1521-009X/44/10/1643–1652$25.00 http://dx.doi.org/10.1124/dmd.116.070243DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 44:1643–1652, October 2016Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

Methylation of the Constitutive Androstane Receptor Is Involved inthe Suppression of CYP2C19 in Hepatitis B Virus–Associated

Hepatocellular Carcinoma s

Xiaojing Tang, Lele Ge, Zhongjian Chen, Sisi Kong, Wenhui Liu, Yingchun Xu, Su Zeng,and Shuqing Chen

Institute of Drug Metabolism and Pharmaceutical Analysis, Zhejiang Province Key Laboratory of Anti-Cancer Drug Research,College of Pharmaceutical Sciences, Zhejiang University, Hangzhou, Zhejiang, China (X.T., W.L., Y.X., S.Z., S.C.); Department ofPharmacy, Sir Run Run Shaw Hospital, College of Medicine, Zhejiang University, Hangzhou, China (L.G.); and Department of

Pharmacy, Zhejiang Cancer Hospital, Hangzhou, China (Z.C., S.K.)

Received February 27, 2016; accepted July 18, 2016

ABSTRACT

Hepatocellular carcinoma (HCC), one of the most dangerous malig-nancies with an increasing incidence and a high mortality rate,represents a major international health problem. HCC progression isknown to involve genome-wide alteration of epigenetic modifica-tions, leading to aberrant gene expression patterns. The activity ofCYP2C19, an important member of the cytochromeP450 superfamily,was reported to be compromised in HCC, but the underlyingmechanism remains unclear. To understand whether epigeneticmodification in HCC is associated with a change in CYP2C19 activity,we evaluated the expression levels of CYP2C19 and its transcriptionfactors by quantitative real-time polymerase chain reaction using

mRNA extracted from both primary hepatocytes and paired tumorversus nontumor liver tissues of patients infected with hepatitis Bvirus (HBV). DNA methylation was examined by bisulfite sequencingand methylation-specific polymerase chain reaction. Our resultsindicated that CYP2C19 could be regulated by e-box methylation ofthe constitutive androstane receptor (CAR). Decreased CYP2C19expression in tumorous tissues of HBV-infected patients with HCCwas highly correlated with suppressed expression and promoterhypermethylation of CAR. Our study demonstrates that aberrant CARmethylation is involved in CYP2C19 regulation in HBV-related HCCand may play a role in liver tumorigenesis.

Introduction

Although tremendous efforts have been made in the fight againsthepatocellular carcinoma (HCC), HCC is the fifth most prevalent cancerworldwide and is currently the third leading cause of cancer mortality(Jemal et al., 2011). Hepatitis B virus (HBV) infection, hepatitis C virusinfection, and alcohol abuse are the main causes of HCC, and HBVinfection accounts for 60% of HCC in the Chinese population (Yanget al., 2014). A recent World Health Organization report estimates that90million people in China (almost 7% of the population) are chronicallyinfected with HBV (Liu et al., 2016).HCC progression was traditionally considered a gradual multistep

process of normal cells evolving into tumor cells caused by mutationsthat led to altered hepatocellular phenotypes (Thorgeirsson andGrisham,2002; Lessel et al., 2014). However, evidence accumulated in the pastfew decades suggests that altered epigenetic modifications also play apivotal role in HCC development (Nishida and Goel, 2011). DNAmethylation, which refers to the addition of amethyl group to the cytosinewithin a CpG dinucleotide, is the most intensively studied epigenetic

mechanism in mammals. The cancerous tissues in HCC are reported topossess a unique methylation landscape that differs from adjacent tissues.In HCC, the methylome feature is roughly characterized by globalhypomethylation and gene-specific DNA hypermethylation, whichcontribute to genomic instability and inactivation of tumor-suppressorgenes, respectively (Herceg and Paliwal, 2011). Since advanced HCC ishighly lethal and incurable, there is an urgent need to understand thetumorigenesis process and to search for new diagnostic biomarkers andnovel therapeutic targets.CYP2C19 is one of the most important members of the cytochrome

P450 (P450) superfamily and is responsible for the metabolism of avariety of clinically important drugs such as mephenytoin, diazepam,and some barbitals (Griskevicius et al., 2003). The genetic polymor-phisms of CYP2C19 have been well studied. Individuals can beclassified as extensive metabolizers, intermediate metabolizers, poormetabolizers, and ultra-rapid metabolizers based on the catalyticcapacity of the enzyme, which largely depends on the different allelesindividuals carry (Uppugunduri et al., 2012). The regulatory mechanismof CYP2C19 is currently characterized by several nuclear receptors andtranscription factors, including the constitutive androstane receptor(CAR), pregnane X receptor, glucocorticoid receptor a, and hepatocytenuclear factor-3g. Of these, CAR seems to play a central role and hasa strong association with the basal expression level of CYP2C19(Wortham et al., 2007). It has been predicted that epigenetics mightalso play a role in CYP2C19 regulation, and some CpG sites exist in the

This research was supported by the National Natural Science Foundation ofChina [Grant 81273578] and the Ministry of Science and Technology of ChinaNatural Major Projects for Science and Technology Development [Grant2012ZX09506001-004].

dx.doi.org/10.1124/dmd.116.070243.s This article has supplemental material available at dmd.aspetjournals.org.

