Effects of structure on binding to the 2,3,7,8-TCDD receptor protein and AHH induction--halogenated biphenyls.

The quantitative structure-activity relationships (QSARs) for polychlorinated biphenyl (PCB) congeners have been determined by comparing the EC50 values for three in vitro test systems, namely, aryl hydrocarbon hydroxylase (AHH) and ethoxyresorufin O-deethylase (EROD) induction in rat hepatoma H-4-II-E cells and competitive binding avidities to the rat cytosolic receptor protein (using 2,3,7,8-tetrachlorodibenzo-p-dioxin as a radioligand). For several PCB congeners that are in vivo inducers of rat hepatic microsomal AHH, there was a linear correlation between the -log EC50 values for receptor and the -log EC50 values for AHH (or EROD) induction; moreover, a comparable linear relationship was observed between the -log EC50 values for AHH and EROD induction. Previous in vivo studies have shown that the most active PCB congeners 3,3',4,4'-tetra-, 3,4,4',5-tetra-, 3,3',4,4',5-penta-, and 3,3',4,4',5,5'-hexachlorobiphenyl, cause many of the biologic and toxic effects reported for the highly toxic halogenated aryl hydrocarbon, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Moreover, the monoortho-substituted homologs of the four coplanar PCBs also elicit comparable in vivo biologic and toxic responses. It was evident from the QSARs for PCBs that there was an excellent correspondence between the in vivo and in vitro potencies of the individual PCB congeners. The effects of substituents on both receptor binding and AHH/EROD induction was determined for a series of 4'-substituted (X)-2,3,4,5-tetrachlorobiphenyls (where X = H, Cl, Br, I, OH, OCH3, NO2, COCH3, F, CF3, CH3, C2H5, i-C3H7, n-C4H9 and t-C4H9). Not unexpectedly, there was a linear relationship between the -log EC50 values for AHH and EROD induction, and these results confirm that both enzymatic oxidations are catalyzed by the same cytochrome P-450 isozyme(s). The effects of substituent structure on receptor binding for 12 substituents was subjected to multiple regression analysis which correlates the relative binding affinities of the compounds with the physical chemical characteristics of the substituents. The analysis gave the following equation: log (1/EC50) = 1.53 sigma + 1.47 pi + 1.09 HB + 4.08 for n = 12, s = 0.18, r = 0.978; where n is the number of substituents, s is the standard deviation, r is the correlation coefficient, and sigma = electronegativity, pi = hydrophobicity (log P) and HB = hydrogen bonding capacity for the substituent groups.(ABSTRACT TRUNCATED AT 400 WORDS)


Introduction
Pharmacogenetic studies have played an important role in understanding the biologic and toxic effects of carcinogens, drugs and related xenobiotics. Nebert and co-workers have investigated the toxic, carcinogenic, and biologic activities of polynuclear aromatic hydrocarbons (PAHs) in genetically inbred mice strains, and their results have suggested a mechanism of action for these compounds (1)(2)(3)(4). The activity of 3-methylcholanthrene (3-MC) and several related PAHs as inducers of hepatic and extrahepatic drug-metabolizing enzymes is remarkably dependent on the strain of mice used. Hepatic microsomal aryl hydrocarbon hydroxylase (AHH), a cytochrome P-448-dependent monooxygenase, is readily induced by 3-MC in several strains of mice, typified by the "responsive" C57BL/6J strain; in contrast, 3-MC does not induce AHH in nonresponsive strains, typified by the DBA/2J mice (5,6). Moreover, responsive mice are more susceptible than nonresponsive mice to the toxic (inflammation, fetotoxicity, primordial oocyte depletion) and carcinogenic effects of PAHs at organs/tissues in direct contact with the applied PAHs (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). The gene complex responsible for mediating the effects of PAHs has been designated the Ah locus which may comprise regulatory, structural, and possibly temporal genes. Extensive studies on genetically inbred responsive and nonresponsive mice (and their backerosses) suggests that the differences in susceptibility are related to the levels of the Ah regulatory gene product, the cytosolic Ah receptor protein. The receptor protein interacts with the PAH ligands and the resultant PAH:receptor ligand complex translocates into the nucleus and presumably initiates the toxic and biologic effects via a process comparable to that proposed for the steroid hormones (1)(2)(3)(4)(5)(6).
