- Email: [email protected]
Metabolic transformation of halogenated and other alkenes — a theoretical approach. Estimation of metabolic reactivities for in vivo conditions
Toxicology
letters
ELSEVIER
Toxicology Letters 75 (1995) 217-223
Metabolic transformation of halogenated and other alkenes a theoretical approach. Estimation of metabolic reactivities for in vivo conditions Gy. A. CsanBdy*“‘*, R.J. LaibbT2, J.G. Filser” aGSF-Institut ftir Toxikologie. Postfach 1129. 85758 Oberschlei$heim. Germany blnstitut fir Arbeitsphysiologie an der Vniversitiit Dortmund, Ardeystr. 67, D-44139 Dortmund, Germany
Received 5 October 1993; revision received 7 July 1994;accepted 7 July 1994 Abstract Olefinic hydrocarbons are metabolized in vivo by cytochrome P450-dependent monooxygenases to the corresponding epoxides. The maximum in vivo metabolic rate, which is an important toxicokinetic parameter, has been used to define the apparent rate constant (k,,,) describing in vivo metabolic reactivity of alkenes. To derive kapp,the metabolic rate normalized per body weight was divided by the corresponding average alkene concentration in the body at saturation conditions of 90%. Toxicokinetic data obtained in rats for 13 compounds (ethene, I-fluoroethene, 1, ldifluoroethene, 1-chloroethene, 1,I-dichloroethene, cis- 1,Zdichloroethene, trans- 1,Zdichloroethene, 1,l ,Ztrichloroethene, perchloroethene, propene, isoprene, 1,3-butadiene and styrene) have been used to calculate k,, values. A theoretical model, based on the assumption that in vivo epoxidation can be described as a cytochrome P450-mediated electrophilic reaction, has been developed. Using the olefinic hydrocarbons as an example it has been shown that kapp can be explained solely by the following molecular parameters: ionization potential, dipole moment and r-electron density. These molecular parameters were calculated by a quantum chemical method or were taken from the literature. Furthermore, the model was tested also by predicting kapp for isobutene, an alkene which was not used for the model development. The predicted value of kapp a g rees with the one derived experimentally, demonstrating that molecular parameters of halogenated and other alkenes can be used to predict in vivo metabolic reactivity. The model presented here is a first contribution to the ultimate goal to predict toxicokinetic parameters for in vivo conditions based on physicochemical parameters of enzymes and compounds exclusively. Toxicokinetics; Quantitative structure activity relationship Ethene; Propene; 1,3-Butadiene; Isoprene; Isobutene; Styrene
Keyword:
(QSAR); Metabolism; Halogenated alkenes;
1. Introduction
* Corresponding author. ’ On leave from: Central Research Institute of Chemistry, Hungarian Academy of Sciences, Budapest, Hungary. 2 Present address: c/o :Uinisterium fiir Umwelt und Gesundheit, Kaiser-Friedrich Str. 7, 55116, Mainz, Germany.
The mutagenic and carcinogenic effects of alkenes are based on their metabolic transformation to reactive intermediates. Biotransformation of these alkenes is generally initiated by a cytochrome P450-dependent oxidation, in which the
0378-4274/95/$09.50 0 1995 Elsevier Science Ireland Ltd. All rights reserved SD1 0378-4274(94)03 174-6
218
Gy.A. Csamidy et al. / Toxicol. Lett. 75 (1995) 217-223
corresponding epoxides are formed. In various theoretical studies, see for example [l-41, attempts have been made to establish a causal relationship between the mutagenic and carcinogenic potential observed from the parent compounds and the chemical structure of metabolically derived epoxides. However, the initial step of biotransformation from the alkenes to the epoxides has not been considered by any theoretical model. A model successfully describing the carcinogenic potential of alkenes should incorporate information about both the reactivity of the epoxides and the metabolic transformation rates of the parent alkenes. The aim of this work was to develop a model based on a quantitative structure activity relationship (QSAR) describing the metabolic rates of fluorinated, chlorinated and other alkenes in dependence on the molecular parameters dipole moment, ionization potential and n-electron density. 2. Methods 2.1. Metabolic reactivity In rats, the metabolic rates of different volatile compounds have been investigated experimentally by inhalation pharmacokinetic procedures based on a two-compartment model [5]. Metabolism of these compounds is saturable and can be described according to Michaelis-Menten. Such a kinetic behaviour is characterized by 2 different regions if the rate of metabolism is plotted over the concentration in the gas phase, under steady state conditions [5]: at low exposure concentration rates of metabolism are directly proportional to the exposure concentrations. In this concentration range physiological factors might limit the rate of metabolism measured in vivo. Therefore, it is difficult to distinguish if the measured metabolic rate results from the enzyme affinity or from transport processes to the metabolizing enzymes. However, at high concentrations the rate of biotransformation reaches asymptotically a maximum value with increasing exposure concentrations. This value depends only on the amount of enzyme and on the rate of disintegration of the enzyme-substratecomplex and therefore can be used as a measure for the reactivity. Consequently, the per body
weight normalized maximum metabolic rate (V,,) is appropriate for a structure activity investigation. In quantum chemical calculations rate constants are connected with energy quantity. Therefore, V,,,,, having unit of amount per time per body weight has to be converted to an apparent first order metabolic rate constant (k,,,) by Eq. 1 dividing V,, by the average concentration of the substrate in the animal compartment (C,). Maximum metabolic rates are reached only at inlinitively high substrate concentrations (C, - Q)), which would lead to zero values for kapp (Eq. 1). To obtain well-defined numerical values for kapp, saturation conditions of 90% have been used in the calculations:
(1) The calculation of the average concentration of the substrate in the animal compartment related to 90% saturation (C$,& is given in detail in the Appendix. The VmX and C,,, values with the corresponding kaPPvalues are listed in Table 1. In the following investigations kaPPwas considered to be the dependent variable. 2.2. Model development The model was based on the assumption that metabolic epoxidation can be described as a cytochrome P450-mediated electrophilic reaction. This is in agreement with experimental data [ 14,151 and also correlates with observations related to non-enzyme mediated epoxidation reactions [16-181. The chemical reactivity, i.e., the interaction between substrate and enzyme can be characterized by the so-called interaction energy (AE) [19], which is in general the sum of electrostatic, solvolitic and covalent terms counting for the different sources of interaction between substrate and enzyme. The influence of the first 2 terms will not be investigated in more detail because the alkenes are not charged and furthermore the reaction takes place in the same environment. The contribution of the covalent term to AE can be approximated as a fraction, where a product of the orbital occupancy and of the orbital availability is in the nominator and a difference of average or-
Gy.A. Csamidy et al. / Toxicol. Left. 75 (1995) 217-223
219
Table 1 Toxicokinetic parameters of alkenes: maximum metabolic rates ( Vmax)of different alkenes experimentally determined in rats, average concentration of the compound in the organism at metabolic saturation of 90% (CS,w, ) and corresponding apparent first order metabolic rate constant (k,,,) Compound
Ethene I-Fluoroethene l,l-Difluoroethene 1Xhloroethene 1,1-Dichloroethene c&1,2Dichloroethene tram- I ,2-Dichloroethene I, I ,2_Trichloroethene Perchloroethene Propene Isoprene 1,3-Butadiene Styrene
Vmlrx
C s.90%
kWP
Refs.
