Microtubules (MTs) are dynamic polymers consisting of α/β tubulin dimers and playing a plethora of roles in eukaryotic cells. Looking at neurons, they are key determinants of neuronal polarity, axonal transport and synaptic plasticity. The concept that MT dysfunction can participate in, and perhaps lead to, Parkinson's disease (PD) progression has been suggested by studies using toxin-based and genetic experimental models of the disease. Here, we first learn lessons from MPTP and rotenone as well as from the PD related genes, including SNCA and LRRK2, and then look at old and new evidence regarding the interplay between parkin and MTs. Data from experimental models and human cells point out that parkin regulates MT stability and strengthen the link between MTs and PD paving the way to a viable strategy for the management of the disease.
The molecular pathways implicated in neurodegenerative disorders are gradually being elucidated and several contributing factors have been identified. To date, aetiopathogenic mechanisms in Parkinson's disease (PD) converge on accumulation of aberrant or misfolded proteins, mitochondrial injury, and oxidative/nitrosative stress, making PD a multifactorial disease [1,2]. However, the primary degenerative events remain unclear, thus making it really hard to develop an efficient therapy for this devastating disorder.
PD is a progressive neurodegenerative disorder that is characterized by tremor, muscular rigidity and bradykinesia, with a prevalence of 2–5% in the population aged 60 years, worldwide. PD can be defined in biochemical terms as a dopamine-deficiency state resulting from loss of dopamine neurons in the substantia nigra pars compacta accompanied by characteristic intraneuronal protein inclusions, termed Lewy bodies. On these grounds, starting in the 1950s, the strategy for treating PD has been to restore the dopamine concentrations in the brain by administering pharmacological treatment. However, thanks to a huge amount of clinical and basic research work, a redefinition of PD as a multiorgan disease has been proposed recently and novel therapeutic strategies are emerging .
In recent years, growing attention has been dedicated to neuronal cytoskeleton dysfunction and increasing evidence suggests a role for the microtubule (MT) system in the pathogenesis of neurodegenerative disorders. Mutations in tubulin, the major constituent of MTs, have been found to induce severe neurological disorders, such as peripheral neuropathy and loss of axons in many kinds of brain neurons  and, very recently, to be associated with familial amyotrophic lateral sclerosis . Moreover, defects in the proper regulation of MT organization and stability are tightly linked to neuronal damages. Indeed, significant impairment in MT-associated proteins has been extensively reported in Alzheimer's disease, frontotemporal dementia and other tauopathies  and, notably, the failure in polyglutamilation of tubulin can dramatically lead to a rapid neuronal cell death in an ataxia mouse model . Besides MT organization and stability, MT-dependent functions, such as overall axonal transport, are increasingly investigated in the field of neurodegeneration. The intracellular transport of organelles along an axon is a complex and crucial process for the maintenance and function of a neuron. Several different mechanisms including defects in the proper organization of MTs, mutations in MT-associated proteins and molecular motors, and activation of MT-targeting kinases act in concert and produce deficits in axonal transport underlying several neurodegenerative diseases, as extensively reviewed by Millecamps and Julien . In addition, recent evidence suggests that axon degeneration underlying PD could depend mainly on the failure of axonal transport .
The question arises as to whether we can reasonably include MT dysfunction among the culprits triggering neurodegeneration in PD or not. Here, we bear in mind such a key question and move from a brief insight into the basis of MT functions in neurons to the evidence that MT dysfunction occurs in experimental parkinsonism and, finally, to the critical discussion on the interplay between parkin and MT system in cellular and animal models and in human tissues.
Microtubules and microtubule-dependent functions in neurons
MTs are non-covalent cytoskeletal filaments, which occur in all eukaryotic cell types from fungi to mammals. They consist of α/β tubulin heterodimers that assemble in a head-to-tail fashion into linear protofilaments whose lateral association forms polarized 25 nm wide hollow cylindrical polymers. MTs are heterogeneous in length and highly dynamic in vivo and in vitro, undergoing cycles of polymerization and rapid depolymerization. This ‘dynamic instability’ property was first described in 1984  as a feature that is crucial to many MT functions. The tight regulation of their organization and dynamics depends on the incorporation of alternative tubulin isotypes, a highly complex and diverse set of MT-interacting proteins, and posttranslational modifications occurring on MTs .