The in vivo structure-activity relationships (SARs) provide qualitative support for the proposed receptormediated hypothesis for the mechanism of action of 2,3,7,8-TCDD and related toxic halogenated aryl hydrocarbons. Quantitative structure-activity relation- ships (QSARs) for the PCBs have been obtained by using interrelated bioassays systems, namely, the competitive binding affinities of PCBs for the 2,3,7,8-TCDD hepatic cytosolic receptor protein from male Wistar rats (57) and the AHH and EROD induction potencies of these compounds in rat hepatoma H-4-II-E cells in culture (34). The former assay measures the important interaction between the toxin or ligand and the receptor protein, whereas the induction assay measures one of the biologic consequences (i. e., the induction of cytochrome P-448-dependent monooxygenases of this initial interaction). Dose-response studies are also readily determined by using in vitro assays, and the pharmacokinetic and metabolic factors which play a role in in vivo studies are minimized. The QSAR for the binding avidities of PCBs to the cytosolic receptor protein have been reported by using the sucrose density gradient assay system (57). Figure  1 illustrates the [3H]-2,3,7,8-TCDD-receptor protein binding peak which sedimented at approximately fraction 25 and corresponded to a sedimentation rate of 8-10S (dependent on ionic concentration of the buffer). Competition experiments with 10 ,uM nonradiolabeled 2,3,7,8-TCDD or 3-MC completed eliminated this peak, whereas competition experiments with PB did not significantly decrease the area of this radioactive binding peak. Figure 2 illustrates the sucrose density gradient proffles observed after competive binding experiments ,5-pentachlorobiphenyl, completely eliminates the radiolabeled binding peak, whereas the less toxic mixed-type inducer only partially reduces the area of this peak. Although 2,2',4,4'-tetrachlorobiphenyl is a relatively nontoxic PCB congener that strictly resembles PB in its mode of induction of the hepatic microsomal monooxygenases, this compound does competitively displace some of the radiolabeled 2,3,7,8-TCDD. Dose-response-competition experiments with several other PCB congeners that do not induce AHH (i.e., 2,3,4,5-tetra-and 2,2',4,4',5,5'hexachlorobiphenyl) showed that at high concentrations there is some competition with [3H]-2,3,7,8-TCDD for the receptor binding protein. The biological significance of this weak binding has not been determined. Table 1 summarizes the binding avidities of 14 PCBs, and the data complement the results illustrated in Figure 1; the coplanar PCBs are the most avid competitors for binding to the receptor protein; the monoortho-substituted analogs of the coplanar PCBs also bind to the receptor but with lower affinities than the coplanar compounds. These in vitro data parallel the in vivo SAR for PCBs and confirm that the most active compounds are substituted in both para and two or more meta positions and that the introduction of a single ortho sub- stituent into the biphenyl ring diminishes but does not eliminate the activity of the resultant compounds. Table 1 also summarizes the effects of structure on the activity of PCBs as inducers of AHH and EROD in rat hepatoma H-4-II-E cells in culture. A plot of receptor binding affinities versus AHH or EROD induction potencies (Figs. 3 and 4) illustrates the linear correlation between these two biologic parameters and thus supports the role of the receptor in mediating the induction of the cytochrome P-448-dependent monooxygenases.

Substituted Halogenated Biphenyls
The in vivo and in vitro studies with structurally diverse PCBs and PBBs support a common receptormediated mechanism of action for halogenated biphen- yls, PCDDs and PCDFs. Despite the critical role played by the Ah receptor protein in the biochemistry and toxicology of halogenated aromatics and PAHs, this protein has not been purified from any species. However, the nature of the protein binding sites for halogenated aromatics have been studied by several groups, and their results indicate the following. Based on the high binding affinity of 2,3,7,8-TCDD, a ligand should be relatively flat or planar, and deviations from this structural feature (e.g., monoortho-substituted halogenated biphenyls) result in diminished binding affinities (21,26,36) of the ligands. The optimal two-dimensional shape of a ligand corresponds to a 3 x 10 A rectangle (17,21,37). Maximum binding affinity requires the substitution of Cl at all four lateral positions for the PCDDs and PCDFs or at four to six lateral positions for PCBs and PBBs (17,21,37,57). Bromo-substituted biphenyls and PCDDs exhibit higher binding affinities than their chlorinated analogs (61,67,68). Polarizability and the electronacceptor properties of the lateral substituents are proposed to be important properties which facilitate binding to the receptor (69)(70)(71).
