Synthesis and structure-activity studies of antofine analogues as potential anticancer agents

Article abstract of DOI:10.1016/j.bmcl.2006.09.080

Due to the profound cytotoxicities and interesting biochemical aspects, phenanthroindolizidine alkaloids have received an attention as potential therapeutic leads. To define the features of the molecule that are essential for cytotoxicity, we have synthes

Full text of DOI:10.1016/j.bmcl.2006.09.080

Bioorganic & Medicinal Chemistry Letters 17 (2007) 97–100
Synthesis and structure–activity studies of antofine analogues
as potential anticancer agents
Ye Fu,b,d Sang Kook Lee,c Hye-Young Min,c Taeho Lee,a
Jaekwang Lee,b Maosheng Chengd and Sanghee Kima,b,
*
aCollege of Pharmacy, Seoul National University, San 56-1, Shilim, Kwanak, Seoul 151-742, Republic of Korea
bNatural Products Research Institute, College of Pharmacy, Seoul National University, 28 Yungun, Jongro,
Seoul 110-460, Republic of Korea
cCollege of Pharmacy, Ewha Womans University, 11-1 Daehyun, Seodaemun, Seoul 120-750, Republic of Korea
dSchool of Pharmaceutical Engineering, Shenyang Pharmaceutical University, Liaoning 110016, PR China
Received 27 July 2006; revised 26 September 2006; accepted 28 September 2006
Available online 30 September 2006
Abstract—Due to the profound cytotoxicities and interesting biochemical aspects, phenanthroindolizidine alkaloids have received
an attention as potential therapeutic leads. To define the features of the molecule that are essential for cytotoxicity, we have syn-
thesized and evaluated a series of phenanthroindolizidine alkaloid, antofine, analogues with different substituents on the phenan-
threne ring. The systematic structure activity relationship studies elucidate the essential functional group requirement of
phenanthrene ring, providing the basis for further development of phenanthroindolizidine alkaloids.
Ó 2006 Elsevier Ltd. All rights reserved.
OMe
Phenanthroindolizidine alkaloids are a small group of
alkaloids isolated mainly from Cynanchum, Pergularia,
Tylophora, and some genera of the Asclepiadaceae
family.1,2 These pentacyclic natural products exhibit a
variety of biological effects including antitumor, anti-
amoebic, antibacterial, and antifungal activities.1–3
Among these interesting biological activities, the most
intriguing property is the profound cytotoxic activity
against various cancer cell lines. For example, (À)-anto-
fine ((À)-1, Fig. 1) has IC50 values in the low nanomolar
range against drug-sensitive KB-3-1 and multidrug-re-
sistant KB-V1 cancer cell lines, comparable to that of
clinically employed cytotoxic drugs.4,5 While (À)-tyl-
ophorine ((À)-2), the representative phenanthroindolizi-
dine alkaloid, showed a decreased cytotoxicity toward
the KB cancer cell lines, 7-desmethyl analogue of tyl-
ophorine 3 exhibited similar cytotoxicity as antofine.4
In addition, (+)-tylophorine ((+)-2), the optical antipode
of natural product, inhibited the growth of KB variants
and HepG2 cells to noteworthy extent with EC50 values
in the low nanomolar range.6 Beyond the potent cyto-
MeO
MeO
H
N
R
(-)-antofine ((-)-1)
R = H
(-)-tylophorine ((-)-2)
R = OMe
R = OH
(-)-7-desmethyltylophorine (-)-3
Figure 1. Chemical structures of compounds 1–3.
toxicity, an additional interesting biological property
of phenanthroindolizidine alkaloids in antitumor activi-
ty is that the mode of action is unique and distinctly dif-
ferent from other known antitumor drugs.5,6
The structure–activity relationship (SAR), obtained by
comparing the cytotoxic activities of the various natural
phenanthroindolizidine alkaloids, indicates that (i) a
rigid phenanthrene structure is a prerequisite for a
high cytotoxic activity, (ii) the unshared electron pair
on nitrogen is important for a high activity, and
Keywords: Antofine; Phenanthroindolizidine; Structure–activity rela-
tionship; Anticancer.
