Exposure to an environmental mixture of polycyclic aromatic hydrocarbons (PAHs) induces hepatic cytochrome P450 enzymes in mice

Abstract

Cytochrome P450 enzymes (CYPs) play an important role in bioactivating or detoxifying polycyclic aromatic hydrocarbons (PAHs), common environmental contaminants. While it is widely accepted that exposure to PAHs induces CYPs, effectively increasing rates of xenobiotic metabolism, dose- and time-response patterns of CYP induction are not well known. In order to better understand dose- and time-response relationships of individual CYPs following induction, we exposed B6129SF1/J mice to single or repeated doses (2–180 μmol/kg/d) of benzo[a]pyrene (BaP) or Supermix-10, a mixture of the top 10 most abundant PAHs found at the Portland Harbor Superfund Site. In hepatic microsomes from exposed mice, we measured amounts of active CYPs using activity-based protein profiling and total CYP expression using global proteomics. We observed rapid Cyp1a1 induction after 6 hr at the lowest PAH exposures and broad induction of many CYPs after 3 daily PAH doses at 72 hr following the first dose. Using samples displaying Cyp1a1 induction, we observed significantly higher metabolic affinity for BaP metabolism (K_m reduced 3-fold), 3-fold higher intrinsic clearance, but no changes to the V_max. Mice dosed with the highest PAH exposures exhibited 1.7 to 5-fold higher intrinsic clearance rates for BaP compared to controls and higher V_max values indicating greater amounts of enzymes capable of metabolizing BaP. This study demonstrates exposure to PAHs found at Superfund Sites induces enzymes in dose- and time-dependent patterns in mice. Accounting for specific changes in enzyme profiles, and relative rates of PAH bioactivation and detoxification, and resulting risk will help translate internal dosimetry of animal models to humans and improve risk assessments of PAHs at Superfund sites.

Graphical Abstract

Introduction

Polycyclic aromatic hydrocarbons (PAHs) are chemicals composed of multiple fused aromatic rings found throughout the environment. Incomplete combustion of organic compounds from fires, volcanic activity, and anthropogenic burning fossil fuels generate PAHs ¹. Inhaling smoke, drinking from contaminated water sources, ingesting smoked and/or grilled meats, and ingesting deposited PAHs on fruits and vegetables expose humans to PAHs ^{2, 3}. Due to the diverse sources of PAH generation and routes of exposure, human exposures to PAHs vary geographically and generally range from 1 to 17 μg/d ^4–7.

Many PAHs are procarcinogens, and cytochrome P450 enzymes (CYPs) oxidize PAHs initiating bioactivation or detoxification pathways. For example, CYPs can transform benzo[a]pyrene (BaP; the prototypical procarcinogen PAH) to reactive metabolites like BaP-7,8-dihydrodiol-9,10-epoxide, which can react with DNA, form adducts, cause DNA mutations, and eventually manifest as cancer ^{8, 9}. CYPs can also detoxify BaP by forming other dihydrodiols, BaP-diones, or hydroxylated metabolites facilitating BaP elimination or further metabolism by Phase I or II enzymes ^10–12. Specific CYP enzymes can contribute to bioactivation, detoxification, or both pathways. For example, in Supersomes expressing human CYPs, Cyp1a1 and Cyp2c19 produced 96% of BaP-7,8-diol (a precursor to BaP-7,8-dihydrodiol-9,10-epoxide) initiating a bioactivation pathway ¹². In the same experiment, Cyp3a4 produced 53% of 3-hyroxy-BaP, a detoxified metabolite of BaP ¹². Cyp1a1 also produced a significant amount of 3-hydroxy-BaP (27%), demonstrating that some CYPs can initiate bioactivation and detoxification pathways. Since specific CYPs can contribute to bioactivation, detoxification, or both pathways, relative amounts of specific CYPs expressed by an individual can determine that individual’s susceptibility PAHs. Relative amounts of specific enzymes that favor bioactivation increases an individual’s susceptibility to PAHs, while relative amounts of specific enzymes that favor detoxification decreases an individual’s susceptibility.

Exposure to PAHs can induce enzymes and change relative amounts of specific enzymes important for bioactivation or detoxification, ultimately affecting susceptibility. Previously, researchers demonstrated that PAHs can induce CYPs via the aryl hydrocarbon receptor (AhR), constitutive androstane receptor, and/or pregnane X receptor ^13–15. Nebert et al. first showed that BaP hydroxylase (aryl hydrocarbon hydroxylase) activity becomes highly induced within 12–24 hr following exposure to various PAHs using mammalian cell cultures ¹⁶. Since this initial discovery, numerous studies with cell and animal models have utilized PAHs to understand the fundamental biology of AhR including transcription, translation, and signal transduction pathways ultimately resulting in induced CYP levels ^{14, 17, 18}.

While researchers have observed CYP induction following PAH exposures in cellular and animal models, several issues make it difficult to translate implications of observed laboratory CYP induction to real-world PAH exposures. First, the majority of these studies investigate inductive effects of a single PAH, while environmental exposures are highly complex mixtures of PAHs with varying carcinogenicity, competing routes of bioactivation/detoxification, and different induction profiles ^{8, 19–22}. Secondly, robust evaluations of CYP induction in respect to both time- and dose-dependence have not been conducted using animal models. Third, CYP induction has primarily been measured by RT-PCR for mRNA quantification and/or western blotting to measure changes in protein expression ^23–25. While such studies effectively measure expression and total amounts of CYPs, these measurements do not necessarily reflect true levels of active CYPs due to RNA translation and post-translational modifications ²⁶. Finally, while quantifying enzyme-specific responses to PAH exposure is important, many of these studies do not measure implications of induction on PAH metabolism, namely resulting effects on metabolism rates of PAHs from induced enzymes.

