Cooperativity in CYP2E1 Metabolism of Acetaminophen and Styrene Mixtures

Abstract

Risk assessment for exposure to mixtures of drugs and pollutants relies heavily on in vitro characterization of their bioactivation and/or metabolism individually and extrapolation to mixtures assuming no interaction. Herein, we demonstrated that in vitro CYP2E1 metabolic activation of acetaminophen and styrene mixtures could not be explained through the Michaelis-Menten mechanism or any models relying on that premise. As a baseline for mixture studies with styrene, steady-state analysis of acetaminophen oxidation revealed a biphasic kinetic profile that was best described by negative cooperativity (Hill coefficient = 0.72). The best-fit mechanism for this relationship involved two binding sites with differing affinities (K_s = 830 µM and K_ss = 32 mM). Introduction of styrene inhibited that reaction less than predicted by simple competition and thus provided evidence for a cooperative mechanism within the mixture. Likewise, acetaminophen acted through a mixed-type inhibition mechanism to impact styrene epoxidation. In this case, acetaminophen competed with styrene for CYP2E1 (K_i = 830 µM and K_si = 180 µM for catalytic and effector sites, respectively) and resulted in cooperative impacts on binding and catalysis. Based on modeling of in vivo clearance, cooperative interactions between acetaminophen and styrene resulted in profoundly increased styrene activation at low styrene exposure levels and therapeutic acetaminophen levels. Current Michaelis-Menten based toxicological models for mixtures such as styrene and acetaminophen would fail to detect this concentration-dependent relationship. Hence, future studies must assess the role of alternate CYP2E1 mechanisms in bioactivation of compounds to improve the accuracy of interpretations and predictions of toxicity.

Graphical abstract

1. Introduction

CYP2E1 plays a central role in the metabolic clearance of low molecular weight (<100 Da) hydrophobic compounds such as endogenous ketone bodies and a wide array of foreign, biologically active (xenobiotic) compounds.¹ The biochemical cost of recognizing a broad array of compounds is a loss of selectivity in which molecules bind to CYP2E1 and what happens during the oxidative catalytic cycle. CYP2E1 reactions may eliminate potentially toxic compounds and paradoxically, activate others into toxins or carcinogens. These metabolic processes contribute to toxicological responses associated with CYP2E1 activity toward drugs (acetaminophen), pollutants (aniline and styrene) and food/beverage constituents (ethanol). An understanding of the correlation between metabolism and toxicity requires knowledge of CYP2E1 metabolic efficiency toward substrates and mechanisms that influence that efficiency.

Localized to the endoplasmic reticulum, CYP2E1 was first identified from hepatic microsomal fractions by its high selectivity toward hydroxylation of 4-nitrophenol.² Kinetic profiles for CYP2E1 substrates are often hyperbolic and thus best fit to the Michaelis-Menten equation in which the metabolic efficiency is constant. Unlike traditional enzymes, a more complex array of mechanistic possibilities are possible for CYP2E1 based on recent steady-state studies on toxicologically important substrates including aniline,³ styrene,⁴ and m-xylene.⁵ The metabolism of those pollutants involved non-hyperbolic kinetic profiles, whereby metabolic efficiency either increased (positive cooperativity) or decreased (negative cooperativity or substrate inhibition) as a function of substrate concentration. Those kinetic profiles could be explained through mechanisms involving catalytic and effector sites for CYP2E1.^{3, 4, 6–11} Their presence and specificities were further characterized through experimental catalytic inhibition^{6, 8, 12} and binding^{6, 8} studies and computational docking and Molecular Dynamics simulations¹¹ using a series of mono- and bi-cyclic azoles. A hallmark of these studies has been the capacity of substrates alone and combined with inhibitors to impact CYP2E1 metabolic efficiency in ways unexpected by the traditional Michaelis-Menten mechanism.

In practice, humans are more likely exposed to mixtures than individual compounds, and thus we investigated the CYP2E1 activation of mixtures of a common drug (acetaminophen) and pollutant (styrene). First, acetaminophen is the most widely used analgesic/antipyretic drug in the world and is generally safe at therapeutic doses. Hepatotoxicity associated with acetaminophen occurs under conditions that favor a minor bioactivation pathway carried out mainly by CYP2E1.^{13, 14} This enzyme catalyzes a two-electron oxidation to a reactive and toxic N-acetyl-p-benzoquinone imine (NAPQI)¹⁵ and contributes to oxidative stress.¹⁴ Second, styrene is a common industrial chemical listed on the 2014 Agency for Toxic Substances and Disease Registry Substance Priority List for Superfund Sites and classified as a class 2B carcinogen.^16–18 In the environment, styrene is also found at low levels from cigarette smoke, diesel exhaust, and degradation of polystyrene food containers. Styrene undergoes metabolism primarily by CYP2E1 into styrene oxide, a reactive and toxic epoxide metabolite. Given their prevalence, it is highly probable that humans undergo exposure to mixtures of acetaminophen and styrene that may influence their potential to induce toxicological outcomes.

