Antioxidants in Disease Mechanisms and Therapy

Antioxidants in Disease Mechanisms and Therapeutic

Strategies

Visit the link below to download the full version of this book:

https://medipdf.com/product/antioxidants-in-disease-mechanisms-and-therapy-antio
xidants-in-disease-mechanisms-and-therapeutic-strategies/

Click Download Now

 Contributors xix

Birgit Heller (629) Clinical Department, Diabetes Research Institute,

Heinrich-Heine University Diisseldorf, D-40225 Diisseldorf, Germany
Karl Huffman (247) Eukarion, Inc., Bedford, Massachusetts 01730
Rocco Zannello (379)Molecular Genetics and Development Group, Institute
of Reproduction and Development, Monash University, Clayton Vic
3168, Australia
Michael 1.Kelner (379) Department of Pathology, University of California,
San Diego, California 92103
Thomas W. Kensler (293) Department of Environmental Health Sciences,
The Johns Hopkins School of Hygiene and Public Health, Baltimore,
Maryland 21205
Mark D. Kilburn (537) Department of Chemistry, Florida International
University, Miami, Florida 33199
Ralf Kinscherf (581) Division of Immunochemistry, Deutsches Krebsfor-
schungszentrum, D-69120 Heidelberg, Germany
Ismail Kola (379) Molecular Genetics and Development Group, Institute
of Reproduction and Development, Monash University, Clayton Vic
3168, Australia
Hubert Kolb (629) Clinical Department, Diabetes Research Institute,
Heinrich-Heine University Diisseldorf, D-40225 Diisseldorf, Germany
Ludmila G. Korkina (151) Institute of Pediatric Hematology, Moscow
117513, Russia
Eberhard Lampeter (629) Clinical Department, Diabetes Research Institute,
Heinrich-Heine University Diisseldorf, D-40225 Diisseldorf, Germany
John T. Landrum (537) Department of Chemistry, Florida International
University, Miami, Florida 33199
Charles S. Lieber (601)Mount Sinai School of Medicine (CUNY), Alcohol
Research and Treatment Center and G.1.-Liver-Nutrition Program, Bronx
Veterans Affairs Medical Center, Bronx, New York 10468
Samuel Louie (457) Division of Pulmonary and Critical Care Medicine,
Department of Internal Medicine, University of California, Davis, Sacra-
mento, California 958 17
Erik L~rvaas(119) The Norwegian College of Fishery Science, University
of T r o m s ~ 9037
, Troms~,Norway
Bernard Malfroy (247) Eukarion, Inc., Bedford, Massachusetts 01730
Catherine B. Marcus (247) Eukarion, Inc., Bedford, Massachusetts 01730
Gregory P . Marelich (491) Division of Pulmonary and Critical Care Medi-
cine and Adult Cystic Fibrosis Program, University of California Davis
Medical Center, Sacramento, California 958 17
xx Contributors

Hiroshi Masumoto (229) Institut fur Physiologische Chemie I, Heinrich-

Heine-Universitat Diisseldorf, D-40001 Diisseldorf, Germany
Susan T. Mayne (657)Department of Epidemiology and Public Health, Yale
University School of Medicine and Yale Cancer Center, New Haven,
Connecticut 06520
Sabine Mihm (581) Division of Immunochemistry, Deutsches Krebsfor-
schungszentrum, D-69120 Heidelberg, Germany
David P. R. Muller (557) Division of Biochemistry and Genetics, Institute
of Child Health, London W C l N lEH, United Kingdom
Wael Musleh (247) Neuroscience Program, University of Southern Califor-
nia, Los Angeles, California 90089
Thomas Nowell (515)Jean Mayer USDA Human Nutrition Research Center
on Aging at Tufts University, Boston, Massachusetts 021 11
Lester Packer (79) Department of Molecular and Cell Biology, University
of California at Berkeley, Berkeley, California 94720
Garth Powis (329) Arizona Cancer Center, University of Arizona, Tucson,
Arizona 85724
Thomas Primiano (293) Department of Environmental Health Sciences,
The Johns Hopkins School of Hygiene and Public Health, Baltimore,
Maryland 21205
Russel J. Reiter (103) Department of Cellular and Structural Biology, The
University of Texas Health Science Center at San Antonio, San Antonio,
Texas 78284
Sashwati Roy (79) Department of Molecular and Cell Biology, University
of California at Berkeley, Berkeley, California 94720
Karin Scharffetter-Kochanek (639)Experimental and Clinical Photoderma-
tology, Department of Dermatology, University of Cologne, 50931
Koln, Germany
Reinhold Schmidt (425)University Clinic of Neurology, University of Graz,
A-8036 Graz, Austria
Michael Schykowski (581)Division of Immunochemistry, Deutsches Krebsf-
orschungszentrum, D-69120 Heidelberg, Germany
Chandan K . Sen (79)Department of Molecular and Cell Biology, University
of California at Berkeley, Berkeley, California 94720
Helmut Sies (229) Institut fur Physiologische Chemie I, Heinrich-Heine-
Universitat Diisseldorf, D-40001 Dusseldorf, Germany
Thomas R. Sutter (293) Department of Environmental Health Sciences,
The Johns Hopkins School of Hygiene and Public Health, Baltimore,
Maryland 21205
 Contributors xxi

