Regulation of Cytochrome P450 2B10 (CYP2B10) Expression in Liver by Peroxisome Proliferator-activated Receptor-β/δ Modulation of SP1 Promoter Occupancy*

Alcoholic liver disease is a pathological condition caused by overconsumption of alcohol. Because of the high morbidity and mortality associated with this disease, there remains a need to elucidate the molecular mechanisms underlying its etiology and to develop new treatments. Because peroxisome proliferator-activated receptor-β/δ (PPARβ/δ) modulates ethanol-induced hepatic effects, the present study examined alterations in gene expression that may contribute to this disease. Chronic ethanol treatment causes increased hepatic CYP2B10 expression inPparβ/δ+/+ mice but not in Pparβ/δ−/− mice. Nuclear and cytosolic localization of the constitutive androstane receptor (CAR), a transcription factor known to regulate Cyp2b10 expression, was not different between genotypes. PPARγ co-activator 1α, a co-activator of both CAR and PPARβ/δ, was up-regulated in Pparβ/δ+/+ liver following ethanol exposure, but not in Pparβ/δ−/− liver. Functional mapping of the Cyp2b10 promoter and ChIP assays revealed that PPARβ/δ-dependent modulation of SP1 promoter occupancy up-regulated Cyp2b10 expression in response to ethanol. These results suggest that PPARβ/δ regulates Cyp2b10 expression indirectly by modulating SP1 and PPARγ co-activator 1α expression and/or activity independent of CAR activity. Ligand activation of PPARβ/δ attenuates ethanol-induced Cyp2b10 expression in Pparβ/δ+/+ liver but not in Pparβ/δ−/− liver. Strikingly, Cyp2b10 suppression by ligand activation of PPARβ/δ following ethanol treatment occurred in hepatocytes and was mediated by paracrine signaling from Kupffer cells. Combined, results from the present study demonstrate a novel regulatory role of PPARβ/δ in modulating CYP2B10 that may contribute to the etiology of alcoholic liver disease.

Chronic consumption of ethanol causes steatosis, hepatomegaly, hepatitis, fibrosis, and cirrhosis collectively referred to as alcoholic liver disease and has become a major health issue because of high morbidity and mortality (1,2). Nutrient deficiencies, impaired fatty acid metabolism, induction of xenobiotic-metabolizing enzymes, and increased oxidative stress are all associated with liver toxicity induced by ethanol (3). However, the precise molecular mechanisms underlying the cause of alcoholic liver disease are not well understood.
Nuclear receptors have key roles in regulating lipid homeostasis and inflammation during the pathogenesis of alcoholic liver disease (4,5). For example, activation of constitutive androstane receptor (CAR) 3 facilitates ethanol metabolism, resulting in enhanced liver damage by increasing oxidative stress, apoptosis, and accumulation of lipids in hepatocytes (6). By contrast, peroxisome proliferator-activated receptor-␣ (PPAR␣) shows a protective role in alcoholic liver disease because ethanol-treated Ppar␣ Ϫ/Ϫ mice exhibit marked liver damage including hepatomegaly, hepatic inflammation, fibrosis, and apoptosis as compared with ethanol-treated Ppar␣ ϩ/ϩ mice (7). A recent study showed that PPAR␤/␦, another PPAR subtype, prevents ethanol-induced hepatic effects by suppressing lipogenesis, modulating amino acid metabolism, and altering pyridoxal kinase activity (8). This is consistent with previous studies showing that PPAR␤/␦ protects against liver damage induced by various hepatotoxicants (9 -11). Although the protective role of PPAR␤/␦ in chemically induced liver toxicity was demonstrated in several models, the detailed molecular mechanism(s) that mediate protection against alcoholic liver disease are not well understood. Thus, the present study focused on identifying and characterizing genes regulated by PPAR␤/␦ and their roles in alcoholic liver disease. that were differentially regulated by ethanol (see NCBI Gene Expression Omnibus database accession number GSE86002). The gene expression profile was markedly different between genotype with distinct changes in gene expression noted in both Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice (Fig. 1A). Of particular note was the PPAR␤/␦-dependent increase in hepatic Cyp2b10 expression in Ppar␤/␦ ϩ/ϩ but not Ppar␤/␦ Ϫ/Ϫ mice (Fig. 1A). This was of interest because previous studies demonstrated a similar PPAR␤/␦-dependent effect on the hepatic expression of CYP2B10 in mice exposed to carbon tetrachloride (10). Overall, ethanol exposure significantly altered expression of 358 genes or 146 genes, in livers of Ppar␤/␦ ϩ/ϩ or Ppar␤/␦ Ϫ/Ϫ mice, respectively (Fig. 1B), with very little overlap between genotypes (76 gene products).
