Kinetic analysis of fluorescein and dihydrofluorescein effluxes in tumour cells expressing the multidrug resistance protein, MRP1

Article abstract of DOI:10.1016/S0006-2952(02)01662-3

Multidrug resistance (MDR) in tumour cells is often caused by the overexpression of two transporters the P-glycoprotein (P-gp) and the multidrug resistance-associated protein (MRP1) which actively pump out multiple chemically unrelated substrates across t

Full text of DOI:10.1016/S0006-2952(02)01662-3

Biochemical Pharmacology 65 (2003) 969–977
Kinetic analysis of fluorescein and dihydrofluorescein effluxes in tumour
cells expressing the multidrug resistance protein, MRP1
Chantarawan Saengkhae, Chatchanok Loetchutinat, Arlette Garnier-Suillerot*
´ ´
Laboratoire de Physicochimie Biomoleculaire et Cellulaire (LPBC-CSSB) UMR CNRS 7033, Universite Paris Nord,
74 rue Marcel Cachin, 93017 Bobigny, France
Received 19 September 2002; accepted 29 November 2002
Abstract
Multidrug resistance (MDR) in tumour cells is often caused by the overexpression of two transporters the P-glycoprotein (P-gp) and the
multidrug resistance-associated protein (MRP1) which actively pump out multiple chemically unrelated substrates across the plasma
membrane. A clear distinction in the mechanism of translocation of substrates by MRP1 or P-gp is indicated by the finding that, in most of
cases, the MRP1-mediated transport of substrates is inhibited by depletion of intracellular glutathione (GSH), which has no effect on their
P-gp-mediated transport. The aim of the present study was to quantitatively characterise the transport of anionic compounds
dihydrofluorescein and fluorescein (FLU). We took advantage of the intrinsic fluorescence of FLU and performed a flow cytometric
analysis of dye accumulation in the wild-type drug sensitive GLC4 that do not express MRP1 and its MDR subline which display high
level of MRP1. The measurements were made in real time using intact cells. The kinetics parameters, ka ¼ VM=Km, which is a measure of
the efficiency of the transporter-mediated efflux of a substrate, was very similar for the two FLU analogues. They were highly comparable
with values for ka of other negatively charged substrates, such as GSH and calcein. The active efflux of both FLU derivatives was inhibited
by GSH depletion.
# 2003 Elsevier Science Inc. All rights reserved.
Keywords: Multidrug resistance; MRP1; Fluorescein; Kinetic parameters
1. Introduction
membranes. The substrate specificity of both transporter
proteins is partly overlapping and for instance, anticancer
drugs, such as anthracyclines, vinca alkaloids and etoposide,
are substrates of both transporters. By pumping these agents
out of the tumour cells, MRP1 causes reduced intracellular
accumulation of drugs, leading to resistance. MRP1 has
been described as an ATP-dependent export GS-X pump for
the endogenous GSH conjugate leukotriene C4 and struc-
turally related anionic amphiphilic conjugates [5,6],
whereas P-gp seems to prefer neutral or positively charged
molecules [7]. A clear distinction in the mechanism of
translocation of these types of substrates by MRP1 or P-
gp is indicated by the finding that their MRP1-mediated
transport is inhibited by depletion of intracellular GSH
which has no effect on their P-gp-mediated transport [8,9].
In any case, little is known on the relative pumping
efficiency by MRP1 of the different types of substrates
and in particular about the mechanism by which GSH
facilitates transport of some compounds. The proposal that
drug efflux by MRP1 might occur after conjugation with
GSH has been questioned [10] and many people are now in
TheMRP1-encoded multidrug resistance protein (MRP1)
and the MDR1-encoded P-gp are both plasma membrane
transporters thought to be responsible in part for the resis-
tance of tumour cells to multiple chemically unrelated drugs
(MDR) [1–3]. Both proteins belong to the superfamily of
the so-called ATP binding cassette transport proteins or
traffic ATPases [4], which are known to be dependent on
ATP hydrolysis for the translocation of substrate across
* Corresponding author. Tel.: þ33-1-48-38-77-48;
fax: þ33-1-48-38-77-77.
E-mail address: garnier@lpbc.jussieu.fr (A. Garnier-Suillerot).
Abbreviations: MDR, multidrug resistance; MRP1, multidrug resistance-
associated protein; P-gp, P-glycoprotein; HRP, horse raddish peroxidase;
GSH, glutathione; DHF-DA, dihydrofluorescein diacetate; FLU, fluorescein;
DHF, dihydrofluorescein; BSO, L-buthionine-(S,R) sulphoximine; PAK-
104P, 2-[4-(diphenylmethyl)-1-piperazinyl]ethyl-5-(trans-4,6-dimethyl-
1,3,2-dioxaphosphorinan-2-yl)-2,6-dimethyl-4-(3-nitrophenyl)-3-pyridine-
carboxylate P oxide; MK571, 3-([{3-(2-[7-chloro-2-quinolinyl]ethenyl)phe-
nyl}-{(3-dimethylamino-3-oxopropyl)-thio}-methyl]thio)propanoic acid;
Ci, concentration of free intracellular FLU (or DHF).
