Suppression of PXR Expression by Endoplasmic Reticulum Stress and Microrna

The pregnane X receptor is a nuclear receptor, functioning as a ligand activated transcriptional factor. PXR is recognized as a major regulator of xenobiotic metabolism. Several important Phase I and Phase II enzymes, and drug transporters are regulated by PXR. PXR is also a master regulator of CYP3A4, which is one of the most important enzymes in drug metabolism. Thus, knowledge of regulation of PXR expression is crucial to understand alterations in drug metabolism. Many liver diseases such as steatosis and inflammatory liver diseases alter the expression of drug metabolizing enzymes and transporters. CYP3A4 is down-regulated in these diseases. Endoplasmic reticulum (ER) stress is recognized as one important cause of the diseases and it also accompanies with other diseases such as viral infections. ER stress is also interconnected with inflammation. CYP3A4 and some other P450 enzymes are also down-regulated in inflammation. Since PXR is the master regulator of CYP3A4, this dissertation investigated the effect of ER stress on the PXR expression and its consequence on PXR-mediated CYP3A4 induction. Thapsigargin and brefeldin A were used to induce ER stress in cell cultures. ER stress down-regulated the expression of PXR in primary hepatocytes and HepG2 cells by repressing transcription. A promoter study revealed that the HNF4α protein level decreased in ER stress whereas C/EBPβ LIP increased. HNF4α and C/EBPβ were bound to the promoter of PXR suggested that both HNF4α and C/EBPβ involved in downregulation of PXR by ER stress. PXR-mediated CYP3A4 induction significantly decreased in ER stress. ER stress also induced the expression of IL-6 in primary hepatocytes. This finding established an interconnection between ER stress and inflammation. Interestingly, IL-6 also repressed PXR expression at the same regulatory sequences as thapsigargin. Furthermore, it has been shown elsewhere that some other nuclear receptors are regulated by microRNAs (miRNAs). miRNAs regulate gene expression at the post-transcriptional level, adding complexity to the gene regulation. This dissertation identified a miRNA that regulated the expression of PXR. miR-30c-1 down-regulated PXR. miR-30c-1 interacted at the 3’-UTR of PXR and decreased the PXR mRNA level. The potential binding site of miR-30c-1 on the 3’-UTR of PXR was located. In summary, ER stress, which is the cause and consequence of several diseases, down-regulated the expression of PXR. As a result, CYP3A4 induction by PXR was reduced. Since CYP3A4 and some others drug metabolizing enzymes regulated by PXR are crucial in drug metabolism, ER stress may alter metabolism of drugs, leading to increasing drug toxicity or decreasing drug efficacy. miR-30c-1 regulated the expression of PXR, adding complexity to the regulation on PXR. Although the regulation on the expression of miR-30c-1 is not known, miR-30c-1 might be related to diseases or biological conditions that could eventually affect PXR expression.


CHAPTER 1 INTRODUCTION 1 Pregnane X receptor
The pregnane X receptor (PXR or SXR) is the member of the nuclear receptor superfamily. PXR is categorized in subfamily 1, group I and member 2 (NR1I2). The human PXR gene is located on chromosome 3q12q13.3. The gene consists of ten exons, including exon 1a, 1b, 2, 3, 4, 5, 6, 7, 8, and 9. PXR is highly expressed in the liver, small intestine, and colon. It is also found in others tissues including, the lung, stomach, peripheral blood monocytes, blood-brain barrier, uterus, placenta, breast, osteoclasts, adrenal gland, ovary, heart, and specific regions of the brain Lamba et al., 2004a, b).
Several decades ago, exogenous steroids were shown to modulate expression of cytochrome P450 enzymes (P450s) that protect human and animals against toxic substances. The expression of the P450 enzymes is induced by several xenobiotic steroids such as pregnenolone-16α-carbonitrile (PCN) and the enzymes metabolize both natural and synthetic compounds including herbal and synthetic drugs. PXR was discovered due to some evidence related to glucocorticoids, a class of steroid hormones, and the glucocorticoid receptor (GR). PXR was first identified as the nuclear receptor that induces CYP3A4 expression. Before the discovery of PXR, it was known that CYP3A and some other P450 enzymes were induced by the glucocorticoids such as dexamethasone. It was believed that this induction was mediated through GR. However, PCN which is an anti-glucocorticoid is shown to induce rat CYP3A (Schuetz and Guzelian, 1984). In addition, treatment with dexamethasone and PCN shows a synergistic effect, inducing CYP3A in rat hepatocytes (Burger et al., 1992). This evidence suggested that there was another signaling pathway that induced CYP3A other than the glucocorticoid receptor pathway and led to the discovery of human PXR (Bertilsson et al., 1998.

Structure of PXR
PXR shares a common structure to the nuclear receptors (NR), containing a DNA binding domain (DBD), a ligand binding domain (LBD), and an activation function domain 2 (AF-2). The DBD of PXR is highly conserved in mammals, sharing more than 94% of their amino acid sequences ( Fig. 1-1) (Kliewer et al., 2002). The DBD contains 2 zinc fingers, approximately 70 amino acids. The DBD of PXR can recognize and bind to a specific DNA sequence so called PXR responsive elements (XREMs). The LBD of PXR contains about 250 amino acids. In contrast to other nuclear receptors, the LBD of PXR is diverse across species. The LBD of human and mouse PXR shares only 76% amino acid identity. The LBD can fold to form a hydrophobic pocket which can bind to a wide range of ligands. The AF-2 domain, located at the C-terminal, is responsible for the dimerization and transcriptional activation. Several isoforms of PXR have been identified. Some certain isoforms are generated due to alternative splicing and usage of an alternative start codon on exon 1a or 1b. hPXR (hPAR-1, SXR), the most abundant isoform of human PXR (about 93.04% of total PXR in the human livers), contains 434 amino acids. The hPXR transcript contains exon 1a but not 1b and uses the start codon in exon 2. Next, hPXR.2 is the most abundant alternative isoform of PXR (approximately 6.7% of total PXR in the human livers), lacking 111 nucleotides and deleting 37 amino acids from the PXR LBD (Lamba et al., 2004). Thus, hPXR.2 contains 397 amino acids. hPAR-2, which is created by alternative splicing using exon 1b but not 1a, uses initiation codon in exon 1b. hPAR-2 is 473 amino acids with an additional 177 nucleotides compared to hPXR (Bertilsson et al., 1998). In the fourth place, PXR.3 lacks 123 nucleotides, deleting 41 amino acids from the LBD compared to hPXR. As a result, hPXR.3 contains 393 amino acids.

Function of PXR
PXR functions as a ligand-activated transcriptional factor. It is found to heterodimerize with the retinoid X receptor-α (RXRα). The PXR/RXR heterodimer can recognize and bind to the PXR responsive element (XREM). XREM contains 2 half sites and spacer sequences on a promoter of a target gene. XREMs include direct repeat 3 (DR-3), direct repeat 4 (DR-4), everted repeat 6 (ER-6), and everted repeat 8 (ER-8) elements. The PXR/RXR heterodimer can also be found with or without promoter binding (Noble et al., 2006). Without a ligand, PXR is associated with transcriptional corepressors such as nuclear corepressor 1 (NCoR1) and NCoR2, resulting in repression of transcription (Ding et al., 2005 andStaudinger et al., 2011).
PXR ligands such as rifampicin (RIF) can bind to the LBD of PXR. Following ligand binding, the AF-2 helix undergoes a conformational change leading to dissociation of corepressors and recruitment of coactivators such as steroid receptor coactiavator 1 (SRC-1). SRC-1 has intrinsic histone acetyltransferase activity, resulting in chromatin remodeling and transcriptional activation . The figure 1-2 shows schematic diagram of the mechanistic function of PXR.
PXR is bound and activated by various xenobiotics and endobiotics. 17-βestradiol, progesterone, and progesterone are examples of endobiotics activating PXR (Blumberg et al., 1998). A number of genes in Phase I and II enzymes, and drug transporters have been identified as PXR target genes. These enzymes and transporters serve as defense mechanisms, transforming and eliminating xenobiotics form bodies.
Therefore, PXR is recognized as a key regulator of xenobiotic metabolism. There are studies proving that PXR is a master regulator of CYP3A. PCN, which is known to induce CYP3A in mice, does not induce CYP3A in mice lacking PXR (Xie et al., 2000). High induction of CYP3A is observed when transgenic mice expressing PXR are treated with rifampicin, proving that PXR regulated CYP3A expression. PXR is a master regulator of CYP3A which is one of the most important Phase I enzyme subfamily. The ensuing discussion will provide more details of the roles of PXR in drug metabolism.

