Melatonin as a mitochondria-targeted antioxidant: one of evolution’s best ideas

Abstract

Melatonin is an ancient antioxidant. After its initial development in bacteria, it has been retained throughout evolution such that it may be or may have been present in every species that have existed. Even though it has been maintained throughout evolution during the diversification of species, melatonin’s chemical structure has never changed; thus, the melatonin present in currently living humans is identical to that present in cyanobacteria that have existed on Earth for billions of years. Melatonin in the systemic circulation of mammals quickly disappears from the blood presumably due to its uptake by cells, particularly when they are under high oxidative stress conditions. The measurement of the subcellular distribution of melatonin has shown that the concentration of this indole in the mitochondria greatly exceeds that in the blood. Melatonin presumably enters mitochondria through oligopeptide transporters, PEPT1, and PEPT2. Thus, melatonin is specifically targeted to the mitochondria where it seems to function as an apex antioxidant. In addition to being taken up from the circulation, melatonin may be produced in the mitochondria as well. During evolution, mitochondria likely originated when melatonin-forming bacteria were engulfed as food by ancestral prokaryotes. Over time, engulfed bacteria evolved into mitochondria; this is known as the endosymbiotic theory of the origin of mitochondria. When they did so, the mitochondria retained the ability to synthesize melatonin. Thus, melatonin is not only taken up by mitochondria but these organelles, in addition to many other functions, also probably produce melatonin as well. Melatonin’s high concentrations and multiple actions as an antioxidant provide potent antioxidant protection to these organelles which are exposed to abundant free radicals.

Introduction

An estimated 2.5 × 10⁹ years ago, molecular oxygen (O₂) began to rise in the Earth’s atmosphere due to its persistent release from photosynthetic bacteria that had evolved an estimated half-billion years earlier [1]. The rise in atmospheric O₂ was a highly selective pressure for the evolution of organisms to use O₂ as the basis of their metabolism. O₂-based metabolism proved to be an enormous metabolic advance for aerobic species given that the total combustion of glucose yields 36 molecules of ATP per molecule of glucose. An estimated 95% of O₂ used by mammalian cells is reduced by the addition of four electrons by cytochrome c to produce two molecules of water.

There is, however, a significant downside to oxidative metabolism. The transfer of electrons between the complexes of the respiratory chain is not flawless with some electrons escaping where they chemically reduce adjacent O₂ molecules [2]. This inappropriate metabolism of O₂ generates free radicals and other reactive oxygen species (ROS) which are often highly toxic to molecules in the vicinity of where they are produced. Free radicals are derivatives of O₂ that contain an odd number of electrons in their valence orbital; they include the superoxide anion radical (O^·−₂), hydroxyl radical (·OH), hydroperoxyl radical (HOO·), peroxyl radical (ROO·), and alkoxyl radical (RO·), among others [3]. Some ROS are not free radicals because they possess an even number of electrons and include hydrogen peroxide (H₂O₂), hypochlorous acid, etc.

O^·−₂ is a consequence of the univalent reduction of O₂ (Fig. 1). Radicals are typically unstable and sometimes highly reactive. While O^·−₂ per se is not highly toxic to neighboring molecules, its most damaging actions stem from its conversion to secondary highly toxic agents, especially the ·OH and the peroxynitrite anion (ONOO⁻) [4, 5]. In an aqueous environment, O^·−₂ is quickly dismutated to H₂O₂. Hence, the formation of O^·−₂ is invariably accompanied by the production of H₂O₂. H₂O₂, a non-radical ROS, is rather stable and only sluggishly interacts with a number of organic molecules. A major aspect of H₂O₂ that makes it destructive is its high lipophilicity; this allows H₂O₂ to readily cross lipid-rich membranes thereby spreading the potential damage inflicted by free radicals. The other danger of H₂O₂ results when it reacts with iron, copper (and other transition metals) or with certain hemoproteins to yield the ·OH [6, 7]. These reactions are termed the Haber–Weiss reactions or, perhaps more appropriately, the O^·−₂ driven Fenton reaction. The ·OH is devastatingly reactive and immediately reacts with any molecule in the immediate vicinity of where it is produced; these reactions occur at diffusion limited rates. As a result, the ·OH is extremely short-lived and its “reaction cage” is very small so the damage it inflicts is site-specific. It is estimated that, of the total free radical/ROS molecular damage that occurs in organisms, the majority may be a consequence of the ·OH [6, 8]. ·OH also interacts with proteins, carbohydrates, nucleic acids, and lipids to produce ROO· as intermediates. ROO·, although less reactive than the ·OH, have a relatively long half-life and, therefore, they damage molecules at some distance from their site of production. The most thoroughly studied reaction of ROO· involves the peroxidation of polyunsaturated fatty acids (PUFA). The destruction of PUFA in cell membranes is a major factor that leads to the functional deterioration of cells and their eventual death. ROO· can also oxidize carbohydrates, proteins and some sulfhydryl components of hemoproteins.

Fig. 1.

A small percentage of oxygen inhaled/utilized by aerobic organisms generates oxygen-based derivatives, often called reactive oxygen species (ROS), that can damage critical molecules within cells. Some of the derivatives are free radicals [with an unpaired electron represented by the dot (·)] and others are not, e.g., hydrogen peroxide. The superoxide anion radical is quickly metabolized by superoxide dismutase (SOD) to hydrogen peroxide which can be removed from the intracellular environment by either catalase (CAT) or glutathione peroxidase (GPx). Oxidized glutathione (GSSG) is converted back to reduced glutathione (GSH) by glutathione reductase (GRd). The most destructive derivatives of oxygen are the hydroxyl radical and peroxynitrite. The hydroxyl radical is formed from hydrogen peroxide during the Fenton reaction and peroxynitrite is generated when the superoxide anion couples with nitric oxide. A large percentage of the ROS formed within cells is produced in mitochondria as a consequence of the leakage of electrons (e⁻) from the electron transport chain

While O₂ is considered poisonous as noted above, the predominate theory to explain its toxicity is a consequence of its chemical reduction to O^·−₂. The most obvious source of O^·−₂ in vivo in aerobic cells is generally considered to be the mitochondrial electron transport chain (ETC) (Fig. 2). In addition to the presence of the ETC in the mitochondria of all mammalian cells (except erythrocytes since they lack mitochondria), it is also present in the membranes of many bacteria; in plant chloroplasts and some other less studied sites [9, 10].

