Skip to main content
NCBI home page
Brain Pathology logoLink to Brain Pathology
. 2010 Mar 11;20(5):926–935. doi: 10.1111/j.1750-3639.2010.00389.x

P‐glycoprotein (ABCB1) and Breast Cancer Resistance Protein (ABCG2) Localize in the Microvessels Forming the Blood‐Tumor Barrier in Ependymomas

Carole Ginguené 1,2, Jacques Champier 1, Said Maallem 1, Nathalie Strazielle 1,3, Anne Jouvet 1,4, Michelle Fèvre‐Montange 1,†, Jean‐François Ghersi‐Egea 1,†,✉
PMCID: PMC8094675  PMID: 20406235

Abstract

Ependymomas are neuroepithelial tumors that arise from the ependymal layer bordering the cerebral ventricles and spinal canal. Intracranial ependymoma represents a major encephalic tumor in children, while spinal ependymoma develops more frequently in adults. To understand the pharmacoresistance that characterizes this tumoral entity, we analyzed the level of expression and localization of three major efflux transport proteins with a multidrug resistance function, P‐glycoprotein, multidrug resistance‐related protein 1 (MRP1) and breast cancer resistance protein (BCRP), in a series of 25 ependymomas from both children and adults. Real‐time‐PCR analysis showed that all three genes were expressed in all tumors, with no apparent correlation between the level of expression and either age or tumor grade. The MRP1 transcript was expressed at a significantly higher level in spinal tumors than in intracranial tumors. The expression of the proteins corresponding to these genes was confirmed by Western blot analysis. In an immunohistochemical study, P‐glycoprotein and BCRP were shown to be associated with the tumoral vessels, where they presented a luminal localization, a prerequisite for their efflux drug activity into the blood. These data indicate that a biochemical, transporter‐dependent blood–tumor barrier may exist in ependymomas, which may reduce the tumoral bioavailability of lipophilic and amphiphilic anticancer drugs.

Keywords: ATP‐binding cassette, BCRP, drug resistance, ependymoma, MRP1, P‐glycoprotein

INTRODUCTION

Ependymomas, neuroepithelial tumors that arise from the ependymal cells of the cerebral ventricles or the central canal of the spinal cord, constitute 3–9% of all intracranial malignancies and are the third most common brain tumors in children (16). The clinical management of these tumors remains difficult. Surgical treatment seems to be a determining factor for a good outcome for these tumors 26, 27, but most childhood ependymomas arise in the posterior fossa and are difficult to resect (24). There is general acceptance that adjuvant therapy is required even when complete resection is achieved. Although recent advances in radiation therapy appear promising, the benefit of this treatment is not clearly established, and radiation is a risk factor for younger patients 32, 36. A variety of chemotherapy protocols has been introduced for the treatment of ependymoma, but with no significant positive contribution to patient outcome 4, 20, 32. The ineffectiveness of brain tumor chemotherapy could be due to intrinsic or acquired tumor cell chemoresistance. It might also result from a low efficiency of drug penetration through the blood–brain and blood–tumor barriers, resulting in reduced, non‐cytotoxic local drug concentrations.

Several efflux transport proteins of the ATP‐binding cassette (ABC) transporter superfamily confer a multidrug resistance phenotype to various tumoral cells and are also constitutively expressed in the cerebral microvessels forming the blood–brain barrier or in the choroidal epithelium, which forms the blood–cerebrospinal fluid (CSF) barrier and is in continuity with the ependyma. These proteins therefore participate in the barrier function attributed to both cellular interfaces. In normal human brain, P‐glycoprotein (Pgp; gene symbol ABCB1) and breast cancer resistance protein (BCRP; gene symbol ABCG2, also known as mitoxantrone resistance protein), which accept many lipophilic drugs as substrates, are expressed at high levels at the blood–brain barrier 9, 10, 12, 17, 23, while multidrug resistance‐associated protein 1 (MRP1; gene symbol ABCC1) is highly expressed at the blood–CSF barrier. MRP1 is involved in multidrug resistance by mediating the transport of different native or glutathione‐conjugated drugs 17, 23. Reports on the level of expression and localization of these chemoresistance‐related proteins in ependymomas are scarce and mostly limited to Pgp. Few real time (RT)‐PCR and immunohistochemical studies have reported Pgp expression in ependymomas, the protein being detected in a proportion of neoplastic cells 8, 18, while other studies have reported that Pgp is only associated with tumoral vessels (3) or with tumoral vessels and some tumor cells in some ependymomas (22). In the last study, surprisingly, the progression‐free survival time of intracranial ependymomas was significantly shorter for immunonegative tumors (22). Only one study on a large number of brain tumors of different types, including a limited group of ependymomas, has described modest MRP1 and BCRP mRNA levels in addition to Pgp mRNA (40).

In this study, we report the expression profile of the Pgp, MRP1 and BCRP genes in a series of ependymomas of different grades and locations. Using double‐labeling immunohistochemical techniques, we also demonstrate the specific localization of Pgp and BCRP in the tumoral vessels, potentially highlighting their role in the blood‐tumor barrier function.

METHODS

Patients, tumors and histological diagnosis

Tumor samples from 29 patients were obtained from surgical resections at the Neurological and Neurosurgical Hospital Pierre Wertheimer (Groupement Hospitalier Est, GPE). The resected tissue was divided into two fragments, one of which was fixed for routine histopathological analysis, while the other was frozen and stored in liquid nitrogen in the Biological Resource Center NeuroBioTec Banques, Lyon, France, for RNA extraction, and, when the size of the fragment was adequate, immunohistochemistry. Histological examination of paraffin sections stained with hemalin‐phloxin saffron was carried out at the Department of Neuropathology (Groupement Hospitalier Est, F‐Bron). Diagnosis and classification into histological subtypes were based on WHO standard diagnostic criteria (25). Twenty‐five tumors were ependymomas, one tumor was a subependymoma and three tumors were myxopapillary ependymomas. Two cortical samples and one choroid plexus sample were obtained at autopsy from one male and one female adult within 24 h after death from accidents, in accordance with the ethical guidelines of INSERM. They were procured by the Neuropathology Department of the Neurological Hospital, centralized via the Biological Resource Center, and kept at 4°C in Krebs‐Ringer buffer prior to processing for RNA extraction. An additional RNA sample from a normal adult whole human brain (single donor, male, 72 years) was purchased from Stratagene Europe (Amsterdam, the Netherlands). The human breast adenocarcinoma cell line MCF7 was obtained from the American Type Culture Collection (ATCC), cultured according to the ATCC protocol, and harvested at subconfluence for use as a positive control for BCRP detection by Western blotting.