0006-2952/03/$ – see front matter # 2003 Elsevier Science Inc. All rights reserved.
doi:10.1016/S0006-2952(02)01662-3

970
C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
favour of a co-transport mechanism involving GSH [11–13].
In the context of our research on the mechanisms of drugs
active efflux in resistant cells, we have undertaken a kinetic
approach to determine the relative pumping efficiency of
two chemically related anionic compounds, FLU and dihy-
drofluorescein (DHF). Small cell lung cancer, GLC4 and
GLC4/ADR, which overexpress MRP1, have been used. We
have determined quantitative parameters for the active
efflux of both compounds and compared them to the values
obtained under similar conditions for other substrates either
anionic or neutral or cationic. Results from this study show
that both FLU and DHF are actively pumped out by MRP1
and that this depends on the presence of GSH. In addition,
the active parameter values (VM/Km) are very close to those
observed for MRP1-mediated efflux of GSH and of calcein
which are also anionic substrates [14,15]. This work repre-
sents the first report, using intact cells, of real-time mea-
surements of the rate of FLU transport.
HEPES plus 132 mM NaCl, 3.5 mM KCl, 1 mM CaCl2,
0.5 mM MgCl2 and 5 mM glucose at pH 7.3. All other
reagents were of the highest quality available. Deionized
double-distilled water was used throughout theexperiments.
2.3. Intracellular GSH measurement
To quantify free GSH inside the cells we have been used
monochlorobimane, which itself is non-fluorescent, but
forms a fluorescent adduct through conjugation to GSH
by glutathione S-transferase [17]. Based on this observation,
we have recently developed a very rapid and sensitive
fluorometric method for GSH measurement [15]. Cells,
2 Â 106suspended in 2 mL of buffer, were disrupted by
sonication on ice (3 Â 10 s, power 2). The rate of mono-
chlorobimane–GSH formation was monitored after addition
of monochlorobimane 100 mM and glutathione S-transfer-
ase 0.5 U/mL, as described [15]. In the absence of glu-
tathione S-transferase, the rate of formation of the
fluorescent derivative was very slow. We checked that
oxidised GSH did not give rise to any modification of the
fluorescence signal.
2. Materials and methods
2.1. Cell culture
2.4. GSH depletion
GLC4 and MRP1-expressing GLC4/ADR cells [16]
were cultured in RPMI 1640 (Sigma Chemical Co) med-
ium supplemented with 10% foetal calf serum (Biomedia)
at 378 in a humidified incubator with 5% CO2. The resistant
GLC4/ADR cells were cultured with 1.2 mM doxorubicin
until 1–4 weeks before the experiments. Cell cultures used
for experiments were split 1:2, 1 day before use in order to
assure logarithmic growth. Cells (106/mL; 2 mL per cuv-
ette) were energy depleted via preincubation for 30 min in
HEPES buffer with sodium azide but without glucose.
In order to examine the effect of GSH depletion by BSO
on FLU derivative efflux, cells were cultured in the pre-
sence of 25 mM BSO for about 18 hr.
2.5. Cellular DHF and FLU accumulation
The methodology for the determination of the kinetics of
active transport of fluorescent drugs or any fluorescent
probe from tumour cells has been used and discussed
before for anthracyclines and calcein [14,18–20]. We have
now adapted this technique to measure the kinetics of
active DHF and FLU transport. Basically, the fluorescence
signal (macrospectrofluorescence) is monitored continu-
ously during incubation of the cells with DHF-DA. DHF-
DA is a non-fluorescent molecule. It enters by passive
diffusion into the cells where it is transformed in DHF by
cytoplasmic esterases. DHF, which is also a non-fluores-
cent molecule, is oxidised by oxidative species to FLU
which is fluorescent (Fig. 1). Since DHF-DA and DHF are
non-fluorescent molecules, the fluorescence which is mea-
sured is always that of the FLU.
The cells were incubated in a quartz cuvette in a suspen-
sion of 106 cells/mL in HEPES buffer with 5 mM glucose as
indicated. The fluorescence was recorded on a Perkin Elmer
LS50B spectrofluorometer at 520 nm (lex ¼ 490 nm). To
determine intracellular FLU accumulation inside the cells,
during the time course of these experiments, aliquots were
taken at various interval of time and used as such for flow
cytometry measurements (Becton-Dickinson, Facscan).
Details of the experimental set-ups are given in the results
section.
2.2. Drugs and chemicals
Dihydrofluorescein diacetate (DHF-DA) was from
Molecular Probes. It was dissolved in DMSO. DHF was
prepared by basic hydrolysis of DHF-DA. Stock solutions
were prepared just before use. 2-[4-(Diphenylmethyl)-1-
piperazinyl]ethyl-5-(trans-4,6-dimethyl-1,3,2-dioxapho-
sphorinan-2-yl)-2,6-dimethyl-4-(3-nitrophenyl)-3-pyridi-
necarboxylate P oxide (PAK-104P) was a gift from Drs.
Shudo, Iwasaki and Akiyama (Nissan Chemical Industries
Ltd.). 3-([{3-(2-[7-Chloro-2-quinolinyl]ethenyl)phenyl}-
{(3-dimethylamino-3-oxopropyl)-thio}-methyl]thio)pro-
panoic acid (MK571) was provided by Dr. R.N. Young
(Merck-Frosst Centre for Therapeutic Research). Triton
X-100 and horse raddish peroxidase (HRP) were from
Sigma and were dissolved in water. Reduced GSH, glu-
tathione transferase from equine liver (GSHT) and L-buthio-
nine-(S,R) sulphoximine (BSO) were from Sigma.
Monochlorobimane was from Molecular Probes. Before
the experiments, the cells were counted, centrifuged and
resuspended in HEPES buffer solutions containing 20 mM

C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
971
The efficiency of the efflux can then be characterised by
calculating ka, the pump-mediated efflux coefficient for the
drug according to the Eq. (2).
3. Results
In order to be able to calculate the kinetic parameters ka
for the active transport of the substrates DHF and FLU, we
have performed a series of experiments designed to deter-
mine at any time (i) the concentration of free internal FLU
and DHF, (ii) the rate constant for outward pumping at
limiting FLU and DHF concentrations, i.e. when the sub-
strate concentrations are much lower than Km and Eq. (2)
can be used.
3.1. FLU formation detected using flow cytometry and
macrospectrofluorescence
Cells, 106/mL, were incubated with 20 mM DHF-DA and
Fig. 2 shows typical records of the fluorescence intensity as
a function of time, using either flow cytometry (Fcyto) or
spectrofluorometry (Fmacro). As can be seen, for sensitive
cells, Fcyto increased during the first hour and then plateaued
for at least for one additional hour. However, for GLC4/
ADR cells, after a first increase Fcyto vanished rapidly
indicating that FLU was pumped out. This was corroborated
by the observation that when energy-depleted GLC4/ADR
cells were used the Fcyto signal was similar to that obtained
with sensitive cells. The Fmacro signal increased as a func-
tion of time, when either sensitive or resistant cells were
used. However, the signal intensity was lower in resistant
cells than in sensitive cells, indicating that less FLU was
formed and, therefore, that DHF-DA was pumped out by
MRP1 as the hydrolysis of DHF-DA to FLU occurred inside
the cells only.
Fig. 1. Structure of dihydrofluorescein diacetate (DHF-DA), dihydro-
fluorescein (DHF) and fluorescein (FLU).
2.6. Mathematical calculations
In the following Va will stand for the rate for outward
pumping of FLU (or DHF) and Ci for the concentration of
free internal FLU (or DHF).The determination of the
kinetic parameters, e.g. the maximum rate (VM) and the
Michaelis constant (Km), characteristic of the transporter-
mediated efflux of any molecule required the measurement
of Va and Ci. When Va can be determined for various
intracellular free substrate concentrations Ci, the maximal
efflux rate (VM) and the apparent Michaelis–Menten con-
stant (Km) can be computed by non-linear regression
analysis of transport velocity Va vs. Ci assuming that the
transport follows the Michaelis equation:
3.2. Quantification of DHF and FLU inside the cells
First, we have measured, under our experimental con-
ditions, the molar fluorescence of FLU: FFLU. For this
purpose, DHF was prepared by basic hydrolysis of DHF-
DA. HRP or HRP þ H2O2 (the result was the same) was
added to a precise concentration of DHF yielding the total
conversion of DHF to FLU and macrospectrofluorescence
was measured.
Second, we have checked that the quantum yield of FLU
was independent of its localisation in the intra- or extra-
cellular medium (see Section 3.3).
Third, sensitive cells, 106/mL, were incubated with
20 mM DHF-DA andtheflow cytometrysignalwas recorded
as a function of time. After 1 hr cells were centrifuged and
suspended in a DHF-DA free buffer. The intensity of the
signal continue to increase despite the absence of DHF-DA
in the extracellular medium. Before centrifugation, the
rate for the appearance of the FLU signal depended on
VM Â Ci
Va ¼
(1)
Km þ Ci
In many cases, the complete curve Va ¼ fðCiÞ cannot be
obtained and, therefore, it is not possible to obtain these
two parameters characteristic of a transporter. However, if
Ci is much lower than Km, Eq. (1) becomes:
VM
Va ꢀ
Ci or Va ¼ ka  Ci
(2)
Km

972
C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
Fig. 2. Uptake of DHF-DA by sensitive and resistant GLC4 cells. Cells, 106/mL either sensitive (&), resistant (*) or resistant in the presence of 10 mM
PAK-104P (*) were incubated with 20 mM DHF-DA and the fluorescence signals, measured using traditional fluorescence (Fmacro) and flow cytometry
(Fcyto), were plotted as a function of the time of incubation. Data points are from a representative experiment.
(1) the rate for the passage of DHF-DA through the plasma
membrane, (2) the rate of hydrolysis of DHF-DA to
DHF and then (3) the rate of oxidation of DHF to the
fluorescent species FLU. The rate of oxidation of DHF to
FLU should be proportional to the DHF concentration. The
observation that after centrifugation and elimination of
DHF-DA in the extracellular medium the rate of FLU signal
apparition was the same that before centrifugation, strongly
suggested that in both cases the FLU concentration was very
similar. Therefore, the limiting step in the formation of FLU
was neither the uptake of DHF-DA by cells, nor its hydro-
lysis to DHF but its intracellular oxidation by oxidising
species.
checked that similar data were obtained with energy-
depleted resistant cells indicating that the conversion of
DHF to FLU was similar in both the sensitive and resistant
cell lines.
As DHF was non-fluorescent, the determination of its
intracellular concentration could only be indirect: cells
were incubated with DHF-DA, Fcyto was measured before
the addition of H2O2, yielding [FLU]i, and after the addi-
tion of H2O2. The increase of Fcyto yielded the intracellular
concentration of DHF which was present inside the cells
and had been converted to FLU by H2O2 addition.
3.3. Control experiments
The oxidation of DHF to FLU being slow, we found a
way to oxidise it rapidly and completely and measured its
intracellular concentration. Cells were incubated for 1 hr
with 20 mM DHF-DA. After 1 hr they were centrifuged
and suspended in a free DHF-DA buffer. According to the
data described earlier, inside the cells DHF and FLU were
present. Then H2O2 was quickly added which yielded the
total and fast conversion of DHF to FLU and the Fmacro
and Fcyto signal very quickly recorded (Fig. 3). Under
these conditions, the fluorescence signal measured with
macrospectrofluorescence (Fmacro) and with flow cytome-
try (Fcyto) were those of FLU inside the cells. From Fmacro
and the value of the molar fluorescence determined ear-
lier, we calculated [FLU]i, the intracellular FLU concen-
tration. For that purpose we have assimilated cells to
spheres having a mean radius equal to 12–13 mm, as
determined with a Coulter counter, and calculated a mean
volume equal to 10À12 L. Therefore, it was possible to
draw a calibration curve by plotting Fcyto as a function of
Fmacro (i.e. [FLU]i) (Fig. 4). It follows that in any ex-
periment, the measure of Fcyto yielded [FLU]i. We have
After having established the principle of the experi-
ments, as explained earlier, a set of control experiments
was performed in order to further validate the use of the
experimental model to analyse the transport kinetics of
FLU and DHF.
First, a control experiment was done in order to establish
that the fluorescence properties of FLU under these con-
ditions are the same inside or outside the cells. Therefore,
after 1 hr of cell incubation with DHF-DA, cells were
centrifuged and suspended in DHF-DA free buffer and
Fmacro recorded. Then, the plasma membranes were dis-
rupted by sonication and we observed that the fluorescence
signal was not significantly modified. This indicates that
there are only minor differences in FLU fluorescence in- or
outside the cells and that there does not seem to be self-
quenching of FLU at the [FLU]i which are reached in the
intracellular volume (see later). In addition, we examined
the GLC4 and GLC4/ADR cells by fluorescence microscopy
after loading with DHF-DA. It appeared that in these cells
the FLU fluorescence was evenly distributed throughout the

C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
973
Fig. 3. Oxidation of the intracellular DHF to FLU by addition of H2O2 Sensitive cells (106/mL) were incubated with 20 mM DHF-DA. After 1 hr cells were
centrifuged and suspended in a free DHF-DA buffer. The fluorescence signal measured using traditional fluorescence (Fmacro) and flow cytometry (Fcyto) was
recorded as a function of the time of incubation. The fluorescence signals (Fcyto)1 and (Fmacro)1 were due to FLU inside the cells. Then 200 mM H2O2 was
quickly added which yielded the total and fast conversion of the intracellular DHF to FLU. The fluorescence signals (Fcyto)2 and (Fmacro)2 were due to the
oxidation of DHF to FLU. Data points are from a representative experiment.
nucleus and cytoplasm. Together these results justify the
assumption made for our calculations that the FLU may be
regarded as evenly distributed without evidence of self-
quenching.
A second control experiment was made with H2O2. The
oxidation of DHF to FLU by addition of H2O2 to the cells
was performed at various concentrations of H2O2 and we
have checked that it depended on the H2O2 concentration
Fig. 4. Flow cytometry and the cytosolic free FLU concentration. The experimental conditions are the same as those in Fig. 3. The signal Fcyto recorded after
the addition of H2O2 has been plotted as a function of [FLU]i, the intracellular FLU concentration. [FLU]i was calculated using the following equation:
½FLUŠ ¼ ðFmacro=FFLUÞ Â 103 were FFLU is the molar fluorescence of FLU determined as described earlier and it is taken into account that 106 cells/mL are
i
used and that the intracellular volume is ꢁ10À12 L.

974
C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
Fig. 6. Efflux of DHF from GLC4/ADR cells. Cells, 106/mL, were
incubated with 20 mM DHF-DA for 1 hr. They were then centrifuged and
resuspended in DHF-DA free buffer. The intracellular DHF concentration
was then determined as a function of time according to the procedure
described in Figs. 2 and 3. Cells were GLC4 (&), GSH-depleted GLC4
(&), GLC4/ADR (*) and GSH-depleted GLC4/ADR (*). The values
represent mean Æ SD of three independent experiments performed on three
different days.
Fig. 5. Efflux of FLU from GLC4/ADR cells. Cells, 106/mL, were
incubated with 20 mM DHF-DA for 1 hr. They were then centrifuged and
resuspended in DHF-DA free buffer. The intracellular FLU concentration
was then determined as a function of time according to the procedure
described in Figs. 2 and 3. Cells were GLC4 (&), GSH-depleted GLC4 (&),
GLC4/ADR (*) and GSH-depleted GLC4/ADR (*). The values represent
mean Æ SD of three independent experiments performed on three different
days.
up to 100 mM and then plateaued. At the 200 mM H2O2
concentration used the cells were viable during the short
time necessary for the measurement, as checked by trypan
blue exclusion.
A third control experiment was made to check the
variation of the FLU fluorescence as a function of pH.
The fluorescence signal increased from pH 3 up to 7 and
then plateaued till pH 10 and then decreased. It follows
that, under our experimental conditions, a variation of the
FLU fluorescence cannot be due to change in pH (the pH of
the cytosol is ꢁ7.3 and FLU which is negatively charged
cannot accumulated inside intracellular acidic organelles
such as lysosomes).
A fourth control was done to check the ATP intracel-
lular level under the different experimental conditions.
The ATP concentration was determined using the luci-
ferin–luciferase test [21]. In both cell lines the presence
of azide under glucose-free conditions yielded 90% ATP
depletion.
Under these experimental conditions (cells incubated for
1 hr with DHF-DA, then centrifuged and suspended in a free
DHF-DA buffer), inside the cells are present DHF-DA, DHF
and FLU. The concentration of DHF-DA should decrease
because of its cleavage by esterase and the concentration of
FLU should increase because of the oxidation of DHF.
However, the concentration of DHF can, on the one hand,
increase because of the cleavage of DHF-DA and, on the
other hand, decrease because of its oxidation to FLU. The
observation (Fig. 6) that, in sensitive cells and resistant cells
plus BSO, the concentration of DHF remains almost con-
stant strongly suggest that there is a steady state for DHF, i.e.
that the rate of cleavage of DHF-DA is very close to the rate
of DHF oxidation to FLU. It follows that, in resistant cells,
the decrease of the DHF concentration is due to its pumping
out by MRP1 only. Our data show that Va ¼ ð2:5 Æ 0:4Þ Â
10À20 mole/cell/swhen½DHFŠ ¼ 41 Æ 4 mM yieldingka ¼
i
ð6:1 Æ 1:4Þ Â 10À16 L/cell/s (ka ¼ Va=Ci).
Now let us consider the FLU concentration in sensitive
cells and GSH-depleted resistant cells. Here, the concen-
tration of the fluorescent molecule can increase only.
Actually, this is what is observed: in Fig. 5 one can see
that during the first 30 min, after resuspension of the cells
in DHF-DA free buffer, the FLU concentration increases
and becomes about 1.5-fold higher than at t ¼ 0, and then
plateaued. Here, also it follows that the decrease of the
FLU concentration in resistant cells is due to its pumping
out by MRP1. Our data show that Va ¼ ð8:8 Æ 1:4Þ Â
3.4. Rate of the MRP1-mediated efflux of
FLU and DHF
GLC4 and GLC4/ADR cells were incubated with 20 mM
DHF-DA. After 1 hr cells were centrifuged and suspended
in DHF-DA free buffer. Then after each 10 min aliquot of
cells was taken and the signal Fcyto was recorded just
before and after addition of 200 mM H2O2. The calibration
curve in Fig. 3 then allowed the determination of [FLU]i
and [DHF]i as a function of time. Figs. 5 and 6 show typical
plots of [FLU]i and [DHF]i, respectively, as a function of
time for sensitive and resistant cells in normal cells and in
GSH-depleted cells.
10À21 mole/cell/s when ½FLUŠ ¼ 14 Æ 1 mM yielding
i
ka ¼ ð6:3 Æ 1:4Þ Â 10À16 L/cell/s (ka ¼ Va=Ci). Therefore,
for these two compounds, the values of the parameters ka
characteristic of the efficiency for the active efflux werevery
similar.

C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
975
3.5. Inhibition of the MRP1-mediated efflux of DHF
and FLU by PAK-104P and MK571
conditions especially not with living cells. We are aware
that ka, which is proportional to VM/Km, contains both the
turnover number of the transporter protein (i.e. the number
of substrate molecules transported per MRP1 molecule per
unit of time) as well as the number of transport proteins in
the cell membrane (VM) and the affinity of the substrate for
the transporter (Km). However, the parameter ka allows a
convenient comparison of the transport efficiency of dif-
ferent substrates in the same cells.
