Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

Phenobarbital (PB) elicits pleiotropic effects on various cellular processes in liver, including growth, apoptosis, communication, tumorigenesis, and induction of drug/steroid-metabolizing enzymes such as cytochromes P450 (P450s) and specific conjugation enzymes. The nuclear receptor CAR is implicated in mediating PB responses (Zelko and Negishi, 2000; Sueyoshi and Negishi, 2001). In response to PB exposure, CAR translocates to the nucleus, forms a heterodimer with the retinoid X receptor (RXR), and activates the 51-bp PB-responsive enhancer module that is conserved in the mouse, rat, and human CYP2B genes (Trottier et al., 1995; Park et al., 1996; Honkakoski et al., 1998a,c; Kawamoto et al., 1999; Ramsden et al., 1999; Sueyoshi et al., 1999; Wei et al., 2000; Zelko et al., 2001). In addition, similar CAR-regulated enhancer sequences have been found in some other PB-inducible genes, such as CYP3A and human bilirubin UDP-glucuronosyltransferase UGT1A1 (Sueyoshi et al., 1999; Smirlis et al., 2001; Sugatani et al., 2001). Increased liver weight and DNA synthesis in response to PB treatment were also reported to be CAR-mediated (Wei et al., 2000).

There are two major questions regarding the role of CAR as a regulator of PB response. 1) How many genes are regulated by CAR? PB induces more than 20 different genes in chicken hepatocytes (Frueh et al., 1997), and as many as 77 differentially displayed cDNAs were detected in PB-induced mouse liver (Garcia-Allan et al., 2000). The number of genes under CAR regulation, however, remains unknown. 2) To what extent does CAR coordinate multifaceted responses to PB? PB concertedly induces various hepatic enzymes to increase the drug/steroid-metabolizing capability, including P450s, P450 reductase (the key enzyme that transfers electrons from NADPH to P450), and such transferases as glucuronosyltransferases, glutathione S-transferases, and sulfotransferases and transporters (Waxman and Azaroff, 1992; Honkakoski and Negishi, 1998b). Moreover, PB also induces aminolevulinate synthase 1 to increase the supply of heme to newly synthesized P450 apoprotein, and proliferates endoplasmic reticulum to provide additional membrane bindings for P450s and other membrane enzymes. To shed light on these questions, we used RT-PCR and cDNA microarrays to compare gene expression profiles derived from livers of PB-treated wild-type and CAR-null mice. PB induced a number of drug/steroid-metabolizing enzymes in a CAR-dependent fashion, although not all genes induced by PB were regulated by CAR. In addition, we have identified a large number of potential new CAR-target genes. CAR plays diverse set of roles (e.g., activator, repressor, and blocker of gene expression) that go beyond drug/steroid metabolism.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References

Animals and RNA Preparation. CAR-null and wild-type mice were produced from a single breeding colony of multiple CAR-heterozygous females with one CAR-heterozygous male (Webb et al., 2001). These mice were fed Teklad 5053 autoclaved rodent chow (Harlan Teklad, Madison, WI) ad libitum and provided with autoclaved drinking water and housed in high-efficiency particulate air (HEPA)-filtered caging in a temperature- and humidity-controlled environment on a 12-h/12-h light/dark cycle. Germline transmission of the disrupted allele was detected by Northern blot analysis or by RT-PCR of liver RNA. Six- to 9-week old CAR-null and wild-type mice were treated by intraperitoneal injection of PB (Sigma, St. Louis, MO) in 10 mg/ml saline solution at a dose of 100 mg/kg of body weight for 12 h before sacrifice. For control mice, 200 µl of saline per 20 g of body weight was injected. Total liver RNA was prepared from each mouse using TRIzol reagent (Invitrogen, Carlsbad, CA), from which mRNA was enriched using Oligotex mRNA kit (QIAGEN, Valencia, CA).