Herein, we assessed how mixtures of acetaminophen and styrene impacted their ability to influence CYP2E1 metabolic efficiency and subsequently activation into reactive and toxic metabolites as a biochemical model for their interactions in the environment. We initially determined the in vitro inhibitory potency of acetaminophen and styrene on their respective reactions through standard IC₅₀ experiments using 7-methoxy-4-trifluoromethylcoumarin (7MFC) as a common reporter substrate.¹⁹ For more detailed studies, we carried out metabolic studies on acetaminophen using recombinant CYP2E1 Supersomes and pooled human liver microsomes (HLM150) due to conflicting reports on the mechanism and corresponding constants in the literature.^20–23 The kinetic profile for the recombinant CYP2E1 reaction was biphasic and indicated negative cooperativity involving two binding sites. Low substrate affinity prevented saturation of the enzyme, and thus detailed kinetic studies were not possible for assessing the impact of styrene on acetaminophen oxidation. We then limited our investigation to determine whether styrene simply competed for acetaminophen during metabolism or interacted through a more complex mechanism. By contrast, CYP2E1 displayed a much higher affinity for styrene such that we could effectively assess the impact of acetaminophen on styrene activation. The resulting data were global fitting to six possible mechanisms involving a single binding site (Michaelis-Menten) or two binding sites and the most statistically probable one identified using DynaFit software²⁴ as described.^{3, 4, 6} Lastly, we assessed the potential impact of acetaminophen on the predicted in vivo metabolic activation of styrene by CYP2E1.

2. Materials and Methods

2.1 Reagents

All chemicals were ACS grade or higher, purchased from Sigma Aldrich (St Louis, MO): styrene (substrate), styrene oxide (product), styrene glycol (product), benzyl alcohol (internal standard), acetaminophen (substrate), and 2-acetoamidophenol (internal standard). Barium hydroxide, zinc sulfate, 1,4-dioxane, HPLC grade methanol and HPLC grade acetonitrile were purchased from Fisher Scientific, Wilmington, MA. The acetaminophen-glutathione (APAP-GSH) and acetaminophen-cysteine (APAP-cys) adducts were synthesized in-house. For purification of adducts, solid-phase extraction Oasis MAX columns were obtained from Waters Corp. (Milford, MA). NAPQI solid was provided as a generous gift from Drs. Dean Roberts and Laura James at Arkansas Children’s Hospital (Little Rock, AR).

2.2 Inhibitory potency of acetaminophen and styrene toward CYP2E1 activity

We estimated the relative binding affinity of acetaminophen and styrene toward CYP2E1 through IC₅₀ studies using the fluorescent marker substrate 7-methoxy-4-trifluoromethylcoumarin (7MFC). For reactions, 25 nM recombinant CYP2E1 (Corning Gentest Supersomes®) was incubated with 25 µM 7MFC and seven concentrations of each inhibitor (plus a negative control containing the inhibitor vehicle only) in 50 mM potassium phosphate buffer, pH 7.4, at 37°C. Inhibitor stocks were prepared in methanol due to the solubility of the substrate and their use in reactions resulted in 0.5% methanol (final), which has been shown to have minimal effect on CYP2E1 activity at that concentration.²⁵ Reaction mixtures were prepared in 96 half-well flat-bottomed plates sealed with ThermoWell Sealing Tape. After pre-incubating at room temperature for 5 min, reactions were initiated with the addition of 1 mM NADPH. After 30 min, the reactions were transferred to a 96-well round-bottom black plate for fluorescence analyses and quenched by addition of an equal volume of ice-cold acetonitrile.

Quenched reactions were then analyzed using a PerkinElmer Victor³V Multilabel Plate Reader with excitation at 410 nm and emission at 510 nm. The resulting rates of product formation were then expressed as a percent of the vehicle-only control rate. Under these experimental conditions, product formation was linear with respect to time and no detectable product formed in the absence of NADPH (data not shown). The data from inhibition studies were normalized to the uninhibited reaction (i.e. 100% activity), plotted as a function of the log inhibitor concentration, and fit to the standard IC₅₀ equation in GraphPad Prism 5.0 (La Jolla, CA). During the fit, the “end point” of each IC₅₀ curve was allowed to vary (above 0%) and the optimal end point was interpreted as residual activity. Each inhibitor was tested with at least three independent experimental replicates.

2.3 Steady-state acetaminophen oxidation studies

We measured steady-state oxidation of acetaminophen by CYP2E1 to its major oxidative metabolite, N-acetyl-p-benzoquinone imine (NAPQI) by trapping the reactive metabolite with cysteine to form a conjugate (APAP-cys). Reaction mixtures contained either 50 nM recombinant CYP2E1 (Corning Gentest Supersomes®) or 0.5 mg/mL pooled human liver microsomes (Corning Gentest UltraPool® HLM150), 12 acetaminophen concentrations between 5 µM and 10 mM, and 2 mM cysteine in 50 mM potassium phosphate buffer, pH 7.4. The experiments were carried out for each substrate concentration in three to six experimental replicates. Reactions were initiated upon the addition of 1 mM NADPH and quenched after 30 min with 0.4 N perchloric acid containing 1 µM 2-acetoamidophenol (internal standard). Excess cysteine present in the reaction was assumed to effectively trap NAPQI and thus the APAP-cys conjugate was used as a measure of total NAPQI formation. The quenched mixture was centrifuged to pellet precipitated proteins and the resulting supernatant was subjected to HPLC analysis to quantify APAP-cys formation. APAP-cys was not detected in reactions lacking NADPH.

2.4 Synthesis and purification of APAP-cys conjugate

APAP-cys was synthesized with modifications to previously published methods.^{26, 27} 0.1 M NAPQI dissolved in 1,4-dioxane was added in a 1:400 dilution to a 2.5 mM solution of L-cysteine in 200 mM triethylammonium buffer containing saline (TEAB/saline) and allowed to stir for 5 min. Meanwhile, an Oasis Max mixed-mode solid-phase extraction column was conditioned with methanol followed by 200 mM TEAB/saline. After loading the reaction mixture onto the column, the column was washed with 200 mM TEAB/water, methanol to remove neutral compounds (such as acetaminophen and phenolics), and finally 1:1 methanol and water to remove TEAB buffer. Free cysteine was then removed by washing with 2% formic acid. Finally, the APAP-cys was eluted with a 1:1 mixture of methanol and 2% formic acid. The eluent was dried down completely in a Savant Speed Vac Concentrator and the dry sample was stored dessicated under argon at −80°C. The purity was estimated by HPLC/UV-vis absorbance at 254 nm to be >98%.

