Pregnane X receptor (PXR, NR1I2) is a ligand-activated nuclear receptor that upon activation forms a heterodimer with retinoid X receptor (RXR) and regulates the expression of a wide range of DMEs (Tolson and Wang, 2010). Apart from being activated by xenobiotics, it also responds to endobiotics including bile acids (Xie et al., 2001) and steroid hormones (di Masi et al., 2009). In mice, PXR activation results in hepatic steatosis due to enhanced sterol regulatory element-binding protein-1c (SREBP-1c)-independent lipogenesis, decreased β-oxidation, and increased uptake of fatty acids via fatty acid translocase (FAT/CD36) activation (Zhou et al., 2006b). The activation of CD36 by PXR in mice occurs directly or via the activation of peroxisome proliferator-activated receptor gamma (PPARγ; Tontonoz et al., 1998; Zhou et al., 2006b). PXR also plays a role in glucose metabolism (Gao and Xie, 2012). It inhibits gluconeogenesis by inactivating hepatocyte nuclear factor-4 (HNF-4) and forkhead box protein O1 (FOXO1), both of which are positive regulators of gluconeogenic genes (Bhalla et al., 2004; Kodama et al., 2004). Genetic association studies in a European NAFLD population indicated an association of the PXR rs2461823[A] and rs7643645[G]-containing haplotypes and disease severity, irrespective of BMI and homeostatic model assessment (HOMA) index (Sookoian et al., 2010). Although PXR activation increases steatosis, PXR-dependent counteraction of inflammation by inhibition of nuclear factor kappa-B (NFκB) has also been documented in human and mouse samples (Zhou et al., 2006a). In support of the previous statement, rats treated with a PXR activator pregnenolone-16α-carbonitrile (PCN) displayed reduced hepatic fibrosis and necrosis in response to a fibrogenesis-inducing agent carbon tetrachloride (CCl₄; Marek et al., 2005). It is thus plausible that PXR polymorphisms associated with NAFLD may affect disease severity by lowering PXR activity, thus resulting in increased susceptibility to NASH. Another link between PXR and NAFLD is through the modulation of DMEs. PXR induces the expression of CYP2C9 (Gerbal-Chaloin et al., 2002), which metabolizes the anti-diabetic drug rosiglitazone known to reduce liver fat. Since rosiglitazone is used in NAFLD patients with hyperglycemia and IR (Ratziu et al., 2008), patients with PXR gene variants may theoretically suffer from aberrant rosiglitazone metabolism. PXR is thus a potential pharmacogenetic marker for thiazolidinedione treatments. Furthermore, PXR is a strong modulator of CYP3A4, the major phase I DME in humans. Several SNPs that affect the expression of CYP3A4 reside within the PXR coding, non-coding, and promoter regions (Zhang et al., 2008). Since many NAFLD patients are treated with drugs metabolized by CYP3A4, further pharmacogenetic evaluation of patients with these PXR variants is required (Table 1).