RT-PCR. cDNAs were prepared from total RNA using SuperScript II reverse transcriptase (Invitrogen) and were amplified using the following sets of primers: CYP2B10 mRNA, 5'-AAAGTCCCGTGGCAACTTCC-3' and 5'-CATCCCAAAGTCTCTCATGG-3'; CYP3A11 mRNA, 5'-CTCAATGGTGTGTATATCCCC-3' and 5'-CCGATGTTCTTAGACACTGCC-3'; P450 reductase mRNA, 5'-GTTTGCTGTGTTTGGTCTCG-3' and 5'-CTCTCAGTGCCTTGGTTCAG-3'. PCR was performed in total volume of 50 µl on cDNA synthesized from 0.12 µg of total RNA using Advantage 2 PCR enzyme (CLONTECH, Palo Alto, CA). The PCR protocol consisted of an initial denaturing step at 95°C for 1 min followed by 19 to 29 cycles of 95°C for 30 s and 68°C for 1 min. The expected sizes of the amplified cDNA were 340 bp, 423 bp, and 393 bp for CYP2B10, CYP3A11, and P450 reductase mRNAs, respectively.

Microarray Hybridization. cDNA microarray chips containing 8736 mouse genes/ESTs were prepared as described previously (DeRisi et al., 1997).¹ Detailed methods are available at http://dir.niehs.nih.gov/microarray. Briefly, mouse clones (gene expression microarray set 1) were purchased as bacterial stocks from Incyte (Palo Alto, CA). Routine re-sequencing of the clones is conducted in our laboratory for validation of spot identity. Updated clone lists are available at the same web site. Purified plasmids were isolated and used as templates for amplification of inserts by PCR. The PCR products were analyzed on 2% agarose gels to ensure quality of the reactions and purified by ethanol precipitation. The purified cDNAs were resuspended in ArrayIt buffer (Telechem, San Jose, CA) and spotted onto poly(L-lysine) coated glass slides using a modified, robotic DNA arrayer (Beecher Instruments, Bethesda, MD). Poly(A⁺) RNA (2-4 µg) was labeled with Cy3 and Cy5-conjugated dUTP (Amersham Biosciences, Piscataway, NJ) using a reverse transcription reaction and hybridized to the mouse cDNA microarray chip (DeRisi et al., 1997). cDNA chips were scanned using the Axon 4000 scanner (Axon Instruments, Union City, CA), and images were analyzed using the Array Suite Software (Scanalytics, Fairfax, VA). The relative fluorescence intensity was measured for each labeled RNA and a ratio of the values for the intensity of each fluor bound to each probe (CAR-null or wild-type) was calculated. The level of autofluorescence was measured and a minimum intensity cut-off was set just above this value. The distribution of the ratio of all of the genes was calculated and intensity ratio values that differed from the median with a confidence interval of greater than 95.0% (Chen et al., 1997) were scored as significant changes. Approximately equal amounts of the same liver RNA samples within each experimental group were labeled and hybridized in four independent reactions to four independent chips; each RNA sample was labeled twice with each fluorophore to account for fluorophore incorporation bias. The data for each array experiment were normalized using the mean of all of the targets on the array (range, 0.8-1.2, indicating a normal distribution). The coefficient of variance for each hybridization was less than 0.1. A database tool, Microarray Project System (MAPS; Bushel et al., 2001) was used to compile the overall list of consistent, significantly changed genes across the multiple hybridizations. The hierarchical clustering method from Eisen et al. (1998) was used to cluster the gene expression changes for visualization.

