Phenylthio-derivatives of α-methylene-γ-lactones as pro-drugs of cytotoxic agents

Article abstract of DOI:10.1016/S0223-5234(99)80100-7

A series of substituted phenylthio-derivatives of grosheimin (1), a natural cytotoxic guaianolide, were investigated with the aim of providing insight into their mechanism of action as cytotoxic agents against KB cell lines. Hydrolysis data, kinetics, in the presence and in the absence of H₂O₂, and the valuation of lipophilicity were correlated with cytotoxicity values and with Hammett-σ-values of substituents (R) at the thiophenol ring. These compounds behave as 'pro-drugs' which release the cytotoxic agent grosheimin by sulphur-oxidation promoted by H₂O₂ and subsequent retro- elimination which depends on the nature and position of the R substituent.

Full text of DOI:10.1016/S0223-5234(99)80100-7

Eur. J. Med. Chem. 34 (1999) 515−523
© Elsevier, Paris
515
Original article
Phenylthio-derivatives of α-methylene-γ-lactones as pro-drugs of cytotoxic
agents§
Giuseppe Fardellaa*, Paolo Barbettia¥, Giuliano Grandolinia, Ione Chiappinia, Valeria Ambrogia,
Vito Scarciab, Ariella Furlani Candianib
aInstitute of Pharmaceutical Chemistry and Technology, University of Perugia, Perugia, Italy
bInstitute of Pharmacology and Pharmacognosy, University of Trieste, Trieste, Italy
(Received 18 November 1998; revised 17 February 1999; accepted 11 March 1999)
Abstract – A series of substituted phenylthio-derivatives of grosheimin (1), a natural cytotoxic guaianolide, were investigated with the aim
of providing insight into their mechanism of action as cytotoxic agents against KB cell lines. Hydrolysis data, kinetics, in the presence and
in the absence of H2O2, and the valuation of lipophilicity were correlated with cytotoxicity values and with Hammett-σ-values of substituents
(R) at the thiophenol ring. These compounds behave as ‘pro-drugs’ which release the cytotoxic agent grosheimin by sulphur-oxidation
promoted by H2O2 and subsequent retro-elimination which depends on the nature and position of the R substituent. © Elsevier, Paris
sesquiterpene lactones / pro-drugs / cytotoxicity
1. Introduction
Appropriate precursors of α-methylene-γ-lactones,
able to release the cytotoxic agent selectively in tumour
cells, can decrease the toxicity of this class of com-
pounds. Our interest in such postulated precursors has
been focused [8] on a series of phenylselenoderivatives
of cytotoxic guaianolides with different functionalized
sesquiterpene skeletons. Their cytotoxicity, evaluated in
vitro against KB cells, showed a generalized increase of
bioactivity (ID50) compared to the parent lactone com-
pound, thus supporting the hypothesis of in situ genera-
tion of the corresponding α-methylene-γ-lactones, prob-
ably via the formation of fast retro-syn-eliminating
selenoxides [9, 10].
Natural and semi-synthetic unsaturated-γ-lactones
have been the subject of continuous research spanning
nearly three decades because of their anti-inflammatory,
antibacterial, anti-hyperlipidemic, antiallergic and espe-
cially because of their cytotoxic activity [1–4]. Sesquit-
erpene α-methylene-γ-lactones are alkylating agents and
have been used as antitumour drugs in some cases [5].
The potential tumour-inhibiting ability of this class of
sesquiterpene γ-lactones derives from a Michael-type
interaction of the α,â-unsaturated moiety with nucleo-
philic endocellular components, such as sulphydril en-
zymes. The usefulness of most of these lactones is limited
because of their toxicity which gives rise to indiscrimi-
nate reactions towards biological nucleophiles. Conse-
quently, a number of efforts were made to achieve lower
toxicity via lipophilicity modulation, or via chemical
derivatization [6, 7].
The activation of such masked α,b-unsaturated lac-
tones by means of intracellular oxidation processes might
occur preferentially in tumour cells, since some of them
are deficient in catalase, peroxidase, superoxidodismutase
and, therefore, more sensitive than normal cells to H2O2
and oxygen radical mediated cytotoxicity [11–15].
§Preliminary results of this work were presented at the Joint
Meeting Society S.D.R./S.C.I. Queen’s College Cambridge UK
14–16 July 1993.