(Fmol kg-‘h-t)
(rmoM)
(ml g-‘h-t)
8.5 7 1.1 110 100 25 I 210 7 40 130 220 224
36.9 16.4 1.5 378.9 86.8 27 38.1 47.4 I323 95.5 226.5 450 360
0.207 0.385 0.132 0.261 1.037 0.833 0.165 3.990 0.005 0.377 0.517 0.440 0.56
9 6,7 637 677 637 6.7 6.7 6-7 637 II 10 8 28
bital energies is in the denominator 1191.Looking at the reaction from the part of the alkenes, the relectrons (i.e., the electrons of the highest occupied molecular orbital; HOMO) and from the part of the enzyme-oxygen complex the electrons of the lowest unoccupied molecular orbital (LUMO) are involved in the epoxidation process. The energy of HOMO can be substituted by the ionization potential (Ip), because the a-orbital of the investigated compounds is identical with their HOMO. The availability of the r-electrons was approximated by the normalized s-electron density (d(r)) calculated by a semi-empirical method of modified neglect of diatomic overlap (MNDO) (201. In the case of asymmetrically substituted ethenes, the carbon atom with the highest rrelectron density was included in the calculation of k app. The quantum chemical parameters (LUMO energy and orbital availability) of the enzyme are not accessible. However, their numerical values are assumed to be constant for a series of homologous compounds and to be related to the 2 parameters a (orbital availability) and b (LUMO energy). The chemical reactivity is assumed to be directly proportional to the interaction energy, and is approximated ;PS (a . d(r)J/( I Ip, - b I). The in vivo metabolic rate of the compounds is
postulated to be ,a product of the chemical (i.e., electrophilic) reactivity and a function describing the binding properties of the compounds to cytochrome P450-dependent enzymes. Due to our hypothesis the biotransformation occurs only if the chemical binds to the active site of the enzyme and the electron structure of the alkene enables the oxidation. The binding properties of the alkenes to the enzyme are described by their dipole moment (p). kapp,i=
Pi
.
alIpi_*d(n)i + 61
c
Taking into consideration further effects as electrostatic, solvolitic processes and also accounting for induced dipole moments, a constant term c was included into the equation. If solvolitic interactions would influence the value of kapp then c would be dependent on the molecular size. Therefore, we have investigated whether c depends on the molecular size using 2 different descriptors, the molecular volume (v) and a topological index, the so-called Wiener index (IV’) [22]. The molecular volume has been calculated for fluids with the aid of density, or for gases with the aid of the additive molar partial volumes [23]. The Wiener index
220
Gy.A. Csamidy et al. / Toxicol. Lett. 75 (1995) 217-223
represents the sum of all lengths of all paths in a given molecule and can be interpreted as a measure of the molecular volume [23]. The Wiener index was determined with an algorithm proposed by Senn [24] approximating the distance matrix with the bond length obtained by the MNDO method. To determine if the volume, the surface or the effective radius are important, different exponents have been introduced in the calculations. Also, a proposed exponent of l/4 [25] has been used for the transformation of the Wiener index. The molecular parameters of 14 alkenes are listed in Table 2. The numerical values of the parameters a, b and c have been determined by the least square method comparing model predictions with experimentally derived kappvalues of the first 13 alkenes. 3. Calculations All statistical calculations have been done with the program package ‘Statgraphics’, version 2.5 using a standard IBM AT computer. The MNDO calculations have been carried out with standard parameters [20] on a VAX 3200 machine.
4. Results and discussion A good correlation (R = 0.92; significantly different from zero at P < 0.01; according to R.A. Fisher) exists between the experimentally derived and calculated kapp values of the alkenes. The mean variance of the predicted values is around 85%, which further supports the derived relationship. The correlation still remains significant (P < 0.10; according to R.A. Fisher) if 1, I ,2-trichloroethene, having the highest k,, value (Fig. l), is deleted from the experimental set. The model parameters, determined by the least square method, are summarized in Table 3. According to the correlation coefficients (R; Table 3) the value of c is independent from the molecular size, because the inclusion of either the molecular volume or the Wiener index did not improve R. The influence of the molecular parameters p, d(r) and Ip on kapp has been investigated, too. The square of correlation coefficient (R2) gives the share of the variances explained by the model. Roughly, a value near to 1 indicates that the variances in kapp are explained by the model parameters. Deleting a single parameter at each time the
Table 2 Physicochemical parameters (c, d(x), Ip) and geometrical descriptors (V and IV) of the alkenes Compound
or (D)
d(x)
V (cm3)
W
IP (ev)
Ethene I-Fluoroethene 1,I-Difluoroethene I-Chloroethene 1,l -Dichloroethene cis-1,2_Dichloroethene rranr-1,2Dichloroethene 1,1,2-Trichloroethene Perchloroethene Propene Isoprene 1,3-Butadiene Styrene Isobutene
0.00
0.500
1.43 1.38 1.44 1.25 1.89 0.00 0.80 0.00 0.36 0.38 o.138 0.37 0.49
0.498 OS10 0.435 0.413 0.350 0.352 0.339 0.308 0.475 0.327 0.308 0.214 0.459
57.0 60.0 68.6 66.6 79.9 75.5 77.0 89.7 102.0 75.6 100.5 84.0 114.6 94.2
1.33 5.35 12.03 6.17 14.53 15.79 15.84 28.92 46.86 5.67 25.68 13.94 90.62 13.1
10.51 [29] 10.57 [30] 10.72 (291 10.18 [29] 10.00 (291 9.83 [29] 9.81 [29] 9.75 [31] 9.58 [31] 10.03 [29) 9.10 (32) 9.03 [32] 8.49 (321 9.85 [29]
Vhe theoretical dipole moment of rran.s-1,3-butadiene is zero; however the rotation barrier is so small (3.5 kcal/mol), that at room temperature 1,3-butadiene exists as a mixture of the cis and trans isomers. Therefore, the experimentally determined c differs from zero, see [33].