ABBREVIATIONS: 5-aza-dC, 5-aza-29-deoxycytidine; BSP, bisulfite sequencing polymerase chain reaction; CAR, constitutive androstane receptor;HCC, hepatocellular carcinoma; HEK293, human embryonic kidney 293; HBV, hepatitis B virus; HK2, human kidkey-2; P450, cytochrome P450.

1643

http://dmd.aspetjournals.org/content/suppl/2016/07/20/dmd.116.070243.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

59-flanking region of the CYP2C19 gene but their functional importanceremains to be further elucidated (Ingelman-Sundberg et al., 2007).CYP2C19 expression is reported to be decreased in advanced cancer,

including breast cancer, lung cancer, and HCC, which is also accompa-nied by compromised enzymatic activity (Tsunedomi et al., 2005; Helsbyet al., 2008). Despite researchers’ extensive efforts to investigate theunderlyingmechanism over the past years, altered expression of CYP2C19in HCC does not appear to be fully accounted for by known polymor-phisms, inflammatory factors, or growth hormones (Helsby et al., 2008).This study aimed to investigate the possible epigenetic mechanism ofdecreased CYP2C19 expression in HBV-infected patients with HCC.Since CYP2C19 is also important in the disposition of a number ofchemotherapeutic agents, including cyclophosphamide, thalidomide, andbortezomib, elucidating the epigeneticmodulation of CYP2C19might alsohelp improve understanding and prediction of chemotherapy outcomes forpatients with HCC.

Materials and Methods

Cell Culture and Tissue Samples. HepG2 and human embryonic kidney293 (HEK293) cells were grown in Dulbecco’s modified Eagle’s medium (LifeTechnologies, Carlsbad, CA), supplemented with 10% fetal bovine serum (LifeTechnologies) and 1% penicillin-streptomycin (Life Technologies), at 37�C with5% CO2. Human kidney-2 (HK2) cells were grown in Dulbecco’s modifiedEagle’s medium/F-12 medium (Life Technologies) supplemented with 18% fetalbovine serum at 37�C with 5% CO. Cryopreserved primary hepatocytes(Research Institute for Liver Disease, Shanghai, China) from one male donorand one female donor (Supplemental Table 1) were plated in InVitroGRO CPmedium (BioreclamationIVT, Baltimore, MD) at 37�C with 5% CO2 beforedrug incubation. Two replicates were conducted for each donor.

Thirty pairs of tumorous and adjacent liver tissues were obtained from freshsurgical specimens of HBV-infected patients with HCC (HBV DNA .1000 copies/ml or positive hepatitis B surface antigen and hepatitis B core antibodytest results) and were then stored in liquid nitrogen before use. The selected patientsdid not receive any radiotherapy or chemotherapy previously. Tissues wereobtained from Zhejiang Cancer Hospital (Hangzhou, China). Information on liverspecies is provided in Supplemental Table 2. The research protocol was carried outin accordance with the Declaration of Helsinki and was approved by the EthicsCommittee of Zhejiang Cancer Hospital.

Treatment of Primary Hepatocytes with 5-Aza-29-Deoxycytidine. Primaryhepatocytes were seeded at a density of 5 � 105 cells/ml in collagen-coated, six-well plates (Life Technologies) with InVitroGRO CP medium. The medium waschanged the next day to remove the unattached cells. After 24 hours of plating,

drug incubation started with 1 mM and 2 mM 5-aza-29-deoxycytidine (5-aza-dC;Sigma-Aldrich, St. Louis, MO) (Christman, 2002) in InVitroGRO HI mediumsupplemented with 25 ng/ml epidermal growth factor (Sigma-Aldrich) to stimulateproliferation. The control group was treated with 0.1% dimethylsulfoxide. Themedium was changed every other day. Cells were harvested after 5 days treatmentfor DNA and RNA isolation.

Semiquantitative Real-Time Polymerase Chain Reaction. mRNA wasextracted from tissue samples and cultured cells using the RNeasy Mini Kit(Qiagen, Valencia, CA). Five-hundred nanograms ofmRNAwas used in a reverse-transcribed polymerase chain reaction (PCR) performed with the PrimeScript RTReagent Kit (Takara, Tokyo, Japan). Quantitative real-time PCR was thenperformed using SYBR Premix Ex Taq II (Takara) and the StepOnePlus system(Applied Biosystems, Foster City, CA). Primers for quantitative real-time PCR arelisted in Table 1. Relative quantitation of gene expression was calculated with the22DDCt method as described previously (Schmittgen and Livak, 2008).

DNA Methylation Analysis. Genomic DNA from tissue samples or cells wasisolated using the DNeasy Blood and Tissue Kit (Qiagen). DNA was consideredqualified with a ratio of absorbance at 260 nm and 280 nm (A260/A280) between1.7 and 1.9. Samples were stored in elution buffer at 280�C before use.

Sodium bisulfite modification of DNA samples was conducted using an EZDNA Methylation-Gold Kit (Zymo, Orange, CA). The location of all CpGsequences examined in this study is shown in Fig. 1. PCR amplification was donewith rTaq polymerase (Takara) in a 50-ml reaction mixture with 2 ml modifiedDNA as the template. PCR amplification started with 10 cycles of touchdownPCR of 30 seconds at 94�C, 30 seconds at 60�C to 55�C (20.5�C per cycle), andthen 45 seconds at 72�C and 30 cycles of 30 seconds at 94�C, 30 seconds at 55�C,and 45 seconds at 72�C, followed by an extension of 10 minutes at 72�C.Amplicons were then purified and subcloned into pMD18-T vector (Takara)according to the manufacturer’s instructions. Ligation products were transformedinto Escherichia coli DH5a cells and 10–15 random colonies with recombinantplasmids were subjected to Sanger sequencing (Sangon, Shanghai, China).

