Regulation of PXR in drug metabolism: chemical and structural perspectives

Abstract

Introduction:

Pregnane X receptor (PXR) is a master xenobiotic sensor that transcriptionally controls drug metabolism and disposition pathways. PXR activation by pharmaceutical drugs, natural products, environmental toxins, etc. may decrease drug efficacy and increase drug-drug interactions and drug toxicity, indicating a therapeutic value for PXR antagonists. However, PXR’s functions in physiological events, such as intestinal inflammation, indicates that PXR activators may be useful in certain disease contexts.

Areas covered:

We review the reported roles of PXR in various physiological and pathological processes including drug metabolism, cancer, inflammation, energy metabolism, and endobiotic homeostasis. We then highlight specific cellular and chemical routes that modulate PXR activity and discuss the functional consequences. Databases searched and inclusive dates: PubMed, January 1, 1980 to January 10, 2024

Expert opinion:

Knowledge of PXR’s drug metabolism function has helped drug developers produce small molecules without PXR-mediated metabolic liabilities, and further understanding of PXR’s cellular functions may offer drug development opportunities in multiple disease settings.

1. Introduction

Nuclear receptors (NRs) are transcription factors with diverse roles in embryonic development and homeostasis in adults [1,2]. The NR1I subfamily contains three of the 48 members of the human nuclear receptor family: pregnane X receptor (PXR), vitamin D receptor (VDR), and constitutive androstane receptor (CAR) [2,3]. All three members can induce the expression of broad-specificity hepatic and intestinal phase I enzymes that metabolize xenobiotics and endogenous compounds. PXR and CAR can also induce the expression of phase II conjugating enzymes and phase III drug transporters [4]. CAR was first described in 1994 and later found to be a transcriptional regulator of cytochrome P450 (CYP)2B6 [5–7], and PXR was first reported in 1998 as the main ligand-dependent regulator of CYP3A4 transcription [8–10]. PXR was originally considered an ‘orphan’ receptor because no ligand had been discovered until Kliewer et al. [10] reported that several steroids, including naturally occurring pregnanes, synthetic glucocorticoids, and antiglucocorticoids, activated PXR. Additional PXR ligands include endobiotics, prescription drugs, environmental chemicals, and herbal supplements [8,9,11]. Based on the broad ligand binding profiles of PXR and CAR, these receptors likely evolved to protect humans against toxic levels of both exogenous and endogenous compounds [4]. Indeed, the observed species-specific differences in PXR ligand specificity suggest that different environments played a role in this evolutionary adaptation [12].

In this review, we cover PXR structure and functions, focusing on the interplay between PXR’s unique structural features, chemical modulators, and posttranslational modifications (PTMs). We then discuss how these modulators and modifications can be manipulated to regulate PXR activity.

2. Overview of PXR functions

Although PXR was initially characterized as a xenobiotic receptor, more recent studies have revealed that PXR has additional functions beyond drug metabolism (Figure 1). The following sections will give a brief summary of PXR’s drug metabolism role and then primarily focus on specific examples of PXR’s roles in cancer and inflammation, as well as its physiological functions in energy metabolism (e.g., glucose metabolism and lipogenesis) and detoxification (e.g., bile acid transport).

Figure 1. Overview of PXR functions.

PXR is involved in many physiological and pathological processes through interacting with other proteins and transcriptionally activating various target genes. Arrows indicate activation or interaction; stop bars indicate suppression or inhibition.

2.1. PXR in drug metabolism

PXR is a ligand-activated transcription factor that regulates the expression of drug-metabolizing enzymes [e.g., CYP3A4, CYP2C9, CYP3A7, and UDP glucuronosyltransferase family 1 member A1 (UGT1A1)] and drug transporters [e.g., multidrug resistance 1 (MDR1) and ATP binding cassette subfamily C member 2 (ABCC2)] (Figure 1) [4,9]. The CYP3A family plays a critical role in the biotransformation of drugs, as it catalyzes the metabolism of more than half of all clinical drugs in use, which is mainly attributed to CYP3A4 [9,13,14]. Commonly used prescription drugs, including dexamethasone, clotrimazole, and rifampicin, and a constituent of the herbal supplement St. John’s wort (hyperforin) induce CYP3A4 expression by activating PXR, which can cause drug-drug interactions in which one drug accelerates the metabolism of a second medication and may cause adverse events [9,15]. For example, co-administration with a PXR activator can increase clearance of the immunosuppressant drug cyclosporine or increase metabolism of the estrogen component of combined oral contraceptives, which could cause organ rejection in an allograft recipient or an unintended pregnancy, respectively [4]. Since the discovery of PXR’s regulatory role in the expression of drug metabolism pathway genes, the pharmaceutical industry has incorporated assays to decrease or eliminate PXR activation by drug candidates to reduce the risk of drug-drug interactions or even liver injury from reactive metabolites [16–18].

