Structure-cytotoxicity relationships of some helenanolide-type sesquiterpene lactones

Article abstract of DOI:10.1021/np960517h

This study deals with the cytotoxicity of helenanolide-type (10α- methylpseudoguaianolide) sesquiterpene lactones. We determined the influence of substitution patterns on the toxicity of 21 helenanolides to a cloned Ehrlich ascites tumor cell line, EN2. Within a series of helenalin esters, the acetate (2) and isobutyrate (3) were more toxic than helenalin itself (1). Esters with larger acyl groups (tiglate 4 and isovalerate 5) exhibited a decreased toxicity compared with the parent alcohol (1). Similar relationships were observed between the 6,8-diastereomer of helenalin, mexicanin I (6) and its acetate (7) and isovalerate (8). In contrast, cytotoxicity within a series of 11α,13-dihydrohelenalin esters (9-12) was shown to be directly related to the size and lipophilicity of the ester side chain, dihydrohelenalin (9) being the least toxic compound in this group. Investigation of several 2,3-dihydrohelenalin derivatives (13-21) with 2α- hydroxy-4-oxo- and 2α,4α-dihydroxy- or -O-acyl-substituted cyclopentane rings (arnifolins and chamissonolides, respectively), for which no pharmacological data have been reported so far, revealed further interesting influences of the substitution pattern on cytotoxicity. The results may be interpreted in terms of lipophilicity and steric effects on the accessibility of the reactive sites considered responsible for biological activity.

Full text of DOI:10.1021/np960517h

252
J . Nat. Prod. 1997, 60, 252-257
Str u ctu r e-Cytotoxicity Rela tion sh ip s of Som e Helen a n olid e-Typ e
Sesqu iter p en e La cton es
Aa¨ron C. Beekman,† Herman J . Woerdenbag,*,† Wim van Uden,† Niesko Pras,† Antonius W. T. Konings,‡
Håkan V. Wikstro¨m,§ and Thomas J . Schmidt|,
Departments of Pharmaceutical Biology and Medicinal Chemistry, University Centre for Pharmacy, Groningen Institute for
Drug Studies, University of Groningen, A. Deusinglaan 1, NL-9713 AV Groningen, The Netherlands, Department of
Radiobiology, Groningen Institute for Drug Studies, University of Groningen, Bloemsingel 1, NL-9713 BZ Groningen,
The Netherlands, and Institut fu¨r Pharmazeutische Biologie der Heinrich-Heine-Universita¨t Du¨sseldorf, Universita¨tsstrasse 1,
D-40225 Du¨sseldorf, Germany
Received J une 14, 1996X
This study deals with the cytotoxicity of helenanolide-type (10R-methylpseudoguaianolide)
sesquiterpene lactones. We determined the influence of substitution patterns on the toxicity
of 21 helenanolides to a cloned Ehrlich ascites tumor cell line, EN2. Within a series of helenalin
esters, the acetate (2) and isobutyrate (3) were more toxic than helenalin itself (1). Esters
with larger acyl groups (tiglate 4 and isovalerate 5) exhibited a decreased toxicity compared
with the parent alcohol (1). Similar relationships were observed between the 6,8-diastereomer
of helenalin, mexicanin I (6) and its acetate (7) and isovalerate (8). In contrast, cytotoxicity
within a series of 11R,13-dihydrohelenalin esters (9-12) was shown to be directly related to
the size and lipophilicity of the ester side chain, dihydrohelenalin (9) being the least toxic
compound in this group. Investigation of several 2,3-dihydrohelenalin derivatives (13-21) with
2R-hydroxy-4-oxo- and 2R,4R-dihydroxy- or -O-acyl-substituted cyclopentane rings (arnifolins
and chamissonolides, respectively), for which no pharmacological data have been reported so
far, revealed further interesting influences of the substitution pattern on cytotoxicity. The
results may be interpreted in terms of lipophilicity and steric effects on the accessibility of the
reactive sites considered responsible for biological activity.