Constitutive androstane receptor (CAR, NR1I3) is also a key regulator of xenobiotic and endobiotic metabolism (Moore et al., 2000; Handschin and Meyer, 2005). Upon ligand activation, CAR is translocated to the nucleus where it binds to DNA elements of DME genes such as CYP2B6, CYP3A4 CYP2Cs, and others (Sueyoshi and Negishi, 2001; Gerbal-Chaloin et al., 2002; Faucette et al., 2006; Chen and Goldstein, 2009), as well as phase II enzymes involved in glucuronidation (Sugatani et al., 2005), sulfation, and drug transport (Tolson and Wang, 2010). Several studies also indicate a role of CAR in energy homeostasis (Wada et al., 2009). Hence, the activation of CAR for regulation of energy homeostasis may affect drug metabolism (Hafner et al., 2011). SREBP-1c, which is upregulated in hepatic steatosis, inhibits CAR and thus, may further contribute to aberrant xenobiotic and endobiotic metabolism (Roth et al., 2008). Car+/+ mice fed with MCD diet, known to induce NASH (Rinella et al., 2008), developed enhanced liver fibrosis due to lipid peroxidation, inducible nitric oxide synthase (iNOS), and increased CYP induction compared to Car−/− mice (Yamazaki et al., 2007). However, no difference in hepatic lipid accumulation was observed between Car+/+ and Car−/− mice, indicating that CAR may be involved in later stages of NAFLD progression and hepatocarcinogenesis (Takizawa et al., 2011). Furthermore, observations in Car−/− mice and human hepatocytes highlight the role of CAR activation in the induction of aberrant hepatic de novo lipogenesis and insulin resistance by enhancing the expression of thyroid hormone-responsive spot 14 protein (THRSP; Anderson et al., 2009; Breuker et al., 2010). Activation of CAR in mice with 1,4-Bis [2-(3,5-dichloropyridyloxy)] benzene (TCPOBOP) increased serum triglycerides and hepatic fatty acid synthesis and repressed adaptation to hyperlipidemia, which is expected to favor the development of NAFLD (Rezen et al., 2009). Contradictory observations arise from other mouse models, where CAR has been linked to improved fatty liver due to decreased lipogenesis, increased β-oxidation, improved glucose tolerance, and insulin sensitivity (Dong et al., 2009; Gao et al., 2009). In mice, CAR also regulates cholesterol and bile acid metabolism by lowering plasma high-density lipoprotein (HDL) and reverse cholesterol transport, possibly via downregulation of apolipoprotein A1 (ApoA1; Masson et al., 2008). CAR polymorphisms have not yet been linked to metabolic diseases; however, its role in glucose and lipid metabolism and its functional redundancy with PXR highlights that the CAR gene is an important candidate for NAFLD association studies (Rezen et al., 2009).

Farnesoid X receptor (FXR) and Liver X receptor (LXR) are not major regulators of xenobiotic metabolism, but they play an important role in the metabolism of cholesterol and bile acids (Rezen et al., 2011). FXR/NR1H4 is the predominant regulator of bile acid synthesis and secretion, thereby lowering hepatic cholesterol levels. The first and rate-limiting step of bile acid synthesis from cholesterol is catalyzed by cholesterol 7α-hydroxylase (CYP7A1). The activation of FXR in primary human and mouse hepatocytes results in decreased transcription of CYP7A1 due to an indirect negative feedback mechanism (Goodwin et al., 2000; Holt et al., 2003). Furthermore, activated FXR upregulates the expression of CYP3A4, which hydroxylates some common bile acids into more soluble forms (Gnerre et al., 2004), as well as the bile salt export pump (BSEP; Ananthanarayanan et al., 2001; Plass et al., 2002; Song et al., 2008). Deficiency of FXR in an Ldlr−/− mouse model of hypercholesterolemia on high-fat diet results in features of NASH, such as macrosteatosis, hepatocyte ballooning, and inflammation (Kong et al., 2009). A common SNP, FXR*1B (-1T), has been identified in the sequence flanking the start codon of FXR in European, African, Chinese, and Hispanic-American populations. It associates with decreased FXR expression and functionality, which may contribute to the pathogenesis of metabolic disorders (Marzolini et al., 2007). However, association of this FXR polymorphism with NAFLD in humans has not yet been identified.

