Phenolic acid derivatives with potential anticancer properties - A structure-activity relationship study. Part 1: Methyl, propyl and octyl esters of caffeic and gallic acids

Article abstract of DOI:10.1016/j.bmc.2004.04.026

The antiproliferative and cytotoxic properties of polyphenolic acid derivatives, structurally related with the natural models caffeic and gallic acids, have been tested in human cervix adenocarcinoma cells (HeLa). Simultaneous structural information was o

Full text of DOI:10.1016/j.bmc.2004.04.026

Bioorganic & Medicinal Chemistry 12 (2004) 3581–3589
Phenolic acid derivatives with potential anticancer properties––
a structure–activity relationship study. Part 1: Methyl, propyl
and octyl esters of caffeic and gallic acids
a
a
a
b,c
d
S. M. Fiuza, C. Gomes, L. J. Teixeira, M. T. Gir ~a o da Cruz, M. N. D. S. Cordeiro,
e,f
N. Milhazes, F. Borges and M. P. M. Marques
a,e
a,b,
*
a
Research Unit ‘Molecular Physical-Chemistry’, Coimbra University, Portugal
b
Biochemistry Department, Coimbra University, Portugal
Centre for Neuroscience and Cellular Biology, Coimbra University, Portugal
c
d
REQUIMTE, Chemistry Department, Faculty of Sciences, University Porto, Portugal
e
Organic Chemistry Department, Faculty of Pharmacy, University Porto, Portugal
f
Institute of Health Sciences-North, Gandra-Paredes, Portugal
Received 13 January 2004; accepted 19 April 2004
Available online 18 May 2004
Abstract—The antiproliferative and cytotoxic properties of polyphenolic acid derivatives, structurally related with the natural
models caffeic and gallic acids, have been tested in human cervix adenocarcinoma cells (HeLa). Simultaneous structural information
was obtained for these compounds through theoretical ab initio methods. This study was conducted for the following esters: methyl
caffeate (MC, 1), propyl caffeate (PC, 2), octyl caffeate (OC, 3), methyl gallate (MG, 4), propyl gallate (PG, 5) and octyl gallate (OG,
6
). A significant growth-inhibition effect was assessed for some of these compounds, clearly dependent on their structural char-
acteristics. Marked structure–activity relationships (SARs)––namely the number of hydroxyl ring substituents––were found to rule
the biological effect of such systems.
ꢀ
2004 Elsevier Ltd. All rights reserved.
5
1. Introduction
cells. CAPE, which is a component of propolis from
honeybee hives, has shown anticarcinogenic and
6
Some naturally occurring phenolic acids and analogues,
namely caffeic and gallic acids, are known to display a
wide variety of biological functions, in addition to their
primary antioxidant activity, which are mainly related to
immunomodulatory properties. It was proposed that
the molecular basis of this action may be connected to
the inhibition of the nuclear transcription factor NK-
7
kappa B.
1;2
modulation of carcinogenesis. Indeed, many phenolic
compounds have been investigated for their potential
Gallic acid and its esters, in turn, are hydroxybenzoic
derivatives used as antioxidant additives in both food
and pharmaceutical industry––E-310 (propyl gallate)
and E-311 (octyl gallate)––which are known to protect
against oxidative damage induced by reactive oxygen
species (ROS), as hydroxyl radicals or hydrogen per-
oxide, and reactive sulfur species (RSS). Synthetic
galloyl esters showed to be potent and selective enzyme
inhibitors, as well as synergistic protectors against per-
oxyl radical damage in membranes. They were also
found to inhibit cytokine-induced nuclear translocation
of NF-kappa B, as well as expression of leukocyte
3
use as cancer chemopreventive agents. Epidemiological
studies indicate that a rich diet in fruits and vegetables
reduces cancer risk in humans, suggesting that certain
dietary constituents may thus be effective in preventing
4
cancer. Cinnamic acid esters, such as caffeic acid
phenethyl (CAPE) and benzyl esters, display selective
antiproliferative activity against some types of cancer
8
;9
Keywords: Caffeates; Gallates; Anticancer activity; Ab initio calcul-
ations.
10
adhesion molecules in vascular endothelium cells.
Moreover, gallic acid derivatives are known to cause
apoptosis in tumour cell lines and to inhibit lymphocyte
*
Corresponding author. Tel./fax: +351-239-826-541; e-mail: pmc@
ci.ui.uc.pt
0
968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.04.026

3582
S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
1
1
proliferation. However, the mechanism by which gallic
acid analogues induce apoptosis in some cell lines is yet
not completely understood, probably involving the
paradoxical generation of ROS, which interfere with the
8 5 4 7 6 5 14 1 6
(H17O C C ), (H18O C C ) and (H20O C C ) equal to
0ꢁ or 180ꢁ (Fig. 1); (ii) position of the aromatic ring and
the carbonyl group relative to the spacer C@C bond, in
the caffeates––either trans or cis, for dihedral (O12
) equal to 0ꢁ or 180ꢁ, respectively; (iii) orientation
of the alkyl chain relative to the benzene ring, in the case
of gallates––(C10
) 0ꢁ or 180ꢁ––and to the spacer
double bond, in the caffeates––dihedral (O13
11
C -
12–14
homeostatic redox balance of the cell. It is also a fact
that these compounds are excellent inhibitors of protein
tyrosinase kinases (PTKs).
