Ability of carbazole salts, inhibitors of Alzheimer β-amyloid fibril formation, to cross cellular membranes

Article abstract of DOI:10.1016/j.ejphar.2007.01.005

Alzheimer's disease is characterized by the presence of β-amyloid fibril formation. The inhibition of this peptide accumulation may be a prevention method for Alzheimer's disease. Several classes of molecules have been reported to inhibit β-amyloid fibril formation and among them carbazoles. However, very few studies have been performed to determine the destination of such molecules in vivo and especially if they can pass the blood brain barrier. The aim of this paper is to study whether carbazoles could pass the blood brain barrier, i.e. if they can circumvent ATP Binding Cassette (ABC) transporters such as P-glycoprotein (P-gp) and Multidrug Resistance-associated protein (MRP1) which efficiently limit drug brain uptake. For this purpose we have synthesized a fluorescent derivative of carbazole benzothiazolium iodide 1,2 disubstituted ethylene (referred as carbazole thiazole: CT), which can be easily detected and followed in the pre-trial study phases in cells or in tissue. We use cellular models overexpressing P-gp and MRP1. Our results show that: i) CT is able to cross membranes and to penetrate rapidly inside the cells, ii) CT is a P-gp substrate and consequently its accumulation in P-gp overexpressing cells is very low, iii) CT is a poor MRP1 substrate. In addition once inside the cells, CT rapidly binds to DNA and is then slowly reduced by intracellular reducing agents. In conclusion, the efficiency of carbazole derivatives in inhibiting the β-amyloid formation in vivo could be highly compromised because, as P-gp substrates, they will probably not cross the blood brain barrier.

Full text of DOI:10.1016/j.ejphar.2007.01.005

European Journal of Pharmacology 559 (2007) 124–131
www.elsevier.com/locate/ejphar
Ability of carbazole salts, inhibitors of Alzheimer β-amyloid fibril
formation, to cross cellular membranes
a
a,
⁎
b
b
Chantarawan Saengkhae , Milena Salerno , Dominique Adès , Alain Siove ,
a
b
a
Laurence Le Moyec , Véronique Migonney , Arlette Garnier-Suillerot
a
Laboratoire de Biophysique Moléculaire, Cellulaire et Tissulaire (BioMoCeTi), UMR CNRS 7033,
Université Paris 13 et Paris 6, 74 rue Marcel Cachin, 93017 Bobigny, France
Laboratoire de Biomatériaux et Polymères de Spécialité, LBPS/B2OA — UMR CNRS 7052, Université Paris 13,
b
9
9 Avenue Jean-Baptiste Clément 93430 Villetaneuse, France
Received 20 October 2006; received in revised form 3 January 2007; accepted 8 January 2007
Available online 19 January 2007
Abstract
Alzheimer's disease is characterized by the presence of β-amyloid fibril formation. The inhibition of this peptide accumulation may be a
prevention method for Alzheimer's disease. Several classes of molecules have been reported to inhibit β-amyloid fibril formation and among them
carbazoles. However, very few studies have been performed to determine the destination of such molecules in vivo and especially if they can pass
the blood brain barrier. The aim of this paper is to study whether carbazoles could pass the blood brain barrier, i.e. if they can circumvent ATP
Binding Cassette (ABC) transporters such as P-glycoprotein (P-gp) and Multidrug Resistance-associated protein (MRP1) which efficiently limit
drug brain uptake. For this purpose we have synthesized a fluorescent derivative of carbazole benzothiazolium iodide 1,2 disubstituted ethylene
(
referred as carbazole thiazole: CT), which can be easily detected and followed in the pre-trial study phases in cells or in tissue. We use cellular
models overexpressing P-gp and MRP1. Our results show that: i) CT is able to cross membranes and to penetrate rapidly inside the cells, ii) CT is
a P-gp substrate and consequently its accumulation in P-gp overexpressing cells is very low, iii) CT is a poor MRP1 substrate. In addition once
inside the cells, CT rapidly binds to DNA and is then slowly reduced by intracellular reducing agents. In conclusion, the efficiency of carbazole
derivatives in inhibiting the β-amyloid formation in vivo could be highly compromised because, as P-gp substrates, they will probably not cross
the blood brain barrier.
©
2007 Elsevier B.V. All rights reserved.
Keywords: Alzheimer; Blood brain barrier; β-amyloid; Inhibitor; Carbazole; P-glycoprotein; MRP1; Membrane
1
. Introduction
(Howlett et al., 1999). These molecules have been mainly
studied in vitro but the possibility that they can be used in vivo
has not yet received much attention. Furthermore, the study of
the behaviour of such molecules with respect to their transport
and destination inside the cells is a major point and a pre-
requisite for further studies.
