Journal of Toxicology and Environmental Health, Part A, 61:209±223, 2000 Copyright 2000 Taylor & Francis 1528-7394/00 $12.00 + .00
PHYSIOLOGICALLY BASED MODELING OF THE MAXIMAL EFFECT OF METABOLIC INTERACTIONS ON THE KINETICS OF COMPONENTS OF COMPLEX CHEMICAL MIXTURES Sami Haddad, Ginette Charest-Tardif, Kannan Krishnan Groupe de recherche en toxicologie humaine (TOXHUM), Faculté de médecine, Université de Montréal, Montréal, Québec, Canada The objective of this study was to predict and validate the theoretically possible, maximal impact of metabolic interactions on the blood concentration profile of each component in mixtures of volatile organic chemicals ( VOCs) [dichloromethane ( DCM) , benzene ( BEN) , trichloroethylene ( TCE), toluene ( TOL), tetrachloroethylene ( PER), ethylbenzene ( EBZ), styrene ( STY), as well as para, ortho-, and meta-xylene ( p-XYL, o-XYL, m-XYL)] in the rat. The methodology consisted of: ( 1) obtaining the validated, physiologically based toxicokinetic ( PBTK) model for each of the mixture components from the literature, ( 2) substituting the Michaelis–Menten description of metabolism with an equation based on the hepatic extraction ratio ( E) for simulating the maximal impact of metabolic interactions ( i.e., by setting E to 0 or 1 for simulating maximal inhibition or induction, respectively), and ( 3) validating the PBTK model simulations by comparing the predicted boundaries of venous blood concentrations with the experimental data obtained following exposure to various mixtures of VOCs. All experimental venous blood concentration data for 9 of the 10 chemicals investigated in the present study ( PER excepted) fell within the boundaries of the maximal impact of metabolic inhibition and induction predicted by the PBTK model. The modeling approach validated in this study represents a potentially useful tool for screening/identifying the chemicals for which metabolic interactions are likely to be important in the context of mixed exposures and mixture risk assessment.
In the past, there have been many reports on toxicokinetic interactions between chemicals applied jointly. An extensive review of the literature (Krishnan & Brodeur, 1991) demonstrated that the vast majority of the toxicokinetic interactions resulted from metabolic induction or inhibition caused by some components of the mixtures. Metabolic interactions can alter tissue dosimetry, and thereby the toxicity of mixture components. Until now, risk assessments of mixtures have not taken into account the possible modulation of tissue dose of chemicals in mixtures. Recently Received 16 February 2000; sent for revision 22 March 2000; revision received 15 May 2000; accepted 15 May 2000. This work represents an initiative undertaken as a part of Dr. Krishnan’s research program on physiological modeling, supported by grants from the Canadian Network of Toxicology Centres (CNTC), Natural Sciences and Engineering Research Council of Canada (NSERC), Fonds de la Recherche en Santé du Québec (FRSQ), and Fonds pour la Formation de Chercheurs et I’Aide a la Recherche (FCAR). K. Krishnan is recipient of a research scholarship from FRSQ (1992–2004). Address correspondence to Kannan Krishnan, Département de médecine du travail et d’hygiène du milieu, Université de Montréal, 2375 Côte Ste-Catherine Bureau 4105, Montréal, PQ H3T 1A8, Canada. E-mail:
[email protected] 209
210
S. HADDAD ET AL.
an approach has been developed by Haddad et al. (1999a) that uses physiologically based toxicokinetic (PBTK) models to account for toxicokinetic interactions in the calculation of biological hazard indices based on target tissue concentrations. The tissue dose of chemicals in mixtures can be accurately predicted with PBTK models when binary interactions between all mixture components are characterized (Haddad et al., 1999b). Unfortunately, there is no easy way to predict a priori the quantitative characteristics of these binary interactions, and therefore they have to be determined following experimentation. Given that the number of binary combinations (N) within a mixture equals n(n – 1)/2, where n is the number of mixture components, it is obvious that a large amount of experiments should be conducted in order to characterize qualitatively and quantitatively the nature of the interactions in mixtures. As an alternative, it could be of interest to use minimal and maximal values of the hepatic extraction ratio (i.e., E = 0 and E = 1) as a predictor of maximal impact of metabolic interactions on the toxicokinetic of chemicals. The minimal value would correspond to maximal inhibition of metabolism, whereas the maximal value would correspond to maximal induction. Considering the complexity of some of the mixtures to which humans are exposed to, it will be nearly impossible to characterize all existing binary interactions in every mixture. In some cases, the mixture composition is ill-defined such that all relevant binary interactions cannot even be identified. One way of addressing this problem, until all binary interactions can be easily characterized, would be to use the theoretical limits of the modulation of tissue dose that would arise form hypothetical metabolic interactions. PBTK modeling is potentially useful to provide predictions of the maximal impact of metabolic interactions, by setting hepatic extraction ratio to 0 and then to 1 (Poulin & Krishnan, 1999). Such predictions of maximal change in the tissue dose or blood concentration of chemicals in mixtures obtained using PBTK models should be validated using appropriate experimental data. The objective of the present study was to predict and validate the theoretically possible, maximal impact of metabolic interactions on the blood concentration profile of each component in rats exposed to various mixtures of volatile organic chemicals (VOCs). The following chemicals were chosen for the study: dichloromethane (DCM), benzene (BEN), trichloroethylene (TCE), toluene (TOL), tetrachloroethylene (PER), ethylbenzene (EBZ), styrene (STY), and para-, ortho-, and meta-xylene (p-XYL, o-XYL, m-XYL). These organic solvents are often found together in various products such as paints, lacquers, inks, and adhesives (Kumai et al., 1983; Seedorf & Olsen, 1990), thus resulting in occupational exposures. Further, exposure of the general population to the mixture of these chemicals is also possible since they have been reported to occur near landfill sites (ATSDR, 2000).
PBTK MODELING OF CHEMICALS IN MIXTURES
211
MATERIALS AND METHODS In this study, a PBTK model for each chemical was used to predict the theoretically possible limits of blood concentration profile arising from metabolic interactions. The PBTK model predictions of the blood kinetics due to maximal interactions among mixture components were then compared with experimental data from exposure to various mixtures of DCM, BEN, TCE, TOL, PER, EBZ, STY, p-XYL, o-XYL, and m-XYL in the rat. PBTK Modeling Model Representation Many PBTK models for VOCs have been developed in the past (Krishnan & Andersen, 1994). Most of them described the rat as a set of four tissue compartments (liver, adipose tissues, poorly perfused tissues, and richly perfused tissues) along with a description of gas exchange at the level of the lung and a description of metabolism in liver (Figure 1). The PBTK model structure shown in Figure 1 was used for all of the 10 chemicals in this study. All model equations, except those describing metabolism, were obtained from Ramsey and Andersen (1984).
FIGURE 1. Conceptual representation of the PBTK model for volatile organic chemicals used in the present study. Qp and Qc refer to alveolar ventilation rate and cardiac output. Cinh, Cexh, Cv, and Ca refer to chemical concentration in inhaled air, exhaled air, venous blood, and arterial blood. Cvi and Qi refer to venous blood concentrations leaving tissue compartments and blood flow to tissues (i.e., f, adipose tissue; s, slowly perfused tissues, r, richly perfused tissues, and l, liver).