Fig. 2.

The mitochondrial respiratory chain utilizes oxygen to generate energy in the form of ATP. Free radicals and reactive oxygen species are formed when electrons that are passed between successive complexes are fumbled and chemically reduce adjacent oxygen molecules. The toxic derivatives of oxygen, since the majority are formed in mitochondria, especially damage mitochondrial DNA, proteins and lipids. Because mitochondria are a primary source of toxic derivatives of oxygen, an antioxidant positioned in mitochondria would be especially important in reducing oxidative damage

While the mitochondrial ETC is efficient in shunting electrons between successive components (which constitute the complexes), some electrons are fumbled and reduce nearby oxygen molecules to O^·−₂. The quantity of O^·−₂ produced is related to the O₂ tension; thus O^·−₂ production increases as the concentration of O₂ rises. Under physiological O₂ levels and when the ETC is functioning optimally, an estimated 1–3% of the O₂ is converted to O^·−₂. In the event of damage to the components of the ETC, they function suboptimally, so electron leakage increases as does ROS formation. This occurs in aged individuals, during toxin exposure, etc. Other sources of ROS include enzymes, e.g., xanthine oxidase, auto-oxidation reactions and haem proteins [11].

Evolution of melatonin’s multiple functions

Melatonin predictably evolved an estimated 3.0–2.5 billion years ago, probably in photosynthetic bacteria [12, 13], where it was specifically designed to neutralize the toxic O₂ derivatives that were produced during photosynthesis [14]. Once melatonin (N-acetyl-5-methoxytryptamine) appeared, for the next 3 billion years evolution never tinkered with the chemical structure of this agent such that the melatonin in cyanobacteria [15, 16] is structurally identical to melatonin that exists in present-day mammals, including the human [17, 18]. While melatonin’s structure has remained stable, its functions have become highly diversified. Thus, its original antioxidant function has been retained and supplemented with a variety of other actions during various stages of evolution (Fig. 3).

Fig. 3.

Melatonin is believed to exist in most, possibly all, animal and plant species. It predictably evolved 3.0–2.5 billion years ago in photosynthetic cyanobacteria as an antioxidant; this function has been retained to the present day including in humans. Other functions of melatonin, many more than are shown in this figure, appeared at later stages of evolution.

Reprinted with permission from Manchester et al. [13]

The first action of melatonin that was identified after its discovery in 1958 [19] was its ability to regulate the reproductive capability of photosensitive mammals [20, 21]. Thus, the duration of nocturnally elevated melatonin levels, which changed seasonally, were shown to drive the waxing and waning of reproductive competence in temperate and arctic species that are seasonal breeders [22, 23]. It soon became apparent that this action of melatonin was not its sole, and likely, not its most important action. Since then, a baffling array of functions have been assigned to this molecule, e.g., oncostatic [24, 25], anti-inflammatory [26–30], circadian rhythm modulation [31, 32], sleep promotion [33, 34], anti-venom [35, 36], body weight regulation [37, 38], anti-diabetic [39, 40], anti-fibrotic [41, 42], and others. The action of melatonin that has the longest history, however, is likely its ability to maintain redox homeostasis.

Melatonin as a free radical scavenger and as an antioxidant

Melatonin is uncommonly effective in reducing oxidative stress because of the number of means it has as a direct free radical scavenger and indirect antioxidant. Thus, melatonin (a) functions in this capacity in both the aqueous and lipid portions of the cell [43], (b) as a result, it protects lipids [44, 45], proteins [46, 47], and DNA [48, 49] from oxidative damage, (c) it is more highly concentrated in the regions of the cells where many of the free radicals are formed, e.g., mitochondria [50, 51], (d) it may be synthesized in the mitochondria [52] and at this site its synthesis may be inducible [53], (e) not only melatonin but a number of its metabolites also function as radical scavengers [54–56], (f) melatonin binds transition metals which reduces the formation of the most aggressive ROS, i.e., ·OH [57, 58], (g) melatonin stimulates the activity of a number of antioxidative enzymes [59–61], and (h) it promotes the synthesis of another important antioxidant, glutathione [62]. Finally, SIRT3—a class III histone deacetylase, which is primarily located in the mitochondrial matrix, has critical functions in protecting these organelles from oxidative stress [63]. While the role of melatonin in impacting mitochondrial SIRT3 is not mechanistically well defined, data indicate that SIRT3 may mediate at least some of the antioxidative actions of melatonin [64–66]. Considering these diverse functions, it is not always possible to determine the relative importance of each of these processes in a given highly oxidizing environment.

The detoxification of ROS/RNS is achieved by melatonin and a number of its metabolic kin in what is later referred to as the antioxidant cascade [67, 68]. Hence, the derivatives of melatonin that are formed when it directly neutralizes a free radical, often by electron donation [69], are equally as effective, and sometimes more so, than melatonin itself in reducing oxidative stress [70]. There are a number of comprehensive reviews that summarize the details by which melatonin functions in the reduction of oxidative damage. Rather than re-iterating these multiple actions here, the reader is directed to the associated publications [13, 15, 61, 71–78].