*
Corresponding author. Tel.: +82 2 740 8913; fax: +82 2 762 8322;
E-mail: pennkim@snu.ac.kr
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.09.080

98
Y. Fu et al. / Bioorg. Med. Chem. Lett. 17 (2007) 97–100
(iii) the cytotoxic potency is highly sensitive to the sub-
stitution types and patterns on the phenanthrene
ring.3,4,7 Although the preliminary SAR of phenanthro-
indolizidine alkaloids has been partially elucidated, the
molecular target site and mechanism of action are still
not clear. Recent mechanism studies by one of us indi-
cated that antofine exhibits an inhibitory activity on cell
proliferation by arresting the G2/M phase of the cell cy-
cle.5 Cheng and co-workers reported that tylophorine
analogues may alter the activity of several transcription
factors, subsequently modulating the expression of tar-
get genes involved in the control of cell growth, differen-
tiation, and survival.6 Very recently, Xi reported that
antofine has favorable molecular interaction with the
bulged regions of DNA and RNA.8
OMe
OMe
MeO
N3
MeO
O
BrPh3P
5
H
KHMDS, THF -78 oC
74% (Z/E, 13:1)
4
6
MeO
N3
MeO
MeO
OMe
OMe
MeO
benzene
reflux
+
N
N
(7/8, 7:3)
7
8
MeO
MeO
NaBH3CN
ZnCl2
THF
OMe
MeO
Due to the profound cytotoxic activity and interesting
biochemical aspects, phenanthroindolizidine alkaloids
have received a significant attention as potential thera-
peutic leads. However, this class of natural products
has not yet been developed for clinical use. The major
drawbacks to the potential therapeutic use of this com-
pound class are the serious central nervous side effects9
and low water solubility. Furthermore, our unpublished
mouse studies revealed that (À)-antofine (1) is metabol-
ically unstable, resulting in a poor pharmacokinetic pro-
file. The methoxy moiety on the phenanthrene ring
proved to be the major cause of the in vivo metabolic
instability. This finding prompted us to perform the sys-
tematic structure activity relationship studies of phenan-
threne ring of antofine to identify the essential structural
elements required for the cytotoxic activity and to opti-
mize the therapeutic index of these compounds. We
present here the synthesis and cytotoxicity evaluation
of a series of antofine analogues with functional group
modifications on the phenanthrene ring.
63%
(from 6)
HCHO, HCl
1
HN
EtOH, reflux
64%
9
MeO
Scheme 1. Synthesis of antofine (1).
aziridine 8 were turned out to be unstable, the crude
reaction mixture of above reaction was directly treated
with ZnCl2 and NaBH3CN15 to give the previously
known pyrrolidine 911,16 in 63% two-step yield. Finally,
the Pictet–Spengler cyclomethylenation of amine 9,
using the previously reported reaction conditions11,16,17
(formaldehyde, HCl, EtOH, reflux), afforded the race-
mic antofine (1)18 in 64% yield.
With the new facile synthetic route, we could properly
prepare the antofine analogues 10–17 (Table 1) from
the appropriate starting materials in similar overall
yield. Further, in order to introduce phenolic hydroxyl
groups in compounds 13–15, the benzyl protecting
group was employed from the beginning and removed
by hydrogenation at the final step of the synthesis.
Due to their exceptional bioactivity and unusual archi-
tecture coupled with low natural abundance, many syn-
thetic methodologies have been developed by medicinal
and synthetic chemist for the synthesis of natural phen-
anthroindolizidine alkaloids and their derivatives.2,3,10
As part of our ongoing program, we have previously
developed two different asymmetric synthetic approach-
es to (À)-antofine.11 Although our previous approaches
are efficient enough to apply for the synthesis of antofine
analogues, a different synthetic route was explored for
the rapid preparation of antofine analogues in racemic
form.