10 9
C C
15
9 3 2
C C C
C
11 10 9
C C )
The recognised cytotoxic activity of phenolic derivatives
depends on their rate of incorporation into cells, directly
related to their lipophilicity, as much as to their anti-
oxidant activity. Both these properties are affected by
the presence of the ring substituent hydroxyl groups and
by the length of the ester moiety, in polyhydroxylated
phenolic esters. Thus, the knowledge of the structural
characteristics of this type of compounds is of the ut-
most importance for the understanding of the structure–
activity relationships (SARs) underlying their biological
activity.
equal to 0ꢁ or 180ꢁ. For all the geometries thus obtained,
calculation of the harmonic vibrational frequencies was
also performed, in order to determine which ones were
real energy minima (conformers) in the potential energy
surface.
For all caffeates and gallates under investigation inves-
tigation––MC, PC, OC, MG, PG and PG––the most
stable conformers (Fig. 1) were found to display a pla-
nar geometry, probably due to the stabilising effect of p-
electron delocalisation between the benzene ring and
both the carbonyl group, in the gallates, or the ethylenic
side chain, in the caffeates, which is favoured when they
all lie in the same plane. Moreover, in the caffeates (1–3),
an E orientation of the ring and the C@O group relative
The present study concentrates on the evaluation of the
growth inhibition and cytotoxic activity towards a hu-
man cancer cell line, of six distinct phenolic esters: me-
thyl trans-3-(3,4-dihydroxyphenyl)-2-propenoate (methyl
caffeate, MC, 1), propyl trans-3-(3,4-dihydroxyphenyl)-
to the C @C10 bond showed to be highly favoured. As to
9
the variable length ester alkyl chain, a zig-zag confor-
mation was determined to yield, in all cases, the most
energetically favoured structures.
2
-propenoate (propyl caffeate, PC, 2), octyl trans-3-(3,4–
dihydroxyphenyl)-2-propenoate (octyl caffeate, OC, 3),
methyl 3,4,5-trihydroxybenzoate (methyl gallate, MG,
4
), and two official antioxidants propyl-3,4,5-trihy-
Furthermore, an identical orientation of the benzene
hydroxyl substituents, always coplanar with the aro-
matic ring, was detected. In fact, this allows the for-
mation of highly stabilising medium strength
intramolecular (O)Hꢀ ꢀ ꢀO interactions (Oꢀ ꢀ ꢀH distances
between 211 (for caffeates) and 218 pm (for gallates),
Fig. 1). As expected, no stable geometries were obtained
when the hydroxyl groups were directed towards each
other. However, some conformers were found for
opposite orientations of these phenolic OHs, although
these were the least stable ones (Oꢀ ꢀ ꢀO distances of ca.
265 pm).
droxy-benzoate (propyl gallate, PG, 5) and octyl-3,4,5-
trihydroxybenzoate (octyl gallate, OG, 6). The results
thus obtained were interpreted in the light of structural
information yielded by theoretical (ab initio) methods.
2
. Results and discussion
Six distinct phenolic esters were studied––MC (1), PC
2), OC (3), MG (4), PG (5) and OG (6) (Fig. 1). These
(
compounds, to be assessed for their antitumour activity
in a human cervix adenocarcinoma cell line (HeLa),
differ mainly in the number of hydroxyl ring substituent
groups and/or in the length of the alkyl side chain, apart
from the nature of the side carbon chain––with or
without an ethylenic spacer (depending on whether they
are caffeates or gallates).
The conformational results presently obtained for the
caffeates and gallates studied are in perfect accordance
with previously reported data on the analogous di-
hydroxylated and trihydroxylated carboxylic acids,
namely trans-caffeic acid (trans-3-(3,4-dihydroxyphen-
16
yl)-2-propenoic acid, CA, 7) and trans-3-(3,4,5-tri-
hydroxyphenyl)-2-propenoic acid.
17
Regarding the
relative orientation of the ring hydroxyls and the C@O
group, for instance, it is worth noticing that the phenolic
esters now investigated present an identical behaviour to
similar phenolic acids: thus, while a trans orientation
was found to be preferred for caffeates 1, 2 and 3––as in
caffeic acid––an S–cis conformation was energetically
favoured for the trihydroxylated esters 4, 5 and 6––
similarly to what was observed for both gallic (8) and
trans-3-(3,4,5-trihydroxyphenyl)-2-propenoic acids. This
is corroborated by single-crystal X-ray studies on similar
2
.1. Structural analysis
In order to determine the effect of the structural char-
acteristics of this kind of phenolic derivatives on their
antiproliferative and/or cytotoxic activity, theoretical
calculations were carried out, using ab initio methods. A
complete geometry optimisation was carried out for
compounds 1–6, in order to obtain the distinct possible
structures for these molecules, as well as their relative
stability. The effect of several geometrical parameters on
the overall stability of the compounds was investigated,
namely: (i) orientation of the phenolic substituents, both
relative to the ring and to the C@O group––dihedrals
18–20
systems, namely propyl, butyl and octyl gallates.
In
fact, both bond lengths and angles reported for the
crystal, as well as intermolecular (O)Hꢀ ꢀ ꢀO distances,
are in close agreement with the presently calculated
values.

S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
3583
Figure 1. Optimised lowest energy geometries (B3LYP/6-31G**) for the phenolic esters studied along this work (The atom numbering is included).
2
.2. Biological assays
was possible to conclude that values above 50 lM did
not correspond to a significant increase in the corre-
sponding growth–inhibition effect (Fig. 2).