Immunocytochemical evidence for melatonin as a mitochondria-targeted antioxidant

That melatonin acts at the level of the mitochondria to prevent ROS toxicity [79, 80] was documented within a decade after the indole was discovered to be a direct potent free radical scavenger [67, 81, 82] and indirect antioxidant [58–60, 62, 83]. Using cyanide, an ETC complex IV inhibitor, Yamamoto and Yang [84] showed that this drug’s ability to cause seizures and kill mice was reversed by melatonin; the implication of these findings was that melatonin entered mitochondria and interfered with the negative actions of cyanide. Similarly, the actions of neurotoxins including 6-hydroxydopamine (6-OHDA) [85], kainic acid [86] and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) [87]-mediated dysfunction of mitochondrial Complex I are also overcome by melatonin. During the same time frame, melatonin was reported to augment the activities of respiratory chain Complexes I and IV and to reduce mitochondrial damage resulting from the treatment of rats with ruthenium red, a molecule that inhibits the mitochondrial Ca²⁺ uniporter leading to oxidative stress and mitochondrial uncoupling [88, 89]. These actions were not duplicated when vitamins C or E were used as replacements for melatonin. Martin and coworkers [90] also verified that melatonin stimulated oxidative phosphorylation and promoted a rise in ATP production in neuronal and hepatic cell mitochondria. Finally, hepatocyte respiratory physiology of aging mice was restored by melatonin, particularly at the levels of Complexes I and IV [91] and it limited ischemia/reperfusion-mediated mitochondrial dysfunction in rat liver [92]. Finally, the group led by Jou [93, 94] found that melatonin curtailed laser irradiation-induced ROS formation in astrocyte mitochondria and protected mitochondrial DNA from mutations/deletions and the cells from apoptosis.

The first comprehensive investigation designed to visualize melatonin’s ability to both quench ROS formed in mitochondria and to elucidate the series of molecular events that culminate in ROS-mediated cellular death was carried out by Jou et al. [95]. Using a combination of time lapse conventional and confocal microscopy with the aid of fluorescent probes, we monitored the actions of melatonin at the mitochondrial level. Initially, melatonin (100 µM) was found to highly effectively resist 10 mM H₂O₂-induced mitochondrial swelling and other cellular changes related to apoptosis of astrocytes as seen using phase contrast microscopy. Moreover, confocal microscopy imaging of the cells after they were treated with fluorescent probes to identify the health of mitochondria (Mito G) and of the nuclei (propidium iodide) documented that H₂O₂-mediated loss of functional mitochondria and nuclear impulsion was prevented by melatonin.

To determine whether the mitochondrial protection associated with melatonin treatment of H₂O₂-exposed astrocytes related to the ability of the indole to neutralize free radicals in these subcellular organelles, ROS levels in mitochondria were evaluated using dichlorofluorescein (DCF) and dichlororhodamine (D-123) [95]. The findings showed that after H₂O₂ treatment, there was a massive rise in free radical fluorescence in mitochondria within 10 min. Concurrent melatonin treatment prevented the rapid increase in mitochondrial ROS concentrations and maintained them at the levels seen in untreated astrocytes (Fig. 4). To achieve this marked inhibition of ROS fluorescence, it is likely that melatonin penetrated to the mitochondrial matrix where it either scavenged the ROS as they were formed or it prevented their formation in the mitochondria.

Fig. 4.

Fluorescence imaging of reactive oxygen species generation in cultured astrocytes, especially in the mitochondria, and the inhibition of ROS by melatonin. a An enhanced pseudocolor image which documents the higher ROS levels in mitochondria (yellow to red) than in other subcellular compartments. b–d Rapid increase in mitochondrial ROS formation in astrocytes exposed to an oxidant, H₂O₂, as visualized using dihydrorhodamine 123; a before exposure to H₂O₂; b at 5 min and c at 10 min following the addition of H₂O₂. In addition to being much brighter, the mitochondria in the H₂O₂-exposed cells are swollen. e, f When melatonin was added to the culture medium simultaneously with H₂O₂, ROS levels in the mitochondria did not increase at either 5 or 10 min, consistent with the ability of melatonin to enter the mitochondria and neutralize the ROS.

Reprinted with permission Jou et al. [95]

The reduction in ROS due to melatonin was further documented using FACS (fluorescence detected by flow cytometry) coupled with the use of ROS sensitive probes [95]. In this case, ROS formation was promoted by the exposure of astrocytes to either H₂O₂ or to short chain oxidants, i.e., tert-butyl hydroperoxide (t-BuOOH) or cumene hydroperoxide (Cu-OOH). Again, using these technologies, melatonin dose-dependently inhibited ROS formation (Fig. 5). Melatonin not only restricted ROS formation in oxidant-treated cells, but also lowered the concentration of ROS in otherwise untreated resting cells. When an equivalent concentration of vitamin E was used as an antioxidant as a replacement for melatonin, it proved far less effective than the indole in reversing the toxic actions of oxidant treatment.

Fig. 5.

Fluorescence detected by flow cytometry (FACS) analysis of ROS generation in astrocytes after exposure to one of several oxidants: H₂O₂, tert-butyl hydroperoxide (t-BuOOH) or cumene hydroperoxide (Cu-OOH). a Illustrates the dose-dependent increase in mitochondrial ROS induced by H₂O₂ (0.1, 1.0 or 10 mM). b The addition of melatonin (100 µM) with H₂O₂ greatly diminished ROS fluorescence. c When 100 µM vitamin E was exchanged for melatonin, it was much less effective than melatonin in reducing mitochondrial ROS. d, e Melatonin also significantly lowered ROS generation in cells exposed to either t-BuOOH or Cu-OOH, respectively. f Melatonin inhibition of ROS-mediated by H₂O₂ exposure.

