Dual inhibitors of P-glycoprotein and tumor cell growth: (Re)discovering thioxanthones

Article abstract of DOI:10.1016/j.bcp.2011.10.004

For many pathologies, there is a crescent effort to design multiple ligands that interact with a wide variety of targets. 1-Aminated thioxanthone derivatives were synthesized and assayed for their in vitro dual activity as antitumor agents and P-glycoprotein (P-gp) inhibitors. The approach was based on molecular hybridization of a thioxanthone scaffold, present in known antitumor drugs, and an amine, described as an important pharmacophoric feature for P-gp inhibition. A rational approach using homology modeling and docking was used, to select the molecules to be synthesized by conventional or microwave-assisted Ullmann C-N cross-coupling reaction. The obtained aminated thioxanthones were highly effective at inhibiting P-gp and/or causing growth inhibition in a chronic myelogenous leukemia cell line, K562. Six of the aminated thioxanthones had GI₅₀ values in the K562 cell line below 10 μM and 1-{[2-(diethylamino)ethyl]amino}-4-propoxy-9H-thioxanthen-9-one (37) had a GI₅₀ concentration (1.90 μM) 6-fold lower than doxorubicin (11.89 μM) in the K562Dox cell line. The best P-gp inhibitor found was 1-[2-(1H-benzimidazol-2-yl)ethanamine]-4-propoxy-9H-thioxanthen-9-one (45), which caused an accumulation rate of rhodamine-123 similar to that caused by verapamil in the K562Dox resistant cell line, and a decrease in ATP consumption by P-gp. At a concentration of 10 μM, compound 45 caused a decrease of 12.5-fold in the GI₅₀ value of doxorubicin in the K562Dox cell line, being 2-fold more potent than verapamil. From the overall results, the aminated thioxanthones represent a new class of P-gp inhibitors with improved efficacy in sensitizing a resistant P-gp overexpressing cell line (K562Dox) to doxorubicin.

Full text of DOI:10.1016/j.bcp.2011.10.004

Biochemical Pharmacology 83 (2012) 57–68
Contents lists available at SciVerse ScienceDirect
Biochemical Pharmacology
journal homepage: www.elsevier.com/locate/biochempharm
Dual inhibitors of P-glycoprotein and tumor cell growth: (Re)discovering
thioxanthones
Andreia Palmeira a,b,c, M. Helena Vasconcelos c,d, Ana Paiva a,b, Miguel X. Fernandes e, Madalena Pinto a,b
,
a,b,
*
´
Emılia Sousa
a
ˆ
ˆ
´
Departamento de Quı´mica, Laborato´rio de Quı´mica Organica e Farmaceutica, Faculdade de Farmacia, Universidade do Porto, Rua Anibal Cunha 164, 4050-047, Porto, Portugal
b
´
Centro de Quımica Medicinal (CEQUIMED-UP), Universidade do Porto, Portugal, Rua Anibal Cunha 164, 4050-047, Porto, Portugal
c Cancer Drug Resistance Group, IPATIMUP – Instituto de Patologia e Imunologia Molecular da Universidade do Porto, Portugal, Rua Dr Roberto Frias s/n, 4200-465, Porto, Portugal
d
´
Departamento de Cieˆncias Biolo´gicas, Laborato´rio de Microbiologia, Faculdade de Farmacia, Universidade do Porto, Portugal, Rua Anibal Cunha 164, 4050-047, Porto, Portugal
e Centro de Quı´mica da Madeira, Universidade da Madeira, Campus da Penteada, 9000-390, Funchal, Portugal
A R T I C L E I N F O
A B S T R A C T
Article history:
For many pathologies, there is a crescent effort to design multiple ligands that interact with a wide
variety of targets. 1-Aminated thioxanthone derivatives were synthesized and assayed for their in vitro
dual activity as antitumor agents and P-glycoprotein (P-gp) inhibitors. The approach was based on
molecular hybridization of a thioxanthone scaffold, present in known antitumor drugs, and an amine,
described as an important pharmacophoric feature for P-gp inhibition. A rational approach using
homology modeling and docking was used, to select the molecules to be synthesized by conventional or
microwave-assisted Ullmann C–N cross-coupling reaction. The obtained aminated thioxanthones were
highly effective at inhibiting P-gp and/or causing growth inhibition in a chronic myelogenous leukemia
Received 12 September 2011
Accepted 4 October 2011
Available online 22 October 2011
Keywords:
Thioxanthones
P-glycoprotein
Multidrug resistance
Anticancer
cell line, K562. Six of the aminated thioxanthones had GI50 values in the K562 cell line below 10
1-{[2-(diethylamino)ethyl]amino}-4-propoxy-9H-thioxanthen-9-one (37) had a GI50 concentration
(1.90 M) 6-fold lower than doxorubicin (11.89 M) in the K562Dox cell line. The best P-gp inhibitor
mM and
Dual ligands
m
m
found was 1-[2-(1H-benzimidazol-2-yl)ethanamine]-4-propoxy-9H-thioxanthen-9-one (45), which
caused an accumulation rate of rhodamine-123 similar to that caused by verapamil in the K562Dox
resistant cell line, and a decrease in ATP consumption by P-gp. At a concentration of 10
mM, compound
45 caused a decrease of 12.5-fold in the GI50 value of doxorubicin in the K562Dox cell line, being 2-fold
more potent than verapamil. From the overall results, the aminated thioxanthones represent a new class
of P-gp inhibitors with improved efficacy in sensitizing a resistant P-gp overexpressing cell line
(K562Dox) to doxorubicin.
ß 2011 Elsevier Inc. All rights reserved.
1. Introduction
Resistance to chemotherapy represents one of the major
obstacles to cancer treatment. Multidrug resistance (MDR) can
be broadly defined as a phenomenon by which tumor cells in vivo,
Abbreviations: ABC, ATP-binding cassette; ATP, adenosine-50-triphosphate; BLAST,
basic local alignment search tool; DNA, deoxyribonucleic acid; GI50, inhibition of
cell growth (the concentration needed to reduce the growth of treated cells to half
that of untreated cells); MDR, multidrug resistance; MFI, mean fluorescence
intensity; MRP, multridrug resistance proteins; MRP-1, multidrug resistance
protein 1; MW, molecular weight; NBD, nucleotide binding domain; P-gp, P-
glycoprotein; Rh123, rhodamine-123; RLU, relative light unit; RMSD, root-mean-
squared error displacement; SE, standard error; SRB, sulphorhodamine-B; TM,
and cultured cells in vitro, show simultaneous resistance to a
variety of structurally and functionally dissimilar cytotoxic and
xenobiotic compounds [1]. P-glycoprotein (P-gp), an ABC (ATP-
binding cassette) super-family member [2,3], is a membrane
transporter that actively extrudes a set of structurally unrelated
compounds, namely chemotherapeutic agents, out of the cells [4]
conferring the MDR phenotype in cancer [5]. Despite promising in
vitro results obtained for several generations of P-gp inhibitors,
disappointing clinical trials demand for new drugs and strategies
to reverse the P-gp-mediated MDR phenotype [6,7].
transmembrane (one
a-helix); TMD, transmembrane domain (six a-helixes).
´
´
ˆ
ˆ
* Corresponding author at: Laboratorio de Quımica Organica e Farmaceutica,
Universidade do Porto, Rua Anibal Cunha 164, 4050-047, Porto, Portugal.
Tel.: +351 222078984; fax: +351 222003977.
E-mail addresses: apalmeira@ff.up.pt (A. Palmeira), hvasconcelos@ipatimup.pt
(M.H. Vasconcelos), apaiva@ff.up.pt (A. Paiva), mxf@uma.pt (M.X. Fernandes),
madalena@ff.up.pt (M. Pinto), esousa@ff.up.pt, esousa@ff.up.pt (E. Sousa).
Simultaneous modulation of multiple biological targets can be
beneficial for treatment of diseases with complex etiologies such
0006-2952/$ – see front matter ß 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2011.10.004

