Mitochondrial dynamics are essential for maintaining organelle stability and function. Through fission, fusion and mitophagic events, optimal populations of mitochondria are retained. Subsequently, alterations in such processes can have profound effects on the individual mitochondrion and the cell within which they reside. Neurons are post-mitotic energy-dependent cells and, as such, are particularly vulnerable to alterations in cellular bioenergetics and increased stress that may occur as a direct or indirect result of mitochondrial dysfunction. The trafficking of mitochondria to areas of higher energy requirements, such as synapses, where mitochondrial densities fluctuate, further highlights the importance of efficient mitochondrial dynamics in neurons. PD (Parkinson's disease) is a common progressive neurodegenerative disorder which is characterized by the loss of dopaminergic neurons within the substantia nigra. Complex I, the largest of all of the components of the electron transport chain is heavily implicated in PD pathogenesis. The exact series of events that lead to cell loss, however, are not fully elucidated, but are likely to involve dysfunction of mitochondria, their trafficking and dynamics.
Integral changes in mitochondria have long been implicated in the pathogenesis of PD (Parkinson's disease), notably in relation to complex I of the electron chain (reviewed in ). Mitochondrial bioenergetics, transmembrane potential, dynamics and trafficking are interrelated. Complex I dysfunction can affect mitochondrial morphology and morphological changes can modulate complex I activity. The combined dysfunction of these two is likely to be key to the understanding of the development of PD.
It is vital for cells to regulate mitochondrial integrity, number and location, to balance energy needs, buffer calcium and modulate various signalling pathways. The distinctive characteristics of dopaminergic neurons within the substantia nigra, known to undergo profound degeneration in PD, may render this population uniquely susceptible to perturbations in mitochondrial motility. The increased oxidative burden placed on dopaminergic neurons through the metabolism of dopamine, alongside their unique calcium-maintained pacemaking activity, are just two individualities that may render this population vulnerable to mitochondrial dysfunction. Furthermore, the pivotal role of this organelle in energy provision, ROS (reactive oxygen species) production and calcium buffering highlights its contribution to neuronal survival in this brain region.
The present review briefly describes key aspects of mitochondrial dynamics, perturbations in PD and how the two may be linked by complex I dysfunction.
Fission and fusion
The notion of mitochondria as lone ellipsoid organelles has long passed. Mitochondria can exist in a multitude of forms, have been shown to be dynamic and are dependent on numerous factors. The transition between elongated reticular networks and punctate structures is dependent on two antagonistic processes, i.e. fission and fusion, both of which are vital for neuronal survival as they directly affect organelle number/shape and location . Fusion events are mediated through the MFNs (mitofusins) of the outer mitochondrial membrane and OPA1 (optic atrophy 1) of the inner membrane. FIS1 (fission protein 1) and DRP1 (dynamin-related protein 1) are heavily involved in fission events (reviewed in ), which have been shown to be preceded by a sustained fall in mitochondrial membrane potential . Furthermore, fusion mixes mitochondrial contents and is known to benefit mtDNA stability and rescue damaged phenotypes . Fission, conversely, generates individual smaller organelles that are necessary for transportation purposes as well as selecting unwanted mitochondria for programmed removal. Changes in both of these processes have been shown to cause certain neurodegenerative conditions and are implicated in the pathogenic pathway of many others .
Fission, fusion and motility are intertwined, and recent work shows that the likeliness of one event is tied into others . Alterations in these processes are therefore rarely independent and subsequently affect many factors controlling mitochondrial stability. Some patients with OPA1 gene mutations, for example, actually show increased numbers of mtDNA mutations, adding to the argument that fusion is necessary for maintenance of mtDNA integrity . In addition, neurons expressing mutant huntingtin (Htt) or tau show impaired mitochondrial movement, specifically suppression of mitochondrial fusion with mitochondrial shortening . In relation to PD, a loss of MFN2 in the nigrostriatal neurons of mice causes retrograde degeneration of dopaminergic neurons, indicating alterations in mitochondrial dynamics can independently cause PD-like nigrostriatal defects . These data highlight the impact of perturbed mitochondrial dynamics on the survival of neurons.
