Overview Articles

The Oxidative Cost of Reproduction:

Theoretical Questions and
Alternative Mechanisms
CARLOS ALONSO-ALVAREZ, TARA CANELO, AND ANA NGELA ROMERO-HARO

The role of reproduction in increasing oxidative stress may be fundamental to understanding aging and the evolutionary trade-off between
survival and fecundity. However, contradictory results among experimental studies have challenged the oxidative cost of reproduction. Limitations
in experimental design may explain inconsistent findings. Nonetheless, some authors have argued that the hypothesis is founded on the faulty
assumptions of an energy-based allocation trade-off and a direct positive link between metabolic rates and the production of reactive oxygen species
(ROS) by the mitochondria. We propose that evolutionary trade-offs do not require the allocation of limiting resources and that reproduction may
result in ROS-induced oxidative stress without increased mitochondrial ROS. We discuss the previously published oxidative-shielding hypothesis
and propose two new hypotheses: hormesis and extortion for reproduction. These hypotheses aim to explain the counterintuitive results indicating
that there is less oxidative damage in animals allowed to breed compared with those prevented from reproducing.

Keywords: hormesis, life history trade-offs, senescence, oxidative stress, terminal investment.

T he cost of reproduction is a cornerstone of and longevity (e.g., De Jong and van Noordwijk 1992). An
evolutionary theory (e.g., Zera and Harshman 2001).
Recently, oxidative stress has been proposed as a new com-
oxidative cost of reproduction would also support evolution-
ary theories linking reproduction and aging (Williams 1957,
ponent of that cost. Oxidative stress can be defined as the Kirkwood 1977). Along these lines, the disposable soma
imbalance between the rate of production of reactive species theory (Kirkwood 1977) suggests that the somatic line of the
(mainly reactive oxygen species; ROS) by cell metabolism and organism (the soma) is maintained (e.g., by antioxidant
the level or efficiency of antioxidant defenses (i.e., including mechanisms) while favoring reproduction (the activity of
repair mechanisms and intrinsic passive defensese.g., cell the germline). When this is not possible, the soma becomes
membrane composition)which leads to damage to lipids, disposable, and death occurs. The oxidative cost of repro-
proteins, or DNA(Halliwell and Gutteridge 2007). We must duction would magnify this effect, accelerating aging and
also consider regulatory processes controlling ROS genera- promoting a trade-off in the allocation of resources (i.e.,
tion and the disruption of redox-signaling mechanisms as energy; see below) between reproduction and survival.
components of oxidative stress (Jones 2006). Redox signal- In the early 2000s, two studies demonstrated that the short-
ing is the specific, usually reversible, oxidation or reduction term (a few hours) survival of fruit flies exposed to paraquat
modification of cellular signaling pathway components by a (an ROS generator) declines when animals are stimulated to
reactive species or by the shift in redox state of a responsive invest in reproduction (Salmon etal. 2001, Wang etal. 2001).
group such as a dithioldisulphide couple (Forman et al. These were the first studies explicitly testing whether repro-
2004). Redox-signaling disruption has often been linked to duction leads to oxidative stress and whether oxidative stress
pathological states (e.g., Go and Jones 2011). induced by reproduction reduces longevity. This could, in
Oxidative stress has traditionally been associated with the turn, explain aging. This idea defines the oxidative cost of
aging process (Sies and Cadenas 1985, but see, e.g., Barja reproduction within the framework of two complementary
2013) but may potentially shape other life-history traits (e.g., hypotheses: (1) reproductive effort increases oxidative stress,
Isaksson et al. 2011). Therefore, the hypothesis that repro- and (2) oxidative stress due to reproductive effort reduces
duction leads to oxidative stress has attracted the attention of lifespan. Since these initial studies, most research has tested
evolutionary biologists, because it may provide a mechanism the first hypothesis, frequently reporting nonsignificant or
for the hypothesized negative link between reproduction even counterintuitive results (i.e., breeding animals may

BioScience XX: 113. The Author(s) 2016. Published by Oxford University Press on behalf of the American Institute of Biological Sciences. All rights
reserved. For Permissions, please e-mail: journals.permissions@oup.com.
doi:10.1093/biosci/biw176 Advance Access publication XX XXXX, XXXX

http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 1

Overview Articles
endure lower oxidative stress than do nonbreeders; e.g.,
Speakman and Garratt 2014, Blount etal. 2016). The second
hypothesis is, however, relevant from an evolutionary point
of view. A cost can only be considered as such when it trans-
lates into a lower contribution to the gene pool of the next
generation (i.e., Darwinian fitness). This can result from
lower survival (reduced longevity), but also from reduced
reproductive success. In fact, the second hypothesis should
be redefined to include the possibility of a reduction in the
breeding output of subsequent reproductive events.
The lack of consistent evidence supporting the first
hypothesis has been explained by some limitations in experi-
mental design (Metcalfe and Monaghan 2013) but also on
the basis of potential theoretical problems (Speakman and
Garratt 2014). In this article, we provide (a) counterargu-
ments to the two main theoretical criticisms of the oxidative-
cost-of-reproduction hypothesis and (b) explanations for
the contradictory evidence suggesting that reproductive
effort may, at least in some cases, reduce oxidative stress. To
address the last point, we discuss several hypotheses. The
recently proposed oxidative-shielding hypothesis (Blount
et al. 2016) and two new alternative or complementary
hypotheses are explained and debated. To conclude, some
final remarks and directions for future research are provided.
First theoretical problem: The requirement of an
energy allocation trade-off
The first theoretical problem argued to explain nonsig-
nificant results obtained when testing the oxidative cost of
reproduction is that there is no trade-off in the allocation of
energy between the maintenance of a good oxidative status
(maintenance of antioxidant or repair mechanisms to attain
low oxidative stress and survive) and reproduction (i.e.,
Speakman and Garratt 2014). This idea was first formulated
as an answer to a previous debate emphasizing that experi-
ments manipulating breeding effort under captivity com-
monly involve ad libitum resources (e.g., food, water, and
temperature; Metcalfe and Monaghan 2013). Under these
supposedly optimal conditions, animals should be able to
simultaneously invest in reproduction and antioxidants or
repair mechanisms, masking the trade-off.
The idea of an absence of an energy allocation trade-off in
the oxidative cost of reproduction was based on intriguing
findings obtained in fruit flies and Caenorhabditis elegans
(see Speakman and Garratt 2014). Animals with ablated
germlines live longer, but those whose whole gonads were
removed did not (Hsin and Kenyon 1999). A complex net-
work of at least four interconnected molecular signaling
pathways generated in somatic tissues, probably including
part of the gonad, promotes increased lifespan but is blocked
by signals from the germline (Hansen etal. 2013). This net-
work suggests that biochemical signals could regulate lon-
gevity and reproduction simultaneously, independent of the
availability of energy resources (Edward and Chapman 2011,
Alonso-Alvarez and Velando 2012). Moreover, reproduc-
tion and longevity seem to be uncoupled in Caenorhabditis
2 BioScience XXXX XXXX / Vol. XX No. X http://bioscience.oxfordjournals.org
and Drosophila mutants that are able to live longer without
apparent fecundity loss or that show a reduction in fecun-
dity without increases in lifespan (Flatt 2011 and references
therein). Nonetheless, we must note that these studies
mostly analyzed effects on egg production, with unknown
impacts on lifetime reproductive success, offspring quality
or fecundity. In fact, we should ask why these mutants are
apparently absent in the wild (Briga and Verhulst 2015).
However, the cited network is intimately related to the
metabolism of energy macromolecules (glucose and lipids;
Hansen etal. 2013), which suggests that signals could play
a role in regulating trade-offs that will only be fully under-
stood when the complete mechanism is revealed.
Social insects are also paradoxical for evolutionary theo-
ries of aging. Queens are long lived compared with other
castes but also invest more in reproduction, at least in terms
of egg production (workers also invest in reproduction via
queen and colony maintenance). Seehuus and colleagues
(2006) proposed that the reproductive regulatory pathways
were remodeled in the honey bee (Apis mellifera) to extend
life. In particular, these authors showed that vitellogenin, a
protein involved in the production of egg proteins in many
taxa including vertebrates (e.g., Olsson et al. 2009), also
functions as an antioxidant prolonging the lifespan. This
implies that queen bees are able to simultaneously invest in
reproduction and survival. However, this mechanism seems
to be restricted to eusocial Hymenoptera. A direct causal
link between vitellogenin and longevity has not been estab-
lished in vertebrates and is unclear in other invertebrates
such as C. elegans and fruitflies, in which high vitellogenin
(Vg) expression is even linked to reduced longevity (e.g., Ren
and Hughes 2014, Seah et al 2016). Moreover, the genetic
background of eusocial insects has led to the interpretation
of the colony as a superorganism, with queens acting as the
germline and workers as its disposable soma (e.g., Flatt etal.
2013). Therefore, testing the OCR would require the com-
parison of different queens or hives on the basis of reproduc-
tive output but not different castes within a hive.
In any case, we propose that the requirement of an
energy-based trade-off in the oxidative-cost-of-reproduc-
tion hypothesis is unnecessary considering two arguments:
(1) nonenergetic resources (micronutrients essential for
antioxidant protection, such as some vitamins, carotenoids,
or cofactors in antioxidant enzymes) could be involved (see
also Mller et al. 2000, Alonso-Alvarez and Velando 2012,
Blount etal. 2016), and (2) the trade-off can also be estab-
lished without any limiting resource at all.
The second argument is probably the most definitive one.
Tatar and Carey (1995) proposed that reproduction is not
an energetic resource sink limiting self-maintenance but
instead causes direct damage, impairing or inhibiting those
processes involved in self-maintenance (see figure 1; Tatar
2001, Zera and Harshman 2001). Resources would not be
necessary for a trade-off if the associated cost is unavoidable
(i.e., an obligatory cost; Speakman 2008, Alonso-Alvarez
and Velando 2012). The same idea was suggested in our first
 Overview Articles