Ethanol-induced Cyp2b10 Expression Is Independent of CAR Signaling-Previous work revealed that ligand activation of CAR caused up-regulation of CYP2B10 in both Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice (10). This suggested that CAR activity was not influenced by PPAR␤/␦. However, whether PPAR␤/␦ modulated the ability of CAR to translocate and activate transcription of Cyp2b10 was not examined in the former study. Although expression of Cyp3a11 mRNA in the liver of Ppar␤/ ␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice fed ethanol was increased compared with controls, the expression of Cyp2b10 mRNA was increased in the liver of Ppar␤/␦ ϩ/ϩ mice but not in Ppar␤/ ␦ Ϫ/Ϫ mice fed ethanol compared with controls ( Fig. 2A). Because Cyp2b10 and Cyp3a11 are known target genes of CAR (12), but activating CAR with a ligand is not influenced by PPAR␤/␦ (10), the translocation of CAR was determined by quantitative Western blotting analysis. Interestingly, ethanol treatment did not influence the relative nuclear to cytosolic ratio of CAR expression compared with controls in either genotype (Fig. 2B). LAMIN and LDH were used as positive controls for nuclear and cytosolic enrichment.
Ethanol-induced Hepatic Cyp2b10 Expression Is Regulated by PPAR␤/␦-dependent Modulation of SP1 Activity-Because the previous results indicate that ethanol-induced Cyp2b10 expression was not mediated by CAR activation, functional mapping of the 5Ј upstream region of the Cyp2b10 gene was performed to identify important regulators that may influence ethanol-induced Cyp2b10 expression. Reporter gene assays revealed that the Ϫ1 to Ϫ500 region of the Cyp2b10 promoter contained critical cis-regulatory elements that were responsive to ethanol exposure (Fig. 4A). The putative trans-acting factors in this region included the octamer-binding transcription factor 1 (OCT1), the CCAAT/enhancer binding protein ␤ (C/EBP␤), and specificity protein 1 (SP1) (Fig. 4B). Mutation of the SP1-binding site in the Cyp2b10 promoter caused decreased luciferase activity following ethanol exposure compared with controls (Fig. 4C). However, this effect was not observed using either mutant OCT1 or mutant C/EBP␤ constructs, indicating that these transcription factors are not required for ethanol-induced Cyp2b10 expression (Fig. 4C). A ChIP assay further confirmed that SP1 occupancy on the Cyp2b10 promoter was higher in ethanol-treated Ppar␤/␦ ϩ/ϩ mouse hepatocytes compared with control (Fig. 4D). By contrast, SP1 occupancy of the Cyp2b10 promoter was not signifi- cantly different in Ppar␤/␦ Ϫ/Ϫ mouse hepatocytes treated with or without ethanol (Fig. 4D).