2.5 HPLC-ECD analysis of acetaminophen reactions

HPLC analysis of APAP-cys adducts was carried out as described previously.²⁸ The analysis was performed at room temperature under isocratic conditions using models 582 solvent delivery system, 5600A CoulArray detector and 540 autosampler from ESA (Chelmsford, MA) for automated injection of samples (50 µL). The mobile phase consisted of 50 mM sodium acetate, pH 4.8 with 8% methanol at a flow rate of 1.2 mL/min. Acetaminophen, APAP-cys, and the internal standard 2-acetoamidophenol were resolved using a Symmetry C18 5um 4.6 × 150mm column from Waters Corp. (Milford, MA). After a 20 min run time, peak areas for APAP-cys and APAP from the chromatogram were normalized to internal standard and quantified relative to known standards.

2.6 Mechanism for acetaminophen oxidation under steady-state conditions

The mechanism for acetaminophen oxidation was analyzed through a two-staged approach. First, we employed the traditional approach pursued by others in the field; kinetic data were fit to the Michaelis-Menten (Equation 1) and Hill equation (Equation 2) and the best fit determined by Akaike Information Criterion as implemented by GraphPad Prism 5.0 (La Jolla, CA). In these equations, ν represents the rate of the reaction, while [APAP] is the concentration of acetaminophen, V_max is the maximal rate and K_m is the substrate concentration required to reach a half-maximal rate. In the Hill equation, the coefficient “n” represents the Hill coefficient.

$$\nu =\frac{{\mathrm{v}}_{\text{max}}[\text{APAP}]}{{\mathrm{K}}_{\mathrm{m}}+[\text{APAP}]}$$ (1)

$$\nu =\frac{{\mathrm{v}}_{\text{max}}{[\text{APAN}]}^{\mathrm{n}}}{{\mathrm{S}}_{50}^{\mathrm{n}}+{[\text{APAP}]}^{\mathrm{n}}}$$ (2)

Second, we investigated more detailed mechanisms involving one (Michaelis-Menten) or two binding site models in which a binary enzyme-substrate (ES) complex, a ternary enzyme-substrate-substrate (ESS) complex, or both complexes are catalytically active (Fig. 1) as described previously.^{3, 4, 6} These binding events were ordered resulting in the observed stoichiometry of CYP2E1 complexes; however, the role of specific sites as catalytic and/or effector sites cannot be distinguished through these types of studies. Steady-state kinetic data were fit to all mechanisms and analyzed globally using the statistical software Dynafit version 3.28 (Biokin Ltd., Watertown, MA) that identified the most probable kinetic mechanism and corresponding parameters for acetaminophen oxidation.²⁴ For this analysis, the reaction rates served as input and all equilibrium constants and kinetic constants were allowed to vary (including K_s, K_ss, k_cat1, and k_cat2, where applicable).

Fig. 1.

Steady-state metabolic reaction mechanisms for CYP2E1 oxidation of acetaminophen; E = enzyme, S = acetaminophen, P = product (NAPQI)

2.7 Impact of styrene on acetaminophen oxidation

The inability to saturate acetaminophen oxidation precluded a detailed analysis of the impact of styrene on its oxidative metabolism; however, we were able to gain insights on their mechanism of interaction. As reported in the Results section, substrate affinity for recombinant CYP2E1 was low enough to prevent saturation of the enzyme, such that the interpretation of inhibitory studies with styrene would not be possible. Complementary studies with human liver microsomes would also not be feasible due to similar low affinity for CYP2E1 and possible contributions from CYP1A2 and 3A4. Nevertheless, it was possible to investigate whether styrene impacted acetaminophen metabolism by recombinant CYP2E1 through a single or two-binding site mechanism. For these studies, 100 µM acetaminophen was incubated with 50 nM recombinant CYP2E1 Supersomes®, 2 mM cysteine, and styrene at one of five concentrations (0, 50, 100, 150, and 500 µM). Each experiment was carried out in three to seven experimental replicates. Styrene stocks were prepared in methanol and final concentrations of methanol in reactions including in the no-styrene control were 0.5% (final). The reaction was initiated with 1 mM NADPH and carried out as described for the steady-state acetaminophen experiments. For comparison, the predicted inhibition for simple competitive inhibition was calculated using the IC₅₀ values determined previously using the 7MFC inhibition experiments. Because the concentration of 7MFC used in those experiments was approximately equal to the K_m, the IC₅₀ should then approximate the K_i. Under those conditions, the traditional equation for competitive inhibition (Equation 3) was applied to estimate how much inhibition would be observed. In this equation, [APAP] represents the concentration of acetaminophen and [STY] represents the concentration of styrene. The 95% confidence intervals for the nonlinear fit of the IC₅₀ equation for acetaminophen were used to give a confidence range for the % APAP (acetaminophen) Activity predicted using this method. A one-way ANOVA test was then applied to compare the measured activity to the observed activity.