Reproducon metabolism, reducing ROS generation

(e.g., Speakman 2008, Salin etal. 2015).
ROS
OS The mechanism allows individuals with
high metabolic rates to reduce ROS pro-
duction (Speakman and Garratt 2014
and references therein).
Acquision Resources Allocaon O
Obligatory cost
Speakman and Garratt (2014) argued
that the apparent lack of a direct link
between metabolic rate and mitochon-
drial ROS production is a key weakness
Oxidave
O damage
of the oxidative-cost-of-reproduction
hypothesis. However, recently, some
Survival
urvival
authors have theoretically (in mathe-
Figure 1. The classical resource-allocation trade-off based on the Y model matical models) and empirically (in an
from de Jong and van Noordwijk (1992) and the obligatory trade-off based on experiment with insects) demonstrated
an inevitable cost (Tatar and Carey 1995). The investment in a trait can not that oxidative damage is more closely
only divert limiting resources for another trait (e.g., energy, macronutrients, related to the biosynthetic rate (which is
and antioxidants) but also trigger an inevitable cost, such as oxidative damage. involved in growth but also in reproduc-
A direct trade-off may thus coexist with a resource allocation trade-off. tion) than to metabolic rates (see Hou
However, this also means that the oxidative cost of reproduction may involve and Amunugama 2015, Amunugama
a trade-off without involving limiting resources. Abbreviation: ROS, reactive et al. 2016). These studies also suggest
oxygen species. that variation in the percent free radical
leak or uncoupling are less relevant. We
also argue that ROS generated from cel-
study on the oxidative cost of reproduction (Alonso-Alvarez lular activities other than respiration can potentially lead to
etal. 2004; see also Wiersma etal. 2004). We proposed that an oxidative cost of reproduction.
the higher metabolic rate during reproduction should inevi- Mitochondrial oxidative phosphorylation is not the only
tably induce higher mitochondrial ROS production and, endogenous source of ROS. Extramitochondrial enzymes
accordingly, oxidative stress. However, this scenario leads us are important producers (figure 2). Among them, the
to the second theoretical problem, which we discuss below. major contributors to ROS production, at least in mam-
malian models, are the nicotinamide adenine dinucleo-
Second theoretical problem: Metabolic rates are tide phosphate (NADPH) oxidases (also NOX), which are
uncorrelated to ROS production involved in superoxide production (Bedard and Krause
Reproductive animals commonly face higher metabolic rates 2007, Covarrubias et al. 2008). They are present in most
compared with nonreproductive ones (e.g., Speakman 2008). body parts, including the immune system (i.e., phagocytes)
As we mentioned above, it was originally proposed that this and reproductive tissues (Bedard and Krause 2007). In fact,
could lead to oxidative stress, explaining the oxidative cost NOX is the only enzyme whose main function is the produc-
of reproduction (Alonso-Alvarez etal. 2004, Wiersma etal. tion of ROS and is involved in redox signaling (Drummond
2004). This assumes that ROS are produced in direct pro- and Sobey 2014). It has been suggested that NOX isoforms
portion to metabolic rate as an unavoidable result of the may generate ROS in some tissues, even at a higher level
activity of the mitochondrial electron transport (respiratory) than mitochondria (e.g., Krause 2007, Brown and Griendling
chain, whose main function is to produce adenosine triphos- 2009, Dymkowska etal. 2015). Furthermore, we should con-
phate (ATP) by oxidative phosphorylation (e.g., Speakman sider that mitochondrial and extramitochondrial ROS might
and Garratt 2014). However, the link between metabolic impose different costs. Accordingly, in some Caenorhabditis
rate and oxidative stress has been seriously questioned. mutants, mild levels of mitochondrial superoxide extend the
Variation in the percent free radical leak (the percentage of lifespan (see mitohormesis below), whereas extramitochon-
total electron flow leading to ROS generation in the respi- drial superoxide may accelerate aging (Schaar etal. 2015).
ratory chain) causes the disproportionality between ROS Interestingly, high levels of energy macronutrients acti-
production and oxygen consumption (Barja 2013). During vate NADPH oxidases. High circulating glucose levels lead
mild aerobic exercise, a high rate of oxygen consumption to an increased ROS production by NOX in rats (Serpillon
leads to low local partial O2 pressure in the mitochondria. etal. 2009). Hypertriglyceridemic mice showed a high activ-
This, in turn, increases the efficiency of the respiratory ity of NOX and xanthine oxidases in the liver and higher
chain reducing ROS production (detailed in Barja 2013). hepatic levels of lipid peroxidation, protein carbonyls, and
Mitochondria also produce heat instead of ATP by uncou- oxidized glutathione (Alberici et al. 2009). In contrast,
pling oxidative phosphorylation. Uncoupling proteins in the the same mice showed elevated body metabolic rates and
inner membrane (UCPs) short-circuit mitochondrial energy mild mitochondrial uncoupling, leading to reduced ROS

http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 3

Overview Articles

Figure 2. A simplified diagram illustrating the extramitochondrial sources of reactive oxygen species (ROS; based on
Bedard and Krause 2007, Covarrubias etal. 2008, Drummond and Sobey 2014). Nicotinamide adenine dinucleotide
phosphate (NADPH) oxidases are considered the main extramitochondrial ROS source within the cell (see the main text).
They have mostly been described in the cell membrane, where they release superoxide (O2-) to the extracellular space. They
can also be found in the nucleus and endoplasmic reticulum membranes. Xanthine oxidases (XO) are other important
sources of ROS. They are located in the cell membrane, peroxisome, and cytoplasm, releasing hydrogen peroxide from
hypoxanthine or xanthine (purine bases). Nitric oxide synthases (NOS) are located in the cytoplasm, releasing nitric oxide
(NO) or superoxide from L-arginine depending on its state (i.e., coupled or uncoupled NOS, respectively). ROS generated
by NADPH oxidases also elicit ROS formation from XO and NOS. Other ROS sources include lipoxygenases (LOX) and
cyclooxygenases (COX), which are located in the cell membrane and generate peroxides from fatty acids (FA). LOX can
also be found in the endoplasmic reticulum. In addition to XO, the peroxisome holds fatty acyl-CoA oxidases, generating
hydrogen peroxide. The cytochrome P450 in the endoplasmic reticulum also contributes to ROS production by releasing
superoxide. Finally, P450, COX, and even NADPH oxidases (NOX4; Sakellariou Jackson and Vasilaki 2014) are involved
in the respiratory chain in the mitochondrion.