Allen Taylor (515) Jean Mayer USDA Human Nutrition Research Center
on Aging at Tufts University, Boston, Massachusetts 021 11
Muret G. Truber (49)Department of Molecular and Cell Biology, University
of California, Berkeley, California 94720
Albert uun der Vliet (491)Division of Pulmonary and Critical Care Medicine,
University of California Davis Medical Center, Sacramento, California
95817
Paul G. Winyard (403) Inflammation Research Group, St. Bartholomew’s
and the Royal London School of Medicine and Dentistry, London E l
lAD, United Kingdom
Ernst J. Woluetung (379) Molecular Genetics and Development Group,
Institute of Reproduction and Development, Monash University, Clayton
Vic 3168, Australia
 Preface

“Oxidative stress,” denoting a disturbance in the prooxidant/antioxi-

dant balance in favor of the prooxidants, leading to potential damage (see
1, 2), has been in focus for some time. The present volume brings together
many of the interesting topics that have developed with regard to potential
clinical implications. High hopes, voiced in the past, of spectacular effects
of single antioxidants as pharmacologically active agents have not fully
materialized; maybe some of the investigators hoped for too much. However,
the novel role of oxidants and antioxidants as part of signaling cascades,
as mediators of adaptive responses, has opened new areas of active research.
Although it is still too early to speak of “free radical diseases,” this volume
brings together evidence for crucial roles of oxidants and antioxidants in
disease states. As most diseases ultimately lead to cell injury or cell death, the
accompanying oxidative damage to DNA, protein, lipids, and carbohydrates
obviously is a feature of widespread interest. It is possible that the time
course, and even the final outcome, of a disease can be critically modulated
by strengthening the antioxidant side of the balance.
This volume focuses on some new strategies of antioxidant defense in
terms of new pharmacologically active agents, presents current knowledge
on known agents, and provides in-depth treatment of some disease states.
It is hoped that the book will serve as a stimulus for further research.
Helmut Sies

References

1. “Oxidative Stress,’’ Academic Press, London (1985).

2. “Oxidative Stress: Oxidants and Antioxidants,” Academic Press, London (1991).

xxiii
 Barry Halliwell
Neurodegenerative Disease Research Centre
Pharmacology Group
University of London King’s College
London SW3 6u<,United Kingdom

Antioxidants: The Basics-What

They Are and How t o
Evaluate Them

1. Introduction

Words such as “antioxidant” and “oxidative stress” are widely used

but surprisingly difficult to define precisely (for a discussion of the latter
term see Sies, 1991). For example, the term “antioxidant” as used in the
literature is often implicitly restricted to chain-breaking antioxidant inhibi-
tors of lipid peroxidation. In particular, food scientists frequently equate
antioxidants to inhibitors of lipid peroxidation because they use antioxidants
largely to prevent rancidity. However, free radicals generated in vivo damage
proteins, DNA, and other molecules in addition to lipids. Hence the author
and his colleagues introduced a broader definition: an antioxidant is any
substance that, when present at low concentrations compared to those of
an oxidizable substrate, significantly delays or prevents oxidation of that
substrate (Halliwell and Gutteridge, 1989; Halliwell, 1990).The term “oxi-
dizable substrate” includes almost everything found in living cells, including

Advonces in Pbarmocology, Volume 38

Copyright 0 1997 by Academic Press, Inc. All rights of reproduction in any form reserved. 3
4 Barry Halliwell

proteins, lipids, carbohydrates, and DNA. Mechanisms of antioxidant action

can include:
Removal of O2 (e.g., the packaging of foodstuffs under N2)
Scavenging reactive oxygednitrogen species (Table I) or their
precursors
Inhibiting ROSRNS formation
Binding metal ions needed for catalysis of ROS generation
Upregulation of endogenous antioxidant defenses.
When ROSRNS are generated in uivo, many antioxidants come into
play. Their relative importance as protective agents depends on:
Which ROSRNS is generated
How it is generated
Where it is generated
What target of damage is measured
For example, if human blood plasma is tested for its ability to inhibit
iron ion-dependent lipid perbxidation, the proteins transferrin and caeru-