The cytotoxic activity of antofine and its analogues in
various cancer cells was determined by the sulforhoda-
mine B (SRB) assay as described previously.5,19 As
shown in Table 1, the naturally occurring antofine
((À)-1) was approximately 2-fold more potent than the
racemic form 1. This result led us to the reasonable
assumption that cytotoxicity is largely due to only one
enantiomer in the racemic mixture without the signifi-
cant interference of the other enantiomer.
Racemic antofine (1) was synthesized by employing the
intramolecular 1,3-dipolar cycloaddition12 as a key step
(Scheme 1). The starting material for our synthesis was
the known phenanthryl aldehyde 4, which was obtained
from the commercially available homoveratric acid and
p-anisaldehyde via the conventional five-step sequence
according to the previously reported procedure.13 Wittig
reaction of aldehyde 4 with the known phosphonium
salt 514 in the presence of KHMDS gave mainly (Z)-
x-azidoalkene 6 as an inseparable Z/E mixture (74%,
13:1 Z/E). Upon heating the azidoalkene 6 in refluxing
benzene, a mixture of imine 7 and aziridine 8 was
obtained in a 7:3 ratio. Since the isolated imine 7 and
By comparing the cytotoxic activities of the racemic
analogues 10–17 it shows that the cytotoxic potency
was highly dependent, as expected, on the substitution
types and patterns on the phenanthrene ring. Replace-
ment of the methoxy group at the 3- or 6-position with
a bulkier isopropoxy group (compounds 11 and 12) did
not significantly alter the cytotoxicities against the can-
cer cells tested. These results indicated that bulky alkoxy
groups are well tolerated at 3- or 6-position of the phen-
anthrene ring and modest alterations of these positions
are seemingly allowed. However, replacing the 2-meth-

Y. Fu et al. / Bioorg. Med. Chem. Lett. 17 (2007) 97–100
99
Table 1. Structures and cytotoxic activities for antofine ((À)-1 and 1) and its analogues (10–17)
R1
2
R2
3
N
R3
6
a
IC50 (nM)
Compound
R1
R2
R3
HCT 116b
HT-1080c
A549d
(À)-1
1
OMe
OMe
OiPr
OMe
OMe
OMe
OiPr
OMe
OMe
OH
OMe
OMe
OMe
OMe
OiPr
OMe
OMe
OH
9.9
29.4
783.2
24.7
19.2
68.3
21.4
1.1
9.6
10.4
25.1
>1000
53.9
40.5
>100
49.8
1.4
27.5
>1000
37.4
23.6
72.6
19.8
1.1
10
11
12
13
14
15
16
17
OMe
OMe
OH
OMe
OMe
OMe
–OCH2O–
OMe
OMe
H
OMe
49.3
364.1
59.1
453.0
>100
747.2
a All values are means of a minimum of three experiments.
b Human colorectal carcinoma.
c Human colorectal adenocarcinoma.
d Human lung carcinoma.
oxy group of antofine with an isopropoxy group resulted
in a complete loss of activity (compound 10), which is
probably caused by unfavorable steric interactions of
the bulky hydrophobic alkyl group at 2-position with
the binding site.
septica and named as ficuseptine C, of which cytotoxic-
ity has not been reported yet.21 Compound 17 was >100
times less potent than racemic antofine. This result
together with the results obtained with compounds 10
and 11 indicates that one or both methyl groups are
important for activity and steric bulkiness is not tolerat-
ed at position 2. Alternatively, If we consider Xi’s pro-
posed binding model of antofine with bulged DNA,8a
the decreased potency of 17 could be explained by the
repulsive steric interactions between the ribose ring of
DNA and the conformationally restricted dioxymethyl-
ene group.