212
S. HADDAD ET AL.
Metabolism is often described as a saturable process [Eq. (1)] characterized by a maximal velocity (Vmax) and a Michaelis–Menten affinity constant (K m), or as a first order process [Eq. (2)] as follows: VmaxCvl RAM = ________ K m + Cvl
(1)
RAM = K fCvlVl
(2)
where RAM, Cvl, Kf, and Vl refer to the rate of the amount of chemical metabolized, chemical concentration in venous blood leaving liver, firstorder metabolism rate constant, and volume of liver, respectively. Metabolism can also be described using the hepatic extraction ratio (E) as follows: RAM = Q lECa
(3)
where Q l and Ca refer to liver blood flow rate and arterial blood concentration, respectively. The mathematical equivalence of Eqs. (1) and (3) was proven (Poulin & Krishnan, 1999). The use of Eq. (3) permits the simulation of the theoretical limits of the impact of metabolic interactions, since the E value can only range between 0 and 1. Considering metabolic interactions, enzyme induction can increase the value of E but to a maximum value of 1 (blood flow limitation). Similarly, inhibition of metabolism decreases the value of E but the value cannot be lower than 0. In the PBTK model for each of the 10 VOCs obtained from the literature, alternate descriptions of metabolism were incorporated within the models: (1) a Michaelis–Menten equation and/or a first-order term [Eqs. (1) and (2)] as described in the original publication, or (2) Eq. (3) facilitating the use of the numerical value of E. Simulations of the model with Eq. (1) and/or (2) predicted the kinetics of a chemical present alone. The simulations obtained using the PBTK model version containing Eq. (3) corresponded to the theoretical limits of blood concentrations attributed to metabolic interactions, since the numerical value of E was set to 0 and then to 1 during the simulations. Model Parameterization The physiological parameters for the rat PBTK models of all 10 mixture constituents were obtained from Arms and Travis (1988) (Table 1). The physicochemical parameters and metabolic rate constants for all 10 chemicals were obtained from the literature (Table 2). Model Simulation and Validation The PBTK model of each mixture constituent was written as a program using Advanced Continuous Simulation Language (ACSL, Pharsight Corporation, Mountain View, CA) and run on an IBM PC. The limits of the impact of metabolic interactions were simulated by setting E = 0 and E = 1, which yielded an envelope of blood
PBTK MODELING OF CHEMICALS IN MIXTURES
213
TABLE 1. Rat Physiological Parameters Used in the PBTK Models Parameter Body weight (kg)
Valuea 0.250
Compartment volumes (% body weight) Liver Adipose tissue Richly perfused tissues Poorly perfused tissues
4 7 5 75
Alveolar ventilation rate (L/h/kg)
15
Cardiac output (L/h/kg)
15
Compartment blood flow (% cardiac output) Liver Adipose tissue Richly perfused tissues Poorly perfused tissues
25 9 51 15
a
From Arms and Travis (1988).
concentrations for each chemical (Poulin & Krishnan, 1999). The simulated envelope was then compared to the simulations of a validated single chemical model as well as to experimental data obtained following 4-h inhalation exposure to mixtures of varying complexities (i.e., 2, 4, 5, 8, and/or 10 VOCs) in rats (Table 3). Data on blood concentrations following exposures to binary, quaternary and quinternary mixtures were obtained from Haddad et al. (1999b, 2000). Blood kinetic data following rat inhalation exposures to 8 or 10 chemicals were obtained in this study (see Experimental). Data on the effect of pretreatment on the kinetics of chemicals were also obtained and compared with the simulated envelope of blood concentration profiles. Experimental Chemicals and Animals DCM, BEN, TCE, TOL, PER, EBZ, STY, o-XYL, m-XYL, and p-XYL (99% purity) were obtained from Aldrich Chemicals (Milwaukee, WI), and used as supplied. Adult male Sprague-Dawley rats (Charles River Canada, St-Constant, Québec) weighing approximately 220–250 g were used in this study. Prior to use and during experiments, the animals were housed in a humidity- and temperature-controlled room with a 12-h light/dark cycle. The animals were not provided with food and water during inhalation exposures. Exposure Scheme and Analyses Two sets of exposures were conducted. The first exposure study was undertaken to evaluate the magnitude of change in the toxicokinetics of chemicals in mixtures during a single 4-h concomitant exposure (i.e., possibly involving inhibition of metabolism). The second set of exposures, focusing on the investigation of the additional impact of possible enzyme induction, involved the char-
