In vitro models of multiple system atrophy from primary cells to induced pluripotent stem cells

Abstract Multiple system atrophy (MSA) is a rare neurodegenerative disease with a fatal outcome. Nowadays, only symptomatic treatment is available for MSA patients. The hallmarks of the disease are glial cytoplasmic inclusions (GCIs), proteinaceous aggregates mainly composed of alpha‐synuclein, which accumulate in oligodendrocytes. However, despite the extensive research efforts, little is known about the pathogenesis of MSA. Early myelin dysfunction and alpha‐synuclein deposition are thought to play a major role, but the origin of the aggregates and the causes of misfolding are obscure. One of the reasons for this is the lack of a reliable model of the disease. Recently, the development of induced pluripotent stem cell (iPSC) technology opened up the possibility of elucidating disease mechanisms in neurodegenerative diseases including MSA. Patient specific iPSC can be differentiated in glia and neurons, the cells involved in MSA, providing a useful human disease model. Here, we firstly review the progress made in MSA modelling with primary cell cultures. Subsequently, we focus on the first iPSC‐based model of MSA, which showed that alpha‐synuclein is expressed in oligodendrocyte progenitors, whereas its production decreases in mature oligodendrocytes. We then highlight the opportunities offered by iPSC in studying disease mechanisms and providing innovative models for testing therapeutic strategies, and we discuss the challenges connected with this technique.

Symptom onset occurs usually in the sixth decade, and the median survival from that time is estimated to be 9.8 years. 9,10 In the majority of cases, MSA appears sporadically in the population. However, some cases have been reported of Japanese, German and American families showing a genetic transmission of the disease. [11][12][13][14][15][16] Despite these findings, no disease-causing mutation could be identified in these studies.
COQ2 is a gene that encodes the enzyme parahydroxybenzoatepolyprenyl transferase, which catalyses one of the final reactions in the biosynthesis of coenzyme Q10 (ubiquinone). Intriguingly, whole genome sequencing of a Japanese family and large case-control series revealed that COQ2 variants may be associated with an increased risk of MSA in East Asia. 17,18 However, other groups failed to report the same correlation between MSA and COQ2 variants in western countries. [19][20][21] These findings notwithstanding, the role of COQ2 in the pathogenesis of MSA remains unclear.
In addition to that, Gaucher disease-causing mutations of GBA (glucocerebrosidase) gene were recently sequenced in 969 MSA patients and in 1509 control subjects, demonstrating an association between many GBA variant and MSA, as it happens in Parkinson's disease (PD). 22 Finally, genomewide association studies (GWAS) have pointed out possible correlation between alpha-synuclein encoding gene (SNCA) polymorphisms and MSA. 23,24 However, the largest MSA GWAS did not find any relevant association. 25 Additional studies with more samples may shed new light on potentially significant associations in future.
At a histopathological level, the main features of MSA are selective neuronal loss and axonal degeneration, alpha-synuclein immunoreactive inclusions and gliosis. The pathologic hallmark of MSA is insoluble GCIs, whose main proteinaceous component is alpha-synuclein. 26 Thus, the presence of GCIs characterizes MSA as a "synucleinopathy," together with PD and Lewy Body Dementia (LBD). In contrast to inclusions in PD and LBD, however, GCIs predominantly accumulate in oligodendrocytes. 27 Other relevant proteins that can be found in aggregates are p25alpha/TPPP, LRRK2 and tau protein. 28,29 GCIs are diffusely distributed in specific anatomical regions of the CNS. More specifically, the pyramidal, striatonigral, corticocerebellar and preganglionic autonomic systems are affected in a preferential manner. 30 GCIs are also widely present in the motor cortex, despite mild levels of cellular degeneration. 31 Nonetheless, glial alterations in the white matter are not limited to oligodendrocytes, but involve also astrocytes and microglia. In fact, the extent of reactive astrocytosis and activated microglia parallels the degree and anatomical distribution of GCIs and neurodegeneration. 32,33 Neuronal cytoplasmic inclusions (NCIs) and neuronal nuclear inclusions (NNIs) can also be found in MSA, although less frequently than GCIs. 34 They are mainly found in cortical, subcortical, cerebellar and brainstem nuclei, being especially prevalent in the pons and inferior olivar nucleus. 35

