Genomic action of permanently charged tamoxifen derivatives via estrogen receptor-α

Article abstract of DOI:10.1016/j.bmc.2010.06.039

Tamoxifen is a selective estrogen receptor modulator widely used in oncology and reproductive endocrinology. In order to decrease its non-desirable effects and elucidate mechanisms of action, permanently charged tamoxifen derivatives (PCTDs) have been rep

Full text of DOI:10.1016/j.bmc.2010.06.039

Bioorganic & Medicinal Chemistry 18 (2010) 5593–5601
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Genomic action of permanently charged tamoxifen derivatives via
estrogen receptor-
a
Claudia Rivera-Guevara a, Víctor Pérez-Alvarez a, Rocío García-Becerra b, David Ordaz-Rosado b,
Martha Sonia Morales-Ríos c, Elizabeth Hernández-Gallegos a, Austin J. Cooney d,
María Elena Bravo-Gómez e, Fernando Larrea b, Javier Camacho a,
*
a Department of Pharmacology, Centro de Investigación y de Estudios Avanzados, Avenida Instituto Politécnico Nacional 2508, México DF 07360, Mexico
b Departamento de Biología de la Reproducción, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Vasco de Quiroga 15, Tlalpan, México DF 14000, Mexico
c Departamento de Química, Centro de Investigación y de Estudios Avanzados, Avenida Instituto Politécnico Nacional 2508, México DF 07360, Mexico
d Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX 77030, USA
e Departamento de Química Inorgánica y Nuclear, Facultad de Química, Universidad Nacional Autónoma de México, Ciudad Universitaria, Coyoacán, México DF 04510, Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
Tamoxifen is a selective estrogen receptor modulator widely used in oncology and reproductive endocri-
nology. In order to decrease its non-desirable effects and elucidate mechanisms of action, permanently
charged tamoxifen derivatives (PCTDs) have been reported. Whether PCTDs have genomic effects
remains controversial. Since the clinical relevance of tamoxifen, the necessity to have new anticancer
drugs, and in order to gain insights into the mechanisms of action of PCTDs, we obtained six quaternary
ammonium salts derived from tamoxifen including three new compounds. We characterized them by
nuclear magnetic resonance, X-ray diffraction, electron microscopy, and/or high performance liquid chro-
matography, and detected them in cell lysates by liquid chromatography coupled to mass spectrometry.
Received 1 April 2010
Revised 10 June 2010
Accepted 14 June 2010
Available online 20 June 2010
Keywords:
Tamoxifen
Tamoxifen derivatives
Estrogen receptor
Breast cancer
We evaluated their binding to estrogen receptor-
iated by ER (gene reporter assays), and the proliferation of cancer cells (MCF-7 and cells from a cervical
cancer primary culture). Structural studies demonstrated the expected identity of the molecules. All
PCTDs did bind to ER , one of them induced ER -mediated transcription while two others inhibited such
a (ERa, their effect on the transcriptional activity med-
a
Genomic action
a
a
genomic action. Accordingly, PCTDs were detected in cell lysates. PCTDs inhibited cell proliferation, note-
worthy, two of them displayed higher inhibition than tamoxifen. Structure–activity analysis suggests that
PCTDs permanent positive charge and the length of the aliphatic chain might be associated to the biolog-
ical responses studied. We suggest genomic effects as a mechanism of action of PCTDs. The experimental
approaches here used could lead to a better design of new therapeutic molecules and help to elucidate
molecular mechanisms of new anticancer drugs.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
gets not related to ER have been suggested. Some examples of such
different targets/mechanisms include inhibition of natural killer
Tamoxifen (IUPAC name 2-[4-[(Z)-1,2-diphenylbut-1-enyl]-
phenoxy]-N,N-dimethylethanamine), is a triphenylbutylene deriva-
tive, widely used for the treatment and prevention of estrogen-
dependent breast cancer.1 It is a selective estrogen receptor
modulator (SERM), displaying estrogen receptor (ER) agonist and
antagonist activities.2 Since tamoxifen has also effects in patients
with negativeER breastcancer,3 other mechanisms of action and tar-
cells activity,4 inhibition of protein kinase C,5 changes in insulin-like
growth factor levels,6 anti-angiogenic activity,7 as well as several ef-
fects on different ion channels.8–18
Several groups have made chemical modifications to the mole-
cule, obtaining permanently charged tamoxifen derivatives
(PCTDs, including tamoxifen ethyl bromide, tamoxifen methyl io-
dide, and tamoxifen butyl bromide), expected to not to cross the
cell membrane.8,13,14,19–21 It was proposed that the permanent
charge would eliminate the intracellular effects of tamoxifen,
remaining only the extracellular actions of PCTDs. Nevertheless,
other studies have suggested that PCTDs are able to enter the cells
without crossing the blood–brain barrier, in such a manner that
these compounds might retain their action in peripheral tissues
and organs, but reducing the side-effects in the central nervous
system.22–26
* Corresponding author. Tel.: +52 55 57473800x5416; fax: +52 55 53430106.
E-mail addresses: jueshalom@gmail.com (C. Rivera-Guevara), vperez@cinvestav.
mx (V. Pérez-Alvarez), rocangeles@hotmail.com (R. García-Becerra), davo21_05@
hotmail.com (D. Ordaz-Rosado), smorales@cinvestav.mx (M.S. Morales-Ríos),
lishergallegos@yahoo.com (E. Hernández-Gallegos), acooney@bcm.edu (A.J.
Cooney), ma_elenab@yahoo.com (M.E. Bravo-Gómez), fernando.larreag@quetzal.
innsz.mx (F. Larrea), fcamacho@cinvestav.mx (J. Camacho).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.06.039

5594
C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
Thus, since the relevance of tamoxifen in oncology, the neces-
placements were measured in ppm (d) relative to internal
tetramethylsilane.
sity to have new anticancer drugs, and in order to gain insights into
the mechanisms of action of PCTDs, we obtained an analogue series
of six quaternary ammonium salts derived from tamoxifen includ-
ing three not previously reported, characterized their molecular
structure, detected them in cell lysates and studied their effect
2.2.1. Structural data from NMR
Tamoxifen propyl bromide: 1H NMR (DMSO):
d 0.82 (t,
J = 7.33 Hz, 3H, CH3CH2C), 0.84 (t, J = 7.17 Hz, 3H, CH3CH2 CH2N),
2.39 (q, J = 7.47 Hz, 2H, CH3CH2), 3.05 [s, 6H, N(CH3)2], 3.25 (q,
J = 7.32, 2H, NCH2CH2CH3), 3.66 (t, J = 4.84 Hz, 2H, CH2CH2N), 4.3
(at, J = 4.1 Hz, 2H, OCH2CH2), 6.66 (d, J = 8.9 Hz, 2H, C6H4OCH2,
meta H), 6.73 (d, J = 8.74 Hz, 2H, C6H4OCH2, ortho H), 7.1–7.4 (m,
10H, 2Ph).
on cell proliferation and transcriptional activity mediated by ERa.
