Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The cytochrome P450 enzymes (P450²) comprise a large family of constitutive and inducible hemoproteins that are involved in the metabolism of endogenous and foreign compounds. The P450 levels are generally low in the brain (Hedlund et al., 2001). Some P450 genes, for instance genes in the CYP1A/B subfamilies, may however be transcriptionally activated in the brain by inducers that act via a receptor-dependent mechanism. Pretreatment with aryl hydrocarbon receptor (AhR)-ligands will for instance increase the CYP1A/B-dependent 7-ethoxyresorufin O-deethylase (EROD) activity in rat brain (Dhawan et al., 1999). Furthermore, the AhR-ligand -naphthoflavone (BNF) has been reported to induce CYP1A1 mRNA in rat brain (Schilter and Omiecinski, 1993; Savas et al., 1994).

The P450-dependent metabolism of drugs and chemicals will generally lead to detoxified metabolites, but in some cases electrophilic, reactive intermediates may be formed. For instance, the AhR-regulated P450 forms are involved in the activation of polycyclic aromatic hydrocarbons (PAHs) to reactive intermediates, which bind irreversibly to protein and DNA. PAHs are potent carcinogens, but these compounds may also initiate atherosclerosis in mice and birds (Revis et al., 1984; Paigen et al., 1986; Zhang and Ramos, 1997). Humans are exposed to PAHs mainly by inhalation (urban air and tobacco smoke) and via the diet (smoked/grilled food and contaminated crops). Both CYP1A1 and 1B1 are known to metabolically activate PAHs to reactive intermediates (Shimada et al., 1996). PAH-DNA adducts have previously been detected in brain homogenates of 7,12-dimethylbenz(a)anthracene (DMBA)-treated rats (Doerjer et al., 1978) and benzo(a)pyrene (B(a)P)-treated rats, mice, rabbits, and monkeys (Das et al., 1985; Stowers and Anderson, 1985; Lu et al., 1993). The cellular localization of protein/DNA adducts in the brain following exposure to PAHs has however not been reported.

When the metabolism of drugs and chemicals is studied in whole brain or in selected brain regions, a cell-specific metabolism will often be masked or diluted by the presence of metabolically inactive cells. The aim of the present study has been to compare the cellular expression of CYP1A1/1B1 and the localization of [³H]DMBA adducts in the brain of mice and rats pretreated with AhR agonists 3,3',4,4',5-pentachlorobiphenyl (PCB 126), BNF, or vehicle. The results demonstrated a differential expression of CYP1A1 and CYP1B1 in blood-brain interfaces; CYP1A1 was preferentially detected in endothelial cells (EC) whereas CYP1B1 was preferentially detected in vascular smooth muscle cells (SMC). Furthermore, a selective localization of [³H]DMBA adducts in EC of some blood-brain interfaces was revealed in mice and rats exposed to AhR-agonists. This localization was in accordance with the expression of CYP1A1 in mice exposed to AhR-agonists. In addition, an increased EROD activity was found in excised choroid plexus, leptomeninges, and dura mater of rats pretreated with PCB 126 as compared with rats pretreated with vehicle.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. [³H]DMBA (22.0, 52.0, 57.0, or 61.0 Ci/mmol, radiochemical purity 93.0-99.5%) was obtained from Amersham Pharmacia Biotech Ltd. (Buckinghamshire, UK). PCB 126 was kindly prepared by Dr. Åke Bergman, Department of Environmental Chemistry, Wallenberg Laboratory, Stockholm University, Sweden. BNF, ellipticine, NADPH, ethoxyresorufin, resorufin, fluorescamine, dimethyl sulfoxide (DMSO), and corn oil were obtained from Sigma Chemical Co. (St. Louis, MO). Fully supplemented Dulbeccos Modified Eagle's Medium (FDMEM), fetal bovine serum, L-glutamine, gentamycin sulfate, and 2-mercaptoethanol were obtained from the Swedish Veterinary Institute (Uppsala, Sweden). Rabbit anti-rat CYP1A1 polyclonal antibody was obtained from Chemicon International (Temecula, CA). According to the manufacturer, this antibody does not cross-react with other P450s. Rabbit anti-human CYP1B1 antibody was obtained from BD Gentest Corporation (Woburn, MA). Western blot confirmed that the CYP1B1 antibody only reacted with one protein in liver and brain microsomes of BNF-treated mice. Biotinylated swine anti-rabbit immunoglobulins and the avidin-biotin peroxidase complex/horseradish peroxidase were obtained from DAKO A/S (Glostrup, Denmark). The Vector VIP peroxidase substrate kit was obtained from Vector Laboratories (Burlingame, CA). Bovine serum albumin (BSA) was obtained from Chemicon. All other chemicals were of analytical reagent grade, obtained from commercial sources.