Resistant cells, 106/mL, were incubated with 20 mM
DHF-DA and either 10 mM PAK-104P or 5 mM MK571.
The variation of the intensity of the flow cytometry signal
as a function of time was very similar to that observed with
sensitive cells indicating that the MRP1-mediated efflux of
DHF and FLU was inhibited by these two compounds.
Data obtained with PAK-104P are shown in Fig. 2.
Several dyes deriving from FLU have recently been
postulated to be handled by MRP1; indeed, 20,70-bis(2-
carboxyethyl)-5(6)-carboxyfluorescein, used to investigate
intracellular pH, andcarboxyfluorescein have been shown to
be actively effluxed by MRP1-overexpressing cells [25,26].
Such carboxyfluorescein-related compounds appear, there-
fore, to constitute a new identified class of substrates of
MRP1 but, up to now, quantitative data are lacking [27,28].
Inthis paper, wepresent data that further characterised the
transport of FLU and DHF. The measurement was made in
real time using intact cells. The findings presented here are
the first to show quantitative information about the kinetics
parameters for MRP1-mediated efflux of FLU derivatives in
intact cells. The conclusions that emerge from our data are
that DHF and FLU are actively pumped, by MRP1, that the
active efflux coefficient is very similar for both compounds,
and that their efflux is inhibited by typical MRP1 inhibitors,
MK571 and PAK-104P [29,30]. Here, it is interesting to
compare the values of the ka parameter for DHF and FLU to
those obtained, using the same cell line, for other anionic
compounds, such as GSH and calcein, which are MRP1
substrates. We have recently determined that ka was equal to
(6:3 Æ 2:9Þ Â 10À16 and (4:4 Æ 2:1Þ Â 10À16 L/cell/s for
calcein and GSH, respectively [14,15], i.e. very close to
thevaluesobtained for DHFand FLU (see Table 1). Also itis
interesting to remark that, GSH depletion brought about by
pretreatment for 24 hr with 25 mM BSO, showed significant
effectsoneffluxoftheseanionic species. As we have already
said, data concerning the effect of GSH on MRP1-mediated
efflux of anionic species are conflicting especially those
obtained with calcein. Feller et al. [24], using GLC4/ADR
cells in which the intracellular level is about 14 mM, have
observed that a GSH depletion to about 18% of the initial
value (i.e. ꢁ2.5 mM) has no effect on calcein efflux,
whereas, Bagrij et al. [23], using COR-L23/R cells, have
observed a decrease of calcein efflux when the intracellular
level of GSH was decreased to 25% of the initial value, i.e.
from ꢁ5.1 to 1.1 mM. The different effects observed were,
therefore, obtained at different intracellular GSH concen-
trations. One hypothesis to explain this apparently conflict-
ing data could be that, after BSO treatment, the GSH
concentration was sufficient to sustain calcein efflux in
GLC4/ADR cells but not in COR-L23/R.
4. Discussion
The mechanism by which GSH facilitates transport of
some compounds by MRP1 is still a matter of debate.
MRP1 is able to transport GSH conjugates, such as dini-
trophenyl glutathione and it was found at an early stage that
MRP1 was also able to transport non-anionic drugs, such as
anthracyclines, vinca alkaloids and epipodophyllotoxins
[11–13,19,22]. Attempts to detect derivatives of these
drugs conjugated to an anionic ligand (GSH, glucuronic
acid, sulphate) have remained unsuccessful and the con-
sensus is now that these drugs are transported as such. So, it
seems now recognised that MRP1 can co-transport unme-
tabolised compounds which are either neutral or cationic
[11–13] with GSH. In addition, it has recently been
demonstrated that the co-transport of DNR and GSH
had a 1:1 stoichiometry [15]. However, the influence of
GSH on anionic substrate MRP1-mediated efflux is far
from being elucidated [23]. For instance, it has been
reported by Feller et al. [24] that MRP1 can transport
the organic anion calcein without the requirement of GSH,
whereas, more recently, Bagrij et al. [23] have shown that a
decrease of the intracellular GSH concentration lead to a
decrease of the MRP1-mediated calcein efflux.
Most of the data found in the literature concerning the
MRP1-mediated efflux of compounds are qualitative and it
is always difficult to compare the efficiency of the efflux of
different substrates by the transporter. For these reasons,
for several years we are involved in the quantitative
determination of the kinetics parameters because measure-
ment of the kinetic characteristics of substrate transport is a
powerful approach for enhancing our understanding of
their function and mechanism. For this purpose we use
the same cell line throughout our experiments. In most of
cases, we characterise this efficiency by calculating the
parameter ka [18–20]. As it is described in the Section 2, ka
is proportional to the ratio VM/Km and is very convenient to
evaluate the efficiency of a transporter. This parameter is
very useful because its value can be estimated from a few
number of measurements while the determination of the
kinetics parameters VM and Km requires (i) a very large
number of measurements, (ii) the use of high substrate
concentration needed to saturate the transporter and reach
the maximal rate. It is not always possible to use such
The relative importance of the charge of molecules for
their transport by MRP1, can be here considered as the four
molecules for which ka has already been determined, i.e.