2. Materials and methods
2.1. Chemical modifications
Tamoxifen isopropyl bromide: 1H NMR (DMSO): d 0.83 (t,
J = 7.4 Hz, 3H, CH3CH2C), 1.28 [d, J = 6.5 Hz, 6H, C(CH3)2], 2.37 (q,
J = 7.4 Hz, 2H, CH3CH2), 3.0 [s, 6H, N(CH3)2], 3.69 (t, J = 4.89 Hz,
2H, CH2CH2N), 3.77 (m, J = 6.5 Hz, 1H, NCHCH3), 4.29 (at,
J = 4.73 Hz, 2H, OCH2CH2), 6.67 (d, J = 8.8 Hz, 2H, C6H4OCH2, meta
H), 6.77 (d, J = 9 Hz, 2H, C6H4OCH2, ortho H), 7.1–7.4 (m, 10H,
2Ph). The rest of the compounds have been already reported.
Melting points (mp) were determined using an electro thermal
J.T.R.-TEMSA melting point apparatus, values were not corrected
(Table 1). Chemical purity was assessed using high performance li-
quid chromatography tandem to mass spectrometry (HPLC–MS).
X-ray diffraction from electron microscopy in a Bruker-Nonius
Tamoxifen free base, alkyl halides and inorganic salts, were pur-
chased from Sigma–Aldrich Chemical Company, (San Louis, MO,
USA); methanol and acetonitrile HPLC grade, used for LC/MS sys-
tem, were obtained from J. T. Baker (NJ, USA). The analogue series
of permanently charged tamoxifen derivatives (PCTDs, Table 1)
was obtained as previously reported26 as follows: (A) Tamoxifen
methyl iodide, 0.5 g of tamoxifen were added to 15 ml of methyl io-
dide and stirred at 0 °C for 10 min, the white precipitate was fil-
tered and recrystallized with methanol; (B) Tamoxifen methyl
bromide, an anion interchange reaction (bromide instead of iodide)
was performed, and we obtained tamoxifen methyl bromide deriv-
ative. The presence of the bromide anion was corroborated by elec-
tron microscopy (with a JEOL JSM-6300 microscope, USA) and the
energy dispersion spectrometry X-ray technique (X-ray detector
for energy dispersion Noral Voyager II, software 1100/1110). (C–E)
Tamoxifen ethyl bromide, tamoxifen propyl bromide, and tamoxifen
butyl bromide: 1 g of tamoxifen was added to 20 or 25 ml of the
alkyl halide (ethyl bromide, propyl bromide and butyl bromide,
respectively) and stirred at room temperature for 5 h. The white
precipitate was filtered and recrystallized with methanol. (F)
Tamoxifen Isopropyl bromide was obtained by refluxing 1 g of
tamoxifen in 20 ml of isopropyl bromide for 72 h at 55 °C. The
white precipitate was recrystallized with a mixture of methanol–
ethyl acetate. To our knowledge tamoxifen isopropyl bromide,
tamoxifen methyl bromide and tamoxifen propyl bromide are
new compounds not previously reported.
CAD4 diffractometer (USA) with Cu-K radiation was used in order
a
to determine the crystal structure of tamoxifen methyl bromide.
SHELXS96 and SHELX97 software (University of Göttingen,
Germany) were used in order to solve and refine these structures.
2.3. Binding and transcriptional studies
2.3.1. Estrogen receptor binding experiments
[2,4,6,7-3H] estradiol ([3H]-E2), sp. act. 72 Ci/mmol was pur-
chased from NEN Research Products (Boston, MA, USA), non-radio-
active estradiol was supplied by Sigma–Aldrich, Inc. (St. Louis, MO,
USA), and the anti-estrogen ICI 182, 780 from Zeneca Pharmaceu-
ticals (Wilmington, DE). Cell culture media and reagents were pur-
chased from Invitrogen life technologies (Carlsbad, CA, USA). Fetal
bovine serum (FBS) was supplied by Hyclone Laboratories, Inc.
(Logan, UT, USA). All reagents and solvents used were of analytical
grade. We used an expression vector for human ER
hER ) containing the coding sequence of the ER and the estrogen
responsive reporter plasmid (ERE-E1b-Luc). The relative receptor
a (pCMV5-
2.2. Structural studies
a
a
The identity of the compounds was determined by 1H and 13C
nuclear magnetic resonance (NMR) on Varian Mercury-300 (USA)
spectrometers at 300 and 75.4 MHz, respectively, chemical dis-
binding affinities for ER
a were determined as previously de-
scribed.27,28 Briefly, cervical cancer HeLa cells (obtained from
American Type Culture Collection, Manassas, VA) were transfected
with 5 lg of ERa expression vector (pCMV5-hERa) using PolyFect
(QIAGEN Inc. Valencia, CA, USA) according to the manufacturer’s
protocol. Forty eight hours later, cells were harvested and the cyto-
solic fraction was obtained in TEDLM buffer (20 mM Tris–HCl, pH
7.4, 1.5 mM EDTA, 0.25 mM dithiothreitol, 10 mg/ml leupeptine,
and 10 mM sodium molibdate). The cytosolic fractions (0.2 mg
protein/ml) were incubated with 1 nM [3H]-E2 and increasing con-
centrations (0.005–5000 nM) of radioinert estradiol, tamoxifen, or
the newly synthetized PCTDs for 18 h at 4 °C. Free steroid was sep-
arated from receptor-bound steroid by addition of dextran-coated
charcoal suspension (250 mg Norit-A and 25 mg Dextran T-70) in
TEDLM buffer and incubated for 10 min at 4 °C. Following centrifu-
gation (800 g at 4 °C for 15 min), bound [3H]-E2 was quantified by
liquid scintillation counting (Betamax (INC, Biomedicals Inc, CA,
USA). The results are expressed as the relative binding affinities
(RBA) and the inhibition constants (Ki) of steroid competitors, as
previously described.27,28
Table 1
Aliphatic chains substituted, melting points and yield of the permanently charged
tamoxifen derivatives
CH3
O
N+
R
H3C
Br- or I-
CH3
Compound
–R
Melting point (°C) Yield (%)
