Glial dysfunction in the pathogenesis of α-synucleinopathies: emerging concepts
- First Online:
- Cite this article as:
- Fellner, L., Jellinger, K.A., Wenning, G.K. et al. Acta Neuropathol (2011) 121: 675. doi:10.1007/s00401-011-0833-z
- 673 Views
Parkinson’s disease (PD), dementia with Lewy bodies (DLB) and multiple system atrophy (MSA) are adult onset neurodegenerative disorders characterised by prominent intracellular α-synuclein aggregates (α-synucleinopathies). The glial contribution to neurodegeneration in α-synucleinopathies was largely underestimated until recently. However, brains of PD and DLB patients exhibit not only neuronal inclusions such as Lewy bodies or Lewy neurites but also glial α-synuclein aggregates. Accumulating experimental evidence in PD models suggests that astrogliosis and microgliosis act as important mediators of neurodegeneration playing a pivotal role in both disease initiation and progression. In MSA, oligodendrocytes are intriguingly affected by aberrant cytoplasmic accumulation of α-synuclein (glial cytoplasmic inclusions, Papp-Lantos bodies). Converging evidence from human postmortem studies and transgenic MSA models suggests that oligodendroglial dysfunction both triggers and exacerbates neuronal degeneration. This review summarises the wide range of responsibilities of astroglia, microglia and oligodendroglia in the healthy brain and the changes in glial function associated with ageing. We then provide a critical analysis of the role of glia in α-synucleinopathies including putative mechanisms promoting a chronically diseased glial microenvironment which can lead to detrimental neuronal changes, including cell loss. Finally, major therapeutic strategies targeting glial pathology in α-synucleinopathies as well as current pitfalls for disease-modification in clinical trials are discussed.
Keywordsα-SynucleinMicrogliaAstrogliaOligodendrogliaNeurodegenerationParkinson’s diseaseMultiple system atrophy
α-Synucleinopathies: general characterisation
α-Synucleinopathies are neurodegenerative diseases characterised by misfolded, hyperphosphorylated and fibrillar α-synuclein-positive cytoplasmic inclusions in the central nervous system (CNS) [17, 51, 211].
α-Synuclein is a protein with 140 amino acids that belongs to a family encoded by three distinct genes for α-, β- and γ-synuclein. In normal brain, α-synuclein is enriched in presynaptic terminals, especially in the neocortex, hippocampus, striatum, thalamus and cerebellum [101, 161]. The normal function of α-synuclein is poorly understood. The absence of α-synuclein expression in knock-out mice is associated with striatal dopamine-dependent dysfunction . α-Synuclein interacts with presynaptic membranes and may regulate synaptic vesicle pools . A recent study of αβγ-synuclein triple knockout mice demonstrated the fundamental role of synucleins in the control of presynaptic terminal size and synaptic structure. Its deletion causes alterations in synaptic transmission, age-dependent neuronal dysfunction and diminished survival . α-Synuclein may directly bind to mitochondrial membranes leading to mitochondrial fragmentation and neuronal cell death . α-Synuclein can also function as a molecular chaperone, important for the folding and re-folding of synaptic proteins .
α-Synucleinopathies include Parkinson’s disease (PD), dementia with Lewy bodies (DLB) and multiple system atrophy (MSA). Their pathological hallmarks are α-synuclein-positive depositions in neurons [Lewy bodies (LBs) and Lewy neurites (LNs)] in PD and DLB, and α-synuclein-positive glial cytoplasmic inclusions (GCIs) in oligodendrocytes in MSA. These inclusions are widely accepted as diagnostic hallmark lesions of α-synucleinopathies [11, 168, 212, 241]. Genetic studies have suggested the causal role of α-synuclein in neurodegenerative diseases [174, 202, 263]. SNCA gene duplications and triplications, as well as three pathogenic point mutations (A53T, A30P, and E46K) have been associated with genetic variants of PD [159, 174, 205, 262], and common SNCA variants influence disease risk of PD . Recent studies confirmed the association between PD and both SNCA SNPs (single nucleotide polymorphisms) and the H1 haplotype of MAPT (microtubule-associated protein tau) . In MSA, familial disease appears rare [209, 257], and up to now mutations have not been detected in the entire coding region of the α-synuclein gene of pathologically confirmed MSA cases . However, candidate SNP studies have demonstrated a significant association between polymorphisms within the SNCA locus and the risk of developing MSA [3, 197]. Post-translational modifications of α-synuclein, including nitration, glycosylation, phosphorylation and ubiquitination are commonly described in relation to pathological accumulation of the protein and may contribute to the pathogenesis of PD, DLB and MSA [6, 73, 74, 239]. Furthermore, the detection of α-synuclein in human cerebrospinal fluid and blood plasma has suggested a possible role of extracellular α-synuclein in neurodegeneration [58, 124, 152]. Finally, prion-like cell-to-cell transmission of α-synuclein has been proposed which may provide a novel insight in disease progression mechanisms [50, 86, 126].
Parallel to α-synuclein accumulation, α-synucleinopathies are characterised by selective neuronal loss in specific brain regions. In PD, dopaminergic neurons in substantia nigra pars compacta (SNpc) and dopaminergic terminals in the striatum are commonly affected [64, 104, 194]. In addition, PD also shows prominent extranigral pathology affecting the mesocortical dopaminergic system, the noradrenergic system with damage to the locus coeruleus, the serotonergic system with involvement of the dorsal raphe nucleus, the cholinergic system involving the nucleus basalis of Meynert and Westphal–Edinger nucleus, the centre–median parafascicular thalamic system, and many limbic structures including the amygdala [104, 105]. MSA has two major clinical subtypes. Cerebellar predominant MSA (MSA-C) presents with gait ataxia reflecting olivopontocerebellar atrophy, while Parkinson-predominant MSA (MSA-P) is characterised by levodopa-unresponsive parkinsonism reflecting striatonigral degeneration . Both motor subtypes feature progressive autonomic failure due to degeneration in autonomic brainstem centres, intermediolateral cell columns and Onuf’s nucleus in the spinal cord [73, 166, 254].
