(This is a sample cover image for this issue. The actual cover is not yet available at this time.

This article appeared in a journal published by Elsevier. The attached

copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Author's personal copy

Chemistry and Physics of Lipids 165 (2012) 638647

Contents lists available at SciVerse ScienceDirect

Chemistry and Physics of Lipids

journal homepage: www.elsevier.com/locate/chemphyslip

Review

Evaluation of antioxidants: Scope, limitations and relevance of assays

I. Pinchuk, H. Shoval, Y. Dotan, D. Lichtenberg
Department of Physiology and Pharmacology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel

a r t i c l e

i n f o

Article history:
Received 23 April 2012
Received in revised form 21 May 2012
Accepted 22 May 2012
Available online xxx
Keywords:
Oxidative stress
Antioxidants
Antioxidative potency
Capacity
Assays

a b s t r a c t
Peroxidation of lipids, particularly polyunsaturated fatty acid residues (PUFA) of phospholipids and
cholesterol esters, is a process of marked implications: it shortens the shelf-life of food and drugs, it
causes fragmentation of DNA, it damages cellular membranes and it promotes the genesis of many human
diseases. Much effort is therefore devoted to a search for potent antioxidants, both synthetic and from
natural sources, mostly plants.
This, in turn, requires a reliable, simple, preferably high throughput assay of the activity of alleged
antioxidants. The most commonly used assays are based on measurements of the total antioxidant capacity (TAC) of a solution, as evaluated either by determining the rate of oxidation of the antioxidant or by
measuring the protection of an easily determined indicator against oxidation by the antioxidants.
The commonly used assays utilized for ranking antioxidants share three common problems:
(i) They usually evaluate the effects of those antioxidants that quench free radicals, which constitute
only a part of the bodys antioxidative network, in which enzymes play the central role. (ii) Both the capacity and potency of antioxidants, as obtained by various methods, do not necessarily correlate with each
other. (iii) Most estimates are based on methods conducted in solution and are therefore not necessarily
relevant to processes that occur at the lipidwater interfaces in both membranes and micro emulsions
(e.g. lipoproteins).
Given this state of art, many researchers, including us, try to develop a method based on the formation of hydroperoxides (LOOH) upon peroxidation of PUFA in lipoproteins or in model membranes, such
as liposomes. In these systems, as well as in lipoproteins, the most apparent effect of antioxidants is prolongation of the lag time preceding the propagation of a free radical chain reaction. In fact, under certain
conditions both water soluble antioxidants (e.g. vitamin C and urate) and the lipid soluble antioxidant
tocopherol (vitamin E), promote or even induce peroxidation.
Based on the published data, including our results, we conclude that terms such as antioxidative capacity or antioxidative potency are context-dependent. Furthermore, criteria of the efcacy of antioxidants
based on oxidation in solution are not necessarily relevant to the effects of antioxidants on peroxidation in
biological systems or model lipid assemblies, because the latter processes occur at water/lipid interfaces.
We think that evaluation of antioxidants requires kinetic studies of the biomarker used and that the
most relevant characteristic of oxidative stress in the biological context is the kinetics of ex vivo peroxidation of lipids. We therefore propose studying the kinetics of lipid-peroxidation in the absence of the
studied antioxidant and in its presence at different antioxidant concentrations. These protocols mean
that antioxidants are assayed by methods commonly used to evaluate oxidative stress. The advantage
of such evaluation is that it enables quantization of the antioxidants efcacy in a model of relevance to
biological systems. In view of the sensitivity of the lag time preceding peroxidation, we propose studying how much antioxidant is required to double the lag observed prior to rapid peroxidation. The latter
quantity (C2lag ) can be used to express the strength of antioxidants in the relevant system (e.g. LDL, serum
or liposomes).
2012 Elsevier Ireland Ltd. All rights reserved.

Corresponding author. Tel.: +972 3 6407305; fax: +972 3 6409113.

E-mail address: physidov@post.tau.ac.il (D. Lichtenberg).
0009-3084/$ see front matter 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.chemphyslip.2012.05.003

Author's personal copy

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

639

Contents
1.

2.

3.

4.

5.

6.

7.

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.
Free radicals, ROS and antioxidants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.
Aim of the present review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.
Signicance of this review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.
Expectation and disappointment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2.
The rational of using antioxidants as food supplements and the need to assay oxidative stress and antioxidative activity . . . .
1.3.3.
The role of antioxidants in prevention of pathogenic processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The most commonly used assays of antioxidants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.
General comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.
Oxygen radical absorbance capacity (ORAC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.
Ferric reducing antioxidant power (FRAP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.
Trolox equivalent antioxidant capacity assay (TEAC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.
Evaluation of antioxidants using more than one criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Testing and comparing antioxidants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.
The need for kinetic data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.
Characterization of lipid peroxidation based on kinetic proles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Oxidative stress and ranking of antioxidants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.
Oxidative stress, denition and quantitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
Terminology used to characterize antioxidative effect(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Problems associated with assaying OS and ranking antioxidants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.
The lack of a universal criterion of OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.
Experimental details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.
The time-dependence of measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.
Comparison of data given in terms of different units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.5.
The difference between reactions in solutions and at interfaces (location, location, location) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ranking antioxidants on the basis of OS methods: the effects of antioxidants on ex vivo peroxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.
Inhibition of peroxidation of PUFA in serum lipoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.
Evaluation of antioxidative potency in model membrane systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions (Take-home messages) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Background
1.1. Free radicals, ROS and antioxidants
An antioxidant is a molecule (or an ion, or a relatively stable
radical) that is capable of slowing or even preventing the oxidation
of other molecules. The continuing growth of the market of antioxidants reects the hope to cure the wide range of diseases that are
believed to be caused or promoted by oxidative stress (Halliwell
and Gutteridge, 2007; Niki, 2011; Dotan et al., 2009a). The latter term, commonly dened intuitively as an imbalance between
pro-oxidative and anti-oxidative factors, is often associated with
high levels of reactive oxygen and nitrogen species (ROS and RNS,
respectively) and with high levels of oxidation products, particularly hydroperoxides and DNA fragments (Sies, 1986; Dotan et al.,
2004).
Harmans discovery of such oxidative damages led to the postulation of the free-radical hypothesis of aging (Harman, 1956,
2009) and of the oxidation-stress hypothesis of atherosclerosis
(Steinberg et al., 1989). Following the latter hypotheses, oxidative
stress has been associated with the pathogenesis of many other
diseases including diabetes mellitus, atherosclerosis, neurodegenerative diseases, different malignant diseases and virus infections,
including AIDS (Halliwell and Gutteridge, 2007).
These hypotheses raised the reasonable expectations that
antioxidants should prolong the shelf life of foodstuff and drugs and
reduce the mortality and/or the morbidity of supplemented people
by neutralizing the harmful free radicals (Frankel, 2007). In fact,
antioxidants prolong the shelf-life of food and drugs, as expected
and this, in turn, helps the ongoing effort to produce more active
antioxidants. It also explains, at least partially, the very large number of publications on antioxidants. Noticeably, the results of the
commonly used assays of antioxidants are, in general, reasonable