Animals. Female NMRI mice (6-11 weeks old; 20-35 g) and female Sprague-Dawley rats (4-6 weeks or 5 months old; 65, 150, or 300 g) were obtained from Charles River (Uppsala, Sweden) and B&K Universal (Sollentuna, Sweden). The animals were housed in a climate-controlled room (20-23°C, 50% humidity) with a 12-h light/dark cycle. They were given tap water and a pellet diet (R36 from Lactamin, Vadstena, Sweden) ad libitum.

Pretreatment of Animals. Mice and rats were pretreated with i.p. injections of BNF (80 mg/kg), PCB 126 (250 µg/kg), or corn oil (10 ml/kg; vehicle). BNF was given 48 and 24 h before the experiment, and PCB 126 or vehicle were given 5 days before the experiment. The rats were killed by exposure to gaseous CO₂ and the mice by cervical dislocation. The studies were approved by the Local Ethics Committee for Research on Animals.

Administration of [³H]DMBA to Rodents. Mice and rats pretreated with PCB 126, BNF, or vehicle were given an i.v. injection of [³H]DMBA (0.5 µmol/kg) dissolved in DMSO (1 ml/kg). The mice were killed 24 h later and subjected to tape-section autoradiography according to Ullberg (1977). Sagittal sections (20 µm) through the skull were collected onto tape (3M, Stockholm, Sweden) in a cryostat microtome. After freeze-drying the tissue sections were apposed to X-ray film (Hyperfilm-³H; Amersham Pharmacia Biotech Ltd.) and exposed at 20°C for 7 or 14 weeks.

Incubation of Precision-cut Brain Slices with [³H]DMBA. Fresh pieces of brain from mice and rats pretreated with PCB 126, BNF, or vehicle were put in ice-cold phosphate-buffered saline (PBS). The pieces were embedded in 3% agarose (40°C), immediately cooled on ice, and sectioned (in cold PBS) in a tissue slicer (MD1100-A2, Alabama Research and Development, Manford, AL). Coronal sections from the middle of the cerebral hemispheres were used. The precision-cut slices (300-µm thick) were placed on titanium inserts and incubated in FDMEM medium (2.5 ml) supplemented with [³H]DMBA (0.5 µM, 64 µCi) dissolved in DMSO (final concentration 0.1%), fetal bovine serum (2%), gentamycin sulfate (0.2%), L-glutamine (2 mM), and 2-mercaptoethanol (50 µM) for 3 h at 37°C (5% CO₂, 95% O₂). After incubation, the brain slices were fixed in formaldehyde and processed for light microscopy autoradiography as described above. The procedures to prepare and incubate precision-cut slices for secondary transversal sectioning have recently been described by Lindhe et al. (2001).

Incubation of Excised Blood-brain Interface Tissues with [³H]DMBA. Dura mater membranes were removed from the parietal and frontal part of the skull, and choroid plexus was excised from the lateral ventricles of mice and rats pretreated with PCB 126 or vehicle. Leptomeninges (composed of pia mater and arachnoid membrane, the innermost pair of membranes surrounding the brain) were prepared from rats by gently scraping the membranes off the brain. Following rinsing in saline (0.9%), the tissues were incubated separately for 90 min (37°C) in FDMEM medium (2.0 ml) supplemented with [³H]DMBA (0.5 µM, 67 µCi) dissolved in DMSO (final concentration 0.1%). To determine the involvement of CYP1A, the inhibitor ellipticine (10 µM) was added to some of the incubation vials. After incubation, the tissue pieces were fixed and processed for light microscopy autoradiography as described above.