Gy.A. Csaruidy et al. / Toxicol. L&t. 75 (1995) 217-223
221
Table 4 Influence of physicochemical parameters on kapP given by the corresponding square of the correlation coefficient (R*) Model
Pi
Pi
fi’i
R*
a d(r)i Abs(Ip, - b)
+’
a Abs(Ipi - b)
+’
a d(rh
a d(*)i 6
kapp (talc.) Iml/g/hl
Fig. 1. Experimentally determined k,, plotted versus the predicted values obtained by the model. The solid line represents an exact linear relationship (R = l), whereas the numbers correspond to the followingchemicals:(1) 1,3-butadiene; (2) cis-1,2-dichloroethene; (3) trans-1,Zdichloroethene; (4) ethene; (5) isoprene; 116)perchloroethene; (7) propene; (8) 1,1,2-trichloroethene; (9) I-chloroethene; (10) l,l-dichloroethene; (11) I-fluoroethene; (12) 1,I-difluoroethene; (13) styrene.
contribution of this Iparameter can be characterized by the differences seen in the values of R2. Table 4 shows that ~1and Ip values contribute much more to the square of the correlation coefflcient than do the d(s) values. According to Table 4, the pi/lIpi - b I part plays the most important role in our model.
Table 3 Dependence of c on molecular size investigated by the least square method using parameters of a = 0.0937 and of b = 9.756 and data of 13 alkenes taken from Tables 1 and 2 Model
X
c=x c=x.v c=x- w
0.126 f (2.67 f (2.36 f (1.15 f 0.049 f 0.094 *
c=x.v~3
c=x c=x.
.
v”3 w’”
R 0.101 1.25)10-3 2.33)10-3 0.55)10_2 0.024 0.050
0.927 0.933 0.915 0.933 0.932 0.928
pi a d(r)/ + c
0.83
0.81
Abs(Ip, - b)
A&Ip, - b)
0.86
+’
0.28
0.03
Introducing the dipole moment as a weighting function to describe the binding properties of the alkenes to cytochrome P450 provides a link between the metabolic elimination rate of the chlorinated olefins and the symmetry rule, proposed by Bonse and Henschler for the epoxides formed [l]. The epoxides of I-chloroethene, 1, Idichloroethene and 1,l ,Ztrichloroethene were characterized by Bonse and Henschler as ‘asymmetric’ and mutagenic. In contrast, the epoxides of perchloroethene and tram- 1,Zdichloroethene were defined as ‘symmetric’ and not mutagenic. Our calculations show that symmetric alkenes have a dipole moment from zero and therefore their metabolic rate, i.e. the formation of the epoxides is limited by the value of c, whereas the metabolic rate of asymmetric alkenes is higher. Consequently, higher amounts of the reactive epoxides per time unit will be produced from the asymmetric chlorinated alkene precursors than from the symmetric ones, counting for differences seen in the mutagenic .potency. The lower metabolic transformation rate of tram- versus cis- 1,2-dichloroethene in rats has been explained with the higher effectiveness of the tram epoxide as a suicide inhibitor of cytochrome P450 [26]. In vitro findings are inconsistent with the suicide inhibitor theory. Costa and Ivanetich
222
Gy.A. Csanddy et al. / Toxicol. L&t. 75 (1995) 217-223
[27] have demonstrated that cis- and trans-1,2dichloroethene show comparable activities with respect to cytochrome P450 inactivation in microsomal incubations. From the in vivo data base used in this study metabolic interactions between the produced epoxide and the parent alkene cannot be excluded unequivocally. However, our calculations predict a much lower metabolic transformation rate for the ‘symmetric’ tram isomer. Therefore, no suicide inhibition of the enzyme by the produced epoxide seems to be necessary to explain differences seen in metabolic rates of cis and tram isomers. Also a reversible competitive inhibition would not influence the numerical value of V,,, and therefore the model predictions are legitimate. The model describes the reactivity of styrene appropriately. There is a factor of 4 between the calculated (based on Eq. 2) and the experimentally determined rates of styrene metabolism. To establish the model, the general assumption was made that all compounds were metabolized by the same enzyme or enzyme family and that their binding properties depend on their dipole moments. Obviously this is also the case for such aromatic compounds as styrene. The model was also used to predict the metabolic rate of isobutene. The calculated kapp of 0.35 ml. h-‘g-l (based on Eq. 2 and the model variables listed in Table 2) is in agreement with the experimentally determined value of 0.11 ml. h-‘g-i (according to = 320 pmol. kg-‘h-l and C, = 2900 pmol/l &. In this work an approach has been presented to calculate in vivo reactivity of alkenes based on a theoretical model and using experimentally determined in vivo kinetic parameters of homologue compounds. Such a quantitative structure activity relationship as presented here might lead to a better understanding of the molecular mechanisms of biotransformation and in the long run might lead to a reduction of animal experiments.
Berufsgenossenschaft der Chemischen Industrie (Heidelberg, La 51511-3) is gratefully acknowledged. Appendix According to the two-compartment model [5] the average concentration in the organism (y2) at steady state is proportional to the concentration in the gas phase (_vl):
y2 = fit.yl
= yl*
Vlk12 V2k21 + V2keP
tit is the bioaccumulation factor at steady state, which depends on kinetic parameters as the constant clearance of inhalation (Vlk12), exhalation (V2k21) and on the concentration dependent clearance of elimination (V2kel*). The metabolic rate at steady state assuming a Michaelis-Menten kinetics can be described as dNel
V,,,.K.st.yl
dt
K,,, + fit-y1
-=
The right side of the equation equals 0.9. Vmaxif metabolic saturation of 90% is achieved. The solution for ~1~~ can be given as, according to [34]:
ylw,, = 0.9.