Removal of damaged or unnecessary mitochondria is a highly regulated event, with current understanding highlighting PINK1 [PTEN (phosphatase and tensin homologue deleted on chromosome 10)-induced putative kinase protein 1], a serine/threonine kinase with a mitochondrion-targeting sequence (MIM 608309), and parkin, an E3 ubiquitin ligase (MIM 602544) encoded by PARK2, as key to facilitating this process. It is worth noting, however, that several factors in this relatively recent discovery remain unclear and it is likely that other components of the pathway or indeed novel pathways are yet to be uncovered. PINK1 localizes to either the outer or inner membrane depending on membrane potential . Upon depolarization of mitochondria following damage/uncoupling, PINK1 accumulates on the outer mitochondrial membrane recruiting parkin, which ubiquitinates numerous outer mitochondrial membrane proteins, including VDAC1 (voltage-dependent anion channel 1) and MFN, that signal autophagic factors and seal the organelle's fate for degradation . Any inhibition of this pathway can easily lead to accumulation of mitochondria that can then harm other cellular components. Useful components that would ordinarily be recycled are confined in these accumulations and the control of apoptotic signals is lost. Knockout of both parkin and PINK1 show various levels of oxidative stress and changes in respiration in a number of mouse models [13–15]. Pink1-null mice, for example, display impaired dopamine release, deficient respiration and increased response to oxidative stress [13,16]. This phenotype increases in severity and distribution in Drosophila , while the modulation of PINK1 and parkin in this model have elegantly displayed that they function in the same pathway . Loss of PINK1 has also been shown to alter mitochondrial morphology to differing degrees dependent on the cell type; for example, a loss of PINK1 from dopaminergic neurons leads to a swollen abnormal mitochondrial phenotype in conjunction with reduction in viability . These experiments highlight further the debilitating impact of changes in this pathway and reiterate the links between neuronal survival, mitochondrial fission, fusion and mitophagy.
Mitochondria not only show dynamic movement within the network, but also show motility throughout neurons with approximately 35% of neuronal mitochondria estimated to be moving at any given time [20–22]. Sophisticated trafficking machinery exists to enable mitochondria to be readily transported to areas of increased energy demand or returned for degradation, a process which is thought to occur predominantly in the cell body, where concentrations of lysosomes are greater. Mitochondria typically have set styles of movement. These can be fairly uniform over long distances or show more ‘stop–start–reconsider’ characteristics, where movements are shorter and often punctuated with a pause that can precede a change in direction. Within neurons mitochondria are largely moved along microtubules, although movement using actin filaments has been reported for smaller distances and within dendritic spines and growth cones . The ‘docking’ and ‘shipping’ of mitochondria on to these cellular tracks is facilitated through motor proteins and a plethora of adaptor proteins such as Milton, Miro (mitochondrial Rho GTPase), myosin and dynactin. Microtubules display uniform polarity in axons with all positive poles aligning at the axonal terminal; the tracking of mitochondria along these is mediated via different motors dependent on direction. For antegrade movement (towards the axon terminal), kinesin motors are employed; conversely, retrograde movement utilizes dynamin motors .
The efficient movement of mitochondria has been shown previously to be compromised in numerous neurodegenerative disorders. The role of mitochondrial motility in PD is arguably one of the strongest cases for the interplay between movement and neurodegeneration, strengthened by the discovery of PINK1 and PARK2 familial gene mutations in autosomal-recessive cases of PD. Alongside their role in degradation discussed above, these two proteins are known to interact with Miro, an adaptor protein that links a kinesin motor to mitochondria [25,26]. Upon damage and the subsequent decrease in membrane potential, both PINK1 and parkin have been shown to show stronger physical interaction with Miro. PINK1 can phosphorylate Miro, which marks it for degradation and in turn arrests mitochondrial movement in all directions. A second role for PINK1/parkin may therefore be with regard to confining unwanted mitochondria, limiting any harm the damaged organelle can do, avoiding its transportation to critical areas and enabling its subsequent encapsulation via autophagosomes. Alterations through mutations in these genes may explain why trafficking is altered in PD and why ultimately neurodegeneration occurs. It is important to also consider that trafficking pathways are not solely dedicated to mitochondria, but serve as transport links for a plethora of cellular components. Mitochondrial ‘swellers’ such as valinomycin have been shown to lead to fast inhibition of organelle movement, suggesting steric hindrance of mitochondria . This block prevents not only their own movement, but also movement of other organelles, which in part may explain how simple deviations from normal trafficking can quickly bring about a collapse of the cellular environment.