Ligand promiscuity and species differences of PXR
A number of compounds, including drugs, herbal medicines, environmental pollutants and endobiotics, have been identified to activate PXR. It has been shown that PXR ligands are structurally diverse (Fig. 1-3 and Table 1-1). Examples of PXR agonists are rifampicin, clotrimazole, phenobarbital, troglitazone, lithocholic acid, hyperforin, and paclitaxel (Kliewer et al., 2002). The ligand promiscuity of PXR is explained by characteristics of PXR binding pocket. The ligand binding pocket of PXR is larger than many other nuclear receptors (Kumar andThompson, 1999 andWatkins et al., 2001). It has been estimated that the binding pocket volume ranges from 1200 Angstroms in the absence of ligands to about 1600 Angstorms in the presence of rifampicin (di Masi et al., 2009 andMoore et al., 2003). Thus, the binding pocket is large and also flexible. The species-specific activation of PXR by certain compounds has been demonstrated. For example, pregnenolone 16α-carbonitrile (PCN) is a strong PXR agonist of mouse and rat PXR but it does not activate human or rabbit PXR. Dexamethasone strongly activates mouse PXR and also moderately activates human PXR. In contrast, rifampicin and phenobarbital activate human PXR but show very little effect on mouse and rat PXR . In addition, transgenic mice possessing human PXR responds to rifampicin, resulting in high induction of CYP3A by rifampicin (Xie et al., 2000). This evidence proves that the species-specific induction of CYP3A by rifampicin is due to differences in the amino acid sequences of PXR. An amino acid mutation in the LBD of PXR can alter activation of PXR by rifampicin and PCN both in mice and human (Ostberg et al., 2002). This study supports that the species-specific induction is due to the differences in the LBD.  (Kliewer et al., 2002).

PXR in drug metabolism and detoxification
Human and animals have been always exposed to xenobiotics, including drugs, environmental pollutants, and naturally-occurring compounds. Our bodies have been evolved to employ various mechanisms to transform and detoxify these foreign compounds. The biotransformation process can be divided into two phases, phase I and II metabolism. Phase I metabolism involves oxidation, reduction, and hydrolysis reactions. The cytochrome P450 (CYP or P450) is the most important enzyme superfamily conducting phase I oxidation reactions. For examples, cytochrome P450 3A subfamily (CYP3A) is responsible for metabolizing approximately 50% of all clinically approved drugs Redinbo, 2004). CYP3A is also the most abundant P450 enzyme in the human liver and intestine (Paine et al., 2006). An important characteristic of CYP enzymes such as CYP2B and CYP3A is that they can be dramatically induced by xenobiotics. This induction is to response to transform and detoxify those foreign compounds. CYP3A are also involved in metabolism of highly toxic and narrow therapeutic window drugs and since many drugs are metabolized by CYP3A, drug interactions among there drugs are life-threatening. PXR is a key regulator of CYP3A expression (Bertilsson et al., 1998, Kliewer et al., 2002. PXR can be activated by a wide range of structurally diverse compounds and then transactivate CYP3A gene, increasing CYP3A expression. CYP3A can metabolize a wide range of compounds which correlates with the promiscuity of PXR. In human, CYP3A4 is the most abundant CYP3A isoforms. PXR also regulates CYP2B6, CYP2C9, and CYP2C19. Therefore, PXR involves in metabolism of a number of xenobiotics, including drugs toxic substances, and supplements. Examples of compounds that activate PXR and are shown in table 1. glucuronic acid, and glutathione (GSH). This is to enhance their excretion into urine or bile and the products from Phase II reactions tend to be less active than the substrates.
Xenobiotic biotransformation is also regulated by uptake and efflux transporters. Some certain drug transporters have been shown to be regulated by PXR.

Target gene Effect on target gene
Induction MRP2 Induction and organic anion transporting polypeptide 2 (OATP2) (Jigorel et al., 2006 andSchrenk et al., 2001). This information, altogether, shows that PXR regulates a number of genes involved in drug adsorption, distribution, metabolism, and elimination ( Fig. 1-4). These genes are controlled by PXR in response to xenobiotic challenge so that our bodies can detoxify and eliminate potentially toxic compounds from the bodies. Since PXR regulates a number of genes involved in drug metabolism, what could happen if the PXR expression level is altered? The ensuing discussion will provide more details of certain conditions and factors that could alter PXR expression.

Fig. 1-4
Following ligand binding, PXR binds to XREM of the target genes resulting in transcriptional activation. The PXR target genes involve in drug adsorption, distribution, metabolism, and elimination .

PXR mediated drug-drug interactions
PXR functions as a xenobiotic sensor binding to its ligand and activating transcription of the target genes which result in increasing drug metabolizing enzymes that can metabolize the ligand and also other compounds. Drug interactions frequently occur when multiple drugs are administered such as anti-HIV and anti-tuberculosis.
Discovery of PXR reveals molecular mechanism of how one drug increases clearance and decrease therapeutic efficacy of a co-administered drug.  (Niemi et al., 2003).
PXR not only involves in drug-drug interaction but also herb-drug interactions.
St John's wort is an herbal medicine that is widely used for depression, sleep disorders, and anxiety. St John's wort contains PXR ligands such as hyperforin and therefore induces CYP3A expression . St John's wort caused serious herb-drug interaction with cyclosporine, an immunosuppressant used in organ transplant patients to decrease the risk of organ rejection. It has been reported that organ transplant failed in patients using St John's wort with cyclosporine. Since the herb induces CYP3A and cyclosporine is metabolized by CYP3A. As a result, cyclosporine blood levels decreased, so it was not effective (Murakami et al., 2006 andMai et al., 2004).

PXR and inter-individual variability in drug response
CYP3A are major enzymes contributing to a large portion of metabolism of drugs and herbal medicines. CYP3A expression in the human liver and intestine varies as much as 40 folds (Lamba et al., 2002). It is predicted that 90% of the variation in CYP3A4 activity is due to genetic factors (Ozdemir et al., 2000). Genetic variation in the CYP3A4 gene has been extensively studied; however, the genetic variation in the CYP3A4 gene alone fails to explain the large variability of CYP3A4 expression (Lamba et al., 2002). Since PXR is the major transcriptional regulator of CYP3A4 it has been thought that the variability in the PXR expression and the PXR gene could contribute to the variability in the CYP3A4 expression. Some studies have shown that there is significant correlation between the PXR and CYP3A4 expression. In the human livers, the PXR mRNA level is associated with CYP3A4 mRNA (Lamba et al., 2010).
The expression of PXR has been associated to metabolism and elimination of certain drugs in human. Atazanavir, a HIV protease inhibitor, is metabolized by CYP3A4 and MDR1, which are the PXR target genes. Plasma clearance of atazanavir is associated with a PXR SNP, rs2472677 (Siccardi et al., 2008, Schipani et al., 2010. The study shows that patients with homozygous T allele exhibit higher atazanavir clearance than those with C allele. In addition, The SNP rs2472677 has previously been reported that it is also associated with PXR and CYP3A4 expression in the human livers (Lamba et al., 2008). Thus, the association between atazanavir clearance and the PXR SNP is likely due to expression of PXR and its target gene CYP3A4. A human study shows that a haplotype of the PXR gene is associated with doxorubicin clearance . The study is conducted in Asian populations (Chinese, Malay and Indian) and the result shows that the haplotype is correlated with the expression of PXR and its downstream target genes CYP3A4 and MDR1. The haplotype is associated with reduced doxorubicin clearance. These studies suggest that PXR involves in inter-individual variability in drug response of certain drugs. The expression of PXR is related to the drug clearance.