The purpose of this study was to quantify dose- and time-dependent induction of active CYPs following PAH exposure. We evaluated two PAH exposures: Supermix-10, a mixture of the top 10 most abundant PAHs found at the Portland Harbor Superfund Site ²⁷, and BaP, the prototypical PAH. The US Environmental Protection Agency and other regulatory bodies use the Relative Potency Factor Approach to assess human risk to mixtures of PAHs ²⁸. The Relative Potency Factor Approach normalizes cancer slope factors of procarcinogneic PAHs to BaP, calculating BaP equivalent doses. As such, we wanted to compare the abilities of an environmental mixture and BaP to cause enzyme induction. We measured induction of active PAHs in hepatic microsomes of exposed mice using activity-based protein profiling ^{29, 30}. Activity-based protein profiling is an approach in which chemical activity-based probes specifically target and label only active enzymes, and provide a route to enrich and quantify active enzymes by proteomics ^{29, 30}. We utilized a mixture of CYP probes previously validated to target numerous individual CYPs in key xenobiotic metabolizing CYP families in the liver ³⁰. We employed global proteomics as an additional method to quantify protein expression. To measure implications of induction, we measured rates of BaP metabolism in mice exposed to PAHs. By measuring enzyme-specific induction of CYP activity and expression by BaP and Supermix-10, as well as the implications of observed induction on PAH metabolism, this study provides a more complete picture of PAH induction at both low versus high doses for a single PAH compared to complex PAH mixtures.

Experimental Procedures

Chemicals

Acetone, ethyl acetate, acetonitrile, methanol, sodium sulfate, sulfuric acid, potassium phosphate salts (dibasic and monobasic), reduced nicotinamide adenine dinucleotide phosphate (NADPH), sucrose, phosphate buffered saline (PBS), trifluoroacetic acid, sodium ascorbate, benzo[a]pyrene, and copper sulfate were purchased from Sigma-Aldrich (St. Louis, MO, USA). Dithiothreitol was purchased from MilliporeSigma. Iodoacetamide was procured from Acros Organics (New Jersey, USA). N-(3-azidopropyl)biotinamide (biotin-azide) was purchased from TCI (Tokyo, Japan). Tris(2-carboxyethyl)phosphine (TCEP) was procured from ThermoScientific (Rockford, IL, USA). TAMRA-azide and tris-hydroxypropyltriazolylmethylamine (THPTA) were purchased from Click Chemistry Tools (Scottsdale, AZ, USA).

Supermix-10 is a mixture of the top 10 PAHs found at the Portland Harbor Superfund site on a mass per volume basis including benzo(a)anthracene, retene, pyrene, phenanthrene, naphthalene, fluorine, fluoranthene, chrysene, acenaphthene, and 2-methylnaphthalene (Table 1). EPA, International Agency for Research on Cancer (IARC), or American Conference of Governmental Industrial Hygienists (ACGIH) have classified three of these compounds as carcinogenic including benzo(a)anthracene, naphthalene, and chrysene (Table 1). Supermix-10 was synthesized as previously described ²⁷. ATW8 and ATW12 activity-based probes were synthesized as previously described ³⁰. Drs. Shantu Amin and Arun Sharma (Pennsylvania State University (State College, PA, USA)) synthesized bibenzo[def,p]chrysene according to previous methods ^31–33.

Table 1.

Polycyclic aromatic hydrocarbons (PAHs) found in Supermix-10. Reported molar ratios are normalized to 2-methylnapthalene, the lowest abundant PAH in the mixture.

Compound	CAS	Molecular Weight (g/mol)	Molar Ratio	Molar Fraction	EPAA	Carcinogenicity Classification IARCB	ACGIHC
benzo(a)anthracene	56–55–3	228.29	2.47	0.05	B2, Probable human	2B, Possible human	A2, Suspected human
retene	483–65–8	234.34	7.43	0.15
pyrene	129–00–0	202.25	14.50	0.28		3, Not classifiable
phenanthrene	85–01–8	178.23	1.72	0.03	D, Not classifiable	3, Not classifiable
naphthalene	91–20–3	128.17	3.48	0.07	C, Possible human	2B, Possible human	A3, Confirmed animal
fluorene	86–73–7	166.22	1.84	0.04	D, Not classifiable	3, Not classifiable
fluoranthene	206–44–0	202.25	14.24	0.28	D, Not classifiable	3, Not classifiable
chrysene	218–01–9	228.29	2.59	0.05	B2, Probable human	3, Not classifiable	A3, Confirmed animal
acenaphthylene	208–96–8	154.21	2.00	0.04	D, Not classifiable
2-methylnaphthalene	91–57–6	142.20	1.00	0.02			A4, Not classifiable

^AUnited States Environmental Protection Agency (EPA), Integrated Risk Information System (IRIS)

^BInternational Agency for Research on Cancer (IARC)

^CAmerican Conference of Governmental Industrial Hygienists (ACGIH)

Animals

Our group has utilized female B6129SF1/J mice as an animal model in prior PAH carcinogenicity, toxicokinetic, and in vitro metabolism studies ^34–38. We developed physiologically based pharmacokinetic models (PBPK) to better interpret various models of PAH toxicity and translate results in animal models to humans ^{34, 39}. As such, we used this animal model to maintain consistency and relevancy with our previous work and allow translation of results to humans.

Adult (8–10 weeks) female B6129SF1/J mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Mice were group-housed (4 per cage) in solid bottom cages using alpha cellulose bedding (Shepherd’s™ ALPHA-dri®, Animal Specialties, Inc., Hubbard, OR, USA) under standard laboratory conditions (temperature ranged 20–24 C, humidity ranged 30–70%, 12 hr light-12 hr dark cycle). LabDiet Certified Rodent Diet rodent chow 5002 (LabDiet, St. Louis, MO, USA) and water were available ad libitum. All procedures involving animals were in accordance with protocols established in the NIH/NRC Guide and Use of Laboratory Animals (NIH/NRC) and were approved by Pacific Northwest National Laboratories Institutional Animal Care and Use Committee.

