Urban areas have been growing rapidly over the last several decades, and it is expected that they will continue to expand at the expense of natural and rural habitats. From an ecological perspective, urban environments present some novel challenges for birds, such as new predators, new flora, human presence, unique food resources, and high levels of chemical, light, and acoustic pollution (Gaston, 2010; Lepczyck and Warren, 2012; Forman, 2013; Gil and Brumm, 2014). However, these novel and potentially stressful conditions generated by urban environment processes have also provided new opportunities for certain species to gain an ecological advantage by exploiting city resources (Lepczyck and Warren, 2012; Costantini et al., 2014). Such resources are characterized by lower predation rates (Evans et al., 2015), higher environmental temperatures (Tryjanowski et al., 2015), lower competition for resources (Kark et al., 2007), and a constant, abundant, and more predictable food resources supply, in comparison with the surrounding non-anthropogenic areas (Shochat, 2004; Oro et al., 2013; Andersson et al., 2015; Tryjanowski et al., 2015; Marzluff, 2016).

The House Sparrow (Passer domesticus L.) is one of these urban exploiter species, meaning that it dominates highly urbanized environments. It is the most ubiquitous and globally distributed bird, probably due to its ecological tolerance that allows them to cope with urban environments: they are generalist, granivorous, gregarious, and sedentary, they nest in cavities, they are unbothered by human presence, and they are able to explore new feeding sources (Anderson, 2006; Kark et al., 2007; Evans et al., 2011). However, in the last few decades, urban populations of House Sparrow have significantly dropped, especially in the highly developed regions of Western Europe (De Laet and Summers-Smith, 2007; Peach et al., 2008; Shaw et al., 2008; De Coster et al., 2015). There is no consensus on the reason for such a decline, but the increasing level of pollution has been highlighted as one of the main driving factors (Shaw et al., 2008; Herrera-Dueñas et al., 2014).

The atmosphere of urban areas tends to accumulate high concentrations of carbon dioxide (CO₂), nitrous oxides (NO_x), sulfur dioxide (SO₂), ozone (O₃), heavy metals and other suspended particulate matters (PM). In cities, these pollutants are mainly derived by burning fossil fuels (Grimm et al., 2008; Gaston, 2010), representing an environmental risk from human to ecosystem levels (Forman, 2013). In relation to human health, a correlation has been described between the level of pollution and higher human mortality rates, mainly associated with cancer, asthma and cardiovascular disorders (Kelly, 2003). To date, deleterious effects on wildlife are less well-known, but pollution has been related to reproductive performance, lower survival rates and shorter lifespans (Isaksson, 2010; Koivula and Eeva, 2010; Salmón et al., 2016). One trait that has been highlighted as responsible for the toxicity of urban pollutants is their high reactivity to biomolecules (Koivula et al., 2011). Therefore, oxidative stress has been described as an underlying mechanism by which urban pollutants cause deleterious effects in living organisms (Isaksson, 2010).

Oxidative stress has been defined as the occurrence of enzymatic and non-enzymatic antioxidants that cannot fully neutralize the free radicals produced in the cells of living organisms. Free radicals are commonly generated as a by-product of normal metabolic processes. Hence all aerobic organisms have evolved multiple defense lines to prevent or counteract the toxicity of reactive by-products, ranging from antioxidant enzymes to low molecular weight endogenous antioxidants (i.e., GSH), complemented by some dietary antioxidant such as carotenoids or vitamins (Halliwell, 2007). Under pro-oxidant conditions, the unquenched free radicals remain in the system long enough to cause further oxidation reactions (Monaghan et al., 2009). Such reactions may damage macromolecules, such as lipids, proteins, and DNA. Severe accumulation of oxidative damage in these molecules leads to cellular dysfunction or apoptosis, which triggers disease and age-related degeneration (Costantini and Verhulst, 2009; Isaksson, 2015). This supports the hypothesis that the resistance to oxidative stress is a key cellular mechanism to cope with urban pollutants, so the capacity to quench the oxidative damage has been highlighted as a critical adaptation to life in the city (Costantini et al., 2014).

