Metformin treatment decreases microglial activation in the in vivo lps model of parkinson's disease.

! I! UNIVERSIDAD DE SEVILLA DEPARTAMENTO DE BIOQUÍMICA Y BIOLOGIA MOLECULAR FACULTAD DE FARMACIA METFORMIN TREATMENT DECREASES MICROGLIAL ACTIVATION IN THE IN VIVO LPS MODEL OF PARKINSON’S DISEASE KHADIJA TAYARA PhD THESIS, 2018. ! II! UNIVERSIDAD DE SEVILLA Departamento de Bioquímica y Biología Molecular Facultad de Farmacia ! !!!!!!!!!!!!!!!!!!!!C!/!Profesor!García!González,!2.! !!!!!!!!!!!!!!!!!!!!41012!Sevilla!(España)! !!!!!!!!!!!!!!!!!!!!Teléfono:!954!55!67!55! (Fax:(954(55(68(54( ! Dr.!José!Luis!Venero!Recio,!Catedrático!de!Universidad!y!Director!del!Departamento!de! Bioquímica!y!Biología!Molecular!de!la!Universidad!de!Sevilla.! INFORMA:) Que!D.a!!Khadija!Tayara!ha!realizada!el!trabajo!experimental!que!ha!llevado!a!la! redacción!de!la!presente!memoria!de!Tesis!Doctoral!bajo!la!dirección!del!Dra.!Rocio! Martinez!De!Pablos!y!el!Dra.!Ana!Maria!Espinosa!Oliva.! Para!que!así!conste!y!surta!los!efectos!oportunos,!firmo!el!presente!informe!en! Sevilla!a!26!de!Junio!de!2018.! ! ) ) Fdo.)Dr.)José)Luis)Venero)Recio) ! III! UNIVERSIDAD DE SEVILLA Departamento de Bioquímica y Biología Molecular Facultad de Farmacia ! !!!!!!!!!!!!!!!!!!!!C!/!Profesor!García!González,!2.! !!!!!!!!!!!!!!!!!!!!41012!Sevilla!(España)! !!!!!!!!!!!!!!!!!!!!Teléfono:!954!55!67!55! (Fax:(954(55(68(54( La Dra Rocio Martinez De Pablos, profesora Contratada Doctora en la Universidad de Sevilla, y la Dra. Ana Maria Espinosa Oliva, profesora Contratada Doctora interina en la Universidad de Sevilla INFORMAN: Que! D.a!Khadija! Tayara! ha! realizado! bajo! nuestra! dirección! el! trabajo! experimental!que!ha!llevado!a!la!redacción!de!la!presente!memoria!de!Tesis!Doctoral! titulada! METFORMIN! TREATMENT! DECREASES! MICROGLIAL! ACTIVATION! IN! THE! IN( VIVO!LPS!MODEL!OF!PARKINSON´S!DISEASE"! Considerando!que!constituye!el!trabajo!de!tesis!doctoral,!autorizamos!su!presentación! para!optar!al!grado!de!Doctor.! Para!que!así!conste!y!surta!los!efectos!oportunos,!firmamos!el!presente!informe! en!Sevilla!a!26!de!Junio!de!2018.! ! Fdo.)Dra.)Rocio)Martinez)de)Pablos)) ) Fdo.)Dra.)Ana)Maria)Espinosa)Oliva) ! IV! Part of the results presented in this thesis was submitted to the Frontiers Journal for peer-review. ! ! 1! Contents ABBREVIATIONS ............................................................................................................................. 4 ABSTRACT ....................................................................................................................................... 11 RESUMEN ......................................................................................................................................... 12 I. INTRODUCTION ...................................................................................................................... 13 1. BASAL GANGLIA ................................................................................................................ 14 1.1. History of basal ganglia discovery .................................................................................. 14 1.2. Anatomy of the basal ganglia .......................................................................................... 15 1.3. Basal ganglia neurotransmisors and neuromodulators .................................................... 17 1.5. Functions of basal ganglia ................................................................................................... 24 2. CLINICAL SYNDROMES RESULTING FROM DAMAGE TO BASAL GANGLIA ....... 26 2.1. Parkinson’s disease .......................................................................................................... 26 2.2. Pathology of Parkinson’s disease .................................................................................... 30 2.3. Neuroinflammation on Parkinson’s disease .................................................................... 31 3. NEURODEGENERATION THEORIES ............................................................................... 35 4. ANIMAL MODELS ............................................................................................................... 38 4.1. Transgenic animals .......................................................................................................... 39 4.2. Models based on neurotoxin injection ............................................................................. 40 2.1.1. Injection of 6-OHDA ................................................................................................ 41 2.1.2. Injection of MPTP .................................................................................................... 41 2.1.3. Injection of pro-inflammatory compounds ............................................................... 43 5. LIPOPOLISACCHARIDE ..................................................................................................... 44 5.1. Interaction between LPS and soluble membrane proteins ............................................... 45 5.2. Transduction of signal induced by LPS ........................................................................... 48 5.3. Mitogen activated protein kinase ..................................................................................... 51 5.4. LPS as an inductor of the inflammasome components expression .................................. 52 ! 2! 6. THE LPS AS A MODEL OF THE PD ................................................................................... 53 7. METFORMIN ......................................................................................................................... 58 7.1. Diabetes mellitus ............................................................................................................. 58 7.2. Neurodegeneration and diabetes mellitus interrelation ................................................... 59 7.3. Treatment of diabetes mellitus ......................................................................................... 60 7.4. Metformin and its history ................................................................................................ 61 7.5. Metformin mechanism of action ...................................................................................... 63 7.6. Anti-inflammatory effects of metformin ......................................................................... 67 II. OBJETIVES ............................................................................................................................... 73 III. MATERIALS AND METHODS ............................................................................................ 76 1. MATERIALS .......................................................................................................................... 77 2. METHODS ............................................................................................................................. 80 1. Cell culture .......................................................................................................................... 80 2.1.1. Thawing cells ........................................................................................................... 80 2.1.2. Passing cells ............................................................................................................. 80 2.1.3. Conservation of cells ................................................................................................ 81 2. Animals and surgery ............................................................................................................ 81 3. Treatment ............................................................................................................................. 82 2.8.1. In vitro treatment ...................................................................................................... 82 2.8.1. In vivo treatment ....................................................................................................... 82 4. Immunohistological evaluation: Tyrosine hydroxylase (TH), glial fibrillary acidic protein (GFAP), and OX-6 ...................................................................................................................... 85 5. Immunohistochemistry data analysis .................................................................................. 86 6. Immunofluorescence ........................................................................................................... 87 7. Real-time RT-PCR .............................................................................................................. 87 8. Western blot ........................................................................................................................ 91 2.8.1. Protein extraction from BV2 cells ............................................................................ 91 ! 3! 2.8.1. Protein extraction from SN ....................................................................................... 91 2.1.1. Western blot analysis ....................................................................................................... 91 9. Production of reactive oxygen species ................................................................................ 92 10. Enzyme-linked immunosorbent assay ............................................................................. 93 11. Statistical analysis ............................................................................................................ 93 IV. RESULTS ............................................................................................................................... 94 1. IN-VITRO EXPERIMENTS ................................................................................................... 95 1.1. Metformin treatment reduces microglial activation in vitro ............................................ 95 1.2. Metformin treatment prevented the induction of NADPH oxidase by LPS .................... 98 1.3. Metformin treatment decreases the production of ROS ................................................ 100 1.4. Metformin treatment decreases the activation of the inflammasome ............................ 102 2. In-vivo experiments ............................................................................................................... 104 2.1. Metformin treatment decreased the number of activated microglial cells in vivo ......... 104 2.2. Metformin treatment does not protect from the LPS-induced damage to the astroglial population ................................................................................................................................. 105 2.3. LPS-activated microglial cells express a M1 phenotype that is reduced by metformin 107 2.4. Metformin treatment decreases the expression levels of several inflammatory mediators after LPS injection in vivo ........................................................................................................ 110 2.5. Metformin effects on microglial activation are mediated by MAPK ............................ 112 2.6. Effect of metformin and LPS on the dopaminergic system ........................................... 114 V. DISCUSSION ........................................................................................................................... 117 VI. CONCLUSION ..................................................................................................................... 132 VII. REFERENCES ...................................................................................................................... 134 ! ! ! ! ! ! 4! ! ! ABBREVIATIONS ! 5! 6-OHDA Ach AGE ALS AMPK APP ATP Aβ BBB Calmodulin CB CCK cDNA CNS COX-2 CR CSF cSNc CTCF CXCR1 DA DAergic DAGs DBS DM DMEM dmSNc 6-hydroxydopamine Acetylcholine Glycosylation end product Amyotrophic lateral sclerosis AMP-activated protein kinase Amyloid precursor protein Adenosine triphosphate Amyloid beta Blood brain barrier Calcium modulated protein Calabinin Cholecystokinin Complementary DNA Central nervous system cyclooxygenase 2 Calretinin Cerebrospinal fluid Caudal substantia nigra pars compacta Corrected total cell fluorescence CX3C chemokine receptor 1 Dopamine Dopaminergic Diacylglycerols Deep brain stimulation Diabetes mellitus Dulbecco's Modified Eagle's médium Dorsomedial substantia nigra pars compacta ! 12! RESUMEN La metformina es un fármaco antidiabético oral ampliamente utilizado con propiedades antiinflamatorias conocidas debido a su acción sobre la AMP cinasa. Esta droga ha mostrado un efecto protector en varios tejidos del cerebro, incluidas las neuronas corticales. Sin embargo, el efecto de la metformina sobre la sustancia negra (SN), la principal estructura afectada en la enfermedad de Parkinson (EP), aún no se ha estudiado. Estudios recientes sugieren que la inflamación puede desempeñar un papel central en la pérdida celular que ocurre en la EP. El objetivo de este trabajo fue determinar el efecto de la metformina sobre la activación microglial de la SN de ratas utilizando el modelo animal de EP basado en la inyección del pro-inflamamógeno lipopolisacárido (LPS). Se usaron experimentos in vivo e in vitro para estudiar la activación de la microglía tanto a nivel celular como molecular. Nuestros resultados indican que la metformina inhibe de manera general la activación de la microglía medida a través de inmunocitoquímica contra OX-6 (un marcador del complejo mayor de histocompatibilidad tipo II), IKKβ (un marcador proinflamatorio) y la arginase (un marcador antiinflamatorio). Además, los experimentos de qPCR revelan que el tratamiento con metformina minimiza los niveles de expresión de varias citocinas pro y antiinflamatorias. Además, este fármaco disminuye las formas fosforiladas de MAPKs así como la generación de ROS a través de la inhibición de la enzima NADPH oxidasa. Sin embargo, el tratamiento con metformina no protege las neuronas dopaminérgicas de la SN. Estos hallazgos sugieren que la metformina podría tener efectos farmacológicos beneficiosos y perjudiciales y plantea la pregunta sobre el posible uso de la metformina para la prevención y el tratamiento de la EP. ! 