Methylation-specific PCR was also conducted using the modified DNA as atemplate. The following touchdown PCR program was applied: 10 cycles oftouchdown PCR of 30 seconds at 94�C, 30 seconds at 58�C to 53�C (20.5�C percycle), 20 seconds at 72�C and 25 cycles of 30 seconds at 94�C, 30 seconds at53�C, and 20 seconds at 72�C, followed by a final extension of 10 minutes at72�C. Ten-microliter PCR products were shown on 2% agarose gel containingethidium bromide. Primers for methylation analysis are listed in Table 2.

Plasmid Construction. pCpGL-CYP2C19 (21977 to +64), pCpGL-CAR-CGI (21446 to +1971), pCpGL-CAR-2E (+175 to +1971), and pCpGL-CAR-1E(+384 to +1971) were constructed with a CpG-free luciferase reporter vectornamed pCpGL-basic (a kind gift from Prof. Michael Rehli, University Hospital,Regensburg, Germany) (Fig. 2A) (Klug and Rehli, 2006). pCpGL-CYP2C19contains an approximately 2.0-kb sequence of the 59-flanking promoter region ofCYP2C19. pCpGL-CAR-CGI contains an approximately 3.5-kb sequence with a

TABLE 1

Primer sequences for quantitative real-time PCR

Primers were designed to amplify either all transcription variants or variants that predominate in the liver.

Gene Accession No. Forward Primer (59–39) Reverse Primer (59–39)

CYP2C19 NM_000769.2 GAAAAATTGAATGAAAACATCAGGATTG CGAGGGTTGTTGATGTCCATCCAR NM_005122.4 GCAGCTGTGGAAATCTGTCA CAGGTCGGTCAGGAGAGAAGGRa NM_000176.2 GGGGAAGAGGGAGATGGA GTGGTCAGAATGGGAGGCPXR NM_022002.2 CAAGCGGAAGAAAAGTGAACG CTGGTCCTCGATGGGCAAGTC

NM_033013.2NM_003889.3

GATA-4 NM_001308094.1 CTGGCCTGTCATCTCACTACG GGTCCGTGCAGGAATTTGAGGNM_001308093.1NM_002052.4

HNF-3g NM_004497.2 ATGCTGGGCTCAGTGAAGAT CAGGTAGCAGCCATTCTCAAHNF-4a NM_178849.2 GGTCAGGTGGGGTGGATGATATAATG TCTAGGTTAATAGGGAGGAAGGGAGG

NM_178850.2NM_001258355.1

GAPDH NM_002046.5 GAAGGTGAAGGTCGGAGTC CAAAGTTGTCATGGATGACCNM_001289745.1NM_001289746.1

GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GR, glucocorticoid receptor; HNF, hepatocyte nuclear factor; PXR, pregnane X receptor.

1644 Tang et al.

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

CpG island on the 59-flanking region and both e-box binding sites close to thetranscription start site and translation start site of CAR, respectively. In pCpGL-CAR-2E, the CpG island region was deleted from the sequence of pCpGL-CAR-CGI; in pCpGL-CAR-1E, another approximately 200-bp sequence that containsthe first e-box binding site was deleted. In addition, a full-length CAR codingDNA sequence amplified from mRNA (primers CDS-F/R) was ligated with the59-flanking region as well as the untranslated exon 1 and intron 1 (amplified byprimers Flanking-F/R) by overlap PCR. The approximately 3-kb ligated fragmentwas then inserted into a CpG-free expression vector, pCpGfree-mcs (Fig. 2B)(InvivoGen, San Diego, CA), to construct pCpGfree-CAR. Primers used forplasmid construction are listed in Table 3. All constructs were confirmed bySanger sequencing (Sangon).

In Vitro Methylation of Plasmid DNA. Plasmids were methylated usingM.SssI (New England Biolabs, Beverly, MA) according to the manufacturer’sinstructions. Two micrograms of plasmid DNA was incubated in a 40-ml volumewith 2 ml (4 U/ml) SssI and 160 mM S-adenosyl methionine at 37�C for 6 hours.The control group of plasmids also underwent the same incubation but in theabsence of SssI. The reactionwas stopped by heating at 65�C for 20minutes. Boththe methylated group and the control group of plasmid DNA were purified withthe Axygen PCR Clean-Up Kit (Axygen, Hangzhou, China).

Transient Transfection Assay. HepG2, HEK293, or HK2 cells were seededin 24-well plates at a density of 2.5 � 105 /ml. The transient transfection wascarried out using Lipofectamine2000 (Life Technologies) when cells reached 70%–

80% confluence. Reporter plasmids and expression plasmids were transfected at500 ng and 250 ng per well, respectively, with 50 ng pRL-TK per well as theinternal control. Medium was refreshed 6 hours after transfection. Luciferaseactivity was assessed using the Dual-Luciferase Reporter Assay System (Promega,Madison, WI) 48 hours after transfection.