Incubation of Excised Blood-brain Interface Tissues with 7-Ethoxyresorufin. We have developed a method to determine EROD in tissues containing low activities. Generally, pooled homogenates from several animals are needed to achieve this. Using the method described below, it is possible to achieve individual data from each animal. Measuring EROD in a multiwell plate reader is a simple procedure, making it possible to quickly determine many samples. EROD and protein content are measured simultaneously, further decreasing the demand for tissue samples. We adapted this method to measure EROD activity and protein content in small intact tissue preparations of choroid plexus, leptomeninges, and dura mater taken from individual rats pretreated as described above. Tissue-pieces from the left and right cerebral hemisphere were assayed as duplicates. The tissue-pieces were placed in a 96-well plate in ice-cold assay medium (130 µl) consisting of NADPH (0.5 mM), MgCl₂ (4.0 mM) and sodium phosphate buffer (0.07 M). The plate was preheated (37°C) in a multiwell plate incubator for 5 min before the reaction was started by addition of 7-ethoxyresorufin solution (20 µl; 10 µM 7-ethoxyresorufin, 1% DMSO, 1% methanol, 0.1 M sodium phosphate buffer). Following incubation for 90 min at 37°C during continuous agitation, the reaction was stopped by addition of fluorescamine in acetonitrile (150 µl; 0.3 mg/ml). During this time (90 min), the conversion of 7-ethoxyresorufin to the fluorescent product resorufin was linear. Fluorescamine reacts with primary amines in proteins forming a fluorescent product. Resorufin and BSA in sodium phosphate buffer (0.1 M, pH 7.8) were used as a combined standard in the EROD/protein assay. Concentrations of resorufin (nM)/BSA (µg/ml) were 0/0, 12.5/200, 25/400, 50/600, and 100/800. Standards (130 µl) were incubated in parallel with the samples. After 90 min, fluorescamine in acetonitrile (150 µl) and 7-ethoxyresorufin solution (20 µl) was added to give similar background fluorescence as for the samples. The plate was placed in an ultrasound bath for 10 s and stored in the dark for 15 min. The fluorescence intensity was then measured using a fluorescence multiwell plate reader (FLUOstar P; SLT Labinstruments GmbH, Grödig, Austria). The excitation/emission wavelengths used were 544 nm/590 nm and 390 nm/460 nm for the EROD and protein determinations, respectively.

Immunohistochemistry of CYP1A1 and CYP1B1. Brain tissue from mice pretreated with PCB 126, BNF, or vehicle were excised and fixed in formaldehyde (4%) in phosphate buffer (pH 7.4) for 5 days. Before embedding in paraffin, the tissues were dehydrated as follows: 70% ethanol (over night), 95% ethanol (4 h), 99.5% ethanol (4 h), xylene (30-40 min). Paraffin-embedded tissues were sectioned (4.5 µm), mounted in distilled water on gelatin-coated slides, and dried over night at 33 to 35°C. Following deparaffinization and rehydration, the sections were washed in PBS (pH 7.4) and PBS with 0.3% Triton X-100 (PBS-T). The sections were treated with 1% hydrogen peroxide in PBS-T to block endogenous peroxidase activity and washed in 3× PBS-T and 3× PBS (10 min), followed by 1 h preincubation with 4% BSA to block nonspecific binding of the antibodies. The brain sections were then incubated with rabbit anti-rat CYP1A1 or rabbit anti-human CYP1B1 antibodies diluted in PBS containing 4% BSA (working dilution 1:300-1:500) overnight at 4°C. The sections were then washed in 3× PBS-T and 3× PBS (10 min) and further incubated with biotinylated second antibody (working dilution 1:300) and avidin-biotin peroxidase complex at room temperature (30 min, respectively). Following washings as above, the sections were incubated with Vector VIP substrate kit for peroxidase to develop color. Finally, the sections were rinsed in tap water for 5 min and mounted in Pertex mounting medium. All samples were stained at least on two separate occasions. Negative control sections were treated in the same way as described above, except that the antibodies against CYP1A1 or CYP1B1 were omitted. The control incubations were conducted simultaneously with the experiment. The negative controls did not show any staining. In addition, no staining was observed in erythrocytes. The sections were examined and photographed in a Leitz DM RXE microscope (Leica) using Nomarski differential interference technique.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Localization of [³H]DMBA Adducts in the Brain in Vivo. Tape-section autoradiography demonstrated that there was a selective localization of radioactivity in the leptomeninges in the brain of PCB 126- and BNF-pretreated mice (Fig. 1) and rats injected i.v. with [³H]DMBA and killed 24 h later. No selective labeling was observed in other brain regions. Negligible, or only weak localization of radioactivity was observed in the leptomeninges of vehicle-treated controls injected with [³H]DMBA.