Endoplasmic reticulum stress in diseases and drug metabolism
Endoplasmic reticulum (ER) is an organelle in eukaryotic cells that serve several functions, including protein synthesis, folding and trafficking, posttranslational modification, calcium storage and regulation, and xenobiotic metabolism.
Newly synthesized proteins translocate to the ER where they are folded into the functional structures. The ER also functions as the calcium storage which responds to stimuli that alter cellular energy level or redox status. Some certain conditions such as oxidative stress, viral infection, and inflammation can disrupt the ER function and cause stress in ER (Zhang and Kaufman, 2008). ER stress occurs when there is accumulation of unfolded or misfolded proteins in ER. Accumulation of these unfolded proteins can be caused by an increase of demand for protein folding or disruption of reactions of protein folding. In mammalian cells, the ER responds to stress by triggering unfolded protein response (UPR), which is signaling pathway aiming to relief the stress from the ER and to resume its function. Generally, the UPR will activate signaling pathways that lead to increasing expression of molecular chaperones involved in protein folding. UPR also increase protein degradation to remove unfolded or misfolded proteins. Protein translation is also reduced to decrease influx of new proteins into the ER . In mammalian cells, the UPR signaling pathways are initiated by ER protein double-stranded RNA-dependent protein kinase-like ER kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring 1α (IRE1α). All three proteins are up-regulated during ER stress and have been used as indicators of stress in ER. Also, IRE1 is a ribonuclease that can splices X box-binding protein 1 (XBP1) mRNA. In ER stress, IRE1 splices XBP1 mRNA, producing spliced XBP1 mRNA that will be translated to XBP1 protein that induces UPR target genes such as chaperones. Thus, the presence of spliced XBP1 mRNA has also been used as an ER stress indicator. When the stress in the ER prolongs and cannot be relived, apoptosis is triggered. The UPR is an adaptive program responding to stimuli. The ER stress-induced apoptosis can be considered as a mechanism to protect tissues and organs by eliminating the stressed cells that produce malfunctioning proteins.
It is recognized that ER stress and UPR implicate in various diseases, including inflammatory diseases, metabolic diseases, neurodegenerative diseases, and cancer Yoshida, 2007). One reason is because ER stress and UPR can cause apoptosis and may damage functioning tissues. Secondly, the UPR alters expression of many genes. Thus, ER stress and UPR have been extensively studied in pathological aspects of several diseases. The liver is the major organ responsible for drug metabolism. Many diseases can impair liver functions such as decreasing capability of xenobiotic detoxification. ER stress has been associated with these diseases such as diabetes and steatosis (Kolwankar et al., 2007). However, studies of ER stress related to drug metabolism are lacking and there are only few studies about P450 enzymes. In rat hepatocytes, ER stress down-regulates CYP3A . Nevertheless, effect of ER stress on many of Phase I and II enzymes and drug transporters has not been investigated. In addition, effect of ER stress on PXR and most of PXR target genes has not been reported.
There is evidence showing that there would be connections between ER stress and drug metabolism. Furthermore, several studies demonstrated that there is cross-talk between UPR and inflammation (Adolph et al., 2012, Zhang andKaufman, 2008).
The link between UPR and inflammation involves NF-κB signaling pathway and inflammatory cytokines. The inflammatory cytokines such as interleukin-1 (IL-1β), interleukin-6 (IL-6), tumor necrosis factorα (TNFα) are pro-inflammatory cytokines that can lead to inflammation. It has been reported that thapsigargin, a natural compound which is widely used to induce ER stress in cell culture, induced IL-6 expression in rat peritoneal macrophages (Ichinowatari et al., 2002). The study showed the link between ER stress and inflammation and the inflammatory cytokine such as IL-6 in rat. In addition, the inflammatory cytokines are known to downregulate P450 enzymes. Therefore, ER stress and UPR could affect P450 enzyme expression. Effect of inflammation on drug metabolism is explained in a following topic.

Effect of Inflammation on PXR and drug metabolism
Inflammation is a response to harmful stimuli such as infection by pathogens or tissue injuries from physical or chemical damages. Inflammation can occur in short term (acute inflammation) about a few days or long term (chronic inflammation) which can last for months or years. Inflammation at organs involving in drug metabolism such as the liver and small intestine can cause a large impact on drug metabolism. This is because the pro-inflammatory cytokines such as IL-1β, IL-6, and TNFα alter expression of drug metabolizing enzymes. Both in humans and animals, inflammation causes changes in the expression levels of several P450 in the liver and small intestine as well as other organs such as the kidney and brain (Morgan, 1997, Renton, 2004. In most cases, the expression of P450 enzymes is suppressed due to inflammation. These decreases in the P450 enzymes significantly alter drug metabolism and can cause serious clinical consequences especially for drugs with a narrow therapeutic index. Therefore, inflammation can increases drug toxicities. It is known that interleukin-6 (IL-6), a pro-inflammatory cytokine, downregulates CYP3A4, 2B6, 2C9, and 2C19 expression in human hepatocytes . PXR which regulates these P450 enzymes is also down-regulated by IL-6 in human hepatocytes . The information of inflammation together with ER stress from the previous topic suggests that ER stress may affect PXR expression and function.

microRNA
A 3'-untranslated region (3'-UTR) of mRNA is involved in several gene regulatory processes, including mRNA stability, transcript cleavage, translation, and mRNA localization. It is a binding site of many regulatory proteins and microRNA (miRNA) . The length of the 3'-UTR in mammalian mRNA varies from 60 -4000 nucleotides. The average length of the 3'-UTR in human genes is about 740 nucleotides which are longer than that in other mammalian and vertebrate genes (420 -500 nucleotides) and in plants and fungi (240-274 nucleotides) (Pesole et al., 1997). The longer 3'-UTR in human suggests that human has evolved to possess more complexity of post-transcriptional control which is another important step to regulate gene expression.
miRNA is a short single strand RNA about 21-23 nucleotides. It is transcribed by RNA polymerase II (pol II) in the nucleus as long primary transcripts (pri-miRNA) (Kim, 2005). These transcripts can be several kilobases in length. miRNA is contained in pri-miRNA as a stem-loop hairpin structure about 60-80 nucleotides. This miRNA hairpin in pri-miRNA are recognized and excised by a microprocessor protein complex. The miRNA hairpin (pre-miRNA) is transported to the cytoplasm and further processed. The hairpin portion of pre-miRNA is excised by an RNase III enzyme dicer, producing a 21-23 nucleotide duplex. The miRNA duplex is incorporated into the RNA-induced silencing complex (RISC). One strand of the miRNA duplex is degraded, leaving the other strand in the RISC complex. miRNA in the RISC complex serves as a guide that directs RISC to target mRNAs. Binding of miRNA-RISC complex to mRNA results in blocking translation or cleaving of the target mRNA . Thus, miRNAs function as post-transcriptional regulators. There is growing evidence showing that many miRNAs can both repress translation and degrade mRNA (Lee et al., 2012). miRNA has been shown to target at the 3'-untranslated region (3'UTR) of mRNA. It has been estimated that human may encode more than 1000 miRNAs . Approximately 70% of miRNAs are located in introns or exons and about 30% are located in intergenic regions (Rodriguez et al., 2004). At present, biological functions of most miRNAs are not known. The role of miRNAs is thought to be for fine-tuning of gene expression (Sevignani et al., 2006).

CHAPTER 2 STATEMENT OF PURPOSE
It has been shown in human that certain liver diseases involving inflammation such as viral hepatitis and steatosis decrease the liver's capability to metabolize a number of drugs. CYP3A4 expression is also decreased in these diseases. PXR is recognized as a xenobiotic sensor, binding to a wide range of compounds and functioning as a ligand-activated transcriptional factor. PXR regulates many important genes in drug metabolism, including drug metabolizing enzymes such as CYP3A and drug transporters such as MDR1. CYP3A4 is a major isoform of the CYP3A subfamily responsible for metabolizing about 50% of marketed drugs. CYP3A4 is known to possess high inter-individual variability in expression due to several contributing factors such as genetic and disease related factors. Variation in CYP3A4 expression has been shown to affect therapeutic efficacy and adverse reaction of drugs. CYP3A4 expression and activity in humans is significantly associated with PXR expression. Therefore, alteration in PXR expression has been accepted as a cause of interindividual variability in CYP3A . Nevertheless, factors causing changes in PXR expression have not been clearly understood. This dissertation studies certain factors altering the PXR expression which may lead to variability in CYP3A expression.
The goal of Chapter 3 of this dissertation is to study effect of ER stress on PXR expression and function. ER stress has been linked to chronic liver diseases such as steatosis and type 2 diabetes. These liver diseases have been shown to alter capacity of drug elimination. CYP3A4 is an example of drug metabolizing enzymes that is known to decrease in the chronic liver diseases. However, the effect of ER stress on PXR, which is the major regulator of CYP3A4, has not been investigated. A mechanism of how ER stress down-regulates CYP3A4 expression has not been elucidated. Thus, Chapter 3 will study the effect of ER stress on expression of PXR and its consequence on induction of CYP3A4. The study will also investigate the mechanism of how ER stress impacts expression of PXR. The effect of ER stress on PXR at the transcriptional level will be studied. Since PXR functions as the liganddependent transcriptional factor binding to its ligand and transactivating PXR target genes, the impact of ER stress on PXR function will be determined by monitoring the transcriptional level of CYP3A4.
The aim of Chapter 4 of this dissertation is to identify a miRNA that regulates expression of PXR. miRNA adds complexity of how a gene is regulated by acting on the mRNA post-transcriptionally. miRNA have been recognized to cause interindividual variation in a gene expression . miRNA is also involved in inflammatory diseases such as inflammatory bowel disease and liver cancer . Other nuclear receptors such as ERα and HNF4α and some genes involving in drug metabolism are regulated by miRNA. Certain P450 enzymes have been showed to be regulated by miRNA .
CYP3A4 is regulated by miR-27b and miR-378 regulates CYP2E1  miR-148a has been identified to target at the PXR mRNA . It is predicted that a gene can be targeted by several miRNAs (John et al., 2005). Identifying a miRNA that regulates PXR will further improve knowledge of how PXR is regulated. A target site of miRNA on PXR will be identified and the effect of miRNA on PXR expression will be studied.
Since ER stress is related to the diseases and could have an impact on drug metabolism, studying the effect of ER stress on PXR, the master regulator of xenobiotic metabolism, could help understand the effect of ER stress on PXR and drug metabolism. The study may reveal the mechanism of regulation of PXR. Identifying miRNA regulating PXR could provide a better understanding of the regulation on PXR. These studies may provide a better understanding in alteration in PXR expression and its effect on drug metabolism. respectively. Over-expression of HNF-4 or liver-enriched activator protein (an activator of C/EBP) restored the expression of PXR. Interestingly, the very same sequence also responded to interleukin-6 (IL-6), and primary hepatocytes treated with thapsigargin significantly increased the level of IL-6 mRNA. These findings establish a functional interconnection between ER-stress and signaling of proinflammatory cytokines in the regulated expression of PXR.