Paracrine Signaling from Intact Kupffer Cells Is Required for PPAR␤/␦-dependent Cyp2b10 Expression in Hepatocytes-To determine the specific cell types that mediate attenuation of ethanol-induced Cyp2b10 expression in liver, co-cultures of primary hepatocytes, primary non-parenchymal cells or Kupffer cells alone isolated from Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice were examined. In the presence or absence of ethanol treatment, ligand activation of PPAR␤/␦ increased Angptl4 mRNA expression in Ppar␤/␦ ϩ/ϩ hepatocytes but not in Ppar␤/␦ Ϫ/Ϫ hepatocytes (Fig. 6A). Surprisingly, ligand activation of PPAR␤/␦ did not suppress ethanol-induced Cyp2b10 mRNA expression in Ppar␤/␦ ϩ/ϩ hepatocytes (Fig. 6B), inconsistent with the results observed in vivo (Fig. 5A). Co-culturing primary hepatocytes with non-parenchymal cells revealed that the decrease in ethanol-induced Cyp2b10 mRNA expression following ligand activation of PPAR␤/␦ was restored and that this rescue was effective when hepatocytes were co-cultured solely with Kupffer cells (Fig. 6, C and D). These changes in Cyp2b10 mRNA expression were not observed in any groups of Ppar␤/␦ Ϫ/Ϫ co-cultures (Fig. 6, C and D).

Discussion
The metabolism of ethanol is mediated by enzymes including alcohol dehydrogenase and cytochrome P450s (CYP2E1 and CYP2B) (13). Oxidation of ethanol produces highly reactive oxygen species, which can cause liver damage (14,15). Variation in the expression of these enzymes influences the sensitivity and the adaption to ethanol consumption (13). PPAR␤/␦ can protect against chemically induced liver injury through multiple mechanisms including inhibition of steatosis, inhibition of NF-kB-dependent signaling, and inhibition of inflammation (10,11,16,17). A previous study revealed that exposure to carbon tetrachloride (CCl 4 ) caused an increase in hepatic Cyp2b10 expression, and this induction was mediated by PPAR␤/␦, which was not due to differences in the relative expression of CAR (10), similar to the results observed in the present study. The present study also showed that ethanol induced Cyp2b10 expression through a mechanism that FIGURE 4. PPAR␤/␦ regulates ethanol-induced Cyp2b10 expression by modulating SP1 activity. A, six luciferase reporter constructs (P3000, P2500, P2000, P1400, P1000, and P570) containing various fragments of 5Ј upstream region of the mouse Cyp2b10 gene were generated (left panel). Primary hepatocytes were isolated from adult male Ppar␤/␦ ϩ/ϩ mice. The cells were co-transfected with luciferase reporter constructs and pCMV-␤-gal plasmid and treated with or without ethanol. Relative luciferase activity was determined and normalized to control group (right panel). B, putative cis-regulatory elements in the Ϫ1 to Ϫ500 bp of 5Јupstream region of Cyp2b10 gene. C, site-directed mutagenesis was performed to generate luciferase reporter constructs carrying mutations in OCT1, SP1, or C/EBP␤ binding sites of the Cyp2b10 5Јupstream region. Primary hepatocytes were transfected with mutant luciferase reporter constructs and treated with or without ethanol. Relative luciferase activity was determined and normalized to control (Con). D, ChIP-qPCR showing PPAR␤/␦-dependent increased SP1 occupancy on the Cyp2b10 promoter in response to ethanol exposure in primary hepatocytes. The values were corrected for background using rabbit IgG controls and represent the means Ϯ S.E. *, significantly different from controls (p Յ 0.05).

Novel PPAR␤/␦-dependent Regulation of Hepatic CYP2B10
NOVEMBER 25, 2016 • VOLUME 291 • NUMBER 48 required PPAR␤/␦. Expression of Cyp2b10 is thought to be mediated by CAR activation (18). However, the present study clearly showed that ethanol treatment did not influence the relative expression of CAR as also observed in a previous study (10) or enhance the nuclear translocation of CAR following treatment with ethanol. Because the translocation of CAR to the nucleus is required to initiate target gene expression (19 -21), this illustrates a unique finding from the present study because the results indicate that PPAR␤/␦, rather than CAR, is required for up-regulation of Cyp2b10 in response to ethanol. This is similar to the PPAR␤/␦-dependent up-regulation of hepatic Cyp2b10 observed following treatment with CCl 4 , and the fact that ligand activation of CAR caused up-regulation of Cyp2b10 in both Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mouse liver (10).