Reprinted with permission from Jou et al. [95]

As a continuation of these studies, we [95] examined the dysregulation of intracellular calcium as an early indication of opening of the mitochondrial permeability transition pore (MPTP) in cells exposed to an oxidant. Under control conditions, astrocyte mitochondrial Ca²⁺ concentrations were low but increased rapidly after H₂O₂ treatment with opening of the MPTP. The events resulting from mitochondrial Ca²⁺ overload were prevented by melatonin. Likewise, mitochondrial membrane potential depolarization induced by H₂O₂ was suppressed by melatonin leading to a reduction in the escape of lethal cytochrome c (Fig. 6) and caspase-mediated apoptosis.

Fig. 6.

The induction of cellular apoptosis after exposure to an oxidant involves the release of cytochrome c from damaged mitochondria with the subsequent activation of caspases leading to programmed cell death. This figure illustrates the localizations of cytochrome c in the mitochondria of astrocytes not exposed to an oxidant (top); middle after 90 min exposure to H₂O₂; bottom after exposure to H₂O₂ plus melatonin. Clearly, H₂O₂ treatment caused a massive release of cytochrome c into the cytosol and much less into the nucleus with melatonin almost totally preventing this escape. The release of cytochrome c occurred simultaneously with retraction of cell processes, shrinkage of the cells and an irregular plasma membrane. Cytochrome c was detected using immunocytochemistry and laser scanning confocal microscopy. Bar 10 µM.

Reprinted with permission from Jou et al. [95]

This series of investigations illustrate that by intervening at the early phases of the mitochondrial-mediated apoptotic process, melatonin prevents all the downstream events associated with the loss of cells resulting from massive free radical damage due to oxidant exposure. The findings also are consistent with melatonin’s ability to quickly gain access to the mitochondrial matrix where it readily soaks up free radicals.

In the studies summarized and reported by others [88, 96], melatonin also was superior to other antioxidants in preventing oxidative damage normally meted out by toxic ROS. The large scale deletion of 4577 bp from mitochondrial DNA is known as the common deletion (CD); this deletion eliminates roughly one-third of the mitochondrial base pairs and severely damages the efficiency of the respiratory chain. This defect greatly augments mitochondrial ROS generation in cells which leads to an elevated rate of apoptosis. When the CD occurs in humans it is responsible for respiratory chain defect-associated diseases; treatments for these conditions have limited efficacy [97].

Using appropriate fluorescent probes, Jou and colleagues [98] and Peng et al. [99] were able to visualize the increased mitochondrial ROS production in single cells suffering with the CD. Considering the high efficacy of melatonin as an ROS scavenger, we [100] tested how effective the indole would be in protecting cells with CD-augmented mitochondrial oxidative stress and apoptosis and especially from secondary oxidative stress (that which occurs when such already-damaged cells are exposed to H₂O₂). In these situations, melatonin lowered basal as well as secondary oxidative stress. Moreover, melatonin prevented mitochondrial ROS-mediated depolarization of the mitochondrial membrane and counteracted opening of the MPTP. Melatonin also reduced cardiolipin depletion and halted apoptosis. Finally, melatonin suppressed mitochondrial Ca²⁺ dysregulation with this protection exceeding that provided by either vitamin E or synthetic mitochondria-targeted co-enzyme Q, i.e., Mito Q. Based on these observations, Jou and coworkers [100] feel melatonin may prove effective as a therapy in clinical situations that involve mitochondrial malfunction.

A follow-up report revealed that melatonin reduced apoptosis of astrocytes in which mitochondrial Ca²⁺ dysregulation was suppressed. Also under some conditions, melatonin effectively attenuates MPTP opening and apoptosis in the cells where its antioxidative actions are prevented, indicating that melatonin may directly target the MPTP [101].

Melatonin in multiple organs and routes of release

Based on the discovery of melatonin in pineal tissue [19] and the fact that surgical removal of this organ (or its sympathetic denervation or decentralization) reduces circulating levels to near zero [102, 103] and eliminates some of its circadian and circannual functions [20, 104–106], for many years the pineal was considered the exclusive source of melatonin in vertebrates. In all mammalian species, the pineal synthesizes and releases melatonin in a circadian fashion with the highest circulating levels at night and much lower values during the day. The details of the circadian control of melatonin synthesis in the pineal gland have been well defined [18].

While it was initially thought that the primary secretory route of pineal melatonin is into the rich vascular network in the gland, we believe that the major pathway of secretion is in fact directly into the cerebrospinal fluid (CSF) [107]. This is consistent with the very large amplitude nocturnal rise in CSF melatonin relative to the much lower nighttime increase in the blood and also with the much more precise rise and fall of melatonin in the fluid present in the third ventricle of the brain [108]. Thus, the circadian regulation of the central biological clock, i.e., the suprachiasmatic nuclei (SCN), is reliant on the CSF melatonin rhythm rather than the cycle of melatonin in the blood [107, 109]. The blood melatonin rhythm, however, along with the autonomic nervous system is presumably left to cue circadian clock genes that exist in peripheral tissues [110].

We also surmise that another important function of the large nightly increases in CSF melatonin is for its antioxidant protection of the brain [107], which has a high metabolic rate and a very high utilization of O₂ which puts the mitochondria of neurons and glial in excessive oxidative jeopardy [111]. The consequences of the high use of O₂ by the brain, because of the generation of partially reduced derivatives, is seen in many neurodegenerative disorders all of which have a prominent oxidative component [112–114]. Besides melatonin of pineal origin providing neuronal/glial mitochondrial protection from oxidative stress, possibly all cells in the central nervous system generate melatonin for their restricted use in resisting the toxicity of ROS/RNS.