$$\%\mathit{\text{ APAP Activity}}=\frac{{\mathit{\text{IC}}}_{50,\mathit{\text{APAP}}}+[\mathit{\text{APAP}}]}{{\mathit{\text{IC}}}_{50,\mathit{\text{APAP}}}(1+\frac{[\mathit{\text{STY}}]}{{\mathit{\text{IC}}}_{50,\mathit{\text{STY}}}})+[\mathit{\text{APAP}}]}$$ (3)

2.8 Impact of acetaminophen on styrene epoxidation

CYP2E1 is relatively efficient at styrene epoxidation and thus provides an adequate dynamic range in activity to assess the impact of acetaminophen on styrene bioactivation from mixtures of the substrates. These studies were carried out with pooled human liver microsomes (HLM150), as an approximation of the average metabolism by an adult toward mixtures. We have previously shown that CYP2E1 dominates styrene epoxidation except at very high concentrations not included in this study.⁴ For these experiments, 0.5 mg/mL HLM150 was incubated with styrene (eight concentrations ranging from 10 to 350 µM) and acetaminophen (control plus four concentrations at 0, 1, 10, 100, and 1000 µM) with 2 mM cysteine in 50 mM potassium phosphate buffer, pH 7.4 at 37°C. Reactions were initiated with 1 mM NADPH and quenched after 30 min using barium sulfate precipitation, as described previously.⁴ Specifically, equal volumes of the reaction, saturated barium sulfate, and 10% zinc sulfate containing the internal standard (10 µM benzyl alcohol) were mixed together to form a fluffy precipitate. Under these conditions, all styrene oxide was converted to styrene glycol and no detectable styrene glycol formed in the absence of NADPH (data not shown). Mixtures were centrifuged and the resulting supernatant transferred to HPLC vials for analysis.

2.9 HPLC-UV/Vis analysis of styrene reactions

The styrene epoxidation was analyzed by our previously described HPLC method.⁴ Briefly, analytes from quenched reactions (styrene, styrene glycol, and the internal standard benzyl alcohol) were resolved using a 4.6 × 150 mm Zorbax 3.5 µm XDB-C18 column (Agilent, Santa Clara, CA) heated to 45°C using a Waters Breeze HPLC System (Waters, Milford, MA) and gradient of 15% to 100% acetonitrile at a flow rate of 1.2 mL/min. Styrene glycol UV absorbance was monitored at 200nm and peak areas were normalized to internal standard and the concentrations determined based on comparison to a standard curve.

2.10 Mechanism for the effect of acetaminophen on styrene epoxidation

The expansion of CYP2E1 metabolic mechanisms to two-site models creates a wide variety of possible mechanisms for mixtures of two compounds to influence metabolism. Data for styrene (S) reactions in the presence of acetaminophen (I) were fit globally and simultaneously to 10 different mechanisms in which acetaminophen bound to one or two sites on CYP2E1 using Dynafit version 3.28 (Biokin, Ltd.), as described previously (Fig. 2). CYP2E1 epoxidation of styrene involves positive cooperativity,⁴ and thus all mechanisms included styrene binding at two sites leading to formation of binary (ES) and ternary (ESS) complexes along with the possibility of a mixed, catalytically active ternary ESI complex.

Fig. 2.

One- and two- binding site inhibition mechanisms for acetaminophen modulation of styrene metabolism by CYP2E1. The uninhibited styrene reaction is shown in bold; possible, but undetectable complexes are shown in grey. E = enzyme, S = styrene, P = product (styrene oxide), I = inhibitor (acetaminophen).

The mechanisms were similar to the traditional inhibition mechanisms, and therefore we used similar nomenclature to describe them. For the competitive inhibition model, the inhibitor could only bind free enzyme, while inhibitor in the uncompetitive inhibition model interacted exclusively with the enzyme-substrate (ES) complex. For the mixed-type mechanisms, inhibitor bound to both free enzyme and the ES complex with the same or different inhibition constants. Finally, an important limitation of this study is that the EII and EIS complexes (shown in grey in Fig. 2) were not measurable using the catalytic inhibition assay, and therefore the K_i,ap and K_{si, ap} were designated as apparent parameters that may include contributions from multiple other complexes. During the analysis, the associated parameters from the uninhibited styrene reaction (K_s, K_ss, k_cat) were held constant while inhibitor constants K_si,ap and k_cat2, where applicable, were allowed to vary. Finally, the inhibitor constant K_i,ap, represented the equilibrium constant for formation of the enzyme-acetaminophen complex and thus was assumed to be equivalent to the K_s value for acetaminophen determined in the steady-state experiment to minimize variables in this analysis.

2.11 Modeling the impact of acetaminophen on styrene metabolic clearance

A readily accessible and powerful approach to modeling metabolic clearance of drugs and pollutants utilizes kinetic parameters obtained through in vitro studies.^{29, 30} As described previously,^{3, 30} we fit our data to mechanistic equations describing the relationship between styrene concentration and its clearance (rate/[styrene]) in the presence and absence of 100 µM acetaminophen, which is within the range of typical maximum plasma concentrations reached after a single therapeutic dose (5–20 mg/L or 33–132 µM).^{31, 32} The metabolic clearance of styrene based on the traditional model was predicted using the parameters derived from fitting the human liver microsomal CYP2E1 data to the Michaelis-Menten equation and applying the results to Equation 4.

