Dose-dependent Induction of Cytochrome P450 (CYP) 3A4 and Activation of Pregnane X Receptor by Topiramate


Address correspondence and reprint requests to Dr. P. B. Desai at College of Pharmacy, University of Cincinnati Medical Center, 3223 Eden Avenue, Mail Location #0004, Cincinnati, OH 45267, U.S.A. E-mail:


Summary: Purpose: In clinical studies, topiramate (TPM) was shown to cause a dose-dependent increase in the clearance of ethinyl estradiol. We hypothesized that this interaction results from induction of hepatic cytochrome P450 (CYP) 3A4 by TPM. Accordingly, we investigated whether TPM induces CYP3A4 in primary human hepatocytes and activates the human pregnane X receptor (hPXR), a nuclear receptor that serves as a regulator of CYP3A4 transcription.

Methods: Human hepatocytes were treated for 72 h with TPM (10, 25, 50, 100, 250, and 500 μM) and known inducers, phenobarbital (PB; 2 mM), and rifampicin (10 μM). The rate of testosterone 6β-hydroxylation by hepatocytes served as a marker for CYP3A4 activity. The CYP3A4-specific protein and mRNA levels were determined by using Western and Northern blot analyses, respectively. The hPXR activation was assessed with cell-based reporter gene assay.

Results: Compared with controls, TPM (50–500 μM)–treated hepatocytes exhibited a considerable increase in the CYP3A4 activity (1. 6- to 8.2-fold), protein levels (4.6- to 17.3-fold), and mRNA levels (1.9- to 13.3-fold). Comparatively, rifampicin (10 μM) effected 14.5-, 25.3-, and a 20.3-fold increase in CYP3A4 activity, immunoreactive protein levels, and mRNA levels, respectively. TPM (50–500 μM) caused 1.3- to 3-fold activation of the hPXR, whereas rifampicin (10 μM) caused a 6-fold activation.

Conclusions: The observed induction of CYP3A4 by TPM, especially at the higher concentrations, provides a potential mechanistic explanation of the reported increase in the ethinyl estradiol clearance by TPM. It also is suggestive of other potential interactions when high-dose TPM therapy is used.

Topiramate [TPM; Topamax; 2, 3:4, 5-bis-0-(1-methylethylidene)-β-d-fructopyranose sulfamate] is a sulfamate-substituted monosaccharide derived from d-fructose and is structurally unrelated to other antiepileptic drugs (AEDs). Compared with many other AEDs, TPM appears to have several distinct advantages. First, it has anticonvulsant activity against a broad spectrum of seizure types and a good safety profile. Second, it exhibits linear disposition over the dosing range used clinically. Third, TPM appears to have few drug–drug interactions (1).

However, one of these drug–drug interactions occurs with oral contraceptives. In a cohort of 12 women with epilepsy receiving stable dosages of valproic acid (VPA) along with a combination norethindrone, 1 mg/ethinyl estradiol, 35-μg tablet, TPM administered at doses of 200 mg/day, 400 mg/day, and 800 mg/day caused a statistically significant dose-related increase in the mean oral serum clearance (CL/F was 14.7–33% higher) of ethinyl estradiol (2). The mean Cmax value for ethinyl estradiol decreased by 15–25.4%, and the area under the curve (AUC0-24) decreased by 18–30% at the 200- to 800-mg/day dose level. The pharmacokinetics of norethindrone remained unchanged. In contrast, a recent study found that TPM at doses of 50–200 mg/day did not significantly affect the clearance of either ethinyl estradiol or norethindrone (3). Thus it appears that the effect of TPM on ethinyl estradiol level may be related to the TPM dose. Because estradiol is a known substrate of cytochrome P450 (CYP) 3A4 isozyme (4), we hypothesized that TPM enhances the activity and expression of CYP3A4 in a dose-proportional manner.

CYP enzymes constitute a superfamily of hemoproteins that catalyze mainly monooxygenation reactions of a wide variety of substrates. As such, these enzymes play a vital role in detoxification and systemic clearance of xenobiotics. Of >55 human CYP isozymes presently known, CYP3A4 is considered to be the drug-metabolizing enzyme of principal importance (4). It is the predominant CYP enzyme present in the liver (∼40%) and intestine, and it participates in the metabolism of >60% of all marketed drugs that are eliminated by enzyme-catalyzed processes (4). As such, modulation of the CYP3A4 activity is known to be a major cause of drug–drug interactions. Recent studies significantly advanced our understanding of the processes that regulate the expression of CYP3A4. Accordingly, an orphan nuclear receptor termed the human pregnane X receptor (hPXR), also known as the steroid xenobiotic receptor (SXR), appears to be the key regulator of the CYP3A4 transcription (5,6).