Sesquiterpene lactones are natural products occurring
in many plant families, but most widely distributed
within the Compositae (Asteraceae). These compounds
are known for their various biological activities, includ-
ing cytotoxicity to tumor cells.1 Alkylation of biological
nucleophiles by R,â-unsaturated carbonyl structures in
a Michael-type addition is considered to be the general
mechanism of action. Covalent binding of sesquiterpene
lactones to free sulfhydryl groups in proteins may
lide-sesquiterpene lactones, only little attention has
been paid to possible effects of the variable molecular
geometry on their activity. Differences in molecular
conformation may affect both the steric accessibility of
Michael addition sites and lipophilicity. It was, there-
fore, the aim of the present study to investigate the
relationships between the cytotoxicity of 21 helenano-
lides and their structures, with special attention to the
above-mentioned aspects. The compounds were divided
into four groups, namely helenalins (1-5), mexicanins
I (6-8), 11R,13-dihydrohelenalins (9-12), and 2,3-
dihydrohelenalins (13,14 arnifolins, 15-21 chamissono-
lides).
interfere with the functions of these macromolecules2-4
.
Consequently, sesquiterpene lactones inhibit a large
number of enzymes involved in key biological processes,
namely DNA and RNA synthesis, protein synthesis,
purine synthesis, glycolysis, citric acid cycle, and the
mitochondrial electron transport chain.3,5
Resu lts a n d Discu ssion
Although the exact cellular targets affected by ses-
quiterpene lactones are not known, it may be speculated
that the cytotoxic effect is a consequence of alkylation
of several enzyme sites. Then, the cytotoxicity may be
directly related to the steric accessibility of the targets.
Because the targets and the sesquiterpene lactones are
both three-dimensional structures with their own spe-
cific environment, the Michael addition that takes place
between these reactants will have a certain degree of
specificity.
EN2 cells were incubated with each compound for 2
h and 72 h. The IC50 values obtained by the 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
(MTT) assay are listed in Table 1. In agreement with
previous findings,1-6 compounds of the helenalin and
mexicanin I series (1-8) were found to be the most toxic,
probably due to their bifunctionality as alkylating
agents. The isobutyryl ester 3 was equally toxic com-
pared to the commonly used antitumor agent cisplatin.
The presence of only one alkylation site still yielded
considerable cytotoxicity [dihydrohelenalins (9-12),
arnifolin 13, and chamissonolides (15-20)]. Com-
pounds 14 (11R,13-dihydroarnifolin) and 21 (11R,13-
dihydrochamissonolide) lacking both the R-methylene-
γ-lactone and the cyclopentenone structure were the
least toxic.
In many studies on the cytotoxic effects of helenano-
* To whom correspondence should be addressed. Phone: 31 50 363
3354. FAX: 31 50 363 7570. E-mail: H.J .WOERDENBAG@FARM.
RUG.NL.
† Department of Pharmaceutical Biology.
‡ Department of Radiobiology.
§ Department of Medicinal Chemistry.
| Heinrich-Heine-Universita¨t.
An increase of incubation time from 2 h to 72 h
resulted in higher cytotoxicity. The relative increase
of toxicity within the compounds 1-8, however, was
Author to whom correspondence concerning the tested compounds
and molecular models should be addressed.
X Abstract published in Advance ACS Abstracts, February 15, 1997.
S0163-3864(96)00517-4 CCC: $14.00
© 1997 American Chemical Society and American Society of Pharmacognogy

Cytotoxicity of Helenanolides
J ournal of Natural Products, 1997, Vol. 60, No. 3 253
size of the ester group (i.e., lipophilicity). The isoval-
erate 12, the most lipophilic compound, was the most
toxic, and the alcohol 9 was the least toxic compound.
Chromatographic behavior determined by HPLC and
TLC showed relative retention time (rtR) and the rela-
tive Rf (rRf) values increasing with the presence of larger
ester groups (Table 2). Higher lipophilicity determined
by the HPLC and TLC systems may result in facilitated
penetration through the lipophilic plasma membrane of
the tumor cells, and subsequently, in higher cytotoxicity.