13! I. INTRODUCTION ! 14! 1. BASAL GANGLIA 1.1. History of basal ganglia discovery In 1664, the English anatomist Thomas Willis was the first to provide a clear identification of the different subcortical structures. What is now known functionally as basal ganglia was then called the corpus striatum. It occupied a central position, connected with a wide variety of cortical fibers and to the brainstem, which at that time was believed to be the "sensorium commune" defined by Aristotle. It was thought that this structure received all the sensorial modalities and initiated all the motor acts. This idea appeared to be anatomically reinforced by its ventral position and by clearly visible ascending and descending fiber systems (Parent and Hazrati 1995). One of the first descriptions of the basal ganglia by Willis in the 17th century intimated a role of the basal ganglia in the control of movement and in neurological disorders: ‘‘To the corpus callosum are attached the corpora striata connecting the cerebrum to the legs of the medulla oblongata. In these corpora there are some striae passing upwards and others downwards and through them the spirits and images of sensible things pass from the medulla oblongata into the cerebrum, while spirits initiating movement descend into the medulla oblongata. In those who suffer or have died from paralysis, I have often observed that these corpora are affected: they became flaccid and their striae are almost obliterated’’ (Smith, et al. 1998) Two facts contributed to relegate the corpus striatum to a dark and less defined position: the attractive histological organization of the cortex, and the possibility of locating higher mental functions in the cortex, which led many neurologists of the 18th and 20th centuries to study the issue. Those who continued to study the corpus striatum suddenly discovered that the neighboring corticospinal pathways, in fact owned many of the functions originally assigned to this structure. As Wilson described in 1914, "the corpus striatum has been depreciated in its physiological significance." At the beginning of the 20th century there were serious attempts to provide a comparative and detailed description of the corpus striatum (Ramón y Cajal, 1911, Wilson, 1914). ! 15! It began to gain importance again with the discovery that lesions of these areas in humans often resulted in disorders of motor functions. The corpus striatum became the main component of the extrapyramidal motor system. This term, in a somewhat indefinite form, groups the corpus striatum with a series of nuclei of the brainstem, and reflects the presumption that this grouping constituted a complete and independent motor unit. Nowadays, the basal ganglia are defined as a group of subcortical telencephalic nuclei spanning the telencephalon, diencephalon, and midbrain at the base of the forebrain that are highly connected with the cortex. 1.2. Anatomy of the basal ganglia The four principal nuclei of the basal ganglia are (Fig. 1): (1) The striatum. The striatum consists of three important subdivisions: the caudate nucleus, the putamen, and the ventral striatum. Except at its most anterior pole, the striatum, where it is known as the neostriatum, is divided into the caudate nucleus and putamen by the internal capsule, a major collection of fibers found between the neocortex and the thalamus in both directions (Gerfen 1992). (2) The globus pallidus (GP). The GP is divided into two segments, the external (GPe) and the internal (GPi); in the rat the latter is called the entopeduncular nucleus. (3) The subthalamic nucleus (STN). The STN of Luys lies inferiorly to the thalamus at the junction of the diencephalon and the mesencephalon (midbrain). (4) The substantia nigra (SN). Although the SN is presenting a more caudal situation, it is included within the structures of the basal ganglia because of its intimate anatomical and functional relationship. It is a mesencephalic nucleus, which appears as a flattened oval structure on the dorsal aspect of the base of the cerebral peduncle. SN lies dorsal to the crus cerebri, a nervous tract that lies between the brain and the medulla oblongata, and extends to the midbrain; it plays a central role in movement control, and acts as the primary output and entrance center of the basal ganglia (Fink ! 16! and Smith 1980; Parent, et al. 1983). SN has been divided cytoarchitectonically into three different parts: 1) SN pars compacta (SNc), formed by a horizontal layer packed with long and medium cells. This part of SN uses dopamine (DA) as neurotransmitter. SN derives its name from the presence of the black pigment neuromelanin, contained in the neurons of the SNc. 2) SN pars reticulate (SNr), a more diffuse and poor cell division containing large neurons that project outside of the basal ganglia (Gulley and Wood 1971; Mink 1996). 3) SN lateralis (SNl), formed by a small group of medial cells extending face-caudally along the lateral border of SNc and SNr (Danner and Pfister 1982; Fallon and Seroogy 1985; McRitchie, et al. 1998; McRitchie, et al. 1995; Yelnik, et al. 1987). Figure 1. Basal ganglia main regions within the brain. The blue part is the striatum, which consists of the caudate nucleus and the putamen. The pink part is the globus pallidus. The green part is the subthalamic nucleus and the yellow part is the substantia nigra (Graybiel 2000). ! 17! 1.3. Basal ganglia neurotransmisors and neuromodulators The main neurotransmitter of the basal ganglia is GABA, which has inhibitory action. In the striatum, GABA is found in the projection neurons, the middle type spiny neurons. These neurons, which have abundant dendritic spines, project their axons to the two GP segments (GPe and GPi) and SNr. In addition to GABA, the middle spiny neurons contain enkephalins or substance P and dynorphins. The projection neurons of the internal and external segments of GP and SNr also contain GABA. Most of the afferents to the SN are GABAergic and at least 70% of the synapses formed on nigral DAergic neurons are GABAergic (Grofova and Rinvik 1970; Gulley and Smithberg 1971). As a result, the basal ganglia exit nucleus is inhibitory. In addition, there are many neuromodulatory substances with slow synaptic actions, such as DA, acetylcholine (ACh), along with Glu (when acting on NMDA receptors) (Mugnaini 1985). Glutamate, considered an excitatory neurotransmitter, is found in the pyramidal neurons, being the main afferent to the striatum (the main input nucleus to the basal ganglia), in the terminals of the thalamic neurons in the striatum and in the projection neurons in the STN (Rajakumar, et al. 1994). ACh is also a normal neurotransmitter in the basal ganglia and is present in the striatal interneurons, where it plays an important role in the neural circuits function (Haber 1986). Its clinical effects are generally antagonistic to those of DA in various movement disorders, especially in Parkinson´s disease (PD) (Alexander and Crutcher 1990). DA is the main neurotransmitter in SN, the structure focus of this work. The number of DAergic neurons in the SN ranges from about 10.000 (bilaterally) in adult mice to about 400.000 in adult humans (Brichta and Greengard 2014; Damier, et al. 1999), and they are predominantly found within the SNc. Within the midbrain DA-releasing catecholaminergic neurons are part of the basal ganglia network, and are crucial for a variety of fundamental brain functions, such as voluntary movement, goal- ! 18! directed behavior and habit formation, motivation, emotion, cognition, reward, memory, associative learning, and decision making (Bjorklund and Dunnett 2007; D'Ardenne, et al. 2012; Gerfen and Surmeier 2011; Pignatelli and Bonci 2015). Dahlström and Fuxe published the first detailed account of the distribution of catecholamines (DA and noradrenaline) and serotonin-containing neurons in the rat brain. Twelve groups of catecholaminergic cells (designated A1–A12) distributed from the medulla oblongata to the hypothalamus were identified (Dahlström and Fuxe 1964). The study provided an evidence for the existence of monoamine-containing neurons in the central nervous system (CNS) (Dahlstroem and Fuxe 1964). Since then this nomenclature has been retained and has proved advantageous because the catecholaminergic cell groups are in most cases not confined to single, defined anatomical structures. In addition, the distribution of cell bodies within each cell group varies markedly between different mammalian species (Fig. 2) (Bjorklund and Dunnett 2007). There are two major DAergic neuron subtypes in the mesencephalon: the nigral A9 neurons and the A10 neurons of the VTA projecting to limbic and cortical areas along mesolimbic and mesocortical pathways. The DA of the SN DAergic neurons is contained in the axons and processes, classically identified as dendrites. The SN DA-containing cell bodies are oval or multipolar, medium-sized that are found rather densely aggregated throughout the SNc and scattered in the SNr (Bjorklund and Lindvall 1975). DA-storing nerve terminals are found in the neostriatum (the caudate nucleus and the putamen), while the DA-containing nerve cells are in the SN (Andén, et al. 1964). ! 19! Figure 2. Schematic representation of DAergic neuron cell groups distribution in the developing (a) and adult (b) rodent brain. The DAergic neurons in the mammalian brain are localized in nine distinctive cell groups, distributed from the mesencephalon to the olfactory bulb, as illustrated in a sagittal view, in (a) the developing and (b) the adult rat brain. The numbering of the cell groups, from A8 to A16, was introduced in the classic study of Dahlstrom and Fuxe in 1964 (Dahlström and Fuxe 1964). (b) The DAergic cell groups principal projections are illustrated by arrows. (Bjorklund and Dunnett 2007). 1.4. Main connections of basal ganglia The basal ganglia may be viewed as two primary input structures, two primary output structures, and two intrinsic nuclei (Fig. 3). The input structures of the basal ganglia are the striatum (caudate and putamen) and the STN. The striatum receives excitatory inputs from virtually all areas of cerebral cortex apart from the primary visual, auditory, and olfactory cortices (Fig. 4). The caudate and putamen are reciprocally interconnected to the SN through the nigrostriatal tract (Grahn, et al. 2008). ! 20! ! Figure 3. Different types of connections between the basal ganglia. D1, D2, striatal neurons preferentially expressing dopamine receptors subtypes D1 and D2. (Chersi, et al. 2013). There are different connections that travels from the cortex to the basal ganglia, and back to the cortex after processing the information. The most important of these connections regarding the subject of this work is the nigrostriatal pathway consisted of DAergic neurons of SN that project to the striatum. This nigrostriatal DAergic pathway, in a restricted sense, is derived from neurons located in both the dorsal and ventral tiers of the SNc, and is the predominant source of DA innervation of the sensorimotor striatum (Bjorklund and Dunnett 2007). ! 21! This nigrostriated projection is mainly ipsilateral, where only 5% of its fibers cross-contralateral neoestriate (Parent, et al. 1983). These DAergic projections establish connections with various neuronal systems, including cholinergic interneurons (McGeer and McGeer 1980). Thus, from 30 to 50% of the caudate nucleus cells project to the SN (Bunney and Aghajanian 1976). In addition, there are also GP neurons that project to the SN (Grofová 1975). Figure 4. Schematic representation of the different cortical areas which receive output from separate basal ganglia thalamocortical circuits. The “motor” circuit is focused on the precentral motor fields, the “oculomotor” circuit on the frontal and supplementary eye fields, the two “prefrontal” circuits on the dorsolateral prefrontal and lateral orbitofrontal cortex, respectively, and the “limbic” circuit on the anterior cingulate and medial orbitofrontal cortex. Taken from (Alexander and Crutcher 1990). ! 28! Figure 7. Proposed timeline from onset to death in classical PD. A 40-year course is assumed although it is inevitable that there will be considerable individual variability. The various steps are presented in linear fashion, but the disease is likely to show an accelerated course in the latter years. Braak pathological staging is indicated in Arabic numerals below the horizontal line. Above this line are the proposed symptoms in the prodromal phase and to the right of this, the major features according to Hoehn & Yahr staging (Roman numerals). The two symbols at the extreme left of the timeline represent two possible causative agents: genetic mutations or viral infections. It is uncertain whether the sympathetic nervous system is involved in Braak stage 1 or 2; hence arrows straddling both of these phases indicate the pathology. Abbreviations: CA2 ¼ second section of the Ammon’s horn; CN X ¼ motor component of cranial nerve X; RF ¼ reticular formation; CN ¼ central subnucleus of the amygdala; Meynert’s n ¼ basal nucleus of Meynert; PPN ¼ pedunculopontine tegmental nucleus; TEC ¼ transentorhinal cortex. (Hawkes, et al. 2010). ! 