Post on 01-Jun-2021
BRAINA JOURNAL OF NEUROLOGY
Adenosine triphosphate-binding cassettetransporters mediate chemokine (C-C motif) ligand2 secretion from reactive astrocytes: relevance tomultiple sclerosis pathogenesisGijs Kooij,1,* Mark R. Mizee,1,* Jack van Horssen,1 Arie Reijerkerk,1 Maarten E. Witte,2
Joost A. R. Drexhage,1 Susanne M. A. van der Pol,1 Bert van het Hof,1 George Scheffer,2
Rik Scheper,2 Christine D. Dijkstra,1 Paul van der Valk2 and Helga E. de Vries1
1 Blood-Brain Barrier Research Group, Department of Molecular Cell Biology and Immunology, VU University Medical Centre, PO Box 7057,
1007 MB Amsterdam, The Netherlands
2 Department of Pathology, VU University Medical Centre, PO Box 7057, 1007 MB Amsterdam, The Netherlands
*These authors contributed equally to this work
Correspondence to: Helga E. de Vries, Blood-Brain Barrier Research Group,
Department of Molecular Cell Biology and Immunology,
VU University Medical Centre,
PO Box 7057, 1007 MB Amsterdam,
The Netherlands
E-mail: he.devries@vumc.nl
Adenosine triphosphate-binding cassette efflux transporters are highly expressed at the blood–brain barrier and actively hinder
passage of harmful compounds, thereby maintaining brain homoeostasis. Since, adenosine triphosphate-binding cassette trans-
porters drive cellular exclusion of potential neurotoxic compounds or inflammatory molecules, alterations in their expression and
function at the blood–brain barrier may contribute to the pathogenesis of neuroinflammatory disorders, such as multiple scler-
osis. Therefore, we investigated the expression pattern of different adenosine triphosphate-binding cassette efflux transporters,
including P-glycoprotein, multidrug resistance-associated proteins-1 and -2 and breast cancer resistance protein in various
well-characterized human multiple sclerosis lesions. Cerebrovascular expression of P-glycoprotein was decreased in both
active and chronic inactive multiple sclerosis lesions. Interestingly, foamy macrophages in active multiple sclerosis lesions
showed enhanced expression of multidrug resistance-associated protein-1 and breast cancer resistance protein, which coincided
with their increased function of cultured foamy macrophages. Strikingly, reactive astrocytes display an increased expression of
P-glycoprotein and multidrug resistance-associated protein-1 in both active and inactive multiple sclerosis lesions, which
correlated with their enhanced in vitro activity on astrocytes derived from multiple sclerosis lesions. To investigate whether
adenosine triphosphate-binding cassette transporters on reactive astrocytes can contribute to the inflammatory process, primary
cultures of reactive human astrocytes were generated through activation of Toll-like receptor-3 to mimic the astrocytic pheno-
type as observed in multiple sclerosis lesions. Notably, blocking adenosine triphosphate-binding cassette transporter activity on
reactive astrocytes inhibited immune cell migration across a blood–brain barrier model in vitro, which was due to the reduction
of astrocytic release of the chemokine (C-C motif) ligand 2. Our data point towards a novel (patho)physiological role for
adenosine triphosphate-binding cassette transporters, suggesting that limiting their activity by dampening astrocyte activation
may open therapeutic avenues to diminish tissue damage during multiple sclerosis pathogenesis.
doi:10.1093/brain/awq330 Brain 2011: 134; 555–570 | 555
Received June 30, 2010. Revised September 14, 2010. Accepted October 2, 2010. Advance Access publication December 22, 2010
� The Author (2010). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: journals.permissions@oup.com
Keywords: adenosine triphosphate-binding cassette transporter; multiple sclerosis; astrocytes; chemokine (C-C motif) ligand 2;blood–brain barrier
Abbreviations: ABC = adenosine triphosphate-binding cassette; BCRP = breast cancer resistance protein; CCL2 = chemokine(C-C motif) ligand 2; GFAP = glial fibrillary acidic protein; MRP = multidrug resistance-associated protein; NAWM = normalappearing white matter
IntroductionMultiple sclerosis is a chronic demyelinating disease of the CNS
and is neuropathologically characterized by multiple focal demye-
linated lesions scattered throughout the CNS (Ewing and Bernard,
1998; Lassmann et al., 2001; Lucchinetti et al., 2001; Frohman
et al., 2006; Polman et al., 2006). Active multiple sclerosis lesions
contain abundant cellular infiltrates, which mainly consist of T cells
and monocyte-derived macrophages (Bruck et al., 1996). The
latter are thought to be responsible for causing damage to the
myelin sheaths that surround axons, resulting in neuronal dysfunc-
tion. In inflammatory demyelinating lesions, foamy macrophages
are present, which acquire their distinctive morphology by inges-
tion and accumulation of vast amounts of myelin-derived lipids
and cellular debris. Foamy macrophages originate from both
resident microglia and infiltrating monocytes (Li et al., 1996)
and are thought to display an anti-inflammatory phenotype
(Boven et al., 2006). In the course of lesion progression, enlarged
proliferative astrocytes become the most predominant cell type.
These reactive astrocytes secrete different neurotrophic factors
for neuronal survival but also contribute to pathology by produc-
tion of proinflammatory cytokines and chemokines (Tani et al.,
1996; Speth et al., 2005; Sofroniew, 2009).
The CNS microenvironment is well protected from the infiltra-
tion of inflammatory cells by the blood–brain barrier. One of the
key protective features of the blood–brain barrier is that it strictly
regulates the efflux of various toxic compounds through specia-
lized membrane pumps, thereby maintaining brain homoeostasis.
Efflux transporters are therefore regarded as key molecules in
protecting the brain from unwanted compounds, enabling
multi-drug resistance (Loscher and Potschka, 2005a). The
ATP-binding cassette (ABC) transporter family consist of a variety
of drug efflux pumps, including P-glycoprotein, breast cancer
resistant protein (BCRP) and the multidrug resistance-associated
proteins-1 and -2 (MRP-1, MRP-2). Interestingly, ABC efflux
pumps are expressed on different cell types, such as brain
endothelial cells and immune cells, and can drive cellular exclusion
of a variety of exogenous compounds and drugs through the cell
membrane against a concentration gradient at the cost of ATP
hydrolysis (Loscher and Potschka, 2005b). Importantly, several
studies have suggested that endogenous substrates for ABC trans-
porters may include inflammatory mediators, such as steroids,
prostaglandins, leucotrienes and cytokines (Drach et al., 1996;
Meijer et al., 1998; Ernest and Bello-Reuss, 1999; Frank et al.,
2001; Raggers et al., 2001; Choudhuri and Klaassen, 2006; van
de Ven et al., 2009). Hence, it is conceivable that besides actively
removing unwanted compounds, ABC transporters at the blood–
brain barrier also mediate the release of inflammatory agents
during (neuro)inflammatory processes, highlighting a potential
new role in multiple sclerosis pathology. To study this, we first
investigated the expression pattern of different ABC transporters
(P-glycoprotein, MRP-1, MRP-2 and BCRP) in well-characterized
multiple sclerosis lesions. We demonstrate here that various CNS
cell types, including endothelial cells, microglia and astrocytes,
express ABC transporter proteins. Importantly, striking differences
in their expression were observed in both active and inactive
multiple sclerosis lesions, which coincided with functional alter-
ations under neuroinflammatory conditions in vitro. These results
implicate a potential novel role for ABC transporters in multiple
sclerosis pathology. Notably, blocking astrocytic P-glycoprotein or
MRP-1 activity severely impairs monocyte migration across an
in vitro model of the blood–brain barrier and we here demonstrate
that P-glycoprotein and MRP-1 are involved in the secretion of
chemokine (C-C motif) ligand 2 (CCL2) by reactive astrocytes.
Together, our findings provide novel insights into the expression
and function of ABC transporters during multiple sclerosis path-
ology and illustrate a potential detrimental role of P-glycoprotein
and MRP-1 in reactive astrocytes.
Materials and methods
Brain tissueBrain tissue from 10 patients with clinically diagnosed and neuropatho-
logically confirmed multiple sclerosis was obtained at rapid autopsy
and immediately frozen in liquid nitrogen (in collaboration with The
Netherlands Brain Bank, coordinator Dr Huitinga). The Netherlands
Brain Bank received permission to perform autopsies, for the use of
tissue and for access to medical records for research purposes from the
Ethical Committee of the VU University Medical Centre, Amsterdam,
The Netherlands. Tissue samples from four control cases without
neurological disease were taken from the subcortical white matter
and corpus callosum. White matter multiple sclerosis tissue samples
were selected on the basis of post-mortem MRI. All patients and
controls, or their next of kin, had given informed consent for autopsy
and use of their brain tissue for research purposes. Relevant clinical
information was retrieved from the medical records and is summarized
in Table 1.
