Cytotoxic activities and structure-cytotoxic relationships of steroidal saponins

Article abstract of DOI:10.1248/bpb.24.1286

We have systematically examined the cytotoxic activities of the steroidal saponins mainly isolated from the Liliaceae plants against HL-60 human promyelocytic leukemia cells and found several structure-activity relationships. Some steroidal saponins evaluated in the assay system showed considerable cytotoxic activities, which were almost as potent as that of etoposide used as a positive control. The activities were found to be sensitive to the monosaccharides constituting the sugar moieties and their sequences, as well as to the structures of the aglycons.

Full text of DOI:10.1248/bpb.24.1286

1
286
Biol. Pharm. Bull. 24(11) 1286—1289 (2001)
Vol. 24, No. 11
Cytotoxic Activities and Structure–Cytotoxic Relationships of
Steroidal Saponins
Yoshihiro MIMAKI,* Akihito YOKOSUKA, Minpei KURODA, and Yutaka SASHIDA
Laboratory of Medicinal Plant Science, School of Pharmacy, Tokyo University of Pharmacy and Life Science, 1432–1,
Horinouchi, Hachioji, Tokyo 192–0392, Japan. Received June 14, 2001; accepted July 16, 2001
We have systematically examined the cytotoxic activities of the steroidal saponins mainly isolated from the
Liliaceae plants against HL-60 human promyelocytic leukemia cells and found several structure–activity rela-
tionships. Some steroidal saponins evaluated in the assay system showed considerable cytotoxic activities, which
were almost as potent as that of etoposide used as a positive control. The activities were found to be sensitive to
the monosaccharides constituting the sugar moieties and their sequences, as well as to the structures of the agly-
cons.
Key words steroidal saponin; cytotoxic activity; HL-60 cell; structure–activity relationship
Steroidal saponins are naturally occurring glycosides that prepared from 5, 6, 23, and 23, respectively.
possess properties such as producing foam, hemolytic activ-
Preparation of 15, 16, and 31 A solution of 5 (27 mg)
ity, toxicity to fish, and complex formation with cholesterol. in 0.2 M HCl (dioxane–H O, 1 : 1, 2 ml) was heated at 95 °C
2
Some of the steroidal saponins isolated recently have been for 30 min under an Ar atmosphere. After cooling, the reac-
1
)
shown to be antidiabetic, antitumor in association with the tion mixture was neutralized by passage through an Amber-
2
)
3)
modification of the immune system, antitussive, and lite IRA-93ZU (Organo, Tokyo, Japan) column, and chro-
4
)
platelet aggregation inhibitors. We systematically examined matographed on silica gel using a gradient mixture of
the cytotoxic activities of the steroidal saponins mainly CHCl –MeOH–H O (60 : 10 : 1; 40 : 10 : 1) and ODS silica
3
2
2
0)
isolated from the Liliaceae plants against HL-60 human gel using MeOH–H O (4 : 1) to give 15 (2.0 mg). By the
2
2
1)
22)
promyelocytic leukemia cells and found several structure– same procedure, 16 (1.8 mg) and 31 (12 mg) were pre-
activity relationships. In this paper, we report the structure– pared from 6 (23 mg) and 23 (50 mg), respectively. The
activity relationships of the steroidal saponins versus HL-60 structures of 15, 16, and 31 were ascertained by analysis of
1
13
cell growth, and also the results of the National Cancer Insti- their IR, H-, and C-NMR spectra.
tute (NCI) 60 cell line assay of a representative saponin.
MATERIALS AND METHODS
Preparation of 21 A mixture of 23 (98 mg) and PtO2
6.9 mg) in CHCl –MeOH (1 : 1, 10 ml) was stirred under an
(
3
H atmosphere at room temperature for 12 h. The reaction
2
mixture, after removal of the catalyst by filtration, was sub-
Materials NMR spectra were recorded on a Bruker jected to an ODS silica gel column eluting with MeOH–H O
2
1
23)
DRX-500 (500 MHz for H-NMR) spectrometer using stan- (2 : 1) to give 21 (19 mg). The structure of 21 was ascer-
dard Bruker pulse programs. Silica gel (Fuji-Silysia Chemi- tained by analysis of its IR, H-, and C-NMR spectra.
