Abstract
The wingless-type MMTV integration site family (WNT)/β-catenin/adenomatous polyposis coli (CTNNB1/APC) pathway has been identified as a regulator of drug-metabolizing enzymes in the rodent liver. Conversely, little is known about the role of this pathway in drug metabolism regulation in human liver. Primary human hepatocytes (PHHs), which are the most physiologically relevant culture system to study drug metabolism in vitro, were used to investigate this issue. This study assessed the link between cytochrome P450 expression and WNT/β-catenin pathway activity in PHHs by modulating its activity with recombinant mouse Wnt3a (the canonical activator), inhibitors of glycogen synthase kinase 3β, and small-interfering RNA to invalidate CTNNB1 or its repressor APC, used separately or in combination. We found that the WNT/β-catenin pathway can be activated in PHHs, as assessed by universal β-catenin target gene expression, leucine-rich repeat containing G protein–coupled receptor 5. Moreover, WNT/β-catenin pathway activation induces the expression of CYP2E1, CYP1A2, and aryl hydrocarbon receptor, but not of CYP3A4, hepatocyte nuclear factor-4α, or pregnane X receptor (PXR) in PHHs. Specifically, we show for the first time that CYP2E1 is transcriptionally regulated by the WNT/β-catenin pathway. Moreover, CYP2E1 induction was accompanied by an increase in its metabolic activity, as indicated by the increased production of 6-OH-chlorzoxazone and by glutathione depletion after incubation with high doses of acetaminophen. In conclusion, the WNT/β-catenin pathway is functional in PHHs, and its induction in PHHs represents a powerful tool to evaluate the hepatotoxicity of drugs that are metabolized by CYP2E1.
Introduction
The liver performs many different functions thanks to the organization of hepatocytes in different functional groups, a remarkable property known as functional zonation. Hepatocytes present highly specialized metabolic functions from the portal space to the centrilobular vein and based on their localization are defined, respectively, as periportal and pericentral hepatocytes. Recently, the wingless-type MMTV integration site family (WNT)/β-catenin-adenomatous polyposis coli (APC) pathway has been identified as the major liver “zonation-keeper” (Benhamouche et al., 2006; Colnot and Perret, 2010).
The expression and localization of β-catenin (CTNNB1) are tightly regulated. In the absence of WNT ligand stimulation, CTNNB1 is phosphorylated by a protein complex that includes glycogen synthase kinase 3β (GSK3β) and APC, leading to its proteasomal degradation. In the presence of WNT ligands, CTNNB1 remains unphosphorylated and accumulates in the cytosol. It then translocates into the nucleus where it transactivates target genes by binding to transcriptional activators of the T-cell factor/lymphoid-enhancing factor (TCF/Lef) family (MacDonald et al., 2009). In parallel to its transcriptional activity, CTNNB1 participates in adherens junction formation by interacting with E-cadherin and α-catenin, linking cadherins to the actin cytoskeleton (Vinken et al., 2006).
Detoxification of exogenous compounds is one of the liver major metabolic functions and is mediated by a large family of proteins, including the phase I cytochrome P450 enzymes (P450). P450 expression is regulated by transcription factors, including hepatocyte nuclear factor-4α (HNF4α) and nuclear receptors such as constitutive androstane receptor (CAR, NR1I3, nuclear receptor subfamily 1, group i, member 3), pregnane X receptor (PXR, NR1I2), and aryl hydrocarbon receptor (AhR) of the basic Helix-Loop-Helix/Per-ARNT-Sim (bHLH/PAS) receptor family, which are the major regulators of drug disposition (Kohle and Bock, 2009).
Histologic analysis of liver sections (Hailfinger et al., 2006; Sekine et al., 2006; Braeuning et al., 2009), microarray and real-time quantitative polymerase chain reaction analysis of enriched periportal and centrilobular hepatocytes (Braeuning et al., 2006; Sekine et al., 2006) have shown that, in the mouse, Cyp2e1, Cyp1a2, and AhR are expressed in pericentral hepatocytes, where the Wnt/β-catenin pathway is mostly activated.
Genetic manipulation of animals is a powerful tool to dissect the complexity of liver zonation. In hepatocyte-specific Ctnnb1 knockout mice, loss of expression of several P450 enzymes, especially Cyp2e1 and Cyp1a2, has been observed (Sekine et al., 2006). Moreover, the basal expression of most drug metabolism-related genes and the response to NR1I3 and AhR agonists are also reduced in these mice (Braeuning et al., 2009; Ganzenberg et al., 2013). Similarly, acetaminophen (APAP)-induced toxicity is abolished in Ctnnb1−/− mice (Sekine et al., 2006). Conversely, several P450 isoenzymes are up-regulated in liver tumors harboring CTNNB1-activating mutations (Loeppen et al., 2005). Liver-specific Apc loss causes the de novo expression of β-catenin–positive pericentral genes and the suppression of β-catenin–negative periportal target genes (Benhamouche et al., 2006).
