Abstract
Over recent years the application of cocktail studies to measure biological markers has become increasingly popular. The current study investigated a novel approach in assessing cytochrome P450 (P450) enzyme induction in an immortalized cell line using a cocktail of five P450 substrate probes compared with the traditional single-probe approach. The findings reported herein support use of a cocktail approach to assess the induction of the major P450s, namely, CYP3A4, CYP1A2, and CYP2C9. CYP2C19 and CYP2D6 could also be followed as part of the cocktail approach reported. Response to prototypical inducers did not differ to those observed in the presence of the specific probes alone. Consequently, this approach requires significantly fewer sample numbers if screening the induction potential of more than one P450. Moreover, these studies highlight the utility of the immortalized cell line Fa2N4 as a robust model system for induction studies. In conclusion, the current experimental setup is an improvement on current approaches used to assess P450 induction, significantly increasing sample throughput.
Drug-drug interactions (DDI) are a particularly important type of adverse drug event that can result in the physiological levels of a drug climbing above or falling below this therapeutic window. A continual increase in plasma concentrations of a drug can ultimately result in toxicity, whereas a decrease results in the failure in efficacy. In this regard, a new chemical entity's (NCE) progression through the discovery/development process is rate-limited, depending on the degree of undesirable properties/actions, which include in part interactions with members of the hepatic drug-metabolizing enzymes, in particular the cytochrome P450 (P450) family. Among the most important adverse interactions that NCE must be assessed for is their propensity to act as perpetrators, in either the inhibition and/or induction of P450 enzymes (see reviews by Lin and Lu, 1998; Lin, 2006), the latter of which is the basis of the current series of studies.
The induction of P450 enzymes results in an increase in the steady-state concentrations of these enzymes following organ exposure to relevant stimuli. This is predominantly observed with hepatic P450s, CYP3A4 and CYP1A2, and to a lesser extent, CYP2C9 and CYP2B6 (Dickins, 2004). Over recent years, there has been a considerable increase in the knowledge surrounding the mechanism behind these processes. The cascade(s) involved in the eventual increase in the steady-state concentration has been shown to involve activation of several nuclear hormone receptors: the pregnane X receptor (PXR), aryl hydrocarbon receptor (AhR), and the constitutive androstane receptor (CAR) (for more detailed descriptions of these pathways, see reviews by Wang and LeCluyse, 2003; Bock and Kohle, 2004). In brief, CAR and PXR regulate distinct but overlapping sets of target genes, which include certain phase I P450 enzymes (e.g., CYP2B, CYP3A, and CYP2C), phase II conjugation enzymes such as UDP-glucuronosyltransferase UGT1A1 and sulfotransferase SULT2A, and phase III transporters such as P-glycoprotein (MDR-1). The AhR receptor has been shown to regulate the expression of CYP1A (Whitlock, 1999).
In the current study, an immortalized human hepatic cell line (Fa2N4 cells) was used as a model system to assess enzyme induction by two established inducers, rifampicin (Waxman, 1999; Roymans et al., 2004) and omeprazole (Raucy, 2003; Roymans et al., 2004). A recent study by Mills et al. (2004) has described the utility of Fa2N4 cells as an in vitro model to assess P450 induction, showing significant retention of activity of P450 enzymes and hepatic transporters such as MDR-1. More recently, this cellular model system has been used to predict clinical induction caused by CYP3A4 (Ripp et al., 2006). Herein we report a novel approach in assessing P450 enzyme induction using a cocktail of substrate probes, allowing the quantification of up to five major P450s in a single assay, thus significantly increasing the throughput for these experiments. Throughput is further increased through the combined assessment of enzyme activity and mRNA in a single sample.
Materials and Methods
Chemicals. Midazolam, 1-OH midazolam, (S)-mephenytoin, 4-OH mephenytoin, and 4-OH diclofenac were generated biosynthetically at Pfizer Inc. (Sandwich, UK). Dextrorphan was purchased from Sigma/RBI (Natick, MA). Dulbecco's phosphate-buffered saline (PBS without sodium bicarbonate) was obtained from Gibco (Paisley, UK). Fa2N4 cells and MFE support (serum-free) and plating media were obtained from MultiCell Technologies (Lincoln, RI). Fetal calf serum (heat-inactivated), penicillin (100 U/ml), and streptomycin (100 μg/ml) were purchased from Invitrogen (Paisley, UK). Collagen I flasks and plates were purchased from BD Biosciences (Cambridge, UK). Organic solvents were obtained from Romil (Cambridge, UK). Promega (Southampton, UK) supplied nuclease free water and SV-96 RNA extraction kits. All the other chemicals were purchased from Sigma-Aldrich (Poole, UK).