After acclimation (1 week), mice (n=4 per dose and time point) were dosed with 0, 2, 7, 20, 60, or 180 μmol/kg/d benzo[a]pyrene or Supermix-10 in corn oil by oral gavage (5 mL/kg body weight). We choose these exposure levels based on two factors. First, we wanted to compare enzyme induction of human-relevant Supermix-10 exposures to equimolar doses of BaP, the prototypical PAH. Allan et al. estimated angler exposures to PAHs at the Portland Harbor Superfund site at ~3 μmol/kg ⁴⁰, which is comparable to the low exposure level for PAHs used here (2 μmol/kg). Secondly, we wanted to evaluate BaP exposures commonly used in mechanistic studies aimed at understanding enzyme induction. Typically, these exposures range from 100–500 μmol/kg/d ^{41, 42}. As such, we anchored our highest exposure at 180 μmol/kg/d. An additional control cohort of mice were gavaged with drinking water at the same dose volume and times as mice receiving corn oil vehicle. Mice were euthanized using CO₂ asphyxiation and cervical dislocation 6 or 24 hr post initial dose. Additional mice were dosed again at 24 and 48 hr after the initial dose for 3 total repeated doses. These mice were euthanized at 72 hr (24 hr post third dose). After euthanization, livers were rapidly removed, rinsed in 0.1 M PBS (pH 7.4), frozen in liquid nitrogen, and stored at −70 C.

Preparation of liver microsomes

Hepatic microsomes were prepared from frozen tissues using differential centrifugation ²⁶. Livers from exposure replicates (n=4) were pooled for processing. Tissues were minced then homogenized using a tissue tearor followed by glass dounce homogenization in a sucrose (250 mM) PBS solution. Next, samples were centrifuged at 10,000 × g for 25 minutes, separating cellular debris and S9 fractions. The S9 fraction was centrifuged further at 100,000 × g for 90 minutes, separating microsome (pellet) and cytosolic (supernatant) fractions. Supernatant was removed, leaving the microsomal pellet. The resulting pellet was gently rinsed and resuspended in sucrose (250 mM) in PBS. Protein concentrations were determined via the bicinchoninic acid assay (Pierce BCA Assay Kit, Thermo Scientific).

In vitro probe labeling and click chemistry

CYPs were labeled using three different activity-based probes. Hepatic microsomes were normalized to 1 mg/mL total protein concentration. For SDS-PAGE analysis via ABPP fluorescence, 50 μg protein was labeled with the 2EN (20 μM), ATW8 (20 μM), and ATW12 (20 μM) activity-based probe mixture to target active CYPs. For samples used for eventual probe and click chemistry mediated streptavidin enrichment, 1.0 mg microsomal protein was probe labeled with 2EN (20 μM), ATW8 (20 μM), and ATW12 (20 μM). Probe and protein mixtures were incubated with NADPH (2 mM) for 1 hr at 37°C on a thermal shaker, shaking at 1000 rpm. Following incubation with protein, copper catalyzed click chemistry was carried out as previously described ³⁸. Samples were incubated at room temperature in the dark for 90 min with rhodamine-azide (60 μM) for SDS-PAGE analysis or biotin-azide for streptavidin enrichment (60 μM), the reducing agent sodium ascorbate (10 mM), the ligand tris-hydroxypropyltriazolylmethylamine (2 mM), and copper sulfate (4 mM) catalyst. To remove excess biotin-azide, proteins were precipitated by adding a 1.6:1 volume of −20°C methanol. To induce further precipitation, samples were placed in a −80°C freezer for 1 hr. Samples were centrifuged at 21,000 × g for 10 min, and the supernatant was discarded. Protein was resolubilized by adding SDS (1.2 %) in PBS solution, heated at 95°C for 2 min, and sonicated with 2 × 1 s pulses.

SDS-PAGE analysis

SDS-PAGE was used to visualize activity-based probe labeling. Precipitated, probe-labeled rhodamine-appended proteins were resolubilized in 25 μL SDS (1.2 %) in PBS and heated at 95°C for 2 min. SDS Running Buffer (25 μL 2X) and reducing agent (5.5 μL 10X, Invitrogen, Carlsbad, CA, USA) were added to the samples. Samples were heated at 95°C for 2 further minutes. After cooling for 10 minutes, samples were quickly centrifuged to pellet any remaining insoluble proteins or other debris, and 15 μL sample was added to wells of 4 – 12 % Bis-Tris gels (Invitrogen). SDS-PAGE was run at 200 V, 35 mA for 45 min. Gels were imaged with General Electric Typhoon FLA 9500. Fluorescence was quantified using ImageQuant software. Gels were fixed for 30 min with a solution containing methanol (50 %), acetic acid (40 %), and MilliQ water (10 %). Fixed gels were quickly rinsed with water and stained with GelCode Blue overnight. Following two washes with MilliQ water, stained gels were imaged using GelDocEZ (BioRad Laboratories, Hercules, CA, USA).

Streptavidin enrichment

Probe labeled, biotin appended proteins were enriched via streptavidin enrichment as detailed previously ³⁸. First, resolubilized protein samples were normalized via the BCA to 0.600 mg/mL protein. Streptavidin agarose resin (100 μL, 50% agarose bead slurry) was washed with SDS (0.5 %) in PBS, urea (6 M) in NH₄HCO₃ (25 mM, pH = 8.0), and PBS solutions. Protein samples were added to the washed resin and incubated at 37°C for 1 hr with rotation. Protein bound resin was subsequently washed under vacuum using fritted columns (BioRad). Wash solutions included SDS (0.5 %) in PBS, urea (6M) in NH₄HCO₃ (25 mM), PBS, and NH₄HCO₃ (25 mM). Beads were then transferred to 1.5 mL tubes and were reduced and alkylated with tris(2-carboxyethyl)phosphine hydrochloride (5 mM) and iodoacetamide (10 mM) incubations at 37°C and 50°C, respectively. Samples were again transferred to fritted columns and subjected to further washes with PBS and NH₄HCO₃ (25 mM). Beads were transferred to new tubes, and 0.25 μg trypsin (Promega, Madison, WI, USA) was added to each sample. On-bead trypsin digestion was conducted overnight in NH₄HCO₃ (25 mM, pH = 8.0) at 37°C with rotation. Beads were centrifuged, and the supernatant containing peptides was dried on a speedvac concentrator. Peptides were resolubilized in 40 μL NH₄HCO₃ (25 mM) and vialed for MS analysis.