639
639
640
640
640
640
640
640
640
641
641
641
641
642
642
642
643
643
643
644
644
644
644
645
645
645
645
646
646
646
646

predictors of the stabilization of foodstuff and drugs but not of their

effects on human health (Bjelakovic et al., 2007; Miller et al., 2005;
Dotan et al., 2009b).
This apparent inconsistency may be a result of a failure of the
hypothesis. However, it may be regarded consistent with the theory if it is attributed to the dependence of peroxidation on the
state of aggregation of lipid substrate. Specically, most of the
commonly used assays briey described below occur in solutions,
whereas in biological systems amphiphilic phospholipids reside
either in membranes or in emulsion or micro-emulsion particles
(lipoproteins) and peroxidation therefore occurs at the lipidwater
interface (Schnitzer et al., 2007).
Notably, the terms used to describe the activity of antioxidants
are context-dependent. For example, the capability of an antioxidant to slow down the oxidation of an oxidizable component
in a soft drink may be very different from its capability to protect biomembranes against oxidative damage in vivo. Moreover,
the assay of an antioxidant to protect a given food supplement
is not necessarily relevant to its potency to prolong the shelf
life of a given drug. This of course means that an assay utilized
to compare antioxidants in the search for improved inhibitors
of peroxidation used as stabilizers of food-stuff, may have to be
considerably different from the assay to be used in the search
for antioxidants that will maximize the shelf life of a given
drug.
Theoretically, to be relevant to the oxidative stress in human, a
reasonable optimal assay has to be based on the effect of externally
added antioxidant on the peroxidation of PUFA either in simple
model membrane or in samples of serum or LDL. We therefore think
that the most commonly used assays are relevant to the search
for antioxidants that stabilize best the peroxidizable water-soluble
compounds in solutions, whereas methods aimed at evaluating biologically relevant antioxidants must involve aggregated substrates.

Author's personal copy

640

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

The methods used to assay the effect of antioxidants on peroxidation of water-soluble substrates have been reviewed extensively
(Prior et al., 2005; Bartosz, 2010). Noticeably, less than a third of
the 60,000 scientic papers published in the last 5 years with the
term antioxidant(s) in their title or abstracts appeared in medical or
pharmacological journals. More than a third appeared in food and
agriculture literature and about a third appeared in chemical and
biochemical literature. Furthermore, only a small minority of newly
developed methods appeared in medical journals, most other being
in Food chemistry, agricultural and analytical journals.
1.2. Aim of the present review
The present, non-comprehensive review reects our point of
view on the available and potential assays of antioxidants. It is
based primarily on our experience and understanding, which will
hopefully contribute to the on-going search for a reliable, biologically relevant, sensitive and easy-to-conduct assay. To achieve this
goal, we rst describe the general features of free radicals chain
reactions and the terminology used to describe the characteristic
factors. Next, we describe the scope and limitations of the commonly used assays. Most of these assays are conducted in solution,
which means that their relevance to the ability of antioxidants to
protect aggregated substrates against peroxidation of lipids (which
occurs at the lipid/water interface) is questionable.
This review leads us to the conclusion that we should evaluate
antioxidants on the basis of the methods used to evaluate another
context-dependent term, oxidative stress (OS). This possibility has
been considered more than a decade ago but rejected on the basis
of experimental difculties (Prior et al., 2005). Re-evaluation of the
idea of using OS-assays to estimate the activity of antioxidants is
presented below.
1.3. Signicance of this review
1.3.1. Expectation and disappointment
For many years, peroxidation of lipids, particularly lipoproteins was considered responsible for CVD and for many other
diseases. Antioxidants have been therefore expected to reduce the
frequency of ROS-dependent diseases. Unfortunately, several large
clinical trials show that indiscriminate supplementation of vitamin E increases the mortality (both cardiovascular and all-cause) of
the vitamin E-treated subjects (Bjelakovic et al., 2007; Miller et al.,
2005). The results of our recent decision analysis supported these
meta-analyses (Dotan et al., 2009b).
The mechanism responsible for these unexpected results may
indicate that free radical chain reactions are not the major cause of
the pathogenesis of arthrosclerosis. Alternatively, the latter results
may be attributed to vitamin E-induced inhibition of the production
of antioxidative enzyme(s) and/or for reducing the concentration
of ROS beyond the level of being sufcient to function normally. In
addition, the unexpected results may reect experimental errors.
Regardless of the mechanism responsible for the unexpected
results, the latter ndings stress the importance of appropriate
analysis of results obtained by methods used to evaluate OS and
antioxidants. Only such methods can enable selective antioxidant
treatment. Meanwhile, several supporters of vitamin E supplementation think that indiscriminate supplementation is justied
because it reduces the risk of many diseases (Pryor, 2000). Other
authors argue that the results of the meta-analyses are questionable, due to serious aws in both the experiments and statistic
analyses (Blumberg and Frei, 2007). The latter view accords with
that of the producers of antioxidants, who still encourage indiscriminate use of their products. This view can explain the popularity
of antioxidants expressed by the presence of antioxidants in most
homes in developed countries. For many years the general attitude

was that if antioxidants do not help, at least they will not harm. The
more recent studies invalidate this approach. We think that many
people may benet from supplementation of antioxidants, including vitamin E (Witztum, 1998; Milman et al., 2008). The challenge
is to establish criteria to predict who is likely to gain from such
supplementation (Niki, 2011).
1.3.2. The rational of using antioxidants as food supplements and
the need to assay oxidative stress and antioxidative activity
As stated above, many investigators, particularly R&D experts
in industrial companies, devote much time, money and effort to
test the possibility that some virtuous antioxidants may have
greater antioxidative potency than the commonly used preparations, containing mostly vitamin E and C. Such a possibility may
be attainable either because new antioxidants may be designed to
be more efcient reducing agents and/or more efcient quenchers
of free radicals or/and due to preferred organ distribution and/or
other pharmacokinetic reasons. Given the expected market for efcient antioxidants, both the industry and university researchers
are looking for safe and potent antioxidants, both synthetic and
extracted from natural sources. This search, of course, requires an
assay that will be easy, reproducible, relatively cheap, preferably
high throughput and, most importantly, closely reect the effect of
the studied antioxidant in vivo.
1.3.3. The role of antioxidants in prevention of pathogenic
processes
A measure of the antioxidative potency and/or capacity can be
used to evaluate the involvement of free radicals in the pathogenesis of a given disease. Specically, if free radicals are involved
in a given pathogenesis, we could expect a correlation between
the efciency of the antioxidants to inhibit peroxidation and their
pharmacological effects. As an example, it is commonly believed
that Alzheimer disease (AD) is associated with aggregation of the
polypeptide Amyloid (A) into typical brils. Several antioxidants inhibit this process and many of them cause dissociation of
pre-formed brils (Shoval et al., 2007).
This does not necessarily mean that oxidative stress plays a
role either in the brilization of A or/and in the dissociation of the
brils. One way to test the question of whether the latter nding is
causal is to test the correlation between the capability of different
antioxidants to inhibit the formation of A brils and their antioxidative activity. This again, requires a reliable way of evaluating
the latter factor. In our previous study, we found that the antioxidative potency of 7 polyphenols, as determined by kinetic studies,
correlates with the potency of the studied polyphenols to cause dissociation of the amyloid brils (Shoval et al., 2007; Bartosz, 2010).
The latter nding does not prove that antioxidants cause dissociation of brils by reducing the concentration of free radicals, but it
strengthens the latter hypothesis
2. The most commonly used assays of antioxidants
2.1. General comments
Many assays have been developed to test and rank antioxidants.
Some of these assays are modications of same basic methods.
Yet, only a small number of methods were considered for standardization in The First International Congress on Antioxidants
Methods. These methods also happened to be the most commonly
used methods (Prior et al., 2005).
The most commonly used assays are not necessarily better, easier, more relevant or more accurate than other assays. Nonetheless,
widely used assays, particularly those conducted with commercial
kits, are less likely than other assays to suffer from irreproducibility due to trivial technical problems. Nonetheless, in the absence