cal, Aichi, Japan) and ODS silica gel (Nacalai Tesque, Cell Culture Assay HL-60 cells were maintained in
1
13
Kyoto, Japan) were used for column chromatography. Com- RPMI 1640 medium containing 10% fetal bovine serum sup-
puter calculations were performed using the molecular-mod- plemented with L-glutamine, 100 units/ml penicillin, and 100
eling software Macro-model 6.0 on a Silicon Graphics work mg/ml streptomycin. The leukemia cells were washed and re-
4
station, COMTEC 4D/O2 10000SC. The following materials suspended in the above medium to 3ϫ10 cells/ml, and 196
and reagents were used for cell culture and assay of cytotoxic ml of this cell suspension was placed in each well of a 96-
activity: microplate reader, Inter Med Immuno-Mini NJ-2300 well flat-bottom plate. The cells were incubated in 5% CO /
2
(
(
Tokyo, Japan); 96-well flat-bottom plate, Iwaki Glass air for 24 h at 37 °C. After incubation, 4 ml of EtOH–H O
2
Chiba, Japan); HL-60 cells, ICN Biomedicals (Costa Mesa, (1 : 1) solution containing the sample was added to give the
CA, U.S.A.); RPMI 1640 medium, GIBCO BRL (Gland final concentrations of 0.1—20 mg/ml and 4 ml of EtOH–H O
2
Island, NY, U.S.A.); MTT, Sigma (St. Louis, MO, U.S.A.). (1 : 1) was added into control wells. The cells were further in-
All other chemicals used were of biochemical reagent grade. cubated for 72 h in the presence of each agent, and then cell
Compounds 1—3, 6, 8—10, 12, 29 and 30 were isolated growth was evaluated using a modified 3-(4,5-dimethylthia-
5
)
from Triteleia lactea, 4 and 17 from Allium narcissiflo- zol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) re-
6
)
7)
8)
24)
rum, 18 from Smilax china, 5 from Tacca esquirolii,
7
duction assay. Briefly, after termination of the cell culture,
9
)
and 11 from Paris polyphylla var. chinensis, 13, 14, 19 and 10 ml of 5 mg/ml MTT in phosphate buffered saline was
0 from Lilium brownii var. colchesteri, 22 and 26 from A. added to each well and the plate was further reincubated in
jesdianum, 23 and 32 from Reineckea carnea, 24 and 28 5% CO /air for 4 h at 37 °C. The plate was then centrifuged
from A. karataviense, 25, 27, 33 and 34 from A. ampelo- at 1500 g for 5 min to precipitate cells and MTT formazan.
1
0)
2
1
1)
12)
2
1
3)
1
4)
15)
prasum, 35 and 36 from Nolina recurvata, 37—39, 45 An aliquot of 150 ml of the supernatant was removed from
1
6)
and 49 from Ruscus aculeatus, 40—44 and 46 from Dra- each well, and 175 ml of dimethyl sulfoxide (DMSO) was
1
7)
18)
caena draco, 47 and 48 from Sansevieria cylindrica, and added to dissolve the MTT formazan crystals. The plate was
1
9)
5
0 from D. surculosa. Compounds 15, 16, 21, and 31 were mixed on a microshaker for 10 min, and then read on a mi-
∗
To whom correspondence should be addressed. e-mail: mimakiy@ps.toyaku.ac.jp
© 2001 Pharmaceutical Society of Japan

November 2001
1287
croplate reader at 550 nm. A dose response curve was plotted
for each compound, and the concentration giving 50% inhibi-
tion (IC ) was calculated. Data are mean values of three ex-
5
0
periments performed in triplicate.
Conformational Analysis 1000-Step systematic Monte
Carlo conformation searches were carried out with the
AMBER* force field as implemented in Macro-model 6.0 to
2
5)
predict the fully optimized lowest energy structure. Ener-
gies were minimized with the Polak–Ribiere (PR) conjugate
gradient minimizer, and convergence was obtained when the
Ϫ1
gradient root mean square was less than 0.001 kJ Å M.