RNA extraction

Total RNA was extracted from the samples using the RNA Plus™ (Qbiogen, Illkirch, France) procedure based on the method of Chomczinski and Sacchi (7) and precipitated with ethanol. The quality of the isolated total RNA was evaluated on nanochips using an Agilent 2100™ bioanalyzer (Agilent Technologies, Massy, France). The samples used in the study showed clear 18S and 28S ribosomal RNA peaks. As another index of mRNA quality, GAPDH mRNA levels were similar in the different samples.

Real‐time PCR

RNA samples (0.5 µg) from tumors and normal brain were heat denatured for 3 minutes at 75°C, then immediately placed on ice. First‐strand DNA was synthesized by incubating the RNA with 0.5 mM of each dNTP, 10 mM DTT, 40 U of RNasin® (Promega, Charbonnières‐les‐Bains, France), 20 µM random hexamers, and 200 U of Moloney murine leukemia virus (M‐MLV) reverse transcriptase (Invitrogen, Cergy‐Pontoise, France) for 90 minutes at 42°C in a final volume of 20 µL of reverse transcriptase buffer (50 mM Tris‐HCl, pH 8.3, 75 mM KCl, and 3 mM MgCl2). The volume was then made to 100 µL with distilled water. Negative controls were performed by replacing the enzyme with water.

PCR was performed on a LightCycler™ instrument (Roche Diagnostics, Mannheim, Germany). cDNA samples (2, 0.2, and 0.02 µL) were diluted in glass capillaries to a volume of 20 µL with PCR mix (LightCycler® Faststart DNA Master Plus SYBR Green 1™, Roche Diagnostics) containing a final concentration of 4 mM MgCl2 and 0.5 µM 3′ and 5′ primers. The oligonucleotide sequences corresponding to the selected genes were designed using Primer 3 software (Infobiogen, Villejuif, France), and are listed in Table S1. The cDNA was denatured for 8 minutes at 95°C, then amplified by 40–50 cycles of 15 s at 95°C, 5 s at 62°C, and 12 s at 72°C. Assessment of single transcript production, negative controls, and quantification of transporter mRNAs relative to the housekeeping gene GAPDH were performed as described previously (15). The presence of a single PCR product of the correct size was assessed by electrophoresis for 1 h at 70 V in 45.5 mM Tris, 45,5 mM borate, 1 mM EDTA buffer, pH 8.3, on a 2% agarose (Tebu‐Bio, Le Perray‐en‐Yvelines, France) gel using a DNA molecular weight standard (100 bp, Promega).

Western blot analysis

Ependymomas were homogenized at 4°C using a Dounce‐type glass‐glass homogenizer in lysis buffer (#9803, Cell Signaling Technology, Danvers, MA, USA), supplemented with 1 mM phenylmethylsulfonyl fluoride (Sigma‐Aldrich, Saint‐Quentin Fallavier, France). The protein content of the samples was determined by the method of Peterson (29), with bovine serum albumin (BSA) as the standard.

The proteins were solubilized at room temperature in 1% sodium dodecyl sulfate (SDS), 10 % glycerol, 0.1 M DTT and 0.01% bromophenol blue (all v/v), then separated by 10% polyacrylamide SDS‐PAGE, using 25 mM Tris, 192 mM glycine, 0.1% sodium dodecyl sulfate, pH 8.5, as migration buffer. The proteins were transferred to nitrocellulose membranes (Protran®, 0.4 µm pores, Whatman, Schleicher & Schuell, Dassel, Germany) at 100 V in migration buffer/methanol/water (5/2/3 by volume) for 2 h at 4°C. Migration and transfer efficacy were assessed by staining with Ponceau red (0.25% in 3% trichloroacetic acid). After saturation for 1 h at room temperature in blocking solution (5% non fat milk, 0.1% Tween‐20 in 50 mM Tris‐buffered saline, pH 7.4), the membranes were incubated overnight at 4°C with 4.8 µg/mL of mouse anti‐Pgp monoclonal antibody C219 (Calbiochem, Darmstadt, Germany, #517310), 2.5 µg/mL of mouse antihuman BCRP monoclonal antibody BXP21 (Alexis Biochemicals, Lausen, Switzerland, #ALX‐801‐029), 1 µg/mL of rabbit antihuman MRP1 polyclonal antibodies A23 (Alexis Biochemicals, #ALX‐210‐841) or 8 µg/mL of rat antihuman MRP1 monoclonal antibody MRPr1 (Alexis Biochemicals, #ALX‐801‐007) in blocking solution. After three rinses, the membranes were incubated for 2 h at room temperature with secondary horseradish peroxidase‐conjugated antimouse, antirabbit, or antirat IgG antibodies (0.06, 0.06 and 0.04 µg/mL in blocking solution, Jackson Immuno Research, West Grove, PA, USA, #115‐36‐003, #111‐36‐003, and #112‐036‐003, respectively), then HRP activity was developed using a chemiluminescence procedure (Millipore™, Billerica, MA, USA) according to the manufacturer's protocol, and visualized on X‐ray films (Biomax®, Kodak™ Rochester, NY, USA). The membranes were stripped in 25 mM glycine, 1% SDS buffer, pH 2, for 30 minutes at room temperature, and analyzed using rabbit anti‐actin polyclonal antibodies (1 µg/mL, Sigma‐Aldrich®, #A2066) as above as a protein loading control.