Author's personal copy

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

of a gold standard assay, the most commonly used assays become

de facto gold standards. Being aware of the limitations of the various common assays, the authors proposed using two parameters
to formulate a gold standard.
The most commonly used assays, their basic features, and their
points of strength and weakness have been studied in detail and
discussed in several thoughtful books and reviews (Ndhlala et al.,
2010; Frankel, 2007; Prior et al., 2005; Re et al., 1999; Bartosz, 2010;
Halliwell and Gutteridge, 2007; Takashima et al., 2012). We limited
the huge number of citations, omitting many important original
investigations, which were summarized in the listed (and many
more) excellent reviews. In the following part of the present review,
we relate to the major characteristics of the most commonly
used specic methods, with emphasis on apparent inconsistencies
between the results obtained by different methods. We also discuss
possible explanations for such inconsistencies.
Before describing these methods, we wish to advise the reader
to become familiar with the large number of abbreviations used for
the different methods and groups of methods. This will help preventing confusion in reading our review and the references given
below.
2.2. Oxygen radical absorbance capacity (ORAC)
Many of the assays used to evaluate antioxidants yield information about the total radical-trapping capacity, expressed in terms
of the, so called, total radical-trapping parameter, TRAP. ORAC is
the most commonly used TRAP assay, and the most widely used
assay for evaluating antioxidants both in the industry and in the
academic institutions. Similar to other TRAP assays, it is based on
measurements of the inhibitory effect of antioxidants, as studied by
monitoring the oxidation of the uorescent protein phycoerytherin
(PE), induced either by copper ions or by AAPH. The kinetics of oxidation of the probe is followed by recording the loss of uorescence
of PE due to its oxidation.
The ORAC test, rst devised by Wayner et al. (1985), has evolved
beyond the limitations of other TRAP assays by measuring both the
effect of the antioxidant on the time preceding oxidation of the
uorescent protein (lag time) and the maximal percentage of this
inhibition. Similar to some of the other methods, it takes the reactions to apparent completion and uses the area under the curve
(AUC) for quantization. The assay can be used to evaluate peroxidation in body uids including plasma, serum as well as tissue
samples and food products.
All this makes the ORAC a very useful tool for research. The major
strong points of this method are that it is quite exible (with respect
to the possibility of using it either manually or automatically) and
relatively rapid (about an hour). Its weakest point is that it is not
suitable for measuring lipid-soluble antioxidants, many of which
are potent antioxidants. An improved ORAC version, based on usage
of uorescein instead of PE has been recently developed (Ou et al.,
2001).
2.3. Ferric reducing antioxidant power (FRAP)
The FRAP test is an example of an antioxidant capacity (AOC)
assay based on the effect of antioxidants on the single electron
transfer (SET) reaction. Originally, it was developed by Benzie and
Strain to measure the reducing power of plasma (Benzie and Strain,
1996). It is based on the capacity of antioxidants to reduce the ferric complex of 2,4,6-tripyridyl-s-triazine (Fe3+ TPTZ) to the colored
ferrous complex FE2+ TPTZ at pH 3.6. Interpretation of the results is
based on the hypothesis that the capability of water soluble antioxidants to reduce ferric ions, reects their ability to reduce ROS.
The problem with this test is that it does not take into consideration important antioxidants, including thiols (most importantly

641

GSH) and albumin but does include reductants that do not act as
antioxidants.
2.4. Trolox equivalent antioxidant capacity assay (TEAC)
The TEAC assay is based on the reduction of the ABTS radicalcation in the presence of ferry myoglobin, activated by hydrogen
peroxide. A newer version is based on decolorization of a preprepared ABTS radical cation, which has a long shelf life (about 2
days). The results have been compared to those observed for Trolox
in similar conditions of concentration and time. The strong points
of this assay are that it can be applied for both water-soluble and
lipid soluble antioxidants. This assay is the basis for more elaborate
assays, which add more time points, more samples and more complex calculations. It has been adapted for clinical studies, based on
more complex conditions (Re et al., 1999).
2.5. Evaluation of antioxidants using more than one criterion
Given the context-dependence of both terms OS and antioxidative capability, it is important to choose the assay of antioxidative
potency and/or capacity according to the aim of the study. Specically, different assays may have to be considered when the goal is to
prevent lipoprotein oxidation or to search for antioxidants that will
maximize the shelf life of a given drug and/or foodstuff. In short,
there are advantages of using several methods, each representing
a different criterion, to rank antioxidants.
Nevertheless, the very large number of assays used by different groups constitutes a major difculty in the search for potent
antioxidants because it interferes with the possibility of comparing results of different laboratories. Even when different assays are
used to evaluate the same criterion, the result may vary considerably. The long list of abbreviations given above is merely a partial
list of the many methods. Several abbreviations denote a group
of methods that often differ from each other in important details
either with respect to the general experimental approach, or the
reagents used in the actual assay and/or the units used to describe
the results. Notably, all the methods listed above measure criteria that are based on the total capacity of antioxidants to reduce
free radicals or to inhibit peroxidation of easily detectable probes
developed to assess the potency of antioxidants in terms of more
specic criteria.
In an attempt to overcome the problem of having too many
methods, several leading scientists in this eld organized in 2004
a meeting entitled The rst International Congress on Antioxidant
Methods. Similar to many other consensus conferences, The First
International Congress on Antioxidants Methods did not yield a
universal assay. Instead, it yielded a consensus to recommend
shortening the list of assays to the three most commonly used
methods. According to the general agreement on this resolution,
using one or more of the three chosen methods covers the whole
spectrum of different antioxidants, from water-soluble and lipidsoluble reducing agents to the level of ROS and the concentrations
of inhibitors of oxidation of uorescent biomarker by ROS.
In retrospect, the First Conference on Antioxidant Methods
(2004) did not stop the effort to develop improved new assays of
antioxidative potency. Furthermore, to the best of our understanding; the agreement on a short list did not prevent the development
of many more methods because in the last 5 years scientists felt that
they can develop better assays. As a result, it appears that The First
International Conference of Antioxidant Methods was also the last
such conference. A Second Conference on Antioxidant Methods
never convened and may never be held in the future (Prior et al.,
2005).
The association between pre-diagnostic serum levels of several oxidative stress indicators and Colorectal Cancer (CRC) was

Author's personal copy

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

Concentration of oxidation products

642

the initial rate of inhibited oxidation and the lag preceding rapid
peroxidation are the basic experimental factors that may be sufcient for comparing antioxidants in meaningful terms, as described
below.

No Antioxidant

3.2. Characterization of lipid peroxidation based on kinetic

proles

Antioxidant A

Antioxidant B

20

40

60

80

100

120

Time
Fig. 1. Schematic description of the accumulation of peroxidation product during
peroxidation in the absence of antioxidants (no antioxidant) and in the presence of
equal concentrations of antioxidants A and B under same conditions (both time and
concentration are given in arbitrary units).

investigated in a multi-center, nested casecontrol study (Leufkens

et al., 2012). The main results were that FRAP was not associated with CRC risk, whereas ROM levels were associated
with increased risk of CRC only in subjects with relatively short
follow-up. Based on these ndings, the authors suggest that the
association results from production of reactive oxygen species by
preclinical tumors. In our view, neither ROM and/or any other kit
of oxidative stress can be used as a diagnostic tool, in spite of the
ongoing aggressive promotions.

