Environmental conditions across a range of habitat types are associated with OS threats. For example, marine environments vary widely in thermal characteristics, partial pressure of oxygen, degree of pollution and light environment, all of which can influence the degree of oxidative insult experienced by an individual, and many marine organisms adjust their AO systems in response to this environmental variability to maintain a stable redox state (reviewed in Lesser 2006; Buttemer, Adele & Costantini 2010). Furthermore, within species, population densities can vary with habitat and, at least in freshwater clams (Sphaerium sp), some of this variation has been tied to OD; clams were most abundant and showed lower levels of nucleic acid damage in relatively more hypoxic environments (Joyner-Matos et al. 2007).
Further environmentally driven variation in an organism’s oxidative exposure and AO system can be a product of season (snails, Helix aspersa; Ramos-Vasconcelos, Cardoso & Hermes-Lima 2005), thermotolerance (grasses; Banowetz et al. 2007), osmotic stress (reeds; Zhao et al. 2008), atrophy (burrowing frogs, Cyclorana alboguttata; Hudson et al. 2006), nutrient starvation (Mycobacterium avium; Gumber et al. 2009) and seed dormancy (Oracz et al. 2009), sometimes with RS playing an essential signalling role (Bailly, El-Maarouf-Bouteau & Corbineau 2008). These few examples demonstrate the wide variety of ecological contexts in which RS affect organismal behaviour, physiology, distribution and fitness, and highlight the breadth of research areas that remain to be explored.
Organismal self-maintenance and survival behaviours
In sports medicine and laboratory studies of exercise, the role of locomotor activity in promoting RS production (Davies et al. 1982), and of regular exercise in nullifying these effects (reviewed in Powers et al. 2004), has received considerable attention (Alessio 1993; Di Meo & Venditti 2001; Gomez-Cabrera, Domenech & Viña 2008). Although physical performance has been linked to aspects of organismal fitness, specifically as a proxy for escape from predators (e.g. flight performance; Veasey, Metcalfe & Houston 1998), the effect of RS and role of AO in maintaining the redox balance during activity remains unclear. Nonetheless, a diversity of ecologically relevant phenomena conducive to this type of research exists, and clearly warrants further investigation. For example, the expression of a number of physically demanding behaviours (e.g. migration, hunting, predator escape, sexual displays) may be evolutionary shaped by OS.
Changes in tissue oxygen levels can directly affect rates of OD, a phenomenon most commonly studied in rodents, rabbits, dogs and pigs during ischaemia-reperfusion as a model for treating injuries in humans (e.g. Aksoy et al. 2004; Hirayama et al. 2005; Ovechkin et al. 2007). However, similar non-pathological changes (e.g. ischaemia) can be found in natural contexts, such as reanimation following estivation, torpor, or hibernation, recovery from extracellular freezing, or temporally variable anoxia conditions commonly found in both invertebrates and vertebrates (Storey 1996; Willmore & Storey 1997; Bickler & Buck 2007; Issartel et al. 2009; Orr et al. 2009). Many of these organisms elevate AO levels during the oxygen-limited state to limit OD upon return to normoxia (Storey 1996; Hermes-Lima, Storey & Storey 1998).
There are also a variety of survival behaviours that are related to OS. For example, degree of torpor decreases with an increased polyunsaturated fatty acid (PUFA) content in the food hoard of chipmunks (Tamias striatus; Munro, Thomas & Humphries 2005), possibly due to the susceptibility of PUFAs to oxidation. Perceived predation risk also seems to have a direct influence on OS, as predator presence can increase oxygen consumption and decrease levels of AO enzymes (Slos & Stoks 2008). Finally, Bize et al. (2008) found that male alpine swifts (Apus melba) with red blood cells that were more resistant to an in vitro free radical attack tended to have greater inter-annual survival rates, providing one of the first clear links between oxidative balance and survival in a wild animal.
Another component of survival is immunocompetence. The study of ecoimmunology has proliferated in recent years, and a large number of studies have examined relationships between AO (particularly carotenoids) and immunity (e.g. Blount et al. 2003; Costantini & Dell’Omo 2006; Hõrak et al. 2007). As this field has been reviewed elsewhere in an evolutionary context (Dowling & Simmons 2009), we do not explore the relationship between RS and immune function in depth here. Briefly, however, during an innate immune response, phagocytic cells use an oxidative burst (RS production) to destroy pathogens. While these molecules effectively neutralize bacterial invaders, they are non-specific and may also destroy self-tissue. Therefore, individuals need to be able to negate the potential immunopathological effects of an immune response to maintain the fitness-related benefit of the immune response (Costantini & Møller 2009; Sorci & Faivre 2009).