FLU, DHF, calcein and GSH, have different net negative

976
C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
Table 1
Rate constant for efflux
a
Compounds
ka ¼ VM/Km (L/cell/s)
Electrostatic charge at pH 7.3
Reference
FLU
(6.1 Æ 1.4) Â 10À16
(6.3 Æ 1.4) Â 10À16
(4.4 Æ 1.5) Â 10À16
(6 Æ 2) Â 10À16
À1
À1
À1
À5
0
This work
This work
[11]
DHF
GSH
Calcein
[10]
[10]
Calcein-AM
Daunorubicin
Tetramethylrosamine
(4 Æ1) Â 10À12
(1.8 Æ 0.5) Â 10À12
þ1
þ1
[15,16]
(0.3 Æ 0.1) Â 10À12
Unpublished data
a The data are the means Æ SEM of at least five determinations
[8] Versantvoort CHM, Broxterman HJ, Bagrij T, Scheper RJ, Twentyman
P. Regulation by glutathione of drug transport in multidrug-resistant
human lung tumour cell lines overexpressing multidrug resistance-
associated protein. Br J Cancer 1995;72:82–9.
charges (Table 1) at pH 7.3. As can be seen, for these
molecules, the ka values are very similar and do not depend
on the net negative charge.
It should be also emphasise here that the efficiency of
MRP1 to efflux anionic or neutral substrates is about 104-
fold lower than its efficiency to pump out neutral or posi-
tively charged substrates. For comparison the ka values
previously determined for daunorubicin, calcein-AM and
the rhodamine derivative, tetramethylrosamine, are reported
in Table 1. The underlying molecular and biochemical
causes for the extreme differences in transporter efficiency
of the different MRP1 substrates as measured by us, remain
unexplained presently.
[9] Hollo Z, Homoloya L, Hegedus T, Sarkadi B. Transport properties of
the multidrug resistance-associated protein (MRP) in human tumour
cells. FEBS Lett 1996;383:99–104.
[10] Deeley RG, Cole SPC. Function evolution and structure of multidrug
resistance protein (MRP). Semin Cancer Biol 1997;8:193–204.
[11] Loe DW, Almquist KC, Deeley RG, Cole SPC. Multidrug resistance
protein (MRP)-mediated transport of leukotriene C-4 and chemother-
apeutic agents in membrane vesicles—demonstration of glutathione-
dependent vincristine transport. J Biol Chem 1996;271:9675–82.
[12] Rappa G, Lorico A, Flavell RA, Sartorelli AC. Evidence that the
multidrug resistance protein MRP functions as a co-transporter of
glutathione and natural produce toxins. Cancer Res 1997;57:15232–7.
[13] Renes J, De Vries EGE, Nuenhuis EF, Jansen PLM, Mu¨ller M. ATP-
and glutathione-dependent transport of chemotherapeutic drugs by the
multidrug resistance protein MRP1. Br J Pharmacol 1999;126:681–8.
Acknowledgments
¨
[14] Essodaıgui M, Broxterman HJ, Garnier-Suillerot A. Kinetic analysis
of calcein and calcein-acetoxymethylester efflux mediated by the
multidrug resistance protein and P-glycoprotein. Biochemistry 1998;
37:2243–50.
´
This research was supported by grants from l’Universite
Paris XIII and CNRS.
[15] Salerno M, Garnier-Suillerot A. Kinetics of glutathione and daunor-
ubicin efflux from multidrug resistance protein overexpressing small-
cell lung cancer cells. Eur J Pharmacol 2001;421:1–9.
References
[16] Zijlstra JG, De Vries EGE, Mulder NH. Multifactorial drug resistance
in an adriamycin-resistant human small cell lung carcinoma cell line.
Cancer Res 1987;47:1780–4.
[1] Cole SPC, Bhardwaj G, Gerlach JH, Mackie JE, Grant CE, Almquist
KC, Stewart AJ, Kurz EU, Duncan AMV, Deeley RG. Overexpression
of a transporter gene in a multidrug-resistant human lung cancer cell
line. Science 1992;258:1650–4.
[17] Fernandez-Checa JC, Kaplowitz N. The use of monochlorobimane to
determine hepatic GSH levels and synthesis. Anal Biochem 1990;190:
212–9.
[2] Bradley G, Juranka PF, Ling V. Mechanisms of multidrug resistance.
Biochem Biophys Acta 1988;948:87–128.
[18] Mankhetkorn S, Dubru F, Hesschenbrouck J, Fiallo M, Garnier-
Suillerot A. Relation among the resistance factor kinetics of uptake,
and kinetics of the P-glycoprotein-mediated efflux of doxorubicin,
daunorubicin, 8-(S)-fluoroidarubicin, and idarubicin in multidrug-
resistant K562 cells. Mol Pharmacol 1996;49:532–9.
