The acylated phloroglucinol, hyperforin, the main active ingredient of St. John’s Wort, exerts antidepressant properties via indirect inhibition of serotonin reuptake by selectively activating the canonical transient receptor potential channel 6 (TRPC6). Hyperforin treatment can lead to drug–drug interactions due to potent activation of the nuclear receptor PXR (NR1I2), a key transcriptional regulator of genes involved in drug metabolism and transport. It was previously shown that synthetic acylated phloroglucinol derivatives activate TRPC6 with similar potency as hyperforin. However, their interaction potential with PXR remained unknown. Here we investigated five synthetic TRPC6-activating phloroglucinol derivatives and four TRPC6-nonactivating compounds compared with hyperforin and rifampicin for their potential to activate PXR in silico and in vitro. Computational PXR pharmacophore modeling did not indicate potent agonist or antagonist interactions for the TRPC6-activating derivatives, whereas one of them was suggested by docking studies to show both agonist and antagonist interactions. Hyperforin and rifampicin treatment of HepG2 cells cotransfected with human PXR expression vector and a CYP3A4 promoter-reporter construct resulted in potent PXR-dependent induction, whereas all TRPC6-activating compounds failed to show any PXR activation or to antagonize rifampicin-mediated CYP3A4 promoter induction. Hyperforin and rifampicin treatment of primary human hepatocytes resulted in highly correlated induction of PXR target genes, whereas treatment with the phloroglucinol derivatives elicited moderate gene expression changes that were only weakly correlated with those of rifampicin and hyperforin treatment. These results show that TRPC6-activating phloroglucinols do not activate PXR and should therefore be promising new candidates for further drug development.
Hyperforin has been identified as the major active compound of St. John’s Wort (SJW; Hypericum perforatum), a plant in use for decades as a self-medication to treat depression (Chatterjee et al., 1998; Müller et al., 1998). Several clinical studies show that extracts of SJW perform superior to placebo and are comparable with standard synthetic antidepressant drugs in treating mild to moderate depression (Kasper et al., 2006; Linde et al., 2008). Moreover, clinical outcome has been correlated with the hyperforin content of SJW (Laakmann et al., 1998).
Hyperforin inhibits the reuptake of serotonin and norepinephrine but does not interact directly with the serotonin transporter such as other selective serotonin reuptake inhibitors (Müller, 2003; Treiber et al., 2005). Leuner et al. (2007) showed that hyperforin specifically activates the canonical transient receptor potential channel 6 (TRPC6), leading to an increased Ca2+-influx into neurons and thereby triggering inhibition of serotonin reuptake by Ca2+-dependent signaling.
Despite a generally favorable side effect profile of SJW (Kasper et al., 2006), there is a well documented potential of SJW to induce clinically relevant drug–drug interactions (DDIs). For example, changes in plasma levels of drugs metabolized by CYP3A4, such as cyclosporine A and indinavir, occurred when patients concomitantly had taken SJW (Piscitelli et al., 2000; Ahmed et al., 2001). SJW-related DDIs were also reported for amitriptyline, irinotecan, digoxin, warfarin, and statins (Madabushi et al., 2006; Vlachojannis et al., 2011). These observations can be explained by the finding that hyperforin is a potent ligand activator of human pregnane X receptor (hPXR), a key regulator of a battery of genes involved in the absorption, distribution, metabolism, and excretion (ADME) of drugs and other xenobiotics (Moore et al., 2000; Chen et al., 2004; Bauer et al., 2006). In addition to hyperforin, PXR is activated by numerous frequently prescribed and structurally diverse drugs such as rifampicin, phenobarbital, and statins (Kliewer et al., 2002; Zhang et al., 2008; Ihunnah et al., 2011). The different available SJW formulations contain variable amounts of hyperforin (0.2%–6%) due to the different types of preparation (Klemow et al., 2011). Furthermore, altered preparation methods led to a strong increase of hyperforin content in SJW extracts in recent years (Erdelmeier et al., 2001a,b). Therapy with SJW therefore implies a considerable risk of potentially dangerous DDIs, which is difficult to predict.