This study had two goals: (a) to examine the effect of graded TPM concentrations on the expression and the activity of CYP3A4 in primary cultures of human hepatocytes; and (b) to investigate the extent to which TPM activates hPXR in a cell-based reporter assay.


Cell culture, chemicals, and reagents

TPM was a gift from the Robert Wood Johnson Pharmaceutical Research Institute. Rifampicin, phenobarbital (PB), testosterone, and 6β-hydroxytestosterone were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). African green monkey kidney fibroblasts (CV-1 cells) were obtained from American Type Culture Collection (Manassas, VA, U.S.A.). Cell-culture media and supplements were purchased from Cell Gro (Herndon, VA, U.S.A.). The plasmids for human PXR (pSG5-hPXRΔATG) and (CYP3A1-DR3)4-tk-CAT chloramphenicol acetyl transferase (CAT) reporter were kindly provided by Dr. Bryan Goodwin, GlaxoSmithKline Inc. (RTP, NC, U.S.A.) (5).

Hepatocyte culture and drug treatment

Primary human hepatocytes, isolated from lobes of liver from four separate donors, were provided by the Liver Tissue Procurement and Distribution System (Pittsburgh, PA, U.S.A.), which was funded by NIH contract N01-DK-9-2310. The use of hepatocytes at our institution was permitted by the University of Cincinnati Institutional Review Board under the “exempt” category, because the identity of the donor is not revealed to us. Hepatocytes were maintained in Williams E medium (BioWhittaker, Walkersville, MD, U.S.A.), as described previously (7). For the determination of CYP3A4 activity and immunoreactive protein levels, hepatocytes were plated in collagen-coated six-well plates (1 × 106 cells/well). In parallel, cells were plated in collagen-coated T-25 cm2 flasks for Northern blot analysis of CYP3A4-specific mRNA, and cells plated in collagen-coated 24-well format (1.25 × 105 cells/well) were used for MTT assay. Drug solutions (×1,000) were prepared in dimethylsulfoxide (DMSO) and diluted before use. Forty-eight hours after isolation and plating, hepatocytes were treated with vehicle, which contained the same amount of DMSO (0.1%), TPM (10, 25, 50, 100, 250, or 500 μM), rifampicin (10 μM), or PB (2 mM) for 72 hrs. Drug-containing medium was replaced every 24 h. At the end of the drug-treatment period, the cells were incubated for 30 min in drug-free medium to facilitate drug elimination. Cells were then washed with buffer and exposed to testosterone-containing medium for assessment of testosterone 6β-hydroxylase activity of intact cells. The media was then collected for high-performance liquid chromatography (HPLC) analysis and the cells processed for protein isolation. Previously reported studies by Kostrubsky et al. (8,9) and from our laboratory (7,10) have shown that DMSO at levels twice as high as those used here do not alter the expression of CYP3A enzymes. Cell viability was assessed at the time of receiving the cells and daily during the course of drug exposure by using the MTT, assay as described earlier (11).

Measurement of CYP3A4 activity

The rate of testosterone conversion to 6β-hydroxytestosterone, a reaction catalyzed by CYP3A4, by intact hepatocytes was used to assess the enzyme activity, as described previously (8,9). At the end of the 72-h drug-treatment period, the drug-containing medium was removed, and the cells were incubated in the drug-free medium for a period of 30 min to facilitate the removal of drug. Cells were then incubated with testosterone containing (250 μM) Williams E medium (1 ml/plate) and incubated for 30 min. Next, 11 α-hydroxyprogesterone (1 μg/10 μl), used as an internal standard for HPLC quantitation, was added to the medium. The medium was then immediately extracted with 3 ml of dichloromethane, and the solvent was evaporated under nitrogen. The samples were reconstituted in 60:40 vol/vol methanol/water and analyzed for 6β-hydroxytestosterone levels with a published HPLC method (12), which was previously used in our studies (7). In brief, Waters 510 pumps were used to elute 60:40 methanol/water mobile phase at 1 ml/min through a C18μbondapak column (3.9 × 30 mm). A Waters 486 UV/VIS detector at a wavelength of 242 nm was used for the detection of 6β-hydroxytestosterone. The total run lasted 20 min, with 6β-hydroxytestosterone, 11α-hydroxyprogesterone (internal standard), and testosterone eluting at 7.5, 11.1, and 18.2 min, respectively. The interday and intraday variability in the HPLC analysis was <5%, and the detection limit was 100 pmol for 6β-hydroxytestosterone.

