The emergence of aerobic organisms was a consequence of oxygen becoming a significant component of the earth’s atmosphere. The evolution of aerobic organisms was probably the result of their gradual adaptation to oxygen. Antioxidant enzymes probably facilitated part of such adaptation, since strict anaerobes have neither superoxide dismutase (SOD) nor catalase, whereas aerotolerant anaerobes have small levels of those enzymes [1]. Aerobic organisms use oxygen to produce the chemical energy they need to survive. Metabolic pathways of aerobic organisms lead to the obligate production of reactive oxygen species (ROS) and reactive nitrogen species (RNS), which are not necessarily bad. Indeed, reactive species are actually needed as signal transduction elements in processes that are essential for life, such as cell growth, differentiation, preconditioning and cell death [2,3]. They also participate in functions such as phagocytosis and stress response [4]. This echoes Barry Halliwell’s [5] “All aspects of aerobic life involve free radicals and antioxidants—you cannot escape them nor should you wish to”.

Free radicals are molecules or atoms characterized by possessing an unpaired electron that is always seeking a counterpart to reside inside the parent home or in the home of the other electron. Free radicals can be produced during cellular processes or as a result of the interaction of free radicals, reactive species, or free radical-generating systems with other molecules [6,7]. Many of the ROS and RNS are free radicals. There are also macromolecular radicals that are formed by the effect of oxygen, drug-induced free radicals, irradiation or free radicals on such macromolecules as protein and lipids. For example, it has been demonstrated that γ-irradiated proteins produce electron paramagnetic resonance (EPR)—detectable superoxide and huydroxyl radicals, in the presence of ferrous iron, via their intermediate conversion to hydroperoxydes [6]. Such macromolecular radicals lead to cellular dysfunction and decreased cell viability [8]. In the presence of hydrogen peroxide, aminoglutethimide (xenobiotic) modified myeloperoxidase, transforming this enzyme in a macromolecular free radical; a reaction that has been suggested to have a role in the development of drug-induced agranulocytosis and has been inhibited by polyunsaturated fatty acids [7]. Generation of lipid-derived free radicals was demonstrated, by in vivo EPR, in the livers of rats treated with the carcinogen diethylnitrosamine [9]. Lipid-derived free radicals have also been demonstrated, through EPR, in culture cells incubated with ferrous iron [10]. Free-radical modified lipids [11] and lipid-derived free radicals may play an important role in the development of atherosclerosis.

According to Halliwell and Gutteridge [12] an antioxidant is “any substance that, when present at low concentrations compared to those of an oxidizable substrate, significantly delays or prevents oxidation of that substance”. Just as reactive species are not necessarily bad, antioxidants are not necessarily good. Antioxidants are a kind of matchmaker midwife trying to give or receive an electron to complete the other’s “pair”, but by making a proper pair, antioxidants can themselves become incomplete, and behave as free radicals; i.e., reactive and looking to pair with an unpaired electron. It is known that those “reactive” antioxidants can be stabilized (recycled) by other antioxidants, which in turn become “reactive” and can be recycled in a cascade of “reactivity mitigation” (Figure 1). It is all a matter of equilibrium. It well is accepted that oxidative stress is produced when the equilibrium between reactive species and antioxidants is tilted in favor of the former. However, the equilibrium can be broken also if antioxidant levels exceed those of the reactive species. Dündar [13] proposed the term “antioxidative stress” for such disequilibrium.

Nature is certainly wise. Aerobic organisms mitigate reactive species not only with antioxidant molecules that donate or receive electrons but also through reactions catalyzed by antioxidant enzymes. Figure 2, below, illustrates the production of superoxide by pro-oxidant enzymes and the intervention of antioxidant enzymes to stabilize such free radical.

Figure 2 is only an example of the different reactions that produce reactive species, which potentially lead to oxidative stress. We have taken this very simple sequence of reactions to illustrate how changes can lead to oxidative stress and make a later parallel with the antioxidant-induced stress. Superoxide radical production can be increased if the activity of NADPHox or XO increases. Indeed, it has been postulated and partially shown, that oxidative stress in hypertension is due, at least in part, to the increase of NADPHox’s activity in vascular smooth muscle cells, fibroblasts and endothelial cells [14,15]. Polymorphisms of the genes codifying SOD and GPx have been related to oxidative stress due to decreased activity of the enzymes and high incidence of coronary artery disease [16], ischemic stroke [17,18] and atherosclerosis [19]. Polymorphisms of GST have been associated with cardiovascular events produced in smokers [20].

