Molecular modeling parameters predict antioxidant efficacy of 3-indolyl compounds

Article abstract of DOI:10.1007/s002040050346

Many dietary constituents, such as indole-3-carbinol, are chemoprotective in toxicity and carcinogenicity bioassays. Indole-3-carbinol and related congeners appear to protect partly via radical and electrophile scavenging. To develop better chemoprotectiv

Full text of DOI:10.1007/s002040050346

Arch Toxicol (1996) 70: 830—834
( Springer-Verlag 1996
ORIGINAL INVESTIGATION
Howard G. Shertzer · M. Wilson Tabor · I.T.D. Hogan
Stephen J. Brown · Malcolm Sainsbury
Mole cular mode ling parame te rs pre dict antioxidant e fficacy
of 3-indolyl compounds
Received: 21 November 1995 / Accepted: 19 March 1996
Abstract Many dietary constituents, such as indole-3-
carbinol, are chemoprotective in toxicity and carcino-
Introduction
genicity bioassays. Indole-3-carbinol and related
The dietary indoles, indole-3-carbinol, indole-3-aceto-
nitrile and 3,3ꢀ-diindolylmethane, occur naturally as
glucosinolate conjugates in Brassica vegetables (Loub
et al. 1975), and are released upon hydrolysis. They
have been shown to protect against chemical carcino-
genesis (Bailey et al. 1985; Wattenberg 1985, 1990) and
chemically induced hepatotoxicity (Shertzer et al.
1987b, c, 1988, 1991). In the course of defining the
mechanism by which indoles protect against chemically
induced tissue damage, the compounds have been
found to both inhibit (Shertzer 1980; Eisele et al. 1983;
Shertzer et al. 1987c) and to induce (Shertzer 1982;
Godlewski et al. 1985; Bradfield and Bjeldanes 1987;
McDanell et al. 1987; Fong et al. 1990) the activities of
several different enzymes. Although changes in the ac-
tivities of enzymes, such as mixed-function oxidases
and phase II enzymes, could alter the biological re-
sponses of tissues exposed to carcinogens and toxi-
cants, indole-3-carbinol and/or its metabolites have
also been shown to be capable of scavenging biolo-
gically reactive electrophiles (Shertzer et al. 1987a;
Shertzer and Tabor 1988) and free radicals (Shertzer
et al. 1988, 1991). Such scavenging provides an addi-
tional and broader-based protection mechanism for
certain indole compounds.
congeners appear to protect partly via radical and
electrophile scavenging. To develop better chemopro-
tective indoles with lower intrinsic toxicity, we per-
formed molecular graphic and quantum-mechanical
analyses of model indolyl compounds to ascertain the
determinant molecular features for antioxidant activ-
ity. We examined eight structurally related 3-indolyl
compounds for relationships between antioxidation
potential (using in vitro lipid peroxidation assays) and
electronic, polar, and steric parameters, including bond
dissociation energies, bond lengths, dipole moments,
electronic charge densities, and molecular size para-
meters. Electronic features of the 3-methylene carbon
and 1-nitrogen were not predictive of antioxidative
potency due to extensive charge delocalization of the
cation radical following electron abstraction from the
nitrogen. Antioxidant efficacy of 3-indolyl compounds
was most strongly predicted by molecular size para-
meters and by the energy of electron abstraction as
calculated from the difference in heat of formation
between the parent compound and its cation radical.
A highly predictive multiple linear regression correla-
tion model (rꢀ"0.97) was obtained using the para-
meters of heat of formation, molecular weight, log P,
and diplole moment.
To both elucidate the molecular and structural basis
for radical quenching by 3-indolyl compounds, and
identify and develop novel 3-indolyl compounds which
exhibit enhanced antioxidant efficacy, we have exam-
ined a number of such compounds for their antioxida-
tion potential . In this paper we report the evaluation of
a set of 3-indolyl congeners for the relationships be-
tween antioxidation potential and electronic, polar,
and steric parameters.