V,,,
+ 10-K,,,. V2k21
Vlk12
The corresponding average concentration in the organism is: C,,g, = ~2~~ = Kitwh. ~1~~. The necessary pharmacokinetic parameters involved in the calculations have been taken from the references listed in Table 2. References
Acknowledgements
111Bonse, G. and Henschler, D. (1976) Chemical reactivity,
We thank Dr. G. Csonka (Budapest) for providing the MNDO program and Drs. P. Politzer (New Orleans) and P.E. Kreuzer (Neuherberg) for valuable discussions. The financial support of the
biotransformation and toxicity of polychlorinated aliphatic compounds. CRC Crit. Rev. Toxicol. 4, 395-409. 121Jones, R.B. and Mackrodt, W.C. (1983) Structuregenotoxicity relationship for aliphatic epoxides. Biochem. Pharmacol. 32, 2359-2362. [31 Politzer, P. and Laurence, P.R. (1984) Halogenated hy-
Gy.A. Csamidy et al. / Toxicoi. Lett. 75 (1995) 217-223
[4]
[5]
[6] [7]
[8]
[9]
[IO]
[l l]
[I21
[13]
[14]
[15]
[16]
[17]
[18]
drocarbon epoxides: factors underlying biological activity. Int. J. Quant. Chem. Quant. Biol. Symp. 11, 155-166. Politxer, P. and Laurence, P.R. (1984) Relationships between the electrostatic potential, epoxide hydrase inhibition and carcinogenicity for some hydrocarbon and halogenated hydrocarbon epoxides. Carcinogenesis 5, 845-848. Filser, LG. (19912)The closed chamber technique uptake, endogenous production, excretion, steady-state kinetics and rate:3 of metabolism of gases and vapors. Arch. Toxicol. 66, l-10. Filser, J.G. and Bolt, H.M. (1979) Pharmacokinetics of halogenated ethylens in rats. Arch. Toxicol. 42, 123-136. Fiber, J.G. and Bolt, H.M. (1981) Inhalation pharmacokinetics based on gas uptake studies. I. Improvement of kinetic models. Arch. Toxicol. 47, 279-292. Kreiling, R., Laib, R.J., Fiber, J.G. and Bolt, H.M. (1986) Species di:fferences in butadiene metabolism between mice and rats evaluated by inhalation pharmacokinetics. Arch. Toxicol. 58, 235-238. Bolt, H.M. and Fiber, J.G. (1987) Kinetics and disposition in toxicology. Example: carcinogenic risk estimate for ethylene. Arclh. Toxicol. 60, 73-76. Peter, H., Wiegand, H.J., Bolt, H.M., Greim, H., Walter, G., Berg, M. and Fiber, J.G. (1987) Pharmacokinetics of isoprene in mice and rats. Toxicol. Lett. 36, 9-14. Golka, K., Peter. H., Denk, B. and Filser, J.G. (1989) Pharmacokinetics of propylene and its reactive metabolite propylene oxide in Sprague-Dawley rats. Arch. Toxicol. Suppl. 13, 2’40-242. Csanady, G.A., Freise, D., Denk, B., Filser, J.G., Cornet, M., Rogiers, V. and Laib, R.J. (1991) Investigation of species differences in isobutene (2-methylpropene) metabolism between mice and rats. Arch. Toxicol. 65, 100-105. Kessler, W., Denk, B. and Fiber, J.G. (1989) Speciesspecific inhalation pharmacokinetics of 2nitropropane, methyl-ethyl-ketone and n-hexane. In: BiologicallyBased Methods for Cancer Risk Assessment, Plenum Publishing Corparation, pp. 123-139. Castellino, A.J. and Bruice, T.C. (1988) Intermediates in the epoxidation of alkenes by cytochrome P-450 Model 1. cis-stilbene as a. mechanistic probe. J. Am. Chem. Sot. 110, 158-162. Garrison, J.M. and Bruice, T.C. (1989) Intermediates in the epoxidation of alkenes by cytochrome P-450 models. J. Am. Chem. Sot. 111, 191-198. Hanzlik, R.P. andt Shearer, G.O. (1975) Transition state structure for peracid epoxidation. J. Am. Chem. Sot. 97, 5231-5233. Plesnicar, B., Tasevski, M. and Azmon, A. (1978) The transition state for the epoxidation of ethylene with performic acid. J. Am. Chem. Sot. 100, 743-746. Swern, D. (1967) Epoxidation of ethylenic compound* Organic React. 7, 378-433.
223
[19] Fukui, K., Yonezawa, T. and Nagata, C. (1954) Theory of substitution in conjugated molecules. Bull. Chem. Sot. Japan 27, 423-427. [20] Dewar, M.J.S. and Thiel, W. (1977) Ground states of molecules 38. The MNDG method. J. Am. Chem. Sot. 99, 4899-4906. (211 Dewar, M.J.S., Zoebisch, E.G., Healy, E. and Stewart, J.F.P. (1985) AMl: a new general purpose quantum mechanical molecular model. J. Am. Chem. Sot. 107, 3902-3909. 1221 Wiener, H.J. (1947) Structural determination of paraffin boiling points. J. Am. Chem. Sot. 69, 17-20. [23] Barton, A.F.M. (1983) CRC Handbook of Solubility Parameters and Other Cohesion Parameters. CRC Press, Inc. Boca Raton, Florida. [24] Senn, P. (1988) The computation of the distance matrix and the Wiener index for graphs of arbitrary complexity with weighted vertices and edges. Comput. Chem. 12, 219-227. [25] RandiC, M., Hansen, P.J. and Jurs, P.C. (1988) Search for useful graph theoretical invariants of molecular structure. J. Chem. Inf. Comput. Sci. 28, 60-68. [26] Gargas, M.L. and Andersen, M.E. (1988) Physiologically based approaches for examining the pharmacokinetics of inhaled vapors. In: Toxicology of Lungs, Raven Press, New York, pp. 449-476. [27] Costa, K.A. and Ivanetich, K.M. (1982) The 1,2dichloroethylenes: their metabolism by hepatic cytochrome P-450 in vitro. B&hem. Pharmacol. 31, 20932102. [28] Fiber, J.G., Schwegler, U., Csanady, Gy.A., Greim, H., Kreuxer, P. and Kessler, W. (1993) Species-specific pharmacokinetics of styrene in rat and mouse. Arch. Toxicol. 67, 517-530. [29] Yamaxaki, T. and Kimura, K. (1975) Some rule consideration of molecular photoelectron spectrum. Substituted ethylenes. Bull. Chem. Sot. Japan 48, 1602-1607. [30] Reinke, D., Baumgartel, H., Cvitas, T., Klasnic, L. and Gusten, H. (1974) Vergleich der Photoelektronspektren und Photoionenspektren von Vinylfluorid. Ber. BunsenGesehschaft Phys. Chem. 78, 1145-l 147. [3l] Sandhu, J.S. (1972) Ionization potentials of chloroethylenes as determined by photoelectron spectroscopy. Indian J. Chem. 10, 667-668. [32] Hemdon, WC. (1980) Perturbation-graph theory III. Resonance energies from photoelectron spectra. Pure Appl. Chem. 52, 1459-1469. [33] Hannay, B. and Smith, C.P. (1943) Hyperconjugation in methyl butadienes as shown by dipole moments. J. Am. Chem. Sot. 65, 1931-1934. [34] Laib, R.J., Tucholski, M., Filser, J.G. and Csanady, Gy.A. (1992) Pharmacokinetic interaction between 1,3butadiene and styrene in Sprague-Dawley rats. Arch Toxicol. 66, 310-314.