Throughout this article, all molecular mechanics (MM) cal-
culations assumed a dielectric constant of 1.0. The molecular
dynamics (MD) simulations were carried out with Macro-
model beginning with the lowest energy structures obtained
by the Monte Carlo conformation search. The following op-
tions were used in the MD calculations; time step: 1 fs, equi-
libration time period: 10 ps, and production run time period:
1
000 ps. Initial kinetic energy was added to all atoms as ran-
dom velocities. Translational and rotational momentum was
reset to zero every 0.1 ps. To maintain a constant tempera-
ture, the system was coupled to an external temperature bath
set at 300 K. Coupling between bath and molecule was up-
dated every 0.2 ps. In the production run time at 300 K the
conformers were sampled every 1 ps, followed by energy
minimizations using the AMBER* force field. The final MM
calculations provided the fully optimized lowest energy
structures as shown in this article.
Fig. 1
RESULTS AND DISCUSSION
A. Structure–Activity Relationships of (25R)-Spirost-
5
-en-3b-ol (Diosgenin) Glycoside Derivatives Diosgenin
b-D-glucoside (1) showed no cytotoxic activity against HL-
0 cells (IC Ͼ20 mg/ml), and the attachment of an a-L-
6
5
0
rhamnosyl group at C-2 of the glucosyl moiety led to the ap-
pearance of considerable activity (2: IC50 1.8 mg/ml). Further
addition of an a-L-rhamnosyl, an a-L-arabinofuranosyl or a
b-D-glucosyl, with the exception of a b-D-galactosyl, to C-3
or C-4 of the inner glucosyl moiety either gave no influence
on the activity or slightly increased the activity (3—7: IC50
0
.5—3.3 mg/ml); the attachment of a b-D-galactosyl at C-3 of
the glucosyl residue led to a decrease in the activity (8: IC
5
0
9
.2 mg/ml). Although the introduction of a C-17a hydroxyl
Fig. 2
group to the aglycon of the active saponins slightly reduced
their cytotoxicities (9—12: IC50 1.5—13.2 mg/ml), that of a
C-27 hydroxyl group led to a considerable decrease in the ac- conformation having a vertical orientation against the steroid
tivity (13, 14: IC50 Ͼ20.0 mg/ml). Among a-L-rhamnosyl- plane of the aglycon, while in 15 and 16, the diglycoside and
(
1→2)-b-D-glucoside (2), a-L-rhamnosyl-(1→3)-b-D-gluco- the steroid skeleton were on the same plane. Furostanol
side (15), and a-L-rhamnosyl-(1→4)-b-D-glucoside (16) of saponins usually exhibit no specific properties of saponins,
diosgenin, only diosgenin a-L-rhamnosyl-(1→2)-b-D-gluco- but the furostanol glycosides bearing an a-L-rhamnosyl-
side (2) exhibited cytotoxic activity. Rotating frame nuclear (1→2)-b-D-glucosyl or an a-L-rhamnosyl-(1→2)-[a-L-rham-
Overhauser and exchange effects (ROE) correlations were nosyl-(1→4)]-b-D-glucosyl at C-3 of the aglycon (17, 18)
observed between the protons of the anomeric center of each were found to be cytotoxic to HL-60 cells. The correspond-
monosaccharide and its linkage site in the ROE correlation ing spirosolan glycoside (19, 20) of 2 and 4 showed no cyto-
spectroscopy (ROESY) spectra of 2, 15, and 16. Comparison toxicity at 20.0 mg/ml.
of the molecular models of 2, 15, and 16, which were con-
B. Structure–Activity Relationships of Saponins with a
structed on the basis of the ROE information, and MM and Lycotetraose Group Lycotetraose, b-D-glucosyl-(1→2)-
MD calculation studies, suggested that the three-dimensional [b-D-xylosyl-(1→3)]-b-D-glucosyl-(1→4)-b-D-galactose,
structure of the diglycoside moiety contributed to the activ- often occurs as a glycoside group of steroidal saponins. The
ity. In the cytotoxic saponin (2), the diglycoside existed in a activity of the A/B trans (H-5a) steroids with lycotetraose

1
288
Vol. 24, No. 11
Fig. 4
Fig. 3. Calculated Preferred Conformations of 2, 15, and 16
(21, 22) was as potent as that of the corresponding C-5(6)
unsaturated saponins (23, 24). Introduction of one hydroxyl
group to the A or B ring of the aglycon slightly enhanced the
activity (22, 24, 25). However, further hydroxylation signifi-
cantly reduced the activity (26—28: IC50 Ͼ20.0 mg/ml). The
furostanol saponins with lycoteraose (32—34) were inactive
in contrast with the furostanols (17, 18) mentioned in the
section A. The terminal xylosyl moiety was revealed to be
essential for the activity as was evident from the fact that the
partial hydrolysate (31) of 23 exhibited no activity. Further
glycosyl formation at C-3 of the terminal glucosyl moiety
with a b-D-glucosyl or an a-L-rhamnosyl group did not affect
the activity (29, 30).
Fig. 5
(GI : 0.68 mM) and NCI-H460 (GI : 0.21 mM), colon cancer
5
0
50
C. Cytotoxicities of Saponins with a Glycosyl Group at SW-620 (GI : 0.85 mM), CNS cancer U251 (GI : 0.55 mM),
C-1 of the Aglycon These saponins showed no activity melanoma LOX IM VI (GI : 0.78 mM), renal cancer UO-31
5
0
50
5
0
(
35—44, 47) with the exceptions of those possessing some (GI : 0.81 mM), and prostate cancer PC-3 (GI : 0.71 mM).
5
0
50
acyl groups at the glycosyl moiety (45, 46, 48, 49). A (25S)- The in vivo antitumor testing for 2 is in progress.
spirost-5-ene-1b,3b-diol derivative (50), whose glycoside
E. Conclusion The cytotoxic activities of some steroidal
moiety is composed of two deoxypyranoses, that is, D-fucose saponins may be due to their non-specific detergent effects
and L-rhamnose, was also cytotoxic. These data led us to as- with changing in membrane architecture, as shown by the ag-
sume that the presence of a certain degree of lipophilicity in gregation of tumor cells at early stages after saponin treat-
2
8)
the sugar moiety is essential for exhibiting the cytotoxic ac- ment. However, the extent of membrane damage induced
2
8)
tivity in this type of saponin.
by different saponins is considerably different, and most
D. NCI 60 Cell Line Screening of 2 Compound 2 was recently we have found that 8 and 12 caused the death of
2
6)
29)
subjected to the NCI 60 cell line assay. The mean concen- L1210 leukemia cells through the apoptotic process. These
trations required to achieve GI , total growth inhibition facts indicate that mechanisms other than membrane damage
5
0
(
2
TGI), and LC50 levels against the panel of cells tested were are also involved. Although Nohara and his co-workers have
2
7)
.4 mM, 6.5 mM, and 13.5 mM, respectively. The differential reported the cytotoxic and anti-herpes virus activities of the
21,30)
cellular sensitivity of 2 was moderate for the cell lines, and it Solanum steroidal glycosides,
this communication more
did not exhibit a specific activity towards any particular sub- systematically describes the cytotoxicities and structure–ac-
panel of cells. However, some cell lines were relatively sensi- tivity relationships of the steroidal saponins.
tive to it. These cell lines included the leukemia RPMI-8226
(
GI : 0.69 mM), non-small cell lung cancer A549/ATCC
Acknowledgment We wish to thank the Developmental
5
0

November 2001
1289
7
)
Sashida Y., Kubo S., Mimaki Y., Nikaido T., Ohmoto T., Phytochem-
istry, 31, 2439—2443 (1992).
8
9
)
)
Yokosuka A., Mimaki Y., Sashida Y., J. Nat. Prod., submitted.
Mimaki Y., Kuroda M., Obata Y., Sashida Y., Kitahara M., Yasuda A.,
Naoi N., Xu Z.-W., Li M.-R., Lao A.-N., Nat. Prod. Lett., 14, 357—
3
64 (2000).
1
1
0) Mimaki Y., Sashida Y., Chem. Pharm. Bull., 38, 3055—3059 (1990).
1) Mimaki Y., Kuroda M., Fukasawa T., Sashida Y., J. Nat. Prod., 62,
1
94—197 (1999).
1
1
1
1
1
2) Kanmoto T., Mimaki Y., Sashida Y., Nikaido T., Koike K., Ohmoto T.,
Chem. Pharm. Bull., 42, 926—931 (1994).
3) Mimaki Y., Kuroda M., Fukasawa T., Sashida Y., Chem. Pharm. Bull.,
4
7, 738—743 (1999).