Key words Antioxidant · Indolyl compounds · Lipid
peroxidation · Molecular modeling · Structure-activity
relationships
H.G. Shertzer ( ) · M. Wilson Tabor · S.J. Brown
Department of Environmental Health,
University of Cincinnati Medical Center,
Cincinnati, OH 45267—0056, USA
I.T.D. Hogan · M. Sainsbury
School of Chemistry, University of Bath,
Claverton Down, Bath BA2 7AY, UK

831
(CDCl ) ppm, 7.8—7.5 (2H,m), 7.3—7.1 (3H,m), 6.72 (2H,s, H-3ꢀ,H-5ꢀ),
ꢂ
Mate rials and me thods
6.35 (1H,s, 2-H), 3.90 (2H,s, CH ), 2.22 (3H,s, p-CH ), 2.18 (6H,s,
ꢀ
ꢂ
2 x o-CH ). [Empirical formula — found: C, 86.9%; H, 7.5%; N,
ꢂ
ꢁꢅ ꢁꢆ
5.5%, C
H N; requires: C, 86.7%; H, 7.7%; N, 5.6%].
Chemicals
All chemicals and solvents used in this study were obtained from
Sigma Chemical Co., St. Louis, Mo., except as noted. Semipurified
asolectin (soybean phospholipids) was supplied by Associated Con-
centrates, Woodside, Long Island, N.Y. Ascorbic acid and asolectin
were purified as described previously (Shertzer et al. 1988). Malonal-
dehyde-bis-(dimethyl acetal), 3-methylindole (skatole), indole-3-car-
binol, and 2,4,6-trimethylbenzyl alcohol were obtained from Aldrich
Chemical Co., Milwaukee, Wis. 3-(4-N,N-Dimethylaminobenzyl)in-
dole (Thesing and Mayer 1954), 3-(4-methoxybenzyl)indole (Pratt
and Botimer 1957), and 3-benzylindole (Biswas and Jackson 1969)
were synthesized according to published procedures. Structures of
the synthesized indoles were verified using both high-resolution
Fourier transformed nuclear magnetic resonance (NMR) and mass
spectrometry (MS) techniques.
Additional indoles were synthesized as described below. In these
syntheses, petrol refers to 60—80°C boiling range petroleum ether. All
reaction and purification solvents were dried and distilled prior to
use. Column chromatography for compound purification and isola-
tion was achieved using Merck 7747 silica gel. For the compounds
synthesized as described below, structural assignments and verifica-
tions were via NMR and MS techniques. ꢁH NMR spectra were
recorded at 270 MHz on a JEOL JNM Fourier transform instru-
ment using trimethylsilylchloride as the internal standard. ꢁꢂC
NMR spectra were obtained under identical conditions at
67.8 MHz. Mass spectra were measured using a VG 707OE mass
spectrometer coupled to a VG 2000 data system.
Lipid peroxidation assay
The cell-free lipid peroxidation assay utilized in this study has been
described previously (Shertzer et al. 1988). Briefly, lipid peroxidation
of purified soybean phospholipid vesicles in aqueous potassium
phosphate buffer (pH 7.4), was initiated by the addition of 10 lM
FeNH (SO ) in the presence of 100 lM ascorbic acid. The reaction
ꢃ
ꢃ ꢀ
was quenched with butylated hydroxytoluene, and the thiobar-
bituric acid-reacting products of lipid peroxidation were standard-
ized using malonaldehyde-bis-(dimethylacetal) (MDA), and ex-
pressed initially as MDA equivalents. For each analysis of inhibition
of lipid peroxidation by indole compounds, at least ten different
concentrations of compound were tested, with at least four different
concentrations above and below the concentration which gave 50%
inhibition. After initial range-finder experiments, each titration was
repeated at least three times. The results were calculated as the
average percentage of control rates (no inhibitor) of lipid peroxida-
tion versus concentration of inhibitor. SigmaPlot (Jandel Scientific)
was used to fit the curve with a third order least-squares regression
line and the concentration of compound required to inhibit lipid
peroxidation by 50% (IC ) was calculated.