An alternative pathway that could be influenced by PPAR␤/␦ that impacts CAR activity is the relative expression of PGC1␣. It is known that PPAR␤/␦ regulates PGC1␣ expression (22) and that CAR requires PGC1␣ as a co-activator to remodel chromatin of target genes (23). The present study demonstrated that nuclear PGC1␣ level was increased by ethanol treatment, and this increased expression of PGC1␣ requires PPAR␤/␦. Indeed, PPAR␤/␦-dependent expression of PGC1␣ is required for the increase in ethanol-induced Cyp2b10 expression because knocking down PGC1␣ prevented this effect in wild-type hepatocytes. This is the first study to demonstrate the essential role of PPAR␤/␦-dependent expression of PGC1␣ in regulating Cyp2b10 expression in response to ethanol exposure. These findings suggest that one mechanism that may mediate the PPAR␤/␦-dependent expression of Cyp2b10 is via modulation of PGC1␣, which is required for CAR activation. The present studies strongly support this notion, but further work is needed to confirm this hypothesis. Interestingly, the basal level of Cyp2b10 mRNA expression was not altered by knocking down PGC1␣ compared with controls, suggesting that PGC1␣ is not critical for regulating constitutive Cyp2b10 expression.
In addition to CAR, the transcription factor SP1 also regulates Cyp2b1 expression in rat hepatocytes (24). The present study revealed that ethanol-induced increased promoter occupancy of SP1 on the Cyp2b10 gene, and this effect required the expression of PPAR␤/␦. Examination of the microarray data did not indicate an increase in the expression of Sp1 mRNA. However, PGC1␣ can also increase SP1-mediated gene expression by either increasing SP1 expression or enhancing the recruitment of SP1 to the transcription complex (25,26). Thus, it is possible that the PPAR␤/␦-dependent expression of PGC1␣ induced by ethanol may also indirectly regulate Cyp2b10 expression by influencing the occupancy of SP1 on the Cyp2b10 promoter.
Because expression of CYPs is known to be involved in alcoholic liver disease in part because of the generation of metabolites such as reactive oxygen species, the finding that ligand activation of PPAR␤/␦ significantly suppressed ethanol-induced hepatic CYP2B10 expression suggests a protective role of PPAR␤/␦ in alcoholic liver disease. Indeed, recent studies have shown that PPAR␤/␦ protects against alcoholic liver disease by inhibiting steatosis, amino acid metabolism, and pyridoxal phosphate activity (8). Similarly, ligand activation of PPAR␤/␦ restores insulin sensitivity and also protects against ethanol-induced liver injury (17). However, ligand activation of PPAR␤/␦ did not prevent ethanol-induced Cyp2b10 mRNA expression in primary hepatocytes alone. By contrast, ligand activation of PPAR␤/␦ repression of Cyp2b10 expression induced by ethanol is restored when primary hepatocytes are co-cultured with Kupffer cells. This suggests that PPAR␤/␦ in the Kupffer cell may function differently than in the hepatocyte. For example, PPAR␤/␦ can modulate gene expression by interacting with other transcription factors and/or by directly regulating target gene expression (reviewed in Ref. 27). Indeed, PPAR␤/␦ expression in Kupffer cells can modulate effects observed during inflammatory insults in the liver that impact the hepatocyte (28). Further studies are needed to delineate the specific activities of PPAR␤/␦ in Kupffer cells and hepatocytes that underlie these differences. To date, this is the first demonstration that Kupffer cell activity is required for a protective effect in hepatocytes by facilitating repression of CYP2B10, whose expression is dependent on PPAR␤/␦. This suggests that paracrine signaling between the Kupffer cell and hepatocyte may be important for preventing alcoholic liver disease through a PPAR␤/␦-dependent mechanism.
