Potential conflict of interest: Nothing to report.
This work was supported, in part, by National Institutes of Health grant nos. DK083952 and ES019629 (to W.X.) and NSFC grant no. 31071027 (to Y.Z.).
Overdose of acetaminophen (APAP), the active ingredient of Tylenol, is the leading cause of drug-induced acute liver failure in the United States. As such, it is necessary to develop novel strategies to prevent or manage APAP toxicity. In this report, we reveal a novel function of the liver X receptor (LXR) in preventing APAP-induced hepatotoxicity. Activation of LXR in transgenic (Tg) mice or by an LXR agonist conferred resistance to the hepatotoxicity of APAP, whereas the effect of LXR agonist on APAP toxicity was abolished in LXR-deficient mice. The increased APAP resistance in LXR Tg mice was associated with increased APAP clearance, increased APAP sulfation, and decreased formation of toxic APAP metabolites. The hepatoprotective effect of LXR may have resulted from the induction of antitoxic phase II conjugating enzymes, such as Gst and Sult2a1, as well as the suppression of protoxic phase I P450 enzymes, such as Cyp3a11 and Cyp2e1. Promoter analysis suggested the mouse Gst isoforms as novel transcriptional targets of LXR. The suppression of Cyp3a11 may be accounted for by the inhibitory effect of LXR on the PXR-responsive transactivation of Cyp3a11. The protective effect of LXR in preventing APAP toxicity is opposite to the sensitizing effect of pregnane X receptor, constitutive androstane receptor, and retinoid X receptor alpha. Conclusion: We conclude that LXR represents a potential therapeutic target for the prevention and treatment of Tylenol toxicity. (HEPATOLOGY 2011)
Overdose of the analgesic and antipyretic, acetaminophen (APAP), is the leading cause of drug-induced acute liver failure.1 APAP usually is well tolerated at recommended therapeutic doses, and the majority of APAP is rapidly metabolized by the phase II conjugating enzymes, UDP-glucuronosyltransferase (UGT) and sulfotransferase (SULT), in the liver to nontoxic compounds,2 which is followed by renal and biliary excretion. Another metabolic pathway is bioactivation by phase I cytochrome P450 (CYP) enzymes to the highly reactive intermediate metabolite, N-acetyl-p-benzoquinone-imine (NAPQI).3 NAPQI has a short half-life under normal conditions and is eliminated by conjugation with glutathione (GSH), a reaction carried out by glutathione S-transferase (GST), and then further metabolized to a mercapturic acid and excreted into the urine.4 In the event of APAP overdose, the glucuronidation and sulfation pathways become saturated, and increasing amounts of APAP undergo P450-mediated formation of NAPQI, as well as depletion of GSH.5 Accumulated NAPQI then binds to cellular macromolecules, leading to structural and metabolic disarray of the cells.6 Furthermore, depletion of intracellular GSH renders the hepatocytes highly susceptible to oxidative stress and apoptosis.
CYP1A2, 2E1, and 3A are the most active P450s that convert APAP to NAPQI.7 Treatment with Cyp1a2 inducers increased APAP hepatotoxicity in rodents.8 Cyp2e1 was found to activate APAP to NAPQI. The implication of Cyp2e1 in APAP toxicity is consistent with the observation that alcohol sensitized rats to fatal APAP hepatotoxicity by inducing Cyp2e1.9 In contrast, inhibition of Cyp2e1 by propylene glycol prevented APAP hepatotoxicity in mice. Cyp2e1 null mice were markedly resistant to APAP-induced lethality,10 and double-null mice lacking both Cyp1a2 and Cyp2e1 were largely resistant to APAP toxicity.11 Inducers of CYP3A potentiated, whereas inhibitors of CYP3A prevented, APAP toxicity.12, 13 For these reasons, it was proposed that inhibitors of P450 enzymes may be of therapeutic value for the treatment of APAP hepatotoxicity.14
At subtoxic doses, NAPQI is inactivated by GST-mediated GSH conjugation, leading to the conversions of NAPQI to APAP cysteine and mercapturate conjugates.4 Treatment of rodents with oltipraz, a GST inducer, was linked to chemopreventive effects against APAP toxicity.15 Among GST isozymes, GST Pi was thought to be particularly important to detoxify NAPQI, based on in vitro conjugation assays.16 However, mice deficient of Gstπ showed a surprisingly increased resistance to APAP hepatotoxicity,17 indicating that Gstπ may not contribute to the formation of GSH conjugates of NAPQI in vivo and could enhance APAP toxicity by accelerating the depletion of GSH. These data suggest that suppression of Gstπ and/or induction of other Gst enzymes may protect mice from APAP-induced hepatotoxicity.