Cell Culture. Fa2N4 cells were grown on Biocoat type I collagen-coated flasks and maintained in MFE support media (serum-free). Flasks were incubated at 37°C, 5% carbon dioxide, and 95% relative humidity. Support media were replaced every 2 days, and cells were passaged at least once a week. For passaging, cells were trypsinized (1 ml/25-cm2 plate surface area) for 5 min, subsequently harvested in MFE plating media (containing serum proteins) (1:3 v/v), and centrifuged at 500g for 5 min, and the pellet was reconstituted in fresh MFE plating media and added into Biocoat type I collagen flasks. After 4-h incubation, MFE plating media were replaced with MFE support media. All the experiments were performed using Fa2N4 cells at passages between 10 and 20. No significant changes in basal activities in probe metabolism were observed between passages 10 to 20.
Enzyme Induction. For induction studies, Fa2N4 cells were plated onto Biocoat type I collagen-coated 24-well plates at 0.5 × 106 cells/well in MFE plating media and supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin. After 4 h, the plating media were replaced with MFE support media. At 48 h after seeding, cells were supplemented with fresh MFE support media containing prototypical inducers rifampicin and omeprazole. A range of inducer concentrations were prepared from 25 mM stock solutions in dimethyl sulfoxide. Aliquots were subsequently diluted in MFE media (1 μl/ml), giving a final concentration range of inducers to be studied as 0, 0.5, 1, 5, 10, and 25 μM. These solutions were supplemented at 0.5 ml into appropriate wells of the 24-well plates. The final organic content per well was equal to 0.1%. This procedure was repeated using fresh inducer solutions every 24 h, up to a total of 96 h.
Analysis of Enzyme Activity. Single-probe studies: CYP3A4, CYP1A2, and CYP2C9 enzyme activities were determined by measuring the extent of metabolism of specific probes—tacrine (CYP1A2), diclofenac (CYP2C9), (S)-mephenytoin (CYP2C19), and midazolam (CYP3A4)—to their respective metabolites—1-hydroxy tacrine, 4-hydroxydiclofenac, 4-hydroxy mephenytoin, and 1-hydroxy midazolam—by mass spectrometry (see “Analytical Procedure”). Midazolam was chosen as the CYP3A4 probe owing to its clinical relevance (Kato et al., 2005).
In brief, test inducer solutions were aspirated from each well, and the cells were washed twice with PBS, pH 7.4. Each probe was prepared in Krebs-Henseleit bicarbonate (KHB, pH 7.4) to give a final concentration of 25 μM. Cells were incubated with 350 μl of each solution for 1 h at 37°C (5% CO2), after which 300-μl aliquots were collected and added to tubes containing 50 μl of ice-cold acetonitrile to precipitate the proteins. These solutions were mixed and stored at –80°C until analyzed.
Cocktail probe studies: the cocktail solution contained the probe substrates tacrine (CYP1A2), diclofenac (CYP2C9), (S)-mephenytoin (CYP2C19), dextromethorphan (CYP2D6), and midazolam (CYP3A4), which were prepared collectively in KHB buffer, each at a final concentration of 25 μM. Test inducer solutions were aspirated from each well, and the cells washed twice with PBS. Cells were then incubated with the cocktail solution for 1 h at 37°C (5% CO2). Aliquots (300 μl) were collected and added to tubes containing 50 μl of ice-cold acetonitrile to precipitate the proteins. These solutions were mixed and stored at –80°C until analyzed. Formation of dextrorphan was followed as a negative control because CYP2D6 is not inducible.