Statistical Analysis. Quantitative PCR results are shown as 22DDCt usingglyceraldehyde 3-phosphate dehydrogenase as a housekeeping gene. The meth-ylation percentage of bisulfite sequencing polymerase chain reaction (BSP)analysis was calculated as the number of methylated cytosines divided by the totalnumber of cytosines on the CpG site in all amplicons. The methylation percentageof a sequence was determined as the average percentage of all CpG sites withinthis region. Comparisons of the methylation percentage and mRNA expressionbetween groups were analyzed by a paired t test using GraphPad Prism software(version 5.0; GraphPad Software Inc., San Diego, CA). The Pearson correlationwas used to determine the correlation between CYP2C19 and CAR expression.Differences or correlations were considered statistically significant at P , 0.05,P , 0.01, and P , 0.001.

Results

Increased CYP2C19 and CAR mRNA in Primary Hepatocytesafter 5-Aza-dC Treatment. To determine whether DNAmethylation isa potential mechanism of CYP2C19 regulation in the liver, a DNAmethyltransferase inhibitor (5-aza-dC) was applied to the primarycultures of humanhepatocytes.Quantitative real-timePCRanalysis revealedan increase in CYP2C19 and CAR expression in a dose-dependentmanner, whereas other transcription factors were not significantly affected(Fig. 3). To investigatewhether demethylation led to the elevated expressionlevel of CYP2C19 and CAR in primary hepatocytes, BSP was performedon the promoter regions as shown in Fig. 1. BSP results revealed amoderatemethylation frequency on the CpG-poor promoter of CYP2C19 (52.5%) aswell as the two e-box binding sites on CAR (50% and 47.5%). After

Fig. 1. Location of the sequences examined byBSP, methylation-specific PCR, and the re-porter gene assay. CpG sites are shown asvertical bars with dots. Transcription start sitesare shown as curved arrows and are marked asposition +1. E-box binding sites are indicatedby downward arrows. Double-arrow bars in-dicate sequences examined by BSP, and thecapped line indicates sequences examined bymethylation-specific PCR. Sequences in boxesindicate CpG islands identified by Methyl PrimerExpress Software (version 1.0). The sequencesanalyzed by the reporter gene assay are shown ashorizontal thick lines marked with starting andending positions. TSS, transcription start site.

TABLE 2

Primer sequences for BSP and MSP

Amplicon Gene ID Forward Primer (59–39) Reverse Primer (59–39)

Primers for BSPCYP2C19 segment 1 1557 GGTAGTATATTTTTGTTAGAGTTTAAAG AATCCACAAATATTTACTAATTATATACACYP2C19 segment 2 1557 GAGAATTGGAAATAATTTTATTAGG ACCACCACAATATATAAACCTTAACCYP2C19 segment 3 1557 GTTTTGTTTTGTTAAAATAAAGTTTTAGT AACCCAAATAATTCCAATACACCCAR segment 4 9970 GTAAATGAAGAAGTAGAAGATATAATAG AATAACTCATACCTATAATCCCAACACAR segment 5 9970 TAGTTATTGAGAGTAATTGGAGGTTATA CCTTAACTTCCCAAAATACTAAAATTACACAR segment 6 9970 TTAGAAGGGATAGAAAAGGGTTAAGG ACTATATCCATCAAACAAACATTCAAC

Primers for MSPCAR-MSP-M 9970 GTTGGGGGTTAGGTATAGTGGTTTACGA CTAAAATTACAAACATAAACTCCCGCGTCAR-MSP-U 9970 TTGGGGGTTAGGTATAGTGGTTTATGA TACTAAAATTACAAACATAAACTCCCACAT

MSP, methylation-specific polymerase chain reaction.

Epigenetic Study of CYP2C19 in Hepatocellular Carcinoma 1645

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

5-aza-dC treatment, the methylation frequency of these three regionsdecreased to 38.3%, 13.3%, and 42.5%, respectively (Fig. 4, A–C). TheCpG island on the CAR promoter, whichwas highlymethylated (92.5%),also showed decreased methylation frequency (72.5%) after exposure to

5-aza-dC (Fig. 4D). BSP results showed an overall decreased DNAmethylation level on the promoter of CYP2C19 and CAR after 5-aza-dCtreatment. These results suggest that DNA methylation is a regulatorymechanism of CYP2C19 in the liver and CAR might also be involved.

Fig. 2. Maps of the CpG-free vectors used in the reporter geneassay.

TABLE 3

Primer sequences for plasmid construction

Lower-case letters indicated the added restriction sites; underlined letters indicated the overlapping part of the primers for overlap PCR.

Amplicon Gene ID Forward Primer (59–39) Reverse Primer (59–39)

pCpGL-CYP2C19 1557 GGactagtATTATTCTAAAGAGAGCAAC CGggatccTTAAGACAACCGTGAGpCpGL-CAR-2E 9970 GGactagtGAAATCTGTTGGGGACCAAAGACTAA CGggatccCTGTGTCCATCAGACAAACATTCAGpCpGL-CAR-1E 9970 GGactagtCCTGTAATCCCAGCACTTTGGGAAGCC CGggatccCTGTGTCCATCAGACAAACATTCAGpCpGfree-CAR 9970 Flanking-F GGactcgtCTCTCTCTCTCTCTTCCCAGCTTG Flanking-R TACTGGCCATGACGTCACGTGTTGGG

NM_005122.4 CDS-F CGTGACGTCATGGCCAGTAGGGAAGATGAG CDS-R GAagatctCCTCAGCTGCAGATCTCCTGGA

CDS, coding DNA sequence.

