Abstract
Cytochrome P450 (P450) enzymes and ATP-binding cassette (ABC) transporters modulate the transport and metabolism of both endogenous and exogenous substrates and could play crucial roles in the human brain. In this study, we report the transcript expression profile of seven ABC transporters (ABCB1, ABCC1–C5, and ABCG2), 24 P450s (CYP1, CYP2, and CYP3 families and CYP46A1), and 14 related transcription factors [aryl hydrocarbon receptor, nuclear receptor (NR)1I2/pregnane X receptor, NR1I3/constitutive androstane receptor and NR1C/peroxisome proliferator-activated receptor, NR1H/liver X receptor, NR2B/retinoid X receptor, and NR3A/estrogen receptor subfamilies] in the whole brain, the dura mater, and 17 different encephalic areas. In addition, Western blotting and immunohistochemistry analysis were used to characterize the distribution of the P450s at the cellular and subcellular levels in some brain regions. Our results show the presence of a large variety of xenobiotic transporters and metabolizing enzymes in human brain and show for the first time their apparent selective distribution in different cerebral regions. The most abundant transporters were ABCC5 and ABCG2, which, interestingly, had a higher mRNA expression in the brain compared with that found in the liver. CYP46A1, CYP2J2, CYP2U1, CYP1B1, CYP2E1, and CYP2D6 represented more than 90% of the total P450 and showed selective distribution in different brain regions. Their presence in both microsomal and mitochondrial fractions was shown both in neuronal and glial cells in several brain areas. Thus, our study shows key enzymes of cholesterol and fatty acid metabolism to be present in the human brain and provides novel information of importance for elucidation of enzymes responsible for normal and pathological processes in the human brain.
The cytochrome P450 (P450) enzymes belonging to families 1 through 3 and many ATP-binding cassette (ABC) transporters are primarily known for their role in xenobiotic transport and metabolism. These proteins exert their activity mainly in the liver but have also been found in other tissues such as lung, kidney, intestine, placenta, and brain. The extrahepatic localization of these proteins suggests important and unrecognized activities toward endogenous substrates. We hypothesize that the roles of these proteins in transport and metabolism both of endogenous and endogenous compounds are important in the human brain, but at present there is a significant lack of knowledge regarding the distribution of P450s and ABC transporters throughout the human brain.
The ABC superfamily is a large and ubiquitous group of proteins. There are 48 human ABC proteins that are organized into seven subfamilies (A–G) (http://www.gene.ucl.ac.uk/nomenclature/genefamily/abc.html). There is an increased interest regarding their roles in the pharmacokinetics of drugs and distribution of endogenous metabolites, particularly in the brain, because several ABC transporters are present in the blood-brain barrier (BBB), where they regulate the access of these compounds into the brain. The ABC transporters that are known to transport drugs include the P-glycoprotein (ABCB1), several multidrug resistance-associated proteins (ABCC1–6), and the breast cancer resistance protein (ABCG2). It has become clear that the main ABC transporters expressed in the healthy human BBB are ABCB1, ABCG2, ABCC4, and ABCC5, whereas ABCC1, ABCC4, and ABCC5 are also expressed in the brain parenchyma (astrocytes or neurons) (Nies et al., 2004; Bronger et al., 2005; Declèves et al., 2006; Dauchy et al., 2008). However, the variability in their expression in different brain regions has not been studied in humans. Three of the CYP gene families (CYP1, CYP2, and CYP3) encode enzymes that metabolize xenobiotics and endogenous substrates. P450s are involved in the metabolism of widespread exogenous compounds such as drugs, pesticides, and other potentially toxic environmental compounds, as well as in the metabolism of cholesterol, bile acids, steroids, arachidonic acid, eicosanoids, vitamin D3, and retinoic acid. The presence of these proteins in the brain and in the BBB could be important for the production of biologically active compounds from centrally active drugs and environmental toxicants that cross the BBB, as well as endogenous substrates.
The main objective of the study was to investigate the expression of these genes in the whole brain, the dura mater, and 17 different encephalic areas. We analyzed the transcript expression profile of seven ABC transporters, 24 distinct P450 isoforms, and 14 different transcription factors known to be of importance in the regulation of these genes by real-time reverse transcriptase-polymerase chain reaction (PCR). In addition, we carried out Western blotting and immunohistochemistry analyses to characterize the distribution of the P450s at the cellular and subcellular levels in some brain regions. Interestingly, we find specific and abundant expression of several of the P450s and propose a role of these enzymes in the homeostasis of the normal brain and during pathological situations.