View larger version (82K): [in this window] [in a new window]   Fig. 1.   Tape-section autoradiogram showing sites of radioactivity in the skull 24 h after an i.v. injection of [3H]DMBA (0.5 µmol/kg) in a BNF- (A) or vehicle- (B) pretreated mouse. There is a selective localization of radioactivity in the leptomeninges (arrows). Black areas correspond to high levels of radioactivity. Autoradiography of freeze-dried tissue sections.

Cellular Localization of [³H]DMBA Adducts in the Brain in Vivo. Light microscopy autoradiography demonstrated that there was high and selective binding of radioactivity in the capillary loop EC of the choroid plexus (Fig. 2, A and B), in EC of cerebral veins and veins in the leptomeninges (Fig. 2, C and D) of BNF-pretreated mice injected i.v. with [³H]DMBA and killed 24 h later. Negligible or only a weak binding was observed in EC of arteries. No binding of radioactivity was observed in EC of cerebral capillaries or in epithelial cells of the choroid plexus, neurones or neuroglial cells in the BNF-pretreated mice. In vehicle-pretreated mice injected with [³H]DMBA, no selective binding of radioactivity in EC of the leptomeninges, choroid plexus, or in other brain regions could be observed (not shown).

View larger version (102K): [in this window] [in a new window]   Fig. 2.   Light microscopy autoradiograms showing the localization of [3H]DMBA adducts in the brain 24 h following an i.v. injection of [3H]DMBA (0.5 µmol/kg) in a BNF-pretreated mouse. There is a selective localization of [3H[DMBA adducts (silver grains) in endothelial cells of the choroid plexus (A, B) and in veins of the leptomeninges (C, D). Left column represents bright-field images (black silver grains), and right column represents dark-field images (white silver grains). Autoradiography of solvent-extracted tissue sections.

Cellular Localization of [³H]DMBA Adducts in Precision-cut Brain Slices in Vitro. Light microscopy autoradiography revealed a high level of bound radioactivity in EC of the choroid plexus capillaries (Fig. 3A), in EC of arteries and veins in the leptomeninges (Fig. 3B) and in EC of certain cerebral veins (Fig. 3C) in [³H]DMBA-incubated fresh brain slices from PCB 126- and BNF-pretreated mice and rats. There was also a high level of binding of [³H]DMBA in smaller arteries such as vasa vasorum EC close to arteries, whereas larger arteries were devoid of [³H]DMBA binding (Fig. 3D). A weak binding was observed in EC of the brain capillaries and epithelial cells of the choroid plexus, whereas no binding was observed in neurones or in neuroglial cells. In [³H]DMBA-incubated brain slices from vehicle-pretreated animals no sites of binding of radioactivity could be observed (not shown).

