Review Article

Antioxidants as therapies: can we improve on nature?

Michael P. Murphy
Medical Research Council Mitochondrial Biology Unit, Wellcome Trust/MRC Building, Cambridge CB2 0XY, UK
a r t i c l e i n f o
Keywords:
Antioxidant
Drug development
Oxidative damage
Pharmacology
Free radicals
a b s t r a c t
Although oxidative damage contributes to many pathologies the use of naturally occurring, small-molecule
antioxidants as therapies for these disorders has not been successful. Here I discuss some of the reasons
this may be so. Paramount among these are the difculties in delivering enough of the antioxidant to the
intracellular location required to decrease pathological oxidative damage and the challenge of assessing
whether the intervention has actually decreased oxidative damage in the patient to a therapeutically useful
extent. To develop effective antioxidant therapies the best strategy may be to create new chemical entities
designed to detoxify a dened reactive oxygen species-dependent process that underlies a particular
pathology, in the same way a conventional drug is designed to modulate a biochemical process, rather than
applying antioxidants in an unfocused manner. In developing new antioxidants it will be useful to utilize
endogenous processes to activate and recycle the molecules in parallel with the targeting of compounds to
cells and organelles in ways that are not limited by the constraints that impair the distribution of
endogenous antioxidants. In short, I suggest that the future development of antioxidant therapies should
be viewed as an arm of drug development, utilizing focused approaches similar to those of medicinal
chemistry and pharmacology, rather than as a branch of nutrition.
& 2013 Elsevier Inc. All rights reserved.
Contents
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Why have natural antioxidants not been successful as therapies?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
How can we assess the efcacy of antioxidant therapies in vivo? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Strategies to design better antioxidants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Introduction
As biomedical researchers in the oxidative stress and free radical
elds, we all have an embarrassing secret. There is compelling
evidence that oxidative damage contributes to cell death and dysfunc-
tion in a wide range of pathologies and that deciencies in endogen-
ous antioxidant defense and repair systems underlie many diseases
in fact it is difcult to think of a pathology in which oxidative stress
does not play a role [1,2]. Consequently it is a short and obvious step to
suggest that antioxidants should be an effective treatment for many
diseases [2,3]. Furthermore, as many endogenous, small-molecule
antioxidants (e.g., vitamin E, vitamin C, coenzyme Q) are cheap, orally
bioavailable, safe in large doses, and absorbed and recycled within our
bodies, they are excellent candidates for translation to the clinic [4].
However, the problem is that when we assess these antioxidants in
controlled clinical trials they show very disappointing outcomes [58].
In fact, the larger and better conducted the clinical trail, the lower the
therapeutic effect tends to be [5]. In this survey I consider some of the
reasons trials of natural antioxidants have been disappointing and
suggest ways of developing new chemical entities as antioxidants that
may help overcome these limitations.
Why have natural antioxidants not been successful as
therapies?
A large number of well-conducted clinical trials have been
carried out using several different antioxidants on a range of
pathologies with little improvement in clinical outcome for the
Contents lists available at SciVerse ScienceDirect
journal homepage: www.elsevier.com/locate/freeradbiomed
Free Radical Biology and Medicine
0891-5849/$ - see front matter & 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.freeradbiomed.2013.04.010
E-mail addresses: michael.murphy@mrc-mbu.cam.ac.uk,
mpm@mrc-mbu.cam.ac.uk
Please cite this article as: Murphy, MP. Antioxidants as therapies: can we improve on nature? Free Radic. Biol. Med. (2013), http://dx.doi.
org/10.1016/j.freeradbiomed.2013.04.010i
Free Radical Biology and Medicine ()
patients [5]. Although it could be that if we hit on the right
combination of disease, antioxidant, and treatment regimen a
positive outcome would result, we should face up to the evidence
and consider why antioxidants have been so disappointing
clinically.
