Introduction

Parkinson’s disease (PD) is a progressive movement disorder characterized pathologically by the relatively selective degeneration of midbrain dopamine (DA) neurons and the presence of prominent cytoplasmic inclusions, termed Lewy bodies [33]. The percentage of affected individuals within a population rises from ∼1% at 65 years to ∼5% at 85 years, making age the main risk factor for PD. The majority of cases are thought to be idiopathic. However, in ∼5–10% of cases, PD is thought to have a genetic component, showing both recessive and dominant modes of inheritance [129]. In these cases, the causative genes that have been identified are: α-synuclein, parkin, nuclear receptor-related 1 (Nurr1), ubiquitin carboxy-terminal hydrolase L1 (UCHL1), PTEN-induced putative kinase 1(PINK1), leucine-rich repeat kinase 2 (LRRK2) and DJ-1 [14, 108].

Although the mechanisms responsible for neurodegeneration in PD are largely unknown, they result in damage and subsequent loss of DA neurons [82]. Both environmental and genetic factors have been implicated in the pathogenesis of PD [103].

While mitochondrial dysfunction has been indirectly linked to idiopathic PD [33, 110], studies of families with rare inherited forms of PD have identified genes involved in regulating mitochondrial function [27, 94, 102]. Recently, clear evidence has been presented for a high burden of mtDNA deletions within the substantia nigra neurons in individuals with PD [9, 67]. Additional support comes from experimental studies with toxins that inhibit Complex I of the mitochondrial respiratory chain and cause selective death of midbrain DA neurons [109]. However, the interpretation of results from experiments with neurotoxins is complicated by the fact that they may have pleiotropic pharmacological effects in DA neurons, effects on non-DA cell types, or both [109]. Although the roles of mitochondrial dysfunction and mtDNA mutations in the pathogenesis of PD remain controversial, parkinsonism has been associated with several mtDNA mutations, including large-scale rearrangements [15, 20, 106] and point mutations or microdeletions [34, 110, 120].

The most compelling evidence yet comes from MitoPark mice. These are mice possessing conditional knockout of mitochondrial transcription factor A (Tfam) in DA neurons. These mice have reduced mtDNA expression with subsequent respiratory chain deficiency in midbrain DA neurons. This in turn leads to a parkinsonism phenotype characterized by adult onset of a progressive impairment of motor function accompanied by formation of intraneuronal inclusions and DA nerve cell death [37]. Further support for mitochondrial involvement in the pathogenesis of PD is supported by post-mortem biochemical studies which show a disease specific and drug independent defect of Complex I in substantia nigra and other tissues in persons with idiopathic PD [95, 99].

Somatic mtDNA mutations and Parkinson’s disease

Recent research has shown that the substantia nigra neurons from individuals with PD have a high level of deleted mtDNA compared to controls [110]. Kraytsberg et al. [67] using a novel single-molecule PCR approach, have quantified the total burden of mtDNA deletions in aged human substantia nigra neurons compared to neurons from younger individuals as controls. They found high levels of mtDNA deletions in the aged neurons. They also found that affected substantia nigra neurons also lost cytochrome c oxidase (COX) expression. In a parallel study, Bender et al. [9] using long range PCR in single cells confirmed the presence of somatic and clonally expanded mtDNA deletions that were associated with respiratory chain deficiency in individuals with PD. These studies suggest direct involvement of mtDNA deletions in the development of respiratory chain deficiency in substantia nigra neurons in individuals with PD. Another study, however, has shown that somatic mtDNA mutations are capable of inducing aging phenotypes without affecting reactive oxygen species (ROS) production suggesting that oxidative stress is not a key modulator of neurodegeneration in aging [121].