Primary human hepatocytes (PHHs) are the most physiologically relevant culture system to study drug metabolism in vitro. However, when hepatocytes are seeded in culture, zonal organization is lost, and P450 expression progressively decreases. As nothing is known about the link between P450 expression and β-catenin pathway activity in PHHs, we decided to investigate this aspect using our in vitro model of PHHs. We demonstrated that the β-catenin pathway can be activated in PHHs, despite the loss of the organization, to restore the expression of its target genes. We identified CYP2E1, CYP1A2, and AhR as CTNNB1 target genes, whereas CYP3A4, HNF4α, and PXR are not directly regulated by this pathway. Moreover, CYP2E1 induction is accompanied by an increase of APAP-induced toxicity in PHHs.
Materials and Methods
Human Liver Samples and Preparation of Primary Human Hepatocytes.
Liver samples were obtained from liver resections performed in adult patients for medical reasons unrelated to our research program. The use of human specimens for scientific purposes was approved by the French National Ethics Committee. Written or oral informed consent was obtained from each patient or family before surgery. The clinical characteristics of the liver donors are presented in Table 1. PHHs were prepared and cultured as described previously elsewhere (Pichard et al., 2006). PHHs were seeded in collagen-coated dishes at 1.7 × 105 cells/cm2 in a hormonally and chemically defined medium consisting of a mixture of William’s E and Ham’s F-12 (1:1 in volume) and additives as described by Ferrini et al. (1997). PHHs were cultured in a 5% CO2 humidified atmosphere at 37°C.
Clinical characteristics of the liver donors
Chemicals.
CHIR99021 (CHIR) [6-((2-((4-(2,4-dichlorophenyl)-5-(4-methyl-1H-imidazol-2-yl)pyrimidin-2-yl)amino)ethyl)amino) nicotinonitrile] was obtained from BioVision (Milpitas, CA). TCDD (2,3,7,8-tetrachlorodibenzo-p-dioxine), rifampicin (RIF), and lithium chloride were purchased from Sigma-Aldrich (Saint-Louis, MO), and mWnt3a was obtained from Peprotech (Neuilly-Sur-Seine, France).
Cell Lines.
The HepaRG cell line was grown and differentiated as recommended by Parent et al. (2004). Briefly, cells were seeded at 2.6 × 104 cells/cm2 and expanded in growth medium for 2 weeks, then differentiated by the addition of 1.5% dimethylsulfoxide for 2 weeks. HepG2-C3 (American Type Culture Collection, Manassas, VA) and HuH7 (Japanese Collection of Research Bioresources Cell Bank, Osaka, Japan) cells were cultured as recommended. CHIR was added to HepG2-C3 and HuH7 cells when they reached 80% confluence and to differentiated HepaRG cells. Cell lines were cultured in a 5% CO2 humidified atmosphere at 37°C.
Small-Interfering RNA Transient Transfection.
Adherent PHHs were transfected with 20 nM nontargeting small-interfering RNA (siRNA) (scrambled, siSC) or siRNA specific for CTNNB1, APC, or AhR (Dharmacon, Lafayette, CO) at day 1 and day 3 after seeding using Lipofectamine RNAiMAX (Life Technologies, Carlsbad, CA). At day 5 after seeding, PHHs were treated for 48 hours.
DNA and Reporter Gene Expression Assays.
PHHs in suspension were transfected with Lipofectamine 2000 transfection reagent (Life Technologies) according to the manufacturer’s instructions before plating. Briefly, 2.7 × 104 PHHs were transfected with either 500 ng of firefly luciferase reporter plasmids SuperTOPflash (Wnt response element, WRE) or SuperFOPflash (mutated WRE) (Staal et al., 1999) and 250 ng of pTK-luc Renilla control vector (Promega, Madison, WI). PHHs were then plated in 24-well plates in DNA:liposome mix in chemically defined medium. After 24 hours, the medium was renewed, and the cells were treated with 3 µM CHIR, 200 ng/ml mWnt3a, or 20 mM LiCl for 24 hours.
RNA Isolation and Reverse Transcription Polymerase Chain Reaction.
After extraction with TRIzol reagent (Invitrogen/Life Technologies, Carlsbad, CA), 500 ng of total RNA was reverse-transcribed using a random hexaprimer and the MMLV Reverse Transcriptase Kit (Invitrogen). Quantitative polymerase chain reaction was performed using the Roche SYBER Green reagent and a LightCycler 480 apparatus (Roche Diagnostic, Meylan, France). Amplification specificity was evaluated by determining the product melting curve. Results are expressed as indicated in the figure legends. The primers used were listed in Table 2. The following program was used: one step at 95°C for 10 minutes, 50 cycles of denaturation at 95°C for 10 seconds, annealing at 65°C for 15 seconds, and elongation at 72°C for 15 seconds.
Primer pair sequences
Protein Analysis.
Total protein extracts were prepared using a radioimmunoprecipitation assay buffer supplemented with a protease inhibitor cocktail (Santa Cruz Biotechnology, Santa Cruz, CA). The protein concentration was determined by the bicinchoninic acid method, according to the manufacturer’s instructions (Pierce Chemical, Rockford, IL). Bovine serum albumin (Pierce Chemical) was used as standard. We separated 20 µg of total proteins on precast SDS-polyacrylamide gels (4–16%) (Bio-Rad Laboratories, Marnes la Coquette, France), then transferred them onto polyvinylidene fluoride membranes (Bio-Rad Laboratories). Membranes were incubated with rabbit polyclonal anti-CYP2E1 (Millipore, Molsheim, France) or anti-CTNNB1 (Abcam, Cambridge, UK), mouse monoclonal anti-CYP1A2 (Santa Cruz Biotechnology), or goat polyclonal anti-actin (Santa Cruz Biotechnology) antibodies. Microsomes from human lymphoblastoid cells transfected with the human CYP2E1 (Gentest, Woburn, MA) were used as standards.