Immunohistochemistry

Unless stated otherwise, all stages were performed at room temperature. Ten micrometer‐thick slices were cryosectioned from 15 tumoral resections and were fixed for 30 s in 1% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.4 (PB) for Pgp, for 10 minutes in 4 % paraformaldehyde in PB for MRP1, or for 8 minutes in acetone for BCRP, then blocked for 1 h in 5% BSA, 5% normal goat serum, 0.3% Triton X‐100 in phosphate‐buffered saline, pH 7.4 (PBS). Primary antibodies were added at a final concentration of 6.025 µg/mL (C219), 2.5 µg/mL (BXP21), 1 µg/mL (A23) 8 µg/mL (MRPr1) in blocking solution and the sections incubated overnight at 4°C, then reacted for 2 h with secondary antimouse or antirabbit Alexa‐fluor 488‐conjugated antibodies (2 µg/mL, Molecular Probe, Invitrogen, Cergy Pontoise, France) in PBS containing 0.3% Triton, and the nuclei stained with DAPI (0.1 µg/mL) in PBS for 10 minutes.

Double immunostaining for ABC proteins and either von Willebrand factor (vWf) or laminin was performed using 6.2 and 12.5 µg/mL of rabbit polyclonal antibodies against vWf (Dako, Glostrup, Denmark; #A0082) or laminin (Novotec, Saint Martin La Garenne, France, #24851), respectively. To avoid spillover of the signal from the vWf or laminin labeling into the ABC protein signal channel, an Alexa‐fluor 555 labeled anti‐mouse or anti‐rat IgG secondary antibody was used to detect the ABC proteins, while the anti‐vWf or anti‐laminin antibodies were detected using an anti‐rabbit Alexa 488‐conjugated secondary antibody. However, for clarity, all Pgp, BCRP and MRP1 labelings are shown in green in 3, 4. The sections were mounted using Fluoroprep™ (Biomerieux, Marcy l'Etoile, France). Negative controls were performed by omitting the first antibody. The sections were examined by fluorescence microscopy using a Zeiss Axioplan II™ microscope (Le Pecq, France), equipped with an F‐View II digital camera and AnalySIS auto software (Olympus, Soft Imaging System™, Münster, Germany).

Figure 3.

Figure 3

Open in a new tab

Immunohistochemical detection of Pgp and BCRP proteins in ependymomas. Pgp and BCRP labeling is shown in green (see Methods). A–G. Pgp immunostaining. A. Low magnification micrograph of Pgp immunodetection showing a strong signal associated with the microvasculature in an intracranial ependymoma. No signal was detected in the tumoral cells. B. A negative control without primary antibody is shown for the adjacent slice. C. High magnification micrograph highlighting a similar labeling of the vasculature in spinal ependymoma. D. Double immunostaining for Pgp and the endothelial marker vWf (red), confirming the localization of Pgp in microvessels. E. In places where the endothelial nucleus is side located, Pgp staining is seen only on the capillary lumen‐facing side of the nucleus (arrow), revealing a luminal localization of the protein. F. Double immunostaining for Pgp and laminin (red), showing the luminal localization of Pgp. In contrast with the Pgp staining of the capillary lumen‐facing side of the nucleus, laminin labeling generates a signal on the other side of the nucleus, as illustrated by the red/green(lumen)/blue/red color sequence along the arrow. G. In one tumor, a few non‐endothelial cells with round nuclei stained positive for Pgp (arrow). H–N. BCRP immunostaining. H. BCRP immunodetection reveals a signal associated with the microvasculature in intracranial ependymoma. I. A negative control without the primary antibody is shown for the adjacent slice. J. Higher magnification highlighting the positive signal associated with microvessels in the spinal tumor. K. This association was confirmed by double immunostaining for BCRP and vWf (red), which also shows that the signal is endothelial rather than perivascular. L. The BCRP signal is visible on one side of side‐located nuclei along the capillaries (arrows), suggesting a typical luminal distribution of the protein. M. Double immunostaining for BCRP and laminin (red) observed at high magnification, confirming this polarized distribution on one side of the endothelial nuclei (E), with the typical red/green(lumen)/blue/red color sequence along the arrow. This also shows that endothelial nuclei can be distinguished from pericyte or perivascular cell nuclei (P), a potential source of misinterpretation of the images, as the latter are separated from the endothelium by a laminin layer. The plain and open arrowheads point to laminin‐labeled basal membranes of vessels and of epithelial tumoral cells, respectively. N. In a few tumors, occasional signals not co‐localized with vWf‐expressing structures can be seen, surrounding the nuclei of individual cells (arrows). Nuclei were labeled with DAPI. The micrographs in panels A and B are from case 5 (Ic), those in panels C and M from case 17 (Sp), those in panels D and G from case 4 (Ic), those in panels E and J from case 25 (Sp), that in panel F from case 15 (Ic), those in panels H, I, K and N from case 13 (Ic), and that in panel L from case 18 (Sp). Pgp = P‐glycoprotein; BCRP = breast cancer resistance protein; MV = microvessels; Ic = Intracranial tumor; Sp = spinal tumor.

Figure 4.