*Correspondence and reprints
¥Deceased March 26, 1994.
None of the predicted guaianolide selenoderivatives
were active in vivo as antitumour agents in leukaemia
screening by standard NCI protocols, at dose levels of
60–240 mg/kg (unpublished data).

516
Table I. Chemical, physico-chemical and pharmacological data of compounds 1–18.
Com-
pounds
Thiephenol type
R-substituents properties
In vitro KB cytotoxicity
% inhib.
Hydrolysis kinetics (I/h)
Hydrophilicity
RRT
R-subsitutents R-position
σ Hammett
Fσ Hammett π hydrophob. ID50
lgki
lgkiH2O2
(M × 10–5
)
5 µg/ml
1
2
3
4
5
6
7
8
-
-
-
-
-
-
-
-
0.56
0.39
0.64
0.86
1.00
1.12
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1.0
3.9
3.4
3.7
3.3
2.4
1.2
2.2
2.9
1.8
1.2
2.3
4.4
4.6
3.2
5.0
5.5
3.5
2,3,5,6 tetra F -
6.05
5.40
5.23
5.18
5.08
4.09
4.12
4.08
6.05
5.69
5.65
5.60
5.59
4.68
4.53
4.40
Me
Me
Me
H
NHCOMe
OMe
OMe
OMe
OH
F
Cl
Cl
Cl
Br
Br
Br
p
m
o
-
p
p
m
o
p
p
p
m
o
p
m
o
- 0.17
- 0.07
-
0.00
0.00
- 0.27
+ 0.11
-
- 0.37
+ 0.06
+ 0.23
+ 0.37
-
+ 0.23
+ 0.39
-
- 0.05
- 0.05
- 0.05
0.00
+ 0.50
+ 0.50
+ 0.50
0.00
+ 0.47
+ 0.41
+ 0.41
+ 0.41
+ 0.49
+ 0.71
+ 0.70
+ 0.70
+ 0.70
+ 0.73
+ 0.73
+ 0.73
- 0.97
- 0.02
- 0.02
- 0.02
- 0.67
+ 0.13
+ 0.76
+ 0.76
+ 0.76
+ 0.94
+ 0.94
+ 0.94
22.25
19.66
19.09
11.91
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
9
10
11
12
13
14
15
16
17
18
4.08
4.32
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
ki = 0
As the high Sharpless retro-elimination reactivity of
the selenoxide intermediates [16] could be responsible
for the in vivo inactiveness of the tested seleno-
guaianolides, we substituted the Se with a sulphur atom,
which has a lower retro-elimination reactivity, by prepar-
ing a series of phenylthio derivatives (2–18) of the natural
cytotoxic guaianolide grosheimin (1) (table I).
According to the mechanism of action reported in
figure 1, the lower reactivity of sulphur to oxidation and
to retro-elimination compared to those of selenium can
provide a more highly selective process, particularly in
the cytosol environment of cancer cells. Moreover, the
change of the nature of the R substituents’ position in the
aromatic ring increases the selectivity of the potential
pro-drugs by modulating their effectiveness on the pro-
cess rate of generating the cytotoxic agent (1).
sulfoxide intermediate, but also increasing the ability of
the sulphinic (-enic) group to undergo retro-elimination
and vice-versa.
The aim of this study was to provide new evidence
supporting the mechanism of action of these postulated
pro-drugs as ‘masked α,â-unsaturated lactones’. Thus,
we have now investigated the effect of both the substitu-
tion of Se/S and of the nature and the position of various
R substituents at the thiophenol ring on the ability of
these lactone precursors to act as selective cytotoxic
agents.
The thiophenols added to the exomethylene group of
the guaianolide grosheimin (1) have different substituted
groups whose electron-withdrawing and polarization ef-
fects are measurable by the Hammett-σ-constants
(position-dependent) and by the π-hydrophobic constants
(position-independent) (table I).
If substituent R is an electron-withdrawing group, it
increases the electrophilicity of the sulphur atom thereby
decreasing its ability to be oxidized in the first step of the
Clearly, compounds 2 (because of the poly-substituted
nature of the aromatic ring) and 5, 10, 15, 18 (owing to

517
Figure 1. Preparation and proposed mechanism of action of compounds 2–18.
the ortho position of their R substituent, which is clearly
responsible for unquantifiable steric valuation) are not
particularly suitable for such a rigorous valuation. How-
ever, they are useful to give further information on this
series of products.
cytotoxic agents, especially in the case of the in vitro test
in cell cultures.