Protein digestion and cleanup

We digested proteins into peptides and cleaned the solution using our standard lab protocols ³⁸. Microsomal proteins from each sample were normalized to 1.000 mg/mL protein. Protein samples (100 μg) were denatured in urea (8 M). Samples were reduced and alkylated with dithiotheitrol (5 mM) and iodoacetamide (40 mM), respectively. Diluted samples (10-fold) were incubated with 2 μg trypsin (Promega) for 3 hr at 37°C with rotation. Peptide cleanup was facilitated with Discovery C18 solid phase extraction columns conditioned with methanol and equilibrated with 0.1% trifluoroacetic acid. Peptides were washed with 5:95 acetonitrile:water and eluted using 1 mL 80:20 acetonitrile:water. Peptides were dried in a speedvac concentrator and resuspended in NH₄HCO₃. Peptide concentration was normalized (0.1 mg/mL) using bicinchoninic acid assay and analyzed (25 μL) using liquid chromatography with tandem mass spectrometry (LC-MS/MS).

LC-MS/MS analysis

Probe-enriched and global peptides were analyzed on a Velos Orbitrap mass spectrometer as described previously ⁴³. An AMT tag approach was used to analyze peptides ⁴⁴. Peptide spectra were searched against the UniProt mus_musculus database ⁴⁵ and rescored using MSGF+ ⁴⁶. Only peptides unique to a single protein were used for further analysis to ensure abundance values related to a specific peptide can be mapped back to one possible protein, eliminating the possibility of overlap between closely related proteins. Peptides were also statistically filtered for those with an MT Uniqueness ≥ 0.5 and MT FDR Threshold ≤ 0.1. The resulting list of peptide abundances were log base 2 transformed and scaled to the protein level using Inferno RRollup software ⁴⁷. The RRollup method scales all peptides for a protein to the peptide with the highest percent coverage across all datasets, with abundance as the secondary parameter for proteins with multiple peptides with 100 % coverage across all samples. After scaling to each reference peptide, the median peptide abundance value is used as the relative protein abundance value. Only proteins with >3 peptides were included in further analysis.

In vitro metabolism studies

Implications of observed PAH induction of hepatic CYPs was measured using in vitro metabolism of benzo[a]pyrene in microsomes of mice from different exposures. We chose to use BaP as the standard for PAH metabolism due to the large body of work surrounding BaP carcinogenicity and metabolism ⁴⁸. All incubations were carried out similar to our previous work using previously optimized conditions ^{35, 37}. Within 0.5 mL total volume, reactions consisted of MgCl₂ (3mM), microsomes (0.3 mg/mL microsomal protein), 0.1 M phosphate buffer (pH 7.4), and excess NADPH (1.5 mM). Control samples absent of NADPH were also included. Microsomes were preactivated via 37 °C incubation on a thermal shaker for 3 min. Microsome solutions were incubated with benzo[a]pyrene (0.625 μM – 20 μM) for 0 or 10 min. Incubations were terminated with H₂SO₄ (0.45 M), and samples were immediately placed on ice until extraction. Each experimental condition (e.g. substrate, concentration, time point, etc.) was conducted in triplicate.

Analyte quantitation

Analytes were extracted from samples with liquid-liquid extraction and measured using high performance liquid chromatography (HPLC). Dibenzo[def,p]chrysene was added to each inactivated sample as an internal standard. Samples were extracted using two 0.750 mL volumes of ethyl acetate. Extracts were dried under a gentle stream of nitrogen and resolubilized in 500 μL methanol. A benzo[a]pyrene (0.031 uM – 25.0 uM) external standard curve was also prepared in inactivated microsomes. Standards were extracted using the same method at the same time as the samples. Analytes were quantified by reverse phase HPLC using an Agilent 1100 HPLC system equipped with a fluorescence detector (Santa Clara, CA, USA). Ten μL of sample was injected into an Ascentis 25 cm x 4.6mm, 5μm C18 column (Sigma-Aldrich, St. Louis, MO, USA). A gradient of water and acetonitrile from 55:45 to 0:100 was employed from 0 to 10 min and then was held at 100% acetonitrile until 22 min at a flow rate of 0.95 mL/min. DBC and BaP were measured via fluorescence detection. Excitation and emission wavelengths were 230 nm and 430 nm, respectively. Elution times for compounds were roughly 20 min for DBC, and 17 min for BaP. Quantification was accomplished using a linear regression fit to an external calibration curve. The approximate limit of reliable quantification (LOQ) for benzo[a]pyrene was ~0.03 μM.

Data analysis

As prior, in vitro metabolism studies conducted in our lab with PAHs ^{35, 37}, metabolism rates were calculated based on substrate disappearance due to further concurrent metabolism of metabolic products. For each substrate concentration, metabolism rates (as initial rates of substrate disappearance; nmol/min/mg microsome) were calculated using an exponential regression of benzo[a]pyrene concentrations (μM) as a function of time (min). Confidence intervals of metabolism rates were calculated using a nonparametric bootstrap, where metabolism rates were resampled with replacement at each concentration, and the regression was repeated on the resampled dataset (n=1000). Metabolism rates as a function of substrate concentration were evaluated using a Michaelis-Menten equation for saturable kinetics. Intrinsic clearance values (Cl_int) of the first order phase were estimated by dividing the Vmax by the respective Km. Confidence intervals of fit parameters were estimated using a bootstrap (n=1000), where each best fit model was re-optimized to sampled metabolism rates from a normal distribution (n=10) at each time point.

Statistics and dose response modeling

A variety of statistical and modeling approaches were used evaluate data. Due to limitations on the amount of available microsomes and the required amount of microsomes for MS and in vitro metabolism assays, we pooled biological replicates. As such, we fit linear regression models with y-intercept fixed at 1 (control level) to measured fluorescence and MS counts normalized to corn oil controls, to statistically test effects of dose (6 exposure groups) on enzyme levels. Slopes were compared using a t-test and an alpha value of 0.05. While analyzing these replicates and using an analysis of variance (ANOVA) to compare results may be favored by some, linear regression analyses across varying doses increases statistical power compared to ANOVA analyses of an equivalent number of samples. Welch’s t-test was used to compare two samples (e.g. corn oil vs. water). Pearson’s correlation coefficient was used to compare global and probe enriched enzymes measured with mass spectrometry. For enzymes that were significantly inducted, we used dose response modeling to identify potency of exposures for inducing enzymes. A Hill Equation (Equation 1, ^49–51) described enzyme levels relative to controls (RE) as a function of daily administered dose of PAHs (x), where h is the Hill Coefficient, max is the maximum induction level, and b is the median dose. We used fit models to calculate the threshold dose required to induce enzymes by 3-fold. This ratio was approximately the level in which we observed significantly different slopes using the regression model. We used R: A language and environment for statistical computing, version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria) for data and statistical analysis and dose response modeling.

$$RE\left(x\right)=1+\frac{max\times x^h}{x^h+b^h}$$ Equation 1

Results

Mice did not display adverse effects following exposure to BaP, Supermix-10, corn oil, or water.