Immunohistochemistry.
Paraffin-embedded human liver tissue sections (4 µm thick) were incubated at 4°C with rabbit antibodies against CYP3A4 (Epitomics, Burlingame, CA), cytokeratin 19 (KRT19, Epitomics), CYP2E1 (Millipore), glutamine synthetase (GLUL) (Abnova Corporation, Taipei, Taiwan), or mouse antibody against CYP1A2 (Santa Cruz Biotechnology). Immunohistochemical staining was performed using the EnVisionTM+ System (Dako, Glostrup, Denmark) according to the manufacturer’s recommendations. Finally, sections were lightly counterstained with Harris hematoxylin. Slides were scanned (Montpellier RIO Imaging Facility, INM Montpellier, France) using a Nanozoomer Slide Scanner (Hamamatsu Photonics, Massy, France), and virtual slides were viewed using the NDP.view software (Hamamatsu Photonics, Massy, France).
Measurement of Chlorzoxazone Hydroxylation.
After 72 hours of treatment with 3 µM CHIR, the PHH culture medium was renewed in the presence of 15 µM chlorzoxazone (Sigma-Aldrich). Six hours later, extracellular medium and cells were collected, and the 6-OH chlorzoxazone content was measured by liquid chromatography with tandem mass spectrometry using an Acquity UPLC System I-Class equipped with a Waters Acquity UPLC BEH C18 column (2.1 mm i.d. × 100 mm length, 1.7-µm particle size) coupled to a Quattro Premier mass spectrometer (all from Waters Corporation, Milford, MA) used in electrospray ion negative mode.
Measurement of Glutathione Content.
After a 72-hour incubation with 3 µM CHIR, the PHH culture medium was renewed, and 20 or 50 mM APAP was added (Sigma-Aldrich). Six hours later, the cells were collected, and the glutathione (GSH) content was measured using a glutathione assay kit (Cayman Chemical, Ann Arbor, MI) according to the manufacturer’s instructions. Results are expressed as the quantity of GSH lost in APAP-treated cells compared with nontreated cells.
Statistical Analysis.
The values of mRNA expression were expressed as the mean ± standard error of the mean (S.E.M.). Statistical analysis between groups was performed by using the paired t test. Analysis of the variance was used to determine statistical differences, and the Tukey’s multiple comparison test was performed to compare three groups. Differences were considered statistically significant when P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001). Data were analyzed by using GraphPad Prism, version 4.0, 2003 (GraphPad Software, San Diego, CA).
Results
The Canonical Wnt/β-Catenin Pathway Is Functional in PHHs.
When human hepatocytes are seeded in culture, zonation is lost, and P450 expression progressively decreases. As in vivo zonation is mainly controlled by the WNT/β-catenin pathway, we first determined whether this pathway can still control gene transcription in PHHs independently of zonal organization. To this end, we monitored the expression of leucine-rich repeat containing G protein–coupled receptor 5 (LGR5 or GPR49), a universal β-catenin target gene (Yamamoto et al., 2003), in PHHs that were cultured according to our established long-term culture conditions (Pichard et al., 2006). LGR5 expression was progressively and strongly reduced during culture. Conversely, the expression of APC and GSK3β, which are involved in CTNNB1 proteasomal degradation, and Dickkopf-3 (DKK3), a member of the DKK family of WNT antagonists, increased over time (Fig. 1A). These results suggest that CTNNB1 signaling is gradually down-regulated during PHH culture.
The β-catenin pathway can be activated in PHHs. (A) RT-qPCR analysis of LGR5, APC, GSK3β, and DKK3 gene expression over time (maximal expression was arbitrary fixed to 100) (n = 3). (B) RT-qPCR analysis of LGR5 mRNA expression after 48 hours of incubation with 10 or 20 mM LiCl, 100 or 200 ng/ml mWnt3a, 1 or 3 µM CHIR, 10 nM TCDD, or 10 µM RIF (n = 4). RT-qPCR analysis of LGR5 mRNA expression in PHHs from different liver donors (PHH365, PHH359, and PHH366) incubated at day 2 after seeding with (C) 1, 3, or 10 µM CHIR (D) 100, 200, or 400 ng/ml mWnt3a for 48 hours. D, day; FIH, freshly isolated hepatocytes; nd, not detected; UT, untreated.
Then, to determine whether the WNT/β-catenin pathway could be activated in PHHs, we exposed the cells or not to recombinant mouse Wnt3a (mWnt3a, the canonical activator of this pathway) or to the GSK3β inhibitors CHIR99021 (Ring et al., 2003) and LiCl (Abu-Baker et al., 2013) for 48 hours (day 2 to day 4 of culture); we then analyzed the mRNA expression of the target gene LGR5 by reverse-transcription real-time quantitative polymerase chain reaction (RT-qPCR). As expected, LGR5 was induced by all these compounds, but its expression was not affected by TCDD or RIF, the prototypical AhR and PXR activators, respectively (Fig. 1B). Moreover, the PHH response to GSK3β inhibition or to mWnt3a showed a significant interindividual variability (Fig. 1, C and D).
