Another class of therapeutics with metal-dependent antioxidant activity are metal chelators. The BBB-permeable (Fredenburg et al., 1996) iron siderophore, deferiprone (DFP), has been investigated as a therapeutic for FRDA in multiple cell and animal models, as well as two clinical trials. DFP treatment of HEK-293 cells stably transfected with a tetracycline-inducible frataxin shRNA, resulted in restoration of mitochondrial functions including decreased mitochondrial redox potential, ROS production and labile iron pools (Kakhlon et al., 2008). DFP also increased survival and improved motor climbing ability in a frataxin-knockdown Drosophila FRDA model (Soriano et al., 2013).
A 6 month clinical trial involving daily DFP administration in young FRDA patients at 20–30 mg·kg−1 improved gait and other neurological signs in all nine patients, and induced a significant progressive reduction in dentate nuclei-localized iron accumulation in eight patients (Boddaert et al., 2007). However, a subsequent 11-month study of combination therapy with DFP and idebenone (IDE), a synthetic analogue of the antioxidant coenzyme Q10, on a cohort of 20 FRDA patients did not show an overall significant improvement in ataxia neurological function scores (Velasco-Sanchez et al., 2011). Nonetheless, as neurological scores are expected to decline with disease progression over the long treatment period (Velasco-Sanchez et al., 2011), it could be inferred that the trial was moderately successful. Moreover, a significant reduction in serum ferritin concomitant with increased transferrin was observed in DFP and IDE co-treated patients (Velasco-Sanchez et al., 2011), suggesting that restoration of iron homeostasis could be the mechanism involved. However, care must be taken with administration of non-selective iron chelators, as these can often have toxic consequences. For example, deferasirox (DFX) is highly toxic in people without iron overload (Singh et al., 1995), and oral administration of dipyridyl is toxic to rats (Groce and Kimbrough, 1982).
Interestingly, the iron affinity of the chelators, DFP, DFX and alicylaldehyde isonicotinoyl hydrazone, was inversely correlated with their ability to improve mitochondrial function in frataxin knockdown cells (Kakhlon et al., 2008). It has been hypothesized that the reason DFP is more effective than other iron chelators (Kakhlon et al., 2008) is its ability to act as a frataxin mimetic ionophore, with functions both in chelation of excess iron and redistribution to iron-deficient subcellular sites (Goncalves et al., 2008). Consistent with this, DFP shuttled iron from iron donor to iron acceptor probes in multiple cellular compartments, including extracellular media, nuclei, mitochondria and endosomes (Sohn et al., 2008). This indicates that DFP is capable of iron mobilization from accumulated stores to sites of low iron bioavailability and therefore acts as a membrane permeable multidirectional iron transporter. Indeed, DFP increased soluble mitochondrial iron and prevented aggregation of insoluble Fe3+ in a frataxin-knockdown Drosophila model (Soriano et al., 2013). However, caution must be exercised with the use of DFP, as excessive depletion of labile mitochondrial iron pools may promote adverse effects including loss of mitochondrial aconitase activity and growth inhibition, as reported with high dose treatments in control and FRDA patient fibroblasts (Goncalves et al., 2008).
Natural plant-based polyphenols
The neuroprotective role of various natural plant-based polyphenolic compounds has been examined in numerous cell and animal models of neurodegeneration. These include baicalein (from skullcap flowers), tocotrienol (from palm oil), epigallocatechin gallate (EGCG; from green tea), rosmarinic acid (from rosemary), tannic acid (from various plant sources), nordihydroguaiaretic acid (from the creosote bush), curcumin (from turmeric), myricetin (from numerous sources including berries, nuts, fruit, vegetables) and kaempferol (from various sources including tea, berries, fruit and vegetables). These compounds are potent metal chelators and free radical scavengers and have been reported to restore multiple parameters of mitochondrial function including mitochondrial lipid peroxidation, membrane potential, mitochondrial swelling and cytochrome c release (Kamat and Devasagayam, 1995; Mandel et al., 2005; Ono and Yamada, 2006; Cheng et al., 2008; He et al., 2009; Mu et al., 2009; Caruana et al., 2011). This review will focus on two of these: the green tea catechin, EGCG and curcumin.