MTs play several essential roles in cell shape acquisition and in the performance of many intracellular processes. Neurons are a striking example of cells in which MTs are essential to achieve a high degree of morphological and functional complexity. Neuronal MTs display different orientation and dynamics in axons and dendrites, and interact with specific associated proteins . In addition, the incorporation of tubulin isotypes and posttranslational modifications of tubulin are selectively combined and distributed among different subcellular compartments, thus generating a tubulin code, that might regulate basic as well as higher-order neuronal functions. Highly dynamic MTs are enriched in tyrosinated tubulin and accumulate a set of factors known as MT plus-end tracking proteins; they are essential for rapid remodelling and reorganization in the growth cone underlying axonal elongation during neuronal differentiation  and synaptic plasticity in mature neurons . On the contrary, a high stability is favoured for MT functions in the shaft of axons and for the preferential binding of MT-based motors transporting membrane-bound organelles and regulatory macromolecular complexes . Neuronal MT stability is related to the accumulation of several posttranslational modifications of tubulin including acetylation, detyrosination, Δ2-tubulin, polyglutamylation and the very recently described polyamination , and to spatial gradient of tau .
Beyond their known conventional roles for supporting neuronal architecture, organelle transport and synaptic plasticity, a novel function as ‘information carriers’ has been attributed to neuronal MTs . This amazing theory posits that both the short, stable and mobile MTs and the highly dynamic ends of longer MTs can act as information carriers in the neuron thanks to their ability to interact with a vast array of proteins. Short MTs, which appear to be unusually stable, move rapidly along axons and presumably in dendrites as well. It is reasonable to assume that they may convey information and signalling molecules with them. In addition, highly dynamic regions would act as scaffolds concentrating MT plus end tracking proteins, which, in turn, interact with many other proteins and structures contributing to the plasticity of the neuron, including kinases and small G proteins that impact the actin cytoskeleton and proteins that reside at the cell cortex .
Microtubule dysfunction in experimental models of Parkinson's disease
The concept that MT dysfunctions can participate in, and perhaps lead to, PD progression has been suggested by studies on toxin-based and genetic experimental models of the disease.
Within the context of studies on PD-inducing neurotoxins, intriguing results have been reported with N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), a neurotoxin widely used as a tool for studies on sporadic PD , and the herbicide rotenone. We showed that 1-methyl-4-phenylpyridinium (MPP+), the toxic metabolite of MPTP, reduces MT polymerization and interferes with dynamic instability of MTs in vitro acting as a destabilizing factor . Then, we confirmed and extended these results reporting that MPP+ leads to MT alteration in neuronal cell and, in turn, to mitochondrial trafficking impairment . Finally, we showed that systemic injection of MPTP to mice induces MT dysfunction that occurs very early, before axonal transport deficit, depletion of tyrosine hydroxylase and, ultimately, dopaminergic neuron degeneration . Moving to the herbicide rotenone, old studies demonstrated its ability to induce MT depolymerization in vitro , whereas more recent data suggest that MT disruption may be an alternative mechanism underlying rotenone-induced dopamine neuron death in cellular models [22,23].
We can find further signs of MT involvement in PD looking at PD-linked genes. Interestingly, several independent GWAS and meta-analysis studies have shown a genome-wide significant association of single nucleotide polymorphisms in the gene coding for α-synuclein (SNCA) and the MT-associated protein tau . α-Synuclein, the first protein associated to familial form of PD , interacts with tubulin with crucial consequences: the promotion of its aggregation in fibrils , the interference with tubulin assembly  and the recycling of monoamine transporter . More recently, MT disruption has been reported in cells overexpressing α-synuclein  or following incubation with extracellular α-synuclein . In addition, the kinase LRRK2 has been shown to interact with and to phosphorylate β-tubulin [31,32] and tubulin-associated tau, whereas a novel role of DJ-1 in the regulation of MT dynamics has been proposed .
Although these studies provide evidences that the MT cytoskeleton could be involved in neuronal damage caused by PD-related proteins or toxins, very little is known about MT dysfunction in patients. Using cybrid cell lines generated from idiopathic PD patients, Esteves et al.  showed significant alterations in MT integrity as compared with healthy subjects. Notably, we have recently analysed primary fibroblasts deriving from patients with idiopathic or genetic PD and disclosed reduction in MT mass and significant changes in signalling pathways related to MT stability .
We believe that it is not a coincidence that tubulin and MTs represent a point of convergence in so many different PD experimental models, thus making the study of MT dysfunction a challenge leading to a better comprehension of PD pathogenesis.