2.2. PXR in cancer

PXR expression was first reported in human liver, colon, and small intestine [9], and PXR has also been found in human breast, stomach, adrenal gland, bone marrow, specific regions of the brain, and at the blood-brain barrier [19–22]. Additionally, PXR is highly expressed in various cancers, such as pancreatic, hepatobiliary, esophagogastric, and colorectal cancers [23]. Given its ability to upregulate CYP3A enzymes and drug transporters, PXR has been shown to regulate drug efficacy and contribute to chemotherapy resistance in prostate and breast cancer, pancreatic ductal adenocarcinoma, and ovarian carcinoma (Figure 1) [24–27]. PXR has also been shown to exert an antiapoptotic effect and promote malignant transformation in colon cancer cells [28–30]. Furthermore, PXR is expressed in cancer stem cells and drives the expression of genes involved in self-renewal and chemoresistance, thus promoting colon cancer relapse [31]. A recent study demonstrated that the anti-helminthic drug niclosamide induces miR-148a expression, decreasing PXR expression and sensitizing colon cancer stem cells to chemotherapy [32]. Although these studies highlighted how PXR expression promotes chemoresistance and tumor progression, PXR’s role in cancer is convoluted by the tissue context and protein associations, such as its reported interaction with the tumor suppressor protein p53 [1]. For example, wild-type p53 was shown to interact with PXR and downregulate PXR transcriptional activity [33]. In addition, PXR activation inhibited apoptosis in HCT116 and LS180 colon cancer cells by downregulating wild-type p53 and other pro-apoptotic genes [28]. However, in contrast to this pro-survival role, PXR activation in breast cancer cells containing wild-type p53 had an antiproliferative effect dependent upon p53 upregulation in response to cellular stress caused by accumulation of reactive nitrogen species [34]. PXR overexpression also suppressed proliferation, but without changing the p53 protein level, in HT29 colon cancer cells that carry mutated p53 [35]. The PXR-p53 interaction leads to mutual inhibition, where p53 increases cancer cell death in response to chemotherapeutics and inhibits PXR to decrease drug metabolism and increase drug efficacy. However, PXR has an oncogenic function by contributing to drug resistance and inhibiting p53 to decrease apoptosis [36]. Thus, the p53 status, PXR protein level, and tissue context are crucial determinants in dictating the cellular outcome of tumor suppression or growth.

2.3. PXR in inflammation

In addition to transcriptionally controlling drug detoxification and clearance genes, PXR acts as a transcriptional suppressor of inflammation in the liver and intestine. Small intestine inflammation is increased in Pxr-null mice [37], and treatment of wild-type mice or primary human hepatocytes with the proinflammatory cytokine interleukin-6 (IL-6) led to decreased levels of PXR and its target genes (Figure 1) [38–40]. Similarly, treating human liver HepG2 cells with the proinflammatory cytokine tumor necrosis factor alpha (TNF-α) attenuated CYP3A4 induction by rifampicin [41]. Mechanistically, PXR and proinflammatory NF-κB pathways display inhibitory crosstalk (Figure 1). Treatment with PXR agonists in human intestinal cells repressed NF-κB-targeted proinflammatory genes, but activation of NF-κB signaling by TNF-α suppressed PXR target gene induction by preventing the PXR-RXR heterodimer from binding target DNA sequences [37,42].

Consistent with PXR’s anti-inflammatory function, altered expression or activity of PXR can contribute to inflammatory diseases like acute kidney injury, atherosclerosis, and inflammatory bowel disease (IBD). Overexpression or activation of PXR was protective against acute kidney injury in rodent models by suppressing inflammation and upregulating Akr1b7 to attenuate mitochondrial dysfunction [43,44]. Increased atherosclerosis was observed in PXR-humanized mice exposed to bisphenol A, a potent human PXR agonist [45,46], and specific deletion of PXR in myeloid cells decreased atherosclerosis [47]. Polymorphisms in PXR are associated with an increased risk of susceptibility to IBD, ulcerative colitis, and Crohn’s disease [48,49]. In ulcerative colitis patients, disease severity was inversely correlated with the expression of PXR and its target genes that detoxify and defend colon cells from injury; thus, dysregulation of PXR activity may contribute to the pathophysiology of IBD [50].

2.4. PXR in energy metabolism

PXR affects energy metabolism through crosstalk with several transcription factors, such as forkhead box O1 (FOXO1), FOXA2, cAMP-response element binding (CREB) protein, and peroxisome proliferator-activated receptor gamma (PPARγ) coactivator-1α (PGC-1α). The outcome is increased lipogenesis and decreased gluconeogenesis and fatty acid oxidation (Figure 1) [51]. Below, we discuss PXR’s role in selected metabolic processes.

PXR activation has been shown to suppress gluconeogenesis by crosstalk with multiple pathways. Interaction between PXR and FOXO1 or CREB prevents transcription of two rate-limiting enzymes in gluconeogenesis, glucose-6-phosphatase (G6Pase) and phosphoenolpyruvate carboxykinase 1 (PEPCK1) [51–53]. PXR activation by rifampicin in a rat model downregulated additional enzymes involved in glucose metabolism: glucose transporter 2, pyruvate dehydrogenase kinase isoenzyme 2, and glucokinase [54]. PXR can also suppress gluconeogenesis by competing for PGC-1α binding with hepatocyte nuclear factor 4α (HNF4α), which requires PGC-1α to positively regulate gluconeogenesis [55]. Though these data suggest that PXR activation may be beneficial in treating type 2 diabetes and obesity [56], another report suggests that PXR antagonists could be useful in treating these diseases, as the ablation of PXR in mice prevented diet-induced and genetic obesity [57].