Attempts to stage the neuropathology of PD, DLB and MSA have been made reflecting the possible initiation and progression of these disorders in association with α-synuclein pathology. According to Braak’s staging in PD, LB pathology may first affect lower brainstem nuclei, including the dorsal motor nucleus of the vagus, the lower raphe nuclei, the locus coeruleus and the olfactory system (Braak stages 1–2). Then, the pathology can progress and lead to changes in SNpc, intralaminar thalamic nuclei, hippocampal CA2 and amygdala (Braak stages 3–4). Ultimately, involvement of the neocortex (cingulate, temporal, frontal and parietal cortex) may occur (Braak stages 5–6) [21–23]. However, the Braak staging scheme has also been contested for multiple reasons including the unclear pathogenic versus protective role of α-synuclein inclusions and the lack of clear correlations with neuronal loss and clinical course [30, 35, 106, 171]. Grading of MSA neuropathology has been proposed by Jellinger and colleagues, based on semiquantitative analysis of brain atrophy, neuronal loss, astrogliosis and GCI pathology in striatum, globus pallidus, SNpc, pontine basis, cerebellum, and inferior olives . Ozawa and colleagues further demonstrated significant correlation between the frequency of GCIs and the severity of neuronal cell loss, both associated with the disease duration, supporting the presumptive major role of oligodendroglial pathology in MSA [166, 253].
Glial pathology is a common finding in α-synucleinopathies and may play a key role in the neurodegenerative process . Neuropathological and in vivo imaging data suggest a role of microgliosis in α-synucleinopathies [71, 72, 100, 137, 164]. In an experimental study aggregated α-synuclein appeared to activate microglia and further, to promote dopaminergic neurodegeneration . The production of reactive oxygen species (ROS) and pro-inflammatory cytokines by activated microglia are linked to disease progression and neuronal cell death in mouse models of MSA and PD [182, 183, 218, 256]. Furthermore, reactive astrogliosis has been demonstrated in α-synucleinopathies and α-synuclein-positive inclusions in astroglia and oligodendroglia have been described in PD [24, 166, 240, 247]. Chemical signals from astrocytes affect microglia and therefore may play a fundamental mediator role in neurodegeneration [43, 231]. Finally, oligodendrocytes have a significant role in the pathogenesis of MSA, because α-synuclein aggregates predominantly in this cell type may lead to neuronal loss, reactive gliosis, iron deposition and myelin degeneration [142, 253]. In this review, we summarise the multiple physiological roles of glia, the effects of ageing on glial function and discuss the role of glia in α-synucleinopathies as well as the involvement of glial cells in the mechanisms of disease progression. Major therapeutic strategies targeting glial pathology in α-synucleinopathies as well as current pitfalls for disease-modification in clinical trials are discussed.
The multiple roles of glial cells in the CNS and their interactions
Glial cells represent around 90% of all cells in the human brain. Astroglia and oligodendroglia are of ectodermal origin, and are often referred to as “macroglial cells”, while microglia are of mesodermal origin .
Microglia and the general concept of microgliosis
With the activation of microglia through injury or disease, their morphology changes and they undergo functional transformations. Activated microglia change the movement of their processes from undirected to targeted towards the injured site . Further, microglia respond to activating stimuli through expression of neurotrophic factors or the release of pro- and anti-inflammatory cytokines . Major histocompatibility complex II (MHCII) and inducible nitric oxide (iNOS) up-regulation can be detected in activated microglia [90, 138]. When activated, because of certain immunological stimuli or injury, microglia can serve diverse beneficial functions to neuron survival, e.g. clearance of damaged or dead cells . Further, they may promote neuron survival through the release of anti-inflammatory or trophic factors  and also through the ability to protect already injured sites . Microglia can provide a physical barrier for the protection of the neighbouring healthy tissue . Trophic factors that are crucial for the viability of neurons include brain-derived neurotrophic factor (BDNF) and glial-derived neurotrophic factor (GDNF) can be released by microglia . The release of anti-inflammatory cytokines, such as transforming growth factor β (TGF-β) and interleukin 10 (IL-10) has also been described [103, 113, 132]. The activation of microglia can lead to an altered expression of cell surface markers, an increased process retraction with change to an amoeboid morphology, a higher proliferative capacity and migration to injured sites .
It is suggested that microglia can become over-activated under specific circumstances, which mostly lead to a chronic microglial activation as observed in different chronic CNS diseases. This up-regulation or accumulation of microglia in the CNS has been termed microgliosis. Microglial over-activation can lead to the release of deleterious and neurotoxic factors which facilitate chronic inflammation of the brain and further, neuronal dysfunction and death can develop . These include pro-inflammatory cytokines, such as tumour necrosis factor alpha (TNF-α) and IL-1β [7, 34, 45, 57, 175, 233] which may be directly toxic to neurons [45, 133, 175]. Furthermore, microglial production of large amounts of superoxide radicals, e.g. ROS and nitric oxide (NO), may contribute to the damage associated with chronic brain inflammation [145, 175, 233].