Impaired mitochondrial transport is implicated in numerous neurodegenerative disorders. Elucidating whether dynamic alterations are primary factors or result from other compounding mitochondrial issues remains difficult. The role of mitochondrial dynamics in apoptosis must also not be overlooked. Morphology is tightly linked to apoptosis demonstrated by the fact MFN2 GTPase activity relies on Bcl2 family members  and OPA1 affects apoptosis through the availability of cytochrome c release . Decreased membrane potential may lead to increased susceptibility to apoptosis, a process which is modulated through different pathways, including changes in the morphology of mitochondria . Calcium also serves an apoptotic stimuli , and one may postulate that areas of increased calcium such as dopaminergic neurons may be quicker to react to mitochondrial changes which diminish their calcium sink capabilities and leave apoptotic cascades unchecked.
Dynamic changes in the pathogenesis of PD
Evidence for a causative role of mitochondrial dysfunction in the development of familial and sporadic PD is compelling (Figure 1). Changes to the ultrastructure of mitochondria have been observed in many relevant models of PD. In fibroblasts derived from patients with PINK1 mutations, fragmented mitochondria have been observed . This phenotype can also be seen in PINK1 knockdowns in human neuronal SH-SY5Y cells . This phenomenon, however, does not appear universal for all cell types; for example, in COS7 cells, elongation of mitochondrial networks have been observed following PINK1 knockdown . Discrepancies in findings among and within different model systems add complexity to understanding the role of mitochondrial dynamics in PD, but could equally reflect the multicomponent nature of the disease and highlight that the pathways are far from fully elucidated. This variation could also explain how it takes a unique cellular environment and chain of events to bring about cell death following deficiency, as highlighted by the preferential loss of SN (substantia nigra) neurons.
Pathogenic factors in PD
Many familial forms of PD have now been shown to occur through mutations in genes that encode proteins that function with or localize to mitochondria, or have some role in maintaining mitochondrial function. In addition, changes in mtDNA and bioenergetic defects have been reported in PD . With regard to mitochondrial biology, one of the first reported examples of a mitochondrial change in PD and now one of the most described is the change in mitochondrial complex I activity and expression in patients with this disease. Pharmacological modulation, for example through toxins that inhibit complex I, are used to create animal models of PD and links between complex I and mitochondrial morphology are tantalizingly linked through PINK1, but as yet not fully understood. In Drosophila, for example, no obvious morphological differences have been observed in PINK1 mutants, but, despite this, synaptic dysfunction occurs. Interestingly PINK1-knockout cells isolated from mouse liver revealed a dramatic decrease in enzymatic activity and complex-I-driven respiration; importantly, this finding was unique for complex I . The authors therefore concluded that ablation of PINK1 results in a primary functional defect in the catalytic activity of complex I. This is made relevant to the human condition using human cells expressing PINK1-carrying mutations, which display a severe complex I dysfunction similar to that observed in mouse PINK1-knockout cells. Furthermore, rescue experiments using human wild-type PINK1 expression affected complex I, but not other electron transport chain complexes .
Complex I and morphological changes
Mitochondrial bioenergetics display a tight bidirectional relationship with mitochondrial morphology and organization of the network. Stimulation of respiration through nutritional modulation, for example, has been shown to result in network lengthening and complexity , whereas the production of energy in HeLa cells has been shown to be compromised following knockdown of the fission protein DRP1 . Inhibition of complex I through the addition of rotenone leads to a notable decline in viability that is accompanied by a dramatic alteration in mitochondrial morphology .