Denition of the effects of antioxidants in meaningful, quantitative terms under specic conditions can be achieved on the basis of
the respective kinetic proles (Denisov, 2000). In empirical terms,
antioxidants affect the kinetics of peroxidation, such that in its
presence (i) the initial rate of lipid peroxidation is lower and (ii)
the lag preceding propagation is longer. Analysis of these factors,
when analyzed as described below, can yield characterization of
antioxidants in terms of two basic factors.
In the presence of an antioxidant AH, an external free radical
X , produced at a constant rate (e.g. during peroxidation induced
by free radical generator such as AAPH) reacts with both AH and a
lipid LH either dispersed or dissolved in the liquid media
AH + X A + HX

(r.1)

LH + X + [O2 ] LO2 + HX

(r.2)

The secondary radical LO2 may, in turn, interact with the

antioxidant
AH + LO2 A + LO2 H

(r.3)

The radical A , formed via (r.1) or/and (r.3) may be rapidly

quenched either by radicals X or by LO2 via very fast reactions
3 and 4.
A + X AX

(r.4)

3. Testing and comparing antioxidants

A + LO2 A-LO2

3.1. The need for kinetic data

Hence, each molecule of an antioxidant quenches two free radicals X and/or LO2

The search for improved antioxidants also requires a quantitative terminology of antioxidative activity. To be able to compare
antioxidants, we need to develop clear denitions of the meaning
of the experimentally observable factors. The following discussion
indicates that to achieve this goal we need kinetic data.
Quite often it is assumed that if studied under the same conditions, in the presence of equal concentrations, antioxidants can be
compared on the basis of the concentration of peroxidation products, as observed at a given (arbitrarily chosen) time. This may be
very misleading because the concentration of peroxidation products, as measured at a given time point, depends on the time of the
assay, as demonstrated in Fig. 1.
In this gure we see that at any time point prior to the lag
observed in the presence of compound A (e.g. at 30 time units),
the concentration of peroxidized lipids is higher in the presence of
compound B than in the presence of A, which means that A protects
the lipid against peroxidation more effectively than B. The opposite
result (B is more efcient than A) is apparent at later time points
(e.g. at 80 time units), exhibiting the misleading nature of one time
point measurements. Subsequently, after all the antioxidants are
consumed, peroxidation occurs rapidly. Finally (after about 120
time units), all the oxidizable lipids became peroxidized, both in
the presence of any antioxidant and in its absence. Measurements
conducted later may lead to the erroneous conclusion that neither
A nor B is an antioxidant.
Measurements of area under the curve (AUC), as proposed for
evaluation of antioxidative effects in several assays (Re et al., 1999),
constitute a partial solution to the ambiguity, because it is based on
continuous recording of the formation of oxidation products. Both

AH + 2X HX + AX

(r.5)

(r.1 + r.4)

AH + 2LO2 LO2 H + A-LO2

(r.3 + r.5)

(r.6)
(r.7)

This again means that a molecule of antioxidant quenches two

radicals. Otherwise, when the production of X is relatively low and
the concentration of AH is relatively high, the radical A , formed
through (r.1) or (r.3), can be quenched either via (r.4) or via (r.5) or
via reaction with another radical A .
In the latter case,
A + A A-A

(r.8)

This means that each molecule of AH quenches only one free

radical X or LO 2 , e.g.
2AH + 2X A-A + 2HX

(2 r.4 + r.8)

(r.9)

Inhibited peroxidation (during the lag time) is characterized by

relatively slow peroxidation of substrate (as compared to the rapid
uninhibited peroxidation period that occurs after consumption
of all the antioxidant). During the inhibited phase, peroxidation
proceeds up to (almost) complete consumption of the antioxidant.
Consequently, inhibited peroxidation is characterized by the factors, namely the (time-dependent) rate of peroxidation and the
duration of the period required for consumption of the antioxidant.
The time required for complete consumption of the antioxidant can be evaluated on the basis of the reasonable assumption
that in the presence of a potent antioxidant, essentially all the
free radicals produced in the studied system are quenched by
the antioxidant ((r.6), (r.7)) and not by double-quenching of X

Author's personal copy

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

or LO2 radicals. When the antioxidant nears complete consumption, inhibited peroxidation becomes un-inhibited. This time-point,
commonly denoted the lag time (tlag ) is given by the established
expression:
tlag = f n [AH]/Ri

(1)

In this equation, [AH] is the concentration of the antioxidant

and Ri is the rate of an apparent zero order production of free
radicals (initiation), as in AAPH-induced peroxidation. The term n
denotes the maximal number of radicals quenched by one molecule
of antioxidant, (n = 2) and the term f reects the actual stoichiometry of quenching. Notably, n depends on the antioxidant structure
(e.g. for most polyphenols n = 2; see (r.6), (r.7)), whereas f is a number between 0 and 1 reects the competition between reactions
(r.4), (r.5) and reaction (r.8), as discussed above. In addition the
term f represents also deviations from ideal stoichiometry.
One implication of our analysis is that the lag preceding rapid
peroxidation (in time units) is proportional to the concentration of
the antioxidant AH (in M), inversely proportional to the rate of free
radical production (in M/t), normalized for two unit factors, namely
on intrinsic factor n (n = 2, 4 and even more) and the extrinsic factor f (f = 01), which reects the efciency of the antioxidant (that
is the average number of free radicals that an antioxidant group
can quench). Notably, the lag is almost independent of the reaction
constant of any stage in the peroxidation (k1 , k2 , kp , etc.).
The lag observed for different antioxidants varies much less than
any of the latter reaction constants and different antioxidants when
exposed at equal concentrations to the same free radical generator, at the same concentration, the accumulation of hydroperoxides
differs only in the two factors n and f. Thus, the capacity of an
antioxidant is given by a product of n and f and the main difference
between different antioxidants is not merely due to their different
capacities.
Experimentally, the initial rate of (inhibited) peroxidation, as
observed during the lag phase, is very sensitive to the presence, the
concentration and the chemical properties of the antioxidant.
The steady-state approximation to the chain peroxidation in
the presence of antioxidant (r.10) implies that the rate of inhibited
peroxidation should obey Eq. (2)
LH + LO2 + [O2 ] LO2 H + LO2
d[LH]/dt = k10 [LH] Ri /(k1 f n [AH])

(r.10)
(2)

Notably, in Eq. (2), [LH] and [AH] are the concentrations of

oxidizable substrate (e.g. lipid) and antioxidant, respectively, k1
denotes the rate constant of inhibition (r.1) and k10 denotes the
rate constant of propagation of peroxidation of LH (r.10).
During the lag phase preceding rapid peroxidation namely, prior
to complete consumption of the antioxidant, the rate of peroxidation reects the potency of the given antioxidant to suppress
the peroxidation of the substrate. Similar to the duration of the
lag phase (characterizing capacity), the potency of an antioxidant
depends on the stoichiometry factors n and f (higher values of
these factors mean that the antioxidant is more potent). Yet, unlike
the duration of the lag phase (compare Eq. (1) and Eq. (2)), the
potency depends also on the rate constant of inhibition (k1 ). Moreover, unlike n and f, the rate constants for different antioxidants
vary within several orders of magnitude (Denisov, 2000). We can,
therefore, conclude that the potency of an antioxidant is primarily determined by k1 . We can also conclude that f and n determine
the capacity (manifesting as duration of the lag), whereas the
potency (manifested as suppressing the rate of inhibited substrate peroxidation) is determined by the rate constant of the
inhibited peroxidation.
In conclusion, when peroxidation is inhibited by an antioxidant
via a mechanism of competitive inhibition, the ratio between the