letters
ELSEVIER
Toxicology Letters 75 (1995) 217-223
Metabolic transformation of halogenated and other alkenes a theoretical approach. Estimation of metabolic reactivities for in vivo conditions Gy. A. CsanBdy*“‘*, R.J. LaibbT2, J.G. Filser” aGSF-Institut ftir Toxikologie. Postfach 1129. 85758 Oberschlei$heim. Germany blnstitut fir Arbeitsphysiologie an der Vniversitiit Dortmund, Ardeystr. 67, D-44139 Dortmund, Germany
Received 5 October 1993; revision received 7 July 1994;accepted 7 July 1994 Abstract Olefinic hydrocarbons are metabolized in vivo by cytochrome P450-dependent monooxygenases to the corresponding epoxides. The maximum in vivo metabolic rate, which is an important toxicokinetic parameter, has been used to define the apparent rate constant (k,,,) describing in vivo metabolic reactivity of alkenes. To derive kapp,the metabolic rate normalized per body weight was divided by the corresponding average alkene concentration in the body at saturation conditions of 90%. Toxicokinetic data obtained in rats for 13 compounds (ethene, I-fluoroethene, 1, ldifluoroethene, 1-chloroethene, 1,I-dichloroethene, cis- 1,Zdichloroethene, trans- 1,Zdichloroethene, 1,l ,Ztrichloroethene, perchloroethene, propene, isoprene, 1,3-butadiene and styrene) have been used to calculate k,, values. A theoretical model, based on the assumption that in vivo epoxidation can be described as a cytochrome P450-mediated electrophilic reaction, has been developed. Using the olefinic hydrocarbons as an example it has been shown that kapp can be explained solely by the following molecular parameters: ionization potential, dipole moment and r-electron density. These molecular parameters were calculated by a quantum chemical method or were taken from the literature. Furthermore, the model was tested also by predicting kapp for isobutene, an alkene which was not used for the model development. The predicted value of kapp a g rees with the one derived experimentally, demonstrating that molecular parameters of halogenated and other alkenes can be used to predict in vivo metabolic reactivity. The model presented here is a first contribution to the ultimate goal to predict toxicokinetic parameters for in vivo conditions based on physicochemical parameters of enzymes and compounds exclusively. Toxicokinetics; Quantitative structure activity relationship Ethene; Propene; 1,3-Butadiene; Isoprene; Isobutene; Styrene
Keyword:
(QSAR); Metabolism; Halogenated alkenes;
1. Introduction
* Corresponding author. ’ On leave from: Central Research Institute of Chemistry, Hungarian Academy of Sciences, Budapest, Hungary. 2 Present address: c/o :Uinisterium fiir Umwelt und Gesundheit, Kaiser-Friedrich Str. 7, 55116, Mainz, Germany.
The mutagenic and carcinogenic effects of alkenes are based on their metabolic transformation to reactive intermediates. Biotransformation of these alkenes is generally initiated by a cytochrome P450-dependent oxidation, in which the
0378-4274/95/$09.50 0 1995 Elsevier Science Ireland Ltd. All rights reserved SD1 0378-4274(94)03 174-6
218
Gy.A. Csamidy et al. / Toxicol. Lett. 75 (1995) 217-223
corresponding epoxides are formed. In various theoretical studies, see for example [l-41, attempts have been made to establish a causal relationship between the mutagenic and carcinogenic potential observed from the parent compounds and the chemical structure of metabolically derived epoxides. However, the initial step of biotransformation from the alkenes to the epoxides has not been considered by any theoretical model. A model successfully describing the carcinogenic potential of alkenes should incorporate information about both the reactivity of the epoxides and the metabolic transformation rates of the parent alkenes. The aim of this work was to develop a model based on a quantitative structure activity relationship (QSAR) describing the metabolic rates of fluorinated, chlorinated and other alkenes in dependence on the molecular parameters dipole moment, ionization potential and n-electron density. 2. Methods 2.1. Metabolic reactivity In rats, the metabolic rates of different volatile compounds have been investigated experimentally by inhalation pharmacokinetic procedures based on a two-compartment model [5]. Metabolism of these compounds is saturable and can be described according to Michaelis-Menten. Such a kinetic behaviour is characterized by 2 different regions if the rate of metabolism is plotted over the concentration in the gas phase, under steady state conditions [5]: at low exposure concentration rates of metabolism are directly proportional to the exposure concentrations. In this concentration range physiological factors might limit the rate of metabolism measured in vivo. Therefore, it is difficult to distinguish if the measured metabolic rate results from the enzyme affinity or from transport processes to the metabolizing enzymes. However, at high concentrations the rate of biotransformation reaches asymptotically a maximum value with increasing exposure concentrations. This value depends only on the amount of enzyme and on the rate of disintegration of the enzyme-substratecomplex and therefore can be used as a measure for the reactivity. Consequently, the per body
weight normalized maximum metabolic rate (V,,) is appropriate for a structure activity investigation. In quantum chemical calculations rate constants are connected with energy quantity. Therefore, V,,,,, having unit of amount per time per body weight has to be converted to an apparent first order metabolic rate constant (k,,,) by Eq. 1 dividing V,, by the average concentration of the substrate in the animal compartment (C,). Maximum metabolic rates are reached only at inlinitively high substrate concentrations (C, - Q)), which would lead to zero values for kapp (Eq. 1). To obtain well-defined numerical values for kapp, saturation conditions of 90% have been used in the calculations:
(1) The calculation of the average concentration of the substrate in the animal compartment related to 90% saturation (C$,& is given in detail in the Appendix. The VmX and C,,, values with the corresponding kaPPvalues are listed in Table 1. In the following investigations kaPPwas considered to be the dependent variable. 2.2. Model development The model was based on the assumption that metabolic epoxidation can be described as a cytochrome P450-mediated electrophilic reaction. This is in agreement with experimental data [ 14,151 and also correlates with observations related to non-enzyme mediated epoxidation reactions [16-181. The chemical reactivity, i.e., the interaction between substrate and enzyme can be characterized by the so-called interaction energy (AE) [19], which is in general the sum of electrostatic, solvolitic and covalent terms counting for the different sources of interaction between substrate and enzyme. The influence of the first 2 terms will not be investigated in more detail because the alkenes are not charged and furthermore the reaction takes place in the same environment. The contribution of the covalent term to AE can be approximated as a fraction, where a product of the orbital occupancy and of the orbital availability is in the nominator and a difference of average or-
Gy.A. Csamidy et al. / Toxicol. Left. 75 (1995) 217-223
219
Table 1 Toxicokinetic parameters of alkenes: maximum metabolic rates ( Vmax)of different alkenes experimentally determined in rats, average concentration of the compound in the organism at metabolic saturation of 90% (CS,w, ) and corresponding apparent first order metabolic rate constant (k,,,) Compound
Ethene I-Fluoroethene l,l-Difluoroethene 1Xhloroethene 1,1-Dichloroethene c&1,2Dichloroethene tram- I ,2-Dichloroethene I, I ,2_Trichloroethene Perchloroethene Propene Isoprene 1,3-Butadiene Styrene
Vmlrx
C s.90%
kWP
Refs.