The liver X receptors (LXRs), LXRα and LXRβ, were isolated as sterol sensors.18, 19 Subsequent characterization revealed that LXRs have diverse physiological functions, ranging from cholesterol18 and lipid metabolism20 to anti-inflammation,21 hepatobiliary diseases,22, 23 and steroid hormone homeostasis.24, 25 We have previously reported that the expression of APAP-detoxifying Sult2a1/2a9 was positively regulated, whereas the expression of protoxic Cyp3a11 was reduced in LXR-activated mice.22 We thus hypothesized that LXR may affect APAP toxicity by regulating the APAP-metabolizing enzymes.
In this study, we showed that activation of LXR relieved APAP-induced hepatotoxicity. The benefits of LXR in preventing APAP toxicity may have resulted from a pattern of metabolic gene regulation that favored a decreased exposure of the host to the parent APAP as well as the toxic APAP metabolites.
ALT, alanine aminotransferase; APAP, acetaminophen; APAP-CYS, APAP-cysteine; APAP-G, APAP-glucuronide; APAP-MER, APAP-mercapulate; APAP-S, APAP-sulfate; AST, aspartate aminotransferase; AUC, area under the curve; CAR, constitutive androstane receptor; CDNB, 1-chloro-2,4-dinitrobenzene; CYP, cytochrome P450; DKO, double knockout; FABP, fatty-acid–binding protein; FXR, farnesoid X receptor; GSH, glutathione; GST, glutathione S-transferase; H&E, hematoxylin and eosin; IP, intraperitoneal; kb, kilobase; LXR, liver X receptor; NAPQI, N-acetyl-p-benzoquinone-imine; PCN, pregnenolone-16α-carbonitrile; PXR, pregnane X receptor; RT-PCR, reverse-transcriptase polymerase chain reaction; RXR, retinoid X receptor; SD, standard deviation; SULT, sulfotransferase; Tg, transgenic; UGT, UDP-glucuronosyltransferase; VP, viral protein 16; Wt, wild type.
Materials and Methods
The creation of fatty-acid–binding protein/viral protein 16 (FABP-VP)-LXRα22 transgenic (Tg) mice and pregnane X receptor (PXR)−/− mice26 has been previously described. The LXRα and β double-knockout (LXR DKO) mice27 were kindly provided by Dr. David Mangelsdorf. The use of mice in this study has complied with all relevant federal guidelines and institutional policies.
APAP Dosing and Hepatotoxicity Analysis.
Mice were fasted overnight before being given an oral administration of APAP (freshly prepared in 0.5% methyl cellulose) at 200 mg/kg. When necessary, wild-type (Wt), LXR DKO, and PXR−/− mice were dosed with the LXR agonist, TO1317 (10 mg/kg, intraperitoneal; IP) or vehicle for 1 week before APAP treatment. Mice were sacrificed 24 hours after APAP administration, and blood and liver tissues were harvested for histology by hematoxylin and eosin (H&E) staining and biochemical analysis by ANTECH Diagnostics (Lake Success, NY).
Northern Blotting and Real-Time Reverse-Transcriptase Polymerase Chain Reaction Analyses.
Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA). Northern hybridization was carried out as previously described.25 In real-time reverse-transcriptase polymerase chain reaction (RT-PCR) analysis, RT was performed with the random hexamer primers and the SuperScript RT III enzyme (Invitrogen). SYBR Green-based real-time PCR was performed with the ABI 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA). Data were normalized against the control of cyclophilin signals. The sequences of real-time PCR primers are listed in Supporting Table 1.
APAP Pharmacokinetic and Metabolic Analyses.
See Supporting Methods.
Measurement of GST Activity.
Total GST activity was measured using 1-chloro-2,4-dinitrobenzene (CDNB) as a substrate as previously described.28 Briefly, GSH (1 mM final) was added to GST (0.5-5 μg of enzyme) in 1 mL of 100 mM of potassium phosphate buffer (pH 6.5). After preincubation for 3 minutes, CDNB (1 mM final) was added, and activity was measured using a spectrophotometer at 25°C. The increase of absorbance resulting from the conjugation of dinitrophenyl with GSH was recorded at 340 nm every 15 seconds for 3 minutes.
Cloning of Mouse Gst Promoters and Luciferase Reporter Gene Assay.
The 2.2-kb (kilobase) mouse Gstμ1 promoter was cloned by PCR using the following primer pair: Gstμ1-p5: 5'-ATCGACGCGTCCTCCAGACCCCAGCTAACTGTG-3', and Gstμ1-p3: 5'-ACCGCTCGAGCTGTGGTCTTCTCAAACTGGCTTCAG-3'. The PCR products were digested with MluI and XhoI and cloned into the same enzyme-digested pGL3-Basic vector from Promega (Madison, WI). The 1.9-kb mouse Gstπ1 promoter was cloned by PCR using the following primer pair: Gstπ1-pF: 5'-ACGGGGTACCACCCCAACTCTCTCATACACAC-3', and Gstπ1-pR: 5'-ACCGCTCGAGTAGACAGAGGGGTACTCAGAGTG-3'. The PCR products were digested with Asp718 and XhoI and cloned into the same enzyme-digested pGL3-Basic vector. To generate the tk-Gstμ1/DR4 reporter, two copies of the sequence 5'-TGTCCTaggaTGAGATggccTCGGCT-3', or its mutant variant, 5'- GGGCCCaggaAGATTTggccTCGGCTT-3', were synthesized and cloned into the tk-Luc reporter gene. The tk-Cyp3a11 report gene contains three copies of the DR-3 type PXRE (TGAACTataCGAACT) found in the mouse Cyp3a11 gene promoter. HepG2 cells were transfected on 48-well plates as previously described.29 When applicable, transfected cells were treated with drugs for 24 hours before harvesting for luciferase and β-gal assays. The transfection efficiency was normalized against the β-gal activity.
Results are expressed as means ± standard deviation (SD). Statistical analysis was performed using the unpaired Student's t test for the comparison between two groups.
Activation of LXR in Tg Mice Conferred Resistance to APAP Hepatotoxicity.
We have previously reported the creation of Tg mice that express the activated LXRα (FABP-VP-LXRα) in the liver (Fig. 1A).22 Created by fusing the VP activation domain of the herpes simplex virus to the aminoterminal of LXRα, VP-LXRα shares the same DNA-binding specificity as its Wt counterpart and constitutively activates LXR target gene expression.22 Gene-expression analysis revealed that activation of LXR in the liver regulated the expression of multiple drug-metabolizing enzymes and transporters,22 which prompted us to examine whether Tg mice had an altered sensitivity to APAP. In vehicle-treated mice, no significant alteration in liver histology was found in Tg mice, when compared to Wt mice (Fig. 1B). After APAP treatment, the Wt liver showed expected typical necrotic liver damage. In contrast, Tg mice showed little signs of liver damage (Fig. 1C), suggesting that this genetic activation of LXR conferred resistance to APAP hepatotoxicity.
Consistent with their reduced histological liver damage, APAP-treated Tg mice showed improved serum chemistry, compared to their Wt counterparts. These included lower serum levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activities, total bilirubin, and alkaline phosphatase (Fig. 1D). Vehicle-treated Tg mice showed little signs of hepatotoxicity, as judged by serum chemistry (Supporting Fig. 1).