214
S. HADDAD ET AL.
TABLE 2. Physico-Chemical and Biochemical Parameters Used in the Rat PBTK Models of DCM, BEN, TCE, TOL, PER, EBZ, STY, and XYL Chemicals in mixtures _________________________________________________________ Parameters
DCMa
BENb
TCEc
TOLb
PERd
EBZb
STYe
XYLb
Partition coefficients Blood:air Liver:air Fat:air RPT:air PPT:air
19.4 14.2 120 14.2 7.92
15 17 500 17 15
21.9 27.16 554.1 27.16 10.07
18 83.52 1020.6 83.52 27.72
18.9 70.3 2300 70.3 20.0
42.7 83.8 1556.0 160.3 26.0
40.0 108 2000 228 40
46.0 90.39 1859 90.9 41.9
Metabolic constants Vmax (mg/kg/h) Km (mg/L) Kf (h–1 kg–1)
4.28 0.40 2.00
11.00 0.25 —
3.44 0.13
Molecular weight (g/mol)
84.94
2.50 0.11 — 78.0
131.39
— 92.14
0.19 0.3 4.14 165.83
6.39 1.04 — 106.17
3.6 0.36 — 104.15
6.49 0.45 — 106.17
a
Obtained from Andersen et al. (1991). Obtained from Haddad et al. (1999b); parameters for all three isomers of xylenes were set equal to that of m-xylene based on observations of Tardif et al. (1997). c Obtained from Fisher et al. (1991). d Obtained from Ward et al. (1988). e Obtained from Ramsey and Andersen (1984). b
acterization of the kinetics of each of the 10 chemicals in rats previously exposed for 3 d to a mixture of the 10 chemicals. The first set of experiments involved placing of a group of 5 rats in an inhalation exposure chamber during 4 h (9:00 a.m.–1:00 p.m.) and exposing the group to one of the following mixtures: 1. A mixture of DCM, TOL, PER, EBZ, STY, o-XYL, m-XYL, and p-XYL (50 ppm each). 2. A mixture DCM, TOL, TRI, EBZ, STY, o-XYL, m-XYL, and p-XYL (50 ppm each). 3. A mixture of DCM, TCE, TOL, PER, BEN, EBZ, STY, o-XYL, m-XYL, and p-XYL (50 ppm each). In the second set of exposures, a group of 5 rats was pretreated by 4-h exposures for 3 consecutive days to the 10-chemical mixture in which the components were present at 50 ppm each. On d 4, the pretreated rats were exposed to the mixture of 10 chemicals (50 ppm each). Following all exposures, serial collection of blood from tail vein of individual animals was performed (0–120 min postexposure). The inhalation exposure system and the analytical method used for the quantitation of unchanged solvents
PBTK MODELING OF CHEMICALS IN MIXTURES
215
in rat blood were the same as those described in Tardif et al. (1997). The data on the blood concentration of each chemical obtained in this set of exposures were compared, using Student’s t-test, with the corresponding data from exposure to the same mixture without the 3-d preexposure. A p value of .05 was used to determine significant differences. RESULTS Figure 2 presents comparisons of PBTK model simulations with experimental data on the blood kinetics of the ten VOCs (DCM, TCE, TOL, PER, BEN, EBZ, STY, o-XYL, m-XYL, and p-XYL) in rats following inhalation exposure to these chemicals in mixtures of varying complexities. For each chemical, three simulations are presented. Two of these simulations were obtained by using Eq. (3) and setting the numerical value of E to 1 and then to 0, such that an envelope of the hypothetical limits of the impact of metabolic interactions on the blood kinetics of each chemical could be simulated. A third simulation was obtained for each chemical (except for o- and p-xylene, for which Vmax and Km were unavailable) using the description of saturable metabolism per Eq. (1). This simulation then corresponds to the blood kinetics of each chemical when no metabolic TABLE 3. Mixtures Investigated in This Study Mixture 1 2 3 4 5
6 7 8 9
10 11 12
a
Composition
Data source
m-XYL (50 ppm) + DCM (100 ppm) EBZ (50 ppm) + DCM (100 ppm) TOL (50 ppm) + DCM (100 ppm) BEN (50 ppm) + DCM (100 ppm) DCM (100 ppm) + BEN (50 ppm) + TOL (50 ppm) + EBZ (50 ppm) + m-XYL (50 ppm) BEN + TOL + EBZ + m-XYL (50 ppm each) DCM (100 ppm) + TOL (50 ppm) + EBZ (50 ppm) + m-XYL (50 ppm) TOL (50 ppm + preexposure) DCM + BEN + TRI + TOL + PER + EBZ + p-XYL + m-XYL + o-XYL + STY (50 ppm each + preexposure) DCM + TOL + PER + EBZ + p-XYL + m-XYL + o-XYL + STY (50 ppm each) DCM + TRI + TOL + EBZ + p-XYL + m-XYL + o-XYL + STY (50 ppm each) DCM + BEN + TRI + TOL + PER + EBZ + p-XYL + m-XYL + o-XYL + STY (50 ppm each)
Haddad et Haddad et Haddad et Haddad et Haddad et
al. al. al. al. al.
Symbola (2000) (2000) (2000) (2000) (2000)
Solid square Solid circle Shaded triangle Shaded diamond Shaded circle
Haddad et al. (1999b)
Asterisk
Haddad et al. (2000)
Plus sign
Present study Present study
Multiplication sign Open triangle
Present study
Open diamond
Present study
Open square
Present study
Open circle
Symbols are those used for representation in Figure 2.