Gel-based Imaging of Active CYP from Mice Exposed to PAHs

Mice exposed to BaP displayed dose- and time-dependent increases in hepatic CYPs. Gel-based imaging resolved a prominent fluorescent band at 52–56 kDa consistent with CYP enzymes labeled with activity-based probes appended with a fluorophore (Figure 1). Fluorescent bands from corn oil and water exposed mice varied 24% and did not demonstrate significant differences (p=0.75). In microsomes from mice exposed to BaP, we did not observe significant changes (p=0.17) in total fluorescence 6 hr post exposure compared to controls. However, after 24 hr post exposure, probe fluorescence intensity increased significantly with dose (p=0.002), increasing over 50% in mice exposed once to 60 or 180 μmol/kg/d BaP compared to control mice receiving corn oil only (Figure 2A). After 72 hr (3 doses), probe fluorescence intensity increased significantly with dose (p=0.007) and exhibited 2-fold higher levels in mice exposed to 60 or 180 μmol/kg/d BaP compared to controls (Figure 2A). Increases in fluorescence of labeled CYPs provide evidence that exposure to BaP induces active CYPs in mice.

Figure 1.

Labeled proteins 52–56 kDa (size range for CYPs) from hepatic microsomes of mice exposed to varying doses benzo[a]pyrene or Supermix-10 separated using an SDS-PAGE gel.

Figure 2.

Fluorescence of labeled proteins from hepatic microsomes from mice treated with varying doses of benzo[a]pyrene (A) or Supermix-10 (B) relative to corn oil (CO) controls assayed with activity-based probes. Control mice exposed to water (H2O) were also included. Measurements represent 4 pooled biological replicates per sample.

We observed less pronounced CYP induction in mice exposed to Supermix-10 than mice exposed to BaP using gel-based techniques. While not statistically significant as a function of dose (p>0.19), CYP fluorescence in hepatic microsomes from mice dosed with 180 μmol/kg/d of Supermix-10 appeared elevated compared to those receiving corn oil (Figure 2B). Since fluorescent imaging of CYPs labeled with activity-based probes is bulk measurement of many CYPs, we used enrichment techniques and mass spectrometry to gain sensitivity and selectivity of possible CYP induction, specifically regarding how individual CYPs respond to PAH exposure.

Active CYP Profiles from Mice Exposed to PAHs

The increased sensitivity and selectivity of mass spectrometry revealed two primary patterns of active CYP induction caused by PAHs including 1) rapid Cyp1a1 induction within 6 hr and 2) broad induction of other CYPs after longer time periods. After labeling active CYPs in microsomes with probes and enriching with biotin/streptavidin, proteomic analysis with mass spectrometry identified 39 total CYPs, 37 of which exhibited enough data for comparisons across PAH dose and time following exposures to BaP or Supermix-10 (Figure 3).

Figure 3.

Relative abundances of active CYPs from hepatic microsomes of mice exposed daily to varying exposures of benzo[a]pyrene (A) or Supermix-10 (B) after 6, 24, or 72 hr. Enzymes were enriched using activity-based probes and measured with mass spectrometry. White blocks indicate the sample was below detection limits. Gray blocks (20 μmol/kg/day benzo[a]pyrene at 24 hr) indicate the sample was missing. Measurements represent 4 pooled biological replicates per sample.

Within 6 hr, both BaP and Supermix-10 caused rapid Cyp1a1 induction in mice (Figure 3). Cyp1a1 was not detected in hepatic microsomes from control mice following exposure to corn oil or water. We detected Cyp1a1 in all mice treated with PAHs at the 6 hr time point, indicating both BaP and Supermix-10 caused rapid Cyp1a1 induction. In later time points, we also detected Cyp1a1 in mice exposed to the high dose of Supermix-10 and all mice exposed to ≥20 μmol/kg/d BaP (Figure 3). However, at lower exposures (<20 μmol/kg/d BaP and <180 μmol/kg/d Supermix-10) detection of Cyp1a1 was sporadic suggesting Cyp1a1 induction attenuated over time. Since control Cyp1a1 levels were below limits of MS detection, we used the lowest detected Cyp1a1 level (mice exposed to 20 μmol/kg/d Supermix-10, 6 hr after exposure) as “control” baseline levels for normalizing Cyp1a1 detected in other doses and time points (Figure 3). Using Cyp1a1 levels from this exposure to normalize data effectively under-biased reported Cyp1a1 induction; however, it allowed us to illustrate and evaluate induction levels in the absence of measurable levels in control samples (Figure 3). Levels of Cyp1a1 significantly increased as a function of dose at every time point evaluated that had at least 2 detections (p ≤ 0.001) using a linear regression model (Tables S1 and S2). BaP-induced Cyp1a1 levels were higher than the respective Supermix-10-induced levels except at one exposure condition (2.2 μmol/kg/d exposure group at 6 hr). This suggests that BaP induced Cyp1a1 more robustly than Supermix-10. Within 6 hr, Supermix-10 also significantly induced Cyp7a1 (p=0.01); however only the high exposure condition (180 μmol/kg/d) exceeded controls by 2-fold. We did not observe any other significantly increased enzymes in 6 hr from BaP or Supermix-10 exposures.