Analytical Procedure. Enzyme activities were measured by liquid chromatography/mass spectrometry using a Sciex API4000 with Turbo IonSpray (Hertfordshire, UK) in conjunction with an Agilent 1100 liquid chromatography binary pump (Cheshire, UK). Samples (80 μl) were injected onto a Phenomenex luna phenyl-hexyl column (50 × 4.6 mm) (Cheshire, UK), and peaks were separated by reversed-phase chromatography (0–1 min 100% A/0% B, 1–6 min 0% A/100% B, 6.1–8 min 100% A/0% B) using the following mobile phases: A, MeOH-ACN/H2O (10:90) containing 2 mM ammonium acetate and 0.025% formic acid (v/v) and B, MeOH-ACN/H2O (90:10) containing 2 mM ammonium acetate and 0.025% formic acid (v/v) at a flow rate of 1 ml/min. The current procedure allows separation of the probe compounds from their respective metabolites. These metabolites are subsequently quantified using authentic standards of each respective metabolite, prepared individually/or as part of a cocktail, depending on the enzyme activity analysis procedure used. Values were used to calculate enzyme activities as nanomoles of metabolite per million cells per minute of incubation.
RNA Isolation. For single-probe studies, Fa2N4 cells were seeded into separate 24-well plates to allow assessment of mRNA induction. Following collection of samples for analysis of enzyme activities, the remaining probe mix was removed, and the cells were washed with sterile water. RNA was then extracted from these samples using the Promega SV 96 Total RNA Isolation System. Lysis buffer containing β-mercaptoethanol was then added to each well to disrupt cellular architecture and release DNA. Following 10-min incubation at 37°C (5% CO2), supernatant was titrated and collected into RNase-free tubes.
Total purity of the RNA extracted was determined by determining the RNA/DNA ratio through absorbance intensity measurements at 260:280 nm. Ratios greater than 1.8 were considered to have negligible DNA contamination. Absorbance values at 260 nm were then used to estimate the concentration of RNA in the sample using the Beer-Lambert equation: A = ebc, where A = absorbance (no units); e = molar absorptivity or extinction coefficient (M/cm); b = path length of the sample (cm); and c = concentration of the compound in solution (M).
Agilent RNA 6000 Nano Assay kits were also used to spot-check the robustness of the automated RNA extraction procedure and confirm the purity of RNA (as estimated by the 260:280 nm ratio). The RNA 6000 Nano LabChip kit provides information on the quality and quantity of RNA samples. The assay kit is designed primarily to check the quality and determine the concentration of total RNA and mRNA samples. The method used followed that recommended by the suppliers. The frequency of extracted RNA with a 260:280 nm ratio of greater than 1.8 was >90% using the current extraction procedure.
Assessment of Changes in mRNA Expression. For assessment of mRNA induction, samples were diluted to a final concentration of 10 ng/μl. mRNA expression was quantified for CYP3A4 only using commercially available INVADER assay kits from Third Wave Technologies (Madison, WI) in accordance with the procedure supplied.
Statistical Analysis. The formation rate data of the respective metabolites were initially log-transformed and subsequently examined by analysis of variance (ANOVA) to measure the influence of inducer concentration on -fold induction. Inspection of the residuals generated following ANOVA indicated that log transformation was required to stabilize the variance. [This transformation was performed on practical grounds because differences between means of log (response) become differences in -fold change.] Statistical differences between each inducer concentration and vehicle control were not calculated. However, all the figures included a line that represented a 2-fold increase in response compared with vehicle control, against which different inducer concentrations could be assessed. This approach is in line with criteria against which a compound may be classed as a potential inducer (Bjornsson et al., 2003). In addition, the 95% confidence intervals (CI) for each data set were also included above and below the 2-fold change.
Results
The induction in catalytic activities of CYP1A2, CYP2C9, and CYP3A4 was studied in an immortalized human hepatocyte cell line, Fa2N4, following exposure to the known inducers rifampicin or omeprazole for 72 or 96 h. These P450s were studied following supplementation of tacrine, diclofenac, and midazolam either alone or as part of a cocktail set. The cocktail set also included (S)-mephenytoin and dextromethorphan as probe substrates for CYP2C19 and CYP2D6. Previous preliminary studies performed in-house have examined the specificity of these five probes for metabolism by the P450s under investigation and confirmed the lack of any interactions (data not shown). Assessment of dextromethorphan metabolism served as a negative control because CYP2D6 is not inducible under the conditions of this study. This was confirmed in the cocktail probe study, in which no changes in the constitutive CYP2D6 activity were observed in the presence of either rifampicin or omeprazole. The average formation rate of dextrorphan was <2 nmol/min/million cells following exposure of either inducer between 0 and 25 μM (data not shown).