643

initial rates of inhibited peroxidation observed in the presence of

two different added antioxidants reects how much the one antioxidant is more reactive than the other. Hence, this ratio relates to
the relative potencies of the antioxidants. We propose using the
term potency specically for comparison of inhibition rates (and,
respectively, the relevant rate constants) of antioxidants. Such denition may reduce the ambiguity in characterizing antioxidants and
the consequent misunderstanding.
Contradistinctively, under conditions of a constant rate of production of free radicals (as in AAPH-induced peroxidation) at any
given concentration of antioxidant, the lag reects only the capacity of the antioxidant, namely the number of free radicals that a
molecule of antioxidant can quench. We think that the appropriate
description of the results given in Fig. 1 is that antioxidant A is a
more potent than antioxidant B but the capacity of B is higher than
that of A.
4. Oxidative stress and ranking of antioxidants
Prior to relating to any method or group of methods, we rst
review the term oxidative stress and the kinetics of reactions used
to assess the oxidative stress and other common methods used to
evaluate it.
4.1. Oxidative stress, denition and quantitation
The term oxidative-stress (OS) is dened intuitively and qualitatively, the most common denition being that it describes an
imbalance between pro-oxidative factors and antioxidative factors
in vivo (Sies, 1986). In recent years, several other denitions have
been proposed. For example, rather than dening oxidative stress
in terms of an imbalance, the authors dene it as being the rate
at which oxidative damage is generated (Costantini and Verhulst,
2009). Oxidative stress, according to this denition, is a continuous variable. The rate of generation of oxidative damage depends in
a complex fashion on the balance between pro-oxidants and antioxidants, which justify the conclusion that antioxidant capacity,
by itself, is not sufcient to make inference about oxidative stress.
To be able to relate to the development of oxidative damage, we
need at least one marker of oxidative damage. We proposed that
the term oxidative stress of any given type (Dotan et al., 2004)
is context-dependent and that OS evaluation should be based on
the use of the most sensitive probe of oxidative damage. According to our experience, the most sensitive biomarkers are lipids.
Hence, lipid peroxidation is likely to reect the balance between
pro-oxidative and antioxidative components of the studied systems.
The use of OS to evaluate the activity of antioxidants requires
understanding of the shortcomings of the existing common assays
and the general problems associated with the methods used to
assess OS. In view of the results of recent investigations that demonstrated the physiological importance of ROS and the dangers of
using excessive concentration of antioxidants, it is essential to be
able to quantitate both OS and antioxidative activity. We propose
using criteria on comparison between the kinetics of peroxidation
observed in the presence and absence of the assayed antioxidant.
Specically, we propose quantifying the antioxidants on the
basis of the prolongation of the lag prior to rapid peroxidation, as
observed in the ratio between the lag in the presence and absence
of the studied antioxidant (Pinchuk et al., 2011).
4.2. Terminology used to characterize antioxidative effect(s)
Comparison between different antioxidants is commonly
ambiguous. The results of most, if not all the methods are based

Author's personal copy

644

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

on measurements of the concentrations of specic products, conducted under well dened conditions with respect to the rate of
production of free radicals, the temperature, the viscosity and the
concentrations of oxidizable lipids and antioxidants. In most assayprotocols, the characteristic measured factor is expressed as a ratio
between the measured characteristic value and the value that characterizes the effect of an arbitrarily chosen inhibitor under the same
conditions.
The choice of an experimental factor to be used for characterization of the effects of inhibitors, as well as the choice of terms to
quantitate the inhibitory effect of antioxidants, is not trivial. At least
ve different terms have been used in relating to the protection
of various molecules against oxidation. In a literature search (ISIS
WOS database) for antioxidant* , we found that the term capacity has been used many more times (6035) than the other terms:
potency (1828), strength (775), effective (625) or active (133).
Quite often, the difference between two antioxidants is
described by ambiguous statements. For example, in relating to the
difference between the inhibitory effect of two antioxidants A and
B, as observed in a given experiment, the results are often described
in ambiguous terms such as A is a better antioxidant than B or A
is a stronger (or a more powerful) antioxidant than B or that A is
more effective than B or that the strength (or efciency, or capability, capacity or potency) of A is higher than that of B. In other
words, no universally accepted criterion has been established to
enable quantitative comparison between antioxidants.
In Fig. 1 described above, antioxidant A prolongs the lag less than
antioxidant B, indicating that A is a weaker antioxidant. Hence,
at the time that A is completely consumed, considerable fraction of
B is left, so that the phase of inhibited peroxidation is longer. Thus,
it can be concluded that the antioxidative capacity of B is higher.
By apparent contrast, accumulation of hydroperoxides during the
lag time is smaller in the presence of A than in the presence of B,
namely, based on the initial rate of lipid peroxidation during the
phase of inhibited peroxidation, A is a more potent antioxidant
than B or the antioxidative potency of A is higher than that of B.
5. Problems associated with assaying OS and ranking
antioxidants
As discussed above, the capacity of antioxidants is contextdependent. The biomedical relevance of the commonly used
methods of evaluating antioxidants is questionable because of the
following considerations.
5.1. The lack of a universal criterion of OS
The use of kinetic assays of OS to evaluate antioxidants is limited
by the lack of a universal criterion of OS. Accordingly, we should rst
consider the major problem of determining the oxidative stress,
which is the lack of a universal criterion. The three most commonly
used assays for the determination of total antioxidant capacity of
human serum (ORAC, TEAC and FRAP) have been compared. The
results demonstrated that there was a weak but signicant linear
correlation between serum ORAC and serum FRAP and no correlation either between serum ORAC and serum TEAC or between
serum FRAP and serum TEAC (Cao and Prior, 1998). Furthermore,
the authors describe the differences between the scope and limitations of the different assays and recall that different extraction
techniques in the ORAC assay also make a gross differentiation
between aqueous and lipid-soluble antioxidants. Since then, many
studies indicated that OS, as assayed by different methods, do nor
correlate with each other and we had reasons to propose (Dotan
et al., 2004) that there are different types of oxidative stress. Nevertheless, the use of the term OS without specifying the method

used to determine it, is very misleading. Several studies show that

even when OS is evaluated on the basis of the same criterion using
the same process but monitored by different biomarkers, the results
may differ considerably. Our interpretation of inconsistencies is
based on the complexity of the different measurements, particularly when the method involves external probes and/or different
partitioning between compartments.
5.2. Experimental details
Another important source of inconsistencies is the reliability of
the methods used to measure biomarkers of OS and antioxidants.
This is particularly serious for the customary, commercially available assays, whose usage increases markedly in the last several
years. In addition, most tests depend prominently on specic experimental details and the quality of measurements of the biomarkers
differs between different labs, adding other problematic factors.
In general terms, most assays are very sensitive to experimental details. Hence, different labs (and even different investigators
within the same lab) might measure the same biomarkers in a
slightly different way, which may result in different outcomes. This
is likely to be especially serious when researchers use kits rather
than manual methods. More than 200 commercial kits, many of
which can be read by Elisa readers, are offered by many producers as indicators of OS. These include assays based on tests of lipid
peroxidation products, DNA fragmentation products, antioxidative
enzymes or genes responsible for over-expression of antioxidative
proteins. Producers include Amsbio, Trevigen, Biovision, Oxford
biomedical research, Life science Neogen, Enzo life sciences Biovision and many more. The different kits in fact assay different
biomarkers. Yet, they are all presented as OS assays, which is quite
problematic due to the lack of correlations between the results of
the various assays with each other.
The advantages of using kits are clear but it is not clean of disadvantages as described in the brilliant presentation of Peter Webster
of the EPA (Webster, unpublished result). In his comparison of the
kits to manual tests, the author raises a cautionary words regarding the need to make sure that the test is conducted within the
linear range.
5.3. The time-dependence of measurements
The concentrations of the products of oxidation that serve as
biomarkers are time-dependent. Hence the concentrations of the
different biomarkers depend on the time of the test. The investigations of OS can be divided to two groups: (i) studies of different
steady-state oxidative status of different populations, such as
healthy people in comparison to patients of different diseases or
people of different races, genders or environments and (ii) the
effect of changes in OS due to acute events such as exposure to
environmental factors, physical activity or unusual dietary events
(e.g. starvation or unusually heavy meal). Both these types of study
require ex vivo monitoring of the increase of the concentration of a
product of oxidation and/or of the decrease of concentration of the
oxidizing species. The concentrations of these biomarkers, as well
as of externally added probes, depend on the time of withdraw of
the relevant biological uids (mostly blood).
In studies of the effects of changes, there is an additional complexity due to the different time-scales of the response to changes.
As an example, the effect of physical activity on the concentration
of different enzymes in the erythrocytes, as measured immediately after exercise is problematic because changes in DNA take a
few minutes, whereas changes in enzymes that reside in the erythrocytes, might take days or even weeks (Nikolaidis et al., 2008).
This difference is independent of the method used to monitor
the production of a given biomarker used to study a given criterion.