INTRODUCTION
The endoplasmic reticulum (ER) is an organelle involved in diverse cellular functions including protein synthesis/transportation, membrane generation, calcium concentration regulation and xenobiotic metabolism (Johnson et al., 2012;. Therefore, ER homeostasis is critical in maintaining the overall cellular functions. On the other hand, many factors such as oxidative stress disrupt ER homeostasis, leading to ER stress (Cali et al., 2011;. While the precise mechanisms of ER-stress remain to be determined, one of the outcomes is the accumulation of unfolded proteins in the ER Haeri and Knox, 2012). The unfolded protein response (UPR), occurring at the initial stage of ER-stress, is triggered to slow down protein synthesis, improve protein folding capacity and enhance degradation of unfolded proteins. Nevertheless, persistent ER-stress has been linked to the development of various conditions such as type 2 diabetes and chronic liver diseases Flamment et al., 2012;He and Chen, 2012;.
The liver is the largest internal organ and plays the primary role in drug metabolism Villarroya et al., 2010). The prevalence of hepatic dysfunction is high and it affects more than 10% of Americans (Liver foundation, 2009). Worldwide, liver cancer and chronic liver diseases are the seventh leading cause of death (IPA, 2007). Many liver diseases are accompanied with ER-stress and exhibit decreased capacity of drug metabolism and detoxification.
Steatotic livers, for example, were found to have significant decreases in CYP3A activity (Kolwankar et al., 2007), the most robust catalytic system in the oxidative metabolism. In cultured primary hepatocytes, lipid-loading significantly decreased the expression of CYP3A4 (Donato et al., 2006). Furthermore, livers from diabetic patients showed significantly lower expression of CYP3A4 (Dostalek et al., 2011).
The expression of CYP3A4 is regulated by several major transcription factors.
Among these proteins, the pregnane X receptor (PXR) has been established to play the primary role . This receptor forms a heterodimer with the retinoid X receptor- and binds to PXR response elements that contain a half-site AG(G/T)TCA or related sequence. We and other investigators have functionally characterized four PXR elements in the CYP3A4 gene and some of the elements operate in a coordinate manner (Goodwin et al., 1999;Toriyabe et al., 2009). While PXR regulates the expression of CYP3A4 and many other drug-eliminating genes, the expression of PXR varies by drugs and disease mediators. Importantly, the expression of PXR directly affects CYP3A induction.
This study was performed to determine whether ER-stress condition decreases the expression of PXR and whether the decrease alters the induction of CYP3A4.
Brefeldin A (BFA) and thapsigargin (Thaps), two well-established ER-stressors, significantly decreased the expression of PXR in both primary hepatocytes and HepG2 cell line (human hepatocellular carcinoma). The decrease led to reduced induction of CYP3A4. The decreased expression of PXR was achieved by transcriptional repression via two adjacent elements recognized by hepatocyte nuclear factor-4 (HNF-4) and cytidine-cytidine-adenosine-adenosine-thymidine enhanced binding proteins (C/EBP proteins), respectively. Over-expression of either protein restored the expression of PXR. Interestingly, the adjacent elements also responded to interleukin-6 (IL-6), suggesting a functional interconnection between ER-stress and signaling of proinflammatory cytokines. The treatment lasted for 24 h and the cells were washed once with phosphate buffered saline and collected by scraping. The reporter enzyme activities were assayed with a Dual-Luciferase Reporter Assay System as described previously .

Electrophoretic mobility shift assay (EMSA)
The EMSA experiment was performed as described previously .
Nuclear extracts of HepG2 cells treated with Thaps (50 nM) for 24 h were prepared with the nuclear and cytoplasmic extraction kit (Pierce, Rockford, IL). The sense and antisense oligonucleotides (Table I) were annealed by heating at 94°C for 5 min followed by gradually cooling to room temperature. The sense strand was synthesized as labeled or non-labeled form (for competition). Nuclear protein (5 μg

Chromatin immunoprecipitation (ChIP)
ChIP experiment was performed, essentially as described previously . HepG2 Cells were treated with DMSO or Thaps (50 nM) for 24 h, washed and underwent cross-linking for 15 min by 1.0% formaldehyde at room temperature, and the cross-linking was terminated with glycine (final concentration of 125 mM). The soluble chromatin preparations were prepared as described previously . For the ChIP experiment, chromatin preparations were pre-cleared for 2 h at 4°C with protein G beads pre-treated with herring sperm DNA (0.2 mg/ml) and BSA (0.5 mg/ml). A fraction of the pre-cleared chromatins was stored at -80°C for later use as an input. An antibody against HNF4 or C/EBP was added into the pre-cleared chromatins, and an overnight incubation at 4°C was performed. As a negative control, incubation was performed with pre-immune IgG. The antibody-bound chromatin and DNA input were analyzed by PCR for the presence of the genomic fragments containing the HNF4 or C/EBP-bound element with primers shown in Table I. The PCR was performed with Platinum Taq DNA polymerase for a total of 32 cycles at 94C for 30 s, 58C for 30 s and 68C for 60 s. A 3-min initial denaturation was performed.

Other analyses
Protein concentrations were determined with BCA assay (Pierce) based on albumin standard. Western blotting was performed as described previously  and the preparation of antibodies were described elsewhere .
Data are presented as mean  SD of at least three separate experiments, except where results of blots are shown in which case a representative experiment is depicted in the figures. Statistical significance between two means was made according to One-way ANOVA followed by a DUNCAN's multiple comparison test (p < 0.05).

Down-regulation of PXR by ER-stressors
ER-stress is a phenomenon in various chronic diseases and many PXR-target genes are down-regulated by disease conditions associated with ER-stress (Cali et al., 2011;Johnson et al., 2012). To test whether the expression of PXR itself is decreased during ER-stress, HepG2 cells were treated with Thaps and BFA. Thaps induces ER-stress by depleting calcium in the ER , whereas BFA by retrogradetransporting proteins from the Golgi apparatus to the ER . As shown in Fig. 1A, treatment with either stressor significantly decreased the level of PXR mRNA. BFA was slightly more potent than Thaps (Fig. 1A). Next, we determined the decrease of PXR mRNA as a function of the amount of Thaps. As shown in Fig. 1B, Thaps at 1 nM caused a 20% decrease of PXR mRNA and at 10 nM caused a 40% decrease. Further increased concentrations of Thaps up to 250 nM caused only a 10% additional decrease (Fig. B). To ascertain the cellular ER-stress level, semiquantitative PCR was performed to detect the presence of spliced XBP1 mRNA (Xbox binding protein 1 gene), a widely used marker for ER-stress . As shown in Fig. 1B (bottom), little spliced XBP1 mRNA was detected in cells treated with solvent or 1 nM Thaps. Comparable levels of spliced and non-spliced XBP1 mRNA were detected in cells treated at 10 nM (Fig. 1B). In contrast, cells treated at 25 or 250 nM exhibited the presence of spliced XBP1 mRNA only. To gain in vivo relevance, primary hepatocytes were treated with Thaps, and the expression of PXR was determined. As shown in Fig. 1C, Thaps decreased PXR at both mRNA and protein levels, and the decrease was even greater than that in HepG2 cells (Figs. 1A and 1B).

Transcriptional repression of PXR by Thaps
The decreases in PXR mRNA pointed to two possibilities: ER-stressors enhanced PXR mRNA degradation and/or reduced PXR transactivation. To shed light on the second possibility, various PXR reporters containing the promoter or along with upstream regulatory sequences at varying length were tested for the repression in response to Thaps. As shown in Fig. 2A, all PXR reporters, compared with the vector control, were significantly repressed. However, the reporter PXR-56Luc, compared with PXR-106Luc, was repressed to a significantly less extent, suggesting that the DNA segment from -106 to -56 nucleotides is critical for the repression. Based on element predication with computer program, this DNA segment contains an HNF4 and a C/EBP binding site. These two elements are spaced by three nucleotides ( Fig.   2A). It should be noted that as many as five transcription start sites (filled triangles in Fig. 2A) are located in the PXR promoter region Kurose et al., 2005;Tompkins et al., 2008).
To test whether these two elements support Thaps-repression, element reporters were prepared to contain this segment or segment with one or both elements disrupted. The resultant reporters were tested for the abolished response to Thaps. As shown in Fig.   2B, disruption of the HNF4 element largely abolished the repressive response to Thaps. In contrast, disruption of the C/EBP element reversed Thaps-repression by 20%. These results suggested that both HNF4 and C/EBP supported the transactivation of PXR with HNF4 playing a greater role for the repression. To shed light on this possibility, cells were transfected with HNF4 or LAP (liver-enriched activator protein: a form of C/EBP), treated with Thaps, and detected for the level of PXR mRNA. As shown in Fig. 2C, transfection of HNF4 or LAP reversed the suppression of PXR mRNA in response to Thaps. Interestingly, LAP showed a greater reversal activity than HNF4.