Senescence is a process in which organisms undergo physiological degeneration with age (Rose 1991). OS has been linked to evolutionary theories of senescence since Harman (1956) first proposed the free radical theory of ageing, which suggests that cumulative OD on cell constituents caused by free radicals produced during aerobic respiration results in ageing and ultimately death. Since this time, the free radical theory has been refined and a number of hypothetical iterations have been presented, including the mitochondrial theory of ageing (Harman 1972) and the OS hypothesis of ageing (Yu & Yang 1996; Sohal, Mockett & Orr 2002). In 2003, Hulbert further developed the hypothesis and proposed the membrane pacemaker theory of ageing, suggesting that it is variation in membrane fatty acid composition, and the consequent differential vulnerability to lipid peroxidation, which determines lifespan: cells richer in PUFA should be more susceptible to OD and will therefore be more prone to cell senescence (Hulbert 2003, 2005; Hulbert et al. 2007). There is considerable support for the free radical theory of ageing in a wide range of taxa. However, much of this evidence is correlative and based on a limited number of laboratory taxa; thus definitive proof that alterations in OS play a role in longevity remains elusive (Wickens 2001; Sohal, Mockett & Orr 2002). Several recent, comprehensive reviews discuss the evidence for the broad free radical theory of ageing in considerable detail (see Beckman & Ames 1998; Ashok & Ali 1999; Wickens 2001; Sohal, Mockett & Orr 2002), and it is not our intention to review such evidence here; instead we present a brief overview of the relationship between lifespan and metabolic rate and discuss the role of OS in this relationship.
Given that RS production is partly a function of metabolic rate (i.e. a higher metabolic rate should lead to greater RS production), the free radical theory of ageing is often thought of synonymously with the rate of living theory: species with higher basal metabolic rates (BMR) are suggested to have shorter maximum lifespan potential (MLSP) as a result of increased RS production (Beckman & Ames 1998; Dowling & Simmons 2009). However, this integrated theory is somewhat erroneous. While there is some empirical support for a relationship between MLSP and BMR (e.g. birds, Cohen et al. 2008; humans, Ruggiero et al. 2008), many studies show no relationship between mortality and BMR (e.g. microchiropteran bats, Filho et al. 2007; colubrid snakes, Robert, Brunet-Rossinni & Bronikowski 2007; zebra finches Taeniopygia guttata, Moe et al. 2009). In addition, several statistical and methodological issues appear to compromise the validity of much of the data supporting this hypothesis; though, at least in one study of small mammals, when these issues are eliminated and a more robust measure of daily energy expenditure is employed (e.g. elimination rates of stable isotopes as opposed to BMR), high energy expenditure remains associated with shorter lifespan (Speakman et al. 2002). Finally, BMR does not appear to be a good predictor of lifespan within a species. For example, experimental manipulation of energy expenditure via cold exposure does not shorten lifespan in field voles (Microtus agrestis; Selman et al. 2008). Moreover, in mice, lifespan is not affected by an increase in energy expenditure induced by either cold exposure (Vaanholt et al. 2009) or aerobic exercise (Vaanholt et al. 2010).
One reason for these mixed results appears to be an uncoupling between metabolism and OS. Specifically, the transition of mitochondria from state 4 (resting) to state 3 (respiratory active, producing ATP) is not accompanied by a proportionate increase of free radical production (Loschen, Flohé & Chance 1971). Consequently, OD cannot be massive during phases of increased metabolic rate because the mitochondrial free radical leak strongly decreases over the states 4 to 3 transition (Herrero & Barja 1997). This uncoupling between metabolism and OS appears to be influenced by uncoupling proteins (Criscuolo et al. 2005) and likely represents an important biochemical adaptation to OS. Importantly, this uncoupling can result in a positive relationship between metabolic rate and lifespan (‘uncoupling to survive hypothesis’Brand 2000; Speakman et al. 2002). However, we suggest that more studies on the link between metabolic rate and RS production/OD (especially among wild animals and understudied taxonomic groups) are necessary to understand if this uncoupling is a general rule or occurs only under certain circumstances. Finally, we recommend further empirical studies of the link between BMR (or daily energy expenditure) and lifespan, but suggest that such studies should incorporate a wider range of taxa and might be best addressed in long-lived organisms or organisms that withstand environmental extremes (e.g. bristlecone pine, Pinus longaeva, MLSP >4500 years, Lanner & Connor 2001; periodical cicadas, Magicicada; Iceland clam, Arctica islandica, MLSP >400 years, Buttemer, Adele & Costantini 2010; waved albatross, Diomedea irrorata; estivating snails or anoxia-tolerant turtles, see Storey 1996).