A set of simple 2-acylphloroglucinol and 2,4-acylphloroglucinol derivatives, designed on the basis of the hyperforin core structure but with higher chemical stability, were recently tested for their bioactivation properties. Five of these molecules were shown to inhibit serotonin reuptake comparable with hyperforin in a TRPC6-mediated and Ca2+ flux–dependent manner (Leuner et al., 2010). In this study, we investigated the PXR interaction potential of these phloroglucinol derivatives using an array of computational PXR agonist and antagonist pharmacophores and docking approaches, as well as luciferase reporter assays and transcriptional analysis of primary human hepatocytes treated with the compounds. Our results indicate that the TRPC6-activating phloroglucinols lack high-potency interaction potential for human PXR and should therefore be promising candidates for further drug development.
Materials and Methods
Source and Preparation of Reagents.
Hyperforin [(1S,5S,7S,8R)-4-hydroxy-8-methyl-3,5,7-tris(3-methylbut-2-enyl)-8-(4-methylpent-3-enyl)-1-(2-methylpropanoyl)-bicyclo[3.3.1]non-3-en-2,9-dion]; rifampicin [5,6,9,17,19,21-hexahydroxy-23-methoxy-2,4,12,16,18,20,22-heptamethyl-8-[N-(4-methyl-1-piperazinyl)formimidoyl]-2,7-(epoxypentadeca-[1,11,13]trienimino)-naphtho[2,1-b]furan-1,11(2H)-dion-21-acetat]; sulforaphane [1-isothiocyanato-4-(methylsulfinyl)-butane]; and Hyp9 [1,1′-(2,4,6-trihydroxy-1,3-phenylene)bis-1-hexanone] were purchased from Sigma-Aldrich (Steinheim, Germany). Phloroglucinol derivatives (Fig. 1) were kindly provided by the Preclinical Research Department of Dr. Willmar Schwabe (Karlsruhe, Germany) (Leuner et al., 2010), with the exception of Hyp9. All substances were dissolved to a stock concentration of 10 mM in pure dimethylsulfoxide (DMSO) except compound Hyp3, which was dissolved in pure ethanol.
Cell Culture and Transient Transfection.
HepG2 cells obtained from American Type Culture Collection (Manassas, VA) were cultured in Dulbecco’s modified Eagle’s medium (Invitrogen, Darmstadt, Germany) supplemented with 100 U/ml 10% fetal calf serum (PAA Laboratories GmbH, Cölbe, Germany), 100 µg/ml penicillin/streptomycin (Invitrogen), and 1 mM sodium pyruvate (Invitrogen) under a 5% CO2 atmosphere at 37°C. Transient transfection was performed reversely in 96-well plates with 20,000 cells/well. Cells were transfected with a cocktail containing 20 µl Dulbecco’s modified Eagle’s medium, 0.4 µl Turbofect (Fermentas Life Science, St. Leon-Rot, Germany) and 2.5 ng of pGL3-CMV-Renilla (Promega, Mannheim, Germany), 80 ng of pGL3-CYP3A4(-7830/Δ7208–364) (Istrate et al., 2010) or pGL3-CYP3A4(-56), 10 ng of pcDhPXR plasmid expressing human PXR or empty pcDNA3 vector. The pGL3-CYP3A4(-56) vector was generated by digesting pGL3-CYP3A4(-1105) (Hustert et al., 2001) with KpnI/PstI, generating blunt ends with T4 DNA polymerase and religation of the isolated 4.9-kb fragment. pUC18 was added to a total amount of 200 ng of DNA/well. Cells were treated 6 hours after transfection with DMSO (0.5%) or ethanol (0.5%) as vehicle controls or with phloroglucinol derivatives, hyperforin, and/or rifampicin at concentrations ranging from 0.001 to 50 µM. Cells were lysed using passive lysis buffer (Promega, Madison, WI) 48 hours after transfection and 42 hours after treatment, respectively. Luciferase activities were determined using beetle juice and Renilla juice (P.J.K. GmbH, Kleinblittersdorf, Germany).