Reactive species participate in different functions and play a role as signal transduction elements in many physiological events. ROS and RNS participate in signal transduction of cytokine receptors, tyrosine receptor, serine/threonine kinases, G protein-coupled receptors, ion-channel linked receptors in response to angiotensin II, cytokines, glutamate, epidermal growth factor, vascular endothelial growth factor, tumor necrosis factor α and platelet derived growth factor [2]. Free radicals and ROS participate in the redox regulation of activator protein 1, nuclear factor κ B (NFκ-B), cyclic response element-binding protein (CREB), nuclear factor E2-related factor 2 (Nrf2) and p53 (“guardian of the genome”) [2,15]. Superoxide and hydrogen peroxide activate extracellular signal-related protein kinase (ERK), protein kinase B, mitogen-activated protein kinases (MAPKs) and insulin receptor kinase [15,21–23].

The effects of reactive species are dose-dependent. Incubation of Jurkat T-cells with increasing concentrations of H₂O₂ produced proliferation (0.7 μM), apoptosis (1.0–1.3 μM) or necrosis (>3 μM) [24]. ROS participate in proteolysis into the proteasome. H₂O₂ increases proteolysis at concentrations of 20–400 μM, whereas it inhibits proteolysis at higher, mM concentrations, producing an accumulation of oxidized proteins [25]. Preconditioning is one of the areas where the dose-dependent physiological effects of reactive species are evident. Repeated or unique application of injurious stimuli, at intensities below the threshold of damage, activates endogenous mechanisms that afford protection when subsequent major injurious stimuli are applied. This phenomenon is known as preconditioning [26]. Hyperbaric oxygen [26], short ischemia [27–29], small doses of lipopolysaccharides [30], restraint stress [31], hypoxia [32], hyperthermia [32] and moderate aerobic exercise [33,34] are some of the stimuli producing preconditioning. We chose the example of exercise to explain the role of reactive species on preconditioning (see Figure 3). An excess of ROS are produced during acute or intense exercise [35,36]. It was thought that mitochondria were the primary site of superoxide production during exercise, because oxygen consumption in respiration is increased. However, it is now known that superoxide is produced mainly through NADPHox, XO and phospholipase A₂ (PLA₂) during intense exercise [34]. Chronic moderate exercise suppresses oxidative stress produced by acute intense exercise [34,35]. The mechanism of such adaptation seems to involve, as has been demonstrated in different studies, activation of NFκ-B by ROS and the consequent expression of nitric oxide synthases and antioxidant enzymes [37–39]. Indeed, the incubation of cells with hydrogen peroxide induces the expression of antioxidant enzymes [40].

As mentioned above, antioxidants become “unstable” and “reactive” when they lose or receive electrons in the presence of reactive species (see Figure 1). In these conditions antioxidants exhibit pro-oxidant effects and can become harmful. Their redox potential (tendency to acquire electrons and thereby be reduced) could be related to the toxic effects of antioxidants. The redox potentials of some antioxidants of interest are: 0.65 V for β-carotene, 0.50 V for Vitamin E, 0.25–0.50 V for flavonoids, 0.25 V for uric acid and 0.01 V for Vitamin C [44–46]. It has been said that a network of antioxidants is necessary. Antioxidants should be interconnected in such a way that they can turn off each other’s reactivity after interacting with reactive species [47]. Precisely; Damiani [48] defined a good antioxidant as one that produces a low oxidant reactivity with a low capacity to produce peroxidation.

α-tocopherol (α-TC, Vitamin E) produces α-tocopheroxyl radical (α-TC.) when it reacts with reactive species such as peroxynitrite [49] or superoxide [50]. α-TC. is then recycled to α-TC by other antioxidants such as ascorbic acid (Vitamin C) and glutathione [48,51,52]. As soon as ascorbic acid recycles Vitamin E, it is transformed to the ascorbyl radical, which has a lower reactivity than α-TC. [48]. α-TC. is also recycled to α-TC by β-carotene [52,53]. Therefore, it is important that Vitamin E be administered with another antioxidant. It is also important to have significant levels of endogenous antioxidants to allow for the recycling of “reactive” antioxidants. Some conditions reduce endogenous antioxidants and could potentially affect the recycling of Vitamin E (or another antioxidant), leaving α-TC. available to produce lipoperoxidation. One such condition is the reduction of ascorbic acid in smokers [48,52]. Ascorbic acid also recycles oxidized glutathione (GSSG) to reduced glutathione (GSH) [51], and carotenoid radicals to carotenoids [53]. Dihydro-lipoic acid (DHLA), the reduced form of lipoic acid (an antioxidant) recycles GSH from GSSG and the ascorbic ascorbyl radical [54].