Green tea catechins
Green tea, which has long been recognized for numerous health benefits, contains a number of polyphenolic flavonoids called catechins, including EGCG, epigallocatechin (EGC), (–)-epicatechin (EC) and (–)-epicatechin-3-gallate (ECG), in order of abundance. These natural compounds are strong antioxidants and have been reported to chelate iron with a similar efficacy as DFP (for a detailed review, see Mandel et al., 2005). Green tea consumption has been linked to increased cognitive performance and is reported to significantly lower the risk of developing PD (Kuriyama et al., 2006; Hu et al., 2007). ECGG is implicated in hippocampal neuroprotection in a gerbil model of ischaemia/reperfusion injury by carotid artery occlusion (Lee et al., 2000), is shown to prevent 6-OHDA-induced neurotoxicity in vitro (Levites et al., 2002) and dopaminergic neuron loss in the SN in the MPP+-induced mouse model of PD (Levites et al., 2001). Moreover, EGCG can both attenuate fibrillar α-synuclein aggregation and induce dissociation of formed fibrils (Ono and Yamada, 2006). Similar to the other antioxidant metal-modulating agents described above, catechins exhibit a biphasic action, where lower concentrations mediate neuroprotection and higher doses induce apoptosis, suggestive of promising therapeutic potential for both neurodegeneration and cancer (Weinreb et al., 2003).
Of the green tea polyphenols, EGCG displays the most robust neuroprotective activity, and hence is the best studied of the catechins (Nie et al., 2002). The molecular mechanism of action of EGCG involves direct antioxidant activity through free radical scavenging (Salah et al., 1995; Nanjo et al., 1996), as well as through potent induction of endogenous antioxidant systems such SOD1 and catalase (Levites et al., 2001). In a study examining the free radical scavenging capacity of EGCG and closely related molecules, it was determined that the ortho-trihydroxyl group on the aromatic ring and the galloyl substituent are primarily responsible for the potent redox properties of EGCG (Nanjo et al., 1996). In the mitochondria, EGCG inhibited 6-OHDA-mediated mitochondrial membrane depolarization and prevented impaired mitochondrial complex I-V function following a hypoxic insult in rats (Sutherland et al., 2005). Moreover, EGCG preferentially targets the mitochondria and displays specific neuroprotective activity to mitochondrial toxic stimuli, but not other apoptosis-inducing agents in rat cerebellar neurons (Schroeder et al., 2009).
In addition to direct effects on mitochondrial function, EGCG activates neuroprotective cellular signalling pathways including the antioxidant response through nrf2-dependent transcription of haeme oxygenase-1 (Romeo et al., 2009). Moreover, in neuronal SH-SY5Y cells, EGCG stimulated the PKC phospho cascades, and attenuated the expression of the pro-apoptotic genes activated by 6-OHDA (Levites et al., 2002). As described for other metal-modulating compounds, EGCG has also been reported to induce ERK activation (Levites et al., 2002; Schroeter et al., 2007), which may contribute to neuroprotection.
Curcumin, a natural plant polyphenol derived from the spice, turmeric, has been shown to exert potent anti-cancer activity. Curcumin is able to bind metal ions and act as a metal ionophore (Dairam et al., 2008; Garcia et al., 2012). Curcumin was demonstrated to exert pleiotropic cellular effects, which include antioxidant activity and direct detoxification of ROS and peroxynitrite (Iwunze and McEwan, 2004). Hence, the neuroprotective efficacy of curcumin has been assessed in various diseases of mitochondrial metal dysfunction including AD, PD and stroke models. For a detailed review on mitochondrial dysfunction and curcumin treatment in AD models see Eckert et al. (2012). Although curcumin is able to cross the BBB, its rapid elimination from the brain has resulted in several drug design approaches to improve bioavailability. Recently, modification of natural compounds such as curcumin by combining individual neuroprotective functionalities into multi-target complexes based on two scaffolds is gaining increasing support. Diester coupling of amino acid moieties has been used as a delivery platform to improve bioavailability, as amino acid transport systems facilitate prodrug influx and intracellular esterase-mediated reactions liberate functional curcumin molecules (Dubey et al., 2008). Naturally, curcumin derivatives display altered metal-binding kinetics, largely dependent on the nature of the aromatic ring substituents (Ferrari et al., 2013).