Another physiological function of PXR is the regulation of lipid homeostasis, particularly the repression of β-oxidation and induction of lipogenesis. In mice, pregnenolone-16α-carbonitrile (PCN)-activated PXR repressed the FOXA2-mediated transcription of carnitine palmitoyl transferase 1A (CPT1A) and 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2), which are key enzymes in β-oxidation and ketogenesis [58]. Activation of PXR in HepG2 cells or primary human hepatocytes promoted steatosis by upregulation of the sterol regulatory element-binding protein (SREBP)1-lipogenic pathway. Interestingly, knockdown of PXR also increased de novo lipogenesis via up-regulation of aldo-keto reductase AKR1B10 [59]. In a humanized mouse model, rifampicin-mediated PXR activation reduced expression of pro-β-oxidative genes (e.g., 3-ketoacyl-CoA thiolase, which catalyzes β-oxidation of long-chain fatty acids), induced expression of PPARγ and the free fatty acid transporter CD36, and resulted in hepatic lipid accumulation (Figure 1) [60,61]. Some lipid homeostasis pathways seem species-specific, as PXR activation in primary human hepatocytes promoted lipogenesis without affecting CD36 by upregulating thyroid hormone-responsive spot 14 protein and fatty acid synthase [62]. Long-term treatment with the PXR agonist rifaximin in a PXR-humanized mouse model led to hepatocellular fatty degeneration caused by PXR-dependent upregulation of lipid accumulation and triglyceride synthesis genes [63]. These data suggest that PXR may promote aberrant hepatic lipogenesis and cause hepatic steatosis.

Besides hepatic steatosis, PXR may be involved in the pathogenesis of metabolic dysfunction-associated fatty liver disease (MAFLD). PXR single nucleotide polymorphisms have been associated with disease severity [64]. Comparative histological analysis of MAFLD liver samples revealed decreased levels of PXR protein and increased levels of SREBP1a and lipogenic target genes [59]. Through PXR-mediated activation of increased NLRP3 inflammasome in macrophages, PXR may aggravate MAFLD severity [65,66]. Furthermore, PXR ablation in two different alcoholic liver disease mouse models led to higher levels of alcohol dehydrogenase 1 (ADH1) and aldehyde dehydrogenase 1A1 (ALDH1A1) [67,68]. However, the role of PXR in alcoholic liver disease is still unclear, as PXR protected mice against alcoholic hyperlipidemia in the binge ethanol model [67] but contributed to alcoholic liver disease development in the chronic ethanol-induced hepatosteatosis model [68].

Cholesterol is an essential lipid component in cell membranes and a precursor for steroid hormones. PXR helps protect the body from excess cholesterol, which can increase the risk of liver and cardiovascular disease [69]. Oxysterols and cholesterol metabolites act as endogenous PXR ligands to induce CYP3A expression [70,71]. In intestinal cells, PXR activation induced expression of CYP27A1, which is required for cleavage and hydroxylation of cholesterol (Figure 1) [72]. Treatment with PXR agonists in rodents led to elevated high-density lipoprotein and apolipoprotein A-I serum levels, suggesting that PXR is involved in the regulation of ‘good cholesterol’ [73,74]. However, more recent studies indicate that PXR activation in rodents increases hepatic cholesterol synthesis and induces hypercholesterolemia and that PXR activation in humans elevates low-density lipoprotein and total cholesterol; thus, PXR-activating drugs may pose a cardiovascular risk and accelerate atherosclerosis [75–79].

2.5. PXR in endobiotic detoxification

PXR’s function as a xenobiotic sensor protects cells from chemical harm by promoting metabolism and excretion of foreign compounds. Likewise, PXR protects cells from toxic endobiotics, such as bile acids. Bile acids, the end products of cholesterol catabolism, are cytotoxic when in excess and may lead to cholestatic liver disease [80]. PXR plays a critical role in preventing this pathophysiology through its regulation of bile acid transporters [e.g., ABCC2, ABCC3, and organic anion transporting polypeptide 2 (OATP2)] and CYP3A enzymes that hydroxylate and excrete bile acids (Figure 1) [81–84]. In addition, PXR regulates the expression of cholesterol 7α-hydroxylase (CYP7A1), the rate-limiting enzyme of cholesterol catabolism and bile acid formation [81]. Similarly, PXR regulates critical proteins involved in detoxifying bilirubin, the neurotoxic breakdown product of hemoglobin [15]. Activation of PXR induces expression of OATP2 to mediate hepatocellular uptake of bilirubin from the blood [81], UGT1A1 to conjugate bilirubin with glucuronide [85], and ABCC2 to facilitate excretion of conjugated bilirubin to the bile canaliculus (Figure 1) [15,83].

3. Perspective: Future directions in regulation of PXR

Protein function can be modulated by numerous mechanisms, such as protein-protein interactions, PTMs, alternative splicing, microRNA binding, and ligand binding. Here, we discuss the two main routes of PXR modulation, ligand binding and PTMs, and highlight potential future directions in each area.

3.1. Structural approaches to modulate PXR ligand binding

Like other NRs, PXR’s modular structure consists of the DNA-binding domain (DBD) that has two zinc-finger motifs to mediate binding to sequence-specific response elements in the promoters of target genes, a hinge region, and the ligand-binding domain (LBD) that mediates ligand binding, ligand-independent repression, and dimerization [2–4]. The LBD also contains the AF-2 ligand-induced activation domain (alpha helix 12 of the LBD) necessary for interaction with transcriptional coregulators [2,3]. Upon ligand binding, PXR follows the basic NR activation pathway of translocation to the nucleus, heterodimerization with RXR, binding to response elements in target gene promoters, and recruitment of coactivator and other transcription machinery [23,86,87]. Coactivators enhance the transcriptional activity of activated NRs by recruiting additional transcriptional machinery proteins for efficient gene expression [88]. Known PXR coactivators, which include members of the p160 family [e.g., steroid receptor coactivator (SRC)-1, SRC-2, and SRC-3], PPARα–binding protein (PBP), and PGC1-α, contain one or more repeats of the LXXLL motif (where L is leucine and X is any other amino acid residue) that is essential for ligand-dependent interaction with the LBD (Figure 2) [10,89–91]. Conversely, PXR transcriptional activity is inhibited by corepressors like nuclear receptor corepressor 1 (NCoR), silencing mediator for retinoid or thyroid hormone receptors (SMRT), receptor-interacting protein 140 (RIP140), and short heterodimer protein (SHP), which contain the (I/L)XX(I/V)I motif that facilitates interaction with PXR (Figure 2) [19,91,92]. When an agonistic ligand binds, NRs undergo a conformational change that disrupts corepressor binding and enables coactivator recruitment [2]. The hormone response elements recognized by PXR are composed of the nucleotide hexamer AG(G/T)TCA with variable nucleotide spacing and may be arranged as direct, inverted, or everted repeats [9,10,15,93,94].