T. methyl iodide
T. methyl bromide
T. ethyl bromide
T. propyl bromide
T. isopropyl bromide –CH–(CH3)2
T. butyl bromide
–CH3
–CH3
–CH2–CH3
–CH2–CH2–CH3
248–250
168–170
158–160
196–198
198–200
94.9
80.0
99.1
90.1
37.2
94.8
2.3.2. Transcriptional activity (luciferase assays)
HeLa cells were cultured in Dulbecco’s Modified Eagle’s Med-
ium (DMEM-HG) without phenol red, supplemented with 5%
–CH2–CH2–CH2–CH3 176–178
T. = Tamoxifen.

C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
5595
stripped fetal bovine serum, 100 U/ml penicillin and 100 mg/ml
Cells were seeded in 96 well plates at 37 °C in a humidified atmo-
streptomycin; and incubated in 5% CO2 at 37 °C. The next day cells
sphere of 5% CO2 and incubated until 60% confluence of more was
were co-transfected with 25 ng pCMV5-hERa, 20 ng TK-Renilla
reached. Cells were incubated during 96 h with PCTDs (6 lM)
Promega Corp (Madison, WI, USA) and 1 g of a luciferase reporter
l
which were dissolved in methanol as vehicle (final concentration,
0.1 v/v%). Cell proliferation was assayed by the colorimetric meth-
od based on the conversion of the tetrazolium salts to formazan
crystals by dehydrogenase activity in active mitochondria (MTT
Cell proliferation Kit I, Boehringer Mannheim GmbH, Germany).
MTT (0.5 mg/ml) was added to the cell culture 4 h before complet-
ing the PCTDs incubation time. Optical density data were obtained
from the resulting colored solution with a microplate photometer
(Sunrise Touchscreen, Tecan, USA).
vector (ERE-E1b-Luc) The transfections were performed in tripli-
cate using PolyFect (Qiagen Inc. Valencia, CA, USA), according to
the manufacturer’s protocol. The plates were incubated for 48 h
at 37 °C in 5% CO2. After incubation, media were replaced with
DMEM-HG containing the compounds of interest and cells were
incubated for 48 h. Ethanol alone was used as vehicle. Luciferase
activity was then measured with the Dual Luciferase Kit Promega
Corp (Madison,WI, USA) according to the manufacturer’s recom-
mendations. Firefly luciferase activity was normalized to the inter-
nal transfection control provided by the Renilla luciferase activity.
2.6. Statistical analysis
2.4. Detection of PCTDs in cell lysates
Experiments were repeated at least three times. Comparisons
were carried out by one-way analysis of variance (ANOVA) fol-
lowed by Dunnett’s test, or a two tailed t-test, differences were
considered statistically significant when P <0.001 or 0.005, respec-
tively. Analysis was made with the software Sigma Stat version 3.5
(U.S.A. or Germany).
2.4.1. Preparation of lysates
HeLa cells were cultured in DMEM without phenol red, and sup-
plemented with 10% fetal calf serum and penicillin/streptomycin
(1%). Next day, cells were incubated with the corresponding com-
pound for 48 h. After treatment, the cells were washed 10 times
with PBS, trypsinized—and centrifuged 15 min at 3000 rpm. Then,
cells were resuspended with PBS, centrifuged (three times), and
sonicated in distilled water. Finally, a mixture of ethanol–acetoni-
trile was added, and the sample was centrifuged 20 min at
15,000 rpm.
2.7. Structure–activity relationship
The association between different parameters was analyzed
using Origin Pro v. 7.0383 2002 (Sony Electronics, Inc.: Northamp-
ton, USA). Log P values were calculated with ALOGPS 2.1.30
2.4.2. LC/MS system
3. Results
Cell lysates were analyzed with a 1100 Rapid Resolution system
(Agilent Technologies USA) containing a binary pump and degas-
ser, a well plate auto sampler with thermostat, and a thermostat-
ted column compartment, coupled to an Agilent 6300 ion trap
quadrupole mass spectrometer system equipped with an electro
spray ionization (ESI) source. The separation was carried out with
3.1. Chemical modifications and identity
Six PCTDs were obtained as described in Section 2. Table 1
shows the aliphatic chain substituted, yield and the resulting melt-
ing point of each compound. We obtained the expected molecular
identity of each PCTD. For illustration purposes, structural analysis
of only the three new, not previously reported compounds is
shown. Figure 1A shows the X-ray elucidation of tamoxifen isopro-
pyl bromide displaying the bromide atom. X-ray energy dispersion
spectrometry shows the expected identity of tamoxifen methyl
bromide (Fig. 1B), and 1H nuclear magnetic resonance (NMR) anal-
ysis shows the identity of tamoxifen propyl bromide (Fig. 1C),
using DMSO as solvent.
a Waters C18 reverse-phase column, 5
analytical column with a mobile phase 85% (v/v) methanol in
l
m, and 4.6 mm ꢀ 150 mm
water (HPLC purity) at a flow rate of 1 ml/min, and injection vol-
ume of 25 ll at 30 °C of column temperature. All spectra were ac-
quired in the positive ion mode, over a mass range of m/z 300–550
at rate of one scan every 2 s. The nebulizer was set to 10–15 psi and
the nitrogen drying gas was set to a flow rate of 4 l/min; drying gas
temperature was maintained at 325 °C.