Author's personal copy

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

It is also true for studies of the production of different biomarkers

of similar nature, as demonstrated by the test of the effect of a
relatively short but extensive exercise on the activity of two antioxidative enzymes immediately after the test as biomarkers of
oxidative stress. In this case, the activity of catalase was almost
unaffected, as could have been expected from the concentration of a
product of a slow process. More problematic is the nding that both
the ratio GSH/GSSG and carbonyl activity in the erythrocytes show
signicant change due to a bout physical exercise when measured
immediately after physical exercise.
We think that the apparent contradiction is due to lysis of the
white blood cells, which do not necessarily reect the OS in the
blood. In the context of the issue discussed in the present communication it is of importance to avoid artifacts.
5.4. Comparison of data given in terms of different units
Different units are used to express OS in terms of the concentration or activity of biomarkers. Even when the same biomarker
has been used to estimate the OS under apparently equal conditions, the reported results often differed not only quantitatively,
but in their units too. For example, the activity of SOD, which is
regarded as a biomarker of OS, was reported in sedentary young
men (age 1822 years) as being 1.28 0.1 U/ml (Pepe et al., 2009),
9.3 0.7 units/l (Brites et al., 1999) and 782.1 59.3 U/g Hb (Chang
et al., 2002). To be able to compare the different measurements, we
must convert the results to common units. The problem of comparing and standardizing the results of different assays for SOD
activity was discussed many years ago (Beyer and Fridovich, 1987;
Greenwald, 1985) and still remains actual. In addition, no comparison is possible between the values normalized for hemoglobin and
absolute enzyme activities given above.
5.5. The difference between reactions in solutions and at
interfaces (location, location, location)
As discussed above, most biomedically important peroxidation processes occur at water/oil interfaces of lipoproteins and
membranes. Such processes depend on more factors, than reactions in solutions. They include the partitioning of the inducer and
inhibitors (antioxidants) between the two phases, the concentration of PUFA in the outer layer and the physical properties of the
latter layers, particularly charge (for peroxidation induced by transition metal ions) and lateral diffusion of the lipids.
In conclusion, evaluation of both factors oxidative stress and
antioxidative capacity is problematic. This general statement is
based on the context-dependent nature of both. Specically, the
possibility of comparing results from different studies is limited
by the large number of different criteria, measured by different
biomarkers. Furthermore, the obtained values are sensitive to multiple experimental details, including the time of measurements
with respect to the onset of the reaction.
The most misleading assays of OS are those assays in which
the studied biomarker is measured at one wavelength, at one time
point (rather than kinetic proles) and using one concentration.
Such measurements are very problematic because (i) it is not clear
whether the measured values represent steady state concentrations of the biomarkers and (ii) it is not certain that the measured
values are within the linear range of the concentration dependence
of absorbance. Aggregation is not rear in dispersions of lipids in
water occurs and such processes result in a major effect on the light
scattering, thus interfering with the reliability of the spectrometric
determination of the employed biomarker.
These considerations are of course valid for determination of
most, if not all, the measurements of different biomarkers by commercial kits (Zembron-Lacny et al., 2008). In these kits, the users are

645

rarely instructed to conduct concentration-dependent measurements. Hence, in spite of their convenience and ease of utilization,
OS kits, particularly HTS assays, should be conducted with caution.
In conclusion, evaluation of both terms oxidative stress and
antioxidative capacity is questionable. These general statements
are based on their context-dependent nature, the large number
of different criteria, measured by methods of relatively different
factors, the sensitivity of the measured criteria to experimental
details including the time of measurements relative to the onset
of the reaction and the consequent inability of comparing results
from different studies. The study of capacity of antioxidants in
solution is further complicated by the difference between the reactions in solution and at interface. In view of these difculties, we
propose assaying the inhibition of lipids peroxidation by different
antioxidants in the given system. Specically, we propose dening
antioxidant criteria on the basis of the concentration dependence
of their inhibition of the substrate peroxidation in the relevant system. This is what we have done in our recent study (Pinchuk et al.,
2011), as summarized below.
At the present time, the whole concept of the good antioxidants
protect us against the bad ROS is being questioned. Specically,
ROS plays important physiological roles, which means that lowering its concentration below a critical value is harmful (Birringer and
Ristow, 2012; Ristow and Schmeisser, 2011).

6. Ranking antioxidants on the basis of OS methods: the

effects of antioxidants on ex vivo peroxidation
In view of the problematic nature of the interpretation of
the terms oxidative capacity and/or oxidative potency particularly with respect to their relevance to the reactions that occur
at lipidwater interfaces, we considered the possibility of using
oxidative stress assays to evaluate the antioxidative potential. In
other words, being interested in inhibiting peroxidation in a given
system, we propose testing antioxidants in that system, as detailed
below.
Specically, when an antioxidant is added to lipid-containing
aggregates (including LDL, HDL, serum or membranes), it reduces
the rate, or even prevents peroxidation. The dependence of the prolongation of the lag on the concentration of the antioxidant can be
used to quantitate the context-dependent antioxidative capacity.
In our recent study, we used this approach to study the activity of
various antioxidants against copper-induced peroxidation of serum
lipids (Pinchuk et al., 2011). A similar approach can be used to assay
the antioxidative capability of different antioxidants to inhibit peroxidation of oxidizable lipids present in other self assemblies. In
these studies antioxidants exert their antioxidative activity against
peroxidation that occurs at the relevant interfaces. The scope, possibilities and limitations of these assays are described and discussed
below.

6.1. Inhibition of peroxidation of PUFA in serum lipoproteins

Our study is based on the kinetics of lipid peroxidation in unfractionated plasma or serum, induced ex vivo by copper and followed
by spectrophotometric recording of the production of hydroperoxides (Schnitzer et al., 1998). A similar method was developed for
organic radical generators-induced peroxidation of serum (Atkin
et al., 2005). For reasons described above, inhibition of peroxidation was expressed in terms of prolongation of the lag. We proposed
expressing the antioxidative capacity in terms of the concentration of the given antioxidant required for doubling the lag. The
results may be compared to that of a standard probe under similar
conditions and concentration (e.g. Trolox).

