Background
Genetic, imaging and environmental studies of Parkinson’s disease (PD) have revealed early problems in synaptic function and connectivity, suggesting that axonal impairment is an early, dominant feature of this disorder [
1]. For example, assessment of available patient positron emission tomography data suggests that at the time of motor symptom onset there is a far greater loss of striatal dopaminergic (DA) terminals than substantia nigra DA neurons [
1]. Moreover, post mortem studies show widespread axonal pathology that precedes the loss of cell bodies [
2,
3]. Such data support the notion that nigral neurons degenerate through a “dying back” axonopathy [
4,
5]. Animal models of PD-linked genes also point to axonal degeneration as an initiating factor. For example, transgenic mice expressing the PD-linked R1441G LRRK2 mutation have decreased DA terminal fields together with increased dystrophic processes and abnormal axonal swellings, findings consistent with DA axonopathy [
6]. In addition, reduced axonal transport is seen with α-synuclein mutants, which accumulate in the cell soma when overexpressed in cortical neurons [
7]. Emerging data also support a role in which the PD-linked genes, PINK1 and Parkin, regulate mitochondrial transport [
8]. Studies in cell lines and hippocampal and cortical neurons show that PINK1 is stabilized on the outer mitochondrial membrane in response to depolarization. Stabilized PINK1 recruits Parkin, which subsequently triggers mitophagy (the autophagy of mitochondria). PD-linked mutations appear to disrupt this process allowing damaged mitochondria to accumulate and then impair axonal transport and initiate neurodegenerative processes [
8].
Studies using Parkinsonian toxins also implicate mitochondrial trafficking and axon integrity in the loss of DA axons. Using specially-designed compartmented chambers and isolated axon preparations derived from transgenic GFP-tagged DA neurons, we discovered that the PD-mimetic toxin MPP
+ rapidly (<1 h) and selectively decreased mitochondrial movement in DA axons [
9,
10]. In support of the notion that damaged mitochondria are re-routed to the cell body for disposal, anterograde traffic was decreased whereas retrograde trafficking was increased [
10]. Temporally, following mitochondrial depolarization and immobility (30–60 min), MPP
+ treatment led to the induction of autophagic markers such as LC3 puncta (microtubule-associated protein 1, light chain 3; also known as ATG8) [
11] (3 h), and then the disruption of microtubule tracks starting at 6 h (beading) peaking between 18–24 h with extensive fragmentation [
10]. Thus in MPP
+-mediated axonal impairment, compromised mitochondria are an early event triggering downstream sequelae leading to autophagy.
6-hydroxydopamine (6-OHDA) is another widely used Parkinsonian toxin that induces degeneration of DA neurons [
12]. 6-OHDA has been shown to disrupt complex I of the mitochondrial electron transport chain and increase generation of reactive oxygen species (ROS) that contributes to an apoptotic form of cell death. However, it is not known how 6-OHDA induces axonal damage.
Using our newly described compartmented microdevices [
9] we studied the effects of 6-OHDA on various processes using murine mesencephalic cultures. Here we show that 6-OHDA decreases mitochondrial and vesicular movement in DA axons and explore potential mechanisms underlying these effects.
Discussion
The use of novel microdevices to isolate axons from cell bodies combined with real time imaging of axonal mitochondria and synaptic vesicles provided new insights into the temporal sequence of cellular changes underlying 6-OHDA-mediated dysfunction (Figure
6C). The present findings demonstrated that (1) 6-OHDA rapidly blocked (<30 min) mitochondrial trafficking in DA axons, a process accompanied by a loss in mitochondrial membrane potential; (2) the effects of 6-OHDA
in vitro were not selective for DA mitochondria as non-DA mitochondria were equally affected; (3) remaining motile mitochondria exhibited decreased movements in anterograde direction; (4) 6-OHDA also decreased axonal transport of synaptic vesicles within 30 min; (5) both mitochondrial and vesicular transport could be rescued by pre-treatment with anti-oxidants, such as NAC; (6) 6-OHDA affected microtubule tracks in axons 6–9 hr after axonal transport ceased and death was observed in cell bodies after 48 hours. (7) 6-OHDA caused the formation of autophagosomes after 9 hr of treatment. Taken together these data demonstrate that 6-OHDA induces cell death via a retrograde dying back process that can be blocked by free radical scavengers.