Figure 2. Regulation of PXR is multifaceted.

PXR activity may be enhanced (arrows) or suppressed (stop bars) by ligands, chemical modulators, cofactors, and PTMs.

Compared to other NRs, the ligand-binding pocket of PXR is large (1,200 to 1,600 ų), flexible, and dynamic, enabling PXR to accommodate a wide range of ligands with diverse structural and physicochemical properties (Figure 3) [12]. Due to these unique properties of the ligand-binding pocket, ligands may adopt multiple binding poses [95], multiple ligands can simultaneously bind to PXR [96,97], and the binding pocket can expand to accommodate large ligands [98,99]. Ligand-binding pocket expansion is unfavorable for PXR binding affinity and activation, a mechanism that may be used to reduce PXR binding by engineering directed rigidity into drug candidates (Figure 4) [98]. In contrast, the flexibility of PXR’s ligand binding pocket may be key to finding molecules that are selective for PXR over other NRs [100]. Ligand binding affects the position of PXR’s AF-2 helix within the LBD, which impacts binding of coactivators or corepressors to ultimately affect the cellular outcome [12,101]. Using PXR mutants coupled with structurally similar ligands, a recent study pinpointed two distinct ligand-induced AF-2 conformations as a mechanism of PXR agonism/antagonism [101]. Manipulating the unique structural features of PXR to achieve ligand specificity is key to producing the desired biological impact without undesirable nonspecific effects [100].

Figure 3.

Chemical structures of PXR agonists and antagonists.

Figure 4. Binding affinity to PXR decreases with ligands that expand the ligand binding pocket.

PXR LBD crystal structures in complex with T0901317 (PDB ID 2O9I), rifamycin S (PDB ID 8E3N), and rifampicin (PDB ID 1SKX) are shown. Binding of the small, high-affinity ligand T0901317 results in a compact ligand-binding pocket while ligands of increasing size expand the pocket, causing structural disorder and exposure of the ligand.. Binding pocket expansion allows PXR to recognize large molecules, but the mechanism comes with a penalty to binding affinity.

3.2. Context-specific benefits of agonists/antagonists

Because of PXR’s multifaceted functions in maintaining physiological homeostasis, both activation and inhibition of PXR can be beneficial, depending on the context (Figure 1). Thus, PXR agonists, antagonists, and inverse agonists are actively being developed to serve as research tools and drug candidates (Figure 3 and Table 1) [102,103]. As PXR activation transcriptionally upregulates drug metabolism processes, possibly decreasing drug efficacy and increasing therapy resistance, inhibiting PXR activity through antagonists or inverse agonists may be beneficial in counteracting these adverse drug reactions [101]. For example, Niu et al. [127] recently demonstrated the therapeutic potential of combining the chemotherapeutic paclitaxel with the PXR antagonist SPA70 [110] for reversing paclitaxel-resistant non-small cell lung cancer. PXR antagonists may also offer therapeutic benefits in atherosclerosis management [66], an idea supported by the decreased atherosclerosis observed in a mouse model with myeloid-specific PXR deficiency [47].

Table 1.

Effect of selected agonists, antagonists, and PTMs on PXR activity

Category	Name	Effect on PXR Activity	References
Agonist	Rifampicin	↑	[9]
	Hyperforin		[104,105]
	SR12813		[106]
	T0901317		[107]
	17α-ethinylestradiol		[108]
	trans-nonachlor		[109]
	SJB7		[110]
	Garcinoic acid		[111]
	Bisphenol A		[112]
Antagonist	SJC2	↓	[110]
	SPA70		[110]
	Compound 89 (SJPYT-310)		[103]
PTMs	Phosphorylation	↓	[113–119]
	SUMOylation	↑	[120,121]
		↓	[120,122,123]
	Acetylation	↓	[123–125]
	Ubiquitination	↓	[114,115,126]

Conversely, activation of PXR has been suggested as a treatment option for inflammatory liver and bowel diseases [128]. By inhibiting the expression of NF-κB target genes, PXR activation served a protective role in a mouse model of IBD [129]. PXR also plays a vital role in maintaining gut barrier integrity. Indole metabolites produced by symbiotic bacteria act as physiological PXR agonists to downregulate inflammatory response and upregulate cell-cell junctional complex markers in the intestinal epithelium [1,130]. Through this mechanism, PXR protects against inflammatory intestinal injury by negatively regulating toll-like receptor 4 (TLR4) [130]. Follow-up studies demonstrated that synthetic indole derivatives interacted with PXR as agonists and reduced colon inflammation in an acute colitis mouse model [131,132]. These studies revealed that PXR is a key player in homeostatic host-microbe interactions in the intestine and demonstrated the therapeutic potential of PXR agonists [133].