OS may have its most dramatic effect on animal fitness by directly impairing reproduction (Harshman & Zera 2007). Age-specific accumulation of OS has been shown to damage oocytes and embryos of mammal mothers later in life (Tarín 1996), but in many taxa, OS can also have age-independent reproductive detriments. For example, mating increases susceptibility to OS in virgin female Drosophila melanogaster (Rush et al. 2007). In domesticated zebra finches, two reproductive phenomena – a greater number of breeding bouts per lifetime (Alonso-Alvarez et al. 2006) and raising more offspring per breeding bout (Alonso-Alvarez et al. 2004; Wiersma et al. 2004) – reduce AO defences. Some nutrients like carotenoids, however, seem to be able to modulate reproduction and trade-offs with other life-history aspects like OS resistance (e.g. birds, Biard, Surai & Møller 2005; Bertrand et al. 2006; mollusks, Petes, Menge & Harris 2008). Maintaining a positive seasonal carotenoid balance also predicts reproductive success in wild barn swallows (Hirundo rustica, Safran et al. 2010). In contrast to emphases on AO, only a handful of studies have investigated OS and female reproduction in wild animals (see de Almedia et al. 2007 for a review in marine bivalves). In a wild, long-lived avian species (alpine swifts), number of eggs laid per attempt (clutch size) was positively correlated with resistance of red blood cells to an in vitro free radical attack (Bize et al. 2008). However, in a wild, long-lived mammal (soay sheep, Ovis aries), lipid peroxidation was unrelated to recent and past reproductive effort (Nussey et al. 2009).
Some aspects of parental care (e.g. nestling food provisioning in birds, Helfenstein et al. 2008; egg fanning in fish, Pike et al. 2007) may also be linked to parental AO status. Given the high growth and metabolic rates in developing young, it is not surprising that OS strongly influences offspring; for example, Costantini et al. (2006) found that nestling Eurasian kestrels (Falco tinnunculus) from larger broods experience higher level of OS. Pre-natal OS can even induce neural damage and cognitive function in young rats (Song et al. 2009). There is also significant interest in how OS impacts fertility in human females (Agarwal, Gupta & Sharma 2005). OS can affect early (e.g. oocyte maturation) and late (e.g. pre-eclampsia, pre-term labour) stages of the female reproductive cycle, and various dietary interventions (e.g. vitamin supplementation) have been implemented before and during pregnancy. In most cases, however, either pre-embryonic or post-hatch/birth OS is studied in isolation; we encourage more comprehensive studies across life-history stages to better understand how OS is transmitted from mother to offspring and which sources are more significant for long-term offspring survival and reproduction. Additionally, investigation of the role of OS in transgenerational effects that are not the result of genetic origin or developmental environment may help us understand if these effects ‘program’ offspring for the future environment or if they simply function as a constraint on phenotypic plasticity (Jablonka & Raz 2009). Finally, we recommend expanding tests of these AO/OS ideas to new taxa, including organisms that are semelparous, very long-lived (and thus infrequently breeding), and that live under extreme environmental conditions (e.g. crowded, inbred, thermal extremes), to permit robust examination of how various reproductive stages and tactics are influenced by oxidative balance.
Sperm cells appear to be particularly vulnerable to attack by RS due to their high PUFA content and metabolic activity (Sikka 2001; Surai et al. 2001). In humans and a limited number of domesticated animals, RS attack has been shown to induce a lipid peroxidation chain event that can lead to decreased motility and viability, and an inability to fuse with the oocyte (Fujihara & Howarth 1978; Wishart 1984; Aitken, Clarkson & Fishel 1989; deLamirande & Gagnon 1992; Baumber et al. 2000; Bilodeau et al. 2002). Importantly, RS may also damage nuclear and mitochondrial DNA of sperm, with negative consequences for fertilizing capacity and post-fertilization embryo survival (Aitken & Krausz 2001). Sperm are also subject to RS attack in the testes, resulting in impaired steroidogenesis, a reduced capacity to differentiate normal spermatozoa, and, ultimately, a reduction in sperm fertilizing ability (Wu et al. 1973; Aitken & Roman 2008). Thus, OS can have a significant impact on male fitness, and may influence a number of evolutionary processes including the evolution of polyandry and mating strategies, sexual conflict, and sperm competition dynamics (Siva-Jothy 2000; Dean, Bonsall & Pizzari 2007; Pizzari et al. 2008; Dowling & Simmons 2009).