TABLE I Reactive Oxygen and Nitrogen Species

Radicals Nonradicals

Reactive oxygen species“

Superoxide, 02.- Hydrogen peroxide, H20Z
Hydroxyl, OH‘ Hypochlorous acid, HOCI
Peroxyl, RO; Ozone, 0,
Alkoxyl, RO’ Singlet oxygen, ’Ag
Hydroperoxyl, HO;
Reactive nitrogen speciesb
Nitric oxide, NO’ Nitrous acid, HNOZ
Nitrogen dioxide, NO; Dinitrogen tetroxide, NzO,
Dinitrogen trioxide, N2O3
Peroxynitrite, ONOO-
Peroxynitrous acid, ONOOH
Nitronium cation, NO;
Alkyl peroxynitrates, ROONO

A collective term that includes both oxygen radicals and certain

nonradicals that are oxidizing agents and/or are easily converted into
radicals (HOCI, 03, ONOO-, ‘ 0 2 , HZOZ).
A collective term including nitric oxide and nitrogen dioxide radicals,
as well as such nonradicals as HN02and N,O+ ONOO- is often
included in both categories. “Reactive” is not always an appropriate
term: HzOZ, NO‘, and 0 2 . - react quickly with few molecules whereas
OH’ reacts quickly with almost everything. RO;, RO’, HOCI, NO;,
ONOO-, and O3 have intermediate reactivities. HOCI could also
be regarded as a “reactive chlorine species.”
 Antioxidants: The Basics 5

loplasmin are the most important protective agents (Gutteridge and Quinlan,
1992). When human blood plasma is exposed to NOz’, uric acid seems to
be a major protective antioxidant (Halliwell et al., 19921, whereas it appears
to play little role as a scavenger of HOCl in plasma (Hu et al., 1992).
Similarly, if the oxidative stress is kept the same but a different target of
oxidative damage is measured, different answers can result. When plasma
is exposed to gas-phase cigarette smoke (CS), lipid peroxidation occurs,
which is inhibited by ascorbate (Frei et al., 1991), whereas ascorbate has
no effect on the formation of plasma protein carbonyls by CS (Reznick et al.,
1992).As an extreme example, some known carcinogens (diethylstilbestrol)
that aggravate oxidative DNA damage in vivo (Roy and Liehr, 1991) are
powerful inhibitors of in vitro lipid peroxidation (Wiseman and Halli-
well, 1993).
The previously mentioned definition emphasizes the importance of the
source of stress and the target (“oxidizable substrate”) measured. However,
there may be cases that the definition does not include. Thus, plasma albumin
may have antioxidant properties, e.g., by binding copper ions and scavenging
HOCl (Halliwell, 1988). Would it therefore be beneficial to broaden the
definition, perhaps to an antioxidant is any substance that inhibits oxidative
damage to a target under the assay conditions being used? The risk is
that every chemical in the laboratory could then be classified as either an
“antioxidant” and/or a “pro-oxidant” on the basis of assays that have little
biological meaning. It is easy to develop some in vitro test to support a
postulate that compound X is an antioxidant, but it is very hard to prove
that compound X actually works by an antioxidant mechanism in vivo.
Antioxidants are of interest to radiation chemists, food scientists, poly-
mer chemists, and even to curators of museums (Daniels, 1989) but this
chapter only discusses the antioxidants known, or proposed, to be important
in aerobic organisms. Because many previous reviews as well as other chap-
ters in the present volume cover the well-established physiological antioxi-
dant roles of vitamin E, ascorbic acid, and such proteins as superoxide
dismutase, glutathione peroxidase, catalase, and caeruloplasmin, these are
not discussed in this chapter.
Many other substances have been proposed to act as antioxidants in
vivo. They include /3-carotene, other carotenoids, xanthophylls, metallothio-
nein, carnosine and related compounds, mucus, phytic acid, taurine and
its precursors, bilirubin, estrogens, creatinine, ergothioneine, dihydrolipoic
acid, ovothiols, ubiquinol, polyamines, retinol, flavonoids, other phenolic
compounds of plant origin, and melatonin. Some drugs already approved
for administration to humans, such as nonsteroidal anti-inflammatory drugs,
angiotensin converting enzyme (ACE) inhibitors, Ca2+antagonists, iron-
chelating agents, and thiol compounds, are frequently suggested to exert
antioxidant properties in vivo (reviewed in Halliwell, 1991).
How can such claims be evaluated? First, one must ask how the putative
antioxidant is supposed to act. Does it act directly (e.g., by scavenging ROS
6 Barry Halliwell