While the derivation of CSF and blood melatonin rhythms are undoubtedly a function of melatonin released from the pineal gland, it is now obvious that melatonin synthesis is not unique to this organ. The first structure, after the pineal, in which melatonin was found to be produced was another neural structure, the retinas [115], where, like the pineal gland, its synthesis is rhythmic. Subsequently, melatonin production has been uncovered in many non-neural tissues.

In some peripheral organs, local melatonin synthesis is inferred because surgical removal of the pineal gland, which depletes circulating melatonin values to barely measureable values, does not decrease values in peripheral cells. For example, in hepatocytes not only does pinealectomy not diminish intracellular concentrations of the indole, but melatonin levels actually rise in some subcellular organelles [50], a presumed compensatory response to the reduction of circulating melatonin. The rise in tissue levels of melatonin after pinealectomy, implies that melatonin synthesis is inducible in animals [53] as it is in plants [13]. Even in the presence of an intact pineal gland, melatonin concentrations in mitochondria and cell membranes of hepatocytes greatly exceed those in the blood [50]. Interestingly, bile, which is produced by hepatocytes, has exceptionally high concentrations of melatonin [116] potentially also originating from the mitochondria of liver-associated cells. The uncommonly high levels of melatonin in bile are predictably for the purpose of protecting the epithelium of the biliary tree from oxidative damage inflicted by highly toxic biliary constituents [117]. The atypically elevated levels of melatonin in the bile may also be augmented due to its recirculation in the enterohepatic circulation. Finally, melatonin produced in the gut microbiome or consumed in the diet is taken up by the capillary bed of the hepatic venous system which may then transfer it to the hepatocyte from where it may be shunted into the bile. Melatonin, which was discovered in land plants in 1995 [118, 119], is in much higher concentrations in plant products that are common in the human and animal diet. Melatonin in plants serves a similar function as in animals, i.e., as an antioxidant [120] and it also functions in growth promotion, not unlike an auxin [121].

The widespread production of melatonin in peripheral organs should not be unexpected considering the evolutionary origin of mitochondria. We recently proposed that the ability of all eukaryotic cells to produce melatonin [12] stems from the likely bacterial origin of mitochondria (and chloroplasts of plants) [122]. According to the endosymbiotic theory, mitochondria and chloroplasts developed from bacteria that were engulfed by ancestral prokaryotic organisms (Fig. 7). Since the devoured bacteria already produced melatonin [15], we speculate that this function was preserved by both the evolving mitochondria and chloroplasts [12]. Since the former exists in most cells of eukaryotes, essentially every cell may have the capability of forming its own melatonin, not for distribution throughout the organism, but use locally in protection against an oxidative challenge (Fig. 8). In cells other than the pinealocytes, the melatonin synthetic pathway in mitochondria and chloroplasts may be functioning at a low level or may be dormant under conditions of minimal oxidative stress, but it is upregulated under circumstances where a compensatory rise in melatonin production would be required to resist an augmented production of ROS/RNS. Such a compensatory stimulation of melatonin formation has already been reported in plants subjected to either abiotic [123] or biotic stress [124] and is consistent with the findings of Venegas and colleagues [50] who reported loss of pineal melatonin is accompanied by a rise in cell membrane melatonin concentrations in cerebrocortical cells and in hepatocytes. Many stresses are known to promote free radical production in cells, e.g., toxic drugs (doxorubicin), chemical toxins (paraquat), heavy metals, ionizing radiation, ultraviolet radiation, hypoxia (stroke/heart attack), hyperoxia, drought, excessive cold or heat, environmental air pollutants (soot/smog), bacterial invasion, excessive exertion, inflammation, allergens, etc. Each of these stresses presumably induces a compensatory rise in intracellular melatonin synthesis in the affected cell for its own protection, i.e., as a firewall against oxidative stress. This is clearly one means among several features by which melatonin differs from other free radical scavengers, i.e., it may be produced at the local level where maximal free radical generation occurs and its synthesis in these cells can be upregulated.

Fig. 7.

The endosymbiotic theory of the origin of mitochondria and chloroplasts. Mitochondria arose from engulfed bacteria that were initially taken in and digested for their nutrients. During evolution, the ingested bacteria developed a symbiotic relationship with the host cell and evolved into mitochondria. Likewise, photosynthetic bacteria were also taken in as food but eventually evolved to form chloroplasts. Since, the ingested bacteria (which formed both mitochondria and chloroplasts) produced melatonin, we proposed this function was retained such that in current day animals and plants, both mitochondria and chloroplasts retain the ability to produce melatonin. Emerging evidence supports this assumption.

Reprinted with permission from Manchester et al. [13]

Fig. 8.

The targeting of melatonin to the mitochondria; evidence suggests that melatonin enters the mitochondria through specific transporters, PETP 1/2 (oligopeptide transporters). The actions of melatonin in mitochondria are multiple. These actions, particularly including its ability to reduce oxidative damage to critical mitochondrial molecules, preserve the function of these organelles and benefit diseases in which mitochondrial malfunction is a feature. Melatonin increases the efficiency of the electron transport chain (I, II, III and IV) and improves ATP production (ATP synthase). Reactive oxygen species (ROS) produced when electrons leak from the ETC are directly scavenged by melatonin and its metabolite [N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK)]. ROS are also metabolized by mitochondria superoxide dismutase (SOD2) and scavenged by glutathione (GSH) and SIRT3. Melatonin also modulates uncoupling protein (UCP2) to maintain an optimal inner mitochondrial membrane potential and prevents opening of the mitochondrial permeability transition pore (MPTP). This limits the escape of cytochrome c when the mitochondrion is damaged by ROS. Recent evidence suggests that melatonin, in addition to quickly entering the mitochondria, may also be synthesized in this organelle (5HT → Mel) where it is also metabolized to AFMK. Not shown in this figure are other actions of melatonin that prevent mitochondrial damage and cell death. These include a reduction in oxidative damage to lipids, proteins and DNA, upregulation of antiapoptotic proteins and downregulation of proapoptotic proteins and a suppression of the activities of caspases which execute the apoptosis pathway