The involvement of microglial activation in neurodegenerative diseases is unequivocal . However, it is evident that microglia can have both neuroprotective and neurotoxic effects. A question that has to be answered is how do microglia switch from supporting neuronal function to a form where the cells become autoagressive effector cells that affect neurons and lead to neurodegeneration . In summary, microglial activation is complex, context dependent and further studies are needed to clearly define the beneficial versus deleterious role of microglia in the CNS.
Astroglia: structure and function
Astrocytes are the most numerous glial cell type in the CNS, and represent about one-third of the brain mass . Astroglia show two types of morphology: (1) protoplasmic astrocytes populating the grey matter , and (2) fibrous astrocytes populating the white matter . Protoplasmic astrocytes are important in wrapping synapses and neuronal cell bodies. In contrast, the fibrous type wraps oligodendroglial cells and nodes of Ranvier . Their bodies are shaped irregularly and they have leaflet-like processes . Astroglia show an even and non-overlapping spread throughout the whole CNS being responsible for many different essential functions in the healthy brain. Both subtypes establish extensive contacts with blood vessels and are therefore, important for energy as well as nutrient support of neurons [78, 118]. Astroglial cells form interconnections by gap junctions  and regulate the activity of large synaptic sets by covering many synapses with their leaflet shaped processes . Astrocytes are involved in the maintenance of the blood–brain barrier  and the regulation of blood flow [37, 118]. Furthermore, astroglia are important in the control of extracellular homeostasis of water, ions and neurotransmitters to optimise the interstitial space for synaptic transmission. Astroglia support synaptic transmission as part of the most frequent CNS synapse—the tripartite synapse. Astroglia can sense transmitter release from the neuronal terminal and modulate/control synaptogenesis and plasticity  and the potency of the synapse [8, 91, 245]. One function of astroglial cells is the control of the extracellular K+ concentration through local K+ uptake, thus acting as buffer of extracellular K+ . Control of water homeostasis through aquaporin (AQP) channels, mainly AQP4, is also an important astroglial task . Another major function is the uptake of glutamate by astrocytes. They can maintain a low concentration of glutamate in the extracellular space through the uptake by high affinity glutamate transporters (GLT1 and GLAST1 subtypes), which are enriched in astroglial processes. This uptake plays a crucial role in neurotransmitter clearance and in the prevention from reaching excitotoxic levels [180, 189].
The general concept of astrogliosis
The definition of reactive astrogliosis lacks uniformity among authors. Different types or categories of severity and intensity have been described. Sofroniew et al. suggest that astrogliosis is a finely graduated continuum of progressive changes (cellular and gene expression) and is regulated by specific mechanisms that can alter the nature and degree of these changes. Furthermore, it is proposed that astroglial changes can involve a gain and loss of function which may be beneficial or detrimental for the surrounding tissue . Others propose two different types: anisomorphic and isomorphic astrogliosis. Anisomorphic astrogliosis includes changes, such as cellular hypertrophy, astroglial proliferation and increased production of GFAP, vimentin and nestin [61, 128, 162]. This response arises after major tissue damaging injuries and astroglia are able to enhance tissue damage through the release of pro-inflammatory cytokines, e.g. the production of TNFα  and ROS or NO. Moreover, this glial response can result in glial scar formation, as described above. The isomorphic astrogliosis is a less severe response to mostly mild brain insults. It is characterised by morphological changes of astroglia, such as nuclear hypertrophy and stellar shape. GFAP expression is more transient and not so prominent. Besides, astroglia enhance the production of antioxidants, cytosolic proteins, soluble trophic and growth factors which are important for the recovery from tissue damage, beneficial for neuronal and glial survival and homeostasis. The isomorphic astrogliosis is considered the more common astroglial response in neurodegenerative diseases. It can adopt a premorbid cytoarchitecture over time, unlike anisomorphic astrogliosis which forms a permanent glial scar [41, 128, 178].
To summarise, astroglia are important for the maintenance of homeostasis and synaptic function in the healthy brain. Astroglia, similar to microglia, can have a beneficial function in the diseased or injured brain via the production of growth and trophic factors [41, 128]. However, they can also be neurotoxic, and enhance tissue damage when releasing ROS, NO and inflammatory cytokines [49, 151, 156]. Astrocytes present with a range of various responses to brain insults and therefore may provide a potential therapeutic target in CNS disease .
Oligodendroglia and oligodendroglial dysfunction
In contrast to microglia and astroglia, which respond to pathological stress by reactive gliosis, as described above, oligodendrocytes are characterised by increased vulnerability which leads to major cellular dysfunction, demyelination and often cell death. Oligodendroglial cells are highly susceptible to oxidative damage, because of their high metabolic rate with toxic by-products, their increased requirement for ATP, the high intracellular iron and low concentrations of the antioxidative glutathione [110, 237]. High oligodendroglial iron levels are very important in the production of myelin ; however, they are related to enhanced free radical formation and lipid peroxidation. The accumulation of intracellular hydrogen peroxide is further increased, because of the low levels of glutathione in oligodendrocytes . Oligodendroglia are also vulnerable to excitotoxic damage and effects of cytokines [131, 143, 264]. Glutamate and cytokines (e.g. TNF-α) released by activated microglia can cause oligodendroglial dysfunction and oligodendroglial apoptosis [53, 109]. LPS may trigger TLR4-mediated signalling by JNK (c-Jun-amino-terminal kinase) phosphorylation in oligodendrocytes  as well as induction of neuronal NOS (nNOS) in OPCs resulting in extensive demyelination .