Potential inhibitors of Alzheimer β-amyloid fibril formation
have to penetrate into the cerebral tissue. The transport of most
compounds from blood to brain is restricted by the presence of
the blood brain barrier. Three type of cells are involved in this
barrier: endothelial cells of brain capillaries, astrocytes and
pericytes (Ballabh et al., 2004). Particular properties of
endothelial cells produce this barrier effect: they present more
complex tight junctions than in other capillary endothelia and
express in their membranes efflux transporters (Abbott, 2002).
β-amyloid fibril formation and accumulation in the brain is
one of the main features of Alzheimer's disease as well as the
neurofibrillar degeneration (Chang and Silverman, 2004). It was
shown that the inhibition of this peptide accumulation may be a
way for preventing Alzheimer's disease. Several series of
compounds inhibiting efficiently β-amyloid fibril formation
have been reported (Gianni et al., 1995; Camilleri et al., 1994;
Sadler et al., 1995; Wood et al., 1996; Salomon et al., 1996;
Tomiyama et al., 1996; Howlett et al., 1997; Sastre et al., 2004;
Yang et al., 2005). Among them are carbazole-type compounds
⁎
Corresponding author. Tel.: +33 1 48 38 77 48; fax: +33 1 48 38 88 88.
E-mail address: m.salerno@smbh.univ-paris13.fr (M. Salerno).
0014-2999/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejphar.2007.01.005

C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
125
These membrane proteins function as pumps which extrude
exogenous compounds out of the brain tissue. Two of these
membrane proteins are the P-glycoprotein (P-gp) and Multidrug
Resistance-associated Protein (MRP1) (Amakrishnan, 2003).
Both transporters belong to the ATP binding cassette (ABC)
protein family, P-gp is also called ABCB1 and MRP1, ABCC1
show that CT is able to penetrate into cells. However, it is a P-gp
substrate which probably compromises its ability to cross the
blood brain barrier.
2. Materials and methods
(
Leslie et al., 2005).
2.1. Drugs and chemicals
In order to test the ability of one class of β-amyloid fibril
formation inhibitors, carbazole, to go through the blood brain
barrier we have synthesized carbazole thiazole (CT), a red
fluorescent salt of carbazole. This fluorescent molecule shows a
maximum emission peak around 590 nm. It makes the agent
easily detected and followed in the pre-trial study phases in cells
or in tissues. To follow the accumulation of this compound
inside cells, we used cellular models overexpressing P-gp and
MRP1, two proteins overexpressed in blood brain barrier. We
Carbazole thiazole (CT) was synthesized using the Knoeve-
nagel condensation between N-ethyl-2-methyl benzothiazolium
iodide (compound A) and N-ethyl-carbazole-3-carbaldehyde
(compound B) (Aldrich, Saint Quentin Fallavier, France) ac-
cording to the reaction scheme described in Fig. 1. The com-
pound Awas prepared at 110 °C, in a sealed tube, during 18 h, by
quaternization of 2-methyl benzothiazole (Aldrich, Saint
Quentin Fallavier, France) by a two-fold excess of ethyl iodide
1
3
3
Fig. 1. Synthesis scheme used to obtain the carbazole thiazole (CT) and C NMR spectrum obtained at 50 MHz in CDCl . (A) N-ethyl-2-methyl benzothiazolium
iodide and (B) N-ethyl-carbazole-3-carbaldehyde.

126
C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
(
reaction 1, Fig. 1). The compound A was recovered as bluish
2.3. Cell nucleus isolation
needles after recrystallization in hot water with a yield of 70%. In
a second step (reaction 2, Fig. 1), the Knoevenagel carbonyl-
methylene condensation between A and B was achieved using
reflux for 1 h in ethyl alcohol in the presence of sodium meth-
ylate as a catalyst. The resulting 1-(N-ethyl-2-benzothiazolium
iodide)-2-(N-ethyl-3-carbazole) ethylene (CT) was obtained as
bright red crystals after recrystallization in chloroform, with a
6
Cells (10 /mL) were suspended in Hepes buffer and
sonicated in a cold bath 3 times for 30 s (50 HZ, 117 v and
80 W) (Vibra Cell, Sonics and Materials Inc., Danbury, CT).
Nuclei were separated form other organelles in the cell by low
speed centrifugation and further purified by repeated washes in
the same buffer (Farrell, 1998).