Liver X receptor plays important roles in lipid biosynthesis as well as cholesterol and bile acid homeostasis (Handschin and Meyer, 2005; Rezen et al., 2011). Typical activators of LXR are oxysterols such as 22(R)-hydroxycholesterol (22(R)-HC), 24(S), 25-epoxycholesterol, and 25-hydroxycholesterol (Jakobsson et al., 2012). LXR activation not only increases cholesterol catabolism but also hepatic lipogenesis through activation of SREBP-1c, PPARγ, or carbohydrate response element-binding protein (chREBP; Lee et al., 2008). LXR and PXR share many target genes (Boergesen et al., 2012) and both regulate the uptake of fatty acids into hepatocytes via expression of FAT CD36 (Zhou et al., 2006b). Although LXR activates Cyp7a1 expression during bile acid synthesis in rodents, it does not have an effect on human CYP7A1 expression (Goodwin et al., 2003). In humans, the LXRα isoform is mainly expressed in tissues involved in lipid metabolism, where it regulates the expression of cholesterol transporter genes, such as the ABC transporters ABC1 (Schwartz et al., 2000), ABCG1 (Sabol et al., 2005), ABCG5 (Repa et al., 2002), and ABCG8 (Repa et al., 2002). LXRβ, the ubiquitously expressed isoform and the only isoform present in pancreatic β-cells, does not play a role in cholesterol homeostasis (Alberti et al., 2001) but has a crucial role in glucose tolerance, insulin secretion, and adipocyte size (Efanov et al., 2004; Gerin et al., 2005). The rs17373080[G] polymorphism in LXRβ associates with a 20–30% lower risk of T2D but with a higher risk of obesity, as observed in two independent studies (Dahlman et al., 2009; Solaas et al., 2010). This is in accordance with data on Lxrβ-null mice that display a lean phenotype with glucose intolerance (Gerin et al., 2005). According to the best of our knowledge, LXR polymorphisms have not yet been linked to NAFLD.

Peroxisome proliferator-activated receptors (PPARs) are transcription factors that are activated by endogenous ligands, such as fatty acids, and synthetic ligands, such as the hypolipidaemic fibrates and the insulin-sensitizing thiozolidinediones. Like PXR and CAR, they form heterodimers with RXR and transactivate numerous target genes with vital roles in metabolism by binding to PPAR response elements (PPRE; Nielsen et al., 2008; van der Meer et al., 2010). The identified subtypes PPARα, PPARγ, and PPARβ/δ have different tissue-specificities and functions (Kallwitz et al., 2008). Whilst PPARγ is highly expressed in adipose tissue and functions in adipocyte differentiation, PPARα functions as a major regulator of lipid and glucose metabolism in the liver. PPARβ/δ is ubiquitously expressed with a wide array of functions. PPARα-agonists, the fibrates, are utilized to treat patients with elevated plasma triglycerides (Sirtori and Franceschini, 1988) due to the ability of PPARα to activate fatty acid oxidation and hepatic lipid hydrolysis by regulating acyl CoA synthetase (Acs), carnitine palmitoyl transferase I (Cpt-I), mitochondrial β-oxidation enzymes, and hepatic lipases in addition to downregulating apoCIII and decreasing hepatic triglyceride secretion (Kersten et al., 1999; Pyper et al., 2010; Rakhshandehroo et al., 2010). Accordingly, Pparα-deficient mice develop hepatic steatosis on a high-fat diet (Abdelmegeed et al., 2011).

A case-control study of NAFLD patients highlighted a potentially protective role for the Val227Ala variant of PPARα against obesity compared to subjects with the wild-type receptor (Chen et al., 2008a). The adipocyte differentiation regulator, PPARγ, plays an important role in lipid homeostasis and insulin sensitivity by enhancing fatty acid and insulin-dependent glucose uptake in adipose tissue (Kallwitz et al., 2008). Moreover, novel PPARγ agonists acting mainly on adipose PPARγ prevent formation of steatotic livers in mice by improving insulin resistance, upregulating adiponectin, and downregulating leptin expression and secretion (Zheng et al., 2011). PPARγ is expressed at low levels in the liver but is upregulated in rodent fatty livers, contributing to hepatic triglyceride accumulation with a protective effect to dyslipidemia and insulin resistance in other tissues (Gavrilova et al., 2003). Upregulation of hepatic PPARγ is also observed in obese NAFLD patients (Pettinelli and Videla, 2011).

