New derivatives of silybin and 2,3-dehydrosilybin and their cytotoxic and P-glycoprotein modulatory activity

Article abstract of DOI:10.1016/j.bmc.2006.01.035

Large series of O-alkyl derivatives (methyl and benzyl) of silybin and 2,3-dehydrosilybin was prepared. Selective alkylation of the silybin molecule was systematically investigated. For the first time we present here, for example, preparation of 19-nor-2,

Full text of DOI:10.1016/j.bmc.2006.01.035

Bioorganic & Medicinal Chemistry 14 (2006) 3793–3810
New derivatives of silybin and 2,3-dehydrosilybin and
their cytotoxic and P-glycoprotein modulatory activity
a,ꢀ
a,ꢀ
c
b
b
´
´
´
´
´
Petr Dzˇubak, Marian Hajdu´ch, Radek Gazˇak, Alena Svobodova, Jitka Psotova,
b
c
c,
*
´
´
ˇ
Daniela Walterova, Petr Sedmera and Vladimır Kren
aLaboratory of Experimental Medicine, Department of Paediatrics, Faculty of Medicine, Palacky´ University and
University Hospital in Olomouc, Pusˇkinova 6, CZ-775 20, Czech Republic
bInstitute of Medical Chemistry and Biochemistry, Faculty of Medicine, Palacky´ University, Hneˇvotı´nska´ 3,
CZ-775 15 Olomouc, Czech Republic
cInstitute of Microbiology, Academy of Sciences of the Czech Republic, Vı´denˇska´ 1083, CZ-142 20 Prague, Czech Republic
Received 13 July 2005; revised 11 January 2006; accepted 17 January 2006
Available online 8 February 2006
Abstract—Large series of O-alkyl derivatives (methyl and benzyl) of silybin and 2,3-dehydrosilybin was prepared. Selective alkyl-
ation of the silybin molecule was systematically investigated. For the first time we present here, for example, preparation of 19-
nor-2,3-dehydrosilybin. All prepared silybin/2,3-dehydrosilybin derivatives were tested for cytotoxicity on a panel of drugs sensitive
against multidrug resistant cell lines and the ability to inhibit P-glycoprotein mediated efflux activity. We have identified effective
and relatively non-cytotoxic inhibitors of P-gp derived from 2,3-dehydrosilybin. Some of them were more effective inhibitors at con-
centrations lower than a standard P-gp efflux inhibitor cyclosporin A. Another group of 2,3-dehydrosilybin derivatives also had
better inhibitory effects on P-gp efflux but a cytotoxicity comparable with that of parent 2,3-dehydrosilybin. Structural requirements
for improving inhibitory activity and reducing toxicity of 2,3-dehydrosilybin were established. Effect of E-ring substitution as well as
an influence of the substituent size at the C-7-OH position of A-ring on P-gp-inhibitory activity was evaluated for the first time in
this study.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
Silymarin obtained from the seeds of Silybum marianum
(milk thistle) has been used since the time of ancient
physicians and herbalists to treat a range of liver and
gallbladder disorders, including hepatitis, cirrhosis,
and jaundice, and to protect the liver against poisoning
from chemical and environmental toxins, including
snakebites, insect stings, mushroom poisoning, and
alcohol. These are typical applications, for which silym-
arin, its main component silybin (1), and their prepara-
tions have been used and prescribed until now.1,2
Abbreviations: P-gp, P-glycoprotein; LRP, lung resistance-related pr-
otein; LNCaP, androgen dependent prostate cancer cell line; HUVEC,
human umbilical vein endothelial cells; VEGF, vascular endothelial
growth factor; VEGR3, VEGF receptor 3; MDR, multidrug resis-
tance; MRP1, multidrug resistance protein 1; DNR, daunorubicin;
MTT assay, 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bro-
mide based cytotoxic assay; TCS, tumor cell survival; CyA, cyclo-
sporin A; EC50, half-maximal effective concentrations; Imax, maximum
efflux inhibition; NBD, nucleotide-binding domain of P-glycoprotein;
CEM, human T-lymphoblastic leukemia cell line CCRF-CEM; CEM-
VCR, vincristine resistant CEM cells; CEM-DNR, daunorubicin res-
istant CEM cells; CEM-DNR 0.3A2, daunorubicin resistant subclone
of CEM-DNR cell line; K562, human myeloid leukemia cell line; K-
562-Tax, paclitaxel resistant K562 cells; L1210, mouse lymphocytic
leukemia cell line; L1210-VCR, vincristine resistant L1210 cells; A549,
human lung adenocarcinoma cell line; MCF7, human breast cancer
cell line; NHL, normal human lymphocytes.
Recently, however, silybin received attention due to its
alternative beneficiary activities that are not directly
bound to its hepatoprotective and/or cytoprotective
actions.3
Keywords: Silymarin; Silybin; Dehydrosilybin; P-glycoprotein inhibi-
tion; Cancerostatic.
One of the most promising activities of this compound is
its anticancer activity that results at least partially from
its cytoprotective, antioxidant, and chemopreventing
*
Corresponding author. E-mail: kren@biomed.cas.cz
These authors contributed equally to this work.
ꢀ
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.01.035

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3794
properties. Moreover silybin has the potential to be ap-
plied as an adjuvant in the anti-cancer chemotherapy
since it diminishes carcinogenic potential of a number
of chemicals.4
P-gp achieved by direct interaction of a compound with
one or more binding sites on P-gp.26
Besides ABC transport proteins there is another MDR
mechanism represented for instance by the lung resis-
tance-related protein (LRP). It is located intracellularly
in vaults and transports toxic agents from the nucleus of
neoplastic cells to the endosomal compartments.27,28
However, silybin also acts on the receptor level affect-
ing various processes involved either in cancerogenesis
and/or tumor proliferation. Modulations of various
mitogenic, signaling, apoptotic, and cell-cycle regula-
tors by silybin and silymarin were observed.5–13 In a
human prostate cancer, silymarin inhibits mitogenic
signaling pathways and alters cell cycle regulators,
leading to growth inhibition and death of advanced
Silymarin increased daunomycin accumulation in P-gp-
positive cells, but not in P-gp-negative cells, in a drug
concentration and P-gp expression level-dependent
manner.29 Silymarin potentiated doxorubicin cytotoxic-
ity in P-gp-positive cells, while it inhibited P-gp ATPase
activity and azidopine photoaffinity labeling of P-gp,
suggesting a direct interaction with P-gp substrate bind-
ing.29 The 1.5 h accumulation of digoxin and vinblastine
in Caco-2 cells was significantly increased by silymarin,
and this effect was concentration dependent.30 In the hu-
man prostate carcinoma DU145 cells, silybin potentiat-
ed doxorubicin-induced growth inhibition and
apoptosis.31 This P-gp inhibitory activity was pro-
nounced in the 2,3-dehydrosilybin (2) derivatives carry-
ing prenyl or geranyl substituents.32 More recently, the
capacity of silybin derivatives to chemosensitize drug-re-
sistant cells with specific MDR transporter profile has
been investigated.33 In addition to P-gp, some flavo-
noids related to silybin were found to inhibit another
important MDR-associated protein, the multidrug resis-
tance protein 1 (MRP1).33,34
and
androgen-independent
prostate
carcinoma
cells.14,15 Some steroid hormone-dependent tumors
are also inhibited by silybin, since both the silymarin
and silybin potentiated activity of antiandrogens in
androgen dependent prostate cancer cell line
LNCaP.16,17
Anticarcinogenic and anti-inflammatory effects of silym-
arin and silybin might be also related to the inhibition of
the transcription factor NF-jB, which regulates and
coordinates the expression of various genes involved in
the inflammatory process, in cytoprotection and carci-
nogenesis. Silymarin was highly active in the inhibition
of NF-jB factor activation by various signals.18–21 New-
ly discovered effects of silymarin and silybin, for exam-
ple, on the cell cycle regulation, apoptosis, nuclear and
estrogenic receptors, and many recently reviewed by
3
ˇ ´
Kren and Walterova.
The aim of our study was to investigate in vitro anti-
cancer and P-gp inhibitory activity of oxidized and
selectively alkylated silybin derivatives in order to iden-
tify the potential anti-cancer and/or chemosensitizing
compounds. Selective alkylation of the silybin hydroxyls
was systematically investigated and a series of novel sily-
bin and 2,3-dehydrosilybin derivatives were prepared
(see Chart 1).
Silybin was also investigated as cytoprotective agent
and/or sensitizer for anticancer chemotherapy using
classical cytostatics such as cisplatin or adriamycin with-
out compromising antitumor activity.22–24 On the con-
trary, potentiation of anticancer activity of those drugs
by silybin was clearly demonstrated.22
One of the plausible explanations for synergistic effects
of silybin and anti-cancer drugs is inhibition of efflux
function of dominant multidrug resistance (MDR)
transporter P-glycoprotein.25 P-gp is a prominent mem-
ber of ABC transporters protein family, that has a sim-
ilar architecture plan comprising four major domains:
two membrane bound domains, each with six transmem-
brane segments, and two cytosolic ATP binding motifs,
commonly known as the Walker A and B domains, that
bind and hydrolyze ATP (also known as the nucleotide-
binding domains, or NBDs). Both ATP binding sites are
essential for drug efflux. P-gp binds its substrates in the
cytosolic membrane leaflet most likely via H-bond for-
mation and moves them to the extracellular leaflet or
to the extracellular aqueous environment at the expense
of one to two molecules of ATP.
2. Results and discussion
2.1. Chemistry
Silybinic acid (3) and 2,3-dehydrosilybinic acid (4) were
prepared by CrO3/H5IO6 oxidation of selectively acety-
lated derivatives of silybin (1) and 2,3-dehydrosilybin
(2) in acetonitrile followed by deacetylation of the oxi-
dized products.35
2.1.1. Preparation of monoacyl derivatives of silybin.
Monoacylations of silybin (1) catalyzed by BF3ÆEt2O in
CH2Cl2/CH3CN led to 7-O-benzoylsilybin (6) and 23-O-
pivaloylsilybin (7) (Scheme 1).
Modulation or inhibition of P-gp activity can be
achieved by: (i) an inhibition of ATP binding, ATP
hydrolysis or coupling of ATP hydrolysis to the translo-
cation of substrates, (ii) inhibition of conformational
changes required for drug extrusion via antibody bind-
ing to certain extracellular loops of the transporter,
and (iii) non-competitive or competitive inhibition of
2.1.2. Preparation of silybin benzyl- and methyl ethers.
Selective alkylation of the C-7 phenolic hydroxyl group
of silybin was accomplished by refluxing 1 with benzyl
bromide or methyl iodide in acetone in the presence of
K2CO3; corresponding 7-O-benzylsilybin (5) or 7-O-
methylsilybin (10) was formed. 20-O-Methylsilybin
(12) was prepared by methylation of 1 with methyl io-