$$\frac{\mathrm{\nu }}{[\text{styrene}]}=\frac{{\mathrm{V}}_{\text{max}}}{{\mathrm{K}}_{\mathrm{m}}+[\text{styrene}]}$$ (4)

where ν is the initial rate of the reaction, V_max the maximal rate of the reaction, and K_m the Michaelis constant. For comparison, the impact of acetaminophen on styrene clearance was predicted assuming simple competition using a derivation of the Michaelis-Menten equation for competitive inhibition shown in Equation 5. The IC₅₀ obtained for acetaminophen was used as an approximation of K_i, which is reasonable given the substrate concentration in the IC₅₀ experiment was approximately equal to the K_m. As in the Michaelis-Menten equation, ν is the initial reaction rate, V_max is the maximal reaction rate, K_m is the substrate concentration at half-maximal reaction rate, and [APAP] and [styrene] are concentrations of acetaminophen and styrene, respectively.

$$\frac{\mathrm{\nu }}{[\text{styrene}]}=\frac{{\mathrm{V}}_{\text{max}}}{{\mathrm{K}}_{\mathrm{m}}(1+\frac{[\text{APAP}]}{{\text{IC}}_{50,\text{APAP}}})+[\text{styrene}]}$$ (5)

Contrasting with the simple models, the impact of cooperativity on the metabolic clearance of styrene by CYP2E1 was predicted using Equation 4 and parameters derived from fitting the microsomal CYP2E1 kinetic data to the Hill equation.

$$\frac{\mathrm{\nu }}{[\text{styrene}]}=\frac{{\mathrm{V}}_{\text{max}}\times {[\text{styrene}]}^{\mathrm{n}-1}}{{\mathrm{S}}_{50}^{\mathrm{n}}+{[\text{styrene}]}^{\mathrm{n}}}$$ (6)

where ν is the initial rate of the reaction, V_max the maximal rate of the reaction, n the Hill coefficient, and S₅₀ the concentration of substrate at half maximal rate for the reaction.

3. Results

3.1 More potent inhibition of CYP2E1 activity by styrene versus acetaminophen

We determined the potency of styrene and acetaminophen inhibition of CYP2E1 oxidation of 7MFC using IC₅₀ studies as an initial measure of their interactions (Fig. 3). Styrene was a relatively potent inhibitor (IC₅₀ = 22 µM; 95% confidence interval, 17 to 28) compared to acetaminophen-mediated inhibition (IC₅₀ = 1300 µM; 95% confidence interval, 450 to 2200). The residual activity (R.A.) was also higher with acetaminophen (R.A. = 30%; 95% confidence interval, 15 to 45%) compared to styrene (R.A. = 19%; 95% confidence interval, 16 to 23%). These end points indicate an incomplete inhibition of 7MFC turnover even at very high acetaminophen concentrations.

Fig. 3.

IC₅₀ plot for acetaminophen and styrene with recombinant CYP2E1 enzyme. Nonlinear plots represent the IC₅₀ nonlinear regression for the normalized activity versus the log of the inhibitor concentration. Individual tracings represent different molecules, specifically acetaminophen (squares) and styrene (circles). Reactions were performed at least six times at 37°C in 50 mM potassium phosphate buffer, pH 7.4. Further reaction conditions are described in Materials and Methods.

3.2 Biphasic steady-state oxidation of acetaminophen by CYP2E1

The oxidation of acetaminophen by recombinant CYP2E1 led to a significantly biphasic, non-hyperbolic kinetic profile (Fig. 4). A preliminary comparison of the fit of the data to the Michaelis-Menten versus Hill equations revealed a strong (>99.99%) preference for the Hill equation. The associated parameters were a Hill coefficient of 0.72 (95% confidence interval, 0.68 to 0.77), indicating negative cooperativity, with a Hill constant of 6900 µM (95% confidence interval, 4700 to 9000 µM) and a maximal rate of 3.8 nmol/min/nmol CYP2E1 (95% confidence interval, 3.3 to 4.2 nmol/min/nmol CYP2E1). Although the Hill equation does allow for distinction between simple and more complex kinetics, it does not provide the mechanism underlying the deviation from a hyperbolic profile.

Fig. 4.

Steady-state kinetics of acetaminophen oxidation by recombinant CYP2E1 and human liver microsomes (inset). For the recombinant enzyme, the solid line represents the fit of the data to the statistically preferred mechanism based on Akaike Information Criterion using DynaFit version 3.28. For the HLMs, due to the contribution of multiple P450s a mechanistic interpretation was not pursued further. For reactions, 50nM recombinant CYP2E1 or 0.5 mg/mL human liver microsomes, acetaminophen (varied from 5 µM to 10 mM), 2 mM cysteine, and 1 mM NADPH were incubated at 37°C in 50 mM potassium phosphate buffer, pH 7.4. The reported values represent the average of at least four experimental replicates.

A more in-depth mechanistic comparison of the steady-state data fit to either the Michaelis-Menten mechanism (single site) or multiple mechanisms involving binding at two sites revealed a strong statistical preference for a cooperative mechanism (>99.99% based on Akaike Information Criterion).³³ The model included formation of catalytically active binary (ES) and ternary (ESS) complexes possessing different catalytic efficiencies (Fig. 1). The first binding event (K_s = 860 µM; 95% confidence interval, 610 to 1100) was more favorable than the second binding event (K_ss = 32 mM, 85% confidence interval, 15.6 to 730 mM) based on the fit parameters. The large equilibrium constant for the second binding event precluded saturation of that site due to acetaminophen solubility limits and therefore, the confidence intervals for the fit had to be set to 85%. Nevertheless, binding at the second site resulted in an overall higher maximal rate for the ternary complex (k_cat2 = 5.6 nmol/min/mg protein; 85% confidence interval, 4.0 to 68) compared to the binary complex (k_cat1 = 1.3 nmol/min/mg protein; 95% confidence interval 1.0 to 1.6). The concentration-dependent shift between these two catalytic efficiencies resulted in the observed biphasic kinetic profile. The oxidation of acetaminophen by human liver microsomes also resulted in a multiphasic profile (Fig. 4, inset) but due to the possible contribution of multiple P450s, the mechanism of oxidation of acetaminophen by CYP2E1 in human liver microsomes was not explored further.