PPARα affects the expression of several phase II enzymes such as SULTs and UGT (Runge-Morris and Kocarek, 2009) as well as of CYP3A4 and several other CYPs in humans (Rakhshandehroo et al., 2009; Klein et al., 2012). Contrasting observations have been made regarding the regulation of DMEs by PPARα in human versus mouse. While treatment of human primary hepatocytes with the selective PPARα agonist WY14 643 resulted in the activation of several drug metabolizing CYPs including CYP3A4, CYP2B6, CYP2C8, and CYP1A2, none of the mouse gene orthologs were regulated (Rakhshandehroo et al., 2009). Downregulation of CYP3A4 in the presence of PPARα variants that result in decreased hepatic PPARα protein levels was also found by genetic association analysis and confirmed in a human atorvastatin volunteer study (Klein et al., 2012). Moreover, activation of PPARα by agonists downregulates the expression of representatives genes of the Sult 1, 2, 3, and 5 families, specifically in female rats (Alnouti and Klaassen, 2008). However, in human hepatocytes activation of PPARα resulted in the upregulation of SULT2A1 via a functional PPRE, further emphasizing species and gender differences in the functionality of PPARα (Fang et al., 2005). Protein and mRNA levels of UGT1A1 increased upon PPARα activation in rat and human hepatocytes (Jemnitz et al., 2000; Richert et al., 2003). Additionally, UGT1A3, UGT1A4, and UGT1A6 are upregulated in human hepatocytes and transgenic mice carrying the human UGT1 locus (Senekeo-Effenberger et al., 2007). UGT2B4 is also enhanced after treatment of human hepatocytes with PPARα agonists (Barbier et al., 2003). Functional PPREs have been identified in the 5′-flanking regions of UGT1A1, UGT1A3, UGT1A6, and UGT2B4 genes, thus providing evidence that these genes are direct targets of PPAR (Barbier et al., 2003; Senekeo-Effenberger et al., 2007). The widespread use of drugs metabolized by SULTs and UGTs such as hormonal contraceptives, acetaminophen, β₂-adrenergic agonists, anti-depressants, and non-steroidal anti-inflammatory drugs highlights the implications of altered PPARα activation on xenobiotic metabolism in NAFLD patients.

Nuclear Factor Erythroid 2-related factor 2 (NRF2) is a transcription factor that responds to oxidative/electrophilic stimuli by releasing from its repressor Kelch-like ECH associating protein 1 (Keap1) in the cytosol, translocating to the nucleus, binding to antioxidant response elements (AREs) upstream of numerous phase II DME genes, and genes involved in redox balance and oxidative stress response [e.g., heme oxygenase-1 (HO-1), NAD(P)H:quinone oxidoreductase-1 (NQO1)] and activating their transcription (Wu et al., 2012). NRF2 also regulates the glutathione synthesis enzyme, glutamate cysteine ligase catalytic (GCLC). The expression of a canalicular biliary efflux transporter, multidrug resistance protein 2 (MRP2/ABCC2) and sinusoidal transporters, MRP3 and MRP4 is also regulated by NRF2 in mouse liver and HepG2 cells in response to oxidative stress and xenobiotics, thus providing further evidence that phase II enzymes and efflux transporters are regulated simultaneously (Vollrath et al., 2006; Aleksunes et al., 2008). MRP2 is involved in the excretion of reduced and oxidized glutathione and hence plays an important role in detoxification and against oxidative stress. Nrf2-null mice on MCD diet exhibit increased hepatic steatosis accompanied by inflammation and oxidative stress (Chowdhry et al., 2010). Similarly, in rats on MCD diet, the NRF2-dependent genes involved in the oxidative stress response were upregulated (Lickteig et al., 2007). Thus, NRF2 appears to have a crucial role in the pathogenesis of NAFLD.

Thus, it is evident that as many of the lipids that accumulate in obesity and steatosis, such as fatty acids, cholesterol, or bile acids, are endogenous ligands of nuclear receptors, their deregulation may not only exacerbate the deregulated metabolic processes in NAFLD patients but also result in deregulated xenobiotic metabolism.