or inhibiting their generation)? This is the most common proposal. Does it

act indirectly (e.g., by upregulating endogenous antioxidant defenses)? For
example, the antioxidant properties of melatonin have received much public-
ity recently (see the chapter by Reiter, this volume). Our data suggest that
melatonin is a mediocre direct antioxidant (Marshall et al., 1996), but this
does not, of course, preclude it from acting by raising the levels of endoge-
nous antioxidant defense enzymes.
In evaluating the likelihood of direct antioxidant action in vivo, it is
important to ask certain questions (Table 11). Simple in vitro experiments
can answer some of these questions, and the results often allow one to
dismiss the proposed antioxidant role: a compound that is a poor direct
antioxidant in vitro will not be any better in vivo. This chapter outlines
some approaches to the characterization of direct antioxidant activity. Two
obvious (but often forgotten) points:
1. A compound should be tested at concentrations achievable
in vivo.
2. In assaying putative antioxidants, biologically relevant ROS/RNS
should be used.
Interest is also growing in reactive nitrogen species in biological systems
(Table I). Low molecular mass antioxidants such as GSH, other thiol com-
pounds, albumin, urate, and ascorbate may scavenge ONOO- (Beckman et
al., 1994) and urate may help protect against NO2‘in the respiratory tract
(Halliwell et al., 1992).

II. Biologically Relevant ROSlRNS

A. Superoxide and Hydrogen Peroxide

Superoxide formed in vivo is largely converted by superoxide dismutase
(SOD)-catalyzedor nonenzymic dismutation into H 2 0 2(Fridovich, 1989).
Some enzymes (e.g., glycollate, monoamine, xanthine oxidases) produce
H202directly (Sies, 1991; Halliwell and Gutteridge, 1989). Unlike 02--,

TABLE II Questions to Ask When Evaluating “Antioxidants” in Vivo

1. What biomolecule is the antioxidant supposed to protect? Does enough antioxidant

reach that target in vim?
2. How does it protect: by scavenging ROSRNS, preventing their formation, or repairing
damage?
3. If the antioxidant acts by scavenging, can the resulting antioxidant-derivedradicals
themselves cause damage?
4. Can the antioxidant cause damage in other biological systems (toxicological studies)?
 Antioxidants: The Basics 7

HzOZis believed to cross membranes easily (Halliwell and Gutteridge, 1989).

Both 02-- and H2O2can damage a few cellular targets (Fridovich, 1989;
Flint et al., 1993; Cochrane, 1991; Bielski, 1985), but, in general, their
reactivity is limited, i.e., few compounds react fast with 02*- or H 2 0 2[excep-
tions for Oz are NO‘ (Huie and Padmaja, 1993) and some iron-sulfur
proteins (Flint et al., 1993)l. Hence it is rather difficult to design an antioxi-
dant that scavenges 0 2 . - or Hz02 at high rates other than by developing
models of such enzymes as SOD or glutathione peroxidase.

I . Measuring Superoxide Scavenging

Superoxide is easily produced by the radiolysis of water in the presence
of Oz and formate, and pulse radiolysis allows the accurate determination
of reaction rate constants as well as examination of the spectrum of products
formed when Oz--reacts with scavengers/SOD mimics (Butler et al., 1988).
However, pulse radiolysis is unsuitable for measuring slow (rate constants
<lo5 M- sec-l) reactions for 02.- in aqueous solution. Stopped-flow meth-
ods can be used to study slower reactions (Bull et al., 1983). However,
approximate rate constants may be obtained using simple “test-tube” meth-
ods. For example, xanthine oxidase plus hypoxanthine (or xanthine) at pH
7.4 can be used to generate 02--, which is detected by its ability to reduce
cytochrome c or nitro blue tetrazolium (NBT) (Fridovich, 1989).If an Oz.--
reactive molecule is added, it decreases the rates of cytochrome c or NBT
reduction, and competition plots allow calculation of approximate rate
constants from the known rate constants for reaction of Oz--with the just-
described molecules (Bielski, 1985; Halliwell, 1985). This approach can be
used with other sources of 02*- [e.g., the potassium salt of superoxide, KOz,
can be used directly (Henry et al., 1976)] but anyone using it should always
perform some essential controls:
1. Check that the “antioxidant” does not directly inhibit Oz.- genera-
tion. Many papers use xanthine oxidase to generate Oz.-, but effects of
putative Oz.- scavengers on the enzyme itself are often not reported.
2. Check that the “antioxidant” does not directly reduce cytochrome
c or NBT, which will deplete their levels and hence decrease their reaction
rates with Oz--.
3. Consider the possibility that a radical formed by the attack or Oz.-
on an “antioxidant” interacts with cytochrome c or NBT.

2. Measuring H202Scavenging
HzOzcan be sensitively measured by peroxidase-based assay systems,
e.g., horseradish peroxidase uses H202to oxidize scopoletin into a nonfluo-
rescent product (Corbett, 1989).If an “antioxidant” is incubated with HzOZ
and the reaction mixture is sampled at various times, the rates of HZO2
disappearance can be calculated. Some points to consider:
8 Barry Halliwell

1. Check that the “antioxidant” is not a substrate for peroxidase, which

could decrease the fluoresence changes by competing with scopoletin for
the enzyme rather than by scavenging H202.
2. Consider adding SOD if 0 2 . - is also being generated. 02.- inhibits
peroxidase (forming compound 111) and may compromise the measurement
of H 2 0 2(Kettle et al., 1994).
If an “antioxidant” interferes with peroxidase-based systems, other
assays for H202can be used, including titration with acidified KMn04,
measuring the 0, release ( 1 mol per 2 mol of H202)when a sample of
the reaction mixture is injected into an 0, electrode containing catalase
and buffer solution, or measuring the release of l4CO2from I4C-labeled 2-
oxoglutarate (Varma, 1989).