ꢄꢇ
Modeling methods
The lowest energy molecular conformations of the 3-indolyl com-
pounds were calculated using HyperChem Molecular Modeling
software (Autodesk), using the Austin model 1, Polak-Ribiere conju-
gate gradient algorithm with UHF spin pairing and a 0.01 conver-
gence limit in vacuo. The log P values were obtained using an
additive-constitutive algorithm with ChemPlus software (Hyper-
cube).
3-(4-Hydroxybenzyl)indole
Trimethylsilyl indole (3.6 ml, 25 mmol) was added slowly to a solu-
tion of 3-(4-methoxybenzyl) indole (2 g, 8.4 mmol) in quinoline
(20ml) stirred in an ice-bath and protected by a nitrogen atmo-
sphere. After the addition, the ice-bath was removed and the reac-
tion mixture was heated to 180°C for 2 h. It was then cooled and
poured onto ice-cold 2 M hydrochloric acid (50 ml) and extracted
with diethyl ether (2;30 ml). The organic layers were combined and
washed successively with 1 M hydrochloric acid (4;50 ml), water
Regression analysis
For the purposes of these quantitative structure-activity relation-
ships (QSAR), the in vitro antioxidant activities were converted to
(2;50ml), brine (1;50 ml), and dried over MgSO . Removal of the
solvent gave a colorless residue, which was dissolved in methanol
ꢃ
the form:!log IC . First order regression analysis between indole
and treated with several drops of diethyl ether saturated with hydro-
gen chloride. The methanol was removed under reduced pressure
and the resultant gum chromatographed. Elution with dichloro-
methane/petrol (1 : 1) gave 3-(4-hydroxybenzyl)indole (720 mg, 38%),
which was crystallized from diethyl ether/petrol as colorless prisms,
ꢄꢇ
antioxidant efficacy and physicochemical properties were computed
using SigmaStat Software (Jandel Scientific). For those parameters
showing relationships with (10% confidence that a linear correla-
tion occurred by chance (P(0.1), multiple linear regression analyses
were performed using all possible combinations of descriptors.
m.p. 142—143°C, j nm (e), 281 (9422), 224 (30, 640); l
ꢀ!ꢁ
cm\ꢁ, 3495 (sh, N-H), 3300 (br, O-H). [Empirical formula — found:
(CHCl )
ꢀ!ꢁ
ꢂ
C, 80.7%; H, 5.9%; N, 6.3%, C
5.95%; N, 6.5%].
H NO; requires: C, 80.3%; H,
ꢁꢄ ꢁꢂ
Re s ults and Dis cus s ion
Free radicals including reactive oxygen and lipid per-
oxidation products have been suggested to be impor-
tant mediating agents in aging and several human
diseases, including cancer, pulmonary and cardio-
vascular disease, cataracts, and neurological dysfunc-
tions such as Parkinson’s disease (Clark et al. 1985;
Ames et al. 1993; Shigenaga et al. 1994; Halliwell and
Cross 1994). While the major intracellular cytosolic
antioxidant, reduced glutathione (GSH), is effective in
preventing radical-induced damage to DNA, GSH is
not effective in directly inhibiting lipid peroxidation
3-(2,4,6-¹rimethylbenzyl)indole
2,4,6-Trimethylbenzyl alcohol (2.5 g, 0.017 mol) in p-cymene (50 ml)
was treated with powdered potassium hydroxide (0.8g) and heated
at reflux in a Dean and Stark apparatus. After 20 min the reaction
mixture was cooled, and indole (1.95 g, 0.017 mol) and nickel metal
(150mg) were added. The mixture was then heated at reflux for
a further 15 h. The solvent was removed and the residue distilled to
afford the title compound as an oil (205 mg, 10%), b.p.