In another variation, we replaced the electron-donating
methoxy group of antofine with a polar phenolic
hydroxyl group. Replacing the methoxy group at 2-posi-
tion with a hydroxyl group (compound 13) caused more
than 2-fold decrease in potency, whereas modification at
3-position (compound 14) did not significantly affect the
activity. Replacement of the 6-methoxy group in anto-
fine by a hydroxyl group, which led to the naturally
occurring 6-O-desmethylantofine 15, caused a significant
increasing activity (IC50, ca. 1 nM) as compared to race-
mic antofine. The pronounced cytotoxicity of natural
(À)-15 toward the KB cancer cell lines has been reported
previously by Stærk.4 Our result is in agreement with
previous studies and indicates the importance of a
hydroxyl group at 6-position for high cytotoxicity.
In conclusion, we have synthesized and evaluated a ser-
ies of antofine analogues with different substituents on
the phenanthrene ring. The present survey revealed that
steric bulkiness is not tolerated at 2-position of antofine,
while modest alteration at position 3 is probably al-
lowed. The substituent at the 6-position seems to be in-
volved in hydrogen bonding as a donor. These results
are of interest for establishing the SAR of phenanthrene
ring of phenanthroindolizidine alkaloids and providing
the basis for further development. Further studies to-
ward more potent and pharmacologically suitable phen-
anthroindolizidine alkaloids, based on these findings,
are currently in progress in our laboratory.
To further examine the influence of substitutions at the
6-position, compound 1620 was prepared and tested.
Compound 16, possessing a hydrogen atom at 6-posi-
tion instead of the methoxy group of antofine, was only
ꢀ2-fold less potent than racemic antofine but >50-fold
less potent than the hydroxyl group containing 15.
These results together with the reported bioactivities of
2 and 3 suggested that the binding site of antofine might
contain a hydrogen-bond acceptor close to the 6-meth-
oxy group of antofine.
Acknowledgments
This study was supported by the National Research
Laboratory Program (2005-01319), NRDP, Ministry
of Science and Technology, Republic of Korea, and
the Korea Health 21 B&D project (02-PJ2-PG6-DC02-
0001), Ministry of Health & Welfare, Republic of Kor-
ea. One of the authors (T. L.) was supported financially
Next, replacement of the 2,3-dimethoxy substituents by
2,3-dioxymethylene group was investigated. The anto-
fine analogue 17 has been previously isolated from Ficus

100
Y. Fu et al. / Bioorg. Med. Chem. Lett. 17 (2007) 97–100
Bezos, A.; Burns, C.; Kruszeknicki, I.; Parish, C. R.;
Su, S.; Sydnes, M. O. Bioorg. Med. Chem. Lett. 2006,
16, 181; (d) Chuang, T.-H.; Lee, S.-J.; Yang, C.-W.;
Wu, P.-L. Org. Biomol. Chem. 2006, 4, 860; (e) Wei,
L.; Brossi, A.; Kendall, R.; Bastow, K. F.; Morris-
Natschke, S. L.; Shi, Q.; Lee, K.-H. Bioorg. Med.
Chem. 2006, 14, 6560.
by 2nd stage of BK21 project for Applied Pharmaceuti-
cal Life Science Research Division.
References and notes
1. Gellert, E. In Alkaloids: Chemical and Biological Perspec-
tives; Pelletier, S. W., Ed.; Academic Press: New York,
1987; pp 55–132.
2. Suffness, M.; Cordell, G. A.. In The Alkaloids Chemistry
and Pharmacology; Brossi, A., Ed.; Academic Press: New
York, 1985; Vol. 25, pp 3–355.
3. For recent reviews, see (a) Michael, J. P. Nat. Prod. Rep.
2005, 22, 603; (b) Michael, J. P. Nat. Prod. Rep. 2004, 21,
625; (c) Michael, J. P. Nat. Prod. Rep. 2003, 20, 458; (d)
Michael, J. P. Nat. Prod. Rep. 2002, 19, 719; (e) Michael,
J. P. Nat. Prod. Rep. 2001, 18, 520; (f) Li, Z.; Jin, Z.;
Huang, R. Synthesis 2001, 2365.