At later timepoints, PAH exposures induced additional CYPs in hepatic microsomes of mice. In 24 hr, BaP and Supermix-10 statistically increased 20 and 2 CYPs, respectively (Tables S1 and S2). Notably, Cyp1a2 levels increased 13- and 7-fold compared to control levels following high BaP and Supermix-10 exposures, respectively. BaP also induced Cyps 2c29 and 4b1 at least 5-fold higher than control levels (Figure 3). After 72 hr and three repeated exposures, BaP significantly induced 36 out of 37 measurable CYPs (Table S1). Supermix-10 significantly induced less CYPs than BaP (17); however, these CYPs still represented the majority of detectable CYPs (Table S2).

Dose-response modeling identified PAH exposure levels that induce CYPs and CYPs susceptible to PAH induction. After fitting a Hill equation to control-normalized levels of PAHs as a function of dose, we calculated exposure levels required to induce various CYPs by 3-fold. Using this approach, BaP more potently induced Cyps 1a1 and 1a2 compared to all CYPs in 6 and 24 hr, as 5–45 μmol/kg/d sufficiently caused 3-fold induction (Tables S1 and S2). Other CYPs required >64 μmol/kg/d to reach the 3-fold induction threshold. At 72 hrs, <50 μmol/kg/d of daily BaP induced Cyps 1a1, 1a2, 2a5, 2b19, 3a11, and 4a10 by 3-fold. Supermix-10 exposure <50 μmol/kg/d daily caused at least 3-fold induction in only one enzyme: Cyp1a1 after 6 hr. All other CYPs required ≥67 μmol/kg/d of daily exposure to cause 3-fold induction regardless of time sampled. Dose-response modeling thresholds demonstrate higher CYP induction potency of BaP compared to Supermix-10 and provides quantitative context of external exposures required to sufficiently induce CYPs.

Global proteomics

Global proteomic measurements with MS also provided evidence of CYP induction following BaP and Supermix-10 exposures. Using global proteomic measurements, BaP significantly increased 1, 2, and 6 enzymes at 6, 24, and 72 hr post exposure, respectively (Figure 4A). Cyps 1a1 and 1a2 were particularly sensitive to BaP induction as demonstrated by low dose-response thresholds required to induce enzymes 3-fold (≤21 μmol/kg/d). Supermix-10 significantly induced 1, 3, and 9 enzymes at 6, 24, and 72 hr post exposure, respectively (Figure 4B), with Cyps 1a1, 1a2, and 2b10 being the most sensitive. Overall, enzymes measured by global proteomic and activity-based approaches were highly correlated for some but not all enzymes (Table 2). For example, global measurements for Cyps 1a1, 1a2, and 2a5 highly corelated (r ≥ 0.86) to activity-based measures, suggesting that induction among these enzymes was a translational event. However, global and activity-based measured Cyp 3a11 and 4a10 poorly correlated (r ≤ 0.26) to each other, suggesting that a non-translational mechanism maybe driving induction of these enzymes by BaP.

Figure 4.

Relative abundances of CYPs from hepatic microsomes of mice exposed daily to varying exposures of benzo[a]pyrene (A) or Supermix-10 (B)after 6, 24, or 72 hr. Enzymes were measured globally with mass spectrometry. White blocks indicate the sample was below detection limits. Gray blocks (20 μmol/kg/day benzo[a]pyrene at 24 hr) indicate the sample was missing. Measurements represent 4 pooled biological replicates per sample.

Table 2.

Pearson’s correlation coefficients (r) of normalized CYP levels of mice exposed to benzo[a]pyrene (BaP) or Supermix-10 measured using activity-based protein profiling or global proteomics.

Enzyme	r
	Both	BaPA	Supermix-10
cyp17a1	0.15	0.22	0.28
cyp1a1	0.89	0.89	0.98
cyp1a2	0.86	0.84	0.95
cyp2c37	−0.48	−0.59	−0.41
cyp2c39	0.12	0.19	0.42
cyp2c40	−0.06	−0.04	0.37
cyp2c54	−0.09	−0.13	−0.01
cyp2c55	−0.21	−0.47	0.22
cyp2c70	0.05	0.04	0.47
cyp27a1	0.06	−0.23	0.48
cyp2a4	0.15	0.17	0.34
cyp2a5	0.89	0.91	0.80
cyp2a12	−0.09	−0.08	−0.04
cyp2b10	0.49	0.26	0.95
cyp2c29	0.22	0.24	0.91
cyp2d9	0.25	0.16	0.24
cyp2d10	0.21	0.18	0.25
cyp2d26	−0.42	−0.47	−0.58
cyp2e1	0.08	−0.05	0.51
cyp2f2	−0.20	−0.27	−0.33
cyp2j5	0.15	0.25	0.31
cyp3a41a	−0.18	−0.27	0.17
cyp3a11	0.15	0.06	0.67
cyp3a13	−0.12	−0.15	−0.17
cyp4a10	0.12	0.23	0.01
cyp4a14	0.48	0.57	0.27
cyp4f14	−0.04	−0.01	−0.14
cyp4v2	0.10	0.14	0.14
cyp51a1	0.39	0.70	0.46

^ABenzo[a]pyrene (BaP)

Metabolism rates

We selected several samples to evaluate implications of observed induction by measuring rates of BaP metabolism in vitro. Residual BaP levels were below detection limits in hepatic microsome samples from treated mice, and as such, did not impede our ability to compare rates of BaP metabolism from various exposure groups. We evaluated mice dosed with 180 μmol/kg/d BaP and Supermix-10 after 72 hr to evaluate effects of maximal induction for both PAH exposures and mice dosed with 6 μmol/kg/d BaP after 6 hr to evaluate implications of rapid Cyp1a1 induction.

In general, mice exposed to high exposures of PAHs for 3 days exhibited faster rates of BaP metabolism than control mice. Hepatic microsomes from treated mice exhibited 2- to 4-fold faster rates of metabolism than control mice at every substrate concentration tested (Figure 5A). Microsomes from mice dosed with 180 μmol/kg/d of BaP for 3 days had a different pattern of BaP metabolism compared to microsomes from mice dosed with Supermix-10. After 72 hr, microsomes from mice dosed with 180 μmol/kg/d of BaP demonstrated a higher V_max and lower K_m compared to control microsomes from mice dosed with corn oil (Figure 5A, Table 3). In contrast, microsomes from mice dosed with 180 μmol/kg/d of Supermix-10 demonstrated only a higher V_max and an equivalent K_m compared to controls (Figure 5A, Table 3). This suggests that exposure to Supermix-10 causes different enzymes to be induced compared to exposure to BaP. BaP induction not only increases enzyme amounts (as evidenced by the higher V_max) but also increases enzymes that have a higher affinity to metabolize BaP (as evidenced by the lower K_m).