Figure 4

Open in a new tab

Detection of MRP1 in ependymomas using antibodies A23 and MRPr1. A. Typical Western blot signals observed in tissue homogenates (30 µg of protein per lane) of spinal (cases 17, 25) and intracranial (cases 5, 13) tumor resections. Human choroid plexus homogenate (hCP, 20 µg protein load) was used as a positive control for antibody MRPr1 reactivity and specificity. Antibody A23 has been previously validated for our Western blot protocol (17). In all tumors, a 190 kDa band of the expected molecular weight for MRP1 was detected by both antibodies. The main additional signals around 90–100 kDa were also observed in control lanes treated with the antirabbit immunoglobulin or anti‐rat immunoglobulin secondary antibody in absence of primary antibody. B–D. MRP1 immunostaining using antibody A23. Immunodetection reveals a signal associated with the tumoral cells in intracranial (B) and spinal (D) ependymoma. C. Negative control without the primary antibody. E–H. MRP1 immunostaining using antibody MRPr1. Immunostaining of a human choroid plexus using MRPr1 antibody showing the expected basolateral signal at the choroidal epithelium (G), which is absent (H) when the primary antibody is omitted. MRP1 immunostaining of ependymomas using antibody MRPr1 (E) is similar to that seen with antibody A23, and double immunolabeling for vWf (red, in E) allows the visualization of tumoral vessels, which display little, or no MRP1 immunostaining (arrows in F). The micrographs in panels B, C, E and F are from case 15 (Ic), and those in panel D from case 20 (Sp). MRP1 = multidrug resistance‐related protein 1.

RESULTS

Clinical data

The clinico‐pathological findings for the 29 patients are shown in Table 1. According to the WHO classification, 16 tumors were intracranial ependymomas (11 pediatric and 5 adult patients), 11 being grade II and 5 grade III. Nine tumors were grade II spinal ependymomas from adult patients, while the other four, also from adults, were grade I and consisted of one subependymoma and three myxopapillary ependymomas. Twenty‐one tumors were completely resected and eight partially removed. Postsurgical recurrence occurred in 11 patients.

Table 1.

Clinical data for the 29 patients included in the study.

Cases Histopathological subtype WHO grade Sex/Age Location Surgery Recurrence Treatment Post‐surgery
1 Ic II F/1 PF + + no
2 Ic II M/2 PF + + CT
3 Ic II M/2 PF + − RT + CT
4 Ic II F/2.5 LO +/− + CT
5 Ic II M/6 PF + − RT
6 Ic II M/14 PF +/− + RT + CT
7 Ic II F/21 ST +/− + CT
8 Ic II M/26 PF +/− − RT + CT
9 Ic II F/33 PF + − no
10 Ic II F/44 ST + − no
11 Ic II F/50 PF + − no
12 Ic III F/1 PF + + CT
13 Ic III M/2.5 PF + + RT + CT
14 Ic recurrence III M/8 PF +/− + no
15 Ic III M/15 PF + − RT
16 Ic recurrence III F/18 PF + + RT
17 Sp II M/27 L, T +/− + RT
18 Sp  II M/29 C +/− − no
19 Sp recurrence II M/30 T +/− + no
20 Sp II F/38 C + − no
21 Sp II M/39 C + − no
22 Sp II M/41 L + − no
23 Sp II M/51 T + − RT
24 Sp II F/62 L + − no
25 Sp II M/63 C, T + − no
26 SubE I M/38 V3, LLV + − no
27 Myx I M/33 CE + − no
28 Myx I F/36 CE + − no
29 Myx I M/44 CE + − no
Open in a new tab

Ic: intracranial ependymoma, Sp: spinal ependymoma, SubE: subependymoma, Myx: myxopapillary ependymoma, PF: posterior fossa, LO: left occipital, C: cervical segment, ST: supratentorial, L: lumbar segment, CE: cauda equina, T: thoracic segment, V3: third ventricle, LLV: left lateral ventricle. For surgery: +: complete, +/−: partial. Post surgical treatment: RT: radiotherapy. CT: chemotherapy as follow: Patients 2,3,4,13, carboplatine or procarbazine, etoposide, cisplatin, cyclophosphamide, and vincristine; patient 6, temozolamide; patients 7 and 8, vincristine, methylprednisolone, hydroxyurea, CCNU, cisplatin, cytosine arabinoside, and decarbazine, then caryolisine, vincristine, procarbasine for patient 7; patient 12, etoposide.

Expression of transcripts for multidrug resistance genes in relation to the primary site of the ependymoma

Relative mRNA levels for the three investigated genes in intracranial and spinal ependymomas and in normal brain are shown in Figure 1. All three genes were detected in all ependymomas analyzed. Levels of Pgp and BCRP mRNAs in both types of ependymomas were low and not significantly higher than those in normal brain. No significant differences were observed in Pgp and BRCP mRNA levels between intracranial and spinal ependymomas. Spinal ependymomas showed significantly higher levels of MRP1 mRNA than intracerebral tumors or normal brain. We found no significant correlation between Pgp, BCRP or MRP1 mRNA levels, and metastatic stage, presence of recurrences or age.

Figure 1.

Figure 1

Open in a new tab

Transcript levels of Pgp, BCRP and MRP determined by quantitative real time‐PCR in intracranial ependymomas (IcEp), spinal ependymomas (SpEp), and normal brain tissue (Nb). The results (arbitrary units, AU) are expressed as the amount of transporter mRNA/amount of GAPDH mRNA (×10−3). The individual data and means ± standard error of the mean are shown. *Significant statistical difference compared to IcEp and Nb (P < 0.05, Kruskal–Wallis variance analysis, followed by multiple comparisons). Pgp = P‐glycoprotein; BCRP = breast cancer resistance protein; MRP1 = multidrug resistance‐related protein 1.

Elevated levels of the three mRNAs were observed in the subependymoma, ranging from 3.5‐ to 4.5‐fold higher than the average level in normal brain. In the three myxopapillary ependymomas, Pgp mRNA levels were very low, while MRP1 and BCRP mRNA levels were close to the levels in spinal ependymomas (data not shown).

Expression and localization of Pgp, BCRP and MRP1 protein in tumoral tissue

Western blot analysis using antibodies C219 and BXP21 revealed a major band with an apparent molecular weight of 170 or 72 kDa corresponding, respectively, to Pgp or BCRP in both intracranial and spinal ependymomas and in the subependymoma (Figure 2). The cellular localization of Pgp and BCRP was then determined by immunohistochemistry in 15 ependymomas using the same antibodies. Pgp was found to be associated with capillaries and larger microvessels in all intracranial and spinal tumors analyzed, although at various labeling intensities (Figure 3A–F, and Table 2). Double immunostaining for Pgp and the endothelial marker vWf confirmed the association of Pgp with tumoral vessels and revealed that Pgp localization was endothelial, rather than perivascular (Figure 3D). All vWf‐labeled vessels were immunoreactive to some extent with anti‐Pgp antibody.

Figure 2.

Figure 2

Open in a new tab

Western blot analysis of Pgp and BCRP expression in ependymomas. Typical signals observed using tissue homogenates (30 µg of protein per lane) of spinal tumor resections from adult patients (cases 17, 18), intracranial tumors from pediatric patients (cases 5, 13) or a subependymoma (case 26) are shown. MCF7 cells (15 µg of protein per lane) were used as a positive control for BXP21 antibody reactivity and specificity. C219 antibody has been previously validated for our Western blot protocol (17). In all tumors, a major band of 170 kDa is detected by C219, and a major strong 72 kDa band is detected by BXP21. 170 kDa and 72 kDa are the expected molecular weights for Pgp and BCRP, respectively. The signal around 100 kDa observed in several samples was also observed in negative control lanes for which the primary antibody was omitted. Pgp = P‐glycoprotein; BCRP = breast cancer resistance protein.

Table 2.

Immunohistochemical detection of Pgp and BCRP in the tumoral neuropil and blood vessel wall in ependymomas including subependymoma and myxopapillary ependymoma.

Case # Pgp BCRP
microvessels tumor cells microvessels tumor cells
4 Ic + Isolated ++ –
5 Ic ++ – ++ –
6 Ic + – ++ Few, in localized areas
9 Ic + – ++ –
13 Ic ++ – ++ Isolated
14 Ic recurrence +/− – ++ –
15 Ic ++ – ++ –
16 Ic recurrence ++ – ++ –
17 Sp ++ – ++ –
18 Sp ++ – ++ –
20 Sp +/− – ++ –
22 Sp + – ++ –
25 Sp + – ++ –
26 SubE ++ Isolated ++ –
29 Myx ++ – ++ –
Open in a new tab

– : no labeling, +/− : weak labeling, + moderate labeling, ++ strong labeling.

Isolated: occasional cells positive for Pgp or BCRP and negative for vWf. The abbreviations are the same as in Table 1.

Upon observation at higher magnification in regions where the endothelial nucleus laterally widened the intracellular endothelial space, double immunostaining for Pgp and laminin, a component of the basement membrane of vessels associated with the abluminal endothelial membrane, revealed a luminal localization for Pgp (Figure 3E and F). In only one case were occasional vWf‐negative tumor cells positive for Pgp (Figure 3G).

BCRP was also detected in the vascular network of all tumors analyzed (Figure 3H–J, and Table 2). Double vWf/BCRP immunostaining confirmed this association and revealed an endothelial, rather than a perivascular, localization (Figure 3K). Vessel sections with side‐located nuclei revealed a typical luminal localization of the protein (Figure 3L), which was confirmed at high magnification by double immunostaining for BCRP and laminin, the latter making it possible to distinguish between nuclei in pericytes/perivascular cells and those in endothelial cells (Figure 3M). In two tumor resections, a few parenchymal cells were positive for BCRP (Figure 3N).

Expression and localization of MRP1 was also evaluated in several tumors using two different antibodies, A23 and MRPr1, which display a good specificity for the protein in human tissue 14, 17, 33. In homogenates of both intracranial and spinal ependymomas, Western blot analysis with either antibody revealed a strong signal with the expected molecular weight of 190 kDa (Figure 4A), demonstrating that the protein was present in these tumors. The other main bands, of lower molecular weight, resulted from non‐specific binding of the secondary antibody to the membrane, as shown by samples tested without the primary antibody. To examine MRP1 localization within the tumoral tissue, five intracranial and spinal tumors were analyzed by immunohistochemistry using both antibodies. As a positive control, the antibodies generated the expected epithelial basolateral signal in the human choroid plexus epithelium (Figure 4G and H). In contrast to Pgp and BCRP, MRP1 was shown to be mainly associated with tumoral cells in all tumors analyzed (Figure 4B–E), and double immunostaining using antibody MRPr1 and anti‐vWf antibodies confirmed that MRP1 was not, or only faintly, associated with tumoral vessels (Figure 4E and F).

DISCUSSION

In an attempt to define the chemotherapeutic resistance of ependymomas, we investigated the expression and localization of ABC transport proteins with multidrug resistance activity in a series of intracranial and spinal ependymomas. Our results showed that these proteins were present in all tumors analyzed, with mRNA levels close to those in normal brain. No significant differences in protein expression were detected between primary and recurrent ependymomas.

Importantly, this work demonstrated that two of these proteins, Pgp and BCRP, were mainly localized in the microvessel endothelium forming the blood–tumor barrier. The distribution of Pgp is in agreement with that previously reported in a study on a collection of tumors from children or young adults, which included eight ependymomas (3). Other authors failed to observe an association of Pgp with the ependymoma microvasculature, instead describing Pgp immunoreactivity in neoplastic cells 8, 18. This discrepancy may arise from the use of different immunohistochemical protocols. In our study, we adapted the fixation procedure for C219 antibody immunostaining by using a very mild aldehyde fixation on frozen sections, as previously described for rodent brain (17), in order to avoid both the nonspecific staining and the loss of epitope recognition that arise from conventional aldehyde fixation 37, 38. Our findings are also in line with the report of Pgp in blood vessels alone or in addition to in tumor cells in other tumors of the central nervous system, such as glioma 3, 13, 21, 39. The localization of BCRP in ependymoma vessels has not been previously reported, and is consistent with the exclusive vascular location of BCRP seen in primary brain tumors from patients diagnosed with epilepsy (1). The immunohistochemical analysis of transporter subcellular distribution in microvessels is difficult due to the close apposition of the luminal and abluminal cell membranes of the endothelial cells, except in certain sections where the intracellular space is greatly widened by the nucleus (17). Analysis of such fields in our study indicated that both Pgp and BCRP were found on the luminal membrane in ependymoma vessels, and thus have the required localization to efficiently mediate the polarized efflux of drugs from the tumoral parenchyma to the blood.

Owing to the fact that intravenously injected MRI contrast agents can penetrate into tumoral tissue, probably as a result of loosened inter‐endothelial tight junctions, brain tumor vessels are frequently considered as “leaky.” However, although this higher paracellular permeability may influence the passive flux of small polar compounds, it will have limited consequences for the bioavailability of lipid‐soluble molecules, such as most anticancer drugs. The bioavailability of these molecules is determined by their diffusion rate across cell membranes and thus depends greatly on the action of transporters, such as ABC transport proteins. Pgp accepts many anticancer drugs, including vinca alcaloids, epidophyllotoxins, anthracyclines, taxanes or imatinib, as substrates (34). Although BCRP was discovered more recently, its list of substrates is growing and includes, in addition to mitoxantrone, other important anticancer drugs, such as topotecan and doxorubicin (31), or the active metabolite of imatinib (2). For both proteins, the functional relevance of their localization at the brain endothelium in preventing their substrates from crossing the blood–brain barrier has been established 2, 35.

Thus, in ependymomas, although the endothelium of the neovessels may not possess as tight interendothelial junctions as in the normal blood–brain barrier, the presence of PgP and BCRP at the luminal membrane of the endothelial cells supports the hypothesis of an efficient biochemical blood–tumor barrier for a large number of lipophilic and amphiphilic anticancer drugs (Figure 5).

Figure 5.

Figure 5

Open in a new tab

Schematic representation of possible mechanisms involved in drug resistance in ependymoma. Intratumoral bioavailability of hydrophilic drugs depends on the degree of interendothelial junctional tightness (1). More lipophilic and amphiphilic drugs will enter the tumoral parenchyma as a function of their diffusion rate across the endothelial membranes (2). Pgp and BCRP can limit the intratumoral concentration by preventing the diffusion of the drug across the blood–tumor barrier (3), and by favoring its elimination from the tumor (4). The tumor cells themselves can be further protected from drugs by other efflux transporters, including multidrug resistance‐related protein 1 (5), or else possibly by enzymatic adaptation that also provides multidrug resistance (6). Pgp =  P‐glycoprotein; BCRP =  breast cancer resistance protein; MDR =  multidrug resistance; en = endothelial cell; ep = ependymoma tumor cell.

In humans, Pgp and BCRP are detectable in brain capillaries prior to birth, and their expression increases during development [reviewed in 11, 19]. Our data, showing that both proteins were expressed in intracranial ependymoma vessels in children, as well as in adults, indicate that young patients are not spared from blood–tumor barrier issues.

Attempts to link the expression levels of ABC transporters, in particular Pgp, in various brain tumors to patient clinical outcome have not been conclusive (5). However, in pharmacoresistant tumors, such as ependymomas, in which we showed that Pgp and BCRP are restricted to the vasculature, representing only a few percent of the tumoral mass, the overall expression of these genes is expected to be moderate. A corollary of this is that low expression levels of ABC proteins determined on overall tumoral tissue may provide little information about the pharmacoresistant phenotype, when the biochemical blood–tumor barrier is able to prevent the accumulation of therapeutic concentrations of drugs in the tumor tissue. This may also explain the lack of any clear correlation between Pgp and BCRP mRNA levels and any of the clinical data evaluated [this study, and (22) for Pgp], as capillary density may vary widely from one tumor to another.

In addition to the tumoral vessels, occasional individual cells were also positive for either Pgp or BCRP. Due to their scarcity, the origin of these cells could not be determined in this work. They may belong to a subset of tumoral ependymal cells, or may be infiltrating immune cells such as macrophages, which express BCRP (41). In fact, cells positive for CD68, a marker of monocytes/macrophages, were found in these tumors (data not shown). The BCRP‐expressing tumoral cells may also be tumoral stem cells, which often express a BCRP‐dependent multidrug resistance phenotype (28), and which appear to be involved in the etiology of ependymomas (30).

In line with a previous study reporting MRP1 expression in various brain tumors, especially in medulloblastomas and a few cases of child ependymomas (40), we also detected mRNA coding for this transporter in all the samples in our ependymoma series, and expression of the protein was confirmed by Western blot analysis. Immunohistochemical studies of several tumor samples showed that MRP1 was localized in tumoral cells, rather than in the tumor vessels. MRP1 mediates the efflux, in some cases by co‐transport with reduced glutathione, of several anticancer drugs, including etoposide, anthracyclines, vincristine and methotrexate 23, 34. Our study does not exclude the possibility that other transport proteins, such as MRP4 and 5, or changes in enzymes such as topoisomerase II, methylguanine methyltransferase or dihydrofolate reductase, are involved in the multidrug resistance phenotype displayed by ependymomas, as suggested for other tumoral tissues 5, 6.

Chemotherapy was initiated for 8 of the 13 patients with intracranial ependymoma, indicating that this second‐line treatment is often considered following surgery, especially in infants in order to avoid, or at least delay, use of radiation that impairs brain maturation. Current therapeutic protocols include drugs that are multidrug resistance efflux protein substrates, such as vincristine and etoposide. Understanding multidrug resistance in ependymomas remains therefore a prerequisite for the development of efficient therapies. Our finding that both Pgp and BCRP are located at the luminal side of tumoral vessels indicates that the blood–tumor barrier constitutes a first line of resistance to such treatment, the efficacy of which may be further impaired by other multidrug resistance proteins, possibly involving MRP1, in the tumoral cells themselves.

Supporting information

Table S1. Sequences of primers used for real‐time RT‐PCR and the predicted size of the PCR product.