Effect of Thaps on the expression of HNF4 and C/EBP (LAP and LIP)
The transfection study demonstrated that HNF4 and LAP were positive regulators of PXR expression. Next we tested whether Thaps decreased PXR expression by downregulating HNF4 and C/EBP. Cells were treated with Thaps and the expression of HNF4 and C/EBP were determined by RT-qPCR and Western blotting. As shown in Fig. 3A, Thaps surprisingly caused a 9-11 fold induction of C/EBP mRNA. It has been well established that C/EBP mRNA produces several in-frame polypeptides including LAP* (38 kDa), LAP (35 kDa) and LIP (liver-enriched inhibitory protein: 20 kDa) . Importantly, these polypeptides are functionally distinct with LAP* and LAP being activators and LIP being an inhibitor. To determine whether the increased C/EBP mRNA by Thaps differentially increases in these polypeptides, Western blotting was performed. As shown in Fig. 3A (Right), little changes were detected on the levels of LAP and LAP*. In contrast, LIP was markedly increased. These results established that induction of C/EBP mRNA by Thaps increased the production of LIP but not activator LAP* and LAP. The level of HNF4 mRNA, in contrast to the level of C/EBP mRNA, was significantly decreased by Thaps (Fig. 3B) and the decrease was less with prolonged treatment.
The 12 h time-point showed a 60% decrease whereas the 24 h time-point showed a 40% decrease (Fig. 3B). Consistent with the decease in HNF4 mRNA, the level of HNF4 protein was drastically decreased (Fig. 3B).

Occupancy of the PXR promoter by HNF4 and C/EBP
The reporter and expression studies collectively suggested that the PXR promoter is  (Fig. 4A). Incubation with the C/EBP probe, on the other hand, led to the detection of several bands (Fig. 4B). Incubation with extracts from Thaps-treated cells produced a new shifted band (probably bound by LIP) and decreased the intensities of two shifted bands (arrowed in column 3). All shifted bands except the top one were competed by non-labeled probe but not by the corresponding non-labeled probe with disrupted C/EBP binding site. Nonetheless, the putative LAP and LIP binding bands were abolished by the antibody against C/EBP (Fig. 3).
The EMSA experiment established that the PXR promoter contained HNF4 and C/EBP binding site. Next we tested whether both proteins occupy the PXR promoter.
ChIP experiment was performed in cells treated with solvent or Thaps. In addition to HNF4 and C/EBP element-containing segment, a segment of the PXR gene containing either element was subjected to PCR-amplification as a control. As shown in Fig. 4C, PCR detected the amplification of both segments with input DNA.
However, PCR detected the HNF4-C/EBP but not the control segment with ChIPed-DNA. The amplification was observed with ChIPed DNA from control but not Thapstreated cells (Fig. 3C). It should be noted that pre-immune IgG for ChIP experiment did not yield any amplification.

Interconnection between Thaps and IL-6 in the suppression of PXR
We have previously showed that PXR was down-regulated by the proinflammatory cytokine IL-6 . To determine whether Thaps and IL-6 use similar genomic sequence in the down-regulation, HepG2 cells were transfected with various PXR reporters, treated with IL-6 as shown, and detected for luciferase activity. BFA, another commonly used ER stressor, was also included in this study. As predicted, both IL-6 and BFA produced a similar responding pattern as Thaps among these reporters ( Figs, 2A and 5A). Two additional experiments were performed to shed slight on the mechanistic connection between ER stress and IL-6. Firstly, the suppression of PXR by IL-6 and Thaps was determined as a function of the time of treatment. Secondly, the expression of IL-6 was determined in Thaps-treated primary hepatocytes. As shown in Fig. 5B, both IL-6 and Thaps significantly decreased PXR mRNA. However, the decrease by IL-6 occurred sooner than that by Thaps (Fig. 5B).
We next tested whether human primary hepatocytes treated with Thaps actually support the induction of IL-6. As shown in Fig. 5C, treatment with Thaps significantly increased IL-6 mRNA (Fig. 5C).

Effect of ER-stress on CYP3A4 induction
The enhanced production of IL-6 by Thaps suggested that IL-6 is a contributor to cells were treated with Thaps, Z-guggulsterone, emodin or in combination, and then the level of PXR mRNA was determined. As expected, Thaps significantly decreased PXR mRNA (Fig. 6A). The decrease, however, was almost completely reversed by Z-guggulsterone but not by emodin. It should be noted that Z-guggulsterone and emodin alone showed no effect on the level of PXR mRNA (Fig. 6A).
Next we tested whether overexpression of PXR itself reverses the effect of Thaps in terms of the induction of CYP3A4. Both transfected and non-transfected HepG2 cells were used, and the transfection was performed with a PXR expression construct or the corresponding vector. The cells were treated with Thaps, rifampicin or both for 24 h and analyzed for the mRNA level of CYP3A4, a prototypical target of PXR . The results were expressed as fold of induction. As shown in Fig. 6B, Thaps significantly decreased the induction of CYP3A4 in both vector-and nontransfected cells. However, the decrease was reversed in PXR transfected cells.

DISCUSSION
ER-stress is recognized as a common theme in the development of metabolic syndrome and other diseases (Johnson et al., 2012;Lin et al., 2012), and emerging evidence has pointed to decreased capacity of metabolism in liver diseases associated with ER-stress (He and . PXR is a master regulator of genes in xenobiotic elimination. In this study, we have shown that Thaps and BFA, two well-characterized ER-stressors significantly decreased the expression of PXR.
The decrease was mediated through transcriptional repression and led to reduced induction of CYP3A4, a prototypical target gene of PXR .
The decrease of PXR expression by Thaps was reversed by Z-guggulsterone, an active ingredient of the hypolipidemic herb guggul .
It is likely that the reversal by Z-guggulsterone was achieved by blocking STAT3 remains to be determined whether increased expression of IL-6 by Thaps represents a general phenomenon among ER-stressors and diseases associated with ER-stress. The connection between Thaps and IL-6, nevertheless, provides a mechanistic understanding of how ER-stress conditions may exert differential effect on the expression of PXR depending on the increased secretion of cytokines such as IL-6.
STAT3 is a DNA-sequence specific transcription factor . However, the PXR promoter regulatory sequence targeted by Thaps and IL-6 does not harbor a consensus STAT3 element. Instead, this sequence contains two adjacent elements that were recognized by HNF4 and C/EBP proteins, respectively. It is therefore assumed that STAT3 decreases the expression of PXR by regulating the expression of HNF4, C/EBP proteins or both. While it is not clear whether STAT3 down-regulates HNF4, it was reported that STAT3 up-regulated the expression of C/EBP . Furthermore, STAT3 was shown to interact directly with C/EBP. Given the fact that co-transfection of LAP increased PXR expression (Fig. 2C), the STAT3-C/EBP complex likely exerts repressive activity. Alternatively, such complex no longer acts on the PXR promoter, thus functioning as a dominant negative in comparison with LAP. The C/EBP family has several members and they all bind to same or similar DNA elements . It is conceivable that other C/EBP members likely participate in the regulated expression of PXR during ERstress.
One of the interesting findings in this study is the unique interplay between C/EBP and HNF4. In the reporter experiment, disruption of the HNF4 element almost completely eliminated the repressive activity in response to Thaps (Fig. 2B). In contrast, disruption of the C/EBP element diminished the repression to a much lesser extent (Fig. 2B). These observations suggested that HNF4 played an essential or a greater role than a C/EBP protein (probably C/EBP) in supporting the expression of PXR. However, transfection of HNF4 surprisingly caused less increases of PXR mRNA than co-transfection of LAP (an active form of C/EBP) (Fig. 2C). One explanation is that LAP functioned as a transactivator of HNF4 and/or LAP enhanced the activity of HNF4. In support of the last possibility, LAP was shown to increase nuclear translocation of HNF4 .
C/EBP mRNA produces several in-frame translated polypeptides including LAP*, LAP and LIP. Under normal conditions, LAP is the most abundant form. While LAP* and LAP are transactivators, LIP acts as a transcriptional repressor . It is generally accepted that the repressive activity of LIP is achieved by forming non-functional dimmer with C/EBP activating members and/or a DNAbinding dominant negative. It is also accepted that the relative abundance of various C/EBP forms (e.g., LAP versus LIP) largely depends on the relative efficiency of the initiation codons for translation. Interestingly, Thaps treatment caused an 11-fold increase of C/EBP mRNA (Fig. 3A), and yet the increase in proteins was detected on LIP but not LAP or LAP* (Fig. 3A). One explanation is that the initiation codon for LIP was more efficient under ER-stress condition induced by Thaps. It has been reported that LAP can be converted into LIP (LAP is bigger than LIP) through proteolytic digestion through an unknown protease. It is likely that such a protease(s) is up-regulated and/or activated by Thaps (Welm at el., 1999). Alternatively, LIP was relatively more stable than LAP in the presence of Thaps (Li et al., 2008).
Nevertheless, EMSA experiment detected increases in DNA binding, apparently by LIP (Fig. 4B). In contrast, the intensity of the shifted bands by LAP and LAP* was slightly decreased in nuclear extracts of cells treated with Thaps. Based on ChIP experiment, however, the increased LIP did not lead to increases in the occupancy of the C/EBP element in the PXR promoter, although the same antibody was used in both EMSA and ChIP experiments. One explanation is that chromatin-bound LIP (ChIP) posed a configuration that hided the epitope from being recognized by this antibody.
Alternatively, LIP normally does not bind to the C/EBP element in the PXR promoter under the native condition (i.e., cell), although it did so under non-cellular context (i.e., EMSA). Nonetheless, cotransfection of LIP indeed conferred potent repressive activity toward the PXR promoter reporter (data not shown).  Kolwankar D, Vuppalanchi R, Ethell B, Jones DR, Wrighton SA, Hall SD, Chalasani N (2007) Association between nonalcoholic hepatic steatosis and hepatic cytochrome P-450 3A activity. Clin Gastroenterol Hepatol. 5:388-93. Asterisk signs indicate statistical significance from the vector (P < 0.05).