1646 Tang et al.

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

Decreased mRNA Expression of CYP2C19 and CAR in TumorousTissues. Expression of CYP2C19 and CAR in tumorous and adjacenttissues is shown as grouped column scatter plots with means 6 S.E. inFig. 5A. The average expression of CYP2C19 in tumorous tissues (0.3960.20) was significantly decreased compared with the normal tissues(2.46 0.40,P, 0.001). In addition, the average expression of CAR alsodecreased significantly (0.40 6 0.085 versus 1.40 6 0.15, P , 0.001).

Alterations of the expression levels of CYP2C19 and CAR werecorrelated, with a Pearson correlation coefficient of 0.31 (R2 = 0.09,P , 0.05) (Fig. 5C).Elevated Methylation Level of CpG Sites on the CAR Promoter

in Tumorous Tissues. BSP analysis of segment 5 on CAR wasperformed using DNA from paired tumor and adjacent tissue samples.Representative BSP results are shown in Fig. 6A and the corresponding

Fig. 3. Expression profiles of CYP2C19 andrelated transcription factors in primary hepato-cytes after 5-aza-dC treatment. The relativeexpression of CYP2C19, CAR, GATA-4,HNF-3g, PXR, GRa, and HNF-4a is shown asmeans 6 S.E. *P , 0.05. The x-axis indicatesthe 5-aza-dC concentration. Data are from fourindependent experiments of two donors. GR,glucocorticoid receptor; HNF, hepatocyte nu-clear factor; PXR, pregnane X receptor.

Epigenetic Study of CYP2C19 in Hepatocellular Carcinoma 1647

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

relative mRNA expression is shown in Fig. 5B. Clearly, in BSP analysis,CpG 3, CpG 5, and CpG 6 in segment 5 around the transcription start siteof CAR had a higher methylated status in tumorous tissue comparedwith adjacent tissue. Among these three sites, CpG 3 is a potential e-boxbinding site, whereas CpG 5 and CpG 6 are two adjacent CpG sites121 bp downstream of CpG 3. In addition, methylation-specific PCRwithin segment 5 also revealed a higher methylation level in tumor-ous tissues (Fig. 6B). Interestingly, in the case of sample 246564,the expression levels of CAR and CYP2C19 were increased whilemethylation status was decreased in tumorous tissues (Fig. 6A). Nosignificant methylation difference was detected between tumorousand adjacent tissues in the other segments analyzed using BSP (datanot shown).In Vitro Methylation on the CAR Promoter Leads to Decreased

CAR-Mediated CYP2C19 Expression in HepG2 Cells. Results ofthe dual-luciferase reporter assay showed that the CpG-poor 2-kbpromoter of CYP2C19 was not affected by methylation (Fig. 7, A, C, D)in all tested cell lines, suggesting that these sporadic CpG sites are irrel-evant to the methylation mechanism. The unmethylated pCpGL-CAR-CGI was 2.5-fold higher in relative luciferase activity compared with themethylated group in HepG2 cells. This fold change remained after thedeletion of the CpG island, as shown by the unmethylated versusmethylated pCpGL-CAR-2E (2.2-fold), but was greatly reduced afterdeletion of the 200-bp fragment containing the e-box binding site near thetranscription start site (in the first intron). Our results indicate that this e-boxbinding site was essential in the methylation-regulating mechanism ofCAR, whereas the e-box close to the translation start site was not involved.However, no difference between the methylated and unmethylated groupswas found in HK2 or HEK293 cells (Fig. 7, C and D).To explore whether the methylation-mediated decrease of CAR

could exert an effect on CYP2C19 expression, either methylated orunmethylated pCpGfree-CARwas cotransfected with pCpGL-CYP2C19in HepG2 cells. A 2.3-fold higher luciferase activity was observedin the unmethylated pCpGfree-CAR group (Fig. 7B), which furthersupported the finding that CYP2C19 could be suppressed throughhypermethylation caused by a decrease in CAR expression.

Discussion

HBV infection remains a key public health issue in China according tothe World Health Organization’s 2014 World Cancer Report (Bernardet al., 2014). HBV-induced hepatocarcinogenesis includes proteomicdisruption, viral integrations, and aberrant epigenetic modifications(Ji et al., 2014). Our study demonstrates that DNA demethylation leadsto elevated CAR and CYP2C19 expression in primary hepatocytes. Thisfinding suggests that DNA methylation is a potential mechanism ofCAR and CYP2C19 regulation in the liver. This notion is furthersupported by suppressed expression of CAR and CYP2C19 in Chinesepatients with HBV-related HCC.Since CAR is a member of the nuclear receptor superfamily and is a

central modulator of oxidative and conjugative enzymes and transportersthat are involved in drug disposition and metabolism, it is conceivablethat the suppression of CAR plays an important role in the alteredexpression of CYP2C19 and possibly other drug metabolism targetinggenes. P450s participate in the activation and metabolism of a largenumber of chemotherapeutic agents such as docetaxel (Shou et al., 1998)and cyclophosphamide (Griskevicius et al., 2003). Thus, CAR suppres-sion mediated by hypermethylation may lead to disturbance of the drugmetabolizing system. Therefore, the methylation status of CAR mightserve as a biomarker of altered pharmacokinetics. Indeed, other studieshave confirmed that several CAR-regulating P450 genes were alsosuppressed along with decreased CAR expression in end-stage HCC

Fig. 4. DNA methylation of CpG sites in segments 1, 2, and 3 on the promoterregion of CYP2C19 (A), segment 5 near the CAR transcription start site (B),segment 6 near the CAR translation start site (C), and segment 4 on the promoterregion of CAR (D). CpG sites of the control group and the group treated with 1 mM5-aza-dC are shown as circles and triangles, respectively. Data are representative ofthe female donor.