250—252°C/0.5—0.7 mmHg, which slowly crystallized to give color-
less prisms, m.p. 134—135°C; j nm (e), 283 (5691), 224 (16, 885); l
ꢀ!ꢁ
ꢀ!ꢁ
(CHCl ) cm\ꢁ, 3500 (NH); m/z 249 (100% [M>]), 130 (82%); d
ꢂ
&

832
(Zhu et al. 1996). Although vitamin E normally serves were obtained for each of the compounds studied. The
to protect lipids from peroxidative damage, radical molar concentrations required to inhibit by 50% the
stress may promote lipid peroxidation during exposure control levels of peroxidation (IC ) were used to calcu-
ꢄꢇ
to pro-oxidative pharmaceuticals, environmental agents, late the !log IC values, and these results are shown
ꢄꢇ
or in the course of inflammatory diseases. Therefore, in Table 1. The !log IC was used as the parameter
ꢄꢇ
the development and evaluation of novel or dietary for the antioxidant potential of the 3-indolyl com-
antioxidants, for the purpose of prevention or treat- pounds on which to regress the parameters from the
ment of such conditions is an important endeavor. This results obtained from molecular modeling.
paper describes the essential features of 3-indolyl com-
pounds that impart antioxidant potency against lipid predict antioxidation efficacy are also reported in Table
peroxidation.
1. Additional parameters that were calculated but not
The calculated physicochemical parameters that best
A typical lipid peroxidation inhibition curve is shown (since they were not predictive of antioxidation
shown for 3-hydroxyethylindole (Fig. 1). Similar curves potency) are the atomic charges on each of the carbon
atoms, and the bond lengths for carbon-hydrogen, car-
bon-nitrogen, and carbon-carbon bonds, and various
other molecular dimensions and dipole moments. Sev-
eral pairs of these most predictive descriptors are
colinear (Table 2). This is to be expected when evaluat-
ing a series of congeners with rather hydrophobic con-
stituents. Since several global molecular properties are
colinear with molecular mass, they are hence colinear
with each other. Linear least squares regression ana-
lyses of the data revealed two parameters to be the best
individual predictors of antioxidant efficacy (Table 2).
These were: (1) the energies of electron abstraction, as
calculated from the differences in heats of formation
between the parent compound and its cation radical
(*Hf); and (2) various parameters of size, for which the
longest dimension (Z-dimension) of the smallest enclos-
ing periodic box was the best predictor. The least linear
regression equations for these data are:
!log IC "3.02 *Hf#1.36 (rꢀ"0.78, P"0.004);
ꢄꢇ
Fig. 1 Lipid peroxidation curve for 3-hydroxyethylindole. The
curve represents the influence of 3-hydroxyethylindole on iron/as-
corbate-initiated lipid peroxidation. Each value represents the aver-
age of three experiments (each performed in duplicate) expressed as
a percentage of the control rates of lipid peroxidation (i.e., those that
occurred in the complete assay systems, but in the absence of the test
indolyl compound)
and
!log IC "0.21 Z-dimensional length
ꢄꢇ
#2.44 (rꢀ"0.71, P"0.009).
Table 1 Lipid peroxidation inhibition constants and physicochemical parameters for 3-indolyl compounds
Parameter
Compound number!
1
2
3
4
5
6
7
8
IC (M;10ꢈ)
1.5
5.82
1.460
13.96
250.3
2.26
11
4.96
1.136
12.32
249.4
3.40
0.78
!0.218
12
4.92
1.066
11.58
223.3
24
4.62
1.166
13.20
237.3
1.74
1.21
!0.217
29
4.54
0.912
8.68
161.2
!0.37
!1.80
!0.216
36
4.44
1.079
11.31
207.3
1.99
0.07
!0.218
100
4.00
0.998
7.27
131.2
0.38
!0.61
!0.218
160
3.80
0.869
7.57
147.2
!0.62
!1.11
!0.215
ꢄꢇ
!Log IC
ꢄꢇ
*Hf (kcal/mol)
Z-dimension (A)
MW (Da)
Log P
Z-dipole (Debyes)
Nitrogen charge
1.71
1.86
!0.225
!0.58
!0.218
Determination of the values shown is as described in the Materials and methods. The Z-dimension is the longest dimension of the smallest
periodic box that may surround a molecule, while Z-dipole is the dipole moment in that dimension. The parameter *Hf is the difference in
the heat of formation between the parent compound and the cation radical produced by electron abstraction. Log P is the log value of the
ꢁꢇ
partition coefficient. IC is the concn. of compound required to inhibit lipid peroxidation by 50%; MW is mol wt.