The rst point to consider is whether these results are telling us
that oxidative damage is not actually that important in disease
pathology and is a poor therapeutic target. This goes against our
instincts, as it is evident that high levels of reactive oxygen species
and oxidative damage lead to cell death in vitro, in vivo, and in
patients. However, do we know for sure that oxidative damage is a
signicant factor in the progression of a particular pathology? Of
course, an increase in oxidative damage during a disease process
could occur because it correlates with, rather than contributes to,
the pathology. Thus, it is important to demonstrate that prevent-
ing oxidative damage actually leads to a clinically signicant
improvement [9,10]. We need to show not just that an antioxidant
therapy decreases oxidative damage, but that the decrease in
oxidative damage has a clinically signicant impact. My view is
that there will turn out to be many disorders in which oxidative
damage does underlie the pathology and for which antioxidant
therapies will be clinically useful. This view is supported by, for
example, animal studies in which removal of antioxidant defenses
increases pathology or overexpression of antioxidant enzymes
improves the outcome [1113]. However, this has to be demon-
strated in patients for each disorder case by case. It is likely that in
many cases oxidative damage will occur but not be directly
causative [14], in which cases antioxidant therapy is unlikely to
ever show benet.
To address this critical point it would be good to measure
changes in oxidative damage in patients and show that this
damage correlated with the disorder and that prevention of the
oxidative damage by the antioxidant treatment in question led to a
clinical improvement [15,16]. At the moment the technical dif-
culties in assessing oxidative damage in vivo mean that we are
often uncertain if the antioxidant therapy has actually led to a
decrease in the extent of oxidative damage [14,17]. A corollary is
that in many clinical trials it is difcult to determine if the lack of a
success of an antioxidant is simply because it has not been
delivered in sufcient amounts to prevent pathological oxidative
damage. Alternatively, it may be that oxidative damage was
prevented but that this had no impact on the pathology. Until
these measurements are available, the outcome of many well-
conducted antioxidant trials will remain ambiguous.
Another potential limitation to the development of antioxi-
dants is whether they can block the many positive sides of
reactive oxygen species, for example, by disrupting bacterial killing
by neutrophils. More generally, the controlled production of
reactive molecules such as hydrogen peroxide can act as a redox
signal, linking metabolic processes by altering protein activity, for
example, by the reversible modication of cysteine residues [18].
Consequently, the indiscriminate use of an antioxidant could do
more harm than good by disrupting essential signaling processes.
However, this concern can be addressed by measuring particular
changes in reactive oxygen species and/or oxidative damage and
by correlating these changes with the clinical outcome.
For an antioxidant to affect oxidative damage caused by a
particular reactive species its activity must increase local antioxidant
defenses sufciently above background, endogenous levels to have
an impact on oxidative damage. For many naturally occurring
antioxidant compounds, such as vitamin E or vitamin C, the amount
present within a particular tissue, cell, or organelle is regulated at
multiple levels by alterations in absorption, distribution, uptake, and
metabolism, with numerous feedback and homeostatic mechanisms
to defend the levels present. Consequently, it can be difcult to
increase many endogenous antioxidants much above the normal
levels. Exceptions to this may occur in the rare cases in which
antioxidant deciency is due to a genetic defect in enzymes
responsible for the synthesis of an antioxidant or in the proteins
responsible for the absorption and distribution of a dietary antiox-
idant. In such cases dietary supplementation may work well to
overcome the genetic defect. A related issue is that antioxidants do
not act in isolation, but as an integrated defense network that
combines the actions of small molecules with protective and repair
enzymatic systems. All of these are subject to multiple layers of
regulation in response to changes in reactive species and oxidative
damage. Consequently, decreases in one antioxidant may be com-
pensated for by upregulation of other defenses; conversely an
exogenous antioxidant may cause a compensatory downregulation
of endogenous protection, with no net increase in antioxidant
defenses. Finally, and critically, the antioxidant is often relatively
evenly dispersed throughout the body and cell, whereas the
oxidative damage may be localized to particular cell types or
organelles such as mitochondria. Consequently the overall antiox-
idant content within a tissue or cell may be adequate, but the local
antioxidant concentration may be insufcient to deal with particular
hotspots of oxidative damage. Thus there are many plausible reasons
to explain the poor outcome of clinical trials of antioxidants.
Howcan we assess the efcacy of antioxidant therapies in vivo?