58
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
as cancer [8,9]. Recently, there has been a paradigm shift in drug
design towards the development of multifunctional drugs, i.e.,
drugs that aim multiple targets [8,9]. The benefits of this approach
include improved compliance, efficacy and often reduced side
effects in comparison with polypharmacy [8]. In addition, some
multifunctional drugs may work better at normalizing the
pathology than a single targeted compound. This strategy has
already been used for the design of multifunctional agents for HIV
[10] and several neurodegenerative diseases [11,12]. Podophyllo-
toxin derivatives were recently described to fight simultaneously
cancer and MDR [13]. Nevertheless, a structure-based design
strategy to obtain antitumor agents which are concomitantly P-gp
inhibitors has never been attempted before. The aim of this work
was the design of P-gp inhibitors which would also behave as
cancer cell growth inhibitors.
introduced an amine side chain to pursuit
approach.
a hybridization
In this work, computational studies, synthesis and subsequent
investigation of cell growth inhibitory effects and P-gp modulation
for a series of new thioxanthonic derivatives are described, as
schematized in Fig. 2. A virtual library of molecules was built by
merging a thioxanthonic scaffold with P-gp inhibitors’ pharma-
cophoric features, i.e., a planar system of rings (thioxanthonic
scaffold) and a ionizable group such as an amine [26] (Fig. 2I).
Those thioxanthonic derivatives were docked into P-gp homology
models (Fig. 2II); the molecules with the best docking scores were
then synthesized (Fig. 2III) and their P-gp inhibition and cell
growth inhibitory effects experimentally assessed using several
techniques (Fig. 2IV–VIII).
Thioxanthones are S-heterocycles with a dibenzo-g-thiopyrone
2. Material and methods
scaffold and they are an important class of molecules showing
interesting biological properties, namely antitumor activity
[14,15]. Also, dibenzo-g-pyrones (xanthones) have already been
2.1. Preparation of a library of virtual thioxanthones and known P-gp
inhibitors/ligands
described as potential P-gp inhibitors, binding in the ATP-binding
site [16]. Therefore, thioxanthones combine structural features
that make them promising as dual ligands.
About one thousand thioxanthones, 22 known P-gp inhibitors
from the 1st, 2nd, and 3rd generation, thirty-six flavonoids that are
known to bind the P-gp ATP-binding site [27] (Supplementary
data, Annex B), were drawn and minimized using AM1 semi-
empirical method, with a gradient energy minimization method
until the energy change between steps was lower than
0.01 kcal molꢀ1 (Hyperchem, Hypercube, USA). The algorithm
used was the Polak-Ribiere (conjugate gradient).
The first thioxanthones described as potential antitumor agents
were hycanthone (1) [17] and lucanthone (2) [18], used in
therapeutics as antischistosomal agents (Fig. 1). However,
mutagenicity [19], probably resulting from the methylene moiety
directly linked to C-4, was reported as a major drawback, leading to
their withdraw [20]. Later, SR233377 (3) [15,21] and SSR271425
(4) [14,22,23] (Fig. 1) were tested in clinical trials as antitumor
agents but showed cardiotoxicity. In contrast, for the photo-
initiators 2-isopropyl-9H-thioxanthen-9-one (ITX, 5) [24], and 1-
chloro-4-propoxy-9H-thioxanthen-9-one (6), in vivo genotoxicity
studies do not indicate a genotoxic potential, and compound 6 was
considered not to be a dangerous substance according to GHS and
to Directive 67/548/EEC [25]. Hence, to continue the search for new
potential anticancer drugs with a thioxanthonic scaffold, we
considered compound 6 to be a suitable chemical substrate and we
2.2. Docking of virtual thioxanthones
Virtual screening was carried out on a commodity PC running
Linux Ubuntu 6.06. The software eHiTS from SimBioSys Inc. [28,29]
was used for active site detection and docking. Docking was
performed on P-gp models built using Sav1866 as template
(Supplementary data, Annex A). No special preparation of the 3D
structures was carried out because eHiTS automatically evaluates
[(Fig._1)TD$FIG]
CH3
CH3
CH3
CH3
CH3
CH3
N
N
N
HN
O
S
HN
O
S
HN
O
S
CH3
OH
NHSO2CH3
3
1
2
CH3
CH3
N
O
O
Cl
HN
O
S
CH3
O
CH3
CH3
O
S
S
O
N
H
H
CH3
6
4
5
Fig. 1. First thioxanthones described as potential antitumor agents: hycanthone (1) and lucanthone (2), SR233377 (3) and SSR271425 (4); thioxanthones used safely as
photoinitiators: 2-isopropyl-9H-thioxanthen-9-one (5) and 1-chloro-4-propoxy-9H-thioxanthen-9-one (6).

[
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
59
Fig. 2. Schematic representation of cheminformatic and laboratorial procedures. (I) Dual ligands design by merging a thioxanthonic scaffold with an amine. (II) Docking
studies. (III) Synthesis. (IV) Sulphorhodamine-B assay (SRB) for investigation of tumor cell growth inhibition in K562 and K562Dox cell lines. (V) Trypan blue exclusion assay
for investigation of cellular viability in a fibroblast cell line (MRC-5). (VI) Rhodamine-123 cellular accumulation assay and (VII) ATPase assay for P-gp inhibition screening.
(VIII) SRB assay in cells treated with a combination of doxorubicin and the test compound.
all of the possible protonation states for ligands and receptor.
eHiTS ran using the residues at: (i) the ATP-binding site on
nucleotide-binding domains (NBD) and (ii) the drug binding
pocket on the interface of the TMD as clip files. The docking is
absorbance at 254 and/or 365 nm. Melting points were obtained in
a Ko¨fler microscope. IR spectra were measured on an ATI Mattson
Genesis series FTIR spectrophotometer (Dekalb, IL, USA), software
WinFirst v.2.10 (Madison, WI, USA) in KBr microplates (cmꢀ1). 1H
and 13C NMR spectra were taken in CDCl3 or DMSO-d6 at room
temperature, on Bruker Avance 300 instrument (300.13 MHz for
1H and 75.47 MHz for 13C) (Blue lion biotech, Snoqualmie, WA,
USA). Chemical shifts are expressed in (ppm) values relative to
tetramethylsilane (TMS) as an internal reference (Supplementary
data, Annex E). Elemental analysis results were obtained in the
services of C.A.C.T.I., Vigo, Spain. The purity of each compound was
determined by HPLC-DAD analysis. All tested compounds, whether
synthesized or purchased, possessed a purity of at least 95%. The
compounds were synthesized and purified by the described
procedures (in Supplementary data D).
˚
performed, by default, within a 7 A margin around those residues.
The input files of the molecules (virtually designed thioxanthones)
were in mol format. The docking accuracy was set to 2, the number
of docking poses was set to 5, and sdf was chosen as the output file
format. Open Babel [30] was used to manipulate the various file
formats of ligands. PyMol from DeLano [31] and Chimera from
UCSF [32] were used for visual inspection of results and graphical
representations.
2.3. Synthesis
1-Chloro-4-propoxy-9H-thioxanthen-9-one (6) and amines (7–
28) were purchased from Sigma–Aldrich (Spain). The solvents used
were products pro analysis or HPLC grade from Sigma–Aldrich and
Fluka. Microwave reactions were performed using a MicroSYNTH
1600 from Millestone (ThermoUnicam, Portugal) synthesizer in
sealed reaction vessels (10 mL). The synthetic procedures and
spectral data are detailed in Supplementary data D. Purifications of
compounds were performed by flash chromatography using Merck
silica gel 60 (0.040–0.063 mm), liquid–liquid extraction
(NMP:water and diethyl ether system) and preparative thin layer
chromatography (TLC) using Merck silica gel 60 (GF254) plates
(Darmstadt, Germany). Reaction progression were controlled by
TLC performed using silica plates HF254 (l = 0.2 mm) (Merck,
Darmstadt, Germany) (ethyl acetate:acetone:TEA 80:20:1 or
90:10:1). Compounds were visually detected on visible light, by
2.4. HPLC chromatographic conditions
The HPLC analysis was performed in a Finnigan Surveyor–
Autosampler Plus and LC Pump Plus (Thermo Electron Corporation,
Waltham, MA, USA), equipped with a diode array detector TSP
UV6000LP, and using a C-18 column (5
m
m, 250 mm ꢁ 4.6 mm
I.D.), from Macherey-Nagel (Deuren, Germany). Xcalibur2.0 SUR 1
software (Thermo Electron Corporation, Waltham, MA, USA)
managed chromatographic data. Acetonitrile was of HPLC grade
from Merck. HPLC ultrapure water was generated by a Milli-Q
system (Millipore, Bedford, MA, USA). The mobile phases were
degassed for 15 min in an ultrasonic bath before use. It is used as an
isocratic elution of MeOH:H2O basified with TEA (1%) or acidified
with CH3COOH (1%) at a constant flow rate of 1.0 mL minꢀ1
.