3. Pharmacology
The cytostatic activity of the compounds 1–18 was
tested in vitro against a cell line of human nasopharinx
carcinoma (KB). The 18 tested compounds can be di-
vided into three pharmacological groups: inactive
(11–18), moderately active (7–10) with low cytostatic
effect at a concentration of 5 µg/mL medium (corre-
sponding to values greater than 10–5 M) and active (1–6)
compounds
2. Chemistry
Phenylthio derivatives 2–18 were obtained by reaction
of grosheimin 1 with the appropriate thiophenol at room
temperature with stirring under nitrogen atmosphere.
Main details of the compounds’ synthesis and character-
ization are reported in tables II and III. The reaction of
grosheimin with the appropriate thiophenols generated
(table I and figure 2, plot 1).
1
compounds (2–18) with a new chiral centre. The H-
4. Results and discussion
NMR signals showed the presence of only one stereoiso-
mer. Unfortunately, the resolution at 200 MHz was not
enough to assign the correct configuration. According to
the results previously obtained for the grosheimin phe-
nylseleno derivatives [8], the presence of the H11-α
stereoisomer can be hypothesized.
Hydrolysis kinetics were carried out at pH 7.4 and
37 °C in the absence and in the presence of a stoichio-
metric excess of H2O2 (ki and kiH2O2 values are reported
in table I, figure 2, plot 3 and figure 3, plot 2).
The measurement of the lipophilic-hydrophilic proper-
ties of all the compounds were carried out by means of an
original HPLC reverse-phase method [17] on the relative
retention time values (RRt) (table I and figure 2, plot 4).
This physico-chemical characteristic is important in the
pharmacokinetic phase of cell-membrane permeation by
Results of in vitro tests, hydrolysis kinetics in the
absence and in the presence of H2O2 and measurements
of the lipophilic-hydrophilic properties of all compounds
were correlated both qualitatively (figure 2) and, in part,
quantitatively (figure 3). These correlations provided con-
sistent suggestions on the role of both the nature and the
position of the R-substituents at the thiophenol ring in the
behaviour of these ‘masked lactones’ which can act as
pro-drugs according to the postulated mechanism of
action. Preliminary results of this study are summarized
in table I and figure 2 in which compounds 2–18 are
plotted in order of increasing cytotoxicity (figure 2, plot
1) and correlated (i) with the respective values of the
Hammett ‘field effect’ of the R substituent (Fσ) (figure 2,
plot 2), (ii) with the values of the kinetic constants of

518
Table II. Compound synthesis and characterization.
Comp.
Thiophenol
(Thi)
Reaction
time
Yield
m.p.* °C
175-78
MS (M/e)
UV (λmax, nm, EtOH)
(% mol.)
added
(hr)
2
2,3,5,6-
tetraF-thi
p-CH3-Thi
m-CH3-Thi
o-CH3-Thi
2
38
441.1 (19.8 %, M+, 32S), 443.0
(2.1 %, M+, 34S)
272 (e = 5010), 206 (e = 13650)
3
4
5
0.5
1
2
78
88
77
164-66
172-74
172-73
385.9 (0.9 %, M+, 32S)
385.9 (2.5 %, M+, 32S)
256 (e = 7420), 204 (e = 14550)
255 (e = 8370), 207 (e = 18240)
251 (e = 3580), 207 (e = 15510)
386.1 (36.7 %, M+, 32S), 387.9
(2.9 %, M+, 34S)
6
Thi
3
63
80
98
85
98
74
42
97
93
84
87
93
87
188-90
205-07
168-70
150-52
216-18
134-36
159-62
180-82
177-79
178-80
212-14
158-60
188-91
372.3 (57.1 %, M+, 32S), 374.3
(8.3 %, M+, 34S)
254 (e = 5160), 205 (e = 12280)
273 (e = 21080), 205 (e = 23930)