1
13
yield of 82%. Both the H and C NMR spectra of CT are in
1
3
agreement with the expected structure. The 50 MHz C NMR
spectrum in deuterated chloroform, is shown in Fig. 1 with the
spectral assignments. Pure CT was dissolved in water and the
concentration was determined by U.V.-visible absorption
2.4. Cellular energy depletion
For energy depletion, cells were incubated for 30 min in a
Hepes buffer in which glucose was substituted by 2-deoxyglu-
cose, in the presence of 10 mM sodium azide (NaN ) (Marbeuf-
−
1
−1
spectroscopy with ε450=24 300 M cm . The fluorescence
3
spectrum of aqueous solution of CT exhibits a signal at around
Gueye et al., 1998) prior to addition of CT. The viability of the
cells, tested by Trypan Blue exclusion, was always greater than
95%.
5
90 nm within the red region of the visible domain.
FCCP and verapamil were purchased from Sigma (Saint
Quentin Fallavier, France). They were dissolved in ethanol. The
-[4-(diphenylmethyl)-1-piperazinyl]ethyl-5-(trans-4,6-di-
2
2.5. Real-time fluorescence and absorption measurements of
drug transport in living cells
methyl-1,3,2-dioxaphosphorinan-2-yl)-2,6-dimethyl-4-(3-nitro-
phenyl)-3-pyridinecarboxylate P oxide (PAK-104P) was a gift
from Drs. Shudo, Iwasaki and Akiyama (Nissan Chemical
Industries, LTD, Japan). Experiments were performed in Hepes
buffer containing 20 mM Hepes, 132 mM NaCl, 3.5 mM KCl,
Fluorescence measurements were carried out using a Perkin
Elmer LS50B spectrofluorometer equipped with a temperature-
controlled sample compartment. Before the experiments, the
cells were counted, centrifuged and resuspended in Hepes
buffer. The appropriate concentration of CT was quickly added
to the cuvette under magnetic stirring. Fluorescence intensity at
590 nm (excitation 470 nm) was measured continuously. During
the time course of these experiments, aliquots were taken at
various time intervals and used as such for flow cytometry
measurements. A Becton Dickinson FACScan Flow cytometer
equipped with a spectra Physics argon-ion laser was used. The
fluorescence signal was gated on the forward angle light scatter
signal to exclude dead cells debris from analysis. The argon-ion
laser was tuned to 488 nm and used at a power of 15 mW. The
emission was detected through an emission filter which collects
radiations from 563 to 607 nm. The cells were not washed
in order to minimize the re-equilibration of the fluorescent
probe in the various intracellular compartments. Re-equilibra-
tion happens when probes from the extra-cellular medium are
removed.
1
mM CaCl , 0.5 mM MgCl at pH=7.3, with or without 5 mM
2 2
glucose. All the reagents were of the highest quality available
and deionized double-distilled water was used throughout the
experiments.
Calf-thymus DNA (Aldrich) was used without purification.
The concentration was given by absorption spectra with ε260
=
1
3200/base pair.
2
.2. Cell lines and cultures
Two cell lines were used: K562 and GLC4. The K562 cell
line is a highly undifferentiated erythroleukemia originally
derived from a patient with chronic myelogenous leukemia
(
Lozzio and Lozzio, 1975). The GLC4 cell line was derived
from pleural effusion of a patient with small cell lung carcinoma
Zijlstra et al., 1987). The K562 leukemia cells and the P-gp
expressing K562/ADR cell were obtained from Prof. J.P. Marie
Hopital Hotel-Dieu, Paris, France). The GLC4 and the MRP1-
(
(
Absorption measurements were carried out using a Varian
Cary 219 double beam spectrophotometer. Cells were treated as
described above for the fluorescent measurements. Absorption
spectra were recorded at various intervals of time.
expressing GLC4/ADR cells were obtained from Prof. E. G. E.
de Vries (Department of Internal Medicine, University Hospital,
Groningen, The Netherlands). The K562 and GLC4 cells were
cultured in RPMI 1640 Medium with GlutaMAX™I (GIBCO)
medium supplemented with 10% fetal calf serum (GIBCO) at
2.6. DNA binding studies
3
7 °C in a humidified incubator with 5% CO . The resistant
2
K562/ADR and GLC4/ADR cells were cultured with 400 nM or
Fluorescence spectroscopy and circular dichroism were used
to monitor the interactions of CT with DNA. Absorption and
circular dichroism measurements were made on a Jobin Yvon
Model Mark CD-6 dichrograph. The instrument was calibrated
1
.2 μM doxorubicine respectively until three weeks before
5
experiments. Cultures initiated at a density of 10 cells/mL grew
exponentially to about 10 cells/mL in 3 days. In order to have
6
−
3
enough cells in the exponential growth phase for assay, culture
using a standard solution of epiandrosterone (3.4×10 ) in a
5
−1
−1
was initiated at 5×10 cells/mL and allowed to grow for 24 h
1 cm cell (Δε=3.3 M cm at 304 nm). All spectra were
recorded using a 1 cm cell and the following parameters:
λ=280–700 nm, step 1 nm, speed 0.25 s, spectra bandwidth
2 nm, number of cycles 1. Water blanks were used as references.
until use. Cultured cells were counted with a Coulter counter
before use. The viability of the cells, tested by Trypan Blue
exclusion, was always greater than 95%.