production by mitochondria (Alberici et al. 2009). These
authors interpreted these findings as a protective response
against a rise in ROS levels due to extramitochondrial
enzymes. We must note that the liver is an important storage
site for lipids in small mammals (which represent the animal
models in most studies testing the hypothesis; e.g., Metcalfe
and Monaghan 2013, Blount etal. 2016), with its mass and
metabolism increasing during reproduction (Speakman
2008). Therefore, we can hypothesize that, rather than an
overall increase in metabolic rates, the increased levels of
macronutrients that allow for reproduction could be the
key to activating extramitochondrial ROS sources. This may
ultimately promote oxidative stress among breeders as a part
of the direct obligatory cost of reproduction (see the section
on the extortion hypothesis below).
Hypotheses to explain lower oxidative damage
among breeders
We have presented two theoretical arguments that may
explain the lack of significance in the results of many studies
testing the oxidative cost of reproduction, but why do breed-
ing animals surpass nonbreeding individuals in antioxidant

4 BioScience XXXX XXXX / Vol. XX No. X http://bioscience.oxfordjournals.org

 Overview Articles

Table 1. Assumptions and predictions in hypotheses explaining lower oxidative damage in breeders compared with that
in nonbreeders.
Oxidative shielding
Summary Breeders preemptively activate a
decrease in oxidative damage to avoid
future fitness reductions
Assumptions A1 Some unknown physiological
mechanism exists and allows for
synchronizing oxidative shielding with
the start of reproductive effort
Hormesis during reproduction
Reproductive effort triggers a
compensatory response reducing
oxidative damage and improving
fitness
Oxidative stress or some unknown
redox signal activated by oxidative
stress triggers the hormetic
mechanism
Extortion for reproduction
In nonbreeders, the maintenance of
large macronutrient amounts required for
reproduction increases oxidative damage,
forcing animals to breed as early as
possible
Some signaling pathway exists and is
activated in nonbreeders to facilitate
reproduction via behavioral changes

A2 Reproductive investment can High levels of macronutrients increase

A3
Predictions P1 Breeders show lower oxidative-damage
levels at the start of reproduction
compared with nonbreeders
P2 Oxidative shielding remains activated
as long as needed to protect offspring
P3 Oxidative shielding should mostly be
present in tissues influencing offspring
quality (i.e., reproductive tissues
including gametes)
P4 The offspring increases quality and
survival prospects
P5 The mechanism becomes inefficient
at older ages because of the
accumulation of oxidative damage
P6
P7
constitute a mild stressor
Breeders transitorily endure mild
oxidative stress levels
Breeders subsequently lower
oxidative stress and show
lower damage values than do
nonbreeders
The decline in oxidative damage
should be present in any tissue
favoring adult fitness
Parents enlarge lifespan and/or
fecundity
The mechanism becomes
inefficient at older ages because
of the accumulation of oxidative
damage
The artificial inhibition of ROS
production or antioxidant
supplementation during
reproduction reduces hormesis
and its fitness consequences
oxidative damage and/or alter redox signaling
Reproductive opportunity can naturally
be constrained while experiencing high
macronutrient availability
Nonbreeders increase oxidative
damage when circulating or storing high
macronutrient levels
Nonbreeders reduce oxidative damage when
breeding, but damage values could still be
higher than those in nonbreeders carrying
lower macronutrient amounts (Fig. 4)
The oxidative damage should first be
present in tissues maintaining high
macronutrient levels
Nonbreeders maintaining high
macronutrient levels reduce lifespan
via oxidative damage
The mechanism should act or be reinforced
at advanced ages as future breeding
opportunities fade
The artificial inhibition of ROS production
or antioxidant supplementation delays
reproduction in nonbreeders circulating
or storing high macronutrient levels
Nonreproductive animals with large
macronutrient levels experience behavioral
changes to increase breeding opportunities

protection? This finding has been relatively frequent, par-
ticularly among studies manipulating reproductive status
(comparing breeders versus nonbreeders) in females of
mammal species (reviewed in Blount etal. 2016), although
examples in male mammals (e.g., Garratt et al. 2012,
Salomon etal. 2013) and birds (Costantini etal. 2014) also
exist. Should we accept that reproduction reduces oxidative
stress, at least under some circumstances? Reduced levels of
oxidative damage in breeders compared with those of non-
breeders would potentially favor longevity, challenging evo-
lutionary theory proposing the classical trade-off between
reproduction and survival (e.g., De Jong and van Noordwijk
1992). To explain these counterintuitive findings, three
hypotheses are discussed (see also table 1, figures 3 and 4).
Oxidative shielding.This hypothesis was formulated on the
basis of results from meta-analyses including both cor-
relational and experimental tests of the oxidative cost
of reproduction in females of homeothermic vertebrates
(Blount etal. 2016). These authors propose that reductions
in oxidative damage are pre-emptive and function to shield
mothers, and in particular their gametes and developing off-
spring, from critical levels of oxidative insults that inevitably
increase as a consequence of reproductive effort (Blount
etal. 2016, p. 495). We assume that, to be preemptive, some
unknown physiological mechanism should synchronize
the oxidative shielding with the start of the reproductive
effort (table 1, A1). Moreover, the cited reductions could be
due to lower ROS production and/or increased antioxidant
protection (Blount et al. 2016). The authors also proposed
that increased UCP activity and estrogen-mediated signals
favoring antioxidant protection may be involved (Blount
etal. 2016). The idea was mostly supported by mammalian
studies showing lower oxidative-damage levels in breeding
females compared with those of nonbreeding ones (Blount
et al. 2016). The statistical effect was driven by decreased

http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 5

Overview Articles

a
1. Nonreproducve stage 2. Prereproducve stage 3. Reproducon

Reproducon Reproducon
Reproducon

OXIDATIVE DAMAGE
Pre-empve protecon (-)

Survival Survival Survival

b
1. Nonreproducve stage 2. Prereproducve stage 3. Reproducon

Reproducon Reproducon Reproducon

OXIDATIVE STRESS
(-) SIGNAL
(+)

Survival Survival Survival

rviva

c
1. Nonreproducve stage 2. Prereproducve stage 3. Reproducon

Reproducon Reproducon Reproducon

SIGNAL
(+) ROS
OS
Oxidized compounds ( )
MN
MN MN
( ) ( )
OBLIGATORY
OBLIGATORY
COST
COST
Survival Survival Survival

Figure 3. Hypotheses aiming to explain the lower oxidative damage among breeders compared with that among
nonbreeders. (a) The oxidative-shielding hypothesis implies that animals preemptively (stage 2) reduce reactive oxygen
species (ROS) production and/or increase antioxidant levels in order to protect themselves and their offspring from
oxidative damage during reproduction. The mechanism might be activated in advance of the reproductive event. (b) The
hormesis-during-reproduction hypothesis suggests that certain intermediate ROS or oxidative-damage levels generated
by reproduction trigger some unknown redox-signaling pathway activating a hormetic (compensatory) response that
should subsequently reduce oxidative damage. The mechanism would be initiated from the start of the reproductive
event. (c) The extortion-for-reproduction hypothesis assumes that animals increase body stores and/or circulating levels
of macronutrients (MN; i.e., carbohydrates, lipids, and proteins) in order to reproduce (1, 2, 3). Macronutrients not
diverted to reproduction would become oxidized or stimulate NADPH oxidases, increasing ROS and inflicting oxidative
damage (2). This process implies an obligatory cost in the trade-off between reproduction and self-maintenance (figure
1). High ROS or oxidative-damage levels would trigger redox signals encouraging reproduction as early as possible.
When reproduction takes place (3), macronutrient amounts decline; therefore, the oxidative damage is reduced. At
that time, ROS, perhaps derived from increased enzyme metabolism in cells (figure 2), would drive the oxidative cost of
reproduction, but circulating oxidized compounds due to resource allocation to reproduction may still contribute to that
cost. We predict that the efficiency of (a) or (b) mechanisms becomes weaker with time because of aging, whereas (c) should
be reinforced with age as opportunities for breeding decline (see also table 1).