Materials and Methods
Human Brain Tissues and RNA. Pooled total RNA prepared from adult total brain, cerebellum, dura matter, and brain regions (cerebral cortex, frontal lobe, parietal lobe, temporal lobe, occipital lobe, paracentral gyrus, postcentral gyrus, insula, corpus callosum, hippocampus, caudate nucleus, nucleus accumbens, putamen, pons, and medulla oblongata) was purchased from BD Biosciences (Ozyme, Saint-Quentin-en-Yvelines, France) for use in this study (Table 1). RNA quality control for all the samples was assessed using electrophoresis on a denaturing gel or high-resolution electrophoresis system (data provided by the manufacturer). The RNA samples were stored at –80°C until use.
Normal brain tissues from two adults were obtained from the Neuropathology Department of Pitié Salpêtrière hospital. After death, the consent of the next of kin was obtained for brain removal following the procedures of the local ethics committee of the Pitié-Salpêtrière hospital. Brain tissue was preserved in 10% neutral buffered formalin when possible; tissue was stored at –80°C until use. These samples were used for Western blot and immunohistochemistry. In addition, liver microsomes, mitochondria, and total RNA were prepared from one normal human liver (Bièche et al., 2007).
Primers. We examined the 24 genes encoding all the known human P450s from families 1, 2, and 3 and CYP46A1 using primers that were previously tested and validated in different tissues (Girault et al., 2005; Bièche et al., 2007). Nucleotide primer pairs are available from Biopredic International (Rennes, France; http://www.biopredic.com). To avoid amplification of contaminating genomic DNA, one of the two primers spanned an exon-exon junction. Amplicons were between 60 and 140 nucleotides long. Specific primers for each gene encoding CYP46A1, ABC transporters (ABCB1, ABCG2, ABCC1–5), and transcription factors [aryl hydrocarbon receptor (AhR), nuclear receptor (NR)1C1/peroxisome proliferator-activated receptor (PPAR)α, NR1C2/PPARδ, NR1C3/PPARγ, NR1H2/liver X receptor (LXR)β, NH1H3/LXRα, NR1H4/farnesoid X receptor (FXR), NR1I2/pregnane X receptor (PXR), NR1I3/constitutive androstane receptor (CAR), NR2B1/retinoid X receptor (RXR)α, NR2B2/RXRβ, NR2B3/RXRγ, NR3A1/estrogen receptor (ER)α, and NR3A2/ERβ] (http://www.ens-lyon.fr/LBMC/laudet/nurebase/nomenclature/nomenclature_table.html) were designed using OLIGO 6.42 software (MedProbe, Oslo, Norway) to avoid amplifying genomic DNA. The specificity of each reaction was also assessed by melting-curve analysis to ensure the presence of only one PCR product. Primer sequences are shown in Supplemental Table S1.
Real-Time Reverse Transcriptase-PCR. Before amplification, the concentration and purity of the total RNA samples were checked spectrophotometrically at 260 nm using the NanoDrop ND-1000 instrument (Thermo Fisher Scientific, Waltham, MA). RNA was reverse-transcribed using a high-capacity cDNA archive kit in a final concentration of RNA corresponding to 20 ng/μl. The samples were incubated at 37°C for 2 h and stored at –20°C. For amplification, all the PCR reactions were performed using an ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City, CA). PCR was performed using the Absolute quantitative PCR ROX Mix (ABGene, Courtaboeuf, France). The thermal cycling conditions comprised an initial Thermus aquaticus polymerase activation step at 95°C for 15 min, 50 cycles at 95°C for 15 s, 65°C for 1 min, and a dissociation stage to identify primer-dimer and specific amplification.
The parameter threshold cycle (Ct) is defined as the fraction cycle number at which the fluorescence generated by SYBR Green dye-amplicon complex formation passes a fixed threshold above baseline. Gene expression was evaluated using the Ct value from each sample. A target gene was considered to be quantifiable when the Ct obtained for the less diluted cDNA sample was lower than 32. A Ct value of 35 was taken as the detection limit.
Concurrently with our target gene analysis, we quantified transcripts of the TBP gene (which encodes TATA box-binding protein, a housekeeping gene) as the endogenous gene control (Girault et al., 2002). Each sample was normalized to its TBP content. The method of calculation is based on the method of the ΔΔCt (Livak and Schmittgen, 2001). The quantification of the results was obtained while successively calculating: ΔCt = average Ct value (triplicate) of gene control (TBP); and ΔΔCt = ΔCt of the studied P450 – ΔCt of the less expressed P450 (the highest value of Ct ≤ 32).
Results are expressed as a percentage [relative expression (%) = (2–ΔΔCt/ Σ2–ΔΔCt) · 100]; the total expression of all the P450s or ABC transporters or transcription factors was set at 100%. Results, expressed as n-fold differences in CYP gene expression relative to the TBP gene and termed “NCYP” were determined by the formula NCYP = 2–ΔCtsample, where the ΔCt value of the sample was determined by subtracting the average Ct value of the CYP gene from the average Ct value of the TBP gene.