CYP3A4. Midazolam was incubated with Fa2N4 cells either alone or as part of a cocktail set. Figure 1A highlights the changes in CYP3A4 enzyme activity (nmol/min/106 cells) as measured by 1-OH midazolam formation following incubation with midazolam only (single probe) after exposure to rifampicin for either 72 or 96 h. A significant induction in CYP3A4 activity was observed as a function of increasing rifampicin concentration following 72- and 96-h exposure (ANOVA; p < 0.05 in each case). Figure 1, B and C, shows the formation rate data following log transformation. The solid line represents a 2-fold increase from vehicle control, whereas the dashed lines represent the 95% CI associated with the experimental data set. Rifampicin exposure for 72 h at 0.5 and 1 μM (Fig. 1B) resulted in a 2-fold induction of CYP3A4 activity, and a >2-fold induction was observed with 5, 10, and 25 μM rifampicin. Similar responses were seen following rifampicin exposure for 96 h (Fig. 1C). However, no significant differences were observed between the levels of CYP3A4 -fold induction following 72- or 96-h exposure at any rifampicin concentration studied.
Assessment of CYP3A4 induction using midazolam as part of a cocktail probe set yielded similar results to those observed using the single probe (Fig. 2, A–C). CYP3A4 activity increased as a function of increasing rifampicin concentration following 72- and 96-h exposure (ANOVA; p < 0.05 in each case). Formation rates were lower than those seen in the single-probe study. However, these differences were not found to be statistically significant. Rifampicin exposure for 72 h at 0.5 and 1 μM (Fig. 2B) resulted in a 2-fold induction of CYP3A4 activity, and a >2-fold induction was observed with 5, 10, and 25 μM rifampicin, consistent with -fold changes observed in the single-probe experiment. Rifampicin exposure for 96 h also resulted in a 2-fold induction of CYP3A4 activity at 0.5 and 1 μM (Fig. 2B) and a >2-fold induction at 5, 10, and 25 μM rifampicin, again consistent with -fold changes observed in the single-probe experiment. Again, no significant differences were observed between the levels of CYP3A4 -fold induction following 72- or 96-h exposure at any rifampicin dose studied. Moreover, no significant differences were observed between single and cocktail studies.
RNA was isolated using the commercially available kits from Promega. Isolated RNA was free from contamination with DNA, as determined by the OD260:280 ratio. Only samples with ratios >1.8 were used in these studies. The robustness of the isolated mRNA was further confirmed using Agilent RNA 6000 Nano Assay kits.
Induction of CYP3A4 mRNA was determined using commercially available kits from Third Wave Technologies. RNA was isolated from cells previously exposed to midazolam, either alone (single probe) or as part of the cocktail probe set. In the case of the single-probe study, induction of CYP3A4 mRNA increased as a function of rifampicin concentration following exposure for either 72 or 96 h (Fig. 3A) (ANOVA; p < 0.05 in each case). Log transformation of the data showed a 2-fold induction of CYP3A4 mRNA with rifampicin at 0.5 μM and >2-fold induction at concentrations from 1 to 25 μM following exposure for 72 h (Fig. 3B). In contrast, 96-h exposure resulted in >2-fold induction of CYP3A4 mRNA at all the rifampicin concentrations studied (Fig. 3C). The -fold increases measured in the cocktail probe study (Fig. 4, A–C) followed the same pattern as those in the single-probe study, with no significant differences between the responses in the two approaches, single versus cocktail studies.
CYP1A2.Figure 5A highlights the changes in CYP1A2 enzyme activity (nmol/min/106 cells) after exposure to the recognized inducer omeprazole as measured by 1-OH tacrine formation (single probe). A significant induction in CYP1A2 activity was observed as a function of increasing omeprazole dose following 72- and 96-h exposure (ANOVA; p < 0.05 in each case). Figure 5B shows the log-transformed data following omeprazole exposure for 72 h. A 2-fold induction of CYP1A2 was measured between 1 and 10 μM, and a >2-fold induction was seen at 25 μM omeprazole. Following 96-h omeprazole exposure (Fig. 5C), a 2-fold induction was measured at 1 μM; however, a >2-fold induction was measured at 5 to 25 μM omeprazole. Responses at 96 h were higher than those seen at 72-h exposure, although this was only significant at 25 μM omeprazole (p < 0.05).