In conclusion, the main role of oligodendroglia in the CNS is associated with the formation, maintenance, and restoration of myelin sheaths and the maintenance of normal axonal function, related to the support of neuronal survival. Multiple factors relevant to the specific properties of oligodendroglia, their environment and interactions with neighbouring cells may compromise the normal function of oligodendrocytes and lead to neurodegeneration.
Glial changes in normal ageing
Age is an important factor that increases the risk of neurodegenerative diseases. α-Synucleinopathies show increasing incidence and prevalence in the elderly population. PD is the most common neurodegenerative movement disorder affecting about 3% of the general population over the age of 65 . In MSA, mean age at onset of first reported symptom is 52–57 years [163, 199], with age-adjusted prevalence of 4.4 new cases per 100,000 year−1 . DLB accounts for up to 20% of all elderly cases of dementia coming to autopsy [146, 147].
To date, different reports evidence that microglia show morphological and phenotypic changes in the course of ageing. Morphological alterations of senescent microglia of the human cerebal cortex of non-demented individuals have been described . The dystrophic microglia are characterised by deramification, spheroid formation, shortened and twisted cytoplasmic processes, and fragmentation of processes . A decrease of total number of microglial cells in the hippocampal field CA1 was shown in 50- to 59-week-old mice . Furthermore, telomere shortening, a sign of cellular senescence, was demonstrated in rat microglia over time in vitro , as well as in vivo in rat cerebellum and cerebral cortex . Senescent microglia show an increase of immunophenotypic expression, including MHCII . Aged brains feature an increase of IL-1 positive microglia  and up-regulation of inflammatory cytokines, such as IL-6 and TNF-α . These reports indicate that microglial cells undergo profound changes during the ageing process and this could contribute to an increased inflammatory response in the brain. Enhanced inflammation may further lead to the progression of neurodegenerative diseases, although a direct correlation between activated microglia and neurodegeneration in human neurodegenerative disorders may be difficult to demonstrate. Age-related loss of microglia  may contribute to limitation of the beneficial effects which microglia exert including diminished phagocytosis and reduced trophic support.
Abnormal astroglial inclusions are common in neurodegenerative disease, but also during the ageing process of healthy elderly. Astroglial tau-inclusions have been reported in the subpial, subependymal and perivascular brain regions of healthy elderly . It is suggested that damaged mitochondria are responsible for the formation of peroxidase-positive inclusions and this can lead to failure in astrocytic ATP-dependent processes. It is hypothesised that brain areas with astroglial inclusion pathology may be prone to oxidative injury . However, mature, old and senescent astrocytes retain their antioxidant capacity in comparison to neonatal murine astroglial cells in cell culture . Ageing-prone astrocytes, cultured from senescence-accelerated mice, show a larger degree of gliosis than normal astrocytes, after amyloid-β and tau treatment. Besides gliosis, ageing-prone astroglia show a decrease in the ability to support co-cultured neurons, in comparison to normal astroglial cells . Changes in oxidative stress, tau phosphorylation and glutamate uptake of senescence-accelerated murine astrocytes have been observed consistent with reduced neuroprotective capacity in the ageing brain . Recently, further evidence in favour of deficient neuroprotection in senescence-accelerated mice was reported. These mice showed a defect in the cytokine-mediated neuroprotectivity of glial cells after excitotoxic hippocampal injury . In conclusion, changes in ageing astroglial cells occur and these changes can lead to a decrease in neuroprotection and further, they may accelerate neurodegeneration or glial degeneration .
Oligodendroglial changes caused by normal ageing mainly concern alterations in myelin dynamics, repair and maintenance, but also changes in total number of oligodendroglial cells. Age-dependent decline of oligodendrocytic fibres in the mouse hippocampal field CA1 has been observed . In an age-accelerated mouse a decrease of myelin basic protein (MBP) and 2′,3′-cyclic nucleotide 3′-phosphodiesterase (CNP), used as oligodendrocyte markers, was shown in the hippocampal CA1 in comparison to an ageing-resistant mouse strain . In contrast, no change of total number of oligodendroglial cells in the human principle inferior olivary nucleus during normal ageing was observed . A disruption of myelin sheath, caused by pro-inflammatory cytokines or by inappropriate protein maintenance within the oligodendroglial cells, has been demonstrated in the ageing rat [117, 229] and monkey [172, 206]. A similar disruption of myelin integrity was also reported in the ageing human brain . The total length of myelinated fibres in a Danish population showed there was a decrease of 10% of total length of myelinated fibres per decade and a decline of 45% across lifetime .
In conclusion, ageing may be an important factor which compromises the normal glial function and predisposes to occurrence or mediates the progression of late onset neurodegenerative disorders like α-synucleinopathies.