2.5. Proliferation studies
3.2. Presence of PCTDs in lysates from HeLa cells incubated with
the compounds
2.5.1. [3H]-thymidine incorporation
Cell proliferation assay was performed with human cervical
cancer cells obtained from previously established primary cul-
tures,29 which were grown in plates with DMEM with 10% FCS pen-
icillin/streptomycin (200 lg/ml). Cells were seeded in 24-well
plates, until 60% confluence of more was reached. Cells were
In order to reveal potential intracellular effects of PCTDs, a very
important issue to solve was to know if the different PCTDs could
be detected in lysates from cells incubated with the compounds.
High performance liquid chromatography coupled to mass spec-
trometry was used for detection. In order to have reliable evi-
dences about the presence and molecular identity of PCTDs,
several parameters were considered, namely, retention time value,
mass–charge ratio, and fragmentation pattern (this pattern must
be the same in a specific molecule, when fixing relative collision
energy is applied, 2.0 V in this case).
Figure 2A shows a cell lysate chromatogram from untreated
HeLa cells. No endogenous peaks that could interfere with the
detection of tamoxifen derivatives were observed. Next, we deter-
mined the different parameters of a tamoxifen butyl bromide stan-
dard solution. Figure 2B shows the retention time value, m/z ratio
and fragmentation pattern of tamoxifen butyl bromide standard
solution. Once we had these controls, HeLa cells were incubated
with the compound during 48 h and the same parameters were
washed and incubated during 24 or 48 h with PCTDs (5 nM or
2
l
M) which were dissolved in methanol as vehicle (final concen-
tration 0.1% v/v). Four hours before completing the treatment, 3H
thymidine (0.5
Ci/well) was added. 3H thymidine was eliminated
l
of culture medium and radioactivity was measured as disintegra-
tions per minute (dpm) with a Beckman Coulter LS 6500 multi pur-
pose scintillation counter (USA).
2.5.2. MTT assay
The human breast adenocarcinoma cell line MCF-7 was ob-
tained from American Type Culture Collection (Manassas, VA)
and cultured according to manufacturer’s instructions. Cells were
grown in RPMI-1640 culture medium with fetal calf serum (10%),
penicillin/streptomycin (1%) and non essential amino acids (1%).

5596
C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
Figure 1. Structural elucidation. (A) Crystal structure analysis by X-ray diffraction of tamoxifen isopropyl bromide (left panel). C, O, N, and Br atoms are indicated in the right
panel. (B) X-ray energy dispersion spectrometry of tamoxifen methyl bromide, showing the presence of the bromide atom, replacing the iodide atom. (C) Tamoxifen propyl
bromide structure elucidation by 1H nuclear magnetic resonance analysis showing the identity of the propyl group.
studied in the cell lysates to know whether the compound had en-
tered the cells. Retention time value, m/z ratio and fragmentation
pattern found in this cell lysate (Fig. 2C) strongly suggest the pres-
ence of tamoxifen butyl bromide inside the cells.
estrogen receptor.26 We selected only some PCTDs for these stud-
ies because PCTDs with substituted methyl differ only in the anion,
having the same cationic molecule. On the other hand, the assess-
ment of isopropyl bromide was aimed to observe the effect of the
branched chain on the affinity, instead of the effect of the linear
A summary of retention time and m/z values from the standard
solutions of all PCTDs as well as PCTDs detected in the cell lysates
are listed in Table 2. The values obtained for the standard solutions
using the compounds in the micromolar range, show high purity of
PCTDs and the absence of contamination with possible unmodified
tamoxifen. Besides, all of the parameters indicate that all the PCTDs
obtained enter the cells emphasizing that the compounds might
have relevant intracellular effects. HPLC–MS analysis showed the
expected m/z values and the absence of a tamoxifen signal in the
PCTDs (Fig. 3), indicating high purity of the compounds, which is
a very important issue in the study and interpretation of the bio-
logical assays.
chain. All of the PCTDs studied did bind to ER
that increasing concentrations of PCTDs decrease the binding of la-
beled estradiol to ER . Inhibition constant (Ki) and Relative Binding
a. Figure 4A shows
a
Affinity (RBA) values for each compound are given in Table 3.
Detection of PCTDs in cell lysates from treated cells and binding
to ER
evant intracellular activity. Thus we performed gene reporter as-
says in HeLa cells transfected with human ER . If the compounds
a strongly suggested that these compounds might have a rel-
a
had an intracellular activity at the transcriptional level, then the
activity of the reporter gene should be affected. Activity induced
by estradiol (10 nM) alone was normalized to 100% (Fig. 4B).
Tamoxifen and most of the PCTDs alone (100 nM) induced only a
modest activity very similar to that induced by either the vehicle
or progesterone (used as a negative control). However, tamoxifen
isopropyl bromide induced up to 50% of the activity (Fig. 4B), sug-
3.3. PCTDs bind to ERa and inhibit its transcriptional activity
Since tamoxifen exerts major antiproliferative effects through
its binding to ER , we wondered whether PCTDs maintained the
a
gesting its role as a new potential ERa-agonist. As expected, tran-
ability to bind to this receptor. Both tamoxifen methyl iodide and
scriptional activity induced by estradiol was decreased in the
tamoxifen ethyl bromide have been already reported to bind to
presence of either tamoxifen or the antiestrogenic ICI 182, 780

C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
5597
with the software ACD/Labs 11.0, Canada). Interestingly, both com-
pounds (1 M) also reduced significantly the transcriptional activ-
ity induced by estradiol (Fig. 4C). All of these results propose that
A
B
l
6
4
2
PCTDs might influence cell physiology by binding to ER
affecting transcriptional activity.
a and
Time (min)
m/z
4
8
12
16
325.2
1.0
0.5
3.4. PCTDs decrease proliferation of cancer cells
200
300
400
Finally, we wondered whether proliferation of cancer cells
known to express ER , might be affected by PCTDs. A breast cancer
a
cell line (MCF-7) and cells from a previously established cervical
cancer primary culture were used. The effect of PCTDs on cell pro-
liferation at different concentrations (5 nM, 2 lM, and/or 6 lM) as
well as different incubation times (24, 48 and 96 h) was tested.