The pro-oxidant or antioxidant effects of some antioxidants depend on their concentration. For example lipoic acid and DHLA inhibit nitro-tyrosine formation by peroxynitrite at a 0.01–0.05 mM concentration, whereas promote mitochondrial permeability transition at 0.1 mM, increase of glucose intake at 0.25 M, and stimulate calcium mitochondria release at 1 mM. β-carotene protects DNA oxidative damage at 1–3 μM, whereas it oxidizes DNA at 4–10 μM. High concentrations of β-carotene inhibit the growth of cancer cells in culture and promote their apoptosis [53].

Some antioxidants exhibit pro-oxidant activities in the presence of transition metals. Such is the case of ascorbic acid that is transformed into an ascorbyl radical [55] and hydroxycinnamic acid whose phenolic groups are transformed into phenoxy reactive groups [56]. Retinal and carotenoid radical are produced by the oxidative cleavage of β-carotene in the presence of oxygen and ferrous ion [57,58]. Oxidative cleavage of β-carotene is also produced by many other oxidative stimuli: ultraviolet light (artificial and sun light), heat, free radicals, and ozone [59]. Carotenoid breakdown products (CBP) were produced in ferrets by cigarette smoke exposure. The effect was prevented by α-TC and ascorbic acid [52].

Reactive species produced by the interaction of an antioxidant with a free radical such as ROS or RNS exhibit potentially toxic pro-oxidant effects. CBP inhibit mitochondrial respiration in the rat lung, brain and liver [59], stimulate apoptosis in neutrophils [59], oxidize DNA [60] and up-regulate advanced glycation endproducts [53].

In summary, the pro-oxidant and potential harmful effects of an antioxidant depend on (a) its concentration; (b) its redox potential; (c) the presence of other antioxidants; (d) the presence of transition metals, (e) the activity and concentrations of endogenous antioxidants. And, even though not yet demonstrated, the genetic background could also contribute to the harmful effects of antioxidants. Some of these factors are further considered in the following section.

The validity of antioxidant therapy has been questioned, having been pointed out that even though several diseases have been related to oxidative stress, antioxidant therapy does not change the natural course of diseases [5,47,61–64]. Some authors have commented that antioxidant therapy has only been validated in experimental models of disease and cell culture [5,61]. However, different factors cause artifacts in cell culture (e.g., high oxygen levels compared to the pO₂ of tissues in the organism, components of the cell culture media, high concentrations of antioxidants) [5,61]. In the clinical arena one of the problems is the biomarkers to evaluate oxidative stress [63]. It has been postulated that a combination of low doses of antioxidants could only be useful in patients suffering from antioxidant deficiencies [63]. It remains unknown if patients with gene polymorphisms affecting the structure and function of antioxidant enzymes could have different responses to antioxidants.

What is interesting is that when no effects or negative results of antioxidants are reported, there is often an alternate conclusion that antioxidants can be dangerous. For example: the method used to evaluate oxidative stress, treatment strategy (preventive strategy, doses, combination of antioxidants, duration), the stage of the disease, the follow up period, and number of patients [65]. Interestingly, in a recent review, Ristow and Schmeisser [66] suggest that life span could be extended by oxidative stress. A summary of clinical trials with antioxidants, as well as possible explanations are shown below.

Even though antioxidants have been considered beneficial, a number of clinical trials have shown harmful effects. The most extreme results evidenced increased mortality in individuals taking antioxidant supplements. Although the mechanisms are not explained in those trials, the results suggest that disrupting the delicate balance between antioxidants and reactive species produces antioxidant-induced stress, if the antioxidants overwhelm the physiological production of reactive species.

Though these conclusions remain speculative, the best advice would be to ingest antioxidants from food sources rather than from self-prescribed supplements. The most probable way of increasing the endogenous antioxidant defense would be by practicing moderate aerobic exercise, as an every day healthy routine. Gene polymorphisms, which condition endogenous antioxidant deficiencies, would be the exception to such advice.