B. Hydroxyl Radical
Much of the damage done by 0 2 . - and HzOzin vivo is thought to be
due to their conversion into more reactive species, including hydroxyl radical
(OH’)(Halliwell and Gutteridge, 1989, 1990). Formation of OH’ in vivo
occurs by at least four mechanisms:
1. Transition metal ion catalysis, especially by iron and copper (Halli-
well and Gutteridge, 1990).Reactive species additional to OH’ [e.g., perfer-
ryl, ferry1 or Cu(111)]may also be formed, although direct chemical evidence
for their existence in Fenton-type systems is lacking.
2. Exposure to radiation (von Sonntag, 1988).
3. Reaction of HOCl with 02.-(Candeias et al., 1993).
4. Possibly, during the decomposition of ONOO- (Beckman et al.,
1994; van der Vliet et al., 1994a).
1. Reactions of Hydroxyl Radical
The hydroxyl radical attacks almost every molecule found in vivo, with
rate constants of ?lo9 M- sec-’ (von Sonntag, 1988).Thus, almost every-
thing in a cell is an OH’ scavenger. Examples are glucose (rate constant
-lo9 M- sec-I, present at millimolar concentrations in body fluids) and
albumin [rate constant >lolo M- sec-’ (Smith et al., 1992)]. Hence, sugges-
tions that natural or synthetic “antioxidants” act by scavenging OH‘ in
vivo are chemically unlikely.
An “antioxidant” that affects OH’-dependent damage in vivo is more
likely to act by blocking OH’ formation, e.g., by removing its precursors
(02.-, H202,ONOO-, HOCI), or by chelating transition metal ions. Because
there is some confusion in the literature about metal-chelating agents, it is
worth listing some basic principles as to how they can work.
1. The binding of a metal ion to the chelator could alter its redox
potential andlor accessibility (e.g., to 02.-
or H L 0 2 )so as to stop it catalyzing
 Antioxidants: The Basics 9

redox reactions. An example is the binding of iron to transferrin or lactofer-

rin (reviewed in Halliwell and Gutteridge, 1989, 1990).
2. Their binding to the “chelator” does not stop OH’ formation, but
because OH’ is formed at the binding site, the “chelator” absorbs it and
“spares” a more important target. For example, copper ions bound to
albumin can still form OH’, and the protein is damaged (Marx and Chevion,
1986). The albumin, by binding copper and targeting damage to itself, is
acting as a “sacrificial antioxidant” that protects more important targets,
such as plasma lipoproteins and cell membranes (Halliwell, 1988). The
binding of copper ions to the amino acid histidine in plasma might also be
a protective mechanism: formation of OH‘ radicals detectable in free solu-
tion is suppressed (Rowley and Halliwell, 1983), but the histidine is de-
stroyed (Uchida and Kawakishi, 1986). Both albumin and histidine might
thus represent safe temporary transport forms for plasma copper ions ab-
sorbed from the gut, until they are cleared from the circulation by the liver.
Histidine-containing dipeptides, found in many mammalian tissues, may
also act as antioxidants by copper ion chelation (Kohen et al., 1988).
2. Measuring Hydroxyl Radical Scavenging
The definitive technique for investigating reactions of “antioxidants”
with OH’ is pulse radiolysis (Bielski, 1985; Butler et al., 1988),but approxi-
mate rate constants can be obtained more easily. For example, the spin-
trap 5,5-dimethyl-l-pyrroline-N-oxide (DMPO) reacts fast with OH’ (rate
constant > l o 9 M- sec-’). An added OH‘ scavenger will compete for OH‘
and decrease the DMPO-OH electron spin resonance (ESR) signal, and its
reaction rate constant can be obtained from a competition plot (Finkelstein
et al., 1980). DMPO also reacts with 02.-, although much more slowly
[rate constant -10 M-’ sec-’ (Finkelstein et al., 1980)] to give a different
adduct, and 02.- scavenging can similarly be measured by examining the
effects on the ESR signal of the DMPO-OOH adduct.
Another example of a detector for OH‘ is the sugar 2-deoxy-~-ribose
(Halliwell et al., 1987).In the “deoxyribose method” for studying reactions
with OH’, the OH‘ is generated by a mixture of ascorbic acid, H2O2,and
Fe3+-EDTA. It attacks deoxyribose, degrading it into fragments that give a
chromogen on heating with thiobarbituric acid (TBA) at low pH. If an
OH‘ scavenger is added, it competes with deoxyribose for OH‘, inhibits
chromogen formation, and competition plots allow the calculation of rate
constants.
Competition methods are widely used, but key controls are often not
reported in the literature. It is essential to show that
1. The “antioxidant” does not interfere with OH- generation, e.g., by
reacting with OH’ precursors in the reaction mixture, such as H 2 0 2or, if
metal ion-dependent systems are used to make OH‘, by chelating them (the
deoxyribose assay avoids this by using iron ions chelated to EDTA).
10 Barry Halliwell