29! As the disease progresses dyskinesia, ataxia, pain and sensory complaints such as orthostatic intolerance and urinary incontinence, and neuropsychiatric manifestations (hallucinations, depression, and dementia) become outstanding. This happens due to the spread of the disease pathology to other parts of the brain (Chaudhuri and Schapira 2009). The diagnosis of PD is generally made based on the clinical data exposed. However, new and modern imaging techniques, such as neurochemical positron emission tomography (PET) and single photo neuropathology studies tomography (SPECT) imaging studies, may play an important role in the differentiation of the various parkinsonian syndromes or in the preclinical diagnosis of this disease. This technique has complemented neuropathology studies by allowing in vivo assessment of the regional distribution and quantitative measurement of cholinergic terminal markers or receptors in the brain of PD patients or related syndromes. These imaging technologies give the opportunity to study cholinergic innervation in vivo at early stages of PD and other neurodegenerative disorders (Bohnen and Albin 2011; Forno 1982). The medical treatment of PD is usually effective and must be continued for the rest of the patient's life. The degree of improvement achieved with the range of drugs used is variable in each case, indicating that these treatments are not curative at present, but palliative. In general, it is based on the following measures: a) Anticholinergics b) Amantadine c) L-DOPA (or levodopa) and other DAergic drugs, d) Other general measures. However, although for most patients these measures maintain a good quality of life, when the disease progresses or when these drugs do not improve mobility or produce significant adverse ! 30! effects, surgical treatment based mainly on pallidotomy and thalamotomy, and high-frequency deep brain stimulation (DBS) of the subthalamic nucleus (STN-HFS) would be the preferred treatment for advanced PD (Benabid, et al. 2009). 2.2. Pathology of Parkinson’s disease In PD there is a severe pathology of the basal ganglia, which starts as an extranigral involvement in the disease. It starts in the lower brainstem (medulla) and the anterior olfactory nucleus regions. These regions include the periamigdaloid nucleus, the locus ceruleus, the dorsal motor nucleus of the vagal nerve, the innominate substance, the raphe nuclei, and the hypothalamus, which accounts for the non-motor manifestations of the disease. However, other authors (Braak, et al. 1996) have suggested that neurodegeneration begins in the entorhinal cortex, so that the cell death cascade begins in this area and continues with the degeneration of the SN. Therefore, the most important neuropathologic hallmark of PD is the loss of DAergic neurons of SNc. Actually, this loss of DA in the basal ganglia appears to produce the most debilitating neurological signs. Therefore, the fundamental neuropathological characteristic in PD is the bleaching of the SN band, usually gray or black by the presence of neuromelanin, due to the loss of myelinated neurons (Fearnley and Lees 1991; Jellinger 1991; Jellinger 2001). It requires the death of at least 50-60% of the SN DAergic neurons for the manifestation of clinical symptoms, which demonstrates the high plasticity of the nigrostriatal DAergic system. Since the disease shows a special predilection for the ventrolateral zone of the SN, it is in this type of cells where the lesion begins, to progress later to the dorsal area of this structure. Cell loss in the SN causes depletion of striatal DA, which results in hyperactivity of striatal projection neurons belonging to the indirect pathway because the corticostriatal glutamatergic excitatory input is no longer inhibited. This might be one cause of hypokinesia. ! 31! At final stages the process enters the neocortex, at which all the clinical minifestations of the disease appears (Braak, et al. 2004). Another evidence of PD is the presence of Lewy bodies in the CNS of individuals who are not yet diagnosed with PD. Although Fredrich Lewy was the first to describe them more than a century ago, their importance in the pathogenesis of PD is still unclear (Bohnen and Albin 2011; Goedert, et al. 2013). These intra cytoplasmic inclusions mainly contain filamentous forms with a dense center consisting of radially oriented filaments. It appears that their formation is the result of an abnormal accumulation of cytoskeletal proteins within the neuron, as well as α-synuclein (α-syn) and ubiquitinated proteins (Lees, et al. 2009). There is a debate about whether Lewy bodies directly cause neuronal death or perhaps isolate small aggregates of cellular toxins to preserve neuronal viability (Goldberg and Lansbury 2000). Some theories indicate that inclusions found in brain tissues with neurodegenerative diseases, including PD, are a mechanism to compartmentalize accumulated defective proteins and prevent them from obstructing the proper functioning of the cell (Goldberg and Lansbury 2000). Another important feature that characterize PD is the presence of neuroinflammation 2.3. Neuroinflammation on Parkinson’s disease Neuroinflammation is a complex process involving cells of the immune system and the CNS, which have the purpose of repairing the damage produced. This inflammation could be both beneficial and detrimental to the brain. On the one hand it is part of a physiological repair process; however, when this process is not controlled, the inflammatory process is extended, thus the inflammation loses its repairing function and can be the cause of the damage. The main players in the inflammatory processes of the brain are the microglia and astrocytes. Microglia are the innate immune cells resident in the CNS. When they are at resting state they have a ramified morphology, but when they become activated upon stimulation, they transform into an amoeboid morphology (Fig. 8). ! 32! Figure 8. Morphological differences between resting and activated microglia Besides, macrophages/microglia have diverse functions that range from fighting bacterial infection to tissue regeneration and wound healing. The diverse functions of microglial cells in the CNS are mirrored by equally diverse phenotypes. A classical model of pro-inflammatory/M1 versus an antiinflammatory/M2 microglia has been extensively used. However, the complex and different functions of microglial cells can only be explained by the existence of varied and plastic microglial phenotypes mediated by distinct gene expression programs and a network of molecular pathways that relay environmental signals via signaling cascades (Guerau-de-Arellano y col., 2017). Therefore, M1 and M2 are just the extremes of a broad spectrum of phenotypes that cover the different functions of microglia. These different phenotypes can be achieved by stimulating microglial cells with different compounds. Hence, when stimulated with LPS and interferon (INF)- γ, macrophages/microglia has long been known as classically activated or M1 (Martínez et al., 2014), while when activated with IL-4 macrophages/microglia show an alternative activated phenotype or M2. In order to standardize the nomenclature and facilitate the communication of macrophage/microglia data, a novel nomenclature has been proposed in which the letter M is ! 33! followed by a parenthesis that includes the stimuli used for activation (Murray et al., 2014). The knowledge of the molecular programs that control the inflammatory phenotypes versus resolution provides a unique opportunity to find new targets that allow modulating these phenotypes and, therefore, controlling the excessively inflammatory responses that accompany neurodegenerative diseases. The knowledge of these molecular mechanisms is greatly advancing in recent years. Functionally, M1 microglia is responsible for fighting infections, for which it adopts a clear proinflammatory phenotype with microbicidal, antigen-presenting and immune-enhancing functions. This type of microglia is characterized by the production of NO by the inducible nitric oxide synthase (iNOS, encoded by the Nos2 gene; MacMicking et al., 1997; Arnold et al., 2014) and by the expression of inflammatory chemokines and cytokines, such as interleukin (IL)-6, IL-12, IL-1β, IL-23 and tumor necrosis factor (TNF)-α that attract new cells of the immune system to the site of infection (Mosser and Edwards, 2008; Murray et la., 2014). In the context of neurodegenerative disease this phenotype produces harmful effects in the neuronal population. When neutrophils undergo apoptosis and microglia switch to a resolution/M2 phenotype, the initial acute inflammation evolves to a resolution phase (Serhan, 2014). This resolution/would healing phase is mediated by lipid mediators, such as classical eicosanoids, phospholipids and sphingolipids, endocannabinoids and specialized proresolving mediators (Chiurchiù and Maccarrone, 2016), that promote the switch to M2 (Bosurgi, 2017). Resolution/M2 microglia suppresses IL-12 secretion and induces the release of IL-10, transforming growth factor (TGB)-β, IL-1R antagonist and decoy IL-R II (Brancato y Albina, 2011). Besides, these microglial cells induce the expression of arginase-1 instead of iNOS, switching arginine metabolism from production of NO to ornithine, and also increase polyamines production for extracellular matrix and collagen synthesis (Gordon y Martínez, 2010). This phenotype promotes the neuroregeneration and tissue repair. ! 34! Uncontrolled M1 activated microglia can produce different substances that are considered neurotoxic, such as pro-inflammatory cytokines, TNF-α, NO, superoxide, and prostaglandin E2, and can produce neuronal death. In turn this neuronal damage cause further microglia activation, known as reactive microgliosis (presence of microglia in nervous tissue secondary to injury), leading to non-stop neuronal damage and death (Block and Hong 2005; Streit, et al. 1999). Thus, inflammatory response in the CNS is thought to favor damage more than reparative function and could explain most CNS pathologies. Therefore, in recent years, numerous therapeutic attempts have been aimed at alleviating it. However, at the moment, it does not seem that any drug has been useful to stop or control neuroinflammation. In the case of DAergic neurons, inflammatory processes (proliferation and activation of microglia) could be a secondary event produced by degenerative processes that occur directly in DAergic neurons, but at the same time it could help the progression of degeneration. In addition, there is the possibility that inflammatory processes by themselves could be harmful to DAergic neurons in the nigrostriatal system (fig. 9). There are several physiological/pathological inducers of neuroinflammation that could be involved in the degeneration of the DAergic system: neuronal injury (stroke or brain trauma) (Lees 1997; Tansey, et al. 2007), immunological challenges (viral or bacterial infections), other inflammatory syndromes (diabetes mellitus type 2, arthrosclerosis, rheumatoid arthritis, and multiple sclerosis), pro-inflammatory compounds (IgG, thrombin, LPS), and environmental toxins (like pesticides). Thus, different toxins, such as 6-OHDA, MPTP/MPP + or rotenone, induce an inflammatory process in addition to its specific toxic mechanisms. Many of these factors can increase the permeability of the blood brain barrier (BBB), thus allowing infiltrations of macrophages, lymphocytes and other toxins in to the brain (Tansey and Goldberg 2010), which could be involved in degeneration of the DAergic system. ! 35! Figure 9. Relationship between microglial activation and neuronal death. Microglia can be a double player in inflammation releasing either neurotoxic or neuroprotector substances (Chen, et al. 2016b). 3. NEURODEGENERATION THEORIES Neurodegenerative diseases are very complex pathologies that have as common characteristics the progressive character of the symptoms and the gradual degeneration of a part or parts of the nervous system that leads to physical and psychic incapacity. In the case of PD, as stated before, a number of key alterations occur in the SN, leading its degeneration. ! 36! Different theories try to elucidate the mechanisms involved in the degeneration of the DArgic system (Tomás-Camardiel y col., 2004): A) Accumulation of both intra and extracellular misfolded proteins. An impairment in the protein quality control and clearance systems such as the ubiquitin-proteosome system (UPS) accounts for the formation of intracellular accumulations of proteins (Jellinger 2010). All this is known as proteinopathies, and is thought to play a major role in the neuronal dysfunction and death that characterizes the development of neurodegenerative diseases, such as Alzheimer's disease, PD, Huntington's, amyotrophic lateral sclerosis, and frontotemporal dementia (Abeliovich and Gitler 2016; Gitler, et al. 2017; Pierre, et al. 2009). If patients with PD have a malfunction of the UPS, the proteins that normally renew through this route will tend to aggregate and form inclusions. B) Oxidative stress. The most important of neurodegeneration theories refers to oxidative stress as a cause of damage to neurons of the SN. This oxidative stress is triggered upon intracellular accumulation of reactive nitrogen species and reactive nitrogen oxygen (ROS) as a result of either reduction in the endogenous anti-oxidant capacity or the ROS overproduction (Lin and Beal 2006). During oxidative or respiratory explosion numerous enzymatic systems are activated among which are the NADPHox, iNOS and myeloperoxidase (MPO), which release into the medium toxic amounts of superoxide (O2-), free radicals of NO (that cause the nitrosylation of certain proteins) or hypochlorous acid (HClO), a powerful oxidant, by converting amines to chloramines, phenols, etc., which in turn can damage macromolecules directly through the production of OH-. The expression of these biocatalytic systems in SN is substantially higher in patients with PD (Fig. 10). ! 