To demonstrate the ability of CTNNB1 to activate the TCF/Lef responsive element after exposure to mWnt3a, LiCl, or CHIR, PHHs were transfected with TOP-luc (black bars) or FOP-luc plasmids (white bars) that harbor consensus or mutated TCF/Lef responsive elements, respectively. After incubation with 3 µM CHIR for 24 hours, a 3-fold increase in luciferase activity was observed in TOP-luc transfected cells compared with untreated cells, but it remained unchanged in FOP-luc transfected cells (Fig. 2A). Induction of the WNT/β-catenin pathway by the canonical CTNNB1 activator mWnt3a or LiCl also stimulated luciferase activity, although to a lesser extent than CHIR.
The β-catenin pathway can be activated in PHHs. (A) Luciferase activity measurement in PHHs transfected with the TK-FOP-luc (white bars) or TK-TOP-luc (black bars) plasmids and incubated 24 hours later with 3 µM CHIR, 20 mM LiCl, or 200 ng/ml mWnt3a for 24 hours. (B) RT-qPCR analysis of LGR5 expression in PHHs after siRNA-mediated CTNNB1 silencing and incubation with 200 ng/ml mWnt3a or 3 µM CHIR for 24 hours (n = 5) or after siRNA-mediated APC silencing (n = 3). (C) RT-qPCR analysis of LGR5 mRNA expression in HepG2-C3A, HuH7 or differentiated HepaRG cells incubated with 3 µM CHIR for 48 hours. nd, not detected.
The involvement of the WNT/β-catenin pathway in Wnt3a- and CHIR-mediated LGR5 up-regulation was confirmed by siRNA-mediated down-regulation of CTNNB1 or APC in PHHs. APC is involved in CTNNB1 degradation, and its inactivation results in CTNNB1 accumulation, thereby mimicking the activation of the WNT/β-catenin pathway (Colnot et al., 2004; Burke et al., 2009). The specific siRNAs efficiently knocked down APC (by 70%) and CTNNB1 (by 90%) mRNA expression in PHHs (data not shown). Wnt3a- and CHIR-mediated LGR5 up-regulation was abolished (Fig. 2B) in CTNNB1-silenced PHHs but not in control cells (scrambled siRNA, siSC). Conversely, LGR5 basal expression was significantly increased in APC-silenced PHHs in comparison with siSC controls (Fig. 2B), confirming that LGR5 is a WNT/β-catenin pathway target and that CHIR-mediated LGR5 up-regulation is dependent on CTNNB1. Similar results were obtained using hepatic cell lines in which the WNT/β-catenin pathway is functional, such as HuH7 cells and differentiated HepaRG cells, or that harbor CTNNB1-activating mutations, such as HepG2-C3 cells (de La Coste et al., 1998). Incubation with 3 µM CHIR for 48 hours increased the expression of LGR5 in HuH7 and HepaRG cells but not in HepG2-C3 cells, confirming that the response to CHIR relies on a functional WNT/β-catenin pathway (Fig. 2C).
These findings indicate that in PHHs the canonical WNT/β-catenin signaling pathway is still functional and can be activated directly by the canonical activator mWnt3a or indirectly by the GSK3β inhibitor CHIR.
P450s Regulation by the WNT/β-Catenin Signaling Pathway.
P450 expression is zonated in human liver tissue. Immunohistochemical analysis of human liver tissue serial sections showed that CYP2E1, CYP1A2, and CYP3A4 are expressed in pericentral hepatocytes (Fig. 3A), similarly to GLUL, a well known pericentral marker. Conversely, cytokeratin 19-positive (KRT19+) biliary cells are indicative of periportal areas. As observed for the CTNNB1 target gene LGR5 (Fig. 1A), P450 expression was progressively down-regulated in cultured PHHs (Fig. 3B). Therefore, we asked whether this resulted from the loss of CTNNB1-mediated gene transcription.
Pericentral expression of CYP2E1, CYP1A2, and CYP3A4. (A) Analysis of CYP2E1, CYP1A2, CYP3A4, GLUL, and KRT19 expression in serial human liver tissue sections by immunohistochemistry. Scale bars = 1 mm. (B) RT-qPCR analysis of CYP2E1, CYP1A2, and CYP3A4 gene expression over time (expression in FIH was arbitrary fixed to 100) (n = 3). CT, control without antibody; D, day; FIH, freshly isolated hepatocytes.
CYP2E1.