Evidence for the uptake of melatonin and its synthesis in mitochondria

The phagocytosis of melatonin-forming bacteria by primitive eukaryotes (the endosymbiotic theory, Fig. 7) theoretically explains the origin of both mitochondria and chloroplasts that exist in present-day unicellular and multicellular organisms. When these engulfed bacteria evolved into these cellular organelles, they retained their melatonin-forming ability and horizontally transferred this machinery to the eukaryotic cell [15]. This inherited capability has persisted throughout evolution.

The ability of mitochondria and chloroplasts to produce melatonin is currently under intensive investigation. With regard to the production of melatonin by mitochondria, it is known that these organelles contain much higher melatonin concentrations than exist in the circulation [50, 125]. This could mean that mitochondria are capable of concentrating melatonin from the blood against a gradient; however, since pinealectomy rather than depleting mitochondrial melatonin levels actually causes them to rise [50] argues against this possibility. An alternative explanation is that the mitochondria produce their own melatonin.

In the pineal gland, where melatonin is abundantly produced on a nightly basis, the immunocytochemical localization of the AANAT at the ultrastructural level seems to be restricted to the mitochondria [126, 127]. Assuming the validity of this finding, it suggests that the melatonin’s synthetic pathway in pinealocytes is at least partially located in mitochondria. If this is the case for pinealocytes, it may extend to other cells as well.

This indirect evidence is supported by a recent observation related to melatonin synthesis in the mammalian oocyte. These critically important cells have been shown to synthesize melatonin to ensure that they are adequately protected from oxidative damage during their maturation, at ovulation and during their fallopian tube transfer and implantation [128–130]. Based on the data of He and coworkers [52], the mitochondria of oocytes are the major site of melatonin production in these cells. When oocyte mitochondria were cultured in a medium containing the melatonin precursor, serotonin, the melatonin concentration in the medium quickly increased and within 15 min, the values were about 12-fold higher than levels in medium from mitochondria not supplemented with serotonin. Similarly, AANAT (alkylamine-N-acetyl transferase) estimated from immunocytochemical images, was shown to exist in both mitochondria and cytosol of mouse oocytes. The production of melatonin in oocyte mitochondria has implications for the more general implications that mitochondria of other cells also produce this indole. Hopefully, these observations will prompt other investigators to examine melatonin synthesis in mitochondria of somatic cells. For chloroplasts, the data relative to their capacity to produce melatonin are more complete than the data for the synthesis of this indole in mammalian mitochondria [131–133].

While the production of melatonin in mitochondria seems likely and evidence of this has emerged [134, 135], the synthesis of melatonin in the cytosol is not precluded. Certainly, in reference to the pineal gland it has always been assumed that melatonin synthesis is primarily confined to the cytosol [18]. However, an analysis of the kinetics of AANAT as well as substrate availability, i.e., acetyl CoA, we predict that mitochondria more efficiently convert serotonin to melatonin than does the cytosol [136]. This is also consistent with the much higher concentrations of melatonin in this organelle than in the cytosol or in the blood, even after pinealectomy which lowers circulating melatonin values to near zero [50, 137].

The high efficacy of melatonin in limiting mitochondrial damage in diseases/disorders that have a high oxidative component [138–141] would generally require elevated levels of this antioxidant, either as a result of its local synthesis or rapid uptake. Even before melatonin was discovered to be a free radical scavenger, we noted that an intensive mitochondrial oxidation-producing situation, i.e., forced swimming, caused the rapid disappearance of melatonin from the pineal and the blood [142–144]. These studies did not prove that melatonin was rapidly extracted from the blood to be concentrated in the mitochondria of cells experiencing high oxidative stress. However, the very high utilization of O₂ in the mitochondria of the stressed cells to produce ATP is known to be associated with elevated O^·−₂ generation and oxidative damage [145–147]. Thus, the increased concentration of melatonin at these sites would provide the necessary reductive equivalents to combat the extensive oxidative damage that would normally occur under such conditions.

Melatonin uptake into cells and eventually into mitochondria was long assumed to be related to its high lipid solubility which would allow it to rapidly diffuse through lipid-rich plasma and mitochondrial membranes [148, 149]. While this explanation may suffice following the exogenous administration of pharmacological amounts of melatonin where blood levels greatly exceed the normally high concentrations in the mitochondria [50], this would not be adequate to explain melatonin uptake by mitochondria that have levels significantly higher than physiological blood concentrations. The implication is that there may be an active uptake mechanism to concentrate melatonin in mitochondria. We recently suggested that melatonin enters cells through the glucose transporters (GLUT1 and GLUT4) [150]. According to this report, melatonin uptake was unexpectedly slow (one-to-several hours) and would seemingly not be consistent with the rapid disappearance of melatonin from the blood of highly stressed animals (seconds to minutes) [144]. The mechanism identified by Hevia and colleagues [150] does not provide information or the uptake of melatonin into mitochondria although the GLUT1 and GLUT4 transporters are located in the mitochondrial membrane.