SYSTEM ATROPHY
To this day, the pathogenic mechanisms that lead to the development of MSA are yet to be unravelled. Nevertheless, there is compelling evidence that MSA is a primary oligodendrogliopathy, which encompasses alpha-synuclein misfolding and aggregation, early myelin dysfunction and axonal disease ( Figure 1). 36,37 Early myelin dysfunction is suggested by the finding that p25alpha (TPPP), usually found in myelin sheaths, relocates in the oligodendrocyte soma in the first stages of the disease. 38 Moreover, the co-localization of p25alpha and MBP is noticeably decreased in MSA brains, and total MBP content is reduced. The presence of p25alpha in the body of the cell could then enhance the aggregation of alphasynuclein. 39 The interaction between p25alpha and alpha-synuclein may result in the formation of GCIs, which in turn interferes with oligodendrocyte survival and neuronal support. Alpha-synuclein is thought to play a major pathogenic role in the disease, although it is unclear whether it is normally expressed in oligodendrocytes. Analysis of postmortem healthy controls' brains yielded contrasting results: in situ hybridization techniques failed to detect alpha-synuclein mRNA expression in glial cells, 40,41 whereas a recent study identified SNCA mRNA in oligodendrocytes using qPCR. 42 Djelloul and colleagues sought evidence of the expression of alpha-synuclein in oligodendrocyte lineages derived from mouse embryonic stem cells (ESC) and human-induced pluripotent stem cells (iPSCs) and in oligodendrocytes from mice postnatal forebrains. They found that alphasynuclein was expressed in oligodendrocytes progenitors, but its levels decreased during maturation and were absent in the final stages of differentiation. 43 The origin of GCIs' alpha-synuclein in MSA therefore is not clear.
One hypothesis is that it results from endogenous overexpression in oligodendrocytes, notwithstanding that many studies contradict this view. The in situ hybridization studies previously described did not retrieve SNCA mRNA in MSA patients' white matter. 40,41 Analysis by qPCR revealed the presence of alpha-synuclein transcripts but the difference with PD patients and healthy controls was not statistically significant. 42 Furthermore, in vitro culturing of MSA-derived iPSCs showed that, as in control and PD-derived lines, alpha-synuclein was expressed in the first stages of oligodendrocytes' development but not in the premyelinating phase. 43 On the contrary, upregulation of synuclein expression has been demonstrated to impair the correct maturation of the oligodendrocyte. 44 Another theory claims that alpha-synuclein might be produced in neurons and then taken up by oligodendrocytes. Several studies advocate this hypothesis, showing that oligodendrocytes are capable to absorb neuronal secreted or exogenously added alpha-synuclein both in vitro and in vivo. 45,46 Under physiological conditions, alpha synuclein is primarily produced by neuronal cells as an unfolded protein. 47 Several studies showed that alpha-synuclein aggregates can transmit from neuron to neuron, 48 astroglial cells 49 and oligodendroglial cells, 47 thus supporting the hypothesis of neuronal ABATI ET AL. | 2537 transmission. Oligodendrocytes are then able to take up alpha-synuclein via endocytotic mechanisms. 50 Hansen and colleagues provided evidence that alpha-synuclein may propagate from cell to cell and exert a seeding effect on the endogenous protein, thereby contributing to the spread of the pathology. 51 This discovery, along with a recent study by Prusiner and colleagues, suggests that alpha-synuclein could disseminate in a prion-like fashion. 52 The connection between alpha-synuclein accumulation and neurodegeneration is still a matter of debate. According to some studies, alpha-synuclein may have a role in activating intrinsic and extrinsic apoptotic pathways within oligodendrocytes. 