216
S. HADDAD ET AL.
FIGURE 2. Comparison of PBTK model simulations (lines) with experimental data (symbols) on venous blood concentrations of dichloromethane (DCM), benzene (BEN), trichloroethylene (TCE), toluene (TOL), and tetrachloroethylene (PER) in rats following a 4-h exposure to these chemicals in various mixtures. The composition, component exposure concentrations and the data sources for all mixtures are provided in Table 3. The PBTK model simulations were obtained using (1) a description of saturable hepatic metabolism (———), (2) a hepatic extraction ratio set to 0 (———), and (3) a hepatic extraction ratio of 1 (– – –).
interactions occur, that is, kinetics of individual chemicals as reported by previous validated modeling efforts of Ramsey and Andersen (1984), Ward et al. (1988), Andersen et al. (1991), Fisher et al. (1991), and Haddad et al. (1999b). For all chemicals, the simulation lines obtained using E = 1 and E = 0 formed the boundary limits, whereas the one obtained using Vmax and Km
PBTK MODELING OF CHEMICALS IN MIXTURES
217
values was somewhat in between. Also, for all chemicals except PER, the kinetic data from mixture exposures were within the simulated boundaries of blood concentrations. For chemicals that have a fair amount of available kinetic data (i.e., BEN, TOL, EBZ, and m-XYL), it can be seen that with increasing mixture complexity, the impact on their blood kinetics is progressively more important (i.e., blood concentrations of unchanged
FIGURE 2. (Continued) Comparison of PBTK model simulations (lines) with experimental data (symbols) on venous blood concentrations of ethylbenzene (EBZ), styrene (STY), and para-, ortho-, and meta-xylene (p-XYL, o-XYL, m-XYL) in rats following a 4-h exposure to these chemicals in various mixtures. The composition, component exposure concentrations and the data sources for all mixtures are provided in Table 3. The PBTK model simulations were obtained using (1) a description of saturable hepatic metabolism (———), (2) a hepatic extraction ratio set to 0 (———), and (3) a hepatic extraction ratio of 1 (– – –).
218
S. HADDAD ET AL.
parent chemicals increase with mixture complexity). These observations are consistent with the occurrence of metabolic inhibition among the chemicals in the mixture. Regardless of the complexity of the mixtures, the blood concentration profile of individual chemicals (except PER) did not exceed their theoretical maximal profile simulated by setting the E value to 0 in the PBTK model (to reflect maximal impact of inhibition of hepatic metabolism). As can be seen, the simulations for PER underestimate the experimental data by a factor of 1.5–2.5.
FIGURE 3. Comparison of experimental data (mean ± SD, n = 5) on blood kinetics of dichloromethane (DCM), benzene (BEN), trichloroethylene (TCE), toluene (TOL), and tetrachloroethylene (PER) in rats following a 4-h inhalation exposure to 50 ppm of each of these chemicals in a 10-component mixture, with (M ) or without (L ) preexposure to the same mixture for 3 d. Statistically significant differences (p < .05) are indicated with an asterisk.
PBTK MODELING OF CHEMICALS IN MIXTURES
219
FIGURE 3. (Continued) Comparison of experimental data (mean ± SD, n = 5) on blood kinetics of ethylbenzene (EBZ), styrene (STY), and para-, ortho-, and meta-xylene (p-XYL, o-XYL, m-XYL) in rats following a 4-h inhalation exposure to 50 ppm of each of these chemicals in a 10-component mixture, with (M ) or without (L ) preexposure to the same mixture for 3 d. Statistically significant differences (p < .05) are indicated with an asterisk.