Knowledge regarding the cellular localization of curcumin is limited. One study demonstrated that curcumin accumulates primarily in the membrane with some nuclear localization (Kunwar et al., 2008), but more research is required to determine the intracellular targeting of curcumin and its derivatives in brain cells. Curcumin is a potent iron chelator (Jiao et al., 2006; 2009) and induces up-regulation of iron regulatory proteins and the transferrin receptor (Jiao et al., 2006), indicative of systemic iron deficiency. Curcumin has been shown to bind Fe2+ ions, which prevents iron-mediated hydroxyl radical production (Dairam et al., 2008). Curcumin also strongly binds and stabilizes Fe3+ (Dairam et al., 2008; Garcia et al., 2012), thereby inhibiting iron redox cycling.
Curcumin and its derivatives have been tested in numerous cell and animal models of PD. Treatment of N27 dopaminergic neurons with a glutamic acid substituted curcumin derivative-induced neuroprotection involving increased GSH synthesis and a reduction in oxidative stress, lipid peroxidation and H2O2 production (Harish et al., 2010). N27 neurons were also protected from MPP+-induced cytotoxicity by Di-glutamoyl curcumin (Mythri et al., 2011). Furthermore, curcumin improved survival of dopaminergic neurons in the SN in MPTP-lesioned C57BL/6 mice via JNK-dependent inhibition of mitochondrial swelling and cytochrome c release (Pan et al., 2012). Curcumin-fed rats that had been 6-OHDA-lesioned displayed elevated striatal dopamine levels and had a decreased number of iron-positive cells in the SN, suggesting that curcumin-mediated iron chelation may be responsible for protection of dopaminergic neurons from Fenton chemistry-dependent neurodegeneration (Du et al., 2012). Curcumin has also been reported to reduce huntingtin accumulation and improve rearing behaviour in HD CAG140 knock-in mice (Hickey et al., 2012).
Similar to CuII(atsm), the cellular mechanism of action of curcumin involves activation of kinase signalling pathways. Pharmacological inhibition of the Akt, but not MAPK, pathways decreased curcumin-stimulated Nrf2 activation and Nrf2-dependent gene expression in primary rat cultured cortical neurons (Wu et al., 2013). Moreover, blocking Akt prevented curcumin-mediated reduction in infarct size and oxidative stress in a rat transient middle cerebral artery occlusion stroke model (Wu et al., 2013). Nrf2 activation was also reported to contribute to curcumin-mediated amelioration of mitochondrial dysfunction in spinal cord astrocytes treated with H2O2 (Jiang et al., 2011).
As mentioned earlier, limitations with bioavailability of curcumin have been overcome by the development of multifunctional drug candidates including CNB-001, which combines the neuroprotective actions of curcumin and cyclohexyl bisphenol A, and was detected in the mouse brain 6 days after a single oral dose (Liu et al., 2008). CNB-001, but not curcumin, was reported to protect PC12 cells from excitotoxicity and glucose starvation, despite a 20-fold lower antioxidant activity (Liu et al., 2008). CNB-001 also restored ATP levels and protected HT22 cells from glutamate toxicity (Lapchak et al., 2011). The in vivo action of CNB-001 in a rabbit small clot embolic stroke model was also associated with induction of the neuroprotective BDNF, PI3K, Akt and calcium-calmodulin-dependent kinase signalling pathways (Lapchak et al., 2011).
Curcumin is widely reported to reduce cellular ROS production (Wang et al., 2010; Jiang et al., 2011; Liu et al., 2011) and prevent toxicity associated with pathological protein aggregation, including both the PD-associated α-synuclein and ALS-associated TDP-43 aggregates (Wang et al., 2010; Lu et al., 2012). Moreover, curcumin and its derivatives exhibit numerous protective effects on mitochondria. Curcumin improved multiple indicators of mitochondrial function, including cytochrome c release, mitochondrial membrane depolarization and caspase activation in a PC12 neuroblastoma cell model of A53T α-synuclein toxicity (Liu et al., 2011). Mitochondrial dysfunction was also attenuated in a model of aluminium-induced cytotoxicity, as demonstrated by enhanced activity of the mitochondrial NADH dehydrogenase, succinic dehydrogenase and cytochrome c oxidase complexes (Sood et al., 2011). Addition of various substituents on the aromatic rings appears to yield protection against mitochondrial dysfunction. Indeed, mitochondrial swelling, membrane potential and NADH dehydrogenase activity of mice isolated brain mitochondria were substantially improved by a glutamoyl diester of curcumin, as compared with curcumin alone (Mythri et al., 2011). Similarly, a dimethoxy curcumin derivative also attenuated TDP-43-induced loss of mitochondrial membrane potential and mitochondrial NADH dehydrogenase activity (Lu et al., 2012).