3.3 Cooperative styrene effects on acetaminophen oxidation

Due to conflicting results in the literature, we initially determined the kinetic mechanism and constants for acetaminophen oxidation in the absence of styrene and then followed up with studies on mixtures of the two CYP2E1 substrates. Based on studies with recombinant CYP2E1, the steady-state CYP2E1 mechanism for acetaminophen metabolism involved initial formation of a high-affinity, low-activity binary complex and then at higher acetaminophen concentrations, a shift to a loweraffinity, higher-activity ternary complex that did not saturate even at high acetaminophen concentrations (10 mM). The weakness of the substrate interactions with CYP2E1 precluded detailed studies with styrene; however, it was possible to determine whether styrene impacted acetaminophen metabolism through a single or two binding site mechanism for recombinant CYP2E1. As shown in Fig. 5, the observed CYP2E1 oxidation of acetaminophen at 100 µM in the presence of increasing concentrations of styrene differed from the effect predicted by simple competition. At every concentration of styrene in the reaction, the remaining activity was greater than would be predicted by simple competition and thus styrene impacts acetaminophen oxidation by CYP2E1 through cooperative mechanism. Unfortunately, the confounding effects of other P450s on acetaminophen oxidation precluded a complementary and interpretable study using human liver microsomes.

Fig. 5.

Impact of styrene on acetaminophen metabolism. For reactions, 50 nM CYP2E1, 100 µM acetaminophen, styrene (varied from 50 to 500 µM), 2 mM cysteine and 1 mM NADPH were incubated at 37°C in 50 mM potassium phosphate buffer, pH 7.4. The reported values (“actual”) represent the average of at least four experimental replicates. The predicted values were calculated based on the expected rate for 100 µM acetaminophen with the various concentrations of styrene assuming simple competition. Further details are provided in Materials and Methods. Asterisks indicate statistical significance; ***, p<0.001.

3.4 Cooperative interaction of acetaminophen during styrene oxidation

We followed up studies on acetaminophen metabolism from mixtures of acetaminophen and styrene with those assessing styrene epoxidation using human liver microsomes. Human liver microsomes provide a suitable model for CYP2E1 interactions in vivo, and we have previously demonstrated that the concentrations used here reflect exclusively CYP2E1 contributions.⁴ The addition of acetaminophen increased the rate of styrene turnover at lower styrene concentrations and decreased those rates at higher styrene concentrations, leading to changes in the shape of the kinetic profile. The shift in the overall profile, especially at higher styrene concentrations, was more dramatic as acetaminophen concentrations increased in the mixed reaction. Among ten possible mechanisms to explain the data, there was significant statistical support for a mixed-type inhibition mechanism wherein acetaminophen interacted with free enzyme or the enzyme substrate-complex to form a binary (EI) or mixed ternary (ESI) complex. The mixed complex was catalytically active for styrene epoxidation but the observed rate differed from that observed in the absence of acetaminophen. All other models had relatively higher AIC values, i.e. ΔAIC_c >>2, and thus are less likely to explain the data. A plot of the data along with the fit from the statistically preferred model is shown in Fig. 6.

Fig. 6.

Impact of acetaminophen on steady-state styrene metabolism. For reactions, 50 nM CYP2E1, styrene (varied from 10 to 350 µM), acetaminophen (varied from 1 to 1000 µM), 2 mM cysteine and 1 mM NADPH were incubated at 37°C in 50 mM potassium phosphate buffer, pH 7.4. The reported values (“actual”) represent the average of at least four experimental replicates. Individual tracings represent different acetaminophen concentrations with increasing concentration represented by increasingly darker shading. The uninhibited styrene reaction is shown with open circles and a dashed line for the fit to the steady-state positive cooperative mechanism. The solid lines represent the fit to the most probable mechanism determined using DynaFit statistical software.

The preferred CYP2E1 mechanism indicated cooperativity between acetaminophen and styrene during binding and catalysis. The presence of styrene in the active site improved the affinity toward acetaminophen; the acetaminophen equilibrium constant for the binary enzyme-styrene complex (K_si,ap = 180 µM; 95% confidence interval, 98 to 420) was about three-fold less than that for the enzyme alone (K_i = K_s = 830 µM). Compared to styrene (K_s=375 µM),⁴ acetaminophen bound the free enzyme with two-fold lower affinity. Similarly, acetaminophen bound four-fold weaker than styrene to the binary enzyme-styrene complex (K_ss=39 µM).⁴ Once bound, acetaminophen caused a one-third drop in the rate of catalysis (k_cat2 = 0.47 nmol/min/mg protein; 95% confidence interval, 0.39 to 0.61) for mixed (ESI) ternary complex when compared to the turnover rate for both binary (ES) and ternary (ESS) styrene complexes at 0.7 nmol/min/nmol.