In an attempt to resolve the issue as to whether there is a means, other than simple diffusion, by which melatonin may enter the mitochondria, Huo and coworkers [151] considered the possibility that either the human oligopeptide transporters (PEPT) 1/2 or the organic anion transporter (OAT) 3 aided this process. The authors used two human cancer cell lines, PC3 and U118, to perform their study. While the OAT3 transporter was not found to be involved, docking analysis of melatonin with PEPT 1/2 which are located in the mitochondrial membrane, showed that melatonin readily embedded into the active site of the transporters. PEPT 1/2 facilitated the transfer of melatonin into the mitochondria which correlated with the intraorganellar concentration of the indole (Fig. 8). The authors concluded that the oligopeptide transporters PEPT 1/2 play a crucial role in determining the high levels of melatonin in mitochondria. As such, melatonin is in a pivotal position to neutralize radicals that are generated by the less-than-perfectly functioning ETC.

Melatonin receptors/binding sites have been described in the plasma membrane, in the cytosol and in the nucleus of cells [26, 152–154]. Additionally, there is a single report claiming that a melatonin receptor also exists in the mitochondrial membrane. This report, published by Wang et al. [155], used cells from an animal model of Huntington disease. If this receptor does exist in the mitochondrial membrane, it could help to explain the high efficiency of melatonin in protecting this organelle from oxidative damage, e.g., by mediating the upregulation of mitochondrial antioxidative enzymes. The presence of this receptor, which has yet to be confirmed, would seemingly have no impact in determining the intramitochondrial concentration of melatonin, but could be related to the indirect antioxidative processes of melatonin in mitochondria, e.g., stimulating antioxidant enzyme activities, enhancing SIRT3 activity, etc.

Diseases where mitochondrial dysfunction occurs and where melatonin is beneficial

Sepsis, severe sepsis, and septic shock are serious medical conditions that are associated with extensive oxidative destruction of key molecules within cells. When several essential organs are damaged to the extent that they functionally fail, the condition is referred to as septic shock and is accompanied by multiple organ failure with a high degree of mortality being the result. A major predicted causative factor for multiple organ failure/septic shock is mitochondrial dysfunction resulting from damage inflicted by locally produced toxic reactive oxygen and nitrogen species [156–158]. In both humans [157, 159] and animals [160, 161] severe mitochondrial malfunction is central in this condition.

In 2001, a report was published by Gitto and colleagues [162] in which melatonin was used to treat sepsis in humans. In that report, we observed that giving septic human neonates intravenously administered melatonin both reduced the degree of oxidative damage (reduced levels of lipid peroxidation products in blood) and significantly limited the rate of mortality in these newborns. Subsequently, there have been numerous similar studies published using experimental animals with all the data pointing to the ability of melatonin to overcome many of the negative molecular consequences of sepsis including preventing the death of animals [163–165], as was shown by Gitto et al. [162] in humans. Since the mitochondria are a central organelle for the site of oxidative damage resulting from the bacterial toxins associated with septic infections, it would be expected that antioxidant supplementation may be beneficial in resisting molecular damage in this critical illness in animal and human sepsis [166, 167].

Perhaps related to the very high degree of damage that occurs or due to the limited ability of classic antioxidants to gain access to mitochondria, conventional free radical scavengers provide little protection, even when they are given in very high doses, against septic infections [168–171]. To overcome this deficiency, a new strategy has been to design antioxidants that target the mitochondria and concentrate in these organelles thereby increasing their effectiveness as a treatment. This therapy would require that the antioxidant be (a) delivered especially to the mitochondria, (b) be located at the proper site in the mitochondria to best scavenge newly formed radicals, and (c) to stimulate other processes that reduce local free radical levels. Examples of this latter function would be the augmentation of the activity of antioxidative enzymes and the reduction of the number of free radicals formed, i.e., radical avoidance [172]. Finally, (d) the optimal antioxidant should also have significant anti-inflammatory actions to quell the cytokine storm that is elicited during a septic event since many cytokines promote free radical generation [153, 173, 174].

The most common means to enhance the cellular and organellar uptake of conventional antioxidants has been to couple them with a lipophilic cation which allows them to more easily permeate lipid bilayers and concentrate in mitochondria. These so-called mitochondria-targeted antioxidants concentrate in the mitochondria up to 500-fold. The triphenylphosphonium cation is most frequently conjugated to an antioxidant to improve its uptake (Fig. 9) [175, 176]. The best known mitochondria-targeted antioxidants are MitoQ (based on the antioxidant co-enzyme Q10 [177]) and MitoE (based on α-tocopherol [178]).

Fig. 9.

The structure of synthetically produced, mitochondria-targeted antioxidants, i.e., MitoE and MitoQ. When vitamin E or co-enzyme Q10 is coupled to the triphosphonium cation, they more readily accumulate in the cytosol and in the mitochondria due to their increased lipid solubility. With regard to mitochondria, these synthetically produced antioxidants accumulate in concentrations of 200-–500-fold greater than unconjugated vitamin E and co-enzyme Q10. Despite this high concentration, when compared under in vivo experimental conditions, melatonin at equimolar concentrations was as good as or better than the fabricated antioxidants in protecting against cellular oxidative stress. For this and other reasons, we consider melatonin to be a mitochondria-targeted antioxidant. IMM inner mitochondrial membrane, OMM outer mitochondrial membrane

Using a sepsis-mediated organ failure model, Lowes et al. [51] compared the ability of melatonin to that of equimolar concentrations of either MitoQ and MitoE in resisting oxidative damage and the inflammatory response of rats given two bacterial endotoxins, lipopolysaccharide and peptidoglycan. Each of the antioxidants was given as a bolus injection followed by a 5-h infusion period. While each of the antioxidants had beneficial actions, melatonin was superior in terms of reducing plasma levels of lipid peroxidation products and hepatic protein damage, in restoring organ dysfunction (estimated using plasma aspartate transaminase and urinary creatinine levels) and in improving mitochondrial oxidative phosphorylation. Of the agents tested, Lowes et al. [51] concluded that melatonin would be the most effective treatment to counteract endotoxin toxicity, i.e., in humans. The findings also speak to the fact that melatonin readily accumulates in the mitochondria, i.e., it is likely a mitochondria-targeted antioxidant. Certainly, the very extensive work of the group of Acuna-Castroviejo [179] is consistent with the high efficacy of melatonin in reducing the consequences of sepsis at the mitochondrial level.