Microgliosis in α-synucleinopathies
All α-synucleinopathies feature microglial activation. In PD, reactive microglia assemble close to LBs [71, 72, 137, 144, 164]. In MSA patients, microglial activation has been detected using [11C](R)-PK11195 PET imaging . Ishizawa et al.  reported an up-regulation of activated microglia in motor-related structures, such as cerebellar input, extrapyramidal motor, and pyramidal motor structures, often associated to GCI pathology. In DLB, a positive correlation between LBs and microglial activation in different brain regions has been demonstrated . The different roles/phases of microglial activation in PD were recently discussed in relation to the disease initiation and progression . It was suggested that microglia are activated by astroglial signals or extracellular α-synuclein before a loss of neurons in the SNpc occurs [84, 227]. Recent evidence suggests that microglial glucocorticoid receptors (GR) may play a role in the regulation of dopaminergic neurodegeneration in parkinsonism . Furthermore, it is considered that microglia in the progression phase of PD display phagocytic features [84, 144, 150]. Various experimental models support the role of α-synuclein and its modified forms as potent activators of microglia [10, 115, 183, 228]. Enhanced dose-dependent release of pro-inflammatory molecules was reported upon α-synuclein treatment of primary microglial cells in vitro [115, 228]. Interestingly, PD-linked mutant forms of α-synuclein (A30P, E46K and A53T) caused increased microglial activation and release of pro-inflammatory cytokines (e.g. IL-6, IL-10) and chemokines (RANTES, MCP-1) as compared to wild type α-synuclein . Aggregated and nitrated α-synuclein induced inflammatory response and neuronal cell death in mesencephalic neuron-microglia co-cultures [182, 263]. α-Synuclein-induced dopaminergic neurotoxicity mediated by microglial cells was associated with activation of NADPH oxidase and ROS production, suggesting that oxidative stress induced by microglia may play a pivotal role in PD pathogenesis . Exposure of microglia to nitrated α-synuclein leads to the secretion of inflammatory, regulatory, redox-active enzymes and cytoskeletal proteins in vitro [115, 183]. Monomeric α-synuclein, but not β- or γ-synuclein, was shown to increase microglial phagocytosis in a dose- and time-dependent manner, while microglial phagocytosis was inhibited when adding aggregated α-synuclein to cell culture . Interestingly, the half-life of α-synuclein in microglia amounts up to 4 h, which is less than half compared to astrocytic and neuronal degradation of internalised α-synuclein .
The effects of α-synuclein on microglial activation have been further confirmed in in vivo models of α-synucleinopathies. Mice over-expressing wild type α-synuclein under control of a rat tyrosine hydroxylase promoter present early microglial activation  and the over-expression of mutant human α-synuclein (A53T and A30P homozygous double-mutants) may enhance this process . Moreover, it was demonstrated that microglia may feature different roles during disease progression in a rat PD model with rAAV-based overexpression of α-synuclein in midbrain . Four different types of microglial morphology were reported upon α-synuclein over-expression. Furthermore, microglial cell numbers in the SNpc correlated with the level of α-synuclein expression in the rat midbrain. α-Synuclein-induced activation of microglia in the absence of dopaminergic cell death was early and transient, while in the presence of dopaminergic cell death, α-synuclein-induced microglial activation was delayed and long-lasting . α-Synuclein alone triggered microglial activation and stimulation of the adaptive immunity in a PD mouse model with rAAV-based overexpression of human α-synuclein . Proteolipid protein (PLP)-α-synuclein mice, presenting a transgenic model of MSA with oligodendroglial overexpression of α-synuclein, developed chronic region-specific and age-dependent microgliosis correlating with progressive neuronal loss in SNpc. The early microglial activation in SNpc of MSA mice was associated with increased expression of iNOS . Early suppression of microglial activation and iNOS expression led to protection of nigral dopaminergic neurons, supporting the role of microglial activation in disease pathogenesis.
The response of microglia to various stimuli depends on different receptors displayed on the cell surface. The pattern-recognition receptors (PRR) Toll-like receptor 2 and 4 (TLR2 and TLR4) have been identified to play a crucial role in microglial activation [96, 130]. Their function is the identification of conserved structural motifs on a wide array of pathogens (pathogen-associated molecular patterns, PAMPs) . TLRs also recognise a wide number of endogenous molecules, which include misfolded proteins, a hallmark feature of neurodegenerative diseases. An up-regulation of TLR2 was identified in an α-synuclein PD/DLB mouse model, although in DLB human postmortem brains no up-regulation was detected . It has been suggested that the differences in the expression of TLR2 between the α-synuclein mouse models and the human disease may reflect higher disease dynamics in animal models . TLR4 up-regulation with NF-κB nuclear translocation was detected in human MSA and in a transgenic α-synuclein mouse model of MSA . Further analysis of the role of microglial TLR4 expression in the PLP-α-synuclein mouse model of MSA indicated its contribution to microglial phagocytosis and clearance of α-synuclein .
To summarise, microglia seem to play a fundamental role in the recognition and clearance of α-synuclein in α-synucleinopathies. However, microglial role in these diseases may also be fatal, because of their over-activation and release of pro-inflammatory cytokines and ROS after α-synuclein recognition . This can lead to a progressive loss of neurons in the diseased brain.
Astrogliosis in α-synucleinopathies
In PD, minimal astrogliosis has been demonstrated [95, 150, 246] as compared to extensive reactive astrogliosis in MSA . However, recent evidence suggests a prominent role of astrogliosis in PD pathogenesis . Furthermore, in PD astroglial cells may feature abnormal α-synuclein inclusion pathology, in contrast to MSA suggesting a different pattern of reactive astrogliosis in these disorders [24, 166, 247]. It appears that astroglial cells are important for the secretion of neurotrophic factors in these diseases . In contrast to beneficial functions, astroglia may also contribute to the production of pro-inflammatory agents. In PD brains, inflammatory response was shown to be mediated by astroglial cells positive for inter-cellular adhesion molecule 1 (ICAM-1) attracting reactive microglia expressing the counter receptor lymphocyte function-associated antigen 1 (LFA-1). In PD postmortem tissue, activated astrocytes displayed enhanced levels of IFN-γ receptor, suggesting that astrocytes activated through IFN-γ may mediate a toxic effect on neighbouring neurons . Myeloperoxidase (MPO), a key enzyme associated with oxidative stress during inflammation, has been found up-regulated in the ventral midbrain astroglia of human PD, suggesting a possible role of astrogliosis in the oxidative insult in PD . Experimental data further strengthen the postmortem evidence on the role of astroglia in the pathogenesis of α-synucleinopathies. Klegeris et al.  found enhanced levels of IL-6 and ICAM-1 in astrocytic cell cultures after α-synuclein activation. However, experimental studies also provide evidence for the production and release of neurotrophic factors by astroglia. Human postmortem data showed astroglial BDNF production, while cell culture experiments demonstrated the secretion of glial cell-line-derived neurotrophic factor (GDNF) under the conditions of oxidative stress . Astroglial GDNF and antioxidant GSH can be important neuroprotectors in PD-like neurodegeneration . Jakel et al.  demonstrated the neuroprotective function of the astroglial factor Nrf-2. Nrf-2, a NF-E2-related factor, can bind to an antioxidant response element and therefore induces antioxidant enzymes with neuroprotective capacity.