Treatment With the LXR Agonist, TO1317, Relieved Mice From APAP Toxicity in an LXR-Dependent Manner.
The hepatoprotective effect of the LXRα transgene prompted us to determine whether a pharmacological activation of LXR would have a similar effect in preventing APAP toxicity. Indeed, the TO1317-treated Wt C57BL/6J mice showed less histological liver damage (Fig. 2A) and lower serum level of AST and ALT (Fig. 2B), compared to their vehicle-treated counterparts.
We then used LXR DKO mice to determine whether the hepatoprotective effect of TO1317 was LXR dependent. The protective effect of TO1317 was abolished in LXR DKO mice, as indicated by the lack of relief in histological liver damage (Fig. 2C), as well as the serum levels of AST and ALT (Fig. 2D) in APAP- and TO1317-treated LXR DKO mice. These results demonstrated that the APAP protective effect of TO1317 required LXRs.
Because TO1317 was also reported to activate PXR,30 we then used PXR−/− mice to determine whether the hepatoprotective effect of TO1317 required PXR. In this experiment, PXR−/− mice were treated with APAP in the presence or absence of TO1317. The hepatic necrosis (Fig. 2E) and serum levels of AST and ALT (Fig. 2F) were both decreased in TO1317-treated PXR−/− mice, compared to their vehicle-treated counterparts, suggesting that the protective effect of TO1317 was independent of PXR.
Activation of LXR in Transgenic Mice Increased APAP Clearance and Decreased the Formation of Toxic APAP Metabolites.
Resistance to APAP toxicity in Tg mice suggested that activation of LXR may promote APAP clearance and/or inhibit the formation of toxicity-indicating metabolites. To test this hypothesis, we examined the in vivo metabolism of APAP. Mice were given a single IP injection of APAP and collected for blood or urine. The pharmacokinetic estimations for the serum level of APAP, and APAP-sulfate (APAP-S) and APAP-glucuronide (APAP-G), are summarized in Table 1 and Fig. 3A, respectively. The decrease in area under the curve (AUC), increase in clearance, and decrease in half-life of parent APAP in Tg mice (Table 1) suggested that activation of LXR reduced the animal's total exposure to the parent drug, which was associated with an increased production of APAP-S (Fig. 3A). The glucuronide metabolite of APAP was unchanged. When the urinary levels of APAP metabolites were measured, we found that the level of APAP-S was increased, whereas the level of APAP-G was unchanged in Tg mice (Fig. 3B). Urinary concentrations of APAP-cysteine (APAP-CYS) and APAP-mercapulate (APAP-MER), two APAP metabolites that indicate the formation of toxic metabolites, were decreased in Tg mice (Fig. 3B).
Table 1. Pharmacokinetic Estimation for the Serum Level of Parent APAP
Values (Mean ± SD)
Abbreviations: APAP, acetaminophen; SD, standard deviation; Wt, wild type; Tg, transgenic; AUC, area under the curve; Cl, clearance; t1/2, half-life.
To understand the mechanism by which activation of LXR relieved APAP toxicity, we measured the messenger RNA expression of major APAP-metabolizing enzymes in Wt and Tg mice. Among phase I enzymes known to facilitate the formation of toxic APAP metabolites, the expression of Cyp3a11 and 2e1 was reduced, whereas the expression of Cyp1a2 remained largely unchanged in Tg mice, as determined by northern blotting analysis (Fig. 4A). The same pattern of P450 regulation was confirmed by real-time PCR analysis (Supporting Fig. 2). The inhibition of Cyp3a11 and 2e1 was consistent with the decreased formation of toxic APAP metabolites in Tg mice.