The phenolic esters investigated were tested in a human
cervix adenocarcinoma cell line––HeLa (epithelial-like
adherent line). In order to determine the degree of
toxicity of these compounds towards healthy cells,
experiments were also carried out in non-neoplastic
cells––fibroblasts from human embryonic lung tissue
The cytotoxic activity of the phenolic derivatives tested
showed to be strongly dependent on their structure.
Actually, for each type of ester––either di or trihydr-
oxylated––the effect of a variation in the length of the
alkyl chain was found to lead to a clear change in the
respective antiproliferative properties: the propyl esters
(2 and 5) displaying a considerably more pronounced
antitumoural effect towards both HeLa and L-132 cell
lines than their octyl and methyl analogues (Figs. 2 and
3). Indeed, the compounds with a shorter alkyl chain––
methyl esters 1 and 4––showed a rather small antipro-
liferative effect, mainly against HeLa cells (Fig. 2). These
results may be explained by the drug uptake process,
which is related to the balance between its lipophilicity
and hydrophilicity. Actually, for a drug to reach its site
of action it must be able to interact with two different
environments: lipophilic (e.g., membranes) and aqueous
(e.g., cytoplasm). The lipophilicity of a phenolic ester,
which is expected to increase with its alkyl chain
(
L-132). All the results presented were obtained by two
independent methods: cell density measurement––
Trypan blue exclusion method––and cellular viability
assessment––MTT colorimetric assay. Cisplatin (cis-
diaminedichloroplatinum(II), CDDP, 9) was used for
comparison purposes in all the experiments carried out
along this work, since it is one of the most widely used
chemotherapeutic drugs in clinical practice. Caffeic and
gallic acids, the di and trihydroxylated precursors of
the esters under study, were also considered for com-
parison.
Figures 2 and 3 represent the cytotoxic effect of the
caffeates and gallates tested, respectively, against the
HeLa and L-132 cell lines––cell viability variation as a
function of the incubation time with the drug, for dif-
ferent drug concentrations. From these time and dose–
response curves, it is possible to relate the structural
characteristics of the compounds to their antitumour
activity, thus learning on factors such as specificity of
action and reversibility of the drug effect. Within the
drug concentration range investigated (25–100 lM), it
21
length, is thus an important factor to consider, once it
leads to a higher affinity for the inner (hydrofobic) part
of the lipid bilayer, thus influencing the localisation of
the compound within the cell. The effect of lipophilicity
in the cytotoxic activity of this kind of compounds was
22
recently reported for caffeate analogues : a remarkable
reduction of cytotoxicity was found upon substitution of

3584
S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
5
Figure 2. Time dependence of the cytotoxic effect of caffeates 1, 2 and 3 against HeLa (A) and L-132 (B) cell lines. Cells (3 · 10 cells/mL) were
incubated with the drugs for periods of 24–72 h. Every 24 h, aliquots of the cell suspensions were removed and cell viability was evaluated by the
MTT colorimetric assay. In addition, the drug was removed 72 h after seeding and the cell viability was assessed following a further incubation of
9
6 h. The data are expressed as a percentage of the control MTT reduction (100%) and represent the average ± mean standard deviation from two
ꢁ
independent experiments carried out in triplicate. Intergroup comparison: p < 0:05. (Values for cisplatin (CDDP, 9) are included for comparison.)
5
Figure 3. Time dependence of the cytotoxic effect of gallates 4, 5 and 6 against HeLa (A) and L-132 (B) cell lines. Cells (3 · 10 cells/mL) were
incubated with the drugs for periods of 24–72 h. Every 24 h, aliquots of the cell suspensions were removed and cell viability was evaluated by the
MTT colorimetric assay. In addition, the drug was removed 72 h after seeding and the cell viability was assessed following a further incubation of
72 h. The data are expressed as a percentage of the control MTT reduction (100%) and represent the average ± mean standard deviation from two
ꢁ
ꢁꢁ
independent experiments carried out in triplicate. Intergroup comparison: p < 0:05; p < 0:01. (Values for cisplatin (CDDP, 9) are included for
comparison.)
the ester group by a hydrophilic moiety (either hydroxy
or amide-type).
Table 1. Calculated partition coefficients (log P) and cytotoxic potency
IC50, lM) of compounds 1–6 towards the human cell lines HeLa and
L-132 (in vitro)
(
a
Table 1 contains the lipophilicity properties––herein
defined by partition coefficients in a two-phase octanol–
water system (log P values)––calculated for the phenolic
compounds under study (1–6). As expected, lipophilicity
increases upon lengthening of the alkyl ester chain, and
was determined to be higher for the caffeates than for
the corresponding gallates. Several studies on different
potential antitumour agents have shown that a maxi-
mum cytotoxic activity is often achieved for intermedi-
ate lipophilicity and water solubility values. This is also
being verified by the present results, the esters yielding a
higher cytotoxic activity––propyl caffeate and gallate––
displaying intermediate log P values (Table 1). A clear
correlation between lipophilicity and anticancer effect,
however, is not always easy to achieve, because the an-
tiproliferative activity also depends upon other factors,
such as reduction potential and structural characteristics
of the compounds––for the esters presently studied, length
of the alkyl chain and degree of OH ring substitution.
Compound
Log P
IC50
Cell line
L-132
HeLa
1
2
3
4
5
6
9
1.21
2.03
4.60
0.77
1.69
4.33
––
––
––
12.0
70.0
––
9.0
22.0
––
8.0
10.0
20.0
2.0
45.0
2.0
IC50 of cisplatin (9) is included for comparison purposes.
a
IC50 values were calculated from dose–response curves for a drug
exposure of 48 h. (Data obtained through independent measurements
of cell density and cell viability, using the Trypan blue method and
the MTT assay.)