Environmental factors and mitochondrial dysfunction in PD

The discovery in 1983 of persons developing typical signs of PD after intravenous injection of drugs contaminated with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), and the subsequent finding that MPTP selectively damages DA cells in the substantia nigra [72], led to the hypothesis that exposure to environmental toxins might be related to the risk of PD. Complex I is inhibited in dopamine-containing neurons by systemic administration of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), which can induce parkinsonism in animal models and in humans [7]. The defect appears to be restricted both to Complex I and the substantia nigra, with other brain areas showing normal activity of the ETC [29]. Studies on platelets of PD patients show a consistent Complex I defect either alone or in combination with a mild defect in other Complexes [95]. The question then arises as to the origin of the Complex I-deficiency in PD. It could result from an environmental toxin or an acquired or inherited mtDNA mutation(s). To define the origin of oxidative phosphorylation deficit in PD, cybrid PD models have been developed. PD platelet mitochondrial genes expressed in cybrids produce reduced Complex I activity [48, 116], suggesting that the Complex I deficiency was determined by the mtDNA derived from PD patient platelets. Since then, many epidemiological studies have been done to examine the association between exposure to pesticides and herbicides, as well as hypothesized surrogate measures, such as farming, living in rural areas, and drinking of well water, and the risk of PD. Evidence fairly consistently points towards a positive association between pesticide exposure and PD risk, although results were statistically significant in only half of the studies [35]. Environmental neurotoxins such as rotenone, paraquart, maneb, and dieldrin are also inhibitors of Complex I and able to induce dopaminergic loss [98].

Genetic factors and mitochondrial dysfunction in PD

Many of the genes associated with familial forms of PD also implicate mitochondria in disease pathogenesis. At least nine nuclear genes have been identified as causing PD or affecting PD risk. Of the nuclear genes, α-synuclein, parkin, DJ-1, tensin homologue (PTEN)-induced kinase 1 (PINK1), leucine-rich-repeat kinase 2 (LRRK2) and high temperature requirement A2 (HTRA2) directly or indirectly impact mitochondrial function [8, 75]. In transgenic mice, α-synuclein overexpression impairs mitochondrial function, increases oxidative stress and enhances the toxicity of MPTP [111]. Parkin is an ubiquitin E3 ligase that can associate with the outer mitochondrial membrane and protects against cytochrome c release and caspase activation [32]; it may also associate with Tfam and enhance mitochondrial biogenesis [71]. When oxidized, DJ-1 translocates to mitochondria (intermembrane space and matrix), downregulates the PTEN-tumour suppressor protein, and protects the cell from oxidative-stress-induced cell death [64]. PINK-1 is a nuclear-encoded kinase localized to the mitochondrial matrix [107] that seems to protect against apoptosis, an effect that is reduced by PD-related mutations or kinase inactivation [97]. About 10% of the kinase LRRK2 is localized to mitochondria, and PD-related mutations augment its kinase activity [128]. Mutations in other nuclear-encoded genes such as HTRA2, UCH-L1, synphilin-1, glucocerebrosidase, POLG, NR4A2 and tau, have been identified in individual patients with idiopathic PD [1, 87]. HTRA2, a serine protease localized to the mitochondrial intermembrane space, degrades denatured proteins within mitochondria and, if released into the cytosol, promotes programmed cell death by binding inhibitor of apoptosis proteins [113, 115]. Table 1 provides a general overview of how the majority of these gene products directly or indirectly implicate mitochondria in the disease pathogenesis. This suggests that mitochondrial dysfunction may be central to the molecular pathogenesis of both idiopathic and familial forms of PD.

Table 1 Summary of the genetic causes of Parkinson’s disease and their association with mitochondria
Full size table

Oxidative stress and PD

Oxidative stress has been demonstrated in PD [51, 126] and evidence also clearly supports the involvement of impaired mitochondrial function in PD [63, 100]. In particular, increased iron, oxidation of proteins, and lipid peroxidation in the SN appear to be common [2, 36, 132]. Reduced levels of glutathione (GSH) in the midbrain may be indicative of increased free radical levels [105]. Dopamine metabolism can also be a source of ROS in nigral neurons. Cytosolic dopamine produces electrophilic semiquinones and quinones which themselves act as oxidants by supporting ROS formation [114]. The possible involvement of oxidative stress as an etiological factor of PD is further supported by studies with specific neurotoxins that are potent inducers of Parkinsonism in humans and animals. MPTP treatment in mice generates hydroxyl radicals in the striatum which in turn leads to oxidant damage and could cause initiation of apoptosis [23]. Activated glial cells were observed at sites of neurodegeneration in PD, and might participate in the mechanism of nerve cell death by producing more ROS and reactive nitrogen species (RNS). Antioxidants have been proposed as a means to accomplish neuroprotection in PD [17]. It has also been recognized that oxidative stress leading to caspase activation and consequent apoptosis are clearly evident in PD [42].