Evidence from patient-derived primary human fibroblasts reveals a modulation of mitochondrial morphology alongside bioenergetic defects. In fibroblasts, an extended reticular network is commonly observed under basal conditions. Dermal fibroblasts from patients with defined isolated complex I deficiency have been analysed with respect to mitochondrial shape, complex I activity, number and ROS production. The degree of mitochondrial fragmentation appears to be correlated with the severity of the complex I mutation, with the most severe complex I deficiency leading to fragmentation of the membrane; less severe cases appear to more closely resemble their control counterparts. Interestingly, a proportion of the less severely affected cell lines revealed an increased in mitochondrial network complexity and length , which is supported by previous work from the same group showing that chronic rotenone treatment of healthy fibroblasts decreased residual complex I activity, but increased mitochondrial branching . The suggested adaptive mechanism is particularly attractive when we consider the progression from ‘normal’ reticular network-increased-complexity-breakdown and fragmentation displayed in parallel with increasing severity of complex I dysfunction . The idea of hyperfusion as an initial protective mechanism is not a new one. Hyperfused mitochondria may be resistant to apoptosis  and mitophagy , becoming too large to engulf, or to act as a means for content mixing to rescue or buffer damage .
Pathogenic implications of complex-I-induced morphological changes
Previous work has shown that respiratory complexes change in activity with age , and this is most noteworthy in complex I where a decrease in age-related activity is most prominent. In Drosophila, these changes in complex I occur in combination with changes in the ultrastructure of mitochondria; however, both of these findings are observed before noticeable changes in viability . Within Caenorhabditis elegans, decreases in complex I activity are also observed and relate to proportions of increased mitochondrial size .
Changes in complex I are likely to have a profound impact on morphology through changes in ATP levels , lipid peroxidation and fluctuations in the production of free radicals, which are able to damage and potentially act as signals, which directly regulate mitochondrial dynamics . Complex-I-induced morphological changes can have numerous effects (Figure 2), and dynamics rely on changes in membrane potential to act as cues and allow mechanistic reformation. Subtle changes may be even more significant in neurons when we consider directionality is strongly governed by the potential of individual organelles, with 90% of high-potential mitochondria moving towards the growth cone and 80% of low-potential mitochondria moving towards the cell body , and that potentiality is varied in dendrites and axons, highlighting clear membrane-dependent functioning . In parallel, morphological changes may cause physical barriers to cellular transport.
Implications of complex-I-induced morphological differences in neurons
Taken together, much experimental evidence points towards a more prominent role for complex I dysfunction and the associated mitochondrial morphological changes in the progression of PD. This dysfunction may occur through many routes, which may go some way to explaining the discrepancy of findings in different experimental models. Conversely, mitochondrial morphology can have an impact on complex I and it is likely that changes in either quickly become self-perpetuating. The fact that many of the morphological effects in experimental models occur outside basal conditions, i.e. when cells become stressed, further highlights the likeliness of a multifactorial model of PD.
One challenge for future research will be deciphering complex pathways that will separate cause from affect. Distinguishing distinct morphological changes from primary mitochondrial dysfunction is difficult. Multiple conditions cause alterations in the movement and form of mitochondria and, subsequently, targeting dynamics may be a limited therapeutic in certain conditions.
Regardless of the timing at which changes in motility arise, their contribution to homoeostatic loss and cell death is unquestionable and is likely to accelerate collapse of the cellular environment in concert with other direct and indirect mitochondrial factors. In line with this, groups have demonstrated that neurons with previous stress are more susceptible to changes in mitochondrial motility. Models of neurodegenerative disease already exist in which mitochondrial phenotype can be rescued by correcting motility . For this reason, perusing modulators of mitochondrial motility may be valuable in assisting restoration of the cellular environment and ultimately prevent neurodegeneration.
5th Conference on Advances in Molecular Mechanisms Underlying Neurological Disorders: A joint Biochemical Society/European Society for Neurochemistry Focused Meeting held at the University of Bath, U.K., 23–26 June 2013. Organized and Edited by Marcus Rattray (University of Bradford, U.K.) and Rob Williams (University of Bath, U.K.).
This work was supported by the Newcastle University Centre for Brain Ageing and Vitality (funded by the Biotechnology and Biological Sciences Research Council, Engineering and Physical Sciences Research Council, Economic and Social Research Council and Medical Research Council [grant number G0700718]), The Wellcome Trust Centre for Mitochondrial Research [grant number G906919] and a UK NIHR Biomedical Research Centre in Age and Age Related Diseases award to the Newcastle upon Tyne Hospitals NHS Foundation Trust.