Immunodetection of CYP3A4 protein

Western blot analysis was performed to compare the CYP3A4 immunoreactive protein levels in control and drug-treated hepatocytes, as described previously (7). Control- and drug-treated hepatocytes were processed for the preparation of S9 fractions. Cells were scraped from the wells and resuspended in 0.1 M potassium phosphate buffer (pH 7.6) and homogenized by using a tissue homogenizer. Homogenates were then centrifuged at 9,000 g for 20 min (10). The S9 fractions (5 μg) were resolved with sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE; 12% acrylamide) and transferred to nitrocellulose membranes. The membranes were then blocked with 3% bovine serum albumin in phosphate-buffered saline supplemented with Tween 20 (0.1 M, pH 7.4, 0.1% Tween 20) for 45 min and then treated with primary anti-CYP3A4 antibody (Gentest, Woburn, MA, U.S.A.) followed by horseradish peroxidase–conjugated anti-mouse secondary antibody. The protein bands were visualized by using enhanced chemiluminescence detection (ECL, Amersham) and quantitated by photodensitometry.

Northern blot analysis of CYP3A4 mRNA

Total cellular RNA from control and drug-treated hepatocytes was isolated by using TRIzol (GIBCO BRL, Rockville, MD, U.S.A.). A 10-μg aliquot was fractionated by electrophoresis in 1% agarose gels containing formaldehyde (2.2 M) and transferred onto a nylon membrane (Millipore, Bedford, MA, U.S.A.). Equal loading per lane was verified by ethidium bromide staining of 18S and 28S rRNA, which was visualized and photographed under UV illumination. The membranes were hybridized as described previously (7,13) with CYP3A4 cDNA probe (780 base pairs, Oxford Biomed. Res., Inc.) labeled with [α32P]-dCTP (NEN, Boston, MA, U.S.A.) by using the random primer method.

Human PXR activation assays

Transient transfection assays in CV-1 cells were performed as described previously (5). In brief, CV-1 cells were plated in 24-well plates in DME medium supplemented with delipidated fetal calf serum at a density of 1.2 × 105 cells per well. After 24 h of plating the cells, overnight transfections were performed by using Lipofectamine and Plus reagent (Invitrogen), exactly as suggested by the manufacturer in a special protocol for CV-1 cells. Transfection mixes contained 100 ng human PXR expression vector (pSG5-humanPXR), 400 ng CAT reporter gene construct, 400 ng pCH110 (an expression vector containing β-galactosidase cDNA under T7 promoter; Amersham). Control transfections with 100 ng pSG5, 400 ng CAT reporter gene construct, 400 ng pCH110 also were maintained and treated with drugs. After transfection, plasmid-containing media were replaced with drug-containing media (triplicates) and incubated for 24 h. The cell layers were washed twice with ice-cold phosphate-buffered saline (pH 7.4) and scraped and collected in 250 μl reporter lysis buffer provided with the CAT ELISA kit (Roche Diagnostics, Indianapolis, IN, U.S.A.). Protein assay was performed on an aliquot of the cell lysate, and 175 μl of the cell lysate was used for determining the CAT enzyme levels by using a CAT ELISA kit (Roche Molecular Diagnostics). An aliquot of cell lysate (50 μl) was used for determining the β-galactosidase activity, exactly as described by the protocol obtained from Promega (Madison, WI, U.S.A.). The CAT protein levels were normalized to the β-galactosidase activity and expressed as fold-activation with respect to the solvent (0.1% DMSO)-treated controls. In additional samples, at the end of drug treatment, cell viability also was determined by MTT assay, performed exactly as described earlier (11).