The toxicokinetics of all 10 chemicals were evaluated in rats exposed to the mixture of these chemicals 24 h after a 3-d preexposure to the same mixture. Since E approaches zero for all these chemicals during coexposures (due to inhibition of metabolism, as can be seen in Figure 2), the effect of enzyme induction is likely to be apparent under such a “sequential plus simultaneous” exposure condition. Figure 3 presents a comparison of the kinetics of each of the 10 chemicals observed with or without the 3-d
220
S. HADDAD ET AL.
preexposure. For all chemicals except PER, there seemed to be a systematic decrease in venous concentrations (i.e., greater metabolism due to induction in the pretreated groups). For all chemicals, there were no statistically significant differences (p < .05) in blood concentrations between the 2 groups during the first sampling point, that is, 30 min postexposure. At the successive sampling points, even though significant differences were observed for some chemicals (Figure 3), the quantitative differences between the pretreated and the non-pretreated groups were minimal. DISCUSSION Toxicokinetic (TK) interactions may result in altered blood and tissue concentration profiles of chemicals in mixtures. The magnitude of such changes depends primarily on the relative exposure concentrations and biochemical rate constants of the mixture components, and the quantitative nature of the interaction mechanisms. A number of TK interactions occur at the metabolism level (Krishnan & Brodeur, 1991), particularly due to an induction or inhibition of cytochrome P-450-mediated metabolism of one substance by another in the mixture. The consequences of metabolic inhibition in binary mixtures have been modeled using a PBTK approach (Krishnan et al., 1994; Simmons, 1996). Recent studies by Tardif et al. (1997) and Haddad et al. (1999b) demonstrated the feasibility of modeling the kinetics of chemicals in more complex mixtures. To develop such mechanistic TK models for complex mixtures, however, all interactions at the binary level should be characterized qualitatively and quantitatively. Due to scarce resources, it is preferable to attempt to identify a priori the theoretical limit to which such TK interactions are likely to alter the blood concentration profiles during mixed exposures. The present study demonstrated that this goal is attainable using PBTK models in which the numerical value of E is set to 1 or to 0. The data collected in the present study indicate that the kinetics of each of the 10 VOCs was modified during coexposure to other chemicals. The extent of modification, reflected by increasing blood concentration of unchanged parent chemicals, was more marked with increasing mixture complexity. This observation is analogous to the impact of several inhibitors on a single substrate, even though in the present study the efficacy of each inhibitor is not known. Available data suggest that all 10 VOCs investigated in the present study are proven or likely substrates of cytochrome P-450 2E1 (Toftgard & Nielsen, 1982; Guengerich et al., 1991; Nakajima et al., 1989, 1990, 1991, 1992, 1993; Liira et al., 1991; Tassaneeyakul et al., 1996). In 9 cases out of 10, the experimental venous blood concentrations never crossed the upper boundary of the simulations (obtained using an E value of 0). This is consistent with the expectation that, with increasing number of metabolic inhibitors, the E of a substance should decrease but cannot be lower than 0. Even though the specific nature of each metabolic
PBTK MODELING OF CHEMICALS IN MIXTURES
221
interaction within this mixture is unknown, it is reasonable to use the PBTK modeling approach to identify the theoretical maximal impact of metabolic interactions among chemicals in mixtures. Conceptually, such an approach is applicable for predicting the impact of maximal change in other model parameters (e.g., breathing rates) likely to occur during mixed exposures on the kinetics of chemicals in mixtures. In the case of PER, the PBTK model underestimated the experimental data obtained from mixture exposure by a factor of 1.5 to 2.5. Since the previously validated PBTK model for this chemical was used, no experimental verification of either the model parameters or the model simulations of single chemical exposures was undertaken in the present study. It is known that the blood:air partition coefficient is the most sensitive parameter in the PER PBTK model (Reitz et al., 1996). Any change, however small it may be, in blood:air partition coefficient is therefore likely to influence the kinetics of PER in exposed animals. The fact that the PER kinetic profiles observed during mixed exposures were outside the simulated boundaries suggests that either the blood:air partition coefficient of PER is modified during mixed exposures or sim ply the individual chemical model for PER needs to be refined for its blood:air partition coefficient to adequately simulate its kinetics in the animals used in the present study. It is possible that the blood:air partition coefficient of certain chemicals is reduced in the presence of other chemicals due to saturation of binding sites in hemoglobin. However, such changes in blood:air partition coefficients of VOCs investigated in the present study are unlikely to occur below a total mixture exposure concentration of several thousand parts per million (Poulin & Krishnan, 1996; Beliveau et al., in press). For highly extracted chemicals (E ³ 0.9), metabolic inhibition will have a greater impact, whereas for more poorly extracted chemicals, such as PER (E = 0.14), metabolic induction is anticipated to have a significant impact on the blood kinetic profiles. Given that most of the VOCs chosen for the present study are highly extracted chemicals, metabolic inhibition and not induction is likely to have a greater impact during mixed exposure scenarios. The results of the present study support these observations, in that the 3-d preexposure caused a significant but very small change in the venous blood concentrations of most VOCs. This observation is also in line with the results of previously published studies on the temporal change in cytochrome P-450 2E1 expression. Bergeron et al. (1999) observed that when rats were exposed daily to an aromatic hydrocarbon (EBZ), the expression of cytochrome P-450-2E1 increased after d 1 of exposure and returned to normal levels after d 2 of exposure. This observation would also explain why the toxicokinetics of all 10 VOCs, particularly the poorly extracted PER, did not change significantly following a 3-d preexposure. Overall, this study has validated a pragmatic way of predicting the maximal impact of metabolic interactions on the kinetics of chemicals in