3.3. PTMs of PXR

Phosphorylation, acetylation, SUMOylation, poly(ADP-ribosyl)ation [134], and ubiquitination of PXR have all been reported (reviewed in [135–137]), and these PTMs can alter subcellular localization, protein stability, and protein interactions to dynamically regulate PXR activity. Multiple mechanisms modify PXR, and PTMs may compete for the same residues, such as lysine [138]. These modifications may contribute to ligand response (i.e., “on” or “off” switch) or may fine-tune NR activity, as seen with RXR Ser265 phosphorylation determining target gene selectivity [139]. Even without a ligand, PTMs can modify the activity of NRs, which could affect other cellular processes and contribute to disease progression [2]. The following subsections will focus on the evidence for phosphorylation, acetylation, SUMOylation, and ubiquitination of PXR, as well as how each modification affects PXR transcriptional activity (Table 1).

3.3.1. Phosphorylation

Phosphorylation plays an important role in NR regulation by enabling integration of various signals and rapid adaptation to environmental and physiological situations [19]. Although phosphorylation of NRs may contribute to their activation or repression [140], phosphorylation of PXR usually represses activity (Figure 2) [135,141]. In one of the earliest studies on PXR phosphorylation, treating primary mouse hepatocytes with okadaic acid (protein phosphatase PP1/PP2A inhibitor) or phorbol 12-myristate 13-acetate [PMA; a protein kinase C (PKC) activator] repressed ligand-dependent PXR activity [142]. Mammalian two-hybrid analysis showed that PMA strengthened PXR’s interaction with the corepressor NCoR and inhibited its interaction with the coactivator SRC-1 [142]. Similarly, activation of PKA signaling by 8-Br-cAMP promoted PXR-NCoR interaction to repress CYP3A4 expression [143]. Unlike the altered cofactor recruitment mediated by PKA and PKC activation, phosphorylation of Thr57 by p70 S6K inhibits PXR transcriptional activity by disrupting its DNA binding [113].

Additional kinases have been shown to phosphorylate PXR in vitro, including glycogen synthase kinase 3 (GSK3), casein kinase 2 (CK2), cyclin-dependent kinase (CDK) 1 [143], CDK5 [144], and CDK2 [116]. Of these reported kinases, phosphorylation of PXR by CDK2 has been studied more extensively. Treating cells with CDK inhibitors (e.g., roscovitine, kenpaullone, and dinaciclib) has been shown to enhance expression of PXR target genes CYP3A4, MDR1, and UGT1A1 [115–118]. Exogenous expression of PXR point mutants indicated that CDK2 likely phosphorylates PXR at Ser350, which abrogates PXR transcriptional activity by inhibiting RXR heterodimerization and coactivator recruitment [116,117,119].

Phosphorylation may also suppress PXR activity by affecting its stability. Sugatani et al. [145] reported that phosphorylation of Thr408 by PKC signals for PXR to enter the CHIP/chaperone-dependent stability check, leading to degradation through autophagy. The kinases involved in PXR phosphorylation may also affect PXR stability through phosphorylating interacting proteins. For example, ligand-activated CK2 phosphorylates HSP90β, which interacts with PXR and increases PXR stability, leading to induction of MDR1 expression [146]. It has been suggested that PXR phosphorylation may serve as an impetus for other PTMs like acetylation and ubiquitination [115,117]. Consistent with this idea, PXR phosphorylation by CDK2 or DYRK2 has been shown to facilitate its subsequent ubiquitination by TRIM21 or UBR5, respectively [114,115].

Although much, if not all, literature on PXR phosphorylation relies on exogenous overexpression of PXR in cell models (partly due to low expression level of endogenous PXR), this method has enabled the identification of several residues that are important for PXR activity [1]. More sensitive techniques are needed to detect and study phosphorylation of endogenous PXR to determine whether the in vitro studies accurately depict effects of PXR phosphorylation in vivo. Further research is essential to dissect the interplay between PXR phosphorylation and other PTMs.

3.3.2. Acetylation

After first confirming the presence of acetylated PXR in vivo, Biswas et al. [124] observed basal acetylation of PXR and decreased PXR acetylation when cells were treated with rifampicin or transfected with the deacetylase SIRT1. Pharmacological inhibition of deacetylation (i.e., with either a SIRT1 inhibitor or a broad inhibitor of class I and II lysine deacetylases) increased PXR acetylation and suppressed induction of PXR target genes (Figure 2) [123,124]. Primary mouse hepatocytes treated with the SIRT1 activator resveratrol exhibited increased lipid accumulation, suggesting that PXR acetylation can regulate lipogenesis independent of ligand activation, as hepatocytes from PXR knock-out mice were unaffected [124]. Another group found that SIRT1 interacted with PXR and was able to attenuate PXR coactivation by PGC-1α [147].

Additional proteins have also been reported to regulate PXR acetylation. Pasquel et al. [125] found that acetylation of PXR at Lys109 by p300 repressed PXR transcriptional activity by interfering with RXR heterodimerization and binding to DNA response elements. The lysine acetyltransferase TIP60 can interact with unliganded PXR and acetylate Lys170 [148]. This TIP60-PXR complex altered PXR intranuclear localization and promoted cell migration and adhesion, and rifampicin disrupted PXR interaction with TIP60 [148]. Consistent with the idea that acetylation impairs PXR activity, the phosphomimetic S350D mutant was detected as acetylated after roscovitine treatment, and acetylation of wild-type PXR likely was not detected because roscovitine treatment increased PXR interaction with the deacetylase HDAC1 [117]. PXR has also been shown to associate with the HDAC3-SMRT corepressor complex, with HDAC3 deacetylating PXR upon ligand activation, and Cui et al. [123] put forth a working model of a “SUMO-acetyl switch” where acetylation of PXR serves as a prerequisite for subsequent SUMOylation.