The levels of oxidative stress in organisms could also vary with developmental stage, environmental condition, and level of activity (Isaksson, 2015). Reproduction could increase oxidative stress as an inevitable consequence of increased reproductive effort (Romero-Haro et al., 2016) and/or because the developing offspring consume antioxidants otherwise allocated to self-maintenance (Alonso-Alvarez et al., 2004; Moller et al., 2010; Berglund et al., 2014). Therefore, oxidative stress may have a dramatic effect on bird fitness during the breeding season.

Regarding the roles of these antioxidants in maintaining the oxidative balance, four different scenarios could be described: (I) low oxidative damage due free radicals being quenched by the excess supply of antioxidants (endogenous or diet-provided); the maintenance of the oxidative balance in this case is not costly for self-maintenance of individuals, and resources can be reallocated to other functions (i.e., reproduction). (II) Oxidative damage is in balance with antioxidant availability; therefore, there is no surplus of resources, but self-maintenance is not yet costly for individuals. (III) Oxidative damage persists and to counteract it, the individual may increase its antioxidant capacity by up-regulating some antioxidant enzymes; this entails a cost because some resources must be diverted from self-maintenance. (IV) Oxidative damage is uncontrolled; the antioxidant defenses have been overwhelmed and the survival of the individual is compromised (Costantini and Verhulst, 2009; Monaghan et al., 2009) (Figure 1).

We predict that the first scenario will correspond to areas with a low degree of urbanization and therefore to lower levels of pollution (such as the rural ones); the second one will correspond to mild stressor environment (such as the suburban areas); whereas the third scenario will correspond to highly urbanized areas, where the high level of pollution may be a challenge for antioxidant defenses. We will explore the oxidative stress balance during winter, but also during the breeding season, when adults are reproducing.

We found that urban sparrows showed higher levels of oxidative damage and a higher activity of antioxidant enzymes, but a lower antioxidant capacity in comparison with the rural birds. Some of these differences are especially increased during the breeding season.

According to our predictions, the low oxidative damage scenarios (I and II) matches the situation found in the rural and suburban populations, respectively. Oxidative damage seems to be low or the level of dietary antioxidants is high, because the oxidative status (total antioxidant capacity) is balanced. In this case, up-regulation of antioxidant enzymes has not been required, and the depletion of antioxidants is low (Costantini and Verhulst, 2009; Monaghan et al., 2009); therefore, they are available for self-maintenance, which reduces the cost of coping with oxidative damage (Isaksson et al., 2011). The higher antioxidant capacity found in the rural populations (scenario I) could reflect a surplus of dietary antioxidants, probably due to the availability of some dietary antioxidant complements such as vitamins or minerals. Based on the antioxidant capacity found in suburban birds, the pool of antioxidants must be lower or, probably, oxidative damage increases with urbanization process (scenario II) (Costantini et al., 2014; Isaksson, 2015). In both cases (rural and suburban areas), the maintenance of the oxidative balance would not be costly for individuals, and a reproductive investment should not represent an additional challenge for coping with the oxidative balance (Alonso-Alvarez et al., 2004; Wiersma et al., 2004).

On the other hand, the increasing oxidative damage scenario (III) is in accordance with the results found in urban areas. The oxidative damage in lipids, proteins, and GPX activity showed higher values than those in rural and suburban populations; and the antioxidant capacity of urban birds was lower, especially during the breeding season. These results were expected and in line with a previous study which also found a lower TAC linked to urban areas (Herrera-Dueñas et al., 2014). Here, we propose that pollution could be responsible for the variation in oxidative stress balance between urban and rural areas. For instance, major atmospheric pollutants, such as particles (PM) or heavy metals, have been described as pro-oxidant elements (Kelly, 2003; Isaksson, 2010; Koivula and Eeva, 2010) and their deleterious effects on oxidative stress balance have been described in some urban birds such as Great Tits (Parus major L.) (Koivula et al., 2011), Feral Pidgeon (Columba livia L.) (Kurhalyuk et al., 2009), White Stork (Ciconia ciconia L.) (Kaminski et al., 2007). Similar results have also been found in humans (Chuang et al., 2007; Moller et al., 2014). Therefore, the quantification of these pollutants in the blood or feathers of our populations of House Sparrow may be a useful parameter to confirm this hypothesis in the future.