It should be noted that there are advantages and disadvantages of using primary cells or co-cultures to study cell-cell interactions (29 -31). For example, global transcriptional regulation is not always consistent in human primary hepatocyte cultures from different individuals (32). This suggests that there can be interindividual variability in response to chemical exposures and illustrates the need to define consistent culture condition when using primary cell cultures. The co-culture of hepatocytes with non-parenchymal cells has been shown to respond differently to chemical exposures, such as exhibiting greater drug metabolism capabilities and drug-induced inflammatory responses, compared with hepatocyte culture alone (33). Although this type of co-culture is designed to model the environmental condition in vivo, challenges remain required to enhance the in vivo-like characteristics of in vitro culture systems. A recent study revealed a unique and stable model because multicell cultures of parenchymal and non-parenchymal cells retain the functional ability of liver cells in response to chemical exposures (34) and may be suitable for studies similar to those performed in the present experiments.
The facts that PPAR␤/␦ protects against alcoholic liver disease (8,17) and that this protection could be related to the repression of Cyp2b10 following exposure to ethanol are of interest. Therefore, further studies are needed to determine whether this PPAR␤/␦-dependent regulation of Cyp2b10 can be used to identify novel markers of alcoholic liver progression and possibly be suitable for targeting for the prevention and/or treatment of this disease.

Experimental Procedures
Animals and Experimental Protocols-Animal usage was approved by the Pennsylvania State University Institutional Animal Care and Use Committee. Wild-type (Ppar␤/␦ ϩ/ϩ ) and Ppar␤/␦-null (Ppar␤/␦ Ϫ/Ϫ ) mice (35) on a C57BL/6 genetic background were housed in a vivarium as previously described (36). Two cohorts of mice were used for the study. For the first cohort, age-matched (8 -10 weeks) male Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice were fed ad libitum daily with either liquid control or ethanol diets (Dyets Inc., Bethlehem, PA) for 16 weeks (n ϭ 10 per group). The ethanol diet contained 4% (v/v) ethanol and was prepared as previously described (8). Samples from the first cohort of experimental mice were used for microarray analyses and hepatic gene/protein expression in response to ethanol treatment.
For the second cohort of mice, age-matched (10 -12 weeks) male Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice were fed ad libitum with liquid control or ethanol diet for 16 weeks (n ϭ 10/group). Mice fed with control or ethanol diet were also given a pellet made with bacon-flavored Transgenic Dough Diet (Bioserv, Inc., Prospect, CT) mixed with vehicle control (0.02% dimethyl sulfoxide) or GW0742 (5 mg/kg/day) (n ϭ 5/group). Pair-fed mice (n ϭ 5/genotype) were included to control for potential difference in average food intake between control and ethanol diet. However, no differences in food intake were noted, so this group was not used for all analyses. Samples from the second cohort of mice were used for analysis of CYP2B10 expression in livers.
Microarray Analysis-Total RNA was isolated as previously described (37). Purified RNA was assessed by GeneChip Mouse Gene 2.0 ST array (Affymetrix, Santa Clara, CA) following the manufacturer's recommended procedures. The robust multichip average approach was applied to normalize microarray data as previously described (37). The p value Ͻ 0.05 and a fold change of 1.5 in the intensity of signals were used to identify genes that were significantly regulated by ethanol treatment. The data were uploaded to the NCBI Gene Expression Omnibus database with the accession number GSE86002.