Regarding the effect of these structural characteristics, it
was found that the presence and number of OH ring

S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
3585
Figure 4. Plot of cell viability for compounds 2, 5, 7 and 8 against HeLa (A) and L-132 (B) cell lines. The cell viability was determined by the MTT
assay, after 72 h of incubation with the drug. In addition, the drug was removed 72 h after seeding and the cell viability was assessed following a
further incubation of 72 h. The data are expressed as a percentage of the control MTT reduction (always taken as 100%), and represent the aver-
ꢁ
age ± mean standard deviation from two independent experiments carried out in triplicate. Intergroup comparison: p < 0:01.
substituents, in polyphenolic derivatives, is determinant
Actually, it
that esters 2 and 5 display a much higher cytotoxic
activity towards the HeLa malignant cell line as com-
pared to the analogous derivatives tested, displaying
shorter (methyl) or longer (octyl) chains. Also, as ex-
pected, the trihydroxylated compound 5 presents a
lower IC50 value than its dihydroxylated analogue 2 (8.0
vs 12.0 lM). Nevertheless, even these IC50 values are still
higher than desirable for these phenolic derivatives to be
considered as efficient antiproliferative agents. Further-
more, the effect of structural changes on the cytotoxicity
of these phenolic esters was found to be much less evi-
dent for the fibroblasts than for the malignant cells
tested, which may reflect a certain specificity towards
neoplastic cells.
23;24
of the corresponding biological activity.
was even reported that methylation of these hydroxyl
groups, which are mainly responsible for the antioxidant
22
characteristics of phenols, abolished cytotoxicity. In
the present study, it was verified that, for the same
length of the alkyl chain, the trihydroxylated esters
displayed a higher antiproliferative and cytotoxic effect
than the dihydroxylated ones––compounds 5 versus 2
(
Fig. 4). This is in total agreement with previous results
on several analogous phenolic acids (namely caffeic (7)
and gallic (8) acids), which presented a similar behaviour
in several human cancer cell lines (from leukaemia to
25
adenocarcinomas).
Furthermore, all the caffeates presently displaying anti-
tumoural properties have an E orientation of the aro-
matic ring and the C@O group relative to the carbon
chain double bond. In fact, it was previously reported by
other authors that an E conformation seems to be
The degree of reversibility of both the antiproliferative
and cytotoxic effects displayed by compounds aimed for
use as antitumour drugs is, understandably, of the ut-
most importance. In fact, consideration of a recovery
period following drug exposure, apart from allowing the
determination of the reversibility of the drug effect, as
well as the measurement of a possible delayed cytotox-
icity, avoids over or underestimation of the level of cell
kill achieved. The present results indicate that, except
for the methyl derivatives (which yield a rather low
cytotoxic effect), both propyl and octyl gallates display a
more reversible cytotoxic activity against HeLa cells (up
to ca. 75%) than the corresponding caffeates (only up to
ca. 60%). However, cell viability recovery after drug
removal was found to be higher for this cancer cell line
(up to 60–75%) than for fibroblasts (recovery up to 40–
50%).
22
essential for the cytotoxic activity of such compounds.
Also, the well-known ability of polyphenolic com-
17
pounds to form dimeric structures should be taken
into account, once their conformational preferences,
and consequently their chemical properties, vary con-
siderably upon dimerisation.
A significant disadvantage of the clinically used che-
motherapeutic agents consists in its rather high toxicity,
mainly due to side reactions with other biomolecules
apart from DNA (e.g., proteins). In order to assess this
factor for the compounds now investigated, experiments
were carried out in non-neoplastic cells––L-132 human
fibroblasts. These assays allowed to conclude that all the
compounds studied present some degree of nonrevers-
ible toxicity against healthy cells, that is quite high for
the propyl esters (2 and 5) for which viability decreases
to about 30% (Figs. 2–4). No significant differences have
been detected between caffeates and gallates.
Despite the significative toxicity of the esters under
study towards fibroblasts, they were found to be far
more cytotoxic towards human cancer cells (almost
threefold) than the corresponding caffeic (7) or gallic (8)
acids (Fig. 4), which is in accordance with previous
5
studies on similar caffeates. In fact, the electrochemical
behaviour of these systems––namely their electron-
donating ability––was previously reported to be affected
by the introduction of alkyl groups in the carboxylate
part of the molecule (ester formation), which may
explain such a marked difference in biological activity
The values of 50% inhibitory concentration (concen-
tration of drug yielding a 50% cell viability decrease,
IC ) measured for the distinct compounds investigated
26
50
are comprised in Table 1. These results clearly evidence

3586
S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
between the acid and the ester form of these phenolic
derivatives.
cient anticancer drugs, structurally based on phenolic
compounds.
3
. Conclusions
4. Experimental
4.1. Chemicals
The phenolic esters investigated for their antitumour
activity in a human adenocarcinoma cell line––methyl,
propyl and octyl caffeates and gallates––varied mainly in
the number of phenolic substituents and/or in the length
of the alkyl side chain, apart from the distance between
the carboxylic group and the aromatic ring (depending
on whether they were caffeates or gallates).