Widely used as an animal model of PD, 6-OHDA quickly oxidizes to form a variety of free radical species which can lead to toxic sequelae, such as DNA damage [
25] and oxidation of proteins [
26‐
28]. Although oxidative protein damage leads to ER stress and the upregulation of the unfolded protein response [
29,
30], this appears to serve as a protective measure in DA neurons [
25]. Instead, DNA damage leads to activation of a p53- and Puma-dependent apoptotic cascade
in vivo and
in vitro; loss of p53 and Puma rescues 6-OHDA-mediated cell death [
25,
31,
32].
How might these studies fit with early organellar transport impairment, retrograde dying back and loss of axonal integrity? Interestingly,
in vivo studies using 6-OHDA to damage the nigrostriatal projection showed that activation of the Akt/mTOR pathway could block apoptosis, preserve DA cell bodies, prevent autophagy and suppress retrograde axon degeneration [
19]. Mechanistically, these data underscore the importance of preserving axonal function. The present
in vitro findings further emphasize very early events that occur in the axonal compartment that set the stage for later events including the loss of connectivity and ultimately cell death. It should be stressed that the direction of degeneration is also an important caveat and differences may exist between anterograde and retrograde models of degeneration, particularly for degeneration in the nigrostriatal region. For example while many
Wlds studies have shown that it delays and protects against axonal loss in anterograde degeneration, it does not confer axonal protection against retrograde degeneration [
33‐
35]. The model and findings of this study are then directly relevant to understanding the retrograde dying back nature of Parkinson's and other neurodegenerative diseases. Akin to the
in vivo results, inclusion of toxin in the somal compartment did not immediately lead to anterograde loss of axonal transport (Figure
1C) whereas axonal transport was rapidly compromised in the retrograde direction (Figure
1). Although we have not yet tested the role of Akt/mTOR, we would predict that these cascades are downstream of ROS generation given the timing by which autophagy is stimulated (9 h; Figure
6) and that microtubules exhibit fragmentation (24 h; Figure
5).
Because the anti-oxidants NAC and SOD1 mimetics rescued 6-OHDA-immobilized mitochondria, it is likely that axonal transport dysfunction and degeneration is due to the increased generation of ROS species affecting general transport processes. The latter might include oxidation of the transport proteins themselves or oxidation of an adaptor protein responsible for connecting the motor protein to the organelle. For example, impairment of motor proteins such as kinesin-1disrupts axonal transport and induces axonal degeneration [
36]. Adaptor proteins such as Miro and Milton can be oxidized but are also regulated by calcium changes that can affect their binding to each other. Given the lack of effect of EGTA (Table
1) and previous experiments showing no change in calcium levels in response to 6-OHDA [
26], that makes this hypothesis less likely to be correct. Alternatively, 6-OHDA-generated ROS might block mitochondrial ATP production leading to a loss of energy required by the motor proteins to function [
37]. Consistent with this notion, a recent report showed that hydrogen peroxide led to the loss of mitochondrial transport in hippocampal neurons, an effect mimicked by blocking ATP synthesis [
38]. Previously we showed that this was not the case in DA axons treated with another widely used PD-mimetic, MPP
+[
10]. Surprisingly, despite being a Complex I inhibitor, MPP
+ also rapidly blocked mitochondrial transport via a redox sensitive process and not via ATP loss [
10]. The extent to which ATP deficiency mediates 6-OHDA effects in the trafficking of mitochondria remains to be tested.
Although 6-OHDA and MPP
+ are often lumped together as PD-mimetics, their effects on neurons and in particular DA neurons are quite unique. Although both toxins lead to the death of DA neurons in a protein synthesis-, p53-, and PUMA-dependent manner [
16,
25,
29,
39], the downstream signaling pathways diverge in many ways [
40]. In terms of axonal impairment, 6-OHDA and MPP
+ both lead to the loss of neurites prior to cell body death [
10,
16,
40,
41] as well as mitochondrial dysfunction and loss of motility in DA axons. In contrast to 6-OHDA, MPP
+ exhibits a more specific effect on mitochondrial movement that cannot be rescued by ROS scavengers, such as MnTBAP (SOD mimetic); MPP
+ could exert its toxicity by disrupting the redox state (e.g. generation of glutathione or hydrogen peroxide) of the mitochondria after internalization whereas 6-OHDA could directly auto-oxidize to ROS, such as hydrogen peroxide both inside and outside of a cell [
10]. The present findings show that 6-OHDA-generated ROS affects many axonal transport processes including mitochondrial and synaptic vesicle trafficking. Taken together, these data further emphasize that 6-OHDA and MPP
+ impair axons and cell bodies by distinct cellular mechanisms.