Both sperm and seminal plasma possess AO systems that protect sperm from RS-induced OS. Specifically, SOD, glutathione peroxidase, and catalase appear to be the major semen AO in a range of taxa, with additional protection provided by vitamins C and E, ubiquinols, and glutathione (Sikka 2001; Surai et al. 2001; Weirich, Collins & Williams 2002). Dietary carotenoid supplementation is associated with improved sperm quality (presumably via prevention of RS-induced damage) in humans (Gupta & Kumar 2002) and fishes (Ahmadi et al. 2006; Pike et al. 2010). Furthermore, carotenoids are present in the ejaculates of birds (Rowe & McGraw 2008) and insects (Heller, Fleischmann & Lutz-Röder 2000), suggesting the AO role of carotenoids in semen deserves further research. Future investigations should also focus on the relative contribution of each AO type to total AO capacity of semen, and examine how these metrics covary with sperm performance.
The impact of OS may be particularly important in species with high levels of sperm competition. In these species, males with a greater capacity to defend sperm against OS should gain a greater share of paternity. Consequently, sexual selection may target cellular mechanisms underlying sperm function aimed at avoiding OS. In a recent comparative study of Mus species, the level of sperm competition was associated with rapid changes in protamine 2 genes, which, in turn, appear to be associated with the efficiency of DNA compaction (Martin-Coello et al. 2009). In sperm nuclei, protamine-induced DNA compaction protects DNA from double-strand breaks by reducing the ability of RS to access DNA (Suzuki et al. 2009), suggesting that the rapid changes observed in Mus species may reflect selection for increased resistance to RS-induced sperm DNA damage. Similarly, protective mechanisms aimed at preventing OD to sperm structures may have evolved in response to sperm competition, though these ideas remain largely unexplored.
AO defences aimed at minimizing OD to sperm should also be particularly important for species with prolonged sperm storage. Stored sperm (i.e. aged sperm) show reduced fertilizing success (Tarín, Pérez-Albalá & Cano 2000; Wagner, Helfenstein & Danchin 2004), which can impact female reproductive success, generate sexual conflict and influence behaviour (e.g. sperm ejection, re-mating strategies; Siva-Jothy 2000; Dean, Bonsall & Pizzari 2007). To date, studies of AO in the sperm storage organs of females are limited to a few species of agricultural importance (i.e. poultry, Brèque, Surai & Brillard 2003; bees, Collins, Williams & Evans 2004). However, a wide range of taxa show extended periods of sperm storage (e.g. bats, up to 198 days; reptiles, up to 2555 days; Birkhead & Møller 1993), and future studies utilizing these species should provide substantial insight into the biological significance of OS-related sperm deterioration in wild organisms. Perhaps the most exciting potential study systems can be found among the insects: in some species, females maintain viable populations of stored sperm for periods as long as thirty years (e.g. narrow-headed ant Formica exsecta, Pamilo 1991). Male extragonadal sperm stores may also experience RS-induced deterioration with important consequences for fitness (Tarín, Pérez-Albalá & Cano 2000) and male behaviour (e.g. sperm wastage; Pizzari et al. 2008). Finally, OS and AO may modulate the relationship between male sexual ornamentation and sperm quality and allow females to choose reproductive partners with high functional fertility (Blount, Møller & Houston 2001; Velando, Torres & Alonso-Alvarez 2008); for example more colourful males appear better able to protect their sperm from OS in both birds (Helfenstein et al. 2010) and fishes (Pike et al. 2010).
The production of RS in semen may also provide a range of beneficial effects for sperm function, including the ability to achieve and sustain hyperactivation and the promotion of capacitation (deLamirande & Gagnon 1993; Zini, deLamirande & Gagnon 1995). Although these initial studies were performed in vitro with human sperm and the mechanisms underlying these effects remain unclear, these reports demonstrate that RS are not always detrimental to the fertilization process. Therefore, understanding the balance between the beneficial and harmful effects of RS and the balance between RS production and AO defences is integral to understanding sperm function. Future ecological studies can benefit by avoiding the oversimplification of RS as purely detrimental and investigating these other potential roles of RS in species with different mating strategies, modes of fertilization (internal vs. external fertilizers), or fertilization environments (marine vs. terrestrial spawners). Finally, as the majority of studies concerning OS in semen have been performed in vitro and in the fields of human infertility and poultry semen cryopreservation, future research across a wider range of taxa (as suggested above) may provide a clearer picture of the influence of OS on gametic performance and individual fitness.