Author's personal copy

646

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647

Our optimized assay relates to the total antioxidative capacity,

due to both lipophilic and hydrophilic antioxidants. Notably, our
assay, similar to peroxidation in vivo and unlike other assays, is
affected in a complex fashion by (1) the partition coefcients of the
antioxidants, (2) the migration of free radicals within and between
lipid particles, and (3) the metal-chelating properties. However,
the sera is a multifunctional environment containing albumin and
other natural antioxidants, all of which may interact with the studied antioxidant in various ways (either inhibiting or promoting the
true antioxidative capacity).
For example, epicatechin is a good antioxidant but at concentrations above 15 milimolar, epicatechin acts as a competitive
inhibitor to copper by binding to albumin and inducing peroxidation. Also, the amount and composition of the serum lipids varies
between individuals, thus affecting the peroxidation. Nonetheless,
the concentration of an antioxidant required to double the lag
appears to be independent of the serum (Pinchuk et al., 2011).
6.2. Evaluation of antioxidative potency in model membrane
systems
In emulsions, micro-emulsions or lipid-surfactant mixed micellar systems, it is possible to monitor oxidation reactions under
varying conditions. In such samples, we can vary the effects of
additives, including antioxidants in a controllable fashion, one at
a time, to yield reliable information on the effects of antioxidants.
By contrast, the dependence of lipid peroxidation and of the effect
of antioxidants on factors such as the composition and physical
properties of biomembranes is hard to study due to the complexity
of the biomembranes. Hence, several investors, including us, look
for a simple model system that will allow spectroscopic evaluation
of the activity of antioxidants without chemical isolation.
Liposomes are relatively simple and convenient model systems for assessing the antioxidative mechanism and potency of
natural and synthetic antioxidants (Schroder et al., 2000; Saija
et al., 2001). The most studied liposomes are those made of
pure polyunsaturated fatty acids (PUFA), alkyl esters of PUFA and
phosphatidylcholine (PC) esteried by PUFA at position sn-2, and
dispersed in aqueous solutions either in the form of phospholipid vesicles (liposomes) (for a review, see Barclay, 1993) or as
PC/detergent mixed micelles (for example, see Rekdal and Melo,
1995).
In an attempt to evaluate the effects of the composition and
physical properties of these model membranes we used liposomes
made of PLPC and other liposomes made of the (non-oxidizable)
palmitoyl-oleyl PC (POPC) as a major component of the liposomal
membrane along with PLPC and/or cholesterol. In an attempt to
mimic in vivo peroxidation, we studied the effects of added negatively charged phospholipid (PS) and phosphatidic acid (PA) to the
well dened liposomal PLPC.
Notably, the use of liposome-based assays for evaluation of
antioxidants is limited by the very high sensitivity of liposome
peroxidation to preparation method, difculties in standardization
of size distribution, structure, impurities and many other poorly
controlled factors (Schnitzer et al., 2007).
Unlike peroxidation of lipoproteins and serum lipids, peroxidation of liposomal lipids, in the absence of antioxidant may occur
without apparent lag phase. Accordingly, for liposomal systems we
have chosen to relate to the antioxidant concentration, doubling the
tmax (time at which the maximal peroxidation rate is achieved), as
a quantitative criterion of antioxidant capacity (Gal et al., 2007).
Notably, tmax correlates with lag value. C2lag parameter, therefore, correlates with C2tmax parameter. Both C2lag and C2tmax can
be further normalized by some standard antioxidant, e.g. Trolox.
The latter normalization yield relative unitless values, which show
how much antioxidant is needed to double the lag (or tmax ), as

compared to how much Trolox is required to double the lag (or

tmax ). In fact, such approach is similar to the concept of Trolox
equivalents widely used in ranking antioxidants.
7. Conclusions (Take-home messages)
The most pronounced effect of antioxidants is prolongation of
the lag prior to rapid peroxidation, which ends only after complete consumption of the antioxidants. Hence, the lag reects
the capacity of the antioxidant. The other meaningful parameter that can be used to characterize antioxidants is the initial rate
of formation of peroxidation products. This factor relates to the
potency of the antioxidant.
Determination of both the capacity and the potency of an antioxidant requires kinetic measurements. The effect of an antioxidant
on the concentration of oxidation products at a given time point
cannot be used to characterize the antioxidant.
Both terms oxidative stress and antioxidant capacity are
context-dependent. Specically, the major effect of antioxidants
on the peroxidation of a probe in solution is not indicative of the
characteristics of the antioxidant at water/oil interfaces. Hence, to
be relevant to biomedical context, antioxidants should be ranked
on the basis of methods currently used to evaluate oxidative
stress.
Abbreviations
One of the most apparent outcomes of the search for an easy
and reliable assay of antioxidative potency is the large number of
abbreviations it yielded. The following abbreviations constitute a
partial list of different criteria used to describe the total capacity of
antioxidant either to reduce free radicals or to inhibit oxidation of
easily detectable oxidizable probes.
AOC antioxidant capacity
TOS total oxidation status, based on inhibition of the oxidation
of Fe2+ to Fe3+
TAS total antioxidant status
TRAP total radical-trapping parameter
TEAC total Trolox equivalent antioxidative capacity
ORAC oxygen radicals absorbance capacity
FRAP ferric reducing antioxidant power
The methods used to test antioxidants can be characterized
according to their mechanism of action. Commonly used abbreviations:
HAT hydrogen atom transfer
ET electron transfer
Other commonly used abbreviations:
ROS reactive oxygen species
PUFA polyunsaturated fatty acid
FR free radical(s)
References
Atkin, M.A., Gasper, A.M., Ullegaddi, R., Powers, H.J., 2005. Oxidative susceptibility of unfractionated serum or plasma: response to antioxidants in vitro and to
antioxidant supplementation. Clinical Chemistry 51, 21382144.
Barclay, L.R.C., 1993. 1992 Syntex Award Lecture Model biomembranes Quantitative studies of peroxidation, antioxidant action, partitioning, and oxidative
stress. Canadian Journal of Chemistry Revue Canadienne de Chimie 71, 116.
Bartosz, G., 2010. Non-enzymatic antioxidant capacity assays: limitations of use in
biomedicine. Free Radical Research 44, 711720.

Author's personal copy

I. Pinchuk et al. / Chemistry and Physics of Lipids 165 (2012) 638647
Benzie, I.F.F., Strain, J.J., 1996. The ferric reducing ability of plasma (FRAP) as a
measure of antioxidant power: the FRAP assay. Analytical Biochemistry 239,
7076.
Beyer, W.F., Fridovich, I., 1987. Assaying for superoxide-dismutase activity some
large consequences of minor changes in conditions. Analytical Biochemistry 161,
559566.
Birringer, M., Ristow, M., 2012. Efcacy and risks of supplementation with antioxidants. Ernahrungs Umschau 59, 1014.
Bjelakovic, G., Nikolova, D., Gluud, L.L., Simonetti, R.G., Gluud, C., 2007. Mortality in
randomized trials of antioxidant supplements for primary and secondary prevention systematic review and meta-analysis. JAMA Journal of the American
Medical Association 297, 842857.
Blumberg, J.B., Frei, B., 2007. Why clinical trials of vitamin E and cardiovascular
diseases may be fatally awed. Commentary on the relationship between dose
of vitamin E and suppression of oxidative stress in humans. Free Radical Biology
and Medicine 43, 13741376.
Brites, F.D., Evelson, P.A., Christiansen, M.G., Nicol, M.F., Basilico, M.J., Wikinski, R.W.,
Llesuy, S.F., 1999. Soccer players under regular training show oxidative stress
but an improved plasma antioxidant status. Clinical Science 96, 381385.
Cao, G.H., Prior, R.L., 1998. Comparison of different analytical methods for assessing
total antioxidant capacity of human serum. Clinical Chemistry 44, 13091315.
Chang, C.K., Tseng, H.F., Hsuuw, Y.D., Chan, W.H., Shieh, L.C., 2002. Higher LDL oxidation at rest and after a rugby game in weekend warriors. Annals of Nutrition
and Metabolism 46, 103107.
Costantini, D., Verhulst, S., 2009. Does high antioxidant capacity indicate low oxidative stress? Functional Ecology 23, 506509.
Denisov, E.T., 2000. Handbook of Antioxidants: Bond Dissociation Energies, Rate
Constants, Activation Energies and Enthalpies of Reactions, 2nd ed. CRC Press,
Florida.
Dotan, Y., Lichtenberg, D., Pinchuk, I., 2004. Lipid peroxidation cannot be used as a
universal criterion of oxidative stress. Progress in Lipid Research 43, 200227.
Dotan, Y., Lichtenberg, D., Pinchuk, I., 2009a. No evidence supports vitamin E indiscriminate supplementation. Biofactors 35, 469473.
Dotan, Y., Pinchuk, I., Lichtenberg, D., Leshno, M., 2009b. Decision analysis supports
the paradigm that indiscriminate supplementation of vitamin E does more harm
than good. Arteriosclerosis Thrombosis and Vascular Biology 29, 13041309.
Frankel, E.N., 2007. Antioxidants in Food and Biology The Oily Press. Bridgwater,
England.
Gal, S., Lichtenberg, D., Bor, A., Pinchuk, I., 2007. Copper-induced peroxidation of
phosphatidylserine-containing liposomes is inhibited by nanomolar concentrations of specic antioxidants. Chemistry and Physics of Lipids 150, 186203.
Greenwald, R.A. (Ed.), 1985. CRC Handbook for Oxygen Radical Research. CRC Press,
Boca Raton, FL.
Halliwell, B., Gutteridge, J., 2007. Free Radicals in Biology and Medicine, 4th ed.
Oxford University Press, New York.
Harman, D., 1956. Aging: a theory based on free radical and radiation chemistry.
Journal of Gerontology 11, 298300.
Harman, D., 2009. Origin and evolution of the free radical theory of aging: a brief
personal history, 19542009. Biogerontology 10, 773781.
Leufkens, A.M., Van Duijnhoven, F.J.B., Boshuizen, H.C., Siersema, P.D., Kunst, A.E.,
Mouw, T., Tjonneland, A., Olsen, A., Overvad, K., Boutron-Ruault, M.C., ClavelChapelon, F., Morois, S., Krogh, V., Tumino, R., Panico, S., Polidoro, S., Palli, D.,
Kaaks, R., Teucher, B., Pischon, T., Trichopoulou, A., Orfanos, P., Goufa, I., Peeters,
P.H.M., Skeie, G., Braaten, T., Rodriguez, L., Lujan-Barroso, L., Sanchez-Perez, M.J.,
Navarro, C., Barricarte, A., Zackrisson, S., Almquist, M., Hallmans, G., Palmqvist,
R., Tsilidis, K.K., Khaw, K.T., Wareham, N., Gallo, V., Jenab, M., Riboli, E., Buenode-Mesquita, H.B., 2012. Educational level and risk of colorectal cancer in EPIC
with specic reference to tumor location. International Journal of Cancer 130,
622630.
Miller, E.R., Pastor-Barriuso, R., Dalal, D., Riemersma, R.A., Appel, L.J., Guallar, E., 2005.
Meta-analysis: high-dosage vitamin E supplementation may increase all-cause
mortality. Annals of Internal Medicine 142, 3746.
Milman, U., Blum, S., Shapira, C., Aronson, D., Miller-Lotan, R., Anbinder, Y., Alshiek,
J., Bennett, L., Kostenko, M., Landau, M., Keidar, S., Levy, Y., Khemlin, A., Radan,
A., Levy, A.P., 2008. Vitamin e supplementation reduces cardiovascular events
in a subgroup of middle-aged individuals with both type 2 diabetes mellitus