ImmunohistochemistryFor immunohistochemical staining, 5 mm cryosections were air-dried
and fixed in acetone for 10 min. Sections were incubated overnight
at 4�C with primary antibodies (Table 2). For the detection of proteo-
lipid protein, major histocompatibility complex class II, P-glycoprotein,
MRP-2 and BCRP, slides were incubated with EnVision Kit rabbit/
mouse-labelled horseradish peroxidase (DAKO, Glostrup, Denmark)
for 30 min at room temperature. For the detection of MRP-1, sections
were incubated with biotin-labelled rabbit anti-rat antibody (DAKO,
Glostrup, Denmark) for 30 min at room temperature and with avidin
556 | Brain 2011: 134; 555–570 G. Kooij et al.
biotin complex (DAKO, Glostrup, Denmark) according to the manu-
facturer’s description. Peroxidase activity was demonstrated with
0.5 mg/ml 3,30-diaminobenzidine tetrachloride (Sigma, St Louis, MO,
USA) in phosphate-buffered saline containing 0.02% hydrogen perox-
ide. Between incubation steps, sections were thoroughly washed with
phosphate-buffered saline. After a short rinse in tap water, sections
were incubated with haematoxylin for 1 min and extensively washed
with tap water for 10 min. Finally, sections were dehydrated with etha-
nol followed by xylol and mounted with Entellan� (Merck, Darmstadt,
Germany). All antibodies were diluted in phosphate-buffered saline
containing 0.1% bovine serum albumin (Boehringer-Mannheim,
Germany), which also served as a negative control.
For double immunofluorescence staining, sections were incubated
for 30 min with 20% normal goat serum. Sections were then incu-
bated overnight at 4�C with primary antibodies for all four transporters
(Table 2). To distinguish between different cell types, sections were
co-incubated with antibodies directed against glial fibrillary acidic
protein (GFAP; astrocytes), and CD11b (microglia/macrophages)
(Table 2), and subsequently labelled with Alexa-488 coupled goat
anti-mouse antibody (for MRP-1, MRP-2, BCRP and P-glycoprotein),
Alexa-633 coupled goat anti-rabbit antibody (for GFAP) and
Alexa-647 coupled goat anti-rat antibody (for CD11b) (all secondary
antibodies from Molecular Probes, Leiden, The Netherlands). After
washing, slides were covered with Vectashield� (Vector laboratories,
Burlington, CA, USA) supplemented with 0.4%
40,6-diamidino-2-phenylindole to stain nuclei. Fluorescence analysis
was performed with a Leica DM6000 microscope (Leica
Microsystems, Heidelberg, Germany).
Quantification ofimmunohistochemistryRelative changes in immunoreactivity were quantified by using ImageJ
software (v. 1.37c, NIH, USA) as described previously (Wang et al.,
2010). In short, three donors encompassing both active and inactive
multiple sclerosis lesions within one section, as well as a considerable
area of normal appearing white matter (NAWM), were selected to
correct for inter-donor variability. Various �40 RGB micrographs
were recorded for each donor, staining and white matter location
(active lesion, inactive lesion and NAWM). Next, images were sub-
jected to colour deconvolution to exclude haematoxylin staining
from analysis. The relative area of specific 3,30-diaminobenzidine de-
position to total area was determined by using threshold segmenta-
tion. The threshold was separately determined for each ABC
transporter and each donor in NAWM micrographs, to include specific
immunoreactivity and exclude background immunoreactivity. The
same threshold was applied to active and inactive multiple sclerosis
Table 1 Clinical information of multiple sclerosis and control patient material
Case Age (years) Type of multiple sclerosis Sex Post-mortem delay (h) Disease duration (years) Cause of death
Patient 1 52 SP F 8:25 n.k. Pneumonia
Patient 2 44 PP F 10:15 8 Decompensation
Patient 3 63 PP M 7:05 24 Cardiac arrest
Patient 4 56 SP M 8:00 27 Pneumonia
Patient 5 66 MS M 7:45 n.k. Sepsis
Patient 6 47 SP M 7:15 7 Urosepsis
Patient 7 79 SP F 14:00 39 CVA
Patient 8 48 PP F 4:50 25 Euthanasia
Patient 9 77 SP M 4:15 26 CVA
Patient 10 41 PP M 7:20 n.k. Pneumonia urosepsis
Control 1 78 – M 4:40 – Pneumonia
Control 2 82 – F 5:10 – Pneumonia by haemothorax
Control 3 57 – M 6:00 – Cardiac arrest
Control 4 77 – F 8:00 – Pneumonia
CVA = cerebral vascular accident; F = female; M = male; MS = MS subtype not determined; n.k. = not known; PP = primary progressive; SP = secondary progressive.
Table 2 Primary antibodies
Primary antibody Dilution Company
CD11b 1:100 Abcam
Proteolipid protein (clone plpc1) 1:500 Serotec Ltd, Oxford, UK
Major histocompatibility complex class II 1:100 DAKO
Glial fibrillary acidic protein 1:20 DAKO
MDR1 P-glycoprotein (clone 15D3) 1:10 Department of Pathology (VUmc, Amsterdam)
MRP1 (clone MRPr1) 1:50 Department of Pathology (VUmc, Amsterdam)
MRP-1 (clone MRPm5) (fluorescence) 1:25 Department of Pathology (VUmc, Amsterdam)
MRP2 (clone M2III-6) 1:50 Department of Pathology (VUmc, Amsterdam)
BCRP (clone BXP-21) 1:50 Department of Pathology (VUmc, Amsterdam)
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 557
lesion micrographs to assess the relative increase in immunoreactivity.
An example of the segmentation procedure is shown in Supplementary
Fig. 2.
Cell culturesPrimary astrocytes from control human brain tissue or multiple sclerosis
lesions and primary monocytes were isolated and cultured as described
previously (De Groot et al., 1997; Elkord et al., 2005). P-glycoprotein
(CEM/VBL), MRP-1 (2008/MRP-1) and BCRP (MCF7) overexpressing
cells and their control cell lines were obtained from and cultured as
described by Oerlemans et al. (2006). The human brain endothelial
cell line hCMEC/D3 was cultured as described previously (Weksler
et al., 2005).
Cell treatmentsPrimary human astrocytes and primary human monocytes were
cultured in 24- or 96-well plates. Astrocytes were subsequently incu-
bated with tumour necrosis factor-� (5 ng/ml; Peprotech, UK) or the
Toll-like receptor-3 ligand polyinosinic–cytidylic acid (50mg/ml,
Amersham Pharmacia Biotech, Piscataway, NJ, USA) for 6 or 24 h in
the presence or absence of the specific P-glycoprotein inhibitor rever-
sin 121 (10mM; Alexis) or the specific MRP-1 inhibitor MK-571
(25 mM; Merck Frosst Canada). Multiple sclerosis lesion reactive astro-
cytes were only cultured in the presence or absence of P-glycoprotein
or MRP-1 inhibitors. Subsequently, supernatants were harvested for
enzyme-linked immunosorbent assay. For migration experiments, all
of the aforementioned treatments were washed away and conditioned
media was collected after 24 h. Primary monocytes were either incu-
bated with human myelin derived from control white matter (Van der
Goes et al., 2005) or latex beads (Polysciences) for different time
points (24 or 48 h).
In vitro assays for ATP-bindingcassette transporter functionP-glycoprotein, MRP1 and BCRP function was determined as described
previously (van der Pol et al., 2003) with minor modifications. Briefly,
after treatment, astrocytes or macrophages were washed three times
with phosphate-buffered saline and subsequently incubated for 45 min
at 37�C with specific substrates in the presence or absence of specific
inhibitors [P-glycoprotein substrate Rhodamine 123 (2mM; Sigma),
inhibitor reversin 121 (10 mM; Alexis); MRP-1 substrate Calc-AM
(500 nM; Molecular Probes), inhibitor MK-571 (25mM; Merck Frosst
Canada); BCRP substrate Bodipy (100 nM; kind gift from Dr G.