257 (e = 13870), 229 (e = 13330)
7
p-NHCO-
CH3-Thi
p-OCH3-Thi
0.3
0.75
2
429.0 (33.6 %, M+, 32S), 431.1
(3.9 %, M+, 34S)
8
402.1 (6.2 %, M+, 32S), 404.0 (1.3 %,
M+, 34S)
9
m-OCH3-
Thi
402.0 (5.1 %, M+, 32S)
286 (e = 2452), 255 (e = 6100), 217
(e = 14680), 204 (e = 17280)
286 (e = 3230), 253 (e = 4770), 205
(e = 17890)
10
11
12
13
14
15
16
17
18
o-OCH3-Thi
p-OH-Thi
p-F-Thi
0.25
2
402.2 (5.6 %, M+, 32S), 404.1 (0.9 %,
M+, 34S)
388.2 (11.3 %, M+, 32S), 390.0
(2.7 %, M+, 34S)
2
390.2 (28.7 %, M+, 32S), 392.1 (3 %,
M+, 34S)
286 (e = 16860)
p-Cl-Thi
m-Cl-Thi
o-Cl-Thi
p-Br-Thi
m-Br-Thi
o-Br-Thi
0.6
0.25
0.1
0.15
1
406.5 (41.8 %, M+, 32S), 408.4
(9.6 %, M+, 34S)
260 (e = 14240), 205 (e = 19190)
406.4 (13.7 %, M+, 32S), 408.5
(2.8 %, M+, 34S)
258 (e = 10010), 212 (e = 18530),
205 (e = 19140)
405.0 (36.4 %, M+-1, 32S), 407.2
(3.9 %, M+-1, 34S)
255 (e = 5220), 204 (e = 16700)
263 (e = 16800), 204 (e = 23160)
259 (e = 10640), 206 (e = 23450)
255 (e = 9630), 207 (e = 26850)
451.0 (19.8 %, M+, 32S), 453.1
(2.1 %, M+, 34S)
451.1 (31.0 %, M+, 32S), 453.2
(2.8 %, M+, 34S)
1
451.1 (29.3 %, M+, 32S), 453.0
(1.8 %, M+, 34S)
* White crystals, from CH2Cl2 for 11 and from MeOH for all the other compounds.
hydrolysis, lgki, at pH 7.4 and 37 °C in the absence of
H2O2 and in the presence of H2O2 (figure 2, plot 3) and
(iii) with the lipophilicity (figure 2, plot 4). These quali-
tative correlations clearly show a parallelism between the
trend of the data of cytotoxicity, that of Fσ values of R
substituents, and that of the two series of lgki values for
all the three groups of: active (2–6), moderately active
(7–10) and inactive (11–18) compounds.
The R-substituent at the thiophenol ring of the mol-
ecule may play a basic role, mainly because of its own
‘electric field effect’ quantified by the Hammett Fσ
values [18, 19]. A dominant ‘field effect’ (Fσ) of the R
substituent would affect the sulphur charge density in the
phase of the proposed mechanism.
the ‘masked lactones’ inside hypothetical normal and
tumour cells.
The rates of hydrolysis, followed by HPLC dosage of
grosheimin (1), demonstrated first that the cytotoxic
agent is actually released during the hydrolysis reaction
of the active compounds (2–10). Second, the rates of
hydrolysis were, as expected, of zero order in all cases.
Third, and most important, the kinetic constants increase
in the presence of H2O2 (figure 2, plot 3 and table I) for
all the active and moderately active compounds. This
demonstrates the role of H2O2 as a promoter of the
hydrolysis reaction rate, that is of ‘grosheimin-release’,
and supports the basic hypothesis of the postulated
mechanism. The only exception is compound 2 which
will be discussed later.
Further insights arise from the data of the hydrolysis
kinetics of the tested compounds in the absence and in the
presence of H2O2 (figure 2, plot 3 and table I). These
hydrolysis processes are the different simplified model-
conditions of a normal and an oxidant environment for
Moreover, in the course of the hydrolysis reactions, the
presence of the ultimate product of oxidation of the
aromatic leaving group, as the totally oxidized R substi-
tuted benzen-sulfonate, was demonstrated by direct TLC

519
1
Table III. Main H-NMR spectroscopy data of compounds 1–18.
Com- H6
pound
H8
H13
H13'
H14
H14'
H15
Arom.