(Fmol kg-‘h-t)
(rmoM)
(ml g-‘h-t)
8.5 7 1.1 110 100 25 I 210 7 40 130 220 224
36.9 16.4 1.5 378.9 86.8 27 38.1 47.4 I323 95.5 226.5 450 360
0.207 0.385 0.132 0.261 1.037 0.833 0.165 3.990 0.005 0.377 0.517 0.440 0.56
9 6,7 637 677 637 6.7 6.7 6-7 637 II 10 8 28
bital energies is in the denominator 1191.Looking at the reaction from the part of the alkenes, the relectrons (i.e., the electrons of the highest occupied molecular orbital; HOMO) and from the part of the enzyme-oxygen complex the electrons of the lowest unoccupied molecular orbital (LUMO) are involved in the epoxidation process. The energy of HOMO can be substituted by the ionization potential (Ip), because the a-orbital of the investigated compounds is identical with their HOMO. The availability of the r-electrons was approximated by the normalized s-electron density (d(r)) calculated by a semi-empirical method of modified neglect of diatomic overlap (MNDO) (201. In the case of asymmetrically substituted ethenes, the carbon atom with the highest rrelectron density was included in the calculation of k app. The quantum chemical parameters (LUMO energy and orbital availability) of the enzyme are not accessible. However, their numerical values are assumed to be constant for a series of homologous compounds and to be related to the 2 parameters a (orbital availability) and b (LUMO energy). The chemical reactivity is assumed to be directly proportional to the interaction energy, and is approximated ;PS (a . d(r)J/( I Ip, - b I). The in vivo metabolic rate of the compounds is
postulated to be ,a product of the chemical (i.e., electrophilic) reactivity and a function describing the binding properties of the compounds to cytochrome P450-dependent enzymes. Due to our hypothesis the biotransformation occurs only if the chemical binds to the active site of the enzyme and the electron structure of the alkene enables the oxidation. The binding properties of the alkenes to the enzyme are described by their dipole moment (p). kapp,i=
Pi
.
alIpi_*d(n)i + 61
c
Taking into consideration further effects as electrostatic, solvolitic processes and also accounting for induced dipole moments, a constant term c was included into the equation. If solvolitic interactions would influence the value of kapp then c would be dependent on the molecular size. Therefore, we have investigated whether c depends on the molecular size using 2 different descriptors, the molecular volume (v) and a topological index, the so-called Wiener index (IV’) [22]. The molecular volume has been calculated for fluids with the aid of density, or for gases with the aid of the additive molar partial volumes [23]. The Wiener index
220
Gy.A. Csamidy et al. / Toxicol. Lett. 75 (1995) 217-223
represents the sum of all lengths of all paths in a given molecule and can be interpreted as a measure of the molecular volume [23]. The Wiener index was determined with an algorithm proposed by Senn [24] approximating the distance matrix with the bond length obtained by the MNDO method. To determine if the volume, the surface or the effective radius are important, different exponents have been introduced in the calculations. Also, a proposed exponent of l/4 [25] has been used for the transformation of the Wiener index. The molecular parameters of 14 alkenes are listed in Table 2. The numerical values of the parameters a, b and c have been determined by the least square method comparing model predictions with experimentally derived kappvalues of the first 13 alkenes. 3. Calculations All statistical calculations have been done with the program package ‘Statgraphics’, version 2.5 using a standard IBM AT computer. The MNDO calculations have been carried out with standard parameters [20] on a VAX 3200 machine.
4. Results and discussion A good correlation (R = 0.92; significantly different from zero at P < 0.01; according to R.A. Fisher) exists between the experimentally derived and calculated kapp values of the alkenes. The mean variance of the predicted values is around 85%, which further supports the derived relationship. The correlation still remains significant (P < 0.10; according to R.A. Fisher) if 1, I ,2-trichloroethene, having the highest k,, value (Fig. l), is deleted from the experimental set. The model parameters, determined by the least square method, are summarized in Table 3. According to the correlation coefficients (R; Table 3) the value of c is independent from the molecular size, because the inclusion of either the molecular volume or the Wiener index did not improve R. The influence of the molecular parameters p, d(r) and Ip on kapp has been investigated, too. The square of correlation coefficient (R2) gives the share of the variances explained by the model. Roughly, a value near to 1 indicates that the variances in kapp are explained by the model parameters. Deleting a single parameter at each time the
Table 2 Physicochemical parameters (c, d(x), Ip) and geometrical descriptors (V and IV) of the alkenes Compound
or (D)
d(x)
V (cm3)
W
IP (ev)
Ethene I-Fluoroethene 1,I-Difluoroethene I-Chloroethene 1,l -Dichloroethene cis-1,2_Dichloroethene rranr-1,2Dichloroethene 1,1,2-Trichloroethene Perchloroethene Propene Isoprene 1,3-Butadiene Styrene Isobutene
0.00
0.500
1.43 1.38 1.44 1.25 1.89 0.00 0.80 0.00 0.36 0.38 o.138 0.37 0.49
0.498 OS10 0.435 0.413 0.350 0.352 0.339 0.308 0.475 0.327 0.308 0.214 0.459
57.0 60.0 68.6 66.6 79.9 75.5 77.0 89.7 102.0 75.6 100.5 84.0 114.6 94.2
1.33 5.35 12.03 6.17 14.53 15.79 15.84 28.92 46.86 5.67 25.68 13.94 90.62 13.1
10.51 [29] 10.57 [30] 10.72 (291 10.18 [29] 10.00 (291 9.83 [29] 9.81 [29] 9.75 [31] 9.58 [31] 10.03 [29) 9.10 (32) 9.03 [32] 8.49 (321 9.85 [29]
Vhe theoretical dipole moment of rran.s-1,3-butadiene is zero; however the rotation barrier is so small (3.5 kcal/mol), that at room temperature 1,3-butadiene exists as a mixture of the cis and trans isomers. Therefore, the experimentally determined c differs from zero, see [33].