Scheffer, VUMC, Amsterdam, The Netherlands), inhibitor KO143
(200 nM; kind gift from Dr G. Scheffer, VUMC, Amsterdam, The
Netherlands)]. After a 45 min incubation, cells were washed three
times with phosphate-buffered saline and fluorescence intensity was
measured using a FLUOstar Galaxy microplate reader (BMG
Labtechnologies, Offenburg, Germany), excitation 485 nm, emission
520 nm or by a FACScan flow cytometer (Becton and Dickinson, San
Jose, CA, USA). Fluoresence-activated cell sorting analysis was
performed on 10 000 viable cells, selected by 7AAD exclusion. ABC
transporter activities are expressed as ratios of drug fluorescence with
inhibitor and drug fluorescence without inhibitor after subtraction of
the fluorescence of the control. Overexpressing cell lines for
P-glycoprotein (CEM/VBL), MRP1 (2008/MRP1) and BCRP (MCF7)
were used for optimizing the functional ABC transporter assays and
indicated high selectivity of all different inhibitors used in our assays.
See Supplementary Fig. 1 for additional information.
Enzyme-linked immunosorbent assayCCL2 or interleukin 1b protein was measured in culture supernatants
of control or polyinosinic–cytidylic acid stimulated astrocytes (24 h) or
multiple sclerosis lesion derived astrocytes using an enzyme-linked
immunosorbent assay (R&D Systems) with a lowest detection level
of 30 pg/ml (CCL2) or 1 pg/ml (interleukin 1b) as described previously
(Tekstra et al., 1996).
RNA isolation and real-time quantitativepolymerase chain reactionMessenger RNA was isolated from control or polyinosinic–cytidylic acid
stimulated astrocytes using a messenger RNA capture kit (Roche) ac-
cording to the manufacturer’s instructions. Complementary DNA was
synthesized with the Reverse Transcription System kit (Promega, USA)
following the manufacturer’s guidelines and real-time quantitative
polymerase chain reaction was performed as described previously
(Garcia-Vallejo et al., 2004). All primer sequences are listed in
Supplementary Table 1 and expression levels of transcripts obtained
with real-time polymerase chain reaction were normalized to
glyceraldehyde 3-phosphate dehydrogenase expression levels.
Monocyte migrationWe used two established protocols for the measurement of human
monocyte migration using the 48-well chemotaxis chamber (Frevert
et al., 1998) and/or a Transwell system with cultured brain endothelial
cells (Viegas et al., 2006) with minor modifications. Briefly, chemotaxis
chamber migration was measured using 48-well chambers with
nitrocellulose filters. Conditioned media from control, in vitro gener-
ated or multiple sclerosis lesion derived reactive astrocytes was added
to the bottom wells (25 ml) of the 48-well plate. Nitrocellulose filters
(Neuro Probe) with a 5 mm pore size were placed between the bottom
and top plates of the chamber assembly and the monocytes (50 ml)
were added to the top wells at a cell density of 4 � 105 monocytes/ml.
The chamber was incubated for 1.5 h (37�C and 5% CO2) and the
non-migrating cells on top of the filter were removed by gentle scraping.
The filter was air-dried, fixed and stained with a modified haematoxylin/
eosin stain. Filters were mounted on glass slides and monocyte migration
was measured visually by counting the number of cells at the leading
front of migration in 10 high-powered fields (�450).
To investigate the influence of conditioned media on the capacity of
monocytes to cross a monolayer of brain endothelial cells, Transwell
migration experiments were performed using human brain endothelial
hCMEC/D3 cells, which were cultured onto collagen (upper side,
Sigma, St Louis, USA) coated Costar Transwell filters (pore size
5 mm; Corning Incorporated, Corning, NY, USA) for four days. At
the start of the experiment, 600 ml of conditioned media from control
or reactive astrocytes was added to the bottom wells. Monocytes
(100 ml) were then added to the top wells at a cell density of
1 � 106 monocytes/ml. Monocytes were allowed to migrate for 8 h
(37�C and 5% CO2). After 8 h, 400ml was collected from the lower
chamber and 20 000 beads (Beckman Coulter, USA) were added to
each sample. Samples were then analysed using a FACScan flow cyt-
ometer (Becton and Dickinson, San Jose, CA, USA) and based on 5000
gated beads, the number of migrated monocytes was determined. In
both assays, monocyte migration was presented as the absolute
558 | Brain 2011: 134; 555–570 G. Kooij et al.
number of migrated monocytes compared with the total number of
monocytes added in the upper chamber.
Statistical analysisData were analysed statistically by means of a single-column t-test.
Statistical significance was defined as *P5 0.05, **P5 0.01 and
***P5 0.001.
Results
Multiple sclerosis lesion classificationClassification of multiple sclerosis lesions was based on standard
immunohistochemical staining for inflammatory cells (anti-major
histocompatibility complex class II) and myelin (proteolipid protein)
as described previously (van der Valk and De Groot, 2000; van
Horssen J. et al., 2006a, b). Based on these findings, 12 lesions
sampled in this study were classified as active with myelin loss
(Figs 1A and 3A) and abundant phagocytic perivascular and
parenchymal macrophages containing myelin degradation
products (Figs 1B and 3B) and seven lesions as chronic inactive
with demyelinated areas (Figs 2A and 4A) containing few major
histocompatibility complex class II-positive cells (Figs 2B and 4B).
Enhanced multidrug resistance-associated protein-1 and -2 expressionin multiple sclerosis lesions andincreased multidrug resistance-associated protein-1 function in foamymacrophagesIn white matter from non-neurological control brain tissue (data
not shown) and normal appearing white matter (NAWM), MRP-1
(Fig. 1C, arrows) and MRP-2 (Fig. 1D, arrows) immunoreactivity is
mainly restricted to glial cells, whereas, endothelial cells that line
the cerebral vasculature only weakly express MRP-1 and MRP-2.
In active demyelinating multiple sclerosis lesions (Fig. 1A and B)
enhanced MRP-1 (Fig. 1E) and MRP-2 (Fig. 1F) staining is
observed in foamy macrophages (arrows) and hypertrophic astro-
cytes (arrowheads). Using double immunofluorescence staining,
we confirmed the cellular localization of these efflux pumps in
active multiple sclerosis lesions and showed that MRP-1
and MRP-2 are expressed by GFAP-positive astrocytes
(Fig. 1G and I) and CD11b-positive macrophages (Fig. 1H and
J). To study whether foamy macrophages are capable of actively
removing substrates for MRP-1, we performed an in vitro func-
tional MRP-1 assay. First, in vitro foamy macrophages were gen-
erated (Van der Goes et al., 2005; Boven et al., 2006) by adding
myelin to human monocytes at different time points (24 and 48 h),
resulting in their characteristic foamy appearance (data not
shown). Subsequently, we determined MRP-1 activity in untreated
or myelin-laden macrophages. Notably, MRP-1 function was
enhanced upon addition of myelin at different time points
(Fig. 1K). In contrast, phagocytosis of latex beads by cultured
macrophages did not result in increased functionality of MRP-1
(Fig. 1K), indicating that myelin specifically induces MRP-1 efflux
transporter activity on macrophages, which correlates with the
increased expression levels of MRP-1 on foamy macrophages in
active multiple sclerosis lesions. In chronic inactive multiple scler-
osis lesions (Fig. 2A and B), hypertrophic astrocytes express
MRP-1 (Fig. 2C, arrowheads) and MRP-2 (Fig. 2D, arrowheads),
whereas microglia also express MRP-2 (arrows) to the same level
as seen in control white matter (Fig. 1D). Brain endothelial cells
express relatively low amounts of MRP-1 and MRP-2 in control or
multiple sclerosis brain tissue (Figs 1 and 2), which coincided with
low MRP-1 activity in cultured human brain endothelial cells using
an in vitro efflux assay (Supplementary Fig. 1). Moreover, no
differences in vascular MRP-1 and MRP-2 expression are observed
between multiple sclerosis lesions and NAWM. To quantitatively
assess MRP-1 and MRP-2 expression in multiple sclerosis lesions,
we quantified MRP-1 and MRP-2 immunoreactivity in NAWM
and different multiple sclerosis lesions (see Supplementary Fig. 2
for detailed description). In line with the immunohistochemical
staining, we observed increased transporter immunoreactivity in
active lesions (MRP-1 and MRP-2) and chronic inactive lesions
(MRP-2) (Fig. 2E).