1648 Tang et al.

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

(Chen et al., 2014), but no significant decrease in CAR and related P450genes was found in HBV cirrhosis samples. This finding suggests thatthe sharp decline in expression of these genes might occur when livercirrhosis evolves into carcinoma. However, our results only offered aPearson correlation coefficient of 0.31 between CYP2C19 and CAR

expression, in contrast with a stronger correlation (0.693) observed inhealthy liver samples (Wortham et al., 2007). Since downregulation ofmetabolism-related genes is demonstrated to be preferentially linkedwith HBV-related HCC rather than hepatitis C virus–related HCC(Okabe et al., 2001; Iizuka et al., 2002), it is possible that the virus itself

Fig. 5. Relative expression of CYP2C19 andCAR in 30 HCC tumorous tissues compared withmatched adjacent normal tissues. (A) Comparisonof CYP2C19 and CAR expression in matchedHCC tumorous tissues and adjacent normaltissues by the paired t test. Data are shown asscatter plots with means 6 S.E. ***P , 0.001.(B) Representative CYP2C19 and CAR expres-sion in matched HCC tumorous tissues andadjacent normal tissues (denoted as C and N,respectively). Data are shown as means. (C)Pearson correlations between mRNA expressionof CYP2C19 and CAR.

Epigenetic Study of CYP2C19 in Hepatocellular Carcinoma 1649

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

Fig. 6. Methylation analysis of segment 5 near the CAR transcription start site. (A) Comparison of methylation status of CpG sites within segment 5 betweenHCC tumorous tissues (circles) and adjacent normal tissues (triangles). (B) Methylation-specific PCR analysis of the e-box binding site within segment 5.DNA samples from tumorous and normal tissue are shown (denoted as C and N, respectively). Segment 5 amplified by PCR was used as either methylatedcontrol DNA (Me, M.SssI treated) or unmethylated control DNA (Un). Methylation-specific and unmethylation-specific primers are shown as M and U,respectively.

1650 Tang et al.

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

might have a direct effect on the expression of certain genes. Further-more, many novel mechanisms have been discovered in recent years.The participation of hsa-miR-29a-3p in CYP2C19 regulation in livercells is one example of an epigenetic mechanism (Yu et al., 2015).MicroRNA may also participate in disrupting the correlation be-tween CAR mRNA and protein levels and may thus lead to the poorcorrelation between CAR and CYP2C19 mRNA levels. Further studyis warranted to build a comprehensive map of how CYP2C19 isregulated in HCC.Since there was a 2-fold decrease in luciferase activity after deletion of

the 200-bp sequence containing the e-box at the first intron as indicatedin the reporter gene assay, this sequence may act as an enhancer elementof CAR. This hypothesis warrants further investigation. We performedthe same in vitro methylation and reporter gene assay in three cell lines(HepG2, HEK293, and HK2). Interestingly, only HepG2 showeddecreased luciferase activity in methylated pCpGL-CAR-2E. No differ-ence between methylated and unmethylated pCpGL-CAR-2E was de-tected in either HEK293 or HK2 cells. Our results indicate that certainfactors specific to HepG2 cells may be involved in this methylation-related regulation. Biobase (http://www.biobase-international.com/)predicted possible binding of C-myc and upstream translation factor 1 tothis sequence, however, quantitative real-time PCR results show that C-mycand upstream translation factor 1 are abundant in all three cell lines (data notshown). Since there are very limited data available regarding the tran-scriptional regulation of CAR, the exact function of this sequence and thecandidate binding proteins remain unclear.An e-box is a special DNA sequence (CANNTG) found in some

promoter regions in eukaryotes that acts as a binding site for regulatingproteins (Massari and Murre, 2000). It has been reported that methyl-ation of the CpG dinucleotide within the e-box sequence can block thebinding of N-Myc to its targeting gene promoters in vivo, thus contrib-uting to the cell type–specific expression pattern of certain genes such ascaspase-8 and the epidermal growth factor receptor (Perini et al., 2005).Since two e-box sequences (CACGTG) are located near the transcriptionstart site and translation start site of CAR, respectively, methylation onthe CpG dinucleotide is likely to play a role in modulating CARexpression. In our study, 29 liver tumor samples showed decreasedCYP2C19 and CAR expression along with elevated DNA methylationaround the e-box binding site compared with adjacent normal liversamples. Only in sample 246564 did the situation reverse. Furtherscrutiny into this patient’s diagnostic information did not reveal anythingunusual. Clearly, this distinct phenomenon awaits more data in orderto be better understood. However, the reversed correlation betweenCYP2C19 and CAR expression levels and DNA methylation was alsoobserved in sample 246564, which confirms that DNA methylationstatus might serve as a potential marker for predicting changes in geneexpression.Nevertheless, other factors that contribute to the epigenetic regulation

of CYP2C19 cannot be excluded, since the distal promoter region ofCYP2C19 is unidentified and more transcription factors might exist.However, reduced CAR expression by e-box methylation at leastpartially accounts for CYP2C19 suppression in the HBV-infectedHCC tissue samples.It has been well documented that DNA methylation in different

genomic contexts exerts different effects on gene transcription (Jones,2012). Much attention has been paid to the CpG-rich promoters duringthe past years, whereas CpG-poor transcription start sites have been