60
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
2.5. Biological activity
resulting in 50% inhibition of cell growth) were calculated from
the plotted results.
2.5.1. Cell lines
The ability of the tested compounds to decrease doxorubicin
GI50 was evaluated in K562 and K562Dox cells by SRB assay, as
previously described [33]. K562 and K562Dox cells were plated
into 96-well tissue culture plates at 5 ꢁ 104 cells mLꢀ1. After 24 h,
cells were treated with serial dilutions of doxorubicin in
K562 (human chronic myelogenous leukemia, erythroblastic;
ECACC, Europe Collection of Cell Cultures, UK), K562Dox (derived
from K562 by doxorubicin stimulated overexpression of P-gp; kind
gift from Prof. J.P. Marie, Paris, France), and MRC-5 (human lung
fibroblasts which undergo between 60 and 70 doublings before
senescence) cell lines were routinely maintained in RPMI-1640
(with Hepes and Glutamax, Gibco1, Invitrogen, Germany), with
10% or 5% fetal bovine serum (FBS, Gibco1, Invitrogen, Germany)
and incubated in a humidified incubator at 37 8C with 5% CO2 in air.
All experiments were performed with cells in exponential growth,
with viabilities over 90% and repeated at least three times.
combination with 1 or 10
mM of the test compound. Following
48 h treatment, cell growth was assayed using the SRB assay. The
GI50 values for doxorubicin (concentration resulting in 50%
inhibition of cell growth) were calculated from the plotted results.
The ratio of doxorubicin GI50 decrease was obtained by calculating
the GI50 of doxorubicin alone/GI50 of doxorubicin in the presence of
the tested compound.
K562Dox cells were maintained by treating them with 1.0
mM
doxorubicin every two weeks in order to maintain the P-gp
overexpression, and all experiments were performed at least four
days after this administration and in doxorubicin-free medium.
2.5.6. Viable cell number by trypan blue exclusion assay
Adherent MRC-5 cells were plated (2 ꢁ 105 cells mLꢀ1 in 96-
well plates). 24 h after plating, cells were treated with complete
medium (blank), with the solvent of the compounds (DMSO,
control), or with the following compounds: 30, 33, 37, 38, 41, 43,
45, and 48 in a concentration corresponding to their GI50 on K562
cell line. Viable cell number was assessed following 48 and 72 h of
incubation with the compounds using the trypan blue exclusion
assay (results not shown).
2.5.2. Compounds
SRB and rh123 were obtained from Sigma Aldrich. Verapamil,
quinidine and mibefradil (known P-gp inhibitors) were used as
controls (Sigma Aldrich, Spain).
2.5.3. Flow cytometry determination of rhodamine-123 accumulation
K562 andK562Dox (5 ꢁ 106 cells mLꢀ1) were incubated for 1 h at
2.5.7. Statistical analysis
37 8C inthe presence of10or20
1
m
M ofthetest compounds, and with
Data was expressed as the mean ꢂ SE and analysed by the
Student’s t-test. P-values below 0.05 were considered statistically
significant.
m
M rh123. K562Dox and K562 cells alone as well as K562Dox cells
in the presence of the known P-gp inhibitors verapamil, mibefradil,
quinidine and hycanthone (2) [33] (10 and 20 M) were used as
m
controls. After the incubation time, cells were washed twice, re-
suspendedinice coldPBSand kept at4 8Cinthe darkuntilanalysisin
the flow cytometer. At least 20,000 cells per sample were counted
and analysed by flow cytometry (Epics XL-MCL, Beckman Coulter,
USA). Cells shown in forward scatter and side scatter were
electronically gated and acquired through the FL1 channel. The
amount of fluorescence was plotted as a histogram of FL1 within the
gate. Data acquisition was performed using WINMDI 2.9 (TSRI, USA)
to determine median fluorescence intensity values (MFI). For simple
interpretation, the ratio of rh123 accumulation was calculated as
(MFIK562Dox+Drug ꢀ MFIK562Dox)/MFIK562Dox. Results represent the
average of at least three independent experiments.
3. Results
3.1. Design and docking of thioxanthones
Docking was performed using a dataset of approximately 1000
virtual new aminated thioxanthones (Table 1A) and two P-gp
models constructed using the homologous Sav1866 from Staphy-
lococcus aureus (Supplementary data, Annex A) as template. The
designed molecules resulted from a merging strategy for dual
ligands [8,9] (represented in Fig. 2I) between a thioxanthonic
scaffold, associated to antitumor activity [15,22,23], with an 4-
alkoxyl or 4-hydroxyl group; and an amine group at C-1, associated
to P-gp inhibitory activity [26,35]. The introduction of an amine
side chain, namely an aminoalkylamino group, on the thiox-
anthonic nucleus seems to be also an important feature for the
expression of cytotoxicity [36–38]. Thus, amine structures [39,40]
such as the N,N-diethylethane-1,2-diamine, present in thiox-
anthones SR233377 (3) and SR271425 (4), and in several DNA
intercalating agents, were also used in the design of 1-aminated
thioxanthonic derivative. Additionally, the presence of an amine in
these derivatives allows the formation of salts, which improves
compound water solubility [41].
Computational filters were applied to the initial library of
aminated thioxanthones, in order to select the molecules that
complied with the Lipinski rules of five [42], and other features
described as important for P-gp inhibition, such as log P value of at
least 2.92 or higher [26,35] (Table 1A). The molecules that
respected these cutoffs were docked into the drug binding pocket
formed by the interface between TMD1 and TMD2 and in the ATP-
binding site located on the NBD of P-gp. Twenty-three aminated
thioxanthones with the best docking scores (on NBD or TMD)
(Table 1B and C) and with a rapid synthetic protocol, were selected
to be synthesized, starting from thioxanthone 6 and amines 7–28.
These molecules were selected based on the following cut-offs:
molecules with NBD docking scores lower than ꢀ3.0 kJ molꢀ1, and/
2.5.4. Determination of ATPase activity
The ATPase activity of P-gp was determined using the
luminescent ATP detection kit (P-gp-Glo Assay Kit, Promega,
Germany) according to the manufacturers’ recommendation [34].
Test compounds at 200
control, noncompetitive inhibitor) or verapamil, mibefradil, and
quinidine at 200 M (positive controls, competitive inhibitors) in
mM or sodium vanadate at 20 mM (positive
m
buffer solution were incubated with 0.5 mg mLꢀ1 P-gp and 5 mM
MgATP at 37 8C for exactly 40 min, and the remaining ATP was
detected after 20 min resting at room temperature, as a luciferase-
generated luminescent signal. Results are presented as the average
of three independent experiments. % Values of RLU (relative light
unit) was calculated in relation to non-treated control (NT). It was
calculated using the following formula that fits the NT control to
zero for easier interpretation:
RLUNT ꢁ 100.
%
RLU = (RLUtest ꢀ RLUNT)/
2.5.5. Sulphorhodamine-B assay
K562 and K562Dox cells were plated into 96-well tissue
culture plates at 5 ꢁ 104 cells mLꢀ1. After 24 h, cells were treated
with serial dilutions of the test compound. Following 48 h
treatment, cell growth was assayed using the SRB assay. The
GI50 values for the thioxanthonic derivatives (concentration

A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
61
Table 1
(A) Scheme of the virtual screening protocol; best scoring virtually designed thioxanthones with (B) a 4-propoxyl or (C) a 4-hydroxyl substituent at C-4.
[TD$INLINE]