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion References

First, we performed RT-PCR on liver RNA samples from control and PB-treated wild-type and CAR-null mice to confirm that CAR, in fact, regulates the Cyp2b10 gene. The CYP2B10 mRNA was measured individually for the four wild-type and six null mice. The mRNA was greatly increased in all wild-type mice treated with PB, whereas no change in message was observed in the null mice (Fig. 1). Thus, the PB induction of Cyp2b10 gene was tightly regulated by CAR in the mice. Although CYP3A11 and P450 reductase showed high levels of the constitutive expression of their mRNA, PB treatment caused a further increase in the wild-type but not in CAR-null mice (Fig. 1). These findings suggested a general role of CAR in regulating a number of drug/steroid-metabolizing enzymes. In particular, the coordinated regulation by CAR of P450 reductase with the P450 induction is particularly important because the reductase is an essential component of P450-dependent metabolic activity.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Absence of PB induction in CAR-null mice. RT-PCR was performed with total RNA from each liver as described under Materials and Methods. The amplified DNA was separated on a 1.2% agarose gel and visualized with staining by ethidium bromide.

To further investigate to what extent CAR regulates hepatic genes, we performed microarray experiments using liver mRNA samples obtained and pooled from three control and three PB-treated wild-type mice and two each of control and PB-treated CAR-null mice. After scanning, microarray images were analyzed to determine relative intensity ratios between the control and treated mice. Subsequently, genes that were determined to be induced or repressed at a 95% confidence interval in three or four of four hybridizations were used to compile a list of 144 significant changes. This list of genes was then used for hierarchical cluster analysis that reiteratively joins the two closest clusters starting from singletons (Fig. 2). The final output revealed relatively high correlation (r > 0.8) in the expression of sets of genes across either of PB-treated wild-type or CAR-null mice. Genes were grouped into six major categories: group A includes genes induced by PB in wild-type but not in CAR-null mice; group B, genes induced by PB in both wild-type and CAR-null mice; group C, genes induced by PB in CAR-null but not in wild-type mice; group D, genes repressed by PB in wild-type but not in CAR-null mice; group E, genes repressed by PB in CAR-null but not in wild-type mice; and group F, genes repressed by PB in both wild-type and CAR-null mice.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Hierarchical clustering of gene expression ratio intensity values. The image shows an average linkage clustering of 144 mouse clones across wild-type (+/+) and CAR-null (/) mice treated with PB. Numerical values are encoded by colors: red and green are used to represent positive and negative values, respectively. Vertical bars indicate group of genes regulated differently: CAR-dependent induction (A), CAR-independent induction (B), CAR-dependent blocking of induction (C) or repression (E), and CAR-dependent repression (D) and -independent repression (F). Fold of the alterations should be taken as semiquantitative due to hybridization conditions intrinsic to the cDNA microarray experiment.

CAR-Dependent PB-Inducible Genes. cDNA microarray analyses revealed a total of 22 genes that were induced by PB in only the wild-type mice (group A). In addition to PB-inducible Cyp2b10 and Cyp3a11, six other Cyp genes (Cyp2b9, -2d10, -2j5, -2f2, -4a10, and -4a14) were contained on the chip, but only Cyp2b10 was identified as being induced by PB in a CAR-dependent fashion. The degree of induction was very low (Fig. 1) and/or high constitutive levels of CYP3A11 mRNA may have masked differential detection of this gene on the chip. A cDNA probe for P450 reductase mRNA was not on the chips. Ten known genes were identified as those regulated by CAR, eight of which encode enzymes involved in xenochemical metabolism (Fig. 2). The other two genes are cAMP-regulated phosphoprotein and semaphorin 3. The 12 remaining genes are ESTs. Together with P450 reductase and CYP3A11, these results indicate that drug/steroid-metabolizing enzymes constitute a majority of PB-inducible genes that are regulated by CAR and the present number of CAR-regulated enzymes could have been underestimated because probes, such as for glucuronosyltransferases, were not on the chip.