Primary Human Hepatocytes.
The use of human hepatocytes for research was approved by the local ethics committees of Berlin and Regensburg, and written informed consent was obtained from all patients. Hepatocytes from one female and two male donors were isolated and cultured essentially as described by Thomas et al. (2013). Cells were treated for 24 hours with 10 µM rifampicin, 1 or 5 µM hyperforin, Hyp1, Hyp5, Hyp7, Hyp8, Hyp9, or Hyp10 and 50 µM Hyp2, Hyp3, Hyp4, Hyp6, or the vehicles 0.5% DMSO or 0.5% ethanol for 24 hours.
Quantitative Real-Time Polymerase Chain Reaction Analysis.
Total RNA was isolated from primary human hepatocytes using the RNeasy Mini Kit. Genomic DNA digestion was performed using the RNase free DNase Set (Qiagen, Hilden, Germany). RNA was reverse transcribed to cDNA with TaqMan Reverse Transcription Reagents (Applera GmbH, Darmstadt, Germany). For quantitative real-time polymerase chain reaction, we used the Fluidigm’s Biomark high-throughput quantitative polymerase chain reaction chip platform (Fluidigm Corporation, San Francisco, CA) with 34 predesigned gene expression assays from Applied Biosystems (Supplemental Table 1) according to the manufacturers’ instructions (Spurgeon et al., 2008). The mRNA expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA expression.
Pharmacophore models for PXR agonists have been described previously (Ekins and Erickson, 2002; Bachmann et al., 2004; Ekins et al., 2007) and their use for predicting molecules has also been described (Yasuda et al., 2008). In addition, a pharmacophore for PXR antagonists was used as previously detailed (Ekins et al., 2007, 2008a). We used these previously published models in Discovery Studio 3.5 (Accelrys, San Diego, CA) to predict the phloroglucinol derivatives. A small three-dimensional conformer database was generated with the FAST conformer generation method and up to 255 conformers were generated. Molecules that fit all features were identified and pharmacophore fit scores were tabulated. Docking studies were performed with Discovery Studio using LibDock with FAST Search, FAST conformations, and steepest decent minimization with CharmM. The structure for agonists used a docking sphere of a 8.4-Å radius around hyperforin in complex in the PXR ligand binding domain (LBD) [Protein Data Bank (PDB) code 1M13]. The PXR steroid receptor coactivator-1 (SRC-1) site was used to evaluate potential for antagonists. This used the PXR protein PDB code 1NRL chain A for rigid docking as described previously (Ekins et al., 2007, 2008a) with the following modifications. The protein was prepared using the “prepare protein” protocol in Discovery Studio. The LibDock program was then used for docking (Rao et al., 2007). The binding site was defined as x = 3.527, y = 12.353, and z = 26.81 and a binding site sphere of 10 Å in diameter was created. A newly proposed antagonist site (Li et al., 2013) was also evaluated. A binding site sphere of 14 Å in diameter was created using S208 as the center and then the phloroglucinol derivatives were docked into the 1NRL structure using LibDock with FAST conformation generation, an energy threshold of 20 kcal/mol, steepest descent minimization, and the CHARMm force-field. Ligand conformations were then manually assessed.