Fig. 7. Effect of DNA methylation on the transcription activity of CYP2C19and CAR promoter constructs. (A, C, and D) The pCpGL plasmids werecotransfected with pRL-TK in HepG2, HEK293, or HK2 cells. The pCpGLplasmids were either methylated in the presence (methylated) or absence(unmethylated) of M.SssI. (B) The CpG-free constructs were cotransfected withpCpGL-2C19 and pRL-TK in HepG2 cells. The CpG-free constructs were eithermethylated in the presence (methylated) or absence (unmethylated) of M.SssI.The relative luciferase activity was calculated as the ratio of firefly luciferase to

Renilla luciferase. Data are shown as means 6 S.D. with three replicate wellsfor each group. Data are representative of three independent experiments.**P , 0.01.

Epigenetic Study of CYP2C19 in Hepatocellular Carcinoma 1651

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from

overlooked. In fact, non-CpG island methylation seems to be moredynamic and therefore more changeable during development orpathogenesis. For example, the CpG-poor LAMB3 promoter was shownto be hypomethylated in primary bladder tumors. In vitro methylation ofthe promoter can directly lead to transcription silencing (Han et al.,2011). Furthermore, enhancers (which can reside at variable distancesfrom promoters and are mostly CpG poor) also tend to show variablemethylation status.Lempiäinen et al. (2011) demonstrated that long-term activation of

CAR in mice can lead to HCC, which is accompanied by a switch in themethylomes. However, whether CAR plays a direct role in human HCCremains controversial. Cancer pathogenesis is a complex process inwhich multiple genetic and epigenetic factors are regulated abnormally.This complexity poses a formidable challenge to holistic integration ofinformation when the investigation is limited to only a certain gene.Traditionally, epigenetic research on cancer has mostly focused ontumor suppressors or oncogenes. Only in recent years has the necessityof building a comprehensive epigenome map been realized. Comparedwith the well established Human Genome Project, the InternationalHuman Epigenome Consortium was launched recently in 2010 (http://ihec-epigenomes.org/) and aims to increase understanding of how theepigenome has shaped human populations over generations and how it isresponds to the environment. The consortium’s goals include generatinghigh-resolution maps of DNA methylation and informative histonemodifications and performing a comparative analysis of epigenomemaps between healthy and diseased states. It is highly anticipated thatadvances in the field of epigenetics will shed light on intricate lifeprocesses such as tumorigenesis and aging.

Acknowledgments

The authors thankMichael Rehli (University Hospital, Regensburg, Germany)for plasmid pCpGL-basic.

Authorship ContributionsParticipated in research design: Tang, Zeng, S. Chen.Conducted experiments: Tang, Ge, Z. Chen, Kong, Liu, Xu.Performed data analysis: Tang, Liu.Wrote or contributed to the writing of the manuscript: Tang, Zeng, S. Chen.

References

Bernard WS and Christopher PW (2014) World cancer report 2014, World Health Organization,Geneva.

Chen H, Shen ZY, Xu W, Fan TY, Li J, Lu YF, Cheng ML, and Liu J (2014) Expression of P450and nuclear receptors in normal and end-stage Chinese livers. World J Gastroenterol 20:8681–8690.

Christman JK (2002) 5-Azacytidine and 5-aza-29-deoxycytidine as inhibitors of DNA meth-ylation: mechanistic studies and their implications for cancer therapy. Oncogene 21:5483–5495.

Griskevicius L, Yasar U, Sandberg M, Hidestrand M, Eliasson E, Tybring G, Hassan M, and DahlML (2003) Bioactivation of cyclophosphamide: the role of polymorphic CYP2C enzymes. Eur JClin Pharmacol 59:103–109.

Han H, Cortez CC, Yang X, Nichols PW, Jones PA, and Liang G (2011) DNA methylation directlysilences genes with non-CpG island promoters and establishes a nucleosome occupied promoter.Hum Mol Genet 20:4299–4310.

Helsby NA, Lo WY, Sharples K, Riley G, Murray M, Spells K, Dzhelai M, Simpson A,and Findlay M (2008) CYP2C19 pharmacogenetics in advanced cancer: compromised functionindependent of genotype. Br J Cancer 99:1251–1255.

Herceg Z and Paliwal A (2011) Epigenetic mechanisms in hepatocellular carcinoma: how envi-ronmental factors influence the epigenome. Mutat Res 727:55–61.

Iizuka N, Oka M, Yamada-Okabe H, Mori N, Tamesa T, Okada T, Takemoto N, Tangoku A,Hamada K, Nakayama H, et al. (2002) Comparison of gene expression profiles between hepatitisB virus- and hepatitis C virus-infected hepatocellular carcinoma by oligonucleotide microarraydata on the basis of a supervised learning method. Cancer Res 62:3939–3944.

