Multiple levels of antioxidative protection by melatonin

Antioxidative protection by melatonin is not just a matter of direct radical scavenging (Fig. 2), as becomes immediately evident from stoichiometry. Although tissue levels of melatonin can be considerably higher than those in the circulation, the quantities of free radicals generated in its metabolism would still be too high for the available amounts of the indoleamine. Our understanding is that direct scavenging by physiological concentrations of melatonin by a non-enzymatic contribution to the kynuric pathway and the subsequent actions of the metabolites formed becomes important. Signaling effects of melatonin, however, are always possible at physiological levels.

Melatonin upregulates several antioxidant enzymes. Most frequently, this has been demonstrated for glutathione peroxidase [7,77,88,104-118] and sometimes glutathione reductase [7,108,112,119], presumably indirectly via GSSG. In some tissues Cu, Zn- and/or Mn-superoxide dismutases [7,108-112,117-123] and, rarely, catalase [112,118,123,124] are upregulated. Stimulation of glutathione peroxidase seems to be widely distributed among tissues and is observed quite regularly in both mammalian and avian brain; upregulations in other organs were more variable. The action of melatonin on glutathione metabolism seems to exceed the effects mentioned. Stimulation of glucose-6-phosphate dehydrogenase [108] and ?-glutamylcysteine synthase [10,112] indirectly supports the action of glutathione peroxidase by providing reducing equivalents (NADPH) for the action of glutathione reductase and by increasing the rate of glutathione synthesis, respectively.

Contrary its effect on the enzymes of glutathione metabolism, the effect of melatonin on superoxide dismutase subforms and catalase strongly depends on organs and species. Stimulation was observed in some tissues, but not in others; in some cases, even decreases were reported. This may not only be a matter of differences in responsiveness of cell types. The complexity in the regulation of the respective enzymes has to be considered. Frequently, they exhibit compensatory rises in response to oxidative stress. When melatonin is counteracting experimentally induced stress, the result may be a normalization of enzyme activity, i.e., lower values, compared to animals treated with oxidotoxins, rather than inductions. Such normalizations were, in fact, described [114,125]. However, in cases of stronger oxidative stress, active centers of enzymes may be destroyed by the free radicals generated and normalization of enzyme activities by melatonin administration appears as an increase [110,124,125].

An additional aspect of melatonin's actions on antioxidant enzymes deserves future attention: In two neuronal cell lines, physiological concentrations of melatonin not only induced glutathione peroxidase and superoxide dismutases at the mRNA level, but concomitantly increased the life-time of these mRNAs [117].

Melatonin also contributes to the avoidance of radical formation in several independent ways. It downregulates prooxidant enzymes, in particular 5- and 12-lipoxygenases [112,126-128] and NO synthases [9,34,77,108,112,129-134]. The widely observed attenuation of NO formation is particularly important in terms of limiting rise in the strongly prooxidant metabolite peroxynitrite and of the free radicals derived from this compound, namely, •NO2, carbonate (CO3•-) and hydroxyl (•OH) radicals. Suppressions of both lipoxygenase and NO synthase may additionally set limits to inflammatory responses, although the immunomodulatory actions of melatonin are certainly more complex and may involve additional effects of melatonin and AMK, too.

Another widely unexplored but potentially important signaling effect of melatonin in antioxidative protection concerns quinone reductase 2 [77,135,136]. This enzyme, which is implicated in the detoxification of potentially prooxidant quinones, binds melatonin at upper physiological concentrations, so that it had originally been presumed to represent a melatonin receptor. Although its precise function under the influence of melatonin is not yet fully understood, the relationship to the indoleamine may become of future interest from the standpoint of nutrition, since quinones are taken up with food, especially, vegetables.

Although less relevant from a nutritional point of view, melatonin also contributes indirectly to radical avoidance, e.g., by its antiexcitatory effects in the central nervous system, and as an endogenous regulator molecule controlling rhythmic time structures. This last action may be particularly important for well-timed alimentary melatonin supplementation in the elderly, who exhibit a strongly reduced amplitude in the circadian melatonin rhythm. The significance of appropriate timing for maintaining low levels of oxidative damage has been overlooked for quite some time. However, temporal perturbations as occurring in short-period or arrhythmic circadian clock mutants lead to enhanced oxidative damage, effects observed in organisms as different as Drosophila and the Syrian Hamster [77,137,138].