62
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
Table 1 (Continued)
[TD$INLINE]
a'
CH3
2''
1''
NH2
3''
4''
2'
3'
HN
O
O
1'
6'
3'
a) -3.755
4'
5'
2'
1'
416 5.45
421 5.90
4'
51 a) -3.644
CH3
b'
43 b) -3.52
451 5.88
N
b) 1.169
5'
O
CH3
c'
19
a'
27
O
CH3
2'
3'
4'
5'
HN
1'
1'
44 a) -4.267
HN
6'
5'
2'
3'
b) -2.638
6'
O
CH3
OH
b'
4'
a) -5.441
O
f'
O
e'
d'
20
b) -1.972 539 3.76
a'
52
NH
2''
HO
b'
OH
H
N
1''
6''
c'
a'
c'
3''
1'
OH
a) -4.981
45
429 4.15
433 5.86
b'
b) 0.317
4''
N
28
5''
R
21
C
O
O
4'
3'
2'
1'
O
b'
a) -5.084
46
S
OH
NH
b) -0.324
6' a'
5'
22
Docking
scores on
a)NBD
a''
CH3
3'
MW LogP
O
b''
CH3
O
2'
1'
b)TMD
(kJ mol-1)
4'
a) -3.217
47
a'
HN
435 5.74
436 5.11
438 5.03
b) -1.214
a'
CH3
O
5'
6'
2'
HN
1'
6'
3'
23
a) -2.493
b) -2.554
4'
O
5' 6'
CH3
b'
55
409 3.86
5'
1' b'
O2N
O
4'
CH3
c'
a) -5.184
48
a'
3' 2'
HN
b) -1.017
27
24
LogP= logarithm of the partition
5'
O2N
6'
4'
3'
coefficient; MW= molecular weight
a'
1'
2'
a) -5.953
49
O
NH
b) -1.505
25
6'
HN
1'
2'
5'
4'
a) -3.428
50
450 4.85
O
a'
O2N
3'
b) -2.19
CH
b'
3
26
or with TMD docking score lower than ꢀ2.0 kJ molꢀ1 (see
amine in excess (two equivalents of amine, 7–28) and thiox-
anthone 6 as limitating reagent, Cu2O [43] in catalytic amounts (5%
equiv.) at 100 8C in a closed vessel for 1–2 days (Scheme 1,
procedure a). However, the reaction was only successful for simple
aliphatic amines (such as 8, with a yield of 30%). Traditional
copper-mediated Ullmann couplings generally have low functional
group tolerance and irreproducible results [44,45], and require
harsh reaction conditions, such as high temperatures and the use of
strong bases [46]. The addition of K2CO3 (Scheme 1, procedure b)
provided a alkaline media and allowed the synthesis of 36 with a
25% yield and of 47 with a 75% yield. As microwave-promoted
examples of discarded molecules in Supplementary data, Annex C).
3.2. Synthesis
Thioxanthones 31–52 (Scheme 1) were synthesized by reaction
between 1-chloro-4-propoxy-9H-thioxanthen-9-one (6) and a
primary or a secondary amine (7–28, Table 1) by C(1)-N Ullmann
cross coupling.
The synthesis of aminated thioxanthones was initially investi-
gated by a classical copper mediated cross-coupling reaction, using

[
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
63
O
S
Cl
O
O
R
a) Cu2O, MeOH,
100 ºC, 2 days; b)
Cu2O, K2CO3, MeOH,
8
5
1
9a
4a
8a
2
7
a, b or c
9
RH
+
6
3
100 ºC, 2 days; c)
S
10a
4
b
Cu2O, K2CO3, NMP,
microwaves, 205 ºC,
50 min
R= 7 - 29
CH3
O
a
30 - 52
CH3
c
30*: R= -OCH3 (29)
31: R= 7
32: R= 8
38: R= 14
39: R= 15
40: R= 16
41: R= 17
42: R= 18
43: R= 19
44: R= 20
45: R= 21
46: R= 22
47: R= 23
48: R= 24
49: R= 25
50: R= 26
51: R= 27
52: R= 28
33: R= 9
34: R= 10
35: R= 11
36: R= 12
37: R= 13
Scheme 1. Copper-catalyzed nucleophilic aromatic substitution of thioxanthones. *Obtained as a secondary product by reaction of 6 with methanol (procedure a).
synthesis is an area of crescent interest in N–C cross-coupling
reaction [47], the heating source was also modified (Fig. 2,
procedure c) [48,49]. Additionally, the solvent was changed to N-
methyl-pyrrolidone (NMP), an aprotic, alkaline solvent, more
appropriate for microwave heating. In some cases, a mixture of
NMP:water [50,51] was used in order to improve the solubility of
the alkaline catalyst and the amine, allowing the reaction to
proceed (amines 11, 21, and 28). Under these conditions, the
amination reaction worked not only for alkylamines (e.g. 12 and
13) but also for cyclic amines (e.g. 14 and 19) and aromatic amines
(e.g. 20, 26, and 27) and allowed to obtain the desirable aminated
thioxanthones (36–38, 43, 44, 50, and 51, respectively).
To obtain compound 55, a dealkylation of 6 with boron
bromide, BBr3 [52] followed by an acetylation was performed
(Scheme 2). Reaction of compound 54 with 3,4,5-trimethoxyani-
line (27) furnished directly the desirable 4-hydroxy-1-aminated
derivative 55 along with compounds 53 and 56 (Scheme 2).
In summary, twenty-three new aminated thioxanthones
predicted from the docking study to be potential P-gp modulating
agents were synthesized in reasonable yields to perform biological
assays. Other thioxanthonic derivatives obtained were also
investigated in the biological assays (30, 53, 54, and 56). The
structure elucidation (Supplementary data, Annexes D and E) of
the synthesized compounds was based on spectral data (IR, 1H
NMR, 13C NMR, HMBC, and HSQC) and on elemental analysis.
3.3. Flow cytometry rhodamine-123 assay
An indirect way to analyse P-gp activity is by the evaluation of
the mean fluorescence intensity (MFI) of cells exposed to
rhodamine-123 (rh123), a known P-gp fluorescent substrate,
together with the potential P-gp inhibitor [53–55]. Quantification
of the relative fluorescence accumulation can be used as an indirect
way to study the individual potency of inhibitors [56]. This flow
cytometry method for P-gp functional evaluation has been used for
decades [57,58] and was chosen to initiate the present study.
However, rh123 is a substrate not only for P-gp but also for MRP
[59] and expression of both P-gp and MRP has been reported to
[(Scheme_2)TD$FIG]
O
S
Cl
O
O
Cl
O
O
S
Cl
acetic anhydride, pyridine
r.t.
1) dry CH2Cl2, -70°C, BBr3
2) MeOH (drop-wise)
S
O
OH
54
98%
53
88%
CH3
6
CH3
CH3
CH3
O
O
O
CH3
Cu2O, NMP
Microwaves, 205°C, 50 min
27
H3C
O
CH3
O
NH2
CH3
O
S
O
S
HN
O
+
O
S
Cl
+
OH
OH
OH
55
53
56
5%
7%
51%
Scheme 2. Synthesis of compound 55.