C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
127
Results were expressed as ε (molar absorption coefficient) and
Δε (differential molar absorption coefficient). The values of ε
and Δε were expressed as molar concentration of CT
chromophore. Uncorrected fluorescence spectra were recorded
at 37 °C on Perkin Elmer LS 50B spectrofluorometer.
∼15 min, the second step (2) is a decrease of fluorescence
intensity during the following ∼100 min. However, the addition
of resistant cells to a CT solution did not give rise to noticeable
variation of the fluorescence intensity. These preliminary data
already suggest that CT enters into sensitive cells but not
resistant cells. To give further support to these data, sensitive
cells were incubated for 15 min with CT in order to reach the
end of step (1). Then cells were centrifuged and new sensitive
cells were added to the supernatant. No increase of the
fluorescence signal was observed. This shows that CT, which
was formerly in solution, was absorbed by the cells during the
first experiment. When the same protocol was performed with
resistant cells, it was observed that the addition of sensitive cells
to the supernatant of resistant cells yields a strong increase of
the fluorescence signal. It proved that almost no CT has entered
the resistant cells.
3
. Results
The CT structure and its NMR 13C spectrum are shown in
the Fig. 1.
3
.1. Time course of CT uptake by K562 and K562/ADR cells
6
Fig. 2A shows typical result obtained when K562 cells, 10 /
mL, either sensitive or resistant were incubated with 5 μM CT.
The free CT has a low fluorescence signal at 590 nm
(
λex =470 nm). Addition of sensitive cells showed a two-step
In a second set of experiments, resistant cells were first
variation of fluorescence intensity as a function of time: the first
incubated for 30 min with 10 mM NaN , 5 mM deoxyglu-
3
step (1) is an increase of fluorescence intensity during the first
cose, in the absence of glucose. It is well-known that, under
6
Fig. 2. Time course of carbazole thiazole (CT) uptake in K562 and K562/ADR cells. Cells (10 /mL) were incubated at 37° with 5 μM CT. Fluorescence intensity at
590 nm (excitation 470 nm) was measured continuously as a function of time. The data points are from one representative experiment out of 4. (A) in Hepes buffer, (B)
in Hepes buffer in the absence of glucose and presence of 10 mM NaN plus 5 mM deoxyglucose, (C) in Hepes buffer in the presence of 10 μM PAK-104P.
3

128
C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
These two last sets of experiments strongly support
the assumption that in resistant cells, CT is pumped out by
P-gp.
3.2. Time course of CT uptake by GLC4 and GLC4/ADR cells
6
The Fig. 3 shows a typical result when GLC4 cells, 10 /mL,
either cells which overexpress MRP1 (GLC4/ADR) or not
GLC4), were incubated with 5 μM CT. The uptake of CT by
(
GLC4/ADR cells was slightly lower than what was obtained
with sensitive ones. In both cases, a first increase of the
fluorescent signal was followed by a decrease of this signal as in
the case for K562 cells.
3.3. Time course of CT uptake by K562 cells under conditions
of decreased membrane potential
Fig. 3. Time course of carbazole thiazole (CT) uptake in GLC4 and GLC4/ADR
cells. Cells (10 /mL) were incubated in Hepes buffer at 37° with 5 μM CT.
Fluorescence intensity at 590 nm (excitation 470 nm) was measured
continuously as a function of time. The data points are from one representative
experiment out of 4.
The equalization of the intra- and extracellular concentra-
tions of potassium was performed by incubating cells in Hepes
buffer in which Na was substituted by K (135 mM). Such
6
+
+
these experimental conditions, the ATP level is very low
(
Marbeuf-Gueye et al., 1998) and P-gp cannot pump out
molecules from the cells. When CT was added to the energy-
depleted K562/ADR cells, the same two steps variation of the
fluorescence intensity as with sensitive cells were obtained.
This treatment did not modify CT uptake in K562 cells
(
Fig. 2B).
In a third set of experiments, K562/ADR cells were incu-
bated with 10 μM PAK-104P together with CT. The PAK-104P
is a very efficient P-gp inhibitor (Marbeuf-Gueye et al., 2000).
A two-steps variation of the fluorescence intensity similar to
that of sensitive cells was obtained. This treatment did not
modify CT uptake in K562 cells (Fig. 2C).