α-Synuclein released from neuronal cells can be endocytosed by astrocytes . This α-synuclein transfer is time-dependent and results in the formation of glial inclusions sharing some common features with LBs. α-Synuclein neuron-to-astroglia transfer further leads to changes in gene expression in astroglia related to drastic increases in TNFα and CXCL1 levels that may mediate enhanced neuroinflammatory response and neurodegeneration . Moreover, selectively expressed mutant A53T α-synuclein in astroglial cells in α-synuclein inducible transgenic mice induced a rapidly progressive paralysis, supposedly resulting from widespread astrogliosis, and midbrain dopaminergic neuron and spinal cord motor neuron degeneration . The transgenic overexpression of mutated α-synuclein has been further associated with mitochondrial damage in astroglia and a defective release of soluble factors supporting neuronal differentiation, which may represent an alternative pathogenic pathway of α-synuclein-related neurodegeneration mediated by astroglia .
Halliday and Stevens  have recently speculated that astrocytes may play a crucial role in the initiation of PD. They presumed that, if α-synuclein deposition is a primary event in the pathogenesis of PD, protoplasmic astrocytes are the first to respond by the accumulation of non-fibrillar α-synuclein in their cytoplasm, however, not undergoing changes typical for a classical reactive astrogliosis as seen in MSA. It is speculated that the pathogenic role of protoplasmic astroglia in PD, independent from a typical reactive response, may be related to interactions of α-synuclein with other proteins including β-synuclein and astrocytic neurotrophic factors. Next, astroglial dysfunction is proposed to mediate late disease progression by the expression of inflammatory molecules and recruitment of reactive microglia . In summary, astroglial cells appear to play a dual role in PD and maybe other α-synucleinopathies. On one hand, there is the production and release of pro-inflammatory cytokines and chemokines and on the other hand, astroglia can provide neurotrophic factors, which may support the survival of neighbouring neurons. Although differences in the histopathology of reactive astrogliosis in MSA versus PD have been observed as mentioned above, there is still insufficient data on the specific facets of astroglial response related to MSA, postulating the need of future studies to identify possible functional differences. Further analyses need to be accomplished to fully understand the specific roles of astroglial cells in α-synucleinopathies, the timing of reactive astrogliosis in these disorders and the relation to normal ageing deficits.
Oligodendroglia in α-synucleinopathies
There is currently no evidence suggesting that oligodendrocytes participate in the initiation of PD , whereas there is some association with the distribution of pathology, particularly in MSA [165, 253].
Filamentous α-synuclein inclusions have been reported in non-myelinating oligodendroglia in clinically overt PD cases [36, 247] suggesting a late secondary involvement. Further oligodendroglial involvement in PD and DLB has been evidenced by the presence of complement-activated oligodendrocytes in the SN of PD cases and in some brain regions of DLB cases [258, 259]. α-Synuclein-affected neurons may show lack or decrease of axon myelination suggesting oligodendroglial involvement [20, 21]. However, oligodendroglial pathology does not seem to be a leading factor in the pathogenesis of PD/DLB.
In contrast, oligodendroglial α-synuclein inclusions (GCIs) are the pathological hallmark lesion in MSA and may represent a primary injury [252, 253]. GCIs or Papp-Lantos inclusions were first described in 1989 by Papp et al. . They can be triangular, sickle, half-moon, oval or conical shaped. In 1998 α-synuclein was found to be the main component of GCIs [211, 249]. Further components of GCIs are ubiquitin and a large number of multifunctional proteins, such as leucin-rich repeat serine/threonine-protein LRRK2, heat shock proteins, microtubule-associated protein tau, prion disease-linked 14-3-3 protein among others . Their distribution in the CNS is extensive: GCIs occur in the pons, medulla, putamen, SN, cerebellum and preganglionic autonomic structures [17, 107, 158, 169]. The formation of α-synuclein aggregation in oligodendroglia has not been completely elucidated, but two hypotheses exist: (1) selective up-regulation of α-synuclein expression and impaired ability to degrade α-synuclein in oligodendroglial cells could lead to oligodendroglial accumulation [153, 184] or (2) active uptake of α-synuclein by oligodendroglia of protein released by neighbouring neurons may lead to dislocated abnormal accumulation of the protein in diseased oligodendrocytes .
Expression of α-synuclein has been shown during the early phases of maturation in rat oligodendrocytes in vitro , and in vivo a low level of α-synuclein in astroglial and oligodendroglial cells has been suggested . However, no α-synuclein mRNA expression has been detected in oligodendroglia of human control and MSA brains [149, 165].