Figure 5.

Rates of enzymatic benzo[a]pyrene (BaP) disappearance in hepatic microsomes from mice 72 hrafter daily exposures to 180 μmol/kg/d BaP (red circle), 180 μmol/kg/d Supermix-10 (SM-10, blue diamond), or corn oil vehicle control (green triangle) (A) or 6 hr after exposure to 20 μmol/kg/d BaP (red circle) or corn oil vehicle control (green triangle) (B). Lines represent Michaelis-Menten model fits to data.

Table 3.

Parameters of benzo[a]pyrene metabolism in hepatic microsomes from mice treated with benzo[a]pyrene (6 or 180 μmol/kg/d) after 6 or 72 hrs, Supermix-10 (180 μmol/kg/d) after 72 hrs, or corn oil vehicle 6 or 72 hrs.

Compound	Time (hr)	Dose (μmol/kg/d)	Vmax (nmol/min/mg)	95% CIA	K m (μM)	95% CI	Clint B (mL/min/mg)	95% CI
Corn oil	6	0	1.57	1.38–1.82	3.81	2.53–5.57	0.41	0.32–0.55
BaPC	6	6	1.44	1.30–1.66	1.16	0.85–1.59	1.24	1.00–1.54
Corn oil	72	0	1.81	1.66–1.97	5.15	4.22–6.48	0.35	0.30–0.40
BaP	72	180	2.41	2.33–2.49	1.38	1.11–1.69	1.75	1.47–2.12
Supermix-10	72	180	3.14	2.89–3.44	5.41	4.52–6.56	0.58	0.52–0.65

^AConfidence Interval (CI)

^BIntrinsic Clearance (Cl_int)

^CBenzo[a]pyrene (BaP)

Microsomes with higher levels of Cyp1a1 induced by BaP increased the intrinsic clearance of BaP but did not increase the V_max. After 6 hr of 20 μmol/kg/d of BaP exposure, we observed 5-fold induction of Cyp1a1 compared to the lowest measured value of Cyp1a1 (mice exposed to 20 μmol/kg/d Supermix-10, 6 hr after exposure) while no other CYPs were elevated (Figure 3A). We evaluated this sample to observe how induction of Cyp1a1 affected overall metabolism rate compared to mice exposed to corn oil. Compared to corn oil controls, mice treated with 20 μmol/kg/d of BaP exposure demonstrated 2-fold higher metabolism rates than control mice at low doses and nearly the same rates at high doses. This exposure to mice significantly reduced the apparent K_m 3-fold without observable changes in V_max, effectively increasing intrinsic clearance by 3-fold (Figure 5A, Table 3). The reduced BaP K_m of these microsomes demonstrates the high affinity of Cyp1a1 for BaP compared to other CYP enzymes. No difference in the V_max values suggest the possibility of product inhibition of Cyp1a1 by BaP metabolites ^{52, 53}.

Discussion

In this study, we quantified dose- and time-dependent induction of active CYPs following PAH exposures. After exposure to Supermix-10 or BaP, we observed rapid induction of Cyp1a1 at the lowest PAH exposures tested in 6 hr and broad scale induction of many CYPs after 3 daily doses and 72 hrs. We measured BaP metabolism using induced samples and observed significantly higher metabolic affinity in samples displaying Cyp1a1 induction and higher overall metabolism rates for broad scale induction of many CYPs. This study provides quantitative dose- and time- dependent induction of an environmental mixture of PAHs (Supermix-10) and a well-studied, prototypical PAH (BaP).

Supermix-10 induced CYPs at doses relevant to humans. The environmental surrogate mixture Supermix-10 contains the top 10 most abundantly measured PAHs at the Portland Harbor Superfund site measured 2010 and 2015 ^{27, 40, 54}. Allan et al. estimated angler exposures to PAHs at the Portland Harbor Superfund site at ~3 μmol/kg ⁴⁰, which is comparable to the low exposure level for Supermix-10 used here (2 μmol/kg). At that Supermix-10 exposure, we observed Cyp1a1 induction in 6 hr and broad scale induction of CYPs at 72 hr (Figure 3).

BaP induced CYPs at exposures commonly used in laboratory animal models used for hazard and risk assessments. EPA’s IRIS assessment of BaP identified benchmark doses, no adverse effect levels (NOAELs), and lowest adverse effect levels (LOAELs) ranging from 0.8 to 120 μmol/kg/d for a variety of non-cancer endpoints in mice and rats ⁴⁸, consistent with doses that caused enzyme induction here. As such, risk assessments that utilize repeated doses in animal models need to consider induced metabolism in order to accurately translate internal dosimetry from animal models to humans. Human exposures to BaP generally range 1.1–4.6 nmol/d for non-smoking adults in US, although BaP exposures can be higher in people that smoke or are occupationally exposed to smoke (e.g. petroleum industry or fire fighters) ^{55, 56}. As such, BaP doses we tested here are higher than typical human BaP exposures but relevant to doses in animal models used for BaP risk assessment.

We measured the effects of CYP induction on overall metabolism of a model PAH, BaP and observed significant changes to the K_m and V_max as a result of exposure to a single PAH and complex PAH environmental mixture. However, these measured changes in the total CYP metabolism of PAHs still do not provide context as to how CYP induction alters the equilibrium between bioactivation and detoxification. Using activity-based protein profiling, we measured enzyme-specific increases in CYP activity as a result of induction. We measured changes in CYP expression via global proteomics to compare to activity-based probe measurements. The metabolic implications of enzyme-specific induction can be garnered by more closely investigating individual cytochrome CYPs induced in this study and their specific roles in PAH metabolism.