Among phase II enzymes, the expression of Gstπ and Gstμ was decreased and increased, respectively (Fig. 4A). The expression of Sult2a1 was induced as expected.26 Among other Sult enzymes, the expression of Sult1d1 and Sult1e1 was also induced, whereas the expression of Sult1a4 and 1b3 was unchanged. Papss2, the primary hepatic enzyme that catalyzes the formation of the sulfonyl group donor, PAPS, was also induced (Fig. 4A). The expression of Ugt1a1 and 1a6 was unaffected (Fig. 4A), consistent with the observation that the APAP-G level was unchanged in Tg mice. Consistent with the pattern of Gst and Sult gene regulation, the liver extract of Tg mice showed increased enzymatic activities of Gst (Fig. 4B) and Sult (data not shown).22
The regulation of Gst and Sult2a1 was confirmed in Wt mice treated with TO1317. The expression of Gst α1, α2, μ1 and μ2, and Sult2a1 was increased, whereas the expression of Gstπ was decreased in TO1317-treated Wt mice (Fig. 4C). The TO1317 effect on the expression of Gst and Sult2a1 was independent of PXR, because a similar pattern of gene regulation was observed in TO1317-treated PXR−/− mice (Fig. 4D).
The Mouse Gst Gene Promoters Are Likely Transcriptional Targets of LXR.
To understand the mechanism by which LXRs regulate Gst, we cloned the mouse Gstμ1 and Gstπ1 gene promoters and characterized their regulation by LXR. The 2.2-kb Gstμ1 promoter report gene, pGL-Gstμ1, was activated by the cotransfection of LXRα, and this activation was enhanced by the addition of LXR agonist 22(R)-hydroxycholesterol or GW3965 (Fig. 5A). Inspection of the Gstμ1 gene promoter revealed several putative DR-4 type LXR response elements. A synthetic reporter, tk-Gstμ1/DR4, that contained two copies of two overlapping DR-4 sites were activated by the cotransfected LXRα in a ligand-dependent manner, whereas the transactivation was abolished when both DR-4s were mutated (Fig. 5B). In contrast, the 1.9-kb Gstπ1 promoter report gene, pGL-Gstπ1, was suppressed by the cotransfection of LXRα (Fig. 5C). The same Gstπ1 reporter gene was activated by the cotransfection of Nrf2, a known positive regulator of Gstπ.31
LXRα Inhibited the Transcriptional Activity of PXR.
To understand the mechanism by which LXR suppressed Cyp3a11 gene expression, we used transient transfection and reporter gene assay to determine whether LXRα could inhibit the transcriptional activity of PXR, the primary transcriptional regulator of Cyp3a11.26 Cotransfection of LXRα inhibited the PXR ligand, pregnenolone-16α-carbonitrile (PCN), induced the activity of PXR on tk-Cyp3a11, a reporter gene that contains the PXR response element found in the Cyp3a11 gene promoter. LXRα alone had little effect on the reporter gene activity, regardless of the treatment of GW3965. These results provided a plausible mechanism by which LXR suppressed the expression of Cyp3a11.
Mounting evidence has suggested that several liver-enriched nuclear receptors, including CAR,32 PXR,12, 33 RXRα,34 and farnesoid X receptor (FXR),35 play pivotal roles in APAP metabolism and toxicity. The nuclear receptor effects on APAP toxicity and their proposed mechanisms are summarized in Table 2. Activation of CAR or PXR was shown to heighten APAP hepatotoxicity. Treatment of mice with the CAR activator, phenobarbital, induced the expression of Cyp1a2 and Cyp3a11 and resulted in increased sensitivity to APAP.32 In contrast, administration of androstanol, a CAR antagonist, blocked hepatotoxicity in Wt mice. CAR−/− mice were resistant to APAP toxicity.32 Pretreatment with PCN, a potent PXR agonist, enhanced APAP hepatotoxicity in Wt, but not in PXR−/−, mice.33 The heightened sensitivity in PCN-treated Wt mice was reasoned to be the result of the induction of Cyp3a enzymes. In contrast, after PCN treatment, PXR−/− mice had lower Cyp3a11 expression, decreased NAPQI formation, and increased maintenance of hepatic GSH content.33 Using mice humanized for both PXR and CYP3A4, Cheng et al. showed that rifampicin-activated hPXR and CYP3A4 induction enhanced APAP toxicity,12 suggesting the human relevance of this effect. Loss of RXRα, the shared heterodimerization partner of CAR and PXR, protected mice from APAP toxicity primarily by regulating the expression of Gst enzymes.34
Table 2. Summary of Nuclear Receptor Effects on APAP Toxicity
Effect on APAP Toxicity When Activated
Abbreviations: CAR, constitutive androstane receptor; PXR, pregnane X receptor; RXR, retinoid X receptor; FXR, farnesoid X receptor; LXR, liver X receptor; APAP, acetaminophen; NAPQI, N-acetyl-p-benzoquinone-imine; GSH, glutathione.