View larger version (98K): [in this window] [in a new window]   Fig. 3.   Light microscopy autoradiograms showing the localization of [3H]DMBA adducts in [3H]DMBA-incubated (0.5 µM; 180 min) fresh precision-cut brain slices (300 µm) from a PCB 126-treated rat (A, B) and from a BNF-treated mouse (C, D). There is a selective localization of [3H]DMBA adducts (black silver grains) in endothelial cells of the choroid plexus (A), in veins of the leptomeninges (B), in a cerebral vein (C), and in vasa vasorum close to a large artery devoid of adducts (D). White arrows indicate endothelial cells, black arrow epithelial cells, and gray arrow smooth muscle cells. Autoradiography (bright-field images) of solvent-extracted tissue sections.

Cellular Localization of [³H]DMBA Adducts in Excised Blood-brain Interfaces in Vitro. Light microscopy autoradiography demonstrated that there was a high level of bound radioactivity in EC in dura mater membrane (Fig. 4A) and choroid plexus (Fig. 4F) of [³H]DMBA-incubated preparations from PCB 126-pretreated mice and rats. A high level of binding was also found in EC in the leptomeninges of PCB 126-pretreated rats (Fig. 4D). The unspecific background labeling in these preparations was fairly high, but no other cellular targets than the EC could be found. It was not possible to distinguish between various types of blood vessels. Addition of the CYP1A1-inhibitor ellipticine to the incubation medium abolished the localization of bound radioactivity in EC (Fig. 4, B, E, and G). In [³H]DMBA-incubated tissue preparations from vehicle-treated control animals, the level of binding in EC was negligible or much lower than that of the PCB 126-pretreated animals (Fig. 4, C and H).

View larger version (178K): [in this window] [in a new window]   Fig. 4.   Light microscopy autoradiograms showing the localization of [3H]DMBA adducts in [3H]DMBA-incubated (0.5 µM; 90 min) explants of dura mater (top row), leptomeninges (middle row) and choroid plexus (bottom row) from a PCB 126-treated mouse (A, B), a vehicle-treated mouse (C), a PBC 126-treated rat (D, E, F, G) and a vehicle-treated rat (H). There is a selective localization of [3H]DMBA binding (black silver grains) in endothelial cells of all blood-brain interfaces from PCB 126-treated animals (left column). Addition of the CYP1A-inhibitor ellipticine (10 µM) to the incubation medium abolished the [3H]DMBA binding (B, E, G), and autoradiograms are similar to those obtained from explants of blood-brain interfaces from vehicle-treated animals (C, H). Autoradiography (bright-field images) of solvent-extracted tissue sections.

EROD Activity in Excised Blood-brain Interfaces. The EROD activity in excised choroid plexus and dura mater membrane was 30 to 45 times higher in the PCB 126-pretreated rats compared with one of the five control rats (Table 1). In the remaining four control rats, no EROD activity could be detected. EROD activity was also found in excised leptomeninges of PCB 126-pretreated rats, whereas no EROD activity could be detected in excised leptomeninges of any of the control rats. The amount of protein per well was 6 to 8, 6 to 8, and 4 to 10 µg for the choroid plexus, leptomeninges, and dura mater membrane, respectively.

Cellular Localization of CYP1A1 in Blood-brain Interfaces. There was a selective staining of EC in cerebral veins (Fig. 5A; Table 2), veins in the leptomeninges (Fig. 5B, Table 2) and in capillary loops of the choroid plexus (Fig. 5E) of PCB 126- and BNF-treated mice. The staining preferentially occurred in the cytoplasm. There was also a positive staining of vasa vasorum EC close to large blood vessels (Fig. 5G). Only a weak, often negligible, staining could be observed in EC of cerebral arteries and of arteries in the leptomeninges (Fig. 5C). No staining was detected in EC of brain capillaries of PCB 126- or BNF-treated mice. Furthermore, the SMC layers in arteries and veins, epithelial cells of the choroid plexus, ependymal cells lining the ventricles or neurones showed no staining in PCB 126- or BNF-treated mice. In vehicle-treated mice no staining of EC in the leptomeninges, cerebral veins, capillary loops of the choroid plexus, or vasa vasorum could be observed (Fig. 5D, F, H). Nor was there any staining in the other regions.