In the last few years, mitochondrial effects of melatonin have been discovered which may turn out to be even more important than the protective actions described above. Mitochondria are the main source of free radicals in the majority of animal cells and are implicated in aging processes. The importance of mitochondrial diseases is increasingly perceived. Mitochondria play a key role in apoptosis. Notably, several of the mitochondrial effects of melatonin were obtained at low pharmacological doses in drinking water [116,139,140] or even at near-physiological concentrations down to 1 nM [141].

Several studies of mitochondrial effects revealed attenuation of mitochondrial lipid peroxidation, prevention of oxidative protein and DNA modifications, preservation of ultrastructure, resistance against toxins etc., findings which were widely in line with previous concepts of protection [10,113,142-146]. Moreover, melatonin was shown to affect redox-active compounds in mitochondria, in particular, to decrease NO [143,147] and to restore normal levels of reduced glutathione [113,144] and coenzyme Q10 [148].

More importantly, beyond these rather conventional findings, with few exceptions, melatonin was found to increase mitochondrial respiration and ATP synthesis, in conjunction with rises in complex I and IV activities [112,141-143,146,147,149,150]. Complex I and IV activities were also found to be increased by melatonin in hepatic mitochondria of senescence-accelerated mice [116,140,151]. Moreover, melatonin was found to enhance gene expression of complex IV components [147].

The improvements of ATP formation and O2 consumption are presumably not decisive for protection, but can serve as good indicators for the reduction of electron leakage from the respiratory chain. Electron transfer to molecular oxygen at the matrix side, largely at iron-sulfur cluster N2 of complex I [152], is a major source of free radicals. This process also diminishes electron flux rates and, therefore, the ATP-generating proton potential. Processes affecting the mitochondrial membrane potential such as calcium overload, either due to overexcitation, to protein misfolding or to damage by free radicals, are antagonized by melatonin. In cardiomyocytes, astrocytes and striatal neurons, melatonin prevented calcium overload [153,154], counteracted the collapse of the mitochondrial membrane potential induced by H2O2 [153], doxorubicin [155] or oxygen/glucose deprivation [154], and also inhibited the opening of the mitochondrial permeability transition pore (mtPTP), thereby rescuing cells from apoptosis. In addition to the antioxidant actions, melatonin directly diminished mtPTP currents, with an IC50 of 0.8 µM [154], a concentration which would require mitochondrial accumulation of melatonin, something which is possible again due to the amphiphilicity of melatonin.

The effects of melatonin on the respiratory chain open new perspectives for diminishing radical formation, instead of seeking only antioxidant effects for the elimination of radicals already formed. We have proposed a model of radical avoidance (Fig. 3) in which electron leakage is reduced by single-electron exchange reactions betwen melatonin and components of the electron transport chain [77,78]. In fact, mitochondrial H2O2 formation was found to be reduced by melatonin [156]. The basic idea of the model is that of a cycle of electron donation to the respiratory chain, eventually to cytochrome c [78], followed by reduction of the formed melatonyl cation radical by electron transfer from N2 of complex I. The cation radical is assumed to act as alternate electron acceptor competing with molecular oxygen, thereby decreasing the rate of O2•- formation. In addition to the electrons being largely recycled, most of the melatonin is also. Therefore, such a mechanism would only require very low, quasi-catalytic amounts of melatonin, in accordance with the effects demonstrated with nanomolar concentrations. Because the recycled electrons are not lost for the respiratory chain, this would also lead to improvements in complex IV activity, oxygen consumption and ATP production. Alternately, the melatonin metabolite AMK, which is also highly reactive and can undergo single-electron transfer reactions [43], may act in the same way. The prediction of our model of mitochondrial protection by AMK was confirmed by other investigators [147]: AMK was shown to exert effects on electron flux through the respiratory chain and ATP synthesis very similar to those observed with melatonin.

A highly attractive aspect of mitochondrial protection results from the small quantities required: experimentally induced mitochondrial damage in rat fetuses was even prevented by maternally administered melatonin [146]. The mechanism as outlined, requiring only low amounts of melatonin or its metabolite AMK, would make these compounds even more interesting from a nutritional point of view. The amounts present in selected food, such as some vegetables, but even more nuts and cereals, could suffice for maintaining tissue levels of the indoleamine capable of safeguarding mitochondrial function, particularly in elderly persons whose nocturnal melatonin maxima in pineal gland and circulation have substantially declined with age. Transient moderate rises in blood melatonin during the day resulting from direct uptake or postprandrial release from the gastrointestinal tract should not be regarded as a problem in terms of circadian timing. The circadian system responds to melatonin according to a phase response curve [157,158]: the so-called silent zone, during which no substantial phase shifts are induced, extends throughout the largest part of the day.