Since the ability to bind effectively to the cytosolic receptor is either enhanced or reduced by various substituent groups on the tetrachlorobiphenyl backbone, the variation in biological activity must be dependent on the properties inherent in the substituents. Quantitative analysis of substituent structure-activity effects can be used to study the intermolecular interactions between ligands and their receptors (74)(75)(76)(77). In general, drug-receptor interactions involve such forces as hydrophobic bonding, hydrogen bonding, van der Waals energy, electrostatic energy, valence-bond energy, and repulsion or strain energies of the bonds (77). The strength of the interaction depends upon the use of these various energies, which for a series of substituted congeners, can be approximated by the use of free-energyrelated physiochemical substituent parameters (74)(75)(76)(77)(78). For the present series of 4'-substituted tetrachlorobiphenyls, the relationships between various substituent parameters and receptor binding constants were examined by means of multiparameter regression analysis.
For each substituent group, a hydrophobic parameter ,Tr, an electronic parameter a, and a hydrogen-bonding accepting parameter HB were determined. The hydrogen-bonding parameter HB is an indicator variable that takes a value of unity for substituents which are hydrogen acceptors but a value of zero for nonhydrogenbonders. For the electronic effect of 4'-substituents, upara (Hammett constant) was used. The up values were obtained from the literature.
The hydrophobic parameter wr, defined as rr = log Pxlog PH where Px and PH are the partition coefficients of 4'substituted-2,3,4,5-tetrachlorobiphenyl and 4'-unsubstituted-2,3,4,5-tetrachlorobiphenyl, respectively, in the n-octanollwater system, was estimated by means of a recently developed method (79). According to this method, -T values for the hydrogen-bondable 4'-substituents were calculated by using the equation: where the subscript X/C6H4C6HC14, denotes wr values for the 4'-substituted-2,3,4,5-tetrachlorobiphenyls and the subscript X/C6H5 refers to values for monosubstituted benzenes. As Eq. (2) illustrates, hydrophobic substituent parameters (IT) were derived from values for monosubstituted benzenes. This derivation is valid because 4'-substituted-2,3,4,5-tetrachlorobiphenyl can be considered as a special case of a para-disubstituted benzene in which one of the substituent groups remains constant while the other is variable. The ir values for disubstituted benzenes have been calculated using parameters from monosubstituted benzenes (79). The factor 0.19px, describes the solubility-modifying effect of the electron-withdrawing 2,3,4,5-tetrachlorobiphenyl group on the hydrogen-bonding interaction of each 4'substituent with solvents. For substituents incapable of hydrogen bonding, such as alkyl and halogen, the p value is zero. The electron-withdrawing effect of the 2,3,4,5-tetrachlorobiphenyl group can be expressed by the cp value, which was estimated as U,p/C6Cc5 X 4/5 = 0.19. The 1MX/c6H, values were taken from the literature. For 15 compounds of the present series of para-substituted biphenyls, electronic, hydrophobic and hydrogen-bonding substituent constants were examined with respect to the EC,, values calculated from competition binding assays (shown in Table 4). Multiple regression analysis of the data led to Eq.    Table 3).
number of compounds is given by n, s is the standard deviation, r is the correlation coefficient, and F is the value of the F-ratio. The figures in parentheses are the 95% confidence intervals. The results of the analysis are shown in Table 4.
The log (1/ECQ) for the 4'-N-acetylamino derivative was most poorly predicted by the Eq. (3). For the 4'nitro compound, the HB value was taken as zero although the nitro group is capable of acting as a hydrogen acceptor. With an HB value of unity, its affinity was predicted to be about ten times higher than the observed value. The hydrogen-bonding interaction may not be insignificant for this compound although the mechanism is not clear. with n = 12, s = 0.18, r = 0.978, F = 65.90 (a < 0.01%). Although the quality of the correlation is much improved, the inferences drawn from the correlation are not significantly changed. The 4'-n-butyl, 4'-tert-butyl and 4'-phenyl compounds were not included in these correlations. Their log(1/EC.) values were much lower than those predicted by the two correlation equations. The van der Waals volumes of these three substituents were among the highest, being 41.8 cm3/mole for the n-butyl/ and tert-butyl groups and 45.8 cm3/mole for phenyl. The value for other substituents used here was less than 35 cm3/ mole (i.e., 34.12 cm3/mole for isopropyl and 33.23 cm3/ mole for N-acetylamino). Although the use of various types of steric parameters did not produce meaningful correlations, there seems to be a certain limiting size in the receptor binding site(s) to accommodate the biphenyl derivatives.