Statistical and data analysis

The difference in the fold increase in CYP3A4 activity, immunoreactive protein content and CYP3A4 mRNA levels, fold activation of hPXR, and cell viability assessed by using the MTT assay between control and treatment groups was analyzed by using a one-factor analysis of variance, followed by a Tukey–Kramer test to compare the mean values at an α= 0.05.


We investigated the effect of graded TPM concentrations (10, 25, 50, 100, 250, and 500 μM) on the CYP3A4 activity and expression in primary cultures of human hepatocytes after 72-hour exposure. Rifampicin and PB, known inducers of CYP3A4, were used as positive controls. Four batches of hepatocytes were used. The information pertaining to the age, gender, ethnicity, and the history of medication of the donors is presented in Table 1. Cell-viability assessment with the MTT test indicated that the viability of hepatocytes treated with TPM (10–500 μM), PB (2 mM), and rifampicin (10 μM) was comparable to vehicle-treated controls.

Table 1. Human liver donor information
BatchAge (yr)SexRaceCause of deathDrugs used in ER/OR
  1. C, Caucasian; ER/OR, Emergency room/Operating room.

HH95222FCMotor vehicle accidentDopamine
HH95462MCIntracranial bleeding
HH97738FCIntracranial bleedingDexamethasone, phenytoin
HH101274FCAortic aneurysmDopamine

The rate of testosterone conversion to 6β-hydroxytestosterone, a reaction catalyzed by CYP3A4, by intact hepatocytes was used to assess CYP3A4 enzyme activity. The net increase in CYP3A4 activity after treatment with TPM, compared with untreated human hepatocytes, is presented in Table 2. As shown, TPM caused a concentration-dependent increase in CYP3A4 activity. Whereas the increase in the 6β-hydroxylation of testosterone was deemed to be statistically insignificant at TPM 10- and 25-μM concentrations (p > 0.05), the increase in CYP3A4 activity (1.6- to 8.2-fold) in the concentration range of 50–500 μM was statistically significant (p < 0.05).

Table 2. Fold-increase (compared with untreated control) in the CYP3A4 activity, immunoreactive protein, and mRNA levels in human hepatocytes after 72-h treatment with various drugs
 Rifampicinb 10 μMPhenobarbitalb 2 mMTopiramate
10 μM25 μM50 μM100 μM250 μM500 μM
  1. Values expressed as mean ± SEM; n = 4.

  2. aTestosterone 6β-hydroxylation assay was used as a marker for CYP3A4 activity in human hepatocytes.

  3. bRifampicin (10 μM) and phenobarbital (2 mM) were used as positive controls for CYP3A4 induction.

  4. cp < 0.05 compared with untreated controls.

CYP3A4 activitya14.5 ± 3c11.9 ± 4c1.5 ± 0.31.4 ± 0.31.6 ± 0.33.6 ± 1.7c5.1 ± 2.5c  8.2 ± 6.2c
CYP3A4 protein levels  25.3 ± 10c  17.9 ± 12c0.9 ± 0.12.3 ± 0.54.6 ± 2c7.8 ± 3.2c  11 ± 2.3c17.3 ± 5c  
CYP3A4 mRNA levels20.3 ± 4c    11.7 ± 3.2c1.2 ± 0.21.7 ± 0.31.9 ± 0.34.3 ± 1.7c9.4 ± 6.2c13.3 ± 0.3c

A typical Western blot of TPM-treated hepatocytes is represented in Fig. 1. The pattern of increase in the immunoreactive CYP3A4 levels mirrored the increase in the testosterone 6β-hydroxylation activity, with marked increase in the levels observed at higher concentrations used. At concentrations ranging between 50 and 500 μM, the CYP3A4-specific immunoreactive protein levels increased 4.6- to 17.3-fold. Photodensitometric analysis of the CYP3A4 immunoreactive protein bands observed in drug-treated hepatocytes in comparison with untreated controls is shown in Table 2.

Figure 1.

Representative Western blots depicting immunoreactive CYP3A4 levels in human hepatocytes treated with topiramate (TPM). A: CYP3A4 protein levels in human hepatocytes treated with phenobarbital (PB; 2 mM), rifampicin (Rif, 10 μM), and TPM (10–250 μM) compared with untreated control hepatocytes (C). B: CYP3A4 protein levels in human hepatocytes treated with PB (2 mM), Rif (10 μM), and TPM (100–500 μM) compared with untreated control hepatocytes (C).