View larger version (113K): [in this window] [in a new window]   Fig. 5.   Immunohistochemical localization of CYP1A1 in the brain of a PCB 126-treated mouse (A, B, C, E, G) and a vehicle-treated mouse (D, F, H). CYP1A1 is expressed (red staining) in endothelial cells in a cerebral vein (A), in veins (B) and arteries (C) of the leptomeninges, in capillary loops of the choroid plexus (E), and in vasa vasorum (G) close to an artery in the PCB 126-treated mouse. In the vehicle-treated mouse no staining is seen in veins and arteries of the leptomeninges (D), in the choroid plexus (F), or in vasa vasorum (H). Bright field images with Nomarski contrast.

Cellular Localization of CYP1B1 in Blood-brain Interfaces. There was an intense staining of the SMC layer in arteries of the leptomeninges (Fig. 6A-B; Table 2), cerebral arteries, arterioles, and in choroid plexus arterioles of PCB 126- and BNF-treated mice as well as of vehicle-treated mice. The staining occurred preferentially in the cytoplasm. No CYP1B1 staining of cerebral capillaries was seen whereas a weak or negligible staining was observed in SMC of cerebral veins and veins in leptomeninges. In EC of blood vessels no or almost negligible staining occasionally was noted (Fig. 6, A-C). No increased staining of EC could be detected in the PCB 126- and BNF-treated mice. Furthermore, there was no staining of EC in capillary loops of the choroid plexus of PCB 126- and BNF-treated mice (Fig. 6D) nor in those of vehicle-treated mice. A moderate staining of ependymal cells lining the ventricles was observed. A weak staining of neurones was also found, but no attempts to evaluate the staining in various brain regions were done. In epithelial cells of the choroid plexus no staining was detected.

View larger version (97K): [in this window] [in a new window]   Fig. 6.   Immunohistochemical localization of CYP1B1 in the brain of a vehicle-treated mouse (A) and a PCB 126-treated mouse (B, C, D). CYP1B1 is expressed (red staining) in the smooth muscle cell layer of arteries in the leptomeninges (A, B). No staining is seen in veins of the leptomeninges (C) nor in the choroid plexus (D). Bright field images with Nomarski contrast.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study revealed a differential expression of CYP1A1 and CYP1B1 in blood-brain interfaces and a metabolic activation of a model PAH at these sites in P450-induced rodents. According to immunohistochemistry, CYP1A1 was inducible in EC of veins in leptomeninges, cerebral veins and vasa vasorum, and in capillary loops of choroid plexus. Notably, the expression was weak or negligible in EC of arteries at these sites. The immunoreactivity preferentially occurred in the cytosol of EC. These results are consistent with a report showing that CYP1A1 mRNA is expressed in EC of the choroid plexus in AhR agonist-treated mice (Dey et al., 1999). Accordingly, bound [³H]DMBA was preferentially found in EC of veins in leptomeninges, cerebral vessels, and capillary loops of choroid plexus. The binding in EC of blood-brain interfaces was observed both after injection, and after incubation of precision-cut brain slices or excised blood-brain interfaces with [³H]DMBA, suggesting an in situ activation of DMBA in the EC. Since the thin tissue sections were repeatedly extracted with organic solvents, we propose that the nonextractable radioactivity in the cytoplasm represented the formation of [³H]DMBA-protein adducts, whereas the radioactivity in the nuclei represented DNA adducts. The irreversible binding of [³H]DMBA in EC of blood-brain interfaces incubated in vitro was efficiently blocked by ellipticine. Ellipticine is known to be a selective inhibitor of CYP1A1 (Kerzee and Ramos, 2000).