CYP2E1 is expressed in pericentral hepatocytes [Fig. 3 for human; Loeppen et al. (2005), Hailfinger et al. (2006), and Sekine et al. (2006) for mouse liver] where the WNT/β-catenin pathway is mostly active (Benhamouche et al., 2006). To assess the effect of the WNT/β-catenin pathway activation on CYP2E1 expression in PHHs, we incubated cultures with mWnt3a (the canonical CTNNB1 activator) or with GSK3β inhibitors (LiCl and CHIR). Upon induction of the WNT/β-catenin pathway, CYPE21 mRNA expression increased in a dose-dependent manner compared with untreated cells. As expected, CYPE21 mRNA expression was not affected by TCDD or RIF (Fig. 4A). Up-regulation of CYP2E1 by mWnt3a- and CHIR was abolished in PHHs in which CTNNB1 was silenced by siRNA down-regulation, indicating that this effect is dependent on CTNNB1 (Fig. 4B). Conversely, APC silencing increased CYP2E1 basal expression (Fig. 4B), as observed for LGR5 (Fig. 2B), compared with control cells (siSC). These findings demonstrate that the canonical activator mWnt3a and the GSK3β inhibitor CHIR can regulate CYP2E1 mRNA expression in a CTNNB1-dependent manner. As CHIR was the most efficient and easy to use activator of the WNT/β-catenin pathway, we preferentially employed it in the subsequent experiments.
CYP2E1 is a CTNNB1 target gene. (A) RT-qPCR analysis of CYP2E1 mRNA expression after incubation with 10 or 20 mM LiCl, 100 or 200 ng/ml mWnt3a, 1 or 3 µM CHIR, 10 nM TCDD, or 10 µM RIF. (B) RT-qPCR analysis of CYP2E1 mRNA expression in PHHs after siRNA-mediated CTNNB1 or APC down-regulation and incubation with CHIR 3 µM or 200ng/ml mWnt3a. ND, not determined.
Analysis of 14 different PHH cultures confirmed that CHIR strongly induced CYP2E1 mRNA expression (mean fold change and median: 12.6 and 10.0) compared with untreated cells (Fig. 5A). Like for LGR5 (Fig. 2C) and as previously described for other human P450s (Gomez-Lechon et al., 2007), a strong interindividual variation was observed. However, there was no correlation between the clinical characteristics of the liver donor and the level of CYP2E1 induction in PHHs. CYP2E1 mRNA (Fig. 5B) and protein (Fig. 5C) expression were followed in PHHs cultured for 4 to 5 days. Basal CYP2E1 expression decreased rapidly over time; however, exposure to 3 µM CHIR from day 2 of culture significantly increased CYP2E1 mRNA and protein levels. CHIR-mediated CYP2E1 protein induction was abolished by siRNA-mediated CTNNB1 down-regulation, but it was markedly increased after silencing of APC (Fig. 5D). Moreover, CYP2E1 enzymatic activity in PHHs was confirmed by adding chlorzoxazone, a CYP2E1-specific substrate, to the cultures. A 5-fold increase in the production of its metabolite 6-OH-chlorzoxazone was observed in cells incubated with CHIR for 72 hours compared with untreated cells (Fig. 5E).
CYP2E1 is a CTNNB1 target gene. (A) CYP2E1 induction in PHHs after incubation with 3 µM CHIR for 48 hours. Results are expressed as fold change relative to untreated cells (UT) (n = 14). CYP2E1 mRNA (B) and protein (C) expression in PHHs before and after addition (or not) of 3 µM CHIR at day 2. One pmol of recombinant CYP2E1 (rCYP2E1) was used as control. (D) CYP2E1 protein expression in siRNA-transfected PHHs after incubation or not with 3 µM CHIR for 48 hours. (E) CYP2E1-mediated 6-OH chlorzoxazone production in PHHs (n = 2). (F) GSH content depletion in PHHs incubated or not with 3 µM CHIR for 72 hours and incubated with 20 or 50 mM APAP for 6 hours. (G) RT-qPCR analysis of CYP2E1 mRNA expression in HepG2-C3A, HuH7 or differentiated HepaRG cells after incubation with 3 µM CHIR for 48 hours. Results are expressed as percentage of the CYP2E1 mRNA expression level in PHHs at day 1 post-seeding. FIH, freshly isolated hepatocytes.
CYP2E1 also contributes to APAP metabolism by catalyzing the production of the reactive intermediate N-acetyl-p-benzoquinoneimine, which is normally rapidly detoxified by glutathione (GSH) conjugation in the liver. Therefore, in case of APAP overdose, this intermediate can lead to GSH depletion and to hepatotoxicity (James et al., 2003). Conversely, Ctnnb1−/− mice are resistant to APAP-induced toxicity (Sekine et al., 2006). Quantification of the GSH content in PHHs stimulated with CHIR for 72 hours and incubated with 20 or 50 mM APAP for the last 6 hours showed a strong GSH depletion compared with controls cells (untreated cells, no CHIR stimulation) (Fig. 5F).
In hepatic cell lines, basal CYP2E1 mRNA expression was low and was up-regulated in response to CHIR exposure only in differentiated HepaRG cells, but not in HepG2-C3 and HuH7 cells (Fig. 5G). This suggests that activation of the WNT/β-catenin pathway is not sufficient per se to induce CYP2E1 expression and underlines the importance of a differentiated hepatic context for full regulation.
These findings demonstrate for the first time that CYP2E1 is transcriptionally up-regulated after CTNNB1 activation in PHHs.
CYP1A2.