Fig. 5. (A) Interaction of carbazole thiazole (CT) with permeabilized cells or
with isolated cell nuclei. The percentage of fluorescence intensity at 590 nm has
6
been plotted as a function of time. ( ) 5 μM CT was added to cells (10 /mL)
▪
suspended in Hepes buffer at 37°, in the presence of 10 μl of triton X-100 4% to
permeabilize plasma membrane. (●) 5 μM CTwas added to nuclei isolated from
cells (10 /mL). (B) Fluorescence titration of carbazole thiazole (CT) interaction
6
Fig. 4. Time course of carbazole thiazole (CT) uptake in K562 cells under
with DNA. DNA was stepwise added to 5 μM CT in Hepes at 37°. The
fluorescence intensity signal is plotted as a function of the b.p. concentration.
These experiments were performed twice. The data points are from one
experiment.
conditions of decreased plasma membrane potential. 5 μM CT was added to
6
+
+
cells (10 /mL) suspended in either Na - or K -rich Hepes buffer at 37°. The data
points are from one representative experiment out of 3.

C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
129
experimental conditions almost completely depolarized the
plasma membrane (Saengkhae et al., 2003). The Fig. 4 shows
that under these experimental conditions CT uptake is higher in
noticeable modification of the extinction coefficient at the
wavelength of this peak was observed (Fig. 6A). Free CT is not
chiral and therefore does not exhibit CD signal. Addition of
DNA gave rise to the appearance of a CD signal (Fig. 6B). Δε at
464 nm has been plotted as function of the [b.p.]/[CT] ratio
value and 50% of the signal increase was obtained for [b.p.]/
[CT]=2.0±0.4.
+
+
cells suspended in Na -rich buffer than in K -rich buffer. This
proves that CT accumulation is sensitive to plasma membrane
potential. It is expected as CT is a lipophilic cation.
6
In a second experiment, K562 cells (10 /mL) were allowed
to accumulate CT in the presence of 1 μM FCCP, a
protonophore, which is known to rapidly depolarize the
mitochondrial membrane (Loetchutinat et al., 2003). The
fluorescence signal was similar to that observed in the absence
of FCCP. From this data one can infer that CT did not
accumulate inside the mitochondria. It suggests that once inside
the cytosol CT has a strong binding site and is therefore not
available for accumulation inside mitochondria.
3
.4. Time course of CT interaction with permeabilized cells and
with isolated cell nuclei
The following experiments were designed to explain the
two-step behavior of the fluorescence signal variation. In
experiment 1, in order to permeabilize the plasma membrane,
CT was added to either K562 or K562/ADR cells together with
1
0 μl of 4% Triton X-100. The fluorescence signal increased
immediately and was followed by a slow decrease of
fluorescence intensity similar to the second step (2) in intact
cells (Fig. 5A). A similar result was obtained when cells were
sonicated before the addition of CT: a fast increase of the
fluorescent signal followed by a slow decrease (data not
shown). In experiment 2, 5 μM CT was added to isolated cell
nuclei suspended in Hepes buffer. The fluorescent signal
increased at once to the value observed at the end of step one
in sensitive cells and then maintained plateaued (Fig. 5A).
These data strongly suggested that the increase of fluorescence
intensity during step one was due to CT interaction with DNA in
the nucleus.
3
.5. Interactions of carbazole thiazole (CT) with DNA: spectral
changes
These experiments were designed to determine if the
increase in the fluorescence signal observed when cells were
incubated with CT was due to its interaction with DNA.
3
.5.1. Fluorescence spectral change
Free CT has a very low fluorescence signal at 590 nm
(
λex =470 nm) and addition of DNA caused a large increase in
fluorescence intensity. Fluorescence titration of 5 μM CT with
DNA (0–100 μM b.p.) is shown in Fig. 5B. The fluorescence
signal intensity reached a plateau for a [b.p.]/[CT] ratio ∼100,
and 50% of the signal increase was obtained for [b.p.]/[CT]=
Fig. 6. Circular dichroism and absorption spectroscopy monitoring of carbazole
thiazole (CT) interaction with DNA. Circular dichroism (A) and absorption (B)
spectra of 30 μM CT in the presence of increasing amount of DNA, in Hepes
buffer at 37 °C: [b.p.]/[CT] is equal to 0 in (1); 5 in (2); 9 in (3); 14 in (4).
2
0±4.
3
.5.2. UV-visible absorption and CD spectral changes
(
(
C) Absorption spectroscopy monitoring of the interaction of carbazole thiazole
CT) with K562 cells. 5 μM CTwas added to cells (10 /mL) suspended in Hepes
Addition of DNA (0–600 μM b.p.) to CT resulted in a shift
of the (50 μM) CT spectrum to longer wavelengths: the
absorption band of CT at 456 nm moved to 480 nm. No
6
buffer at 37 °. The absorption spectrum was recorded after time intervals ranging
from 0 to 120 min.

130
C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
From these data, we can already assert that the first step (1),
which is observed when CT is incubated with cells, is due to its
interaction with DNA in the nucleus.