Interestingly, the simple relationship between ester
size and cytotoxicity could not be found within the
helenalin series (1-5), despite similar increases in
lipophilic behavior (Table 2). Although the acetate 2
and the isobutyrate 3 were more toxic than helenalin
1, decreased toxicity was observed in compounds with
larger ester groups (tiglate 4 and isovalerate 5). The
same was found for the mexicanin I derivatives. The
acetate 7 was more toxic and the isovalerate 8 less toxic
than the parent alcohol 6. These findings indicate that
for compounds possessing an R-methylene-γ-lactone
group a size optimum of an adjacent ester group exists.
In compounds not bearing this alkylating center (9-
12) an increase in lipophilicity causes the predicted
increase in cytotoxicity. In the helenalin and mexicanin
I derivatives with larger ester groups, the favorable
effect of higher lipophilicity is obviously being out-
matched by steric hindrance of the exocyclic methylene’s
approach towards a target structure by the bulky ester
side chains. In 11R,13-dihydrohelenalin and its esters,
cytotoxicity is caused by the cyclopentenone moiety,
with the alkylation site at C-2. The distance of this
center from the ester group at C-6 is considerably larger
so that its approach to a target molecule should be less
affected.
Although still strongly active, the mexicanin I deriva-
tives (6-8) were slightly, but significantly, less toxic
than their stereoisomers of the helenalin group (1, 2,
5). These discrepancies may be caused by different
stereochemistry and the resulting differences in the
three-dimensional shape. The helenalin molecule has
been shown to possess a rather high degree of confor-
mational flexibility and is subject to a very fast confor-
mational exchange at room temperature,7,8 whereas
mexicanin I, due to the 7,8-trans fused lactone ring, is
a rather rigid molecule. This difference may affect the
bioactivities. The possibility of reacting with sulfhydryl
groups in different environments may increase with
steric flexibility. Moreover, the solubility of helenalin
is higher than mexicanin I in polar as well as in
nonpolar solvents, which may facilitate the molecule in
reaching its targets inside a living cell. Comparing
mexicanin I ester 8 to the parent compound 6, we found
that toxicity was reduced about two-fold, similar to the
corresponding pair, 5 and 1. As stated above, hindered
approach to a target molecule might be the reason for
the decreased toxicity of the isovalerate.
The steric hindrance caused by the tigloyl group in
the 2-deacetylchamissonolide series is clearly illustrated
by the fact that 2-deacetyl-4-O-tigloylchamissonlide (20),
in which the tigloyl rest is farther away from the
alkylating center, was twice as toxic as the 6-O-tigloyl
derivative 19. Moreover, the unesterified parent alco-
hol, 2-deacetylchamissonolide (16), although expected
to be less toxic than its esters because of its higher
smaller than that of 9-21; comparing 2-h to 72-h
incubation, the IC50 values of 1-8 decreased about six-
fold; those of 9-21, about 13-fold. This may be related
to the presence of two alkylating centers in compounds
1-8 exerting their cytotoxic effect faster.
Within the 11R,13-dihydrohelenalin group (9-12),
cytotoxicity appeared to be directly proportional to the

254 J ournal of Natural Products, 1997, Vol. 60, No. 3
Beekman et al.