The correlation (developed in the above equations) means that both increasing substituent electron with- drawing activity and hydrophobicity, increased binding avidity of the ligands for the receptor. Hydrogen bondaccepting substituents also favor this binding. The coefficient of the HB term, 1.10, means that the molarity of the hydrogen donor in the receptor region is 12to 13-fold that of the octanol phase which is used as a model of the hydrophobic receptor binding site (81 with n = 12, s = 0.259, r = 0.983. The linear correspondence between the relative activities of these two induced enzyme activities provides strong evidence that both enzymes are induced via the same mechanism and that the oxidation of both substrates is catalyzed by the same cytochrome P-450 isozymes. Equation (7) has been developed to relate the receptor binding affinities to the AHH induction potencies of the substituted halogenated biphenyls: log RA (AHH) = 2.132 log (1/EC50) (  with n = 11, s = 0.557, r = 0.942. The value AB5 iS one of the steric parameters that represent the maximum width of the substituents from the axis connecting the 4'-substituent with the rest of the molecule. AB, is the value relative to that of hydrogen [B5(X) -B5(H)]. Although the standard deviation is large, each term is statistically significant (> 95%). The results also show that the AHH induction activity is linearly related to receptor binding (i.e., see Figs. 8 and 9) and the magnitude (2.132) of the coefficient of log (1/EC50) indicates that enzyme induction is about two times more sensitive than receptor binding to the structural variations of the 4'-substituents. The AHH induction is also dependent parabolically with the maximum width of the substituent X, the optimum width being 1.4, 'Which is close to the value of the NO2and CF3 substituents. This suggests that the interaction of the ligand:receptor protein complex with the nuclear receptor site is dependent on the conformation of this complex which in turn is governed by the size of the 4'-X substituent.
The importance of the size of the substituents is evident in the receptor binding affinities of a series of halogenated biphenyls which contain two vicinal meta and Para (4-) substituents in one of the phenyl rings. Figures 10 and 11 sumnmarize the competitive binding affinities of the 3,4-dichloro-, 3,4-dibromo-, and 3,4diiodo-4'-biphenyltrifluorides and the 4'-bromo-3,4dichloro-, 3,4,4'-dibromo-, and 4'-bromo-3,4-diiodo-3'biphenyltrifluorides. The CF3, group was used in both sets of ligands since this functional group in the 4'substituted series of compounds facilitated binding to the rat cytosolic receptor protein. For the 4'-substituted halogenated biphenyls, the order of binding avidities for the substituents was I>Br>Cl, and these data could be rationalized by the contributions of their respective electronegativities, lipophilicities and HB capacity.
However, for the 3,4-dihalosubstituted biphenyls, the order of binding avidities for both series was Br>Cl>I. This would suggest that steric effects may also play a 120 (I) QB~~~r ICF, role in facilitating ligand substituent interactions with the receptor protein and that optimal steric, electronic and lipophilic substituent parameters all contribute to binding affinity of a ligand for the 2,3,7,8-TCDD cytosolic receptor protein. This is also illustrated by the activity of 3,4,4'-tritrifluoromethylbiphenyl in competitively displacing rH]-2,3,7,8-TCDD from the receptor protein. This ligand contains three substituent groups that are highly electronegative (a = 0.54) compared to the Cl, Br and I groups (a = 0.23, 0.23, and 0.18, respectively); moreover, there are only small differences in their hydrophobic r values. However, the competitive binding affinity of this ligand for the receptor protein was unexpectedly low. Current research in our laboratory is focused on in vitro QSAR studies with other substituted halogenated biphenyls, dibenzofurans and dibenzo-p-dioxins as probes for investigating the structure and function of the 2,3,7,8-TCDD cytosolic receptor protein. This approach is also being used to compare the relative binding affinities of ligands to cytosolic receptor protein from species with different susceptibilities to the effects of toxic halogenated aryl hydrocarbons.
The financial assistance of the National Institutes of Health (ES02795 and ES02937), the Natural Sciences and Engineering Research Council of Canada, the Center for Comparative Medicine and the Texas Agricultural Experiment Station (Grant No. 6376) is gratefully acknowledged.