[
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
Fig. 3. Relative accumulation ratios of rh123 after 1 h of incubation in K562 or K562Dox cell lines. K562Dox alone (no drug) is represented as zero for easier interpretation.
Accumulation ratio superior to zero correspond to rh123 accumulation superior to K562Dox, i.e., potential P-gp inhibitors. Results are the mean ꢂ SE of three independent
experiments. Results are the average of three independent experiments ꢂ SE. Statistical significance was tested by paired t-test using the untreated K562Dox (second entry) as
control. *** Indicates P < 0.001; ** indicates 0.001 < P ꢃ 0.01; * indicates 0.01 < P ꢃ 0.05; ns indicates not significant, i.e., P > 0.05 (n = 3 independent experiments).
occur in leukemic cells [60]. We have previously failed to detect
MRP-1 by Western Blot in either K562 or K562Dox cell lines,
whereas we confirmed that P-gp is expressed in K562Dox cell line
but not in K562 cell line [33]. The results from the effect of the test
compounds 2, 6, 30–56 in the cellular accumulation of rh123 are
shown in Fig. 3.
indirectly by blocking an allosteric site relevant for P-gp activity
(i.e., noncompetitive inhibitors) [62,63]. Therefore, to elucidate if
the thioxanthonic derivatives that caused accumulation of rh123
were either competitive inhibitors (substrates for transport by P-
gp), or P-gp noncompetitive inhibitors, effects on the ATPase
activity of P-gp were measured using human P-gp membranes
[33,64]. The results for the fifteen potential P-gp inhibitors
(compounds 30, 33, 35, 37, 38, 41–45, 48–51, and 55) along with
the building block 6 and four controls (sodium orthovanadate,
verapamil, mibefradil and quinidine) are shown in Fig. 4.
Sodium orthovanadate is an inhibitor of P-gp ATPase activity
[65] and as expected caused significant increase in % relative light
unit (% RLU), whereas verapamil, mibefradil and quinidine are P-gp
substrates and competitive inhibitors, showing decreased % RLU
(Fig. 4).
Compounds 30, 33, 35, 38, 41, 42, 45,and 48–51 caused a
significant increase in the % RLU in relation to the untreated (NT),
i.e., they were found to be P-gp ATPase inhibitors (noncompetitive
P-gp inhibitors). They may prevent ATP hydrolysis by targeting the
P-gp ATP binding site directly [16,66] or by binding to an allosteric
residue [67,68] which causes inhibition of P-gp function with
resulting inhibition of ATPase activity. Compounds such as 37, 43,
44, and 55 behave similarly to verapamil, mibefradil, and
quinidine, i.e., they increased the hydrolysis of ATP through P-
gp, indicating that they are competitive inhibitors of P-gp, possibly
being themselves transported by the pump. In summary, from the
sixteen compounds that had caused accumulation of rh123 in the
cells, four are acting as competitive and eleven as noncompetitive
inhibitors of P-gp (docking poses of two of the best P-gp inhibitors,
43 and 45, are described on Supplementary data, Annex F).
Furthermore, there was a correlation between docking scores on
NBD and experimental results for noncompetitive ATPase inhibi-
tors, as well as between docking scores on TMD and experimental
results for competitive ATPase inhibitors (Supplementary data,
Annex G).
After K562 cells were incubated with rh123, accumulation ratio
increased when compared with K562Dox, as it was expected for a
cell line which does not express P-gp (Fig. 3). Untreated K562Dox
cells rh123 accumulation was considered residual (Fig. 3).
Verapamil, quinidine and mibefradil, known first-generation P-
gp inhibitors [61], and hycanthone (2), were used in this study as
controls. As expected, they caused a rh123 accumulation on
K562Dox cells similar to the one observed on the sensitive cell line,
K562 (Fig. 3). After K562Dox cells were incubated in the presence
of the test compounds (2, 6, 30–56), the accumulation ratios
varied. In the case of fifteen of the investigated compounds
(compounds 30, 33, 35, 37, 38, 41–45, 48–51, and 55, Fig. 3),
together with the controls (verapamil, quinidine, mibefradil and
compound 2), an increase in the accumulation ratio of rh123 was
observed, an effect compatible with P-gp inhibition. Particularly, 1-
[2-(1H-benzimidazol-2-yl)ethanamine]-4-propoxy-9H-thiox-
anthen-9-one (45), 1-(4-acetylpiperazin-1-yl)-4-propoxy-9H-
thioxanthen-9-one (38) and 1-{[2-(4-nitrophenyl)ethyl]amino}-
4-propoxy-9H-thioxanthen-9-one (48) showed an effect similar to
that elicited by the known P-gp inhibitor, quinidine, and higher
than the known thioxanthone hycanthone (2). A dose-dependent
response was clearly observed for compound 48, whereas there
was no clear evidence of
a dose-dependent response for
compounds 38 and 45. In contrast, twelve derivatives (compounds
31, 32, 34, 36, 39, 40, 46, 47, 52–54, and 56, Fig. 3) showed a
significant decrease in the accumulation ratio, an effect compatible
with P-gp activation. Although the building block 6 had no effect
on P-gp function (Fig. 3), the side product 30 revealed an
interesting rh123 accumulation rate.
3.4. P-gp ATPase assay
3.5. Growth inhibition assay
The rate of P-gp ATP hydrolysis allows to discriminate between
compounds that increase the ATP consumption by P-gp (i.e.,
competitive inhibitors or substrates) and compounds that reduce
the use of ATP either by acting directly on ATP-binding site or
The cell growth inhibitory effect of twenty seven thioxanthones
(30–56) was examined in the K562 and K562Dox cell lines by the
sulphorhodamine-B assay (SRB) and results are presented in Table
2A and B, respectively. On K562 cell line, the GI50 values for

[
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
65
30
20
10
0
***
***
***
***
***
**
**
**
**
*
*
*
*
*
-10
-20
**
**
ns
**
***
***
Compounds
Fig. 4. P-gp ATPase assay. Results are % of RLU (relative light unit) in relation to non-treated (NT) control. NT control luminescence is represented as zero for easier
interpretation. % Values of RLU superior to zero correspond to inhibitors of P-gp ATPase, i.e., noncompetitive P-gp inhibitors. % Values of RLU inferior to zero correspond to
substrates of P-gp, i.e., competitive P-gp inhibitors. Results are the mean ꢂ SE of three independent experiments. Na3VO4, verapamil, mibefradil, and quinidine were used as
controls. Statistical significance was tested by paired t-test using the NT-control as a negative control. *** Indicates P < 0.001; ** indicates 0.001 < P ꢃ 0.01; * indicates
0.01 < P ꢃ 0.05; ns indicates not significant, i.e., P > 0.05 (n = 3).
twenty-four of the compounds (30, 33–39, 41–56) were in the
range 1.9–104.7
GI50 < 10 M (30, 33, 37, 41, 43, 55). The most potent cell growth
inhibitor thioxanthone was 1-{[2-(diethylamino)ethyl]amino}-4-
propoxy-9H-thioxanthen-9-one (37), with a GI50 of 1.9 M. Eight
of the thioxanthones that presented the best P-gp and cell growth
inhibitory activity (30, 33, 37, 38, 41, 43, 45, and 48) were further
investigated on the K562Dox cell line (Table 2B). As expected, the
GI50 values in K562Dox cell line were generally higher than in the
K562 cells. Six of the compounds investigated on K562Dox cell line
mM (Table 2A). Six compounds showed a
Table 2
m
Cytotoxic effects of thioxanthones 30–56 on sensitive K562 (A) and resistant
K562Dox cells (B).
m
(A)
Compound
GI50 (K562) (mM)
37
1.90 ꢂ 0.15
3.00 ꢂ 0.48
43
33
3.72 ꢂ 1.47
55
4.38 ꢂ 0.44
showed a GI50 value < 20
mM (30, 33, 37, 41, 43, and 45),
30
4.47 ꢂ 1.93
indicating that they might have both P-gp and cell growth
inhibitory activity. Noteworthy, compound 37 was 9-fold more
potent than doxorubicin in this cell line.
The growth inhibitory effect of the thioxanthonic derivatives
was also assayed on a nontumor cell line model, MRC-5. There was
no significant decrease in viable cell number of cells treated with
compounds 30, 33, 37, 38, 41, 43, 45, and 48, when compared to
DMSO solvent control, 48 and 72 h after treatment (results not
shown).
To investigate the ability of these newly designed dual
inhibitors to reverse the MDR phenotype, a doxorubicin sensitiza-
tion assay in the P-gp overexpressing cell line (K562Dox) was
performed (Table 3A). To understand if the growth inhibitory effect
on this cell line was due to their dual actions, a sensitization assay
was also performed on K562 cell line (Table 3B). We hypothesized
that if a compound was able to decrease doxorubicin GI50 in
K562Dox, without decreasing doxorubicin GI50 in K562 cell line,
then its effect should be mainly due to P-gp inhibition. On the other
hand, if a compound decreases doxorubicin GI50 on both K562 and
K562Dox cell lines, then its effect should be due to both cell growth
inhibitory activity and P-gp inhibition. Results of the SRB assay on
cells treated with a combination of the thioxanthonic derivative
(30, 33, 37, 38, 41, 43, 45, or 48) and doxorubicin are presented in
Table 3. Two concentrations of the thioxanthonic derivatives were
41
4.81 ꢂ 4.21
50
12.98 ꢂ 0.36
13.57 ꢂ 2.96
15.57 ꢂ 3.15
16.22 ꢂ 0.48
16.50 ꢂ 3.06
16.99 ꢂ 2.33
18.13 ꢂ 4.35
19.23 ꢂ 0.98
20.96 ꢂ 2.08
21.47 ꢂ 2.61
22.73 ꢂ 0.64
29.79 ꢂ 3.02
52.95 ꢂ 1.47
59.45 ꢂ 2.77
60.58 ꢂ 2.01
74.32 ꢂ 7.16
92.92 ꢂ 3.33
104.71 ꢂ 7.29
54
51
35
39
45
49
44
48
36
53
38
34
52
42
56
46
47
31
>150
32
>150
>150
>150
40
Verapamil
Doxorubicin
0.06 ꢂ 0.01
(B)
Compound
GI50 (K562Dox) (mM)
37
1.95ns ꢂ 0.34
11.89* ꢂ 1.17
12.59* ꢂ 0.69
15.38* ꢂ 0.53
19.43ns ꢂ 0.76
19.60* ꢂ 2.73
37.31* ꢂ 2.08
41.56* ꢂ 5.48
>150
43
used, 1 and 10
tested alone.
mM, which are below the GI50 of the compounds
33
41
At 10
mM, verapamil caused a 5.9-fold decrease in doxorubicin
45
GI50 on K562Dox cell line. At the same concentration, compound
45 decreased the doxorubicin GI50 by 12.6-fold, followed by
compound 38 with a 8.9-fold decrease, and by compound 48 with a
30
38
48
Verapamil
Doxorubicin
7.5-fold decrease (Table 3A). Compound 45 at 10
mM was
11.62* ꢂ 1.14
approximately 2-fold more potent than verapamil, and compounds
38 and 48 were approximately 1.5-fold more potent than
verapamil, probably due to a dual activity as P-gp and cell growth
inhibitors, since these compounds also reduced doxorubicin GI50 in
the sensitive K562 cell line (Table 3B).
Results are mean ꢂ SE of at least three independent experiments, except for
compounds with GI50 > 150 mM (two independent experiments).
*
Values of K562Dox GI50 are statistically different (P < 0.05) from the K562 GI50
.
ns: values of K562Dox GI50 are not statistically different (P > 0.05) from the K562
GI50
.