2. The “antioxidant” does not interfere with product measurement. It

should not inhibit when added to the reaction mixture at the end of the
incubation, e.g., with the TBA reagents during the deoxyribose assay, or
after the DMPO-OH ESR spectrum has developed. Many compounds, e.g.,
ascorbate and certain thiols, such as captopril (Bartosz et al., 1996), reduce
DMPO-OH to an ESR silent species.
3. In chromogenic assays such as the deoxyribose assay, the attack of
OH’ on the “antioxidant” does not generate a false chromogen, e.g., omit-
ting deoxyribose from the reaction mixture should eliminate color devel-
opment.
3. Inhibition of Metal Ion-Dependent Hydroxyl
Radical Formation
Some “antioxidants” may block OH’ formation by chelating metal ions.
The deoxyribose method can also be used to test this possibility. When iron
ions are added to the reaction mixture as FeCI3 (not EDTA chelated), some
of them bind to deoxyribose. They still appear to catalyze OH’ formation,
but because the OH’ immediately attacks deoxyribose, OH’ scavengers (at
moderate concentrations) do not inhibit chromogen formation (Gutteridge,
1984).However, an “antioxidant” can inhibit in this version of the deoxyri-
bose assay if it chelates iron ions away from the deoxyribose and renders
them inactive or poorly active in generating OH‘. For example, citrate, a
poor OH’ scavenger but a good iron chelator, is a very effective inhibitor
in this version of the deoxyribose assay but not when EDTA is present
(Aruoma and Halliwell, 1988). Hence, this version of the deoxyribose assay
(subject to the controls described earlier) indicates the potential ability of
a compound to interfere with the “site-specific’’generation of OH‘ (Gutte-
ridge, 1984; Aruoma and Halliwell, 1988).
If an “antioxidant” binds metal ions and decreases the amount of OH’
detected, two possibilities exist:
1. The “antioxidant”-metal ion complex cannot catalyze OH‘ for-
mation.
2. OH‘ is still made but is largely intercepted by the antioxidant. To
distinguish between these mechanisms, the fate of the antioxidant in the
reaction mixture can be examined; it will be chemically modified if it reacts
with OH’.

C. Peroxyl Radicals
The formation of peroxyl radicals (RO,‘) is a key step in lipid peroxida-
tion (Halliwell and Gutteridge, 1989), but they can also be formed from
DNA and proteins and when thiyl (RS’)radicals combine with oxygen (Dean
et al., 1993; Sevilla et al., 1989; Willson, 1985). Peroxyl radical scavengers
may be water soluble (e.g., dealing with radicals from DNA, thiols, proteins)
 Antioxidants: The Basics II

or lipid soluble (e.g., the chain-breaking antioxidant inhibitors of lipid perox-

idation).
1. Measuring Peroxyl Radical Scavenging
Peroxyl radicals can be generated by pulse radiolysis (Willson, 1985)
and their reactions with “antioxidants” studied. ESR spin-trapping studies
of peroxyl radical reactions can also be carried out (e.g., Davies et al., 1993).
Another approach is the total (peroxyl) radical-trapping antioxidant
parameter (TRAP) assay (Wayner et al., 1987), which is much used (in
various versions) to study antioxidants in biological fluids. Peroxyl radicals
are generated at a controlled rate by the thermal decomposition of a water-
soluble “azo initiator,” such as 2,2 -azobis(2-amidinopropane)hydrochlor-
ide (AAPH). Decomposition produces carbon-centered radicals, which react
fast with O2to give peroxyl radicals that then attack a lipid to cause peroxida-
tion. By analyzing the effect of an “antioxidant” on the lag time to onset
and the rate of peroxidation, information about its mechanism of action
(Wayner et al., 1987) and a relative rate for its reaction with R 0 2 ‘(Darley-
Usmar et al., 1989) can be obtained. Lipid “targets” can be endogenous
lipids in biological fluids or added lipids, often linoleic acid/ester. Studies
of the ability to protect against AAPH-induced peroxidation have been used
to show, for example, that ascorbate is an excellent scavenger of water-
soluble ROz’ (Wayner et al., 1987; Darley-Usmar et al., 1989), whereas
desferrioxamine is not as good (Darley-Usmar et al., 1989). AAPH-derived
radicals also inactivate the enzyme lysozyme, providing a protein target for
studies of protection by “antioxidants” (Lissi and Clavero, 1990; Paya et
at., 1992).
The carbon-centered radicals produced by AAPH decomposition can
do direct damage [e.g., to DNA (Hiramoto et a/., 1993)] and can deplete
antioxidants (Soriani et al., 1994). Thus, one must ensure that reaction
mixtures contain enough O2 to convert them completely into peroxyl rad-
icals.
2. Lipid-Soluble Peroxyl Radicals
It is difficult to generate “clean” lipophilic peroxyl radicals in uitro.
One exception is trichloromethylperoxyyl,formed by exposing a mixture of
CC14, propan-2-01, and buffer to ionizing radiation (Alfassi et al., 1993).
Rate constants for the reaction of several “antioxidants” with CC1302‘
have been determined (reviewed in Aruoma et al., 1995). However, because
CC1302’is more reactive than nonhalogenated peroxyl radicals, the results
should be taken only as approximations of relative reactivity with the peroxyl
radicals formed during lipid peroxidation.