The fold-increase in CYP3A4 mRNA levels after various concentrations is presented in Table 2. The Northern blots shown in Fig. 2 represent the observed expression of CYP3A4 mRNA at lower concentration (batch HH1012) and at higher concentration (batch HH952) of TPM. CYP3A4 mRNA levels increased with increasing TPM concentrations, again reflecting the pattern observed with the testosterone 6β-hydroxylation activity. At low TPM concentration (10–50 μM), a 1.2- to 1.9-fold increase was noted in CYP3A4 mRNA levels (p > 0.05). However, at concentrations ranging from 100 to 500 μM, the increase in the mRNA levels ranged from 4.3 to 13.3-fold (p < 0.05).

Figure 2.

Representative Northern blots depicting CYP3A4 mRNA levels in human hepatocytes treated with topiramate (TPM). A: CYP3A4 mRNA levels in human hepatocytes treated with phenobarbital (PB; 2 mM), rifampicin (Rif; 10 μM), and TPM (10–250 μM) compared with untreated control hepatocytes (C). B: CYP3A4 mRNA levels in human hepatocytes treated with PB (2 mM), Rif (10 μM), and TPM (100–500 μM) compared with untreated control hepatocytes (C).

Because TPM increased transcription of CYP3A4, we tested whether this was associated with the activation of hPXR, an orphan nuclear receptor involved in the regulation of CYP3A4 expression by structurally diverse compounds such as rifampicin and PB. The activation of hPXR by TPM was examined by transient transfection in CV-1 cells by using a reporter gene construct harboring hPXR-responsive regions of the CYP3A gene. In parallel transfection experiments, we compared the fold activation of hPXR by TPM (10 to 500 μM) to rifampicin (10 μM) and PB (2 mM) (Fig. 3). We observed an ∼6-fold activation of hPXR after treatment with rifampicin (10 μM) and PB (2 mM), which is consistent with previous observations (5). Statistically significant (p < 0.05) increase in the hPXR activation was observed at TPM concentration ≥50 μM. TPM activated hPXR in a concentration-dependent manner with a maximal activation (threefold) at 500 μM.

Figure 3.

Activation of human pregnane X receptor (hPXR) by topiramate (TPM) in transient transfection assays. Fold-activation of hPXR after 24-h treatment with various concentrations of TPM (white bars, 10–500 μM), rifampicin (black bar, 10 μM), and phenobarbital (striped bar, 2 mM). The data are expressed as fold-activation of normalized CAT expression compared with untreated controls (mean ± SEM, n = 3). *p < 0.05, **p < 0.01, ***p < 0.005 compared with untreated controls.

Overall, TPM-mediated fold-increases in the CYP3A4 activity and CYP3A4 immunoreactive protein content of the hepatocytes were well correlated (R2= 0.95, p = 0.0004, Fig. 4A). The fold-increase in CYP3A4 immunoreactive content and CYP3A4 mRNA levels, in TPM treatments, also correlated very well (R2= 0.97, p = 0.0006, Fig. 4B). When TPM-mediated increase in the CYP3A4 activity and expression as well as hPXR activation were compared with those observed with the positive controls such as rifampicin, we observed a fairly consistent pattern. For instance, the ratio of an observed effect at a given TPM concentration (50, 100, 250, and 500 μM) to that of rifampicin (10 μM), ranged from 0.11 to 0.56 for CYP3A4 activity, 0.18 to 0.68 for CYP3A4 immunoreactive protein levels, 0.095 to 0.66 for CYP3A4-specific mRNA levels, and 0.21 to 0.5 for hPXR activation. Overall, this indicates that when indexed against rifampicin, TPM at 50–500 μM may exhibit ∼10 and 50% induction capacity relative to the 10 μM rifampicin, which is known to cause marked CYP3A4 induction.

Figure 4.

Correlation between fold-increase in activity and expression of CYP3A4 in hepatocytes after treatment with topiramate (10–500 μM), phenobarbital (2 mM), and rifampicin (10 μM). A: Fold-increase in immunoreactive CYP3A4 levels plotted in relation to the fold-increases in CYP3A4 activity measured as 6β-hydroxylation of testosterone (means, n = 4). B: Fold-increase in CYP3A4 mRNA levels plotted in relation to the fold-increase in immunoreactive CYP3A4 protein levels (means, n = 4). Fold-increase responses for drug treatments were calculated with respect to untreated controls.