Antibiotics (penicillin–streptomycin 100· solution), trans-
caffeic acid ([trans-3-(3,4-dihydroxyphenyl)-2-propenoic
acid)], 3,4-dihydroxycinnamic acid), cisplatin (cis-di-
aminedichloroplatinum(II)), DMEM-HG culture med-
ium, EDTA (ethylenediamine–tetraacetic acid, disodium
salt, dihydrate), gallic acid (3,4,5-tri-hydroxybenzoic
acid), glutamine, HEPES (N-[2-hydroxy-ethyl]pipera-
A structural analysis of these compounds was carried
out by theoretical (ab initio) methods, once structurally
different esters were found to yield distinct cell density
and viability results, thus allowing the establishment of
structure–activity relationships. The geometrical pref-
erences of the compounds studied are mainly deter-
mined by electrostatic factors, as well as by the
formation of both intra and intermolecular hydrogen
bonds (e.g., (O)Hꢀ ꢀ ꢀO or even (C)Hꢀ ꢀ ꢀO). Conse-
quently, the most stable structures calculated for these
systems displayed a clear preference for a planar
geometry. In the case of caffeates, the presence of an
ethylenic spacer allows the formation of a conjugated
system, strongly stabilised through p-electron delocali-
sation.
0
zine-N -[4-butane-sulfonic acid]), methanol, methyl
gallate (methyl 3,4,5-trihydroxybenzoate), MTT (3-
[4,5-dimethyl-thiazol-2-yl]-2,5-diphenyltetrazolium bro-
mide), phenol red (phenol-sulfonphthalein), n-propanol,
propyl gallate (propyl 3,4,5-trihydroxybenzoate), octyl
gallate (octyl 3,4,5-trihydroxybenzoate), n-octanol,
Trypan blue (0.4% solution, prepared in 0.81% sodium
chloride and 0.06% dibasic potassium phosphate),
trypsin, trisodium citrate, inorganic salts and acids (of
analytical grade) were purchased from Sigma-Aldrich
Chemical Co. (Madrid, Spain). Dimethylsulfoxide-d
6
(99.8%) and tetramethylsilane (TMS) were obtained
from E. Merck, Darmstadt, Germany. Fetal calf serum
was obtained from Biochrom KG, Berlin. All other re-
agents and solvents were pro analysis grade, purchased
from Merck (Lisbon, Portugal).
The final aim of this type of studies is the investigation
of new anticancer agents, since chemotherapeutic drugs
presently used in clinical practice display a high toxicity
towards healthy cells and often lead to the development
of resistance. The present work allowed to conclude on
the clearly higher growth inhibition and cytotoxic
activity of phenolic esters as compared to their corre-
sponding acids. Also, the effect of slight structural
variations in this kind of systems was found to cause a
marked change in their biological activity. In fact, di
and trihydroxylated propyl esters––both caffeates and
gallates––displayed a far more pronounced growth-
inhibition activity towards the cell lines tested than their
methyl and octyl analogues. Although these particular
compounds presented some drawbacks for its use as
antitumour agents––namely their hardly reversible
cytotoxicity towards non-cancer cells––they also dis-
played a good water solubility. Knowing that propyl
gallate (5) and octyl gallate (6) are commonly used
additives in food and pharmaceutical industry, the
results presently described––evidencing a significant
cytotoxic activity of compound 5 against healthy cells––
should not be disregarded.
The HeLa cell line was purchased from the European
Collection of Cell Cultures (ECACC, United Kingdom),
while the L-132 cells were made available by the Centre
of Experimental Medicine and Surgery of the University
General Hospital Gregorio Mara n~ on, Madrid (Spain).
4.2. General synthetic procedure
Methyl caffeate (methyl trans-3-(3,4-dihydroxyphenyl)-
2-propenoate, MC) and propyl caffeate (propyl trans-3-
(3,4-dihydroxyphenyl)-2-propenoate, PC). The propyl
and methyl esters of caffeic acid were synthesised by
Fischer esterification, following the procedure described
27;28
by Borges and coworkers, with slight modifications:
Caffeic acid (1.0 g) was refluxed for ca. 5 h with 150 mL
4
of methanol or n-propanol, containing 2 mL of H SO ,
2
until the caffeic acid was no longer detected by TLC.
After cooling, the solvent was partially evaporated un-
der reduced pressure and the solution was neutralised
2 3
with 10% Na CO . The mixture was then extracted with
diethyl ether (3 · 150 mL). The organic phases were
combined, washed with water (3 · 100 mL), dried over
When studying the potential antiproliferative role of this
type of polyhydroxylated phenols, one should attend to
properties such as size, degree of ring hydroxyl substi-
tution and length of the alkyl chain, as well as lipophi-
licity. The understanding of the structure–activity
relationships ruling their biological function lies in more
specific assays, that target particular binding sites and
mechanisms of action. This knowledge is of the utmost
importance for the development of new and more effi-
4
MgSO and concentrated under reduced pressure. The
residue was purified by column chromatography (silica
gel, petroleum ether/ethyl ether). The structural data for
methyl and propyl caffeates is in accordance with the
28
results reported in the literature.
Octyl caffeate (octyl trans-3-(3,4-dihydroxyphenyl)-2-
propenoate, OC). Caffeic acid (5 g) and anhydrous

S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
3587
dimethylformamide (0.1 mL) were dissolved in anhy-
4.5. Preparation of solutions
All compounds studied were water soluble in the con-
centration range used––from 2.5 · 10 to 1.0 · 10 M.