The CYP3A sub-family of DMEs plays a predominant role in the metabolism of statins. Statins, in monotherapy and in combination with other lipid-lowering drugs or antioxidants, are beneficial in NAFLD patients by improving dyslipidemia (Athyros et al., 2011; Fon Tacer and Rozman, 2011). The inter-individual variability in the response to statins varies in NAFLD patients based on their risk for cardiovascular diseases (Maroni et al., 2011). Moreover, the CYP3A4 drug metabolizing activity is also a factor influencing inter-individual variability and hence, is relevant to NAFLD patients undergoing statin therapy. The level of CYP3A protein correlates negatively with the severity of steatosis in humans (Kolwankar et al., 2007). No changes were found in the CYP3A4 mRNA level in human fatty liver samples at various stages of NAFLD progression, however a trend of decreasing activity and protein levels was observed (Fisher et al., 2009b). In another study, CYP3A4 activity significantly decreased in macrosteatotic fatty livers and cultured human hepatocytes treated with fatty acids (Donato et al., 2006, 2007). CYP3A4 also displays sexual dimorphism with approximately twofold elevated expression in premenopausal women (Wolbold et al., 2003), who display a more favorable lipid profile compared to men (Williams, 2004). An intron 6 polymorphism in CYP3A4 (rs35599367[T]) results in decreased expression and activity of CYP3A4, with carriers of the T allele requiring significantly lower doses of statins (Elens et al., 2011; Wang et al., 2011). In accordance with the reduced CYP3A4 expression in NAFLD, studies to determine the association of the rs35599367 CYP3A4 polymorphism in NAFLD cohorts will enable the elucidation of statin dose selection in these patients. Moreover, genetic variants in other factors implicated in NAFLD, endobiotic, and xenobiotic metabolism such as PXR, PPARα, and POR have also been associated with altered CYP3A4 expression and activity (Zhang et al., 2008; Gomes et al., 2009; Klein et al., 2012). These studies emphasize the high level of variability in responses to statin treatments and may provide a basis for dose selection in NAFLD patients based on CYP3A4 status (Table 2).

CYP2C9 is the most abundant CYP of the CYP2C sub-family in human liver microsomes, accounting for the metabolism of a large number of clinically important drugs, especially some with a narrow therapeutic index, such as warfarin. The expression of CYP2C9 is coordinated by nuclear receptors such as CAR and PXR in association with nuclear factors and coactivators such as hepatocyte nuclear factor-4 alpha (HNF-4α) and PPARγ coactivator-1 alpha (PGC-1α), which is also involved in energy homeostasis (Chen and Goldstein, 2009). CYP2C9 has been closely associated with adverse drug reactions. Its mRNA and enzyme activity increase with NAFLD progression, hypoxia, and at later stages of NASH in humans (Fisher et al., 2009b). Previous observations linking CYP2C9 with arachidonic acid metabolism and vasocontraction in hypoxic conditions (Pokreisz et al., 2006) may possibly provide an explanation for elevated CYP2C9 in progressive NAFLD. Approximately 50 variants have been identified in the CYP2C9 gene to date, with the CYP2C9*2 and CYP2C9*3 loss-of-function alleles as the most important. Heterozygotes and homozygotes for these polymorphisms are common in Caucasians, with frequencies of approximately 10–17 (CYP2C9*2) and 7% (CYP2C9*3). Both polymorphic alleles were associated with increased response to anti-diabetic sulfonylurea drugs and an increased glycemic response in T2D patients (Zhou et al., 2010). As NAFLD patients are treated with sulfonylureas, genotyping is clinically relevant. Further studies are needed to identify the association of the CYP2C9*2 and CYP2C9*3 variants with adverse drug reactions such as hypoglycemia and weight gain resulting from sulfonylurea treatment.