37! C) Mitochondrial energy failure. There are numerous data that point to the involvement of mitochondrial defects in the pathogenesis of PD (Beal, 2003). Although this does not seem to be the primary cause of the disease it could be an important factor in its development. D) Inflammation. Since 1988, when McGeer et al. showed the presence of inflammatory markers in the CNS of PD patients, the implication of inflammatory events in the pathogenesis of this neurodegenerative disease was postulated (McGeer, et al. 1988a; McGeer, et al. 1988b). This was supported by the presence of activated microglia, accumulation of cytokines, activation of Nuclear Factor kappa β (NF-κβ) pathway, and oxidative damage to the proteins in the brains of PD individuals (McGeer et al., 1988; Hirsch and Hunot, 2009) and animal models (Gao, et al. 2002; Gao, et al. 2008; Herrera, et al. 2000). These authors showed the presence of T lymphocytes cytotoxic suppressors CD+ and microglial cells with the activated major histocompatibility complex class II (MHC II), in the tissue of patients with PD. In addition, different inflammatory markers have been described in parkinsonian brains (Hunot and Hirsch 2003): (a) Significant proliferations of ameboid macrophages and reactive microglia (Hirsch, et al. 1998; McGeer, et al. 1988a); (b) Activated glial cells expressing different pro-inflammatory cytokines such as TNF-α, IL-1β and IFN-γ also appear in the SN of this type of patients (Hunot, et al. 1999), together with an expression of iNOS (Herrera, et al. 2000; Hunot, et al. 1997; Mogi, et al. 1998); c) Increases in the expression of IL-1β, IL-6 and TNF-α in the cerebrospinal fluid of patients with PD. ! 44! mediators of inflammation. In addition, thrombin injection produces selective destruction of the SN DAergic neurons, leaving unaltered others neuronal phenotypes which are present in this structure (Carreno-Muller, et al. 2003). 5. LIPOPOLISACCHARIDE LPS represents the main component of the outer membrane of Gram-negative bacteria and some cyanobacteria. It plays a key role in the development of infections and sepsis (Rietschel and Brade 1992; Schletter, et al. 1995). The LPS was discovered about 100 years ago by Richard Pfeiffer and, unlike the exotoxin secreted by the cholera bacteria, it is a heat-stable toxin (Rietschel and Cavaillon 2003). Although there is a great variation in the composition of the LPS, they all show a common structure. They are mainly composed of three parts: lipid A, core, and specific O-chain. Lipid A is the toxic part of LPS and the hydrophobic component. When bacterial cells are lysed by the immune system, fragments of membrane containing lipid A are released into the circulation, causing fever, diarrhea, and possible fatal septic shock (Lerouge and Vanderleyden 2002). The polysaccharide and hydrophilic part comprises two regions: the O-specific chain or O antigen, and the central region or core. The O antigen is attached to the core and consists of repetitive units of oligosaccharide. The core, which consists of short chain of sugars, in turn is subdivided into an external core (formed by hexoses), through which it binds to the O antigen, and the internal core (formed by heptoses). Lipid A is generally composed of a phosphorylated disaccharide, linked to two D-glucosamines in position β1,6 and carrying a maximum of six or seven acyl residues. It is attached to the internal core by a residue called 2-keto-3-deoxyoctanoic acid (KDO) (Fig. 12). ! 45! Figure 12. General chemical structure of bacterial lipopolysaccharide (Pupo and Hardy 2009). 5.1. Interaction between LPS and soluble membrane proteins A prerequisite for the recognition of LPS by the innate immunity cells and thus activate them is its interaction with specific LPS-binding molecules on the surface of the target cells. Several LPS binding structures have been described, but physiological relevance has only been demonstrated for some of them (Fig. 13). The activation starts when LPS binds to a 60 kDa serum glycoprotein called LPS-binding protein (LBP), which in turn catalyzes the transfer of LPS to membrane-bound CD14 (mCD14) or soluble (sCD14) (Triantafilou and Triantafilou 2002). CD14 is the most important cell surface protein related to LPS binding and cell activation (Tanaka, et al. 2000). CD14 is a 55 kDa glycoprotein that exists as a membrane protein (CD14m) anchored to a glycosylphosphatidylinositol (GPI) in monocytic cells, epithelial cells, polymorphonuclear leukocytes and some B lymphocytes (Gu, et al. ! 46! 1998). The binding of LPS to CD14m in monocytes is necessary for the stimulation of these cells, which elicits strong pro-inflammatory responses causing production and release of immune mediators. As mentioned before, LPS binding to the CD14 receptor is strongly increased by LBP (Staal and Sonsalla 2000; Tanaka, et al. 2000). During the acute phase response, the concentration of LBP increases from 5-10 µg/ml to 200 µg/ml (Staal and Sonsalla 2000). LBP reduces the concentration of LPS required for the activation of monocytes by forming LBP-LPS complexes, which are recognized by CD14 (Fig. 13). In addition to CD14m, there are different types of soluble CD14 (48, 53 and 55 kDa) present in concentrations around 2-6 µg/ml in serum (Frey, et al. 1992), and which are released by monocytes or secreted by free forms of GPI (McNaught and Olanow 2003). It is known that LPS is capable of binding directly to the soluble forms of CD14 (sCD14), a process highly facilitated by LBP, although LBP is not present in the LPS-sCD14 complexes. The LPS-sCD14 complexes can activate some mCD14-negative cells, such as endothelial cells, fibroblasts and smooth muscle cells, to produce cytokines (Frey, et al. 1992; Haziot, et al. 1993; Pugin, et al. 1993). Thus, it has been postulated the existence of a specific receptor for the LPS-sCD14 complexes that would be expressed in these cells and that would mediate the activation by LPS of cells lacking CD14. In conclusion, CD14 as a soluble membrane receptor would play an extremely important role in the molecular recognition of LPS in several cells. However, it appears that additional membrane molecules are required for cellular activation (Rietschel, et al. 1996). ! 47! Figure 13. Schematic view of the LPS signaling pathway through TLR4 (Diamond, et al. 2015). A family of proteins called Toll-like receptors (TLRs) has been identified in humans and mice. These are transmembrane proteins with an extracellular domain with leucine-rich repeats (LRR) and a cytoplasmic receptor Toll/IL-1 (TIR) with a domain of great structural similarity to the IL-1 receptor (Anderson, et al. 2000; Fukata and Abreu 2012; O'Neill and Greene 1998; Shah 2008). There is evidence that these TLRs are the most important pathogen recognition molecules and they are used in the start of the signaling cascades in response to infections with different classes of pathogens (Lester and Li 2014). ! 48! Both of LBP and CD14 are considered key accessory molecules for the efficient transfer of LPS to the Toll like receptor (TLR)4-MD2 (MD2 is an accessory molecule required by the TLR4 to capture its ligand, LPS) and increasing the sensitivity of LPS detection (Hailman, et al. 1994; Ryu, et al. 2017). 5.2. Transduction of signal induced by LPS Once the LPS stimulate the target cells a series of intracellular events occurs. The most important and best characterized is the activation of the NFκB, which plays a central role in the regulation of inflammatory and immune responses (Ghosh, et al. 1998; May and Ghosh 1998). NFκB represents a group of structurally related proteins, including p65, p50, p52, RelB and c-Rel. Under normal conditions, NFκB is sequestered in the cytosol as an inactivated homo or heterodimeric form, with non-covalent interactions with its inhibition proteins, called IκBs. After stimulation with the appropriate agonist, IκB is phosphorylated, ubiquitinated and degraded by the proteasome (Israel 2010). Thus, NFκB is released and translocated to the nucleus initiating gene expression phenomena. The target genes of NFκB include those that code for cytokines, chemokines, adhesion molecules, acute phase proteins, antimicrobial peptides, iNOS and cyclooxygenase 2 (COX-2). Together, these mediators provide immediate protection to the host and induce the development of the adaptive immune response (Ghosh, et al. 1998; May and Ghosh 1998). In recent years there has been a significant expansion of our knowledge about the signaling mechanisms by which LPS induces the activation of NFκB. In fact, several of the molecular components involved in this process have been discovered (Fig. 13 and 14). The TLR4 is a primary transmembrane signaling receptor for the LPS. Once bound to CD14, LPS must be transferred to TLR4, resulting in the homodimerization of TLR4 and a conformational change of its cytosolic domain (Kagan 2017). Subsequently, an adaptor protein called MyD88 is recruited on the receptor, ! 49! which is followed by the interaction with the kinase associated with IL-1 receptor (IRAK)-1, -2 or - M. The IRAK dissociates from the receptor complex and recruits TNF receptor-associated factor 6 (TRAF6), resulting in the assembly and activation of the IKKα/β/γ complex and the subsequent phosphorylation and degradation of IκB and translocation to the nucleus of NFκB (Kawai, et al. 1999; Lu, et al. 2008; Rosadini and Kagan 2017). The mechanism by which TRAF6 activates IKKs is not fully understood, although there is evidence that show that TRAF6 could interact with additional molecules that would help to activate kinase proteins in cascade, leading to phosphorylation and activation of IKKs. Three TRAF6 proteins have been identified, named evolutionarily conserved signaling intermediate in Toll pathways (ECSIT), TAK-1 binding proteins (kinase activated by transforming growth factor β) -2 (TAB2), and TRAF6 binding protein (Fig. 14) (Wi, et al. 2014). There are other signal transduction pathways of the LPS. In several in vitro studies with monocytes and fibroblasts, G proteins and small G proteins have been implicated in the activation of tyrosine kinases (TK) (Cattaneo, et al. 2014; Mayeux 1997; Tanke, et al. 1991), phospholipase C (PLC) and (A2) (Chang, et al. 1990; Fleming, et al. 1996), as well as calcium modulated protein (calmodulin) (Mattson and Rydel 1996; Nakano, et al. 1993). Also sphingomyelitis has been considered another second messenger, which would hydrolyze sphingomyelin into ceramide, which would activate different protein phosphatases and protein kinase (PK), as the PK activated by ceramide (Joseph, et al. 1994; Liu, et al. 1994), which could intervene in the LPS signal by activating or inhibiting various enzymes such as PK C, phospholipases PLC and PLA2, and COX-2 (Hayakawa, et al. 1996; Liu, et al. 1999). ! 50! Figure 14. Molecules implicated in the transduction of signal induced by LPS (Wi, et al. 2014). Another way is formed by PKs, in which several groups are involved: the serine-threonine PK A and C (Kozak, et al. 1997) and a large set of TK (Ruetten and Thiemermann 1997). The set of mitogen-activated TKs, MAP kinases (MAPKs), also participate to a large extent in the intracellular signals of the LPS. ! ! 51! 5.3. Mitogen activated protein kinase This large family consists of PK that phosphorylates serine-threonine and tyrosine-threonine residues. They act in cascade form, in which signal is spread by sequential phosphorylation and activation of kinases (Plotnikov, et al. 2011). According to different in vitro studies performed on macrophages, leukocytes, endothelial cells, smooth muscle cells and other cell types (Baydoun, et al. 1999; Downey and Han 1998), there are at least four subgroups of MAPK, of which it has been described that three are related to the responses induced by the LPS. One of them is formed by kinases regulated by extracellular signals (ERK). ERK1/2 is responsible for proliferation and differentiation, neuronal plasticity, and apoptosis. The other pathways related to MAPKs include the set of proteins that make up the PK subfamily of transcription factor c-Jun, called JNK. JNK is a stress-associated MAPK pathway that can be activated by inflammatory cytokines, oxidative stress, and UV radiation. It phosphorylates a large number of nuclear and cytoplasmic substrates, which in turn regulates many genes involved in cellular processes such as apoptosis and neuronal and immunological activities (Haeusgen, et al. 2009; Rincon and Davis 2009; Yoon and Seger 2006). Dysregulation of JNK has an important role in a number of neurodegenerative diseases, mainly Alzheimer’s disease, PD, amyotrophic lateral sclerosis, and other diseases such as diabetes and cancer. P38 MAPK pathway is known as a stress-activated protein kinase pathway. It plays a crucial role in regulating the response with cytokines after stimulation with LPS. In fact, it is known that a specific inhibitor of these kinases, SB203580, completely prevents the release of cytokines after stimulation of monocytes with LPS (Lee, et al. 1994). For this role in inflammation induction, it is considered a therapeutic target for inflammatory and autoimmune diseases. P38 also has an important role in the management of immunological effects, cellular and apoptotic aging, and cell cycle checkpoints. Its importance in such processes gives an overview about the pathological manifestations resulting ! 52! from its dysregulation. It also plays a role in cardiovascular and neurodegenerative diseases, diabetes, and cancer (Plotnikov, et al. 2011). Other important pathway is the PI3K/AKT, which is an intracellular signaling pathway involved in cell cycle regulation. PI3K phosphorylates and thus activates AKT. AKT is a Ser/Thr kinase, which regulates cell survival, proliferation, migration, and differentiation through the mTorc1 pathway, p53 pathway, Forkhead O family (FOXO) transcription factors and other substrates. AKT activation is considered the main output of growth factor-induced PI3K signaling (Dinsmore and Soriano 2018). 