53,54 Oligodendrocytes' dysfunction then affects neuronal survival, 55 for example through a reduction in glial derived neurotrophic factor (GDNF). 56 Moreover, in vitro studies demonstrated that alpha-synuclein aggregates directly induce neuronal dysfunction and apoptosis. 48,57,58 Desplats and colleagues demonstrated the formation of inclusion bodies in neurons after alpha-synuclein uptake, possibly through lysosomal dysfunction, and apoptosis of involved neurons. 48 Another study showed that exogenous alpha-synuclein fibrils induce pathological alpha-synuclein accumulation, neuron loss and diminished levels of synaptic proteins. 58 Klucken and colleagues found that inhibition of autophagy with bafilomicin A1 increased toxicity, measured as the release of adenylate kinase, in neurons transfected with C-terminal modified a-synuclein. 59 Impairment of autophagic pathways has already been reported in nigral neurons of PD patients' brains 60 and evidence for a potential role of autophagy in MSA pathogenesis is emerging from in vitro and in vivo studies. 61,62 Furthermore, there is evidence that also inflammatory response, whose main actors are microglia and astrocytes, plays an active role in perpetuating and extending brain damage. Activated Iba-1-positive microglia and GFAP-positive astrocytes are shown to colocalize with GCIs. 63 Moreover, treatment of primary astrocytes with alpha-synuclein triggered astrogliotic changes, whereas extracellular alpha-synuclein is phagocytosed by microglia inducing microgliosis and production of reactive oxygen species (ROS). 64 In particular, the Tolllike receptors 2 and 4 are reported to interact with alpha-synuclein and exhibit upregulation in MSA patients. [65][66][67] Finally, it is suggested that the release of cytotoxic products by activated glia may favour alpha-synuclein misfolding and aggregation. 68 In the light of the recent progress on the pathophysiological mechanism, we now have a better understanding of how oligodendrocytes' dysfunction and alpha-synuclein accumulation develop in the human CNS. However, as symptoms in MSA patients appear to be caused by neuronal degeneration, and not by demyelination in oligodendrocytes, the molecular interactions between the degenerated oligodendrocytes have to be better elucidated.
Among the hypothesized mechanisms, oligodendrocytes' dysfunction might cause neuronal death through the activation of neuroinflammatory mechanisms [63][64][65][66][67] and the loss of neurotrophic support. 56 Neuronal dysfunction because of a-synuclein inclusions 48,57,58 and autophagy impairment 60-62 also act synergistically, leading to neuronal death in the striatonigral, olivopontocerebellar and central autonomic pathways. 31 This secondary neurodegeneration may explain the typical symptoms observed in MSA patients, the lack of response to L-DOPA and the rapid progression of this disease.
Demyelination plays an important role in advanced MSA, 33 and recent studies found that intracellular alpha-synuclein delays oligodendrocytes maturation and myelination by downregulating myelingene regulatory factor and myelin basic protein (MBP). 44,69 Myelin dysregulation is often followed by axonal degeneration, 36 as demonstrated by transgenic animal models. 70,71 Several other mechanisms have been examined as potentially pathogenic, such as proteasome system inhibition. 72 A recent field of investigation is focusing on exosomes, small extracellular vesicles which are involved in the reciprocal communication between oligodendrocytes and neurons, in neural trophic support and in the regulation of microglial response. Exosomes appear able to catalyse and accelerate the nucleation of alpha-synuclein. Exosomes are also suspected of playing a role in the prion-like spread of proteins, such as alpha-synuclein, in neurodegenerative diseases. [73][74][75] In conclusion, many pathways, from gene expression to protein transport and inflammatory response, appear to be involved and to interact over the course of the disease. However, further investigations are needed to establish their precise role and weight in the pathogenesis of MSA.