The NCYP values of the samples were subsequently normalized to a “basal transcript level,” that is, to the smallest quantifiable amount of CYP gene transcripts (CYP gene Ct value = 32, and it is then scored as the smallest NCYP value, i.e., NCYP value = 1). For each target gene, normalized transcript levels reflect the variations of its transcript content from tissue to tissue. By using the same threshold for the baseline (i.e., 0.1 arbitrary fluorescence unit), we compared the P450 transcript content from one tissue with all the others (Table 2).
Western Blot Analysis. The preparation of the subcellular fractions (microsomal and mitochondrial fractions) was the same for all the tissues and was adapted from a procedure described by Voirol et al. (2000). To minimize the cross-contamination between mitochondria and microsomal fractions, two additional centrifugations were performed at 20,000g for 20 min. Mitochondrial and microsomal fractions were resuspended in specific buffer [100 mM KH2PO4, 0.2 mM EDTA, 0.5 mM dithiothreitol, 20% (v/v) glycerol, pH 7.4] and then frozen at –80°C.
Monoclonal mouse anti-CYP2J2 and anti-CYP46A1 antibodies were gifts from Dr. Kristopher Krausz (National Cancer Institute, National Institutes of Health, Bethesda, MD) (Xiao et al., 2004) and Dr. David W. Russell (Department of Molecular Genetics, Dallas, TX) (Ramirez et al., 2008), respectively. The anti-CYP2U1 rabbit antiserum used has been described previously (Karlgren et al., 2004). Polyclonal rabbit anti-CYP1B1 antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal rabbit anti-CYP2D6 (Guitton et al., 1997) antibody and its specificity were previously checked in our laboratory. Yeast microsomes expressing CYP1B1, CYP2D6, and CYP2U1 were used as positive controls for the anti-CYP2D6 antibody and anti-CYP2U1 antiserum, whereas purified human CYP2J2 and recombinant human 24-hydroxylase proteins were used as positive controls for CYP2J2 and CYP46A1, respectively. Yeast microsomes expressing recombinant CYP1B1, CYP2D6, and CYP2U1 proteins were previously developed in our laboratory. Human CYP2J2 protein and recombinant human CYP46A1 protein (HΔ46A1) were gifts from Drs. Pierre Lafite and Patrick Dansette (Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, Centre National de la Recherche Scientifique, Unité Mixte de Recherche 8601, Université Paris Descartes, Paris, France) and Dr. Irina Pikuleva (Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, TX), respectively.
Samples (10 μg protein/lane) were boiled in Laemmli buffer, subjected to electrophoresis on a 10% SDS-polyacrylamide gel electrophoresis, and blotted onto Hybond-C transfer membrane (GE Healthcare, Little Chalfont, Buckinghamshire, UK). Membranes were blocked in phosphate-buffered saline buffer containing 0.1% Tween 20 and 0.2% I-Block (Applied Biosystems). Appropriate dilutions of primary antibodies (Supplemental Table S2) were then applied overnight at 4°C. Second-step alkaline phosphatase-conjugated anti-rabbit (Applied Biosystems), anti-mouse (Applied Biosystems), or anti-rat (Sigma-Aldrich, St. Louis, MO) antibodies were used and visualized using CDP-Star chemiluminescent substrate (Sigma-Aldrich).
Immunohistochemistry. Brain tissue was preserved in buffered 10% formal saline and frozen at –80°C. Tissue blocks were taken, when available, from representative areas, including the frontal lobe (middle frontal gyrus), hippocampus (CA2), and cerebellum. Ten-micrometer-thick sections were fixed for 15 min in 4% paraformaldehyde in 0.1 mM phosphate buffer, pH 7.4. Endogenous peroxidase activity was quenched for 30 min in a Tris-buffered saline/Tween (0.5%) solution containing 0.3% H2O2 and 20% methanol. Nonspecific binding was blocked by incubating sections for 1 h in 10% normal goat serum in Tris-buffered saline/Tween. Appropriate dilutions of primary antibodies (Supplemental Table S2) were then applied overnight in a humidified chamber at room temperature. Immunoreactivity was detected using the avidin-biotin VECTASTAIN ABC Elite kit (CliniSciences, Montrouge, France). The horseradish peroxidase activity was revealed using 3,3′-diaminobenzidinetetrahydrochloride. The peroxidase substrate was used according to the manufacturer's instructions (Dako France, Trappes, France). Nonspecific labeling by different P450 antibodies was assessed on directly adjacent sections by incubation with normal goat serum instead of the primary antibody or by omitting the primary antibody.
Statistical Analysis: Principal Component Analysis. The normalized matrix of P450 expression was imported into MATLAB (MathWorks, Natick, MA). Principal component analysis (PCA) analysis reduces the number of variables while retaining most of the variance in the data set. The data are recalculated into a number of principal components (PCs) where the first PC represents the largest variance in the data, the second PC represents less variance, and so on (Bakken and Jurs, 2000).