Differences in gene expression and promoter activity between treatments were analyzed using repeated measurement one-way analysis of variance (ANOVA). Those with Bonferroni-adjusted ANOVA P values < 0.05 were further examined by Dunnett’s multiple comparisons test, only comparing the conditions versus the respective control. Results from cotreatment experiments were analyzed using two-way ANOVA adjusted for multiple testing (Bonferroni). Those with a Bonferroni-adjusted ANOVA P value < 0.05 were further examined by the paired t test (also Bonferroni corrected) comparing the conditions versus the respective control. Statistical analyses and nonlinear curve fitting (variable slope, four parameters) were performed using GraphPad Prism 5.04 software (GraphPad Software, Inc., La Jolla, CA).
To explore the potential of agonistic and antagonistic interactions of the phloroglucinol derivatives with PXR, we used an array of computational methods including three agonist pharmacophores and an antagonist pharmacophore (Table 1). With the Bristol-Myers Squibb pharmacophore (Ekins et al., 2007), Hyp3 was suggested to have a higher fit score than hyperforin. The original PXR pharmacophore (Ekins and Erickson, 2002) also indicated that Hyp3 had a similar fit score to hyperforin. Hyp3 is also the most lipophilic of all of the molecules, as indicated by its highest AlogP value (Table 1). Hyp7 was the only molecule mapping to the PXR antagonist pharmacophore (Table 1). We also used docking into the PXR LBD (Table 2). This showed that all molecules had docking scores lower than hyperforin (135.85), the closest being Hyp7 (129.28) suggesting that few were likely to be agonists. Docking into the two separate proposed antagonist sites (Li et al., 2013) (Table 3) suggested that Hyp3 might be a candidate for binding the SRC-1 coactivator site with a score similar to ketoconazole. Hyp3 and Hyp7 were predicted to bind the S208 site with docking scores greater than for ketoconazole (129.4), indicating potential for inhibitory interactions. In summary, ligand-based pharmacophores and structure-based methods for identifying PXR agonists and antagonists suggest that most of the phloroglucinol derivatives should neither interact as agonists (using hyperforin as a comparator) nor as antagonists (using ketoconazole as a comparator). With respect to the interactions of Hyp3 and Hyp7, the computational predictions pointed to both agonist and antagonist potential.
Effects of Phloroglucinol Derivatives on PXR-Mediated CYP3A4 Promoter Activity.
To investigate the potential of the phloroglucinol derivatives (Fig. 1) to activate PXR, HepG2 cells were cotransfected with a CYP3A4/xenobiotic responsive enhancer module promoter-based luciferase reporter system and hPXR cDNA expression plasmid and treated with the substances or with hyperforin or rifampicin as a positive control. EC50 values were determined in a range from 0.001 to 50 µM (Fig. 2, A and B). It should be noted that TRPC6-activating phloroglucinols including hyperforin showed cytotoxic effects above 5 µM, as previously reported for hyperforin treatment of human hepatocytes and CV-1 cells (Moore et al., 2000; Komoroski et al., 2004). Hyperforin (Fig. 2A) and rifampicin (Fig. 2B) showed dose-dependent activation of the promoter, with maximal induction of 19.8- and 11.7-fold, respectively. Surprisingly, only the TRPC6-nonactivating compound Hyp4 showed a dose-dependent activation of the promoter, with the highest induction of 5.8-fold observed at 50 µM (Fig. 2B). Nonlinear curve fitting revealed EC50 values of 0.59 and 1.9 µM for hyperforin and rifampicin, respectively. All other phloroglucinol derivatives did not activate the promoter in a dose-dependent manner (Fig. 2, A and B).
We next evaluated whether the activation of the CYP3A4 promoter by hyperforin, rifampicin, or Hyp4 was PXR dependent. To this end, HepG2 cells cotransfected with the luciferase reporter system and pcDhPXR or pcDNA3 (empty vector) were treated with hyperforin, rifampicin, Hyp4, or DMSO (Fig. 2C). Hyperforin (1 µM), rifampicin (10 µM), and Hyp4 (50 µM) showed significant induction of the CYP3A4 promoter only in the presence of hPXR. The weak remaining induction of the promoter in the absence of hPXR is most likely due to endogenous PXR.