Gy.A. Csaruidy et al. / Toxicol. L&t. 75 (1995) 217-223
221
Table 4 Influence of physicochemical parameters on kapP given by the corresponding square of the correlation coefficient (R*) Model
Pi
Pi
fi’i
R*
a d(r)i Abs(Ip, - b)
+’
a Abs(Ipi - b)
+’
a d(rh
a d(*)i 6
kapp (talc.) Iml/g/hl
Fig. 1. Experimentally determined k,, plotted versus the predicted values obtained by the model. The solid line represents an exact linear relationship (R = l), whereas the numbers correspond to the followingchemicals:(1) 1,3-butadiene; (2) cis-1,2-dichloroethene; (3) trans-1,Zdichloroethene; (4) ethene; (5) isoprene; 116)perchloroethene; (7) propene; (8) 1,1,2-trichloroethene; (9) I-chloroethene; (10) l,l-dichloroethene; (11) I-fluoroethene; (12) 1,I-difluoroethene; (13) styrene.
contribution of this Iparameter can be characterized by the differences seen in the values of R2. Table 4 shows that ~1and Ip values contribute much more to the square of the correlation coefflcient than do the d(s) values. According to Table 4, the pi/lIpi - b I part plays the most important role in our model.
Table 3 Dependence of c on molecular size investigated by the least square method using parameters of a = 0.0937 and of b = 9.756 and data of 13 alkenes taken from Tables 1 and 2 Model
X
c=x c=x.v c=x- w
0.126 f (2.67 f (2.36 f (1.15 f 0.049 f 0.094 *
c=x.v~3
c=x c=x.
.
v”3 w’”
R 0.101 1.25)10-3 2.33)10-3 0.55)10_2 0.024 0.050
0.927 0.933 0.915 0.933 0.932 0.928
pi a d(r)/ + c
0.83
0.81
Abs(Ip, - b)
A&Ip, - b)
0.86
+’
0.28
0.03
Introducing the dipole moment as a weighting function to describe the binding properties of the alkenes to cytochrome P450 provides a link between the metabolic elimination rate of the chlorinated olefins and the symmetry rule, proposed by Bonse and Henschler for the epoxides formed [l]. The epoxides of I-chloroethene, 1, Idichloroethene and 1,l ,Ztrichloroethene were characterized by Bonse and Henschler as ‘asymmetric’ and mutagenic. In contrast, the epoxides of perchloroethene and tram- 1,Zdichloroethene were defined as ‘symmetric’ and not mutagenic. Our calculations show that symmetric alkenes have a dipole moment from zero and therefore their metabolic rate, i.e. the formation of the epoxides is limited by the value of c, whereas the metabolic rate of asymmetric alkenes is higher. Consequently, higher amounts of the reactive epoxides per time unit will be produced from the asymmetric chlorinated alkene precursors than from the symmetric ones, counting for differences seen in the mutagenic .potency. The lower metabolic transformation rate of tram- versus cis- 1,2-dichloroethene in rats has been explained with the higher effectiveness of the tram epoxide as a suicide inhibitor of cytochrome P450 [26]. In vitro findings are inconsistent with the suicide inhibitor theory. Costa and Ivanetich
222
Gy.A. Csanddy et al. / Toxicol. L&t. 75 (1995) 217-223
[27] have demonstrated that cis- and trans-1,2dichloroethene show comparable activities with respect to cytochrome P450 inactivation in microsomal incubations. From the in vivo data base used in this study metabolic interactions between the produced epoxide and the parent alkene cannot be excluded unequivocally. However, our calculations predict a much lower metabolic transformation rate for the ‘symmetric’ tram isomer. Therefore, no suicide inhibition of the enzyme by the produced epoxide seems to be necessary to explain differences seen in metabolic rates of cis and tram isomers. Also a reversible competitive inhibition would not influence the numerical value of V,,, and therefore the model predictions are legitimate. The model describes the reactivity of styrene appropriately. There is a factor of 4 between the calculated (based on Eq. 2) and the experimentally determined rates of styrene metabolism. To establish the model, the general assumption was made that all compounds were metabolized by the same enzyme or enzyme family and that their binding properties depend on their dipole moments. Obviously this is also the case for such aromatic compounds as styrene. The model was also used to predict the metabolic rate of isobutene. The calculated kapp of 0.35 ml. h-‘g-l (based on Eq. 2 and the model variables listed in Table 2) is in agreement with the experimentally determined value of 0.11 ml. h-‘g-i (according to = 320 pmol. kg-‘h-l and C, = 2900 pmol/l &. In this work an approach has been presented to calculate in vivo reactivity of alkenes based on a theoretical model and using experimentally determined in vivo kinetic parameters of homologue compounds. Such a quantitative structure activity relationship as presented here might lead to a better understanding of the molecular mechanisms of biotransformation and in the long run might lead to a reduction of animal experiments.
Berufsgenossenschaft der Chemischen Industrie (Heidelberg, La 51511-3) is gratefully acknowledged. Appendix According to the two-compartment model [5] the average concentration in the organism (y2) at steady state is proportional to the concentration in the gas phase (_vl):
y2 = fit.yl
= yl*
Vlk12 V2k21 + V2keP
tit is the bioaccumulation factor at steady state, which depends on kinetic parameters as the constant clearance of inhalation (Vlk12), exhalation (V2k21) and on the concentration dependent clearance of elimination (V2kel*). The metabolic rate at steady state assuming a Michaelis-Menten kinetics can be described as dNel
V,,,.K.st.yl
dt
K,,, + fit-y1
-=
The right side of the equation equals 0.9. Vmaxif metabolic saturation of 90% is achieved. The solution for ~1~~ can be given as, according to [34]:
ylw,, = 0.9.