Fig. 4. Characterization of HNF4 and C/EBP elements by EMSA and ChIP (A)
EMSA analysis of HNF4 element Nuclear extracts (5 μg) from HepG2 cells treated with DMSO or Thaps (50 nM) for 24 h were incubated with a biotinylated HNF4 probe for 20 min. In the competition assay, nuclear extracts were pre-incubated with the unlabeled element (100x), for 20 min, and then incubated with the biotinylated probe. In disruption assay, nuclear extracts were incubated first with an antibody against HNF4 on ice for 20 min and then with the biotinylated probe. The protein-DNA complexes were electrophoretically resolved, transferred to a Biodyne® nylon membrane and located with streptavidin-conjugated horseradish peroxidase and chemiluminescent substrate. (B) EMSA analysis of C/EBP element Incubations were performed as characterization of the HNF4 element. However, the C/EBP probe and an antibody against C/EBP were used. (C) ChIP analysis HepG2 cells were treated with DMSO or Thaps at 50 nM for 24 h, washed and underwent crosslinking for 15 min by 1% formaldehyde, and the cross-linking was terminated with 125 mM glycine. The soluble chromatins were prepared, pre-cleared with protein G beads and incubated with an antibody against HNF4 or C/EBP. As a control, the antibody was replaced with pre-immune IgG. The antibody-bound chromatins and DNA input (1/20 of the antibody-bound chromatins) were analyzed by PCR for the presence of the genomic fragment containing the HNF4 and C/EBP adjacent elements. The location of the primers is shown in the diagram and the sequences of primers are shown in Table I. All experiments in this figure were performed three times.

ABSTRACT
The pregnane X receptor (PXR) is a master regulator of genes involved in drug elimination. Activation of PXR is linked to the development of a spectrum of tumor behaviors, particularly chemoresistance. MicroRNAs (miRs) have emerged as important molecular species involved in tumor progression or suppression. This study was undertaken to test a large number of oncogenically implicated miRs for their ability to regulate PXR expression. A total 58 miRs (tumor enhancers or suppressors) were tested and miR-30c-1 was identified to suppress PXR expression. The suppression was achieved by targeting the 3'-untranslated region, 438 nucleotides from the stop codon. The suppression was detected in multiple cell lines from different organ origins. In addition, miR-30c-1 altered the expression of cytochrome P450 3A4 (CYP3A4), a prototypical target gene of PXR. The alteration varied depending on the time and amounts of miR-30c-1. CYP3A4 is responsible for the metabolism of more than half of drugs in the market. Interestingly, miR-30a, sharing the seed sequence with miR-30c-1 in the guide strand, failed to suppress PXR, suggesting that the suppression miR-30c-1 is mediated with the passenger strand. It has been reported that the expression of miR-30c is low in chemoresistant cell lines.
The interconnection between miR-30c-1 and PXR likely plays an important role in regulating tumor behaviors, particularly related to chemoresistance.

INTRODUCTION
All organisms are exposed constantly to toxic chemicals from both foreign and endogenous sources. Organisms such as humans have evolved several defensive systems against chemical insults . In mammals, these systems are generally referred to as phase I , phase II phase III (van Waterschoot andSchinkel, 2011). Phase I and II consist of drug-metabolizing enzymes, whereas phase III is drug transporters. The expression of these genes undergoes constant changes in response to chemical stimuli. The pregnane X receptor (PXR, NR1I2) is established as a master transcription factor intimately involved in the regulated expression of these genes .
Structurally, PXR belongs to the nuclear hormone receptor superfamily . Like other nuclear receptors, PXR consists of a variable N-terminal domain, a highly conserved DNA-binding domain, a hinge region and a multifunctional C-terminal ligand-binding domain .
The DNA-binding domain recognizes the canonical sequence AGG/TTCA (Modica et al., 2009). The major portion of the ligand-binding domain is helical in structure, and the C-terminal helix (helix 12) is directly involved in switching from repressing to activating status of a target gene . Binding to an agonist induces conformational changes of this helix, leading to a platform favoring association with coactivators, namely transactivation.
The expression of PXR itself, like its target genes, is drastically altered by certain xenobiotics and disease conditions . For example, the hypolipidemic agent clofibrate and synthetic glucocorticoid dexamethasone have been shown to induce PXR . The induction synergistically increased the expression of cytochrome P450 3A genes (CYP3A) , the prototypical targets of PXR . Dexamethasone induced PXR in both rodents and humans . Proinflammatory stimuli, on the other hand, have been shown to suppress the expression of PXR . The level of PXR mRNA was rapidly decreased in rodents treated with lipopolysaccharide, a potent immunostimulant . In human hepatocytes, pro-inflammatory cytokine interleukin-6 markedly reduced the levels of PXR mRNA . The suppression was accompanied by reduced induction on the expression of PXR-regulated genes such as CYP3A23 .
While transactivation and repression are recognized the major mechanisms in the regulated expression of PXR , post-transcriptional mechanisms have been increasingly implicated in the regulated expression of this nuclear receptor. MicroRNA (miR)-148a reportedly downregulated PXR post-transcriptionally, manifested by a significant induction of PXR mRNA . On the other hand, miRs are important regulators in a wide spectrum of diseases including malignancies and metabolic syndrome . These diseased conditions are inherently interconnected with inflammatory process, although the magnitude of inflammation may vary depending on the stage of diseases .
Interestingly, the majority of studies with human malignant tissues indicated increases in PXR expression, and activation of PXR has been associated with the development of a wide spectrum of tumor behaviors, particularly chemoresistance .
In this study, we took a comprehensive approach and tested a large number of miRs for their ability to regulate PXR expression. These miRs are implicated in the development of cancer, inflammation or both. A total 58 miRs were tested and miR-30c-1 was identified to suppress PXR. The suppression was achieved by targeting the 3'-untranslated region (UTR) and detected in multiple cell lines from different organ origin. Importantly, miR-30c-1 was shown to alter the expression of CYP3A4, a prototypical target gene of PXR. CYP3A4 is responsible for the metabolism of more than half medicines and many chemotherapeutic agents are PXR activators. Therefore, the PXR and miR-30c-1 connection is likely a major determinant in regulating tumor behaviors, particularly related to chemosensitivity.

Plasmid constructs
All miR precursor clones were purchased from System Biosciences Inc (Mountain View, CA). The CYP3A4-DP-Luc reporter was described in our previous publication . The PXR cDNA reporters harboring a 3'-UTR segment were prepared with the pGL3 promoter vector (Promega, Madison, WI) through Xba I and Fse I restriction endonuclease sites. The 3'-UTR segments were amplified by PCR with high fidelity Platinum Taq DNA polymerase (Life Technology Co., Carlsbad, CA). A cDNA clone encoding human PXR, used as the PCR template, was described elsewhere . The PXR 3142/3691Luc and 3690/4408Luc reporters were prepared initially and the PXR 3142/3691Luc reporter was used as the template for preparing deletion mutants at 5' end. The primers for PCR amplification are listed in Table I. All reporter constructs were subjected to sequence analysis.