We have previously reported that the food and tobacco carcinogen 3-amino-1,4-dimethyl-5H-pyrido-[4,3-b]indole (Trp-P-1) is metabolically activated and irreversibly bound in EC in choroid plexus and large cerebral veins of BNF-treated mice, whereas no irreversible binding occurs in EC of vehicle-treated control mice (Brittebo, 1994). Also Trp-P-1 is a substrate for CYP1A1 and CYP1B1 (Shimada et al., 1996). The preferential binding of [³H]DMBA and [³H]Trp-P-1 in EC of some blood-brain interfaces in AhR agonist-treated rodents suggests a CYP1A1-dependent activation of these compounds to reactive intermediates. EROD is generally used for quantification of CYP1A-dependent metabolism since the CYP1B1-dependent EROD activity is low (Murray et al., 2001). We found a marked induction of EROD activity in small intact tissue preparations of leptomeninges and choroid plexus from PCB 126-pretreated rats. This observation is consistent with a report of Morse and coworkers showing induction of CYP1A1 and EROD in homogenates of meninges and choroid plexus from BNF-treated rats (Morse et al., 1998). No expression of CYP1A1 in EC of noninduced animals has been noted in previous immunohistochemical studies (Kapitulnik et al., 1987; Warner et al., 1988; Riedl et al., 1996). Based on these observations, we conclude that EC in veins in the leptomeninges and capillaries in the choroid plexus of rodents express a constitutively low but highly inducible and catalytically active CYP1A1 enzyme.

CYP1B1 protein has not previously been reported in mouse brain tissues although it is known to be present in human brain (Savas et al., 1994). Our immunohistochemical investigation revealed a distinct expression of CYP1B1 in the SMC of arteries in the leptomeninges, cerebral arteries, arterioles and choroid plexus arterioles whereas the expression of CYP1B1 in EC was negligible. An intense staining of the cytosol in SMC was noted. In contrast to the marked induction of CYP1A1 in EC of blood-brain interfaces, no increased levels of CYP1B1 in EC or SMC in blood-brain interfaces could be detected in mice treated with AhR agonists. Zhao and coworkers have reported an induction of CYP1B1 in cultured cells from human vascular wall (Zhao et al., 1998) whereas CYP1B1 was not detected in human EC (Kerzee and Ramos, 2001). Also in human cerebral microvasculature, expression of CYP1B1 has been observed (Rieder et al., 2000).

Although DMBA is known to be metabolized by CYP1B1 (Otto et al., 1992), selective localization of [³H]DMBA adducts was not observed in SMC of arteries in the leptomeninges or cerebral arteries, either in vehicle-treated nor in AhR agonist-treated rodents. Similarly, no irreversible binding of [³H]Trp-P-1 in SMC of blood-brain interfaces has been observed (Brittebo, 1994). There are several possible explanations for the different fate of [³H]DMBA in EC and SMC; CYP1B1 could be catalytically less active than CYP1A1 toward DMBA or its metabolites; the two enzymes could form different metabolites; or other enzymes, differentially expressed in EC and in SMC, could compete for DMBA or its reactive intermediates. Further studies are needed to clarify the lack of DMBA and Trp-P-1 adduct formation in cerebro-vascular SMC.

A weak staining of CYP1B1 was also detected in ependymal cells lining the ventricles whereas there was no immunoreactivity in the choroid plexus epithelium. Constituting one of the blood-cerebrospinal fluid (CSF) barriers, the choroid plexus epithelium acts both by controlling the influx of blood-borne substances into the CSF and by clearing substances from the CSF (Ghersi-Egea and Strazielle, 2001). The present results suggest that neither CYP1B1 nor CYP1A1 is responsible for metabolism of foreign compounds in choroid plexus epithelium. Little attention has been paid to the role of ependymal cells in the clearing of foreign compounds in CSF although a localization of a member of the CYP2 family has been reported in these cells (Volk et al., 1991). No localization of [³H]DMBA adducts at this site was however observed in the present study.