Using the cocktail probe approach, a similar induction in CYP1A2 activity was observed, as a function of increasing omeprazole dose following 72- and 96-h exposure (ANOVA; p < 0.05 in each case) (Fig. 6A). No induction was observed at 72-h exposure with omeprazole at 0.5 to 5 μM (Fig. 6B). However, a 2-fold induction was observed at 10 μM, and a >2-fold induction was seen at 25 μM omeprazole. Exposure of omeprazole for 96 h resulted in a 2-fold induction at 5 μM and >2-fold at 10 and 25 μM (Fig. 6C). Overall, no significant differences were observed between responses observed at 72- and 96-h omeprazole exposure. Comparisons between single- and cocktail-probe studies showed no differences between each of the approaches, with the exception of omeprazole exposure for 96 h at 25 μM, for which the responses were found to be higher when supplementing tacrine alone.
CYP2C9.Figure 7A highlights the changes in CYP2C9 enzyme activity (pmol/min/106 cells) after exposure to the recognized inducer rifampicin, as measured by 4-OH diclofenac formation (single probe). A significant induction in CYP2C9 activity was observed as a function of increasing rifampicin dose following 72- and 96-h exposure (ANOVA; p < 0.05 in each case). Figure 7, B and C, shows the log-transformed activity data following rifampicin exposure for 72 and 96 h, respectively. A 2-fold induction was measured at 0.5 to 25 μM at both exposure times.
Using the cocktail probe approach, a significant induction in CYP2C9 activity was also observed as a function of increasing rifampicin dose following 72- and 96-h exposure (ANOVA; p < 0.05 in each case) (Fig. 8A). Figure 8, B and C, shows the log-transformed activity data following rifampicin exposure for 72 and 96 h, respectively. A 2-fold induction was observed with 5 to 25 μM rifampicin following 72-h incubation. In comparison, 2-fold induction was only observed at 25 μM rifampicin following 96-h incubation. However, no significant differences between the responses using single versus cocktail approaches were observed.
CYP2C19. In the current series of experiments no induction of CYP2C19 was observed using either the cocktail or single-probe approach (data not shown). The average formation rate of 4-OH mephenytoin was <1 nmol/min/million cells following rifampicin exposure between 0 and 25 μM.
Discussion
The current study compared the robustness of a cocktail probe approach to investigate the induction of the major drug-metabolizing P450 enzymes and CYP3A4 mRNA levels with the routinely used single-probe approach. The specificity of these probes toward their targeted P450s has previously been reported (Hakala et al., 2005; Mohutsky et al., 2005; Turpeinen et al., 2005). Studies were performed in an immortalized human hepatocyte cell line: Fa2N4 (Mills et al., 2004), which has been reported to exhibit phenotypic properties, commonly expressed in hepatocytes, including inducibility of CYP1A2, CYP2C9, CYP3A4, UGT1A, and MDR1 mRNA in response to known inducers. These observations were in agreement with studies performed in cultured primary human hepatocytes (Madan et al., 2003), confirming their suitability as a substitute in vitro model to fresh human hepatocytes and other reported hepatic cell model systems (Gomez-Lechon et al., 2003; Trubetskoy et al., 2005; Kafert-Kasting et al., 2006). Moreover, these cells are readily available and can be cultured as experimental needs demand, in contrast to fresh human hepatocytes, whose use is reliant on the availability of a suitable liver tissue donor.
The induction of P450 enzymes, following hepatic exposure to NCE, results in an increase in the amount of protein. This is predominantly observed with CYP3A4 and CYP 1A2 and to a lesser extent with CYP2C9 and CYP2B6 (Dickins, 2004). The mechanism involved in the up-regulation of these proteins has been shown to involve activation of nuclear hormone receptors PXR, AhR, and/or CAR (Wang and LeCluyse, 2003; Bock and Kohle, 2004). The effect of activation of these nuclear receptors can influence disposition of many xenobiotics and endogenous substances. For example, a drug that activates CAR or PXR (Wei et al., 2002) may induce changes in the metabolism or transcellular transport of one or more coadministered drugs, reducing their therapeutic efficacy (e.g., oral contraceptives containing ethinyl estradiol). Conversely, activation of PXR may lead to drug-induced toxicity, as seen with acetaminophen (Guo et al., 2004). Clearly, early identification of any undesirable properties is an important feature in the drug discovery process.