The second hypothesis has become increasingly relevant in light of the recent findings on cell-to-cell propagation of α-synuclein [50, 86, 126]. α-Synuclein is released in culture media of neural cells overexpressing the protein, experimentally confirming its release in the extracellular space [60, 86]. Yet, the hypothesis of cell-to-cell transfer of α-synuclein specifically to oligodendroglia has not been proven in any in vivo graft experiment to date [86, 214].
Cytoplasmic α-synuclein aggregates may significantly affect the normal functions and viability of oligodendrocytes. In vitro evidence has indicated that overexpression of α-synuclein in glial cells leads to increased susceptibility to oxidative stress and TNF-α [215, 219]. Oligodendroglial cells overexpressing α-synuclein show decreased adhesion to fibronectin, relating the oligodendroglial cytotoxicity of α-synuclein to altered cell–extracellular matrix interactions . In a recent in vitro study, the role of α-synuclein and tau protein in oligodendroglia was examined . The formation of inclusion bodies with aggregated α-synuclein, tau and αB-crystallin under oxidative stress was demonstrated. However, in the absence of α-synuclein, no aggregation of tau could be found under the same conditions, suggesting a primary role for α-synuclein as pro-aggregatory protein in oligodendrocytes . p25α also known as tubulin polymerisation promoting protein (TPPP) is an oligodendroglia-specific phosphoprotein functionally linked to myelination . After phosphorylation of p25α its binding-affinity to microtubule decreases . Interestingly, p25α was shown to induce α-synuclein aggregation in vitro . Moreover, in MSA p25α was found to co-localise with α-synuclein-positive GCIs. Oligodendroglial cells within affected fibre tracts in MSA show abnormal redistribution and accumulation of p25α in the cell soma in contrast to the normal distribution in a thin perinuclear cytoplasmic sheet, as well as in the surrounding myelin. The dislocation of oligodendroglial p25α in MSA may represent early oligodendroglial pathology preceding GCI formation . The experimental coexpression of human α-synuclein and p25α in a rat oligodendroglial cell line resulted in a disorganisation of the microtubular cytoskeleton and induction of apoptosis . The response was dependent on Ser 129 phosphorylation of α-synuclein. Its inhibition abrogated the formation of phosphorylated α-synuclein oligomers, microtubule retraction and apoptosis .
Transgenic mouse models with overexpression of α-synuclein in oligodendroglial cells, using selective oligodendroglial promotors have been major contributors to evidence the role of oligodendroglial α-synucleinopathy . This approach has reproduced the hyperphosphorylation and insolubility of human α-synuclein aggregates . Mouse models have provided further evidence that neurodegeneration may be promoted by the accumulation of α-synuclein in oligodendrocytes, either through mitochondrial dysfunction, myelin disruption which causes axonal degeneration, or through inducing microglial activation [203, 218, 261]. In the PLP-α-synuclein mouse model, overexpression of human α-synuclein in oligodendroglia leads to progressive degeneration of brain areas, such as SNpc, locus coeruleus, nucleus ambiguus, pedunculopontine tegmental nucleus, laterodorsal tegmental nucleus and Onuf’s nucleus [217, 222]. In a recent study, it was demonstrated that the neurodegeneration in the MBP-α-synuclein mouse model is associated with altered expression of neurotrophic factors, especially GDNF released from oligodendroglial cells  providing a new insight in the possible pathogenic mechanisms of oligodendroglial α-synucleinopathy.
In summary, oligodendroglial cells may play a major part in the pathogenesis of MSA and rather a secondary role in other α-synucleinopathies. To date the origin of α-synuclein in MSA oligodendroglia is not clear; however, the pathogenic role of α-synuclein-positive GCIs is beyond question.
Glial dysfunction and therapeutic target definition in α-synucleinopathies
The impact of glia has been increasingly recognised in α-synucleinopathies. Therefore, possible therapeutic targets to regulate glial responses in α-synucleinopathies are of high interest. However, up to date, the origin and progression of glial dysfunction in α-synucleinopathies remain, to a great extent, unclear. The insufficient knowledge on the chronology of gliosis and on understanding the patterns of glial dysfunction in α-synucleinopathies, like in other neurodegenerative disorders, may be a main reason for the divergence between pre-clinical neuroprotection data and the limited success in a clinical setting.
Neuroinflammatory responses of activated glia related to up-regulation of pro-inflammatory mediators and pathways have been of interest in α-synucleinopathies like in other neurodegenerative disorders with reactive gliosis. Early studies indicated that non-steroidal anti-inflammatory drugs (NSAIDs) might protect against toxin-induced dopaminergic lesions in rodents [9, 62, 235]. Furthermore, NSAIDs, such as ibuprofen, indomethacin, aspirin and ketoprofen, were shown to inhibit the formation of α-synuclein fibrils in vitro in a dose-dependent manner as well as destabilise pre-formed α-synuclein fibrils . Inhibitors of iNOS and NADPH oxidase were found neuroprotective in transgenic models of PD featuring overexpression of A53T mutant or wild type α-synuclein and LPS-induced microglial activation [68, 69]. An interesting strategy to suppress the activation of glial cells and their inflammatory properties is the targeting of peroxisome proliferator-activated receptor γ pathway (PPARγ) achieved with the PPARγ agonist pioglitazone . In the MPTP mouse model, pioglitazone crossed the blood–brain barrier, attenuated MPTP induced micro- and astroglial activation and partially prevented dopaminergic cell loss [26, 46]. Pioglitazone was also effective to decrease microglial activation and protect dopaminergic neurons in rats after intrastriatal LPS injection which normally leads to inflammation-induced neurodegeneration . Myeloperoxidase (MPO), a key enzyme which is involved in the production of ROS has been shown up-regulated in both experimental and human PD and MSA associated with gliosis accompanying the neurodegeneration [40, 213]. Suppression of MPO activity by genetic ablation  or pharmacological inhibition  in experimental models of PD and MSA proved efficient to protect against neurodegeneration and α-synuclein aggregation, associated with reduced oxidative and nitrative stress. As aforementioned, TLR4 may play an important role in microglial activation related to α-synucleinopathies. Therefore, therapeutic implications to control TLR4 activation are of great interest in α-synucleinopathies. Naloxone, an opioid receptor antagonist and effector of TLR4 mediated up-regulation of TNFα, NO and IL-1β , provoked a reduction of dopaminergic neurodegeneration in the SNpc of PD models . However, recent observations in TLR4 knock-out mice with α-synucleinopathy suggest that in addition to mediating pro-inflammatory microglial response, TLR4 may be crucial in the microglial clearance of pathogenic α-synuclein, thus postulating caution in the use of TLR4 blockers in these disorders .