Animal models

Several animal models of PD, such as the rotenone [11, 43] Drosophila DJ-1 mutants, [84, 85], and 6-hydroxydopamine (6-OHDA) [26] and MPTP [31], have been shown to have multiple mitochondrial dysfunctions including increased ROS generation and striking sensitivity to stressors. Of these, the mammalian models have activated migroglia.

Already substantial and still accumulating evidence shows that lipopolysaccharide (LPS)-induced microglial activation causes DA neurodegeneration in vitro and in vivo [4, 45, 57, 79]. Also, inducible nitric oxide synthase (iNOS) is generally found to be upregulated in experimental PD models [60, 74] and inhibition of iNOS reduces the toxicity of LPS or LPS and interferon γ (IFN- γ-activated microglia on DA neurons in vitro by approximately 75%) [56, 73]. Evidence now clearly indicates that inflammatory cytokines, such as Tumor necrosis factor a (TNF-α), interleukin-1 and 6 (IL-1 and -6)and the signaling molecule NO are toxic to neurons [3, 41, 45, 77, 81, 112]. Inactivation of the genes involved in the synthesis of proinflammatory molecules such as COX-2 [40], nicotinamide adenine dinuleotide phosphate (NADPH) oxidase [130] and both TNF-α receptors [112] were shown to protect DA neurons against MPTP-induced neurotoxicity, thus indicating that inflammation plays an important role in MPTP-mediated as well as other types of nigrostriatal neurodegeneration. Interestingly, it has been observed that MPTP and LPS act synergistically to mediate nigral DA neurotoxicity, probably by stimulating release of the superoxide radical [44].

Cell models of PD

It is difficulty to assess the rate of neuronal loss in PD since it is very slow. In vitro studies, in spite of their limitations, are therefore a better alternative to explore degenerative pathways involved in PD. Most in vitro studies have been conducted in SHSY5Y, MN9D, NB41, and PC12 cell lines, in primary mesencephalic cultures or in cerebellar granule cells.

Cybrids bearing mitochondrial DNA from patients with idiopathic PD produce Lewy body inclusions [122]. They have decreased Complex I activity that is associated with increased apoptosis via the p38 and JNK pathways as compared to age and sex matched control cybrids [93]. They are also more susceptible to oxidative stress [92]. This suggests that mtDNA, if not causative, modulates susceptibility to PD [123].

MPP+ was found to cause apoptotic cell death in PC12 and SH-SY5Y cells and primary midbrain (mesencephalic) cell cultures through the activation of caspase-3 [39, 53], treatment of SH-SY5Y cells with MPP+ induced ROS production, lactate release, inhibition of ETC, p53 expression, cleavage of caspase-3 and PARP, and apoptotic cell death with DNA fragmentation. bcl-2 over expression protected these cells against MPP+ toxicity whereas decreased bcl-2 levels enhanced MPP+-induced cell death [39, 66]. It has now been shown that upon in vivo inhibition of Complex I, p53 mediates Bax transcription and translocation to mitochondria [96].

6-OHDA induces an early increase in p53 cellular content in PC12 cultures [13]. Activation of caspase-3 and caspase-9 was also demonstrated in 6-OHDA induced apoptosis in SHSY5Y cells [28], which was inhibited by caspase inhibitor [117].

Chronic low-grade Complex I inhibition by rotenone exposure induces accumulation and aggregation of α-synuclein and ubiquitin, as well as progressive oxidative stress and caspase-dependent apoptotic death in human neuroblastoma cells [104]. In PC 12 cells inhibition of Complex I by rotenone caused apoptotic cell death [52]. They found both rotenone and MPTP induced apoptosis at low concentrations and necrosis at high concentrations.

Dopamine induced apoptosis in SH-SY5Y neuroblastoma cells through activation of caspase-9 and caspase-3 and cleavage of PARP. In this model nuclear condensation was mediated by the activation of p38 mitogen activated protein (MAP) kinase and mitochondrial cytochrome-c release [62]. These studies suggest a role for apoptosis in PD neurodegeneration although the signaling events involved are yet to be clearly elucidated. Unfortunately, the strategy of inhibiting apoptosis by preventing caspase activation has failed clinically [12].