The development of molecules to be used as inhibitors of
Alzheimer β-amyloid fibril formation was a goal of research for
several years. Numerous series of compounds were described
that effectively inhibit β-amyloid fibril formation. These
include quaternary ammonium salts (Wood et al., 1996),
sulphonates dyes (Sadler et al., 1995), nicotine (Salomon et
al., 1996), rifamycins (Tomiyama et al., 1996), cyclodextrin
(Camilleri et al., 1994), haemin (Howlett et al., 1997),
anthracyclines (Gianni et al., 1995), curcumin (Yang et al.,
2005), carbazoles (Howlett et al., 1999), cystatin C (Sastre et al.,
2004). Most of these molecules are efficient in vitro but very
few studies have been performed in vivo because of the poor
passage of most of molecules through the blood brain barrier.
The blood brain barrier is located between the blood and the
brain extracellular space, in the endothelial cells of brain
capillaries joined by tight junctions (Ballabh et al., 2004). The
blood brain barrier is a major impediment to the entry of many
therapeutic drugs into the brain. A very significant number of
molecules, including many useful therapeutic drugs, have lower
brain permeability because these molecules are substrates for
the ABC-transporters which are present in the brain capillary
endothelial cells. These transporters efficiently remove the drug
from the central nervous system, limiting drug brain uptake
(Begley, 2004). The P-gp was the first of these ABC
transporters to be described, followed by the MRP1, and
more recently breast cancer resistance protein (BCRP). All are
expressed in the blood brain barrier and combine to reduce the
brain penetration of many drugs (Leslie et al., 2005;
Amakrishnan, 2003). The P-gp and MRP1 are located in the
luminal membranes of endothelial cells. The P-gp is involved in
the efflux of many amphipathic cationic drugs while MRP1
extrudes neutral and anionic compounds (Scherrmann, 2005).
Consequently, these ABC proteins help to protect the brain from
the potential toxicity of exogenous compounds. Drugs target-
ting brain should be designed to overcome these mechanisms.
In the context of research on Alzheimer's disease, we are
interested on the ability to go through the blood brain barrier of
β-amyloid fibril formation inhibitors and β-amyloid markers
(Darghal et al., 2006). As it was previously reported that
carbazoles can inhibit β-amyloid fibril formation (Howlett et
al., 1999), we wondered whether carbazoles would be able to
pass the blood brain barrier and if these compounds would be
substrates for P-gp and MRP1.
3
.6. Time course of CT uptake by K562 cells monitored by
absorption spectroscopy
6
The K562 cells (10 /mL) were incubated in Hepes buffer at
3
7° with 5 μM CTand the absorption spectrum was recorded as a
function of time (Fig. 6C). One observed a shift of the visible
absorption band which occurred during the first ∼20 min. This
can be assigned to the uptake of CT by cells and its interaction
with nucleus (step (1)). Then as time elapsed a decrease of the
visible absorption band together with the appearance of a new
band in the UV at ∼250 nm was observed. About 50% of the
signal modification was obtained after 1 h. We can already infer
that these spectral modifications can be related to the second step
(
2). The decrease of the visible band, concomitantly with the
appearance of a new one in the UV, is probably due to reduction
of the ethylene double bond between the benzothiazolium and
carbazole groups. It disrupts the electronic conjugation across
the whole molecule. Under these conditions the absorption in the
UV part of the spectrum arises from the carbazole moiety.
Actually, the addition of 15 μM sodium dithionite to 4 μM CT
gave rise, within ∼1 h, to the disappearance of the 460 nm
absorption band and the appearance of the 270 nm one.
3
.7. Reaction of CT with intracellular reducing agents
These experiments aimed to determine which intracellular
molecules were able to reduce CT. In the cytosol the most
abundant reducing agents are glutathione GSH and NADH. The
incubation of CT with 10 mM GSH did not give rise to any
spectral modification. However, when similar experiments were
performed with 90 μM NADH and 4 μM CT, in Hepes buffer at
7°, the CT visible absorption band decreased as time elapsed.
This was concomitant with the appearance of the band at
70 nm. When 1 CT molecule was reduced, ∼1.5 NADH
molecules were oxidized.
3
2
4
. Discussion
In developed countries the number of elderly people is
growing rapidly and, therefore, an increase in neurodegenera-
tive and cerebrovascular disorders causing dementia is ex-
pected. Alzheimer's disease is the most common cause of
dementia. Early recognition and intervention helps optimal care
of Alzheimer's patients. It delays the morbidity with this
progressive illness (Chang and Silverman, 2004). Alzheimer's
disease develops as a result of the progressive degeneration of
specific pathways in the brain. β-amyloid is a major protein
component of senile plaques in Alzheimer's disease, and is
neurotoxic when aggregated (Selkoe, 2003). The size of
aggregated β-amyloid responsible for the observed neurotox-
icity and the mechanism of aggregation are still under inves-
tigation. However, prevention of fibril formation still holds
promise as a mean to reduce β-amyloid neurotoxicity.
