The first and most robust animal models of Parkinson's disease have been toxin-based models using the complex 1 inhibitors rotenone and MPTP (Betarbet et al., 2000; Forno et al., 1986
). Such models recapitulated selective dopaminergic neuronal loss, α-synuclein-positive aggregates, and the clinical features of parkinsonism. However, the use of complex 1 toxins precludes the study of the natural pathophysiological events within mitochondria prior to neuronal death. Genetic models of the autosomal recessive PD have been less successful at reproducing the clinicopathological features of PD in vivo mammalian models (Kitada et al., 2007
). However, in vitro models of PINK1 deficiency have produced phenotypes consistent with PD, namely age-related dopaminergic neuronal loss with mitochondrial dysfunction and oxidative stress (Wood-Kaczmar et al., 2008
). Moreover, the in vitro models enable the physiologic events occurring within neurons, directly due to a mutation causing PD, to be characterized.
A reduction of complex 1 activity by 30% has been described in brain, muscle, and platelets of patients with sporadic PD (Schapira et al., 1990
). Here, we demonstrate that the absence of PINK1 is associated with inhibition of respiration, with concomitant reduced oxygen consumption and an altered redox state. The lowered activity of the respiratory complexes is insufficient to maintain the Δψm, and hence results in a decrease in Δψm. As a result, the mitochondria switch from the production of ATP to the consumption of ATP by the F1
-ATPase in order to maintain their Δψm (Campanella et al., 2008
). Interestingly, this phenomenon could be reversed by the provision of additional respiratory chain substrates: the increase in respiration in the presence of additional pyruvate resulted in a concomitant switch in the mechanism of Δψm maintenance from hydrolysis of ATP to production of ATP. The ability to reverse the respiratory chain inhibition and repolarize the Δψm has not been previously demonstrated in PD. These data strongly suggest that the respiratory complexes in PINK1 deficiency are intact and that their functional inhibition is in fact secondary to reduced substrate supply. In keeping with other reports in the literature (Scheele et al., 2007
), we found that PINK1 loss of function was associated with reduction in glucose uptake at the plasmalemmal membrane in human and mouse neurons. Thus, we conclude that reduced substrate delivery causes the impairment of respiration and reduced Δψm in cells that lack PINK1.
One of the major functions of the mitochondria is the maintenance of calcium homeostasis within the cell. Abnormal calcium homeostasis has been implicated in a range of diseases such as Alzheimer's disease (Abramov et al., 2003, 2004
), Huntington's disease, amyotrophic lateral sclerosis, and stroke (Mattson, 2007
). In PD, one important observation is that adult dopaminergic neurons are uniquely dependent on calcium channels (rather than sodium channels) to maintain autonomous pacing activity. As a result these neurons are exposed to frequent large influxes of cytosolic calcium, which must be buffered by the mitochondria (Chan et al., 2007
). Mitochondrial dysfunction and in particular an inability to handle these calcium loads may render dopaminergic neurons particularly vulnerable to injury. Furthermore, it has been reported that SN and VTA neurons that express higher levels of calcium-binding proteins such as calbindin are spared in both sporadic PD and MPTP toxicity (Damier et al., 1999; Liang et al., 1996
). However, the nature of calcium dysregulation and its contribution to neuronal death is largely uncharacterized in PD. We report that human and mouse neurons lacking PINK1 have a higher basal [Ca2+
than control neurons. Furthermore, stimuli that induce a rise in cytosolic calcium cause mitochondrial calcium overload with greatly impaired recovery. As a consequence of this, the mitochondrial calcium uptake capacity is dramatically reduced in the absence of PINK1. Thus, repetitive rises in cytosolic calcium result in mitochondrial calcium overload, leading to premature opening of the PTP, profound mitochondrial depolarization, and ultimately cell death (Crompton, 1999; Nicholls and Budd, 2000
). In addition, the opening of the PTP by calcium is modulated by factors such as lowered basal Δψm and raised mROS production (described below), which also occur in PINK1 deficiency.