CYP1A2 mRNA expression also was increased in PHHs after incubation with mWnt3a exposure, suggesting that CTNNB1 participates in the regulation of CYP1A2 expression in PHHs (Fig. 6A). We evaluated the role the Wnt/β-catenin pathway in CYP1A2 regulation also by siRNA-mediated down-regulation of CTNNB1, APC, and AhR (as a positive control because CYP1A2 is a known AhR target gene). CYP1A2 basal expression was not affected by CTNNB1 down-regulation (Fig. 6B), but CYP1A2 mRNA and protein expression (Fig. 6B and 6C) were strongly increased upon APC silencing compared with the cells transfected with scrambled siRNA (siSC). As previously described elsewhere (Lee et al., 2011), AhR down-regulation significantly decreased basal CYP1A2 mRNA expression (Fig. 6B). Similarly, CYP1A2 up-regulation upon incubation with TCDD (dioxin) was not affected by siRNA-mediated CTNNB1 silencing, but it was increased upon APC down-regulation compared with control cells (siSC). As expected, CYP1A2 induction was strongly reduced in AhR-silenced cells (Fig. 6D) because TCDD needs to bind to AhR to act as a transcriptional regulator (Fujii-Kuriyama and Mimura, 2005). Finally, we assessed whether AhR expression was also regulated by CTNNB1 in PHHs. AhR mRNA expression was increased in a dose-dependent manner by incubation with mWnt3a compared with untreated cells. A similar effect was observed after APC silencing (Fig. 6F), but CTNNB1 down-regulation had no effect compared with control siRNA.
CYP1A2 regulation in PHHs. (A) RT-qPCR analysis of CYP1A2 expression in PHHs after incubation, or not, with 100 or 200 ng/ml mWnt3a (n = 4). (B) CYP1A2 mRNA expression after CTNNB1, APC, or AhR silencing. (C) CYP1A2 protein expression in PHHs after siRNA-mediated APC silencing. (D) CYP1A2 mRNA expression after CTNNB1, APC, or AhR silencing and incubation with 10 nM TCDD for 24 hours. Results are expressed as fold change relative to untreated hepatocytes (UT). (E) RT-qPCR analysis of AhR expression in PHHs after incubation or not, with 100 or 200 ng/ml mWnt3a. (F) AhR mRNA expression after CTNNB1, APC, or AhR silencing. siSC, scrambled siRNA.
CYP3A4.
Differently from CYP2E1 and CYP1A2, CYP3A4 mRNA expression in PHHs was not induced by incubation with mWnt3a, the canonical activator of the CTNNB1 pathway. Conversely, and as expected, it was strongly induced by RIF, a well known PXR activator, compared with untreated cells (Fig. 7A). Thus, although in vivo CYP2E1 and CYP3A4 showed similar zonal expression (Fig. 3), the basal expression of CYP3A4 in PHHs does not seem to be regulated by the CTNNB1 pathway. Similarly, CYP3A4 induction by RIF was not affected by CTNNB1 down-regulation or activation via APC siRNA silencing (Fig. 7B). PXR mRNA expression also was not affected by CTNNB1 or APC down-regulation, or incubation with Wnt3a (not shown). Moreover, expression of HNF4, another factor involved in CYP3A4 regulation (Jover et al., 2001), was not affected by CTNNB1 or APC silencing or stimulation with Wnt3a (not shown). These results indicate that CYP3A4 basal expression and its PXR-mediated induction are independent of CTNNB1 and APC in our model of PHH culture.
CYP3A4 regulation in PHHs. (A) RT-qPCR analysis of CYP3A4 expression in PHHs after incubation or not with 100 or 200 ng/ml mWnt3a or 10 µM RIF (n = 4). (B) CYP3A4 mRNA expression after CTNNB1 or APC silencing and treatment with 10 µM RIF for 24 hours. UT, untreated.
Discussion
We evaluated the impact of the WNT/β-catenin pathway, which has been described as the main zonation keeper in the liver lobule (Benhamouche et al., 2006), on P450 gene expression in PHHs. When hepatocytes are seeded in culture, zonation is lost, and the expression of WNT/β-catenin target genes and P450s is concomitantly reduced. We thus hypothesized that by restoring the activity of the WNT/β-catenin pathway we could restore P450 expression in PHHs, despite the loss of the zonal organization. Here, we show that the expression of the CTNNB1 target gene LGR5 can be restored in PHHs upon activation of the WNT/β-catenin signaling pathway. Moreover, we report for the first time that CYP2E1 expression is transcriptionally regulated by CTNNB1 and that is a true marker of CTNNB1 activity in PHHs.
The activity of the WNT/β-catenin pathway has been studied in various in vitro cell models, such as colon cancer cell lines (Verma et al., 2003), embryonic stem cells (Kielman et al., 2002; Sato et al., 2004), or isolated mouse hepatocytes (Hailfinger et al., 2006). Here, we show that the WNT/β-catenin pathway can also be reactivated in PHHs by stimuli that trigger this signaling cascade at different levels. First, we used the canonical activator WNT. However, the effect of WNT depends on the expression level of positive/negative effectors of the WNT/β-catenin pathway, which can vary with time in PHHs. For instance, the progressive increase of DKK3 expression, an antagonist of the WNT/β-catenin pathway (Veeck and Dahl, 2012), during PHH culture could explain the lower ability of mWnt3a to activate the WNT/β-catenin pathway in comparison with other compounds used in this work. Then we inhibited GSK3β by using CHIR99021, which is considered one of the most specific GSK3β inhibitors (Cohen and Goedert, 2004) and has been widely used in stem cell studies (Li et al., 2012). Finally, we down-regulated APC. APC inhibits β-catenin–dependent transcription by providing a scaffold for the destruction complex, by promoting its export from the nucleus and by reducing its interaction with TCF. Moreover, APC expression in cancer cells causes nuclear and cytoplasmic translocation of β-catenin to the cell membrane (Aoki and Taketo, 2007). The increase of APC mRNA expression in parallel with PHHs adhesion and polarization (Gondeau et al., 2014) could contribute to WNT/β-catenin pathway down-regulation in PHHs (Hanson and Miller, 2005).