D. Studies of Lipid Peroxidation

To test lipid antioxidant activity directly, the ability of an “antioxidant”
to inhibit the peroxidation of lipoproteins, tissue homogenates, fatty acid/
I2 Barry Halliwell

ester emulsions, liposomes, or membranes (e.g., erythrocytes, liposomes,

microsomes) can be examined. Such studies are widespread and cannot be
discussed in detail here, but a few points can be made.
1. The lipid systems are usually kept under ambient PO*, but some
antioxidants [e.g., &carotene (Burton and Ingold, 1984)] work better at
lower pOz. Variable results can arise if rapid peroxidation depletes Ozduring
the reaction. This can be a special problem in ESR studies using narrow-
bore ESR tubes.
2. Accurate measurement of peroxidation is not easy (reviewed in
Packer, 1994). One must ensure that an apparent “antioxidant” action is
not caused by interference with the assay. For example, much of the “lipid-
peroxidation inhibitory effect” of carnosine and anserine in microsomes is
due to interference with the TBA test (Aruoma et al., 1989).
3 . How lipid peroxidation is started. If azo initiators (e.g., AAPH) are
used, it can be difficult to distinguish whether an antioxidant is acting by
direct scavenging of the initiator-derived peroxyl radicals or by scavenging
the chain-propagating peroxyl radicals generated from the lipid substrate.
Another problem is that lipophilic antioxidants added to reaction mix-
tures seldom partition completely into membranes or lipoproteins (e.g., a -
tocopherol added to low density lipoprotein (LDL) suspensions enters the
LDL very inefficiently).
Lipid peroxidation is often started by adding metal ions, e.g., as CuS04
(for LDL), FeS04, FeCL3 plus ascorbate, or FeCL,-ADP plus NADPH (for
microsomes). In these cases, an “antioxidant” effect could occur not only
by peroxyl radical scavenging, but also by metal ion chelation. However,
these two possibilities can be distinguished. If the antioxidant is acting only
by chelation, it will not be consumed during the reaction, as shown by, for
example, HPLC analysis. A chain-breaking antioxidant is consumed as it
scavenges peroxyl radicals.
1. Microsomal Peroxidation Assays
These are extremely popular, and some authors appear to think that
“microsomes” are subcellular organelles! In fact, “rnicrosomes” prepared
by differential centrifugation of tissue homogenates are a complex mixture
of vesicles from endoplasmic reticulum, plasma membrane, and other cell
membranes. They contain variable amounts of endogenous antioxidants,
such as a-tocopherol. Hence an “antioxidant” inhibiting microsomal lipid
peroxidation could (in addition to the mechanisms discussed earlier) be
acting by “recycling” endogenous antioxidants. For example, dihydrolipoate
does not inhibit irodascorbate-dependent peroxidation in liposomes (Scott
et al., 1994), but it recycles the vitamin E radical in microsomes to inhibit
peroxidation (Scholich et al., 1989). If microsomal lipid peroxidation is
started by adding NADPH plus Fe3+-ADP, a control (usually measuring
NADPH consumption) is needed to check that the “antioxidant” does not
 Antioxidants: The Basics 13

inhibit the enzymic reduction of Fe3+-ADP.Because the addition of NADPH

to microsomes activates cytochromes P450,added “antioxidants” could be
metabolized to products more (or less) active in inhibiting peroxidation.
Microsomal lipid peroxidation is often started by adding irodascorbate.
This avoids problems with P450 activation. However, ascorbate may be
capable of reducing “antioxidant” radicals (generated as they scavenge per-
oxyl radicals) back to the antioxidant, thus enhancing their action. Such a
reaction will only occur if the antioxidant-derived radicals become accessible
for reduction at the membrane surface. The classic example is the recycling
of the a-tocopheryl radical by ascorbate, but the same effect may occur
with other antioxidants, such as certain flavonoids. Hence, the antioxidant
activity of some lipid-soluble, chain-breaking antioxidants may appear to
be greater if ascorbate is present.
Overall, it is wise to evaluate anti-lipid peroxidation ability using several
different lipid substrates, with peroxidation started by different mechanisms.