(1) Induction of phase I enzymes Cyp1a2 and Cyp3a11, which can convert APAP to cytotoxic NAPQI; (2) induction of phase II enzyme Gstπ, which may deplete GSH used for NAPQI detoxification
Zhang et el., 2002
(1) Induction of Cyp3a and Cyp1a2; (2) enhanced depletion of GSH
Cheng et al., 2009; Guo et al., 2004
Regulation of Gst expression and APAP-GSH conjugation
Dai et al., 2005
(1) Induction of the Gstα class; (2) induction of Gclm and Gpx1 that are involved in GSH homeostasis
Lee et al., 2010
(1) Suppression of Cyp3a11 and Cyp2e1; (2) induction of the detoxifying phase II enzyme, Sult2a1; (3) isoform-specific Gst regulations that include induction of Gstα and Gstμ and suppression of Gstπ; (4) increased APAP clearance and decreased the formation of toxic APAP metabolites
Our current results showed that unlike CAR and PXR, activation of LXR was beneficial in relieving APAP hepatotoxicity. The hepatoprotective effect of LXR may have resulted from the combined suppression of protoxic P450s and induction of antitoxic phase II enzymes Gst and Sult. Suppression of Cyp3a11 by LXR was opposite to the induction of the same enzyme in CAR/PXR-activated mice.32, 33 Induction of Cyp1a2, observed in CAR/PXR-activated mice,32, 33 was absent in LXR Tg mice. Suppression of Cyp3a by LXR was previously reported,22 which was proposed to be the result of the cross-suppression of CAR by LXR.36 We now provide evidence suggesting that LXR may also suppress Cyp3a11 by antagonizing the positive regulation of Cyp3a11 by PXR. The suppression of Cyp2e1 by LXR has not been reported. Cyp2e1 is better known for its post-transcriptional regulation. LXR has recently been shown to regulate the E3 ubiquitin ligase-inducible degrader of the LDLR (Idol).37 It remains to be determined whether LXR can regulate the expression or activity of Cyp2e1 through a post-transcriptional mechanism.
Among the LXR responsive phase II enzymes, the activation of Sult2a1 gene expression and lack of Ugt1a1 regulation by LXR have been reported.22 The isoform-specific regulation of Gst was intriguing. We reasoned the combined induction of Gstα and Gstμ classes and suppression of Gstπ may have contributed to the hepatoprotective role of LXR. The suppression of Gstπ in LXR-activated mice was consistent with the previous report that mice that lacked Gstπ were resistant to APAP hepatotoxicity.17 In contrast, an induction of Gstπ in CAR-activated mice was associated with the sensitizing effect.32 Our promoter analysis suggested that Gstμ1 and Gstπ1 gene promoters were positively and negatively regulated by LXR, respectively. The induction of Gstα and Gstμ isoforms was reminiscent of the effect of FXR, whose activation has recently been linked to protection against APAP-induced hepatic toxicity.35
In summary, the current study demonstrated that LXR may represent a potential therapeutic target for the prevention and treatment of APAP overdoses via induction of APAP-detoxifying/clearance enzymes and suppression of protoxic P450 enzymes.
The authors thank Dr. David Mangelsdorf for LXR DKO mice and Dr. Song Li for synthesizing TO1317.