Results
In this study, we first assessed, by quantitative reverse transcriptase-PCR, the expression profile of all the P450s belonging to families 1, 2, and 3 and CYP46A1, which has already been described as a specific brain enzyme. In addition to P450 transcripts we quantified seven ABC transporters and 14 transcriptional factors (TFs). For all the genes of interest, the whole brain, dura mater, and 17 encephalic regions [cerebral cortex, four cerebral lobes (frontal, temporal, occipital, and parietal), paracentral and postcentral gyri, insula, corpus callosum, hippocampus, caudate nucleus, nucleus accumbens, putamen, substantia nigra, pons, and medulla oblongata] were studied. No significant differences on the Ct value of the TBP gene were observed between the different samples (p = 0.46), making it possible to use this gene as reference and compare the gene expression of each gene of interest in all the brain regions.
Relative P450 Transcript Expression in the Human Whole Brain. In Fig. 1A, the P450 transcript expression is calculated as a percentage of the total P450 transcript expression, and the CYP gene expression relative to the TBP gene is shown in Table 2. The main P450 transcripts expressed in total human brain were CYP46A1, CYP2J2, and CYP2U1, representing 45, 20, and 11%, respectively. CYP2D6, CYP2E1, and CYP1B1 corresponded to 5 to 10% of total P450s. Other P450s (CYP1A1, CYP2B6, CYP2C8, CYP2R1, CYP2S1, CYP2W1, CYP3A5, and CYP3A43) were present at lower levels. Finally, CYP2C9 and CYP3A43 were not quantified, and CYP1A2, CYP2A6, CYP2A7, CYP2A13, CYP2C18, CYP2C19, CYP2F1, CYP3A4, and CYP3A7 were not detected in human total brain.
Total Transcripts of Each P450 in the Different Human Brain Areas. The expression of P450 mRNAs was investigated in the dura mater and the 17 encephalic regions. CYP46A1 was found to be highly expressed in all the brain structures. Although the majority of isoforms represented by CYP2D6, CYP2E1, CYP2J2, and CYP2U1 were largely distributed throughout the brain, most P450 mRNAs display variable distribution in the regions studied (Table 2). For CYP1A1, CYP2U1, and more markedly for CYP1B1, the highest expression was found in the dura mater. CYP2A6 is only expressed in the corpus callosum, caudate nucleus, and medulla oblongata. CYP2B6 and CYP2C8 are detected in very small quantities in most locations. CYP2S1 is mainly expressed in the substantia nigra. Only weak levels of CYP3A4, CYP3A5, and CYP3A43 mRNAs were detectable in some brain locations. As observed in the whole brain, CYP2C9 and CYP3A7 were not quantified, and CYP1A2, CYP2A7, CYP2A13, CYP2C18, CYP2C19, and CYP2F1 were not detected in the different brain locations.
In Fig. 1B, levels of expression of P450 mRNAs in the brain are compared with their levels in the liver. Two P450s (CYP46A1 and CYP1B1) are found at greater quantities in the brain, 60- and 2-fold higher, respectively, whereas CYP2U1 has an equal expression. Several other P450 mRNAs, such as those for CYP2E1, CYP2D6, CYP2C8, CYP2R1, CYP2B6, and CYP1A1, are quantitatively much lower in the brain compared with that in the liver.
Regional Selective Distribution of P450s throughout the Human Brain Using PCA. We analyzed the data for differences in P450 transcript patterns between total brain and other brain regions using PCA. In the first PCA of all the regions, we observed that the dura mater was clearly separated of all the other structures because of its singular pattern expression of P450s with a high level of CYP1B1 transcripts (data not shown). In this way, the dura matter was not included in the second analysis. The inspection of the PCs reveals that the first two PCs (T1 and T2) accounted for approximately 69% of the variance of independent variables (Fig. 1C). The T1 contribution was dominated by CYP46A1 (0.9) and CYP2D6 (0.4), whereas CYP2C8, CYP1A1, and CYP2B6 also contributed to T2 variation (Fig. 1D). PCA analysis allowed for the discrimination of several patterns in P450 expression, thus indicating the existence of comparable P450 contents in distinct brain regions. In Fig. 1C, three regions (putamen, nucleus accumbens, and caudate nucleus) are clearly separated from all the others on the first component, indicating a similar P450 pattern for these basal ganglia structures. Total brain, cerebral cortex, frontal lobe, and, to a lesser extent, hippocampus qualitatively contain the same P450 transcripts corresponding to a high number of distinct isoforms. We observed that the medulla oblongata and pons share a very similar pattern of P450 expression. Finally, the cerebellum, corpus callosum, and substantia nigra appear to have a distinct pattern of P450 expression, which is unlike that seen in other regions.