4. Stærk, D.; Lykkeberg, A. K.; Christensen, J.; Budnik, B.
A.; Abe, F.; Jaroszewski, J. W. J. Nat. Prod. 2002, 65,
1299.
5. Lee, S. K.; Nam, K.-A.; Heo, Y.-H. Planta Med. 2003, 69,
21.
6. Gao, W.; Lam, W.; Zhong, S.; Kaczmarek, C.; Baker, D.
C.; Cheng, Y. C. Cancer Res. 2004, 64, 678.
7. Huang, X.; Gao, S.; Fan, L.; Yu, S.; Liang, X. Planta
Med. 2004, 70, 441.
8. (a) Xi, Z.; Zhang, R.; Yu, Z.; Quyang, D.; Huang, R.
Bioorg. Med. Chem. Lett. 2005, 15, 2673; (b) Xi, Z.;
Zhang, R.; Yu, Z.; Quyang, D. Bioorg. Med. Chem. Lett.
2006, 16, 4300.
11. (a) Kim, S.; Lee, J.; Lee, T.; Park, H.-g.; Kim, D. Org.
Lett. 2003, 5, 2703; (b) Kim, S.; Lee, T.; Lee, E.; Lee, J.;
Fan, G.-j.; Lee, S. K.; Kim, D. J. Org. Chem. 2004, 69,
3144.
12. For a previous application of intramolecular 1,3-dipolar
cycloaddition in synthesis of phenanthroindolizidine alka-
loids, see Pearson, W. H.; Walavalkar, R. Tetrahedron
1994, 50, 12293.
13. Bremmer, M. L.; Khatri, N. A.; Weinreb, S. M. J. Org.
Chem. 1983, 48, 3661.
´
14. Vaultier, A. C. M.; Carrie, R. Tetrahedron Lett. 1989, 30,
4953.
15. (a) Kim, S.; Oh, C. H.; Ko, J. S.; Ahn, K. H.; Kim,
Y. J. J. Org. Chem. 1985, 50, 1927; (b) Kim, S.; Yoon,
J.-Y. Synthesis 2000, 1622; (c) Bytschkov, I.; Sven, D.
Tetrahedron Lett. 2002, 43, 3715.
16. Lebrun, S.; Couture, A.; Deniau, E.; Grandclaudon, P.
Tetrahedron 1999, 55, 2659.
17. Nordlander, J. E.; Njoroge, F. G. J. Org. Chem. 1987, 52,
1627.
18. Spectroscopic data (1H and 13C NMR) for 1 were identical
with those of the natural antofine. see Refs. 4,11 and 16.
19. (a) Lee, S. K.; Cui, B.; Mehta, R. R.; Kinghorn, A. D.;
Pezzuto, J. M. Chem. Biol. Interact. 1998, 115, 215;
(b) Lee, S. K.; Nam, K. A.; Hoe, Y. H.; Min, H.-Y.; Kim,
E.-Y.; Ko, H.; Song, S.; Lee, T.; Kim, S. Arch. Pharmacol.
Res. 2003, 26, 253.
9. Suffness, M.; Douros, J. In Anticancer Agents Based on
Natural Product Models; Cassady, J. M., Douros, J. D.,
Eds.; Academic Press: London, 1980; pp 465–487.
10. For representative recent examples, see (a) Banwell, M.
G.; Sydnes, M. O. Austr. J. Chem. 2004, 57, 537;
(b) Jin, Z.; Li, S. P.; Wang, Q. M.; Huang, R.
Chin. Chem. Lett. 2004, 15, 1164; (c) Banwell, M. G.;
20. Chauncy, B.; Gellert, E. Austr. J. Chem. 1970, 23, 2503.
21. Damu, A. G.; Kuo, P.-C.; Shi, L.-S.; Li, C.-Y.; Kuoh,
C.-S.; Wu, P.-L.; Wu, T.-S. J. Nat. Prod. 2005, 68,
1071.