To improve the outcome of antioxidant trials in pathologies,
there is a pressing need for methods to assess whether a particular
antioxidant actually improves detoxication of a particular reac-
tive species at a given stage in the pathology, at the right location,
by increasing antioxidant defenses sufciently above endogenous
levels. This information, in conjunction with an assessment of the
clinical outcome, is essential to determine if an antioxidant
therapy is unlikely ever to be successful or whether improving
the antioxidant efcacy is worth pursuing. To achieve this, the
development of better markers of reactive oxygen species and
oxidative damage in vivo are essential. This is a major weakness in
our current investigations in experimental animal studies and is
an even greater limitation in patient trials in which our options to
assess antioxidant efcacy are far more limited. Ideally, we would
like robust readouts of the levels of the particular reactive species
within various tissues, cell types, and subcellular compartments
and how these change over the course of the pathology. In
addition we would like to be able to relate these changes to the
levels of the exogenous antioxidants, the changes in posttransla-
tional modications to proteins, the levels of metabolites, the
expression of genes at the transcriptional and protein levels, and
also the amounts and types of oxidative damage. Clearly, we are a
very long way from this ideal scenario, even in experimental
animals, and the situation in patients is even less informative as
analyses are restricted to a few oxidative damage markers in
accessible body uids. Even so, we need to assess as well as we can
the effects of antioxidant administration so as to answer the
critical questions: does the antioxidant intervention decrease a
dened reactive species or type of damage at a particular location
and time and does this improve the clinical outcome? There is an
old saying about Christopher Columbus: he didn't know where he
was going when he set out, didn't know where he was when he
got there, and didn't know where he had been when he got back.
Our technical limitations in assessing oxidative damage and
reactive oxygen species in vivo mean that we are often in a
similar, uncertain situation. The development of better biomarkers
of oxidative damage is essential for the development of antiox-
idant therapies.
M.P. Murphy / Free Radical Biology and Medicine () 2
Please cite this article as: Murphy, MP. Antioxidants as therapies: can we improve on nature? Free Radic. Biol. Med. (2013), http://dx.doi.
org/10.1016/j.freeradbiomed.2013.04.010i
Strategies to design better antioxidants
A critical point in designing effective antioxidant therapies is
that they should be focused on a particular reactive oxygen species
or form of oxidative damage that contributes to the disease by a
specic mechanism in a dened tissue, organ, or organelle and
that occurs at a critical point in a pathological process. The past
development of antioxidant therapies has tended to focus on
endogenous antioxidants. This was because these were generally
regarded as safe. Furthermore, the accumulation of oxidative
damage that occurred when these antioxidants were lacking
indicated that they performed an antioxidant role in vivo,
although whether their levels could be increased was not gen-
erally assessed. A concern is that many of these antioxidants are of
broad specicity in the reactive oxygen species they detoxify and
in their intracellular location; consequently, many of these inter-
ventions are unfocused. For comparison, we would not consider a
trial of a general protein kinase inhibitor or of assessing average
protein phosphorylation level as particularly useful. Instead, we
would expect the kinase inhibitor to be selectively targeted to a
particular pathway, which was chosen based on an understanding
of the pathological mechanisms that led to the disease. In much
the same way we should focus on the development of novel
antioxidants designed and produced by medicinal chemistry to
intervene in a particular oxidative process, just as we would for
the development of any other drug.
In designing an ideal new chemical entity to be antioxidant,
we can adapt many of the positive aspects of endogenous
antioxidants. For example, the antioxidant should be designed to
intercept and detoxify biologically important reactive species and
to be activated and recycled by endogenous processes. By incor-
porating these traits into novel antioxidants their activity and
duration of effectiveness in vivo can be enhanced. Furthermore,
the use of articial antioxidants opens up the possibility of
bypassing many of the limitations of endogenous compounds, in
particular by designing compounds for delivery to particular
organs, tissues, cells, and subcellular locations. For example, there
are a number of approaches that have been developed to deliver
compounds selectively to mitochondria in vivo that have facili-
tated the development of more effective antioxidants [19].
An antioxidant molecule can be effective in a number of ways.