66
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
Table 3
Cytotoxic effects of doxorubicin in combination with newly synthesized thioxanthones (at 1 or 10
mM) on resistant K562Dox (A) and sensitive K562 cells (B).
(A) K562Dox
(B)K562
GI50
(m
M)
Ratio of doxorubicin GI50 decrease
GI50 (nM)
Ratio of doxorubicin GI50 decrease
Doxorubicin (Dox)
11.62 ꢂ 1.14
1.97 ꢂ 0.03***
0.92 ꢂ 0.16***
1.30 ꢂ 0.25***
1.56 ꢂ 0.16***
–
61.67 ꢂ 1.27
55.76 ꢂ 1.12*
15.89 ꢂ 0.6***
6.34 ꢂ 0.02***
9.65 ꢂ 1.07***
–
Dox + 10
Dox + 10
Dox + 10
Dox + 10
m
m
m
m
M Verapamil
M 45
5.90
12.6
8.94
7.45
1.11
3.88
9.73
6.39
M 38
M 48
Dox + 1
Dox + 1
Dox + 1
Dox + 1
Dox + 1
Dox + 1
Dox + 1
Dox + 1
Dox + 1
m
M Verapamil
M 37
4.46 ꢂ 0.21**
3.15 ꢂ 0.32**
7.17 ꢂ 0.20*
7.36 ꢂ 0.16*
8.50 ꢂ 0.33*
9.60 ꢂ 0.27*
9.64 ꢂ 0.21*
11.57 ꢂ 0.37ns
11.70 ꢂ 0.34ns
2.61
3.69
1.62
1.58
1.37
1.21
1.21
1.00
0.99
58.45 ꢂ 1.17ns
4.71 ꢂ 0.40***
58.31 ꢂ 0.9ns
57.16 ꢂ 1.80ns
51.84 ꢂ 0.52*
49.11 ꢂ 0.62**
13.00 ꢂ 0.88***
16.37 ꢂ 1.00***
42.20 ꢂ 1.45**
1.05
13.09
1.06
1.08
1.19
1.26
4.74
3.77
1.46
m
m
m
m
m
m
m
m
M 45
M 48
M 38
M 41
M 43
M 33
M 30
The selected concentrations of test compounds are below the GI50 of the tested compounds on each cell line when tested alone. Statistical significance was tested by paired t-
test using the doxorubicin alone GI50 of each cell line as the control. *** Indicates P < 0.001; ** indicates 0.001 < P ꢃ 0.01; * indicates 0.01 < P ꢃ 0.05. ns indicate not significant,
i.e., P > 0.05 (n = 3 independent experiments). The ratio of doxorubicin GI50 decrease was calculated as follows: GI50 of doxorubicin alone/GI50 of doxorubicin in the presence
of the test compound.
When compounds were tested at 1
and 38 were the ones which had a greater effect in reducing the
GI50 of doxorubicin in the K562Dox cell line. In contrast, in K562
cell line, compounds 45 and 48 at 1
m
M, compounds 37, 45, 48,
due to the fact that the rh123 accumulation assay uses cells and the
ATPase assay uses membrane fractions, and therefore, the latter
does not suffer interferences by factors related to drug permeabil-
ity [69,70].
mM did not cause any
alteration in the GI50 of doxorubicin, suggesting that the
sensitization effect of those two compounds seen in the
K562Dox cell line is due to their P-gp inhibitory activity. On the
The study of the cell growth inhibitory effect of twenty seven
thioxanthones (30–56) revealed that six compounds present GI50
values < 10
mM (30, 33, 37, 41, 43, 55) in the K562 cell line. In most
other hand, compound 37 at 1
m
M caused a 13-fold decrease on
cases, the GI50 values in the K562Dox cell line were higher than in
the K562 cells. Nevertheless, the most potent cell growth inhibitor
thioxanthone, 1-{[2-(diethylamino)ethyl]amino}-4-propoxy-9H-
doxorubicin GI50 on the K562 cell line (GI50 = 4.71 ꢂ 0.40 nM),
suggesting that the effects of compound 37 are not only due to P-gp
inhibition but also and mainly due to cell growth inhibition.
Additionally, compound 38 at 1 mM caused a 1.2 fold-reversal on
this cell line, which implies that, besides the P-gp inhibitory effect of
this compound, which seems predominant, a significant antileukemic
effect is also observed. Compounds 30, 33, and 43, sensitized K562
cells more than K562Dox cells to the effect of doxorubicin, suggesting
that their effect is mostly on cell growth.
thioxanthen-9-one (37), showed similar GI50 values (1.9
mM) in
both tumor cell lines, being 9-fold more potent than doxorubicin in
K562Dox cells. Indeed, compound 37 which is structurally similar
to the known antitumor thioxanthones 1–4, was shown in the
present work to be a potent inhibitor of tumor cell growth. Further
studies are needed in order to investigate the effect of compound
37 on cellular proliferation, cell cycle and apoptosis. Our future
work will focus on these aspects as well as on the molecular
mechanism of action of some of the most promising compounds,
such as compound 37. The possible mechanisms that will be
investigated will include intercalation with DNA, as has been
previously reported for the known lucanthone and hycanthone
[71].
4. Discussion
The synthesis of 1-aminated thioxanthones was performed by
Ullman cross-coupling, using conventional or microwave heating.
Both methods yielded the desirable compounds. The presence of a
base was favorable and increased the yields of reaction. After
purification and structural elucidation, a rh123 accumulation
assay was performed in order to select the thioxanthones which
were interfering with the efflux of rh123 by P-gp (Fig. 3).
Noteworthy, the most potent compounds (potential P-gp inhibi-
tors) present two nitrogen atoms (amine or nitro groups) separated
by an alkyl (37, 38) or alkyl-aryl chain (45 and 48).
From the sixteen compounds that had caused accumulation of
rh123 in the cells, four are acting as competitive inhibitors
(increased P-gp ATPase activity compared to non-treated control)
and eleven as noncompetitive inhibitors of P-gp (decreased P-gp
ATPase activity comparing to non-treated control). When compar-
ing these results (Fig. 4) with the ones obtained on the rh123
accumulation assay (Fig. 3), general agreement between the
outcomes of the two assays could be observed. Indeed, most of the
compounds with the highest accumulation ratio (>1.0) in the
rh123 accumulation assay, such as 30, 38, 45, and 48 also showed a
high activity in the ATPase assay. Nevertheless, compound 41,
which presented a high % value of RLU in the ATPase assay, showed
an accumulation ratio of rh123 <0.5. This discrepancy in the
activity of some compounds measured by two assays is probably
Concerning the capacity of MDR reversal of P-gp overexpressing
cell line K562Dox, compound 45 at 10
mM was approximately 2-
fold more potent than verapamil, and compounds 38 and 48 were
approximately 1.5-fold more potent than verapamil (Table 3A),
probably due to a dual activity as P-gp and cell growth inhibitors,
since these compounds also reduced doxorubicin GI50 in the
sensitive K562 cell line (Table 3B). In spite of the lower rh123
accumulation on K562Dox in comparison to verapamil (Fig. 3),
compounds 38 and 48 at 10
mM succeed to sensitize resistant
K562Dox cells due to its dual activity as P-gp and cell growth
inhibitors.
These results suggest that the dual effect of some 1-aminated
thioxanthones have a sensitization effect to a traditional che-
motherapeutical agent such as doxorubicin, which might be a
valuable strategy to explore further in the design of efficient dual
activity molecules.
5. Conclusion
The rational merged scaffolds approach rendered a new class
of dual activity agents (cell growth and P-gp inhibitors),