V,,,
+ 10-K,,,. V2k21
Vlk12
The corresponding average concentration in the organism is: C,,g, = ~2~~ = Kitwh. ~1~~. The necessary pharmacokinetic parameters involved in the calculations have been taken from the references listed in Table 2. References
Acknowledgements
111Bonse, G. and Henschler, D. (1976) Chemical reactivity,
We thank Dr. G. Csonka (Budapest) for providing the MNDO program and Drs. P. Politzer (New Orleans) and P.E. Kreuzer (Neuherberg) for valuable discussions. The financial support of the
biotransformation and toxicity of polychlorinated aliphatic compounds. CRC Crit. Rev. Toxicol. 4, 395-409. 121Jones, R.B. and Mackrodt, W.C. (1983) Structuregenotoxicity relationship for aliphatic epoxides. Biochem. Pharmacol. 32, 2359-2362. [31 Politzer, P. and Laurence, P.R. (1984) Halogenated hy-
Gy.A. Csamidy et al. / Toxicoi. Lett. 75 (1995) 217-223
[4]
[5]
[6] [7]
[8]
[9]
[IO]
[l l]
[I21
[13]
[14]
[15]
[16]
[17]
[18]
drocarbon epoxides: factors underlying biological activity. Int. J. Quant. Chem. Quant. Biol. Symp. 11, 155-166. Politxer, P. and Laurence, P.R. (1984) Relationships between the electrostatic potential, epoxide hydrase inhibition and carcinogenicity for some hydrocarbon and halogenated hydrocarbon epoxides. Carcinogenesis 5, 845-848. Filser, LG. (19912)The closed chamber technique uptake, endogenous production, excretion, steady-state kinetics and rate:3 of metabolism of gases and vapors. Arch. Toxicol. 66, l-10. Filser, J.G. and Bolt, H.M. (1979) Pharmacokinetics of halogenated ethylens in rats. Arch. Toxicol. 42, 123-136. Fiber, J.G. and Bolt, H.M. (1981) Inhalation pharmacokinetics based on gas uptake studies. I. Improvement of kinetic models. Arch. Toxicol. 47, 279-292. Kreiling, R., Laib, R.J., Fiber, J.G. and Bolt, H.M. (1986) Species di:fferences in butadiene metabolism between mice and rats evaluated by inhalation pharmacokinetics. Arch. Toxicol. 58, 235-238. Bolt, H.M. and Fiber, J.G. (1987) Kinetics and disposition in toxicology. Example: carcinogenic risk estimate for ethylene. Arclh. Toxicol. 60, 73-76. Peter, H., Wiegand, H.J., Bolt, H.M., Greim, H., Walter, G., Berg, M. and Fiber, J.G. (1987) Pharmacokinetics of isoprene in mice and rats. Toxicol. Lett. 36, 9-14. Golka, K., Peter. H., Denk, B. and Filser, J.G. (1989) Pharmacokinetics of propylene and its reactive metabolite propylene oxide in Sprague-Dawley rats. Arch. Toxicol. Suppl. 13, 2’40-242. Csanady, G.A., Freise, D., Denk, B., Filser, J.G., Cornet, M., Rogiers, V. and Laib, R.J. (1991) Investigation of species differences in isobutene (2-methylpropene) metabolism between mice and rats. Arch. Toxicol. 65, 100-105. Kessler, W., Denk, B. and Fiber, J.G. (1989) Speciesspecific inhalation pharmacokinetics of 2nitropropane, methyl-ethyl-ketone and n-hexane. In: BiologicallyBased Methods for Cancer Risk Assessment, Plenum Publishing Corparation, pp. 123-139. Castellino, A.J. and Bruice, T.C. (1988) Intermediates in the epoxidation of alkenes by cytochrome P-450 Model 1. cis-stilbene as a. mechanistic probe. J. Am. Chem. Sot. 110, 158-162. Garrison, J.M. and Bruice, T.C. (1989) Intermediates in the epoxidation of alkenes by cytochrome P-450 models. J. Am. Chem. Sot. 111, 191-198. Hanzlik, R.P. andt Shearer, G.O. (1975) Transition state structure for peracid epoxidation. J. Am. Chem. Sot. 97, 5231-5233. Plesnicar, B., Tasevski, M. and Azmon, A. (1978) The transition state for the epoxidation of ethylene with performic acid. J. Am. Chem. Sot. 100, 743-746. Swern, D. (1967) Epoxidation of ethylenic compound* Organic React. 7, 378-433.
223
[19] Fukui, K., Yonezawa, T. and Nagata, C. (1954) Theory of substitution in conjugated molecules. Bull. Chem. Sot. Japan 27, 423-427. [20] Dewar, M.J.S. and Thiel, W. (1977) Ground states of molecules 38. The MNDG method. J. Am. Chem. Sot. 99, 4899-4906. (211 Dewar, M.J.S., Zoebisch, E.G., Healy, E. and Stewart, J.F.P. (1985) AMl: a new general purpose quantum mechanical molecular model. J. Am. Chem. Sot. 107, 3902-3909. 1221 Wiener, H.J. (1947) Structural determination of paraffin boiling points. J. Am. Chem. Sot. 69, 17-20. [23] Barton, A.F.M. (1983) CRC Handbook of Solubility Parameters and Other Cohesion Parameters. CRC Press, Inc. Boca Raton, Florida. [24] Senn, P. (1988) The computation of the distance matrix and the Wiener index for graphs of arbitrary complexity with weighted vertices and edges. Comput. Chem. 12, 219-227. [25] RandiC, M., Hansen, P.J. and Jurs, P.C. (1988) Search for useful graph theoretical invariants of molecular structure. J. Chem. Inf. Comput. Sci. 28, 60-68. [26] Gargas, M.L. and Andersen, M.E. (1988) Physiologically based approaches for examining the pharmacokinetics of inhaled vapors. In: Toxicology of Lungs, Raven Press, New York, pp. 449-476. [27] Costa, K.A. and Ivanetich, K.M. (1982) The 1,2dichloroethylenes: their metabolism by hepatic cytochrome P-450 in vitro. B&hem. Pharmacol. 31, 20932102. [28] Fiber, J.G., Schwegler, U., Csanady, Gy.A., Greim, H., Kreuxer, P. and Kessler, W. (1993) Species-specific pharmacokinetics of styrene in rat and mouse. Arch. Toxicol. 67, 517-530. [29] Yamaxaki, T. and Kimura, K. (1975) Some rule consideration of molecular photoelectron spectrum. Substituted ethylenes. Bull. Chem. Sot. Japan 48, 1602-1607. [30] Reinke, D., Baumgartel, H., Cvitas, T., Klasnic, L. and Gusten, H. (1974) Vergleich der Photoelektronspektren und Photoionenspektren von Vinylfluorid. Ber. BunsenGesehschaft Phys. Chem. 78, 1145-l 147. [3l] Sandhu, J.S. (1972) Ionization potentials of chloroethylenes as determined by photoelectron spectroscopy. Indian J. Chem. 10, 667-668. [32] Hemdon, WC. (1980) Perturbation-graph theory III. Resonance energies from photoelectron spectra. Pure Appl. Chem. 52, 1459-1469. [33] Hannay, B. and Smith, C.P. (1943) Hyperconjugation in methyl butadienes as shown by dipole moments. J. Am. Chem. Sot. 65, 1931-1934. [34] Laib, R.J., Tucholski, M., Filser, J.G. and Csanady, Gy.A. (1992) Pharmacokinetic interaction between 1,3butadiene and styrene in Sprague-Dawley rats. Arch Toxicol. 66, 310-314.