The most straightforward mechanism is by reacting directly with a
particular reactive species, thereby diverting it from damage, but
to design a molecule to do this requires a clear understanding of
the chemistry involved and of the alternative fates for the target
reactive species in vivo. For example, superoxide is a critical
pathological reactive oxygen species whose production underlies
much oxidative stress. Consequently antioxidants that selectively
degrade superoxide should be useful therapeutically. However, to
do this an antioxidant designed to degrade superoxide would have
to compete with the endogenous superoxide dismutase enzymes,
which can very rapidly dismutate the free radical at close to the
diffusion limit and which are also present in vivo at very high
concentrations. It is a major challenge for an articial antioxidant
to enhance the rate of degradation of superoxide within a tissue
above background levels. Therefore it is important to consider that
antioxidants designed to degrade superoxide, and indeed all other
reactive species, may operate effectively in vitro, but may not be
able to alter the levels of the target reactive oxygen species in vivo
because of competing endogenous processes. A further vital point
is that once an antioxidant has intercepted a reactive species
the product of this reaction needs to be unreactive and, ideally,
rapidly recycled back to its active antioxidant form. Endogenous
antioxidants have evolved to do this by coupling with metabolic
pathways. For example, within mitochondria the antioxidants
coenzyme Q and vitamin E are recycled by the respiratory chain,
and these processes have been utilized for articial antioxidants
[20]. Similarly, the recycling of the selenium-containing antiox-
idant ebselen occurs by coupling with the endogenous thioredoxin
system [21,22]. Antioxidant compounds can also act by targeting
the production of the reactive species. For example, now that we
understand how the avin site of complex I produces superoxide
[23], it is in principle possible to design compounds that bind to
this site to decrease pathological ROS production. Another strategy
is that instead of itself sequestering the reactive species, a
compound can upregulate the expression of endogenous antiox-
idant defenses, for example, by increasing the activity of transcrip-
tion factors that control the levels of antioxidant enzymes [24]. Of
course, in these scenarios the importance of demonstrating the
selectivity and antioxidant efcacy of the intervention is just as
important as for antioxidants that directly react with damaging
species.
Conclusion
The poor results in clinical trials have dampened enthusiasm
for the development of antioxidant therapies. My view is that this
response is too pessimistic and that the evidence for oxidative
damage and disrupted redox processes in pathology compels us to
revisit this issue, but with better designed antioxidants. To do this
it may be helpful to stop thinking of antioxidant therapy as a
branch of nutrition and instead view it as a drug discovery and
development program following the same approaches that we use
for other new medicines. In this it should be possible to design a
molecule with the appropriate chemistry to selectively decrease
the concentration of a dened damaging reactive species at a
specic stage of the pathology within the appropriate organ/cell/
subcellular location, sufciently to bring about a clinically signi-
cant improvement. To put it another way, we need to move away
from seeing antioxidants as a general, nonspecic panacea and
instead use them to target specic druggable redox processes
using the tools of medicinal chemistry and pharmacology.
Acknowledgments
Work in the author's laboratory is supported by the Medical
Research Council (UK). The author holds patents and has nancial
interests in the development of mitochondria-targeted antioxi-
dants and therapies.
References
[1] Halliwell, B. Oxidative stress and neurodegeneration: where are we now? J.
Neurochem. 97:16341658; 2006.
[2] Finkel, T. Opinion: radical medicine: treating ageing to cure disease. Nat. Rev.
Mol. Cell. Biol. 6:971976; 2005.
[3] Balaban, R. S.; Nemoto, S.; Finkel, T. Mitochondria, oxidants, and aging. Cell
120:483495; 2005.
[4] Halliwell, B.; Gutteridge, J. M. C.; Cross, C. E. Free radicals, antioxidants, and
human disease: where are we now? J. Lab. Clin. Med. 119:598620; 1992.
[5] Bjelakovic, G.; Nikolova, D.; Gluud, L. L.; Simonetti, R. G.; Gluud, C. Antioxidant
supplements for prevention of mortality in healthy participants and patients
with various diseases. Cochrane Database Syst. Rev. 3:CD007176; 2012.
[6] Iannitti, T.; Palmieri, B. Antioxidant therapy effectiveness: an up to date. Eur.
Rev. Med. Pharmacol. Sci. 13:245278; 2009.
[7] Ueda, S.; Yasunari, K. What we learnt from randomized clinical trials and
cohort studies of antioxidant vitamin? Focus on vitamin E and cardiovascular
disease Curr. Pharm. Biotechnol. 7:6972; 2006.
[8] Halliwell, B. The antioxidant paradox. Lancet 355:11791180; 2000.
[9] Murphy, M. P.; Partridge, L. Toward a control theory analysis of aging. Annu.
Rev. Biochem. 77:777798; 2008.
[10] Gutteridge, J. M.; Halliwell, B. Antioxidants: molecules, medicines, and myths.
Biochem. Biophys. Res. Commun. 393:561564; 2010.