Considering the obviously high production of ROS in mitochondria, it seems obvious that a free radical scavenger specifically designed to localize in mitochondria would be highly beneficial in reducing the total oxidative burden that cells sustain. Free radical scavengers and enzymatic antioxidants are essential in maintaining redox homeostasis so as to reduce excessive damage to critical molecules.

While the idea of designing mitochondria-targeted antioxidants has been around for at least two decades [168, 170, 171, 180] and since melatonin is highly effective in reducing oxidative stress at the mitochondrial level [88, 91, 92, 163, 166, 181–184], we recently proposed that in fact melatonin is a mitochondria-targeted antioxidant [185]. The rationale for this classification was further elaborated in two subsequent publications [77, 167].

An example of a neurodegenerative condition that is clearly linked to aberrant mitochondrial structure and function is multiple sclerosis (MS) [186]; in fact, this progressive disease is believed to be primarily a mitochondria-related condition [187]. In an experimental model of this disease, Kashani and colleagues [188] reported that the use of melatonin as a treatment of MS-like pathology in mice was followed by an obvious improvement of mitochondrial function and a reduced disease progression. More recently, we [189] reported on an MS patient who was treated with 5–300 mg melatonin daily for 4 years. Melatonin treatment was initiated after the patient had failed to respond to glucocorticoid medications and was judged to be at the Expanded Disability Status Scale (EDSS) 8.0 of the disease (patient restricted to bed or wheelchair) [189]. During the use of melatonin daily for 4 years, the patient showed steady improvement and was diagnosed as being at EDSS 6.0 (walks with cane, crutch or brace up to 100 m without resting). While the symptoms of this disease spontaneously wax and wane in the short term, we feel that melatonin definitely reduced the severity of this condition because of the degree of improvement and the duration of the beneficial effects (4 years). Given that MS is generally considered a condition in which mitochondrial physiology is compromised, coupled with the known action of melatonin at the mitochondrial level, the findings suggest a more comprehensive evaluation of melatonin in MS patients should be considered.

A physiological deficiency of a number of progressive neurodegenerative conditions also involves mitochondrial malfunction as a primary or secondary disturbance in neurons or glia, e.g., Alzheimer, Parkinson, and Huntington disease [190–192]. In addition to a genetic predisposition, the onset of these conditions typically occurs when endogenous melatonin levels have deteriorated to chronically low values. Moreover, melatonin treatment of models of these diseases in experimental animals have often shown benefits in terms of prevention or slowed progression of both the neuropathology and behavior. Based on the outcomes reported, melatonin supplementation would likely be more useful in the prevention/slowing of these diseases rather than as a treatment of these conditions after they are at an advanced stage. Finally, a patient suffering with Duchenne muscular dystrophy, a condition that has a major mitochondrial dysfunction component, also responded to melatonin treatment with a substantial slowing of the disease [193].

The findings summarized above represent only a small percentage of the studies that have described an improvement of mitochondrial function when they are exposed to melatonin. Collectively, the data from both experimental and clinical publications, which are numerous, attest to melatonin having specific beneficial effects at the mitochondrial level.

Concluding remarks

Melatonin seems to meet the criteria as a mitochondria-targeted antioxidant. Abundant immunocytochemical evidence supports the conclusion that melatonin has ready access to the intermembrane space and matrix of mitochondria (Fig. 8). Certainly, this ubiquitously acting antioxidant, particularly in situations where mitochondrial ROS production is exaggerated, has the capability of reducing intramitochondrial free radical levels, visualized using appropriate fluorescent probes, as well as minimizing the potential molecular damage that is a normal consequence of enhanced ROS generation. Moreover, melatonin’s ability to achieve these healthful actions may be aided by its local production in the mitochondria.

We have surmised that mitochondria originated in early prokaryotes when these cells engulfed melatonin-producing bacteria as a source of nutrients. Eventually, the engulfed bacteria established a symbiotic relationship with these cells and evolved into mitochondria; when they did so they retained the melatonin-producing ability that their precursors possessed.

The likelihood of mitochondria being a source of melatonin is supported by other data as well. Thus, mitochondria have melatonin levels that greatly exceed those in the blood and, additionally, surgical removal of the pineal gland which lowers circulating melatonin to near zero and deprives cells of exogenously produced melatonin, does not cause a concomitant drop in mitochondrial melatonin concentrations. In fact, the loss of exogenously available melatonin may cause a compensatory rise in melatonin production. Based on kinetic considerations and substrate availability, melatonin production is calculated to more likely take place in the mitochondria rather than in the cytosol. Finally, incubation of isolated mitochondria with serotonin, a necessary precursor of melatonin, is followed by a large rise in melatonin concentrations in this organelle.

The number of reports proving that melatonin reduces molecular damage under severely enhanced oxidative experimental and clinical situations has accumulated for two decades and are now numerous. A plethora of toxins and disease models which cause or are accompanied by greatly exaggerated free radical generation have been used to challenge the ability of melatonin to prevent or forestall the massive damage that would normally occur. In these situations, melatonin has never failed to be protective.

Given the numerous findings that document the highly significant protective actions of melatonin in very high oxidative stress conditions, this endogenously produced indoleamine has proven itself as an essential and worthy “firewall” against toxic free radicals. Considering its high efficacy along with its mitochondria targeting, melatonin should be more extensively exploited for its mitochondrial protective actions and its antioxidant potential, particularly at the clinical level.