Relative ABC Transporter Transcript Expression in the Whole Human Brain. In Fig. 2A, ABC transporter transcript expression was calculated as a percentage of the total transcript expression. The n-fold differences in ABC transporter gene expression relative to the TBP gene are reported in Table 3. The main transcripts detected in total brain are ABCC5 and ABCG2, representing around 55 and 30%, respectively (Fig. 2A). ABCC1, ABCC4, and ABCB1 transcript levels corresponded to 10, 6, and 3% of total ABC transporter transcripts investigated, respectively. ABCC2 and ABCC3 transcripts were not detected in human total brain extract.
Total Transcripts of ABC Transporter in the Different Brain Localizations. Although ABCC5, ABCG2, and ABCC1 are broadly found in different brain areas, ABC transporters are not expressed identically in all the brain structures. Indeed, whereas ABCB1 transcripts are mainly detected in putamen, substantia nigra, and pons, they appear in smaller quantities in other areas. In Fig. 2B, the levels of ABC transporter mRNA expression in the brain are compared with those in the liver. Compared with the levels observed in the liver, several ABC transporters (ABCC1, ABCC4, ABCC5, and ABCG2) are expressed more in the brain, ranging from around 1.5-fold greater for ABCG2 and ABCC1 to 10-fold more for ABCC5.
Relative TF Transcript Expression in the Whole Human Brain. The TF transcript expression was calculated as a percentage of the total TF transcript expression (Fig. 2C), and the n-fold differences in TF transporter gene expression relative to the TBP gene are reported in Table 4. The main TF transcripts expressed in human total brain are NR1H2/LXRβ, NR2B2/RXRβ, NR1C2/PPARδ, NR2B1/RXRα, NR1C1/PPARα, and NR1C3/PPARγ, representing around 30, 25, 20, 10, 6, and 4%, respectively. NR2B3/RXRγ, NR3A2/ERβ, AhR, and NR1H3/LXRα corresponded to 2.4, 1.3, 0.7, and 0.5% of total TF transcripts, respectively. NR3A1/ERα and NR1I2/PXR are present at the threshold for quantification, indicating weak expression in the whole brain. NR1I3/CAR and NR1H4/FXR are detected under the limit of quantification.
Total Transcripts of Each TF in the Different Locations. Whereas NR1H2, NR1C2, NR1C3, NR2B1, NR2B2, and NR2B3 mRNAs were found broadly in all the brain regions, some other TFs were localized in specific regions (Table 4). Thus, AhR, NR3A1, and NR1H3 were mainly expressed in the dura mater and weakly detected in other areas. It is important to note that TFs not quantifiable in the whole brain were found at high levels in some brain areas. It is the case for NR1I3 quantified in the nucleus accumbens, the caudate nucleus, and the putamen and for NR1H4 in the corpus callosum, the nucleus accumbens, the putamen, and the dura mater.
In Fig. 2D, the levels of expression of TFs in the brain are compared with those of the liver. The results showed that NR1H2/LXRβ, NR2B2/RXRβ, and NR1C2/PPARγ expression was 10, 40, and 70% higher in the brain than in the liver, respectively. On the other hand, we observed weaker expression of NR2B1, NR1C1, and NR1C3 (5-, 7-, and 5-fold, respectively) in the brain compared with the liver. The transcripts of AhR, NR1H3, NR3A1, and NR1I2 were also less expressed in the brain than in the liver (14-, 38-, 67-, and 160-fold, respectively). Another isoform, NR2B3, which was secondarily expressed in the brain, was expressed quantitatively much higher (20-fold) in the brain compared with the liver (Table 4).
Analysis of P450 by Western Blotting. Because of the high levels of the mRNAs for CYP1B1, CYP2D6, CYP2E1, CYP2J2, CYP2U1, and CYP46A1 detected, we investigated the corresponding protein expression in both microsomal and mitochondrial purified fractions from human liver and brain samples, including frontal lobe, hippocampus, substantia nigra, and cerebellum. The quality assessment of the method separating the cellular fractions is explained in the Supplemental Data. We detected CYP2D6 and CYP2J2 in liver microsomes and mitochondria (data not shown). CYP1B1 was only found in liver microsomes, whereas we did not detect CYP2U1 and CYP46A1 in our liver sample. Moreover, as shown in Fig. 3, we were able to detect CYP1B1, CYP2D6, CYP2J2, CYP2U1, and CYP46A1 in all the regions analyzed. Although we used three different polyclonal anti-CYP2E1 antibodies (two homemade and one commercial), we were unable to detect CYP2E1 in our two normal human brains (data not shown). It is interesting to note that, whereas CYP2U1 and CYP46A1 transcripts are present both in brain and liver, our data indicate the presence of the corresponding proteins only in the brain.
