Screening of a chemical library reveals novel PXR-activating pharmacologic compounds

Highlights

• Screening of 1120-compound library for PXR activators.

• 19 PXR activators upregulated CYP450 expression in HepG2 cells.

• Several novel PXR activators active at clinically significant concentrations.

• Significant correlation between the strength of CYP3A4 induction and selected molecular descriptors.

Abstract

The pregnane X receptor (PXR) is one of the master regulators of xenobiotic transformation. Interactions between pharmacologic compounds and PXR frequently result in drug-to-drug interactions, drug-induced hepatotoxicity, and the development of drug-resistant phenotypes in cancer cells. Potential PXR-mediated effects on drug metabolism can be predicted using high-throughput methods to detect PXR transactivation. We used the reporter cell line nhrtox-hepg2 to screen an 1120-compound library of pharmacologic substances. Using a three-stage screening process combined with a quantitative structure-activity relationships (QSAR) analysis, we detected 16 novel, previously unreported PXR activators capable of upregulating CYP450 expression. For some of these compounds such as mycophenolic acid, leflunomide, and trifluridine, the observed interactions with PXR occurred at clinically significant concentrations and could provide potential mechanistic explanations for observed drug-to-drug interactions and drug-induced toxicity. A parallel QSAR analysis revealed significant correlation between the experimentally measured PXR-dependent bioactivity and the calculated molecular descriptors of the PXR activators.

Introduction

The pregnane X receptor (PXR) gene belongs to the nuclear hormone receptor (NHR) superfamily (Blumberg et al., 1998, Kliewer et al., 1998). The PXR product, similar to several other NHRs, functions as an endobiotic- and xenobiotic-sensing transcription factor capable of regulating the metabolism and disposition of recognized compounds (Bertilsson et al., 1998, Goodwin et al., 1999, Kliewer et al., 1998, Lehmann et al., 1998, Waxman, 1999). This receptor recognizes a broad spectrum of structurally diverse compounds, including endogenous metabolites such as bile acids (Staudinger et al., 2001) and estrogens (Kliewer et al., 1998); environmental toxins such as organochlorines (Mikamo et al., 2003), phthalate esters (Hurst and Waxman, 2004), and mycotoxins (Ratajewski et al., 2011); and different pharmacologic compounds including rifampicin, dexamethasone, clotrimazole, phenobarbital, RU486, and SR 12,813 (Moore et al., 2000). PXR is recognized as one of the master regulators of liver metabolism, especially liver-based processes of xenobiotic transformation (Francis et al., 2003). Among the genes that are regulated by PXR are multiple genes that encode enzymes, such as CYP3A4 (Lehmann et al., 1998), CYP2B6 (Goodwin et al., 2001), CYP2C9 (Gerbal-Chaloin et al., 2002), CYP7A1 (Staudinger et al., 2001), UGTs (Gardner-Stephen et al., 2004), SULTs (Duanmu et al., 2002); and transporters, such as ABCB1 (Geick et al., 2001, Synold et al., 2001) and OATP2 (Staudinger et al., 2001), that are involved in xenobiotic metabolism and disposition. Among many PXR regulated genes that are involved in drug metabolism CYP3A4 is of special interest because its product is responsible for metabolism of up to 37% of the hepatically cleared drugs (Zanger et al., 2008). Interactions between pharmacologic compounds and PXR are important because recognition by PXR may significantly increase the rate of transformation of other xenobiotics in the liver, resulting in drug-to-drug interactions (Fuhr, 2000, Kliewer et al., 1998, Lehmann et al., 1998). Thus, the ability to interact with PXR is an important feature of a pharmacologic compound that significantly increases the risk of this compound’s interactions with other drugs. There are many examples of compounds capable of mediating PXR-dependent drug-to-drug interaction processes, which have been determined by in vitro studies and studies employing animal models (Luo et al., 2002, Xie and Evans, 2002). There are also published clinical observations supporting the importance of PXR-mediated drug-to-drug interactions, such as interactions between rifampicin and antivirals, including efavirenz (Lopez-Cortes et al., 2002) and nevirapine (Ribera et al., 2001), and the antihypertensive drug verapamil (Fuhr, 2000); or interactions between St. John’s wort extracts and the HIV protease inhibitor indinavir (Piscitelli et al., 2000), the immunosuppressant cyclosporin (Ruschitzka et al., 2000), and the antineoplastic agent irinotecan (Yang et al., 2010). An assessment of the overall risk of such PXR-dependent drug-to-drug interactions may be based on the observation that approximately 2% of currently registered drugs are capable of inducing the expression of the CYP3A4 (Smith, 2000) gene, which is known to be predominantly regulated by a PXR-dependent pathway. Thus, the hypothesis that PXR-mediated changes are important for clinically relevant drug-to-drug interactions (DDI) is broadly accepted and supported by experimental and clinical data. Additional clinically relevant aspects of the interactions between pharmacologic compounds and PXR that have emerged from recent reports include the PXR-mediated induction of xenobiotic-metabolizing enzymes and transporters in cancer cells, which may lead to the development of drug-resistant phenotypes (Chen et al., 2007, Mensah-Osman et al., 2007).