-
Others
-
1
4.00
3.92
m
3.60
m
3.67
m
3.68
m
3.64
m
3.70
m
3.67
m
3.40
m
3.72
m
3.53
m
3.57
m
3.63
m
6.31
d (2)
3.56
dd (4,13)
3.58
dd (4,14)
3.64
dd (4,14)
3.62
dd (4,12)
3.64
dd (4,14)
3.51
d (4)
3.58
dd (4,14)
3.68
dd (4, 14)
3.40
d (4)
3.42
dd (4,13)
3.66
dd (4,13)
3.61
6.38
d (2)
3.38
dd (4,13)
3.37
dd (4,14)
3.40
dd (4,14)
3.38
dd (4,12)
3.43
dd (4,14)
5.11
s
4.82
s
4.73
s
4.74
s
4.75
s
4.74
s
4.69
s
4.80
s
4.75
s
4.59
s
4.65
s
4.75
s
4.76
s
4.73
s
4.77
s
4.71
s
5.87
s
5.08
s
5.02
s
5.02
s
5.02
s
5.03
s
5.04
s
5.10
s
5.03
s
4.93
s
4.97
s
5.02
s
5.04
s
5.02
s
5.03
s
5.25
s
1.17
d (8)
1.20
d (7)
1.22
d (7)
1.22
d (7)
1.23
d (7)
1.23
d (8)
1.22
d (8)
1.20
d (7)
1.25
d (8)
1.10
d (7)
1.22
d (7)
1.23
d(7)
1.27
d (7)
1.23
d (7)
1.22
d (7)
1.22
d (7)
1.23
d (7)
1.25
d (7)
dd (10,10)
3.87
2
7.06
tt (8,15)
7.04-7.10
dd (9,9)
3.89
3
2.30 arom-Me
s
2.33 arom-Me
s
2.42 arom-Me
s
-
dd (10,10)
3.90
4
7.35-7.0
7.5-7.1
7.5-7.2
7.6-7.4
6.7-7.6
6.7-7.7
6.7-7.5
6.6-7.5
6.8-7.5
7.2-7.4
7.1-7.5
7.1-7.6
7.3-7.5
7.1-7.7
7.1-7.7
dd (8,8)
3.90
5
dd (9,9)
3.90
6
dd (9,9)
3.98
7
2.13 acetyl-Me
s
3.60-OMe
s
3.82-OMe
s
3.88-OMe
s
-
dd (9,9)
3.95
8
3.30
dd (4,14)
3.42
dd (4,14)
3.38
d (4)
3.25
dd (4,13)
3.47
dd (4,13)
3.45
dd (4, 14)
3.50
dd (4,11)
3.50
dd (4,13)
3.54
d (4)
3.58
d (4)
3.65
dd (4,14)
dd (9,9)
3.92
9
dd (9,9)
3.82
dd (8,8)
3.84
dd (9,9)
3.87
dd (10, 10)
3.93
dd (9,9)
3.92
dd (9,9)
3.93
dd (9,9)
3.98
dd (9,9)
3.92
10
11
12
13
14
15
16
17
18
-
-
-
-
-
-
-
3.74
ddd (5, 8, 13) dd (4, 14)
3.72
3.62
dd (4,11)
3.63
ddd (4,9,14)
3.72
ddd (5,12,12) dd (4, 13)
3.68 3.67
ddd (5, 11,11) d (4)
3.75
m
3.70
m
3.52
d (4)
3.50
dd (4,14)
4.72
s
4.74
s
5.03
s
5.04
s
dd (9,9)
3.95
dd (9,9)
δ-values in ppm.
Coupling constants in Hz are in parentheses.
Solvent: CDCl3.
comparison with the corresponding products of H2O2
oxidation obtained from commercial substituted thiophe-
nols [20]. As further confirmation of our expectations,
biologically inactive compounds (11–18) did not undergo
the hydrolysis process in either of the experimental
conditions.
Hydrolysis reactions carried out in the absence of
H2O2 confirmed, by means of the test of free sulphydril
groups [21] on compounds 2–10, that the secondary
product of the reaction was actually the substituted
thiophenol in the reduced state. This was confirmed by
TLC comparison with the respective commercial prod-
ucts used in the preparation of all the phenylthio deriva-
tives.