Brain CYP1B1, CYP2D6, CYP2J2, CYP2U1, and CYP46A1 Protein Immunohistochemical Staining. We determined the topographical and cellular distribution of CYP1B1, CYP2D6, CYP2J2, CYP2U1, and CYP46A1 in three different brain regions (frontal lobe, hippocampus, and cerebellum) as illustrated in Fig. 4 and Supplemental Figs. S1 and S2. The regional and cellular expression and the intensity of staining in the sections examined from two normal brains are summarized in Table 5.
In the frontal cortex sections (Fig. 4, A–F, and Supplemental Fig. S1, A–C), the detection of CYP2D6 (Supplemental Fig. S1A) and CYP2U1 (Supplemental Fig. S1B) proteins was compatible with low expression in neuronal cells, whereas staining for CYP46A1 revealed its exclusive and robust expression in the soma and dendrites (Fig. 4B). No immunostaining of CYP1B1 protein was seen in neurons. Astrocytes (including glia limitans) were intensely stained with the CYP1B1, CYP2D6 (Fig. 4C), CYP2J2 (Fig. 4D), and CYP2U1 (Fig. 4E and Supplemental Fig. S1C) antibodies. Indeed, CYP2J2 (Fig. 4D) and CYP2U1 (Supplemental Fig. S1C) were strongly detected in the astrocyte foot processes that contact microvessels. CYP1B1 was detected in microglia in sections from frontal cortex (Fig. 4F), hippocampus (Supplemental Fig. S1H), and cerebellum, which all showed high levels of staining.
We observed similar results in sections from hippocampus (Fig. 4, G and H, and Supplemental Fig. S1, D–H). Dentate gyrus cells, CA1 through CA3, and CA4 pyramidal neurons showed only moderate staining for CYP2D6, CYP2J2 (Supplemental Fig. S1F), and CYP2U1 (Supplemental Fig. S1G) compared with the strong signal for CYP46A1 (Fig. 4H). Moreover, staining of CYP1B1 (Supplemental Fig. S1D) and CYP2J2 (Supplemental Fig. S1E) proteins was seen in dentate gyrus cells. Intense staining in astrocytes was observed for CYP1B1, CYP2D6, CYP2J2, and CYP2U1.
Cerebellar sections (Fig. 4, I–L, and Supplemental Fig. S2, A–J) stained intensely with CYP2D6 (Fig. 4J), CYP2J2 (Supplemental Fig. S2A), and CYP46A1 (Supplemental Fig. S2B) in the soma and dendritic trees of many Purkinje cells. No CYP1B1 and CYP2U1 (Fig. 4L and Supplemental Fig. S2I) staining was seen in this cell type. Weak and moderate staining of CYP2J2 (Supplemental Fig. S2C) and CYP2U1 (Supplemental Fig. S2D) was observed in the cell body of the granule cells. Moreover, intense staining for CYP1B1 (Supplemental Fig. S2E) compared with moderate staining for CYP2D6 (Supplemental Fig. S2F), CYP2J2, and CYP46A1 (Supplemental Fig. S2G) was observed in glomeruli. Astrocytes that comprise the molecular layer, white matter, and glia limitans displayed high staining for CYP1B1, CYP2D6 (Supplemental Fig. S2H), CYP2J2 (Fig. 4K), and CYP2U1 (Fig. 4I and Supplemental Figs. S1I and S2J). As previously shown in the two other regions, CYP2J2 and CYP2U1 enzymes were strongly detected in the astrocytes that contact microvessels. No immunostaining was seen in control sections incubated without primary antibody (Fig. 4, A, G, and I).