The interactions of numerous chemical compounds with the PXR molecule have been investigated using different experimental and theoretical models. One of the striking features of this receptor is its ability to bind structurally diverse molecules. Two model ligands frequently employed in studies of PXR function are SR 12,813 and rifampicin. These compounds, which differ significantly in molecular weight (504 and 822, respectively), have been used to obtain crystals of the ligand-binding domain of PXR (Chrencik et al., 2005, Watkins et al., 2001). Crystallographic studies have resulted in a hypothesis explaining the structural basis of the ligand promiscuity of PXR, linking this feature of PXR to the flexible structure of the PXR ligand-binding domain, which results from the presence of polypeptide loops that acquire disorderly structures upon ligand binding (Chrencik et al., 2005, Watkins et al., 2001). This mechanism of ligand-receptor interaction may explain the difficulties in developing predictive models for the relationship between the structures of molecules and their PXR-binding capacities. Although several such models have shown significant correlations between selected molecular descriptors and PXR binding or transactivation, these models alone do not suffice for screening drug candidates in the development process (Sinz et al., 2008). Therefore, the prediction of potential PXR-mediated DDI interactions requires additional experimental data, which could be provided by high-throughput methods of detecting PXR activation. Among the methods developed and proposed for implementation as a tool to predict a drug candidate’s potential to activate PXR is the detection of the direct binding of a tested molecule to a recombinant PXR protein (Jones et al., 2000, Zhu et al., 2004). A potential downside of this approach is the difficulty in distinguishing agonists from antagonists. In vitro cell-based assays detecting the transactivation of PXR using reporter gene technique provide additional information on the functional consequences of interactions of tested compounds with PXR. These methods frequently employ the transient cotransfection of a hepatocyte-like cell line with PXR-expressing and PXR-dependent reporter gene constructs (Goodwin et al., 1999, Kliewer et al., 1998, Luo et al., 2004, Sinz et al., 2006, Zhu et al., 2004). Alternatively, a cell line expressing endogenous PXR could be stably transfected with a PXR-dependent reporter gene transgene to obtain a reporter cell line. We previously used such an approach to develop the reporter cell line nhrtox-hepg2, which was employed for the screening of mycotoxins and experimental antineoplastic compounds for their capability to activate PXR (Ratajewski et al., 2011, Niemira et al., 2013). Now we have employed this reporter cell line to screen a 1120-compound chemical library consisting predominantly of pharmacologic substances. As will be shown in this report, a series of screening experiments has detected a number of PXR activators, including novel, previously unreported PXR-transactivating compounds that are capable of upregulating CYP450 expression in HepG2 cells.