The lower response to TAC in the urban population is also in line with previous studies that reported poor quality of urban food in terms of antioxidant levels such as carotenoids, vitamins, and minerals (Isaksson and Andersson, 2007; Isaksson, 2015; Tryjanowski et al., 2015). In other urban species like Blackbirds (Turdus merula L.), it has been reported that rural populations showed higher concentrations of vitamin E and carotenoids than their urban conspecifics, mainly caused by differences in their diet (Moller et al., 2010). Additional studies about the quality of diet in urban birds and its effects on oxidative stress balance could contribute to verifying this theory.

Under pro-oxidant conditions, the up-regulation of antioxidant enzymes may be required and dietary antioxidants may be depleted; both mechanisms entail some costs (Dowling and Simmons, 2009; Isaksson et al., 2011), which may reduce the fitness and reproductive performance of urban birds (Isaksson, 2015). This is a costly status for individuals, as they must allocate all available resources to counteract the oxidative stress imbalance (Costantini and Verhulst, 2009; Monaghan et al., 2009; Isaksson, 2015). The lack of dietary antioxidants during the breeding season in a challenging environment such as an urban area could lead to potentially negative carry-over effects, in nestlings, linked to early-life nutritional constraints (Metcalfe and Monaghan, 2001). It has been widely reported that the poor condition of House Sparrows nestlings in urban areas is linked to a nutritional deficit (Liker et al., 2008; Bókony et al., 2010; Seress et al., 2012; Meillère et al., 2015). Furthermore, an early-life diet that is poor in antioxidants can result in a long-term impairment in the capacity to assimilate dietary antioxidants, like the carotenoids, vitamin A and E in adulthood (Blount et al., 2003; Monaghan et al., 2009; Costantini et al., 2014). Nutritional constraints in early-life further interact to shape the organization of the redox system (Costantini et al., 2014), which could may explain the lack of the antioxidant capacity that we have found in the adults from our urban populations which could be due to the important role that epigenetic mechanism seem to play in the final modulation of oxidative status (Isaksson, 2015).

In our study, we found that rural birds showed better body condition in comparison with their conspecifics in other areas, especially the urban individuals. This result is consistent with the hypothesis of Shochat (2004) in which birds from urban environments show a poorer condition than conspecifics rural birds; and according to previous studies that found these differences between urban and rural populations of House Sparrow (Vincent, 2005; Liker et al., 2008; Bókony et al., 2010; Seress et al., 2012; Herrera-Dueñas et al., 2014). The weak condition of urban birds may be also linked to the nutritional constraints that they suffer during early-life, which could negatively affect body condition in adulthood (Bókony et al., 2010; Seress et al., 2012), even if the restriction diminish at later life stages (Liker et al., 2008). Body condition, especially during development, is often positively correlated with pre- and post-fledgling survival rates (Lamb et al., 2016); consequently, poor condition is often observed in a declining population such as urban House Sparrows (Dulisz et al., 2016).

Apart from the direct effects of poor diet and pollution on oxidative stress, we should also consider that natural selection may have favored different phenotypes in urban and rural populations. For instance, it has been reported that expression of genes that code for metal detoxification enzymes are upregulated in urban populations of Great Tits (Parus major L.) in comparison with rural ones. Therefore, epigenetic mechanisms could generate life-long changes in gene expression, and subsequently, phenotypic traits between urban and rural populations (Watson et al., 2017).

Our study has revealed the fragile oxidative stress balance of the urban House Sparrow, promoted either by an excess of harmful free radicals due to pollution, or insufficient antioxidants due to scarce availability of dietary antioxidants, or both in urban environments. Therefore, birds need to invest resources in maintaining the oxidative stress balance, which is costly especially during the breeding season, when the constraints of antioxidants could entail negative consequences not only on the body condition of the individuals, but also on the viability and future fitness of their offspring. The accumulative effects of lower quality individuals and carry over effects at the population level may be one of the causes underlying the population decreases of urban House Sparrow populations across European cities.

JA, MA, JP, and AH conceived and designed the study. JA, JP, and AH performed the fieldwork. JP and AH conducted lab work. AH performed statistical analyses with support from JP. AH drafted the manuscript and all the authors contributing to the final text.