3.3.3. SUMOylation

SUMOylation is like ubiquitination in that an enzyme cascade culminates in the attachment of a small ubiquitin-like modifier (SUMO) protein to lysine residues in target proteins [137]. Of the three SUMO isoforms widely expressed in humans, SUMO2/3 can form polymeric or mixed ubiquitin chains, while SUMO1 is typically added as a single moiety [149,150]. SUMOylation may stabilize proteins and modulate protein-protein interactions, subcellular localization, or DNA binding to regulate the activity of target proteins [149]. In the case of NRs, SUMOylation has been shown to either repress or potentiate transcriptional activity, depending on the particular NR being investigated [151–155].

There are mixed reports on the impact of SUMOylation on PXR activity (Figure 2). In an early study, PXR was shown to serve as a substrate for all three SUMO proteins in vitro, and subsequent cell-based experiments revealed that SUMO3 chains preferentially modified PXR in response to rifampicin treatment [156]. SUMOylated PXR level also increased with TNF-α treatment, which repressed NF-κB reporter gene activity without affecting CYP3A4 reporter gene activity, leading the authors to speculate that SUMOylation of PXR represses inflammatory response in hepatocytes [156]. Cui et al. [120] later identified PIAS1 and PIAS4 as the SUMO-E3 ligases that most efficiently SUMOylated PXR. Interestingly, PIAS1-mediated SUMO(3)ylation enhanced PXR induction of CYP3A genes, while PIAS4-driven SUMO(1)ylation repressed PXR transcriptional activity by disrupting recruitment of the coactivator PGC-1α [120]. De-SUMOylation of PXR by SENP2 significantly decreased PXR protein levels, indicating that SUMOylation may stabilize PXR protein [120,137]. It has also been shown that SUMO(1)ylation of rifampicin-activated PXR enhanced expression of target genes (i.e., CYP3A4, CYP2C9, MDR1, and UGT1A1) without affecting PXR protein stability [121]. In contrast to these findings, Tan et al. [122] reported that PXR SUMO(3)ylation attenuated rifampicin-induced expression of CYP3A4 and MDR1 in LS174T cells. To more precisely and directly determine how SUMO1 or SUMO3 modifications affected PXR transcriptional activity, Cui et al. [123] tested PXR-SUMO fusion proteins and found that covalent attachment of SUMO1 or SUMO3 drastically reduced rifampicin-inducible CYP3A4 reporter gene activity. Many studies on PXR SUMOylation relied on the exogenous expression of PXR and components of the SUMO enzyme cascade (i.e., SUMO1, SUMO3, PIAS1, PIAS4), which could increase SUMOylation of many other co-regulatory proteins and contribute to the different effects on PXR activity that have been observed.

3.3.4. Ubiquitination

Ubiquitination is mediated by E1 ubiquitin-activating enzymes, E2 ubiquitin-conjugating enzymes, and E3 ubiquitin ligases, with E3s directly or indirectly facilitating transfer of ubiquitin from the E2 to lysine residues in target proteins [157,158]. In humans, there are two E1 enzymes, 35 E2 enzymes, and over 600 E3 ligases; this increasing specificity facilitates precise regulation of ubiquitination [159]. The type of ubiquitin modification dictates the fate of the ubiquitinated substrate. Polyubiquitinated proteins with K48- and K11-linkages are usually directed to the proteasome for degradation, whereas monoubiquitination is associated with nonproteolytic functions, like subcellular localization and endocytosis [157–159]. E3 ubiquitin ligases are subdivided into three categories: really interesting new gene (RING) E3s, homologous to the E6-associated protein carboxyl terminus (HECT) E3s, and RING-between-RING (RBR) E3s [160]. The RING E3s, the most abundant group, function as scaffolds that facilitate transfer of ubiquitin from an E2 to the target protein [161]. In contrast, HECT E3s possess intrinsic catalytic cysteines that accept ubiquitin from an E2 by forming a covalent thioester intermediate before conjugating ubiquitin to the substrate [162]. The RBR E3s employ a ‘hybrid’ mechanism of ubiquitin transfer, where the RING1 domain binds the E2 in a RING E3-like fashion, and a catalytic cysteine in the RING2 domain forms a reactive HECT-like E3-ubiquitin thioester intermediate that facilitates ubiquitin transfer to the target protein [163].

Ubiquitination of NRs can alter their protein levels and transcriptional activity, particularly by helping the NR detach from the promoter it regulates [159]. This ubiquitin-mediated promoter clearance mechanism has been reported for other NRs PPARγ and LXRα/β [164,165], so ubiquitination of PXR may also serve as a promoter clearance mechanism [136]. In support of this idea, pharmacological inhibition of the proteasomal degradation pathway prevented PXR transcriptional activity as measured by CYP3A4 reporter gene activity [136]. Degradation of PXR may be differentially affected by agonists. Suppressor for gal 1 (SUG1), a subunit of the 26S proteasome, interacted with PXR in the presence of progesterone but no other ligands, leading to differential degradation of PXR [166,167]. Like these observed agonist-specific effects on PXR degradation, ubiquitination of exogenously expressed PXR markedly increased upon treatment with TNF-α or the PKA activator 8-Br-cAMP [120,136]. Although rifampicin treatment led to increased ubiquitination of overexpressed human PXR in mouse hepatocytes [120], no changes were observed in PXR ubiquitination or protein levels upon rifampicin treatment in HeLa, AD-293, or HepG2 cells [115,136,168]. The differential effects of rifampicin on PXR ubiquitination may be due to the species (i.e., mouse vs. human) or cell models used.