[3] Broxterman HJ, Giaccone G, Lankelma J. Multidrug resistance pro-
teins and other drug transport-related resistance to natural product
agents. Curr Opin Oncol 1995;7:532–40.
[4] Higgins CF. ABC transporters—from microorganisms to man. Annu
Rev Cell Biol 1992;8:67–113.
[19] Marbeuf-Gueye C, Broxterman HJ, Dubru F, Priebe W, Garnier-
Suillerot A. Kinetics of anthracycline efflux from multidrug resistance
protein-expressing cancer cells compared with P-glycoprotein-expres-
sing cancer cells. Mol Pharmacol 1998;53:141–7.
[5] Leier I, Jedlitschky G, Buchholz U, Cole SPC, Deeley RG, Keppler D.
The MRP gene encodes an ATP-dependent export pump for leuko-
triene C-4, and structurally related conjugates.
J Biol Chem
[20] Marbeuf-Gueye C, Ettori E, Priebe W, Kozlowski H, Garnier-Suillerot
A. Correlation between the kinetics of anthracycline uptake and the
resistance factor in cancer cells expressing the multidrug resistance
protein or the P-glycoprotein. Biochim Biophys Acta 1999;1450:
374–84.
1994;269:27807–10.
[6] Mu¨ller M, Meijer C, Zaman GJR, Borst P, Scheper RJ, Mulder NH, de
Vries EGE, Jansen PLM. Overexpression of the gene encoding the
multidrug resistance-associated protein results in increased ATP-de-
pendent glutathiones S-conjugate transport. Proc Natl Acad Sci USA
1994;91:13033–7.
[21] Kimmich G, Randles J, Brand J. Assay of picomole amounts of ATP,
ADP and AMP using the luciferase enzyme system. Anal Biochem
1975;69:187–91.
[7] Broxterman HJ, Schuurhuis GJ, Lankelma J, Oberink JW, Eekman
CA, Claessen AME, Hoekman K, Poot M, Pinedo HM. Highly
sensitive and specific detection of P-glycoprotein function for heme-
tological and solid tumor cells using a novel nucleic acid stain. Br J
Cancer 1997;76:1029–34.
[22] Zaman GJR, Flens MJ, Van Leusden MR, De Haas M, Mu¨lder HS,
Lankelma J, Pinedo HM, Scheper RJ, Baas F, Broxterman HJ, Borst P.
The human multidrug resistance-associated protein MRP is a plasma
membranedrug-effluxpump. ProcNatlAcadSciUSA1994;91:8822–6.

C. Saengkhae et al. / Biochemical Pharmacology 65 (2003) 969–977
977
[23] Bagrij T, Klokouzas A, Hladky SB, Barrand MA. Influences of
glutathione on anionic substrate efflux in tumour cells expressing
the multidrug resistance-associated protein, MRP1. Biochem Phar-
macol 2001;62:199–206.
[27] Laupeze B, Amiot L, Courtois A, Vernhet L, Drenou B, Fauchet R,
Fardel O. Use of the anionic dye carboxy-20,70-dichlorofluorescein for
sensitive flow cytometric detection of multidrug resistance-associated
protein activity. Int J Oncol 1999;15:571–6.
[24] Feller N, Broxterman HJ, Wahrer DC, Pinedo HM. ATP-dependent
efflux of calcein by the multidrug resistance protein (MRP): no inhibi-
tion by intracellular glutathione depletion. FEBS Lett 1995;368:385–8.
[25] Draper MP, Martell RL, Levy SB. Active efflux of the free acid form of
the fluorescent dye 20,70-bis(2-carboxyethyl)-5(6)-carboxyfluorescein
in multidrug-resistance-protein-overexpressing murine and human
leukemia cells. Eur J Biochem 1997;243:219–24.
[28] Sun H, Johnson DR, Finch RA, Sartorelli AC, Miller DW, Elmquist
WF. Transport of fluorescein in MDCKII-MRP1 transfected cells and
mrp1-knockout mice. Biochem Biophys Res Commun 2001;284:
863–9.
[29] Leier I, Jedlitschky G, Buchholz U, Center M, Cole SPC, Deeley RG,
Keppler D. ATP-dependent glutathione disulphide transport mediated
by the MRP gene-encoded conjugate export pump. Biochem J 1996;
314:433–7.
[26] Van der Kolk DM, De Vries EG, Koning JA, van den Berg E, Muller
M, Vellenga E. Activity and expression of the multidrug resistance
proteins MRP1 and MRP2 in acute myeloid leukemia cells tumor cell
lines and normal hematopoietic CD34þ peripheral blood cells. Clin
Cancer Res 1998;4:1727–36.
[30] Marbeuf-Gueye C, Salerno M, Quidu P, Garnier-Suillerot A. Inhibi-
tion of the P-glycoprotein-and multidrug resistance protein-mediated
efflux of anthracyclines and calceinacetoxymethyl ester by PAK-104P.
Eur J Pharmacol 2000;391:207–16.