5.4. LPS as an inductor of the inflammasome components expression Inflammasomes are cytosolic multiprotein complexes in the CNS that enable autocatalytic activation of inflammatory caspases upon different stimuli, resulting in the production of proinflammatory cytokines including IL-18 and IL-1β (Lamkanfi and Dixit 2012). These multiprotein complexes constitute a part of the innate immune response, functioning as intracellular sensors and responding to infectious agents, pathogenic stimuli, and host-derived danger signals associated with neurological diseases. Inflammasomes consist of three components: caspase-1, a cytosolic patternrecognition receptor, and an adapter protein enabling interaction between them. Upon activation, the cytosolic receptors, including members of the NOD-like receptor (NLR) family (Nucleotidebinding oligomerisation domain receptors), undergo oligomerisation results in the assembly of the inflammasome complex and increase in pro-IL-1β and NLRP3 levels. This acts as a recruitment and activation way for pro-caspase-1. Activation of caspase-1 leads to the cleavage and release of IL-1β and IL-18, resulting in inflammatory reactions. Both the ligands and their cognate receptors are expressed in the CNS, where they modulate homeostatic and neuroinflammatory processes (Fig. 15) (Lamkanfi and Dixit 2012; Machado, et al. 2016; Schroder and Tschopp 2010; Walsh, et al. 2014). ! 53! Figure 15. Schematic representation of inflammasome activation process (Shao, et al. 2015). ROS is produced in response to initiating factors, triggering the inflammasome mediated inflammatory cascade. Components oligomerization results in inflammasome assembly. As a consequence it activates IL-1β and IL-18 through caspase-1, which causes oxidative DNA damage. Alongside with inflammation, they cause pyroptosis (Harijith, et al. 2014). 6. THE LPS AS A MODEL OF THE PD As already mentioned, and due to the importance of neuroinflammation role in the pathology of PD, there are several models that allow us to study the molecular mechanisms involved in ! 60! The probability of getting PD in DM patients increases with age. Therefore, researchers found that both PD and DM are age-related, and that they share similar pathways of mitochondrial dysfunction (Santiago and Potashkin 2013b; Yang, et al. 2017) and have a common feature of hippocampal volume loss. It was reported that PD might predispose to DM because of mutations of mitochondria-associated proteins, such as DJ-1 and Parkin, which have been implicated in familial PD. These alterations may promote mitochondrial dysfunction in beta islet cells, thereby leading to T1D development. Inflammation appears to play an important role in both DM and PD. In DM, CNS neuronal loss is focused in the hippocampus, while in PD neurodegeneration is centered in the SN; since midbrain DAergic neurons have some projections into the posterior hippocampus, it is possible that DA neuronal loss in PD with DM could start in the hippocampus and progress to the midbrain (Miller, et al. 2017). Recently the neuronal protein named amyloid precursor protein (APP), which was previously linked to neurodegeneration and insulin regulation, has been identified and described as a potential biomarker for PD in T2DM patients. It was found that APP expression is upregulated in PD, which suggests that increased expression of APP in blood may alter the neurodegenerative phenotype in type 2 DM (Santiago and Potashkin 2013a). 7.3. Treatment of diabetes mellitus At the beginning of diabetes, it can be usually controlled by changing lifestyle that includes weight loss, a normal or hypocaloric healthy diet and daily physical exercise. Exercise reduces the need for insulin even when all other factors remain the same, since working muscle has some ability to take up glucose without insulin help. The capillary blood glucose monitoring is very important whether insulin is being used as a treatment or not. If these methods (diet and exercise) do not help achieve ! 61! good metabolic control, oral antidiabetic drugs will be used as initial therapy. Normally these drugs cannot give long-term control, thus insulin will be added to maintain glycemic control. Most of the antidiabetic drugs are orally administrated. Selection of the optimal compound depends mainly on the nature of diabetes, age and patients situation. There are different types of oral antidiabetics that can be classified depending on their mechanism of action: Secretagogues: Sulfonylureas and Meglitinides (repaglinide) Sensitizers: Biguanides (metformin) and thiazolidinedione (rosiglitazone and pioglitazone) Alpha-glucosidase inhibitors and peptide analogs: (acarbose) 7.4. Metformin and its history Commercially named metformin hydrochloride, is a member of the biguanide class of oral antihyperglycemic agents (3- (diaminomethylidene) -1,1-dimethylguanidine). Its empirical formula is C4H11N5 * HCl. It is a white to off-white, crystalline powder, which is easily soluble in water and practically insoluble in acetone, ether and chloroform, and whose structural formula is shown in Fig. 16. Figure 16. Chemical structure of Metformin. ! 62! Discovered in 1922, metformin was accepted clinically for the treatment of hyperglycemia in England in 1958, in Canada in 1972, and in the USA in 1995, for its antihyperglycaemic effect by reducing hepatic glucose production and decreasing insulin resistance, as well as fasting plasma insulin levels. Insulin resistance has been the central mechanism underlying various pathologies including DM type2, by decreasing this resistance. Hence, metformin imparts its beneficial role in a broad spectrum of diseases associated with insulin resistance (Viollet, et al. 2009). In addition, it has therapeutic potential in other conditions among which include diabetic nephropathy, cardiovascular diseases, polycystic ovarian disease and has even been investigated as antiviral and anticancer (Ben Sahra, et al. 2010; Rotella, et al. 2006). In fact, today, it is the most widely prescribed treatment to treat hyperglycemia in individuals with DM type 2 and is recommended as first-line oral therapy. Its beneficial effects are shown by acting on different organs such as liver and intestine involving complex mechanisms in which mitochondria is its main target (Foretz and Viollet 2015). It lowers blood glucose levels by suppressing hepatic glucose output, increasing insulin-mediated glucose disposal and intestinal glucose use, and decreasing fatty acid oxidation metformin is effective in reducing high blood glucose levels as many other antidiabetics but, unlike many of them, by itself, does not produce hypoglycemia. It also reduces insulin levels, inflammation and thrombosis, and the risks of metabolic syndrome and diabetes in healthy persons (Li, et al. 2015). It also has beneficial effects on circulating lipids linked to cardiovascular diseases (Wu, et al. 1990). More interestingly it has shown neuroprotective effects. Some recent analyses have shown a reduction in PD risk in DM type2 when used in combination with sulfonylurea (Bayliss, et al. 2016; Wahlqvist, et al. 2012). ! 63! 7.5. Metformin mechanism of action The main action of metformin is to decrease glucose concentrations without causing hypoglycemia. This is mainly achieved by inhibiting gluconeogenesis thus decreasing hepatic glucose production (Natali and Ferrannini 2006). This inhibitory action on hepatic gluconeogenesis could be due to either reduction in the hepatic uptake of gluconeogenic substrates or changes in the enzyme activities (Argaud, et al. 1993; Radziuk, et al. 1997). It exerts its direct metabolic action on the liver (Fig. 17), apparently due to the predominant expression of organic cation transporters 1 (OCT 1) in hepatocytes, which in turn facilitates the cellular entry of metformin, due to its high dissociation constant. Under physiological conditions it exists in a positively charged protonated form and can only cross the plasma membrane by passive diffusion (Shu, et al. 2007). The interest in the therapeutic use of metformin has been increased since the recognition of its pleiotropic actions in various tissues, which are affected by insulin resistance and /or hyperinsulinemia (Palomba, et al. 2009). This improvement in insulin sensitivity is associated with its positive effect on the expression of insulin receptors (Gunton, et al. 2003). According to Zhou et al. (Zhou, et al. 2001), the pleiotropic actions of metformin are intimately associated with the activation of AMP-activated protein kinase (AMPK), an important protein kinase enzyme in the body energy balance signaling and a master regulator of metabolism (Gruzman, et al. 2009). It acts as a fuel gauge monitoring systemic and cellular energy status controlling cellular functions under restrictive energy conditions (Galdieri, et al. 2016). It responds to metabolic requirments either by stimulating energy production including glucose and lipid catabolism or by inhibiting energy consuming pathways. It is known that impairment of AMPK activity can induce insulin resistance (Salminen, et al. 2011a). AMPK is a heterotrimeric protein consisting of an α catalytic subunit and two β and γ regulatory subunits, where each subunit has at least two isoforms. AMPK is activated by an increase in the AMP/ATP ratio and/or intracellular ADP/ATP resulting from the imbalance between consumption ! 64! and production of ATP. This activation involves the binding of AMP to the regulatory sites of the γ subunit. This causes conformational changes that allosterically activate the enzyme and inhibit the dephosphorylation of Thr172 within the α subunit (activation of AMPK requires phosphorylation of Thr172 by kinases such as LKB1 and CaMKKβ). Once AMPK is activated by ATP depletion, it inhibits the pathways that consume ATP and stimulates alternative pathways for the regeneration of ATP (Carling 2004); that is, it changes the cells from an anabolic state (ATP consuming) to a catabolic one (ATP producing), closing the synthetic pathways that consume ATP and restoring the energy balance. As a result, the synthesis of glucose, lipids, and protein, and cell growth are inhibited while the oxidation of the fatty acids and the uptake of glucose are stimulated (Fig. 17). Figure 17. Mechanism of action of metformin (He and Wondisford 2015). ! 65! But the activation of AMPK by metformin does not seem to occur directly, but is secondary to its effect on the mitochondria, which is the drug primary target. As reported in early 2000, metformin induces specific inhibition of mitochondrial respiratory chain complex I (Fig. 18). This was initially observed in perfused livers and hepatocytes isolated from rodents (El-Mir, et al. 2000; Owen, et al. 2000), but later also in other tissues, such as skeletal muscle (Brunmair, et al. 2004), endothelial cells (Detaille, et al. 2005), pancreatic beta cells (Hinke, et al. 2007) and neurons (El-Mir, et al. 2008). Figure 18. Metformin primary target is the mitochondrial respiratory chain complex 1. (Viollet, et al. 2012). ! 66! Although the exact mechanism by which metformin inhibits complex I of the respiratory chain remains unknown, it has been shown that this effect does not require AMPK. Therefore, the activation of AMPK by metformin in the liver, and probably in other tissues, is the direct consequence of a transient reduction in the cellular energy state induced by the specific inhibition of the respiratory chain complex I by the drug (Stephenne, et al. 2011), which results in low oxygen consumption by the cells and an increase in the NADH/NAD ratio. The β-oxidation of the fatty acids is carried out by the enzyme L-3 hydroxy acyl CoA dehydrogenase, which uses NAD as a cofactor. When there is an increase in the NADH/NAD ratio, the β-oxidation of the fatty acids is blocked, increasing the AMP/ATP ratio and resulting in the activation of AMPK (Aljada and Mousa 2012) (Fig. 19). Figure 19. Metformin interrelation with AMPK (Aljada and Mousa 2012). LBK1% ! 67! The demonstration that complex I of the respiratory chain and not AMPK is the main target of metformin was reinforced by showing that the metabolic effect of the drug is preserved in the liver of mice deficient in AMPK, so that other mechanism (s) independent of AMPK could be involved. Changes in AMPK signaling are associated with human health, and implicated in T2DM, inflammatory disorders, stroke and neurodegenerative diseases (Salminen, et al. 2011b). 7.6. Anti-inflammatory effects of metformin As it has already been demonstrated in different studies, inflammation seems to be implicated in several pathological mechanisms, such as, for example, those that lead to the death of DAergic neurons in PD. Several results from clinical and experimental studies suggest that metformin, in addition to its hypoglycaemic action, may attenuate both peripheral and central inflammation. The anti-inflammatory potential of metformin has been studied in particular in experimental models of peripheral inflammation. It has been shown that it attenuates the pro-inflammatory response in endothelial cells (Isoda, et al. 2006), decreases the proliferation of smooth muscle cells of human aorta (Li, et al. 2005) and reduces the activation of macrophages (Mamputu, et al. 2003). Isoda et al. (Isoda, et al. 2006) investigated these anti-inflammatory effects in cells of the human vascular smooth muscle, where they observed that metformin exerts an anti-inflammatory. On the other hand, in endothelial cells, the activity of AMPK is associated with the phosphorylation and activation of endothelial nitric oxide synthase (eNOS), which results in an anti-inflammatory action in the vascular wall (Zou, et al. 2002). Anti-inflammatory effects of metformin have also been described in human monocytes in which a reduction in the production of TNF and tissue factor is observed through the inhibition of ERK1/2- Egr.1 pathway (Arai, et al. 2010). Recent studies have also shown that metformin significantly reduces inflammatory cytokines such as TNF-α, monocyte chemoattractant protein 1 (MCP-1), IL- 1β, macrophage inflammatory proteins (MIP-1α), IL-6, leptin, and IL-18 in uveitis induced by ! 