CYP2E1, a fatty acid (Ω-1)-hydroxylase, catalyzes the oxidation of many low molecular weight molecules, including ethanol and acetone, a product of fatty acid oxidation. An important catalytic feature of CYP2E1 is the generation of ROS such as superoxide anion radical and hydrogen peroxide as a result of uncoupling of oxygen consumption with NADPH oxidation and as a by-product of lipid peroxidation (Robertson et al., 2001; Caro and Cederbaum, 2004). It is also involved in the biotransformation of xenobiotics such as acetaminophen, resulting in the generation of toxic reactive metabolites (Aubert et al., 2011). NAFLD and NASH patients and the MCD diet-fed rat model of NASH display enhanced expression of CYP2E1, which is in contrast to all other drug metabolizing CYPs, and elevated lipid peroxidation (Weltman et al., 1996, 1998; Videla et al., 2004) with increased localization to hepatic areas with oxidative stress injuries. Obese females with steatosis and NASH display elevated CYP2E1 protein levels and a positive correlation between the c2 allele of Rsa1/Pst1 polymorphisms in CYP2E1 and liver injury (Varela et al., 2008). Mice with silenced diacylglycerol acyltransferase 2 (Dgat2) on MCD diet display elevated Cyp2e1 expression that correlates with increased lipid peroxidation and oxidative damage, thus highlighting the role of CYP2E1 in the progression to NASH in response to increased hepatic free fatty acids (Yamaguchi et al., 2007). An upregulation in CYP2E1 has also been associated with leptinemia, reduced adiponectin levels, and insulin resistance in NAFLD (Aubert et al., 2011). This phenomenon is reversed in patients who have undergone bariatric surgery with resulting decreases in weight and hepatic steatosis (Bell et al., 2010). With the robust cellular protection mechanisms intact, increases in pro-oxidant molecules and CYP2E1 are counteracted by increased levels of glutathione (GSH). However, most NASH rodent models display lower GSH, indicating defects in the oxidative stress response pathways in progressive NAFLD. Nitrosylation of antioxidant enzymes superoxide dismutase (SOD) and catalase (CAT) is crucial in a Cyp2e1-overexpressing mouse model of NAFLD (Kathirvel et al., 2010) because increased levels of iNOS generates reactive nitrogen species (RNS), which nitrosylate antioxidant enzymes and decrease their activity. Thus, CYP2E1 polymorphisms that associate with the progression of NAFLD to NASH may possibly trigger the combined detrimental effects of both ROS and RNS, which in combination with toxic metabolites from xenobiotic biotransformation may result in further aggravated liver injury in NAFLD patients.

Observations in a NAFLD pediatric population have indicated a direct correlation between lipid peroxidation and disease severity irrespective of CYP2E1 levels; however, the small sample size of this study and the possibility of alternative mechanisms of lipid peroxidation in early onset hepatic steatosis cannot be excluded (Bell et al., 2011). Other groups have also previously indicated decreased CYP2E1 mRNA and protein levels and no changes in CYP2E1 activity at progressive stages of NAFLD (Fisher et al., 2009b; Mitsuyoshi et al., 2009). Thus, no conclusive role of CYP2E1 in NAFLD can be described. Interestingly, CYP4A genes seem to compensate for microsomal lipid oxidation in the absence of CYP2E1 as observed in Cyp2e1-null mice that display lipid peroxidation and increased expression of Cyp4a10 and Cyp4a14 (Leclercq et al., 2000; Hardwick et al., 2009). This observation may possibly explain the absence of changes in CYP2E1 activity in some NAFLD populations.

CYP4A enzymes ω-hydroxylate fatty acids into dicarboxylic acids that are preferentially oxidized by peroxisomes. Genes of the CYP4A sub-family are induced by PPARα-agonists and in conditions of fasting. The enhanced activity of CYP4A results in increased production of ROS, thus contributing to steatohepatitis. In contrast, PPARα agonists prevent NASH by increasing β-oxidation. Moreover, Pparα-null mice on MCD diet are more prone to developing NASH in the absence of Cyp4a induction (Ip et al., 2003). These observations suggest that the anti-steatotic effects of PPARα may be more potent than its activation of CYP4A genes, hence overriding the ROS-generating effects of CYP4A. In the absence of PPARα alternative oxidative stress mechanisms may act as causal factors.