These incubation conditions (concentration and treatment dura-
tion) were selected after some experiments with 1 nM or 1 lM
1.2
0.8
0.4
Time (min)
428.5
4
8
12
16
1.2
0.8
concentrations, in which no effect was observed. As shown in
Figure 5A, all of the compounds decreased breast cancer cell prolif-
eration (assessed by a metabolic activity assay) in comparison to
untreated cells (Vehicle). Interestingly, tamoxifen isopropyl bro-
mide and tamoxifen butyl bromide decreased cell proliferation
more than tamoxifen. Cervical cancer cells from a primary culture
previously characterized to express human papilloma virus onco-
0.4
m/z
200
200
300
300
400
500
300
100
genes, cytokeratins, and ERa 29,31
were also studied. Both tamoxi-
,
m/z
400
fen and PCTDs decreased thymidine incorporation only at the
lowest concentration (5 nM, Fig. 5B). The potential mechanisms
explaining the effect of PCTDs on cancer cell proliferation might
be related to the inhibition of the transcriptional activity mediated
C
8
5
by ERa.
2
Time (min)
16
428.5
4
8
12
1.0
3.5. Structure–activity relationship
0.5
0
600
400
200
In order to gain some insight into structure–activity relation-
ship, we analyzed the charge and length of the aliphatic chain of
PCTDs and their effect on cell proliferation and binding to ER
200
300
300
400
m/z
a
m/z
(Ki and RBA), finding interesting trends. Regarding metabolic activ-
ity (Fig. 6A), the values are grouped into two clusters when plotted
against calculated Log P (from 4-OH tamoxifen, data not shown).
The difference between these two groups is the positive charge
of the nitrogen of PCTDs but absent in tamoxifen. This charged
nitrogen, plus the length increase on the hydrocarbonated chain,
seems to enhance the activity on MCF-7 cells. However, when
these data are plotted against the hydrophobicity descriptor
the substituent on N ( N), the values display a linear relationship
with R = ꢁ0.90 (Fig. 6B). Since describes the hydrophobicity only
m/z
400
200
Figure 2. HPLC–ESI MS analysis of tamoxifen butyl bromide in cell lysates. (A) Cell
lysates from untreated cells shows no endogenous signal potentially interfering
with the detection of PCTDs. (B) Retention time, total ion chromatography (TIC), and
pattern of fragmentation of tamoxifen butyl bromide standard solution (200 nM,
compound diluted in methanol). (C) Same parameters as in (B) but from HeLa cell
p for
p
lysates incubated with tamoxifen butyl bromide (48 h, 2 lM).
p
for the substituent on N and not for the whole molecule, this leads
to propose that there exist a local effect on this N moiety, probably
associated with the interaction of the molecule with its receptor.
This suggests that local effects on specific molecular regions are
determinant for molecular interactions and subsequent receptor
conformational changes and biological responses. Analysis of RBA
showed a similar behavior (Fig. 6C and D). The combination of
charge on quaternary ammonium and the length increase on the
hydrocarbonated chain favors the affinity for ERa. These similar
data for these two biological effects strongly suggest that the de-
crease on metabolic activity is related to molecular interactions
with ERa; however, more data are required to fully demonstrate it.
Table 2
PCTDs retention times and m/z values from either standard solutions or treated HeLa
cell lysates (determined by HPLC tandem mass spectrometry analysis)
Compound
Standard solutions
HeLa Cell lysates
Retention
time (min)
[M+H]+ Retention
[M+H]+
m/z
time (min) m/z
Tamoxifen
Tamoxifen methyl
bromide
Tamoxifen ethyl
bromide
Tamoxifen propyl
bromide
Tamoxifen isopropyl 11–14
bromide
Tamoxifen butyl
bromide
8–10
5–8
372.3
386.3
9–12
5–10
372.3
386.3
8–10
8–12
400.5
414.5
414.5
428.5
8–12
400.5
414.5
414.5
428.5
9–13
10–13
10–15
4. Discussion
12–16
Tamoxifen is a selective estrogen receptor modulator widely
used in the treatment of estrogen-dependent breast cancer. In or-
der to decrease its non-desirable effects and to elucidate mecha-
nisms of actions, permanently charged tamoxifen derivatives
including tamoxifen ethyl bromide, tamoxifen methyl iodide and
(Fig. 4C). Then we tested the effects of the two compounds with
lower theoretical lipophilicity than tamoxifen, namely, tamoxifen
methyl bromide and tamoxifen methyl iodide (Log P calculated

5598
C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
Figure 3. HPLC–MS analysis of PCTDs. The analysis shows the m/z relationships for PCTDs in standard solution. (A) Tamoxifen, (B) tam methyl iodide, (C) tam ethyl bromide,
(D) tam butyl bromide, (E) tam isopropyl bromide. The expected m/z values and the absence of a tamoxifen signal in the PCTDs indicate high purity of the compounds. Such
purity is also observed for Tam propyl bromide in a cell lysate (F).
tamoxifen butyl bromide, have been produced by several
groups.8,13,14,19–21 Whether or not PCTDs display intracellular and
genomic actions remains controversial. Some authors assume that
the charged molecules would not enter the cells, which would
eliminate the potential genomic effects of PCTDs via interaction
with intracellular estrogen receptors. In this manner, either extra-
cellular or non-genomic mechanisms of actions of PCTDs have
been proposed to explain the effect of different PCTDs on vol-
ume-activated chloride currents in HeLa cells, voltage-gated cat-
ionic channels in rat embryonic hypothalamic neurons, apoptosis
in acutely damaged mammary epithelial cells, smooth muscle cal-
cium-activated large-conductance potassium channels, and inhibi-
tion of intestinal and uterine muscles.8,13,14,19–21 It has been also
suggested that tamoxifen methyl iodide has a poor penetration
(5%) lacking in vitro activity.26 Nevertheless in vivo studies suggest
that tamoxifen methyl iodide displays intracellular actions,
authors claim that the discrepancies between in vitro and in vivo
results may be related to the dynamics of distribution of the com-
pound since in whole animal studies the PCTD was present for
longer times.22,23
Here we demonstrate that PCTDs can be detected in cell lysates
from HeLa cells incubated with the compounds, and that transcrip-
tional activity mediated by ERa is affected by PCTDs. By using mass
spectrometry and considering retention times, m/z ratio and frag-
mentation patterns, we were able to detect and identify all PCTDs
in HeLa cell lysates. Such parameters were evaluated in six PCTDs
at different incubation times but detectable amounts were
achieved only after 48 h of incubation. This observation of the re-
quired time to detect PCTDs in cell lysates is in accordance with
the in vivo findings which suggest that in whole animal studies
the PCTD was present for longer times.22,23 In the cell lysate anal-
ysis, broad peaks were observed in the retention times, which can
be explained by the low concentration of the compound present in
the lysates, or by additional complex matrix properties from the
cell lysates. The differences observed in the fragmentation patterns
of Figure 2B and C may be due to the presence of several

C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
5599
MCF-7 6 µM
A
100
A
120
100
80
*
*
*
80
60
40
20
,
*
,
*
*
**
**
60
40
20
0.01 0.1
1.0
10.0 102 103 104
0
Competitor (nM)
B
C
B
Cervical cancer primary culture
150
(2 µM)
100
50
0
200
Con Veh Met I Et B Iso B But B Tmx
150
100
50
0
(5 nM)
Con Veh
Et B Iso B But B Tmx
Met I
Figure 5. Effect of PCTDs on cancer cell proliferation. (A) All PCTDs significantly
decreased proliferation of MCF-7 breast cancer cells (6 M, 96 h treatment) in
l
comparison to untreated cells (control). Besides tamoxifen isopropyl bromide and
tamoxifen butyl bromide show a stronger antiproliferative effect than tamoxifen.
(B) 3H thymidine incorporation in cervical cancer cells from a primary culture was
not affected by 2 lM PCTDs however, a lower concentration (5 nM) significantly
Figure 4. Binding and transcriptional activity of PCTDs. (A) All of the compounds
shifted the binding of labeled estradiol in a similar manner to that of tamoxifen.
(B) Transcriptional activity of ERa (analyzed by gene reporter assays) was induced by
decreased thymidine incorporation. Con = control, Veh = vehicle, Met B = tamoxifen
methyl bromide, Met I = tamoxifen methyl iodide, Et B = tamoxifen ethyl bromide,
Prop B = tamoxifen propyl bromide, Iso B = tamoxifen isopropyl bromide, But
B = tamoxifen butyl bromide, Tmx = tamoxifen. Columns, mean; bars, s.e.m. *P
<0.001 in comparison to vehicle. **P <0.001 in comparison to tamoxifen.
tamoxifen isopropyl bromide in almost 50% of the activity induced by estradiol
(normalized to 100%). Neither tamoxifen nor other PCTDs displayed such agonist
property. (C) Gene reporter activity induced by estradiol was significantly decreased
in the presence of tamoxifen, the antiestrogenic ICI 182, 780 or PCTDs. V = vehicle,
E2 = estradiol, Tmx = tamoxifen, Met I = tamoxifen methyl iodide, Met B = tamoxifen
methyl bromide, Et B = tamoxifen ethyl bromide, But B = tamoxifen butyl bromide, Iso
B = tamoxifen isopropyl bromide, ICI = ICI 182, 780, P4 = progesterone. Columns,
mean; bars, s.d. *P <0.005.
might display high passive diffusion provided that the charge can
be spread by several aromatic rings. Actually, they suggest that
there is not an apparent association between partition coefficient
values (Log P) and permeation properties.
Binding studies demonstrated that PCTDs affinity to ER
a is
Table 3
similar to that of tamoxifen, as previously reported for tamoxifen
Inhibition constant (Ki) and relative binding affinities (RBA) of PCTDs to ERa (obtained
from binding experiments)
methyl iodide and ethyl bromide.26 Among PCTDs, methyl
tamoxifen displayed the lower affinity to ER
tamoxifen showed the higher. The structure–activity analysis
(Fig. 6) suggests that PCTDs permanent charge and the incre-
ment on the hydrocarbonated chain, favors the affinity for ER
Besides, the decrease on metabolic activity of MCF-7 cells might
be associated to molecular interactions with ER . It would be
a while butyl
Compound
Ki (nM)
RBA (%)
E2
0.16
31.19
46.29
13.66
25.43
13.01
100
Tamoxifen
0.51
0.35
1.17
0.63
1.23
a.
T. methyl iodide
T. ethyl bromide
T. isopropyl bromide
T. butyl bromide
a
very interesting to synthesize more PCTDs in order to perform
a complete QSAR analysis.
E
2 = estradiol, T = tamoxifen.
Certain key regions of tamoxifen are thought to be important
for its binding to ER and to block estrogen action,33 the amino eth-
oxy side chain of the a0 ring is critical for its antiestrogenic activ-
ity.34 These findings together with our results suggest that the
alkylaminoethoxy chain of PCTDs is an important region both for
the activity of the molecule as well as for the affinity to ERa. Of
course also the status of the receptor is important in the response
to tamoxifen. For example, it has been reported that tamoxifen can
components present in very low concentrations in the cell lysates,
making difficult to clearly appreciate fragments of the correspond-
ing PCTD.
Fisher et al.32 estimated permeation properties of several posi-
tively charged compounds (from parallel artificial membrane per-
meability assays) concluding that permanently charged molecules

5600
C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
Figure 6. Structure–activity relationship. (A) Partition coefficient calculated with ALOGPS 2.1 versus metabolic activity (%) of MCF-7 cells (from Fig. 4). (B) Hydrophobicity
descriptor
descriptor
p
p
for the substituent on N (
for the substituent on N (
p
p
N) versus metabolic activity (%) of MCF-7 cells. (C) Partition coefficient (calculated Log P) versus RBA (%) to ER
a. (D) Hydrophobicity
N) versus RBA (%) to ER
a.
switch from ER
lated by PKA.35
a
-antagonist to ER
a
-agonist if ER
a
is phosphory-
-medi-
tration is saturable. It would be also important to study the effect
of PCTDs on ERb.