The blood-brain barrier, mainly the cerebral capillaries, serves to protect the central nervous system from blood-borne toxic chemicals. In the present study neither CYP1A1 nor CYP1B1 immunoreactivity was detected in cerebral capillaries of BNF-, PCB 126- or vehicle-treated mice. This negative finding is consistent with studies showing lack of EROD activity and CYP1A1 in homogenates of isolated brain capillaries from control or BNF-treated rats (Morse et al., 1998). Our light microscopic autoradiography revealed no signs of irreversible [³H]DMBA-binding in cerebral capillary EC of mice and rats pretreated with AhR agonists. The lack of [³H]DMBA binding in cerebral capillary EC therefore seems to be because of a lack of CYP1A1 in this cell type. Given the marked expression of CYP1A1 in EC of veins in leptomeninges and capillary loops in choroid plexus, it was surprising that we found no expression of CYP1A1 in cerebral capillary EC. There are, however, differences in endothelial function throughout the vascular tree. The underlying causes for the heterogeneity of EC are not clear. One factor may be the interaction with surrounding cells. Another factor is the expression of the drug transporter P-glycoprotein (P-gp) in brain capillary EC (for review see Schinkel, 1999; Tsuji and Tamai, 1999). P-gp regulates the efflux of PAHs such as DMBA and B(a)P from human intestine and breast cancer cells (Phang et al., 1993; Penny and Campbell, 1994). A P-gp-mediated elimination of the CYP1A1-inducers BNF and PCB 126 from brain capillary EC cannot be ruled out.

Injury to the blood vessel endothelium is often proposed as an early lesion in cardiovascular disease. We have reported earlier that there is an induction of CYP1A1 and irreversible binding of [³H]DMBA or [³H]Trp-P-1 also in human umbilical vein EC (HUVEC) following pretreatment with BNF (Annas et al., 2000b). Interestingly, shear stress has been demonstrated to induce CYP1A1 and CYP1B1 in HUVEC (McCormick et al., 2001). The induction of these enzymes in human EC may be an important determinant for the susceptibility of EC toward protoxicants and procarcinogens. The formation of reactive intermediates and subsequent reaction with vital macromolecules may result in a changed or impaired function leading to detrimental consequences for the cell. We have recently observed B(a)P-induced DNA damage in HUVEC exposed to BNF (Annas et al., 2000a). In addition, PAHs such as DMBA and B(a)P can induce or accelerate development of atherosclerotic plaques in the aorta of pigeons and mice (Revis et al., 1984; Paigen et al., 1986). PAH-induced DNA adducts have been detected in mouse aortic SMC and in arteries of atherosclerotic human subjects (Zhang et al., 1998; Moorthy et al., 2002). It has been suggested that the proliferation of SMC represents a critical event in the development of atherosclerotic lesions (Kerzee and Ramos, 2001). Taken together these studies demonstrate that exposure to PAH may induce adverse effects in both EC and vascular SMC. The relative roles of EC and SMC in PAH-induced vascular disease merits further investigation.

Both tobacco smoking and combustion effluents are important sources for PAH exposure. Epidemiological studies indicate common risk factors to be involved in the etiology of both atherosclerosis and cancer, including tobacco smoking and exposure to combustion effluents from burning of fuel (Ross et al., 2001). Air pollution has recently been associated with ischemic stroke mortality (Hong et al., 2002). Furthermore both active and passive smoking represent risk factors for atherosclerosis and increases the risk of acute stroke (Bonita et al., 1999). Smokers are also known to have higher expression of CYP1. Our present results, demonstrating the presence of [³H]DMBA adducts in EC of some blood-brain interfaces of P450-induced rodents, and previous results, showing PAH-adducts and B(a)P-induced DNA damage in HUVEC, suggest that PAH-exposure should be considered as a potential risk factor with regard to cerebrovascular disease (Annas et al., 2000a,b).

In conclusion, the results of this study showed a cell-specific expression of CYP1A1 and CYP1B1 in EC or vascular SMC of some blood-brain interfaces, whereas no expression occurred in the blood-brain barrier. This implies that these enzymes may play an important role in a cell-specific metabolism of endogenous vasoactive substrates in EC and vascular SMC. The cell-specific expression of CYP1A1 and CYP1B1 may also be part of the blood-CSF barrier regulating the passage of foreign compounds. Studies on the fate of a model PAH, [³H]DMBA, in the brain, revealed that the localization of [³H]DMBA adducts corresponded to the expression of CYP1A1 in EC of AhR agonist-treated rodents. The potential role of environmental pollutants, such as AhR agonists and PAHs, in the etiology of smoking- and air pollution-related cerebrovascular disease merits further consideration.