Hydrolysis kinetics show that, as expected, the R
substituent at the thiophenol moiety of the compounds
affects the rate of reaction in the presence of H2O2. The
strongly deactivating electron attracting groups, such as
F, Cl and Br, may prevent sulfur oxidation and conse-
quently further retroelimination. Conversely, the moder-
ate electron-attracting substituents, such as OMe and
NHCOMe, are sufficient for a moderate rate of
S-oxidation and they may also assist consequent retro-
elimination. In the absence of S-oxidation they are not
opposed to retroelimination, and this process proceeds
more slowly.
Finally, weak electron-donating groups, such as Me or
the indifferent H-atom, help or permit a higher rate of

520
Figure 2. See text.
sulphur oxidation because they are not opposed to the
consequent fast step of retroelimination which can pro-
ceed at a good rate although not activated by the previous
S-oxidation.
In all cases, the sulphur-oxidation step is important in
the total process of ‘grosheimin-release’ by hydrolysis as
evidenced by all the reactive and unreactive phenylthio
derivatives.

521
Figure 3. See text.
Compounds 11 and 2 showed a particular behaviour.
Compound 11 (R = OH) was totally inactive as a cyto-
toxic agent and it was also inert to hydrolysis, even in the
presence of H2O2 despite its positive, but moderately
high, Fσ value (figure 2, plot 2 and table II). Probably the
OH group is deprotonated at pH = 7.4 so the retroelimi-
nation is prevented by a strong negative mesomeric effect
(σ = –0.52; σ– = –0.81) [18].
Compound 2 (R = 2,3,5,6-tetra F) resulted in the most
active cytotoxic agent and it is surprising that it is even
more active than natural grosheimin 1. This result might
be explained by a possible better ability of 2 to penetrate
into the cells in comparison to 1. The behaviour of 2 in
the reactions of hydrolysis was surprising too. Under the
experimental conditions, compound 2 was the fastest
‘releaser’ of grosheimin, and the two ki values, in the
absence and in the presence of H2O2, were identical
(figure 2, plot 3 and table II). This shows its absolute
insensitivity towards the oxidant agent. Probably the very
high electron-attracting effect of the tetrafluoro benzene
ring (Hammett-σ-constants not available) promotes effi-
caciously and rapidly a simple base-assisted retroelimi-
nation process without the need for any preliminary
oxidative activation at the sulfur centre (figure 4). Fur-
thermore, this oxidation is improbable for the same
reason.
The last question to be elucidated is whether, and how,
the characteristics of lipophilicity/hydrophilicity of the
compounds can affect the in vitro cytotoxicity by modu-
lating pharmacokinetic phases of action and, if so, what
specific role they would play. To this end, on the basis of
our previous results obtained for the same series of
products [17], we chose to utilize RRT ratio values which
are based on HPLC retention factors. The data reported in
figure 2, plot 4, show that the differences in activity
cannot be explained by differences in lipophilicity be-
cause no elements of parallelism in trend can be found
between lipophilicity evaluation and bioactivity. These
findings indicate that the lipophilic character of these
substances seems to be less significant for cytotoxicity

522
Figure 4. Base-assisted retroelimination process of compound 2.
than the primary and secondary prerequisites of their
chemical features.
recorded with a Bruker Spectrospin 200 MHz (TMS as an
internal standard). Mass spectra were carried out with a
spectrometer MAT-311-A, AEI-MS-902. UV spectra
were obtained with a spectrophotometer Perkin Elmer
5515. HPLC apparatus consisted of a Perkin Elmer
410-LC Pump, Detector Perkin Elmer LC90J. Spectro-
photometric UV, Integrator Trivector TRIO. Column
Chromatography was performed on Merck Silicagel 60,
TLC on DC Alufolien kieselgel F254 (Merck) analytical,
detectors = H2SO4 20%, UV lamp (λ = 254 nm) and Stra-
tochrom SIF-254 (Carlo Erba) preparative. TLC eluent
was CH2Cl2/MeOH = 90/10. Elemental analyses (C, H,
N, S for compound 7 and C, H, S for all the others) were
carried out with Carlo Erba Elemental Analyzer model
1106 and were within ± 0.4% of theoretical values.
Grosheimin 1 was extracted according to [6], other re-
agents were purchased from Carlo Erba (Italy) and
Aldrich (Italy).