Ta ble 1. Mean IC50 (µM) Values ( the Standard Deviation (n g 3) of the Arnica Sesquiterpene Lactones as Measured by the MTT
Assaya
compound
2 h
72 h
helenalin (1)
2.4 ( 0.49
1.3 ( 0.12
1.3 ( 0.49
5.3 ( 0.10b
3.2 ( 0.17
0.39 ( 0.09
0.22 ( 0.07
0.17 ( 0.02b
1.00 ( 0.18b
0.72 ( 0.08b
6-O-acetylhelenalin (2)
6-O-isobutyrylhelenalin (3)
6-O-tigloylhelenalin (4)
6-O-isovalerylhelenalin (5)
mexicanin I (6)
6-O-acetylmexicanin I (7)
6-O-isovalerylmexicanin I (8)
2.7 ( 0.67
2.0 ( 0.24
5.4 ( 1.4
0.56 ( 0.10b
0.36 ( 0.06c
0.99 ( 0.15c
11R,13-dihydrohelenalin (9)
128 ( 17
6.5 ( 2.2
6-O-acetyl dihydrohelenalin (10)
6-O-tigloyl dihydrohelenalin (11)
6-O-isovaleryl dihydrohelenalin (12)
43 ( 5.6d
19 ( 1.9d
20 ( 2.7d
3.5 ( 0.70d
2.6 ( 0.61d
1.7 ( 0.33d
arnifolin (13)
11R,13-dihydroarnifolin (14)
62 ( 7.4
>100
4.0 ( 0.3
76 ( 5.7
2-desacetyl-6-deoxychamissonolide (15)
2-deacetylchamissonolide (16)
6-deoxychamissonolide (17)
n.de
n.d
15.2 ( 1.1f
13.6 ( 0.5g
10.6 ( 1.7
7.6 ( 0.6
102 ( 3.3
Chamissonolide (18)
106 ( 17
2-deacetyl-6-O-tigloylchamissonolide (19)
2-deacetyl-4-O-tigloylchamissonolide (20)
11R,13-dihydrochamissonolide (21)
cisplatin
>100
20.7 ( 0.5h
10.4 ( 1.6
116 ( 5.7
>100
1.4 ( 0.10
>100
0.18 ( 0.04
b
d
aIncubation times were 2 h or 72 h. Significantly different from 1. c Significantly different from 6. Significantly different from 9.
e n.d. stands for not determined. f Significantly different from 17. Significantly different from 18. Significantly different from 20.
g
h
Ta ble 2. Chromatographic Behavior of the Helenanolides
under Study using HPLC and TLC
(18, and 20, both with free hydroxyl functions at C-6)
may very well be explained by its higher polarity (Table
2).
a
b
a
compound
rtR
hRf
rRf
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
0.752
1.012
2.655
3.527
4.860
0.675
1.187
3.502
0.581
0.903
2.933
4.252
1.721
1.285
d
24
48
68
66
75
24
42
78
18
43
61
71
28
1.60
3.20
4.53
4.40
5.00
1.60
2.80
5.20
1.20
2.87
4.07
4.73
1.87
1.67
(0.52e)
(0.39e)
1.40
1.00
0.40
1.60
0.66
Within the tested group of 2,3-dihydrohelenalin de-
rivatives (13-21), arnifolin (13) showed the strongest
cytotoxic effect. It was found to be five times more toxic
than 2-deacetyl-6-O-tigloylchamissonolide (19). It has
been demonstrated that the two compounds adopt
completely different molecular conformations.10 Cha-
missonolide derivatives have been shown to possess a
twist boat conformation,9-11 whereas arnifolin adopts
a twist chair conformation.10 The steric hindrance of
the exomethylene group by the bulky ester side chain
would be stronger in a twist boat form, where the ester
side chain is in an equatorial position coming closer to
the molecule’s sole alkylating center than in the twist
chair conformer. This is illustrated by Figure 1, which
shows molecular models of 13, 19, and 20 from two
different angles. The five-fold difference in IC50 values
between 13 and 19, however, is not likely to be caused
by this steric effect alone, because the cytotoxicity of
19 and its isomer 20 differed only by a factor 2. The
hydroxyl at C-4 of 19 is replaced by a carbonyl function
in 13, and apart from possible steric effects, 13 might
to some extent be transformed into 6-O-tigloylhelenalin
(4) by metabolic dehydratation within a living cell,
which would create an additional Michael addition site
and enhance its cytotoxicity. Compound 14 has no
Michael addition site. Therefore, an equivalent conver-
sion of 11R,13-dihydroarnifolin 14 into 11 could explain
the very low, but still measurable, cytotoxicity of 14;
however, it has been demonstrated that R,â-unsaturated
ester groups have a low alkylating potency of their
own.12 Thus, it cannot be excluded that the tiglic ester
moiety itself causes the weak cytotoxicity. As expected
from its chemical structure, 6-deoxychamissonolide (17)
was more lipophilic than chamissonolide (18) (Table 2).
Yet, 17 turned out to be somewhat less toxic (Table 1).