Mills et al. (2004) reported that Fa2N4 cells provide a suitable model to assess induction of P450 enzyme activities and mRNA transcripts. In their studies, two separate assays were performed to collect material to measure these markers. Owing to a dramatic increase in the number of NCE in the field of drug discovery, there has been a drive by the pharmaceutical industry to improve efficiency and screening techniques. Traditional assays generate large numbers of samples, requiring cumbersome, labor-intensive, and cost-ineffective analytical procedures. To this end, the application of cocktail studies to measure biological markers has become increasingly popular. With respect to drug metabolism and in particular in vitro DDI, numerous cocktail procedures have been reported, allowing simultaneous determination of an NCE on the activity of the major drug-metabolizing enzymes (Dierks et al., 2001; Zhou et al., 2004; Hakala et al., 2005; Kim et al., 2005; Mohutsky et al., 2005; Turpeinen et al., 2005). The majority of these studies report higher throughput in vitro P450 techniques for evaluating the potency of putative inhibitors (Dierks et al., 2001; Zhou et al., 2004; Hakala et al., 2005; Kim et al., 2005; Mohutsky et al., 2005; Turpeinen et al., 2005). These assays have been shown to use a variety of analytical techniques and differing numbers of substrates. A recent review by Zhou et al. (2004) described drug cocktail studies having used between two- and six-probe substrates simultaneously. More recently, procedures measuring between seven- (Dierks et al., 2001) and nine-probe molecules simultaneously (Kim et al., 2005) have been reported.
Within the context of the current study, few cocktail approaches have examined the P450-inductive potential of NCE in vitro. Mohutsky et al. (2005) reported a cassette strategy to measure induction of CYP1A2, CYP2C9, and CYP3A4 enzyme activities in primary hepatocytes. They found that induction in the presence of prototypical inducers were similar to those measured using the single-probe substrate approach. The current study also investigated the induction of CYP1A2, CYP2C9, and CYP3A4. However, the cocktail probe set also included probes for CYP2C19 and CYP2D6, the latter of which served as a negative control because this P450 isoform is not inducible. It is important to note that induction is not limited to phase I drug-metabolizing enzymes and that other phase II enzymes could also be induced, such as UGT enzymes (Soars et al., 2004). However, the use of well known prototypical inducers of P450 enzymes supports the use of the probes used in these studies, namely, tacrine (CYP1A2), diclofenac (CYP2C9), (S)-mephenytoin (CYP2C19), dextromethorphan (CYP2D6), and midazolam (CYP3A4), despite some potential for phase II metabolism of their metabolites.
Observations in the current study confirm the robustness of the Fa2N4 cell line as a model to assess P450 induction. CYP3A4 exhibited the largest response following exposure to the prototypical inducer, rifampicin, followed by CYP1A2 (omeprazole) and CYP2C9 (rifampicin). Furthermore, the cocktail approach described can be used to determine the effects of an NCE on the activity of the major drug-metabolizing CYP3A4, CYP1A2, and CYP2C9. The induction responses to the prototypical inducers did not differ from those measured using single probes. This confirmed that the probe substrates when presented as a cocktail did not interact adversely in anyway to affect any activities. The induction pattern (CYP3A4 > CYP1A2 > CYP2C9) was in agreement with previous studies (Soars et al., 2004), having shown that induction via the PXR pathway is more pronounced than that via CAR or AhR pathways (LeCluyse, 2001).
Although both CYP3A4 and CYP2C9 are induced via the PXR pathway, differences in the sequences of their respective DNA promoter regions have been shown to explain the observed lower induction response of CYP2C9 to rifampicin compared with CYP3A4 (Goodwin et al., 1999; Gerbal-Chaloin et al., 2001). Differences in the architecture of the promoter regions of CYP2C19 have also been reported, explaining its reduced response to induction by rifampicin (Gerbal-Chaloin et al., 2001). However, the current study failed to identify induction of CYP2C19 with rifampicin at any concentration studied. Although induction of CYP2C19 using in vitro models systems have been reported previously (Wang et al., 2004), induction is significantly lower than that of CYP3A4 and CYP1A2. The absence of CYP2C19 induction may simply be indicative of the current cell model. Few, if any, clinical DDI have been reported as a consequence of CYP2C19 induction. However, its induction should not be dismissed, despite the ability to measure basal activities.