Quantitative Real Time Polymerase Chain Reaction (qPCR)-Expression of mouse Cyp2b10, Cyp3a11, and Angptl4 in response to ethanol treatment was determined by qPCR analysis as previously described (38). Briefly, total RNA was isolated using RiboZol RNA extraction reagent (AMRESCO, Solon, OH) following the manufacturer's recommended procedures. cDNA was synthesized using 1.25 g of total RNA as template mixed with Moloney MLV reverse transcriptase and random primers (Promega, Madison, WI). DNA amplification was carried out in 25-l volumes containing SYBR Green PCR Supermix (Quanta Biosciences, Gaithersburg, MD) using the iCycler iQ5 PCR thermal cycler (Bio-Rad) with 45-55 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s. The sequences of primers used to detect mRNA were Cyp2b10 (NM_009999.4): forward, 5Ј-TTCTGCGCATGGAGAAGGAGAAGT-3Ј and reverse, 5Ј-TGAGCATGAGCAGGAAGCCATAGT-3Ј; Cyp3a11 (NM_007818.3): forward, 5Ј-GACAAACAAGCAGGGATG-GAC-3Ј and reverse, 5Ј-CCAAGCTGATTGCTAGGAGCA-3Ј; and Angptl4 (NM_020581.2): forward, 5Ј-TTCTCGCC-TACCAGAGAAGTTGGG-3Ј and reverse, 5Ј-CATC-CACAGCACCTACAACAGCAC-3Ј. The expression of glyceraldehyde-3-phosphate dehydrogenase (Gapdh; BC083149) was quantified as an internal control using the forward and reverse primers: 5Ј-GGTGGAGCCAAAAGGGTCAT-3Ј and 5Ј-GGTTCACACCCATCACAAACAT-3Ј. The primers were designed that spanned exon-exon junctions, which prevents genomic DNA amplification. Each assay included a standard curve and a non-template control that were performed in trip-licate. Four to six representative tissue samples/treatment group were randomly chosen for each analysis. Relative mRNA levels were normalized to Gapdh because there was no difference in expression between groups as determined by ANOVA and post hoc testing (p Ն 0.05).
Western Blotting Analysis-Hepatic microsomal protein was extracted from Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mouse livers as previously described (39). Nuclear and cytosolic fractions were isolated from Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mouse livers using NE-PER nuclear and cytoplasmic extraction reagents (Pierce) following the manufacturer's recommended procedures. Quantitative Western blotting analysis using radioactive detection techniques, the gold standard for quantifying protein expression, was performed as previously described (40). The following primary antibodies were used: anti-LAMIN A/C (sc7293, lot no. L0909, mouse monoclonal; Santa Cruz Biotechnologies, Santa Cruz, CA); anti-ACTIN (sc47778, lot no. K1414, mouse monoclonal, Santa Cruz Biotechnologies); anti-LDH (200-1173, lot no. 11538, goat polyclonal; Rockland Immunochemicals, Inc., Limerick, PA); anti-CYP2B1 (clone 2-66-3 (41), mouse monoclonal, produced at the National Cancer Institute as previously described (41,42)); anti-PGC1␣ (ab51365, clone PPARAH6, mouse monoclonal; Abcam, Cambridge, MA), and anti-CAR (PP-N4111-00, lot no. A-1, mouse monoclonal; R & D Systems, Minneapolis, MN). Specificity of the antibodies was confirmed through different methods including: 1) confirming relative mobility for each protein (69/62 kDa for LAMIN A/C, respectively; 43 kDa for ACTIN; 36 kDa for LDH; 57 kDa for CYP2B10; 89 kDa for PGC1␣; or 40 kDa for CAR); 2) confirming relative mobility and subcellular localization in the nucleus versus cytosol (CAR, LDH, or LAMIN A/C); 3) confirming relative mobility and knockdown of protein expression (PGC1␣); or 4) confirming relative mobility and lack of inducibility in knock-out mice (CYP2B10). The relative expression level of each microsomal or cytosolic protein was normalized to the value of LDH or ACTIN. The relative expression level of each nuclear protein was normalized to the value of LAMIN. Statistical analyses of hybridization signals for LDH, ACTIN, and LAMIN revealed no significant differences between treatment groups as assessed by ANOVA and post hoc testing (p Ն 0.05). A minimum of three mice per group was analyzed.
siRNA Knockdown of Mouse PGC1␣ in Primary Hepatocytes-Primary hepatocytes (2 ϫ 10 5 ) were seeded in 12-well collagen-coated culture plates and transiently transfected with 10 M non-targeting scrambled siRNA or mouse Pgc1␣ siRNA (Invitrogen) using Lipofectamine TM 2000 following the man-ufacturer's recommended procedures. The cells were treated with or without ethanol (Յ100 mM) for 12 h. Quantitative Western blotting analysis was performed as described above to confirm nuclear PGC1␣ expression.