In light of the above results, an attempt was made to
find a quantitative correlation between bioactivity and
molecular physico-chemical properties in the groups of
active phenylthio derivatives 2–6. ID50 values were
plotted versus: (i) Hammett-constant values of the sub-
stituents; and (ii) hydrolysis kinetic ki and kiH2O2 con-
stant values (figure 3, plots 1, 2). Sufficient linear corre-
lation appears between ID50 and σ values for H (6), m-Me
(4) and p-Me (3) substituents (r = 0.988) and good
linearities were obtained by plotting ID50 versus lgki
(r = 0.992) and lgkiH2O2 (r = 0.983). The anomaly of
compound 2 in this trend may derive from the above-
mentioned presumed difference in its mechanism of
action. These linear correlations confirm our expectations
and reinforce the hypothetical mechanism of action of
these cytotoxic agents.
6.2. General procedure for phenylthio derivatives 2–18
(table I)
5. Conclusion
In conclusion, the present study contributes to the
understanding of the mechanism of action of substituted
phenylthio derivatives of α,â-unsaturated cytotoxic lac-
tones which seem to behave as ‘masked lactones’, thus
probably acting as pro-drugs in bioactivity. They release,
in vitro, the cytotoxic agent by sulphur-oxidative promo-
tion and by subsequent retroelimination of the thiosul-
phenic moiety.
These results also focus on the effect of both the nature
and the position of R substituents at the thiophenol ring
upon the chemical and physico-chemical properties pro-
viding correlations with cytotoxic activity.
Grosheimin 1 (200 mg) was dissolved in MeOH con-
taining triethylamine at room temperature and the appro-
priate thiophenol (stoichiometric excess) was added un-
der N2 and stirring. For compounds 2, 7 and 11, the
reaction mixture was evaporated in vacuo and the crude
residue chromatographed on preparative TLC. For the
other phenylthio derivatives, the reaction mixture was
acidified with aq. HCl 1 N, then the addition of cold water
caused the precipitation of the crude product. It was
filtered, washed with H2O/MeOH:1/1 and dried.
6.3. Ki values of hydrolysis of compounds 2–18
6. Experimental protocols
A quantity corresponding to 150 µmol of each com-
pound (2–18) was dissolved in MeOH (1 mL, RS-HPLC)
and 9 mL phosphate buffer pH 7.4 (H2O RS-HPLC) was
added at 37 °C, under stirring. Then, 20 µL reaction
solution was directly injected in the HPLC pump. HPLC
6.1. Chemistry
Melting points were determined with a Kofler hot stage
instrument and are uncorrected. H-NMR spectra were
1

523
conditions: Column Viosil C-18 (10 µ) l = 25 cm,
φ = 4 mm, 29.000 t.p., mobile phase: MeOH/H2O =
68/32 isocratic, flux = 0.8 mL/min, attenuation = 0.5, Rt
(Grosh.) = 120 (± 5) s.
Acknowledgements
This work was supported by grants from the Ministero
dell’Università e della Ricerca Scientifica e Tecnologica
(M.U.R.S.T.) and the Consiglio Nazionale delle
Ricerche-Rome (C.N.R.).The authors are grateful to Mr
Costantino Moriconi for his technical assistance.
From the integral values (I) of Grosheimin peak versus
time, for the compounds 2–10, ki values of the resulting
zero-order straight line were obtained. d(I Grosh.)/dt = ki.
References
[1] El-Sayed K.A., Hamann M.T., J. Nat. Prod. 59 (1996) 687–689.
6.4. KiH2O2 values of hydrolysis of compounds 2–18
[2] Abdel-Sattar E., Galal A.M., Mossa G.S., J. Nat. Prod. 59 (1996)
403–405.
[3] Goren N., Woerdenbag H.J., Bozok-Johansson C., Planta Med. 62
(1996) 419–422.
These values were measured according to the same
procedure used for the ki values determination, after
addition of H2O2 35% m/v (40 mL ca. 460 mmol) to the
phenylthioderivatives solutions.
[4] Calera M.R., Soto F., Sanchez P., Bye R., Hernandez-Bautista B.,
Anaya A.L., Lotina-Hennsen B., Mata R., Phytochemistry 40 (1995)
419–425.
[5] Pettit G.R. (Ed.), Biosynthetic Products for Cancer Chemotherapy,
vol. I, Plenum Press, New York, 1977.