6 BioScience XXXX XXXX / Vol. XX No. X http://bioscience.oxfordjournals.org

 Overview Articles

levels in organs to remain low? Hormone-

Nonreproductive Prereproductive Reproduction related mechanisms (estrogens) or mito-
stage stage chondrial uncoupling are not stable
processes. Perhaps we should exclusively
predict a preemptive increase in protec-
tive mechanisms (via antioxidant levels
or reduced ROS production) and then a
Extortion subsequent reduction in oxidative dam-
Oxidative stress

age (e.g., figure 3a). Second, although

Hormesis reduced levels of oxidative damage in
follicles may support the protection of
the offspring, why should we expect
the same shielding process in somatic
Oxidative shielding
tissues, such as the heart or liver (table
1, P3)? Third, why would somatic non-
reproductive tissues in breeders show
lower (not merely equal) damage levels
than those in nonbreeders? Finally, the
hypothesis is difficult to distinguish from
a hormetic mechanism (below).
Time On the other hand, recent correla-
Figure 4. The proposed dynamics of oxidative-damage values throughout tional evidence from the same research
nonreproductive and reproductive periods in the life of an animal. The team (Vitikainen et al. 2016) seems to
extortion-for-reproduction hypothesis (dotted line) predicts an increase in support the hypothesis. Wild female
the oxidative damage associated with the accumulation of macronutrients banded mongooses (Mungos mungo)
in blood and/or storage tissues. The oxidative-shielding hypothesis (dashed showed lower plasma levels of oxida-
line) predicts a preemptive decline in oxidative damage, which should imply a tive damage in lipids (malondialde-
decrease in advance of reproduction effort. This could be triggered by hormones hyde, MDA, levels) during pregnancy,
(e.g., estrogens; see main text) and could start abruptly. The duration could in contrast to other mammalian species
subsequently vary depending on reproductive strategies (e.g., mammals could (below). Moreover, contrary to previous
shield pups from the embryo stage until weaning, whereas birds or reptiles findings in other species (Blount et al.
should do so only until oviposition). Finally, the hormesis-during-reproduction 2016), Vitikainen and colleagues (2016)
hypothesis (solid line) implies a transient rise in oxidative damage after the did not find changes in protein carbonyls
start of reproductive effort. In the latter two cases, the damage level is lower but rather found a negative correlation
at the end of the reproductive event than at the prereproductive stage. This in pregnant females between the num-
is not predicted in extortion for reproduction, where damage at the end of ber of fetuses in the uterus and plasma
reproduction can remain higher than damage in the non-reproductive stage. MDA values. Furthermore, plasma MDA
of adults predicted survival in a 2- to
protein carbonyl levels in the heart, liver, and follicles. This, 3-year lapse. These results suggest that females may have
however, contrasted with higher levels of the same bio- shielded offspring by reducing their own damage levels,
marker in the blood of breeding animals (Yang etal. 2013, although these results were based on blood analyses and not
Xu etal. 2014). Reduced oxidative-damage values in breed- on reproductive tissues. Moreover, we cannot fully discard
ers compared with those of nonbreeders would be required the possibility that a reduction in lipid intake or fat reserves
for shielding offspring because embryos are more sensitive during pregnancy could influence circulating lipid levels
to disturbances than adults are. The hypothesis is appealing and, accordingly, plasma MDA values (see the section on the
as it considers fitness returns derived from offspring protec- extortion-for-reproduction hypothesis below).
tion (via reproductive success).
Some weaknesses in this hypothesis can, nonetheless, Hormesis during reproduction. Intermediate levels of exposure to
be argued. First, if the mechanism is preemptive, the lower a stressor may trigger a physiological compensatory response
oxidative-damage levels found in the organs of animals dur- that could, theoretically, not only maintain fitness returns
ing the last stages of reproduction (i.e., the main empirical at suitable levels but even increase fitness. This response is
support; Yang et al. 2013, Xu et al. 2014) should also have known as hormesis (e.g., Costantini 2014). If the breeding
been present just before or at the beginning of reproduction effort is considered a mild stressor (table 1, A2), lower oxida-
(figure 4). In other words, the low value detected at the end tive-damage levels among breeders may be due to hormesis
of reproduction was interpreted as evidence of a precedent (figure 3b). However, hormesis should only be triggered
decline. However, why should we expect oxidative-damage when animals are exposed to reproductive stress (Costantini

http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 7

Overview Articles
2014), whereas the preemptive response proposed in the
oxidative-shielding hypothesis should take place in advance
of stress (figure 4; Blount et al. 2016). This complicates the
decision of when sampling should be performed.
In trying to explain aging (i.e., not the cost of repro-
duction), the mitohormesis hypothesis (i.e., Tapia 2006)
proposes that a longer lifespan after caloric (mostly glu-
cose) restriction in the diet is due to a moderate increase
in mitochondrial ROS production. This, in turn, stimu-
lates a compensatory (hormetic) response improving long-
term oxidative stress resistance. This response would be
induced by certain redox signals (reviewed in Ristow and
Schemeisser 2014). In addition to reduced glucose metabo-
lism (Schulz etal. 2007), moderately high levels of oxidative
damage (Tapia 2006), including lipid peroxidation, could
also activate protection via mitohormesis (e.g., Ristow etal.
2009, Labuschagne et al. 2013). In that case, the involve-
ment of mitochondria may not be strictly necessary, because
other pro-oxidant sources (figure 2; see also Ristow and
Schemeisser 2014) could trigger protection.
In a reproductive context, Gago-Domnguez and col-
leagues (2005) suggested that reduced breast cancer preva-
lence among parous women is due to the protective (perhaps
hormetic) effect of high lipid peroxidation levels during
pregnancy. Indeed, a rise in the level of oxidative damage to
blood lipids in pregnant compared with that in nonpregnant
females has been demonstrated in humans (reviewed in Little
and Gladen 1999) and other mammals, including rodents
(e.g., Upreti etal. 2002, Castillo etal. 2005, but see Vitikainen
etal. 2016). The same pattern has been described for protein
carbonyls (e.g., Mihu et al. 2012). The increase seems to
be transient and limited to the gestation period, at least for
lipid peroxides (Gago-Domnguez etal. 2005). However, the
ephemeral nature of these changes does not imply positive
effects on fitness. In fact, as far as we know, no studies have
reported lower lipid peroxidation levels in mothers after birth
compared with levels in nonreproducing women. Moreover,
if hormesis indeed acts on reproducing females, their lifes-
pans or future fecundities should be increased compared
with those of nonbreeders (Schulz etal. 2007). Therefore, the
hormesis hypothesis also challenges the traditional evolu-
tionary theory considering reproduction as a cost.
Extortion for reproduction.The possibility that reproduc-
tion reduces oxidative damage is not as paradoxical if
we consider environmental conditions and life history.
Experiments preventing breeding are difficult to perform in
the field (e.g., it requires animal sterilization or the removal
of nest sites or mates), whereas captivity (laboratory) con-
ditions are usually designed for optimal reproduction. In
terms of life history, most laboratory studies have involved
animals without reproductive experience (Speakman and
Garratt 2014, Blount etal. 2016). We suggest that prevent-
ing reproduction under optimal conditions and for more or
less long periods can affect the animals under study. Under
these circumstances, the organism would be better suited to
8 BioScience XXXX XXXX / Vol. XX No. X http://bioscience.oxfordjournals.org
a reproductive situation. We believe that the focus should
therefore be placed on the control, nonreproducing animals.
Importantly, we assume that experimental controls showing
high oxidative damage did not suffer social stress due to iso-
lation or stocking density. These situations lead to changes in
glucocorticoid secretion that, in turn, could increase oxida-
tive stress (Costantini 2016).
We propose that animals prevented from breeding under
optimal conditions during prolonged periods would, in
some way, be forced to reproduce by high oxidative dam-
age (figure 3c). This mechanism may also have evolved
in the wild (table 1, A3). Wild animals can experience
high resource availability, allowing breeding, but may be
unable to mate if the population (operational) sex ratio is
biased. Macronutrients not diverted to reproduction can
become oxidized (Romero-Haro et al. 2015) or stimulate
NADPH oxidases (above), recreating a terminal situation
(see terminal investment in Clutton-Brock 1984). This
should not be limited to species assigned to the traditional
classification of capital breeder (i.e., species mostly relying
on energy stores for reproduction; Jonsson 1997), because
high circulating levels of macromolecules (also present in
income breeders) could be sufficient to trigger the mecha-
nism. High amounts of oxidized macromolecules or ROS
would alter redox-signaling pathways, promoting hormonal
and behavioral changes leading to reproduction with as
little delay as possible (figure 3c; table 1, A1, P7). Under
free-living conditions, this could be attained by dispersal
to other populations. In summary, animals prevented from
reproducing would be spurred on by the perception that life
is advancing too rapidly. This hypothesis is attractive, but is
there any supporting evidence?
Cancer epidemiology suggests higher levels of oxidative
damage in women without reproductive experience. We
must note that cancer development is strongly linked to high
oxidative stress levels and high activity of NADPH oxidases
(Bedard and Krause 2007). Nulliparity and late age at first
birth in women are associated with a higher prevalence of
breast cancer (e.g., Ma et al. 2006), which is also the most
frequent cancer in women worldwide (www.wcrf.org/int/
cancer-facts-figures/worldwide-data). The same association
was found in other reproductive and nonreproductive can-
cers (e.g., ovarian, brain, pancreas, colon, and liver; Martnez
etal. 1997, Merrill etal. 2005, Chang etal. 2010, Wu etal.
2011, Schler et al. 2013). Late parturition is, nonetheless,
related to a lower prevalence of uterine cancers, mostly cer-
vical cancer (Merrill etal. 2005). However, cervical cancer is
primarily due to human papillomavirus (Bosch etal. 2006).
We assume that women in most demographic studies on
cancer have a sufficiently high availability of energy macro-
nutrients to allow for reproduction. However, we recognize
that demographic studies in humans are basically correla-
tions in which socioeconomic factors and other covariates
cannot be well controlled. Therefore, causality should be
inferred with caution. Moreover, we could argue that cancer
is mostly detected at relatively older ages, with its impact