The same relationship was found between the corre-
25
(24e)
(18e)
21 (60e)
15 (46e)
6
f
1.707
1.000
2.004
0.941
g
24
10
a
Relative retention time compared to 18 as determined by
HPLC. 100 × Rf. c Relative Rf compared to 18. Not available.
b
d
e TLC-system: toluene-EtOAc (20:80). f HPLC-rtR 16 in another
g
system (1.8 mL/min, 25% MeOH isocratic): 0.670. UV not
detectable.
polarity, was still more active than 19. Interestingly,
19 and 20 showed divergent chromatographic behavior
in TLC and HPLC9 (Table 2). According to the HPLC-
rtR value, 19 was relatively lipophilic, while the hRf
value suggested a comparatively high polarity. Com-
pound 20 showed the opposite behavior having a
relatively low rtR value, but a high hRf. Consequently,
based on these chromatography data, we cannot predict
which of these two compounds is the most lipophilic,
with respect to penetration through a biological mem-
brane. However, the fact that 16 was somewhat less
toxic than the 2-O-acetyl and the 4-O-tigloyl derivatives

Cytotoxicity of Helenanolides
J ournal of Natural Products, 1997, Vol. 60, No. 3 255
F igu r e 1. Molecular models of 13 (a, d), 19 (b, e), and 20 (c, f), viewed from two different angles (see Experimental Section).
Oxygen atoms are hatched for clarity; R-methylene carbons involved in Michael addition are colored black. The lateral view of 19
(e) shows the possible hindrance of the exocyclic methylene by the bulky tigloyl ester side chain.
sponding 2-deacetyl derivatives 15 and 16. This finding
is in agreement with Kupchan et al.13 who found
enhanced activities for compounds bearing an oxygen
function at C-6. We suppose that the oxygen at C-6 may
form a hydrogen bond with amino acid residues of a
target protein adjacent to the reactive cysteine. This
possibly increases the reactivity to the Michael addition
site and the toxic effect to some extent.
Although the target molecules affected by the ses-
quiterpene lactones in the cell line used in this study
are not known, our findings clearly demonstrate that
differences in substitution pattern and molecular ge-
ometry of sesquiterpene lactones should be considered
in the interpretation of their biological activity. Results
obtained in a particular cytotoxicity testing system
depend on parameters, such as the cell line and the
assay chosen to measure cytotoxicity. These parameters
may account for differences between some of our results
and data previously published for helenalin esters.5,12,14,15
Recently, we have demonstrated for two sesquiterpene
lactones each containing a Michael addition site (arte-
misitene, eupatoriopicrin) that the MTT and the clas-
sical clonogenic assay yielded comparable cytotoxicity
data indicating actual cell death.16 It is desirable to
extend the present investigations to other cell line
systems and to other classes of sesquiterpene lactones
in order to draw more general conclusions.
Exp er im en ta l Section
Sesqu iter pen e Lacton es-Natu r al P r odu cts. Most
of the compounds under study were isolated as natural
products from different Arnica species as reported
previously: helenalin (1) and 11R,13-dihydrohelenalin
(9);17 helenalin esters and 11R,13-dihydrohelenalin
esters (2-5 and 10-12);18 mexicanin I (6);19 6-deoxy-
chamissonolide (17);20 and arnifolin (13), 11R,13-dihy-
droarnifolin (14), chamissonolide (18), 2-deacetyl-6-O-
tigloylchamissonolide (19), 2-deacetyl-4-O-tigloyl-
chamissonolide (20), 11R,13-dihydrochamissonolide (21).9
Sesqu iter p en e La cton es-Sem isyn th etic Der iva -
tives. Mexica n in I Aceta te (7). Ten mg (38.2 µmol)
of mexicanin I was refluxed for 3 h with 5 mL of a 10%
solution of Ac2O in pyridine. The solvent was evapo-
rated in vacuo and the residue purified by vacuum liquid
chromatography (VLC) on 15 g silica with hexane-
EtOAc (6:4) to give 6 mg (19.7 µmol) mexicanin I acetate.