Accumulation of calcium within the mitochondria matrix depends on both calcium uptake into the mitochondria through an electrogenic uniporter, as well as extrusion of calcium from the mitochondria through Na+
antiporters (Szabadkai et al., 2006
). Our data suggest that the cause of mitochondrial calcium accumulation in PINK1 deficiency is a direct impairment of calcium efflux from the mitochondria secondary to dysfunction of the Na+
exchanger. The reasons for this are (1) following an increase in [Ca2+
there was no mitochondrial efflux of Ca2+
and concomitant influx of Na+
(which would normally enable recovery of [Ca2+
), (2) application of external Na+
was able to activate the Na+
exchanger in control neurons but not in PINK1 KD neurons, and (3) there was no apparent abnormality of calcium influx in PINK1 KD/KO cells, To date, the existence of the calcium uniporter and antiporters has been established only functionally. The Na+
antiporter was first recognized in heart and brain mitochondria (Carafoli et al., 1974; Crompton et al., 1977, 1978
). Although one putative candidate was reported to be isolated from beef heart mitochondria (Li et al., 1992
), the proteins that truly account for the activity of these transporters have not yet been identified and confirmed. We have shown physiologically that PINK1 regulates the Na+
exchange mechanism in mitochondria, and that this is fundamental to the role of PINK1 in cell physiology. We are unable to prove biochemically that the Na+
exchanger is a direct substrate or interactor of PINK1 because the molecular identity of this protein remains unclear. In order to fully clarify the relationship between PINK1 and Na+
exchange, it is necessary to first identify this protein, which is the subject of ongoing research.
One interesting observation in these experiments was the significant difference in the mitochondrial calcium capacity of mouse compared to human neurons. Control mouse neuronal mitochondria depolarized at much lower calcium concentrations than control human neurons. Control mouse neurons did not accumulate mitochondrial calcium in the same way as normal human neurons due to immediate opening of the PTP.
Oxidative stress has long been implicated in sporadic PD: there is evidence of elevated levels of lipid peroxidation markers (4-hydroxynonenal and malondialdehyde) and protein nitration in the substantia nigra and in Lewy bodies of patients with PD (Andersen, 2004
). However, it is less clear whether the oxidative stress is causal in PD or a consequence of dysfunctional neurons. At a cellular level, overproduction of ROS would theoretically be able to inhibit the mitochondrial Na+
exchanger, causing mitochondrial calcium overload, as well as inhibiting the plasmalemmal glucose transporter and reducing respiration (Jornot et al., 1999
). In our models, we demonstrate that in the absence of PINK1, there was a significant increase in ROS production from two separate sources. There was an increase in mROS that may be secondary to the increase in mitochondrial calcium, as well as the impairment of respiration. Alternatively, it is recognized that PTP opening per se results in an increase in ROS production through a conformational change in complex I (Batandier et al., 2004
). In addition, there was overproduction of cROS, in the form of superoxide by NOX. Increased activity of NOX has been reported to be responsible for the oxidative stress seen in other neurodegenerative diseases and is a major source of ROS production in MPTP-induced cell toxicity (Anantharam et al., 2007
). As calcium is known to increase the activation of NOX-2 (Abramov et al., 2005
), we postulate that cytosolic calcium may result in an increase in cROS production. Manipulation of ROS production using inhibitors of NOX, NOX-2 (gp91phox
) siRNA, and ROS scavengers enabled us to distinguish the primary abnormalities from the secondary consequences of ROS production in PINK1 KD/KO neurons. Reduction of ROS production was able to reverse the impaired glucose uptake in PINK1-deficient cells, and thus oxidative stress plays an important role in causing the impaired respiration seen in PINK1 KD/KO. Provision of substrates was also able to overcome the impaired respiration and mitochondrial depolarization. However neither reduction of ROS nor substrate provision was able to affect mitochondrial calcium overload or calcium-induced mitochondrial depolarization in PINK1 KD/KO cells.
In summary, we have characterized the mitochondrial pathophysiology that occurs due to PINK1 loss of function and have attempted to define mechanisms by which PINK1 deficiency renders neurons vulnerable. Within mitochondria there is considerable crosstalk between the bioenergetic function and calcium homeostasis. Our data suggest that as a result of PINK1 deficiency there is primarily an impairment of mitochondrial calcium efflux resulting in mitochondrial calcium overload. This induces a rise in ROS that may further impair calcium efflux and also inhibit glucose uptake, resulting in reduced substrate delivery and impaired respiration. Ultimately, the synergistic action of increased mROS and mitochondrial calcium overload induces opening of the PTP. Opening of the PTP, occurring either as an early event (calcium-induced) or as a late event, has several consequences that exacerbate the mitochondrial pathophysiology and promote cell death: (1) PTP opening will further increase ROS production in the mitochondria, (2) PTP opening will result in reduced rate of maximal respiration through pyridine nucleotide depletion (Di Lisa et al., 2001
), and (3) PTP opening will cause cytochrome c
release and neuronal apoptosis. Neurons of the substantia nigra, where there is increased oxidative stress and large calcium influxes, will be particularly susceptible to mitochondrial apoptosis via these mechanisms.