on individual fitness perhaps being minor. Nonetheless,
fitness returns in older women may also arise via parental
care devoted to adult offspring or grandchildren (Hawkes
2004). In any event, we note the lack of knowledge regarding
the impact of cancer on wild populations of other species
(Vittecoq etal. 2015).
Potential evidence may be drawn from results of an exper-
iment manipulating precopulatory reproductive investments
in zebra finches (Romero-Haro et al. 2015). We predicted
that male zebra finches caged in view of females should show
an increase in MDA levels in plasma as a result of investing
carotenoid pigments (antioxidants) in reproduction (sexual
coloration). We found the opposite, with males housed in
view of other males showing an increase in plasma MDA
values. Our result was, however, explained by an increase in
the level of circulating triglycerides. Plasma MDA and even
hydroperoxide levels, another oxidative-damage marker
(derivatives of reactive oxygen metabolities, d-ROMS; e.g.,
Costantini etal. 2014) are positively correlated to circulating
lipid (triglycerides) levels, at least in birds (Prez-Rodrguez
etal. 2015) and humans (e.g., Toescu etal. 2004). The same
correlation is present for isoprostanes, another popular lipid
oxidation marker, in human blood (Halliwell and Lee 2011).
This means that recent food intake, fatness, or lipid mobi-
lization from stores may potentially affect these measures.
Therefore, we urge caution in direct interpretations derived
from these parameters. Male zebra finches housed with other
males showed increased body mass and plasma triglyceride
levels, probably because of lower rates of social interactions,
and therefore less energy consumption (Romero-Haro etal.
2015). The results, therefore, suggest that sexually mature
animals with abundant energy macromolecules, circulating
in blood or being stored, may suffer oxidative damage when
no mate is available (table 1, A2).
In addition, recent studies testing somatic and ger-
mline signaling networks leading to longevity extension
and reduction, respectively, suggest that the first is, at least
partially, attained by reorganizing lipid stores (Hansen
et al. 2013). Germline ablation (mostly in invertebrates)
favors transcription factors promoting lipase activity and
autophagy, lipid unsaturation, and beta-oxidation (Hansen
etal. 2013, Lapierre etal. 2013). Accordingly, some authors
have suggested that these signals may eliminate the excess
of those fatty acids used for reproduction and convert them
into forms that favor a long life (Ratnappan etal. 2013). This
perspective agrees with the idea that high lipid accumulation
for reproduction leads to higher oxidative stress in prebreed-
ing animals.
Although the hypothesis may apparently reconcile empir-
ical findings with life-history theory, one potential problem
is that no study testing the oxidative cost of reproduction has
reported (or tested) reduced longevity among nonreproduc-
ing captive animals showing higher oxidative-damage levels
(table 1, P4). This would be expected if higher oxidative
damage among nonbreeders leads to reduced longevity. We
must nonetheless consider that, most studies have involved
http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 9
Overview Articles
young individuals that may be better able to face the chal-
lenge, at least under laboratory conditions (Speakman and
Garratt 2014, Blount et al. 2016). In fact, we predict that the
mechanism would be reinforced when age advances (see the
next section).
Testing the hypotheses. Experiments testing reproductive costs
are usually performed by randomly allocating animals to
reproduction or nonreproduction groups or by manipulat-
ing reproductive effort only among breeders (e.g., Alonso-
Alvarez and Velando 2012, Metcalfe and Monaghan 2013).
To test the three hypotheses using these approaches, an
accurate sampling schedule is required involving repeated
measures taking into account the predicted changes in oxi-
dative damage (table 1, P1 and P2; figure 4).
In the case of oxidative shielding, at least three sampling
events are required: two events before reproduction to docu-
ment a preemptive decline in damage (i.e., to discard horme-
sis) and another sampling after breeding to demonstrate that
the potential impact was not only avoided but also reversed
(i.e., less damage among breeders; figure 4). Only noninva-
sive techniques such as blood analyses, or perhaps semen
sampling, allow for repeated measures, precluding the study
of internal organs. The question is important if we predict
that oxidative shielding should be present in those tissues
that mostly influence offspring quality (i.e., the reproductive
system; table 1, P3). Alternatively, different animals could
be sacrificed at different phases, but this reduces the reli-
ability of results compared with that of longitudinal analy-
ses. In addition, experiments should include the analysis of
the individual quality of descendants. Parents generating
a stronger preemptive decline in oxidative damage should
produce higher-quality descendants (i.e., lower oxidative
damage and higher survival probabilities; table 1, P4).
In the case of hormesis during reproduction, tissues
should be sampled before and during reproduction to detect
a transient rise in oxidative stress (figure 4). A third meas-
ure should be carried out to detect the subsequent decline,
which should attain lower damage levels than those found in
control nonbreeding individuals. The positive impact of this
mechanism should be tested in terms of parental longevity
but may also affect current and future reproductive success
(i.e., fitness; table 1, P4).
Testing extortion for reproduction requires the detection
of increased levels of both macronutrients (lipids and/or
proteins) and oxidized compounds (e.g., MDA or protein
carbonyls) among fully sexually mature and healthy individ-
uals prevented from breeding. A decline in both parameters
when reproduction is allowed to occur must also be seen
(figure 4). However, such a decrease may not result in less or
equal oxidative damage in breeding individuals compared
with that of animals in a nonreproductive stage (figure 4). In
the latter two hypotheses, the type of tissue is not necessarily
related to the quality of the offspring (table 1, P4).
Experimental approaches can involve not only manipu-
lations of breeding effort but also of oxidative stress. This
Overview Articles
would allow the testing of predictions related to life history
traits. For instance, in the case of the hormesis-during-
reproduction hypothesis, we should predict that reproduc-
ing animals supplied with dietary antioxidants will mount
a weaker compensatory response and suffer a reduction in
fitness (shorter lives or reduced fecundity; table 1, P6). In
the extortion-for-reproduction hypothesis, we predict that
fully sexually mature nonreproducing animals with ad libi-
tum access to macronutrients and without previous breed-
ing experience should delay their first reproduction when
antioxidant supplements are provided (table 1, P6). The last
two predictions are apparently counterintuitive considering
our current knowledge on reproductive constraints (e.g.,
Dowling and Simmons 2009) and the oxidative stress theory
of aging (Sies and Cadenas 1985). Nonetheless, they can be
tested by correctly adjusting antioxidant doses. As alterna-
tives to dietary antioxidants, mitochondrial uncouplers
and NOX inhibitors developed to combat cancer and other
human diseases may be promising methods to reduce ROS
generation (e.g., Salin et al. 2012, Altenhfer et al. 2015).
The simultaneous use of these two methods could allow
for testing the relative contribution of mitochondrial and
extramitochondrial ROS sources to the oxidative cost of
reproduction.
To conclude, the three hypotheses are compatible. They
formulate predictions for different reproductive groups.
Hormesis and oxidative shielding should act among breed-
ers, whereas extortion for reproduction should affect
nonbreeders (table 1, P1). This, in turn, should magnify
differences between groups. Nonetheless, we also predict
that hormesis during reproduction could weaken over the
lifespan as we expect an aging-related decline in the capaci-
ties for hormesis (table 1, P5; Costantini 2014). The same
could apply to oxidative shielding, assuming an increased
accumulation of oxidative damage with age (i.e., Sies and
Cadenas 1985). However, extortion for reproduction should
mostly act or be reinforced at advanced ages as opportunities
for breeding decline (table 1, P5). Therefore, experiments on
comparatively young animals would be more likely to detect
hormesis, whereas manipulations in older individuals would
be more likely to support the extortion hypothesis. In sup-
port of the latter, studies have shown that male and female
rats prevented from reproducing throughout their lifetime
undergo higher levels of oxidative damage compared with
reproducing animals, but only at older ages (about 2 years
old; Alabarse etal. 2011, Da Silva etal. 2013).
Conclusions
In our view, the oxidative-cost-of-reproduction hypoth-
esis does not present insurmountable theoretical problems.
Efforts should be redirected to refine experimental designs
and address new hypotheses (figure 3). The fact that virtu-
ally all of the experiments reporting counterintuitive results
compare reproducing versus nonreproducing animals, and
not reproductive animals enduring different breeding effort
(but see Garratt et al. 2013), suggests that the first type of
10 BioScience XXXX XXXX / Vol. XX No. X http://bioscience.oxfordjournals.org
study must be improved. Importantly and not previously
highlighted, it cannot be discarded that some animals
prevented from breeding may be low-quality individuals
unable to reproduce. These individuals may endure greater
oxidative damage, buffering the statistical effect. Moreover,
some animals will be unable to breed quickly in the relatively
short time period of an experimental work. In fact, experi-
mental studies often mention that those animals allocated
to reproduction that did not breed during the study period
were excluded (e.g., Yang etal. 2013, Costantini etal. 2014),
but not all the works describe this detail. The excluded
animals may never breed (i.e., may be sick individuals) or
may require more time to reach a certain threshold in terms
of energy, micronutrients, or signaling (hormones) allow-
ing reproduction. They may also be prone to physiological
corticoid-based stress under captivity conditions. Although
these biases are probably subtle, they would act in opposing
directions (i.e., by including and excluding the worst animals
in nonbreeders and breeders, respectively) increasing the
likelihood of detecting null effects or even oxidative-damage
reduction among breeders. The only solution is to test and
control for differences between treatments in oxidative stress
levels before the experiment. This is, however, infeasible
when animals are sacrificed for analyzing internal tissues.
To demonstrate a cost in evolutionary terms, any manipu-
lation of reproductive investment should have a consequence
in terms of reduced fitness (e.g., Alonso-Alvarez and Velando
2012). This has rarely been tested for most studies testing the
OCR and is a difficult task, particularly in free-living animals
(but see Vitikainen et al. 2016). For instance, Santos and
Nakagawa (2012) recently performed a meta-analysis review-
ing evidence of a cost of brood or clutch size manipulations
in wild birds, detecting weak effects on survival (only among
male parents experiencing increased breeding effort). An
absence of effective evolutionary costs would make discus-
sions about the proximate mechanisms irrelevant.
The same argument can be applied to the choice of ana-
lytical techniques. Much effort has been placed on oxidative-
damage markers, because they are supposedly unambiguous.
However, the evidence of a connection between levels of
these parameters and longevity in longitudinal studies is still
scarce. Furthermore, the search for a link to fitness should
not be limited to longevity but should consider future
fecundity (e.g., Costantini et al. 2015). In fact, it has been
suggested that oxidative stress may act as a reproductive
constraint (reviewed in Stier etal. 2012; see also Costantini
etal. 2016). If oxidative stress constrains reproduction, oxi-
dative stress due to reproduction could also be costly by lim-
iting future reproduction. We suggest focusing on oxidative
damage in germline DNA, which mostly promotes deleteri-
ous mutations but may also lead to infertile adult offspring
(Velando etal. 2008). All parental investment in these defi-
cient descendants would be lost. Therefore, germline oxida-
tive damage might have a greater impact on parental fitness
than does oxidative damage in somatic cells. Unfortunately,
no work has yet addressed this question. Finally, the study of