68! endotoxin in rats (Kalariya, et al. 2012). Similarly, a recent study showed upon metformin treatment a reduction in macrophage infiltration and reduction in TNF-α, MCP-1 and leptin in adipose tissue of high-fat fed rats with insulin resistance (Lu, et al. 2016a). Another study showed that metformin treatment inhibited the cardiac expression of proinflammatory cytokines such as TNF-α, IL-1β, and IL-6 in endotoxin-induced acute myocarditis mice through the AMPK activation (Liu, et al. 2017a). Jing et al. showed similar findings concerning metformin anti-inflammatory role in obese mice and palmitate induced activation of macrophages. In vivo and in vitro studies showed the ability of metformin to reduce the pro-inflammatory cytokines, TNF-α and IL-6, and alteration in macrophages polarization, decreasing the pro-inflammatory M1 macrophage proportion while increasing the anti-inflammatory M2 proportion (Jing, et al. 2018). Likewise, the therapeutic benefits of metformin have been described in endotoxin-induced hepatic injury, which suggests its pharmacological potency in inflammation disorders. Other data supporting the anti-inflammatory effect of metformin were obtained by Koh et al. (2014) to find that metformin suppresses the activation of NFκB in intestinal epithelial cells and amyloid colitis in mice, suggesting that metformin could be used as a potential therapeutic agent for the treatment of inflammatory bowel diseases (Koh, et al. 2014). Cameron et al. in 2016 concluded that the anti-inflammtory activity of metformin is exerted irrespective of the diabetes status, where metformin specifically blunted secretin of proinflammatory cytokines in macrophages without inhibiting M1/M2 activation or differentiation (Cameron, et al. 2016). Vasamsetti et al. showed the ability of metformin to inhibit monocyte-to-macrophage differentiation in THP-1 cells, a human monocytic leukemia cell line, which can be considered a novel mechanism of metformin anti-inflammatory activity. This inhibition was AMPK dependent, ! 69! in a way by reducing STAT3 activity due to AMPK activation, thus reducing the production of proinflammatory cytokines. This inhibition of monocytes-macrophage differentiation by metformin could be a promising treatment in cases of vascular injury, atherosclerosis, cancer, and insulin resistance (Vasamsetti, et al. 2015) (Fig. 20). Figure 20. Anti-inflammatory effect of metformin through macrophage-targeting mechanism (Hattori, et al. 2015). ! 76! III. MATERIALS AND METHODS ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 77! 1. MATERIALS 0.45µm nitrocellulose membrane Bio-Rad Laboratories, Hercules, CA (USA) 30% Acrylamide solution Bio-Rad Laboratories, Hercules, CA (USA) Amonium persulfate Sigma Aldrich, San Luis, MO (USA) Anti- AKT Santa Cruz Biotechnology, Dallas (USA) Anti-arginase Santa Cruz Biotechnology, Dallas (USA) Anti-GAPDH Santa Cruz Biotechnology, Dallas (USA) Anti-GFAP Chemicon International Inc, (USA) Anti-Iba-1 Wako, Osaka, Japan Anti-IKKβ Santa Cruz Biotechnology, Dallas (USA) Anti-JNK Santa Cruz Biotechnology, Dallas (USA) Anti-p38 Santa Cruz Biotechnology, Dallas (USA) Anti-TH Sigma Aldrich, San Luis, MO (USA) ATP Sigma Aldrich, San Luis, MO (USA) Biotinylated anti-IgG Vector (UK) CellROX® Deep Red Reagent Invitrogen, Thermo Fischer Scientific (USA) Clarity™ and Clarity Max™ western ECL Bio-Rad Laboratories, Hercules, CA (USA) DAB Sigma Aldrich, San Luis, MO (USA) DMEM high glucose Invitrogen, Carlsbad, CA (USA) DMSO Sigma Aldrich, San Luis, MO (USA) DPX mountant VWR (USA) ExtraAvidin®-Peroxidase Sigma Aldrich, San Luis, MO (USA) Heat-inactivated fetal bovine serum (FBS) Sigma Aldrich, San Luis, MO (USA) ! 78! Histolemon Carlo Erba®, France Hydrogen peroxide Sigma Aldrich, San Luis, MO (USA) Ibidi® plates Invitrogen, Carlsbad, CA (USA) IL-4 Sigma Aldrich, San Luis, MO (USA) LPS from Escherichia coli serotype 026: B6 Sigma Aldrich, San Luis, MO (USA) Metformin Dianben® 1000mg, MERK, (USA) Monastral Blue inert tracer Sigma Aldrich, San Luis, MO (USA) Mouse IL1β ELISA Ready-SET-GO Invitrogen, Thermo Fischer Scientific (USA) OX-6 Serotec, Oxford, (UK) Paraformaldehyde Carlo Erba®, France PBS Sigma Aldrich, San Luis, MO (USA) Pierce™ BCA protein assay kit Thermo Fischer Scientific (USA) Protease inhibitor Sigma Aldrich, San Luis, MO (USA) Revert Aid First Strand cDNA Synthesis Kit Thermo Fisher Scientific (USA). RIPA buffer Sigma Aldrich, San Luis, MO (USA) RNeasy® kit QIAGEN, Gmbh, Hilden (Germany) SDS Sigma Aldrich, San Luis, MO (USA) SensiFASTTM SYBR NO-ROX KIT Bioline,(USA) Streptomycin/penicillin Biowest, Nuaillé (France) TBS Sigma Aldrich, San Luis, MO (USA) TEMED Bio-Rad Laboratories, Hercules, CA (USA) Tris Sigma Aldrich, San Luis, MO (USA) TRIsure™ Bioline, (USA) ! 79! Trypan blue Sigma Aldrich, San Luis, MO (USA) Trypsin-EDTA Biowest, Nuaillé, France TWEEN 20 Sigma Aldrich, San Luis, MO (USA) Trypan blue Sigma Aldrich, San Luis, MO (USA) DMSO (Sigma Aldrich, San Luis, MO, USA) ! 80! 2. METHODS 1. Cell culture For all cell culture experiments, murine microglial BV2 cell line was used. These cells were cultured in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Sigma-Aldrich, San Luis, MO, USA), streptomycin (100 mg/ml), and penicillin (100 IU/ml) (Biowest,!Nuaillé,!France), under 100% humidity and 5% CO2. Experiments were performed in reduced 5% FBS media.!The medium was changed every 3 days and allowed to grow up to 80% confluence. All experiments were carried out in 3-20 passage cells. The passing and treatment of cells were always carried out in a laminar flow hood.! ! !!!!! 2.1.1. Thawing cells The thawing of the vials to recover the conserved cells was done as fast as possible, according to the following protocol: Vials were placed in a 37°C bath until the cell suspension was thawed. Then the thawed cell suspension was transferred to a centrifuge tube with 9 ml of culture medium preheated to 37 ºC. Cells were centrifuged at low speed (300 xg) for 5 min at 4 ºC, the precipitate resuspended in preheated culture medium and seeded in a flask (Sarstedt AG & Co, Germany). 2.1.2. Passing cells The separation of the crops was carried out when their growth reached a 80% confluence, by the following procedure. Cells were washed with phosphate buffer (PBS) 1X, deattached from the flasks by adding enough 1X trypsin-EDTA (Biowest,!Nuaillé, France) to cover the crop surface, and incubated for 5 min at 37 ºC. To stop the trypsinization, pre-heated culture medium at 37 ºC was added. Then cells were collected, centrifuged at 500 xg, for 5 minutes at room temperature, and the precipitate resuspended in a desired volume and reseeded. To determine the number of living cells, a Neubauer chamber and Trypan blue (Sigma Aldrich, San Luis, MO, USA) were used. Trypan blue is a colloid that is introduced inside the cells that have ruptured membranes, thus dead cells will ! 81! have blue color while white cells are the viable ones. The used procedure was as follows. Cell suspension was homogenized and 10 µL transferred into an eppendorf tube where an equal volume of a 0.4% solution of Trypan blue was added. The compartment of the Neubauer chamber was filled with 10 µL of the suspension and an inverted microscope (Leica DMIL) was used to count the number of cells. The number of cells/ml was calculated using the following formulas: Cells/ml = mean of the number of cells counted in the four grids × 104 × 2 (Dilution factor) Total cells = (Cells / ml) × original volume of cell suspension. 2.1.3. Conservation of cells For the purpose of conserving the cells, the following protocol was followed. 10 ml of a cell suspension was collected and centrifuged at 500 xg for 5 min at room temperature. Then the precipitate was resuspended in 5-10 ml of freezing medium consisting of 10% (w/v) dimethyl sulfoxide (DMSO) (Sigma Aldrich, San Luis, MO, USA), as a cryoprotectant, and 90% DMEM supplemented with 10% FBS. The cell suspension was aliquotted in sterile freezing cryogenic vials. Since DMSO is toxic at room temperature, it was convenient to proceed quickly to the cooling of the vials. However, freezing must be carried out slowly. For this purpose, the vials were put in a MR. Frosty™ (ThermoFischer) freezing box containing isopropanol to allow gradual decrease of temperature and kept at -80 ºC for two days, then moved to liquid nitrogen for long storage. 2. Animals and surgery Male albino Wistar rats (200–270 g) were used for these studies. The rats were kept at constant room temperature of 22 ± 1 ºC and relative humidity of 60% with a 12-h light–dark cycle with free access to food and water. Rats were anaesthetized with isoflurane and positioned in a stereotaxic apparatus (Kopf Instruments, Tujunga, CA, USA) to conform to the brain atlas of Paxinos and Watson (1986). Injections into the SN were made 5.5 mm posterior, 1.8 mm lateral and 8.3 mm ventral to the bregma point at day 1. ! 82! Experiments were carried out in accordance with the Guidelines of the European Union Directive (2010/63/EU) and Spanish regulations (BOE 34/11370-421, 2013) for the use of laboratory animals; the study was approved by the Scientific Committee of the University of Seville. 3. Treatment 2.8.1. In vitro treatment For testing the effect of metformin on the pro- and anti-inflammatory phenotypes of microglial cells, BV2 cells were treated with LPS (pro-inflammatory phenotype inducer, 1 µg/ml, Sigma- Aldrich, San Luis, MO, USA) or IL-4 (anti-inflammatory phenotype inducer, 20 ng/ml, Sigma- Aldrich, San Luis, MO, USA), with and without metformin (1 mM, Dianben® 1000mg, MERK) for 12 h (Fig. 22 A). The effect on production of ROS was measured by treating cells with LPS and/or metformin for 24 h (Fig. 22 B). The effect of metformin on inflammasome activation was tested by Western Blot and ELISA, cells were treated with LPS and/or metformin for 3 h then ATP (1 mM, Sigma-Aldrich, San Luis, MO, USA) was added for 1 h as the second signal for inflammasome activation (Martinon et al., 2002) (Fig. 22C). At least four experiment were done for each technique. 2.8.1. In vivo treatment Animals were divided into four groups according to two different variables (Fig. 23): solution administered orally by a plastic enteral feeding tube (tap water or 150 mg/kg of metformin dissolved in tap water, 4 ml/kg) and product injected in the SN (2 µl of saline solution or 2 µg of LPS dissolved in 2 µl of saline solution). In group I (the control group), animals received a single intranigral injection of saline solution (Monastral Blue inert tracer, 1% in saline solution; Sigma- Aldrich, St Louis, MO, USA) the first day. ! 83! Figure 22. Time course of different in vitro treatments. A. Real-time PCR treatment, B. ROS measurement treatment, C. Inflammasome activity treatment. (LPS W/O with or without metformin) Twice a day, animals received 4 ml/kg of tap water through the plastic tube for seven days. Group II received a single intranigral injection of saline the first day and two daily doses of metformin dissolved in tap water for seven days. Group III received a single intranigral injection of LPS (from LPS/IL4'W/O' Me,ormin' End'of' treatment' 0h''''''''''''''''''''''''''''''''''''''''''''''''12h' LPS'W/O' Me,ormin' ATP' End'of' treatment' 0h''''''''''''''''''''''''''''''''''3h'''''''''''4h' C" A" B" LPS'W/O' Me,ormin' Probe' End'of' treatment' 0h'''''''''''''''''''''''''''''''''24h'''''''''''' 30min' ! 84! E.coli, serotype 026:B6; Sigma Aldrich, San Luis, MO, USA) the first day and 4 ml/kg of tap water through the plastic tube twice a day for seven days. Group IV received a single intranigral injection of LPS the first day and two daily doses of metformin dissolved in tap water for seven days. The first dose of metformin or water was administrated 1 h before intranigral injection. All animals were sacrificed seven days after the initiation of treatment, unless otherwise stated. At least four animals were used for each group. Figure 23. Schematic representation of the experimental design: groups and treatments in vivo. Tap water or metformin were administrated orally, twice a day for 7 days to the corresponding groups. Saline solution or LPS were injected intranigrally to the corresponding groups the first day of the experiment, before tap water or metformin were administered. ! 85! 