A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
67
[19] Cioli D, Pica-Mattoccia L, Archer S. Antischistosomal drugs: past, present and
future? Pharmacol Ther 1995;68:35–85.
aminoalkylthioxanthones. We identified 1-{[2-(diethylamino)
ethyl]amino}-4-propoxy-9H-thioxanthen-9-one (37) as a potent
cell growth inhibitor. We also recognized 1-[2-(1H-benzimidazol-
2-yl)ethanamine]-4-propoxy-9H-thioxanthen-9-one (45) and 1-
{[2-(4-nitrophenyl)ethyl]amino}-4-propoxy-9H-thioxanthen-9-
one (48) as dual noncompetitive inhibitors of P-gp and inhibitors of
cell growth, being more active than verapamil in sensitizing a P-gp
overexpressing cell line to doxorubicin.
[20] Hartman PE, Hulbert PB, Bueding E, Taylor DD. Microsomal activation to
mutagens of antischistosomal methyl thioxanthenones and initial tests on
a possibly non-mutagenic analogue. Mutat Res 1975;31:87–95.
[21] LoRusso PM, Foster BJ, Wozniak A, Heilbrun LK, McCormick JI, Ruble PE, et al.
Phase I pharmacokinetic study of the novel antitumor agent SR 233377. Clin
Cancer Res 2000;6:3088–94.
[22] Campone M, Isambert N, Bourbouloux E, Maury S, Monin-Baroille P, Berille J,
et al. Phase I dose-escalation study of a novel antitumor agent, SR 271425,
administered intravenously in split doses (d1–d2–d3) in patients with refrac-
tory solid tumors. Cancer Chemother Pharmacol 2007;59:689–95.
[23] Goncalves PH, High F, Juniewicz P, Shackleton G, Li J, Boerner S, et al. Phase I
dose-escalation study of the thioxanthone SR 271425 administered intrave-
nously once every 3 weeks in patients with advanced malignancies. Invest
New Drugs 2008;26:347–54.
[24] Anton R, Barlow S, Toldra´ F. Opinion of the scientific panel on food additives,
flavourings, processing aids and materials in contact with food on a request
from the commission related to 2-isopropyl thioxanthone (ITX) and 2-ethyl-
hexyl-4-dimethylaminobenzoate (EHDAB) in food contact materials. EFSA J
2005;293:1–15.
[25] 1-Chloro-4-propoxy-9H-thioxanthen-9-one safety datasheet. S-AC Ltd.; 2010.
[26] McDevitt CA, Callaghan R. How can we best use structural information on P-
glycoprotein to design inhibitors? Pharmacol Ther 2007;113:429–41.
[27] Chan KF, Zhao Y, Burkett BA, Wong IL, Chow LM, Chan TH. Flavonoid dimers as
bivalent modulators for P-glycoprotein-based multidrug resistance: synthetic
apigenin homodimers linked with defined-length poly(ethylene glycol)
spacers increase drug retention and enhance chemosensitivity in resistant
cancer cells. J Med Chem 2006;49:6742–59.
Acknowledgements
This work is funded through national funds from FCT –
ˆ
a Ciencia e a Tecnologia under the project
Fundac¸a˜o para
CEQUIMED – PEst-OE/SAU/UI4040/2011, by FEDER funds through
the COMPETE program under the project FCOMP-01-0124-FEDER-
011057, and by U. Porto and Santander-Totta. The authors would
like to thank Sara Cravo for technical support in microwave
methodology. IPATIMUP is an Associate Laboratory of the
Portuguese Ministry of Science, Technology and Higher Education
and is partially supported by FCT, the Portuguese Foundation for
Science and Technology.
[28] Zsoldos Z, Reid D, Simon A, Sadjad BS, Johnson AP. eHiTS: an innovative
approach to the docking and scoring function problems. Curr Protein Pept
Sci 2006;7:421–35.
[29] Zsoldos Z, Reid D, Simon A, Sadjad SB, Johnson AP. eHiTS: a new fast, exhaustive
flexible ligand docking system. J Mol Graph Model 2007;26:198–212.
[30] O’Boyle NM, Morley C, Hutchison GR. Pybel: a Python wrapper for the Open-
Babel cheminformatics toolkit. Chem Cent J 2008;2:5.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bcp.2011.10.004.
[31] Ordog R. PyDeT, a PyMOL plug-in for visualizing geometric concepts around
proteins. Bioinformation 2008;2:346–7.
[32] Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, et al.
UCSF Chimera—a visualization system for exploratory research and analysis. J
Comput Chem 2004;25:1605–12.
[33] Palmeira A, Rodrigues F, Sousa E, Pinto M, Helena Vasconcelos M, Fernandes
MX. New uses for old drugs: pharmacophore-based screening for the discov-
ery of P-glycoprotein inhibitors. Chem Biol Drug Des 2011;78(1):57–72.
[34] Promega. Technical Bulletin—Pgp-GloTM Assay Systems. https://promega.es/
tbs/tb341/tb341.pdf.
[35] Wang RB, Kuo CL, Lien LL, Lien EJ. Structure-activity relationship: analyses of
p-glycoprotein substrates and inhibitors. J Clin Pharm Ther 2003;28:203–28.
[36] Kostakis I, Ghirtis K, Pouli N, Marakos P, Skaltsounis AL, Leonce S, et al.
Synthesis and cytotoxic activity of 2-dialkylaminoethylamino substituted
xanthenone and thioxanthenone derivatives. Farmaco 2000;55:455–60.
[37] Kolokythas G, Kostakis IK, Pouli N, Marakos P, Kletsas D, Pratsinis H. Synthesis
and cytotoxic activity of some new azapyranoxanthenone aminoderivatives.
Bioorg Med Chem 2003;11:4591–8.
[38] Kolokythas G, Kostakis IK, Pouli N, Marakos P, Skaltsounis AL, Pratsinis H.
Design and synthesis of some new pyranoxanthenone aminoderivatives with
cytotoxic activity. Bioorg Med Chem Lett 2002;12:1443–6.
[39] Patchett AA, Nargund RP. Privileged structures—an update. In: Annual reports
in medicinal chemistry. Academic Press; 2000. p. 289–298, Chapter 26.
[40] DeSimone RW, Currie KS, Mitchell SA, Darrow JW, Pippin DA. Privileged
structures: applications in drug discovery. Comb Chem High Throughput
Screen 2004;7:473–94.
[41] Serajuddin AT. Salt formation to improve drug solubility. Adv Drug Deliv Rev
2007;59:603–16.
[42] Lipinski CA. Chris Lipinski discusses life and chemistry after the Rule of Five.
Drug Discov Today 2003;8:12–6.
References
[1] Higgins CF. Multiple molecular mechanisms for multidrug resistance trans-
porters. Nature 2007;446:749–57.
[2] Teodori E, Dei S, Scapecchi S, Gualtieri F. The medicinal chemistry of multidrug
resistance (MDR) reversing drugs. Farmaco 2002;57:385–415.
[3] Higgins CF. The multidrug resistance P-glycoprotein. Curr Opin Cell Biol
1993;5:684–7.
[4] Lage H. MDR1/P-glycoprotein (ABCB1) as target for RNA interference-mediat-
ed reversal of multidrug resistance. Curr Drug Targets 2006;7:813–21.
[5] Clarke R, Leonessa F, Trock B. Multidrug resistance/P-glycoprotein and breast
cancer: review and meta-analysis. Semin Oncol 2005;32:S9–15.
[6] Baguley BC. Novel strategies for overcoming multidrug resistance in cancer.
BioDrugs 2002;16:97–103.
[7] Takara K, Sakaeda T, Okumura K. An update on overcoming MDR1-mediated
multidrug resistance in cancer chemotherapy. Curr Pharm Des 2006;12:
273–86.
[8] Morphy R, Kay C, Rankovic Z. From magic bullets to designed multiple ligands.
Drug Discov Today 2004;9:641–51.
[9] Morphy R, Rankovic Z. Designing multiple ligands—medicinal chemistry strat-
egies and challenges. Curr Pharm Des 2009;15:587–600.
[10] Zhan P, Liu X. Designed multiple ligands: an emerging anti-HIV drug discovery
paradigm. Curr Pharm Des 2009;15:1893–917.
[11] Cavalli A, Bolognesi ML, Minarini A, Rosini M, Tumiatti V, Recanatini M, et al.
Multi-target-directed ligands to combat neurodegenerative diseases. J Med
Chem 2008;51:347–72.
[12] Youdim MB, Buccafusco JJ. Multi-functional drugs for various CNS targets in
the treatment of neurodegenerative disorders. Trends Pharmacol Sci 2005;26:
27–35.
[13] Hu C, Xu D, Du W, Qian S, Wang L, Lou J, et al. Novel 4 beta-anilino-
podophyllotoxin derivatives: design synthesis and biological evaluation as
potent DNA-topoisomerase II poisons and anti-MDR agents. Mol Biosyst
2010;6:410–20.
[14] Corbett TH, Panchapor C, Polin L, Lowichik N, Pugh S, White K, et al. Preclinical
efficacy of thioxanthone SR 271425 against transplanted solid tumors of
mouse and human origin. Invest New Drugs 1999;17:17–27.
[15] Stevenson JP, DeMaria D, Reilly D, Purvis JD, Graham MA, Lockwood G, et al.
Phase I/pharmacokinetic trial of the novel thioxanthone SR 233377
(WIN33377) on a 5-day schedule. Cancer Chemother Pharmacol 1999;44:
228–34.
[43] Lang F, Zewge D, Houpis IN, Volante RP. Amination of aryl halides using copper
catalysis. Tetrahedron Lett 2001;42:3251–4.
[44] Consiglio G, Frenna V, Mezzina E, Pizzolato A, Spinelli A. Catalysis in aromatic
nucleophilic substitution, Part 12. Kinetics of the reactions of some 2-phe-
noxy- and 2-(p-nitrophenoxy)-3-nitro-5-X-thiophenes with benzylamine and
N-benzylmethylamine in benzene. J Chem Soc Perkin Trans 1998;2:325–33.
[45] Emokpae T, Uwakwe P, Hirst J. The effect of ortho substituents on the
mechanism of aromatic nucleophilic substitution reactions in dipolar aprotic
solvents. J Chem Soc Perkin Trans 1993;2:125–32.
[46] Tao C, Li J, Fu Y, Liua L, Guo Q. Copper-catalyzed synthesis of primary
arylamines from aryl halides and 2,2,2-trifluoroacetamide. Tetrahedron Lett
2008;49:70–5.
[47] Caddick S, Fitzmaurice R. Microwave enhanced synthesis. Tetrahedron 2009;65:
3325–55.
[48] Kidwai M, Sapra P, Dave B. A facile method for nucleophilic aromatic substi-
tution of cyclic amine. Synth Commun 2000;30:4479–88.
[16] Tchamo DN, Dijoux-Franca MG, Mariotte AM, Tsamo E, Daskiewicz JB, Bayet C,
et al. Prenylated xanthones as potential P-glycoprotein modulators. Bioorg
Med Chem Lett 2000;10:1343–5.
[17] Pica-Mattoccia L, Cioli D. Studies on the mode of action of oxamniquine and
related schistosomicidal drugs. Am J Trop Med Hyg 1985;34:112–8.
[18] Rosi D, Peruzzotti G, Dennis EW, Berberian DA, Freele H, Tullar BF, et al.
Hycanthone, a new active metabolite of lucanthone. J Med Chem 1967;10:
867–76.
[49] Salmoria G, Dall’Oglio E, Zucco C. Aromatic nucleophilic substitutions under
microwave irradiation. Tetrahedron Lett 1998;39:2471–4.
[50] Prim D, Kirsch G. Convenient amination of weakly activated thiophenes, furans
and selenophenes in aqueous media. Tetrahedron 1999;55:6511–26.