E. Hypochlorous Acid
Neutrophils contain myeloperoxidase, which uses H 2 0 2to oxidize C1-
into HOCl (Weiss, 1989). Eosinophils contain a similar enzyme, which
prefers to oxidize bromide (Br-) ions and presumably produces HOBr (May-
eno et al., 1989). Hypohalous acids contribute not only to phagocyte killing
of foreign organisms (although the extent of the contribution is uncertain),
but also to tissue damage. For example, HOCl inactivates a,-antiproteinase,
an important inhibitor of certain serpins, such as elastase (Weiss, 1989).
HOCl is a powerful oxidizer of -SH groups on cell surfaces and can inhibit
membrane transport systems, as well as leading to chlorination of tyrosine
residues (Weiss, 1989; Kettle, 1996).

I. Hypochiorous Acid Scavenging Assays

“Antioxidants” that prevent HOCI-mediated damage could scavenge
HOCl directly and/or inhibit its production by myeloperoxidase. Myeloper-
oxidase can be assayed by standard tests of peroxidase activity (Halliwell
and Gutteridge, 1989) (e.g., oxidation of guaiacol to a chromogen in the
presence of H202)or by measuring its production of HOCl (Kettle and
Winterbourn, 1988). Often the former type of assay is easier when looking
for inhibitors because the latter type can give confusing results if compounds
that also scavenge HOCl are tested. If an apparent inhibition of myeloperoxi-
dase is found, it should be checked whether the “antioxidant” is really
inhibiting myeloperoxidase or is simply acting as a competing substrate,
e.g., several thiols are not only HOCl scavengers but also substrates for
myeloperoxidase (Svensson and Lindvall, 1988). The plant phenol
4-hydroxy-3-methoxyacetophenone(apocynin) inhibits neutrophil 02-- re-
14 Barry Halliwell

lease in vitro, apparently because it is oxidized by myeloperoxidase to gener-

ate the “real” inhibitor (Hart et al., 1990).
Once established that an “antioxidant” does not inhibit myeloperoxi-
dase, HOCl scavenging can then be examined using myeloperoxidase/H202/
CI- as an HOCl source. More simply, HOCl can be made by acidifying
sodium hypochlorite (Green et al., 1985). If a physiologically relevant con-
centration of an “antioxidant” is mixed with al-antiproteinase, a good
HOCl scavenger should protect the a1-antiproteinase against inactivation
when HOCl is added (Wasil et al., 1987). Controls are needed to show that
the “antioxidant” does not interfere with this assay by:
1. inactivating elastase directly,
2. stopping al-antiproteinase from inhibiting elastase, or
3. reactivating crl-antiproteinase after inactivation by HOCl.
If an “antioxidant” fails to protect al-antiproteinase against HOCl, it
may be that
1. It reacts too slowly (if at all) with HOC1.
2. It reacts with HOCl to form a “long-lived” oxidant that is itself
capable of inactivating al-antiproteinase (Weiss, 1989). Such
products are formed, for example, when taurine reacts with
HOCl. An alternative screening assay for HOCl scavenging
involves testing to see if the “antioxidant” can prevent HOCl-
dependent oxidation of 5-thio-2-nitrobenzoic acid (Ching et al.,
1994).
Even if an “antioxidant” could act by scavenging HOCl in vivo, the
possibility of forming toxic reaction products should be considered
(Uetrecht, 1983).

F. Heme ProteinslPeroxides
Mixtures of H 2 0 2with heme proteins (including hemoglobin and myo-
globin) oxidize many substrates and catalyze lipid peroxidation, apparently
by generating amino acid radicals and heme-associated oxo-iron species
(Rao et al., 1994; Kelman et al., 1994; Evans et al., 1994; Giulivi and
Cadenas, 1994). Cytochrome c can also accelerate oxidative damage in the
presence of peroxides (Evans et at., 1994). Such reactions may contribute
to ischernia-reperfusion injury, atherosclerosis, neurodegenerative disease,
muscle injury, and chronic inflammation (Kelman et al., 1994; Evans et al.,
1994; Giulivi and Cadenas, 1994; Rice-Evans et al., 1989). The ability of
a substance to react with activated heme proteins can be examined spectro-
photometrically by looking for loss of the ferry1 myoglobin (or hemoglobin)
spectrum as the compound reduces it to the ferrous or ferric state (Giulivi
and Cadenas, 1994; Rice-Evans et al., 1989). “Antioxidants” can also be