P-glycoprotein and breast cancerresistance protein expression in controlwhite matter and multiple sclerosislesions and increased breast cancerresistance protein in foamymacrophagesP-glycoprotein immunoreactivity is predominantly localized to the
cerebral microvasculature in NAWM (Fig. 3C, arrows) and control
brain tissue (data not shown) and only weakly expressed by astro-
cytes (arrowheads). Notably, in active demyelinated multiple scler-
osis lesions (Fig. 3E, arrows) and chronic inactive lesions (Fig. 4C,
arrow), decreased vascular P-glycoprotein immunoreactivity was
observed. Surprisingly, hypertrophic GFAP-positive astrocytes were
markedly decorated with anti-P-glycoprotein in active and chronic
inactive multiple sclerosis lesions (Figs 3E, 3G and 4C, arrowheads).
Quantification of relative P-glycoprotein immunoreactivity showed a
significant increase in active and chronic inactive multiple sclerosis
lesions as compared with NAWM (Fig. 4E). BCRP expression is re-
stricted to the brain microvasculature (arrows) and microglial cells
(arrowheads) in NAWM (Fig. 3D) and control brain tissue (data
not shown). In active demyelinating multiple sclerosis lesions a
marked increase in BCRP staining is observed on CD11b-positive
foamy macrophages (Fig. 3F and H), which correlated with enhanced
BCRP functionality upon myelin phagocytosis of human macro-
phages at different time points (Fig. 3I). In line with MRP-1 (Fig.
1K), enhanced BCRP activity in foamy macrophages appeared to
be a myelin-specific effect, as phagocytosis of latex beads did not
alter its efflux capacity (Fig. 3I). Brain endothelial cells express high
amounts of P-glycoprotein and BCRP in control tissue, which coin-
cides with high P-glycoprotein and BCRP activity in human brain
endothelial cells in vitro (Supplementary Fig. 1). However, in contrast
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 559
Figure 1 MRP-1 and MRP-2 expression in normal appearing white matter and multiple sclerosis lesions and increased MRP-1 function on
foamy macrophages. (A) Loss of proteolipid protein (PLP) immunoreactivity in a subcortical lesion, with (B) enhanced expression of major
histocompatibility complex class II (MHCII) (magnification �10). Boxed sites are representative areas of the �40 magnification of
adjacent sections stained for MRP-1 and MRP-2. In normal appearing white matter (NAWM) MRP-1 (C) and MRP-2 (D) immunor-
eactivity is observed on microglial cells (arrows) and faintly on endothelial cells. Within an active demyelinating multiple sclerosis lesion,
MRP-1 (E) and MRP-2 (F) immunoreactivity is highly increased on hypertrophic astrocytes and astrocyte processes (arrowheads) and
(continued)
560 | Brain 2011: 134; 555–570 G. Kooij et al.
to P-glycoprotein, no differences in endothelial BCRP expression
were observed between multiple sclerosis lesions and normal
appearing white matter. The increase in relative BCRP immunoreac-
tivity observed in active multiple sclerosis lesions as compared with
NAWM (Fig. 4E) can be explained by the presence of BCRP-positive
macrophages, which are absent in chronic inactive multiple sclerosis
lesions.
Increased expression and function ofP-glycoprotein and multidrugresistance-associated protein-1 inreactive astrocytes in vitroIncreased protein expression of P-glycoprotein and MRP-1 on
reactive astrocytes in multiple sclerosis lesions suggests an altered
function of these efflux pumps under neuroinflammatory condi-
tions. To study this, we first isolated astrocytes from multiple scler-
osis lesions and control white matter and determined the
messenger RNA expression level of the (reactive) astrocyte
marker GFAP (Pekny and Nilsson, 2005) by means of quantitative
polymerase chain reaction. Interestingly, multiple sclerosis
lesion-derived astrocytes display increased transcription levels of
GFAP (Fig. 5A), illustrating a reactive phenotype. Next, in vitro
functional assays for P-glycoprotein and MRP1 were performed on
control and multiple sclerosis lesion-derived astrocytes. Notably,
reactive astrocytes isolated from multiple sclerosis lesions display
increased functionality of both efflux pumps (Fig. 5B and C),
which correlates with the enhanced expression levels of astrocytic
P-glycoprotein and MRP-1 in human multiple sclerosis lesions (Figs
1–4). To investigate whether inflammatory mediators could affect
astrocytic P-glycoprotein and MRP-1 activity, we treated primary
human astrocytes with inflammatory mediators like tumour necro-
sis factor-� and/or polyinosinic–cytidylic acid, a double-stranded
RNA mimetic ligand for Toll-like receptor-3, to mimic a
pro-inflammatory environment as observed during neuroinflam-
mation (Obata et al., 2008). Interestingly, both inflammatory me-
diators increased MRP-1 and P-glycoprotein efflux capacity by
astrocytes (Fig. 5B and C), with Toll-like receptor-3 activation
being the most potent inducer (Fig. 5B and C). To determine
whether Toll-like receptor-3 activation on astrocytes leads to a
reactive astrocyte phenotype, we verified messenger RNA expres-
sion levels of various reactive astrocytic markers, such as GFAP,
S100b, vimentin and interleukin 6 (Ridet et al., 1997) and the ABC
transporters P-glycoprotein (MDR1) and MRP-1 by means of
real-time quantitative polymerase chain reaction. Notably, polyi-
nosinic–cytidylic acid-treated astrocytes display increased transcrip-
tion levels of GFAP, interleukin 6, S100b, MDR1 and MRP-1
compared with control astrocytes (Fig. 5D), whereas vimentin
levels remained unaltered. These results illustrate an in vitro
model for the generation of reactive astrocytes by Toll-like
receptor-3 activation. Together, these results show that
P-glycoprotein and MRP-1 expression and function are highly
increased in inflammatory reactive astrocytes.
P-glycoprotein and multidrugresistance-associated protein-1 onreactive astrocytes mediate monocytemigration across a blood–brain barriermodelReactive astrocytes contribute to the inflammatory process by the
production and secretion of proinflammatory cytokines and
chemokines (Tani et al., 1996; Speth et al., 2005). In particular
chemokines, such as chemokine (C-C motif) ligand 2 (CCL2), are
known to attract leucocytes and monocyte-derived macrophages
into multiple sclerosis lesions, which in turn results in severe tissue
damage (Tani and Ransohoff, 1994). As ABC transporters have
been suggested to be involved in the secretion of inflammatory
mediators, we investigated whether P-glycoprotein and MRP-1
are capable of regulating the efflux of the astrocyte-derived
chemokine CCL2. Our results show that polyinosinic–cytidylic
acid-treated astrocytes secrete (Fig. 6A) and produce (Fig. 6B)
high levels of CCL2 compared with control astrocytes. Notably,
blocking P-glycoprotein or MRP-1 activity with specific inhibitors
(see Supplementary Fig. 1) such as reversin 121 (Koubeissi et al.,
2006) and MK-571 (van de Ven et al., 2006), respectively,
significantly reduced CCL2 secretion from reactive astrocytes
(Fig. 6C), whereas CCL2 messenger RNA expression levels
remained unaffected (Fig. 6D). Moreover, P-glycoprotein and
MRP-1 on reactive astrocytes did not affect the secretion of the
proinflammatory cytokines interleukin 1b (Fig. 6E) or interferon-�
(data not shown), indicating that these transporters are selectively
involved in CCL2 secretion, but not the production of CCL2 from
reactive astrocytes.
As CCL2 is a potent chemokine involved in leukocyte migration,
we hypothesize that P-glycoprotein and MRP-1 on reactive
astrocytes contribute to the inflammatory process by mediating
CCL2 efflux and induce immune cell migration. To assess this, we
determined the potential role for these transporters in mediating
monocyte migration in a chemotaxis assay. Conditioned media
from polyinosinic–cytidylic acid-treated astrocytes or multiple scler-
osis lesion-derived astrocytes significantly increased monocyte
migration across filters (Fig. 7A) compared with conditioned media
from control astrocytes. Interestingly, blocking MRP-1 and, in
particular, P-glycoprotein activity strikingly reduced the migration
capacity of monocytes (Fig. 7A), indicating that both
P-glycoprotein and MRP-1 on reactive astrocytes are involved in
leukocyte migration processes. To mimic the blood–brain barrier in
Figure 1 Continuedfoamy macrophages (arrows). These figures show representative images observed in all patient material. Colocalization of MRP-1 and
MRP-2 immunoreactivity (in green) with GFAP (G and I) and CD11b (H and J) immunoreactivity (in red) confirms the morphological
observations. Myelin phagocytosis for 24 or 48 h by primary human monocytes results in increased MRP-1 function (K), whereas the
phagocytosis of latex beads did not affect MRP-1 functionality. A value of 100% corresponds to a ratio of 1.07 � 0.04. Experiments were
performed in triplicate using three different human donors and were presented as the mean � SEM. *P5 0.05, **P50.01 by Student’s
t-test.