In this study we observed that TPM treatment of human hepatocytes resulted in marked increase in the levels of CYP3A4-specific mRNA, in immunoreactive levels, and in the catalytic activity of the enzyme. The increase in the CYP3A4 activity and expression was a function of TPM concentration, with most pronounced effects occurring only at the higher TPM concentrations. In a previously reported clinical study, TPM administration caused a reduction in ethinyl estradiol concentrations when oral contraceptives were administered along with TPM (2) at a dose range of 200–800 mg/day. As indicated earlier, the mean Cmax value for ethinyl estradiol decreased by 25.4%, and the AUC0-24 decreased by 30% at the 800-mg/day dose level. However, in a recently published comparative evaluation, Doose et al. (3) did not observe a statistically significant induction of ethinyl estradiol clearance in the presence of TPM, whereas concomitant carbamazepine (CBZ) therapy notably induced the clearance. The daily TPM doses used in this study were 50, 100, and 200 mg. The investigators also did not observe an increase in the urinary 6β-hydroxycortisol/cortisol ratio in patients treated with TPM at these doses, which is reflective of lack of CYP3A4 induction in the human subjects.

The results of our in vitro study may provide an explanation for these seemingly contradictory observations. As seen in our studies, distinct CYP3A4 induction occurs only at TPM concentrations >50 μM. It is unlikely that the TPM concentrations in subjects receiving TPM doses <200 mg/day would yield such concentrations. However, higher doses of TPM (≥800 mg/day; 400 mg b.i.d) are likely to yield peak and steady-state plasma levels in excess of concentrations (50 μM) where we see distinct CYP3A4 induction (14,15). Moreover, given the large interindividual variability in TPM pharmacokinetics, it is likely that some individuals, especially those with impaired renal function, may achieve such levels even at lower TPM doses. To the best of our knowledge, assessment of CYP3A4 activity in patients with such approaches as the erythromycin breath-test analysis or the urinary ratio of 6β-hydroxycortisol/cortisol, at TPM doses >200 mg/day have not been reported. The extrapolation of data from our in vitro study to the in vivo situation should be done with consideration of several factors. Given that TPM has high water solubility (9 mg/ml) and low plasma protein binding (13–17%) (16,17) and that the hepatocyte cell-culture medium used in our study does not contain proteins, it is conceivable that the interstitial TPM levels achieved in human livers are reflected in the concentration range used in our study. TPM also has a relatively long elimination half-life in humans (∼16 h), and with the usual b.i.d. regimen clinically used (14,16), the relative fluctuations between peak and trough levels during steady state are minimal. This low degree of fluctuation coupled with the fact that TPM must be taken continuously for many years, it can be reasoned that the in vivo exposure of hepatocytes to TPM occurs for an extremely long period. These factors underscore the potential for enzyme induction with the clinical use of TPM. As perspective, the induction capacity of TPM relative to rifampicin (10 μM) over the concentration range of 50–500 μM is ∼10–50%. Taken together, this implies that although the overall induction potential of TPM is low, it can reach clinical significance at the higher end of clinically used dosing regimens.

The induction of hepatic CYP3A4 by TPM observed here suggests that the reduction in systemic bioavailability of ethinyl estradiol reported by Rosenfeld et al. (2) may have been contributed by both the systemic metabolism and the first-pass hepatic metabolism of the compound. The possibility that TPM also induces the intestinal metabolism, which was not addressed in our studies, cannot be ruled out. Whereas most compounds that induce hepatic CYP3A4 also induce the intestinal enzyme, a few notable exceptions exist to this observation (18). Currently the unavailability of an appropriate in vitro tool to study intestinal CYP induction is a major limitation to investigating this possibility. Another limitation of our study is that CYP3A4 and CYP3A5, a closely related analogue often coexpressed with CYP3A4 in hepatocytes, are indistinguishable based on testosterone 6β-hydroxylase activity, the polyclonal antibody used for Western blot analysis and the oligonucleotide probe from Northern blot analysis used in this study. Therefore it is likely that the measured enzyme activity and expression ascribed to CYP3A4 may include a contribution by CYP3A5. Usually the expression of CYP3A5 ranges only from 2 to 5% relative to CYP3A4, and therefore the overall CYP3A5 contribution may be marginal (4). However, recent studies show that CYP3A5 expression exhibits genetic polymorphism, and in a small number of individuals, the levels of the enzyme may be considerably higher (19), especially in patients of Chinese origin (20). Because CYP3A5 is considered to be a noninducible enzyme, in these subjects, TPM may not modulate the enzymatic activity.