68
A. Palmeira et al. / Biochemical Pharmacology 83 (2012) 57–68
[51] Cherng Y. Efficient nucleophilic substitution reaction of aryl halides with
amino acids under focused microwave irradiation. Tetrahedron 2000;56:
8287–9.
[52] Weissman S, Zewge D. Recent advances in ether dealkylation. Tetrahedron
2005;61:7833–63.
[53] Tatu CA, Paunescu V, Stanescu DI, Schneider F. P-glycoprotein mediated
multidrug resistance assessment by flow-cytometry in malignant hemopa-
thies. Anticancer Res 1997;17:2583–6.
[54] Wang EJ, Casciano CN, Clement RP, Johnson WW. In vitro flow cytometry
method to quantitatively assess inhibitors of P-glycoprotein. Drug Metab
Dispos 2000;28:522–8.
[61] Kim RB. Drugs as P-glycoprotein substrates, inhibitors, and inducers. Drug
Metab Rev 2002;34:47–54.
[62] Garrigos M, Mir LM, Orlowski S. Competitive and non-competitive inhibition
of the multidrug-resistance-associated P-glycoprotein ATPase—further exper-
imental evidence for a multisite model. Eur J Biochem 1997;244:664–73.
[63] Litman T, Zeuthen T, Skovsgaard T, Stein WD. Competitive, non-competitive
and cooperative interactions between substrates of P-glycoprotein as mea-
sured by its ATPase activity. Biochim Biophys Acta 1997;1361:169–76.
[64] Belli S, Elsener PM, Wunderli-Allenspach H, Kramer SD. Cholesterol-mediated
activation of P-glycoprotein: distinct effects on basal and drug-induced
ATPase activities. J Pharm Sci 2009;98:1905–18.
[55] Akazawa Y, Kawaguchi H, Funahashi M, Watanabe Y, Yamaoka K, Hashida M,
et al. Effect of interferons on P-glycoprotein-mediated rhodamine-123 efflux
in cultured rat hepatocytes. J Pharm Sci 2002;91:2110–5.
[56] Coburger C, Wollmann J, Baumert C, Krug M, Molnar J, Lage H, et al. Novel
insight in structure-activity relationship and bioanalysis of P-glycoprotein
targeting highly potent tetrakishydroxymethyl substituted 3,9-diazatetraas-
teranes. J Med Chem 2008;51:5871–4.
[57] Ludescher C, Thaler J, Drach D, Drach J, Spitaler M, Gattringer C, et al. Detection
of activity of P-glycoprotein in human tumour samples using rhodamine 123.
Br J Haematol 1992;82:161–8.
[58] Petriz J, O‘Connor JE, Carmona M, Garcia-Lopez J. Is rhodamine 123 an
appropriate fluorescent probe to assess P-glycoprotein mediated multidrug
resistance in vinblastine-resistant CHO cells? Anal Cell Pathol 1997;14:
129–40.
[59] Denis-Gay M, Petit JM, Ratinaud MH. Rhodamine 123: is it an appropriate dye
to study P-glycoprotein activity in adriamycin-resistant K562 cells? Antican-
cer Res 1995;15:121–6.
[65] Urbatsch IL, al-Shawi MK, Senior AE. Characterization of the ATPase activity of
purified Chinese hamster P-glycoprotein. Biochemistry 1994;33:7069–76.
[66] Dayan G, Jault JM, Baubichon-Cortay H, Baggetto LG, Renoir JM, Baulieu EE,
et al. Binding of steroid modulators to recombinant cytosolic domain from
mouse P-glycoprotein in close proximity to the ATP site. Biochemistry
1997;36:15208–15.
[67] Maki N, Hafkemeyer P, Dey S. Allosteric modulation of human P-glycoprotein.
Inhibition of transport by preventing substrate translocation and dissociation.
J Biol Chem 2003;278:18132–9.
[68] Martin C, Berridge G, Higgins CF, Callaghan R. The multi-drug resistance
reversal agent SR 33557 and modulation of vinca alkaloid binding to P-
glycoprotein by an allosteric interaction. Br J Pharmacol 1997;122:765–71.
[69] Eytan GD, Regev R, Oren G, Assaraf YG. The role of passive transbilayer drug
movement in multidrug resistance and its modulation. J Biol Chem 1996;271:
12897–902.
[70] Dongping M. Identify P-glycoprotein substrates and inhibitors with the rapid,
HTS Pgp-GloTM assay system. Promega Notes; 2007.
[60] Slapak CA, Mizunuma N, Kufe DW. Expression of the multidrug resistance
associated protein and P-glycoprotein in doxorubicin-selected human mye-
loid leukemia cells. Blood 1994;84:3113–21.
[71] Bailly C, Waring MJ. Preferential intercalation at AT sequences in DNA by
lucanthone, hycanthone, and indazole analogs. A footprinting study. Biochem-
istry 1993;32:5985–93.