Please note: Wiley‐Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Supporting info item

Click here for additional data file. (27.5KB, doc)

ACKNOWLEDGEMENTS

We are grateful to Neurobiotec Banques and to the neurosurgeons at the Groupement Hospitalier Est, Bron, who supplied the tumor samples. We thank Sylvie Cavagna and Chantal Watrin for their excellent technical help and Tom Barkas for linguistic help.

MFM and JFGE contributed equally to the coordination of the work.

REFERENCES

  • 1. Aronica E, Gorter JA, Redeker S, Van Vliet EA, Ramkema M, Scheffer GL et al (2005) Localization of breast cancer resistance protein (BCRP) in microvessel endothelium of human control and epileptic brain. Epilepsia 46:849–857. [DOI] [PubMed] [Google Scholar]
  • 2. Bihorel S, Camenisch G, Lemaire M, Scherrmann JM (2007) Influence of breast cancer resistance protein (Abcg2) and p‐glycoprotein (Abcb1a) on the transport of imatinib mesylate (Gleevec) across the mouse blood‐brain barrier. J Neurochem 102:1749–1757. [DOI] [PubMed] [Google Scholar]
  • 3. Billson AL, Palmer JB, Walker DA, Lowe J (1994) Multidrug resistance gene (MDR 1) expression in neuro‐axial tumours of children and young adults. Br J Neurosurg 8:585–591. [DOI] [PubMed] [Google Scholar]
  • 4. Bouffet E, Foreman N (1999) Chemotherapy for intracranial ependymomas. Childs Nerv Syst 15:563–570. [DOI] [PubMed] [Google Scholar]
  • 5. Bredel M (2001) Anticancer drug resistance in primary human brain tumors. Brain Res Brain Res Rev 35:161–204. [DOI] [PubMed] [Google Scholar]
  • 6. Bronger H, Konig J, Kopplow K, Steiner HH, Ahmadi R, Herold‐Mende C et al (2005) ABCC drug efflux pumps and organic anion uptake transporters in human gliomas and the blood‐tumor barrier. Cancer Res 65:11419–11428. [DOI] [PubMed] [Google Scholar]
  • 7. Chomczynski P, Sacchi N (1987) Single‐step method of RNA isolation by acid guanidinium thiocyanate‐phenol‐chloroform extraction. Anal Biochem 162:156–159. [DOI] [PubMed] [Google Scholar]
  • 8. Chou PM, Barquin N, Gonzalez‐Crussi F, Ridaura Sanz C, Tomita T, Reyes‐Mugica M (1996) Ependymomas in children express the multidrug resistance gene: immunohistochemical and molecular biologic study. Pediatr Pathol Lab Med 16:551–561. [PubMed] [Google Scholar]
  • 9. Cooray HC, Blackmore CG, Maskell L, Barrand MA (2002) Localisation of breast cancer resistance protein in microvessel endothelium of human brain. Neuroreport 13:2059–2063. [DOI] [PubMed] [Google Scholar]
  • 10. Cordon‐Cardo C, O'Brien JP, Casals D, Rittman‐Grauer L, Biedler JL, Melamed MR, Bertino JR (1989) Multidrug‐resistance gene (P‐glycoprotein) is expressed by endothelial cells at blood‐brain barrier sites. Proc Natl Acad Sci U S A 86:695–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Daood M, Tsai C, Ahdab‐Barmada M, Watchko JF (2008) ABC transporter (P‐gp/ABCB1, MRP1/ABCC1, BCRP/ABCG2) expression in the developing human CNS. Neuropediatrics 39:211–218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Dauchy S, Dutheil F, Weaver RJ, Chassoux F, Daumas‐Duport C, Couraud PO et al (2008) ABC transporters, cytochromes P450 and their main transcription factors: expression at the human blood‐brain barrier. J Neurochem 107:1518–1528. [DOI] [PubMed] [Google Scholar]
  • 13. Fattori S, Becherini F, Cianfriglia M, Parenti G, Romanini A, Castagna M (2007) Human brain tumors: multidrug‐resistance P‐glycoprotein expression in tumor cells and intratumoral capillary endothelial cells. Virchows Arch 451:81–87. [DOI] [PubMed] [Google Scholar]
  • 14. Fernetti C, Pascolo L, Podda E, Gennaro R, Stebel M, Tiribelli C (2001) Preparation of an antibody recognizing both human and rodent MRP1. Biochem Biophys Res Commun 288:1064–1068. [DOI] [PubMed] [Google Scholar]
  • 15. Fevre‐Montange M, Champier J, Szathmari A, Wierinckx A, Mottolese C, Guyotat J et al (2006) Microarray analysis reveals differential gene expression patterns in tumors of the pineal region. J Neuropathol Exp Neurol 65:675–684. [DOI] [PubMed] [Google Scholar]
  • 16. Figarella‐Branger D, Civatte M, Bouvier‐Labit C, Gouvernet J, Gambarelli D, Gentet JC et al (2000) Prognostic factors in intracranial ependymomas in children. J Neurosurg 93:605–613. [DOI] [PubMed] [Google Scholar]
  • 17. Gazzin S, Strazielle N, Schmitt C, Fevre‐Montange M, Ostrow JD, Tiribelli C, Ghersi‐Egea JF (2008) Differential expression of the multidrug resistance‐related proteins ABCb1 and ABCc1 between blood‐brain interfaces. J Comp Neurol 510:497–507. [DOI] [PubMed] [Google Scholar]
  • 18. Geddes JF, Vowles GH, Ashmore SM, Cockburn HA, Darling JL (1994) Detection of multidrug resistance gene product (P‐glycoprotein) expression in ependymomas. Neuropathol Appl Neurobiol 20:118–121. [DOI] [PubMed] [Google Scholar]
  • 19. Ghersi‐Egea JF, Gazzin S, Strazielle N (2009) Blood‐Brain Interfaces and Bilirubin‐induced neurological diseases. Cur Pharm Des 15:2893–2907. [DOI] [PubMed] [Google Scholar]