3.5 Impact of acetaminophen on cooperative styrene metabolic clearance

Styrene epoxidation is preferentially catalyzed through a cooperative mechanism; however, we employed the kinetic parameters from fits of styrene reaction data with and without acetaminophen to either the Michaelis-Menten or cooperative mechanisms in order to assess the effects of alternate mechanisms on styrene metabolic clearance. A fit of the microsomal kinetic profile for styrene to the Michaelis-Menten equation yielded V_max and K_m equal to 1.1 nmol/min/mg protein (0.96 to 1.3, 95% confidence interval) and 210 µM (150 to 260, 95% confidence interval), respectively. These values were used to simulate a clearance curve using Equation 2 (solid line in Fig. 7, panel A). We then modeled if acetaminophen had acted like a simple competitor by calculating the apparent Michaelis-Menten parameters for steady-state oxidation of styrene using the IC₅₀ for acetaminophen to predict styrene clearance in the presence of 100 µM acetaminophen using Equation 3 (dashed line in Fig. 7, panel A). For the preferred cooperative mechanism, values from the fit of the data to the Hill equation were used to plot a curve based on Equation 4 (solid line in Fig. 7, Panel B). The reliance on traditional Michaelis-Menten based models resulted in log plots with little difference in predicted metabolic clearance of styrene in the presence or absence of acetaminophen. By contrast, the use of the statistically preferred cooperative mechanisms yielded significantly different clearance plots for styrene when styrene was metabolically activated in the presence of acetaminophen especially at low exposure levels.

Fig. 7.

Styrene clearance in the absence (solid line) or presence (dashed line) of acetaminophen considering either simple competition (panel A) or cooperative (panel B) kinetics. GraphPad Prism software (San Diego, CA) was used to simulate clearance curves for styrene based on microsomal kinetic parameters for the respective mechanisms and equations published by Houston and Kenworthy³⁰.

4. Discussion

Risk assessment for exposure to mixtures of drugs and pollutants relies heavily on in vitro characterization of their bioactivation and/or metabolism individually and extrapolation to mixtures assuming no interaction. In this study, we demonstrated that in vitro CYP2E1 metabolic activation of mixtures between acetaminophen and styrene involved changes in metabolic efficiencies for reactions that could not be explained through the Michaelis-Menten mechanism or any models relying on that premise. As a baseline for mixture studies with styrene, steady-state analysis of acetaminophen oxidation revealed a biphasic kinetic profile that was best described by a negative cooperative (two-site) mechanism. The introduction of styrene inhibited that reaction but to a lesser extent than would be predicted by simple competition as evidence for a cooperative mechanism of turnover even with mixtures. Likewise, during the positively cooperative steady-state epoxidation of styrene, addition of acetaminophen modulated styrene bioactivation to a genotoxic metabolite styrene oxide that was best fit to a mixed-type inhibition mechanism. In this case, acetaminophen competed with styrene for either the catalytic or effector site on CYP2E1 and resulted in cooperative impacts on binding and catalysis. The CYP2E1 mechanism for metabolism of mixtures revealed profound effects of acetaminophen on the predicted in vivo metabolic activation of styrene that would not be anticipated given the reliance on traditional models involving Michaelis-Menten kinetics.

Given its prevalent use, acetaminophen has been the subject of numerous in vitro metabolic studies to assess potential health risks, yet there is no consistent kinetic data for CYP2E1 oxidation of acetaminophen to the reactive quinoneimine metabolite NAPQI. Kinetic profiles from those studies were fit only to Michaelis-Menten mechanism and reported different affinities (K_m) between CYP2E1 and acetaminophen.^20–22 We broaden the analysis of our data to include multiple possible mechanisms and identified negative cooperativity as the most statistically plausible one. This mechanism indicated a decrease in metabolic efficiency as CYP2E1 shifted from a catalytically active form to a less active one at higher substrate concentrations. Based on our previous experimental^{3, 4, 6–8, 12} and computational studies,^{6, 11} we explained different functional forms of CYP2E1 through concentration-dependent occupancy of two binding sites that modulated the rate of turnover (Fig. 8, blue circles). This negatively cooperative mechanism for acetaminophen was also observed for CYP2E1 hydroxylation of aniline.³ This observation may reflect similar structures for aniline and acetaminophen that favor common binding interactions between an amino substituted phenyl ring and CYP2E1 binding sites. The presence of the acetyl group and/or phenolic group for acetaminophen significantly perturbed CYP2E1 binding based on an over 10-fold weaker affinity for CYP2E1 compared to aniline.³

Fig. 8.

Stylized depiction of the CYP2E1 cooperative mechanism for styrene and acetaminophen interaction in a binary mixture. Catalytic cycles highlighted in blue represent the homotropic binary (ES) and ternary (ESS) complexes formed during acetaminophen metabolism alone. Catalytic cycles highlighted in blue represent the catalytic cycles formed during styrene metabolism alone. The purple-highlighted catalytic cycles are when product is formed while styrene and acetaminophen are simultaneously bound to CYP2E1.

Weak affinity between CYP2E1 and acetaminophen is a common observation among our studies and others;^20–23 however, there are significance differences among them. The first binding event for acetaminophen in our studies had a dissociation constant of 860 µM, when employing CYP2E1 Supersomes ® (Corning Gentest, Corning, NY). By contrast, a recent study using the same enzyme system did not report any binding interactions between CYP2E1 and acetaminophen, because the kinetic profile was linear kinetics up to 1 mM substrate concentration indicating a lack of enzyme saturation.²³ Nevertheless, the rates at high substrate concentration exhibited very high error bars that compromised the analysis of the data. Our studies lack such uncertainties and included a much larger assessment of initial rates at different acetaminophen concentrations. These experimental properties allowed us to better define the kinetic profile and accurately measure kinetics for CYP2E1 metabolism of acetaminophen. Moreover, our results are more consistent with those reported by other groups using different CYP2E1 preparations. Human CYP2E1 heterologously expressed in HepG2 cells or purified CYP2E1 in a reconstituted system displayed similar K_m values for acetaminophen of ~600 µM)^{20, 21} and, a later study involving a similarly reconstituted CYP2E1 system reported a K_m of 1.3 mM.²² Consequently, steady-state CYP2E1 metabolism of acetaminophen is saturable under in vitro conditions, yet our more detailed analysis revealed a more complex mechanism of action that explained the metabolism of mixtures in this study.