Mitogen activated signaling cascades and mitochondria in PD

Kinase signaling pathways impact major mitochondrial functions including oxidative phosphorylation, antioxidant protein expression, mitochondrial fission, and execution of survival-death decisions. Extracellular signal regulated kinase (ERK) has also been implicated in promoting oxidative neuronal injury [24] in PD [47, 68, 69]. Phospho-ERK was found at high labeling densities within a subset of mitochondria in degenerating neurons from patients with Parkinson’s disease and Lewy body dementia [25, 134], corresponding to a distinct granular cytoplasmic pattern of staining not observed in age-matched control patients [135]. It is interesting to note that mitochondrial ROS and the permeability transition have both been implicated upstream of JNK activation under both physiologic and pathologic conditions [18, 89]. In a similar manner, p38 MAPK is activated by mitochondrially derived ROS [70, 92]. Thus, these kinases, like ERK, may signal to influence mitochondrial functions as well as to communicate mitochondrial signals to the rest of the cell.

The proteasome system and PD

A significant feature of PD pathology is the presence of Lewy bodies, which contain a variety of proteins including α-synuclein, ubiquitin, proteasome subunits, chaperone proteins, and neurofilament proteins [125]. This is indicative of incomplete clearance of the target proteins by the ubiquitin-proteasome system (UPS). This may arise from increased protein oxidation products overwhelming the proteolytic capacity of the proteasome [38, 91]. Furthermore, mitochondrial inhibition by neurotoxins can deplete intracellular ATP levels, thereby adversely affecting ATP-dependent proteasomal degradation. Oligomerization of α-synuclein following exposure to toxins, DA, DA metabolites, and upregulation of α-synuclein expression inhibit proteasomal function [22, 46, 76, 91]. Nitrosylation, altered solubility or covalent modification of Parkin diminishes its ubiquitin ligase activity, and Uch-L1 oxidation reduces its hydrolase activity [49, 91]. These findings suggest that interactions between the UPS and mitochondrial function promote the degenerative processes in DA neurons.

Inflammation and PD

A large cohort study of patients has shown that the risk of developing PD in regular non steroidal anti inflammatory drugs (NSAID) users (for cardiovascular protection) was decreased by up to 45% compared with those who take NSAIDs on a non-regular basis [21]. Thus, it is suggested that the use of NSAIDs may lead to neuroprotection in PD [50]. However, the neuroprotective effect of COX-2 inhibitors against MPTP in vivo may not be due to the reduced microglial activation, but rather has been linked to inhibition of COX-induced DA oxidation [118, 119].

Inflammation also has been proposed to contribute to PD pathogenesis, [83] in part through upregulation of inflammatory cytokines such as TNF-α [58, 88]. Various PD models demonstrate inflammation in neurodegeneration, and anti-inflammatory drugs attenuate toxin-induced PD [43, 55, 124]. The well known association between encephalitis and parkinsonism [10, 16] and a report of parkinsonism induced by accidental exposure to lipopolysaccharide (LPS) from Salmonella minnesota [90] have supported the role of inflammation in the etiology of PD. Intranigral and intrapallidal LPS induce microglial activation and DA neuronal death [4, 5, 19, 60, 78, 80, 133]. Microglial activation initiates or perpetuates neuronal loss by increasing cytotoxic molecules like superoxide, NO, various pro-inflammatory cytokines, and prostaglandins [6, 65, 86]. Recently, LPS-induced mitochondrial dysfunction has been demonstrated in vitro [131], where mitochondrial dysfunction precedes cell death [30, 127]. In these studies, LPS toxicity was associated with respiratory chain dysfunction. In mitochondria isolated from the striatum, there was evidence of oxidative damage to mitochondrial components, which suggests that mitochondria may be a target of free-radical stress initiated by activated microglia. The fact that both celecoxib and pioglitazone can reduce mitochondrial dysfunction suggests that mitochondrial impairment may be secondary to inflammation [59].

Future directions

Mitochondrial dysfunction and oxidative stress are thought to play an important role in the pathogenesis of idiopathic PD [54, 61, 101]. Although the mechanisms by which this leads to neurodegeneration in DA neurons are still unknown, recent advances in recessive PD implicate parkin, PINK1, LRRK2 and DJ-1 in mitochondrial function. Further studies in both in both toxin based models and genetic based models of the disease will help elucidate the molecular relationships between the two models and hopefully lead to development of therapeutic interventions that prevent, reduce or ameliorate PD associated mitochondrial dysfunction and oxidative stress.