  • 20. Grill J, Kalifa C, Doz F, Schoepfer C, Sainte‐Rose C, Couanet D et al (1996) A high‐dose busulfan‐thiotepa combination followed by autologous bone marrow transplantation in childhood recurrent ependymoma. A phase‐II study. Pediatr Neurosurg 25:7–12. [DOI] [PubMed] [Google Scholar]
  • 21. Henson JW, Cordon‐Cardo C, Posner JB (1992) P‐glycoprotein expression in brain tumors. J Neurooncol 14:37–43. [DOI] [PubMed] [Google Scholar]
  • 22. Korshunov A, Sycheva R, Timirgaz V, Golanov A (1999) Prognostic value of immunoexpression of the chemoresistance‐related proteins in ependymomas: an analysis of 76 cases. J Neurooncol 45:219–227. [DOI] [PubMed] [Google Scholar]
  • 23. Leslie EM, Deeley RG, Cole SP (2005) Multidrug resistance proteins: role of P‐glycoprotein, MRP1, MRP2, and BCRP (ABCG2) in tissue defense. Toxicol Appl Pharmacol 204:216–237. [DOI] [PubMed] [Google Scholar]
  • 24. Maksoud YA, Hahn YS, Engelhard HH (2002) Intracranial ependymoma. Neurosurg Focus 13:e4. [DOI] [PubMed] [Google Scholar]
  • 25. McLendon RE, Wiestler OT, Kros JM, Korshunov A, Ng HK (2007) Ependymoma. In: WHO Classification of Tumours of the Central Nervous System, Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds), pp. 74–80. IARC: Lyon. [Google Scholar]
  • 26. Merchant TE, Fouladi M (2005) Ependymoma: new therapeutic approaches including radiation and chemotherapy. J Neurooncol 75:287–299. [DOI] [PubMed] [Google Scholar]
  • 27. Nejat F, El Khashab M, Rutka JT (2008) Initial management of childhood brain tumors: neurosurgical considerations. J Child Neurol 23:1136–1148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Nicolis SK (2007) Cancer stem cells and “stemness” genes in neuro‐oncology. Neurobiol Dis 25:217–229. [DOI] [PubMed] [Google Scholar]
  • 29. Peterson GL (1977) A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal Biochem 83:346–356. [DOI] [PubMed] [Google Scholar]
  • 30. Poppleton H, Gilbertson RJ (2007) Stem cells of ependymoma. Br J Cancer 96:6–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Robey RW, Polgar O, Deeken J, To KW, Bates SE (2007) ABCG2: determining its relevance in clinical drug resistance. Cancer Metastasis Rev 26:39–57. [DOI] [PubMed] [Google Scholar]
  • 32. Sangra M, Thorp N, May P, Pizer B, Mallucci C (2009) Management strategies for recurrent ependymoma in the paediatric population. Childs Nerv Syst 25:1283–1291. [DOI] [PubMed] [Google Scholar]
  • 33. Scheffer GL, Kool M, Heijn M, De Haas M, Pijnenborg AC, Wijnholds J et al (2000) Specific detection of multidrug resistance proteins MRP1, MRP2, MRP3, MRP5, and MDR3 P‐glycoprotein with a panel of monoclonal antibodies. Cancer Res 60:5269–5277. [PubMed] [Google Scholar]
  • 34. Schinkel AH, Jonker JW (2003) Mammalian drug efflux transporters of the ATP binding cassette (ABC) family: an overview. Adv Drug Deliv Rev 55:3–29. [DOI] [PubMed] [Google Scholar]
  • 35. Schinkel AH, Wagenaar E, Mol CA, Van Deemter L (1996) P‐glycoprotein in the blood‐brain barrier of mice influences the brain penetration and pharmacological activity of many drugs. J Clin Invest 97:2517–2524. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Shim KW, Kim DS, Choi JU (2009) The history of ependymoma management. Childs Nerv Syst 21:21. [DOI] [PubMed] [Google Scholar]
  • 37. Soontornmalai A, Vlaming ML, Fritschy JM (2006) Differential, strain‐specific cellular and subcellular distribution of multidrug transporters in murine choroid plexus and blood‐brain barrier. Neuroscience 138:159–169. [DOI] [PubMed] [Google Scholar]
  • 38. Thiebaut F, Tsuruo T, Hamada H, Gottesman MM, Pastan I, Willingham MC (1989) Immunohistochemical localization in normal tissues of different epitopes in the multidrug transport protein P170: evidence for localization in brain capillaries and crossreactivity of one antibody with a muscle protein. J Histochem Cytochem 37:159–164. [DOI] [PubMed] [Google Scholar]
  • 39. Toth K, Vaughan MM, Peress NS, Slocum HK, Rustum YM (1996) MDR1 P‐glycoprotein is expressed by endothelial cells of newly formed capillaries in human gliomas but is not expressed in the neovasculature of other primary tumors. Am J Pathol 149:853–858. [PMC free article] [PubMed] [Google Scholar]
  • 40. Valera ET, Lucio‐Eterovic AK, Neder L, Scrideli CA, Machado HR, Carlotti‐Junior CG et al (2007) Quantitative PCR analysis of the expression profile of genes related to multiple drug resistance in tumors of the central nervous system. J Neurooncol 85:1–10. [DOI] [PubMed] [Google Scholar]
  • 41. Van Der Heijden JW, Oerlemans R, Tak PP, Assaraf YG, Kraan MC, Scheffer GL et al (2009) Involvement of breast cancer resistance protein expression on rheumatoid arthritis synovial tissue macrophages in resistance to methotrexate and leflunomide. Arthritis Rheum 60:669–677. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1. Sequences of primers used for real‐time RT‐PCR and the predicted size of the PCR product.

Please note: Wiley‐Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Supporting info item

Click here for additional data file. (27.5KB, doc)

Articles from Brain Pathology are provided here courtesy of Wiley

RESOURCES