Besides ligands, other PTMs may affect or even precede ubiquitination of PXR. Compared to wild-type PXR, the phosphomimetic T408D mutant accumulated more in the cytoplasm after proteasomal inhibition with MG-132, suggesting Thr408 phosphorylation may act as a signal for PXR to be ubiquitinated and degraded by the proteasome [117]. Two studies have reported that PXR protein stability is negatively regulated by kinases (i.e., DYRK2 and CDK2) that phosphorylate PXR, facilitating its subsequent ubiquitination [114,115]. In addition, overexpressing PIAS4 and SUMO1 with His-tagged ubiquitin and PXR increased both SUMOylation and ubiquitination of PXR, possibly due to formation of mixed SUMO-ubiquitin chains or SUMO-dependent ubiquitination [120].

As has been demonstrated with two different E3 ligases mediating ubiquitination of the nuclear receptor Rev-erbα [169], multiple E3 ligases have also been reported for PXR. RBCK1 (HOIL-1L) was the first reported E3 ligase for PXR. Initially identified as a PXR-interacting protein from a yeast two-hybrid screen, RBCK1 was subsequently shown to modulate PXR protein level by ubiquitination that likely directed PXR to the proteasome for degradation [168]. In 2014, a small interfering RNA (siRNA)-based screen identified the kinase DYRK2 as a negative regulator of PXR activity [114]. Further analysis revealed that PXR phosphorylation by DYRK2 facilitates PXR ubiquitination by E3 ubiquitin ligase UBR5 [114]. The E3 ubiquitin ligase TRIM21 was identified as a PXR-interacting protein from an immunoprecipitation with mass spectrometry (IP-MS) experiment in the presence of rifampicin [170], and a separate group later investigated the regulation of PXR by TRIM21 in connection to CDK2 [115]. Most recently, F-box-only protein 44 (FBXO44) was identified as an E3 ligase for PXR [126]. The F-box associated domain of FBXO44 interacted with PXR LBD, leading to PXR’s ubiquitination and proteasomal degradation [126].

While PXR inhibition may be achieved through antagonists or inverse agonists, an alternative approach to decreasing PXR activity is decreasing PXR protein levels via targeted protein degradation. The first attempt to design PXR proteolysis targeting chimeras (PROTACs) used a derivative of PXR antagonist SPA70 conjugated to ligands of the E3 substrate receptor cereblon (CRBN) [171]. Although one such molecule, SJPYT-195, successfully reduced PXR protein level, further analysis revealed that SJPYT-195 acted as a molecular glue degrader for the translation termination factor G1 To S phase transition protein 1 homologue (GSPT1) and that GSPT1 degradation resulted in subsequent reduction of PXR protein [171]. The observation that derivatives of thalidomide also induced degradation of GSPT1 [172,173] suggests that developing ligands for other E3 ligases would prevent the degradation of unintended protein targets. Thus, finding and developing ligands that bind to the reported PXR E3 ligases (i.e., FBXO44, UBR5, RBCK1, and TRIM21) may be an approach for specifically inducing PXR degradation.

4. Conclusions

Since its discovery nearly three decades ago, PXR has been established as a primary regulator of the liver detoxification system. Though PXR protects against harmful levels of exogenous and endogenous substances, PXR activation may also lead to unexpected effects, including drug-drug interactions and altered energy metabolism. The large, flexible, and dynamic ligand-binding pocket of PXR that allows it to respond to many different ligands may also be key to developing agonists and antagonists with PXR specificity. While PXR inhibition by inverse agonists or antagonists may be beneficial for combating the PXR-activating effects of drug candidates, PXR activation by agonists may be desirable for treating IBD. Further research is needed to develop PXR agonists, inverse agonists, and antagonists for clinical use, and tuned activities may be the optimal approach (e.g., partial agonists rather than full agonists). Besides ligands, PTMs of PXR also affect its transcriptional activity. Considering the growing research into targeted degradation therapeutic approaches, perhaps the most exciting and clinically applicable PTMs include phosphorylation and ubiquitination, as these modifications have been shown to affect PXR protein level, which in turn affects PXR activity. It will be intriguing to see whether PXR targeted degradation becomes a reality.

5. Expert Opinion

Throughout our lives, we are exposed to countless diverse molecules, such as cosmetics, pesticides, microplastics, pharmaceuticals, dietary compounds, natural products, etc. The biological impacts are as diverse as the chemicals, and in 2007, the United States Environmental Protection Agency described the “ToxCast” program to develop methods and assays to forecast toxicity of environmental chemicals [174]. In 2010, PXR was found to be associated with significant in vivo perturbations by these chemicals [175,176]. Then, in 2021, a screen of the Tox21 10,000 compound library found a nearly 20% hit rate for PXR agonists [177]. The diversity of agonists increases when considering that mixtures of more than one chemical may activate PXR at concentrations where individual components do not [96,97]. The ligand promiscuity and variability of PXR functions are further compounded by chemical modifications carried out by PXR’s downstream targets, such as CYP3A4. Thus, chronic activation or inactivation of PXR can have sweeping physiological consequences.