647

and the haptoglobin 22 genotype A prospective double-blinded clinical trial.

Arteriosclerosis Thrombosis and Vascular Biology 28, 341347.
Ndhlala, A.R., Moyo, M., Van Staden, J., 2010. Natural antioxidants: fascinating or
mythical biomolecules? Molecules 15, 69056930.
Niki, E., 2011. Do free radicals play causal role in atherosclerosis? Low density
lipoprotein oxidation and vitamin E revisited. Journal of Clinical Biochemistry
and Nutrition 48, 37.
Nikolaidis, M.G., Jamurtas, A.Z., Paschalis, V., Fatouros, I.G., Koutedakis, Y., Kouretas,
D., 2008. The effect of muscle-damaging exercise on blood and skeletal muscle
oxidative stress: magnitude and time-course considerations. Sports Medicine
38, 579606.
Ou, B.X., Hampsch-Woodill, M., Prior, R.L., 2001. Development and validation of an
improved oxygen radical absorbance capacity assay using uorescein as the
uorescent probe. Journal of Agricultural and Food Chemistry 49, 46194626.
Pepe, H., Balci, S.S., Revan, S., Akalin, P.P., Kurtoglu, F., 2009. Comparison of oxidative
stress and antioxidant capacity before and after running exercises in both sexes.
Gender Medicine 6, 587595.
Pinchuk, I., Shoval, H., Bor, A., Schnitzer, E., Dotan, Y., Lichtenberg, D., 2011. Ranking antioxidants based on their effect on human serum lipids peroxidation.
Chemistry and Physics of Lipids 164, 4248.
Prior, R.L., Wu, X.L., Schaich, K., 2005. Standardized methods for the determination
of antioxidant capacity and phenolics in foods and dietary supplements. Journal
of Agricultural and Food Chemistry 53, 42904302.
Pryor, W.A., 2000. Vitamin E and heart disease: basic science to clinical intervention
trials. Free Radical Biology and Medicine 28, 141164.
Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., Rice-Evans, C., 1999.
Antioxidant activity applying an improved ABTS radical cation decolorization
assay. Free Radical Biology and Medicine 26, 12311237.
Rekdal, K., Melo, T.B., 1995. UV-initiated autoxidation of methyl linoleate in micelles
studied by optical-absorption. Chemistry and Physics of Lipids 75, 127136.
Ristow, M., Schmeisser, S., 2011. Extending life span by increasing oxidative stress.
Free Radical Biology and Medicine 51, 327336.
Saija, A., Tomaino, A., Pellegrino, M.L., Giuffrida, N., Trombetta, D., Castelli, F., 2001.
In vitro evaluation of the antioxidant activity and biomembrane interaction of
the lazaroid U-74389G. Life Sciences 68, 13511366.
Schnitzer, E., Pinchuk, I., Bor, A., Fainaru, M., Samuni, A.M., Lichtenberg, D., 1998.
Lipid oxidation in unfractionated serum and plasma. Chemistry and Physics of
Lipids 92, 151170.
Schnitzer, E., Pinchuk, I., Lichtenberg, D., 2007. Peroxidation of liposomal lipids.
European Biophysics Journal with Biophysics Letters 36, 499515.
Schroder, H., Navarro, E., Tramullas, A., Mora, J., Galiano, D., 2000. Nutrition antioxidant status and oxidative stress in professional basketball players: effects of
a three compound antioxidative supplement. International Journal of Sports
Medicine 21, 146150.
Shoval, H., Lichtenberg, D., Gazit, E., 2007. The molecular mechanisms of the antiamyloid effects of phenols. Amyloid-Journal of Protein Folding Disorders 14,
7387.
Sies, H., 1986. Biochemistry of Oxidative Stress. Angewandte Chemie International
Edition in English 25, 10581071.
Steinberg, D., Parthasarathy, S., Carew, T.E., Khoo, J.C., Witztum, J.L., 1989. Beyond
cholesterol modications of low-density lipoprotein that increase its atherogenicity. New England Journal of Medicine 320, 915924.
Takashima, M., Horie, M., Shichiri, M., Hagihara, Y., Yoshida, Y., Niki, E., 2012. Assessment of antioxidant capacity for scavenging free radicals in vitro: a rational basis
and practical application. Free Radical Biology and Medicine 52, 12421252.
Wayner, D.D.M., Burton, G.W., Ingold, K.U., Locke, S., 1985. Quantitative measurement of the total, peroxyl radical-trapping antioxidant capability of
human-blood plasma by controlled peroxidation the important contribution
made by plasma-proteins. FEBS Letters 187, 3337.
Webster., P. Analytical procedures and method validation. IPPC Discharges Monitoring Workshop. Available at: http://www.epa.ie/downloads/advice/water/
quality/Method%20Assessment%20&%20Validation.pdf.
Witztum, J.L., 1998. To E or not to E How do we tell? Circulation 98, 27852787.
Zembron-Lacny, A., Ostapiuk, J., Slowinska-Lisowska, M., Witkowski, K., Szyszka,
K., 2008. Pro-antioxidant ratio in healthy men exposed to muscle-damaging
resistance exercise. Journal of Physiology and Biochemistry 64, 2735.