Investigation of Antagonistic Properties of the Phloroglucinol Derivatives.
Because antagonist properties have been described for some PXR ligands (Ekins et al., 2007, 2008a), we investigated whether the phloroglucinols could compete or antagonize rifampicin-mediated PXR activation at the CYP3A4 promoter. We treated HepG2 cells transfected with the luciferase reporter system with rifampicin (10 µM) in combination with the different phloroglucinol derivatives or sulforaphane, a known PXR antagonist (Fig. 3). As expected, sulforaphane showed dose-dependent reductions in promoter activity of 24 and 55% at 5 and 10 µM, respectively (Fig. 3). No reduction of the rifampicin-induced promoter activity was found for any of the TRPC6-activating or TRPC6-nonactivating phloroglucinol derivatives (Fig. 3, A and B). Cotreatment with hyperforin (0.5 and 1 µM), Hyp8 (0.5 µM) (Fig. 3A), or Hyp6 (50 µM) (Fig. 3B) resulted in a significant activation of the promoter.
Effects of Phloroglucinol Derivatives on the Expression of ADME Genes in Primary Human Hepatocytes.
To assess whether the phloroglucinol derivatives have other, potentially PXR-unrelated effects on gene expression in human liver, we analyzed the mRNA expression of a set of 33 ADME genes (Supplemental Table 1), including the PXR target genes CYP3A4, ABCB1, and UGT1A1, in primary human hepatocytes from three individual donors. Spearman correlation analysis was performed to compare mRNA expression changes obtained by treatment with rifampicin and those caused by the different phloroglucinol derivatives (Fig. 4). Gene expression changes upon treatment with 1 µM hyperforin were highly correlated (rs = 0.96; P < 0.0001) with those of rifampicin (10 µM). Treatment of hepatocytes with 5 µM hyperforin led to a weaker correlation (rs = 0.63; P < 0.0001) with the rifampicin profile, which may be explained by cytotoxic or other less selective effects of hyperforin at higher concentrations. The correlations of the rifampicin expression profile with all other phloroglucinol derivatives in the different concentrations used were much weaker (rs values ≤ 0.5), with the exception of Hyp4 (50 µM; rs = 0.73) (Fig. 4).
In particular, rifampicin and hyperforin (1 µM) both led to a significant and comparable induction of CYP3A4 of 24- and 16-fold, respectively, whereas all other phloroglucinols did not affect CYP3A4 expression in the three donors tested (Fig. 5; Supplemental Tables 2 and 3). Treatment with 5 µM hyperforin led to a 5-fold weaker induction of CYP3A4 expression compared with treatment with 1 µM hyperforin, probably indicating onset of toxicity (Fig. 5A).
The PXR target genes CYP2B6, ABCB1 (MDR1), UGT1A1, CYP2C9, CYP3A5, and ALAS1 showed significant induction by hyperforin (2.9-, 2.4-, 2.9-, 3.9-, 5.3-, 5.8-, and 2.7-fold, respectively) and by rifampicin (3.8-, 2-, 2.9-, 4.1-, 5.3-, 5.8-, and 2.7-fold, respectively) (Fig. 4; Supplemental Tables 2 and 3). In contrast, CYP7A1 mRNA expression was significantly downregulated by rifampicin (6.3-fold) and by hyperforin (6.7-fold) (Fig. 4; Supplemental Figs. 1 and 2; Supplemental Table 2). Hyperforin also significantly upregulated POR expression (1.6-fold) (Fig. 4; Supplemental Table 2).