Among the various CYPs, Cyp1a1 demonstrated the most potent induction in response to PAH exposure. The AhR highly regulates Cyp1a1, and many PAHs induce this receptor ^{14, 57}. Previous studies observed increased Cyp1a1 expression with high doses of PAHs or other AhR inducers. While we did find that Cyp1a1 was greatly induced at high doses and later time points, it was induced rapidly (6 hr) at the lowest administered exposures for BaP and Supermix-10 (2 μmol/kg/d). Other researchers observed low AhR binding affinities of fluoranthene and pyrene, which make up ~50% of Supermix-10 ⁵⁸. Assuming AhR mediates observed Cyp1a1 induction, these results may explain our observation that BaP demonstrated higher Cyp1a1 induction potency than Supermix-10.

Using samples with higher levels of Cyp1a1 only, we observed higher affinity for BaP metabolism compared to vehicle control (lower K_m) and higher rates (5-fold) of metabolism at lower substrate concentrations (0.6 μM), suggesting Cyp1a1 metabolizes BaP with higher affinity than other enzymes. Consistent with our observation, Sulc et al. observed BaP 3-hydroxide, a detoxification pathway, as the primary product of recombinant Cyp1a1 ¹². Sulc et al. also observed considerable amount of benzo[a]pyrene-7,8-dihydrodiol formation, a precursor to BaP diol epoxide, suggesting that induced Cyp1a1 also contributes to BaP bioactivation ¹². Rats with induced Cyp1a1 demonstrated higher BaP bioactivation and elevated BaP–DNA adduct levels in ex vivo incubations of microsomes compared to non-induced Cyp1a1 rats ⁵⁹, suggesting induced Cyp1a1 may increase cancer risk. Hepatic constitutive expression of Cyp1a1 is very low ^{60, 61}, and we did not observe Cyp1a1 in control samples here. Cyp1a1 induction of the magnitude observed in our study likely results in higher BaP detoxification and BaP bioactivation compared to control samples.

In Cyp1a1 induced microsomes, we observed higher BaP metabolism rates at lower doses without an increase in V_max, which is consistent with product inhibition of Cyp1a1 through allosteric or steric substrate mechanisms. Lin et al. observed similar phenomena with the related Cyp1a2 enzyme ⁶². Others have also observed oxidized BaP metabolites inhibiting BaP metabolism ^{52, 53}. It is reasonable that substrate and/or product inhibition may serve a protective role during exposure to PAHs, with bioactivation by Cyp1a1 being reduced at increased PAH concentrations. This phenomenon could reduce carcinogenic risk as the maximum rate of BaP metabolism was expected to be much higher than the corn oil controls without observed inhibition; however, product inhibition would require sustainable exposures to achieve internal concentrations of ~5–15 μM, where this phenomenon appears. As such, it may only be relevant for interpreting high exposure models of PAH mechanisms rather than those relevant to typical human exposures.

PAHs induced Cyp1a2 the second most of all CYPs measured. Large fold change increases in active CYP enzymes closely correlate with increases in expression as measured by global proteomics, indicating that the increases in activity are largely driven by increases in expression. Cyp1a2 is known to be highly inducible by AhR and is known to have high constitutive expression in the liver ^{14, 63}. Bauer et al. demonstrated that Cyp1a2 bioactivates various PAHs and their diols slower than Cyp1a1 ⁶⁴. Human liver Cyp1a2 is also known to exhibit a high amount of variability between individuals (up to 60-fold) ⁶⁵. This high amount of variability, in combination with its propensity for the bioactivation of PAHs, could have significant implications on an individual’s susceptibility to PAH exposures.

BaP and Supermix10 robustly induced Cyp2a5 activity after 72 hr. Activity-based probe measurement correlated well with global measurements, suggesting protein expression caused increased activity. The human ortholog of mouse Cyp2a5 is Cyp2a6. Previous studies show that 2a6 is regulated by an AhR-dependent pathway and is inducible by AhR ligands ⁶⁶. Cyp2a6 is known to produce procarcinogenic products of PAH metabolism, and observed Cyp2a6 induction may result in increased cancer risk ⁶⁷.

Observed induction of CYP enzymes by BaP were not accompanied by similar induction of glutathione transferases (GSTs), Phase II enzymes that detoxify PAHs. After 3 days of 180 μg/kg/d BaP in these same mice, Stoddard et al. observed modest induction of some GSTs; however, at lower doses, no major induction was observed ⁶⁸. This suggests that changes in CYP enzyme activity from repeated BaP exposures will drive toxicity at lower BaP exposures.

Regulatory agencies currently do not consider enzyme induction when assessing risk of PAHs. Currently, the US Environmental Protection Agency uses the Relative Potency Factor Approach to assess human risk to mixtures of PAHs ²⁸. This approach compares potencies of PAHs to BaP in animal models, and relative potency factors are calculated from external daily doses of PAHs and BaP that cause toxicity ²⁸. Internal dose metrics of PAHs or the reference BaP are currently not considered when translating relative potency factors to humans. Since our research has demonstrated that rates of PAH metabolism differ among animal models and humans ^{35, 37}, we propose that tools like physiologically based pharmacokinetic (PBPK) models could be used to translate internal dose metrics of PAHs from animal models to humans for more accurate relative potency factor calculations ^{34, 39}. Integrating enzyme induction and metabolism rate data measured here into PBPK models will allow us to better predict consequences of enzyme induction in relative rates of bioactivation and detoxification of PAHs in animal models and translate those results to humans for better risk assessment.

Supplementary Material

SI

Table S1. Statistics for activity-based probe analysis of mice exposed to 2–180 μmol/kg/d benzo[a]pyrene and sampled at various times.

Table S2. Statistics for activity-based probe analysis of mice exposed to 2–180 μmol/kg/d Supermix-10 and sampled at various times.

Table S3. Statistics for global proteomic analysis of mice exposed to 2–180 μmol/kg/d benzo[a]pyrene and sampled at various times.

Table S4. Statistics for global proteomic analysis of mice exposed to 2–180 μmol/kg/d Supermix-10 and sampled at various times.

Abbreviations

ABP

activity-based probe

ABPP

activity-based protein profiling

AhR

aryl hydrocarbon receptor

AMT

accurate mass and time

BaP

benzo[a]pyrene

BPDE

anti-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene

CAR

constitutive androstane receptor

LC-MS

liquid chromatography-mass spectrometry

PAH

polycyclic aromatic hydrocarbons