Clinical PXR impacts are generally thought of in terms of drug cocktails in which one drug is a known PXR agonist. Patients of various pathologies are commonly prescribed drug cocktails for effective treatment. For example, first-line tuberculosis therapy consists of four drugs: isoniazid, rifampicin, ethambutol, and pyrazinamide [178]. A common prescription medicine for human immunodeficiency virus infections is Biktarvy, a three-drug combination [179]. The most common first-line pancreatic cancer treatment is FOLFIRINOX, a four-drug combination [180]. The simultaneous use of medicines becomes more common as people age and require treatment for various maladies. It was recently estimated that ~40% of adults >65 years old consume five or more medications at a time, with 12% consuming at least ten medications [181]. Drug-drug interactions among these regimens are common and may certainly benefit from metabolic modulators. For example, PXR activation substantially reduces circulating concentrations of oral contraceptive components and type 2 diabetes medications [182,183]. Importantly, combination of direct treatments with metabolism-altering drugs has been clinically successful. For example, the SARS-CoV-2 drug Paxlovid is a combination of a direct acting antiviral and the CYP3A inhibitor ritonavir [184]. PXR antagonists may be of great therapeutic value because they would prevent upregulation of the entire drug metabolism program rather than inhibiting a single enzyme.

Since its discovery as a ligand-dependent regulator of CYP3A4 transcription in 1998, PXR assessment has been widely incorporated into drug development programs. This has resulted in fewer PXR-mediated drug-drug interactions, with only 7 out of the 103 new molecular entities approved by the FDA between 2013 and 2016 having clinically relevant PXR-related drug-drug interactions [185]. Development of PXR antagonists has been much slower, which may be attributed to PXR’s evolutionary propensity to be activated by diverse molecules to protect the body from toxicological harm. Indeed, initial derivatization of the antagonist SPA70 generated only weaker antagonists or even agonists [102]. Though the functional impacts of subtle chemical changes have been studied [101], the mechanisms of activation vs. inhibition by chemically similar molecules are yet to be revealed. Because of the apparent subtleties involved, it may be beneficial to consider an alternative approach, such as chemically targeted protein degradation.

As with all medicinal interventions, context must be carefully considered. One example of a therapeutic dichotomy is the use of estrogen mimics for oral contraceptives but anti-estrogens for breast cancer treatment. Likewise, though PXR antagonists could be beneficial to aid in the areas of bioavailability and drug-drug interactions, agonists may be useful for treating certain diseases, such as IBD. Interestingly, although there are no approved drugs that utilize PXR as the intended target, the mechanism of the IBD drug rifaximin has been suggested to be at least partly through PXR activation. The poor oral bioavailability of rifaximin makes it a particularly useful IBD drug, as it can perform its therapeutic function in the intestine without systemic effects [186–188]. Thus, non-orally bioavailable PXR agonists may be potential solutions to treat IBD without upregulation of metabolic programs in liver and other tissues.

The mechanisms of PXR regulation are clearly much more complicated than simple ligand binding (Figure 2). In the sections above, we described specific examples of PTMs that can alter PXR activity both in the presence and absence of exogenous ligand. Ligand binding combined with PTMs, protein-protein interactions, and additional regulatory processes, such as miRNA binding, diversifies the cellular activities and functions of PXR. These functions may extend beyond transcriptional control, as non-genomic effects of PXR activation were recently reported in platelets, which lack nuclei [189]. Non-canonical signal crosstalk through PXR heterodimerization with CAR was also recently reported, which further diversifies routes of PXR regulation and function [190]. Cell-type specific functions may also be a factor for PXR, similar to the role of CAR in protecting intestinal T cells from toxic concentrations of bile acids [191]. Further study of these cellular mechanisms will give valuable insight into PXR’s complete physiological roles.

We believe that PXR modulators have broad medicinal applications. However, due to the inherent difficulty of inhibiting PXR, the first fully validated antagonist, SPA70, was not reported until 2017 (Figure 3) [110]. This antagonist has already been shown to have promising effects in multiple models, including protection from hemorrhagic shock-induced liver injury [192]. Because of the large number of clinical drugs that have known PXR-mediated drug-drug interactions, including chemotherapies (e.g., paclitaxel) [193], antivirals (e.g., efavirenz) [75], and antibiotics (e.g., rifampicin) [8,9,194], there is a high likelihood that PXR antagonists could be used to enhance bioactivity or prevent drug-drug interactions. We speculate that, in the coming years, antagonists will be functionally validated in more models and that more therapeutically attractive agonists will be developed for IBD treatments. Because of the success in mitigating PXR-related drug-drug interactions by standardizing PXR assays in drug development [185], directly targeting PXR might be the next logical step. Furthermore, increasing numbers of studies on various biological roles of PXR will identify new diseases for which PXR-directed therapies may be beneficial.

Article highlights:

• PXR plays significant roles in drug metabolism, cancer, inflammation, energy metabolism, and endobiotic detoxification.

• Inhibiting PXR activity through antagonists or inverse agonists may be beneficial in counteracting adverse drug reactions, while PXR activation may have therapeutic value in patients with inflammatory bowel disease.

• Manipulating the unique flexibility of PXR’s large ligand-binding pocket to achieve ligand specificity is key to producing the desired biological impact without undesirable nonspecific effects.

• PXR phosphorylation, acetylation, SUMOylation, and ubiquitination modify PXR activity, typically by inhibition.

• The mechanisms of PXR regulation are more complicated than simple ligand binding, as combinations of posttranslational modifications, protein-protein interactions, and additional regulatory processes diversify the cellular activities and functions of PXR.