4. Immunohistological evaluation: Tyrosine hydroxylase (TH), glial fibrillary acidic protein (GFAP), and OX-6 Animals used for immunohistochemistry completed 7 days of treatment with water/metformin. One week after the surgical procedure in the SN, animals were perfused through the heart under deep anesthesia (isofluorane) with 150-200 ml of 4% paraformaldehyde in phosphate buffer, pH 7.4. The brains were removed and then cryoprotected serially in sucrose dissolved in phosphate-buffered saline (PBS), pH 7.4, first (24 h) in 10% sucrose, next (24 h) in 20% sucrose, and finally in 30% sucrose until they sank (2–5 days). Brains were then frozen in isopentane at -20 ºC, and sections of 20 µm thickness for each were cut on a cryostat at -20 ºC, and mounted in gelatine-coated slides. Primary antibodies used were rabbit-derived anti-tyrosine hydroxylase (anti-TH, Sigma, 1:300), mouse-derived antiglial fibrillary acidic protein (anti-GFAP; Chemicon International Inc, USA.; 1: 300) and mouse-derived OX-6 (Serotec, Oxford, UK; 1: 200). OX-6 is directed against a monomorphic determinant of the rat MHCII antigens, expressed by activated microglia but not for the resting cells. Incubations and washes for all the antibodies were in Tris-buffered saline (TBS), pH 7.4. All work was performed at room temperature. Sections were washed and then treated with 0.3% hydrogen peroxide in methanol for 15 min, washed again, and incubated in a solution containing TBS and 1% horse (for GFAP and OX-6) or goat (for TH) serum for 60 min in a humid chamber. Slides were drained and further incubated with the primary antibody in TBS containing 1% horse serum/goat and 0.25% Triton-X-100 for 24 h. Sections were then incubated for 2 h with biotinylated horse anti-mouse (for GFAP and OX-6) or biotinylated goat anti-rabbit (for TH) IgG (Vector, 1:200). The secondary antibody was diluted in TBS containing 0.25% Triton-X-100, and its addition was preceded by three 10-min rinses in TBS. Sections were then incubated with ExtrAvidin®-Peroxidase solution (Sigma Aldrich, San Luis, MO, USA, 1:100) for 1h. The peroxidase was visualized with a standard diaminobenzidine/hydrogen reaction for 5 min. ! 92! USA) onto 0.45 µm nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA, USA) for 7 min. After blocking with 5% skimmed milk diluted in TBS-Tween (TBST) (0.05% Tween 20 in TBS) for 1 h, membranes were incubated with a primary antibody against phosphorylated and nonphosphorylated forms of the MAPKs family proteins P38 and JNK (Santa Cruz Biotechnology) for in vivo experiments and against IL-1β (R&D Systems, Minneapolis, USA) for in vitro experiment. Antibodies of non-phosphorylated forms of MAPKs and the housekeeping GAPDH were used at a dilution of 1:1000 in 5% skimmed milk dissolved in TBST, while the antibodies of the phosphorylated forms of MAPKs and IL-1β were used at a dilution of 1:500 in 5% skimmed milk dissolved in TBST. Membranes were incubated with these antibodies for 24 h at 4 ºC. Membranes were washed 4 times (10 min each) with TBST, then incubated for 1 h with the secondary antibody (1:5000) in 5% skimmed milk dissolved in TBST against the specific primary antibody at room temperature. Another 4 washes with TBST (10 min each) were done before visualizing the bands with Clarity™ and Clarity Max™ western ECL Blotting Substrates (Bio-rad, Laboratories, Hercules, CA, USA) using Amersham Imager 600 Imagers (general electric, USA). These images were later analyzed using the ImageJ software. 9. Production of reactive oxygen species The level of oxidative stress produced in viable BV2 cell cultures was determined by seeding 35 × 103 cells/well in Ibidi® plates (Invitrogen, Carlsbad, CA, USA), followed by incubation for 24 h at 37 ºC. Then, cells were treated with metformin (1 mM) and/or LPS (1 µg/ml) for 24 hours At the end of the treatment a membrane permeable reagent (CellROX® Deep Red Reagent) was added to each well. This reagent is a fluorogenic probe that exhibits fluorescence signal in the presence of ROS in the cell cytoplasm. ! 93! The test was performed following the manufacturer’s instructions. The reactive was added to the treated cells at a final concentration of 5 mM, and incubated at 37 ºC for 30 min. Subsequently, the media was removed and cells were washed with PBS three times. Confocal microscope (Zeiss LSM 7 DUO) was employed to acquire images of the treated cells. Images of five fields per condition were captured. These images were later analyzed using the ImageJ software. In each image the total number of cells was calculated, and the area integrated intensity for fluorescent and non-fluorescent cells was measured. Consequently, the following formula was used to compute the corrected total cell fluorescence (CTCF). CTCF = Integrated density – (Area of selected cell × Mean fluorescence of background readings) 10. Enzyme-linked immunosorbent assay BV2 cell cultures were seeded in 24-well plates at a concentration of 75 × 103 cells/well. Then, cells were treated with metformin (1 mM) and/or LPS (1 µg/ml) and incubated at 37 ºC for 3 hours. Afterwards, ATP (1mM) was added to each for 1 h. At the end of the treatment, the supernatant was collected and centrifuged at 400xg , 5 min, 4 ºC to remove cell debris. After centrifugation, the supernatant was stored at -80 ºC for subsequent use. 100 µl of the supernatant was used for the measurement of IL-1β production by enzyme-linked immunosorbent assay (ELISA) following the manufacturer’s instructions (Invitrogen, Mouse IL1β ELISA Ready-SET-GO! Thermo Fisher Scientific). All standards and samples were run in duplicates. Samples were diluted in a proportion of 1:1 in ELISA/ELISASPOT Diluent (1×), except LPS containing samples, which were diluted in a proportion of 1:2. 11. Statistical analysis Results are expressed as mean ± SD. Means were compared by One-Way ANOVA followed by the LSD test for post hoc multiple range comparisons. An alpha level of 0.05 was used. The Statgraphics Plus 3.0 statistical package was used for the analyses. ! 94! IV. RESULTS ! 95! 1. IN-VITRO EXPERIMENTS 1.1. Metformin treatment reduces microglial activation in vitro Activated microglia can be categorized into the classically pro-inflammatory and neurotoxic M1 phenotype, and the alternative activated anti-inflammatory M2 phenotype involved in the resolution of inflammation, phagocytosis and tissue repair. Depending on the inductor of their activation, when microglia are activated they produce different cytokines and pro or anti-inflammatory substances with different actions. When LPS activates microglia, these cells produce proinflammatory cytokines such as IL-6, TNFα and IL-1β, and oxidative metabolites like nitric oxide; while IL-4 treatment induces microglia activation toward M2 phenotype, expressing lower levels of these previously mentioned markers, but higher levels of IL-10, IL-4, and arginase (Fig. 24) (Kim, et al. 2016). Figure 24. Microglia phenotypes. Activated microglia has two activation phenotypes: classically activated (M1) and alternatively activated (M2). Taken from(Liu, et al. 2017b). ! 96! In this work, BV2 cells were treated with LPS to induce a M1 phenotype, or with IL-4 to induce a M2 phenotype. When BV2 cells were treated with LPS, RTPCR analysis showed a strong induction of the four pro-inflammatory mediators studied, ranging from 348% for TNF-α to 1438% for IL-1β with respect to the control (p <0.001; Fig. 25). Metformin treatment reduced the expression levels of the mRNA of IL-1β (56 ± 11.5% with respect to the LPS; p <0.001; Fig. 25A). No statistical difference was found in the amounts of iNOS, TNF-α, and CXCL10 mRNAs when metformin was added to LPS-treated cells. When BV2 cells were treated with IL-4 a strong induction of arginase was observed (20000-fold compared to the control group, p <0.001; Fig. 25D). This induction was decreased by 37.7% by metformin treatment (p <0.001; Fig. 25E). However no statistical difference was found concerning the levels IL-10 mRNA when metformin was added to IL-4 treated cells (Fig. 25F). ! 97! Figure 25. Quantification of mRNA expression levels of different pro and anti-inflammatory markers in BV2 cell-line. BV2 cells were treated with LPS or IL-4 for 12 hours. Total RNA was extracted from the cells to measure the change in the expression levels of mRNA of (A) IL-1β, (B) iNOS, (C) TNFα, and (D) CXCL10, when the used treatment was LPS; (E) Arginase and (F) IL-10, when IL-4 was the treatment. Results are mean ± SD of at least 3 independent experiments, normalized to β-actin and expressed as percentage relative to the LPS/IL4 group. Statistical significance (one-way ANOVA followed by the LSD post hoc test for multiple comparisons): a, Arginase (% of control) 0 20 40 60 80 100 120 IL-10 (% of control) 0 50 100 150 200 TNF-α (% of control) 0 20 40 60 80 100 120 iNOS (% of control) 0 20 40 60 80 100 120 IL-1 beta (% of control) 0 20 40 60 80 100 120 A B ab a a ab a C aa a C Metf LPSMetf LPS C Metf LPSMetf LPS C Metf LPSMetf LPS C Metf LPS LPSMetf D a!a E F C Metf LPS LPSMetf C Metf LPS LPSMetf! ! 98! compared with the control group; b, compared with the LPS/IL4 group; p<0.001. (C, control cells without treatment; Metf, cells treated with metformin; LPS, cells treated with LPS; LPSMetf, cells treated with LPS and metformin). 1.2. Metformin treatment prevented the induction of NADPH oxidase by LPS In microglial cells, ROS are mainly derived from NADPH oxidase (Block 2008; Gao, et al. 2012). NAPDH oxidase have seven described homologs which differ in their structure, expression levels in different tissues, and their activation mechanisms. NAPDH oxidase is a multi-subunit enzyme complex, consists of the membrane bound cytochrome b558 (p22phox and the heme binding enzymatic subunit gp91phox), cytosolic subunits (p40phox, p47phox and p67phox) and the Rac G-protein (Block 2008; Dohi, et al. 2010). During stimulation by a pathogenic trigger, the cytosolic subunits and Rac associate with the membrane-bound subunits resulting in activation of the complex (Block 2008; Gao, et al. 2012). Therefore, to study the effect of metformin on the oxidative stress induced by LPS treatment, the expression levels of the p22phox, p47phox and gp91phox subunits of the NADPH oxidase enzyme were measured using RT-PCR. In our experimental conditions LPS treatment increased the expression levels of p22phox, gp91phox and p47phox subunits (251.1± 74.2, 249.7 ± 96.4 and 524.8 ± 222.7 % respect to the control respectively, Fig 26A, B and C, p < 0.05), whereas metformin reduced this elevation in the expression level of p22phox, and gp91phox subunits caused by LPS treatment (173.2 ± 21.7 and 308.7 ± 136.9, p < 0.05, Fig 26A and C), no effect was observed in the expression level of p47phox subunit upon metformin treatment (Fig. 26B). These results indicate that metformin inhibited NADPH oxidative activation, which mediate the production of ROS, which is involved in microglia activation by LPS and the resultant inflammatory process. ROS is produced by NAPDH-dependent one-electron reduction of oxygen to superoxide, and its production is central to many inflammatory diseases, acting as both signalling molecule and inflammatory mediator. ! 99! Figure 26. Effect of Metformin on the expression levels of NADPH oxidase subunits. Expression levels of (A) p22phox, (B) p91phox and (C) p47phox mRNAs were measured by RT-PCR in BV2 cells from the different treatments assayed (C, control cells without treatment; Metf, cells treated with metformin; LPS, cells treated with LPS; LPSMetf, cells treated with LPS and metformin). Results are mean ± SD of at least 3 independent experiments, normalized to β-actin and expressed as percentage relative to the control group. Statistical significance (one-way ANOVA followed by the LSD post hoc test for multiple comparisons): a, compared with the control group; p<0.05. gp91phox (% of control) 0 200 400 600 800 p22phox (% of control) 0 50 100 150 200 250 300 350 a a C Metf LPSMetf LPS A B C Metf LPSMetf LPS p47phox (% of control) 0 100 200 300 400 C C Metf LPS LPSMetf a a! ! 100! 1.3. Metformin treatment decreases the production of ROS Oxidative stress has an important role in the degeneration of the DAergic neurons occurring in PD. During oxidative stress there is a state of disequilibrium between the levels of produced ROS and their clearance. Activated microglia are source of superoxide and nitric oxide, which contribute to the oxidative stress in the brain. Neurons are sensitive to oxidative stress due to their membrane composition of high polyunsaturated fatty acid content, high oxygen consumption and weak antioxidant defense (Liu, et al. 2017c). Taking in consideration that the main source of ROS in microglial cells is the NADPH oxidase as well as metformin effect on the expression of this enzyme, we next sought to measure the level of oxidative stress in BV2 cells after LPS treatment and the effect of metformin on ROS production after LPS treatment. Using a fluorogenic probe that exhibits fluorescence signal in the presence of ROS in the cell cytoplasm, a two-fold increase of ROS production in BV2 cells after treatment with LPS (197.1 ± 31.6% with respect to the control, p < 0.001; Fig 27C and E) was observed. Whereas metformin treatment decreased ROS production in LPS-treated BV2 cells by approximately 30% (1388 ± 9.7% respect to the control, p < 0.001; Fig 27D and E). ! 101! Figure 27. Effect of LPS and metformin on ROS production in BV2 cells. ROS were measured 24 h after the different treatments assayed: (A) Control; (B) Metformin; (C) LPS; (D) LPS plus metformin. (E) Quantification of ROS production. ROS production was measured after adding CellROX flurogenic probe. Images of different conditions were captured using confocal microscope. Results are mean ± SD of at least 3 independent experiments, expressed as percentage of control. Statistical significance (one-way ANOVA followed by the LSD post hoc test for multiple comparisons): a, compared with the control group; b, compared with the LPS group; p<0.001. Scale bar, 50 µm. ROS production (% of control) 0 50 100 150 200 250 C LPSMetf Metf LPS a ab A B C D CMetf LPSMetf LPS E ! 108! pro-inflammatory (IKKβ) or an anti-inflammatory (arginase) marker was performed. Iba-1, is a calcium binding protein that is specifically expressed in microglia and is up regulated during microglia activation. LPS injection is known to evoke M1 microglial activation, thus after LPS treatment most of microglial cells were co-localized with IKKβ (Fig.30A-C), whereas only scare co-localization appeared with arginase (Fig. 30G-I). Thus microglial cells activated with LPS displayed a proinflammatory phenotype. When animals were treated orally with metformin, most microglial cells expressed IKKB (Fig. 30J-L), whereas no co-localization appeared with arginase (Fig. 30J-L). Therefore, LPS also induces a pro-inflammatory phenotype even in metformin-treated animals. Arginase (panels H and K) is not induced in Iba-1-labelled microglial cells, except for a few ones (G, H, J and K; merge panels I and L) in the animals treated with LPS and LPsMetf. Images in the inserts in panels C, F, I and L are high-magnification photographs of the small white boxes in their respective panels. Scale bars: A-L, 200 µm; insert, 25 µm. ! 