A negatively cooperative mechanism critically impacts the effects of compound mixtures on activity in ways inconsistent with the traditional competitive mechanism predicted by the Michaelis-Menten model. We demonstrated that the presence of styrene at multiple concentrations inhibited acetaminophen bioactivation yet the effects were muted compared to that predicted by simple competition. Styrene acted as a mixed inhibitor indicating styrene mediates its concentration-dependent effects on acetaminophen oxidation through more than one active form of CYP2E1 as we proposed in our mechanism for acetaminophen metabolism (Fig. 8, blue and purple circles). In this case, styrene interacts with CYP2E1 like acetaminophen but with different outcomes on activity. Unfortunately, we were not able to fully saturate the second functional form of CYP2E1 during acetaminophen oxidation and thus not able to carry out an in depth analysis to determine actual mechanistic constants describing the effects of styrene on acetaminophen metabolism. Despite this shortcoming, our evidence clearly showed that styrene impacts acetaminophen activation to the toxic NAPQI metabolite through a cooperative mechanism, such that the traditional reliance on Michaelis-Menten models in toxicity estimates will underestimate the influence of alternate, competing substrates on CYP2E1 oxidation of acetaminophen.

Unlike acetaminophen, styrene undergoes efficient metabolic activation by CYP2E1 through a positively cooperative mechanism (Fig. 8, red circles),^{4, 7} and thus it was possible to fully interrogate the impact of acetaminophen on styrene conversion to a reactive epoxide metabolite. Acetaminophen acted as a mixed-type inhibitor on the styrene reaction, which was similar to inhibition mechanisms reported for many other five- and six- member aromatic and heterocyclic compounds.^{4, 7, 8, 12} In particular, there were many similarities between the effects of acetaminophen on styrene epoxidation to that by 4-methylpyrazole.⁴ First, acetaminophen bound to the binary (enzyme-styrene) complex, yet CYP2E1 remained active toward styrene, albeit with a slightly lower rate (Fig. 8, red and blue circles). Second, the binding of styrene to CYP2E1 increased affinity toward acetaminophen. However, an important distinction in this comparison is that 4-methylpyrazole is strictly a CYP2E1 inhibitor whereas acetaminophen is an alternate substrate of CYP2E1. Any effects on acetaminophen interactions with CYP2E1 during styrene metabolism would be reflected in the metabolic activation of acetaminophen. The increased affinity toward acetaminophen could have improved its rate of turnover resulting in more efficient activation to NAPQI as observed in our previous experiments assessing acetaminophen oxidation in mixtures with styrene.

The alteration of styrene epoxidation by acetaminophen could reflect inactivation of CYP2E1 during the course of the reaction. Oxidation of acetaminophen yields a highly reactive electrophile, NAPQI, which is known to readily from protein adducts. The covalent modification of CYP2E1 could reduce its observed activity and contribute to the kinetic profile for styrene epoxidation in the presence of acetaminophen as a competing substrate. Nevertheless,, it has been previously shown that CYP2E1 oxidation of acetaminophen does not impact its activity for as long as 30 minutes.²² This reaction window encompasses the time frame for our experimental studies suggesting CYP2E1 remains active over the course of the reactions. In addition, all experiments carried out in this study included an excess quantity of the nucleophile cysteine; those conditions for the NAPQI trap would further minimize the likelihood of any protein modification and subsequent impact on observed rates of turnover. Therefore, it is unlikely that any irreversible inhibition by NAPQI impacted the results for the metabolism of single and combined mixtures of acetaminophen and styrene.

We revealed the potential in vivo implications of the cooperative interactions between acetaminophen and styrene by predicting the intrinsic clearance of styrene by CYP2E1. In the absence of acetaminophen, styrene clearance was much lower than predicted by Michaelis-Menten kinetics at low styrene concentrations, as reported previously.⁴ Under those same conditions, the presence of a therapeutic level of 100 µM acetaminophen induced the most pronounced effects on styrene metabolic clearance as reflected in a dramatic increase in the efficiency of styrene epoxidation. The impact of these differences on the relative toxicity of styrene could be significant. Styrene is bioactivated by CYP2E1 to a potentially genotoxic epoxide metabolite, styrene oxide, and therefore acetaminophen may increase mutagenicity of styrene at low styrene exposures. This scenario is highly probable given the higher likelihood for exposure to chronic levels of pollutants in everyday life. Importantly, current toxicological models predicting toxicity associated with acetaminophen and styrene would typically rely on Michaelis-Menten based models and thus fail to accurately predict the concentration-dependent relationship determining styrene genotoxicity.

Concluding Remarks

Despite the critical role of CYP2E1 in the bioactivation and/or detoxification of many small molecule drugs, pollutants, dietary compounds, and endogenous compounds, its reactions are still not fully characterized. Importantly, the role of CYP2E1 cooperativity in the metabolism of mixtures of these compounds has never been reported. Herein, we have demonstrated the first example of a kinetic characterization of interactions in a mixture of common CYP2E1 substrates. Knowledge of the mechanisms underlying the bioactivation of these compounds is necessary to characterize and predict the link between exposure, metabolism, and toxicity, especially in the case of mixtures.