Discussion
We present here the first study of transcript analysis of all 24 P450s belonging to families 1 through 3 and CYP46A1, seven ABC transporters, and 14 TFs in human whole brain, cerebellum, dura mater, and in a high number of other brain regions. Our results show the presence of a large variety of xenobiotic transporters and metabolizing enzymes in human brain and show for the first time their apparent selective distribution in different cerebral regions. Our results clearly showed the presence of the major transcripts corresponding to CYP1B1, CYP2D6, CYP2E1, CYP2J2, CYP2U1, and CYP46A1 in human brain across all the brain structures with an unequal distribution in different brain regions. The presence of P450 transcripts in human whole brain was previously reported (Nishimura et al., 2003), but the distribution of these transcripts throughout the brain and the presence of the corresponding proteins was unknown or poorly documented (Nishimura et al., 2003). Despite similar results for the main P450s (CYP46A1, CYP2E1, and CYP2J2) at the mRNA level, some discrepancies with this previous Japanese study that reported lower levels of CYP1B1 and CYP2D6 than our results can be noted. Concerning the lack of CYP1B1 and CYP2D6 mRNA in most brain regions reported in other studies, this could be because of the sensitivity or specificity of the methods (PCR versus quantitative PCR) (McFayden et al., 1998). At the protein level, we confirmed the presence of CYP2D6 in human frontal lobe, hippocampus, and cerebellum as described elsewhere (Chinta et al., 2002; Miksys et al., 2002). Moreover, it appears that CYP2U1 protein is present in the frontal lobe at higher levels than in the cerebellum, similar to a previous report regarding the rat brain (Karlgren et al., 2004). For the first time, CYP2J2, CYP2U1, and CYP46A1 were characterized at the cellular level in different human brain locations. The subcellular localization in the mitochondrial and microsomal fractions of P450 proteins in the human frontal lobe, hippocampus, substantia nigra, and cerebellum was shown for the main P450s (CYP1B1, CYP2D6, CYP2J2, CYP2U1, and CYP46A1). This level of detail was not available for the majority of P450s. Finally, we assessed the presence of cytochrome P450 reductase (CRP) and ferredoxin reductase (FDXR) enzymes, which is a prerequisite for a functional enzymatic system (Bhagwat et al., 2000; Lavandera et al., 2007). Our results support the existence of active enzymes, as previously documented in human and rodent brain by measurement of microsomal monooxygenase activities and enzyme inhibition method (Miksys and Tyndale, 2009).
Previous studies have focused on the three major enzymes, CYP46A1, CYP2J2, and CYP2U1, with reports published about their catalytic activities toward endogenous substrates underlying the requirement of the knowledge of their distribution throughout the brain. CYP46A1 is a cholesterol 24-hydroxylase that is selectively expressed in the brain and responsible for cholesterol turnover in the central nervous system. Ramirez et al. (2008) have shown that 24-hydroxylase is highly expressed in pyramidal neurons of the hippocampus and cortex, in Purkinje cells of the cerebellum, and in hippocampal and cerebellar interneurons. In accordance with these findings, our results showed exclusive neuronal expression of CYP46A1 in human brain. CYP2J2 metabolizes arachidonic acid to 20-hydroxyeicosatetraenoic acid and epoxyeicosatrienoic acids, which play a critical role in the regulation of renal, pulmonary, cardiac, and vascular function. CYP2U1 metabolizes arachidonic acid and other long-chain fatty acids (Chuang et al., 2004) and could play a role in fatty acid signaling processes. The substrate specificity of these enzymes is consistent with their high expression in the mitochondria, an organelle devoted to fatty acid and steroid metabolism.
Other major P450s (CYP2E1, CYP2D6, CYP1B1) have been previously described to be involved or to have potential impact on brain function. First, CYP2E1 can generate reactive oxygen species that represent toxic molecules that have been implicated in the pathogenesis of neurodegenerative disorders such as Parkinson's disease. Furthermore, recent data suggest that CYP2E1 could modulate dopamine release and free radical production in agreement with its presence in the substantia nigra (Shahabi et al., 2008). In this study, we were unable to show the presence of CYP2E1 protein, which has been reported to be expressed at very low levels in the brains of nonalcoholics (Howard et al., 2003). The lack of detection of CYP2E1 protein, even though high levels of CYP2E1 mRNA were present, could be explained by the rapid ubiquitination of the enzyme in the absence of substrates (Banerjee et al., 2000). CYP2D6 has been implicated in the metabolism of endogenous compounds such as anandamide (Snider et al., 2008) and neurotransmitters (Yu et al., 2003, 2004) and reported as a risk factor in Parkinson's disease (Elbaz et al., 2004) and psychiatric disorders (Llerena et al., 2007). Finally, CYP1B1 is overexpressed in a variety of central nervous system malignancies, and increased CYP1B1 expression in glial tumors was associated with decreased patient survival (Barnett et al., 2007).
Our study showed the existence of the regional expression pattern of P450s, suggesting similar metabolic pathways involving P450 in different brain regions. For example, the caudate nucleus and nucleus accumbens share an identical expression profile of P450s, including a high level of CYP2D6. It is known that the caudate nucleus and nucleus accumbens contain high quantities of different neurotransmitters, such as dopamine, noradrenalin, and serotonin (Tong et al., 2006), and the role of CYP2D6 in the metabolism of these neurotransmitters is now well documented (Yu et al., 2003, 2004). Our results confirm the presence of CYP2D6 in Purkinje cells and cortical neurons (Gilham et al., 1997).