Solutions were prepared in phosphate buffered saline
drous ethyl ether (30 mL). Oxalyl chloride (7.3 mL) was
added dropwise and the solution was stirred for 1 h at
room temperature and another hour at 50 ꢁC. Octanol
ꢂ5
ꢂ4
(
6 mL) was then added and the mixture stirred for 8 h.
ꢂ3
ꢂ3
After cooling, the solvent was partially evaporated
under reduced pressure giving a yellowish-brown oil
residue, which was dissolved in chloroform (75 mL).
The organic phase was washed with 5% NaH-
solution (PBS): 132.0 · 10 M NaCl, 4.0 · 10 M KCl;
ꢂ3
ꢂ3
M
1.2 · 10
M
NaH
2
PO
4
;
1.4 · 10
MgCl
2
;
6.0 ·
ꢂ3
ꢂ2
10 M glucose; 1.0 · 10 M HEPES (N-[2-hydroxy-
ethyl]piperazine-N -[4-butane-sulfonic acid]). Fresh
0
CO
MgSO
3
(3 · 100 mL) and water (3 · 100 mL) and dried over
. After solvent evaporation, the crude product
solutions were prepared monthly and kept from light, in
order to prevent oxidation. Trypan blue was used as a
0.04% (w/v) solution in PBS. MTT (3-[4,5-dimethylthi-
azol-2-yl]-2,5-diphenyl-tetrazolium bromide) was pre-
4
was purified by column chromatography (silica gel,
chloroform). The fractions containing the product were
collected and crystallised (petroleum ether/chloroform)
pared, in a concentration of 0.5 mg/mL, in PBS solution
ꢂ3
giving a white solid compound. Yield 35%; FTIR tmax
containing 1.0 · 10 M CaCl
2
.
ꢂ1
(
cm ): 3489, 3315, 2954, 2979, 2850, 1683, 1624, 1604,
530, 1472, 1441, 1309, 1281, 1241, 1177, 1154, 1107,
1
1
1
1
031, 974, 815; H NMR d: 0.85 (3H, t, J ¼ 7:1, CH ),
.25 (10H, s, CH ), 1.61 (2H, m, CH ), 4.09 (2H, t,
3
4.6. Ab initio MO calculations
2
2
J ¼ 6:6, OCH
2
), 6.25 (1H, d, J ¼ 15:9, H(a)), 6.75 (1H,
Full geometry optimisation and calculation of the har-
monic vibrational frequencies––were performed using
d, J ¼ 8:0, H(5)), 6.99 (1H, dd, J ¼ 8:2, 1.9, H(6)), 7.04
29
(
(
1H, d, J ¼ 1:9, H(2)), 7.46 (1H, d, J ¼ 15:9, H(b)), 9.20
the GAUSSIAN 98W program, within the density func-
tional theory (DFT) approach, in order to properly ac-
count for the electron correlation effects (particularly
important in this kind of systems). The widely employed
13
1H, s, OH), 9.50 (1H, s, OH); C NMR d: 13.9 CH
3
2
,
,
2
3
1
2.1 CH
1.2 CH
2 2 2 2
, 25.4 CH , 28.3 CH , 28.6. CH , 28.7 CH
2
, 114.0 C(2), 114.8 C(a), 115.7 C(5), 121.3 C(6),
30–35
25.5 C(1), 145.0 C(4), 145.5 C(b), 148.4 C(3), 166.6
hybrid method denoted by B3LYP,
which includes a
þꢃ
(
C@O); EI-MS m=z (%): 292 (M , 60), 180 (100); 163
mixture of HF and DFT exchange terms and the gra-
dient-corrected correlation functional of Lee, Yang and
(
57), 136 (27), 123 (16), 89 (25); mp 110–112 ꢁC.
3
6;37
38;39
Parr,
was used, along with the double-zeta split valence basis
as proposed and parametrised by Becke,
40;41
set 6-31G**.
mised by the Berny algorithm, using redundant internal
Molecular geometries were fully opti-
4
.3. Apparatus
42
coordinates : The bond lengths to within ca. 0.1 pm and
the bond angles to within ca. 0.1ꢁ. The final root-mean-
square (rms) gradients were always less than
Infrared spectra were recorded on an ATI Mattson
Genesis Series FTIR spectrophotometer using potas-
sium bromide disks. Only the most significant absorp-
tion bands are reported (tmax, cm ). H and C NMR
data were acquired (at room temperature) on a Br u€ ker
AMX 300 spectrometer, operating at 300.13 and
ꢂ4
ꢂ1
ꢂ1
3
· 10 hartree bohr or hartree radian . No geomet-
ꢂ1
1
13
rical constraints were imposed on the molecules under
study.
7
6
5.47 MHz, respectively. Dimethylsulfoxide-d was used
as a solvent; chemical shifts are expressed in d (ppm)
values relative to tetramethylsilane (TMS) as an internal
reference; coupling constants (J) are given in Hz. Elec-
tron impact mass spectra (EI-MS) were carried out on
a VG AutoSpec instrument. The data are reported as
m=z (% of relative intensity of the most important
4
.7. Log P calculations
Estimated partition coefficients were calculated accord-
43
ing to Parham et al.
fragments). Melting points were obtained on
K o€ fler microscope (Reichert Thermovar) and are
uncorrected.
a
4
.8. Biological assays
Cell culture: Stock cultures of cells were maintained at
7 ꢁC, under 5% CO . HeLa, and L-132 (grown in
3
2
monolayers) were kept in Dulbecco’s modified Eagle’s
high glucose (4500 mg/L) medium (DMEM-HG), sup-
plemented with 10% heat-inactivated fetal calf serum,
glutamine (1.168 g/L) and antibiotics (100 units of pen-
icillin and 100 mg streptomycin). The cell lines were
subcultured twice a week. HeLa cells were harvested
4
.4. Other conditions
Thin-layer chromatography (TLC) was carried out on
precoated silica gel 60 F254 and cellulose plates (E.
Merck). The layer thickness was 0.2 and 0.1 mm,
respectively. The following systems were used for ana-
lytical control: silica gel, petroleum ether/diethyl ether/
formic acid (5:5:0.1); silica gel, chloroform/methanol/
formic acid (9:1:0.5); cellulose, water/acetic acid (9:1).
The spots were visualised under UV detection (254 and
using
a
dissociation
medium
ꢂ3
composed
of
M
ꢂ3
ꢂ3
136.9 · 10 M NaCl, 2.7 · 10 M KCl, 8.2 · 10
ꢂ3
ꢂ4
Na HPO , 1.5 · 10 M KH PO , 4.0 · 10 M EDTA
2
4
2
4
(ethylene–diaminetetraacetic acid, disodium salt, dihy-
drate) (pH 7.4) and containing 0.0004% (w/v) phenol
red. L-132 cells were harvested with a trypsin/
EDTA solution (0.05% trypsin, 0.35 mM EDTA.4Na,
366 nm) and iodine vapour. Solvents were evaporated in
a Buchi Rotavapor.

3588
S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
2
þ
reconstituted in balanced salt solution without Ca or
þ
2. Leu, T. H.; Maa, M. C. Curr. Med. Chem. Anti-Canc.
Agents 2002, 2, 357.
2
Mg ).
3
. Tapiero, H.; Tew, K. D.; Ba, G. N.; Mathe, G. Biomed.
Pharmacother. 2002, 56, 200.
4
. Mahmoud, N. N.; Carothers, A. M.; Grunberger, D.;
Bilinski, R. T.; Churchill, M. R.; Martucci, C.; Newmark,
H. L.; Bertagnolli, M. M. Carcinogenesis 2000, 21, 921.
. Nagaoka, T.; Banskota, A. H.; Tezuka, Y.; Saiki, I.;
Kadota, S. Bioorg. Med. Chem. 2002, 10, 3351.
4
.9. Toxicity and cell growth inhibition evaluation
Cytotoxicity and cell density evaluation after drug
ꢂ5
5
exposure––for drug concentrations between 2.5 · 10 to
ꢂ4
1
.0 · 10 M––was assessed with use of standard assays.
6. Nomura, M.; Kaji, A.; Ma, W.; Miyamoto, K.; Dong, Z.
Mol. Carcinog. 2001, 31, 83.
7. Etzenhouser, B.; Hansch, C.; Kapur, S.; Selassie, C. D.
Bioorg. Med. Chem. 2001, 9, 199.
5
Cells were plated at 5 · 10 cell/mL. Twenty four hours
after seeding, drug solutions were added to the medium
and the cultures were incubated at 37 ꢁC. Cells were
harvested and analysed (both in controls and in drug-
treated cultures) every 24 h, for a total period of 3 days.
Reversibility of the drug effect was tested by removing
the drug and adding fresh culture medium in the last day
of incubation with the drug, and assessing the cell via-
bility following three to four more days of incubation.
Cell density and viability were determined by Trypan
blue exclusion on single-cell suspensions obtained from
the monolayer cultures. Cell viability was further as-
sessed by mitochondrial dehydrogenase activity––MTT
8
. Masak, H.; Okamoto, N.; Sakaki, S.; Sakurai, H. Biol.
Pharm. Bull. 1997, 20, 304.
. Klein, E.; Weber, N. J. Agric. Food Chem. 2001, 49, 1224.
9
1
1
0. Murase, T.; Kume, N.; Hase, T.; Shibuya, Y.; Nishizawa,
Y.; Tokimitsu, I.; Kita, T. Arterioscler. Thromb. Vasc.
Biol. 1999, 19, 1412.
1. Serrano, A.; Palacios, C.; Roy, G.; Cespon, C.; Villar, M.
L.; Nocito, M.; Gonzalez-Porque, P. Arch. Biochem.
Biophys. 1998, 350, 49.
12. Inoue, M.; Suzuki, R.; Koide, T.; Sakaguchi, N.; Ogihara,
Y.; Yabu, Y. Biochem. Biophys. Res. Commun. 1994, 204,
898.
44–47
assay.
Measurements were carried out each 24 h and
1
1
1
3. Inoue, M.; Sakaguchi, N.; Isuzugawa, K.; Tani, H.;
Ogihara, Y. Biol. Pharm. Bull. 2000, 23, 1153.
4. Isuzugawa, K.; Ogihara, Y.; Inoue, M. Biol. Pharm. Bull.
the results expressed as a percentage of the control (non-
treated) cells, which was always taken as 100%. All
experiments were performed in triplicate.
2
001, 24, 249.
5. Palacios, C.; Cespon, C.; Martin de la Vega, C.; Roy, G.;
Serrano, A.; Salinas, M.; Gonzalez-Porque, P. J. Enzyme.
Inhib. 2001, 16, 527.
The 50% inhibitory concentration (concentration of
drug required to inhibit cell growth by 50%, IC ) was
5
0
calculated, for each compound tested, from dose-
response curves (for an incubation period of 48 h).
16. Van Besien, E.; Marques, M. P. M. J. Molec. Struct.
(Theochem) 2003, 625, 265.
1
7. Fiuza, S. M.; Van Besien, E.; Milhazes, N.; Borges, F.;
Marques, M. P. M. J. Molec. Struct. 2004, 693(1–3), 103.
8. Jeffrey, G. A.; Yeon, Y. Acta Cryst. 1990, B46, 519.
9. Okabe, N.; Kyoyama, H. Acta Cryst. 2002, E58, 245.
0. Okabe, N.; Kyoyama, H. Acta Cryst. 2002, E58, 442.
1. M ꢀa saki, H.; Okamato, N.; Sakaki, S.; Sakurai, H. Biol.
Pharm. Bull. 1997, 20, 304.
1
1
2
2
4
.10. Statistical analysis
All experiments were performed in triplicate. The results
are expressed as mean values ±SD, (the corresponding
error bars being displayed in the graphical plots). Sta-
tistical analysis were performed using ANOVA, fol-
lowed by post hoc test of Fisher’s Protected Least
Significant Difference. Statistical comparison between
the data was based on the Pearson correlation coeffi-
cient, values less than 0.05 being considered as signifi-
cant.
2
2. Nam, N.; You, Y.; Kim, Y.; Hong, D.; Kim, H.; Ahn, B.
Z. Bioorg. Med. Chem. Lett. 2001, 11, 1173.
23. Abe, I.; Kashiwagi, Y.; Noguchi, H. Bioorg. Med. Chem.
Lett. 2000, 10, 2525.
2
2
2
4. Huang, M. T.; Ma, W.; Yen, P.; Xie, J. G.; Han, J.;
Frenkel, K.; Grunberger, D.; Conney, A. H. Carcinogen-
esis 1996, 17, 761.
5. Gomes, C. A.; Gir ~a o da Cruz, M. T.; Andrade, J. L.;
Milhazes, N.; Borges, F.; Marques, M. P. M. J. Med.
Chem. 2003, 46, 5395.
6. Gunckel, S.; Santander, P.; Cordano, G.; Ferreira, J.;
Munoz, S.; Nunes-Vergara, L. J.; Squella, J. A. Chem.
Biol. Interact. 1998, 114, 45.
Acknowledgements
L.J.T. acknowledges financial support from FCT––PhD
fellowship SFRH/BD/8580/2002. The authors thank the
Centre of Experimental Medicine and Surgery of the
University General Hospital Gregorio Mara n~ on, Ma-
drid, Spain (in the person of Dr. Maria Concepci oꢀ n
Guisasola Zululeta), for the kind offer of the L-132 cell
line.
2
2
7. Borges, M. F. M.; Pinto, M. M. M. J. Liq. Chromatogr.
1
994, 17, 1125.
8. Silva, F. A. M.; Borges, F.; Guimaraes, C.; Lima, J. L. F.
C.; Matos, C.; Reis, S. J. Agric. Food Chem. 2000, 48,
2122.
~
29. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G.
E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.;
Montgomery, J. A. Jr.; Stratmann, R. E.; Burant, J. C.;
Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.;
Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.;
Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.;
Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko,
References and notes
1
. Rao, C. V.; Desai, D.; Rivenson, A.; Simi, B.; Amin, S.;
Reddy, B. S. Cancer Res. 1995, 55, 2310.

S. M. Fiuza et al. / Bioorg. Med. Chem. 12 (2004) 3581–3589
3589
A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R.
38. Becke, A. Phys. Rev. 1988, A38, 3098.
39. Becke, A. J. Chem. Phys. 1993, 98, 5648.
40. Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28,
213.
41. Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.;
Gordon, M. S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys.
1982, 77, 3654.
L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J.
A. GAUSSIAN 98, Revision A.9, Gaussian, Pittsburgh PA,
USA, 1998.
3
0. Russo, T. V.; Martin, R. L.; Hay, P. J. J. Phys. Chem.
42. Peng, C.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J.
Comp. Chem. 1996, 17, 49.
43. Parham, M.; Hall, L.; Kier, L. In Accurate Prediction of
LogP Using E-Sate Indices with Neural Network Analysis.
220th ACS National Meeting, Washington DC, USA,
2000.
1
995, 99, 17085.
1. Ignaczak, A.; Gomes, J. A. N. F. Chem. Phys. Lett. 1996,
57, 609.
3
2
3
3
2. Cotton, F. A.; Feng, X. J. Am. Chem. Soc. 1997, 119, 7514.
3. Ignaczak, A.; Gomes, J. A. N. F. J. Electroanal. Chem.
1
997, 420, 209.
44. Slater, T. F.; Sawyer, B.; Strauli, U. D. Biochim. Biophys.
Acta 1963, 77, 383.
45. Mosmann, T. J. Immunol. Meth. 1983, 65, 55.
46. Ramanathan, R.; Das, N. P.; Tan, C. H. Free Rad. Biol.
Med. 1994, 16, 43.
3
3
4. Wagener, T.; Frenking, G. Inorg. Chem. 1998, 37, 1805.
5. Cotton, F. A.; Feng, X. J. Am. Chem. Soc. 1998, 120,
3
387.
3
3
6. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. 1988, B37, 785.
7. Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys.
Lett. 1989, 157, 200.
47. Rao, C. V.; Rivenson, A.; Simi, B.; Reddy, B. S. Cancer
Res. 1995, 55, 259.