Finally, we observed that PCTDs decrease proliferation of cancer
cells in a cell-type and concentration-dependent manner. PCTDs
were tested on the proliferation of MCF-7 at different concentrations
Potential genomic action of PCTDs was assessed by ER
a
ated transcription. Interestingly, tamoxifen isopropyl bromide
alone induced transcriptional activity, while tamoxifen methyl
bromide and tamoxifen methyl iodide significantly decreased the
gene reporter activity in the presence of estradiol. Then, substitu-
tions on the nitrogen of a branched chain, switch tamoxifen from
an antiestrogen to an estrogen agonist. These gene reporter assays
strongly suggest that PCTDs might influence cell physiology by
binding to ERa. Since the inhibition constants observed is in the
nanomolar range (Table 3), and compounds are tested at micromo-
lar concentrations, then even penetrations below 1% would be
needed in order to interfere with the genomic actions of ERa. This
scenario could be possible if relatively long incubation times are
considered. Actually, detection of PCTDs in cell lysates and effects
on transcriptional activity were only observed after incubating the
cells with the compounds during 48 h, in contrast to tamoxifen
which could be detected in cell lysates after 24 h of incubation
(data not shown). The required time to reach relevant concentra-
tions might indeed explain differences between the in vitro and
in vivo findings obtained by different researchers. It would be very
interesting to quantify the intracellular concentration of PCTDs
when incubating the cells at different extracellular concentrations
of the compounds and to know whether such intracellular concen-
(5 nM, 2
96 h). Significant effects were only observed with the treatment at
M and during 96 h. Noteworthy, the effect of propyl tamoxifen
lM, and 6 lM) and different incubation times (24, 48 and
6
l
and tamoxifen butyl bromide was higher in comparison to that of
tamoxifen alone. A higher anticancer activity in vivo was already ob-
served for tamoxifen methiodide (tamoxifen methyl iodide) in com-
parison to tamoxifen.22 Interestingly, PCTDs decreased proliferation
of cervical cancer cells only at the lowest concentration (5 nM). Dif-
ferential concentration effects of tamoxifen on cervical cancer cells
have been already reported. For example, tamoxifen treatment
(5 lM, six day exposure) resulted in 66–74% growth inhibition of
cervical cancer cell lines including HeLa cells;36 but in the cervical
cancer cell line SFR lacking ERs, low concentrations of tamoxifen
(1 ꢀ 10ꢁ9 or 1 ꢀ 10ꢁ11 M, five day treatment) stimulated HPV-16
gene expression and cell proliferation.37 Clinical studies showed
that 16% of cervical cancer biopsies from patients treated with
tamoxifen (given orally during 10 days) had a significant decrease
in the number of mitotic figures.38 In summary, the response of
cervical cancer cells to tamoxifen is highly variable depending on
different aspects including treatment conditions, differences in

C. Rivera-Guevara et al. / Bioorg. Med. Chem. 18 (2010) 5593–5601
3. Swain, S. M. J. Clin. Oncol. 2001, 19, 93.
5601
HPV copy number, different amounts of ERs and tamoxifen
concentration (probably tamoxifen has different targets at higher
concentrations avoiding its antiproliferative effects). Thus, it would
be very important to test the effects of PCTDs on cell proliferation of
different cervical cancer cells, both in cancer cell lines and primary
cultures from different patients. Whether or not such variable re-
sponse would be eliminated by PCTDs remains to be elucidated. It
will be also very important to discover additional targets of PCTDs.
Actually, the effect of PCTDs on cell proliferation might be also ex-
plained by the inhibition of some ion channels for example the ether
à-go-go related gene potassium channel (Kv11.x), which is known to
be inhibited by tamoxifen and to be an advantage for proliferation of
several cancer cells.39,18
4. Berry, J.; Green, B. J.; Matheson, D. S. Eur. J. Cancer Clin. Oncol. 1987, 23, 517.
5. Gundimeda, U.; Chen, Z.; Gopalakrishna, R. J. Biol. Chem. 1996, 271, 13504.
6. Pollak, M.; Costantino, J.; Polychronakos, C.; Blauer, S. A.; Guyda, H.; Redmund,
C.; Fisher, B.; Margolese, R. J. Natl. Cancer Inst. 1990, 82, 1693.
7. Gagliardi, A.; Collins, D. C. Cancer Res. 1993, 53, 533.
8. Sahebgharani, M.; Hardy, S. P.; Lloyd, A. W.; Hunter, A. C.; Allen, M. C. Cell.
Physiol. Biochem. 2001, 11, 99.
9. Allen, M. C.; Newland, C.; Valverde, M. A.; Hardy, S. P. Eur. J. Pharmacol. 1998,
354, 261.
10. Sartor, P.; Vancher, P.; Mollard, P.; Dufy, B. Endocrinology 1998, 123, 534.
11. Valverde, M. A.; Mintenig, G. M.; Sepúlveda, F. V. Pflüg Arch. 1993, 425,
552.
12. Zhang, J. J.; Jacob, T. J.; Valverde, M. A.; Hardy, S. P.; Mintenig, G. M.; Sepulveda,
F. V.; Gill, D. R.; Hyde, S. C.; Trezise, A. E. O.; Higgins, C. F. J. Clin. Invest. 1994, 94,
1690.
13. Hardy, S. P.; de Felipe, C.; Valverde, M. A. FEBS Lett. 1998, 434, 236.
14. Allen, M. C.; Gale, P. A.; Hunter, A. C.; Lloyd, A.; Hardy, S. P. Biochem. Biophys.
Acta Biomembr. 2000, 1509, 229.
15. Sha, Y.; Tashima, T.; Mochizuki, Y.; Toriumi, Y.; Adachi-Akahane, S.; Nonomura,
T.; Cheng, M. S.; Ohwada, T. Chem. Pharm. Bull. 2005, 53, 1372.
16. Bolanz, K. A.; Hediger, M. A.; Landowski, C. P. Mol. Cancer Ther. 2008, 7, 271.
17. Perez, G. J. J. Biol. Chem. 2005, 280, 21739.
Despite a quantitative study of the intracellular concentration
of PCTDs is necessary and the precise mechanism by which PCTDs
enter the cells remains elusive, here we show that PCTDs display
genomic action by affecting transcriptional activity of ER
a and
influence cancer cell proliferation. The values obtained (retention
time value, m/z ratio and fragmentation pattern) for the standard
solutions using the compounds in the micromolar range, and the
agonistic effects of two PCTDs alone on the transcriptional activity
show that the compounds are not contaminated with possible
unmodified tamoxifen. To our knowledge, three of these PCTDs,
namely, tamoxifen isopropyl bromide, tamoxifen methyl bromide
and tamoxifen propyl bromide, are new compounds not previously
reported.