256 J ournal of Natural Products, 1997, Vol. 60, No. 3
Beekman et al.
1H NMR (200 MHz, CDCl3): δ 7.58 (1H, dd, J ) 2, 6
Hz, H-2), 6.28 (1H, d, J ) 4 Hz, H-13a), 6.12 (1H, dd, J
) 3, 6 Hz, H-3), 5.94 (1H, d, J ) 5 Hz, H-6), 5.69 (1H,
d, J ) 3 Hz, H-13b), 4.80 (1H, ddd, J ) 12, 9, 3 Hz, H-8),
3.26 (1H, m, H-7), 2.77 (1H, ddd, J ) 2, 2, 11 Hz, H-1),
2.56 (1H, ddd, J ) 13, 5, 3 Hz, H-9b), 2.15 (1H, m, H-10),
2.07 (3H, s, CH3 Ac), 1.36 (1H, ddd, J ) 13, 12, 10 Hz,
H-9a), 1.25 (3H, d, J ) 7 Hz, CH3-14), 1.22 (3H, s, CH3-
15).
Mexica n in I Isova ler a te (8). Fifty-one mg (0.195
mmol) of mexicanin I, 25 mg (0.245 mmol) of isovaleric
acid, 43 mg (0.208 mmol) of dicyclohexylcarbodiimide,
and 3 mg (24.6 µmol) of 4-(dimethylamino)pyridine were
dissolved in 13 mL of dry CH2Cl2 and stirred overnight.
The reaction was quenched by addition of 25 mL of
aqueous NaCl solution. The aqueous layer was ex-
tracted with 2 × 20 mL of CH2Cl2, the CH2Cl2 layers
were combined and evaporated to dryness. The residue
was subjected to VLC on 20 g of silica with hexane-
EtOAc (6:4) to yield 18 mg (51.7 µmol) of mexicanin I
isovalerate. 1H NMR (250 MHz, CDCl3): 7.58 (1H, dd,
J ) 2, 6 Hz, H-2), 6.28 (1H, d, J ) 4 Hz, H-13a), 6.11
(1H, dd, J ) 3, 6 Hz, H-3), 5.97 (1H, d, J ) 5 Hz, H-6),
4.81 (1H, ddd, J ) 12, 9, 3 Hz, H-8), 3.26 (1H, m, H-7),
2.78 (1H, ddd, J ) 2, 2, 11 Hz, H-1), 2.57 (1H, ddd, J )
13, 5, 3 Hz, H-9b), 2.3-2.0 (4H, m, H-10, CH ival, CH2
ival), 1.43 (1H, ddd, J ) 13, 12, 10 Hz, H-9a), 1.26 (3H,
d, J ) 7 Hz, CH3-14), 1.22 (3H, s, CH3-15), 0.962 [6H,
(CH3)2 ival].
2-Dea cetyl-6-d eoxych a m isson olid e (15). Ten mg
(32.5 µmol) of 6-deoxychamissonolide (17) was dissolved
in 5 mL of a 1% aqueous solution of NaOH and stirred
at 60 °C for 2 h. After neutralization with dilute HCl,
the aqueous solution was extracted with 3 × 20 mL CH2-
Cl2, and the organic layer was evaporated to dryness
yielding 8 mg (30.1 µmol) of pure 2-deacetyl-6-deoxy-
chamissonolide. 1H-NMR findings were identical to
those reported.21
2-Dea cetylch a m isson olid e (16). Fifteen mg (46.3
µmol) of chamissonolide (18) was dissolved in 5 mL of a
1% aqueous solution of NaOH and stirred at 60 °C for
2 h. After neutralization with dilute HCl the aqueous
solution was extracted with 3 × 20 mL CH2Cl2, and the
organic layers were evaporated to dryness. The residue
was subjected to column chromatography on 10 g of
silica with cyclohexane-EtOAc (6:4) to yield 5 mg (17.7
µmol) of pure 2-deacetylchamissonolide. 1H-NMR find-
ings were identical to those reported.21
TLC system : Silica 60F254 (Merck); toluene-EtOAc
60:40 (v/v); the relative Rf (rRf) values were based on
that of 18 (RfCham): rRf ) RfX/RfCham
.
