in the Supporting Information
section at the end of the article.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
3

BULLETIN OF THE
Note
KOREAN CHEMICAL SOCIETY
13. M. Khaled, Z. Z. Jiang, L. Y. Zhang, J. Ethnopharmacol.
2013, 149, 24.
REFERENCES
14. N. Sakakibara, S. Suzuki, T. Umezawa, M. Shimada, Org.
Biomol. Chem. 2003, 1, 2474.
1. C. Rivera, Int. J. Clin. Exp. Pathol. 2015, 8, 11884.
2. W. L. Dissanayaka, G. Pitiyage, P. V. Kumarasiri,
R. L. Liyanage, K. D. Dias, W. M. Tilakaratne, Oral. Surg.
Oral. Med. Oral. Pathol. Oral. Radiol. 2012, 113, 518.
3. S. Koontongkaew, J. Cancer 2013, 4, 66.
4. D. M. Cognetti, R. S. Weber, S. Y. Lai, Cancer 2008,
113, 1911.
5. H. -N. Oh, H.-J. Kim, J. -I. Chae, J. -H. Shim, G. Yoon, Kor.
J. Pharmacogn 2017, 48, 213.
15. R. Ikeda, T. Nagao, H. Okabe, Y. Nakano, H. Matsunaga,
M. Katano, M. Mori, Chem. Pharm. Bull. 1998, 46, 871.
16. T. Terada, K. Fujimoto, M. Nomura, J. Yamashita,
T. Kobunai, S. Takeda, K. Wierzba, Y. Yamada,
H. Yamaguchi, Chem. Pharm. Bull. 1992, 40, 2720.
17. A. Kamal, N. Laxman, G. Ramesh, Bioorg. Med. Chem. Lett.
2000, 10, 2059.
18. (a) J. Xiao, X. W. Cong, G. Z. Yang, Y. W. Wang, Y. Peng,
Org. Lett. 2018, 20, 1651. (b) E. Roulland, P. Magiatis,
P. Arimondo, E. Bertounesque, C. Monneret, Bioorg. Med.
Chem. 2002, 10, 3462.
19. Z. X. Chen, W. Y. Ma, C. N. Zhang, Chinese Chem. Lett.
2000, 11, 505.
20. M. Kuhn, A. Wartburg, Helv. Chim. Acta 1968, 51, 163.
21. W. L. Xi, Q. Cai, Y. B. Tang, H. Sun, Z. Y. Xiao, Chinese
Chem. Lett. 2010, 21, 1153.
6. Y. H. Lim, M. J. Leem, D. H. Shin, H. B. Chang,
S. W. Hong, E. Y. Moon, D. K. Lee, S. J. Yoon, W. S. Woo,
Arch. Pharm. Res. 1999, 22, 208.
7. J. J. Chen, Y. L. Chang, C. M. Teng, I. S. Chen, Planta Med.
2000, 66, 251.
8. K. Sudo, K. Konno, S. Shigeta, T. Yokota, Antivir. Chem.
Chemother. 1998, 9, 263.
9. R. Gao, C. F. Gao, X. Tian, X. Y. Yu, X. D. Di, H. Xiao,
X. Zhang, Pest Manag. Sci. 2004, 60, 1131.
22. Y. J. You, Y. Kim, N. H. Nam, S. C. Bang, B. Z. Ahn, Eur.
J. Med. Chem. 2004, 39, 189.
23. A. Dorléans, B. Gigant, R. B. G. Ravelli, P. Mailliet,
V. Mikol, M. Knossow, PNAS 2009, 106, 13775.
24. J. Li, J. Feng, C. Luo, H. Y. S. Herman, R. W. Jiang, Chin.
J. Nat. Med. 2015, 13, 59.
10. S. H. Lee, M. J. Son, H. K. Ju, C. X. Lin, T. C. Moon,
H. G. Choi, J. K. Son, H. W. Chang, Biol. Pharm. Bull. 2004,
27, 786.
11. Y. Kiso, C. Konno, H. Hikino, Y. Yagi, I. Hashimoto,
J. Pharmacobiodyn 1982, 5, 638.
12. E. Roulland, P. Magiatis, P. Arimondo, E. Bertounesque,
C. Monneret, Bioorg. Med. Chem. 2002, 10, 3463.
25. A. Daina, O. Michielin, V. Zoete, Sci. Rep. 2017, 7, 42717.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
4