18. Thomas, D.; Gut, B.; Karsai, S.; Wimmer, A. B.; Wu, K.; Wnedt-Nordhal, G.;
Zhang, W.; Kathöfer, S.; Schoels, W.; Katus, H. A.; Kiehn, J.; Karle, C. A. Naunyn-
Schmiedeberg’s Arch. Pharmacol. 2003, 368, 41.
19. Marrero-Alonso, J.; García Marrero, B.; Gómez, T.; Díaz, M. Eur. J. Pharmacol.
2006, 532, 115.
20. Dietze, E. C.; Troch, M. M.; Bean, G. R.; Heffner, J. B.; Bowie, M. L.; Rosenberg, P.;
Ratliff, B. Oncogene 2004, 23, 3851.
21. Dick, G. M.; Hunter, A. C.; Sanders, K. M. Mol. Pharmacol. 2002, 61, 1105.
22. Biegon, A.; Brewster, M. E.; Hadassa, D.; Emil, P.; Dalia, S.; Alvin, M. K. Cancer
Res. 1996, 56, 4328.
23. Somjen, D.; Waisman, A.; Kaye, A. M. J. Steroid. Biochem. Mol. Biol. 1996, 59, 389.
24. Brewster, M. E.; Yael, P.; Edna, R.; Biegon, A.; Emil, P.; Hadassa, D. Int. J. Pharm.
1997, 153, 147.
5. Conclusions
25. Biegon, A.; Brewster, M. E. PCT Int. Appl. 1998, 44.
26. Jarman, J.; Leung, O. T.; Leclercq, G.; Devleeschouwer, N.; Stoessel, S.; Coombes,
R. C.; Skilton, R. A. Anti-Cancer Drug Disc. 1996, I, 259.
27. Reel, J. R.; Humphrey, R. R.; Shih, H.; Windsor, B. L.; Sakowski, R.; Creger, P. L.;
Edgren, R. A. Fertil. Steril. 1979, 31, 552.
Due to the relevance of tamoxifen in oncology, the necessity to
have new anticancer drugs, and in order to gain insights into the
mechanisms of action of permanently charged tamoxifen deriva-
tives, we synthesized, characterized, and studied the genomic ac-
tions of these compounds including three new derivatives. Our
results strongly suggest genomic effects as a mechanism of action
of permanently charged tamoxifen derivatives. Detection of com-
pounds in cell lysates and gene reporter studies provide more
accurate information on their potential effects and mechanisms
of action. This approach could lead to a better design of new ther-
apeutic molecules and help to elucidate molecular mechanisms of
new anticancer drugs.
28. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099.
29. Farias, L. M. B.; Ocaña, D. B.; Díaz, L.; Larrea, F.; Avila-Chávez, E.; Cadena, A.;
Hinojosa, L. M.; Lara, G.; Villanueva, L. A.; Vargas, C.; Hernández-Gallegos, E.;
Camacho-Arroyo, I.; Dueñas-González, A.; Pérez-Cárdenas, E.; Pardo, L. A.;
Morales, A.; Taja-Chayeb, L.; Escamilla, J.; Sánchez-Peña, C.; Camacho, J. Cancer
Res. 2004, 64, 6996.
30. Tetko, I. V.; Gasteiger, J.; Todeschini, R.; Mauri, A.; Livingstone, D.; Ertl, P.;
Palyulin, V. A.; Radchenko, E. V.; Zefirov, N. S.; Makarenko, A. S.; Tanchuk, V. Y.;
Prokopenko, V. V. J. Comput. Aided Mol. Des. 2005, 19, 453.
31. Díaz, L.; Ceja-Ochoa, I.; Restrepo-Angulo, I.; Larrea, F.; Avila-Chávez, E.; García-
Becerra, R.; Borja-Cacho, E.; Barrera, D.; Ahumada, E.; Gariglio, P.; Alvarez-Rios,
E.; Ocadiz-Delgado, R.; Garcia-Villa, E.; Hernandez-Gallegos, E.; Camacho-
Arroyo, I.; Morales, A.; Ordaz-Rosado, D.; Garcia-Latorre, E.; Escamilla, J.;
Sánchez-Peña, L. C.; Saqui, M.; Gamboa-Dominguez, A.; Vera, E.; Uribe-
Ramirez, M.; Murbartian, J.; Ortiz, C. S.; Rivera-Guevara, C.; De Vizcaya, A.;
Camacho, J. Cancer Res. 2009, 69, 3300.
Acknowledgments
32. Fisher, H.; Kansy, M.; Avdeef, A.; Senner, F. Eur. J. Pharmacol. Sci. 2007, 31, 32.
33. Jordan, V. C. J. Natl. Cancer Inst. 2007, 99, 350.
34. Jordan, V. C. Pharmacol. Rev. 1984, 36, 245.
We thank Verónica Cadena for her secretarial work, and Guada-
lupe Montiel and Rosa Elena Flores for their technical assistance.
This work was partially supported by Consejo Nacional de Ciencia
y Tecnología (Grant 45753 to J.C.).
35. Michalides, R.; Griekspoor, A.; Balkenende, A.; Verwoerd, D.; Janssen, L.; Jalink,
´
K.; Floore, A.; Velds, A.; vant Veer, L.; Neefjes, J. Cancer Cell 2004, 5, 597.
36. Grenman, S.; Shapira, A.; Carey, T. E. Gynecol. Oncol. 1988, 30, 228.
37. Hwang, J. Y.; Lin, B. Y.; Tang, F. M.; Yu, W. C. Y. Cancer Res. 1992, 52, 6848.
38. Roig, L. M. V.; Lotfi, H.; Olcese, J. E.; Castro, G. L.; Ciocca, D. R. Anticancer Res.
1993, 13, 2457.
References and notes
39. Bianchi, L.; Wible, B.; Arcangeli, A.; Taglialatela, M.; Morra, F.; Castaldo, P.;
Crociani, O.; Rosati, B.; Faravelli, L.; Olivotto, M.; Wanke, E. Cancer Res. 1998, 58,
815.
1. Jordan, V. C. Nat. Rev. Drug Disc. 2003, 2, 205.
2. Singh, M. N.; Strigfellow, H. F.; Paraskevaidis, E.; Martin-Hirsch, P. L.; Martin, F.
L. Cancer Treat. Rev. 2007, 33, 91.