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3795
R1 R2 R3
R4
1
3
5
6
7
10
12
19
H
H
H
H
H
H
H
H
H
H
Bn
Bz
H
Me
H
H
H
H
H
H
H
CH OH
2
16
B
4
10
R
O
COOH
D
CH OH
2
2
19
20
17
R O
OCH3
7
2
3
O
CH OH
2
CH OPiv
2
CH OH
2
14
11
O
E
A
C
3
OR
4
OH
5
Me CH OH
2
1
OR
O
Me Me CH OH
2
23 Me Me Me CH OH
2
O
CH OH
2
OR
OH
HO
O
O
OH
OH
O
16 R = Bn
17 R = H
R1 R2
R3
H
H
H
H
R4
R5
H
H
H
2
4
8
9
H
H
H
H
H
H
Bn
H
CH OH
2
COOH
CH OH
2
4
R
O
O
CH OH Me
2
11
13
18
20
21
22
H
H
H
H
H
H
Me
H
Me
H
CH OH
H
H
2
2
R O
OCH3
O
Me CH OH
H
2
CH OH Me
2
3
OR5
OR
Me Me CH OH
H
Me Me CH OH Me
H
2
1
Me CH OH Me
2
OR
O
2
24 Me Me Me CH OH
H
2
Chart 1.
23
16
B
4
10
O
CH2OR
Compound
1
2
3
4
(procedure)
D
R
R
R
R
2
19
17
OCH3
7
2
3
R O
O
5 (a)
6 (b)
7 (c)
10 (d)
12 (e)
19 (f)
23 (g)
H
Bn
Bz
H
Me
H
H
H
H
H
Me
H
H
Piv
H
H
H
14
a - g
11
O
E
A
H
H
H
H
H
C
3
20
OR
4
5
OH
1
OR
O
3
4
1
2
Me Me
1 R = R = R = R = H
Me Me Me
H
Scheme 1. Reagents and conditions: (a) BnBr, K2CO3, acetone, reflux, 3.5 h, 81%; (b) PhCOCl, BF3ÆEt2O, CH2Cl2/CH3CN (1:1, v/v), rt, 42 h, 90%;
(c) (CH3)3COCl, BF3ÆEt2O, CH2Cl2/CH3CN (1:1, v/v), rt, 12 h, 60%; (d) MeI, K2CO3, acetone, reflux, 3.5 h, 53%; (e) MeI (1.2 equiv), NaH (3 equiv),
DMF, rt, 2 h, 54%; (f) MeI (2 equiv), NaH (3 equiv), DMF, 1 h at 0 ꢁC, then 1 h at rt, 21%; (g) dimethyl sulfate, K2CO3, acetone, reflux, 2 h, 56%.
dide (1 equiv) in the presence of an excess of NaH in
DMF at 0 ꢁC. The reaction of 1 with 2 equiv of methyl
iodide under the same conditions led to 7,20-di-O-meth-
ylsilybin (19). 5,7,20-Tri-O-methylsilybin (23) formation
was achieved by the action of dimethyl sulfate and
K2CO3 on silybin in boiling acetone. For all described
alkylation methods of silybin, see Scheme 1.
iodide and KOH in DMSO afforded 3,20-di-O-methyl-
2,3-dehydrosilybin (21) and 3,7,20-tri-O-methyl-2,3-
dehydrosilybin (22) (Scheme 2).
2.1.4. Preparation of 19-O-demethyl-2,3-dehydrosilybin.
The preparation of 19-O-demethyl-2,3-dehydrosilybin
(17) was based on the radical-coupling reaction of quer-
cetin with (E)-3-O-benzyl-3,4-dihydroxy-cinnamyl alco-
hol (15) (Refs. 36,37) and followed catalytic
hydrogenolysis of the benzyl group at C-19 of 16
(Scheme 3). Compound 15 can couple to quercetin in
two possible ways, leading either to silybin- (16) or iso-
silybin-type (16a) products. Because of poor solubility
2.1.3. Preparation of 2,3-dehydrosilybin methyl ethers by
direct alkylation of 2. 3-O-Methyl-2,3-dehydrosilybin (9)
was prepared analogously as 20-O-methylsilybin (12) by
methylation of 2,3-dehydrosilybin (2) with methyl iodide
(1 equiv; NaH, DMF, 0 ꢁC). Selective formation of 3,7-
di-O-methyl-2,3-dehydrosilybin (18) was observed in the
reaction of 2 with dimethyl sulfate and K2CO3 in DMF
at room temperature. The alkylation of 2 with methyl
1
of compound 17, it was converted to its peracetate. H
and 13C NMR spectra of this substance were experimen-
tally (COSY, ROESY, TOCSY, HMQC, and gHMQC)

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3796
O
O
CH2OH
3
OCH3
R O
O
O
a
c
1
2
3
4
9 R = Me, R = R = R = H
4
OR
1
OR
2
OR
1
2
3
4
2 R = R = R = R = H
b
2
3
1
4
21 R = R = H, R = R = Me
+
2
4
1
3
2
1
3
4
18 R = R = H, R = R = Me
22 R = H, R = R = R = Me
Scheme 2. Reagents and conditions: (a) MeI (1.1 equiv), NaH (1.2 equiv), DMF, 1.5 h at 0 ꢁC, then 12 h at rt, 45%; (b) dimethyl sulfate, K2CO3,
DMF, rt, 12 h, 32%; (c) MeI, KOH, DMSO, rt, 1 h, 25% of 21 and 24% of 22.
O
O
CH2OH
HOCH2
OH
OH
OR
OH
HO
HO
O
O
O
O
OBn
OH
a
+
OH
OH
OH
OH
15
16 R = Bn
17 R = H
quercetin
b
OH
O
O
OR
O
O
HO
CH2OH
OH
OH
16a R = Bn
Scheme 3. Reagents and conditions: (a) Ag2CO3 (1 equiv), benzene/acetone (2:1, v/v), reflux, 36 h, 10%; (b) H2–Pd/C, THF, 12 h, 96%.
fully assigned. According to the coupling of H-11 to C-
12a, observed by gHMQC, this compound belongs to
the 2,3-dehydrosilybin series (the coupling of H-11 to
C-16a would indicate an isosilybin analogue).
Moreover, Merlini and Zanarotti39 reported the self-
coupling reaction of some derivatives of coniferyl alco-
hol. Thus, the result of this radical-coupling reaction
strongly depends on the reactivity of both starting com-
pounds; the best yields are achieved when their reactivity
is almost the same, which is not our case.
Published yields of an analogous reaction of taxifolin or
luteolin with coniferyl alcohol are usually higher (40–
70%). However, both the above flavonoids are consider-
ably weaker radical scavengers than quercetin.38 Indeed,
we have observed the formation of side products, related
to quercetin radical degradation (e.g., quercetin qui-
none). Thus, the main reaction is not radical coupling
of quercetin with 15 but an elimination of radical species
generated from quercetin leading to its degradation. As
a proof of this speculation behavior of both starting
compounds during reaction can serve, where quercetin
is completely consumed, but 15 remains present in the
reaction mixture (even after an addition of an excess
of the fresh quercetin). Partial radical decomposition
of the product formed could be another reason for its
low yield (due to its rather high antiradical activity).
The starting (E)-3-O-benzyl-3,4-dihydroxycinnamyl
alcohol (15) was prepared from the commercially avail-
able caffeic acid by its esterification leading to the methyl
ester 14, followed by the selective benzylation of the
hydroxyl group at C-3 and finally the reduction of ben-
zylated methyl ester to the respective alcohol 15 by
LiAlH4 (Scheme 4).
2.1.5. Oxidation of silybin derivatives to 2,3-dehydrosily-
bin derivatives. All silybin derivatives with free C-3
hydroxyl group can be oxidized to the corresponding
2,3-dehydrosilybin derivatives using the iodine in glacial
acetic acid in the presence of potassium acetate40
(Scheme 5)—7-O-benzyl-2,3-dehydrosilybin (8), 7-O-

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3797
HOOC
MeOOC
HOCH2
OH
a
OH
OH
b, c
OBn
OH
OH
caffeic acid
14
15
Scheme 4. Reagents and conditions: (a) MeOH, H2SO4, rt, 12 h, 98%; (b) BnCl (1.2 equiv), NaH (2.0 equiv), DMSO, rt, 1 h; (c) LiAlH4, THF, 1 h at
0 ꢁC, then 3 h at rt, 42%.
O
O
CH2OH
O
O
CH2OH
R2O
OCH3
OR3
O
O
R2O
OCH3
OR3
O
a
OH
OH
OR1
OR1
O
1
3
2
8 R = R = H, R = Bn
1
3
2
5 R = R = H, R = Bn
1
3
2
11 R = R = H, R = Me
1
3
2
10 R = R = H, R = Me
1
2
3
13 R = R = H, R = Me
1
2
3
12 R = R = H, R = Me
1
2
3
20 R = H, R = R = Me
1
2
3
19 R = H, R = R = Me
1
3
2
24 R = R = R = Me
1
2
3
23 R = R = R = Me
Scheme 5. Reagents and conditions: (a) I2, glacial AcOH, anhydrous AcOK, reflux for 4 h, then EtOH/HCl, reflux for 1 h, 46–90%.
methyl-2,3-dehydrosilybin (11), 20-O-methyl-2,3-dehy-
drosilybin (13), 7,20-di-O-methyl-2,3-dehydrosilybin
(20), and 5,7,20-tri-O-methyl-2,3-dehydrosilybin (24)
were prepared by this way.
H-21, H-22; C-12a vs C-16a; C-19 vs C-20; C-4 vs carbo-
nyls in the 2-series).
2.2. Biological studies
2.1.6. Structure determination. Since the starting natural
silybin was a nearly equimolar mixture of the A and B
isomers, some NMR signals were doubled (Tables 1
and 2). The correct proton signal multiplicity was there-
fore obtained from HOM2DJ and COSY. Fortunately,
the chemical shift differences were in the ppb (part per
billion) range so that common crosspeaks for both iso-
mers were usually observed in the heterocorrelated
experiments. This problem does not exist with 2,3-dehy-
drosilybins as the signals of both stereoisomers are
undistinguishable. The unambiguous NMR signal
assignment is essential for the structure elucidation.
Therefore, the measurements were mostly performed
in DMSO, in which the observation of OH signals is
possible. The assignment of all hydroxyl groups repre-
sents an indirect way of the site of substitution determi-
nation: it provides the information on sites that are not
affected. Some substituents (OCH3 and OCH2Ph) offer
direct means of their localization: NOE to their neigh-
bor protons (OCH3), H,H-long range couplings
Analysis of the P-gp efflux activity in the presence or ab-
sence of silybin and 2,3-dehydrosilybin derivatives was
based on flow cytometry measurement of content of
the intracellular daunorubicin (DNR) (Fig. 2), a fluores-
cent substrate of P-gp pump.41,42 Herein, we show that
selected silybin derivatives effectively inhibited the P-gp
mediated efflux of DNR from drug resistant and exclu-
sively P-gp expressing K562-tax cells. The inhibition was
concentration-dependent and displayed a classical sig-
moidal dose–response curve (Fig. 3), which suggests a
competitive inhibition with a constant concentration of
DNR. The 50% effective drug concentration (EC50) for
silybin derivatives ranged from 0.1 to 20.1 lM, and thus
in selected cases exceeded potency of standard P-gp
inhibitor cyclosporin A (4.09 lM) (Fig. 3).
According to P-gp inhibitory activity of the compounds
tested (Fig. 3), we can classify them into the following
categories: (a) strong inhibitors: compounds 8, 9, 13,
18, 20, and 22 (Imax > 40% and EC50 < 2 lM). Those
are compounds, which reach at least 50% inhibition
activity of CyA (79%) at concentration lower than half
of the CyA EC50 value (4.09 lM). Compound 9 is
included due to extremely high Imax (130%) at compara-
ble EC50 as CyA; (b) medium inhibitors represented by
the compounds 2, 7, and 21 (Imax > 40% and
EC50 > 2 lM) and (c) the compounds without signifi-
cant inhibitory activity (1, 5, 11, 17, and 24; Imax < 40%).
Biological activity of compounds 3, 4, and 6 was difficult
to analyze or interpret due to solubility problems.
3
(OCH3, OCH2Ph), JC,H (substituents’ protons to the
carbon at the site of substitution, Fig. 1) or even JC,H
4
(AcO protons to the carbon at the site of substitution;
seen in the gHMQC experiment only).
Except for acetates, the position of other acyls was de-
duced from the chemical shift changes in the assigned
spectra. The HMBC experiment also resolves ambigui-
ties arising during the assignment in 1 and 2 derivatives
(C-2 vs C-3; H-6 vs H-8; H-13, H-15, H-16 vs H-18,