[6] Barbetti P., Fardella G., Chiappini I., Scarcia V., Furlani A., Il
Farmaco Ed. Sci. 40 (1985) 755–769 (and references herein).
6.5. In vitro cytostatic activity
[7] Beekman A.C., Woerdenbag H.J., van Uden W., Pras N., Konings
A.W., Wikstrom H.V., Schmidt T.J., J. Nat. Prod. 60 (1997)
252–257.
The previously described method [22] was followed.
Minimal Eagle’s Medium (MEM) with non-essential
aminoacids and 10% newborn calf serum was used. A
total of 105 KB cells, an established human tumour line,
were incubated at 37 °C in Leighton tubes. After 24 h, the
cells were attached to the glass and the medium was
changed with MEM containing the compounds to be
tested. The compounds were dissolved immediately be-
fore use in sterile dimethylsulfoxide (DMSO). First the
percentage of growth inhibition at 5 µg/mL was evalu-
ated. For the most active compounds, the ID50 (defined as
the molar drug concentration required to inhibit cell
growth by 50%) was determined. Further dilutions were
carried out with culture medium to the required concen-
tration. For each compound, at least five concentration
levels were used. Each compound was tested on at least
two separate occasions. Incubation was carried out at
37 °C for 72 h, the time interval in which exponential
growth occurs. Cell growth was estimated by counting
the cells which were detached from the glass surface with
trypsin. The cytostatic activity was evaluated as percent-
age of growth inhibition in the treated tubes in respect to
the controls. The significance of the results was evaluated
by the Student’s t-test (P < 0.01). The ID50 values were
determined by linear regression analysis. We have chosen
an ID50 value of about 10–5 M as our upper-limit criterion
for a significant and promising level of activity in order to
reduce the number of false positives [23]. Pharmacologi-
cal data are reported in table I.
[8] Barbetti P., Fardella G., Chiappini I., Scarcia V., Furlani A., Eur. J.
Med. Chem. 24 (1989) 299–306 (and literature herein cited).
[9] Clive D.L.J., Tetrahedron 3 (1978) 1049–1132.
[10] Groutas W.C., Theodorakis M.C., Tomkins W.A.F., Herro G.,
Gaynor T., J. Med. Chem. 27 (1984) 548–550 (and literature herein
cited).
[11] Dionisi O., Galeotti T., Terranova T., Assi A., Biochim. Biophys.
Acta 403 (1975) 292–301.
[12] Oberlay L.W., Buetner G.R., Cancer Res. 39 (1979) 1141–1149.
[13] Cederbaum A., Wassermann L.R., Proc. Am. Assoc. Cancer Res. 21
(1980) 26.
[14] Mavelli I., Rotilio G., in: Icosanoids and Cancer, Raven Press, New
York, 1984, pp. 1–10.
[15] Galeotti T., Masotti L., Borrello S., Casali E., Xenobiotica 21 (1991)
1041–1051.
[16] Sharpless K.B., Lauer R.F., J. Am. Chem. Soc. 95 (1973)
2697–2697.
[17] Barbetti P., Chiappini I., Fardella G., Grandolini G., Ambrogi V.,
Polenzani P., Acta Tecn. Legis Medicam IV (1993) 153–168.
[18] Gordon A.J., Ford R.A., The Chemist’s Companion, John Wiley and
Sons, New York, 1983, pp. 144–155.
[19] Lowry T.H., Richardson K.S., Mechanism and Theory in Organic
Chemistry, Harper and Row Publ., New York, 1981, pp. 130–250.
[20] Capozzi G., Modena G., in: Patai S. (Ed.), The Chemistry of the
Thiol Group, Part 2, John Wiley and Sons, London-New York-
Sidney-Toronto, 1974, pp. 789–790 (and literature herein cited).
[21] Van Door R., Leijdekkers C.M., Boss R.P., Brouns R.M.E., Hend-
erson P.T., Ann. Occup. Hig. 24 (1981) 77–92.
[22] Furlani A., Scarcia V., Faraglia G., Sindellari L., Trincia L., Nicolini
M., Eur. J. Med. Chem.-Chim. Ther. 21 (1986) 261–265.
[23] Hakala M.T., Rustum Y.M., in: De Vita V.T. Jr., Bush (Eds.),
Methods in Cancer Research, XVI, Academic Press, NewYork,
1979, p. 247.