the entire transcriptome (e.g., Rey etal. 2016) could reveal
unknown signaling molecules involved in the reproductive
trade-off, which could open the possibility of manipulations
allowing the testing of the proposed hypotheses.
Acknowledgments
We thank Thomas Flatt, Malene Hansen, and Hugo Aguilaniu
for fruitful discussions on the role of reorganization of lipid
stores and germline signals in longevity. Dr. Chen Hou also
contributed to clarifying the link the biosynthetic rate and
oxidative stress. We acknowledge Professor Neil Metcalfe
for comments on previous versions of the text. We are also
grateful to Mara Benfato, Pawel Brzek, and David Costantini
for providing details on their experimental designs and to
Sarah Young for her English review. We also acknowledge
comments from the five reviewers that helped to improve
the first versions of the text. Ana Angela Romero-Haro
was funded by a Formacin de Personal de Investigacin
grant (no. BES-2010-035013; Ministerio de Economa y
Competitividad, MINECO, Spanish Government) and pro-
ject nos. CGL2012-40229-C02-01 and CGL2015-69338-C2-
2-P (MINECO, Spain).
References cited
Alabarse PVG, Hackenhaar FS, Medeiros TM, Mendes MFA, Viacava PR,
Schller AK, Salomon TB, Ehrenbrink G, Benfato MS. 2011. Oxidative
stress in the brain of reproductive male rats during aging. Experimental
Gerontology 46: 241248.
Alberici LC, Oliveira HCF, Paim BA, Mantello CC, Augusto AC, Zecchin
KG, Gurgueira SA, Kowaltowski AJ, Vercesi AE. 2009. Mitochondrial
ATP-sensitive K+) channels as signals to liver mitochondria in response
to hypertriglyceridemia. Free Radical Biology Medicine 47: 14321439.
Alonso-Alvarez C, Velando A. 2012. Benefits and costs of parental care.
Pages 4054 in Klliker M, Smiseth PT, Royle NJ, eds. Evolution of
Parental Care. Oxford University Press.
Alonso-Alvarez C, Bertrand S, Devevey G, Prost J, Faivre B, Sorci G. 2004.
Increased susceptibility to oxidative stress as a proximate cost of repro-
duction. Ecology Letters 7: 363368.
Altenhfer S, Radermacher KA, Kleikers PWM, Wingler K, Schmidt
HHHW. 2015. Evolution of NADPH oxidase inhibitors: Selectivity and
mechanisms for target engagement. Antioxidants and Redox Signaling
23: 403427.
Amunugama K, Jiao L, Olbricht GR, Walker C, Huang Y-W, Nam PK, Hou
C. 2016. Cellular oxidative damage is more sensitive to biosynthetic rate
than to metabolic rate: A test of the theoretical model on hornworms
(Manduca sexta larvae). Experimental Gerontology 82: 7380.
Barja G. 2013. Updating the mitochondrial free radical theory of aging: An
integrated view, key aspects, and confounding concepts. Antioxidants
and Redox Signaling 19: 14201445.
Bedard K, Krause K-H. 2007. The NOX family of ROS-generating NADPH oxi-
dases: Physiology and pathophysiology. Physiological Reviews 87: 245313.
Blount JD, Vitikainen EIK, Stott I, Cant MA. 2016. Oxidative shielding and
the cost of reproduction. Biological Reviews 91: 483497.
Bosch FX, Qiao Y-L, Castellsagu X. 2006. The epidemiology of human
papillomavirus infection and its association with cervical cancer.
International Journal of Gynecology and Obstetrics 94: S8S21.
Briga M, Verhulst S. 2015. What can long-lived mutants tell us about mech-
anisms causing aging and lifespan variation in natural environments?
Experimental Gerontology 71: 2126.
Brown DI, Griendling KK. 2009. NOX proteins in signal transduction. Free
Radical Biology and Medicine 47: 12391253.
http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 11
Overview Articles
Castillo C, Hernandez J, Bravo A, Lopez-Alonso M, Pereira V, Benedito JL.
2005. Oxidative status during late pregnancy and early lactation in dairy
cows. Veterinary Journal 169: 286292.
Chang CC, Chiu HF, Yang CY. 2010. Parity, age at first birth, and risk
of death from pancreatic cancer: Evidence from a cohort in Taiwan.
Pancreas 39: 567571.
Clutton-Brock TH. 1984. Reproductive effort and terminal investment in
iteropaous animals. American Naturalist 123: 212229.
Costantini D. 2014. Oxidative Stress and Hormesis in Evolutionary Ecology
and Physiology: A Marriage between Mechanistic and Evolutionary
Approaches. Springer.
. 2016. Oxidative stress as a cost of reproduction: Beyond the sim-
plistic trade-off model. Frontiers in Ecology and Evolution 4 (art. 10).
Costantini D, Casasole G, Eens M. 2014. Does reproduction protect against
oxidative stress? Journal Experimental Biology 217: 42374243.
Costantini D, Goutte A, Barbraud C, Faivre B, Sorci G, Weimerskirch H,
Delord K, Chastel O. 2015. Demographic responses to oxidative stress
and inflammation in the wandering albatross (Diomedea exulans).
PLOS ONE 10 (art. e0133967).
Costantini D, Casasole G, AbdElgawad H, Asard H, Eens M. 2016.
Experimental evidence that oxidative stress influences reproductive deci-
sions. Functional Ecology 30: 11691174. doi:10.1111/1365-2435.12608
Covarrubias L, Hernndez-Garca D, Schnabel D, Salas-Vidal E, Castro-
Obregn S. 2008. Function of reactive oxygen species during animal
development: Passive or active? Developmental Biology 320: 111.
Da Silva, AC, Salomon TB, Behling CS, Putti J, Hackenhaar FS, Alabarse
PVG, Schller AK, Benfato MS. 2013. Oxidative stress in the kid-
ney of reproductive female rats during aging. Biogerontology 14:
411422.
De Jong G, van Noordwijk AJ. 1992. Acquisition and allocation of resources:
Genetic (co) variances, selection, and life histories. American Naturalist
139: 749770.
Drummond GR, Sobey CG. 2014. Endothelial NADPH oxidases: Which
NOX to target in vascular disease? Trends in Endocrinology and
Metabolism 25: 452463.
Dymkowska D, Drabarek B, Podszywaow-Bartnicka P, Szczepanowska J,
Zabocki K. 2014. Hyperglycaemia modifies energy metabolism and
reactive oxygen species formation in endothelial cells in vitro. Archives
in Biochemistry and Biophysiology 542: 713.
Edward DA, Chapman T. 2011. The evolution and significance of male mate
choice. Trends in Ecology and Evolution 26: 647654.
Flatt T. 2011. Survival costs of reproduction in Drosophila. Experimental
Gerontology 46: 369375.
Flatt T, Amdam GV, Kirkwood TBL, Omholt SW. 2013. Life-history evolu-
tion and the polyphenic regulation of somatic maintenance and sur-
vival. Quarterly Review of Biology 88: 185218.
Forman HJ, Fukuto JM, Torres M. 2004. Redox signaling: Thiol chemistry
defines which reactive oxygen and nitrogen species can act as second
messengers. American Journal of Physiology: Cell Physiology 287:
C246C256.
Gago-Dominguez M, Castelao JE, Pike MC, Sevanian A, Haile RW. 2005.
Role of lipid peroxidation in the epidemiology and prevention of
breast cancer. Cancer Epidemiology, Biomarkers and Prevention 14:
28292839.
Garratt M, McArdle F, Stockley P, Vasilaki A, Beynon RJ, Jackson MJ, Hurst
JL. 2012. Tissue-dependent changes in oxidative damage with male
reproductive effort in house mice. Functional Ecology 26: 423433.
Garratt M, Pichaud N, King EDA, Brooks RC. 2013. Physiological adapta-
tions to reproduction. I. Experimentally increasing litter size enhances
aspects of antioxidant defence but does not cause oxidative damage in
mice. Journal Experimental Biology 216: 28792888.
Go YM, Jones DP. 2011. Cysteine/cystine redox signaling in cardiovascular
disease. Free Radical Biology and Medicine 50: 495509.
Halliwell B, Gutteridge JMC. 2007. Free Radicals in Biology and Medicine.
Oxford University Press.
Overview Articles
Halliwell B, Lee CYJ. 2010. Using isoprostanes as biomarkers of oxidative
stress: Some rarely considered issues. Antioxidants and Redox Signaling
13: 145156.
Hansen M, Flatt T, Aguilaniu H. 2013. Reproduction, fat metabolism, and
life span: What is the connection? Cell Metabolism 17: 1019.