[11] Schriner, S. E.; Linford, N. J.; Martin, G. M.; Treuting, P.; Ogburn, C. E.; Emond,
M.; Coskun, P. E.; Ladiges, W.; Wolf, N.; Van Remmen, H.; Wallace, D. C.;
M.P. Murphy / Free Radical Biology and Medicine () 3
Please cite this article as: Murphy, MP. Antioxidants as therapies: can we improve on nature? Free Radic. Biol. Med. (2013), http://dx.doi.
org/10.1016/j.freeradbiomed.2013.04.010i
Rabinovitch, A. T. Extension of murine life span by overexpression of catalase
targeted to mitochondria. Science 308:19091911; 2005.
[12] Li, Y.; Huang, T. -T.; Carlson, E. J.; Melov, S.; Ursell, P. C.; Olson, J. L.; Noble, L. J.;
Yoshimura, M. P.; Berger, C.; Chan, P. H.; Wallace, D. C.; Epstein, C. J. Dilated
cardiomyopathy and neonatal lethality in mutant mice lacking manganese
superoxide dismutase. Nat. Genet. 11:376381; 1995.
[13] Ames, B. N.; Shigenaga, M. K.; Hagen, T. M. Oxidants, antioxidants, and the
degenerative diseases of aging. Proc. Natl. Acad. Sci. USA 90:79157922; 1993.
[14] Cochem, H. M.; Murphy, M. P. Can antioxidants be effective therapeutics?
Curr. Opin. Invest. Drugs 11:426431; 2010.
[15] Murphy, M. P.; Holmgren, A.; Larsson, N. G.; Halliwell, B.; Chang, C. J.;
Kalyanaraman, B.; Rhee, S. G.; Thornalley, P. J.; Partridge, L.; Gems, D.;
Nystrom, T.; Belousov, V.; Schumacker, P. T.; Winterbourn, C. C. Unraveling
the biological roles of reactive oxygen species. Cell Metab. 13:361366; 2011.
[16] Winterbourn, C. C. Reconciling the chemistry and biology of reactive oxygen
species. Nat. Chem. Biol. 4:278287; 2008.
[17] Halliwell, B.; Whiteman, M. Measuring reactive species and oxidative damage
in vivo and in cell culture: how should you do it and what do the results
mean? Br. J. Pharmacol. 142:231255; 2004.
[18] Janssen-Heininger, Y. M.; Mossman, B. T.; Heintz, N. H.; Forman, H. J.;
Kalyanaraman, B.; Finkel, T.; Stamler, J. S.; Rhee, S. G.; van der Vliet, A.
Redox-based regulation of signal transduction: principles, pitfalls, and pro-
mises. Free Radic. Biol. Med. 45:117; 2008.
[19] Smith, R. A.; Hartley, R. C.; Cochem, H. M.; Murphy, M. P. Mitochondrial
pharmacology. Trends Pharmacol. Sci. 33:341352; 2012.
[20] James, A. M.; Cochem, H. M.; Smith, R. A.; Murphy, M. P. Interactions of
mitochondria-targeted and untargeted ubiquinones with the mitochondrial
respiratory chain and reactive oxygen species: implications for the use of
exogenous ubiquinones as therapies and experimental tools. J. Biol. Chem.
280:2129521312; 2005.
[21] Zhao, R.; Holmgren, A. A novel antioxidant mechanism of ebselen involving
ebselen diselenide, a substrate of mammalian thioredoxin and thioredoxin
reductase. J. Biol. Chem. 277:3945639462; 2002.
[22] Zhao, R.; Masayasu, H.; Holmgren, A. Ebselen: a substrate for human
thioredoxin reductase strongly stimulating its hydroperoxide reductase activ-
ity and a superfast thioredoxin oxidant. Proc. Natl. Acad. Sci. USA 99:8579-
8584; 2002.
[23] Kussmaul, L.; Hirst, J. The mechanism of superoxide production by NADH:
ubiquinone oxidoreductase (complex I) from bovine heart mitochondria. Proc.
Natl. Acad. Sci. USA 103:76077612; 2006.
[24] Hayes, J. D.; McMahon, M.; Chowdhry, S.; Dinkova-Kostova, A. T. Cancer
chemoprevention mechanisms mediated through the Keap1Nrf2 pathway.
Antioxid. Redox Signaling 13:17131748; 2010.
M.P. Murphy / Free Radical Biology and Medicine () 4
Please cite this article as: Murphy, MP. Antioxidants as therapies: can we improve on nature? Free Radic. Biol. Med. (2013), http://dx.doi.
org/10.1016/j.freeradbiomed.2013.04.010i