Treatment with the TRPC6-activating phloroglucinol derivatives (Hyp1, Hyp5, Hyp7, Hyp8, and Hyp9) did not significantly change the expression of CYP2B6, CYP7A1, CYP1A1, CYP1A2, CYP2C8, ABCB1, or UGT1A1, whereas CYP2C9 and CYP2B6 were significantly induced 1.9- and 1.7-fold by treatment with 1 µM Hyp7 (Fig. 4; Supplemental Table 2), respectively. Hyp7 (1 µM) also significantly induced CYP2E1 expression. Hyp9 (5 µM) was found to significantly induce delta-aminolevulinate synthase 1 expression. For the TRPC6-noninducing phloroglucinols, the only significant expression change observed was 2.7- and 2-fold induction of CYP1A2 by 10 µM Hyp2 and Hyp4, respectively (Fig. 4; Supplemental Table 3).
TRPC6 and PXR are structurally unrelated proteins with highly distinct physiologic functions. The fact that hyperforin, a potent PXR activator, is also a TRPC6 activator, is thus very surprising and may be a coincidence rather than biologically meaningful. Regarding their structural heterogeneity, it should be possible to separate the activator functions of TRPC6 from those of PXR and develop ligands that activate TRPC6 but not PXR. The current study used both computational and in vitro approaches to indicate that the TRPC6-activating phloroglucinol derivatives are unlikely to have interactions with PXR.
Ligand-based pharmacophores use known information on agonists and antagonists to identify key features for interactions. Structure-based methods such as docking enable one to determine whether molecules will fit and have favorable interactions in crystal structures and homology models. Both of these approaches have been widely used for identifying PXR agonists and antagonists (Ekins et al., 2007, 2008a, 2009; Yasuda et al., 2008; Biswas et al., 2009; Kortagere et al., 2009, 2012; Li et al., 2013) for which crystal structures exist (Watkins et al., 2001, 2002, 2003a,b; Chrencik et al., 2005; Noble et al., 2006; Xue et al., 2007a,b; Teotico et al., 2008). The phloroglucinol derivatives appear structurally distinct from hyperforin (Fig. 1) and the physicochemical parameters would also be expected to differ, which would suggest that their protein interactions would also likely differ. For example, the lipophilicity parameter AlogP of hyperforin is considerably higher compared with the majority of phloroglucinol derivatives, with only the TRPC6-nonactivating Hyp3 being higher (Table 1). This could also explain why Hyp3 was frequently retrieved by both pharmacophores and docking. It is widely known from our previous work and the many crystal structures (Watkins et al., 2001, 2002, 2003a,b; Chrencik et al., 2005; Noble et al., 2006; Xue et al., 2007a,b; Teotico et al., 2008) that hydrophobicity is important for interaction in the LBD and at the SRC-1 antagonist site (Ekins et al., 2007). The majority of the phloroglucinol derivatives were found to have docking scores lower than the comparator compounds hyperforin and ketoconazole, suggesting that they were unlikely to behave as agonists or antagonists, respectively. The pharmacophores retrieved few of the phloroglucinols also, suggesting that they were in general less likely to interact with PXR.
The in vitro studies showed that only high concentrations of Hyp4 could induce CYP3A4 (Fig. 2B) and that this only occurred in the presence of hPXR (Fig. 2C). We hypothesized the lack of agonist activity was due to potential antagonist activity of the phloroglucinol derivatives, yet no reduction of the rifampicin-induced promoter activity was found for any of the phloroglucinol derivatives (Fig. 3). Instead, some of the molecules including hyperforin showed increased promotor activity. The reason for this is currently unclear. It is difficult to rationalize how two very large molecules such as rifampicin and hyperforin could bind the LBD together. It may be the result of an allosteric mechanism that requires further investigation.