Cytotoxicity Testin g. The murine EN2, a cloned
Ehrlich ascites tumor cell line, was grown in suspension
culture in RPMI 1640 medium (Gibco, Paisley, Scotland)
supplemented with 10% heat-inactivated fetal bovine
serum (Gibco) plus 0.2 mg/mL streptomycin and 200 IU/
mL penicillin G. The cell line was cultured at 37 °C in
a shaking incubator. The doubling time of the cells was
about 12 h. Exponentially growing cells were used for
all experiments. In all experiments over 95% of the cells
excluded trypan blue.
The principle of the cytotoxicity method is the reduc-
tion of the soluble MTT into a blue-purple formazan
product, mainly by mitochondrial reductase activity
inside living cells.22 The number of cells was found to
be proportional to the extent of formazan production.
The test compounds were made as a 10-mM stock in
96% EtOH, except mexicanin I and its esters, which
were dissolved in 100% DMSO. As a control, helenalin
was dissolved in either EtOH or DMSO, which did not
lead to differences in cytotoxicity.
The compounds were pipetted into 96-wells microtiter
plates (Nunc, Roskilde, Denmark). Equal volumes of
cell suspensions containing 500 cells and the compounds
were mixed in the wells. The cells were incubated with
the compounds for either 2 h or 72 h. After 2 h of
exposure at 37 °C in a humidified incubator with 5%
CO2, the test compounds were washed away in three
steps, always leaving the cells behind in 50 µL of
medium. After each washing step, the plates were
centrifuged for 10 min at 20 °C, 195g. The cells were
further incubated for 72 h at 37 °C in a humidified
incubator with 5% CO2. Then, 20 µL of a 5-mg/mL stock
solution of MTT (Sigma, St Louis, Mo) in phosphate-
buffered saline (PBS: 1.6 mM KH2PO4, 6.5 mM Na2-
HPO4‚12H2O, 0.137 mM NaCl, 2.7 mM KCl, pH 7.4) was
added, and the plates were incubated for another 3 h
45 min. The medium was removed after centrifugation
(15 min, 195g, 20 °C) leaving the insoluble formazan
product behind, which was subsequently dissolved in
200 µL of 100% DMSO. After mixing, the absorbance
(A) was measured at 550 nm using a spectrophotometer
(Titertek Multiscan, Flow Laboratories, Irvine, Scot-
land). Cell growth was calculated using the formula:
[A550 (treated cells) - A550 (culture medium)]100
[A550 (control cells) - A550 (culture medium)]
For all compounds, purity was confirmed by 1H-NMR
spectroscopy. The purity of all compounds was con-
firmed by HPLC and GC to be >95%.
A computer program (Graphpad) was used to calcu-
late the test compound concentration resulting in 50%
growth inhibition (IC50), a parameter for cytotoxicity.
Sta tistics. The independent T-test was used for
statistics, which is part of the SYSTAT software.
Com p u ter Mod els. The depicted computer models
of compounds 13, 19, and 20 represent the conforma-
tions recently determined by X-ray crystallographic
analyses.10 In Figure 1 they are depicted from two
different viewing angles. For a-c the carbon atoms C-6
and C-10 were aligned with a hypothetical y-axis in the
paper plane, while C-6 and C-10 were aligned with the
corresponding z-axis in pictures d-f.
Refer en ce Com p ou n d . Cisplatin (cis-dichlorodiam-
mineplatinum II, Aldrich, Milwaukee, WI) was used as
a reference compound. It was dissolved in H2O (1 mg/
mL) and stored at -80 °C for no longer than 1 month.
Ch r om a togr a p h ic Assessm en t of Lip op h ilicity.
Chromatographic behavior was used to estimate the
order of lipophilicity of the sesquiterpene lactones under
study. It was tested by HPLC retention times (tR) and
TLC hRf () 100 × Rf) values.
HP LC system : Hypersil ODS (RP 18) 5mm, 125 ×
4.6 mm, flow 1.80 mL/min, isocratic elution with 52.5%
MeOH-H2O; detection at 220 nm; the relative retention
times (rtR) were based on that of chamissonolide 18
(tRCham) using the formula rtR ) tRX - t0/tRCham - t0.