F I G U R E 1 Hypothetical features of multiple system atrophy pathophysiology. Early in the course of the disease, p25alpha relocalizes into the oligodendroglial soma. Subsequently, altered expression or uptake of alpha-synuclein in oligodendrocytes and interaction with p25alpha causes the formation of glial cytoplasmic inclusions, which eventually determine oligodendroglial dysfunction and loss of neurotrophic support. Misfolded alpha-synuclein can also be taken up by neurons, with the formation of neuronal cytoplasmic and nuclear inclusions. Defective autophagic clearance mechanisms promote the accumulation of intracellular alpha-synuclein at an increased rate. Together with microglial activation, these factors ultimately lead to neurodegeneration and neuronal death

IN N EURODEGENERATIVE D ISEASES
In 2006, Takahashi and Yamanaka for the first time successfully reprogrammed somatic cells, mouse fibroblasts at the beginning, into embryonic like cells, the so-called iPSCs. 76 The induction of pluripotency in vitro can be achieved by overexpression of defined cocktail of pluripotency-associated genes, or reprogramming factors, namely Oct3/4, Sox2, Klf4 and c-Myc (OSKM), that are transduced in somatic cells, usually fibroblasts, with a retroviral or lentiviral system. 77,78 This procedure allows to obtain cells similar to ESCs in morphology, proliferation, surface antigens, gene expression, epigenetic status and telomerase activity.
The mechanisms by which reprogramming factors exert their action have not been completely elucidated. Oct3/4 and Sox2 upregulate stemness genes and suppress differentiation-associated genes, acting synergistically. 79 The role of Klf4 and c-Myc is less clear, but it has been proposed that Klf4 favours epithelial transition by binding to specific genes, while c-Myc seems to be involved in the regulation of cellular proliferation, metabolism, and biosynthetic pathways. 80,81 The advantages of iPSCs over ESCs are numerous and relevant.
Firstly, while the use of ESCs is limited by ethical concerns because of their embryonal origin, iPSCs are produced from adult cells and thus they overcome this issue. Secondly, they retain the same genetic makeup as the original source. Therefore, they allow the production of patient-specific disease models and provide hope for the develop- The group also found out that OPC content is increased in the striatum of MSA-P patients and that mice overexpressing human wild-type alpha-synuclein also display an increased number of striatal OPCs. Thus, they hypothesized that alpha-synuclein may impair adult oligodendrogenesis, preventing OPCs from remyelination and contributing to MSA pathogenesis. Furthermore, the group found that BDNF mRNA is significantly reduced in the striatum of MBP transgenic mice, and that the supplementation of BDNF in vitro to transfected oligodendroglial cells partially rescues early OPCs' maturation, but lacks the potential to induce myelination.
Recently, Valera and colleagues generated a rodent oligodendroglial cell line, CG4, co-infected with Lentivirus expressing human alpha-synuclein or control and microRNA-101 (miR-101a-3p) or control vector. 61 The aim of the study was to investigate the potential

| Achievements and limitations of primary cell models
These studies played a fundamental role in defining some key events in MSA pathogenesis. An important result that was observed is that overexpression of a-synuclein in a human and rat primary mixed glial culture is sufficient to produce widespread fibrillar a-synuclein aggregates and to trigger cellular stress and degeneration. 106 In addition to that, it was demonstrated that the accumulation of alphasynuclein in OPCs may downregulate myelin-associated genes and impair adult oligodendrogenesis. 69  Presence of widespread fibrillar a-syn aggregates, more numerous in cells expressing the C-terminally truncated form; increased cell death rates; increased susceptibility to treatment with TNFa. 106 Kragh et al (2009) Oligodendroglial cell line (OLN-93) derived from primary Wistar rat brain glial cultures expressing human WT a-syn or S129A or S129D mutant asyn with human WT p25a.
Coexpression of a-syn and p25a causes microtubule relocalization to the perinuclear region; p25a-mediated microtubule retraction requires low levels of a-syn; asyn-dependent microtubule retraction induces apoptotic markers with activation of caspase-3 and nuclear chromatin condensation.

| Differentiating iPSCs into oligodendrocytes
In this regard, one of the main challenges is represented by the nature of the cells targeted by MSA. In fact, oligodendrocytes' maturation in vitro is longer and more complex than neuronal one. 114,115  NSCs and to upregulate Olig2, thereby generating the highest yield of Olig2+ progenitors. 115,117 This first part of differentiation, which includes the induction of neuroepithelium and then of Olig2-expressing progenitors, generally takes place in adherent cultures.
