Influence of lipid lowering fibrates on P-glycoprotein activity in vitro

Article abstract of DOI:10.1016/j.bcp.2003.09.008

Statin/fibrate combinations are frequently used to treat mixed dyslipidemia. However, these combinations may cause life-threatening drug interactions (e.g. rhabdomyolysis) possibly induced by modifications of cytochrome P450 isozyme activities. Some statins are also transported by P-glycoprotein (Pgp) and may act as inhibitors of this drug efflux pump. So far, nothing is known about possible Pgp modulating effects of fibrates. We tested whether gemfibrozil, fenofibrate, fenofibric acid, and bezafibrate inhibit Pgp in vitro using a calcein acetoxymethylester (calcein-AM) uptake assay and confocal laser scanning microscopy with bodipy-verapamil as substrate in L-MDR1 cells, which overexpress human Pgp. In uptake assays in cells with (L-MDR1) and without (LLC-PK1) human Pgp we also investigated whether these compounds are transported by Pgp. Intracellular concentrations were measured by liquid chromatography tandem mass spectrometry. Of the tested fibrates, only fenofibrate increased calcein-AM uptake into cells indicating an inhibition of Pgp mediated transport by this compound. The potency of fenofibrate (mean±SD: 7.1±3.2μM), evaluated by calculating the concentration needed to double baseline fluorescence (f2), was similar to that of simvastatin (5.8±1.5μM), lovastatin (10.1±1.0), and verapamil (4.7±0.8μM). For simvastatin and fenofibrate Pgp inhibition was confirmed with confocal laser scanning microscopy. Fenofibrate, fenofibric acid, gemfibrozil, and bezafibrate showed no difference in the cellular uptake between LLC-PK1 and L-MDR1, indicating that the tested fibrates are not Pgp substrates. In conclusion, this study demonstrates that fenofibrate inhibits Pgp in vitro with a potency similar to simvastatin.

Full text of DOI:10.1016/j.bcp.2003.09.008

Biochemical Pharmacology 67 (2004) 285–292
Influence of lipid lowering fibrates on P-glycoprotein activity in vitro
Manuela Ehrhardt, Heike Lindenmaier, Juergen Burhenne,
*
Walter Emil Haefeli, Johanna Weiss
Department of Internal Medicine VI, Clinical Pharmacology and Pharmacoepidemiology,
University of Heidelberg, Bergheimer Strasse 58, D-69115 Heidelberg, Germany
Received 28 February 2003; accepted 19 September 2003
Abstract
Statin/fibrate combinations are frequently used to treat mixed dyslipidemia. However, these combinations may cause life-threatening
drug interactions (e.g. rhabdomyolysis) possibly induced by modifications of cytochrome P450 isozyme activities. Some statins are also
transported by P-glycoprotein (Pgp) and may act as inhibitors of this drug efflux pump. So far, nothing is known about possible Pgp
modulating effects of fibrates. We tested whether gemfibrozil, fenofibrate, fenofibric acid, and bezafibrate inhibit Pgp in vitro using a
calcein acetoxymethylester (calcein-AM) uptake assay and confocal laser scanning microscopy with bodipy-verapamil as substrate in L-
MDR1 cells, which overexpress human Pgp. In uptake assays in cells with (L-MDR1) and without (LLC-PK1) human Pgp we also
investigated whether these compounds are transported by Pgp. Intracellular concentrations were measured by liquid chromatography
tandem mass spectrometry. Of the tested fibrates, only fenofibrate increased calcein-AM uptake into cells indicating an inhibition of Pgp
mediated transport by this compound. The potency of fenofibrate (mean ꢀ SD: 7:1 ꢀ 3:2 mM), evaluated by calculating the concentration
needed to double baseline fluorescence (f2), was similar to that of simvastatin (5:8 ꢀ 1:5 mM), lovastatin (10:1 ꢀ 1:0), and verapamil
(
4:7 ꢀ 0:8 mM). For simvastatin and fenofibrate Pgp inhibition was confirmed with confocal laser scanning microscopy. Fenofibrate,
fenofibric acid, gemfibrozil, and bezafibrate showed no difference in the cellular uptake between LLC-PK1 and L-MDR1, indicating that
the tested fibrates are not Pgp substrates. In conclusion, this study demonstrates that fenofibrate inhibits Pgp in vitro with a potency similar
to simvastatin.
#
2003 Elsevier Inc. All rights reserved.
Keywords: Fibrates; P-glycoprotein; In vitro; Drug interactions; Calcein-acetoxymethylester; Inhibition
1
. Introduction
reduce cholesterol synthesis fibrates increase catabolism
of triglyceride rich lipoproteins and increase HDL-choles-
terol. For both, statins and fibrates, drug-induced myopathy
is a rare but well documented adverse reaction [1] and the
risk for progression to life-threatening rhabdomyolysis is
increased when statins and fibrates are used in combination
[1–3]. The incidence of these interactions depends on the
individual drug and the drug combination used as shown in
an analysis of all fatal cases of rhabdomyolysis reported to
the American Food and Drug Administration (FDA) until
June 2001. In this series no case of fatal rhabdomyolysis for
fluvastatin was reported whereas this rate was 3.16 per
million cerivastatin prescriptions [3]. Twelve of the 31
cases reported to the FDA occurred during combined use of
cerivastatin and gemfibrozil. The most common source for
such drug–drug interactions is thought to be the cyto-
chrome P450 (CYP) system. Whereas pravastatin is not
metabolized by the CYP system simvastatin, lovastatin,
and atorvastatin are mainly metabolized by the CYP3A4
Lipid lowering fibrates are an established class of drugs
for the treatment of hypertriglyceridemia. Combined with
-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA)
3
reductase inhibitors (statins), which are used primarily
to treat hypercholesterolemia, they represent an attractive
approach for the treatment of mixed dyslipidemia because
of their different mechanisms of action. While statins
*
Corresponding author. Tel.: þ49-6221-565582;
fax: þ49-6221-564642.
E-mail address: johanna_weiss@med.uni-heidelberg.de (J. Weiss).
Abbreviations: APCI, atmospheric pressure chemical ionization; cal-
cein-AM, calcein-acetoxymethylester; CYP, cytochrome P450; DMSO,
dimethyl sulfoxide; ESI, electrospray ionization; f2, concentration needed
to double baseline fluorescence; HBSS, Hanks’ balanced salt solution;
HHBSS, HBSS supplemented with 10% HEPES; IC50, concentration
inducing half maximal effect; LC/MS/MS, liquid chromatography tandem
mass spectrometry; LDH, lactate dehydrogenase; MRP, multidrug
resistance-associated protein; Pgp, P-glycoprotein.
0006-2952/$ – see front matter # 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2003.09.008

286
M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
isoform, fluvastatin by CYP2C9, and cerivastatin by
CYP3A4 and CYP2C8 [4,5]. Also gemfibrozil, fenofibrate,
and bezafibrate are metabolizes via CYP3A4 [6]. Even
though gemfibrozil elevates the risk to develop statin-
induced muscle disorders it does not inhibit CYP3A4 in
human liver microsomes suggesting a different mechanism
of drug interaction [7]. Another possible interaction site is
Pgp, a member of the superfamily of ATP binding cassette
acid for analysis, hydrochloric acid, and sodium hydroxide
were from Merck. Cell culture medium M199, fetal calf
serum (FCS), glutamine, 100 U/mL penicillin with 100 mg/
mL streptomycin sulfate, HBSS, HEPES, and versene were
from Invitrogen. Collagen-R was obtained from Serva.
DMSO and Triton X-100 were purchased from Appli-
Chem, calcein-AM from MoBiTec, 48-well plates from
2
Falcon, and 96-well plates and 80 cm culturing bottles
from Nunc.
(ABC) transporters. This efflux transporter is mainly
expressed at the blood–brain barrier [8] and the intestinal
barrier [9] and participates in the renal and biliary elim-
ination of drugs [10,11]. It prevents intracellular drug
accumulation by actively removing compounds from the
cell. Therefore, Pgp plays a role in absorption, distribution,
and excretion of certain drugs and may be an important
site of drug–drug interactions. Indeed, in vitro several
statins are inhibitors of Pgp [12–15], and many of them
are also substrates of the efflux pump [13,16] suggesting
that drug interactions might also occur at the level of
Pgp. Thus far, the interaction of fibrates with Pgp has
not been evaluated.
2.3. Isolation procedures
2.3.1. Simvastatin
According to Bogman et al. [14] 50 tablets Zocor forte
containing 40 mg simvastatin from Merck Sharpe &
Dohme Limited were mortared and stirred 2 hr with
100 mL of an extraction solution containing methanol,
hexane, and methylene chloride in equal volumetric parts.
The extraction solution was filtered and evaporated to
dryness under a stream of nitrogen at 508. The remaining
residue was weighed and redissolved in acetonitrile/water
1
The aim of the present study was to assess the Pgp
modulating effect of fibrates and to compare it with the
known Pgp inhibitor effects of simvastatin and lovastatin.
Therefore, we investigated the Pgp inhibitory potency of
lipid lowering fibrate derivatives in vitro in a porcine
kidney epithelial cell line overexpressing human Pgp (L-
MDR1) [17] and the corresponding parental cell line
and deuterated DMSO (d -DMSO). Subsequently, the
3
solutions were analyzed to determine purity and identity
by HPLC and spectroscopic measures.
2.3.2. Fenofibric acid
An amount of 1.06 g of fenofibrate was dissolved in a
mixture of 8 mL acetonitrile and 4 mL water. After addi-
tion of 40 pellets sodium hydroxide the mixture was heated
to 908 in a water bath for 90 min. After 1 hr, 2 mL water
and 40 mL 1 M hydrochloric acid were added. For separ-
ating the two phases the mixture was allowed to cool down
and further 20 mL hydrochloric acid were added. Feno-
fibric acid was precipitated from the aqueous phase by
acidification with 1 M hydrochloric acid (200 mL). The
precipitate was centrifuged for 10 min at 3000 g and was
dried at 908. The remaining residue was also weighed and
(LLC-PK1) as a negative control without the human
ABC transporter. In particular we studied the lipid low-
ering active parent compounds gemfibrozil and bezafi-
brate, and the inactive prodrug fenofibrate with its active
metabolite fenofibric acid. For this assay we used the
fluorogenic calcein-AM as Pgp substrate and confirmed
the results using a different technique (confocal laser
scanning microscopy) with an independent Pgp substrate
(bodipy-verapamil). Furthermore, intracellular drug accu-
mulation of each fibrate in L-MDR1 and LLC-PK1 cells
was measured directly by LC/MS/MS to assess whether
test compounds are Pgp substrates.
redissolved in acetonitrile/water and d
mine purity and identity.
-DMSO to deter-
3
2.4. Purity and identity
2
. Material and methods
Purity of isolated simvastatin and fenofibric acid was
determined by HPLC/UV (215 and 254 nm) using a pump
gradient program (100% aqueous eluent to 100% organic
eluent). The purity was >99% for either compound. Iden-
tity of both isolated compounds was confirmed by ESI LC/
MS/MS in the daughter ion scanning mode and nuclear
magnetic resonance (NMR) spectroscopy (one-dimen-
sional H and 13C, as well as two-dimensional HH-Cosy
and CH-Cosy).
2
.1. Drugs
Gemfibrozil, fenofibrate, bezafibrate, lovastatin, loper-
amide, and verapamil hydrochloride were obtained from
Sigma-Aldrich. Vincristine sulfate was purchased from
Calbiochem and LY335979 was a generous gift from Eli
Lilly Co.
1
2
.2. Solvents, chemicals, and cell culture accessory
2.5. Stock solutions
Methanol Uvasol, hexane Uvasol, methylene chloride
Uvasol, acetonitrile LiChrosolv, water LiChrosolv, formic
Stock solutions of test compounds were prepared fol-
lowing the manufacturers’ instructions. All compounds

M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
287
were dissolved in DMSO. The DMSO concentration in the
assays never exceeded 1% (v/v), a concentration that was
found not to influence the results of the assay in pilot
experiments.
detached after 2, 10, 30, or 60 min (sextuplet determina-
tion). The cells were washed twice with precooled (48)
HHBSS and lyzed with 1% Triton X-100 within 30 min.
After adding internal standard the cell lysates were directly
analyzed by LC/MS/MS.
2
.6. Calcein uptake assay
2.9. Analytical methods
The assay was used to determine the Pgp inhibitor
potency of the test compounds and was performed as
previously described [18] with calcein-AM in a final
concentration of 0.5 mM.
Quenching test and cytotoxicity assay were performed
for each test compound as previously described [18,19].
For the quantification of intracellular drug concentra-
tions in the uptake assays methods depending on LC/MS/
MS were developed. The mass spectrometric analyses
were performed on a triple stage quadrupole instrument
(TSQ 7000, Thermo Finnigan). Sample admission was
done by an autosampler Type 232XL from Gilson Abimed
and an LC gradient pump (P4000 Thermo Finnigan) was
used with a degasser. Chromatography was performed on a
Kromasil RP18 CC 125/3; 100-3.5 analytical column
(Macherey und Nagel) using an isocratic eluent (flow
0.35 mL/min) consisting of water with 1% formic acid
(eluent A) and acetonitrile (eluent B) in different volu-
metric parts depending on the drug (Table 1). For detection
of fenofibrate, fenofibric acid, bezafibrate, and loperamide
[20] single reaction monitoring (MS/MS) mode with elec-
2
.7. Confocal laser scanning microscopy
Intracellular accumulation of the dye bodipy-verapamil
in the presence of the test compounds was analyzed with a
DM IRE 2 TCS SP II confocal laser scanning microscope
from Leica as described earlier [19]. Bodipy-verapamil
instead of calcein-AM was chosen as a substrate to confirm
the interaction with calcein with an independent Pgp
substrate. In addition, bodipy-verapamil provides the
advantage to be more resistant to photobleaching effects
caused by the laser.
trospray ionization (ESI, 4.5 kV, Sheath/Aux gas N ) and
2
argon as collision gas (3 mbar) was used. Gemfibrozil was
detected in single ion monitoring mode with atmospheric
pressure chemical ionization (APCI, 5.0 mA, 5008, Sheath/
2
.8. Uptake assay
Aux gas N ). Details of the LC and mass spectrometric
parameters are shown in Table 1.
2
To evaluate the substrate characteristics L-MDR1 and
LLC-PK1 cells were incubated with the test substance
100 mM for gemfibrozil, fenofibric acid, and bezafibrate
and 25 mM for fenofibrate) with and without Pgp inhibitor
LY335979). To demonstrate the suitability of the assay,
The calibration range for gemfibrozil, fenofibric acid,
and bezafibrate was 0.1–200 mM, for fenofibrate 0.025–
50 mM and 0.0025–0.5 mM for loperamide. The lower limit
of quantification was 0.1 mM for gemfibrozil, fenofibric
acid, and bezafibrate, 0.025 for fenofibrate, and 0.0025 mM
for loperamide.
(
(
loperamide (1 mM) was used as a Pgp substrate in a
separate experiment in both cell lines. One day before
performing the assay culture medium was replaced by
vincristine-free medium. The assay was conducted in
collagen-coated 48-well microtiter plates (150,000 cells/
2.10. Statistical analysis
2
cm ). All incubation steps and cell lysis were performed at
Due to cytotoxic influences, plateau effects and thus IC50
values (the concentration leading to half-maximal inhibi-
tion of the calcein-AM transport) could not be determined
for all test compounds. Based on the concentration–
response curve we therefore also calculated the concentra-
tion needed to double baseline fluorescence (f2) according
to Weiss et al. [18]. Unless indicated otherwise data are
expressed as mean ꢀ SD. For the f2 values P-values were
room temperature on a rotary shaker at 450 rpm. The cells
were washed with prewarmed HHBSS and preincubated
for 1 hr with HHBSS. Subsequently, HHBSS as a control
or LY335979 (final concentration 5 mM in uptake experi-
ments with fibrates and 1 mM for uptake experiments with
loperamide) as a Pgp inhibitor was added for 15 min. The
test compound was applied and the supernatant was
Table 1
LC/MS/MS parameters for quantification of the intracellular drug uptake
Analyte
Eluent, %A/%B
Rt (min)
Parent mass (m/z)
Center mass (m/z)
Ionization
Collision energy (V)
Gemfibrozil
Fenofibrate
Fenofibric acid
Bezafibrate
Loperamide
30/70
30/70
50/50
60/40
55/45
2.01
3.69
3.45
3.31
2.55
249.2
316.2
316.9
362.2
477.2
–
APCI negative
ESI positive
ESI positive
ESI positive
ESI positive
–
22
22
28
28
233.1
230.8
316.1
266.3
A ¼ 1% formic acid; B ¼ acetonitrile; Rt ¼ retention time.

288
M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
determined by ANOVA with Dunnett’s multiple compar-
ison test for post hoc pairwise comparison with the control
results obtained with verapamil. For the uptake assay, P-
values were determined by ANOVA with Bonferroni’s
multiple comparison test for post hoc pairwise comparison.
P-values for the experiments with the confocal laser
scanning microscope were calculated by one-tailed t-test.
All statistical analysis was performed with GraphPad
InStat, version 3.05, GraphPad Software. A P-value of
14
12
Simvastatin
Lovastatin
10
8
6
4
2
10-3 10-2 10-1 100 101 102 103
ꢁ
0.05 was considered significant.
Compound [µM]
3
. Results
Fig. 2. Concentration response curve of simvastatin and lovastatin
obtained with calcein assay in L-MDR1 cells. Data are expressed as
mean ꢀ SD for N ¼ 8 wells.
3
.1. Pgp inhibitor potency
None of the test substances showed any quenching effect
potent than verapamil (P < 0:01). Except for simvastatin,
which showed a minor effect also in LLC-PK1 cells, all
other compound tested had no effect in the control cell line.
For simvastatin and fenofibrate the results obtained in
the calcein assay were further evaluated with an indepen-
dent assay based on confocal laser scanning microscopy
with bodipy-verapamil as a substrate confirming the Pgp
inhibitor activity of both compounds (Fig. 3). In this assay
fenofibrate significantly increased basal fluorescence by
1:46 ꢀ 0:12 (25 mM) and 1:69 ꢀ 0:19 (50 mM) in L-MDR1
cells (P < 0:05) whereas in the control cell line LLC-PK1
no effect was observed (Fig. 3). Simvastatin (25 mM)
significantly increased basal fluorescence in L-MDR1 cells
by 1:79 ꢀ 0:34 (P < 0:05) (Fig. 4), but also in LLC-PK1
cells by 1:19 ꢀ 0:06 (P < 0:05). The effect seen in LLC-
PK1 cells was clearly smaller than in the overexpressing
cell line (P < 0:05).
in the concentration range tested on the fluorescence of
calcein and no autofluorescence at the excitation wave-
length used to measure calcein fluorescence (485 nm) was
observed. Moreover, no self-quenching effects of calcein
were observed in the concentration range applied.
The two paradigm Pgp inhibitors verapamil and
LY335979 and the fibrates were not cytotoxic in the
concentration range used in the assays whereas the statins
were cytotoxic in concentrations higher than 100 mM
(
lovastatin) and 35 mM (simvastatin) in the LDH detection
assay and simvastatin as well as in confocal laser scanning
microscopy (manifesing in deformations of the cells).
Of all tested compounds simvastatin, lovastatin, and
fenofibrate showed an inhibitory effect on Pgp in the
calcein assay. In contrast to fenofibrate, where a plateau
in the concentration–response curve was reached (Fig. 1),
for simvastatin and lovastatin (Fig. 2) this was impossible
(due to cytotoxic effects), and thus no IC50 could be
3.2. Pgp substrate characteristics
calculated. On the basis of f2-values the potency of feno-
fibrate was similar to the potency of the statins and
verapamil, a well-known Pgp inhibitor (Table 2). Compar-
ing the IC50 values, fenofibrate was even 1.5 times more
To investigate whether the fibrates are transported by
Pgp an uptake assay in LLC-PK1 and L-MDR1 cells was
performed. First, the suitability of the uptake assay for
Table 2
Inhibition of Pgp by lipid lowering fibrates and typical Pgp inhibitors in L-
1.00
MDR1 cells
Fenofibrate
Test compound
N
f2 (mM)
IC50 (mM)
0
.75
Fenofibric acid
Gemfibrozil
Fenofibrate
Fenofibric acid
Bezafibrate
Simvastatin
Lovastatin
4
4
3
4
3
3
3
4
n.d.
n.d.
*
7.1 ꢀ 3.2
n.d.
5.83 ꢀ 2.3
n.d.
0.50
n.d.
n.d.
5.8 ꢀ 1.5
10.1 ꢀ 1.0
4.7 ꢀ 0.8
n.d.
*
*
n.d.
20.0 ꢀ 3.1
0.25
Verapamil
*
*
*
LY335979
0.02 ꢀ 0.01
0.14 ꢀ 0.06
1 -
0
3
10-2
10-1
Compound [µM]
100
101
102
N ¼ number of experiments, each performed in octuplet. Values are
given as mean ꢀ SD. P-values are determined by ANOVA with Dunnett’s
multiple comparison test for post hoc pairwise comparison of the results
with the verapamil control. f2 ¼ concentration needed to double baseline
fluorescence calculated according to [21]; n.d.: not definable.
Fig. 1. Concentration response curve of fenofibrate and fenofibric acid
obtained with calcein assay in L-MDR1 cells. Data are expressed as
mean ꢀ SD for N ¼ 8 wells.
*
**
P < 0:01; P < 0:05.

M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
289
Fig. 3. Confocal laser scanning microscopy: bodipy-verapamil uptake into L-MDR1 (A and B) and LLC-PK1 (C and D) cells. On the left: without
fenofibrate. On the right: after exposure for 20 min with fenofibrate (50 mM) as Pgp inhibitor.
evaluating Pgp substrate characteristics was tested using
the known Pgp substrate loperamide. The loperamide
uptake was significantly higher in the parental cell line
compared to L-MDR1 cells (P < 0:01) (Fig. 5A). Pgp
inhibition with LY335979 completely abolished drug resis-
tance of L-MDR1 cells leading to similar concentration–
time profiles as in the parental cell line (P < 0:001,
compared to L-MDR1 cells without LY335979). In
LLC-PK1 cells LY335979 had no statistically significant
effect on the loperamide uptake (Fig. 5A).
In this assay there was no difference between the two
cell lines in the intracellular uptake of fenofibrate (Fig. 5B),
Fig. 4. Confocal laser scanning microscopy: bodipy-verapamil uptake into L-MDR1 (A and B) cells. On the left: without simvastatin. On the right: after
exposure for 20 min with simvastatin (25 mM) as Pgp inhibitor.

290
M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
0
.5
.4
.3
.2
.1
.0
32
L-MDR1
L-MDR1 + LY335979
LLC-PK1
0
25
0
LLC-PK1 + LY335979
LLC-PK1 + LY335979
LLC-PK1
18
0
}
p<0.01
L-MDR1
L-MDR1 + LY335979
}
p<0.001
11
4
0
0
0
10 20 30 40 50 60 70
0
10 20 30 40 50 60 70
(A)
Time [min]
(B)
Time [min]
6
5
4
3
2
1
3
2
1
0
L-MDR1
L-MDR1
L-MDR1 + LY335979
L-MDR1 + LY335979
LLC-PK1
}
p<0.05 (30 min)
LLC-PK1
LLC-PK1 + LY335979
LLC-PK1 + LY335979
0
10 20 30 40 50 60 70
0
10 20 30 40 50 60 70
(C)
Time [min]
(D)
Time [min]
Fig. 5. Uptake assay of loperamide (1 mM) (A), fenofibrate (25 mM) (B), bezafibrate (100 mM) (C), and gemfibrozil (100 mM) (D) in LLC-PK1 cells and in
L-MDR1 cells with and without Pgp inhibition by LY335975 (1 mM for the uptake with loperamide; 5 mM for the uptake with the fibrates). Data are shown as
mean ꢀ SD for N ¼ 6 wells. P-values were determined by ANOVA with Bonferroni’s multiple comparison test for post hoc pairwise comparison.
bezafibrate (Fig. 5C), gemfibrozil (Fig. 5D) or fenofibric
acid (data not shown).
The aim of this in vitro study was to elucidate whether
fibrates are transported by Pgp and whether they inhibit this
drug transporter. An evaluation of both properties was
necessary because some Pgp substrates do not act as
Pgp inhibitors (e.g. digoxin) [18,26,27] and some inhibi-
tors are not transported [27]. All assays were performed
with the cell line L-MDR1 overexpressing human Pgp and
the corresponding native cell line LLC-PK1 without
human Pgp as a control. Thus, effects only seen in L-
MDR1 cells can be attributed to functional human
Pgp. Except simvastatin all compounds tested had no effect
in LLC-PK1 cells. The fact, that simvastatin not only
increases the uptake of bodipy-verapamil and of calcein-
AM in L-MDR1 cells, but also revealed a minor effect in
LLC-PK1 cells and of calcein-AM in LLC-PK1 (data not
shown) indicates that simvastatin not only inhibits Pgp, but
possibly also other transporters like multidrug resistance-
associated proteins (MRPs).
In agreement with earlier investigations simvastatin was
more potent an inhibitor of Pgp than lovastatin [14,15].
Moreover, our study demonstrates that gemfibrozil and
bezafibrate do not interact with Pgp, whereas fenofibrate
acts as a Pgp inhibitor. In line with the model established by
Seelig [28], all fibrates contain electron donor patterns,
which might interact with Pgp. However, whereas the other
fibrates are anionic and hydrophilic compounds which are
normally weak Pgp substrates if at all [10] only fenofibrate
is lipophilic and neutral which predicts high affinity to
Pgp. Therefore, our finding that fenofibrate inhibits Pgp
could be expected. It is, however, surprising that fenofi-
4
. Discussion
The activity of the efflux transporter Pgp affects the
pharmacokinetics of many drugs and contributes to numer-
ous pharmacokinetic drug–drug interactions [21]. Hitherto,
the role of Pgp for the bioavailability, distribution, and
excretionoflipidloweringfibrateshasnotbeeninvestigated.
Especially with statins severe interactions have occurred. A
well known example is the increased risk for rhabdomyo-
lysis emerging from the combination of cerivastatin with
gemfibrozil, which recently lead to the withdrawal of cer-
ivastatinfromtheEuropeanandUSmarket. Interestinglythe
incidence of these adverse drug reactions depends on the
individual statin and fibrate and the particular combination
chosen. For instance in the published cases of rhabdomyo-
lysis after a fibrate/statin combination between 1989 and
July 2000 gemfibrozil was incriminated in each case [22].
Furthermore, about 25% of the rhabdomyolysis related
mortality with cerivastatin was associated with cerivasta-
tin–gemfibrozil combination therapy [23]. Whereas part of
this interaction is likely caused by inhibition of CYP2C8
[7,24] cerivastatin and many other statins also interact with
Pgp [13–16,25]. A further explanation for the accumulation
of statin metabolites during co-administration of fibrates
could therefore be an interaction at the level of Pgp by which
statins are actively transported [13,16].

M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
291
[
[
[
3] Staffa JA, Chang J, Green L. Cerivastatin and reports of fatal rhabdo-
myolysis. N Engl J Med 2002;346:539–40.
brate appeared not to be transported, because no difference
between the uptake into the parental cell line (LLC-PK1)
and the overexpressing cell line (L-MDR1) was found
4] Bays HE, Dujovne CA. Drug interactions of lipid-altering drugs. Drug
Saf 1998;19:355–71.
(
Fig. 5B). Therefore, similar to verapamil, fenofibrate
5] Jacobsen W, Kirchner G, Hallensleben K, Mancinelli L, Deters M,
Hackbarth I, Benet LZ, Sewing KF, Christians U. Comparison of
cytochrome P-450-dependent metabolism and drug interactions
of the 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors lovas-
tatin and pravastatin in the liver. Drug Metab Dispos 1999;27:
might exhibit a high intrinsic permeability and pass
through the cell membrane faster than it is removed by
Pgp [27]. In this case, the transport will not be detectable in
an uptake assay.
1
73–9.
In agreement with the lack of Pgp inhibition there was no
difference in the uptake of gemfibrozil and bezafibrate
between the two cell lines (Fig. 5C and D), indicating that
these compounds are not transported by Pgp.
[
6] Miller DB, Spence JD. Clinical pharmacokinetics of fibric acid
derivatives (fibrates). Clin Pharmacokinet 1998;34:155–62.
[7] Backman JT, Kyrklund C, Kivisto
¨
KT, Wang JS, Neuvonen PJ. Plasma
concentrations of active simvastatin acid are increased by gemfibrozil.
Clin Pharmacol Ther 2000;68:122–9.
The present results implicate that the interactions occur-
ring between drugs and the fibrates gemfibrozil and bezafi-
brate, as well as the known differences in their interaction
pattern [29] are not due to interactions with Pgp. We have
not studied all known fibrates and not all of their metabolites
and hence cannot rule out that some of them might interact
with human Pgp. However, a major active metabolite
[
8] Schinkel AH, Smit JJ, van Tellingen O, Beijnen JH, Wagenaar E, van
Deemter L, Mol CA, van der Valk MA, Robanus-Maandag EC, te
Riele HP. Disruption of the mouse mdr1a P-glycoprotein gene leads to
a deficiency in the blood–brain barrier and to increased sensitivity to
drugs. Cell 1994;77:491–502.
[
9] van Asperen J, van Tellingen O, Beijnen JH. The pharmacological role
of P-glycoprotein in the intestinal epithelium. Pharmacol Res 1998;37:
4
29–35.
(
(
fenofibric acid) and the lipid lowering parent compounds
gemfibrozil, bezafibrate) clearly did not interact with Pgp.
Based on our results only for fenofibrate an in vivo
[10] Kusuhara H, Suzuki H, Sugiyama Y. The role of P-glycoprotein and
canalicular multispecific organic anion transporter in the hepatobiliary
excretion of drugs. J Pharm Sci 1998;87:1025–40.
[
11] Chiou WL, Chung SM, Wu TC. Potential role of P-glycoprotein in
affecting hepatic metabolism of drugs. Pharm Res 2000;17:903–5.
12] Boyd RA, Stern RH, Stewart BH, Wu X, Reyner EL, Zegarac EA,
Randinitis EJ, Whitfield L. Atorvastatin coadministration may in-
crease digoxin concentrations by inhibition of intestinal P-glycopro-
tein-mediated secretion. J Clin Pharmacol 2000;40:91–8.
interaction on the level of Pgp is conceivable. Because
fenofibrate proved to be an inhibitor of Pgp it appears
possible that the bioavailability of statins and other Pgp
substrates will increase during co-administration of feno-
fibrate. However, if occurring at all these interactions will
be limited to gastrointestinal Pgp because fenofibrate is
undergoing a nearly complete presystemic metabolism
[
[
13] Wu X, Whitfield LR, Stewart BH. Atorvastatin transport in the Caco-2
cell model: contributions of P-glycoprotein and the proton-monocar-
boxylic acid co-transporter. Pharm Res 2000;17:209–15.
[6,30]. Whether this interaction occurs and whether it is
[14] Bogman K, Peyer AK, T o¨ r o¨ k M, K u¨ sters E, Drewe J. HMG-CoA
clinically relevant has to be elucidated in vivo. However,
one of the very few interaction studies performed with
fenofibrate suggests that it does not affect the pharmaco-
kinetics of oral cyclosporine A [31] whose absorption is
modulated by intestinal Pgp [32].
reductase inhibitors and P-glycoprotein modulation. Br J Pharmacol
2
001;132:1183–92.
[
[
15] Wang E, Casciano CN, Clement RP, Johnson WW. HMG-CoA
reducatase inhibitors (statins) characterized as direct inhibitors of
P-glycoprotein. Pharm Res 2001;18:800–6.
16] Jacobsen W, Djuve S, Hirai S, Benet LZ, Christians U. Human liver
microsomal metabolism and P-glycoprotein mediated transport
through MDCK-MDR1 cell monolayers of HMG-CoA reductase
inhibitors in comparison. Drug Metab Rev 2001;33(Suppl 1):181.
17] Schinkel AH, Wagenaar E, Mol CA, van Deemter L. P-glycoprotein
in the blood–brain barrier of mice influences the brain penetration
and pharmacological activity of many drugs. J Clin Invest 1996;97:
2517–24.
Acknowledgments
[
The authors would like to thank Stephanie Fuchs for her
excellent technical assistance, Dr. Gerd Mikus for all his
support, Dr. M. Maurer (University of Kassel) for acquisi-
tion of the NMR data of simvastatin and fenofibric acid, Dr.
Alfred H. Schinkel (The Netherlands Cancer Institute) for
generously providing the cell line L-MDR1, and Eli Lilly
company for providing LY335979. This study was sup-
ported by the German Ministry for Education and Research
[
[
[
18] Weiss J, Dormann S-MG, Martin-Facklam M, Kerpen CJ, Ketabi-
Kiyanvash N, Haefeli WE. Inhibition of P-glycoprotein by newer
antidepressants. J Pharmacol Exp Ther 2003;305:197–204.
19] Weiss J, Kerpen CJ, Lindenmaier H, Dormann S-MG, Haefeli WE.
Interaction of antiepileptic drugs with human P-glycoprotein in vitro. J
Pharmacol Exp Ther 2003; 307: 262–7.
20] Ganssmann B, Klingmann A, Burhenne J, Tayrouz Y, Aderjan R,
Mikus G. Simultaneous determination of loperamide and its demethy-
lated metabolites in plasma and urine by high pressure liquid chro-
matography/API mass spectrometry. Chromatographia 2001;53:
(BMBF) grant 01EC9902.
6
56–60.
References
[
21] Yu DK. The contribution of P-glycoprotein to pharmacokinetic drug–
drug interactions. J Clin Pharmacol 1999;39:1203–11.
[
1] Shepherd J. Fibrates and statins in the treatment of hyperlipidaemia:
an appraisal of their efficacy and safety. Eur Heart J 1995;16:5–13.
2] Farmer JA. Learning from the cerivastatin experience. Lancet 2001;
[22] Shek A, Ferrill MJ. Statin-fibrate combination therapy. Ann Pharmac-
other 2001;35:908–17.
[
[23] Evans M, Rees A. The myotoxicity of statins. Curr Opin Lipidol
2002;13:415–20.
358:1383–5.

292
M. Ehrhardt et al. / Biochemical Pharmacology 67 (2004) 285–292
[
[
[
[
24] Wang JS, Neuvonen M, Wen X, Backman JT, Neuvonen PJ. Gemfi-
brozil inhibits CYP2C8-mediated cerivastatin metabolism in human
liver microsomes. Drug Metab Dispos 2002;30:1352.
markedly increased by gemfibrozil but not by bezafibrate. Clin
Pharmacol Ther 2001;69:340–5.
[30] Balfour JA, McTavish D, Heel RC. A review of its pharmacodynamic
and pharmacokinetic properties and therapeutic use in dyslipidaemia
fenofibrate. Drugs 1990;40:260–90.
25] Williams D, Feely J. Pharmacokinetic–pharmacodynamic drug inter-
actions with HMG-CoA reductase inhibitors. Clin Pharmacokinet
2002;41:343–70.
[31] deLorgeril M, Boissonnat P, Bizollon CA, Guidollet J, Faucon G,
Guichard JP, Levy-Prades-Sauron R, Renaud S, Dureau G.
Pharmacokinetics of cyclosporine in hyperlipidaemic long-term
survivors of heart transplantation. Lack of interaction with the
lipid-lowering agent fenofibrate. Eur J Clin Pharmacol 1992;43:
161–5.
26] Barecki-Roach M, Wang E-J, Johnson WW. Many P-glycoprotein
substrates do not inhibit the transport process across cell membranes.
Xenobiotica 2003;33:131–40.
27] Schwab D, Fischer H, Tabatabeai A, Poli S, Huwyler J. Comparison of
in vitro P-glycoprotein screening assays: recommendations for their
use in drug discovery. J Med Chem 2003;46:1716–25.
[32] Lown KS, Mayo RR, Leichtman AB, Hsiao HL, Turgeon DK,
Schmiedlin-Ren P, Brown MB, Guo W, Rossi SJ, Benet LZ, Watkins
PB. Role of intestinal P-glycoprotein (mdr1) in interpatient variation
in the oral bioavailability of cyclosporine. Clin Pharmacol Ther
1997;62:248–60.
[
[
28] Seelig A. A general pattern for substrate recognition by P-glycopro-
tein. Eur J Biochem 1998;251:252–61.
29] Kyrklund C, Backman JT, Kivist o¨ KT, Neuvonen M, Laitila J,
Neuvonen PJ. Plasma concentrations of active lovastatin acid are