222
S. HADDAD ET AL.
mixtures. The information on the altered kinetics and tissue dose of chemicals can be used in mixture risk-assessment methodologies based on internal doses (Haddad & Krishnan, 1998; Haddad et al., 1999a). In light of the fact that it is impractical to assess all possible interactions occurring in mixtures, this modeling methodology should be useful for identifying the chemicals for which metabolic interactions are likely to be important in the context of mixed exposures and mixture risk assessment. REFERENCES Agency for Toxic Substances and Disease Registry. 2000. Internet HazDat Database. Andersen, M. E., Clewell, H. J. III, Gargas, M. L., MacNaughton, M. G., Reitz, R. H., Nolan, R. J., and McKenna, M. J. 1991. Physiologically based pharmacokinetic modeling with dichloromethane, its metabolite, carbon monoxide, and blood carboxyhemoglobin in rats and humans. Toxicol. Appl. Pharmacol. 108:14–27. Arms, A. D., and Travis, C. C. 1988. Reference Physiological Parameters in Pharmacokinetic Modeling. Washington, DC: Office of Risk Analysis, U.S. EPA. NTIS PB 88-196019. Beliveau, M., Charest-Tardif, G., and Krishnan, K. In press. Blood:air partition coefficients of individual and mixtures of trihalomethanes. Chemosphere. Bergeron, R. M., Desai, K., Serron, S. C., Cawley, G. F., Eyer, C. S., and Backes, W. L. 1999. Changes in the expression of cytochrome P450s 2B1, 2B2, 2E1, and 2C11 in response to daily aromatic hydrocarbon treatment. Toxicol. Appl. Pharmacol. 157:1–8. Fisher, J. W., Gargas, M. L., Allen, B. C., and Andersen, M. E. 1991. Physiologically based pharmacokinetic modeling with trichloroethylene and its metabolite, trichloroacetic acid, in the rat and mouse. Toxicol. Appl. Pharmacol. 109:183–195. Guengerich, F. P., Kim, D. H., and Iwasaki, M. 1991. Role of human cytochrome P-450 IIE1 in the oxidation of many low molecular weight cancer suspect. Chem. Res. Toxicol. 4:168–179. Haddad, S., and Krishnan K. 1998. Physiological modeling of toxicokinetic interactions: Implications for mixture risk assessment. Environ. Health Perspect. 106 (suppl. 6):1377–1384. Haddad, S., Tardif, R., Viau, C., and Krishnan, K. 1999a. A modeling approach to account for toxicokinetic interaction in the calculation of biological hazard index for chemical mixtures. Toxicol. Lett. 108:303–308. Haddad, S., Charest-Tardif, G., Tardif, R., and Krishnan, K. 1999b. Physiological modeling of the toxicokinetic interactions in a quaternary mixture of aromatic hydrocarbons. Toxicol. Appl. Pharmacol. 161:249–257. Haddad, S., Charest-Tardif, G., Tardif R., and Krishnan, K. 2000. Validation of a physiological modeling approach for extrapolating the occurrence of interactions from simple to complex chemical mixtures. Toxicol. Appl. Pharmacol., in press. Krishnan, K., and Andersen, M. E. 1994. Physiologically based pharmacokinetic modeling in toxicology. In Principles and methods in toxicology, 3rd ed., ed. W. Hayes, pp. 149–187. New York: Raven Press. Krishnan, K., and Brodeur, J. 1991. Toxicological consequences of combined exposure to environmental pollutants. Arch. Complex Environ. Stud. 3:1–106. Kumai, M., Koizumi, A., Saito, K., Sakurai, H., Inoue, T., Takeuchi, Y., Hara, I., Ogata, M., Matsushita, T., and Ikeda, M. 1983. A nationwide survey on organic solvent components in various solvent products: Part 2. Heterogeneous products such as paints, inks, and adhesives. Ind. Health 21:185–197. Liira, J., Elovaara, R., Raunio, H., Riihimaki, V., and Engstrom, K. 1991. Metabolic interaction and disposition of methylethyl ketone and m-xylene in rats at single and repeated exposures. Xenobiotica 21:53–65. Nakajima, T., Elovaara, E., Park, S. S., Gelboin, H. V., Hietanen, E., and Vainio, H. 1989. Immuno-