CYP2E1 has been extensively studied because of its implication in many toxicologic and carcinogenic processes (Butura et al., 2009). CYP2E1 metabolizes chemicals, including APAP, carbon tetrachloride, dimethylsulfoxide, and ethanol. We demonstrate for the first time that CYP2E1 expression and metabolic activity can be restored through activation of the WNT/β-catenin pathway in PHHs, as indicated by the production of 6-OH chlorzoxazone and GSH depletion after exposure to APAP. This can thus constitute a useful model for studying the role of CYP2E1 in the hepatotoxicity induced by APAP and other drugs.
CYP2E1 protein and activity are often induced by its own substrates through posttranscriptional mechanisms (Gonzalez, 2007). Posttranscriptional regulation involves also CYP2E1 mRNA stabilization (Woodcroft et al., 2002). In parallel, CYP2E1 gene transcription is under the control of HNF1α in rat hepatocytes (Liu and Gonzalez, 1995; Woodcroft et al., 2002), or interleukin-4 in a human hepatoma cell line (Lagadic-Gossmann et al., 2000). Here, we show that CYP2E1 can also be transcriptionally regulated by the WNT/β-catenin pathway in PHHs. Specifically, CYP2E1 expression was significantly up-regulated in response to GSK3β inhibition (CHIR, LiCl) or CTNNB1 stimulation (mWnt3a, APC silencing), demonstrating that CYP2E1 is a true CTNNB1 target gene. This is in agreement with data obtained in Ctnnb1−/− mice where CYP2E1 gene expression is lost (Sekine et al., 2006; Tan et al., 2006) but its expression is increased in mouse (Loeppen et al., 2005) and human (Schmidt et al., 2011) liver tumors harboring Ctnnb1-activating mutations. Moreover, CYP2E1 expression is induced in primary mouse hepatocytes incubated in conditioned medium from mWnt3a-producing 3T3 fibroblasts (Hailfinger et al., 2006).
However, the direct effect of β-catenin on the CYP2E1 promoter and the presence of a functional TCF/Lef responsive element need to be demonstrated. Indeed, direct binding of TCF4 (coactivator for β-catenin) on the 5000-bp region spanning the CYP2E1 promoter could not be demonstrated (Liu et al., 2012). Recently, using Chip-seq analysis, Gougelet et al. (2014) identified a WRE motif in a 100-kpb region in the upstream and intragenic regions of mouse Cyp2e1 but did not demonstrate its functionality. An indirect effect via HNF1α transcriptional activity was also proposed (Gonzalez, 2006).
Moreover, we demonstrated that CYP2E1 transcriptional regulation by the WNT/β-catenin pathway is dependent on the cell context. Its expression is induced in PHHs and in differentiated HepaRG cells but not in hepatoma cell lines (HepG2-C3 and HuH7). The β-catenin pathway is constitutively active in HepG2-C3 (very high LGR5 expression) without significant expression of CYP2E1 (Fig. 5G). Moreover, HNF1α expression in HepG2-C3 cells and PHHs is comparable (Funakoshi et al., 2011).
Thus, CYP2E1 is transcriptionally regulated by CTNNB1 in a direct or indirect manner, and the presence of other coregulators is certainly needed and has to be investigated. Whether these observations obtained in PHHs can be extrapolated to the physiologic condition in human liver remains to be proven, but the fact that CYP2E1 is expressed according to the β-catenin activation in the liver lobule is in favor of this statement.
CYP1A2 expression also is induced by mWnt3a in PHHs (2- to 4-fold) and is strongly up-regulated upon APC down-regulation (100-fold), demonstrating a strong repressive activity mediated by APC. Cyp1a2 mRNA expression can be increased by mWnt3a in isolated mouse hepatocytes and its expression is reduced in Ctnnb1−/− mouse liver (Sekine et al., 2006; Braeuning et al., 2011).
CYP1A2 expression and induction are under the control of nuclear receptors and transcription factors. AhR mRNA expression in PHHs was increased by incubation with mWnt3a and after APC silencing. Centrilobular expression of AhR was previously reported (Lindros et al., 1997), and AhR expression is increased in Apc−/− and decreased in Ctnnb1−/− mouse livers (Torre et al., 2011; Gougelet et al., 2014). As AhR participates in CYP1A2 basal and induced expression, the positive effect of β-catenin activation on CYP1A2 expression could therefore be direct or indirect via AhR up-regulation, as recently suggested (Gougelet et al., 2014). AhR may therefore participate in β-catenin–mediated CYP1A2 regulation and pericentral zonation. Moreover, several studies suggest a cross-talk between β-catenin and AhR that seems to be very complex. Specifically, β-catenin may act as a coactivator of AhR in mouse hepatocytes (Braeuning et al., 2011), and AhR participates in β-catenin degradation in mouse intestine (Kawajiri et al., 2009). Moreover, sustained activation of AhR by TCDD reduces the level of β-catenin targets genes in WB-F344 cells (Prochazkova et al., 2011).