Refer en ces a n d Notes
(1) Picman, A. K. Biochem. Syst. Ecol. 1986, 14, 255-281.
(2) Kupchan, S. M. Fed. Proc. 1974, 33, 2288-2295.

Cytotoxicity of Helenanolides
J ournal of Natural Products, 1997, Vol. 60, No. 3 257
(3) Lee, K. H.; Hall, I. H.; Mar, E. C.; Starnes, C. O.; ElGebaly, S.
A.; Waddell, T. G.; Hadgraft, R. I.; Ruffner, C. G.; Weidner, I.
Science 1977, 196, 533-535.
(14) Page, J . D.; Chaney, S. G; Hall, I. H.; Lee, K. H.; Holbrook, D.
J . Biochem. Biophys. Acta 1987, 926, 186-194.
(15) Lee, K. H.; Meck, R.; Piantadosi, C.; Huang, E. S. J . Med. Chem.
1973, 16, 299-301.
(16) Beekman, A. C.; Woerdenbag, H. J .; Kampinga, H. H.; Konings,
A. W. T. Phytother. Res. 1996, 10, 140-144.
(17) Skibinski, A.; Willuhn, G. Planta Med. 1988, 54, 578.
(18) Willuhn, G.; Ro¨ttger, P.-M.; Matthiesen, U. Planta Med. 1983,
49, 226-231.
(19) Leven, W. “Qualitative und quantitative chromatographische
und photometrische Analyse der Sesquiterpenlactone in ‘Arni-
kablu¨ten DAB 9’ und Untersuchungen zur Sesquiterpenlacton-
Fu¨hrung in Blu¨ten von Arnica acaulis”; Ph.D. Dissertation,
Universita¨t Du¨sseldorf, 1990.
(20) Willuhn, G.; Pretzsch, G.; Wendisch, D. Tetrahedron 1981, 37,
773-776.
(21) Willuhn, G.; J unior, I.; Kresken, J .; Pretzsch, G; Wendisch, D.
Planta Med. 1985, 51, 398-401
(22) Carmichael, J .; DeGraff, W. G.; Gazdar, A. F.; Minna, J . D.;
Mitchell, J . B. Cancer Res. 1987, 47, 936-942.
(4) Hall, I. H.; Lee, K. H.; Mar, E. C.; Starnes, C. O.; Waddell, T. G.
J . Med. Chem. 1977, 20, 333-36.
(5) Grippo, A. A; Hall, I. H.; Kiyokawa, H.; Muraoka, O.; Shen, Y.
C.; Lee, K. H. Drug Des. Delivery 1992, 8, 191-206.
(6) Woerdenbag, H. J .; Merfort, I.; Passreiter, C. M.; Schmidt, T.
J .; Willuhn, G.; Van Uden, W.; Pras, N.; Kampinga, H. H.;
Konings, A. W. T. Planta Med. 1994, 60, 434-437.
(7) Rychlewska, U.; Budesinsky, M.; Szczepanska, B.; Bloszyk, E.;
Holub, M. Collect. Czech. Chem. Commun. 1995, 60, 276-299.
(8) Schmidt, T. J . J . Mol. Struct. 1996, 385, 99-112.
(9) Schmidt, T. J .; Willuhn, G.; Wendisch, D.; Steigel, A. Planta Med.
1995, 61, 544-550.
(10) Schmidt, T. J .; Fronczek, F. R.; Liu, Y.-H. J . Mol. Struct. 1996,
385, 113-121.
(11) Wiebcke, M.; Kresken, J .; Mootz, D.; Willuhn, G. Tetrahedron
1982, 38, 2709-2714.
(12) Kupchan, S. M.; Fessler, D. C.; Eakin, M. A.; Giacobbe, T. Science
1970, 168, 376-378.
(13) Kupchan, S. M.; Eakin, M. A.; Thomas, A. M. J . Med. Chem.
1971, 14, 1147-1152.
NP960517H