PBTK MODELING OF CHEMICALS IN MIXTURES
223
chemical characterization of P-450 isozymes responsible for benzene oxidation in the rat liver. Carcinogenesis 10:1713–1717. Nakajima, T., Elovaara, E., Park, S. S., Gelboin, H. V., Hietanen, E., and Vainio, H. 1990. Monoclonal antibody directed characterization of benzene, ethoxyresorufin and pentoxyresorufin metabolism in rat liver microsomes. Biochem. Pharmacol. 40:1255–1261. Nakajima, T., Wang, R., Elovaara, E., Park, S. S., Gelboin, H. V., Hiltman, E., and Vainio, H. 1991. Monoclonal antibody directed characterization of cytochrome P-450 isozymes responsible for toluene metabolism in rat liver. Biochem. Pharmacol. 41:395–404. Nakajima, T., Wang, R. S., Elovaara, E., Park, S. S., Gelboin, H. V., and Vainio, H. 1992. A comparative study on the contribution of cytochrome P450 isozymes to metabolism of benzene, toluene and trichloroethylene in rat liver. Biochem. Pharmacol. 43:251–257. Nakajima, T., Wang, R. S., Elovaara, E., Park, S. S., Gelboin, H. V., and Vainio, H. 1993. Cytochrome P450-related differences between rats and mice in the metabolism of benzene, toluene and trichloroethylene in liver microsomes. Biochem. Pharmacol. 45:1079–1085. Poulin, P., and Krishnan, K. 1996. A mechanistic algorithm for predicting blood:air partition coefficients of organic chemicals with the consideration of reversible binding in hemoglobin. Toxicol. Appl. Pharmacol. 136:131–137. Poulin, P., and Krishnan, K. 1999. Molecular structure-based prediction of the toxicokinetics of inhaled vapors in humans. Int. J. Toxicol. 18:7–18. Ramsey, J. C., and Andersen, M. E. 1984. A physiologically-based description of the inhalation pharmacokinetics of styrene in rats and humans. Toxicol. Appl. Pharmacol. 73:159–175. Reitz, R. H., Gargas, M. L., Mendrala, A. I., and Schumann, A. 1996. In vivo and in vitro studies of perchloroethylene metabolism for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol. Appl. Pharmacol. 136:289–306. Seedorf, L., and Olsen, E. 1990. Exposure to organic solvents—I. A survey on the use of solvents. Ann. Occup. Hyg. 34:371–378. Simmons, J. E. 1996. Application of physiologically based pharmacokinetic modelling to combination toxicology. Food Chem. Toxicol. 34:1067–1673. Tardif, R., Charest-Tardif, G., Brodeur, J., and Krishnan, K. 1997. Physiologically-based pharmacokinetic modeling of a ternary mixture of alkyl benzenes in rats and humans. Toxicol. Appl. Pharmacol. 144:120–134. Tassaneeyakul, W., Birkett, D. J., Edwards, J. W., Veronese, M. E., Tassaneeyakul, W., Tukeg, R. H., and Miners, J. O. 1996. Human cytochrome P450 isoform specificity in the regioselective metabolism of toluene and o-, m- and p-xylene. J. Pharmacol. Exp. Ther. 276:101–108. Toftgard, R., and Nielsen, O. G. 1982. Effect of xylene and xylene isomers on cytochrome P-450 and in vitro enzymatic activities in rat liver, kidney and lung. Arch. Toxicol. 23:197–212. Ward, R. C., Travis, C. C., Hetrick, D. M., Andersen, M. E., and Gargas, M. L. 1988. Pharmacokinetics of tetrachloroethylene. Toxicol. Appl. Pharmacol. 93:108–117.