The experimental evidence on anti-neuroinflammatory strategies still need appropriate translation as a therapeutic approach in α-synucleinopathies under clinical conditions. While to date the successful experimental treatments in animal models of α-synucleinopathies have aimed primarily at early detrimental glial changes to provide neuroprotection, in patients with longer standing α-synuclein disease, downstream glial events should be considered. This is one of the issues which may cause the divergence between the outcomes of pre-clinical and clinical studies.
Epidemiological studies suggested possible neuroprotection by anti-inflammatory medication (NSAIDs), related to reduced risk of PD ; however, this hypothesis could not be confirmed by a recent study using data of PD cases and controls from the UK primary care , and by meta-analysis of observational studies evaluating NSAID use and the risk of PD [173, 191]. To date there has been no clinical trial confirming unequivocally disease-modification through anti-inflammatory treatment. A prime example for an anti-inflammatory drug which has gone through pre-clinical studies and clinical trials in α-synucleinopathies is minocycline. Minocycline is a tetracycline antibiotic that crosses the blood–brain barrier and has anti-inflammatory along with its anti-apoptotic activity . This drug has proven at least partial neuroprotective efficacy in various experimental models of PD and MSA [38, 216, 218, 256]. However, clinical trials in PD and MSA had inconsistent outcomes regarding the disease-modifying potential of minocycline therapy [52, 177]. Interestingly, in a prospective, randomised, double-blind clinical trial of minocycline in MSA, microglial activation was significantly reduced after 24 weeks therapy as shown by [11C](R)-PK11195 PET, but no clinical effect on symptom severity could be achieved probably due to the late start of the therapy .
Immunomodulation has been an alternative approach proposed to target glial dysfunction in α-synucleinopathies. Adaptive T cell-mediated immunity may play a part in the modulation of neuroinflammatory response in α-synucleinopathies. CD8+ and CD4+ T cells were shown to invade the PD brain. Experimental data suggested that T cell-mediated dopaminergic toxicity was mainly controlled by CD4+ T cells in a FAS-dependent way . Although these data need further confirmation, they suggest a possible rationale for targeting adaptive immunity in PD which may involve developing vaccines for antigens that promote cell-mediated anti-inflammatory responses. In support of this strategy, preclinical studies showed that adoptive transfer of copolymer-1 immune cells leads to T cell accumulation within SNpc, suppression of microglial activation, and increased local expression of astrocyte-associated GDNF [16, 181]. Alternatively, the adoptive transfer of T cells from mice immunised with nitrated α-synuclein resulted in a robust neuroinflammatory response with accelerated dopaminergic cell loss . The immunisation approach has been further evaluated in experiments applying transgenic mice overexpressing human α-synuclein. Active immunisation with recombinant α-synuclein in this mouse model of PD induced the production of antibodies against human α-synuclein and resulted in modulation of microglial and astroglial response, reduction of the inclusion pathology and neuroprotection . Immunomodulatory strategies await clinical confirmation and approval in α-synucleinopathies.
Neurotrophic factor delivery and modulation is considered a further major candidate therapeutic strategy in α-synucleinopathies. Glial dysfunction in these disorders may directly relate to the reduction of trophic support. For example, in a transgenic mouse model of MSA, GDNF reduction was directly associated with α-synuclein inclusion pathology in oligodendroglia and intracerebroventricular infusion of GDNF ameliorated neurodegeneration . Alternatively, glial reactivity, typical for the neurodegenerating brain in α-synucleinopathies, may be used as a tool to supply targeted delivery of trophic factors and provide neuroprotection. The adoptive transfer of CD3-activated regulatory T cells to MPTP-intoxicated mice triggered protection of the nigrostriatal system and modulation of microglial responses, associated with up-regulation of astrocyte-derived GDNF . Further, microglial activation following MPTP intoxication has been successfully used to confer neuroprotection by targeted experimental brain delivery of GDNF through genetically modified macrophages . However, AAV gene delivery of neurturin, a structural and functional analogue of GDNF, in the putamen of PD patients failed to improve the motor performance from baseline to 12 months versus placebo .
In summary, based on the wide range of pre-clinical experimental data, glial dysfunction seems to be a prominent target for disease-modification in α-synucleinopathies. Unfortunately, clinical translation of these strategies is still lacking success due to many undefined variables associated with temporal and spatial initiation and progression of glial pathology, regulatory mechanisms as well as definition of the transitional border between beneficial and detrimental properties of glia in α-synucleinopathies.
This review was supported by grants of the Austrian Science Funds (FWF) P19989-B05 and SFB F44-B19.