The interplay between parkin and microtubules
Exonic deletions in the Parkin gene were first reported in Japanese families with autosomal recessive juvenile-onset parkinsonism with onset frequently occurring before the age of 20 . The Parkin gene encodes for a member of the E3 ligase family that catalyses the addition of ubiquitin to numerous target proteins . The molecular understanding of the regulation of parkin E3 ligase activity is emerging . However, it has been suggested that parkin, in addition to its ligase activity, has a number of other roles including the regulation of mitochondria dynamics and quality control designed to preserve mitochondria integrity . Most of the supporting observations derive from mammalian cell lines overexpressing parkin, but endogenous parkin does not induce mitophagy in induced-pluripotent stem cell (iPSC)-derived human neurons . This raises the issue of whether parkin involvement in this process is actually relevant in neurons or in PD pathogenesis . Very recently, it has also been reported that parkin interacts with the kainate receptor GluK2 subunit and regulates the receptor function in vitro and in vivo .
Parkin interaction with tubulin and MTs has been proposed many years ago and remained largely neglected for a long time. Interestingly, parkin binds and increases the ubiquitination and degradation of both α- and β-tubulin , whose complex folding reactions are prone to produce misfolded intermediates. In addition to its E3 ligase activity on tubulin, however, Yang et al.  proposed that parkin strongly binds tubulin/MTs through three redundant interaction domains resulting in MT stabilization. At the moment, we can simply speculate that the anchorage of parkin to MTs could provide an important environment for its E3 ligase activity on misfolded substrates that are usually transported on MTs themselves. Further work demonstrates that parkin protects midbrain dopaminergic neurons against PD-causing substances, as rotenone and colchicine, by stabilizing MTs . This process seems to be mediated by the regulation of the MAP kinase pathway, which, interestingly, is a direct regulator of MT stability via the modulation of tubulin posttranslational modifications.
Bringing into focus the impact of parkin on MT-dependent functions, a reliable consequence of the alteration of MT stability could be the dysregulation of axonal transport. Indeed, previously, parkin has proved to regulate the trafficking of mitochondria in hippocampal neurons, especially when they are damaged and have to be degraded. This process was found to be dependent on the Miro phosphorylation .
Striking data coming from human cells have recently contributed to our understanding of the interplay between parkin and MTs strengthening interest in this aspect. We reported that PD-patient skin fibroblasts bearing Parkin mutations display reduced MT mass and imbalance in the pattern of tubulin posttranslational modifications, and that MT pharmacological stabilization or the overexpression of wild-type parkin rescue control phenotype . This is not restricted to skin cells from patients but, interestingly, has been confirmed in iPSC-derived neurons. Ren et al.  found that the complexity of neuronal processes was greatly reduced in both dopaminergic and non-dopaminergic neurons from PD patients with parkin mutations and that MT stability was significantly decreased as demonstrated by the reduction in MT mass. Overexpression of parkin, but not its PD-linked mutants, restored the complexity of neuronal processes and MT mass. Notably, the MT depolymerizing agent colchicine mirrored the effect of parkin mutations by decreasing neurite complexity in control neurons while the MT stabilizing drug taxol mimicked the effect of parkin overexpression. These results strongly support the concept that the interaction of parkin with MTs in neurons may have an important physiological role. Thus, although the hypothesis of the interaction of parkin with MTs is supported mainly by studies in cellular models, it seems to be a promising theory, which provides a mechanistic explanation for the multiple intracellular functions and, possibly, dysfunctions of parkin. Indeed, we are currently undergoing the analyses of brain samples from Parkin knockout mice; our preliminary results have shown an early alteration of MT stability, thus confirming and expanding the importance of parkin in modulating the MT system.
A growing body of evidence from experimental models and human cells indicates that parkin regulates MT stability and strengthens the link between MTs and PD. Indeed, the MT cytoskeleton represents a point of convergence in the action of various proteins mutated in PD and of PD-inducing neurotoxins, suggesting that it has a major role in the onset of the disease and providing the rationale for novel therapeutic interventions. Thus, MT stabilizing strategies may offer an opportunity for treating neurodegenerative diseases [48–50]. Importantly, we have recently demonstrated that this may be true also in PD showing that Epothilone D, a MT stabilizer drug, exerts neuroprotective effects in a toxin-based murine model of PD .
This work was supported by the Fondazione Grigioni per il Morbo di Parkinson, Milan, Italy (2011–2015 grants to G.C.) and the ‘Dote ricerca’, FSE, Regione Lombardia (to D.C.).
PINK1-Parkin Signalling in Parkinson's Disease and Beyond: Held at Charles Darwin House, London, U.K., 2 December 2014.