CAR-Dependent PB-Repressed Genes. A total of 30 genes were found to be down-regulated in response to PB treatment in wild-type but not CAR-null mice (group D). Of the 30 genes, 11 are ESTs. Interestingly, the majority of repressed genes encode enzymes and proteins that are involved in basic liver function. Carnitine palmitoyltransferase and enoyl-CoA isomerase are key enzymes in fatty acid oxidation. Phosphoenolpyruvate carboxykinase is involved in glucose synthesis and catalyzes the formation of phosphoenolpyruvate from oxaloacetate in the presence of GTP. Fibronectin promotes cell adhesion and angiogenin regulates neovascularization required for normal physiology such as embryonic development, reproduction, and wound repair. Angiogenin is also up-regulated in various tumor tissues. In addition, our present microarray experiments showed that peptidyl isomerase, keratins, metallothionein, translocator of inner mitochondrial membrane 4, liposaccharide binding protein, myoglobin, DNase inhibitor inhibited by DNA fragment factor, lipocalin, inter--trypsin inhibitor, fibrinogen 1, histone H1, and pleiotropic regulator 1 were all in this group. These enzymes and proteins are involved in signal transduction, fatty acid oxidation, energy metabolism, and/or cell surface communication, indicating that CAR may mediate PB effects on many critical processes in liver physiology.

Genes under CAR-Dependent Blocking. Interestingly, this group contains genes that are either induced or repressed by PB in only the CAR-null mice, not in wild-type mice (Fig. 2). Apparently, the presence of CAR blocked these genes from being either induced or repressed by PB. Among the 24 genes that fall into these two categories, only nine genes are known. Including Cyp4a10 and Cyp4a14, four genes are induced only in the CAR-null mice. Consistent with the results for the CYP4A10 mRNA obtained from real-time PCR (Fig. 3A), these Cyp genes were clearly induced by PB in only the CAR-null mice, whereas these were slightly repressed in the PB-treated wild-type mice. Both Cyp4a genes are highly inducible by the peroxisome proliferator methylclofenapate in normal mouse liver (Heng et al., 1997). CYP4A10 and CYP4A14, major microsomal lipid peroxidases, are up-regulated in Cyp2e1-null mice with nonalcoholic steatohepatitis, thus identified as possible initiators of oxidative stress in the liver (Leclercq et al., 2000). Intriguingly, CAR is absent in the fatty liver of the obese Zucker rat (K. Yoshinari and M. Negishi, unpublished observations), in which lipid peroxidation products were increased (Koneru et al., 1995). Thus, CAR can be viewed as an endogenous determinant suppressing oxidative stress by preventing the induction of Cyp4a genes by xenochemical exposures. This function of CAR as a repressor for oxidative stress is consistent with the CAR-mediated induction of superoxide dismutase 3 by PB (Sugatani et al., 2001). Eight genes, of which three are ESTs, are repressed only in the CAR-null mice.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Quantitative measurement of mRNA levels. Real time PCR was performed with the pooled mRNA samples that were used for the cDNA microarray experiments. Closed and dotted bars represent folds of induction in the PB-treated wild-type and CAR-null mice, respectively. Bars indicate experimental errors from multiple experiments using the same RNA samples. A, CYP4A10 mRNA. B, ALAS, 7-hydroxycholesterol reductase (Dhcr7) and squalene epoxidase (Sqle) mRNAs.

CAR-Independent PB-Inducible or -Repressed Genes. A total of 37 genes were induced in both wild-type and CAR-null mice (group B), whereas 23 genes were repressed (group F) (Fig. 2). Several known genes were induced by PB in a CAR-independent manner, and aminolevulinate synthase 1 (ALAS-1) was one of the 14 known genes that was induced by PB. ALAS-1, the key enzyme in regulating heme biosynthesis was clearly induced by PB both in the wild-type and CAR-null mice. In addition, real time PCR confirmed the induction of ALAS-1 mRNA in both types of mice (Fig. 3B). The concerted induction of the enzymes is considered as a way to effectively supply heme to newly synthesized P450 apoenzyme. The mechanism by which liver cells coordinate the induction of both P450 and ALAS has been an unsolved question. The present results clearly indicate that CAR does not play a role in the coordinated induction of the two enzymes systems.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Function-based representation of PB-modulated genes. Genes in each group are considered based on their function and a possible function that may represent the group is drawn.