ꢄꢇ
! Compound 1, 3-(4-N,N-dimethylaminobenzyl)indole; compound 2, 3-(2,4,6-trimethylbenzyl)indole; compound 3, 3-(4-hydroxybenzyl)in-
dole; compound 4, 3-(4-methoxybenzyl)indole; compound 5, 3-hydroxyethylindole (tryptophol); compound 6, 3-benzylindole; compound 7,
3-methylindole (skatole); compound 8, 3-hydroxymethylindole (indole-3-carbinol)

833
Table 2 Colinearity analysis of
the major physicochemical
parameters used in this study
(All parameters are as defined in
Table 1)
!Log IC
*Hf
Z-dimension
MW
Log P
Z-dipole
ꢄꢇ
!Log IC
*Hf
1.00
—
—
—
—
0.78
1.00
—
—
—
0.71
0.73
1.00
—
—
—
0.68
0.62
0.96
1.00
—
0.43
0.51
0.69
0.78
1.00
—
0.41
0.81
0.62
0.54
0.61
1.00
ꢄꢇ
Z-dimension
MW
Log P
Z-dipole
—
—
—
Linear least squares regression analyses among all pairs of major parameters used for the multiple
linear regression analyses. The data are of rꢀ values
Fig. 3 The regression plane and data points for !log IC as
ꢄꢇ
a function of *Hf and Z-dimension length. These two parameters
represent the best pair of predictors (multiple linear regression) of
antioxidant efficacy
parameters to predict !log IC did not significantly
ꢄꢇ
improve the predictive power (rꢀ"0.71, P"0.045).
Fig. 2 The regression line and data points for *Hf and Z-dimension
This was not an expected result, since lipid peroxida-
tion occurred in a hydrophobic liposomal membrane
and it was expected that log P alone would have been
a better predictor of antioxidation efficacy.
length versus !log IC . These parameters represent the best
ꢄꢇ
individual predictors (simple linear regression) of antioxidant effi-
cacy. The first order regression fit of the data and 95% confidence
intervals are shown. (*Hf Difference in heat of formation between
parent compound and cation radical produced by electron abstrac-
tion)
The charge on individual atoms and groups of atoms
in the parent compounds and their cation radicals was
investigated in depth. Of particular interest was the
charge on the indolic nitrogen (Table 1), since electron
abstraction from this site by lipid radicals would pro-
vide a reasonable mechanism for the chain-breaking
antioxidant efficacy of these compounds. No significant
correlation between any parameter of charge and anti-
oxidant potential was found (data not shown). The lack
of significant differences in the charge on any single
atom or group of atoms between one indole and an-
other after the loss of an electron from the nitrogen in
the indole suggests an extensive delocalization of
charge throughout the indole ring system. Although
charge delocalization of the cation radical prevents
atom-specific static charge from being used as a pre-
dictor of antioxidant efficacy, it serves two very import-
ant functions in enhancing radical chain-breaking
These regression plots are shown in Fig. 2. Further-
more, the multiple linear regression of these two inde-
pendent variables yielded an rꢀ value that is better than
either of the simple linear regressions whereas the P-
value remained significant. The resultant regression
plane is described by the following equation and pre-
sented graphically in Fig. 3
!log IC "2.09 *Hf#0.08 Z-dimension length
ꢄꢇ
#1.54 (rꢀ"0.81, P"0.016).