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 561
more detail, we next used a Transwell system using human brain
endothelial cells cultured on Transwell filters. Notably, conditioned
media from reactive astrocytes significantly enhanced monocyte
migration across brain endothelial cells (Fig. 7B), which could be
blocked by using specific P-glycoprotein or MRP-1 inhibitors
(Fig. 7B). Together these results indicate that reactive astrocytes
are actively involved in the inflammation process during multiple
sclerosis lesion formation and point to a novel role of ABC transport-
ers on reactive astrocytes in inducing monocyte migration across
brain endothelial cells under pathological conditions.
DiscussionIn this study, we provide for the first time a comprehensive over-
view of ABC transporter expression in multiple sclerosis brain
tissue and we illustrate the potential contribution of ABC trans-
porters to neuroinflammation. The predominant cell types involved
in multiple sclerosis pathology, including brain endothelial cells,
reactive astrocytes and infiltrated foamy macrophages, display
marked alterations in their ABC transporter expression, which
coincides with functional changes in vitro under inflammatory
Figure 2 MRP-1 and MRP-2 expression in chronic inactive multiple sclerosis lesions. (A) Loss of proteolipid protein (PLP) immunor-
eactivity in a subcortical lesion, with (B) a low number of major histocompatibility complex class II (MHCII) positive cells (magnification
�10). Boxed site is a representative area of the �40 magnification of adjacent sections stained for MRP-1 and MRP-2. In chronic inactive
multiple sclerosis lesions, MRP-1 (C) and MRP-2 (D) immunoreactivity is observed on hypertrophic astrocytes (arrowheads) and resting
microglia (arrows). These figures show representative images observed in all patient material. (E) Quantification of the relative difference
in immunoreactivity for both MRP-1 and MRP-2 in NAWM, active lesions and inactive lesions. Presented as mean fold change from
NAWM, �SEM. *P50.05, ***P50.001 by Student’s t-test.
562 | Brain 2011: 134; 555–570 G. Kooij et al.
Figure 3 P-glycoprotein and BCRP expression in normal appearing white matter and multiple sclerosis lesions and increased BCRP function
on foamy macrophages. (A) Loss of proteolipid protein (PLP) immunoreactivity in a subcortical lesion, with (B) enhanced expression of
major histocompatibility complex class II positivity (magnification �10). Boxed sites are representative areas of the �40 magnification of
adjacent sections stained for P-glycoprotein (P-gp) and BCRP. In NAWM, P-glycoprotein immunoreactivity (C) is observed on endothelium
(continued)
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 563
conditions. Moreover, we show here that Toll-like receptor-3 ac-
tivation in astrocytes induces enhanced expression and function of
P-glycoprotein and MRP-1. We provide evidence that the astro-
cytic ABC transporters may play a role in the neuroinflammatory
process by mediating the efflux of the inflammatory molecule
CCL2, thereby promoting immune cell migration across brain
endothelial cells.
In different stages of multiple sclerosis lesions, we observed an
altered expression pattern of various ABC transporter proteins,
such as P-glycoprotein, MRP-1, MRP-2 and BCRP. In particular,
reactive astrocytes, which are abundantly present in multiple scler-
osis lesions, display enhanced expression of P-glycoprotein, MRP-1
and MRP-2. So far, enhanced astrocytic expression of
P-glycoprotein and MRP-1 has been reported in brain tissue of
patients with epilepsy (Sisodiya et al., 2002; Marroni et al.,
2003), which has been suggested to be a result of seizures or
drug treatment. Moreover, we observed enhanced expression of
BCRP, MRP-1 and MRP-2 on infiltrated foamy macrophages in
active multiple sclerosis lesions. Notably, enhanced function of
BCRP and MRP-1 in myelin-laden macrophages was found in
our in vitro model. Although the exact role of BCRP and MRP-1
function in this process is still unknown, it has been described that
macrophages rely on cholesterol efflux mechanisms to maintain
cellular cholesterol homoeostasis by means of ABC transporters
ABCA1 and ABCG1 (Out et al., 2008). Moreover, as both
MRP-1 and BCRP participate in cellular detoxification (Borst and
Elferink, 2002; Krishnamurthy and Schuetz, 2006), it is likely that
both ABC transporters may be involved in the removal of phago-
cytosed myelin components such as cholesterol from foamy
macrophages and thereby control cellular homoeostasis.
Recently, BCRP and MRP-1 expression has been detected on
rheumatoid arthritis synovial tissue macrophages, which was sug-
gested to be a result of drug treatment (van der Heijden et al.,
2009). We demonstrate here that, upon myelin phagocytosis,
macrophages establish increased MRP-1 and BCRP activity,
which is in line with the increased expression pattern of these
ABC transporters in inflammatory multiple sclerosis lesions.
In control white matter and NAWM, we observed a cerebro-
vascular expression of P-glycoprotein and BCRP, while MRP-1 and
MRP-2 are weakly expressed by brain endothelial cells. In vitro
functional assays confirmed that both P-glycoprotein and BCRP
are highly active on human brain endothelial cells, whereas
MRP-1 function is nearly absent. Notably, we detected a
decreased endothelial P-glycoprotein expression in multiple scler-
osis lesions, whereas no differences in endothelial BCRP, MRP-1
and MRP-2 were observed. We have previously shown that vas-
cular P-glycoprotein expression and function is strongly decreased
during multiple sclerosis pathology and identified a crucial role for
activated CD4 + T cells in endothelial P-glycoprotein regulation via
intracellular adhesion molecule 1 and nuclear factor � B signalling
(Kooij et al., 2010). Since no changes in vascular expression for
BCRP, MRP-1 and MRP-2 were observed, our results indicate
differential ABC transporter regulatory mechanisms during
pathological conditions and further research is warranted to
define these underlying differences. Cerebrovascular expression
has previously been shown for BCRP (Krishnamurthy and
Schuetz, 2006), MRP-2 (Potschka et al., 2003) and to a lesser
extent for MRP-1 (Nies et al., 2004). In contrast, other groups
did not detect the MRP-1 protein on the microvasculature when
analysed by immunohistochemistry (Rao et al., 1999; Aronica
et al., 2003), which might be explained by the use of different
antibodies to MRP-1. Together, our results demonstrate the
expression of P-glycoprotein, BCRP, MRP-1 and MRP-2 in the
cerebral vasculature in NAWM, of which P-glycoprotein is select-
ively affected during multiple sclerosis pathology.
In both active and inactive multiple sclerosis lesions, we
observed an increased astrocytic P-glycoprotein and MRP-1
expression, which correlated with enhanced P-glycoprotein and
MRP-1 activity of lesion-derived astrocytes compared with
astrocytes isolated from non-affected white matter. Activation of
astrocytes has been implicated in the pathogenesis of a variety of
neurodegenerative diseases, including Alzheimer’s disease, inflam-
matory demyelinating diseases and human immunodeficiency
virus-associated dementia (Eng and Ghirnikar, 1994). Conversely,
an equal body of evidence suggest that astrocyte activation can
also exert beneficial effects (see Sofroniew, 2009 for a review), as
reactive astrocytes can secrete neurotrophic factors. However,
severe activation might augment an inflammatory response, lead-
ing to neuronal death and brain injury (Tani et al., 1996). In spite
of the ubiquitous presence of reactive astrocytes at various sites of
CNS pathology, their potential contribution to pathology and
underlying mechanisms are still poorly understood. To study the
potential contribution of reactive astrocytes in multiple sclerosis
pathology in more detail, we first generated reactive astrocytes
in vitro by Toll-like receptor-3 activation of primary human astro-
cyte cultures, which resulted in high messenger RNA expression
levels of the reactive astrocyte marker GFAP to a similar extent as
observed in astrocytes isolated from multiple sclerosis lesions.
Furthermore, Toll-like receptor-3 activation enhanced the expres-
sion of S100b and interleukin 6, which are well-known markers for
Figure 3 Continued(arrows) and faintly on astrocytes (arrowhead). BCRP immunoreactivity in NAWM (D) is prominent on endothelial cells (arrows) as well as
on resting microglial cells (arrowheads). Within an active demyelinating multiple sclerosis lesion, P-glycoprotein immunoreactivity (E) is
highly increased on hypertrophic astrocytes (arrowheads) and decreased on the endothelium (arrow). BCRP immunoreactivity within an
active lesion is unaltered on the endothelium (arrow) and present on foamy macrophages (arrowheads). These figures show representative
images observed in all patient material. Co-localization of P-glycoprotein (P-gp) immunoreactivity (in green) with GFAP (in red) (G) and
BCRP immunoreactivity (in green) with CD11b (in red) (H) confirms the morphological observations. Myelin phagocytosis for 24 or 48 h by
primary human monocytes results in increased BCRP function (I), whereas the phagocytosis of latex beads did not affect BCRP functionality.