The PD-linked genes, Pink1 and Parkin appear to play important roles in regulating mitochondrial dynamics such as movement and morphology as well as mitochondrial removal after damage [
42‐
45]. Many studies especially in neuroblastoma cells show that mitochondrial membrane depolarization stabilizes Pink1 on the outer mitochondrial membrane leading to the recruitment of Parkin, cessation of movement and the rapid induction of autophagy [
46]. Previously we showed that MPP
+ depolarized DA mitochondria and blocked trafficking within 1 hr following treatment; autophagy was observed shortly thereafter (3 hr; [
10]). Despite the rapid depolarization and cessation of mitochondrial movement in 6-OHDA-treated axons, autophagy was observed after 9 hrs (Figure
6). It is unclear why this delay for non-DA neurons or even less for DA neurons exists since damaged mitochondria could serve as a source for leaking ROS that can further exacerbate the oxidative damage to the axon. The role of autophagy in 6-OHDA has been inconsistent in the literature [
47,
48]; one study showed that blocking autophagy helped protect SH-SY5Y cells against 6-OHDA toxicity, whereas the other study showed that regulation of 6-OHDA induced autophagy had no effect on the death of SK-N-SH cells derived from SH-SY5Y cells, a human neuroblastoma cell line. Although not significant, there was a clear trend towards autophagosome formation in DA neurons. Also, we noted differences in the appearance of LC3 puncta between DA and nonDA neurons, which calls for further investigation to determine the characteristics of autophagy in primary DA neurons.
Many additional questions must be addressed, such as could ROS generated from mitochondrial damage or 6-OHDA oxidation limit intra-axonal recruitment of Pink1 to the mitochondria or its stabilization? Perhaps, as suggested above, it is a loss of ATP that impairs organelle movement and Pink1/Parkin are only involved at later time points if at all. Other pathways exist that trigger autophagy, and it may be that these represent alternative, yet slower mechanisms to ensure axonal removal of damaged mitochondria or vesicles [
49,
50]. In any case, the delay in the onset of autophagy suggests that damaged mitochondria are remaining within the axons and are not being removed which may contribute to further axonal impairment due to steric hindrance. Moreover, just the appearance of LC3 puncta is not indicative of the successful removal of damaged organelles, since the formation of an autolysosome is required for complete removal of damaged mitochondria. Excessive autophagosome formation without proper trafficking could also lead to transport blocks.
It is clear that axonal transport disruptions play an early and important role in 6-OHDA induced axonal degeneration. While differences exist between 6-OHDA’s and MPP
+’s effects on axonal transport, the observation that these two widely used toxin models converge on early dysregulation of mitochondrial transport prior to other events such as microtubule fragmentation points to the importance of maintaining the health of the axonal compartment. While it remains to be seen whether other PD toxin models, such as paraquat or rotenone induce similar patterns of axonal impairment in midbrain DA axons, maintenance of mitochondrial transport could bridge the gap between different causes of axonal degeneration and suggest a common therapeutic strategy. Improper trafficking of vital organelles, such as mitochondria and other signaling vesicles may lead to energy deficits, exacerbate oxidative stress, ionic disruption, accumulation of misfolded proteins, or the inability of retrograde signaling molecules to reach their somal targets. All of these processes could lead to the activation of axonal death pathways. The discovery of Sarm1, a protein required for the activation of injury-induced axonal degeneration points to the existence of one such axonal death signaling pathway [
51]. Whether Sarm1 or an axon regenerative pathway, such as mTOR [
52,
53], is applicable to axonal impairment in PD remains to be addressed. The development of microdevices provides a tool to rigorously characterize cell populations such as neurons whose extended, compartmented morphology renders previously intractable problems solvable. These new technologies continue to enhance and expand the available toolset for understanding key biological processes in order to develop better therapies for patients suffering from major neurological disorders.
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
XL, JSK, KOM, and SSE were involved in the design of experiments. SH performed all animal procedures. XL and JSK performed experiments and data analysis, while XL drafted the manuscript. All authors participated in revising, editing and approving the final manuscript.