The importance of the Z-dimension length may be
related partly to its colinearity with the log of the
partition coefficient (log P) (rꢀ"0.69, P"0.008).
A multiple linear regression using both of these

834
potency. Firstly, charge delocalization increases the Fong AT, Swanson HI, Dashwood RH, Williams DE, Hendricks
JD, Bailey GS (1990) Mechanisms of anti-carcinogenesis by in-
dole-3-carbinol. Studies of enzyme induction, electrophile-scav-
ease of electron abstraction, and secondly, stabilization
of the resulting radical reduces its ability to propagate
the radical chain reaction. The stabilizing effect is pre-
dicted on a molecular level by the *Hf, which increases
with antioxidant efficacy. This relationship is best ex-
plained by the fact that as *Hf increases, the relative
stability of the cation radical increases in relation to the
parent compound, making the loss of an electron by the
parent compound more favored.
enging, and inhibition of aflatoxin B activation. Biochem Phar-
ꢁ
macol 39: 19—26
Godlewski CE, Boyd JN, Sherman WK, Anderson JL, Stoewsand
GS (1985) Hepatic glutathione S-transferase activity and
aflatoxin B -induced enzyme altered foci in rats fed fractions of
ꢁ
Brussels sprouts. Cancer Lett 28: 151—157
Halliwell B, Cross CE (1994) Oxygen-derived species: their relation
to human disease and environmental stress. Environ Health
Perspect 102 (Supp 10): 5—12
Loub WD, Wattenberg LW, Davis DW (1975) Aryl hydrocarbon
hydroxylase induction in rat tissues by naturally occurring in-
doles of cruciferous plants. J Natl Cancer Inst 54: 985—988
McDanell R, McLean AEM, Hanley AB, Heaney RK, Fenwick GR
(1987) Differential induction of mixed-function oxidase (MFO)
activity in rat liver and intestine by diets containing processed
cabbage: correlation with cabbage levels of glucosinolates and
glucosinolate hydrolysis products. Food Chem Toxicol 25:
363—368
Pratt EF, Botimer LW (1957) Disproportionative condensations.
IV. The 3-alkylation of indoles by primary and secondary alco-
hols. J Am Chem Soc 79: 5248—5250
Shertzer HG (1980) Effects of indole derivatives on mono-oxygenase
activity in rabbit liver microsomes. In: Coon MJ, Conney AH,
Estabrook RW, Gelboin HV, Gillette JR, O’Brien PJ (eds)
Microsomes, drug oxidations and chemical carcinogenesis, vol 2.
Academic Press, NY, pp 889—892
Shertzer HG (1982) Indole-3-carbinol and indole-3-acetonitrile in-
fluence on hepatic microsomal metabolism. Toxicol Appl Phar-
macol 64: 353—361
Shertzer HG, Tabor MW (1988) Nucleophilic index value: implica-
tion in the protection by indole-3-carbinol from N-nitroso-
dimethylamine cyto and genotoxicity in mouse liver. J Appl
Toxicol 8: 105—110
Shertzer HG, Berger ML, Tabor MW (1987a) Protection from
N-nitrosodimethylamine mediated liver damage by indole-3-car-
binol, and correlation with nucleophilic index value. In: Nygaard
OF, Simic M, Cerutti P (eds) Anticarcinogenesis and radiation
protection. Plenum Press, NY, pp 241—245
Shertzer HG, Niemi MP, Reitman FA, Berger ML, Myers BL,
Tabor MW (1987b) Protection against carbon tetrachloride
hepatotoxicity by pretreatment with indole-3-carbinol. Exp Mol
Pathol 46: 180—189
The best set of parameters for multiple linear regres-
sion to describe the antioxidant efficacy of the indolyl
congeners is:
!log IC "4.85 *Hf!0.49 Z-dipole#0.04 log P
ꢄꢇ
#0.004 MW!1.48 (rꢀ"0.97, P"0.013).