Ingelman-Sundberg M, Sim SC, Gomez A, and Rodriguez-Antona C (2007) Influence of cyto-chrome P450 polymorphisms on drug therapies: pharmacogenetic, pharmacoepigenetic and clin-ical aspects. Pharmacol Ther 116:496–526.

Jemal A, Bray F, Center MM, Ferlay J, Ward E, and Forman D (2011) Global cancer statistics. CACancer J Clin 61:69–90.

Ji X, Zhang Q, Du Y, Liu W, Li Z, Hou X, and Cao G (2014) Somatic mutations, viral integrationand epigenetic modification in the evolution of hepatitis B virus-induced hepatocellular carci-noma. Curr Genomics 15:469–480.

Jones PA (2012) Functions of DNA methylation: islands, start sites, gene bodies and beyond. NatRev Genet 13:484–492.

Klug M and Rehli M (2006) Functional analysis of promoter CpG methylation using a CpG-freeluciferase reporter vector. Epigenetics 1:127–130.

Lempiäinen H, Müller A, Brasa S, Teo SS, Roloff TC, Morawiec L, Zamurovic N, Vicart A,Funhoff E, Couttet P, et al. (2011) Phenobarbital mediates an epigenetic switch at the constitutiveandrostane receptor (CAR) target gene Cyp2b10 in the liver of B6C3F1 mice. PLoS One 6:e18216.

Lessel D, Vaz B, Halder S, Lockhart PJ, Marinovic-Terzic I, Lopez-Mosqueda J, Philipp M, SimJC, Smith KR, Oehler J, et al. (2014) Mutations in SPRTN cause early onset hepatocellularcarcinoma, genomic instability and progeroid features. Nat Genet 46:1239–1244.

Liu R, Li Y, Wangen KR, Maitland E, Nicholas S, and Wang J (2016) Analysis of hepatitis Bvaccination behavior and vaccination willingness among migrant workers from rural Chinabased on protection motivation theory. Hum Vacc Immunother 12:1155–1163.

Massari ME and Murre C (2000) Helix-loop-helix proteins: regulators of transcription in eucaryoticorganisms. Mol Cell Biol 20:429–440.

Nishida N and Goel A (2011) Genetic and epigenetic signatures in human hepatocellular carci-noma: a systematic review. Curr Genomics 12:130–137.

Okabe H, Satoh S, Kato T, Kitahara O, Yanagawa R, Yamaoka Y, Tsunoda T, Furukawa Y,and Nakamura Y (2001) Genome-wide analysis of gene expression in human hepatocellularcarcinomas using cDNA microarray: identification of genes involved in viral carcinogenesis andtumor progression. Cancer Res 61:2129–2137.

Perini G, Diolaiti D, Porro A, and Della Valle G (2005) In vivo transcriptional regulation of N-Myctarget genes is controlled by E-box methylation. Proc Natl Acad Sci USA 102:12117–12122.

Schmittgen TD and Livak KJ (2008) Analyzing real-time PCR data by the comparative C(T)method. Nat Protoc 3:1101–1108.

Shou M, Martinet M, Korzekwa KR, Krausz KW, Gonzalez FJ, and Gelboin HV (1998) Role ofhuman cytochrome P450 3A4 and 3A5 in the metabolism of taxotere and its derivatives: enzymespecificity, interindividual distribution and metabolic contribution in human liver. Pharmaco-genetics 8:391–401.

Thorgeirsson SS and Grisham JW (2002) Molecular pathogenesis of human hepatocellular carci-noma. Nat Genet 31:339–346.

Tsunedomi R, Iizuka N, Hamamoto Y, Uchimura S, Miyamoto T, Tamesa T, Okada T, Takemoto N,Takashima M, Sakamoto K, et al. (2005) Patterns of expression of cytochrome P450 genes inprogression of hepatitis C virus-associated hepatocellular carcinoma. Int J Oncol 27:661–667.

Uppugunduri CR, Daali Y, Desmeules J, Dayer P, Krajinovic M, and Ansari M (2012) Tran-scriptional regulation of CYP2C19 and its role in altered enzyme activity. Curr Drug Metab 13:1196–1204.

Wortham M, Czerwinski M, He L, Parkinson A, and Wan YJ (2007) Expression of constitutiveandrostane receptor, hepatic nuclear factor 4 a, and P450 oxidoreductase genes determinesinterindividual variability in basal expression and activity of a broad scope of xenobiotic me-tabolism genes in the human liver. Drug Metab Dispos 35:1700–1710.

Yang P, Markowitz GJ, and Wang XF (2014) The hepatitis B virus-associated tumor microenvi-ronment in hepatocellular carcinoma. Natl Sci Rev 1:396–412.

Yu D, Green B, Tolleson WH, Jin Y, Mei N, Guo Y, Deng H, Pogribny I, and Ning B (2015)MicroRNA hsa-miR-29a-3p modulates CYP2C19 in human liver cells. Biochem Pharmacol 98:215–223.

Address correspondence to: Shuqing Chen, Institute of Drug Metabolism andPharmaceutical Analysis, Zhejiang Province Key Laboratory of Anti-Cancer DrugResearch, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou,Zhejiang 310058, China. E-mail: chenshuqing@zju.edu.cn

1652 Tang et al.

at ASPE

T Journals on D

ecember 2, 2020

dmd.aspetjournals.org

Dow

nloaded from