Subsequently, continuous stimulation of SHH pathway is necessary to induce the formation of Olig2+, Nkx2.2+ pre-OPCs. In contrast to the previous stage, transition from neuroepithelium to pre-OPCs is best carried out as floating aggregates. 115 Differentiation to Sox10+, PDGFR+ OPCs is a long process that may take up to 10 weeks. In order to boost the transition, most protocols introduce between the third and the fifth week a cocktail of factors known to drive oligodendrocyte differentiation or to promote oligodendrocyte survival, namely platelet-derived growth factor (PDGF), neurotrophin 3 (NT3), triiodo-L-thyronine (T3), insulin-like growth factor 1 (IGF-1) and hepatocyte growth factor (HGF). 114,115,117 Last steps of differentiation require the withdrawal of mitogenic factors and result in the production of O4+ immature oligodendrocyte and finally in the appearance of MBP+ ramified, mature oligodendrocytes. 115

| Generation of an iPSC-based MSA model
MSA oligodendrocytes have been successfully generated by Djelloul and colleagues (Table 1). 43 In their work, they generated iPSCs from

| Relevance of the first iPSC-based MSA model
This study suggests that oligodendrocytes might be able to produce alpha-synuclein in vivo, at least during the first stages of maturation. Mixed oligodendroglial and neuronal cultures could also represent a useful tool to study trophic interaction between neurons and oligodendrocytes. A reduction in GDNF was observed in transgenic mouse models of MSA, and it is believed to play a role in neurodegeneration. 56  was observed that the excessive production of alpha-synuclein at an oligodendroglial level causes the formation of fibrillary alpha-synuclein aggregates and favours cell death. 106 In another study, the concomitant expression of p25alpha was shown to induce alphasynuclein aggregation and the subsequent activation of the apoptotic cascade. 105 Furthermore, it was seen that overexpression of alphasynuclein in OPCs impaired their normal maturation and myelination, possibly through a downregulation of myelin-associated genes. 69 In addition to that, the presence of a dysregulation of autophagic pathways and its contribution to the intracellular deposition of alphasynuclein was analysed. 61 However, these models have several flaws, as they are not based on patient-derived cultures and thus they do not share the same genetic background as patients. Moreover, it is not known whether alpha-synuclein is primarily overexpressed in oligodendrocytes or if MSA neuropathological features originate from the interaction between neurons and oligodendrocytes. Therefore, models based on primary cell cultures are of limited use in drug testing.
Conversely, the advent of iPSC technology opens up thrilling possibilities for the study of neurodegenerative diseases such as MSA. iPSC-based models are human relevant, and they retain the same genetic inheritance as the patient. Thus, they raise hopes on the possibility of testing drugs in a safe and reliable manner, and developing treatments based on autologous stem cell transplantation. Furthermore, the use of stem cells allows researchers to observe neurons and glial cells during their differentiation and maturation, making it possible to identify early events that could trigger late neurodegeneration. Djelloul and colleagues were the first to generate oligodendrocytes from iPSCs of patients with MSA. Their work shows that alpha-synuclein is produced in oligodendrocytes progenitors and registers a significant decrease during maturation, but no differences were observed between healthy and MSA cell lines. 43 However, it was noted that disease lines generated a higher yield of O4+ progenitors at an accelerated rate. Although these results suggest the hypothesis that GCIs' alpha-synuclein might not originate from oligodendrocytes, further models are needed to confirm these findings.
iPSC technology also poses a variety of challenges. The absence of known genetic or environmental culprits makes it difficult to determine whether MSA is one disease or many. As a consequence, the validity and significance of in vitro results obtained from patients' iPSCs needs to be confirmed with a great number of observations.
In conclusion, although we are far from the understanding of disease mechanisms, future research may take advantage from the uncountable opportunities offered by IPSCs. For example, co-cultures of neurons and oligodendrocytes may shed light on the cellular origin of alpha-synuclein. Furthermore, iPSC-based tridimensional cellular cultures, or organoids, may provide an excellent insight into MSA pathogenesis, as they reproduce complex cellular interaction in a near-physiological environment. 119 In addition to that, MSAderived iPSC cultures may allow scientists to focus on molecular changes that occur in patients prior to neurodegeneration or symptoms onset. The identification of markers of subclinical disease may be the first step towards early diagnosis and effective pharmacological interventions.