Cell transfection and luciferase assay
Three cell lines were used in this study including 293T (human embryonic kidney), HepG2 (human hepatocellular carcinoma) and LS180 (human colon adenocarcinoma).
HepG2 and LS180 lines were purchased from American Type Culture Collection (Manassas, VA), but the 293T line was from GenHunter Corporation (Nashville, TN).
All cell lines were maintained in Dulbecco's modified eagle medium (DMEM) containing 10% fetal bovine serum, penicillin and streptomycin, 1x non-essential amino acids. Unless otherwise indicated, cells were plated in 48 well-plates and transiently transfected by GenJet version II from SignaGen Laboratories (Rockville, MD). For reporter assays, the transfection mixture typically contained 50 ng of a reporter, 50 ng of a miR construct and 5 ng of the null-Renilla luciferase plasmid.
After incubation at 37C for 24 h, cells were extensively washed, collected and assayed for luciferase activities with the Dual-Luciferase Reporter Assay System as described previously . The reporter luciferase activity was normalized with Renilla luciferase activity, and the vector-transfected cells served as the basal reporter activity for miRs-transfected cells. It should be noted that the same amount of total plasmids were used in all reporter assays.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
The LS180 cell line was primarily used for the RT-qPCR and because this cell line expresses high-levels of PXR and has been shown to robustly support PXR-directed transactivation . Cells were typically plated in 24-well plates for overnight and then transfected with the miR-30c-1 construct or the corresponding vector. Cells were harvested 72 or 96 h after the transfection. In some cases, cells were treated with DMSO or rifampicin (10 M) after transfection. Harvested cells were used for the preparation of total RNA. For the determination of PXR or CYP3A4 mRNA, total RNA (1 g) was subjected to the synthesis of the first strand cDNA as described previously . cDNAs were then diluted 8 times and RT-qPCR was conducted with TaqMan Gene Expression Assay (Applied Biosystems, Foster City, CA). The TaqMan probes were: PXR, Hs00243666_m1; CYP3A4, Hs00604506_m1; glyceraldehyde-3-phosphate dehydrogenase (GAPDH), conducted in a total volume of 20 l containing universal PCR master mixture (10 l), gene-specific TaqMan assay mixture (1 µl), and cDNA template (6 l). The mRNA levels were normalized according to the level of GAPDH and the normalization of selected samples was confirmed based on the signal of RNA polymerase II.
Amplification and quantification were done with the Applied Biosystems 7500 Real-Time PCR System.
3.1. Identification of miR-30c-1 as a suppressive miR of PXR PXR is a master regulator of gene expression and has been known to modulate the tumor behaviors, particularly related chemoresistance .
Like its target genes, we and other investigators have shown that the expression of PXR is regulated by factors such as age and therapeutic agents . While transactivation and repression are involved in regulated expression of PXR, emerging evidence suggests an involvement of post-transcriptional mechanisms, particularly through miR species . miRs constitute a superfamily of small RNA species, and many of them are implicated in the development of various types of tumor behaviors . To shed light on the missing link between PXR expression and miRs, we tested a large number of miRs (cancer enhancers or suppressors) for their ability to regulate the expression of PXR.
Some of the miRs are known to regulate inflammatory reaction. We took the advantage of miRs as predominantly post-transcriptional regulators , mRNA-based PXR reporters (cDNA reporters) were used for the initial study.
While there are exceptions, miRs usually target the 3'-UTR sequences. Therefore, we constructed two luciferase reporters that together harbor the entire 3'-UTR sequence of the dominant human PXR transcripts (Fig. 1A) (NM_003889). These two reporters were designated PXR3142/3691Luc and PXR3690/4408Luc, respectively.
Cotransfection was performed in 293T cells to determine the effect of a miR on the reporter activity. In addition, a renilla luciferase construct was included in the transfection. As shown in Fig. 1B, all miRs, with an exception of miR-30c-1, affected the activity of both reporters to a similar extent. Transfection of miR-30c-1, compared with the vector, resulted in decreased activity of PXR3142/3691Luc by 50%, but only 10% decrease on the activity of PXR3690/4408Luc. Some miRs such as miR-26A-2, compared with the vector, caused significantly increases in the activity of both reporters. Some others such as miR-433 significantly decreased the activity of both reporters (Fig. 1B). These miRs generally affected the renilla luciferase activity, which was used for normalizing transfection efficiency.

Sequence-specific targeting by miR-30c-1 independently of cell types
The screening study clearly demonstrated that human PXR mRNA is a sequencespecific target of miR-30c-1. To locate the sequence that supports the action of this miR, a serial of deletions from the 5' end were made on the PXR3142/3691Luc reporter and the resultant reporters were tested for the lost activity toward miR-30c-1.
As shown in Fig. 2A, all deletion mutants, except PXR3592Luc and PXR3642Luc, were repressed by miR-30c-1. These results established that the 50 base pairs, namely from 3542 to 3592 support the repression of miR-30c-1 on PXR expression. We are in the process of specifying the precise sequence and nucleotides that support the action of miR-30c-1 by site-directed mutagenesis.
It is well known that some miRs are processed in a cell-specific manner . We next tested whether the repression by miR-30c-1 occurs in HepG2 and LS180 cell line. HepG2 was derived from hepatocellular carcinoma whereas LS180 from colon adenocarcinoma. Importantly, both organs abundantly express PXR Zhang et al., 1999). In addition, the cotransfection was performed with various amounts of miR-30c-1 to establish the concentration-response relationship.
As expected, the PXR3542Luc but not PXR3592Luc reporter was repressed in both cell lines (Fig. 2B). HepG2 cells supported greater repression than LS180 cells. Overall, the magnitude of the repression occurred in a concentration-dependent manner (Fig. 2B).

Effect of miR-30c-1 on the expression of CYP3A4
The study with the reporters established that the action of miR-30c-1 on PXR suppression is sequence-specific but not cell-specific . Next we tested whether the repressive activity of the reporter can be recaptured on the expression of the endogenous gene of PXR. Initially, cells were transfected with miR-30c-1 or the corresponding vector, cultured for various lengths of time, and the levels of PXR mRNA were determined. Fig. 3A shows representative results of this study. The level of PXR mRNA was decreased in miR-30c-1 transfected cells, and the 96 h time-point showed less decrease than the 72 h timepoint (Left of Fig. 3A). The decrease in cells transfected with 20 ng plasmid of miR-30c-1 was less profound than that in cells transfected with 200 and 800 ng.
Nonetheless, 200 and 800 ng caused a comparable decrease in both time-points. Next we tested whether miR-30c-1 alters the mRNA level of CYP3A4, a prototypical target of PXR . While miR-30c-1 caused changes in the level of CYP3A4 mRNA, the changes varied depending on the amount of miR-30c-1 plasmid used for the transfection as well as the time after the transfection (Fig. 3A, Right) We next tested whether miR-30c-1 also alters the induction of CYP3A4. Cells were transfected with miR-30c-1 or the vector, and then treated with DMSO or rifampicin, a prototypical activator of human PXR. As shown in Fig. 3B (Left), the overall expression of CYP3A4 mRNA was decreased at both basal and induced level with an exception of the basal level in cells transfected with 20 ng plasmid of miR-30c-1. As described above, marked decreases in basal expression were detected in cells transfected with miR-30c-1 at 200 and 800 ng. As a result, these cells showed no changes in terms of fold of induction (Fig. 3B, Right). To further establish the role in reduced induction of CYP3A4, a CYP3A4 reporter was tested for the reduced activation in response to miR-30c-1 . Consistent with the result on the level of CYP3A4 mRNA, cotransfection of miR-30c-1 decreased the activation of this reporter by as much as 60% (Fig. 3C).