4) Mimaki Y., Kuroda M., Sashida Y., Natural Medicines, 53, 134—137
1999).
(
5) Mimaki Y., Takaashi Y., Kuroda M., Sashida Y., Nikaido T., Phyto-
chemistry, 42, 1609—1615 (1996).
6) Mimaki Y., Kuroda M., Kameyama A., Yokosuka A., Sashida Y.,
Chem. Pharm. Bull., 46, 298—303 (1998); idem, Phytochemistry, 48,
4
85—493 (1998).
1
1
1
2
2
2
2
7) Mimaki Y., Kuroda M., Ide A., Kameyama A., Yokosuka A., Sashida
Y., Phytochemistry, 50, 805—813 (1999).
8) Mimaki Y., Kuroda M., Yokosuka A., Sashida Y., Natural Medicines,
5
2, 374 (1998).
9) Yokosuka A., Mimaki Y., Sashida Y., J. Nat. Prod., 63, 1239—1243
2000).
(
0) Rumyantseva G. N., Fonin V. S., Kalunyants K. A., Shain S. S., Ko-
dash A. G., Prikl. Biokhim. Mikrobiol., 14, 759—766 (1978).
1) Nakamura T., Komori C., Lee Y.-Y., Hashimoto F., Yahara S., Nohara
T., Ejima A., Biol. Pharm. Bull., 19, 564—566 (1996).
Fig. 6
2) Hirai Y., Konishi T., Sanada S., Ida Y., Shoji J., Chem. Pharm. Bull.,
3
0, 3476—3484 (1982).
3) Yahara S., Ura T., Sakamoto C., Nohara T., Phytochemistry, 37, 831—
Therapeutics Program, National Cancer Institute, Bethesda,
MD, for performing the cytotoxic screening studies.
835 (1994).
24) Sargent J. M., Taylor C. G., Br. J. Cancer, 60, 206—210 (1989).
2
5) Goodman J. M., Still W. C., J. Comput. Chem., 12, 1110—1117
1991).
(
REFERENCES AND NOTES
2
6) Monks A., Scudiero D., Skehan P., Shoemaker R., Paull K., Vistica D.,
Hose C., Langley J., Cronise P., Vaigro-Wolff A., Gray-Goodrich M.,
Campbell H., Mayo J., Boyd M., J. Natl. Cancer Inst., 83, 757—766
1
)
)
Nakashima N., Kimura I., Kimura M., Matsuura H., J. Nat. Prod., 56,
45—350 (1993).
Wu R.-T., Chiang H.-C., Fu W.-C., Chien K.-Y., Chung Y.-M., Horng
L.-Y., Int. J. Immunopharmacol., 12, 777—786 (1990).
3
(
1991).
2
2
7) The LC50 is the concentration at which only 50% of the cells are vi-
able, the GI50 value is the concentration that yields 50% cell growth,
and the TGI is the concentration at which no growth is observed.
3
4
)
)
Miyata T., J. Trad. Sino-Jpn. Med., 13, 276—281 (1992).
Niwa A., Takeda O., Ishimaru M., Nakamoto Y., Yamasaki K., Kohda
H., Nishio H., Segawa T., Fujimura K., Kuramoto A., Yakugaku
Zasshi, 108, 555—561 (1988); Dong J.-X., Han G.-Y., Planta Medica,
2
8) Furuya S., Takayama F., Mimaki Y., Sashida Y., Satoh K., Sakagami
H., Anticancer Res., 20, 4189—4194 (2000); idem, ibid., 21, 959—
9
64 (2001).
5
7, 460—462 (1991).
2
3
9) Candra E., Matsunaga K., Fujiwara H., Mimaki Y., Kuroda M.,
Sashida Y., Ohizumi Y., J. Pharm. Pharmacol., accepted.
0) Ikeda T., Ando J., Miyazono A., Zhu X. -H., Tsumagari H., Nohara T.,
Yokomizo K., Uyeda M., Biol. Pharm. Bull., 23, 363—364 (2000).
5
)
)
Mimaki Y., Nakamura O., Sashida Y., Nikaido T., Ohmoto T., Phyto-
chemistry, 38, 1279—1286 (1995).
Mimaki Y., Satou T., Ohmura M., Sashida Y., Natural Medicines, 50,
6
3
08 (1996).