Hawkes K. 2004. The grandmother effect. Nature 428: 128129.
Hou C, Amunugama K. 2015. On the complex relationship between energy
expenditure and longevity: Reconciling the contradictory empirical
results with a simple theoretical model. Mechanisms of Ageing and
Development 149: 5064.
Hsin H, Kenyon C. 1999. Signals from the reproductive system regulate the
lifespan of C. elegans. Nature 399: 362366.
Isaksson C, Sheldon BC, Uller T. 2011. Oxidative stress into life-history
biology. BioScience 61: 194202.
Jones D. 2006. Redefining oxidative stress. Antioxidants and Redox
Signaling 8: 18651879.
Jonsson KI. 1997. Capital and income breeding as alternative tactics of
resource use in reproduction. Oikos 78: 5766.
Kirkwood TBL. 1977. Evolution of ageing. Nature 270: 301304.
Krause K-H. 2007. Aging: A revisited theory based on free radicals gener-
ated by NOX family NADPH oxidases. Experimental Gerontology 42:
256262.
Labuschagne CF, Stigter EC, Hendriks MMWB, Berger R, Rokach J,
Korswagen HC, Brenkman AB. 2013. Quantification of in vivo oxida-
tive damage in Caenorhabditis elegans during aging by endogenous
F3-isoprostane measurement. Aging Cell 12: 214223.
Lapierre LR, Silvestrini MJ, Nuez L, Ames K, Wong S, Le TT, Hansen M,
Melendez A. 2013. Autophagy genes are required for normal lipid levels
in C. elegans. Autophagy 9: 278286.
Little RE, Gladen BC. 1999. Levels of lipid peroxides in uncomplicated
pregnancy: A review of the literature. Reproductive Toxicology 13:
347352.
Ma H, Bernstein L, Pike MC, Ursin G. 2006. Reproductive factors and
breast cancer risk according to joint estrogen and progesterone recep-
tor status: A meta-analysis of epidemiological studies. Breast Cancer
Research 8 (art. R43).
Martnez ME, Grodstein F, Giovannucci E, Colditz GA, Speizer FE,
Hennekens C, Rosner B, Willett WC, Stampfer MJ. 1997. A prospective
study of reproductive factors, oral contraceptive use, and risk of colo-
rectal cancer. Cancer Epidemiology, Biomarkers and Prevention 6: 15.
Merrill RM, Fugal S, Novilla LB, Raphael MC. 2005. Cancer risk associated
with early and late maternal age at first birth. Gynecologic Oncology
96: 583593.
Metcalfe NB, Monaghan P. 2013. Does reproduction cause oxidative stress?
An open question. Trends in Ecology and Evolution 28: 347350.
Mihu D, Sabu L, Costin N, Ciortea R, Mluan A, Mihu CM. 2012.
Implications of maternal systemic oxidative stress in normal pregnancy
and in pregnancy complicated by preeclampsia. Journal Maternal and
Neonatal Medicine 25: 944951.
Mller AP, Biard C, Blount JD, Houston DC, Ninni P, Saino N, Surai
PF. 2000. Carotenoid-dependent signals: Indicators of foraging effi-
ciency, immunocompetence or detoxification ability? Avian and Poultry
Biology Reviews 11: 137159.
Olsson M, Wilson, M, Uller T, Mott B, Isaksson, C. 2009. Variation in levels
of reactive oxygen species is explained by maternal identity, sex and
body-size-corrected clutch size in a lizard. Naturwissenschaften 96:
2529.
Prez-Rodrguez L, Romero-Haro AA, Sternalski A, Muriel J, Mougeot F,
Gil D, Alonso-Alvarez C. 2015. Measuring oxidative stress: The con-
founding effect of lipid concentration in measures of lipid peroxidation.
Physiological and Biochemical Zoology 88: 345351.
Ratnappan R, Amrit FRG, Chen S-W, Gill H, Holden K, Ward J, Yamamoto
KR, Olsen CP, Ghazi A. 2014. Germline signals deploy NHR-49 to
modulate fatty-acid -oxidation and desaturation in somatic tissues of
C. elegans. PLOS Genetics 10 (art. e1004829).
Ren Y, Hughes KA. 2014. Vitellogenin family gene expression does not
increase Drosophila lifespan or fecundity. F1000Research 3 (art. 125).
12 BioScience XXXX XXXX / Vol. XX No. X http://bioscience.oxfordjournals.org
Rey B, Dgletagne C, Bodennec J, Monternier P-A, Mortz M, Roussel D,
Romestaing C, Rouanet J-L, Tornos J, Duchamp C. 2016. Hormetic
response triggers multifaceted anti-oxidant strategies in immature king
penguins (Aptenodytes patagonicus). Free Radical Biology and Medicine
97: 577587.
Ristow M, Schmeisser K. 2014. Mitohormesis: Promoting health and
lifespan by increased levels of reactive oxygen species (ROS). Dose
Response 12: 288341.
Ristow M, Zarse K, Oberbach A, Nora K, Birringer M, Kiehntopf M,
Stumvoll M, Kahn CR, Blher M. 2009. Antioxidants prevent health-
promoting effects of physical exercise in humans. Proceedings of the
National Academy of Sciences 106: 86658670.
Romero-Haro AA, Canelo T, Alonso-Alvarez C. 2015. Early development
conditions and the oxidative cost of social context in adulthood: An
experimental study in birds. Frontiers in Ecology and Evolution 3: 115.
Sakellariou GK, Jackson MJ, Vasilaki A. 2014. Redefining the major con-
tributors to superoxide production in contracting skeletal muscle: The
role of NAD(P)H oxidases. Free Radical Research 48: 1229.
Salin K, Luquet E, Rey B, Roussel D, Voituron Y. 2012. Alteration of mito-
chondrial efficiency affects oxidative balance, development and growth
in frog (Rana temporaria) tadpoles. Journal of Experimental Biology
215: 863869.
Salin K, Auer SK, Rudolf AM, Anderson GJ, Cairns AG, Mullen W, Hartley
RC, Selman C, Metcalfe MB. 2015. Individuals with higher metabolic
rates have lower levels of reactive oxygen species in vivo. Biology Letters
11 (art. 20150538).
Salmon AB, Marx DB, Harshman LG. 2001. A cost of reproduction in
Drosophila melanogaster: Stress susceptibility. Evolution 55: 16001608.
Salomon TB, Hackenhaar FS, Almeida AC, Schller AK, Gil Alabarse PV,
Ehrenbrink G, Benfato MS. 2013. Oxidative stress in testis of animals
during aging with and without reproductive activity. Experimental
Gerontology 48: 940946.
Santos ESA, Nakagawa S. 2012. The costs of parental care: A meta-analysis
of the trade-off between parental effort and survival in birds. Journal of
Evolutionary Biology 25: 19111917.
Schaar CE, Dues DJ, Spielbauer KK, Machiela E, Cooper JF, Senchuk M,
Hekimi S, Van Raamsdonk JM. 2015. Mitochondrial and cytoplas-
mic ROS have opposing effects on lifespan. PLOS Genetics 11 (art.
e1004972).
Schler S, Ponnath M, Engel J, Ortmann O. 2013. Ovarian epithelial tumors
and reproductive factors: A systematic review. Archives of Gynecology
and Obstetrics 287: 11871204.
Schulz TJ, Zarse K, Voigt A, Urban N, Birringer M, Ristow M. 2007. Glucose
restriction extends Caenorhabditis elegans lifespan by inducing mito-
chondrial respiration and increasing oxidative stress. Cell Metabolism
6: 280293.
Seah NE, de Magalhaes Filho CD, Petrashen AP, Henderson HR, Laguer
J, Gonzalez J, Dillin A, Hansen M, Lapierre LR. 2016. Autophagy-
mediated longevity is modulated by lipoprotein biogenesis. Autophagy
12: 261272.
Seehuus S-C, Norberg K, Gimsa U, Krekling T, Amdam GV. 2006.
Reproductive protein protects functionally sterile honey bee workers
from oxidative stress. Proceedings of the National Academy of Sciences
103: 962967.
Serpillon S, Floyd BC, Gupte RS, George S, Kozicky M, Neito V, Recchia
F, Stanley W, Wolin MS, Gupte SA. 2009. Superoxide production
by NAD(P)H oxidase and mitochondria is increased in genetically
obese and hyperglycemic rat heart and aorta before the development
of cardiac dysfunction: The role of glucose-6-phosphate dehydroge-
nase-derived NADPH. American Journal of Physiology: Heart and
Circulatory Physiology 297: H153H162.
Sies H, Cadenas E. 1985. Oxidative stress: Damage to intact cells and organs.
Philosophical Transactions of the Royal Society B 311: 617631.
Speakman JR. 2008. The physiological costs of reproduction in small mam-
mals. Philosophical Transactions of the Royal Society B 363: 375398.
Speakman JR, Garratt M. 2014. Oxidative stress as a cost of reproduction:
Beyond the simplistic trade-off model. BioEssays 36: 93106.