For this purpose, we have synthesized a carbazole derivative,
carbazole thiazole, a slightly fluorescent compound which
fluorescence is increased through interaction with cellular
components. Thanks to this property, we were able to follow its
evolution inside the cells.
The carbazole rings also present DNA binding properties.
The binding affinity depends on the substitution of the
carbazole ring (Sajewicz and Dlugosz, 2000). In some cases,
(pyrimidyl groups), the affinity and the effects may be com-
parable to those obtained with doxorubicin (Tanious et al.,
2000). Consequently, compounds derived from carbazole were
proposed to be potential anticancer agents (Chang and Silver-
man, 2004). Some of the carbazole derivative may present
fluorescence properties.

C. Saengkhae et al. / European Journal of Pharmacology 559 (2007) 124–131
131
Gianni, L., Bellotti, V., Gianni, A.M., Merlini, G., 1995. New drug therapy of
amyloidoses: resorption of AL-type deposits with 4′-iodo-4′-deoxydoxor-
ubicin. Blood 86, 855–861.
Howlett, D., Cutler, P., Heales, S., Camilleri, P., 1997. Hemin and related
porphyrins inhibit beta-amyloid aggregation. FEBS Lett. 417, 249–251.
Howlett, D., George, A., Owen, D., Ward, R., Markwell, R., 1999. Common
structural features determine the effectiveness of carvedilol, daunomycin
and rolitetracycline as inhibitors of Alzheimer β-amyloid fibril formation.
Biochem. J. 343, 419–423.
Our data show that the lipophilic cation carbazole thiazolium
enters rapidly into the cells by passive diffusion and that its
accumulation depends on the plasma membrane potential. Once
inside the cell CT does not accumulate inside the mitochondria,
as this could be expected, but binds to DNA in the nucleus,
yielding a strong increase of the fluorescence signal. The
binding of CT to DNA is fast. The limiting step in the increase
of the fluorescence signal is the crossing of the molecule
through the plasma membrane. The binding of CT to DNA was
expected. It is well documented that carbazoles bind to DNA
most of the time in the minor groove (Tanious et al., 2000). This
is corroborated by our findings with isolated nuclei and DNA.
In a second step CT is slowly reduced by probably NADH. This
leads to the disappearance of the absorption band at 470 nm due
to reduction of the double bond between carbazole and thiazole.
Concomitantly, an absorption band appears at 260 nm,
characteristic of unconjugated carbazole. Our data also show
that CT is a good P-gp substrate as CT is not detected in P-gp-
overexpressing cells. The addition of P-gp inhibitor allows CT
uptake by these cells. However, CT is not an MRP1 substrate.
We conclude that even if carbazole derivatives are, in vitro,
efficient in inhibiting β-amyloid formation, this efficiency
could be highly compromised: being P-gp substrates they will
probably not pass through the blood brain barrier. Of course this
result should be corroborated using an in vivo model such as
transgenic mice overexpressing β-amyloid.
Leslie, E.M., Deeley, R.G., Cole, S.P.C., 2005. Multidrug resistance proteins:
role of P-glycoprotein, MRP1, MRP2, and BCRP (ABCG2) in tissue
defense. Toxicol. Appl. Pharmacol. 204, 216–237.
Loetchutinat, C., Saengkhae, C., Marbeuf-Gueye, C., Garnier-Suillerot, A.,
2
003. New insights into the P-glycoprotein-mediated effluxes of rhoda-
mines. Eur. J. Biochem. 270, 476–485.
Lozzio, C.B., Lozzio, B.B., 1975. Human chronic myelogenous leukemia cell-
line with positive Philadelphia chromosome. Blood 45, 321–334.
Marbeuf-Gueye, C., Broxterman, H., Dubru, F., Priebe, W., Garnier-Suillerot,
A., 1998. Kinetics of anthracycline efflux from mutidrug resistance protein-
expressing cancer cells compared with P-glycoprotein-expressing cancer
cells. Mol. Pharmacol. 53, 141–147.
Marbeuf-Gueye, C., Salerno, M., Quidu, P., Garnier-Suillerot, A., 2000.
Inhibition of the P-glycoprotein-and multidrug resistance protein-mediated
efflux of anthracyclines and calceinacetoxymethyl ester by PAK-104P. Eur.
J. Pharmacol. 391, 207–216.
Sadler, I.I., Hawtin, S.R., Tailor, V., Shearman, M.S., Pollack, S.J., 1995.