Although it is clear that *Hf and to a lesser extent Z-
dipole (the dipole moment along the Z-axis) are domin-
ant (for these two parameters alone, rꢀ"0.91, P"
0.002), log P and mol.wt. also contribute to the best
predictive equation. In summary, we have shown that
molecular modeling in conjunction with multiple linear
regression can be an important tool in predicting the
antioxidant efficacy of 3-indolyl compounds. The anti-
oxidant potential is highly predicted by *Hf in con-
junction with dipole moment, log P, and molecular size.
Although extensive charge delocalization masks the
single-atom effects of electron abstraction, the overall
stability of the cation radical is predictive of the anti-
oxidant potential. The multiple linear regression equa-
tions developed in this paper may be used to predict the
antioxidant efficacies of potentially useful chemo-
protective compounds.
Acknowledgements Supported by NIH grants P30-ES06096, P42-
ES04908, and the Cancer Research Campaign, UK. The authors
thank Dr. Eugene Coats for helpful discussions in the early phases of
this work.
Shertzer HG, Tabor MW, Berger ML (1987c) Protection from
N-nitrosodimethylamine mediated liver damage by indole-3-car-
binol. Exp Mol Pathol 47: 211—218
Shertzer HG, Berger ML, Tabor MW (1988) Intervention in free
radical mediated hepatotoxicity and lipid peroxidation by in-
dole-3-carbinol. Biochem Pharmacol 37: 333—338
Re fe re nce s
Shertzer HG, Sainsbury M, Graupner PR, Berger ML (1991) Mech-
anisms of chemical mediated cytotoxicity and chemoprevention
in isolated rat hepatocytes. Chem Biol Interact 78: 123—141
Shigenaga MK, Hagen TM, Ames BN (1994) Oxidative damage and
mitochondrial decay in aging. Proc Natl Acad Sci USA 91:
10771—10778
Thesing J, Mayer H (1954) Die umsetzung von gramin und vervan-
dten verbindungen mit aromatischen aminen. Uber die Hoff-
mann-Martius-Umlagerung beim methyl-phenyl-amin. Chem
Berl 87: 1084—1092
Ames BN, Shigenaga MK, Hagen TM (1993) Oxidants, anti-
oxidants, and the degenerative diseases of aging. Proc Natl Acad
Sci USA 90: 7915—7922
Bailey G, Goeger D, Hendricks J, Nixon J, Pawlowski N (1985)
Indole-3-carbinol promotion and inhibition of aflatoxin B car-
ꢁ
cinogenesis in rainbow trout. Proc Am Assoc Cancer Res 26: 115
Biswas KM, Jackson AH (1969) Electrophilic substitution in indoles
V. Indolenines as intermediates in the benzylation of 3-sub-
stituted indoles. Tetrahedron 25: 227—241
Bradfield CA, Bjeldanes LF (1987) Structure-activity relationships of
dietary indoles: a proposed mechanism of action as modifiers of
xenobiotic metabolism. J Toxicol Environ Health 21: 311—323
Clark IA, Cowden WB, Hunt NH (1985) Free radical-induced patho-
logy. Med Res Rev 5: 297—332
Wattenberg LW (1985) Chemoprevention of Cancer. Cancer Res 45:
1—8
Wattenberg LW (1990) Inhibition of carcinogenesis by naturally-
occurring and synthetic compounds. In: Kuroda Y, Shankel DM,
Waters MD (eds) Anticarcinogenesis mechanisms II (Basic life
sciences, vol 52). Plenum Press, NY, pp 155—166
Eisele TA, Bailey GS, Nixon JE (1983) The effect of indole-3-car-
binol, an aflatoxin B hepatocarcinoma inhibitor, and other
indole analogs on the rainbow trout hepatic mixed function
oxidase system. Toxicol Lett 19: 133—138
Zhu H, He M, Bannenberg GL, Moldeus P, Shertzer HG (1996)
Effects of glutathione and pH on the oxidation of biomarkers of
cellular oxidative stress. Arch Toxicol (in press)
ꢁ