Treatment of human hepatocytes with hyperforin confirmed induction of previously described PXR target genes CYP2B6, CYP3A4, CYP2C9, and ABCB1 (Fig. 4) (Moore et al., 2000; Goodwin et al., 2001; Chen et al., 2004; Haslam et al., 2008). In addition, we observed induction of CYP3A5, ALAS1, POR, and UGT1A1, which to our knowledge had not been previously reported to be induced by hyperforin. The high correlation of expression changes of a broad set (n = 33) of ADME genes after hyperforin (1 µM) or rifampicin treatment of human hepatocytes (rs = 0.96) is in agreement with the assumption that both substances induce gene expression only via PXR activation (Fig. 4). Although this finding may not be surprising, it has not been reported before and it helps to further specify the DDI potential of hyperforin. Treatment with the phloroglucinol derivatives also led to expression changes of the investigated ADME genes, which were, however, more modest compared with hyperforin and appeared to be PXR independent because most PXR target genes were not affected, except for an approximately 2-fold induction of CYP2C9 and CYP2B6 by Hyp7 (1 µM). This is further supported by the weak correlations with the changes caused by rifampicin treatment (Fig. 4). Whether other ligand-dependent nuclear receptors such as constitutive androstane receptor, glucocorticoid receptor, farnesoid X receptor, liver X receptor, or vitamin D receptor are involved in this response appears unlikely because the gene expression changes did not appear to match their known gene target profiles (Zanger and Schwab, 2013).
Only Hyp4 showed a moderate correlation (rs = 0.73) with the rifampicin expression profile at higher concentration (50 µM), although none of the ADME genes, except CYP1A2, was significantly regulated by this compound (Fig. 4). In contrast to the results obtained from the reporter assays, Hyp4 did not induce CYP3A4 expression in the primary human hepatocytes (Figs. 4 and 5B). Given the high concentrations needed to activate the CYP3A4 promoter in HepG2 cells, it is conceivable that these concentrations were not reached in hepatocytes (e.g., due to differences in transporter function in HepG2 compared with hepatocytes or due to metabolic degradation).
In conclusion, our results show that a set of TRPC6-activating phloroglucinols, which are capable of activating the TRPC6 channel and therefore likely possess antidepressant properties, are clearly unable to activate or antagonize PXR. These phloroglucinols thus represent promising new candidates for further drug development as antidepressants with improved safety because they lack the DDI potential of hyperforin and SJW. Our results also strongly indicate that the pharmacophore for TRPC6 activation differs from that required for PXR agonism or antagonism. Nevertheless, further optimization of these compounds should pay close attention to the PXR pharmacophores as well as other ADME/toxicity properties (Ekins et al., 2010; Williams et al., 2012) that may limit their viability as drugs.
The authors gratefully acknowledge Oliver Burk for providing plasmids, Stefan Winter for statistical advice, and Igor Liebermann for expert technical assistance (all from Stuttgart, Germany). The authors also thank Accelrys, Inc. for providing Discovery Studio. The authors especially thank Maria Hauner and Barbara Donabauer (Munich), Theresa Schulz and Georg Damm (Berlin), and Thomas Weiss and Susanne Heyn (Regensburg) for cell isolation and culture and contribution of human hepatocytes. Furthermore, the authors are indebted to the Charitable Human Tissue and Cell Research Foundation (Regensburg) for making human tissue available for research.
Participated in research design: Kandel, Ekins, Leuner, Harteneck, Zanger.
Conducted experiments: Kandel, Ekins, Thasler.
Contributed new reagents or analytic tools: Leuner, Thasler.
Performed data analysis: Kandel, Ekins.
Wrote or contributed to the writing of the manuscript: Kandel, Ekins, Zanger.
- Received September 23, 2013.
- Accepted November 18, 2013.
This study was supported by the German Federal Ministry of Education and Research Virtual Liver Network [Grant 0315755]; and the Robert Bosch Foundation.
- absorption, distribution, metabolism, excretion
- analysis of variance
- drug–drug interaction
- heme oxygenase
- human pregnane X receptor
- ligand binding domain
- multidrug-resistance protein
- cytochrome P450
- Protein Data Bank
- pregnane X receptor
- St. John’s Wort
- steroid receptor coactivator
- thiopurine S-methyltransferase channel
- UDP glucuronosyltransferase
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics