The role of Galectin-3 in the pathogenesis of Parkinson's Disease
Abstract
En 1817, James Parkinson describió por primera vez un conjunto de síntomas motores en el ensayo titulado “Shaking Palsy” (Parálisis temblorosa). Este conjunto de síntomas fue posteriormente renombrado como Enfermedad de Parkinson (EP). Durante estos dos siglos, los avances en la investigación de la EP han revelado un papel fundamental para la proteína neuronal α-sinucleína (αSyn). Esta proteína, cuyas funciones fisiológicas están aún por determinar, forma inclusiones citoplasmáticas características de la EP llamadas Cuerpos de Lewy. Junto a la aparición de estos Cuerpos de Lewy, la EP se caracteriza por la degradación específica de la substancia nigra, un área dopaminérgica del mesencéfalo cerebral involucrada en el control del movimiento. En la actualidad, las diversas hipótesis sobre la etiología de la EP no están logrando transformarse en mejores tratamientos o un diagnóstico más temprano, por ello, la búsqueda de nuevas moléculas implicadas en el origen y la progresión de la EP se revela indispensable para desarrollar nuevas estrategias terapéuticas. Recientemente se ha relacionado a la proteína Galectina-3 (Gal3) con los Cuerpos de Lewy en cerebros post-mortem. Gal3 es una proteína de unión a galactósidos sin actividad catalítica y expresada principalmente por células de la microglía, las células inmunes del sistema nervioso central, en procesos neuroinflamatorios. En la presente tesis doctoral, tratamos de describir el papel de Gal3 en la patogénesis de la EP basándonos en muestras post-mortem, in vitro e in vivo. En estudios humanos post-mortem, hemos podido asociar esta proteína al proceso de agregación que da lugar a los Cuerpos de Lewy. Los estudios in vitro nos han permitido detallar los procesos de agregación de αSyn y captación neuronal de Gal3. Por su parte, hemos desarrollado un modelo in vivo de sobreexpresión de αSyn humana durante 6 meses en ratones salvajes y carentes de Gal3 (Gal3 knockout). Nuestros resultados confirman la presencia de Gal3 en muestras post-mortem tanto de pacientes con Demencia de Cuerpos de Lewy como de EP. Como observación relevante, hemos relacionado la presencia de Gal3 preferentemente con acumulaciones menos densas de αSyn pero no con los Cuerpos de Lewy. Igualmente, hemos descrito un papel fundamental de Gal3 durante el proceso de agregación de αSyn in vitro, donde Gal3 bloquea y revierte dicho proceso de agregación. Además, hemos demostrado que las neuronas son capaces de incorporar Gal3 del medio extracelular y que esta Gal3 se une rápidamente a las acumulaciones de αSyn. Por último, nuestras observaciones in vivo corroboraron estos datos, demostrando un gran impacto de la deleción de Gal3 en el modelo animal de sobreexpresión de αSyn humana mediada por adenovirus. Los animales carentes de Gal3 desarrollaron un mayor número de agregados de αSyn, pero no mostraron neurodegeneración de la substancia nigra ni desarrollaron síntomas motores, lo que supone que la deleción de Gal3 no solo preservó la supervivencia de las neuronas dopaminérgicas sino también preservó su funcionalidad. Asimismo, hemos identificado la aparición de un fenotipo microglial basado en los marcadores TNFα, Trem2 y MerTK que podría ser específico en la EP. Este fenotipo se ve abolido en ausencia de Gal3, lo que pone de manifiesto la importancia de esta proteína en la activación microglial asociada a neurodegeneración. En definitiva, nuestro trabajo demuestra un importante papel de Gal3 en la patogénesis de la EP y en la viabilidad de las neuronas dopaminérgicas. Los datos que se muestran sugieren que Gal3 está inherentemente relacionada al proceso de agregación de αSyn, su toxicidad y al proceso neuroinflamatorio que desencadena.
Full text
JUAN GARCIA REVILLA Departamento de Bioquímica y Biología Molecular Facultad de Farmacia. Universidad de Sevilla Instituto de Biomedicina de Sevilla
INDEX RESUMEN ............................................................................................................................ 1 ABSTRACT ........................................................................................................................... 3 INTRODUCTION ................................................................................................................... 4 Parkinson’s Disease ................................................................................................................... 4 Historical view ....................................................................................................................... 4 Pathophysiology .................................................................................................................... 8 Epidemiology ....................................................................................................................... 17 Treatments .......................................................................................................................... 18 α-synuclein: The killer queen .................................................................................................. 19 Historical perspective .......................................................................................................... 19 Structure and physiological functions ................................................................................. 20 Lewy Bodies, aggregation and clearance ............................................................................ 22 Synucleinopathies ............................................................................................................... 24 Microglia: The catcher in the brain ......................................................................................... 25 Historical perspective .......................................................................................................... 25 Microglia homeostatic functions......................................................................................... 26 Neuroinflammation ............................................................................................................. 28 New activation patterns ...................................................................................................... 30 Microglia in Parkinson’s Disease ......................................................................................... 31 Galectin-3: The Theory of Everything...................................................................................... 32 Family of galectins ............................................................................................................... 32 Galectin-3 ............................................................................................................................ 33 Galectin-3 in the central nervous system ........................................................................... 35 Galectin-3 in Parkinson’s Disease and neurodegeneration ................................................ 37 METHODS.......................................................................................................................... 39 Animals and cultures ............................................................................................................... 39 Genotyping .......................................................................................................................... 39 Intracranial injection ........................................................................................................... 40 Treatment and Sacrifice ...................................................................................................... 41 Cell lines .............................................................................................................................. 42 PFFs optimisation ................................................................................................................ 44
Behavioural Tests .................................................................................................................... 45 Open Field ........................................................................................................................... 45 Rotarod ................................................................................................................................ 46 Cylinder test ........................................................................................................................ 46 Immunofluorescence .............................................................................................................. 46 Stereological cell counting .................................................................................................. 47 Dendritic tree measure ....................................................................................................... 48 LB-like counting ................................................................................................................... 48 Human immunofluorescence .............................................................................................. 48 Immunofluorescence of cell cultures .................................................................................. 49 Protein Analysis ....................................................................................................................... 49 Protein extraction and quantification ................................................................................. 49 Western Blot ....................................................................................................................... 50 Cytokine quantification ....................................................................................................... 51 ThT Aggregation Assay ............................................................................................................ 51 Gene expression analysis ........................................................................................................ 52 RNA extraction .................................................................................................................... 52 RT-PCR ................................................................................................................................. 53 Statistical analysis ................................................................................................................... 54 RESULTS ............................................................................................................................ 55 Galectin-3 is present in αSyn aggregates from DLB and PD patients ..................................... 55 Neurons can actively uptake exogenous Gal3 in vitro. Association to pathological αSyn aggregates ............................................................................................................................... 59 Gal3 inhibits and reverse αSyn aggregation in vitro ............................................................... 61 AAV5-αSyn nigral injection as a model for PD study .............................................................. 62 Analysis of hSYN in the AAV5-αSyn model .............................................................................. 66 Quantification of monomeric and tetrameric hSYN in SN and STR .................................... 66 Analysis of hSYN distribution .............................................................................................. 67 Striatum analysis ................................................................................................................. 68 Analysis of pSyn in the AAV5-αSyn model .............................................................................. 71 Quantification of pSyn in SN ............................................................................................... 71 Similar expression pattern and relation with degeneration ............................................... 71 Analysis of the integrity of dopaminergic system ................................................................... 73 Substantia nigra ................................................................................................................... 73 Striatum ............................................................................................................................... 75
GFP control vector ................................................................................................................... 75 Substantia Nigra .................................................................................................................. 75 Minor inflammatory response and transient Gal3 release ..................................................... 76 Cytokine profile ................................................................................................................... 77 Galectin-3 transient release ................................................................................................ 78 Trem2-MerTK alternative activation ....................................................................................... 80 Autophagy analysis.................................................................................................................. 82 Absence of Gal3 protects from ipsilateral motor impairment ................................................ 83 Open Field ........................................................................................................................... 84 Rotarod ................................................................................................................................ 84 Cylinder test ........................................................................................................................ 85 DISCUSSION ...................................................................................................................... 87 Gal3 is related with αSyn in LB pathologies in an antagonistic manner ................................. 87 Identifying Gal3-αSyn society in vitro ..................................................................................... 88 Adenovirus-mediated overexpression of αSyn as a model for PD study ................................ 89 hSYN accumulates in intracytoplasmic inclusions in Gal3KO mice ......................................... 90 Gal3 deletion reduces pSyn levels and protects from neurodegeneration ............................ 91 Minor inflammatory response and transient Gal3 expression ............................................... 92 Signs of alternative activation pathway by Gal3 ligands TREM2 and MerTK.......................... 93 Gal3-mediated autophagy as a potential player in αSyn managing ....................................... 94 Motor impairments are prevented in Gal3KO mice................................................................ 96 CONCLUSIONS ................................................................................................................... 99 REFERENCES .................................................................................................................... 100
1 RESUMEN En 1817, James Parkinson describió por primera vez un conjunto de síntomas motores en el ensayo titulado “Shaking Palsy” (Parálisis temblorosa). Este conjunto de síntomas fue posteriormente renombrado como Enfermedad de Parkinson (EP). Durante estos dos siglos, los avances en la investigación de la EP han revelado un papel fundamental para la proteína neuronal α-sinucleína (αSyn). Esta proteína, cuyas funciones fisiológicas están aún por determinar, forma inclusiones citoplasmáticas características de la EP llamadas Cuerpos de Lewy. Junto a la aparición de estos Cuerpos de Lewy, la EP se caracteriza por la degradación específica de la substancia nigra, un área dopaminérgica del mesencéfalo cerebral involucrada en el control del movimiento. En la actualidad, las diversas hipótesis sobre la etiología de la EP no están logrando transformarse en mejores tratamientos o un diagnóstico más temprano, por ello, la búsqueda de nuevas moléculas implicadas en el origen y la progresión de la EP se revela indispensable para desarrollar nuevas estrategias terapéuticas. Recientemente se ha relacionado a la proteína Galectina-3 (Gal3) con los Cuerpos de Lewy en cerebros post-mortem. Gal3 es una proteína de unión a galactósidos sin actividad catalítica y expresada principalmente por células de la microglía, las células inmunes del sistema nervioso central, en procesos neuroinflamatorios. En la presente tesis doctoral, tratamos de describir el papel de Gal3 en la patogénesis de la EP basándonos en muestras post-mortem, in vitro e in vivo. En estudios humanos post-mortem, hemos podido asociar esta proteína al proceso de agregación que da lugar a los Cuerpos de Lewy. Los estudios in vitro nos han permitido detallar los procesos de agregación de αSyn y captación neuronal de Gal3. Por su parte, hemos desarrollado un modelo in vivo de sobreexpresión de αSyn humana durante 6 meses en ratones salvajes y carentes de Gal3 (Gal3 knockout). Nuestros resultados confirman la presencia de Gal3 en muestras post-mortem tanto de pacientes con Demencia de Cuerpos de Lewy como de EP. Como observación relevante, hemos relacionado la presencia de Gal3 preferentemente con acumulaciones menos densas de αSyn pero no con los Cuerpos de Lewy. Igualmente, hemos descrito un papel fundamental de Gal3 durante el proceso de agregación de αSyn in vitro, donde Gal3 bloquea y revierte dicho proceso de agregación. Además, hemos demostrado que las neuronas son capaces de incorporar Gal3 del medio extracelular y que esta Gal3 se une rápidamente a las acumulaciones de αSyn. Por último, nuestras observaciones in vivo corroboraron estos datos, demostrando un gran impacto de la deleción de Gal3 en el modelo animal de sobreexpresión de αSyn humana mediada por adenovirus. Los animales carentes de Gal3 desarrollaron un mayor número de agregados de αSyn, pero no mostraron neurodegeneración de la substancia nigra ni desarrollaron síntomas motores, lo que supone que la deleción de Gal3 no solo preservó la supervivencia de las neuronas dopaminérgicas sino también preservó su funcionalidad. Asimismo, hemos
2 identificado la aparición de un fenotipo microglial basado en los marcadores TNFα, Trem2 y MerTK que podría ser específico en la EP. Este fenotipo se ve abolido en ausencia de Gal3, lo que pone de manifiesto la importancia de esta proteína en la activación microglial asociada a neurodegeneración. En definitiva, nuestro trabajo demuestra un importante papel de Gal3 en la patogénesis de la EP y en la viabilidad de las neuronas dopaminérgicas. Los datos que se muestran sugieren que Gal3 está inherentemente relacionada al proceso de agregación de αSyn, su toxicidad y al proceso neuroinflamatorio que desencadena.
3 ABSTRACT In 1817, James Parkinson described a serial of motor symptoms in the essay Shaking Palsy. These symptoms were later renamed as Parkinson’s Disease (PD). In the following two centuries after the discovery of the disease, knowledge has being accumulated and pointing out a pathological role of the neuronal protein α-synuclein (αSyn). This protein can form intracytoplasmic inclusion called Lewy body along and dopaminergic degeneration of substantia nigra. Both established hallmarks of the disease. However, no hypothesis about the PD etiology has been confirmed in terms of resulting in an effective treatments that could prevent or stop the progression of the disease. For that reason, discovery of new molecules implied in PD development and progression is indispensable for new investigation strategies. Recently, the inflammatory protein Galectin-3 (Gal3) has been related to Lewy body in post-mortem studies. Gal3 is a galactosebinding protein without known catalytic activity and expressed mostly by activated microglial cells, the immune cell of the brain. No major expression has been detected in neurons in basal conditions. Role of this protein in disease development is unknown. In my thesis, we unravel the role of Gal3 in the pathogenesis of PD based on; human postmortem identification of the protein related with Lewy body aggregation process; in vitro studies of aggregation and neuronal uptake of Gal3; and in vivo model of αSyn overexpression for 6 months in wild type and Gal3-knockout mice. Our results confirmed the presence of Gal3 in post-mortem samples from Dementia of Lewy Body and PD patients. Strikingly, Gal3 accumulation was related to non-dense accumulation of αSyn rather than dense Lewy body. Indeed, neuronal Gal3 protein expression was correlated with lower αSyn staining. Our data showed a critical role for Gal3 in aggregation pattern of αSyn with a marked inhibition and disruption of aggregation process in vitro. Furthermore, we demonstrate the ability of neurons to uptake extracellular Gal3 and incorporate it to αSyn accumulations. In vivo data followed the same pattern and demonstrated a huge impact for Gal3 deletion in our model of adenovirus-mediated overexpression (AAV5) of human αSyn. Gal3 knockout mice displayed higher number of aggregates but absence of neuronal damage and lack of behavioural impairment meaning that not only surveillance of dopaminergic neurons were maintained intact but also functionality of these neurons were preserved. Our work demonstrate a role of Gal3 in PD pathogenesis and its detrimental role for dopaminergic neurons. The present data suggest that Gal3’s predominant role is inherently associated to the aggregation process of αSyn and inflammation in mediating neurotoxicity.
4 INTRODUCTION Parkinson’s Disease “Parkinson’s disease occurs when brain cells that make dopamine, a chemical that coordinates movement, stop working or die”. This definition is given by The Michael J. Fox Foundation for Parkinson’s Research and highlights the basis of a disease. A neurodegenerative disease where there at present are no cure or treatment that could halt its progression. Historical view Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by severe motor symptoms triggered by a specific loss of dopaminergic neurons in a structure called substantia nigra (SN). It was first described in 1817 as a “Shaking Palsy” by the British scientist James Parkinson in his work “An Essay on the Shaking Palsy” (Parkinson, 1817). However, we can find earlier reports of the most characteristic symptoms of the disease back in the ancient India and China about 30 centuries ago (Zhang, et al., 2006; Manyam, 1990). James Parkinson described for the first time a new disorder that included symptoms like involuntary tremor and motion as well as lessened muscular power leading to rigidity with uninjured intellect and senses. He reported six cases, three of them being his own patient and three more observed in the streets of London. We had to wait additional 50 years for the the term Parkinson’s disease to be first mentioned in the literature. This name was given by Jean-Martin Charcot, a French neurologist and anatomy professor that became one of the most important pioneers in the field of clinical neurology (Goetz, 2005). Charcot established a clinical method for identification of neurological diseases unifying symptomatic and post-mortem signals. Charcot description of diseases were extremely detailed and his contributions to neurological disease were of great clinical value. In fact, the description of bradykinesia, one of the classical PD triad symptoms along with tremor and rigidity is attributed to Charcot. Among all his work, Charcot and his student Leopold Ordenstein separated for the first time two of the main tremor-associated diseases: multiple sclerosis and PD, giving the actual name to the last in honour of James Parkinson. In addition, in Ordenstein’s work, a detailed recapitulation of symptoms was provided for both diseases including symptom like tremor and gait disorders followed by rigidity and palsy. He even detailed a hallmark for diagnosis in rest tremor that persists with voluntary movements. Ordenstein related for the first time the progressive characteristic of the disease and a late age of onset as a key feature of PD (Lehmann, et al., 2007). Meanwhile, Charcot continued his work detailing more clinical cases
11 deficits in key CMA proteins like LAMP2A and Hsc70 have been identified (Campbell, et al., 2018). In fact, PD animal models show accumulation of αSyn following CMA deficiency and, in the other way, mutations and post-translational modification in αSyn leads to aberrant CMA (Campbell, et al., 2018). e) Prion-like propagation of αSyn: at the beginning of the century, Braak and colleagues proposed that PD lesions started in the olfactory bulb or the dorsal motor nuclei of the vagal and glossopharyngeal nerves to further spread to other areas of the brain. In this study, the authors established a six-stage system where stage 3 denotes SN affection and clinical symptoms appearance (Braak, et al., 2003). Despite this system has not been completely accepted (Kalaitzakis, et al., 2008), it has been useful and valid in some cases (Dickson, et al., 2010). Nevertheless, the theory that pathogenic αSyn can propagate from one anatomical region to another has led to the idea that some forms of αSyn can act in a prion-like way and that neuron-to-neuron or cell-to-cell transport is an essential pathway for disease progression and LB formation (Angot, et al., 2010). This idea gathered great attention when grafted fetal neurons were found to contain LB after 11- 16 years in PD host (Li, et al., 2008). Furthermore, extending Braak´s theory, gut has been proposed as the initiator for αSyn misfolding and PD onset and several stimuli including microbiome, intestinal inflammation and toxins (Klingelhoefer and Reichmann, 2015). In a prior-like fashion after enteric nervous system is affected by αSyn misfolding, vagus nerve will serve as a highway to central nervous system were the disease will spread (Kim, et al., 2019). Despite several hypothesis has been proposed, none has been able to give a unifying explanation of the disease. While it is true that all of them have evidence of support, a more global view including several detrimental factors is warranted. However, an unsolved point is the fact that αSyn starts to aggregate with no apparent cause. This could support the prion-like theory since this hypothesis claims that the disease can potentially start in a single cell. However, even prionlike propagation do not explain the origin of the initial prion protein. Indeed, it needs to incorporate any other hypothesis previously mentioned as key factor for disease onset (Brundin and Melki, 2017). In addition, assumption that LB formation leads to neuronal death has also been questioned, as they can be a neuroprotective mechanism of infected neurons by concentrating detrimental proteins at one intracellular place. (Chartier and Duyckaerts, 2018).
12 Figure 4. Schematic view of the main hypotheses for PD pathogenesis. Note all factors are based on either neurotoxicity or aggregation as the main features of PD. Symptoms & Diagnosis PD has three classical symptoms called ‘the triad of PD’: tremor, bradykinesia and rigidity. However, a big subset of non-motor symptoms has been identified since PD first description and some of them appear before motor symptoms and can be a clue for early diagnosis of the disease (Figure 5A). According to Braak´s stage system, a prodromal phase precedes the clinical manifestations that lead to the correct diagnosis of PD (Figure 5B) (Braak, et al., 2003). These are named stages I-II where affection has not reached SN yet and symptoms are restricted to non-motor symptoms. On the other hand, Hoehn and Yahr Scale, that is still predominant in clinical world, lacks from a specific stage for prodromal phase (Hoehn and Yahr, 1967). For that reason, the Unified Parkinson's Disease Rating Scale (UPDRS) was first proposed and later updated by the Movement Disorder Society to MDS-UPDRS (Goetz, et al., 2007), which establish a score rating rather than a stage system that encompass motor and non-motor symptoms. This system let to follow the disease progression with the increase in score rating. Motor symptoms include classical rest tremor, bradykinesia, muscular rigidity, postural instability or gait impairment. They often appear in a unilateral manner that progress to also affect the contralateral side (Kalia and Lang, 2015). These are also related with SN degeneration and according to Braak´s staging appears in stage III of the disease. Motor symptoms are also the most common diagnosis criteria for PD and the most accurate so far, despite the neuronal damage in this stage is already very high.
13 Tremor is the most recognized symptom of PD, and it is present in about two thirds of PD patients and it is not a clinical feature for non-tremor-dominant subtype of PD. Parkinsonian tremor is easily identified in hands but also in toes, limbs and jaws. It is characterised by an exacerbation under stressful situations and a decrease with voluntary movement and sleep (DeMaagd and Philip, 2015a). However, mixed tremor can be present and identification of parkinsonian tremor from other types of tremors are difficult. Indeed, tremor does not correlate with dopaminergic cell loss seen by imaging studies of PD patients and other motor symptoms are preferable for PD diagnosis (Xia and Mao, 2012). Bradykinesia defines the loss of speed, gait and amplitude of voluntary movements. It is the most common motor symptom and a key diagnosis criterion (DeMaagd and Philip, 2015a). Bradykinesia is commonly present as a delay or difficulty in the initiation of voluntary movement and as abnormalities in gait and posture. Other features like dysarthria, hypomimia and decrease blinking has been detected. In fact, in advance disease, episodes of transient freezing can be observed (Xia and Mao, 2012). In contrast with tremor, bradykinesia seems to correlate with dopaminergic degeneration (Jankovic, 2008). Rigidity is the third major component of motor symptoms. Comprise resistance to movement in what is known as “cogwheel rigidity” because of resistance to a range of motion. Rigidity can affect the limbs, or the face contributing to hypomimia also named “masked expression” (Garcia Ruiz, et al., 2011). Rigidity can also be associated with joint pain with shoulder pain being the most common. Rigidity, however, is not accompanied by muscle strength decrease, as this feature is not affected in PD (DeMaagd and Philip, 2015a). Long-term rigidity can lead to postural deformities especially in neck and trunk were it leads to abnormal posture and deformities of hands and toes (Jankovic, 2008). The fourth most common motor dysfunction is postural instability. It is related with previously described motor symptoms and is commonly observed in advance PD. However, it can appear before disease diagnosis as deficit in balance and insecurity when walking (Garcia Ruiz, et al., 2011). Other motor symptoms can be secondary to the previously mentioned and include difficulties in handwriting and speech (hypophonia) (DeMaagd and Philip, 2015a), dysphagia and sialorrhea, ocular problems derived from decreased blink rate and appearance of some reflexes movements (Jankovic, 2008). A feature of particular relevance for PD patients is what is known as on-off or wearing-off phenomenon. That is the re-emergence of motor and non-motor symptoms even in worsen manifestation during periods between treatment doses (Pahwa and Lyons, 2009). Non-motor symptoms can appear even before motor manifestations and can help in the diagnosis of the disease. However, disease cannot be diagnosed based only in non-motor symptoms. All PD patients suffer for at least one non-motor symptom and they are related with different anatomical areas that are affected through disease progression.
14 Sensory symptoms are one of the most common non-motor symptoms and in some cases represent the first detectable PD symptom. Olfactory disturbance (hyposmia or anosmia) are one of the first symptoms to appear and is related to possible olfactory bulb origin of the disease (Schapira, et al., 2017). It is difficult to detect and many patients are unaware of it despite 90% of the diagnosed patients have this symptom (Bohnen, et al., 2008). PD patients also complain about vision problems. Visual hallucinations and double vision are the most relevant and disabling symptoms and appears with disease progression (Pfeiffer, 2016) but in prodromal phase of the disease, symptoms like impair colour discrimination, reduced contrast sensitivity or dry eye syndrome are the most common (Nowacka, et al., 2014). These visual disturbances either can have a dopaminergic neurotransmission origin or be caused by αSyn deposition in retina (Schapira, et al., 2017). Sensory symptoms also include pain as one of the most notified by patients. Pain can be caused by motor symptoms like rigidity but also has been described lower pain threshold in PD patients due to impaired dopaminergic neurotransmission (Schapira, et al., 2017). Other sensory disturbances have been described like paraesthesia or burning sensation. Psychological symptoms are also very common in PD within all stages of the disease with particular incidence of anxiety and depression. Anxiety can exhibit in form of panic attacks or social phobias and is very sensitive to on-off periods of treatment due to strong association to dopamine levels (Pfeiffer, 2016). On the other hand, depression can involve apathy and anhedonia and can precede motor symptoms. Indeed, depression is associated with a worse prognosis of PD (Dissanayaka, et al., 2011). Apathy can appear independently from depression and is thought to be present in at least 40% of PD patients (den Brok, et al., 2015). Fatigue is also a common symptom of PD patients. It can derived from motor dysfunction but also from depression and apathy or can be a consequence of sleep disturbances (Schapira, et al., 2017). Dementia and cognitive impairments are the latest psychological symptoms. Despite PD patients were initially described by James Parkinson to have “uninjured intellect” (Goetz, 2011), dementia has emerge as a high disabling and common feature of late-stage and severe PD (Hely, et al., 2008). Indeed, severity of dementia is associated with presence of LB in cortical areas (Schapira, et al., 2017). In general, cognitive impairments worsen with disease progression but can appear at mild levels before PD diagnosis (Pfeiffer, 2016). Sleep disorders can make it easier to diagnose PD. They have a high prevalence (up to 90%) and comprise a form of insomnia characterised by sleep fragmentation with frequent awakenings (Pfeiffer, 2016). Insomnia in PD is hard to treat because of its diverse origin: motor symptoms, on-off periods or REM sleep behaviour disorder (RBD) (Schapira, et al., 2017). In addition, RBD has emerged as a potent prodromal marker of PD (Tekriwal, et al., 2017). RBD consist in exceptional movement ability during REM periods of sleep. PD patients show increased motor function during these episodes. These episodes resemble in some aspects those described as kinesia paradoxa, a sudden (and transient) recovery from motor disability. Kinesia paradoxa has been described under some extreme stimuli, including emotional or physical stress (Banou, 2015).
15 Other non-motor symptoms can be described as autonomic dysfunction. Gastrointestinal dysfunction is one of the most common non-motor symptoms including dysphagia, gastroparesis, bacterial overgrowth and, importantly, constipation (Pfeiffer, 2016). According to the prion-like theory, ENS origin of the disease is responsible for these symptoms that frequently appear even several years before PD diagnosis. However, use of intestinal αSyn as a biomarker have not achieved a consensus so far (Shannon, et al., 2012; Antunes, et al., 2016) and origin of these symptoms can be due to the role of dopamine in gut neurotransmission (Mittal, et al., 2017). Urinary symptoms are also frequent in PD patients. These include nocturia, increased frequency and urgency and sphincter relaxation. Urinary dysfunction progress along with the disease what implies an important role for dopamine in urinary processes (Schapira, et al., 2017). Other less known motor symptoms includes sexual problems with erectile dysfunction and low sexual desire as the most representative; thermoregulatory dysfunction that includes sweating problems; and cardiovascular problems where orthostatic hypotension is the most recognized with a high prevalence among PD patients. Orthostatic hypotension is frequently accompanied by episodes of nocturnal hypertension or higher heart rate (Schapira, et al., 2017). Control of this symptomatology is essential for higher life-quality of patients and prevention of clinical complications.
16 Figure 5. PD symptoms progression and Braak stage system. (A) Non-motor symptoms appear before motor symptoms whose onset is considered the diagnosis criteria (year 0). Prodromal phase can last over 20 years before PD diagnosis. Late phase of the disease curse with disabling motor symptoms and cognitive failures. Reproduced from Kalia and Lang, 2015 (Kalia and Lang, 2015). (B) Six stages of PD pathology according to Braak system. Disease would developed in anterograde axis starting from the the olfactory bulb, the anterior olfactory nucleus and/or the dorsal motor nuclei of the vagal and glossopharyngeal nerves in the brainstem. In stages I-II (B i) motor symptoms should be absence and they will start in stages III-IV (B ii) with SN affection. Final progression of the disease would reach the cortex in late stages V-VI (B iii). Reproduced from Goedert et al., 2013 (Goedert, et al., 2013) Differentiating PD from other diseases can be difficult because of the heterogeneity of PD symptoms and similarities with other diseases, especially in the early stages of PD, like multiplesystem atrophy, Dementia with LB (DLB) or essential tremor (DeMaagd and Philip, 2015a). Up to 30% of the initially diagnosed PD patients are later re-diagnosed with other disease. Techniques like MRI, despite not being able to distinguish PD anatomical features, let the exclusion of other pathologies like multisystem atrophy, progressive supranuclear palsy or corticobasal degeneration (Tolosa, et al., 2006) Motor symptoms are still the unequivocal signal for PD diagnosis. In particular, bradykinesia is the reference symptom when accompanied by rest tremor, rigidity or postural instability. Unilateral onset and levodopa treatment response are features that support the diagnosis (Kalia
17 and Lang, 2015). However, definitive diagnosis is made upon histopathological hallmarks in autopsy (Tolosa, et al., 2006). Incorporation of non-motor symptoms to diagnosis criteria and other technological techniques like neuroimaging has been claimed for earlier diagnosis (Berg, et al., 2013). Difficulties for that lie in the poor specificity of non-motor symptoms, the absence of reliable biomarkers, the doubts about relevance of gastrointestinal biopsies and the inefficacy of neuroimaging for early diagnosis. However, advances in these fields would let not only an earlier diagnosis but also a better understanding of the disease. Epidemiology PD is the most common neurodegenerative disorder after AD. The number of people that are diagnosed with PD are presently increasing due to higher life expectancy. Prevalence of the disease is in the rage from 100 to 200 per 100,000 people with an annual incidence of 15 per 100,000 people. However, prevalence of the disease dramatically increases in elder population over 1,000 patients per 100,000 people (Tysnes and Storstein, 2017). This increase is due to the age of onset of the disease situated at an age of 65-70 years. Earlier onset is seen in less than 5% of the cases and caused in great majority by genetic factors (Van Den Eeden, et al., 2003). Incidence is in average 14 new cases per 100,000 people (Ascherio and Schwarzschild, 2016) with some increased incidence in men compared with women. In Spain, general prevalence for PD is around 200 patients per 100,000 inhabitants but elevates up to 1500 in population over 65 years (von Campenhausen, et al., 2005). In Sweden, the incidence of the disease was noted to be as high as 2.9% (2,900 per 100,000) in people over 89 years (Linder, et al., 2010). Global incidence is similar for Asia, Europe and America but nor for Africa (Okubadejo, et al., 2006). However, despite incidence in black community is lower than in other types of communities like Hispanic (Van Den Eeden, et al., 2003), it has to be determined if this is caused to genetic, lifestyle or socioeconomic factors. Several risk factors have been identified to be related with disease onset. Supporting data, however, is sparse in some cases. High consumption of milk, alcohol or fatty diet have been proposed but not confirmed as risk factors, but have provided some clues that point towards gut microbiome as a potential risk factor. At the same time, traumatic brain injuries, diabetes and pesticides have substantial data supporting their role as inducers of PD because of wellstablished mechanisms of physiological damage(Ascherio and Schwarzschild, 2016). However, genetic factors are undoubtedly the best-known risk factor for PD incidence (Noyce, et al., 2012), but the genetic contribution is believed to be stronger for AD compared to PD In contrast, protective factors are better established including consumption of tobacco and coffee at the top of the list because of the neuroprotective effect of nicotine and caffeine respectively. High uric acid serum concentration is another protective factor probably due to its antioxidant properties (Oropesa-Ruiz, et al., 2016). Consumption of non-steroidal antiinflammatory drugs, in particular ibuprofen, is also related with lower PD incidence supporting
18 the key role of neuroinflammation in PD pathogenesis. Physical activity is also thought to be a protective factor, despite mechanism of action is diverse, anti-oxidant properties and liberation of neurotrophic factors could be behind this effect. Other potential factors are still under investigation like calcium channel blockers, flavonoids or statins (Ascherio and Schwarzschild, 2016). Treatments Since the first studies, PD researchers have worked for finding effective treatments for the disease. Charcot first used anticholinergic drugs and vibratory therapy against motor symptoms including vibratory chair and helmet. Gowers was the first to introduce cannabis as a potential treatment for PD symptoms (Goetz, 2011). But it was not until 1961 when levodopa was first used as treatment for PD (Birkmayer and Hornykiewicz, 1961). Levodopa is still the gold standard treatment for PD (Salat and Tolosa, 2013). It is a prodrug of dopamine but is able to cross BBB via the large neutral amino acid transporter in contrast to dopamine that is unable to cross BBB due to its ionization state (DeMaagd and Philip, 2015b). The mechanism of action of levodopa is simple, when crossing the BBB can be internalised by neurons and converted to dopamine. This increases the dopamine release in presynaptic terminals thus compensating the effect of lost dopaminergic terminals/neurons due to the neurodegeneration (DeMaagd and Philip, 2015b). Levodopa has been proved to be a safety drug in terms of toxicity. In peripheral tissues, levodopa is quickly decarboxylated to dopamine increasing dopamine peripheral effect but greatly reducing its bioavailability in the CNS. In addition, short half-life of levodopa is challenging and provokes on-off episodes that can be diminished with special levodopa formulations with sustained release (Salat and Tolosa, 2013). In addition, increase bioavailability of levodopa is obtained with use of a dopa-decarboxylase inhibitor like carbidopa (or benserazide) or catechol-O-methyltransferase (COMT) inhibitor like entacapone. These two drugs reduce the peripheral catabolism of levodopa, thus increasing levodopa half-life and quantity of levodopa that cross the BBB. While these drugs do not cross the BBB, the effect of levodopa in the CNS is not altered (Salat and Tolosa, 2013). Unfortunately, the effectiveness of levodopa is better in early stages of PD and decreases in late stages with more recurrence of on-off symptoms and dyskinesias. A probable explanation is that loss of dopaminergic neurons dramatically decrease the therapeutic effect of levodopa because of the loss of most of the target cells. In addition, loss of dopaminergic cells produces lower dopamine endogenous levels which increases the dependence of levodopa levels for neurotransmission leading to acute on-off episodes (Stocchi, et al., 2010). Levodopa is a symptomatic treatment without affecting the progression of the disease. However, control of levodopa dosage is still the only efficient pharmacological treatment in PD (Salat and Tolosa, 2013) as other drugs has variable efficacy and are used only in combination with levodopa or in very early stages of the disease. Among these drugs, we can find dopamine agonists and MAO inhibitors. Dopamine agonists can be administered alone (in first stages) or in combination with levodopa/carbidopa
19 and can directly activate the dopamine receptors in the CNS. Non-ergot class of dopamine agonist are the most used because of their selectivity and safety profile (DeMaagd and Philip, 2015b). Inhibition of monoamine oxidase type B (MAOB) is another field of action for PD treatment as this enzyme is responsible for degradation of dopamine release to the synaptic cleft (Poewe, et al., 2017). Seleginine and rasalagine present lower adverse effects than dopamine agonist and can be the election in early stages of PD (Poewe, et al., 2017). Meanwhile, management of non-motor symptoms is variable and non-specific and adverse effect of levodopa treatment are managed with the use of amantadine a N-methyl-D-aspartate receptor inhibitor (Seppi, et al., 2011). Diverse surgeries have been proposed for treatment of symptoms and modifying the course of the disease. Deep brain stimulation (DBS) is the best established surgery. DBS consists in the electric stimulation of different brain areas through the surgical colocation of an electrode. Despite its robust therapeutic effect, the mechanism of action is not fully elucidated. DBS can change brain activity in a controlled manner and it has been proved to be very effective in the treatment of motor symptoms (Poewe, et al., 2017). It is usually an option in late PD when levodopa treatment is inefficient (Kalia and Lang, 2015). Moreover, surgical lesioning of basal ganglia was an experimental treatment frequently applied before the development of levodopa. Nowadays this therapy has been replaced by new techniques like MR-guided focused ultrasound which are based on the same principles (Strauss, et al., 2014). New therapies are being tested with little or no success in modifying disease progression. Biological therapies with trophic factors like BDNF or CDNF, adenovirus mediated overexpression of enzymes mediating dopamine production or control and dopaminergic cell transplant have been tested in clinical trials. The latest approaches have failed to show clinical improvement but it has been confirmed that implanted cells can survive in brain of patients. Altogether, development of new therapies are based on the different approaches to the pathogenesis of the disease. A need for a therapy that modulate or even stop disease progression is urgent, which largely depends on preclinical research (Elkouzi, et al., 2019). α-synuclein: The killer queen Historical perspective Synuclein was first identified in the late 1980s by Maroteaux and colleagues as a neuronalspecific protein in presynaptic terminals. The name given by Maroteaux was based on protein location: synapsis and nucleus, despite the first was later identified to be the predominant location. Authors identified synuclein as 140 amino acid protein in rats that includes an 11 amino
20 acid sequence (6 of them forming a well-preserved core) repeated for 7 times (Maroteaux, et al., 1988). Same authors proposed synuclein as a family of proteins with at least three different genes that shared the repetition core (Maroteaux and Scheller, 1991). Other groups discovered similar proteins and human homologous were identified giving them the names α, β and γ synucleins (Clayton and George, 1998). Mutations in αSyn human gene were related with familiar forms of PD (Nussbaum and Polymeropoulos, 1997) while αSyn was identified in LB (Spillantini, et al., 1997) (Figure 6A) and in AD amyloid accumulation (Uéda, et al., 1993). All this led to an increasing interest in this protein in the next years with a better description of its properties and roles in the CNS and, particularly, in neurodegeneration. Despite role of αSyn in AD was shown to be at least minor (Culvenor, et al., 1999), growing knowledge has confirmed a huge role in PD and DLB. In parallel, αSyn inclusions were also found in glial cells in multiple system atrophy (Wakabayashi, et al., 1998). Inclusions seen in these diseases (latter known as synucleopathies) led to investigation about αSyn properties where it was found to be natively unfolded (Weinreb, et al., 1996). Later, it was discovered that in association with membranes αSyn could change its structure increasing its alpha-helicity (Davidson, et al., 1998). In addition, normal αSyn was discovered to fibrillate under certain circumstances (Hashimoto, et al., 1998) and fibrillar αSyn was found to be the core of LB (Spillantini, et al., 1998). Structure and physiological functions In terms of structure, αSyn is composed by three domains: the N-terminal with lipid binding properties and α-helix structure; a central domain with amyloid-binding capacities; and a C- terminal acidic domain (Figure 6B). The N-terminal domain (1-60) is highly preserved among the synuclein family and includes the 11 amino acid repetition and have some similarities with apolipoproteins. The central domain (61-95) is involved in fibril formation and can form cross-β structures. The C-terminal domain (96-140) present a random coil structure and may interact with central domain to modify aggregation tendency (Emamzadeh, 2016). Among the six mutations currently known to cause familiar PD, all of them affect the N-terminal domain: A30P, E46K, H50Q, G51D, A53E, and A53T. However, lipid binding and aggregative properties vary among the distinct mutations and exact mechanism (or mechanisms) that underlies the aggregation process remains deeply unknown (Ruf, et al., 2019). Knowledge about aggregation process is undoubtedly hampered by lack of certainty about structure and function of αSyn. Indeed, some authors have claimed that αSyn in physiological conditions tend to form α-helix tetramers that resist aggregation process (Bartels, et al., 2011). However, other authors points toward monomers as the only physiological form of αSyn (Burré, et al., 2013). Different pieces suggest an important role for tetramers in physiological conditions and aggregative properties of αSyn. In addition, a balance between tetramer:monomers seems to exists and to be determinant for PD onset. A decrease in tetramers has been shown to lead to αSyn accumulation, aggregation and PD resembling symptoms (Nuber, et al., 2018).
27 attention of many researchers, functions from what was described as ‘resting’ microglia are currently accumulating (Figure 7C). Figure 7. Pío-Hortega original drawings of microglia and their diversity. (A) Pío-Hortega illustration of microglia diverse morphology. Note multiple branches and high contact surface for all type. Pío-Hortega differentiate several morphologies including: A, bipolar, rod-shaped cell; B, elongated cell with two nuclei; C, multipolar cell with branched expansions; D, multipolar cell with spiny appendages; E, isogenic couple; F, bipolar corpuscle with forked appendages. Reproduced from Sierra et al., 2016 (Sierra, et al., 2016). (B) Pío-Hortega original draw from activated microglia two days after injury. Note engrossed and shortened branches turning to amoeboid shape. Reproduced from Sierra et al., 2016 (Sierra, et al., 2016). (C) Representation of microglia diversity in physiological conditions and disease. Currently, diversity of microglia leads to a heterogeneous population of microglia with different functions and specializations. Reproduced from Stratoulias et al., 2019 (Stratoulias, et al., 2019). One of the first non-immune roles of microglia was detected in connection with synapses. Microglia were observed to extend their branches and prolongations and make contact with synapses being responsible for disappearance of inefficient or useless synapses (Stevens, et al., 2007; Wake, et al., 2009). This process is known as synaptic pruning and thought to be mediated by microglia-neuron communications, in particular with the interaction of the microglial fractalkine receptor CX3CR1 and its neuronal ligand CX3CL1 (Paolicelli, et al., 2011; Hoshiko, et
28 al., 2012). However, microglial role in synapse function seem to go beyond pruning. Microglial roles in synapses formation have also been described and suggested to be mediated though secretion of the neurotrophic factor BDNF (Parkhurst, et al., 2013) or by direct contacts (Miyamoto, et al., 2016). Actually, synaptic plasticity is controlled through microglial neurotransmitter receptors that constantly monitor neuronal functionality (Pocock and Kettenmann, 2007). Microglia are also determinant during CNS development in prenatal stage. Thus, microglia actively participate in axonal grown (Rigato, et al., 2011) through diverse activation pathways (Squarzoni, et al., 2014) thus raising the potential existence of a specific activation phenotype (Stratoulias, et al., 2019). Microglia have also active roles during neurogenesis and include removal of neuronal progenitors to regulate the number of newborn cells, a process thought to be done by primary phagocytosis (or phagoptosis) (Luo, et al., 2016). Importantly, microglia also contribute to CNS vasculature formation (Arnold and Betsholtz, 2013) probably through membrane receptors (Chen, et al., 2017) together with other glial cells during embryogenesis. Interestingly, microglia also play a role in neurotransmission. In particular, microglia can prevent excitotoxicity that occurs when neurotransmitter are released in excess, which can create a sustained polarization state and neurotoxicity. Thus, microglia have been shown to be important in glutamate-mediated neurotoxicity, which could also imply a role in long-term potentiation (LTP) and/or long-term depression (LTD), two important features of synaptic plasticity associated to learning and memory (Colonna and Butovsky, 2017). Moreover, importance of microglia in dopamine neurotransmission has been hypothesised as microglia can degrade dopamine through the COMT enzyme (Redell and Dash, 2007). Microglial cells are equipped with a great variety of receptors that allow them to constantly scan the microenvironment but also let them to establish numerous crosstalks with other cells including neurons and astrocytes. These cells have shown to control microglial activity by the secretion of transforming growth factor beta (TGFβ) that seems to mediate the homeostatic phenotype of microglial cells (Butovsky, et al., 2014). However, pathways of cell-to-cell communication are greatly increasing and include secreted factor, cytokines, direct ligandreceptor interaction or extra cellular vesicles such as exosome (Szepesi, et al., 2018). Neuroinflammation Microglial cells are the leading actors regulating immune responses within the CNS. Inflammation has long viewed as a vital process against invasive pathogens. However, in the CNS, sterile inflammation has progressively gained attention. It occurs after tissue damage and in absence of pathogens and it has been related with many neurodegenerative diseases. Protein aggregates, neuronal debris or nucleic acids are among the signals that can trigger sterile inflammation (Chen and Nuñez, 2010).
29 Typically, after activation, microglia lose their homeostatic morphology and functions withdrawing its ramifications and acquiring a rounded shape. This morphological transformation goes along with transcriptional and receptor changes (Haynes, et al., 2006). Microglia have a wide range of membrane receptors that are able to detect several dangerous signals. In particular, damage associated molecular patterns (DAMPs) and pathogen associated molecular patterns (PAMPs) are unequivocal signals of danger that are recognized by pattern recognition receptors (PRRs). Toll-like receptors (TLRs) are the main membrane proteins that can detect these signals and induce transcriptional changes that govern the response and phenotype of microglia (Chen and Nuñez, 2010). This includes the secretion of inflammatory cytokines like tumour necrosis factor (TNFα), interleukins (ILs), interferon-γ and chemokines, as well as producing reactive oxygen and nitrogen species (ROS and RNS). Importantly, many of these factors can act in a paracrine and autocrine way. This means that inflammatory process is a feedback mechanism that can propagate through the brain (Cherry, et al., 2014). Neuroinflammation is also highly cytotoxic as its main goal is to destroy any invader pathogen but it may also produce damage to endogenous cells, especially for neurons. The Inflammatory phenotype is tightly controlled and can be reversed when the process is no longer needed. In this case, anti-inflammatory cytokines like IL-4 or IL-10 are secreted (Gadani, et al., 2012). However, in some cases like PD, the inflammatory process can turn chronic forming an endless cycle of cytotoxicity and inflammation. A common feature of neuroinflammatory processes is the production and secretion of proinflammatory ILs, for instance, IL-1β and IL-18. These are secreted in a tightly controlled process mediated by the two-step formation of the molecular platform known as inflammasomes that ultimately triggers caspase-1 activation and ILs processing (Broz and Dixit, 2016). Inflammasome formation is trigger by DAMPs or PAMPs signals. Among these signals, αSyn has been postulated as an inflammasome inducer since PD patients showed unequivocal signals of inflammasome assembly (Gordon, et al., 2018). Most of the common animal models of PD produce a marked neuroinflammation. For instance, MPTP, 6-hydroxidopamine and lipopolysaccharide (LPS) show a marked microglial reactivity that enhance neurodegeneration (Barcia, et al., 2004; Castaño, et al., 1998). It is remarkable that in the model of LPS, the associated neurodegeneration is provoked by microglial activation as demonstrated by the fact that anti-inflammatory compounds prevent neuronal loss (Tomás- Camardiel, et al., 2004; Herrera, et al., 2005) and also because LPS has been proven to be harmless to neurons in absence of microglial cells (Lehnardt, et al., 2003). This suggests that neuroinflammation can not only be secondary to neuronal death but also a key player in the onset of neurodegeneration. New models of overexpression of Syn and injection of preformed fibrils (PFFs) also point in the same direction, as PFFs increase reactive microglia prior to onset of neurodegeneration (Duffy, et al., 2018). Inflammatory process has been studied for long, however, microglia was traditionally considered as macrophages in the brain. Thus, most authors gave microglia the same stages of activation than macrophages in a black-or-white classification: M1 and M2 (Tang and Le, 2016;
30 Moehle and West, 2015; Orihuela, et al., 2016). This view has radically changed in the last few years with the advent of single-cell transcriptomic analysis. These studies have revealed the presence of multiple microglia phenotypes that depends not only of the stimuli but also of the CNS localisation, stage of life and microenvironment (Stratoulias, et al., 2019). New activation patterns Microglia is a multiple-task cell present in all regions of the CNS. It is logical to assume that microglia possess a whole range of phenotypes (Figure 8). Thus, classification of these phenotypes is challenging and maybe useless in term of physiology, but it can help to better understand the pathways that are implied in different situations and thus help in understanding the role of microglia in neurodegenerative diseases. The recent discovery of a specific disease-associated microglial phenotype (DAM) in contrast to the homeostatic phenotype (M0) has shed light to better understand the role of reactive microglia under conditions of neurodegeneration (Krasemann, et al., 2017; Keren-Shaul, et al., 2017). These single-cell analysis identified a specific microglia transcription profile associated to neurodegenerative diseases like AD, amyotrophic lateral sclerosis and multiple sclerosis. In particular, homeostatic microglia overexpress some genes that where widely known by its homeostatic functions like Cx3cr1 or P2ry12 and maybe driven by TGFβ signalling. In contrast, DAM microglia seem to be driven by the axis formed by apolipoprotein E (APOE) and triggering receptor expressed on myeloid cells 2 (TREM2). Activation of this pathway triggers transcriptional machinery that leads to overexpression of several genes including Clec7a, Lpl, Gpnmb, Itgax (CD11c) and, importantly, Lgals3 (galectin-3). This phenotype leads to effective phagocytosis and enhance cytokine production, but can be perpetually maintained by protein aggregation and neurodegeneration. So far, a distinctive microglia phenotype associated to PD is missing. Distinct features among diseases have to be considered, as DAM microglia have been primarily described in AD where deposition of protein aggregates occurs extracellularly rather than intracellularly like in PD. Nevertheless, PD and AD share some common features, like progressive neurodegeneration or ageing as the main risk factor. Altogether, research aimed at elucidating PD-associated microglia has yet to be done in this field in order to elucidate potential variations and similarities between DAM microglia and PD microglia. For instance, Gpnmb is one of the most upregulated genes in DAM microglia, while is also one of the identified genetic risk factor for PD (Nalls, et al., 2014). In addition, other DAM gene like Lpl, that codes for protein lipoprotein lipase, has been shown to have a determinant role in αSyn aggregation (Yang, et al., 2015). Further, Lgals3 has been proved to be secreted by microglia after αSyn oligomer stimuli (Boza-Serrano, et al., 2014). Moreover, genetic loss of homeostatic genes like Cx3cr1 have shown to exacerbate neuroinflammation in mouse models of PD (Castro-Sánchez, et al., 2018).
31 Figure 8. Microglia diversity is controlled by multiple receptor signalling. Traditionally, microglia activation was thought to be a black-or-white system governed by few receptors that lead to a pro- or anti-inflammatory role (upper part). Currently, multiple receptors have been identified to trigger specific functions of microglia. Indeed activation of single receptor pathway is unlikely to happen in vivo. Instead, mixed phenotypes leads to the diversity observed and the specific response for each stimulus or disease. The importance of establishing a microglial phenotype relevant for PD pathogenesis is critical to identify novel relevant proteins implied in the disease. However, some limitations exists for this as PD has shown to be variable in location, progression and protein aggregation. In addition, microglia has proved to be a rapidly reactive cell and several pathways may be implied at the same time. Microglia in Parkinson’s Disease As already stated, microglia is known to play an important role in PD progression. Neuroinflammation is one of the factors that have been hypothesized to be behind PD onset and clinical data supports the importance of microglia in PD. Importantly, microglia density is higher in SN compared to other brain regions including other dopaminergic nuclei. In particular, SN pars reticulata is extremely rich in microglial cells and some differences can be found in microglial cells from distinct parts of the SN (McCarthy, 2017). Different microglia concentration in certain brain regions is not the only factor supporting the existence of different microglia subtypes. Another factor is age, which is well documented to have an effect on glial populations. For instance, increasing microgliosis has been detected in human SN after mid-age (Jyothi, et al.,
32 2015). Implication of homeostatic microglia in PD onset has been considered as minor while all the relevance has been given to the subsequent neuroinflammation process. However, microenvironment can be critical for neuronal homeostasis, and thus, disease onset. Microglia’s ability to maintain healthy neuronal microenvironment is crucial and can be a related to PD onset hypothesis (Figure 4). In fact, microglia depletion has been shown to exacerbate PD symptoms and neuronal degeneration in an MPTP model (Yang, et al., 2018). Moreover, pharmacological depletion of microglia increase cell-to-cell propagation of αSyn in mouse models, highlighting its role in αSyn clearance. Similar results were obtained when microglia were activated with LPS suggesting that loss of homeostatic phenotype is highly relevant for prion-like progression of the disease (George, et al., 2019). Neuroinflammatory process alters neuronal microenvironment because of microglial loss of homeostatic functions. Chronic alterations like emotional stress are known to be a risk factor for PD while also affecting microglia phenotype by promoting neuroinflammation (Herrera, et al., 2015). Indeed, systemic inflammation is known to trigger microglial activation and dopaminergic loss in animal models of sepsis (Qin, et al., 2007) and ulcerative colitis (Machado, et al., 2011). In addition, familiar forms of PD are known to elicit microglial activation in SN (Sanchez- Guajardo, et al., 2015). Therefore, the identification and modulation of the microglial activation pattern in PD turns essential for the new therapeutic strategies. Galectin-3: The Theory of Everything Family of galectins Galactose-binding lectin 3 also known as galectin-3 (Gal3) is a 263 amino acid protein that belongs to the family of lectins. So far, 15 members have been identified in mammals. Galectins binds to β-galactoside structures, a relatively common structure in glycans which explains the diversity of functions of these proteins. This sugar structure is recognised by a specific part of the galectin, a protein domain named carbohydrate recognition domain (CRD). Galectins are divided into three groups depending on its structure: prototype, tandem and chimeric. Prototype galectins includes those galectins that possess only one CRD domain and includes galectins -1, -2, -5, -7, -10, -13, -14 and -15. The second group of galectins possess two CRD domains like galectins -4, -6, -8, -9 and -12. Finally, the chimeric group of galectins is only composed by Gal3 which possesses three domains: an N-terminal domain with oligomerisation ability, a proline-rich domain and the CRD domain (Suarez and Meyerrose, 2014) (Figure 9).
33 Figure 9. Galectin family and function (A) Representation of the three types of galectins including the prototype, with a single CRD domain and ability to form dimers; the tandem type, with two different CRD domains; and chimera type, with ability to form oligomers. (B) Ligand type define galectins structure and functionality. Formation of lattice let galectins to be implied in protein sequestration, protein interaction and scaffolding function in molecular platforms. The different structures that may acquire each galectin may have influence on its activity and binding properties. Prototype galectins can be found in monomeric or dimeric forms but only recognise one ligand. On the other hand, tandem galectins can bind to two different ligands at the time. While Gal3 can oligomerise and bind to distinct ligands establishing a lattice-like network (Rabinovich and Toscano, 2009). Additionally, it is known that galectins can act both intra and extracellularly. Hence, due to their particular properties, galectin family is involved in multiple functions like inflammation, immune response, cell migration or autophagy. Such target diversity makes location of galectins, glycosylation state of the cell and cell subtype the determinant factors for galectins functions (Johannes, et al., 2018). Galectin-3 Gal3 is normally present in small quantities and diffuse manner in the cytosol. However, it can be secreted to the extracellular space upon certain stimuli. In addition, Gal3 can be found in body fluids and interestingly, high serum levels of Gal3 predict prognosis for some diseases like cardiac failure (Suarez and Meyerrose, 2014) or inflammation associated to atherosclerotic plaque (Papaspyridonos, et al., 2008).
34 Functions of Gal3 are, as well as the rest of the family, varied and can be classified upon the cellular location of Gal3. a) In the cell nucleus, Gal3 has been associated with mRNA splicing and cellular proliferation (Patterson, et al., 2015). Mechanistically, Gal3 binds to U1 snRNP (small nuclear ribonucleoproteins) that lately attracts pre-mRNA forming an active molecular platform called spliceosome that can process pre-mRNA into mature mRNA (Haudek, et al., 2016). Moreover, β-catenin has been shown to interact with Gal3 and consequently induce nuclear transcription in cancer cells (Shimura, et al., 2004) b) At the cytoplasm, Gal3 can be phosphorylated and its main target seems to be Bcl-2. Gal3 binds Bcl-2 through homolog domain NWGR (placed in the C-terminal domain of Gal3). Therefore, Gal3 can recruit Bcl-2 and, by doing so, prevent its apoptotic activity (Yang, et al., 1996). In contrast, Gal3 was identified to be bound to the internal part of CD95 receptor, leading to caspase-8 activation and apoptosis (Fukumori, et al., 2004). In tumours, Gal3 seems to play a major role in cell proliferation and adhesion (Vladoiu, et al., 2014). Recently, Gal3 has been shown to interact with damaged lysosomes that expose its internal glycans and also participate in an unknown manner in autophagy pathway (Weng, et al., 2018). c) At the cell membrane is where Gal3 plays more diverse roles. For the immune system is probably the most important localisation. Increasing number of studies have demonstrated the ability of Gal3 to interact with multiple membrane receptors (Rabinovich and Toscano, 2009). However, specific function is highly dependent on the cell subtype. In T-lymphocytes, Gal3 can have an anti-inflammatory effect by forming N- glycan complexes with T cell receptors (TCR) immobilizing them and diminishing their functionality (Demetriou, et al., 2001). This action is preceded by the addition of N- glycans to these receptors by enzyme Mgat5 (a glycan transferase). In contrast, in leukocytes, Mgat5 promoted cytokine signalling and migration (Partridge, et al., 2004), features impaired in Gal3 deficient leukocytes (Gittens, et al., 2017). Importantly, in microglial cells, Gal3 has been proved to interact with TLR4 (Burguillos, et al., 2015) and triggering receptor expressed on myeloid cells 2 (TREM2) (Boza-Serrano, et al., 2019). d) At extracellular level, it has been shown that Gal3 can be secreted through unconventional ways like exosomes (Bänfer, et al., 2018; Théry, et al., 2001). In the case of some tumour cells, Gal3 can inactivate the immune system by immobilizing TCR and avoiding its approach to co-receptors like CD8 (Demotte, et al., 2008). Indeed, Gal3 has been detected to promote apoptosis in T lymphocytes in vitro (Fukumori, et al., 2003). However in other cell types like neutrophils, Gal3 has demonstrated to enhance activation (Fermino, et al., 2011) and recruitment (Bhaumik, et al., 2013). Gal3 has also been identified as an opsonin since it promotes macrophage clearance of apoptotic cells by binding glycoproteins in the apoptotic surface (Karlsson, et al., 2009) e) Gene deletion models have been a very useful tool for studying the role of Gal3, but its pleiotropic character made it difficult to interpret the data obtained from them. Gal3 knockout mice are fertile and present no distinct phenotype. However, it has been
35 reported that Gal3 deficient mice present an attenuated immune response in several diseases models like autoimmune encephalitis, asthma or parasitosis (Jaquenod De Giusti, et al., 2011; Zuberi, et al., 2004; Breuilh, et al., 2007). It has been described that Gal3 deletion affects T cell maturation as it has been detected in thymus promoting cell death (Silva-Monteiro, et al., 2007). Similar proinflammatory effect has been seen in models of acute colitis where Gal3 promotes inflammasome signalling in macrophages (Simovic Markovic, et al., 2016). Remarkably, glycosylation is a variable process that can create a different glycosylation environment depending not only of the cell subtype but also of the biological process (Marth and Grewal, 2008; van Kooyk and Rabinovich, 2008). Altogether, different mechanisms explain the different and in some cases conflicting roles for galectins. Galectin-3 in the central nervous system Gal3 has been identified to be expressed predominantly in activated microglia within the CNS. However, it has also been detected in other cell types including neurons, oligodendrocytes and astrocytes. In the adult rat brain, neuronal Gal3 has been detected predominantly in cortical areas and hypothalamus but not in basal ganglia (Yoo, et al., 2017). Evidence of transitional expression of Gal3 has also been found in neurons exposed to certain stimuli like global ischemia in the gerbil brain (Hisamatsu, et al., 2016). However, physiological functions of Gal3 in neurons remain elusive. Yet, it has been suggested to act as an opsonin and facilitate phagocytosis of damaged neurons. Glial cells are the most common source of Gal3 in the CNS. Remarkably, all three subtypes (astrocytes, oligodendrocytes and microglia) have been identified to express Gal3 under certain conditions. Oligodendrocytic Gal3 seems to have a clear role in differentiation and myelinisation (Pasquini, et al., 2011). However, the source of this Gal3 is doubtful as some authors have used exogenous Gal3 and efficiently provoke oligodendrocyte differentiation by inducing cytoskeleton changes (Thomas and Pasquini, 2019) suggesting that Gal3 can be secreted elsewhere. Meanwhile, Gal3 expression has been well-established in astrocytes (Yang, et al., 2006). Data about a functional role for Gal3 in astrocytes is scarce but it has been suggested to facilitate cell motility (Comte, et al., 2011). In contrast, increased expression has been detected in models of AD associated to inflammatory response (Ramírez, et al., 2019), reactive astrocytes after virusinduces encephalitis (Jaquenod De Giusti, et al., 2011) and in ischemic brain (Yan, et al., 2009). Indeed, Gal3 gene (Lgals3) is one of the most upregulated genes in astrocytes after ischemia (Rakers, et al., 2019) and stab wound injury (Sirko, et al., 2015). However, we and other have demonstrated that under multitude of conditions that involves neuroinflammatory processes,
36 microglial cells are the primary source of Gal3 over astrocytes (Bonsack and Sukumari-Ramesh, 2019; Yip, et al., 2017; Boza-Serrano, et al., 2019). Little is known about role of Gal3 in physiological conditions. Some authors have proposed that microglial cells exposed to exogenous Gal3 increases ramification and motility in vivo and in vitro without production of cytokines (Rahimian, et al., 2019), while others proposed an angiogenic role (Wesley, et al., 2013). In fact, role of Gal3 has been widely debated with some authors claiming for a pro-inflammatory effect (Jeon, et al., 2010) and others claiming for an antiinflammatory effect (Lerman, et al., 2012). In order to clarify this point, our group has identified Gal3 as an endogenous ligand of TLR4 receptor (Burguillos, et al., 2015). Moreover, microglia can secrete Gal3 and enhance activation via TLR4 similar to LPS stimulus. Notably, similar effect was seen in relation of Gal3 and TLR2 (Jouault, et al., 2006). Nevertheless, Gal3 seems to exacerbate microglial activation under a broad spectrum of stimuli within the CNS. Recently, Gal3 has been proposed to control microglial phenotype, specially phagocytosis, through regulation of cytoskeleton (Reichert and Rotshenker, 2019) what may imply Gal3 in some of the physiological functions of microglia. Despite increasing data about the potential roles of Gal3 in the CNS, few molecular pathways have been identified to be mediated or at least influenced by Gal3. Inflammasome is one of the most important and controlled pathways during innate immune system activation. Gal3 has been shown to interact with NLRP3 inflammasome in peripheral macrophages (Tian, et al., 2016; Simovic Markovic, et al., 2016). However, if this is a direct interaction with the molecular platform or, in contrast, Gal3 acts as a priming signal is still undetermined. Interestingly, Gal3 has been proposed to be secreted after pyroptosis induction and pore formation in macrophages (Chen, et al., 2019). In addition, in the CNS, Gal3- inflammasome relationship has been observed in models of Huntington’s Disease where Gal3 suppression impairs inflammasome signalling (Siew, et al., 2019). Autophagy-lysosome pathway and Gal3 have been closely related since Gal3 was identified as a reliable marker of lysosomal membrane permeabilization (Aits, 2019). This method was efficiently used by Weng et al. (Weng, et al., 2018) to detect lysosomal rupture after bacterial infection. Authors detected that Gal3 associated to lysosome impairs autophagy and promote bacterial surveillance. Moreover, it has been recently published that lysosomal rupture is a mechanism for prion-like invasion of amyloid proteins including αSyn (Flavin, et al., 2017; Jiang, et al., 2017). However, this may contrast with several studies that indicates that Gal3 interacts with TRIM16 protein to detect defective cytoplasmic membranes thus forming a molecular platform that recruits autophagy inductors and triggers autophagy-lysosome pathway (Kumar, et al., 2017; Chauhan, et al., 2016). Hence, association of Gal3 with membrane impairment has been well identified in lysosomes, but can be extended to other cytoplasmic membranes. Nevertheless, a role of Gal3 in autophagy-lysosome pathway is still unknown. Overall, Gal3 seems to play major roles in the immune system with pleiotropic functions that have been recently associated to inflammasome and autophagy. Indeed, increasing evidence anticipates Gal3 as a key player in several neurodegenerative diseases. Supporting this, many studies have shown that Gal3 is a highly inducible protein under disease conditions. In contrast,
43 optimisation section). Length and homogeneity of fibrils were the two determinant factors. Final conditions were established in a total of 40 pulses at 30% power (total of 20 sec and 0.5 sec on/off cycle). For treatment, PFFs were added to the media at a final concentration of 1 µg/ml. Implication of Gal3 in aggregation was determined by adding exogenous Gal3 at a final concentration of 50 nM 0, 24 and 48 hours after PFF addition. Cells were fixed in 4% PFA 3 hours after last Gal3 treatment. Apoptotic neurons To induce apoptosis, we followed the protocol published by Butovsky and colleagues (Krasemann, et al., 2017). Firstly, N27 neurons were carefully detached from the plate surface by repeated washes with PBS. Later, neurons were irradiated with UV light (302 nm) with an intensity of 6 x 15 W for 20 min. Neurons were rapidly chilled on ice and centrifuged at 300g for 5 min at 4 °C. Apoptotic neurons were resuspended in DMEM and counted before being used as treatment. For transcriptional signature determination, BV2 cells were challenge with 100 nM of Gal3, 5000 apoptotic N27 cells or both for 18 hours. After the experiment, cells were lysed in TRisure (Bioline) extraction buffer. RNA was extracted following the manufacture’s protocol (see gene expression analysis section) and gene expression was analysed. Table 3. Product used for cell lines culture. PRODUCT COMPANY HEADQUARTERS REFERENCE RPMI 1640 Medium, GlutaMAX Supplement GibcoTM-Thermo Fischer Scientific Waltham, MA, USA 61870010 DMEM High Glucose Medium Biowest Nuaillé, France L0102 Fetal Bovine Serum (FBS) Biowest Nuaillé, France S181H Penicilin- Streptomicin Solution Biowest Nuaillé, France L0022 Menzel™ Microscope Coverslips 12 mm2 MenzelTM-Thermo Fischer Scientific Waltham, MA, USA 11708701 Phosphate Buffer Saline Thermo Fischer Scientific Waltham, MA, USA 10010015
44 PFFs optimisation αSyn PFFs were sonicated as indicated in Volpicelli-Daley study (Volpicelli-Daley, et al., 2014b). Up to 6 conditions were tested using a UP50H ultrasonic processor (Hielscher Ultrasonic, Teltow, Germany): 1. No sonication 2. Incubation in ice for 1 hour 3. 20 pulses at 30% power (total of 20 sec and 0.5 sec on/0.5 sec off cycle) 4. 20 short pulses at 30% power (total of 20 sec and 0.25 sec on/0.75 sec off cycle) 5. 60 pulses at 30% power (total of 60 sec and 0.25 sec on/0.75 sec off cycle) 6. 20 short pulses at 50% power (total of 20 sec and 0.5 sec on/0.5 sec off cycle) Electron microscopy Sonicated PFFs were then processed at the Microscopy Service (CITIUS) at the University of Seville. A conventional protocol for negative staining was applied. Firstly, a 5 µl drop of PFFs diluted in PBS solution was placed on top of a 100-mesh cupper grid (Electron Microscopy Science, Hatfield, PA, USA). After incubation for 20 min at room temperature (RT), adhered sample was counterstained with 2% uranyl acetate in aqueous solution for 15 min. Later, excess of uranyl acetate was eliminated and sample was allowed to dry at RT. Images were obtained in ZEISS Libra 120 electron microscope (Zeiss, Oberkochen, Germany). Figure 12. Electron microscopy of PFF. Condition 1 (no sonication) is shown in left panel and condition 3 sonication is shown in right panel. Scale bar 200 nm. 4% Paraformaldehyde Panreac Applichem Castellar del Vallés, Barcelona, Spain A3697 Triton-X100 Sigma Aldrich St Louis, MO, USA X100
45 ThS aggregation assay ThS (Sigma Aldrich) was first prepared at a 10 mM stock solution in distilled water and filterer with a 0.2 µm syringe filter. PFFs were diluted at 50 µg/ml for the assay. Up to 50 µl of a mixture of ThS and PFFs (from different sonication conditions) were placed along with their respective controls and blank (distilled water) in a 96-well dark plate with clear flat bottom (Sigma Aldrich) and later introduced in a Synergy HT microplate reader (BioTek Winooski, VT, USA). Microplate reader was previously warmed at 37 °C. Samples were maintained in agitation between reading measures. Emissions at 480 nm were recorded every 5 min, upon excitation at 440 nm. Condition 3 was selected for the in vitro assays due to constant size observed in electron microscopy (between 100-200 nm) along with low presence of aggregates, and also because of higher aggregation capacity showed in the ThS assay. PFFs were first filtered with Q-Max 0.22 µm syringe filter (Frisenette ApS, Knebel, Denmark), thus avoiding the presence of big aggregates. PFFs were tested in N27 cell line prior to use in any experimental procedure. Behavioural Tests Behavioural tests were all performed in the Faculty of Pharmacy from University of Seville. Animals were placed in the appropriate room at least one hour prior experiment in order to habituate them to the new environment. Mice were housed individually the week before to start the tests. No other change was made in animal routine unless indicated. WT GFP (n=14), WT hSYN (n=14), Gal3KO GFP (n=13) and Gal3KO hSYN (n=14) were tested 8 weeks and 24 weeks after injection. Open Field
46 For this test, mice were placed in a clean, open space. 40x55x40 cm white open field were used and a camera recorded up to 5 min without previous habituation to the cage. Spontaneous movement was recorded. Trajectory, activity, distance or speed were among the measured parameters by an automated tracking system (SMART 3.0, Panlab®, Harvard Apparatus, Holliston, MA, USA), so that we can study exploratory and locomotor behaviour. Rotarod Mice were first habituated to rotarod apparatus with a 5 min session at a constant speed of 5 rpm. To test motor coordination, animals faced 4 session at constant acceleration from 5 rpm to 40 rpm during 5 min. Mice took a 15 min break between each session. The day after, animals faced another 5 min session at constant acceleration to evaluate coordination memory and learning ability. When an animal fell off the rotarod, it was re-located in the roller by the experimenter. Number of total falls in each session was measured. Cylinder test Ipsilateral motor capacities were measured by cylinder test. For that, mice were placed in a 15 cm diameter, 30 cm high glass cylinder during 5 min without previous habituation. Because of natural explorative behaviour, mice inspect the cylinder by touching the glass walls. The number of times the mice touch the glass with each front paw was counted and ratio between right and left paw was measured. Immunofluorescence Figure 14. Video capture of cylinder test. Image show WT hSYN mice with ipsilateral preference while exploring the cylinder. Figure 13. Trajectory map (left) and Activity map (right). Trajectory and activity map from control mice showing common tigmotaxis. These maps represent a graphical view of some of the recorded parameters like distance travelled and activity.
47 Fixed brains obtained from animals were cryoprotected in 30% sucrose and later frozen in cold isopentane (-40 °C). After freezing, a microtome was employed to cut samples at 30 µm thickness in coronal plane. Sections were collected serially and quickly submerge in antifreeze buffer composed by 30% glycerine and 30% ethylene glycol diluted in phosphate buffer. For immunofluorescence staining, free-floating sections were first washed from antifreezing solution with treated with PBS (Nzytech, Lisbon, Portugal) and later treated with citrate buffer (Nzytech) for antigen retrieval by incubating the sections at 80 °C for 30 min. Secondly, sections were rinse at RT for 1-2 hours in PBS with the addition of 1% Triton-X100 (PBS-T) for cell permeabilization and 5% Bovine Serum Albumin (BSA) (Sigma Aldrich) for tissue blockade. Once sections were permeabilized, primary antibody recognition was let to occur overnight at 4 °C (See Table 4). The following day, unbound primary antibody was properly removed with sequential PBS washes. After that, 1 hour incubation with conjugated secondary antibodies from InvitrogenTM (Thermo Fischer) diluted 1:300 in PBS was carried out. Sections were finally washed and mounted on a clean slide and some drops of DAKO fluorescence mounting medium (DAKO were applied to preserve tissue fluorescence and seal the coverslip to the slide. Confocal fluorescent images were obtained from ZEISS LSM 7 DUO (Zeiss, Oberkochen, Germany). All images were captured under same conditions for unbiased comparison. Table 4. Antibodies and dilution used for immunofluorescence. ANTIBODY COMPANY HEADQUARTERS REFERENCE DILUTION Sheep anti-TH NOVUS Biologicals, Bio-techne Minneapolis, MN, USA NB300-110 1:500 Rabbit anti-hSYN Abcam Cambridge, UK AB138501 1:1000 Goat anti-Gal3 R&D Systems, Biotechne Minneapolis, MN, USA AF-1197 1:500 Rabbit anti-pSyn P129 Abcam Cambridge, UK AB168381 1:500 Mouse anti-Iba1 Millipore, Merck Darmstadt, Germany MABN92 1:250 Rat anti-CD11b AbD serotec (Bio- Rad) Hercules, CA, USA MCA74G 1:200 Mouse anti- LAMP1 DSHB Iowa City, IO, USA 1D4B-s 1:150 Stereological cell counting
48 Single immunofluorescent staining for Tyrosine Hydroxylase (TH) was employed for effective counting of total number of dopaminergic neurons in SN. A total of 6-8 serial slices containing the mesencephalon were used. Samples were place under an OLYMPUS BX61 (Olympus, Tokio, Japan) epifluorescence microscope. SN pars compacta was delimited by experimenter expertise and the help of mouse anatomical map. Stereological analysis was carried out using newCAST software (Visiopharm, Hørsholm, Denmark), especially designed for unbiased stereological studies. Delimited volume was calculated with the software using Cavallieri Points approach and 63x immersion oil objective was employed for clear cell counting. Microscope positions were unbiased selected by the software. At least 200 cells were counted for each individual substantia nigra following stereology guidelines and estimation of total number was later calculated based on counted neurons, counted volume and total estimated volume. In the case of our study, because of ipsilateral damage, data is presented as fold to non-injected side to decrease deviation due to individual variability. Dendritic tree measure Single immunofluorescent staining for TH was employed for measure of dopaminergic arborisation. A total of 6-8 serial slices containing the mesencephalon were used. Whole SN images were taken in OLYMPUS BX61 (Olympus) epifluorescence microscope. SN pars reticulata was delimited by experimenter expertise and the help of mouse anatomical map. ImageJ® software was used for TH area measure. In the case of our study, because of ipsilateral damage, data is presented as fold to non-injected side to decrease deviation due to individual variability LB-like counting Double immunofluorescence for TH and hSYN was used to detect the percentage of neurons that present aggregate-like staining. A total of 6-8 serial slices were also used for LB-like counting. Confocal images obtained from ZEISS LSM 7 DUO were employed. SNpc was delimited based on TH staining using ImageJ software. The area was calculated and unbiased counting position selected. At least 200 neurons were counted for each animal. Rounded intense accumulation of hSYN was determined as positive by experimenter. Data is express as percentage of positive accumulation per total dopaminergic neurons counted. Human immunofluorescence
49 DLB samples were provided by Skåne University Hospital (Lund) and processed in Lund University (Sweden) while PD samples were provided by Hospital Universitario Virgen del Rocio and were processed in University of Seville. Post-mortem samples from DLB and PD patients were received as paraffin embedded samples. Thus, tissue was first deparaffinised with xylene for 10 min followed by rehydration with decreasing concentration of ethanol: 100%, 90%, 70% and 50% until final distilled water step. Antigen retrieval was applied to the samples in order to improve antibody linking. Samples were incubated in a citrate solution and heated in a microwave for 5 min. After repeating previous step, samples were let to recover RT for 20 min. A blocking step consisting in sample incubation in PBS-T 0.05% and 10% BSA was included to improve signal. Incubation with primary antibody was made overnight at 4 °C diluting it in PBS-T 0.05% and 5% BSA until desired concentration. After 6 washing steps with PBS, we proceed with secondary antibody incubation for 1 hour at RT. Finally, samples were rinse in PBS for three times and 50% glycerol was applied as mounting media, coverslip placed and sealed. Images were obtained from ZEISS LSM 7 DUO (Zeiss, Oberkochen, Germany) confocal microscope. Immunofluorescence of cell cultures For cell visualisation of primary cultures, immunofluorescence procedure was as follows. Cells were first blocked with 5% BSA and 0.1% Triton in PBS for 1 hour under soft agitation. Then, cells were incubated with primary antibody diluted in PBS-T 0.1% at 4 °C overnight. After repeated washes, cultures were incubated with secondary antibody diluted in PBS for 1 hour at RT under soft agitation. Finally, cell nuclei were counterstained with Hoechst (Sigma Aldrich) at 1:5000 dilution in PBS. Coverslips containing the attached cells were mounted with DAKO fluorescent mounting medium. Protein Analysis For protein analysis, mice brains were carefully dissected to obtain fresh striatum and mesencephalon that were quickly frozen in liquid nitrogen (for more information see Treatment and Sacrifice section). All following procedures were carried out in Experimental Neuroinflammation Laboratory (BMC B11) from Lund University (Sweden). Protein extraction and quantification
50 Protein extraction was the first step before any additional procedure. Initially, samples were embedded in cold RIPA buffer (Sigma Aldrich) supplemented with proteinase and phosphatase inhibitor (Roche, Basel, Switzerland) and later sonicated for 10 sec for membrane disruption. Samples were then centrifuged for 10 min at 10000 rpm to isolate total proteins that remained in the supernatant. Protein concentration was calculated using BCA Protein Assay-Kit (Thermo Fischer). As recommended by the manufacturer, BSA standard dilutions were carefully made in RIPA buffer in a working range of 2000-20 µg/ml. Samples were then incubated with supplied reagent at 37 °C for 30 min. Later, 280 nm wavelength was measure in a Benchmark Plus microplate reader (Bio-Rad, Hercules, CA, USA). Samples were diluted to a final concentration of 1 µg/µl with RIPA buffer. Western Blot Quantified proteins were first denaturalized by the addition of Laemmli buffer (Bio-Rad) supplemented with β-mercaptoethanol (Sigma Aldrich). After that, proteins were separated by SDS-PAGE using 4% pre-cast acrylamide gels (Bio-Rad) in TGS buffer (Bio-Rad). After electrophoresis, proteins were transferred from the gel to a nitrocellulose membrane (Bio-Rad) using a TransBlot turbo system from Bio-Rad. Membranes were then blocked for 1 hour with 3% skim milk (Sigma Aldrich) in PBS. Membranes were subsequently washed in 0.1% Tween 20 (Sigma Aldrich) diluted in PBS (PBS-Tw) prior to overnight incubation with primary antibodies (Table 5). After several washes with PBS-Tw, we incubated with peroxidase-conjugated secondary antibody for 2 hours at 1:10000 dilution. After final washes with PBS-Tw, blots were developed using ECL Clarity and ChemiBlot XRS+ system from Bio-Rad. β-actin was used as housekeeping protein and data is expressed as percentage of fold to actin. Table 5. Antibodies used for Western Blot analysis and dilution used. ANTIBODY COMPANY HEADQUARTERS REFERENCE DILUTION Rabbit antihSYN Abcam Cambridge, UK AB138501 1:5000 Rabbit anti-pSyn P129 Abcam Cambridge, UK AB168381 1:3000 Goat anti-Gal3 R&D Systems, Biotechne Minneapolis, MN, USA AF-1197 1:1000 Mouse anti- LAMP1 DSHB Iowa City, IO, USA 1D4B-s 1:2000 anti-Hsp70 Santa Cruz Biotechnology Dallas, TX, USA sc-32239 1:3000
51 Mouse anti- LC3B Cell Signaling Technology Danvers, MA, USA 2775 1:2000 Goat anti-Cat.B R&D Systems, Biotechne Minneapolis, MN, USA AFG965 1:3000 Mouse anti-β- actin-Peroxidase Sigma Aldrich St Louis, MO, USA A3854 1:10000 Cytokine quantification To measure cytokine levels in SN and striatum we took advantage of ELISA multiplex technology supplied by Meso Scale Diagnostics (Rockville, MD, USA). We selected a proinflammatory panel composed by IFNγ, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12 and TNFα. 50 µg of total protein extract diluted in RIPA buffer from SN and STR of animals were used. Plates were incubated for 2 hours at RT and constant shaking. Later, samples were rinsed 3 times with PBS-Tw 0.1%. Then, secondary antibodies were added to the plate and incubated for 2 hours at RT and constant shaking. After incubation, three PBS-Tw 0.1% washes were applied again. The plates were developed using the 2X reading buffer diluted to a factor of 1× with MiliQ water, and the plates were read using the QuickPlex Q120 reader from Meso Scale. The detection ranges of the different cytokines measured were as follows: IL1β (1670–0.408 pg/ml), IL4 (1660–0.405 pg/ml), IL12 (32,200–7.86 pg/ml), IL10 (3410–0.833 pg/ml), IFN-γ (938–0.229 pg/ml), IL2 (2630–0.642 pg/ml), IL5 (967–0.236 pg/ml), IL6 (5720–1.40 pg/ml) and TNF-α (627–0.153 pg/ml). ThT Aggregation Assay αSyn aggregation assay was performed by Dr. Franscesco Aprile in University of Cambridge, UK as previously described in published articles (Aprile, et al., 2015). Briefly, 20 µM αSyn monomer was incubated in buffer (50 mM Tris pH 7.4, 150 mM KCl, 5 mM MgCl2), 20 µM ThT, 5% αSyn PFFs, and 5 mM ATP at 37° °C in quiescence supplemented with 40 µM of Gal3 or Gal3 R186S molecules. Emissions at 480 nm were recorded every 20 min, upon excitation at 440 nm, using a CLARIOstar plate reader (BMG Labtech, Allmendgruen, Germany). Plates were sealed to prevent evaporation. For this experiment, αSyn PFFs were prepared as follows: First generation fibrils were created by incubating 300 µL of 70 µM monomeric αSyn in buffer at 37 °C for 4 days shaking at 200 rpm. The resulting fibrils were centrifuged at 16000g in 5 min, the pellet was washed twice with 300 µL of buffer and then sonicated for 1 min at 10% maximum power with 0.3 sec on/0.7 sec off cycles using a probe sonicator (Bandelin, Sonoplus HD 2070). Second generation fibrils were then created by incubating 100 µM monomeric αSyn with 10 µM of the first-generation fibrils in 500 µL of buffer for 13–14 hours under quiescent conditions. Finally,
52 the second-generation fibrils were sonicated for 20 s, 10% power, 0.3 sec on/0.7 sec off cycles to generate the seeds for the assay. For the disaggregation assay, purified aggregates were obtained after 7 days of aggregation following the previously described process. 20 µM of αSyn aggregates were incubated at 37 °C in buffer, 20 µM and 5 mM ATP under quiescent conditions in the presence of Gal3 and Gal3 R186S at molar ratios of 1:2 (αSyn:Gal3). Again, plates were sealed to prevent evaporation and emissions at 480 nm were recorded every 20 min, upon excitation at 440 nm, using a CLARIOstar plate reader. Gene expression analysis RNA extraction Cell lines For RNA extraction, BV2 cells were first washed 2 times with PBS. After that, cells were lysed with the addition of 800 µl per well of TRisure buffer (Bioline). As recommended, cells were pipetted several times for complete disruption. Next, all the volume was placed in Eppendorf tubes and 200 µl of chloroform (PanReac-AppliChem) was added and tubes were vigorously agitated for 15 min to achieve a complete mix. Samples were incubated for 3 min and later centrifuged at 12000g for 15 min. After that, an aqueous phase was form in the upper part containing the RNA. This phase was carefully removed and place in another tube containing 500 µl of cold isopropanol (PanReac-AppliChem) for RNA precipitation. After 10 min of incubation in ice, samples were centrifuged at 12000g for 10 min. Isopropanol was then removed and RNA pellet was washed with 1 ml of cold 75% ethanol (PanReac-AppliChem) and again centrifuged at 7500g for 5 min. Finally, ethanol was completely removed and RNA was resuspended in nuclease-free water (Nzytech). Mouse brain sections Serial sections from SN were used for RNA extraction and RT-PCR analysis. After mechanical disaggregation, FFPE-RNA Purification Kit (Norgen, Thorold, ON, Canada) was employed to obtain total RNA. Following manufacturer guidelines, provided digestion buffer was first used for lysate preparation supplemented with proteinase K for reversing formalin-induced crosslink between RNA and proteins. Samples were incubated in this buffer at 55 °C for 15 min followed by incubation at 80° °C for 15 min. After incubation, samples were placed on silica-membrane spin column and centrifuged at >3500g for 1 min until entire lysate volume had passed through the spin column. Next, column was washed 2 times with provided washing solution and centrifuged at >3500g for 1 min. Finally, RNA was eluted with nuclease-free water.
59 (A). Control Patient staining with low Gal3 and hSYN reactivity in highly neuromelanin marked neurons. Note isolated Gal3 dots in the proximities of the neuron (white arrows). Scale bar 10 µm (B). Neurons can actively uptake exogenous Gal3 in vitro. Association to pathological αSyn aggregates Previous anatomical studies dealing with the distribution of Gal3 has shown reactivity in neurons in some brain areas but not in SN or dopaminergic nuclei (Yoo, et al., 2017). Previously, we have demonstrated that Gal3 is preferentially expressed by reactive but not homeostatic microglia and involved in microglial response when activated by αSyn. TLR stimulation and fibrillary αSyn are among the stimuli that have been shown to increase Gal3 production (Burguillos, et al., 2015; Boza-Serrano, et al., 2014). For that reason, we explored if dopaminergic neurons have the ability to take up extracellular Gal3 in vitro. To this end, we treated N27 dopaminergic cells with 100nM exogenous Gal3 during 24h. Confocal immunofluorescence analysis demonstrated the unequivocal intracellular presence of Gal3 within N27 dopaminergic cells (Figure 18). Interestingly, intracellular Gal3 staining was visualized in terms of puncta accumulations within the cytoplasm. It has been suggested that under conditions of intact vesicular membrane and artificial lentiviral-associated Gal3 expression, Gal3 staining is diffuse as β-galactoside targets remain unavailable. Instead, ruptured vesicles, particularly lysosomes, expose β-galactosidase motifs to the cytosol, thus attracting Gal3 binding (Flavin, et al., 2017). Consequently, we analysed whether intracellular Gal3 was related to lysosomal membranes by using LAMP1 immunofluorescence. This analysis demonstrated that Gal3 was not present in LAMP1-labelled lysosomes (Figure 18) and alternative Gal3 ligand(s) can be present in dopaminergic neurons.
60 Figure 18. N27 dopaminergic cell line incorporates extracellular Galectin-3 to the cytosolic compartment after 24 hours exposure. N27 dopaminergic cell line were exposed to exogenous Gal3 for 24 hours. Immunofluorescence analysis of Gal3 and LAMP1 lysosomal marker was performed. Note that Gal3 (blue arrows) does not co-localise with LAMP1 lysosomal marker. Scale bar 10 µm To determine association of Gal3 with αSyn we designed an in vitro experiment where N27 neuronal cells were exposed to αSyn pre-formed fibrils (PFF) for 48 hours. This led to formation of αSyn aggregates.. For a better study of the impact of Gal3 in αSyn aggregation process, N27 were simultaneously stimulated with exogenous low dose of Gal3 (50nM) for different duration times: 48h, 24h and 3h. Immunofluorescence analysis of these cells revealed an unequivocal association of Gal3 with αSyn at every time point studied and, importantly, a clear impact in the aggregation process of αSyn (Figure 19). Cells exposed to Gal3 alone for 48h showed a puncta pattern as we previously described. Cells exposed to PFFs alone for 48h showed intense and widespread αSyn accumulations acquiring a punta-like staining. N27 neurons treated with PFF and Gal3 for 48h showed a precise correlation between Gal3 and αSyn aggregates, which showed a more diffused staining pattern thus suggesting a disaggregation effect. Indeed, we found a time-dependent effect. Thus, when PFF-treated cells were exposed for 24 hours to Gal3, lower colocalisation was observed. Finally, when PFF-treated cells were exposed for only 3 hours to Gal3, colocalisation was even lower and αSyn accumulations were less diffused. This data confirms Gal3 association with αSyn also in vitro, while a role of Gal3 in aggregation/disaggregation of αSyn seems to have a time-dependent effect.
61 Figure 19. Gal3 interacts with αSyn in vitro.N27 cell line treated for 48h with PFF 1µg/ml and, simultaneously, with Gal3 50nM for 48h, 24h or 3h. Immunostaining for Gal3 and αSyn revealed a role for Gal3 in αSyn aggregation along with a quick interaction of Gal3 and αSyn after only 3 hours that maintains with time. Gal3 was found to be associated with more diffused αSyn in a time dependent effect (blue arrows). αSyn negative for Gal3 was observed at shorter time points (white arrows). Scale bar 10 µm. Gal3 inhibits and reverse αSyn aggregation in vitro The presence of Gal3 in LB from PD and DLB (present study) anticipates a direct interaction between Gal3 and αSyn and raises the attractive hypothesis that Gal3 may affect αSyn aggregation. Our in vitro data shows Gal3 internalisation and rapid interaction with pre-
62 pathological αSyn, thus supporting this view. To test this hypothesis, we performed an in vitro Thioflavin-T aggregation (ThT) assay to analyse the effect of Gal3 in αSyn fibrillation (Figure 20A & 20B). As expected, αSyn monomers in absence of Gal3 displayed a classical aggregation kinetics; they quickly aggregate to form amyloid fibrils until saturation. Strikingly, exogenous Gal3 almost completely prevented αSyn fibrillation (Figure 20A). Indeed, this effect was full dependent on the carbohydrate recognition domain (CRD) as demonstrated by the usage of the Gal3 R186S mutant protein, which lacks CRD functionality. The mutation of the CRD reversed the effect seen with the non-mutated form of Gal3 and αSyn fibrillation was similar to control conditions (Figure 20A & 20B). More importantly, when Gal3 was added once aggregation had started, the aggregation rate was impaired to ultimately reduce the aggregation of αSyn (Figure 20B). In contrast, the addition of mutant Gal3 lacking the CRD function (Gal3 R186S) failed to show alter the normal αSyn aggregation process (Figure 20). These results confirm the predicted role of Gal3 in αSyn aggregation and the potential disaggregation effect seen in PD patients tissue. Overall, our data strongly suggest that Gal3 is implied in LB shaping. Figure 20. Gal3 prevents and retrace aggregation of αSyn. (A) ThT assay for αSyn aggregates formation in presence of Gal3 or mutant Gal3 R186S. Gal3 blocks aggregation in vitro in a CRD dependent manner. (B) ThT disaggregation assay where Gal3 is added to preformed αSyn aggregates. Gal3 but not Gal3 R186S retrace αSyn aggregation. AAV5-αSyn nigral injection as a model for PD study So far, the role(s) of Gal3 in PD has not been elucidated in any in vivo model. Hence, we evaluated the effect of Gal3 deletion in an adenovirus-based PD mouse model overexpressing human αSyn (hSYN) or GFP (control). We established a method of intranigral injection with extremely thin capillary glass (50-100 µm diameter) that highly reduces injection-associated inflammation. Adenovirus-associated overexpression of hSYN is a well established method for in vivo study of αSyn. However, several changes can derive in significant differences; i.e. animal selected,
63 injection method or titration of virus. We selected mice as experimental animals and inflammatory-reduced injection method with capillary glass while using an elevated virus titration. We first determine the efficacy of our in vivo model to express hSYN within dopaminergic neurons in the SN along with its anterograde transport to the striatum. Thus, we analysed two times points for model validation. These time points were established at 2 and 4 weeks after adenovirus injection. Expression of hSYN and control GFP proteins were robustly detected in the ventral mesencephalon with great expression in majority of nigral dopaminergic neurons 2 weeks after injection (Figure 21). At this postinjection time, no significant expression of both proteins was found in the striatum (Figure 21). Figure 21. Ipsilateral expression of adenovirus-overexpressed proteins (GFP and hSYN) in Substantia Nigra and Striatum 2 weeks after injection. Mice were injected with adenovirus in the left mesencephalon. Immunostaining of WT mice 2 weeks after injection revealed a clear ipsilateral expression in SN for both proteins but not in STR. Scale bar 100 µm. Scale bar of amplified images 20 µm.
64 Four weeks after injection, high expression in the SN for both hSYN and control GFP proteins with high selectivity for dopaminergic neurons was evident (Figure 22A). A clear difference was evidenced in the striatum respect to earlier time point as we found complete anterograde transport of both proteins from the SN to the dopaminergic nerve terminals (Figure 22B). Finally, long-term validation was performed 6 months after injection prior to experimental analysis. As expected, expression of both hSYN and control GFP proteins in both SN and striatum were maintained at similar levels than that observed at the earlier time point examined.
65 Figure 22. Long-lasting overexpression and effective anterograde transport of adenovirus-associated proteins to the striatum 4 weeks after the injection. (A) Mice were injected with adenovirus in the left mesencephalon. Immunostaining of WT mice 4 weeks after injection revealed a maintained expression of proteins, preferentially in dopaminergic neurons. Scale bar 20 µm. (B) Effective expression of both proteins in striatal terminals was observed 4 weeks after the injection confirming an anterograde transport of the protein. Scale bar 500 µm. Scale bar of amplified images 20 µm. Altogether, this data represent the validation of the injection model, confirming not only the successful expression of the human form of αSyn within dopaminergic neurons with relatively high selectivity, but also confirms the effective anterograde transport from the neuronal soma to the dopaminergic terminals.
66 Analysis of hSYN in the AAV5-αSyn model Our model represents a unique opportunity to study the in vivo effect of LGALS3 gene deletion in several αSyn features like accumulation or aggregation. To achieve this, we quantified and analysed the distribution and location of different forms of hSYN. Quantification of monomeric and tetrameric hSYN in SN and STR In order to test if the observed expression of hSYN was different within our genotypes, we quantified the total amount of hSYN present in the ventral mesencephalon and striatum by Western Blot. We obtained a 17kDa band for total monomer form of hSYN and around 60kDa band for total tetramer form. Once normalised with actin values, total amount of hSYN monomers were higher in Striatum of Gal3KO mice but not in WT mice (Figure 23A & 23C). Whether this could mean a more efficient anterograde transport by mice lacking Gal3 requires further investigation. However, in SN, monomers and tetramers were present in both of our genotypes at a similar expression level (Figure 23B-D). This implies that we face an unbiased model with equal expression of the protein in the ventral mesencephalon. Interestingly, the tetramer/monomer ratio showed a significant elevation in SN of Gal3KO animals, what could mean that hSYN is more prone to aggregation in Gal3KO as compared with WT (Figure 23E). Figure 23. Western Blot quantification shows similar hSYN production but a higher aggregation ratio in SN samples from Gal3KO mice. Western Blot analysis of distinct forms of hSYN in (A) Striatum and (B) SN. (C) Quantification of blots showed a higher level of hSYN monomers in Gal3KO Striatum. p<0.01. No significant difference was observed
67 between genotypes in SN for (C) monomers and (D) tetramers. Data is expressed as percentage fold to actin. (E) Tetramer/monomer ratio showed an increase in aggregated species in SN from Gal3KO mice. p<0.05. Analysis of hSYN distribution To further investigate the state of the overexpression of hSYN in our model, we performed double immunofluorescence against hSYN and TH and discovered a remarkably different pattern in the expression of hSYN within SN of both animals (Figure 24A). While WT animals presented a classical puncta pattern corresponding to localization in synaptic processes and mild soma reactivity with diffuse intracytoplasmic accumulation, Gal3KO mice presented strong rounded intracytoplasmic inclusions of hSYN in nigral dopaminergic cell bodies reminiscent of Lewy bodies. Intracytoplasmic accumulation of hSYN was later quantified following stereological criteria (Figure 24B) and confirmed that 42.83% ± 4.10 of the dopaminergic neurons in the SN of Gal3KO mice present this type of accumulation, almost twice than that obtained in WT animals (23.01% ± 2.45). These results suggest a previously unidentified role of Gal3 in αSyn aggregation.
68 Figure 24. Increased hSYN accumulations in TH neurons in Gal3KO mice. (A) TH/hSYN immunostaining revealed a different expression pattern for WT and Gal3KO mice. Note intracellular rounded accumulations present in Gal3KO but not in WT (white arrows and amplification below). Scale bar 50 µm. (B) TH+/hSYN+ double positive stereological counting of the SNpc (n=5). hSYN intracellular inclusions are present predominantly in Galectin-3 knockout mice. Data is expressed as fold to total TH+ neurons. p<0.01. Striatum analysis Supporting the results observed in SN, a different accumulation pattern was also evident in the striatum of Gal3 KO animal. Thus, a widespread presence of fibrillar structures of hSYN not associated to dopaminergic nerve terminals was seen in Gal3KO but not in WT mice (Figure 25A). These species are full reminiscent of those described by Baekelandt and colleagues (Peelaerts, et al., 2015). In their study, injection of different pre-formed αSyn strains including fibrils and
75 cell counting of the SNpc (n=5). Data is expressed as left hemisphere (LH) fold to right hemisphere (RH) (p<0.0001). (C) Images from arborisation of ipsilateral SNpr. Note decreased arborisation in WT. Scale bar 100 µm. (D) Quantification of TH+ area in SNpr for dendritic tree arborisation. Dendritic arborisation was decreased in WT mice compared to Gal3KO mice. Data is expressed as fold to contralateral side. p<0.05. (E-G) mRNA expression levels were measured in brain sections including mesencephalon. Dopaminergic markers TH, DAT and Gucy2C showed no differences in any condition. Striatum We have previously demonstrated an effective anterograde transport of hSYN from the SN to the striatum. Consequently, we also analysed the effect of Gal3 deletion in the striatum. For that purpose, we analysed TH fibres density within striatum of WT and Gal3KO. GFP control vector To discard any effect provoked by the injection process and/or the adenovirus, we carried out the same analysis in the animals injected with the GFP-control vector. Substantia Nigra Firstly, we examined the integrity of the SN after the long-lasting expression of GFP protein. Previous studies have shown TH staining loss in rats injected with GFP-carrying adenovirus. However, we saw no evidence of degeneration in neither the WT nor the Gal3 KO animals 6 months after the injection of the control vector (Figure 28A). In addition, we run stereological analysis of the dopaminergic cell bodies within the SN structure and no sign of neuronal loss was found (Figure 28B). Thus, we can confirm that the degeneration observed in WT animals is intrinsically due to hSYN overexpression.
76 Figure 28. Dopaminergic neuron viability is not compromised by control GFP-AVV5 injection. (A) TH immunostaining of the SN showing GFP injected side (right) and non-injected side (left). No apparent loss of dopaminergic neurons was observed. Scale bar 500 µm. (B) TH+ stereological cell counting of the SNpc (n=3). Data express as left hemisphere (LH) fold to right hemisphere (RH). Minor inflammatory response and transient Gal3 release
77 Neuroinflammation has traditionally been linked to PD (Joshi and Singh, 2018) and Gal3 has been linked to neuroinflammation in several neurodegenerative diseases including AD or Multiple Sclerosis (Krasemann, et al., 2017; Boza-Serrano, et al., 2019). Further, we have demonstrated a significant role of Gal3 in driving αSyn-induced proinflammatory response (Boza-Serrano, et al., 2014). Cytokine profile Consequently, we measured the molecular neuroinflammatory status in the ventral mesencephalon 6 months after hSYN overexpression in both WT and Gal KO mice. We performed MesoScale ELISA for quantification of main cytokines in soluble fractions including TNFα and interleukins (IL) 1β, 2, 4, 5, 6, 10 and 12. Most cytokines failed to show significant variations in response to Gal3 deletion in both striatum and SN (Figure 29A-H). However, levels of TNFα, a key inflammatory cytokine, were significantly higher in WT mice as compared with Gal3KO mice overexpressing hSYN, especifically in the SN (Figure 29I). There is strong evidence involving TNFα in the pathophysiology of PD (Tansey and Goldberg, 2010), whose levels are elevated in PD patient (Mogi, et al., 1994). Our mice model mimic this feature but points out that the TNFα increase is strictly related to neurodegeneration.
78 Figure 29. Selective increase of TNFα levels in SN of WT animals (A-I) Cytokine quantification in SN and striatum was performed by MesoScale plates of brain extracts from AAV5-αSyn injected SN and striatum. p<0.05. Galectin-3 transient release We next analysed the expression of Gal3 in the ventral mesencephalon of our experimental WT mice. We performed a quantitative and qualitative analysis by Western Blot (total amount) and by immunohistochemistry (microglia-related expression) in WT mice overexpressing hSYN. Our Western Blot analysis of Gal3 in the ipsilateral and contralateral nigra 6 months after injections demonstrated i) there is a basal constitutive expression of Gal3 in the ventral mesencephalon, and ii) no significant differences were found between the injected and the non-injected sides (Figure 30A & 30B). Later, we wondered if Gal3 basal expression could be sufficient for a determinant effect inside the neuron. To further investigate this, we performed immunofluorescence analysis in the injected area 2 and 4 weeks after injection and discovered a transient Gal3 release in injection area of WT animals that was present 2 weeks after injection (Figure 30C) but disappeared at 4
79 weeks. This Gal3 release could have acted as a priming effector in our model, as suggested by Gal3lo immunostaining of some hSYN positive cell bodies (Figure 30C lower panel). Indeed, 4 weeks after the injection we were still able to detect some Gal3 inside neurons (Figure 30D). Importantly, in this case we were able to distinguish between neurons with high Gal3 and low expression of hSYN and neurons Gal3lo/hSYNhi. This correlation is somehow reminiscent of what we had previously seen in human patients, what demonstrate that Gal3 in a critical player in our model. In addition, this rapid could have been maintained over time with accumulation of transient single microglia secretion of Gal3 upon extracellular hSYN stimuli. Since reactive microglia is the cellular phenotype supposedly expressing highest levels of Gal3 (Burguillos, et al., 2015; Boza-Serrano, et al., 2014), we next performed tissue immunofluorescence against CD11b (highly expressed by microglia), Gal3 and pSyn in both WT and Gal3 KO mice. We were able to confirm low Gal3 reactivity within the SN, but even in that context, we were able to detect some highly reactive Gal3 positive microglia in SN 6 months after αSyn overexpression (Figure 30E).Microglial Gal3 is known to be upregulated upon several extracellular stimuli, including αSyn (Boza-Serrano, et al., 2014). Despite presence of αSyn in the extracellular space is known to occur through several pathways including neuronal death and exosomes (Emmanouilidou and Vekrellis, 2016), whether extracellular αSyn can interact with microglia in vivo has not been described yet. Our model showed clear pSyn internalization by microglia that was not dependant on the presence of Gal3 (Figure 30E). Taken together, we provide evidence that Gal3 is expressed in the ventral mesencephalon. Under conditions of hSYN overexpression, we found small clusters of DAM expressing high levels of Gal3. We have demonstrated that Gal3 is released by reactive microglia (Burguillos, et al., 2015; Boza-Serrano, et al., 2014). In addition, we demonstrate that there is efficient uptake of exogenous Gal3 by dopaminergic cells in vitro.
80 Figure 30. Early Gal3 expression in the ventral mesencephalon in response to injection of adenovirus expressing hSYN. (A) Western Blot of Gal3 from brain homogenates from WT and Gal3KO mice revealed constitutive expression of Gal3 in WT mice. (B) Western Blot quantification of total Gal3 protein in WT mesencephalon samples. No difference was found between contralateral and ipsilateral hemisphere. Data is expressed as percentage fold to actin. (C) hSYN/Gal3 double immunofluorescence from injection area of WT brains 2 weeks after injection revealed no colocalisation and big release of Gal3. Scale bar 50 µm. Gal3lo/hSYN colocalisation (white arrow) can be found near highly reactive Gal3+ cell indicating Gal3 release and neuronal internalisation. Scale bar 10 µm. (D) hSYN/Gal3 double immunofluorescence of WT brains 4 weeks after adenovirus injection revealed neuronal Gal3 staining. Note that higher neuronal Gal3 correlated with lower hSYN reactivity and vice versa. Scale bar 10 µm. (E) Double immunofluorescence 6 months after adenovirus injection showed clusters of microglial cells highly reactive for Gal3. Microglia pSyn internalisation did not depend on Gal3 presence. Microglial marker (CD11b) colocalise with pSyn in both genotypes. Scale bar 20 µm. Trem2-MerTK alternative activation
81 Gal3 is one of the most upregulated genes in DAM microglia, and identified previously in AD. Unlike AD, PD presents intracellular aggregates, which can derive in a different microglia polarization state. Trem2 is a specific myeloid receptor, whose mutations have been linked to increased AD risk. However, the role of Trem2 in PD is still not defined. Our group has previously identified a precise interaction of Gal3 with Trem2 in an AD in vivo model (Boza-Serrano, et al., 2019). Consequently, we wondered if Trem2 levels were altered in our PD model. mRNA analysis revealed an Trem2 induction in animals overexpressing hSYN for 6 months (Figure 31A). Interestingly, Gal3KO failed to induce Trem2 expression, suggesting that Trem2 induction is somehow related to the presence of Gal3. Gal3 has also been shown to act as an opsonin of MerTK, (Nomura, et al., 2017; Caberoy, et al., 2012) a myeloid receptor implied in phagocytosis that has been described as an homeostatic gene. Surprisingly, transcription levels of MerTK revealed an induction in MerTK expression in WT animals injected with hSYN adenovirus (Figure 31B), suggesting a possible phagocytic phenotype in microglia. In contrast, levels of MerTK mRNA were significantly lower in Gal3KO mice overexpressing hSYN revealing a significant effect of Gal3 deletion in MerTK expression. In an interesting approximation of the effect of Gal3 in MerTK function, we developed an in vitro experiment where microglia BV2 cells were exposed to apoptotic neurons (N27 cell line), Gal3 or both. Apoptotic neurons were obtained after exposure of N27 cells to UV light for 20 minutes. 100 nM of Gal3 were used for this experiment. We measure the transcription levels of main homeostatic and DAM genes 18 hours after stimulation (Figure 31C). Our preliminary data suggests three hypothesis involving MerTK: i) MerTK was upregulated by apoptotic neurons stimuli, ii) MerTK was downregulated by Gal3 in absence of apoptotic neurons, iii) MerTK was synergistically upregulated in presence of Gal3 and apoptotic neurons. Additionally, our study seems to indicate a strong feedback signal for Gal3 that induced its own expression alone and in combination with apoptotic neurons. Finally, other genes like Igf1r, ApoE and Trem2 also showed moderate induction in presence of Gal3 and apoptotic neurons, but failed to replicate the induction pattern of homeostatic vs DAM genes. In contrast, the experiment show that different transcriptional signatures have to be taken into account. Altogether, upregulation of both Trem2 and MerTK suggests an activation of microglia that could resemble that identified in other neurological diseases as DAM microglia. In contrast, Gal3 deletion prevented this induction, thus indicating a potent role of Gal3 for the induction or maintaining of this phenotype. Moreover, preliminary data supports an important role for Gal3 in MerTK induction during phagocytosis of apoptotic neurons but failed to replicate the DAM phenotype, and instead, homeostatic genes (MerTK and Igf1r) and DAM genes (Gal3, ApoE and Trem2) were induced in this Gal3-phagocytic microglial phenotype.
82 Figure 31. Alternative activation implying Trem2 and MerTK is observed. (A) Expression analysis of Trem2 and MerTK in whole brain sections. Trem2 expression is significantly elevated in WT mice overexpressing hSYN (WT asyn). p<0.05. (B) MertK expression is significantly elevated in WT mice overexpressing hSYN. p<0.05. (C) Transcriptomic analysis of main DAM and homeostatic genes in BV2 cells challenged with Gal3 or apoptotic neurons. Autophagy analysis Alteration of autophagy has also been related to PD pathogenesis in the context of αSyn accumulation (Xilouri, et al., 2016). Discovery of LRRK2 mutations as cause of some forms of autosomal PD (Funayama, et al., 2002) and the implication of LRRK2 in macroautophagy and chaperon-mediated autophagy (CMA) has emphasised the potential role of this pathway. Consequently, we measured macroautophagy by measurement of autophagy flux (Yoshii and Mizushima, 2017). Firstly, we measured the transcriptional levels of main components of the autophagy-lysosome pathway. While nor autophagy (LC3) nor lysosomal (LAMP1) receptors showed any difference, Cathepsin B (Cat.B), one of the main lysosomal proteases, was
83 significantly decreased in Gal3KO brains (Figure 32A). However, Western blot analysis of total proteins showed a different pattern. LC3 maturation forms (LC3-I and LC3-II) were increased in Gal3KO mice along with Cat.B, while LAMP1 levels were lower (Figure 32B). We also tested levels of Hsp70, the family of chaperones that are known to participate in other form of autophagy known as chaperon-mediated autophagy (CMA). The levels of Hsp70 proteins were similar in WT and Gal3KO mice indicating no disturbance in this pathway. The effects shown in all these markers were observed also in the contralateral side of the brain (data not shown) meaning that LGALS3 gene deletion alters macroautophagy flux. Indeed this alteration is independent on hSYN overexpression. Our data suggest an impairment in autophagosome-lysosome fusion or in lysosome availability, what has to be considered as a potential cause for hSYN accumulation. Figure 32. Autophagy flux is different in Gal3KO mice. (A) Macroautophagy flux was measured by comparing transcriptional and protein levels. mRNA analysis of brain sections containing ventral mesencephalon of injected mice revealed no difference in lysosomal or autophagosomal markers but a difference in the protease Cathepsin B. P<0.001. (B) Western Blot analysis of protein extracted from SN of injected mice showed a different pattern for lysosomal and autophagosomal markers and proteases. ** p<0.01, *p<0.05. Absence of Gal3 protects from ipsilateral motor impairment
84 Neurodegeneration of dopaminergic neurons in the SN is known to be the main cause of motor symptoms in PD (Kalia and Lang, 2015). For that reason, we performed a combination of motor behavioural tests 8 weeks and 6 months after injection of the adenovirus expressing hSYN to address motor performance of mice along the experiment. Open Field Open Field is a widely use test to measure general motor impairment. In this test, mice are placed in an open space and able to freely move around. Measures like distance travelled, resting time, activity or tigmotaxis are recorded. As previously published by other authors (Caudal, et al., 2015), no differences were found in terms of travelled distance or total activity in response to overexpression of hSYN (Figure 33A & 33B). We conclude that, in our experimental conditions, ipsilateral overexpression of hSYN fail to alter basic locomotor activity or spontaneous movement. Figure 33. No general locomotor impairment was observed in mice overexpressing hSYN. Mice were placed in an open space and let to move around. Several parameters were analysed including distance travelled, resting time, time in border and center, mean speed or total activity. Data of distance travelled (A) and total activity (B) are shown. Data is expressed as fold change of individual data from mice unilaterally injected with adenovirus expressing hSYN 24 weeks after injection and compared with individual data obtained from 8 weeks after injection. Rotarod Rotarod test is used to detect any motor coordination impairment. To perform the test, mice are forced to keep their balance over a moving roller. Number of total falls are then measured over 5 days of experiments. WT overexpressing hSYN (WT hSYN) were the group that showed more deterioration over time, greatly increasing number of falls throughout the 6 months of
91 propagation view of the disease. However, if that is the case, lack of Gal3 seems to rather increase the release of hSYN from infected neurons, or, for instance, reduce the uptake of hSYN by other neurons or glial cells. Nevertheless, this concept needs further research and has not been addressed in this study. Gal3 deletion reduces pSyn levels and protects from neurodegeneration Phosphorylation of Serine129 is one of the best-known markers for pathological αSyn aggregates, which is abundantly found in LB prom PD patients. Yet, the real function of this posttranslational modification is still in debate. For instance, in vitro phosphorylation has been shown to impair αSyn aggregation (Paleologou, et al., 2008). However, Vekrellis and colleagues have recently discovered that despite S129 phosphorylation is not needed for fibril formation and seeding effect in mice, it enhances neurotoxicity (Karampetsou, et al., 2017). Here we show that in our in vivo model pSyn is increased in WT but not Gal3KO mice inversely to hSYN aggregation state. In addition, despite both mice genotypes displayed neurons with an evident accumulation of pSyn, WT neurons presented weaker TH staining and reduced surrounding dendrites. In contrast, Gal3KO neurons that presented pSyn accumulations looked apparently healthy with normal TH staining and abundant dendrite arborisation. The apparent high toxicity of pSyn in WT was later proved by stereological cell counting of SN dopaminergic neurons. WT mice showed a consistent 25% loss of nigral neurons while Gal3KO did not show any degeneration. It has to be noted that phosphorylation of αSyn is not the only factor that influences its toxicity and several studies point towards αSyn structure and oligomerisation as the determinant factor for toxicity (Cremades, et al., 2012; Lam, et al., 2016). Interestingly, GFP overexpression did not promote any degeneration nor TH staining loss as has been noticed to occur in rats by the provider of the viral vector in rats (The Michael J. Fox Foundation for Parkinson’s Research). The demonstrated neurodegeneration in WT but not in Gal3KO mice, along with the fact that Gal3KO mice developed a higher number of hSYN intracytoplasmic accumulations, supports the theory that αSyn aggregation is not a toxic process itself but rather a protective mechanism for the cell. In either case, our data supports a deleterious role of Gal3 in αSyn-associated damage to the dopaminergic system that goes beyond that already known as a lysosomal-rupture marker. Indeed, we provide evidence supporting a novel role of Gal3 in the aggregation/disaggregation process of αSyn, one of the main features of PD pathology. Furthermore, the presented data shed light about the role of LB-like aggregation forms of αSyn, which remains as a stable long-term response to the accumulation of the protein that does not correlate with the severity of neurodegeneration. Interestingly, pSyn levels correlated with dopaminergic neurotoxicity but not with the aggregation state of hSYN what supports that both are separate processes. Finally, despite further experiments are needed to address the final fate of LB-like inclusions seen in Gal3KO mice, and with the basis of our in vitro experiments, our data suggest that Gal3 may play a role in a late process of αSyn disaggregation from pathological aggregates.
92 Disaggregation of stable αSyn aggregates would be inherently associated with the release of toxic oligomeric intermediate, which have been proposed as the most toxic species associated with αSyn (Cremades, et al., 2012). Minor inflammatory response and transient Gal3 expression Neuroinflammation has long been related to PD (Ransohoff, 2016). Brains from PD patients were first shown to present reactive microglia. Furthermore, anti-inflammatory drugs have been associated with a lower incidence of PD. Within this context, microglia, the main innate immune cells of the CNS, are responsible for key tasks like phagocytosis, apoptosis induction, cytokine liberation and, remarkably, Gal3 expression and release. Our results showed a minor inflammatory response in both WT and Gal3KO mice long after hSYN injections. Main cytokines were maintained at basal levels including IL-1β that has been linked to PD through activation of the inflammasome pathway discovered in some PD patients (von Herrmann, et al., 2018). Yet, levels of TNFα were significantly higher in mesencephalon of WT mice as compared with Gal3KO 6 month after the injection of adenovirus. TNFα is a neurotoxic cytokine (Prajapati, et al., 2015) that has been long associated with neurodegenerative diseases. In PD, TNFα was shown long ago to be increased in post-mortem analysis of patient brains and CSF (Mogi, et al., 1994) and activation of TNFR1 is also known to provoke neurodegeneration in PD models (Prajapati, et al., 2015; Fischer and Maier, 2015). Similarly, neurodegeneration can stimulate microglial TNFα expression through DAMPs release and associated sterile inflammation (Banjara and Ghosh, 2017). In our model, we found that TNFα increase is associated to the neurodegenerative process. However, sustained neuroinflammation does not seem to occur in this model as other classical cytokines studied were maintained at basal levels. Nevertheless, it is important to highlight that this analysis was performed 6 months after the adenovirus injection what suggest that microglia-associated immune response probably happened in previous stages of the process. Previously, we have demonstrated that neurons can internalise exogenous Gal3 from the extracellular space in vitro and have shown intraneuronal Gal3 in PD brain samples but the origin of the Gal3 observed in these samples is unknown. Hence, we wondered if reactive microglia was the source of this Gal3 in our mouse model. Firstly, 2 weeks after the injection of adenovirus, we discovered a bigger cluster of Gal3 overexpressing microglia in the vicinities of the injection site. In this exacerbated conditions we were able to observe the phenomenon of some neurons that presented a weak Gal3 staining when surrounded by highly reactive microglia. Secondly, 4 weeks after the injection, microglial overexpression was abolished but we were able to find some neurons that presented the Gal3 puncta-like pattern that we had previously described in PD brain sections. Indeed, higher levels of neuronal Gal3 correlated with lower levels of cytoplasmic hSYN. Those observations support and extend the current knowledge from Gal3 that was identified to be secreted in a paracrine manner by microglial cells to promote inflammatory response (Burguillos, et al., 2015), but in this precise case, are neurons the receptor of the Gal3 secretion. Additionally, 6 months after
93 the injection we found small clusters of microglia with high expression of Gal3 that was related to pSyn capture. However, Gal3KO microglia also showed pSyn capture, which means that Gal3 does not mediate pSyn internalisation. It is known that microglia can secrete Gal3 upon αSyn stimulus (Boza-Serrano, et al., 2014), for that reason we analysed the total expression of Gal3 in ventral mesencephalon 6 months after the injection Remarkably, our western blot analysis revealed a clear expression of Gal3 in the ventral mesencephalon irrespective of the adenovirus injection what strongly supports a constitutive expression of Gal3 in the substantia nigra. Thus, the detection of basal Gal3 expression supported by transient and local microglial Gal3 reactivity suggests that Gal3 is constantly available for neurons that can internalise Gal3 and interact with αSyn. Overall, we conclude that the long-term inflammatory response under conditions of hSYN expression is rather low. Yet, we found an increase in TNFα levels in WT but not in Gal3KO mice 6 months after adenovirus injection that correlated with dopaminergic cell death, thus raising as a potential candidate mediating increased neurotoxicity. Microglia is the main source of TNFα in the brain and its elevation suggests a change in their phenotype. A more exhaustive study of precise microglial phenotype has to be done to discover a possible PD-specific DAM phenotype. However, the fact that αSyn accumulation is an intracellular process and the neurodegeneration is a progressive feature could lead to the formation of transient clusters of locally activated microglia that can secrete Gal3 to be further incorporated by neurons in a paracrine manner. Regardless of this consideration, our data demonstrate an unequivocal expression of Gal3 in the ventral mesencephalon, a prerequisite for playing a potential role in αSyn-associated damage. Signs of alternative activation pathway by Gal3 ligands TREM2 and MerTK Discovery of a TREM2-APOE pathway governing the recently described microglial DAM phenotype in neurodegenerative diseases and particularly AD (Krasemann, et al., 2017) has shed light on the importance of TREM2 in neurodegeneration. Currently, divergence exists regarding the implications of TREM2 mutations in PD onset (Ortega-Cubero, et al., 2015; Walton, et al., 2016). Nevertheless, TREM2 expression is known to modulate microglial activation in several PD models including MPTP and overexpression of hSYN (Ren, et al., 2018; Guo, et al., 2019). In both cases, TREM2 expression is related to an attenuated microglial response and lower associated neurodegeneration. Strikingly, in AD, TREM2 is thought to play the inverse role supporting the neuroinflammation process and driving microglial plaque-associated DAM phenotype (Jay, et al., 2015; Wang, et al., 2015). Our results showed an increase in the TREM2 transcriptional expression in WT mice overexpressing hSYN. Interestingly, Gal3KO mice failed to increase TREM2 transcriptional levels suggesting that either TREM2 is strictly related to neurodegeneration (only observed in WT mice) or Gal3 mediates TREM2 response. This overexpression and the previously described mild neuroinflammation situate our model in a middle term respect previous studies in PD and the role for TREM2 in AD. In fact, TREM2 increase does not seem to mediate an important inflammatory response but is present during the
94 neurodegeneration process. The last view is supported by our previous study demonstrating a functional interaction between these two proteins (Boza-Serrano, et al., 2019). In this study, we demonstrated that Gal3 is needed for full microglial activation in response to fibrillary Aβ and that Gal3 acts as an endogenous ligand of TREM2. The long-term upregulation of TREM2 in response to sustained adenovirus-associated hSYN expression supports the view that TREM2 plays an important role in our model and hence the presence of a PD-associated DAM phenotype. MerTK is a phagocytic receptor known for mediating phagocytosis of apoptotic and stressed dying cells by binding to the known as ‘eat-me signals’. Its role in neurodegenerative disease is poorly documented but it has been suggested to be one of the microglial homeostatic genes (Keren-Shaul, et al., 2017; Krasemann, et al., 2017). MerTK along with Tyro3 and Axl form part of the TAM receptors family of protein tyrosine kinases that regulates physiological phagocytosis specially that of apoptotic neurons. In a recent study, Lemke and colleagues demonstrated that the absence of MerTK or Axl leads to the accumulation of apoptotic neurons and debris (Fourgeaud, et al., 2016). Indeed, in a transgenic model of PD using mutant A53T synucleinexpressing mice discovered that Axl expression was increased in the spinal cord despite no variation was found in MerTK expression (Fourgeaud, et al., 2016). However, despite the role of MerTK in PD still awaits for further characterization, a relation between MerTK and Gal3 has been recently proposed. Initially, Gal3 was described as an ‘eat-me signal’ that enhances MerTK mediated phagocytosis in macrophages (Caberoy, et al., 2012). Recently, Brown and colleagues demonstrated that activated microglia release Gal3 to further bind to MerTK and glycoproteins from stressed neurons promoting phagocytosis (Nomura, et al., 2017). Recent massive microglial transcriptomic studies under disease conditions have revealed a consistent downregulation of homeostatic genes along with robust upregulation of disease-associated genes (Keren-Shaul, et al., 2017; Krasemann, et al., 2017). Intriguingly, our results showed a clear induction of MerTK expression in the ventral mesencephalon of WT mice upon hSYN overexpression. A plausible explanation for this effect would be an increment in phagocytosis of apoptotic neurons inherent to the dopaminergic neurodegeneration. Interestingly, Gal3KO mice failed to increase MerTK expression after hSYN overexpression what Overall, TREM2 and MerTK transcriptional induction support the presence of an alternative microglial phenotype. Our results should bring attention to these two markers and their potential and poorly studied roles in PD. Moreover, we confirmed that the induction of both receptor is impaired in Gal3KO and, more importantly, confirmed their relationship with Gal3 in an in vivo model of PD. Gal3-mediated autophagy as a potential player in αSyn managing Glycosylation is a post-translational modification that commonly occur to membrane proteins and is a key process for extracellular proteins. However, it is known that glycosylation of intracellular proteins is an age-related process. Indeed, glycosylated proteins like amyloid-β, tau
95 or prions have been identified in several neurodegenerative diseases (Vicente Miranda and Outeiro, 2010). Advanced glycation end-products (AGEs) can appear as a result of glycosylation. AGEs have been identified in PD patients associated with LB (Castellani, et al., 1996), while Gal3 has been proposed as a ligand for AGEs (Castellani, et al., 1996). The effect of glycation in αSyn models was recently investigated and suggested to cause impaired clearance and promote the accumulation of αSyn by autophagy dysfunction (Vicente Miranda, et al., 2017; Wani, et al., 2017). The role of Gal3 in AGEs-mediated accumulation of αSyn is still unknown but data strongly suggest a critical role for Gal3 in this process. Indeed, our results show autophagy flux deregulation in Gal3 knockout mice affecting LC3B and LAMP1 markers compared to WT conditions which indicate a different role for autophagy in both animals. Two major autophagy mechanisms have been related to PD and αSyn. Macroautophagy is able to degrade large quantity of structures and biomolecules by engulfing them in doublemembrane vesicles called autophagosomes that are further fused with lysosomes to form the autolysosome with low pH and high degradation capacity. In contrast, chaperon-mediated autophagy (CMA) is based on the ability of some chaperones to directly introduce target proteins into lysosomes (Xilouri, et al., 2016). While macroautophagy is recognised as a clearance method for misfolded αSyn, it has also been proved to be ineffective against Lewy-like αSyn aggregates (Tanik, et al., 2013). In fact, these aggregates impair this process. However, upon fibrils accumulation, induction of autophagy prevents from formation of αSyn inclusions (Gao, et al., 2019). Thus, the role of autophagy seems to be specific for toxic less-aggregated species, despite no clear mechanism has been well established. In our mice model, autophagy flux seems to be impaired. In WT mice autophagy flux did not vary between the intact and the hSYN-overexpressing hemisphere. The same was true in Gal3KO mice, thus indicating that no change in autophagy is due to animal treatment. However, differences were detected between both genotypes indicating a basal change in autophagy flux. For instance, Gal3KO mice presented an accumulation of autophagy marker species LC3B-I and LC3B-II and a decrease in the lysosomal marker LAMP1 at the protein but not at mRNA level. This may suggest a lower autophagosome-lysosome fusion in Gal3KO mice due to lower availability of LAMP1. This can be caused by increased autophagy flux triggered by the other affected autophagy pathway, CMA. However, protein levels of Hsp70, the family that includes Hsc70 chaperone that is a known mediator of this pathway, remained unchanged in Gal3KO and WT mice in response to hSYN overexpression, thus discarding an increase in CMA autophagy. Another possibility that arises is a higher degradation of lysosomes in Gal3KO mediated by αSyn. This is supported by the data that states that αSyn can promote lysosomal rupture (Flavin, et al., 2017). In fact, Tanaka and colleagues claim that αSyn aggregation starts in the lumen of lysosomes mediated by cathepsin B and rapidly moves to the cytosol (Tsujimura, et al., 2015). Lower LAMP1 and accumulation of cathepsin B is a feature found in Gal3KO mice but not in WT what may be caused by this mechanism. Overall, Gal3 is undoubtedly related to the autophagy-lysosome pathway, probably related to damaged membranes, either in lysosomes or organelles. Despite the precise role of Gal3 is still
96 undetermined, lysosomal compartment and autophagy pathways emerge as a candidate location for Gal3 interaction with αSyn. Moreover, the implication of autophagy in PD is also well-known, and our data suggest that Gal3KO mice developed a different autophagy flux. Despite more research is needed to know the precise relation of Gal3 with autophagy, our data states that the different regulation of autophagy flux in Gal3KO mice can be related to higher αSyn accumulation. Motor impairments are prevented in Gal3KO mice Motor symptoms are still the hallmark for PD diagnosis. Correlation between motor symptoms and dopaminergic degeneration is not precise but it is estimated that motor impairments can be detected after 30-60% of dopaminergic loss (Kordower, et al., 2013). Identification of motor symptoms in mice can be challenging as essential motor symptoms like tremor or bradykinesia are measured by general tests like the open field. Literature shows that behavioural tests efficiently detect major motor impairments. Most of the published data were obtained from neurotoxic PD animal models; for instance, MPTP and 6-hydroxydopamine models. In these models, the open field test can detect a decrease in motor activity after 80% loss of striatal dopamine content (Taylor, et al., 2010). Our results showed no differences between both experimental groups in any of the parameters obtained from the open field test including global locomotor abilities with features like total distance and global activity. The open field is, however, highly influenced by an anxious behaviour of mice (Brooks and Dunnett, 2009). However, we failed to detect differences in parameters of anxiety between both experimental groups. This states that overall motor behaviour was similar between genotypes with no major locomotor impairments. This analysis fits with the modest loss of nigral dopaminergic neurons (about 25%) seen in the ipsilateral side of WT mice overexpressing hSYN. Rotarod is an easy test that measures motor coordination and balance. It has been widely used to test motor performance in animal PD models. Indeed, it has been effective in detecting behavioural disturbances after either MPTP or αSyn oligomer injections (Sedelis, et al., 2001; Fortuna, et al., 2017). In our model, we performed the test 8 and 24 weeks after injection of the viral vector. Our analysis showed that WT mice injected with adenovirus present a higher and progressive motor impairment at 24 weeks compared to the earlier time point analysed (8 weeks). All groups studied performed the test slightly worse at 24 weeks (vs. 8 weeks) after injection. This can be easily attributed to aging. However, WT mice overexpressing hSYN showed an additional motor impairment that can be attributed to hSYN-associated dopaminergic damage. As previously stated, neurodegeneration is considered as the main cause of motor impairment, but accumulation of αSyn has also been proposed to be determinant. Recently, Romero-Ramos and colleagues detected motor impairments in the early stages of hSYN overexpression in rats before neurodegeneration appeared (Phan, et al., 2017). Interestingly,
97 they detected synaptic dysfunction in the dopaminergic terminals from the striatum. Despite synaptic function has not been tested in this study, is remarkable that Gal3KO mice injected with hSYN did not show any motor impairment what suggests that Gal3 role in αSyn aggregation can be extended to the striatum. Regardless of these observations, our data strongly support a precise relation between neurodegeneration and motor impairment. Moreover, rotarod is an easy and reproducible test that can be performed either at constant speed or in acceleration. We used the accelerating version in order to obtain more sensitivity (Brooks and Dunnett, 2009). Under these conditions, fatigue can be a determinant factor beyond any motor impairment. While it is known that fatigue is a common symptom of PD in humans (Schapira, et al., 2017), it is poorly studied in mice models of PD. In a recent study by Saligan and colleagues, the authors defined the fatigue behaviour based on lower activity in the open field, and lower time on running wheels (Wolff, et al., 2018). In the same study, the authors failed to observe any change in the rotarod test. However, fatigue is unlikely to appear in our model as the open field failed to detect any change between groups or any depression-like behaviour that is thought to accompany fatigue symptom. Indeed, we performed the horizontal bar test (data not shown) where mice are placed on a horizontal bar only grasp by their forepaws. Latency to fall is measured. This test requires not only motor coordination but also dependence on strength and muscle tone (Deacon, 2013). No differences were found in this test, which strongly suggests no change in neither strength nor muscle fatigue. Lastly, as our model consisted of an ipsilateral injection, we performed the cylinder test. This is a simple and reliable test highly sensitive for ipsilateral impairment (Brooks and Dunnett, 2009). In this test, mice forepaws voluntary movement based on the exploratory instinct of mice are measured and major PD symptoms like bradykinesia or rigidity can be observed. Actually, it has been widely used for the study of ipsilateral impairment in PD models like 6-hydroxydopamine or MPP+ (Iancu, et al., 2005; Mendieta, et al., 2009). Because of its great sensitivity, it is especially appropriated to detect ipsilateral deficits. Our model setting of ipsilateral injection makes the cylinder test the most suitable behavioural test for our study. Interestingly, this analysis demonstrated an unequivocally ipsilateral impairment in WT mice overexpressing hSYN. In contrast, we failed to detect such an ipsilateral impairment in Gal3KO mice, full validating the deleterious role of Gal3 in dopaminergic integrity. The impairment in the cylinder test performance can only be attributed to ipsilateral damage. In addition, no other group gave signs of ipsilateral damage meaning that injection-derived mechanical damage was absent or minimal and results can only be attributed to hSYN overexpression. Overall, this data along with rotarod and other complementary behavioural tests strongly suggest that overexpression of hSYN in WT SN promoted selective motor impairment that worsens motor coordination without affecting mice basic behaviour in full agreement with ipsilateral dopaminergic neurodegeneration and loss of dendritic arborisation. Our data fit well with initial stages of PD with evident but still no disabling motor symptoms. Importantly, Gal3KO mice did not develop any motor symptoms what assures that observed neuroprotection is also accompanied by preserved functionality. Thus, Gal3 emerges as a disease-changing molecule,
98 as the deletion of Gal3 breaks the link between the hallmarks of PD disease: αSyn accumulation, neurotoxicity, and motor impairments. Our study presents strong experimental data supporting an important role of Gal3 in the pathogenesis of PD. Our mice model showed that the deletion of Gal3 prevents from dopaminergic degeneration and motor impairment. Besides, we provide evidence supporting the existence of a specific PD-associated microglial phenotype involving TREM2 and MerTK as important receptors for this phenotype. Moreover, our data reinforce the view that αSyn aggregation is not a neurotoxic process by itself. Remarkably, Gal3 appears as a key driver of αSyn-associated neurotoxicity. Indeed, we provide compelling evidence demonstrating the ability of Gal3 to interact with αSyn to further inhibit and retrace its aggregation. Finally, the identification of Gal3 in human samples from PD patients and its association with pathological αSyn aggregates supports the importance of Gal3 in PD pathogenesis. Overall, Gal3 emerges as a promising therapeutical target for PD treatment.
99 CONCLUSIONS In the present study, we have performed a wide array of experimental data to analyse the potential and promising role of Gal3 in PD pathogenesis. Collectively, we considered that our experimental data allow us to conclude the following: I. Gal3 is associated with αSyn in synucleopathies including DLB and PD. II. Gal3 is associated with non-dense accumulation of αSyn in PD patients. III. Gal3 can be internalised by neurons from the extracellular space in a lysosomeindependent manner. IV. Gal3 inhibits and retrace αSyn aggregation in vitro. V. In a mice model of overexpression of human αSyn, Gal3 deletion promotes αSyn aggregation. VI. Aggregation of αSyn in Gal3KO mice is neuroprotective and maintains neuronal functionality to ultimately preserve motor impairment. VII. Microglia display a PD-like phenotype with mild inflammatory response and highlighted by robust upregulation of TREM2 and MERTK. VIII. Gal3 deletion impairs microglial activation. Thus, Gal3 appears as a key mediator of this response. IX. Gal3KO mice present a different autophagy flux that may be involved in αSyn clearance and toxicity. X. There is a constitutive basal expression of Gal3 in the ventral mesencephalon that together with transient and local expression of microglial Gal3 can be sufficient to promote neuronal Gal3 internalisation and Gal3-αSyn interaction.
100 REFERENCES 1 Aits, S. Methods to Detect Loss of Lysosomal Membrane Integrity. Methods Mol Biol 1880, 315-329, doi:10.1007/978-1-4939-8873-0_21 (2019). 2 Aits, S. et al. Sensitive detection of lysosomal membrane permeabilization by lysosomal galectin puncta assay. Autophagy 11, 1408-1424, doi:10.1080/15548627.2015.1063871 (2015). 3 Alafuzoff, I. & Hartikainen, P. Alpha-synucleinopathies. Handb Clin Neurol 145, 339-353, doi:10.1016/b978-0-12-802395-2.00024-9 (2017). 4 Alfaro, J. F. et al. Tandem mass spectrometry identifies many mouse brain O- GlcNAcylated proteins including EGF domain-specific O-GlcNAc transferase targets. Proc Natl Acad Sci U S A 109, 7280-7285, doi:10.1073/pnas.1200425109 (2012). 5 Anderson, J. P. et al. Phosphorylation of Ser-129 is the dominant pathological modification of alpha-synuclein in familial and sporadic Lewy body disease. J Biol Chem 281, 29739-29752, doi:10.1074/jbc.M600933200 (2006). 6 Angot, E., Steiner, J. A., Hansen, C., Li, J. Y. & Brundin, P. Are synucleinopathies prionlike disorders? Lancet Neurol 9, 1128-1138, doi:10.1016/s1474-4422(10)70213-1 (2010). 7 Antunes, L. et al. Similar α-Synuclein staining in the colon mucosa in patients with Parkinson's disease and controls. Mov Disord 31, 1567-1570, doi:10.1002/mds.26702 (2016). 8 Anwar, S. et al. Functional alterations to the nigrostriatal system in mice lacking all three members of the synuclein family. J Neurosci 31, 7264-7274, doi:10.1523/jneurosci.6194- 10.2011 (2011). 9 Aprile, F. A., Sormanni, P. & Vendruscolo, M. A Rational Design Strategy for the Selective Activity Enhancement of a Molecular Chaperone toward a Target Substrate. Biochemistry 54, 5103-5112, doi:10.1021/acs.biochem.5b00459 (2015). 10 Arnold, T. & Betsholtz, C. The importance of microglia in the development of the vasculature in the central nervous system. Vasc Cell 5, 4, doi:10.1186/2045-824x-5-4 (2013). 11 Ascherio, A. & Schwarzschild, M. A. The epidemiology of Parkinson's disease: risk factors and prevention. Lancet Neurol 15, 1257-1272, doi:10.1016/s1474-4422(16)30230-7 (2016). 12 Azeredo da Silveira, S. et al. Phosphorylation does not prompt, nor prevent, the formation of alpha-synuclein toxic species in a rat model of Parkinson's disease. Hum Mol Genet 18, 872-887, doi:10.1093/hmg/ddn417 (2009). 13 Bänfer, S. et al. Molecular mechanism to recruit galectin-3 into multivesicular bodies for polarized exosomal secretion. Proc Natl Acad Sci U S A 115, E4396-E4405, doi:10.1073/pnas.1718921115 (2018). 14 Banjara, M. & Ghosh, C. Sterile Neuroinflammation and Strategies for Therapeutic Intervention. Int J Inflam 2017, 8385961, doi:10.1155/2017/8385961 (2017). 15 Banou, E. Kinesia paradoxa: a challenging Parkinson's phenomenon for simulation. Adv Exp Med Biol 822, 165-177, doi:10.1007/978-3-319-08927-0_18 (2015). 16 Barcia, C. et al. Evidence of active microglia in substantia nigra pars compacta of parkinsonian monkeys 1 year after MPTP exposure. Glia 46, 402-409, doi:10.1002/glia.20015 (2004). 17 Bard, J. A. M. et al. Structure and Function of the 26S Proteasome. Annu Rev Biochem 87, 697-724, doi:10.1146/annurev-biochem-062917-011931 (2018).
107 137 Holtz, P., Heise, R. & Lüdtke, K. Fermentativer Abbau von l-dioxyphenylalanin (Dopa) durch Niere. Naunyn-Schmiedeberg's Archives of Pharmacology 191, 87-118 (1938). 138 Hornykiewicz, O. The action of dopamine on the arterial blood pressure of the guineapig. Br J Pharmacol Chemother 13, 91-94, doi:10.1111/j.1476-5381.1958.tb00197.x (1958). 139 Hornykiewicz, O. in Dopaminergic Systems and their Regulation (eds G. N. Woodruff, J. A. Poat, & P. J. Roberts) 3-18 (Palgrave Macmillan UK, 1986). 140 Hoshiko, M., Arnoux, I., Avignone, E., Yamamoto, N. & Audinat, E. Deficiency of the microglial receptor CX3CR1 impairs postnatal functional development of thalamocortical synapses in the barrel cortex. J Neurosci 32, 15106-15111, doi:10.1523/jneurosci.1167-12.2012 (2012). 141 Hostiuc, S., Drima, E. & Buda, O. Shake the Disease. Georges Marinesco, Paul Blocq and the Pathogenesis of Parkinsonism, 1893. Front Neuroanat 10, 74, doi:10.3389/fnana.2016.00074 (2016). 142 Iancu, R., Mohapel, P., Brundin, P. & Paul, G. Behavioral characterization of a unilateral 6-OHDA-lesion model of Parkinson's disease in mice. Behav Brain Res 162, 1-10, doi:10.1016/j.bbr.2005.02.023 (2005). 143 Ikeuchi, T. et al. Patients homozygous and heterozygous for SNCA duplication in a family with parkinsonism and dementia. Arch Neurol 65, 514-519, doi:10.1001/archneur.65.4.514 (2008). 144 Jankovic, J. Parkinson's disease: clinical features and diagnosis. J Neurol Neurosurg Psychiatry 79, 368-376, doi:10.1136/jnnp.2007.131045 (2008). 145 Jaquenod De Giusti, C. et al. Galectin-3 is upregulated in activated glia during Junin virusinduced murine encephalitis. Neurosci Lett 501, 163-166, doi:10.1016/j.neulet.2011.07.007 (2011). 146 Jay, T. R. et al. TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer's disease mouse models. J Exp Med 212, 287-295, doi:10.1084/jem.20142322 (2015). 147 Jellinger, K. A. Dementia with Lewy bodies and Parkinson's disease-dementia: current concepts and controversies. J Neural Transm (Vienna) 125, 615-650, doi:10.1007/s00702-017-1821-9 (2018). 148 Jeon, S. B. et al. Galectin-3 exerts cytokine-like regulatory actions through the JAK-STAT pathway. J Immunol 185, 7037-7046, doi:10.4049/jimmunol.1000154 (2010). 149 Jiang, P., Gan, M., Yen, S. H., McLean, P. J. & Dickson, D. W. Impaired endo-lysosomal membrane integrity accelerates the seeding progression of α-synuclein aggregates. Sci Rep 7, 7690, doi:10.1038/s41598-017-08149-w (2017). 150 Johannes, L., Jacob, R. & Leffler, H. Galectins at a glance. J Cell Sci 131, doi:10.1242/jcs.208884 (2018). 151 Joshi, N. & Singh, S. Updates on immunity and inflammation in Parkinson disease pathology. J Neurosci Res 96, 379-390, doi:10.1002/jnr.24185 (2018). 152 Jouault, T. et al. Specific recognition of Candida albicans by macrophages requires galectin-3 to discriminate Saccharomyces cerevisiae and needs association with TLR2 for signaling. J Immunol 177, 4679-4687, doi:10.4049/jimmunol.177.7.4679 (2006). 153 Jyothi, H. J. et al. Aging causes morphological alterations in astrocytes and microglia in human substantia nigra pars compacta. Neurobiol Aging 36, 3321-3333, doi:10.1016/j.neurobiolaging.2015.08.024 (2015). 154 Kalaitzakis, M. E., Graeber, M. B., Gentleman, S. M. & Pearce, R. K. Controversies over the staging of alpha-synuclein pathology in Parkinson's disease. Acta Neuropathol 116, 125-128; author reply 129-131, doi:10.1007/s00401-008-0381-3 (2008). 155 Kalia, L. V. & Lang, A. E. Parkinson's disease. Lancet 386, 896-912, doi:10.1016/s0140- 6736(14)61393-3 (2015).
108 156 Karabiyik, C., Lee, M. J. & Rubinsztein, D. C. Autophagy impairment in Parkinson's disease. Essays Biochem 61, 711-720, doi:10.1042/ebc20170023 (2017). 157 Karampetsou, M. et al. Phosphorylated exogenous alpha-synuclein fibrils exacerbate pathology and induce neuronal dysfunction in mice. Sci Rep 7, 16533, doi:10.1038/s41598-017-15813-8 (2017). 158 Karlsson, A. et al. Galectin-3 functions as an opsonin and enhances the macrophage clearance of apoptotic neutrophils. Glycobiology 19, 16-20, doi:10.1093/glycob/cwn104 (2009). 159 Keeney, P. M., Xie, J., Capaldi, R. A. & Bennett, J. P. Parkinson's disease brain mitochondrial complex I has oxidatively damaged subunits and is functionally impaired and misassembled. J Neurosci 26, 5256-5264, doi:10.1523/jneurosci.0984-06.2006 (2006). 160 Keren-Shaul, H. et al. A Unique Microglia Type Associated with Restricting Development of Alzheimer's Disease. Cell 169, 1276-1290.e1217, doi:10.1016/j.cell.2017.05.018 (2017). 161 Kierdorf, K. et al. Microglia emerge from erythromyeloid precursors via Pu.1- and Irf8- dependent pathways. Nat Neurosci 16, 273-280, doi:10.1038/nn.3318 (2013). 162 Kim, S. et al. Transneuronal Propagation of Pathologic α-Synuclein from the Gut to the Brain Models Parkinson's Disease. Neuron 103, 627-641.e627, doi:10.1016/j.neuron.2019.05.035 (2019). 163 Klingelhoefer, L. & Reichmann, H. Pathogenesis of Parkinson disease--the gut-brain axis and environmental factors. Nat Rev Neurol 11, 625-636, doi:10.1038/nrneurol.2015.197 (2015). 164 Konno, T., Ross, O. A., Puschmann, A., Dickson, D. W. & Wszolek, Z. K. Autosomal dominant Parkinson's disease caused by SNCA duplications. Parkinsonism Relat Disord 22 Suppl 1, S1-6, doi:10.1016/j.parkreldis.2015.09.007 (2016). 165 Kordower, J. H. et al. Disease duration and the integrity of the nigrostriatal system in Parkinson's disease. Brain 136, 2419-2431, doi:10.1093/brain/awt192 (2013). 166 Kosaka, K. Diffuse Lewy body disease. Neuropathology 20 Suppl, S73-78, doi:10.1046/j.1440-1789.2000.00301.x (2000). 167 Kosaka, K., Oyanagi, S., Matsushita, M. & Hori, A. Presenile dementia with Alzheimer-, Pick- and Lewy-body changes. Acta Neuropathol 36, 221-233, doi:10.1007/bf00685366 (1976). 168 Krasemann, S. et al. The TREM2-APOE Pathway Drives the Transcriptional Phenotype of Dysfunctional Microglia in Neurodegenerative Diseases. Immunity 47, 566-581.e569, doi:10.1016/j.immuni.2017.08.008 (2017). 169 Kumar, S. et al. Galectins and TRIMs directly interact and orchestrate autophagic response to endomembrane damage. Autophagy 13, 1086-1087, doi:10.1080/15548627.2017.1307487 (2017). 170 Kuusisto, E., Parkkinen, L. & Alafuzoff, I. Morphogenesis of Lewy bodies: dissimilar incorporation of alpha-synuclein, ubiquitin, and p62. J Neuropathol Exp Neurol 62, 1241- 1253, doi:10.1093/jnen/62.12.1241 (2003). 171 Lafora, G. R. & Glueck, B. Beitrag zur Histopathologie der myoklonischen Epilepsie. Zeitschrift für die gesamte Neurologie und Psychiatrie 6, 1-14, doi:10.1007/BF02863929 (1911). 172 Lam, H. T., Graber, M. C., Gentry, K. A. & Bieschke, J. Stabilization of α-Synuclein Fibril Clusters Prevents Fragmentation and Reduces Seeding Activity and Toxicity. Biochemistry 55, 675-685, doi:10.1021/acs.biochem.5b01168 (2016). 173 Langston, J. W., Ballard, P., Tetrud, J. W. & Irwin, I. Chronic Parkinsonism in humans due to a product of meperidine-analog synthesis. Science 219, 979-980, doi:10.1126/science.6823561 (1983).
109 174 Lassen, L. B., Reimer, L., Ferreira, N., Betzer, C. & Jensen, P. H. Protein Partners of α- Synuclein in Health and Disease. Brain Pathol 26, 389-397, doi:10.1111/bpa.12374 (2016). 175 Lee, H. J., Bae, E. J. & Lee, S. J. Extracellular α--synuclein-a novel and crucial factor in Lewy body diseases. Nat Rev Neurol 10, 92-98, doi:10.1038/nrneurol.2013.275 (2014). 176 Lee, H. J., Khoshaghideh, F., Patel, S. & Lee, S. J. Clearance of alpha-synuclein oligomeric intermediates via the lysosomal degradation pathway. J Neurosci 24, 1888-1896, doi:10.1523/jneurosci.3809-03.2004 (2004). 177 Leestemaker, Y. et al. Proteasome Activation by Small Molecules. Cell Chem Biol 24, 725- 736.e727, doi:10.1016/j.chembiol.2017.05.010 (2017). 178 Lehmann, H. C., Hartung, H. P. & Kieseier, B. C. Leopold Ordenstein: on paralysis agitans and multiple sclerosis. Mult Scler 13, 1195-1199, doi:10.1177/1352458507076979 (2007). 179 Lehnardt, S. et al. Activation of innate immunity in the CNS triggers neurodegeneration through a Toll-like receptor 4-dependent pathway. Proc Natl Acad Sci U S A 100, 8514- 8519, doi:10.1073/pnas.1432609100 (2003). 180 Lerman, B. J. et al. Deletion of galectin-3 exacerbates microglial activation and accelerates disease progression and demise in a SOD1(G93A) mouse model of amyotrophic lateral sclerosis. Brain Behav 2, 563-575, doi:10.1002/brb3.75 (2012). 181 Leverenz, J. B. et al. Proteomic identification of novel proteins in cortical lewy bodies. Brain Pathol 17, 139-145, doi:10.1111/j.1750-3639.2007.00048.x (2007). 182 Levine, P. M. et al. α-Synuclein O-GlcNAcylation alters aggregation and toxicity, revealing certain residues as potential inhibitors of Parkinson's disease. Proc Natl Acad Sci U S A 116, 1511-1519, doi:10.1073/pnas.1808845116 (2019). 183 Lewy, F. Paralysis agitans. I. Pathologische anatomie. Handbuch der neurologie (1912). 184 Li, J. Y. et al. Lewy bodies in grafted neurons in subjects with Parkinson's disease suggest host-to-graft disease propagation. Nat Med 14, 501-503, doi:10.1038/nm1746 (2008). 185 Linder, J., Stenlund, H. & Forsgren, L. Incidence of Parkinson's disease and parkinsonism in northern Sweden: a population-based study. Mov Disord 25, 341-348, doi:10.1002/mds.22987 (2010). 186 Logan, T., Bendor, J., Toupin, C., Thorn, K. & Edwards, R. H. α-Synuclein promotes dilation of the exocytotic fusion pore. Nat Neurosci 20, 681-689, doi:10.1038/nn.4529 (2017). 187 Luo, C., Koyama, R. & Ikegaya, Y. Microglia engulf viable newborn cells in the epileptic dentate gyrus. Glia 64, 1508-1517, doi:10.1002/glia.23018 (2016). 188 Machado, A. et al. Peripheral inflammation increases the damage in animal models of nigrostriatal dopaminergic neurodegeneration: possible implication in Parkinson's disease incidence. Parkinsons Dis 2011, 393769, doi:10.4061/2011/393769 (2011). 189 Machiya, Y. et al. Phosphorylated alpha-synuclein at Ser-129 is targeted to the proteasome pathway in a ubiquitin-independent manner. J Biol Chem 285, 40732- 40744, doi:10.1074/jbc.M110.141952 (2010). 190 Manyam, B. V. Paralysis agitans and levodopa in "Ayurveda": ancient Indian medical treatise. Mov Disord 5, 47-48, doi:10.1002/mds.870050112 (1990). 191 Maroteaux, L., Campanelli, J. T. & Scheller, R. H. Synuclein: a neuron-specific protein localized to the nucleus and presynaptic nerve terminal. J Neurosci 8, 2804-2815 (1988). 192 Maroteaux, L. & Scheller, R. H. The rat brain synucleins; family of proteins transiently associated with neuronal membrane. Brain Res Mol Brain Res 11, 335-343, doi:10.1016/0169-328x(91)90043-w (1991). 193 Marth, J. D. & Grewal, P. K. Mammalian glycosylation in immunity. Nat Rev Immunol 8, 874-887, doi:10.1038/nri2417 (2008). 194 McCarthy, M. M. Location, Location, Location: Microglia Are Where They Live. Neuron 95, 233-235, doi:10.1016/j.neuron.2017.07.005 (2017).
110 195 Medeiros, A. T., Soll, L. G., Tessari, I., Bubacco, L. & Morgan, J. R. α-Synuclein Dimers Impair Vesicle Fission during Clathrin-Mediated Synaptic Vesicle Recycling. Front Cell Neurosci 11, 388, doi:10.3389/fncel.2017.00388 (2017). 196 Mendieta, L. et al. The carboxyl-terminal domain of the heavy chain of tetanus toxin prevents dopaminergic degeneration and improves motor behavior in rats with striatal MPP(+)-lesions. Neurosci Res 65, 98-106, doi:10.1016/j.neures.2009.06.001 (2009). 197 Menzies, F. M. et al. Autophagy and Neurodegeneration: Pathogenic Mechanisms and Therapeutic Opportunities. Neuron 93, 1015-1034, doi:10.1016/j.neuron.2017.01.022 (2017). 198 Mittal, R. et al. Neurotransmitters: The Critical Modulators Regulating Gut-Brain Axis. J Cell Physiol 232, 2359-2372, doi:10.1002/jcp.25518 (2017). 199 Miyamoto, A. et al. Microglia contact induces synapse formation in developing somatosensory cortex. Nat Commun 7, 12540, doi:10.1038/ncomms12540 (2016). 200 Moehle, M. S. & West, A. B. M1 and M2 immune activation in Parkinson's Disease: Foe and ally? Neuroscience 302, 59-73, doi:10.1016/j.neuroscience.2014.11.018 (2015). 201 Mogi, M. et al. Tumor necrosis factor-alpha (TNF-alpha) increases both in the brain and in the cerebrospinal fluid from parkinsonian patients. Neurosci Lett 165, 208-210 (1994). 202 Mollenhauer, B. et al. α-Synuclein in human cerebrospinal fluid is principally derived from neurons of the central nervous system. J Neural Transm (Vienna) 119, 739-746, doi:10.1007/s00702-012-0784-0 (2012). 203 Nalls, M. A. et al. Large-scale meta-analysis of genome-wide association data identifies six new risk loci for Parkinson's disease. Nat Genet 46, 989-993, doi:10.1038/ng.3043 (2014). 204 Nikodemova, M., Small, A. L., Smith, S. M., Mitchell, G. S. & Watters, J. J. Spinal but not cortical microglia acquire an atypical phenotype with high VEGF, galectin-3 and osteopontin, and blunted inflammatory responses in ALS rats. Neurobiol Dis 69, 43-53, doi:10.1016/j.nbd.2013.11.009 (2014). 205 Nimmerjahn, A., Kirchhoff, F. & Helmchen, F. Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. Science 308, 1314-1318, doi:10.1126/science.1110647 (2005). 206 Nissl, F. Über einige Beziehungen zwischen Nervenzellerkrankungen und gliösen Erscheinungen bei verschiedenen Psychosen. Arch f Psychiatr 32, 656-676 (1899). 207 Nomura, K., Vilalta, A., Allendorf, D. H., Hornik, T. C. & Brown, G. C. Activated Microglia Desialylate and Phagocytose Cells via Neuraminidase, Galectin-3, and Mer Tyrosine Kinase. J Immunol 198, 4792-4801, doi:10.4049/jimmunol.1502532 (2017). 208 Nowacka, B., Lubinski, W., Honczarenko, K., Potemkowski, A. & Safranow, K. Ophthalmological features of Parkinson disease. Med Sci Monit 20, 2243-2249, doi:10.12659/msm.890861 (2014). 209 Noyce, A. J. et al. Meta-analysis of early nonmotor features and risk factors for Parkinson disease. Ann Neurol 72, 893-901, doi:10.1002/ana.23687 (2012). 210 Nuber, S. et al. Abrogating Native α-Synuclein Tetramers in Mice Causes a L-DOPA- Responsive Motor Syndrome Closely Resembling Parkinson's Disease. Neuron 100, 75- 90.e75, doi:10.1016/j.neuron.2018.09.014 (2018). 211 Nussbaum, R. L. & Polymeropoulos, M. H. Genetics of Parkinson's disease. Hum Mol Genet 6, 1687-1691, doi:10.1093/hmg/6.10.1687 (1997). 212 Okubadejo, N. U., Bower, J. H., Rocca, W. A. & Maraganore, D. M. Parkinson's disease in Africa: A systematic review of epidemiologic and genetic studies. Mov Disord 21, 2150- 2156, doi:10.1002/mds.21153 (2006). 213 Orihuela, R., McPherson, C. A. & Harry, G. J. Microglial M1/M2 polarization and metabolic states. Br J Pharmacol 173, 649-665, doi:10.1111/bph.13139 (2016). 214 Oropesa-Ruiz, J. M. et al. Low serum uric acid levels in progressive supranuclear palsy. Mov Disord 31, 402-405, doi:10.1002/mds.26466 (2016).
111 215 Ortega-Cubero, S. et al. TREM2 R47H variant and risk of essential tremor: a crosssectional international multicenter study. Parkinsonism Relat Disord 21, 306-309, doi:10.1016/j.parkreldis.2014.12.010 (2015). 216 Ostrerova, N. et al. alpha-Synuclein shares physical and functional homology with 14-3- 3 proteins. J Neurosci 19, 5782-5791 (1999). 217 Pahwa, R. & Lyons, K. E. Levodopa-related wearing-off in Parkinson's disease: identification and management. Curr Med Res Opin 25, 841-849, doi:10.1185/03007990902779319 (2009). 218 Paleologou, K. E. et al. Phosphorylation at Ser-129 but not the phosphomimics S129E/D inhibits the fibrillation of alpha-synuclein. J Biol Chem 283, 16895-16905, doi:10.1074/jbc.M800747200 (2008). 219 Pampalakis, G. et al. KLK6 proteolysis is implicated in the turnover and uptake of extracellular alpha-synuclein species. Oncotarget 8, 14502-14515, doi:10.18632/oncotarget.13264 (2017). 220 Panov, A. et al. Rotenone model of Parkinson disease: multiple brain mitochondria dysfunctions after short term systemic rotenone intoxication. J Biol Chem 280, 42026- 42035, doi:10.1074/jbc.M508628200 (2005). 221 Paolicelli, R. C. et al. Synaptic pruning by microglia is necessary for normal brain development. Science 333, 1456-1458, doi:10.1126/science.1202529 (2011). 222 Papaspyridonos, M. et al. Galectin-3 is an amplifier of inflammation in atherosclerotic plaque progression through macrophage activation and monocyte chemoattraction. Arterioscler Thromb Vasc Biol 28, 433-440, doi:10.1161/atvbaha.107.159160 (2008). 223 Parent, A. in Handbook of Behavioral Neuroscience Vol. 24 (eds Heinz Steiner & Kuei Y. Tseng) 33-44 (Elsevier, 2016). 224 Parent, A. Felix Vicq d'Azyr: anatomy, medicine and revolution. Can J Neurol Sci 34, 30- 37, doi:10.1017/s0317167100018722 (2007). 225 Parkhurst, C. N. et al. Microglia promote learning-dependent synapse formation through brain-derived neurotrophic factor. Cell 155, 1596-1609, doi:10.1016/j.cell.2013.11.030 (2013). 226 Parkinson, J. An essay on the shaking palsy. 1817. J Neuropsychiatry Clin Neurosci 14, 223-236; discussion 222, doi:10.1176/jnp.14.2.223 (1817). 227 Parkkinen, L., Pirttilä, T., Tervahauta, M. & Alafuzoff, I. Widespread and abundant alphasynuclein pathology in a neurologically unimpaired subject. Neuropathology 25, 304- 314, doi:10.1111/j.1440-1789.2005.00644.x (2005). 228 Partridge, E. A. et al. Regulation of cytokine receptors by Golgi N-glycan processing and endocytosis. Science 306, 120-124, doi:10.1126/science.1102109 (2004). 229 Pasquini, L. A. et al. Galectin-3 drives oligodendrocyte differentiation to control myelin integrity and function. Cell Death Differ 18, 1746-1756, doi:10.1038/cdd.2011.40 (2011). 230 Patterson, R. J., Haudek, K. C., Voss, P. G. & Wang, J. L. Examination of the role of galectins in pre-mRNA splicing. Methods Mol Biol 1207, 431-449, doi:10.1007/978-1- 4939-1396-1_28 (2015). 231 Paumier, K. L. et al. Behavioral characterization of A53T mice reveals early and late stage deficits related to Parkinson's disease. PLoS One 8, e70274, doi:10.1371/journal.pone.0070274 (2013). 232 Peelaerts, W. et al. α-Synuclein strains cause distinct synucleinopathies after local and systemic administration. Nature 522, 340-344, doi:10.1038/nature14547 (2015). 233 Peng, X. et al. Alpha-synuclein activation of protein phosphatase 2A reduces tyrosine hydroxylase phosphorylation in dopaminergic cells. J Cell Sci 118, 3523-3530, doi:10.1242/jcs.02481 (2005). 234 Pfeiffer, R. F. Non-motor symptoms in Parkinson's disease. Parkinsonism Relat Disord 22 Suppl 1, S119-122, doi:10.1016/j.parkreldis.2015.09.004 (2016).