DISCUSSION
PXR has been established to play a major role in the expression of genes involved in drug elimination . However, emerging evidence has suggested that this nuclear receptor, in addition to drug elimination, is integrally connected with endobiotic signaling and homeostasis of energy balance . While PXR is recognized as a master regulator of gene expression, we and other investigators have reported that the expression of PXR is regulated as well . Multiple mechanisms are reportedly involved in the regulation of PXR expression including transactivation, repression and miR-148a silencing . This study identified and characterized miR-30c-1 as a silensor of PXR. A set of experiments have shown that miR-30c-1 targeted the PXR 3'-UTR and decreased the level of PXR mRNA, suggesting that miR-30c-1 suppresses PXR by decreasing the mRNA stability of this nuclear receptor. Importantly, this miR altered the expression of CYP3A4, a prototypical target of PXR , pointing to the functional consequences of miR-30c-1 drug elimination.
miR-30c-1 belongs to the miR-30 family, and the human genome has six miR-30 genes . However, these genes produce only four distinct mature guide strands. Importantly, these guide strands have the identical seed sequence, allowing them to regulate the expression of the same target genes, at least through the guide strand. This study, however, has demonstrated that miR-30a, a member of this family, failed to suppress the PXR reporters (Fig. 1B), suggesting that miR-30c-1 uses the passenger strand to target the PXR sequence. In addition to miR-30c-1, the miR-30c class has another member, namely miR-30c-2 . Both miRs are derived from intronic sequences of other genes. The miR-30c-1 gene is located at chromosome 1 and the miR-30c-2 gene is located at chromosome 6.
Importantly, miR-30c-1 and miR-30c-2 differ in the sequence of their passenger strand (Fig. 4A). Interestingly, the passenger strand of miR-30c-2 matches better than that of miR-30c-1 with the PXR 3-UTR region (Fig. 4A). Therefore, it is likely that miR-30c-2 is more potent than miR-30c-1 in silencing PXR, although the relative expression of these two miRs likely determines their contribution to PXR silencing.
In addition to miR-30c-2, miR-148a, another miR, reportedly silenced PXR . Based on the recognition sequences (Fig. 4B), both miR-30c and miR-148a target the same RNA species of PXR, although their recognition elements are 218 nucleotides apart (Fig. 4B). Nevertheless, it remains to be determined whether and how miR-148a networks with miR-30c in terms of regulating PXR expression.
The suppression of PXR by miR-30c-1 may have important clinical consequences. In this study, we have shown that miR-30c-1 decreased the level of PXR mRNA accompanied by altered expression of CYP3A4, a prototypical gene of PXR . The altered expression of CYP3A4, however, varied depending on time and the amounts of miR-30c-1. The overall expression of CYP3A4 mRNA was decreased under both basal and induction conditions (Fig. 3B, Left), but the decrease was not evident until later time-point (Fig. 3A, Right). Many investigators reported delayed response in miR-mediated suppression (Haenishch et al., 2011). The altered expression of CYP3A4 by miR-30c-1 was in particular as it represented a mechanism secondarily to the suppression of PXR. Such an indirect mechanism required longer time to achieve the anticipated effect. On the other hand, the basal level of CYP3A4 mRNA at the early time-point was actually increased when small amounts of miR-30c-1 were used (Fig. 3A, Right). The precise mechanism on the increase remains to be determined. Nuclear receptors including PXR are known to interact with corepressors and coactivators . It is the presence of a ligand that induces conformational changes from repressing to activating status of a target gene.
The functionality of PXR has been linked to a wide range of behavior changes of malignancy and some of them are opposing to each other . In some cases, activation of PXR up-regulates the expression of proapoptotic genes thus shows tumor suppressive activity. In other cases, PXR is linked to upregulation of antiapoptotic genes and favors tumor progression. Nevertheless, upregulated expression of genes involved in drug eliminations is generally considered to be a major contributing factor to the development of chemoresistance. Interestingly, miR-30c has been shown to promote cell apoptosis, inhibit cell proliferation, reduce tumor clonogenicity and suppress metastatic potentials . In addition, the expression of miR-30c was significantly decreased in many chemoresistant cell lines . It remains to be determined whether miR-30c members can overcome PXR-directed chemoresistance.
Interestingly, miR-148a, another silencer of PXR, has also been down-regulated in advanced cancer .
In summary, our study presents several important conclusions. Firstly, identification of miR-30c-1 as a negative regulator of PXR, along with the previous reporter on miR-148a, points the existence of a miR-network in regulating the functionality of PXR. Secondly, miR-30c-1 but not miR-30a suppressed PXR, underscoring the importance of the passenger strand in gene silencing. Thirdly, activation of PXR has been closely linked to the development of a spectrum of tumor behaviors, particularly in chemo resistance. miR-30c and miR-148a, on the other hand, have been associated with less aggressive behaviors of malignancy. These findings suggest that miR-30c/miR-148a-PXR represents favorable outcomes of chemotherapy, particularly those with a potent activating activity of PXR.

Fig. 4. Sequence targeted by miR-30c-1 and its location in the 3'-UTR of PXR
transcript (A) Sequence targeted by miR-30c-1 The sequence potentially targeted by miR-30c is identified by xx. Both miR-30c-1 and miR-30c-2 are shown as the passenger strand. (B) Location of the sequence targeted by miR-30c In addition to the location of the sequence targeted by miR-30c, the location of the sequence targeted by miR-148a is also shown. Specifically, miR-30c (both 30c-1 and 30c-2) likely targets the sequence from nucleotide 3583 to nucleotide 3660, whereas miR148a from 3359 to 3386, respectively. The sequence is numbered according to NM_003889.

CHAPTER 5 CONCLUSION
PXR is the master regulator of xenobiotic clearance. PXR regulates the expression of important genes in xenobiotic metabolism. CYP3A4, which metabolizes approximately 50% of marketed drugs, is mainly regulated by PXR Redinbo, 2004). Various diseases such as steatosis, diabetes, liver cancer, and viral infections are known to affect the expression of CYP3A4 and other PXR target genes. As a result, drug metabolism and elimination are altered in these conditions. ER stress and unfolded protein response is profoundly involved in the diseases. Thus, the effect of ER stress on PXR was studied to understand the consequences of ER stress on the expression of PXR and CYP3A4.

Effect of ER stress on PXR expression
ER stressors thapsigargin and brefeldin A caused ER stress in human hepatocytes and HepG2 cells, leading to down-regulation of PXR and reduced induction of CYP3A4. The expression of PXR was repressed at the transcriptional level. The promoter study revealed that HNF4α and C/EBPβ bound to the promoter region of PXR. Thapsigargin decreased HNF4α protein level whereas increased C/EBPβ LIP level. It is known that HNF4α generally functions as a positive transcriptional regulator and C/EBPβ LIP represses transcription of target genes (Cereghini 1996, Schrem et al., 2002. Therefore, downregulation of HNF4α and up-regulation of C/EBPβ LIP contributed to reduced transcription of PXR. The reporter assay showed that mutation at the HNF4α binding site abolished the repressive effect of thapsigargin. Overexpression of HNF4α or C/EBPβ LAP alone in the HepG2 cells treated with thapsigargin returned PXR mRNA to the normal level. The EMSA experiment showed that HNF4α bound to the promoter region of PXR and the binding decreased once HepG2 cells were treated with thapsigargin. The experiment proved that HNF4α bound to the binding site and thapsigargin-induced ER stress decreased HNF4α protein level and thus decreased bounded HNF4α. The ChIP assay also showed that HNF4α bound to the binding site in living cells and the binding of HNF4α decreased in thapsigargin-treated cells. Both the EMSA and ChIP assay showed that C/EBPβ bound to the binding site next to the HNF4α binding site. Thus, C/EBPβ would also contribute to the expression of PXR.
Thapsigargin-induced ER stress up-regulated IL-6 expression in human hepatocytes. IL-6 repressed the expression of the reporters at the same sequences as thapsigargin. This finding suggested that IL-6 involved in repression of PXR by thapsigargin. IL-6 is a pro-inflammatory cytokines, inducing acute or chronic inflammation which involves in several diseases such as viral infections and liver cancer . Many P450 enzymes are known to down-regulated during inflammation. Thus, the evidence that thapsigargin induced IL-6 showed that there were connection between ER stress and inflammation which led to down-regulation of PXR and CYP3A4.
The study showed that ER stress induced by ER-stressors decreased PXR expression and PXR-mediated induction of CYP3A4. The reduction of CYP3A4 induction in ER stress could diminish capability to metabolize and eliminate drugs, which could affect drug toxicity and efficacy. ER stress was connected to inflammation through the pro-inflammatory cytokine IL-6. Thus, diseases such as steatohepatitis, viral infections, and cancer in the state that involves ER stress and inflammation would decrease PXR expression.

miRNA regulating PXR
miRNAs are recognized as important regulators of gene expression at the posttranscriptional level. In human, more than 1000 miRNAs have been identified.
However, target genes of many of these miRNAs are not known and one miRNA tends to interact with more than one mRNA. In addition, a gene can be regulated by several miRNAs. Thus, miRNAs play important roles in regulating protein coding gene. Identifying miRNA that targets a gene of interest is challenging because factors that dictate binding between miRNA and the gene are not well understood. In this study, miR-30c-1 was identified to interact with the 3'-UTR of PXR. The potential binding site of miR-30c-1 was located based on the computational analysis and the reporter assay. Overexpression system showed that miR-30c-1 decreased PXR mRNA level, suggesting that miR-30c-1 decreased stability of PXR mRNA. In the reporter system, miR-30c-1 reduced CYP3A4 induction by PXR. The miRNA also altered the basal expression of CYP3A4.
The present study showed that PXR was regulated by miR-30c-1. The miRNA decreased the expression of PXR and the CYP3A4 induction. Thus, miR-30c-1 may contribute to alteration in drug metabolism. Since the miRNA regulated the expression of PXR, the expression of miR-30c-1 is likely to affect the PXR expression. At present, the regulation on the expression of miR-30c-1 is not known. miR-30c-1 might be related to diseases or biological conditions that could eventually affect PXR expression.
In summary, the studies showed that ER stress down-regulated the expression of PXR and reduced PXR-mediated CYP3A4 induction. The transcription of PXR was repressed. Decreasing of HNF4α and increasing of C/EBPβ LIP were shown to be the mechanism of repression. Since HNF4α and C/EBPβ bound to the promoter region of PXR, alteration in the levels of HNF4α and C/EBPβ in other conditions would also affect the expression of PXR. The transcriptional activation of CYP3A4 was decreased in ER stress due to reduction of PXR. Inflammation is known to downregulated CYP3A4 and some other P450s enzymes and impair drug metabolism.
Many of them are the PXR target genes. The pro-inflammatory cytokine IL-6 was also induced during ER stress, proving the link between ER stress and inflammation. Thus, ER stress could affect drug toxicity and efficacy. PXR was also regulated by miR-30c-1. The miRNA down-regulated the expression of PXR at the specific binding site on the 3'-UTR. Both the basal expression and induction of CYP3A4 were altered by miR-30c-1. Thus, miR-30c-1 affected both the expression and function of PXR and may contribute to alteration in drug metabolism.