Glycosaminoglycans and sulphated polyanions attenuate the neurotoxic
effects of beta-amyloid. Biochem. Soc. Trans. 23, 106S.
Saengkhae, C., Loetchutinat, C., Garnier-Suillerot, A., 2003. Kinetic analysis of
rhodamines efflux mediated by multidrug resistance protein (MRP1).
Biophys. J. 85, 2006–2014.
Sajewicz, W., Dlugosz, A., 2000. Cytotoxicity of some potential DNA
intercalators (carbazole, acridine and anthracene derivatives) evaluated
through neutrophil chemiluminescence. J. Appl. Toxicol. 20, 305–312.
Salomon, A.R., Marcinowski, K.J., Friedland, R.P., Zagorski, M.G., 1996.
Nicotine inhibits amyloid formation by the beta-peptide. Biochemistry 35,
Acknowledgements
We would like to thank the CNRS, Institut Fédératif de
Recherche-Paris Plaine de France and University Paris 13 for
financial support in the framework of the PPF-chimie program.
1
3568–13578.
Sastre, M., Calero, M., Pawlik, M., Mathews, P.M., Kumar, A., Danilov, V.,
Schmidt, S.D., Nixon, R.A., Frangione, B., Levy, E., 2004. Binding of
cystatin C to Alzheimer's amyloid beta inhibits in vitro amyloid fibril
formation. Neurobiol. Aging 25, 1033–1043.
References
Scherrmann, J.M., 2005. Expression and function of multidrug resistance
transporters at the blood–brain barriers. Expert Opin. Drug Metab. Toxicol.
1, 233–246.
Abbott, N.J., 2002. Astrocyte–endothelial interactions and blood–brain barrier
permeability. J. Anat. 200, 629–638.
Amakrishnan, P., 2003. The role of P-glycoprotein in the blood–brain barrier.
Einstein Q. J. Biol. Med. 19, 160–165.
Ballabh, P., Braun, A., Nedergaard, M., 2004. The blood–brain barrier: an
overview: structure, regulation, and clinical implications. Neurobiol. Dis.
Selkoe, D.J., 2003. Folding proteins in fatal ways. Nature 426, 900–904.
Tanious, F., Ding, D., Patrick, D., Bailly, C., Tidwell, R., Wilson, W., 2000.
Effects of compound structure on carbazole dication-DNA complexes: tests
of the minor-groove complex models. Biochemistry 39, 12091–12101.
Tomiyama, T., Shoji, A., Kataoka, K., Suwa, Y., Asano, S., Kaneko, H., Endo,
N., 1996. Inhibition of amyloid beta protein aggregation and neurotoxicity
by rifampicin. Its possible function as a hydroxyl radical scavenger. J. Biol.
Chem. 271, 6839–6844.
Wood, S.J., MacKenzie, L., Maleeff, B., Hurle, M.R., Wetzel, R., 1996.
Selective inhibition of Abeta fibril formation. J. Biol. Chem. 271,
4086–4092.
Yang, F., Lim, G.P., Begum, A.N., Ubeda, O.J., Simmons, M.R., Ambegaokar,
S.S., Chen, P.P., Kayed, R., Glabe, C.G., Frautschy, S.A., Cole, G.M., 2005.
Curcumin inhibits formation of amyloid beta oligomers and fibrils, binds
plaques, and reduces amyloid in vivo. J. Biol. Chem. 280, 5892–5901.
Zijlstra, J.G., De Vries, E.G.E., Mulder, N.H., 1987. Multifactorial drug
resistance in an adriamycin-resistant human small cell lung carcinoma cell
line. Cancer Res. 47, 1780–1784.
16, 1–13.
Begley, D.J., 2004. ABC transporters and the blood–brain barrier. Curr. Pharm.
Des. 10, 1295–1312.
Camilleri, P., Haskins, N.J., Howlett, D.R., 1994. Beta-cyclodextrin interacts
with the Alzheimer amyloid beta-A4 peptide. FEBS Lett. 341, 256–258.
Chang, C.Y., Silverman, D., 2004. Accuracy of early diagnosis and its impact on
the management and course of Alzheimer's disease. Expert Rev. Mol.
Diagn. 4, 63–69.
Darghal, N., Garnier-Suillerot, A., Salerno, M., 2006. Mechanism of thioflavin
T accumulation inside cells overexpressing P-glycoprotein or multidrug
resistance-associated protein: role of lipophilicity and positive charge.
Biochem. Biophys. Res. Commun. 343, 623–629.
Farrell Jr., R.E., 1998. Analysis of nuclear RNA. In: Farrell Jr., R.E. (Ed.), RNA
Methodologies: A Laboratory Guide for Isolation and Characterization.
Academic Press, San Diego, pp. 406–437.