112 235 Phan, J. A. et al. Early synaptic dysfunction induced by α-synuclein in a rat model of Parkinson's disease. Sci Rep 7, 6363, doi:10.1038/s41598-017-06724-9 (2017). 236 Pieri, L., Madiona, K. & Melki, R. Structural and functional properties of prefibrillar αsynuclein oligomers. Sci Rep 6, 24526, doi:10.1038/srep24526 (2016). 237 Pissadaki, E. K. & Bolam, J. P. The energy cost of action potential propagation in dopamine neurons: clues to susceptibility in Parkinson's disease. Front Comput Neurosci 7, 13, doi:10.3389/fncom.2013.00013 (2013). 238 Pocock, J. M. & Kettenmann, H. Neurotransmitter receptors on microglia. Trends Neurosci 30, 527-535, doi:10.1016/j.tins.2007.07.007 (2007). 239 Poewe, W. et al. Parkinson disease. Nat Rev Dis Primers 3, 17013, doi:10.1038/nrdp.2017.13 (2017). 240 Polymeropoulos, M. H. et al. Mutation in the alpha-synuclein gene identified in families with Parkinson's disease. Science 276, 2045-2047, doi:10.1126/science.276.5321.2045 (1997). 241 Prajapati, P., Sripada, L., Singh, K., Bhatelia, K. & Singh, R. TNF-α regulates miRNA targeting mitochondrial complex-I and induces cell death in dopaminergic cells. Biochim Biophys Acta 1852, 451-461, doi:10.1016/j.bbadis.2014.11.019 (2015). 242 Qin, L. et al. Systemic LPS causes chronic neuroinflammation and progressive neurodegeneration. Glia 55, 453-462, doi:10.1002/glia.20467 (2007). 243 Qualman, S. J., Haupt, H. M., Yang, P. & Hamilton, S. R. Esophageal Lewy bodies associated with ganglion cell loss in achalasia. Similarity to Parkinson's disease. Gastroenterology 87, 848-856 (1984). 244 Rabinovich, G. A. & Toscano, M. A. Turning 'sweet' on immunity: galectin-glycan interactions in immune tolerance and inflammation. Nat Rev Immunol 9, 338-352, doi:10.1038/nri2536 (2009). 245 Rahimian, R. et al. Delayed Galectin-3-Mediated Reprogramming of Microglia After Stroke is Protective. Mol Neurobiol 56, 6371-6385, doi:10.1007/s12035-019-1527-0 (2019). 246 Rakers, C. et al. Stroke target identification guided by astrocyte transcriptome analysis. Glia 67, 619-633, doi:10.1002/glia.23544 (2019). 247 Ramírez, E. et al. Neuroinflammation induced by the peptide amyloid-β (25-35) increase the presence of galectin-3 in astrocytes and microglia and impairs spatial memory. Neuropeptides 74, 11-23, doi:10.1016/j.npep.2019.02.001 (2019). 248 Ramon y Cajal, S. Histologie du système nerveux de l'homme et des vertébrés. Maloine, Paris 2, 153-173 (1911). 249 Ramon y Cajal, S. Textura del sistema nervioso del hombre y des los vertebrados. Histologie du systeme nerveux de l’homme et des vertébrés 2, 1909-1911 (1899). 250 Ransohoff, R. M. How neuroinflammation contributes to neurodegeneration. Science 353, 777-783, doi:10.1126/science.aag2590 (2016). 251 Redell, J. B. & Dash, P. K. Traumatic brain injury stimulates hippocampal catechol-O- methyl transferase expression in microglia. Neurosci Lett 413, 36-41, doi:10.1016/j.neulet.2006.11.060 (2007). 252 Reichert, F. & Rotshenker, S. Galectin-3 (MAC-2) Controls Microglia Phenotype Whether Amoeboid and Phagocytic or Branched and Non-phagocytic by Regulating the Cytoskeleton. Front Cell Neurosci 13, 90, doi:10.3389/fncel.2019.00090 (2019). 253 Rekas, A., Ahn, K. J., Kim, J. & Carver, J. A. The chaperone activity of α-synuclein: Utilizing deletion mutants to map its interaction with target proteins. Proteins 80, 1316-1325, doi:10.1002/prot.24028 (2012). 254 Ren, M. et al. TREM2 overexpression attenuates neuroinflammation and protects dopaminergic neurons in experimental models of Parkinson's disease. Exp Neurol 302, 205-213, doi:10.1016/j.expneurol.2018.01.016 (2018).
113 255 Rigato, C., Buckinx, R., Le-Corronc, H., Rigo, J. M. & Legendre, P. Pattern of invasion of the embryonic mouse spinal cord by microglial cells at the time of the onset of functional neuronal networks. Glia 59, 675-695, doi:10.1002/glia.21140 (2011). 256 Ruf, V. C. et al. Different Effects of α-Synuclein Mutants on Lipid Binding and Aggregation Detected by Single Molecule Fluorescence Spectroscopy and ThT Fluorescence-Based Measurements. ACS Chem Neurosci 10, 1649-1659, doi:10.1021/acschemneuro.8b00579 (2019). 257 Salat, D. & Tolosa, E. Levodopa in the treatment of Parkinson's disease: current status and new developments. J Parkinsons Dis 3, 255-269, doi:10.3233/jpd-130186 (2013). 258 Sanchez-Guajardo, V., Tentillier, N. & Romero-Ramos, M. The relation between αsynuclein and microglia in Parkinson's disease: Recent developments. Neuroscience 302, 47-58, doi:10.1016/j.neuroscience.2015.02.008 (2015). 259 Sano, I. et al. Distribution of catechol compounds in human brain. Biochim Biophys Acta 32, 586-587, doi:10.1016/0006-3002(59)90652-3 (1959). 260 Scott, D. & Roy, S. α-Synuclein inhibits intersynaptic vesicle mobility and maintains recycling-pool homeostasis. J Neurosci 32, 10129-10135, doi:10.1523/jneurosci.0535- 12.2012 (2012). 261 Schapira, A. H. V., Chaudhuri, K. R. & Jenner, P. Non-motor features of Parkinson disease. Nat Rev Neurosci 18, 435-450, doi:10.1038/nrn.2017.62 (2017). 262 Sedelis, M., Schwarting, R. K. & Huston, J. P. Behavioral phenotyping of the MPTP mouse model of Parkinson's disease. Behav Brain Res 125, 109-125, doi:10.1016/s0166- 4328(01)00309-6 (2001). 263 Seppi, K. et al. The Movement Disorder Society Evidence-Based Medicine Review Update: Treatments for the non-motor symptoms of Parkinson's disease. Mov Disord 26 Suppl 3, S42-80, doi:10.1002/mds.23884 (2011). 264 Shannon, K. M., Keshavarzian, A., Dodiya, H. B., Jakate, S. & Kordower, J. H. Is alphasynuclein in the colon a biomarker for premotor Parkinson's disease? Evidence from 3 cases. Mov Disord 27, 716-719, doi:10.1002/mds.25020 (2012). 265 Shen, Y. F. et al. The change of plasma galectin-3 concentrations after traumatic brain injury. Clin Chim Acta 456, 75-80, doi:10.1016/j.cca.2016.02.029 (2016). 266 Shimura, T. et al. Galectin-3, a novel binding partner of beta-catenin. Cancer Res 64, 6363-6367, doi:10.1158/0008-5472.can-04-1816 (2004). 267 Sierra, A. et al. The "Big-Bang" for modern glial biology: Translation and comments on Pío del Río-Hortega 1919 series of papers on microglia. Glia 64, 1801-1840, doi:10.1002/glia.23046 (2016). 268 Siew, J. J. et al. Galectin-3 is required for the microglia-mediated brain inflammation in a model of Huntington's disease. Nat Commun 10, 3473, doi:10.1038/s41467-019- 11441-0 (2019). 269 Silva-Monteiro, E. et al. Altered expression of galectin-3 induces cortical thymocyte depletion and premature exit of immature thymocytes during Trypanosoma cruzi infection. Am J Pathol 170, 546-556, doi:10.2353/ajpath.2007.060389 (2007). 270 Simovic Markovic, B. et al. Galectin-3 Plays an Important Pro-inflammatory Role in the Induction Phase of Acute Colitis by Promoting Activation of NLRP3 Inflammasome and Production of IL-1β in Macrophages. J Crohns Colitis 10, 593-606, doi:10.1093/eccojcc/jjw013 (2016). 271 Sirko, S. et al. Astrocyte reactivity after brain injury-: The role of galectins 1 and 3. Glia 63, 2340-2361, doi:10.1002/glia.22898 (2015). 272 Spillantini, M. G., Crowther, R. A., Jakes, R., Hasegawa, M. & Goedert, M. alpha- Synuclein in filamentous inclusions of Lewy bodies from Parkinson's disease and dementia with lewy bodies. Proc Natl Acad Sci U S A 95, 6469-6473, doi:10.1073/pnas.95.11.6469 (1998).
114 273 Spillantini, M. G. & Goedert, M. Neurodegeneration and the ordered assembly of αsynuclein. Cell Tissue Res 373, 137-148, doi:10.1007/s00441-017-2706-9 (2018). 274 Spillantini, M. G. et al. Alpha-synuclein in Lewy bodies. Nature 388, 839-840, doi:10.1038/42166 (1997). 275 Squarzoni, P. et al. Microglia modulate wiring of the embryonic forebrain. Cell Rep 8, 1271-1279, doi:10.1016/j.celrep.2014.07.042 (2014). 276 St Martin, J. L. et al. Dopaminergic neuron loss and up-regulation of chaperone protein mRNA induced by targeted over-expression of alpha-synuclein in mouse substantia nigra. J Neurochem 100, 1449-1457, doi:10.1111/j.1471-4159.2006.04310.x (2007). 277 Stevens, B. et al. The classical complement cascade mediates CNS synapse elimination. Cell 131, 1164-1178, doi:10.1016/j.cell.2007.10.036 (2007). 278 Stocchi, F., Jenner, P. & Obeso, J. A. When do levodopa motor fluctuations first appear in Parkinson's disease? Eur Neurol 63, 257-266, doi:10.1159/000300647 (2010). 279 Stratoulias, V., Venero, J. L., Tremblay, M. & Joseph, B. Microglial subtypes: diversity within the microglial community. EMBO J 38, e101997, doi:10.15252/embj.2019101997 (2019). 280 Strauss, I., Kalia, S. K. & Lozano, A. M. Where are we with surgical therapies for Parkinson's disease? Parkinsonism Relat Disord 20 Suppl 1, S187-191, doi:10.1016/s1353-8020(13)70044-0 (2014). 281 Suarez, G. & Meyerrose, G. Heart failure and galectin 3. Ann Transl Med 2, 86, doi:10.3978/j.issn.2305-5839.2014.09.10 (2014). 282 Sulzer, D. & Edwards, R. H. The physiological role of α-synuclein and its relationship to Parkinson's Disease. J Neurochem 150, 475-486, doi:10.1111/jnc.14810 (2019). 283 Sun, J. et al. Functional cooperation of α-synuclein and VAMP2 in synaptic vesicle recycling. Proc Natl Acad Sci U S A 116, 11113-11115, doi:10.1073/pnas.1903049116 (2019). 284 Surmeier, D. J., Obeso, J. A. & Halliday, G. M. Selective neuronal vulnerability in Parkinson disease. Nat Rev Neurosci 18, 101-113, doi:10.1038/nrn.2016.178 (2017). 285 Szepesi, Z., Manouchehrian, O., Bachiller, S. & Deierborg, T. Bidirectional Microglia- Neuron Communication in Health and Disease. Front Cell Neurosci 12, 323, doi:10.3389/fncel.2018.00323 (2018). 286 Tang, Y. & Le, W. Differential Roles of M1 and M2 Microglia in Neurodegenerative Diseases. Mol Neurobiol 53, 1181-1194, doi:10.1007/s12035-014-9070-5 (2016). 287 Tanik, S. A., Schultheiss, C. E., Volpicelli-Daley, L. A., Brunden, K. R. & Lee, V. M. Lewy body-like α-synuclein aggregates resist degradation and impair macroautophagy. J Biol Chem 288, 15194-15210, doi:10.1074/jbc.M113.457408 (2013). 288 Tansey, M. G. & Goldberg, M. S. Neuroinflammation in Parkinson's disease: its role in neuronal death and implications for therapeutic intervention. Neurobiol Dis 37, 510- 518, doi:10.1016/j.nbd.2009.11.004 (2010). 289 Tao-Cheng, J. H. Activity-related redistribution of presynaptic proteins at the active zone. Neuroscience 141, 1217-1224, doi:10.1016/j.neuroscience.2006.04.061 (2006). 290 Tatebe, H. et al. Extracellular neurosin degrades α-synuclein in cultured cells. Neurosci Res 67, 341-346, doi:10.1016/j.neures.2010.04.008 (2010). 291 Taylor, T. N., Greene, J. G. & Miller, G. W. Behavioral phenotyping of mouse models of Parkinson's disease. Behav Brain Res 211, 1-10, doi:10.1016/j.bbr.2010.03.004 (2010). 292 Tekriwal, A. et al. REM sleep behaviour disorder: prodromal and mechanistic insights for Parkinson's disease. J Neurol Neurosurg Psychiatry 88, 445-451, doi:10.1136/jnnp-2016- 314471 (2017). 293 Thaisetthawatkul, P. Pure Autonomic Failure. Curr Neurol Neurosci Rep 16, 74, doi:10.1007/s11910-016-0673-2 (2016).
115 294 Théry, C. et al. Proteomic analysis of dendritic cell-derived exosomes: a secreted subcellular compartment distinct from apoptotic vesicles. J Immunol 166, 7309-7318, doi:10.4049/jimmunol.166.12.7309 (2001). 295 Thomas, L. & Pasquini, L. A. Extracellular Galectin-3 Induces Accelerated Oligodendroglial Differentiation Through Changes in Signaling Pathways and Cytoskeleton Dynamics. Mol Neurobiol 56, 336-349, doi:10.1007/s12035-018-1089-6 (2019). 296 Tian, J. et al. Galectin-3 regulates inflammasome activation in cholestatic liver injury. FASEB J 30, 4202-4213, doi:10.1096/fj.201600392RR (2016). 297 Tolosa, E., Wenning, G. & Poewe, W. The diagnosis of Parkinson's disease. Lancet Neurol 5, 75-86, doi:10.1016/s1474-4422(05)70285-4 (2006). 298 Tomás-Camardiel, M. et al. Minocycline reduces the lipopolysaccharide-induced inflammatory reaction, peroxynitrite-mediated nitration of proteins, disruption of the blood-brain barrier, and damage in the nigral dopaminergic system. Neurobiol Dis 16, 190-201, doi:10.1016/j.nbd.2004.01.010 (2004). 299 Tremblay, M., Lecours, C., Samson, L., Sánchez-Zafra, V. & Sierra, A. From the Cajal alumni Achúcarro and Río-Hortega to the rediscovery of never-resting microglia. Front Neuroanat 9, 45, doi:10.3389/fnana.2015.00045 (2015). 300 Tsujimura, A. et al. Lysosomal enzyme cathepsin B enhances the aggregate forming activity of exogenous α-synuclein fibrils. Neurobiol Dis 73, 244-253, doi:10.1016/j.nbd.2014.10.011 (2015). 301 Tysnes, O. B. & Storstein, A. Epidemiology of Parkinson's disease. J Neural Transm (Vienna) 124, 901-905, doi:10.1007/s00702-017-1686-y (2017). 302 Uéda, K. et al. Molecular cloning of cDNA encoding an unrecognized component of amyloid in Alzheimer disease. Proc Natl Acad Sci U S A 90, 11282-11286, doi:10.1073/pnas.90.23.11282 (1993). 303 Van Den Eeden, S. K. et al. Incidence of Parkinson's disease: variation by age, gender, and race/ethnicity. Am J Epidemiol 157, 1015-1022, doi:10.1093/aje/kwg068 (2003). 304 van Gijn, J. Félix Vicq d'Azyr (1748-1794). J Neurol 256, 1384-1385, doi:10.1007/s00415- 009-5211-6 (2009). 305 van Kooyk, Y. & Rabinovich, G. A. Protein-glycan interactions in the control of innate and adaptive immune responses. Nat Immunol 9, 593-601, doi:10.1038/ni.f.203 (2008). 306 Vargas, K. J. et al. Synucleins Have Multiple Effects on Presynaptic Architecture. Cell Rep 18, 161-173, doi:10.1016/j.celrep.2016.12.023 (2017). 307 Vicente Miranda, H. & Outeiro, T. F. The sour side of neurodegenerative disorders: the effects of protein glycation. J Pathol 221, 13-25, doi:10.1002/path.2682 (2010). 308 Vicente Miranda, H. et al. Glycation potentiates α-synuclein-associated neurodegeneration in synucleinopathies. Brain 140, 1399-1419, doi:10.1093/brain/awx056 (2017). 309 Vladoiu, M. C., Labrie, M. & St-Pierre, Y. Intracellular galectins in cancer cells: potential new targets for therapy (Review). Int J Oncol 44, 1001-1014, doi:10.3892/ijo.2014.2267 (2014). 310 Volpicelli-Daley, L. A. et al. Formation of α-synuclein Lewy neurite-like aggregates in axons impedes the transport of distinct endosomes. Mol Biol Cell 25, 4010-4023, doi:10.1091/mbc.E14-02-0741 (2014a). 311 Volpicelli-Daley, L. A., Luk, K. C. & Lee, V. M. Addition of exogenous α-synuclein preformed fibrils to primary neuronal cultures to seed recruitment of endogenous αsynuclein to Lewy body and Lewy neurite-like aggregates. Nat Protoc 9, 2135-2146, doi:10.1038/nprot.2014.143 (2014b). 312 von Campenhausen, S. et al. Prevalence and incidence of Parkinson's disease in Europe. Eur Neuropsychopharmacol 15, 473-490, doi:10.1016/j.euroneuro.2005.04.007 (2005).
116 313 von Herrmann, K. M. et al. expression in mesencephalic neurons and characterization of a rare. NPJ Parkinsons Dis 4, 24, doi:10.1038/s41531-018-0061-5 (2018). 314 Wakabayashi, K., Tanji, K., Mori, F. & Takahashi, H. The Lewy body in Parkinson's disease: molecules implicated in the formation and degradation of alpha-synuclein aggregates. Neuropathology 27, 494-506 (2007). 315 Wakabayashi, K., Yoshimoto, M., Tsuji, S. & Takahashi, H. Alpha-synuclein immunoreactivity in glial cytoplasmic inclusions in multiple system atrophy. Neurosci Lett 249, 180-182, doi:10.1016/s0304-3940(98)00407-8 (1998). 316 Wake, H., Moorhouse, A. J., Jinno, S., Kohsaka, S. & Nabekura, J. Resting microglia directly monitor the functional state of synapses in vivo and determine the fate of ischemic terminals. J Neurosci 29, 3974-3980, doi:10.1523/jneurosci.4363-08.2009 (2009). 317 Walton, R. L. et al. TREM2 p.R47H substitution is not associated with dementia with Lewy bodies. Neurol Genet 2, e85, doi:10.1212/nxg.0000000000000085 (2016). 318 Wang, Y. et al. TREM2 lipid sensing sustains the microglial response in an Alzheimer's disease model. Cell 160, 1061-1071, doi:10.1016/j.cell.2015.01.049 (2015). 319 Wani, W. Y. et al. O-GlcNAc regulation of autophagy and α-synuclein homeostasis; implications for Parkinson's disease. Mol Brain 10, 32, doi:10.1186/s13041-017-0311-1 (2017). 320 Weinreb, P. H., Zhen, W., Poon, A. W., Conway, K. A. & Lansbury, P. T. NACP, a protein implicated in Alzheimer's disease and learning, is natively unfolded. Biochemistry 35, 13709-13715, doi:10.1021/bi961799n (1996). 321 Weng, I. C. et al. Cytosolic galectin-3 and -8 regulate antibacterial autophagy through differential recognition of host glycans on damaged phagosomes. Glycobiology 28, 392- 405, doi:10.1093/glycob/cwy017 (2018). 322 Wesley, U. V., Vemuganti, R., Ayvaci, E. R. & Dempsey, R. J. Galectin-3 enhances angiogenic and migratory potential of microglial cells via modulation of integrin linked kinase signaling. Brain Res 1496, 1-9, doi:10.1016/j.brainres.2012.12.008 (2013). 323 Wolff, B. S., Raheem, S. A. & Saligan, L. N. Comparing passive measures of fatigue-like behavior in mice. Sci Rep 8, 14238, doi:10.1038/s41598-018-32654-1 (2018). 324 Xia, R. & Mao, Z. H. Progression of motor symptoms in Parkinson's disease. Neurosci Bull 28, 39-48, doi:10.1007/s12264-012-1050-z (2012). 325 Xilouri, M., Brekk, O. R. & Stefanis, L. Autophagy and Alpha-Synuclein: Relevance to Parkinson's Disease and Related Synucleopathies. Mov Disord 31, 178-192, doi:10.1002/mds.26477 (2016). 326 Yan, J. et al. Increased Expressions of Plasma Galectin-3 in Patients with Amyotrophic Lateral Sclerosis. Chin Med J (Engl) 129, 2797-2803, doi:10.4103/0366-6999.194656 (2016). 327 Yan, Y. P., Lang, B. T., Vemuganti, R. & Dempsey, R. J. Galectin-3 mediates post-ischemic tissue remodeling. Brain Res 1288, 116-124, doi:10.1016/j.brainres.2009.06.073 (2009). 328 Yang, H. et al. Lipoprotein lipase deficiency leads to α-synuclein aggregation and ubiquitin C-terminal hydrolase L1 reduction. Neuroscience 290, 1-10, doi:10.1016/j.neuroscience.2014.12.068 (2015). 329 Yang, J. W. et al. Mass spectrometrical analysis of galectin proteins in primary rat cerebellar astrocytes. Neurochem Res 31, 945-955, doi:10.1007/s11064-006-9100-4 (2006). 330 Yang, R. Y., Hsu, D. K. & Liu, F. T. Expression of galectin-3 modulates T-cell growth and apoptosis. Proc Natl Acad Sci U S A 93, 6737-6742, doi:10.1073/pnas.93.13.6737 (1996). 331 Yang, X. et al. Depletion of microglia augments the dopaminergic neurotoxicity of MPTP. FASEB J 32, 3336-3345, doi:10.1096/fj.201700833RR (2018).