Stier A, Reichert S, Massemin S, Bize P, Criscuolo F. 2012. Constraint
and cost of oxidative stress on reproduction: Correlative evidence in
laboratory mice and review of the literature. Frontiers in Zoology 9
(art. 37).
Tapia PC. 2006. Sublethal mitochondrial stress with an attendant stoi-
chiometric augmentation of reactive oxygen species may precipitate
many of the beneficial alterations in cellular physiology produced by
caloric restriction, intermittent fasting, exercise and dietary phytonu-
trients: Mitohormesis for health and vitality. Medical Hypotheses 66:
832843.
Tatar M. 2001. Senescence. Pages 128141 in Fox CW, Roff DA, Fairbaim
DJ. Evolutionary Ecology: Concepts and Case Studies. Oxford University
Press.
Tatar M, Carey JR. 1995. Nutrition mediates reproductive trade-offs with
age-specific mortality in the beetle Callosobruchus maculatus. Ecology
76: 20662073.
Toescu V, Nuttall SL, Martin U, Nightingale P, Kendall MJ, Brydon P, Dunne
F. 2004. Changes in plasma lipids and markers of oxidative stress in
normal pregnancy and pregnancies complicated by diabetes. Clinical
Science 106: 9398.
Upreti K, Chaki SP, Misro MM. 2002. Evaluation of peroxidative stress and
enzymatic antioxidant activity in liver and kidney during pregnancy
and lactation in rats. Health and Population: Perspectives and Issues
25: 177185.
Velando A, Torres R, Alonso-Alvarez C. 2008. Avoiding bad genes:
Oxidatively damaged DNA in germ line and mate choice. Bioessays 30:
12121219.
Vittecoq M, etal. 2015. Animal behaviour and cancer. Animal Behaviour
101: 1926.
http://bioscience.oxfordjournals.org XXXX XXXX / Vol. XX No. X BioScience 13
Overview Articles
Wang Y, Salmon AB, Harshman LG. 2001. A cost of reproduction: Oxidative
stress susceptibility is associated with increased egg production in
Drosophila melanogaster. Experimental Gerontology 36: 13491359.
Wiersma P, Selman C, Speakman JR, Verhulst S. 2004. Birds sacrifice oxida-
tive protection for reproduction. Biology Letters 271: S360S363.
William GC. 1957. Pleiotropy, natural selection, and the evolution of senes-
cence. Evolution 11: 398411.
Wu C-H, Chan T-F, Changchien CC, Yang CY. 2011. Parity, age at first birth,
and risk of death from liver cancer: Evidence from a cohort in Taiwan.
Journal of Gastroenterology and Hepatology 26: 334339.
Xu YC, Yang DB, Speakman JR, Wang DH. 2014. Oxidative stress in
response to natural and experimentally elevated reproductive effort is
tissue dependent. Functional Ecology 28: 402410.
Yang D-B, Xu Y-C, Wang D-H, Speakman JR. 2013. Effects of reproduction
on immuno-suppression and oxidative damage, and hence support or
otherwise for their roles as mechanisms underpinning life history trade-
offs, are tissue and assay dependent. Journal of Experimental Biology
216: 42424250.
Zera AJ, Harshman LG. 2001. The physiology of life history trade-offs in
animals. Annual Reviews in Ecology and Systematics 32: 95126.
Carlos Alonso-Alvarez (carlos.alonso@csic.es) is affiliated with the
Departamento de Ecologa Evolutiva at the Museo Nacional de Ciencias
Naturales (National Museum of Natural History) which is part of the Consejo
Superior de Investigaciones Cientficas (CSIC; the Spanish National Research
Council), in Madrid, Spain. Tara Canelo and Ana ngela Romero-Haro are
affiliated with the Instituto de Investigacin en Recursos Cinegticos (IREC;
Institute for Game and Wildlife Research), in Ciudad Real, Spain, which is
also part of CSIC and University of Castilla La Mancha (UCLM).