On the other hand, our results indicate that, despite its centrilobular expression in vivo (Ratanasavanh et al., 1991), CYP3A4 transcriptional regulation in PHHs is independent of β-catenin and APC expression and/or stimulation. This is in agreement with the finding that in early Ctnnb1−/− mice, Cyp3a11 mRNA expression is not affected (Sekine et al., 2006), while Cyp3a is slightly induced in male mice (Braeuning et al., 2009). Conversely, Cyp3a expression is reduced in the liver of male Ctnnb1loxP/loxP/TTR-Cre+ mice when Ctnnb1 is invalidated in adulthood by injection of tamoxifen (Ganzenberg et al., 2013). As observed for CYP3A4 induction by RIF in PHHs, Cyp3a induction by pregnenolone-α-carbonitrile was not affected in Ctnnb1−/− mice (Braeuning et al., 2009). CYP3A4 expression in PHHs might therefore be regulated through other mechanisms. Indeed, CYP3A4 mRNA expression is up-regulated by the medium flow in PHHs (Vinci et al., 2011), and the oxygen tension in rat hepatocytes (Allen and Bhatia, 2003). Alternatively, Cyp3a localization could also be regulated in a RAS/MAPK/ERK (Hailfinger et al., 2006; Braeuning et al., 2007), Dicer- (Sekine et al., 2009) or morphogen-dependent manners (Gebhardt and Hovhannisyan, 2010). Like for CYP3A4, expression of NR1I2 is independent of β-catenin and APC in PHHs, but can be moderately induced by mWnt3a in mouse hepatocytes (Braeuning et al., 2011) and is inhibited in females Ctnnb1−/− mice (Braeuning et al., 2009).
Finally, in contrast to observations in Ctnnb1−/− mice (Sekine et al., 2006), CTNNB1 silencing in PHHs did not seem to have an effect on CYP2E1 and CYP1A2 basal expression. CTNNB1 silencing by siRNAs is a 4-day-long process, and during this time the basal expression of CYP2E1 and CYP1A2 progressively decreases in culture. Therefore, CTNNB1 invalidation has no further effect. However, AhR mRNA expression, which did not change during PHH culture, was also unaffected by CTNNB1 silencing. As this gene is ubiquitously expressed in most mammalian tissues (Le Carrour et al., 2010), its basal expression is likely to be β-catenin independent, and other transcription factors might contribute to its transcriptional regulation. Similar observations were reported using rat WB-F344 progenitors cells (Prochazkova et al., 2011).
In conclusion, the Wnt/β-catenin pathway is functional in PHHs and constitutes a new regulatory network of drug metabolism in human hepatocytes. CYP2E1, CYP1A2, and AhR are transcriptionally regulated by this pathway, while CYP3A4, PXR, and HNF4α are not. Induction of the Wnt/β-catenin pathway in PHHs represents a powerful tool to evaluate hepatotoxicity of drugs that are metabolized by CYP2E1.
Authorship Contributions
Participated in research design: Gerbal-Chaloin, Daujat-Chavanieu.
Conducted experiments: Gerbal-Chaloin, Dume, Briolotti, Klieber, Raulet, Duret.
Contributed new reagents or analytic tools: Fabre, Ramos.
Performed data analysis: Gerbal-Chaloin, Daujat-Chavanieu.
Wrote or contributed to the writing of the manuscript: Gerbal-Chaloin, Maurel, Daujat-Chavanieu.
Footnotes
- Received July 18, 2014.
- Accepted September 16, 2014.
This work was supported by the Innovative Medicine Initiative Joint Undertaking (IMI JU) [Grant 115001] (MARCAR project, URL:http://www.imi-marcar.eu/).
Abbreviations
- AhR
- aryl hydrocarbon receptor
- APAP
- acetaminophen
- APC
- adenomatous polyposis coli
- CHIR99021
- CHIR, 6-((2-((4-(2,4-dichlorophenyl)-5-(4-methyl-1H-imidazol-2-yl)pyrimidin-2-yl)amino)ethyl)amino) nicotinonitrile
- CTNNB1
- β-catenin
- DKK-3
- Dickkopf-3
- GLUL
- glutamine synthetase
- GSH
- glutathione
- GSK3β
- glycogen synthase kinase 3β
- HNF4α
- hepatocyte nuclear factor-4α
- KRT19
- cytokeratin 19
- LGR5
- leucine-rich repeat containing G protein–coupled receptor 5
- P450
- cytochrome P450
- PHH
- primary human hepatocyte
- PXR
- pregnane X receptor
- RIF
- rifampicin
- siRNA
- small-interfering RNA
- TCF/Lef
- T-cell factor/lymphoid-enhancing factor transcriptional activator
- RT-qPCR
- reverse transcription real-time quantitative polymerase chain reaction
- TCDD
- 2,3,7,8-tetrachlorodibenzo-p-dioxine
- WNT
- wingless-type MMTV integration site family
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics
References
In this issue
β-Catenin Regulates Drug Metabolism