This study also demonstrated that at relatively high concentrations, TPM is an efficacious activator of the hPXR, a recently cloned orphan nuclear receptor. Rapid advances in our understanding of the regulation of CYP enzymes has revealed that PXR is activated by a number of CYP3A4 inducers such as rifampicin, PB, and hyperforin in a ligand-dependent manner (5,6,21–23). In the presence of an activating ligand, PXR forms a heterodimer with retinoid X receptor (RXRα). This heterodimer binds to the xenobiotics response element (XRE) in the promoter sequence of CYP3A4, leading to increased gene transcription. Typically, PXR ligands are lipophilic agents that bind to hydrophobic binding sites on hPXR (22,23). TPM appears to be one of the few water-soluble compounds that activate hPXR.

Our findings underscore the potential use of cell-based reporter assays such as the one used here to screen TPM-related compounds for CYP3A4 induction. Because the use of human hepatocytes has several limitations, the use of high-throughput methods for screening CYP3A4 is rapidly gaining wide acceptance. Although the approach for transient transfection of CV-1 cells used here has been extensively used in earlier studies (5,23), a potential limitation of the model is that it does not express receptors and the complement of other factors involved in transcriptional activation of hepatic genes, which include receptor coactivators and corepressors. In this regard, it may be physiologically more relevant to use primary hepatocytes or hepatic cell lines, which also allow a higher level of hPXR activation. Further, we did not test whether TPM also activates the constitutive androstane receptor (CAR), another nuclear receptor that, as a heterodimer with RXRα, binds to the XRE in the CYP3A4 promoter. It has been postulated that PXR and CAR cross-regulate target genes (24–26). Further studies using CAR constructs, dominant negative receptor (PXR/CAR) constructs, or PXR/CAR-null mice are warranted to clarify the role of CAR, if any, in TPM-mediated induction of CYP3A4.

Several AEDs known to be associated with multiple drug–drug interactions activate hPXR (21). Although both PB and CBZ activate hPXR, CBZ is comparatively weaker (27). Because pronounced CYP3A4 induction and hPXR activation by TPM occurred in our model systems only at higher concentration range, TPM may be associated with fewer drug–drug interactions. However, in addition to the interaction with ethinyl estradiol, TPM may be involved in other interactions. For instance, a modest but statistically significant reduction in the mean AUC (∼12%) and increase in the oral clearance of valproic acid (VPA) was observed with concomitant 800-mg/day TPM administration (14,28). Although the clinical significance of this interaction was not apparent, this observation also indicates the potential for drug–drug interactions at high-dose TPM therapy. The involvement of CYP3A4 in VPA is currently unclear, but previous studies indicated that it undergoes oxidation catalyzed by CYPs 2A6, 2C9, and 2C19 and glucuronidation by UDP glucuronosyltransferases (UGTs) 1A6, 1A9, and 2B7 (29). CYP2C9 and, to a lesser extent, CYP2C19 and UGTs, also are regulated by PXR (30). As seen in our study, because TPM activates PXR, it is likely that TPM also may induce PXR target genes other than CYP3A4, although this may occur only at high doses of TPM.

In conclusion, this study found that TPM induces CYP3A4 activity and expression at concentrations likely to be achieved in patients receiving high-dose TPM therapy. CYP3A4 induction by TPM may result from the ability of TPM to activate hPXR. Our study provides the molecular basis for the clinical observation that TPM administration results in enhanced clearance of ethinyl estradiol. Moreover, it alerts us to the potential for CYP3A4-based drug–drug interactions with TPM, especially when TPM doses ≥400 mg/day are used. In relation to the other AEDs such as CBZ and PB, the overall CYP3A4 induction potential of TPM may be low. However, with high doses of TPM, drug–drug interaction may be anticipated, and precautionary measures such as adjusting the dose of ethinyl estradiol or that of similarly affected other CYP3A4 substrates, can be exercised.


Acknowledgment:  This study was supported in part by grant NS40261 from the National Institutes of Health. TPM was a gift from the Robert Wood Johnson Pharmaceutical Research Institute.