A value of 100% corresponds to a ratio of 1.19 � 0.13. Experiments were performed in triplicate using three different human donors and
were presented as the mean � SEM. **P5 0.01 by Student’s t-test.
564 | Brain 2011: 134; 555–570 G. Kooij et al.
astrocyte activation (Ridet et al., 1997), indicating that Toll-like
receptor-3 activation is a suitable method for the generation of
reactive astrocytes in vitro. Notably, tumour necrosis factor-�
treatment and Toll-like receptor-3 activation of astrocytes led to
an increased P-glycoprotein and MRP-1 activity, indicating that
these inflammatory agents are involved in the regulation of ABC
transporter expression and function in astrocytes. It has been
described that at the transcriptional level ABC transporters are
under the control of the orphan nuclear receptors such as steroid
and xenobiotic receptor (or pregnane X receptor in rodents)
(Loscher and Potschka, 2005b). Furthermore, their expression
and function are regulated by environmental stimuli that evoke
stress responses (Sukhai and Piquette-Miller, 2000), such as the
excitatory neurotransmitter glutamate (Zhu and Liu, 2004) or the
inflammatory cytokines (Bauer et al., 2007). So far, only regulation
of brain endothelial P-glycoprotein expression and function has
been reported (Goralski et al., 2003; Bauer et al., 2007). We
here extend these results by demonstrating that inflammatory
Figure 4 P-glycoprotein and BCRP expression in chronic inactive multiple sclerosis lesions. (A) Loss of proteolipid protein (PLP) immu-
noreactivity in a subcortical lesion, with (B) a low number of major histocompatibility complex class II (MHCII) positive cells (magnification
�10). Boxed site is a representative area of the �40 magnification of adjacent sections stained for P-glycoprotein and BCRP. In chronic
inactive lesions P-glycoprotein (P-gp) immunoreactivity (C) is prominent on hypertrophic astrocytes (arrowheads) and faintly present on
the endothelium (arrow). BCRP immunoreactivity (D) in chronic inactive lesions is present on the endothelium (arrow) and resting
microglial cells (arrowheads). These figures show representative images observed in all patient material. (E) Quantification of the relative
difference in immunoreactivity for both P-glycoprotein (P-gp) and BCRP in NAWM, active lesions and inactive lesions. Presented as mean
fold change from NAWM � SEM. *P50.05 by Student’s t-test.
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 565
mediators can affect both MRP-1 and P-glycoprotein expression
and function on human astrocytes.
In this study, we revealed a novel pathophysiological role for
P-glycoprotein and MRP-1 on reactive astrocytes in mediating
immune cell migration across brain endothelial cells, which may
aggravate the inflammatory attack during multiple sclerosis lesion
progression. Active multiple sclerosis lesions are characterized by
the presence of infiltrated leukocytes, and chemokines such as
CCL2 play a key role in the attraction of immune cells into
these multiple sclerosis lesions (Tani and Ransohoff, 1994).
Notably, mice lacking the receptor for CCL2 (CCR2) did not
develop experimental autoimmune encephalomyelitis, an animal
model for multiple sclerosis (Fife et al., 2000; Izikson et al.,
2000), indicating its in vivo relevance. Moreover, CCL2 expression
appeared to be restricted to reactive astrocytes in multiple sclerosis
lesions (van der Voorn et al., 1999). These in vivo observations
were confirmed in vitro in our study, as reactive astrocytes
produce and secrete high levels of CCL2 upon Toll-like receptor-3
activation. Moreover, we identified a novel role for P-glycoprotein
and MRP-1 in the modulation of CCL2 secretion from reactive
astrocytes. These results support the hypothesis that endogenous
substrates for ABC transporters may include inflammatory
mediators, such as prostaglandins, leukotrienes and cytokines as
observed in studies using immune cells (Drach et al., 1996; Meijer
et al., 1998; Ernest and Bello-Reuss, 1999; Frank et al., 2001;
Choudhuri and Klaassen, 2006) and now possibly chemokines
such as CCL2. It still remains unknown whether ABC transporters
themselves are capable of transporting chemokines or if they are
Figure 5 Increased P-glycoprotein and MRP-1 expression and function in reactive astrocytes. (A) Increased GFAP messenger RNA
expression on primary human astrocytes isolated from active multiple sclerosis (MS) lesions compared with astrocytes isolated from control
white matter. GFAP expression was determined by real-time quantitative polymerase chain reaction and presented as relative expression
compared with glyceraldehyde 3-phosphate dehydrogenase. Enhanced MRP-1 (B) and P-glycoprotein (C) functionality in primary human
astrocytes isolated from active multiple sclerosis lesions (grey bars) and in control astrocytes stimulated with either tumour necrosis
factor-� (TNF-�) (5 ng/ml) or polyinosinic–cytidylic acid (poly I:C) (50 mg/ml) for 6 h, compared with untreated astrocytes. Control
astrocytes have a ratio of 1.03 � 0.18 (P-glycoprotein) or 1.15 � 0.07 (MRP-1). GFAP, interleukin 6 (IL-6), S100b, vimentin, MDR1
(P-glycoprotein) and MRP-1 transcripts from control or polyinosinic–cytidylic acid-treated (24 h) astrocytes (D) were detected by real-time
quantitative polymerase chain reaction and presented as relative expression compared with glyceraldehyde 3-phosphate dehydrogenase.
Experiments were performed in triplicate using three different human donors and were presented as the mean � SEM. *P50.05,
**P50.01, ***P50.001 by Student’s t-test.
566 | Brain 2011: 134; 555–570 G. Kooij et al.
involved in the secretion of other relevant more lipophilic inflam-
matory substrates such as platelet activating factor (Raggers et al.,
2001) that in turn may affect CCL2 secretion (Huang et al., 1999)
as a secondary effect. Nevertheless, increased ABC transporter
expression and function on reactive astrocytes may result in local
enhanced efflux of inflammatory mediators in multiple sclerosis
lesions, amplifying the inflammatory response. To reveal such a
novel pathophysiological role for the ABC transporters
P-glycoprotein and MRP-1 on reactive astrocytes, conditioned
media from either in vitro generated or multiple sclerosis
lesion-derived reactive astrocytes markedly enhanced monocyte
migration across an in vitro model of the blood–brain barrier.
Moreover, blocking P-glycoprotein or MRP-1 on these reactive
astrocytes severely inhibited this monocyte migration capacity,
illustrating a novel detrimental role for these ABC transporters
on reactive astrocytes by facilitating immune cell migration
Figure 6 P-glycoprotein and MRP-1 mediate CCL2 secretion from reactive astrocytes. Primary human astrocytes were treated with or
without polyinosinic–cytidylic acid (poly I:C) (50 mg/ml) for 24 h and CCL2 secretion was determined in cell supernatants by enzyme-linked
immunosorbent assay (A). CCL2 transcripts were determined by real-time quantitative polymerase chain reaction and presented as relative
expression (FI = fold induction) compared with glyceraldehyde 3-phosphate dehydrogenase (B). Astrocytes were treated with polyino-
sinic–cytidylic acid (50 mg/ml) for 24 h in the presence or absence of the P-glycoprotein inhibitor (P-gp inh) reversin 121 (10mM) or the
MRP-1 inhibitor (MRP-1 inh) MK-571 (25 mM), after which CCL2 secretion (C) and expression (D) or interleukin 1b (IL-1b) secretion (E)
was determined by enzyme-linked immunosorbent assay (C and E) or real-time quantitative polymerase chain reaction (D). One hundred
percent corresponds to 18.0 � 0.46mg/ml CCL2 (C) or 0.032 � 0.003 CCL2 expression relative to glyceraldehyde 3-phosphate de-
hydrogenase (D). Experiments were performed in triplicate using three different human donors and were presented as the mean � SEM.
*P5 0.05, **P50.01, ***P5 0.001 by Student’s t-test.
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 567
across brain endothelial cells. Recently, we reported that
P-glycoprotein knockout mice developed reduced clinical signs of
experimental autoimmune encephalomyelitis (Kooij et al., 2009),
illustrating that ABC transporters may be considered as a potential
therapeutic target. However, as ABC transporters such as
P-glycoprotein or MRP-1 are widely expressed on a variety of
cells, including cells of the immune system, it is not feasible to
specifically inhibit ABC transporters solely on CNS cells, such as
astrocytes. Therefore, further research is warranted to generate
mice that lack astrocyte-specific ABC transporters to unravel the
role of astrocytic ABC transporters in multiple sclerosis pathogen-
esis in detail.
In conclusion, we show here that ABC transporter expression
is markedly altered in multiple sclerosis brain tissue. In particu-
lar, hypertrophic reactive astrocytes and infiltrating foamy
macrophages show high expression levels of different ABC trans-
porters, which coincide with increased transporter activity in vitro
under inflammatory conditions. Moreover, the ABC transporters
P-glycoprotein and MRP-1 were shown to mediate CCL2 secretion
from reactive astrocytes, thereby controlling monocyte migration
across a blood–brain barrier model. Our study provides first
evidence for a novel detrimental role of ABC transporters on
reactive astrocytes under pathological conditions, and may open
therapeutic avenues to diminish the neuroinflammatory attack
during multiple sclerosis pathology.
FundingNetherlands Organization of Scientific Research (grant
016.046.314 to G.K.); the Dutch foundation of Multiple Sclerosis
Research (grant MS 08-652 to HE.dV and G.K.; grant MS 07-615
to M.M.; grant MS 05-567, J.vH. and J.D.; grant MS 05-358c,
J.vH.); Top Institute Pharma (A.R.); the Dutch brain foundation.
Figure 7 P-glycoprotein and MRP-1 on reactive astrocytes mediate monocyte migration. Conditioned media from control astrocytes,
polyinosinic–cytidylic acid (poly I:C) treated astrocytes or multiple sclerosis (MS) lesion derived astrocytes [cultured in the presence or
absence of specific P-glycoprotein (P-gp inh) or MRP-1 (MRP-1 inh) inhibitors] was used to assess its capacity to attract monocytes in a
chemotaxis assay (A) or in a Transwell migration assay (B), which represents an in vitro model of the blood–brain barrier. Data is presented
as absolute numbers of migrated monocytes compared with the total number of monocytes added to each well. Experiments were
performed in triplicate using three different human donors (both control and multiple sclerosis donors) and are presented as the
mean � SEM. *P50.05, **P50.01, ***P50.001 compared with vehicle treated control astrocytes; ###P5 0.001 compared with
vehicle polyinosinic–cytidylic acid-treated astrocytes; $P50.05, $$$P50.001 compared with vehicle multiple sclerosis lesion astrocytes as
determined by Student’s t-test.
568 | Brain 2011: 134; 555–570 G. Kooij et al.
Supplementary materialSupplementary material is available at Brain online.
ReferencesAronica E, Gorter JA, Jansen GH, van Veelen CW, van Rijen PC,
Leenstra S, et al. Expression and cellular distribution of multidrugtransporter proteins in two major causes of medically intractable epi-
lepsy: focal cortical dysplasia and glioneuronal tumors. Neuroscience
2003; 118: 417–29.
Bauer B, Hartz AM, Miller DS. Tumor necrosis factor alpha and
endothelin-1 increase P-glycoprotein expression and transport activityat the blood-brain barrier. Mol Pharmacol 2007; 71: 667–75.
Borst P, Elferink RO. Mammalian ABC transporters in health and disease.
Annu Rev Biochem 2002; 71: 537–92.
Boven LA, Van Meurs M, Van Zwam M, Wierenga-Wolf A, Hintzen RQ,
Boot RG, et al. Myelin-laden macrophages are anti-inflammatory, con-sistent with foam cells in multiple sclerosis. Brain 2006; 129: 517–26.
Bruck W, Sommermeier N, Bergmann M, Zettl U, Goebel HH,
Kretzschmar HA, et al. Macrophages in multiple sclerosis.
Immunobiology 1996; 195: 588–600.
Choudhuri S, Klaassen CD. Structure, function, expression, genomic or-ganization, and single nucleotide polymorphisms of human ABCB1
(MDR1), ABCC (MRP), and ABCG2 (BCRP) efflux transporters. Int J
Toxicol 2006; 25: 231–59.
De Groot CJ, Langeveld CH, Jongenelen CA, Montagne L, van der
Valk P, Dijkstra CD. Establishment of human adult astrocyte culturesderived from postmortem multiple sclerosis and control brain and
spinal cord regions: immunophenotypical and functional characteriza-
tion. J Neurosci Res 1997; 49: 342–54.
Drach J, Gsur A, Hamilton G, Zhao S, Angerler J, Fiegl M, et al.
Involvement of P-glycoprotein in the transmembrane transport ofinterleukin-2 (IL-2), IL-4, and interferon-gamma in normal human
T lymphocytes. Blood 1996; 88: 1747–54.
Elkord E, Williams PE, Kynaston H, Rowbottom AW. Human monocyte
isolation methods influence cytokine production from in vitro gener-
ated dendritic cells. Immunology 2005; 114: 204–12.Eng LF, Ghirnikar RS. GFAP and astrogliosis. Brain Pathol 1994; 4:
229–37.
Ernest S, Bello-Reuss E. Secretion of platelet-activating factor is mediated
by MDR1 P-glycoprotein in cultured human mesangial cells. J Am Soc
Nephrol 1999; 10: 2306–13.Ewing C, Bernard CC. Insights into the aetiology and pathogenesis of
multiple sclerosis. Immunol Cell Biol 1998; 76: 47–54.
Fife BT, Huffnagle GB, Kuziel WA, Karpus WJ. CC chemokine receptor 2
is critical for induction of experimental autoimmune encephalomyelitis.J Exp Med 2000; 192: 899–905.
Frank MH, Denton MD, Alexander SI, Khoury SJ, Sayegh MH,
Briscoe DM. Specific MDR1 P-glycoprotein blockade inhibits human
alloimmune T cell activation in vitro. J Immunol 2001; 166: 2451–9.
Frevert CW, Wong VA, Goodman RB, Goodwin R, Martin TR. Rapidfluorescence-based measurement of neutrophil migration in vitro. J
Immunol Methods 1998; 213: 41–52.
Frohman EM, Racke MK, Raine CS. Multiple sclerosis–the plaque and its
pathogenesis. N Engl J Med 2006; 354: 942–55.
Garcia-Vallejo JJ, Van Het HB, Robben J, Van Wijk JA, Van Die I,Joziasse DH, et al. Approach for defining endogenous reference
genes in gene expression experiments. Anal Biochem 2004; 329:
293–9.
Goralski KB, Hartmann G, Piquette-Miller M, Renton KW.
Downregulation of mdr1a expression in the brain and liver duringCNS inflammation alters the in vivo disposition of digoxin. Br J
Pharmacol 2003; 139: 35–48.
Huang YH, Schafer-Elinder L, Wu R, Claesson HE, Frostegard J.
Lysophosphatidylcholine (LPC) induces proinflammatory cytokines by
a platelet-activating factor (PAF) receptor-dependent mechanism. Clin
Exp Immunol 1999; 116: 326–31.
Izikson L, Klein RS, Charo IF, Weiner HL, Luster AD. Resistance to
experimental autoimmune encephalomyelitis in mice lacking the CC
chemokine receptor (CCR)2. J Exp Med 2000; 192: 1075–80.
Kooij G, Backer R, Koning JJ, Reijerkerk A, van Horssen J, van der
Pol SM, et al. P-glycoprotein acts as an immunomodulator during
neuroinflammation. PLoS One 2009; 4: e8212.
Kooij G, van Horssen J, de Lange EC, Reijerkerk A, van der Pol SM,
van het Hof B, et al. T lymphocytes impair P-glycoprotein function
during neuroinflammation. J Autoimmun 2010; 34: 416–25.Koubeissi A, Raad I, Ettouati L, Guilet D, Dumontet C, Paris J. Inhibition
of P-glycoprotein-mediated multidrug efflux by aminomethylene and
ketomethylene analogs of reversins. Bioorg Med Chem Lett 2006; 16:
5700–3.
Krishnamurthy P, Schuetz JD. Role of ABCG2/BCRP in biology and medi-
cine. Annu Rev Pharmacol Toxicol 2006; 46: 381–410.
Lassmann H, Bruck W, Lucchinetti C. Heterogeneity of multiple sclerosis
pathogenesis: implications for diagnosis and therapy. Trends Mol Med
2001; 7: 115–21.
Li H, Cuzner ML, Newcombe J. Microglia-derived macrophages in early
multiple sclerosis plaques. Neuropathol Appl Neurobiol 1996; 22:
207–15.Loscher W, Potschka H. Blood-brain barrier active efflux
transporters: ATP-binding cassette gene family. NeuroRx 2005a; 2:
86–98.
Loscher W, Potschka H. Drug resistance in brain diseases and the role of
drug efflux transporters. Nat Rev Neurosci 2005b; 6: 591–602.
Lucchinetti C, Bruck W, Noseworthy J. Multiple sclerosis: recent devel-
opments in neuropathology, pathogenesis, magnetic resonance ima-
ging studies and treatment. Curr Opin Neurol 2001; 14: 259–69.
Marroni M, Agrawal ML, Kight K, Hallene KL, Hossain M, Cucullo L,
et al. Relationship between expression of multiple drug resistance
proteins and p53 tumor suppressor gene proteins in human brain
astrocytes. Neuroscience 2003; 121: 605–17.Meijer OC, de Lange EC, Breimer DD, de Boer AG, Workel JO,
de Kloet ER. Penetration of dexamethasone into brain glucocorticoid
targets is enhanced in mdr1A P-glycoprotein knockout mice.
Endocrinology 1998; 139: 1789–93.
Nies AT, Jedlitschky G, Konig J, Herold-Mende C, Steiner HH,
Schmitt HP, et al. Expression and immunolocalization of the multidrug
resistance proteins, MRP1-MRP6 (ABCC1-ABCC6), in human brain.
Neuroscience 2004; 129: 349–60.Obata K, Katsura H, Miyoshi K, Kondo T, Yamanaka H, Kobayashi K,
et al. Toll-like receptor 3 contributes to spinal glial activation and tact-
ile allodynia after nerve injury. J Neurochem 2008; 105: 2249–59.
Oerlemans R, van der HJ, Vink J, Dijkmans BA, Kaspers GJ, Lems WF,
et al. Acquired resistance to chloroquine in human CEM T cells is
mediated by multidrug resistance-associated protein 1 and provokes
high levels of cross-resistance to glucocorticoids. Arthritis Rheum 2006;
54: 557–68.Out R, Jessup W, Le GW, Hoekstra M, Gelissen IC, Zhao Y, et al.
Coexistence of foam cells and hypocholesterolemia in mice lacking
the ABC transporters A1 and G1. Circ Res 2008; 102: 113–20.
Pekny M, Nilsson M. Astrocyte activation and reactive gliosis. Glia 2005;
50: 427–34.
Polman CH, O’Connor PW, Havrdova E, Hutchinson M, Kappos L,
Miller DH, et al. A randomized, placebo-controlled trial of natalizumab
for relapsing multiple sclerosis. N Engl J Med 2006; 354: 899–910.
Potschka H, Fedrowitz M, Loscher W. Multidrug resistance protein
MRP2 contributes to blood-brain barrier function and restricts antiepi-
leptic drug activity. J Pharmacol Exp Ther 2003; 306: 124–31.
Raggers RJ, Vogels I, van MG. Multidrug-resistance P-glycoprotein
(MDR1) secretes platelet-activating factor. Biochem J 2001; 357:
859–65.Rao VV, Dahlheimer JL, Bardgett ME, Snyder AZ, Finch RA,
Sartorelli AC, et al. Choroid plexus epithelial expression of MDR1 P
glycoprotein and multidrug resistance-associated protein contribute to
ATP-binding cassette transporters in multiple sclerosis Brain 2011: 134; 555–570 | 569
the blood-cerebrospinal-fluid drug-permeability barrier. Proc Natl AcadSci USA 1999; 96: 3900–5.
Ridet JL, Malhotra SK, Privat A, Gage FH. Reactive astrocytes: cellular
and molecular cues to biological function. Trends Neurosci 1997; 20:
570–7.Sisodiya SM, Lin WR, Harding BN, Squier MV, Thom M. Drug resistance
in epilepsy: expression of drug resistance proteins in common causes
of refractory epilepsy. Brain 2002; 125: 22–31.
Sofroniew MV. Molecular dissection of reactive astrogliosis and glial scarformation. Trends Neurosci 2009; 32: 638–47.
Speth C, Dierich MP, Sopper S. HIV-infection of the central nervous
system: the tightrope walk of innate immunity. Mol Immunol 2005;42: 213–28.
Sukhai M, Piquette-Miller M. Regulation of the multidrug resistance
genes by stress signals. J Pharm Pharm Sci 2000; 3: 268–80.
Tani M, Glabinski AR, Tuohy VK, Stoler MH, Estes ML, Ransohoff RM.In situ hybridization analysis of glial fibrillary acidic protein mRNA re-
veals evidence of biphasic astrocyte activation during acute experi-
mental autoimmune encephalomyelitis. Am J Pathol 1996; 148:
889–96.Tani M, Ransohoff RM. Do chemokines mediate inflammatory cell inva-
sion of the central nervous system parenchyma? Brain Pathol 1994; 4:
135–43.
Tekstra J, Visser CE, Tuk CW, Brouwer-Steenbergen JJ, Burger CW,Krediet RT, et al. Identification of the major chemokines that regulate
cell influxes in peritoneal dialysis patients. J Am Soc Nephrol 1996; 7:
2379–84.van de Ven R, de Jong MC, Reurs AW, Schoonderwoerd AJ, Jansen G,
Hooijberg JH, et al. Dendritic cells require multidrug resistance protein
1 (ABCC1) transporter activity for differentiation. J Immunol 2006;
176: 5191–8.van de Ven R, Scheffer GL, Scheper RJ, de Gruijl TD. The ABC of dendritic
cell development and function. Trends Immunol 2009; 30: 421–9.
Van der Goes A, Boorsma W, Hoekstra K, Montagne L, De Groot CJ,
Dijkstra CD. Determination of the sequential degradation of myelinproteins by macrophages. J Neuroimmunol 2005; 161: 12–20.
van der Heijden JW, Oerlemans R, Tak PP, Assaraf YG, Kraan MC,
Scheffer GL, et al. Involvement of breast cancer resistance protein
expression on rheumatoid arthritis synovial tissue macrophages in
resistance to methotrexate and leflunomide. Arthritis Rheum 2009;
60: 669–77.
van der Pol MA, Feller N, Ossenkoppele GJ, Weijers GW, Westra AH,
van Stijn A, et al. Minimal residual disease in acute myeloid leukemia is
predicted by P-glycoprotein activity but not by multidrug resistance
protein activity at diagnosis. Leukemia 2003; 17: 1674–7.
van der Valk P, De Groot CJ. Staging of multiple sclerosis (MS) lesions:
pathology of the time frame of MS. Neuropathol Appl Neurobiol
2000; 26: 2–10.
van der Voorn P, Tekstra J, Beelen RH, Tensen CP, van der Valk P,
De Groot CJ. Expression of MCP-1 by reactive astrocytes in demyeli-
nating multiple sclerosis lesions. Am J Pathol 1999; 154: 45–51.
van Horssen J, Bo L, Dijkstra CD, de Vries HE. Extensive extracellular
matrix depositions in active multiple sclerosis lesions. Neurobiol Dis
2006a; 24: 484–91.
van Horssen J, Vos CM, Admiraal L, van Haastert ES, Montagne L,
van der Valk P, et al. Matrix metalloproteinase-19 is highly expressed
in active multiple sclerosis lesions. Neuropathol Appl Neurobiol 2006b;
32: 585–93.
Viegas P, Chaverot N, Enslen H, Perriere N, Couraud PO, Cazaubon S.
Junctional expression of the prion protein PrPC by brain endothelial
cells: a role in trans-endothelial migration of human monocytes. J Cell
Sci 2006; 119: 4634–43.
Wang Q, Symes AJ, Kane CA, Freeman A, Nariculam J, Munson P, et al.
A novel role for Wnt/Ca2 + signaling in actin cytoskeleton remodeling
and cell motility in prostate cancer. PLoS One 2010; 5: e10456.
Weksler BB, Subileau EA, Perriere N, Charneau P, Holloway K,
Leveque M, et al. Blood-brain barrier-specific properties of a human
adult brain endothelial cell line. FASEB J 2005; 19: 1872–4.
Zhu HJ, Liu GQ. Glutamate up-regulates P-glycoprotein expression in rat
brain microvessel endothelial cells by an NMDA receptor-mediated
mechanism. Life Sci 2004; 75: 1313–22.
570 | Brain 2011: 134; 555–570 G. Kooij et al.