Whereas Fa2N4 cells exhibit inductive responses to prototypical inducers, the dynamic ranges over which these responses are observed are also important. -Fold changes in induction were identical in both the single- and cocktail-probe approaches, further supporting the utility of this approach to measure induction of P450 enzymes. It has been recommended that the minimum induction for CYP3A4 and CYP1A2 should be 2-fold (http://www.fda.gov/cder/drug/druginteractions/guidance.htm). Using the cocktail approach, the enzyme activities of CYP3A4, CYP1A2, and CYP2C9 were induced by at least 2-fold. In each case, the level of induction was higher following incubation with inducers for 96 h compared with 72 h, albeit not significant, suggesting the shorter incubation time would suffice. Even shorter incubation times may provide a suitable -fold increase in the activities of these P450 enzymes. However, this was not investigated here. Interestingly, the current experimental model showed a 2-fold induction of CYP2C9, in contrast to recent studies, in which no significant induction of CYP2C9 was observed (Madan et al., 2003; Mohutsky et al., 2005). This may reside in the differences in models used, immortalized compared with primary hepatocytes, and/or the experimental conditions (e.g., the length of time inducers were incubated with the cells) and the P450-specific probes used.
Although enzyme activity represents the gold standard measure of in vitro induction, greater emphasis is being paid toward assessment in changes in P450 mRNA expression. The current study reports that material collected from a single experimental well could be used to assess both enzyme activities and mRNA, helping minimize confounding interwell influences if having to run separate experiments. Owing to recent advances in technologies and techniques, such as reverse transcription-polymerase chain reaction, promoter gene assays, and gene chips, growing emphasis is being paid toward measurement of mRNA induction. Although cocktail approaches can help significantly reduce sample numbers to be analyzed, procedures to assess mRNA provide another tool whereby the pharmaceutical industry can significantly increase throughput. Changes in mRNA levels of P450s can provide an early screen of the inductive potential of compounds and related chemical series, with shorter analysis time than analytical techniques such as liquid chromatography/mass spectrometry. Moreover, there is a greater dynamic range across which changes in P450 induction can be studied. Moreover, measurement of changes in mRNA together with enzyme activities can provide significantly more information than with either alone. In this regard, the inclusion of mRNA assessment would reduce false-negative findings, as might been seen with potent P450 inhibitors, including mechanism-based inhibitors, which would significantly reduce P450 metabolism of probe molecules but would not necessarily affect transcriptional up-regulation of the P450 mRNA.
In summary, the observations made in the current study support the application of Fa2N4 cells as an in vitro model system to assess P450 induction. Moreover, the findings support use of a cocktail approach to assess the induction of CYP3A4, CYP1A2, and CYP2C9. Response to prototypical inducers did not differ to those observed in the presence of the specific probes alone. As such, this cassette approach reduces the sample numbers if screening the induction of more than one P450. Furthermore, the current approach allows for simultaneous collection of samples to measure changes in enzyme activities and the mRNA expression. In conclusion, the current experimental setup is an improvement on previous approaches used to assess P450 induction.
Acknowledgments
We thank Dr. Claudia Garcia-Allan and Dr. Maurice Dickins (Pfizer, PGRD, UK) and Dr. Odette Fahmi (Pfizer, PGRD, CT) for their help in the collection of data and preparation of the manuscript.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.106.012864.
-
ABBREVIATIONS: DDI, drug-drug interaction(s); NCE, new chemical entity; P450, cytochrome P450; PXR, pregnane X receptor; AhR, aryl hydrocarbon receptor; CAR constitutive androstane receptor; PBS, phosphate-buffered saline; KHB, Krebs-Henseleit bicarbonate; ANOVA, analysis of variance; CI, confidence interval(s).
- Received September 11, 2006.
- Accepted November 10, 2006.
- The American Society for Pharmacology and Experimental Therapeutics