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3798
Table 1. 1H NMR data (399.89 MHz, 30 ꢁC) of silybin derivatives
Proton 5a
6a
7a
10b
12b
19b
23b
2
3
5.002 d (11.8) 5.090 d (12.2) 4.984 d (11.9)
4.549 d (11.9) 4.632 d (12.2) 4.543 d (11.9)
5.136 d (11.4)
5.080d (11.3)
5.139 d (11.4)
5.046 d (11.1)
4.668 dd (11.4, 5.4) 4.605 dd (11.3, 6.2) 4.671 dd (11.4, 6.3) 4.381 dd
(11.1, 4.8)
4.655 dd (11.4, 5.4) 4.596 dd (11.3, 6.2) 4.660 dd (11.4, 6.3) 4.369 dd
(11.1, 4.8)
4.536 d (11.9) 4.619 d (12.2) 4.526 d (11.9)
6
6.199 d (2.2) 6.516 d (2.1) 6.035 d (2.2)
6.201 d (2.2)
6.188 d (2.2)
5.915 d (2.1)
6.120 d (2.3)
6.223 d (2.3)
6.033 d (2.2)
8
6.115 d (2.2) 6.461 d (2.1) 5.967 d (2.2)
6.127 d (2.2) 6.447 d (2.1) 5.958 d (2.2)
6.100 d (2.2)
6.094 d (2.2)
5.873 d (2.1)
5.868 d (2.1)
6.099 d (2.3)
6.093 d (2.3)
6.177 d (2.3)
6.171 d (2.3)
10
4.064 m
4.049 m
4.062 ddd
(8.3, 4.1, 2.6) (8.1, 4.1, 3.0)
4.057 ddd 4.227 ddd
(8.3, 4.1, 2.6) (8.1, 4.1, 3.0)
4.236 ddd
4.176 ddd
4.204 ddd
4.206 ddd
4.199 ddd
(7.9, 4.7, 2.5)
4.166 ddd
(7.9, 4.7, 2.5)
(7.8, 4.0, 2.5)
4.189 ddd
(7.8, 4.0, 2.5)
(7.7, 4.6, 2.5)
4.196 ddd
(7.7, 4.6, 2.5)
(7.8, 4.8, 2.6)
4.190 ddd
(7.8, 4.8, 2.6)
11
13
15
4.962 d (8.3) 4.969 d (8.3) 4.894 d (8.1)
4.887 d (8.1)
4.915 d (7.9)
4.967 d (7.8)
4.974 d (7.7)
4.969 d (7.8)
7.210 d (1.8) 7.217 d (2.0) 7.183 d (2.0)
7.188 d (1.9) 7.199 d (2.0) 7.170 d (2.0)
7.099 d (1.8)
7.092 d (1.8)
7.093 d (2.1)
7.087 d (2.1)
7.111 d (2.2)
7.105 d (2.2)
7.081 d (1.9)
7.072 d (1.9)
7.098 dd
(8.3, 1.9)
7.079 dd
(8.3, 1.8)
7.111 dd
(8.3, 2.0)
7.097 dd
(8.3, 2.0)
7.080 dd
(8.5, 2.0)
7.069 dd
(8.5, 2.0)
7.028 dd (8.3, 1.8) 7.011 dd (8.3, 2.1) 7.035 dd
7.017 dd
(8.3, 1.9)
7.010 dd
(8.3, 1.9)
(8.3, 2.2)
7.031 dd
(8.3, 2.2)
16
18
21
22
7.058 d (8.3) 7.066 d (8.3) 7.031 d (8.3)
7.053 d (8.3) 7.059 d (8.3) 7.023 d (8.3)
6.977 d (8.3)
6.974 d (8.3)
6.984 d (8.3)
7.051 m
6.983 d (8.3)
6.980 d (8.3)
6.971 d (8.3)
6.967 d (8.3)
6.942 d (1.8) 6.939 d (1.4) 6.878 d (2.0)
6.873 d (2.0)
7.014 d (2.0)
7.056 m
7.058 d (2.0)
7.056 d (2.0)
6.971 m
6.961 m
6.950 d (7.7)
6.948 d (7.7)
6.803 d (8.1)
6.977 d (8.3)
6.973 d (8.3)
7.007 d (8.1)
6.983 d (8.1)
6.946 m
6.961 m
6.893 dd (7.7, 2.0) 6.866 dd (8.1, 2.0) 7.021 dd (8.3, 2.1) 6.987 dd
(8.1, 2.0)
6.891 dd (7.7, 2.0)
7.004 dd
(8.1, 2.0)
23d
23u
3.811 dd
(12.5, 2.8)
3.809 dd
(12.5, 2.8)
3.815 dd
(12.2, 2.6)
4.395 dd (12.3, 3.0) 3.546 ddd
(12.2, 3.4, 2.5)
4.383 dd (12.3, 3.0)
3.550 dm (11.7)
3.554 ddd
(12.3, 5.1, 2.5)
3.550 ddd
(12.3, 5.2, 2.6)
3.569 dd
(12.5, 4.0)
3.579 dd
(12.2, 4.1)
3.565 dd
(12.2, 4.1)
3.934 dd (12.3, 4.1) 3.351 ddd
3.340 m
3.351 ddd
(12.3, 5.7, 4.6)
3.347 ddd
(12.3, 5.2, 4.8)
(12.2, 4.7, 3.4)
3.920 dd (12.3, 4.1)
5-OMe
7-OMe
—
—
—
—
—
—
3.798 s
3.795 s
3.788 s
3.787 s
—
—
3.789 s
3.787 s
3.798 s
—
19-OMe 3.962 s
20-OMe
3.926 s
3.907 s
3.780 s
3.774 s
3.776 s
3.774 s
—
—
3.767 s
3.765 s
3.770 s
3.768 s
3.768 s
3.765 s
3-OH
5.890 d (5.4)
11.841 s
5.798 d (6.2)
11.875 s
5.854 d (6.3)
11.844 s
5.322 d (4.8)
—
5-OH
11.171 s
11.166 s
11.106 s
11.109 s
11.197 s
5.749 s
7-OH
—
—
—
—
—
—
—
—
20-OH 5.782 s
23-OH
5.740 s
9.122 s
4.934 t (3.4)
4.954 dd (5.7, 5.1) 4.950 dd
(5.2, 5.2)
a CDCl3.
b DMSO-d6. Additional signals—5: 5.081 (2H, m, –CH2–), 7.386 (5H, m, Ph); 6: 7.523 (2H, m, 2· meta-Ph), 7.663 (1H, m, para-Ph), 8.166 (2H,
2· ortho-Ph); 7: 1.221 (9H, s, CMe3), 1.223 (18H, s, 2· CMe3).

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3799
Table 2. 13C NMR data (100.55 MHz, 30 ꢁC) of silybin derivatives
Carbon
2
5a
83.07
6a
83.22
7a
83.05
10b
82.68
12b
82.63
19b
82.71
23b
82.18
83.09
83.18
83.02
82.72
82.59
82.67
3
72.32
72.40
72.73
72.65
72.34
72.30
71.57
71.51
71.54
71.48
71.58
71.52
72.55
72.48
4
195.89
195.90
197.42
197.40
195.75
195.74
198.30
101.40
163.03
94.95
197.73
197.71
198.33
101.40
163.04
94.96
190.13
103.60
161.70
93.00
4a
5
100.98
104.14
100.83
100.82
100.52
163.37
96.16
163.61
163.62
162.89
162.87
163.84
163.83
6
96.41
103.77
159.58
102.29
97.13
97.11
7
167.92
165.62
165.61
167.61
93.87
166.94
95.11
167.61
93.87
165.74
93.63
8
95.38
95.40
95.96
95.94
8a
10
11
12a
13
14
15
16
16a
17
18
19
162.90
162.42
162.40
163.17
163.16
162.41
78.14
162.54
162.52
162.41
78.05
163.76
78.07
78.34
78.37
78.39
78.36
76.05
76.02
78.12
78.09
76.39
76.44
76.43
76.39
76.63
75.85
75.75
75.70
75.70
143.94
143.97
143.98
143.96
143.78
143.28
143.26
143.23
143.21
143.69
143.66
143.18
143.15
116.46
116.61
116.64
116.52
116.54
116.40
116.65
116.58
116.67
116.58
116.67
116.61
116.53
116.43
129.46
129.49
129.02
129.33
129.32
129.94
129.89
130.18
130.14
129.99
129.94
130.29
130.24
121.15
121.12
120.92
121.25
121.08
121.33
121.17
121.42
121.27
121.39
121.25
121.20
121.01
117.20
117.32
117.35
117.25
117.46
117.36
116.37
116.32
111.74
116.40
116.36
116.39
116.34
144.14
144.17
144.27
144.24
144.13
144.12
143.71
143.68
143.69
143.67
143.19
143.18
143.56
143.53
127.86
109.56
146.97
146.48
127.84
127.53
127.52
127.53
129.10
129.08
129.11
109.61
109.58
109.49
111.82
111.76
111.34
111.29
111.31
111.26
111.30
111.26
146.98
146.97
147.05
147.04
147.67
149.22
149.19
149.18
20
21
146.49
146.64
147.07
115.36
148.88
148.84
111.70
148.85
111.70
114.67
114.68
114.69
114.68
114.80
114.79
116.44
116.39
22
120.83
120.89
120.84
120.86
120.54
120.34
120.28
120.29
23
61.70
—
61.91
—
62.71
—
60.21
—
60.21
—
60.17
—
60.18
55.99
55.79
5-OMe
7-OMe
19-OMe
—
—
—
55.97
—
55.98
56.08
56.07
56.08
56.07
56.04
56.03
55.75
—
55.68
—
55.64
—
55.64
—
20-OMe
a CDCl3.
—
—
—
—
55.63
55.59
55.59
b DMSO-d6. Additional signals—5: 70.48 (–CH2–), 127.43 (2· ortho-Ph), 128.40 (para-Ph), 128.74 (2· meta-Ph), 135.56 (ipso-Ph); 6: 128.70 (2· meta-
Ph), 130.32 (2· ortho-Ph), 134.06 (para-Ph), 163.88 (C@O); 7: 27.18 (3· Me), 38.97 (C), 178.25 (piv-CO).

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3800
OH
CH2
8, 9, 11, and 18) and (ii) the compounds selectively toxic
to tumor cells but not for the normal human lympho-
cytes (13, 17, 20, 21, 22, and 24). Interestingly, the com-
pound 22 displayed a substantially higher cytotoxicity in
MDR cells compared to drug sensitive counterparts,
with the only exception of Cem-VCR cell line for which
was compound 22 rather non-toxic (Table 5). Interest-
ingly, drug resistant subline CEM-VCR is the most
LRP expressing cell line in our collection43 and this
may suggest that LRP mediated mechanisms of MDR
participate in resistance to compound 22. Finally, no
significant cytotoxicity was found for silybin itself.
O
O
OCH3
OH
HO
O
O
OH
OH
Figure 1. Diagnostic heteronuclear couplings in 2 observed by HMBC.
Concluding biological part of experimentation, we have
identified new silybin based substances as effective P-gp
inhibitors with no or low cytotoxic capacity against nor-
mal human lymphocytes (7, 13, 20, 21, and 22). Impor-
tantly, all these compounds (except 7) are highly potent
inhibitors at lower doses than the lowest TCS50 value on
the most sensitive cell line. In contrast, the compounds
2, 8, 9, and 18 are effective inhibitors of P-gp efflux
but they also exhibit significant intrinsic cytotoxicity
against all tested cell lines including normal human
lymphocytes.
In accordance with reported findings,25,44 our results
demonstrate the importance of double bond in the posi-
tions C-2 and C-3 of C ring (1 vs 2 or 5 vs 8) for the P-
gp-inhibitory activity of flavonoids. Further important
structural motifs for this activity in simple flavones are
as follows: the presence of free C-3-OH group of ring
C and C-5-OH group of ring A in close proximity to
the carbonyl group in the position C-4 of ring C was
found responsible for high binding affinity to mouse P-
gp NBDs (Ref. 45), while free C-7-OH group did not
produce any effect. On the contrary, we demonstrated
that the substitution of the position C-3-OH of 2,3-
dehydrosilybin (2) by methyl group led in almost all
cases (9, 18, 21, and 22) to a significant improvement
of in vitro anti-cancer activity and capacity to block
functional activity of P-gp (compound 21 exhibits a
slightly lower activity than 2 but at 3.5 times lower con-
centration). A substitution of C-3-OH position does not
reduce cytotoxicity of these compounds (2 vs 9) but the
methylation of C-20-OH decreases cytotoxicity to a
great extent (9 vs 21), while the blocking of C-3-OH
together with C-7-OH by methyl groups does not show
this effect (9 vs 18).
Figure 2. The effects of CyA (6.25 lM) and 2,3-dehydrosilybin
derivative 9 (6.25 lM) on inhibition of P-gp function measured as
daunorubicin efflux in K562-Tax cells. Note increased accumulation of
daunorubicin in CyA/9 treated cells in comparison with cells with no
added inhibitor. Maternal K562 cells are included for illustration.
A comparison of the P-gp inhibitors at one given con-
centration enables to define the order of their inhibition
activity. Thus, the sequence of the activities for good
inhibitors is 22 > 13 > 8 > 20 ꢀ 18 (at the arbitrary ref-
erence concentration 1 lM). Although the compound
9 is a strong inhibitor (Imax = 130%) it exhibited a bio-
logical activity at higher EC50 value (4.8 lM), so it can-
not be compared with the rest of the compounds at the
reference concentration. The order of inhibitory activity
of the second group of compounds was determined as
21 > 2 > 7 (at the arbitrary reference concentration
1 lM).
A relatively high intrinsic cytotoxicity at doses close to
P-gp inhibitory activity of many effective MDR cell
modulators is a major obstacle of their therapeutic
use. Results of the in vitro cytotoxicity measurements
using the MTT assay (Table 5) showed that the drug
concentration lethal to 50% of the cells (TCS50) ranged
from 2.44 to 250 lM in our compounds. Preferential
activity against tumor cell lines compared to resting nor-
mal human lymphocytes was demonstrated, which is an
important observation with potential impact for further
development (Table 5).
Effects of A-ring substitution in the compound 2 (mainly
its C-6 and C-8 positions) to the binding affinity to P-gp
transporters were extensively studied in recent years. An
introduction of either geranyl- or prenyl-group to the
positions C-6 or C-8 of 2 led to significant increase in
binding affinity toward recombinant NBD2 of mamma-
lian MDR cancer cells.25 The same authors showed that
geranylation exhibits higher effect on the binding affinity
than introduction of a smaller prenyl group and that the
substitution of C-8 position had stronger effect on
the binding to the NBD2 than a substitution of the
C-6 one. Similar effect on binding affinity to NBD of
Leishmania tropica P-gp-like transporter was also dem-
onstrated but the prenyl group was more influential than
Tested compounds can be further categorized based on
their intrinsic cytotoxicity (Table 5) into two basic
groups: (i) The compounds toxic for both the tumor cell
lines and also for the normal human lymphocytes (2, 5,

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3801
Figure 3. Inhibition of P-gp mediated daunorubicin efflux in K562-Tax cells by silybin/2,3-dehydrosilybin derivatives compared to reference inhibitor
cyclosporin A. Data are expressed as the percentage of relative fluorescence intensity in untreated cells. The average number of cells per analysis was
5000.
the geranyl group in this case. (Ref. 44) However, an
explanation of higher binding affinity of these com-
pounds compared to 2 consists in a potentiation of their
interaction with hydrophobic steroid-interacting region
vicinal to the ATP-binding site of cytosolic NBD due
to the presence of hydrophobic residue at the A-ring.45
Flavonoids inhibit a number of ATP-utilizing enzymes
by competitive interaction at their ATP site46 but the

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3802
prenylation (or geranylation) of their A-ring significant-
ly increases their selectivity to P-gp-transporter, which
results in the lowering of their cytotoxicity.44
the level of highly toxic 2 but the effect on P-gp inhibi-
tion activity was less remarkable (approximately the
same value of Imax but at 8 times lower concentration
than 1).
Based on these findings we investigated the effect of 2,3-
dehydrosilybin alkylation in the position C-7-OH, which
is between two positions critical for the activity (C-6 and
C-8). Results obtained were somewhat controversial:
introduction of methyl group led to diminished P-gp
inhibition (2 vs 11), while a substitution by the benzyl
group increased the activity (2 vs 8). The cytotoxicity
of 7-O alkyl derivatives was comparable with the cyto-
toxicity of 2 but it was reduced by additional substitu-
tion of C-20-OH (8, 11, and 18—toxic vs 20, 22, and
24—non-toxic).
We have also observed the effect of C-5-OH substitu-
tion: P-gp inhibitory activity of 24 was markedly low-
ered compared to compound 20 (which is consistent
with the observation of Conseil et al.45) while the influ-
ence on the cytotoxicity was not so significant.
Some flavone and flavanon derivatives containing N-
´
benzylpiperazine side chain were prepared by Ferte
et al.49 In agreement with our results, this study con-
firmed that the replacement of the C-7-OH position of
simple flavones by methoxylated N-benzylpiperazine
groups strongly improves their P-gp-inhibitory activity,
while the substitution of C-5-OH leads to the loss of this
ability.
These results indicate that free C-7-OH position is criti-
cal for the binding to ATP-site of the NBD: its blocking
even by a small methyl group in 11 deteriorates this
affinity (lower inhibitory activity and similar cytotoxici-
ty compared to those of 2), while introduction of a spa-
tially bulkier benzyl group in 8, which can probably
reach the hydrophobic binding site (similarly as C-6
and C-8 alkylated derivatives), leads to increasing of
the inhibitory activity of 2. Nevertheless, an improve-
ment of inhibitory activity of 7-O methyl derivatives
was observed in the cases when the second methyl group
was present in the structure (11 vs 18 and 20). Addition-
al increase in the activity was achieved in the case of
3,7,20-tri-O-methyl derivative 22, which was the best
inhibitor from our tested series.
The effect of the E ring substitution was noted in our
study as well. Selective methylation of C-20-OH group
of 2 (yielding 13) was connected with a significant in-
crease of P-gp inhibition and a decrease of cytotoxicity
for normal human lymphocytes. Indeed, almost all
derivatives of 2 with 20-OMe (13, 20, 21, and 22)
showed better P-gp inhibitory activity together with
lower cytotoxicity for normal human lymphocytes
compared to the parent compound 2. Maitrejean
et al.25 demonstrated significance of monolignol unit
for strengthening of binding affinity of flavonoids to
recombinant cytosolic NBD2 (nucleotide-binding do-
main) of P-gp. We proved that substitution of mono-
lignol unit (ring E of 2) by its methylation further
improves this binding affinity but also decreases the
intrinsic cytotoxicity of new derivatives. Additionally,
we showed that the presence of methoxyl group at
the C-19 position of 2 is also essential for P-gp inhib-
itory activity (2 vs 17).
This effect can be well explained by comparing the struc-
ture of 2,3-dehydrosilybin (2) with that of another P-gp
modulator GF 120918 that is at present already under
clinical testing.47,48 Rather high structural resemblance
of this compound with 2 that is further potentiated by
its methylation in positions 3, 7, and 20 well correlates
with the favourable effect of these modifications. The
very best fit gives our best dehydrosilybin derivative
22—Chart 2 (for molecular modeling of the compounds
involved, see also Supplementary material to this paper).
Effect of bulky pivaloyl group at C-23-OH position of 1
on its P-gp inhibitory activity was rather surprising:
compound 7 showed four times higher activity than 1
at approximately two times lower concentration.
Interestingly, benzylation of C-7-OH of 1 (yielding 5)
changed the toxicity of this non-toxic compound to
This improvement of silybin activity can be explained by
an increase of its lipophilicity. The above-mentioned ef-
fect of monolignol unit, together with importance of its
substitution may suggest the presence of a new hydro-
phobic binding region on the NBD. However, steric
proximity of bulky lipophilic pivaloyl group to the E
ring can be another reason for its improved P-gp inhib-
itory activity due to a reinforcement of this hydrophobic
interaction. Interestingly, the compound 7 is cytotoxic
for most of cancer cell lines but not for non-malignant
human lymphocytes.
O
OH
O
O
O
O
O
O
O
OH
22
O
NH
O
N
O
O
O
N
The P-gp inhibitory activity also correlates with lipo-
philicity of the compounds quite well, however, with
some exceptions, for example, 2,3-dehydrosilybin (2)
versus its trimethylderivative 24. Zanden et al. have
recently demonstrated that the inhibition activity on P-
gp homologous MRP-associated proteins depends rath-
.
HCl
GF 120918
Chart 2.

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3803
er on the planarity of their molecules (represented here
3. Experimental
3.1. General methods
by the dihedral angle between the B and C rings) and
the number of hydroxyl and methoxyl groups than on
lipophilicity of the flavonoid derivatives.50 They also
confirmed that the methoxyl group instead of free
hydroxyl group improves the inhibition activity.
The reactions were monitored by TLC on silica gel F254
(Merck) and the spots were visualized by UV light and
by charring with 5% H2SO4 in ethanol.
However, the design of pharmacologically more active
structures also requires reasonable solubility of the com-
pounds and this aspect must be taken into account since
flavonolignans and especially 2,3-dehydrosilybin (2) suf-
fer from low solubility in both lipophilic and hydrophilic
solvents. In those cases, the appropriate substitution of
selected positions with OH groups seems to improve at
least solubility in lipophilic environment.
NMR spectra were recorded on a Varian INOVA-400
spectrometer (399.89 MHz for 1H, 100.55 MHz for
13C) in CDCl3 or DMSO-d6 (see text) at 30 ꢁC. Chemical
shifts were referenced to the residual solvent signal (dH
7.265, dC 77.00; dH 2.50, dC 39.60). Digital resolution
used justified reporting the proton and carbon chemical
shifts to three and two decimal places, respectively. All
2D NMR experiments (HOM2DJ, gCOSY, TOCSY,
HMQC, and HMBC) were performed using standard
manufacturers’ software. The sequence for 1D-TOCSY
experiments was obtained through Varian User Library,
while the sequence gHMBC was obtained from Varian
Application Laboratory in Darmstadt (D).
2.3. Conclusion
We have found a new group of non-toxic inhibitors of
P-gp, based on O-alkylated derivatives of 2,3-dehydro-
silybin (13, 20, 21, and 22). These compounds are very
effective inhibitors at lower doses than the lowest
TCS50 value. The compounds 2, 8, 9, and 18 are effec-
tive inhibitors of P-gp efflux but they exhibit high
intrinsic cytotoxicity to all tested cell lines. On the
other hand, derivatives with inhibitory activity lower
than that of parent 2 were also identified (11, 17,
and 24).
Positive-ion electrospray ionization (ESI) mass spectra
were recorded on a double-focusing instrument Finni-
gan MAT 95 (Finnigan MAT, Bremen, FRG) with BE
geometry. Samples dissolved in methanol:water (2:1, v/v)
were continuously infused through a stainless capillary
held at 3.3 kV into Finnigan ESI source via linear
syringe pump at a flow rate of 40 lL/min.
SAR study of these derivatives disclosed several struc-
tural requirements important for improving P-gp inhib-
itory activity of parent 2,3-dehydrosilybin (2):
3.2. Chemistry
Silybin (natural, mixture of A + B diastereomers—in all
following synthetic sections) (1) was kindly donated by
the IVAX Co., Opava, Czech Republic. 2,3-Dehydrosi-
lybin (2) was prepared as described previously.35 Sily-
binic acid (3) and 2,3-dehydrosilybinic acid (4) were
prepared by CrO3/H5IO6 oxidation of selectively acety-
lated derivatives of silybin/2,3-dehydrosilybin in aceto-
nitrile with following deacetylation of oxidized
products.35
(a) Methylation of C-20-OH position (the most impor-
tant modification).
(b) Methylation of C-3-OH position.
(c) Presence of free C-5-OH position.
(d) introduction of spatially bulkier benzyl group at the
C-7-OH position.
The factors with deteriorative effect on P-gp inhibitory
activity of 2 were also defined:
3.2.1.
7-O-Benzylsilybin
(5).
K2CO3
(210 mg;
(a) Selective methylation of C-7-OH position only (with-
out no other substituted position)
(b) Absence of C-19 methoxyl group.
(c) Methylation of C-5-OH position.
1.520 mmol) and benzyl bromide (50 lL, 0.421 mmol)
were added to the solution of silybin (1, 105 mg,
0.218 mmol) in dry acetone (30 mL). The mixture was
stirred and refluxed for 3.5 h under argon. The reaction
was quenched by addition of concd HCl (1 mL) and
after short stirring the reaction mixture was diluted with
water (50 mL) and extracted with ethyl acetate
(2· 30 mL). The combined organic layers were dried
(Na2SO4) and evaporated; the residue afforded after
flash chromatography (chloroform/acetone/toluene/for-
mic acid 85:10:5:1) the title compound 5 (101 mg,
81%) as a white amorphous solid. MS-ESI (m/z) 573
[M+H]+. For 1H and 13C NMR data see Tables 1 and 2.
In addition, impact of hydroxyl group substitution on a
reduction of cytotoxicity of 2 was evaluated. The meth-
ylation of C-20-OH group was the most influential,
while the effects of further substitutions were less
important.
Compounds with markedly higher cytotoxicity (mainly
for cancer cells) than 2 were found as well (8, 11, 20,
and 24). It seems substitution of C-7-OH (either by
methyl or benzyl group) or C-5-OH leads to increase
in cytotoxicity of 2,3-dehydrosilybin derivatives in can-
cer cell lines.
3.2.2. 7-O-Benzoylsilybin (6). Benzoyl chloride (0.2 mL,
1.721 mmol) and BF3ÆEt2O (0.1 mL, 50% solution in
diethyl ether) were added to a solution of 1 (198 mg,
0.410 mmol) in dichloromethane/acetonitrile (90 mL,
1:1 v/v) at 0 ꢁC and the mixture was stirred for 42 h at
room temperature. The reaction mixture was then dilut-
Interestingly, C-7-OH benzylation of silybin substantial-
ly increases its general cytotoxicity.

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3804
ed with saturated aqueous solution of NaHCO3 and
extracted with ethyl acetate (2· 30 mL). The organic lay-
ers were combined, dried (Na2SO4), and evaporated; the
residue afforded after flash chromatography (ethyl ace-
111.84 (C-18), 115.39 (C-21), 116.24 (C-13), 116.86 (C-
16), 120.61 (C-22), 121.42 (C-15), 123.70 (C-14),
127.28 (C-17), 127.85 (2· C-ortho), 128.13 (C-para),
128.54 (2· C-meta), 136.19 (C-ipso), 136.71 (C-3),
143.48 (C-12a), 145.21 (C-16a), 146.24 (C-2), 147.15
(C-20), 147.11 (C-19), 156.11 (C-8a), 160.42 (C-5),
164.06 (C-7), 176.18 (C-4).
tate/petroleum ether 1:1) the title compound
6
(220 mg, 90%) as a white amorphous solid. MS-ESI
(m/z) 587 [M+H]+. For 1H and 13C NMR data, see
Tables 1 and 2.
3.2.5. 3-O-Methyl-2,3-dehydrosilybin (9). Methyl iodide
(0.034 mL, 0.550 mmol) was added to a cooled (0 ꢁC)
mixture of 2 (240 mg, 0.50 mmol) and NaH (47 mg;
60% w/w in mineral oil, 0.596 mmol) in dry DMF
(2.5 mL) and the reaction mixture was stirred for 1.5 h
at 0 ꢁC and then 12 h at room temperature. The reaction
was quenched by addition of concd HCl (0.2 mL) and
20 mL of water. The yellow precipitate was filtered off,
dried, and dissolved in ethyl acetate. After evaporation,
flash chromatography (chloroform/acetone/toluene/for-
mic acid, 85:10:5:1) gave the title compound (9)
(110 mg, 45%) as a yellow amorphous solid. MS-ESI
3.2.3. 23-O-Pivaloylsilybin (7). Pivaloyl chloride
(0.6 mL, 4.872 mmol) and BF3ÆEt2O (0.6 mL, 50% solu-
tion in diethyl ether) were added to a cooled (0 ꢁC) solu-
tion of 1 (530 mg, 1.042 mmol) in dichloromethane/
acetonitrile (90 mL, 1:1, v/v). After 1 h of stirring at
0 ꢁC, the mixture was warmed up to the room tempera-
ture and stirred for another 12 h. Then it was diluted
with water and extracted with dichloromethane
(2· 50 mL). Organic layers were combined, dried
(Na2SO4), and evaporated. Flash chromatography
(chloroform/acetone/formic acid, 9:1:0.1) afforded 7
(350 mg, 60%) as a white amorphous solid. MS-ESI
(m/z): 567 [M+H]+. For 1H and 13C NMR data,
see Tables 1 and 2.
1
(m/z): 495 [M+H]+. For H and 13C NMR data see Ta-
bles 3 and 4.
3.2.6. 7-O-Methylsilybin (10). K2CO3 (210 mg;
1.520 mmol) and MeI (0.1 mL, 1.606 mmol) were added
to a stirred mixture of 1 (530 mg, 1.099 mmol) in dry
acetone (45 mL) and the reaction mixture was refluxed
for 3.5 h under argon. The reaction was quenched by
the addition of concd HCl (1 mL) and after short stir-
ring the reaction mixture was diluted with water
(150 mL) and extracted with ethyl acetate (2· 50 mL).
The organic layer was dried (Na2SO4) and evaporated.
3.2.4. 7-O-Benzyl-2,3-dehydrosilybin (8). Iodine solution
(250 mg I2 dissolved in 15 mL of glacial acetic acid,
0.984 mmol) was added dropwise within 30 min to a stir-
red solution of 5 (250 mg, 0.437 mmol) and anhydrous
potassium acetate (1.5 g, 15.306 mmol) in glacial acetic
acid (15 mL). The reaction mixture was refluxed under
stirring for 4 h. After cooling to room temperature
and diluting with 150 mL of ice-cold water, the excess
iodine was removed by addition of solid Na2S2O3 (until
the bleaching of the solution). Fine yellow precipitate
was filtered off, washed with water to neutral pH and
after drying it was dissolved in ethyl acetate. The solvent
was evaporated, the evaporated residue was re-dissolved
in CH2Cl2 (10 mL), and 50 mL of ethanolic HCl (42 mL
EtOH + 8 mL concd HCl) was added. The resulting
mixture was refluxed under stirring for 1 h to cleave ace-
tates formed during the oxidation in acetic acid. The
mixture was then diluted in 150 mL of water and
extracted with ethyl acetate (2· 75 mL). The organic
layer was dried (Na2SO4) and evaporated.
The evaporated residue afforded after flash chromatog-
raphy
85:10:5:1) the title compound (10) (290 mg, 53%) as a
(chloroform/acetone/toluene/formic
acid,
white amorphous solid. MS-ESI (m/z): 497 [M + H]+.
1
For H and 13C NMR data see Tables 1 and 2.
3.2.7. 7-O-Methyl-2,3-dehydrosilybin (11). Compound
10 (200 mg, 0.403 mmol) and anhydrous potassium ace-
tate (1.2 g, 12.245 mmol) were dissolved in glacial acetic
acid (12 mL). To a stirred mixture, iodine solution
(200 mg I2 dissolved in 12 mL of glacial acetic acid,
0.787 mmol) was dropwise added within 30 min. The
reaction and workup were performed analogously as
for 7-O-benzyl-2,3-dehydrosilybin (8). The evaporated
residue afforded after flash chromatography (chloro-
form/acetone/toluene/formic acid, 85:5:10:1) the title
compound (11) (180 mg, 90%) as a yellow amorphous
solid. MS-ESI (m/z): 495 [M+H]+. For 1H and 13C
NMR data, see Tables 3 and 4.
The crude product afforded after flash chromatography
(chloroform/acetone/toluene/formic acid, 85:5:10:1) the
title compound (8) (210 mg, 84%) as a yellow amor-
phous solid. MS-ESI (m/z): 571 [M+H]+. 1H NMR
(DMSO-d6): 3.377 (1H, ddd, J = 12.2, 5.9, 4.6 Hz, H-
23u), 3.580 (1H,ddd, J = 12.2, 4.8, 2.5 Hz, H-23d),
3.792 (3H, s, OCH3), 4.279 (1H, ddd, J = 7.8, 4.6,
2.5 Hz, H-10), 4.971 (1H, d, J = 7.8 Hz, H-11), 4.979
(1H, dd, J = 5.9, 4.8 Hz, 23-OH), 5.212 (2H, s, –CH2–),
6.421 (1H, d, J = 2.2 Hz, H-6), 6.819 (1H, d,
J = 8.1 Hz, H-21), 6.890 (1H, d, J = 2.0 Hz, H-8),
6.894 (1H, dd, J = 8.1, 2.0 Hz, H-22), 7.050 (1H, d,
J = 2.0 Hz, H-18), 7.129 (1H, m, H-16), 7.348 (1H, m,
H-para), 7.409 (2H, m, 2· H-meta), 7.465 (2H, m,
2· H-ortho), 9.141 (1H, s, 20-OH), 9.673 (1H, s, 3-
OH), 12.386 (1H, s, 5-OH). 13C NMR (DMSO-d6):
55.79 (OCH3), 60.14 (C-23), 70.01 (CH2), 75.92 (C-11),
78.61 (C-10), 93.02 (C-8), 98.13 (C-6), 104.26 (C-4a),
3.2.8. 20-O-Methylsilybin (12). MeI (0.160 mL,
2.579 mmol) was added to a solution of 1 (1.032 g,
2.139 mmol) and NaH (257 mg; 60% w/w in mineral
oil, 6.417 mmol) in DMF (15 mL) and the mixture was
stirred for 2 h at room temperature. The reaction was
quenched by the addition of concd HCl (1 mL) and then
the mixture was diluted with water (75 mL). The white
precipitate was filtered off, dried, and dissolved in ethyl
acetate. After evaporation, flash chromatography (chlo-
roform/acetone/toluene/formic acid, 85:10:5:1) gave the

Table 3. 1H NMR data (399.89 MHz, DMSO-d6, 30 ꢁC) of 2,3-dehydrosilybin derivatives
Proton
9
11
13
18
20
21
22
24
6
6.192 d (2.1)
6.474 d (2.1)
4.309 ddd
6.336 d (2.2)
6.784 d (2.2)
6.183 d (2.0)
6.444 d (2.0)
6.356 d (2.2)
6.785 d (2.2)
6.345 d (2.2)
6.795 d (2.2)
6.195 d (2.1)
6.445 d (2.1)
6.368 d (2.2)
6.800 d (2.2)
4.350 ddd
6.462 d (2.2)
6.866 d (2.2)
4.293 ddd
8
10
4.276 ddd (7.8, 4.7, 2.5) 4.301 ddd (7.8, 4.5, 2.6) 4.316 ddd (8.0, 4.5, 2.5) 4.308 ddd (7.9, 4.4, 2.4) 4.333 ddd
(7.9, 4.6, 2.5)
5.027 d (7.9)
(7.9, 4.6, 2.4)
4.970 d (7.9)
7.617 d (2.2)
(7.9, 4.6, 2.4)
5.037 d (7.9)
7.699 d (2.2)
(7.9, 4.6, 2.5)
5.025 d (7.9)
5.778 d (2.2)
11
13
15
4.967 d (7.8)
7.816 d (2.2)
5.025 d (7.8)
4.975 d (8.0)
5.029 d (7.9)
7.778 d (2.2)
7.754 dd (8.6, 2.2)
7.688 d (2.2)
7.699 dd (8.5, 2.2)
7.828 d (2.2)
7.819 dd (8.7, 2.2)
7.626 d (2.2)
7.643 dd (8.4, 2.2)
7.639 dd (8.5, 2.2) 7.810 dd (9.2, 2.2)
7.708 dd (9.0, 2.2) 7.790 dd
(9.2, 2.2)
16
18
21
22
7.136 d (8.5)
7.051 d (1.9)
6.824 d (8.2)
7.122 d (9.2)
7.047 d (2.0)
6.818 d (8.1)
7.121 d (8.6)
7.087 d (1.8)
7.146 d (8.5)
7.055 d (2.0)
7.133 d (8.7)
7.089 d (1.7)
7.138 d (8.4)
7.094 d (1.8)
7.159 d (9.0)
7.096 d (1.8)
7.003 d (8.3)
7.036 dd
7.118 d (9.2)
7.088 d (1.8)
6.998 d (8.3)
7.029 dd
6.998 d (8.3)
7.028 dd (8.3, 1.8)
6.824 d (8.1)
6.898 dd (8.1, 2.0)
7.000 d (8.3)
7.030 dd (8.3, 1.7)
7.002 d (8.3)
7.033 dd (8.3, 1.8)
6.896 dd (8.2, 1.9) 6.892 dd (8.1, 2.0)
(8.3, 1.8)
(8.3, 1.8)
3.581 ddd
23
3.580 dd
(12.4, 2.4)
3.374 dd
(12.4, 4.6)
3.800 s
3.577 ddd
3.581 dd (12.4, 2.6)
3.371 dd (12.4, 4.5)
3.586 ddd
(12.2, 5.0, 2.5)
3.378 ddd
(12.2, 5.0, 4.5)
3.825 s
—
3.585 dd
3.591 dd (12.3, 2.5) 3.593 ddd
(12.4, 4.8, 2.4)
3.375 dd (12.3, 4.6) 3.378 ddd
(12.4, 4.8, 4.6)
3.828 s
(12.3, 4.4, 2.5)
3.376 ddd
(12.3, 4.7, 4.4)
(12.5, 5.1, 2.4)
3.375 ddd
(12.5, 5.1, 4.4)
(12.3, 5.1, 2.5)
3.372 ddd
(12.3, 5.8, 4.6)
3-OMe
5-OMe
7-OMe
3.804 s
3.875 s
3.857 s
3.782 s
3.782 s
8.918 s
3.845 s
3.792 s
3.844 s
3.793 s
—
3.850 s
3.783 s
3.783 s
9.662 s
12.381 s
3.850 s
3.786 s
3.782 s
19-OMe 3.792 s
20-OMe
3-OH
3.783 s
3.783 s
3.785 s
3.785 s
9.650 s
12.379 s
—
12.571 s
—
5-OH
7-OH
12.604 s
9.211 s
8.306 s
—
12.396 s
12.605 s
12.572 s
20-OH
23-OH
9.148 s
4.989 t (4.4)
9.170 s
4.993 t (5.0)
5.008 t (4.8)
4.987 dd (5.8, 5.1)

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3806
Table 4. 13C NMR data (100.55 MHz, DMSO-d6, 30 ꢁC) of 2,3-dehydrosilybin derivatives
Carbon
9
11
13
18
20
21
22
24
2
154.63
138.21
177.93
104.20
161.26
98.79
146.19
136.69
176.19
104.13
160.38
97.55
145.64
136.38
175.99
102.96
160.70
98.41
154.92
138.45
178.14
105.31
160.92
97.87
146.16
136.72
176.21
104.13
160.37
97.54
154.63
138.21
177.94
104.28
161.26
98.71
154.92
138.46
178.14
105.31
160.91
97.86
141.29
138.06
171.09
106.31
160.14
95.73
3
4
4a
5
6
7
164.72
93.91
165.08
92.14
164.58
93.64
165.27
92.46
165.08
92.15
164.46
93.86
165.27
92.48
163.79
92.92
8
8a
156.49
78.61
75.97
156.20
78.61
75.92
156.30
78.46
75.75
156.37
78.63
75.96
156.19
78.50
75.76
156.45
78.51
75.80
156.37
78.52
75.79
158.10
78.41
75.76
10
11
12a
13
143.58
116.74
122.77
121.96
117.07
146.01
127.17
111.96
147.74
147.23
115.43
120.72
60.13
143.48
116.25
123.73
121.42
116.85
145.19
127.29
111.84
147.72
147.15
115.40
120.62
60.15
143.31
116.22
123.88
121.28
116.87
144.98
128.84
111.32
148.88
149.24
111.71
120.36
60.09
146.15
116.78
122.63
122.06
117.07
143.63
127.14
111.94
147.71
147.22
115.41
120.69
60.10
143.39
116.24
123.77
121.46
116.88
145.15
128.82
111.31
148.87
149.26
111.69
120.36
60.10
143.48
116.74
122.78
122.00
117.06
145.97
128.71
111.44
148.89
149.32
111.75
120.44
60.07
143.53
116.79
122.66
122.10
117.09
146.11
128.68
111.41
148.86
149.30
111.72
120.44
60.05
144.29
115.56
124.12
120.76
116.83
143.43
128.89
111.30
148.56
149.24
111.68
120.35
60.12
14
15
16
16a
17
18
19
20
21
22
23
3-OMe
5-OMe
7-OMe
19-OMe
20-OMe
59.81
59.78
59.78
59.78
—
56.22
55.98
55.66
55.591
56.05
55.80
56.10
55.80
56.05
55.66
55.59
56.10
55.67
55.59
55.81
55.65
55.59
55.61
55.67
—
title compound (12) (560 mg, 54%) as a white amor-
9.329 (2H, br s, 30-OH + 40-OH); 13C NMR (DMSO-
d6, 30 ꢁC): 51.17 (OMe), 113.74 (C-2), 114.81 (C-20),
115.76 (C-50), 121.37 (C-60), 125.52 (C-10), 145.13 (C-
3), 145.59 (C-30), 148.43 (C-1).
1
phous solid. MS-ESI (m/z): 497 [M+H]+. For H and
13C NMR data, see Tables 1 and 2.
3.2.9. 20-O-Methyl-2,3-dehydrosilybin (13). Iodine solu-
tion (200 mg I2 dissolved in 12 mL of glacial acetic acid,
0.787 mmol) was dropwise added within 30 min to a
mixture of the compound 12 (200 mg, 0.403 mmol)
and anhydrous potassium acetate (1.2 g, 12.245 mmol)
in glacial acetic acid (12 mL). The reaction and workup
were performed analogously as for 8. The crude product
afforded after flash chromatography (chloroform/ace-
tone/toluene/formic acid, 85:5:10:1) the title compound
13 (176 mg, 88%) as a yellow amorphous solid. MS-
3.2.11. (E)-3-O-Benzyl-3,4-dihydroxycinnamyl alcohol
(15). NaH (206 mg, 5.167 mmol, 60% w/w in mineral
oil) was dissolved at 0 ꢁC in DMSO (8 mL) and the solu-
tion was stirred for 5 min under argon. Methyl ester 14
(500 mg, 2.575 mmol) was added and the mixture was
stirred additional 5 min followed by benzyl chloride
(0.357 mL, 3.088 mmol) addition and stirring for 1 h at
room temperature. The reaction was quenched by addi-
tion of concd HCl (1 mL) and 50 mL of water. Aqueous
solution was extracted with CH2Cl2 (2· 50 mL), organic
layers were combined, dried (Na2SO4) and evaporated;
the residue was re-dissolved in dry THF (30 mL), the
solution was cooled to 0 ꢁC and LiAlH4 (500 mg,
13.193 mmol) was added portionwise. Reaction mixture
was stirred 1 h at 0 ꢁC and then 3 h at room tempera-
ture. The excess of LiAlH4 was destroyed under cooling
by addition of water (3 mL). Celite (2 g) was added to
the resulting solution and after short stirring the insolu-
ble impurities were filtered off. The filter cake was thor-
oughly washed with ethyl acetate and then with acetone.
All filtrates were combined and evaporated to a yellow-
ish syrup. Flash chromatography (chloroform/acetone/
toluene/formic acid, 85:10:5:1) gave 15 (280 mg, 42%)
as a yellowish oil. MS-ESI (m/z): 257 [M + H]+. 1H
NMR (DMSO-d6, 30 ꢁC): 4.281 (2H, dd, J = 6.0,
1.5 Hz, 2 · H-1), 5.114 (2H, s, –CH2O–), 5.795 (1H, br
1
ESI (m/z): 495 [M+H]+. For H and 13C NMR data,
see Tables 3 and 4.
3.2.10. Caffeic acid methyl ester (14). Caffeic acid (2.5 g,
13.873 mmol) was dissolved in dry methanol (150 mL)
and H2SO4 (2.5 mL, 96%, v/v) was added. Reaction mix-
ture was stirred at room temperature for 12 h. Reaction
mixture was then evaporated to approximately 50 mL,
treated by ice-cold saturated solution of NaHCO3
(150 mL) and after short stirring it was extracted with
ethyl acetate (3· 75 mL). Organic layers were combined,
dried (Na2SO4) and evaporated to afford pure 14
1
(2.65 g, 98%). MS-ESI (m/z): 195 [M+H]+. H NMR
(DMSO-d6, 30 ꢁC): 3.686 (3H, s, OCH3), 6.264 (1H, d,
J = 15.9 Hz, H-2), 6.760 (1H, d, J = 8.2 Hz, H-50),
6.995 (1H, dd, J = 8.2, 2.1 Hz, H-60), 7.049 (1H, d,
J = 2.1 Hz, H-20), 7.481 (1H, d, J = 15.9 Hz, H-3),

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3807
s, 40-OH), 6.186 (1H, dt, J = 15.8, 6.0 Hz, H-2), 6.514
(1H, dt, J = 15.8, 1.5 Hz, H-3), 6.892 (1H, d,
J = 8.2 Hz, H-50), 6.925 (1H, dd, J = 8.2, 1.7 Hz, H-
60), 6.999 (1H, d, J = 1.7 Hz, H-20), 7.373 (1 H, m,
para-Ph), 7.417 (2H, m, 2· meta-Ph), 7.429 (2H, m,
2· ortho-Ph); 13C NMR (DMSO-d6, 30 ꢁC): 63.77 (C-
1), 71.21 (–CH2O–), 110.06 (C-20), 114.76 (C-50),
120.63 (C-60), 126.24 (C-2), 127.83 (2· C-ortho), 128.45
(C-para), 128.75 (2· C-meta), 129.26 (C-10), 131.26 (C-
3), 136.23 (C-ipso), 145.88 (C-40), 145.92 (C-30).
7.079 (1H, d, J = 8.6 Hz, H-16), 7.299 (3H, m, H-18,
H-21, H-22), 7.318 (1H, d, J = 2.2 Hz, H-8), 7.459
(1H, dd, J = 8.6, 2.2 Hz, H-15), 7.511 (1H, d,
J = 2.2 Hz, H-13). 13C NMR (CDCl3): 20.54 (23-Ac),
20.59 (9-Ac, 10-Ac), 20.65 (7-Ac), 21.00 (3-Ac), 21.11
(5-Ac), 62.25 (C-23), 75.81 (C-10, C-11), 108.87 (C-8),
113.64 (C-6), 114.73 (C-4a), 117.40 (C-13), 117.56 (C-
16), 122.37 (C-15), 122.63 (C-18), 122.99 (C-14),
124.02 (C-21), 125.38 (C-22), 123.52 (C-3), 133.95 (C-
17), 142.59 (C-20), 142.91 (C-19), 143.33 (C-12a),
145.64 (C-16a), 150.37 (C-5), 154.11 (C-7), 154.66 (C-
2), 156.84 (C-8a), 167.75 (3-CO), 167.79 (7-CO),
167.83 (9-CO), 167.86 (10-CO), 169.21 (5-CO), 170.09
(C-4), 170.25 (23-CO).
3.2.12. 19-O-Demethyl-19-O-benzyl-2,3-dehydrosilybin
(16). The alcohol 15 (420 mg, 1.641 mmol) and querce-
tinÆ2H2O (555 mg, 1.641 mmol) were dissolved in the
mixture of benzene (100 mL) and acetone (50 mL).
Resulting mixture was stirred for 10 min at 60 ꢁC and
then Ag2CO3 (451 mg, 1.640 mmol) was added. Reac-
tion mixture was stirred under reflux for 30 h, then left
to cool to a room temperature and filtered through the
Celite pad. The filter cake was thoroughly washed with
acetone. After evaporation of combined filtrates, flash
3.2.14. 3,7-Di-O-methyl-2,3-dehydrosilybin (18). K2CO3
(345 mg, 2.500 mmol) and dimethyl sulfate (0.178 mL,
1.870 mmol) were added to a solution of 2 (300 mg,
0.624 mmol) in dry DMF (5 mL) and the reaction mix-
ture was stirred for 12 h at room temperature under ar-
gon. The reaction mixture was poured into 75 mL of
solution of concd HCl in water (1/14, v/v). The yellow
precipitate was filtered off, washed with water, dried,
and dissolved in ethyl acetate. After flash chromatogra-
chromatography
(chloroform/acetone/toluene/formic
acid, 85:10:5:1) gave 16 (90 mg, 10%) as a yellow amor-
phous solid. MS-ESI (m/z): 557 [M+H]+. 1H NMR
(DMSO-d6): 3.333 (1H, m, H-23), 3.548 (1H, dm,
J = 12.4 Hz, H-23d), 4.238 (1H, ddd, J = 7.8, 4.6,
2.6 Hz, H-10), 4.948 (1H, d, J = 7.8 Hz, H-11), 4.989
(1H, t, J = 4.6 Hz, 23-OH), 5.115 (2H, s, –CH2–),
6.196 (1H, d, J = 2.1 Hz, H-6), 6.459 (1H, d,
J = 2.1 Hz, H-8), 6.851 (1H, d, J = 8.3 Hz, H-21),
6.906 (1H, dd, J = 8.3, 1.9 Hz, H-22), 7.116 (1H, m,
H-16), 7.145 (1H, d, J = 1.9 Hz, H-18), 7.317 (1H, m,
H-para), 7.383 (2H, m, 2· H-meta), 7.493 (2H, m,
2· H-ortho), 7.762 (2H, m, H-15, H-13), 9.222 (1H, s,
20-OH), 9.554 (1H, s, 3-OH), 10.811 (1H, s, 7-OH),
12.410 (1H, s, 5-OH). 13C NMR (DMSO-d6): 60.08
(C-23), 70.18 (CH2), 75.78 (C-11), 78.53 (C-10), 93.59
(C-8), 98.31 (C-6), 103.13 (C-4a), 113.99 (C-18), 115.75
(C-21), 116.22 (C-23), 116.86 (C-16), 121.09 (C-22),
121.27 (C-15), 123.82 (C-14), 127.23 (C-17), 127.75 (C-
para), 127.89 (2· C-ortho), 128.29 (2· C-meta), 136.37
(C-3), 137.23 (C-ipso), 143.38 (C-16a), 145.03 (C-12a),
145.79 (C-2), 146.59 (C-19), 147.54 (C-20), 156.28 (C-
8a), 160.74 (C-5), 164.12 (C-7), 176.05 (C-4).
phy
(chloroform/acetone/toluene/formic
acid,
85:5:10:1), the title compound 18 was obtained as a yel-
low amorphous solid (100 mg, 32%). MS-ESI (m/z): 509
[M+H]+. For 1H and 13C NMR data, see Tables 3 and 4.
3.2.15. 7,20-Di-O-methylsilybin (19). MeI (0.129 mL,
2.071 mmol) was added to a solution of 1 (500 mg,
1.036 mmol) and NaH (124 mg; 60% w/w in mineral
oil, 3.100 mmol) in DMF (7 mL), the mixture was stir-
red under argon for 1 h at 0 ꢁC and then 1 h at room
temperature. The reaction was quenched by addition
of HCl (2 mL) and water (50 mL). The white precipitate
was filtered off, dried and dissolved in ethyl acetate.
After evaporation, flash chromatography (chloroform/
acetone/toluene/formic acid, 85:10:5:1) gave 19
(110 mg, 21%) as a white amorphous solid. MS-ESI
1
(m/z): 511 [M+H]+. For H and 13C NMR data, see
Tables 1 and 2.
3.2.16. 7,20-Di-O-methyl-2,3-dehydrosilybin (20). Iodine
solution (200 mg I2 dissolved in 12 mL of glacial acetic
acid, 0.787 mmol) was added dropwise within 30 min
to a solution of the compound 19 (200 mg, 0.403 mmol)
and anhydrous potassium acetate (1.2 g, 12.245 mmol)
in glacial acetic acid (12 mL). The reaction and workup
were performed analogously as for 8. The crude product
afforded after flash chromatography (chloroform/ace-
tone/toluene/formic acid, 85:5:10:1) the title compound
20 (162 mg, 81%) as a yellow amorphous solid. MS-
3.2.13. 19-O-Demethyl-2,3-dehydrosilybin (17). The com-
pound 16 (80 mg, 0.144 mmol) was dissolved in THF
(10 mL), Pd/C (40 mg, 10% w/w) was added and the
reaction mixture was stirred for 12 h at room tempera-
ture under hydrogen. The mixture was filtered through
Celite pad, which was washed by acetone (2· 10 mL).
Evaporation of the combined filtrates gave pure 17
(64 mg, 96%) as a yellow amorphous solid.
1
ESI (m/z): 509 [M+H]+. For H and 13C NMR data,
The structural proof was performed on its hexaacetate
(obtained after standard Ac2O/Py peracetylation proce-
see Tables 3 and 4.
1
dure). MS-ESI (m/z): 467 [M+H]+. H NMR (CDCl3):
3.2.17. 3,20-Di-O-methyl-2,3-dehydrosilybin (21) and
3,7,20-tri-O-methyl-2,3-dehydrosilybin (22). Powdered
KOH (360 mg, 6.429 mmol) was stirred for 5 min in
DMSO (4 mL). 2,3-Dehydrosilybin (2, 300 mg,
0.624 mmol) in 4 mL of DMSO was added and the mix-
ture was stirred for 10 min. MeI (0.200 mL, 3.210 mmol)
was added and the reaction mixture was stirred for 1 h
2.063 (3H, s, 23-Ac), 2.308 (3H, s, 20-Ac), 2.313 (3H,
s, 19-Ac), 2.337 (3H, s, 7-Ac), 2.347 (3H, s, 3-Ac),
2.430 (3H, s, 5-Ac), 4.030 (1H, dd, J = 12.4, 4.1 Hz,
H-23u), 4.300 (1H, ddd, J = 7.9, 4.1, 3.6 Hz, H-10),
4.442 (1H, dd, J = 12.4, 3.6 Hz, H-23d), 5.016 (1H, d,
J = 7.9 Hz, H-11), 6.857 (1H, d, J = 2.2 Hz, H-6),

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3808
at room temperature. The reaction was quenched by
addition of concd HCl (1 mL) and diluted in water
(50 mL). The yellow precipitate was filtered off, dried,
and dissolved in ethyl acetate. After evaporation, flash
3.3.2. Cytotoxic MTT assay. Cell suspensions were pre-
pared and diluted according to the particular cell type
and the expected target cell density (2500–30,000 cells/
well based on cell growth characteristic). Cells were add-
ed by pipette (80 lL) into 96-well microtiter plates. Inoc-
ulates were allowed a preincubation period of 24 h at
37 ꢁC and 5% CO2 for stabilization. Fourfold dilutions,
in 20 lL aliquots, of tested compounds were added to
the microtiter plate wells. All test compound concentra-
tions (250–0.2441 lM) were examined in duplicate.
Incubation of the cells with the test compounds lasted
for 72 h at 37 ꢁC, in a 5% CO2 atmosphere at 100%
humidity. At the end of the incubation period, the cells
were assayed using MTT. Aliquots (10 lL) of the MTT
stock solution were pipetted into each well and incubat-
ed for a further 1–4 h. After this incubation period, for-
mazan produced was dissolved by the addition of
100 lL/well of 10% aqueous SDS (pH 5.5), followed
by a further incubation at 37 ꢁC overnight. The absor-
bance (A) was measured at 540 nm with a Labsystem
iEMS Reader MF. Tumor cell survival (TCS) was calcu-
chromatography
(chloroform/acetone/toluene/formic
acid, 85:10:5:1) gave 21 (80 mg, 25%) and 22 (80 mg,
24%) as yellow amorphous solids. MS-ESI (m/z): 509
[M+H]+ and 523 [M+H]+, respectively. For 1H and
13C NMR data see Tables 3 and 4.
3.2.18. 5,7,20-Tri-O-methylsilybin (23). K2CO3 (960 mg;
6.638 mmol) and dimethyl sulfate (0.474 mL,
4.979 mmol) were added to a solution of 1 (400 mg,
0.829 mmol) in dry acetone (30 mL) and the mixture
was stirred and refluxed for 2 h under argon. The reac-
tion was quenched by addition of concd HCl (2 mL)
and water (100 mL). The mixture was extracted with
ethyl acetate (2· 50 mL). The organic layer was dried
(Na2SO4), and evaporated to afford after flash chroma-
tography (chloroform/acetone/toluene/formic acid,
85:10:5:1) the title compound 23 as a white amorphous
solid (245 mg, 56%). MS-ESI (m/z): 525 [M+H]+. For
1H and 13C NMR data see Tables 1 and 2.
lated using the following equation: TCS = (Adrug-exposed
/
mean Acontrol) · 100%. The TCS50 value, the drug con-
centration lethal to 50% of the tumor cells, was calculat-
ed from appropriate dose–response curves.
3.2.19. 5,7,20-Tri-O-methyl-2,3-dehydrosilybin (24). Io-
dine solution (120 mg I2 dissolved in 15 mL of glacial
acetic acid, 0.472 mmol) was dropwise added within
30 min to a solution of 23 (120 mg, 0.229 mmol) and
anhydrous potassium acetate (700 mg, 7.143 mmol) in
glacial acetic acid (15 mL). The reaction and workup
were performed analogously as for 8. The crude product
gave after flash chromatography (chloroform/acetone/
formic acid, 90:10:1) the title compound 24 (55 mg,
46%) as a yellow amorphous solid. MS-ESI (m/z): 523
[M+H]+. For 1H and 13C NMR data see Tables 3 and 4.
3.3.3. Modulation of P-gp efflux function. Paclitaxel resis-
tant derivative of human myeloid leukemia cell line
K562, K562-Tax cells previously shown to overexpress
exclusively functional P-gp and not another MDR pro-
teins were used to investigate P-gp modulatory activity
of silybin/2,3-dehydrosilybin derivatives.43 Cells were
cultured in RPMI 1640 medium supplemented with
sodium bicarbonate and FCS 10% only. Following an
overnight incubation, 106 of the cells were seeded in
six-well plates with or without the reference P-gp inhib-
itor (cyclosporin A, CyA) or tested derivatives for 1 h
and labeled with P-gp fluorescent substrate daunorubi-
cin (DNR, 1 lM) for another 1 h. Both the CyA and
tested compounds were tested in wide concentration
range (100–0.0977 lM) in fourfold dilutions and six
concentrations. Following this period, the cells were
washed in PBS and re-suspended in the complete medi-
um with an identical concentration of cyclosporin A or
tested drugs but no DNR. Treated cellular suspensions
were allowed to culture for 1 h at 37 ꢁC in order to
pump out the DNR by functional P-gp. Finally, K562-
Tax myeloblasts were put on ice until the red fluores-
cence data of DNR accumulation were collected
through a 585/42 bandpass filter, using a linear amplifi-
cation (FL-2 height) on the flow cytometer. The ratio of
the mean DNR fluorescence of treated versus untreated
cells was calculated in order to evaluate the capacity of
reference inhibitor cyclosporin A or tested substances to
block the functional activity of P-gp.
3.3. Biological studies
3.3.1. Cell cultures. Cell lines CCRF-CEM (CEM, hu-
man T-lymphoblastic leukemia), K562 (human myeloid
leukemia), L1210 (mouse lymphocytic leukemia),
L1210-VCR (vincristine resistant mouse lymphocytic
leukemia), A549 (human lung adenocarcinoma), and
MCF7 (human breast cancer) were purchased from
the American Tissue Culture Collection (ATTC,
Manassas, VA, USA). The drug resistant sublines of
CEM and K562 cells were selected in our laboratory
by the cultivation of maternal cell lines in increasing
concentrations of daunorubicin (Cem-DNR and clone
Cem-DNR 0.3A2), vincristine (Cem-VCR) or paclit-
axel (K562-Tax), respectively.43 The cells were main-
tained in Nunc/Corning 80 cm2 plastic tissue culture
flasks and cultured in cell culture medium (RPMI
1640 with 5 g/L glucose, 2 mM glutamine, 100 U/mL
penicillin, 100 mg/mL streptomycin, 10% fetal calf ser-
um, and NaHCO3).
3.3.4. Statistical analysis. With exception of kinetic
parameters, all values are expressed as means and stan-
dard deviations (SD). Kinetic and statistical analyses
were calculated using PRISM 4 (GraphPad Software
Inc., San Diego, CA). Kinetic values, for example,
half-maximal effective concentrations (EC50) and maxi-
mum efflux inhibition (Imax), were estimated from the
Phenotype characteristics of the drug resistant cell lines
were previously published43 and are as follows: CEM-
DNR (P-gp+, MRPꢁ, and LRP+), CEM-DNR 0.3A2
(P-gp+, MRPꢁ, and LRPꢁ), CEM-VCR (P-gp+, MRPꢁ,
and LRP+), and K562-Tax (P-gp+, MRPꢁ, and LRPꢁ)
(Table 5).

Table 5. Cytotoxic activity of silybin/2,3-dehydrosilybin derivatives on a panel of drug sensitive versus drug resistant cell linesa compared to normal human lymphocytes
Compound
Cytotoxicity (TCS50b,c, lM)
CEM
Cem-VCR
>250 [0]
17.28 [0.445] 28.89 [1.164]
Cem-DNR
Cem-DNR 0.3A2 K 562
K 562-Tax
L1210
L1210-VCR
MCF7
A 549
Normal human
lymphocytes
1
2
155 [3.871]
>250 [0]
6.18 [1.863]
>250 [0]
>250 [0]
>250 [0]
136.5 [12.23] >250 [0] >250 [0]
10.99 [2.711] 34.73 [13.372] >250 [0]
>250 [0]
>250 [0]
214.83 [16.555] 156.5 [18.75]
16.49 [1.309]
>250 [0]
>250 [0]
31.09 [18.246]
>250 [0]
>250 [0]
20.1 [2.047]
>250 [0]
>250 [0]
8.46 [1.072]
>250 [0]
>250 [0]
22.45 [1.8]
>250 [0]
>250 [0]
3
>250 [0]
>250 [0]
>250 [0]
>250 [0]
>250 [0]
>250 [0]
8.203 [0.056] 16.65 [0.76]
>250 [0]
>250 [0]
>250 [0]
>250 [0]
>250 [0]
4
5
7.83 [0.929] 16.09 [2.751] 26.2 [4.874]
26.1 [2.05] 57.23 [8.504] 88.74 [7.405]
22.55 [1.849] 86.27 [10.591] 68.25 [4.876] 78.23 [7.069]
13.01 [3.211]
41.5 [1.545]
24.2 [6.437] 18.02 [1.139]
160.31 [2.424] 42.51 [10.07]
105.33 [24.372] 93.86 [3.953]
13.41 [1.711]
71.22 [13.605]
34.79 [2.612]
4.24 [1.4]
16.37 [0.7656]
158.6 [7.838]
>250 [0]
6
34.92 [2.159] 52.77 [25.466] >250 [0]
17.07 [0.841] 28.07 [3.089]
7
>250 [0]
4.82 [2.64]
8
2.83 [0.05]
12.19 [1.89]
8.34 [0.47]
9.29 [1.43]
11.81 [0.96]
10.29 [0.78]
3.24 [0.43]
10.73 [1.07]
130.57 [5.2]
3.06 [0.32]
11.79 [1.2]
24.40 [9.33]
10.34 [1.47]
11.54 [1.45]
9.22 [0.39]
11.77 [0.5]
9.09 [0, 52]
10.05 [0.39]
7.93 [1.74]
11.55 [0.82]
5.57 [1.73]
10.38 [0.29]
42.24 [1.2]
9.03 [2.08]
3.02 [0.6]
3.3 [1.25]
4.93 [0.83]
10.52 [1.09]
3.38 [0.26]
9.64 [0.57]
35.91 [1.5]
7.89 [0.57]
2.81 [0.07]
3.47 [0.73]
2.56 [0.1]
9.64 [0.89]
14.68 [0.82]
10.21 [0.55]
12.69 [1.44]
48.85 [4.29]
11.59 [0.58]
8.63 [2.72]
10.05 [0.54]
12.50 [2.12]
49.77 [2.88]
16.66 [7.03]
204.99 [53.08]
>250 [0]
9
13.23 [1.66]
8.25 [1.44]
9.85 [0.54]
12.88 [0.95]
8.96 [0.36]
3.6 [0.4]
29.68 [4.71] 27.14 [14.95]
8.46 [1.1] 8.07 [4.93]
11.41 [1.61] 10.46 [0.56]
11.42 [1.04]
8.75 [2]
11
13
17
18
20
21
22
24
9.56 [0.71]
11.67 [1.45]
10.08 [2.93]
6.49 [4.63]
9.02 [2.88]
201.56 [55.84] 42.83 [2.45]
43.73 [2.5]
9.17 [0.64]
4.42 [0.89]
9.48 [1.79]
58.7 [2.82]
3.03 [0.39]
32.62 [13.64]
8.87 [0.69]
2.44 [0.13]
6.62 [3.7]
11.45 [0.94]
4.84 [2.17]
8.92 [2.84]
10.04 [0.55]
27.37 [3.5]
8.45 [1.04]
68.31 [34.56]
>250 [0]
7.54 [2.54]
19.18 [11.67]
3.45 [0.43]
10.61 [1.41]
242.95 [14.1]
>250 [0]
187.42 [29.65]
203.75 [32.37] 19.95 [4.51]
4.58 [2.67] 2.44 [0.12]
116.84 [7.89] 75.14 [16.02]
4.48 [0.66] 2.56 [0.19]
27.58 [7.14]
6.54 [3.86]
109.27 [15.77] 182.17 [76.31]
11.26 [2.95] 3.3 [0.24]
a CEM—human T-lymphoblastic leukemia cell line CCRF-CEM; CEM-VCR vincristine resistant CEM cells; CEM-DNR—daunorubicin resistant CEM cells; CEM-DNR 0.3A2—daunorubicin resistant
subclone of CEM-DNR cell line; K562—human myeloid leukemia cell line; K562-Tax—paclitaxel resistant K562 cells; L1210—mouse lymphocytic leukemia cell line; L1210-VCR—vincristine resistant
L1210 cells; A549—human lung adenocarcinoma cell line; MCF7—human breast cancer cell line.
b Data are expressed as mean values from six measurements in three independent experiments and standard deviations in brackets [SD].
c Maximum/minimum tested concentrations were 250/0.2441 lM, respectively. The TCS50 values indicate the drug concentration lethal to 50% of the cells.

´
P. Dzˇubak et al. / Bioorg. Med. Chem. 14 (2006) 3793–3810
3810
22. Giacomelli, S.; Gallo, D.; Apollonio, P.; Ferlini, C.;
Distefano, M.; Morazzoni, P.; Riva, A.; Bombardelli, E.;
Mancuso, S.; Scambia, G. Life Sci. 2002, 70, 1147.
23. Soose, M. Eur. J. Protistol. 1994, 30, 394.
24. Bokemeyer, C.; Fels, L. M.; Dunn, T.; Voigt, W.;
Gaedeke, J.; Schmol, H.-J.; Stolte, H.; Lentzen, H. Br. J.
Cancer 1996, 74, 2036.
non-linear regression equation with the method of least
squares and the EC50 value is the concentration at which
half-maximal efflux was achieved.
Acknowledgments
25. Maitrejean, M.; Comte, G.; Barron, D.; El Kirat, K.;
Gwenaelle, C.; Di Pietro, A. Bioorg. Med. Chem. Lett.
2000, 10, 157.
26. Seelig, A.; Gatlik-Landwojtowicz, E. Mini-Rev. Med.
Chem. 2005, 5, 135.
27. Kickhoefer, V. A.; Rajavel, K. S.; Scheffer, G. L.; Dalton,
W. S.; Scheper, A. J.; Rome, L. H. J. Biol. Chem. 1998,
273, 8971.
28. Kong, L. B.; Siva, A. C.; Rome, L. A.; Stewart, P. H.
Structure 1999, 7, 371.
This work was supported by the Grant Agency of the
Czech Republic (Grant No. 303/02/1097), the Czech
Ministry of Education (Research concepts No. MSM
6198959216, MSM 0021620835, and AV0Z50200510)
and KONTAKT project No. 7-2006-16. Authors are
´
´
indebted to Drs. K. Hofbauerova and V. Kopecky Jr.,
Institute of Physics, Charles University, Prague, for cal-
culation of molecular models of some compounds in-
volved in this study.
29. Zhang, S. H.; Morris, M. E. J. Pharmacol. Exp. Ther.
2003, 304, 1258.
Supplementary data
30. Zhang, S. Z.; Morris, M. E. Pharm. Res. 2003, 20, 1184.
31. Tyagi, A. K.; Singh, R. P.; Agarwal, C.; Chan, D. C.;
Agarwal, R. Clin. Cancer Res. 2002, 8, 3512.
32. Rodriguez, R. J.; Miranda, C. L.; Stevens, J. F.; Deinzer,
M. L.; Buhler, D. R. Food Chem. Toxicol. 2001, 39, 437.
33. Trompier, D.; Baubichon-Cortay, H.; Chang, X. B.;
Maitrejean, M.; Barron, D.; Riordon, J. R.; Di Pietro,
A. Cell. Mol. Life Sci. 2003, 60, 2164.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2006.01.035.
References and notes
34. Leslie, E. M.; Mao, Q.; Oleschuk, C. J.; Deeley, R. G.;
Cole, S. P. C. Mol. Pharmacol. 2001, 59, 171.
1. Flora, K.; Hahn, M.; Rosen, H.; Benner, K. Am. J.
Gastroenterol. 1998, 93, 139.
2. de Groot, H.; Rauen, M. Fundam. Clin. Pharmacol. 1998,
12, 249.
´
35. Gazˇak, R.; Svobodova, A.; Psotova, J.; Sedmera, P.;
Prikrylova, V.; Walterova, D.; Kren, V. Bioorg. Med.
´
´
ˇ
´
´
ˇ
Chem. 2004, 12, 5677.
´
36. Antus, S.; Baitz-Gacs, E.; Bauer, R.; Gottsegen, A.;
´
ˇ
´
3. Kren, V.; Walterova, D. Biomed Papers 2005, 149, 29.
4. Dorai, T.; Aggarwal, B. B. Cancer Lett. 2004, 215, 129.
5. Ahmad, N.; Gali, H.; Javed, S.; Agarwal, R. Biochem.
Biophys. Res. Commun. 1998, 248, 294.
6. Zi, X.; Mukhtar, H.; Agarwal, R. Biochem. Biophys. Res.
Commun. 1997, 239, 334.
Seligmann, O.; Wagner, H. Liebigs Ann. Chem. 1989,
1147.
37. Guz, N. R.; Stermitz, F. R. J. Nat. Prod. 2000, 63, 1140.
38. Pietta, P. G. J. Nat. Prod. 2000, 63, 1035.
39. Merlini, L.; Zanarotti, A. Tetrahedron Lett. 1975, 16,
3621.
40. Pelter, A.; Ha¨nsel, R. Chem. Ber. 1975, 108, 790.
41. Jiang, X.; Macey, M. G.; Collins, P.-W.; Newland, A. C.
Br. J. Hematol. 1994, 86, 547.
7. Zi, X.; Feyes, D. K.; Agarwal, R. Clin. Cancer Res. 1998,
4, 1055.
8. Zi, X.; Agarwal, R. Biochem. Biophys. Res. Commun.
1999, 263, 528.
42. Lee, J. S.; Scala, S.; Matsumoto, Y.; Dickstein, B.; Robey,
R.; Zhan, Z.; Altenberg, G.; Bates, S. E. J. Cell Biochem.
1997, 65, 513.
9. Zi, X.; Zhang, J.; Agarwal, R.; Pollak, M. Cancer Res.
2000, 60, 5617.
10. Singh, R. P.; Tyangi, A. K.; Zhao, J.; Agarwal, R.
Carcinogenesis 2002, 23, 499.
11. Singh, R. P.; Agarwal, R. Antioxid. Redox Signal. 2002, 4,
655.
12. Kang, S. N.; Lee, M. H.; Kim, K.-M.; Cho, D.; Kim, T. S.
Biochem. Pharmacol. 2001, 61, 1487.
´
ˇ
´
43. Noskova, V.; Dzubak, P.; Kuzmina, G.; Ludkova, A.;
Stehlık, D.; Trojanec, R.; Janost’akova, A.; Korınkova,
´
´
ˇ ´ ˇ´
´
´
´
´
G.; Mihal, V.; Hajduch, M. Neoplasma 2002, 49, 418.
´
´
44. Perez-Victoria, J. M.; Perez-Victoria, J. F.; Conseil, G.;
Maitrejean, M.; Comte, G.; Barron, D.; Di Pietro, A.;
Castanys, S.; Gamarro, F. Antimicrob. Agents Chemother.
2001, 45, 439.
13. Jiang, C.; Agarwal, R.; Lu, J. X. Biochem. Biophys. Res.
Commun. 2000, 276, 371.
45. Conseil, G.; Baubichon-Cortay, H.; Dayan, G.; Jault, J.
M.; Barron, D.; Di Pietro, A. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 9831.
46. Middleton, E., Jr.; Kandaswami, C.; Theoharides, T. C.
Pharmacol. Rev. 2000, 52, 673.
14. Zi, X.; Agarwal, R. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
7490.
15. Bhatia, N.; Agarwal, R. The Prostate 2001, 46, 98.
16. Zhu, W.; Zhang, J.-S.; Young, Y. F. Carcinogenesis 2001,
22, 1399.
47. Hyafil, F.; Vergely, C.; Du Vignaud, P.; Grand-Perret, T.
Cancer Res. 1993, 53, 4595.
17. Thelen, P.; Jarry, H.; Ringert, R.-H.; Wuttke, W. Planta
Med. 2004, 70, 397.
48. Planting, A. S. T.; Sonneveld, P.; Gaast, A.; Sparreboom,
A.; Burg, M. E. L.; Luyten, G. P. M.; Leeuw, K.; Boer-
Dennert, M.; Wissel, P. S.; Jewell, R. C.; Paul, E. M.;
Purvis, N. B., Jr.; Verweij, J. Cancer Chemother. Pharma-
col. 2005, 55, 91.
18. Manna, S. K.; Mukhopadhyay, A.; Van, N. T.; Aggarwal,
B. B. J. Immunol. 1999, 163, 6800.
19. Saliou, C.; Valacchi, G.; Rimbach, G. Methods Enzymol.
2001, 335, 380.
20. Dhanalakshmi, S.; Singh, R. P.; Agarwal, C.; Agarwal, R.
Oncogene 2002, 21, 1759.
´
49. Ferte, J.; Kuhnel, J. M.; Chapuis, G.; Rolland, Y.; Lewin,
¨
G.; Schwaller, M. A. J. Med. Chem. 1999, 42, 478.
50. Zanden, J. J.; Wortelboer, H. M.; Bijlsma, S.; Punt, A.;
Usta, M.; Bladeren, P. J.; Rietjens, I. M. C. M.; Cnubben,
N. H. P. Biochem. Pharmacol. 2005, 69, 699.
21. Yoo, H. Y.; Jung, S. N.; Hwang, Y. S.; Park, J. S.; Kim,
M. H.; Jeong, M.; Ahn, S. J.; Ahn, B. W.; Shin, B. A.;
Park, R. K.; Jung, Y. D. Int. J. Mol. Med. 2004, 13,
81.