Luciferase Assay-To identify the cis-regulatory element responsible for ethanol-induced Cyp2b10 expression in hepatocytes, six reporter gene constructs driven by 5Ј upstream region of Cyp2b10 gene with serial deletion were generated by PCR. Primers were designed to contain overhanging sites for restriction enzymes, which allowed for directional cloning into the multiple cloning site of pGL3-basic luciferase reporter vector (Promega). The composition of each construct was confirmed by restriction endonuclease digestion and DNA sequencing.
Primary hepatocytes were transiently co-transfected with 0.5 g of pCMV-␤-galactosidase plasmid (Promega) and 3.5 g of either Cyp2b10 promoter-luciferase reporter constructs or the control plasmid (pGL3-basic vector) using Lipofectamine LTX reagent (Invitrogen) following the manufacturer's recommended procedures. The cells were treated with or without ethanol (Յ100 mM) for 12 h, and cell lysates were prepared in passive lysis buffer (Promega). Luciferase and ␤-galactosidase activities were measured using Luciferase and Beta-Glo assay systems, respectively, following the manufacturer's recommended procedures. Relative luciferase activity was normalized to ␤-galactosidase activity.
Site-directed Mutagenesis-DpnI-mediated site-directed mutagenesis was performed to generate mutation in Cyp2b10 promoter-luciferase reporter construct as previously described (44). The sequence-specific primers were overlapping with and flanking the transcription factor binding sites for identified trans-acting factors. The successful mutagenesis was confirmed by DNA sequencing. Primary hepatocytes were transiently transfected with mutant luciferase reporter constructs followed by ethanol treatment. Relative luciferase activity was determined as described above.
ChIP-qPCR-To confirm the occupancy of SP1 on the Cyp2b10 promoter, ChIP-qPCR was performed to quantify relative promoter occupancy following ethanol exposure in primary hepatocytes from Ppar␤/␦ ϩ/ϩ and Ppar␤/␦ Ϫ/Ϫ mice as previously described (38). The following primary antibody was used: anti-SP1 (sc59, lot no. 0915, rabbit polyclonal; Santa Cruz Biotechnologies). Rabbit IgG was used a negative control.
Co-cultures of Primary Hepatocytes with Primary Kupffer Cells or Non-parenchymal Cells-Primary hepatocytes were seeded in 12-well collagen-coated culture plates. Primary Kupffer cells or non-parenchymal cells were seeded in Transwell inserts (Corning Inc., Corning, NY). Primary hepatocyte-Kupffer cell co-cultures (5:1 ratio) or primary hepatocyte-nonparenchymal cell co-cultures (1:1 ratio) were pretreated with or without GW0742 (1 M) for 6 h and then treated with or without ethanol (Յ100 mM) for 12 h.
Statistical Analysis-The data were subjected to either Student's t test or a parametric one-way ANOVA followed by Tukey test for post hoc comparisons (Prism 5.0; GraphPad Software Inc., La Jolla, CA). The correlation between fold changes of Cyp2b10 mRNA expression and nuclear PGC1␣ protein expression was determined by Pearson correlation method with a two-tailed p value (Prism 5.0).
Author Contributions-T. K., P.-L. Y., M. G., G. H. P., A. J. F., F. J. G., and J. M. P. conceived and coordinated the study and wrote the paper. T. K., G. B., P.-L. Y., and I. A. M. performed and analyzed experiments. All authors reviewed the results and approved the final version of the manuscript.