In addition to P450s, we reported the presence of five ABC transporters (ABCB1, ABCC1, ABCC4, ABCC5, and ABCG2), confirming previously published data (Dauchy et al., 2008). In this recent study, ABCB1 and ABCG2 were predominantly found in the BBB, whereas ABCC1, ABCC4, and ABCC5 were distributed equally in microvessels and the parenchyma of human cortex, indicating the strong capability of this barrier to limit the diffusion of xenobiotics into neuronal and glial cells. Therefore, we can hypothesize that ABC transporters and P450s cooperate in detoxication pathways to eliminate xenobiotics. Our results must be interpreted carefully because of the heterogeneity of the samples analyzed. The relative distribution of ABC transporters between the BBB and parenchyma is probably different and remains to be clearly established.
Characterization of transport and metabolism properties of the human brain requires the elucidation of signaling mechanisms involved in gene regulation. The presence of TF transcripts in human whole brain has been previously reported (Nishimura et al., 2004), but we described for the first time the distribution of these transcripts in various brain regions. Our findings are in agreement with those of Nishimura et al. (2004) concerning the main TFs (NR1H2/LXRβ, NR1C2/PPARδ, NR1C1/PPARα, NR2B2/RXRβ, and NR2B1/RXRα) and bring new data about the presence of AhR. In our study, NR1I3/CAR and NR1I2/PXR, which are able to induce the expression of CYP2B, CYP2C, CYP3A, and ABCB1, were found at low levels in the human brain, in agreement with previous studies (Lamba et al., 2004a,b; Nishimura et al., 2004). According to the brain region, our results allowed us to distinguish several major regulation pathways involving the AhR, NR1H2/LXRβ, and NR1C/PPAR subfamilies, respectively, in agreement with the presence of the corresponding P450 target genes. An interesting finding is the presence of high amounts of AhR transcripts in the dura mater, in which high levels of CYP1A1, CYP2S1, and CYP1B1, known to be inducible in an AhR-dependent manner, are also detected.
The epoxygenase CYP2J2 metabolized epoxyeicosatrienoic acids and other metabolites that have been found to activate both NR1C1/PPARα and NR1C3/PPARγ. Finally, recent studies clearly showed that PPAR and LXR are nuclear receptors activated by fatty acid and cholesterol derivates, respectively, that control the expression of a range of genes involved in the lipid metabolism and inflammation (Bensinger and Tontonoz, 2008; Hong and Tontonoz, 2008).
In conclusion, the present study encompasses original data concerning P450, ABC transporters, and related TF distribution in the whole human brain, cerebellum, dura mater, and 16 different brain regions. Several P450s were detected in specific regions possibly in relation with their involvement in brain physiology. Although extremely informative, this work constitutes a descriptive approach required to establish a regional distribution. This work needs to be extended to investigate the substrates of these enzymes, the metabolic end products, and their biological activities. However, the limits of this study include the use of pooled specimens, which does not allow for the evaluation of interindividual variability and the lack of information about smoking habits, drug consumption, and other clinical data from the brain donors. The identification of highly expressed xenobiotic transporters and metabolizing enzymes in the brain reinforces the hypothesis of a potential impact of environmental exposure on cerebral function. However, our results also bring new significant insights into the distribution of P450 enzymes involved in endogenous metabolic pathways, including cholesterol, fatty acids, and neurotransmitters. Altogether, these data provide original candidates for generating biochemical hypotheses during neurodegenerative and psychiatric disorders and consequently for developing new therapeutic strategies.
Acknowledgments
We thank Dr. Kristopher Krausz (National Cancer Institute, National Institutes of Health, Bethesda, MD), Dr. David W. Russell (Department of Molecular Genetics, Dallas, TX), Drs. Pierre Lafite and Patrick Dansette (Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, Centre National de la Recherche Scientifique, Unité Mixte de Recherche 8601, Université Paris Descartes, Paris, France), and Dr. Irina Pikuleva (Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, TX) for providing the monoclonal mouse anti-CYP2J2 antibody, monoclonal mouse anti-CYP46A1 antibody, human purified CYP2J2 protein, and recombinant human CYP46A1 protein (HΔ46A1), respectively.
Footnotes
-
This work was supported by a grant from Servier Technology.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.109.027011.
-
ABBREVIATIONS: P450, cytochrome P450; ABC, ATP-binding cassette; BBB, blood-brain barrier; PCR, polymerase chain reaction; AhR, aryl hydrocarbon receptor; NR, nuclear receptor; PPAR, peroxisome proliferator-activated receptor; LXR, liver X receptor; FXR, farnesoid X receptor; PXR, pregnane X receptor; CAR, constitutive androstane receptor; RXR, retinoid X receptor; ER, estrogen receptor; Ct, threshold cycle; PCA, principal component analysis; PC, principal component; TF, transcriptional factor; CPR, cytochrome P450 reductase; FDXR, ferredoxin reductase.
-
↵ The online version of this article (available at http://dmd.aspetjournals.org) contains supplemental material.
- Accepted April 7, 2009.
- Received February 2, 2009.
- The American Society for Pharmacology and Experimental Therapeutics