109! Figure 30. Effect of metformin and LPS on Iba-1, IKKβ and arginase immunostaining in the SN after LPS (panels A-C and G-I) and LPSMetf (panels D-F and J-L) treatments. Iba-1 immunofluorescence in panels A and D, and IKKβ immunofluorescence in panels B and E show an induction of IKKβ in Iba-1-labelled microglial cells (merge panels C and F). ! 110! 2.4. Metformin treatment decreases the expression levels of several inflammatory mediators after LPS injection in vivo Activated microglia produces pro-inflammatory cytokines and several anti-inflammatory mediators. To examine metformin effect on these different cytokines, mRNA levels of pro-inflammatory cytokines (TNF-α, IL-1β, IL-6 and iNOS) in the SN were measured after treatment as described above. Beside the afore mentioned cytokines the expression level of CD200, CX3CR1, and MCP1 were quantified. Analysis by real-time PCR showed that the mRNAs for the pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 were induced in the SN up to 6 hours after LPS injection (496.3% ± 53 and 243, 3 ± 96% compared to the control group, respectively; p <.0.01 for TNF-α and IL-1β; Fig.318A, B, C). Moreover, LPS treatment induced the expression of MCP-1 (369 ± 87% with respect to the control; p <0.001; Fig. 31E). CD200 and CX3CR1 have shown to have inhibitory actions on brain microglia. Hence, we decided to study these molecules in our experimental conditions to seek further explanations of how metformin induces an attenuated effect after LPS treatment. Our PCR analysis showed that the expression of CD200 and CX3CR1 mRNAs decrease in respect to the control group (40 ± 2.45% and 71 ± 12.7%, respectively; p <0.001 for CD200 and p <0.05 for CX3CR1; Fig. 31F and G). Metformin significantly reduced the LPS-induced increase in the mRNA levels of inflammatory markers compared to the control group (322.94 ± 67.6%; 38.1 ± 13.5% and 122.47 ± 18.76 for TNF-α, IL-1β and IL-6 of control values, respectively; p <0.01; Fig. 31A, B and C). On the contrary, metformin treatment increased the levels of MCP-1, CD200 and CX3CR1 mRNAs (619.7 ± 59.2%, p <0.001; 68 ± 10.50%, p <0.001; and 110 ± 12.09% of the control values, p <0.05, respectively; Fig. 31E, F and G). ! 111! Figure 31. Effect of metformin and LPS on the expression of TNF-α (A), IL-1β (B), IL-6 (C), iNOS (D), MCP-1 (E), CD200 (F) and CX3CR1 (G) mRNAs in the SN of rats from the different treatments assayed, measured by RT-PCR. Sal, animals injected with saline solution in the SN plus oral administration of water; Metf, animals injected with saline solution in the SN plus CX3CR1 (% of control) 0 20 40 60 80 100 120 140 MCP-1 (% of control) 0 200 400 600 800 CD200 (% of control) 0 20 40 60 80 100 120 %α )( 0 100 200 300 400 500 600 a" ac" ab" IL-6 (% of control) 0 50 100 150 200 250 ab" c" iNOS (% of control) 0 50 100 150 200 250 %β )( 0 100 200 300 400 ab" Sal" SalMe)" LPSMe)" LPS" ab" ac" ab" c" ab" abc" c" A B C E G D F ! 112! oral administration of metformin; LPS, animals injected with LPS in the SN plus oral administration of water; LPSMetf, animals injected with LPS in the SN plus oral administration of metformin. Results are mean ± SD of at least four independents experiments, normalized to β-actin and expressed as percentage relative to the control (Sal) group. Statistical significance (one-way ANOVA followed by the LSD post hoc test for multiple comparisons): a, compared with the Sal group; b, compared with the Sal/Metf group; c, compared with the LPS group; p < 0.01. 2.5. Metformin effects on microglial activation are mediated by MAPK The MAPK signaling pathway plays a vital role in microglial activation (Zhao, et al. 2017) Metformin is known as an activator of AMPK, where AMPK is involved as an upstream element of MAPKs signaling pathways in different cell types (Wu, et al. 2011). Thus, the involvement of JNK, p38 in the anti-inflammatory role of metformin was examined by Western blot analysis. Specific antibodies against the phosphorylated and non-phosphorylated forms of JNK and p38 were used. As expected, LPS treatment increased the ratio P-JNK/JNK (183.68 ± 42.82%; p<0.001 compared with control animals; Fig. 32A), whereas metformin treatment abolished this effect (39.93 ± 21.98%). Metformin treatment also reduced the ratio P-p38/p38 (57.89 ± 29.16%; p < 0.05 compared with control animals; Fig. 32B). ! 113! ! Figure 32. Effect of metformin on the expression of JNK, P-JNK, p38, and P-p38 proteins in the SN. Proteins from the SN of rats under the different treatments assayed were separated by electrophoresis, transferred to nitrocellulose membranes and stained using anti-JNK, anti-P-JNK, anti-p38 and anti-P-p38 antibodies. Total optical density of each band was calculated. Results are mean ± SD of at least 4 independent experiments, expressed as P-JNK/JNK (A) and P-p38/p38 (B) intensity ratios (normalized to GAPDH expression) relative to the control (Sal) group. Statistical significance (one-way ANOVA followed by the LSD post hoc test for multiple comparisons): a, compared with the control group; b, compared with the LPS group; p < 0.01. Bands corresponding to the Sal and SalMetf treatments can be seen in Figure 33. (Sal, animals injected with saline solution in the SN plus oral administration of water; Metf, animals injected with saline solution in the SN plus oral administration of metformin; LPS, animals injected with LPS in the SN plus oral administration of water; LPSMetf, animals injected with LPS in the SN plus oral administration of metformin). ! 114! Figure 33. Effect of metformin on the expression of JNK, P-JNK, p38, and P-p38 proteins in the SN. Bands corresponding to the Sal and SalMetf treatments 2.6. Effect of metformin and LPS on the dopaminergic system Herrera et al. reported that a single intranigral injection of LPS results in microglial activation and strong inflammatory response in SN causing a specific damage to the DAergic neurons without ! 115! affecting other neuronal types. These findings suggest that microglial activation may cause a decrease in TH immunoreactivity (Herrera, et al. 2000). Same findings were published by Gao et al in 2002 (Gao, et al. 2002). Therefore, the integrity of the nigro-striatal DAergic neurons was evaluated at the cellular level by immunohistochemistry against TH, which is the rate-limiting enzyme in the synthesis of DA and widely used as an index of DAergic neurons. Oral administration of metformin (Fig. 34B and E) had no effect on the number of DAergic neurons with respect to the control group (2882±685 and 2294±530 for control and metformin, respectively; Fig.34 A, B and E). Intranigral administration of LPS reduced the number of neurons to 66.1% of control values (Fig 34C and E; p <0.01). Interestingly, metformin not only failed to protect against the LPS-induced death of DAergic neurons, but even exacerbated the damage (29.6% of control value; Fig 34D and E). ! 116! Figure 34. Effect of metformin and LPS on dopaminergic neurons. (A) Coronal section showing that TH immunreactivity was not affected by the injection of saline solution in the SN of rats . (B) Coronal section showing the TH immunreactivity was not affected after vehicle injection in the SN of rats receiving metformin treatment twice daily. (C) Coronal section showing a decrease in TH immunoreactivity around the injection area after the injection of 2 µg of LPS in comparison with the control group. (D) Coronal section showing that the loss of TH immunoreactivity induced by LPS was increased in the animals treated orally with metformin. Scale bar: 500 µm. (E) Quantification of the number of TH-positive cells of the four treatment groups. Results are expressed as mean ± SD of 4 independent experiments, expressed as TH-positive cells within the bounded area of the SN. Statistical significance (one-way ANOVA followed by the LSD post hoc test for multiple comparisons): a, compared with the control group; b, compared with the LPS group; p < 0.01. ! 117! V. DISCUSSION % % % % ! 124! studies explore the possible protective effect of metformin on DAergic neurons after treatment with MPTP (Bayliss, et al. 2016; Ismaiel, et al. 2016; Lu, et al. 2016b; Patil, et al. 2014a). Here, we study for the first time the effect of metformin on the in vivo model of PD based on the injection of LPS within the SN. We have chosen this model for several reasons: i) it is a well-established and extensively used model of PD; ii) inflammation is the main mechanism of cell death involved; iii) the effect of metformin on this model of PD is not well documented in the literature. Therefore, we firstly measured microglial activation after the injection of LPS within the SN. OX-6 immunostaining and PCR revealed that metformin decreased the number of activated microglial cells and the expression levels of several pro-inflammatory cytokines and mediators. In this sense, metformin treatment decreased the mRNA levels of TNF-α, IL-6 and especially IL-1β. These results corroborate the in vitro results, showing that IL-1β is especially downregulated by metformin. To shed more light into this issue, we also studied the mRNA expression of CD200, a surface molecule expressed on the microglial cell membrane that keeps the microglia in a quiescent state when it interacts with its receptor CD200R, and the mRNA expression of CX3CR1, also known as fractalkine receptor, which has been associated with inhibitory actions on brain microglia. Microglia-deficient of CX3CR (Kierdorf and Prinz 2013), as well as MCP-1 (a potent chemokine involved in different neurological disorders) (Conductier, et al. 2010), has an over-activated phenotype. As expected, intranigral LPS injection led to a significant down-regulation of CD200 and CX3CR1 mRNA expression in the ventral mesencephalon. However, metformin treatment turned the down-regulation of CD200 and CX3CR1 expression induced by LPS into a significant up-regulation of these microglia regulators, thus pointing up metformin as a molecule responsible for inactivating microglia. This finiding suggest a beneficial role of metformin treatment in neurodegenerative disease. ! 125! Inflammatory processes start by the attraction of certain immune cells to the site of inflammation. Chemokines are the main cells responsible for creating a concentration gradient inducing inflammatory cells from the circulation to the injury site. MCP-1 is one of the main chemokines which regulates microglia movement, and the most potent at activating signal transduction pathways leading to monocyte transmigration (Sozzani, et al. 1994). On the contrary to several authors have demonstrated that metformin can damp the induction of this chemokine by LPS (Chen, et al. 2018b; Han, et al. 2018), our results showed that metformin treatment increased the mRNA levels of MCP-1 in the SN in our experimental conditions. These findings suggest that metformin may not a protective effect against MCP-1. As stated above, microglial cells can adopt different phenotypes, named traditionally M1 (a classical pro-inflammatory phenotype), an initial defense mechanism characterized by the production of inflammatory cytokines and reactive oxygen species, and M2 (an alternative antiinflammatory phenotype) that has a repairing role after the inflammation. Several studies have proposed that metformin could exert its anti-inflammatory actions inducing a shift in microglia towards a M2 phenotype. This modulation has been suggested as a novel promising therapeutic strategy in neurodegenerative diseases (Pena-Altamira, et al. 2017) Although our PCR data from BV2 cells have indicated that metformin decreased both the M1 and M2 phenotypes in vitro, we sought to corroborate these results in vivo. We took advantage of colocalization techniques and found that almost all Iba-1 microglial cells (a general microglia marker) co-localized with the M1 marker IKKβ, whereas none of them co-localized with the M2 marker arginase in both the LPS and LPS+metformin groups. Hence, our results showed again that metformin treatment decreased both the classical and alternative activation forms of microglia in our experimental conditions. ! 126! One of the possible mechanisms by which metformin could decrease inflammatory mediators is related to the activation of MAPKs and the PI3/Akt pathways. The mitogen-activated protein kinases (MAPKs) in mammals include c-Jun NH2-terminal kinase (JNK), p38 MAPK, and extracellular signal-regulated kinase (ERK). These enzymes are serine-threonine protein kinases that regulate various cellular activities including proliferation, differentiation, apoptosis or survival, inflammation, and innate immunity. The compromised MAPK signaling pathways contribute to the pathology of diverse human diseases including cancer and neurodegenerative disorders such as AD, PD, and amyotrophic lateral sclerosis (Kim and Choi 2015). Therefore, in a next step we analyzed the phosphorylated and total amounts of JNK and p38 MAPK and found that metformin was able to damp the increase of the activated state of this proteins induced by LPS, metformin suppressed the phosphorylation of the MAP kinases studied here, p38 and JNK. These results in microglial cells agree with those found in models such as the human monocytic leukemia cell line THP-1 (Chen, et al. 2018b), the rat cardiac myoblast H9c2 cells (Chen, et al. 2018a), and colitis mice models. The PI3K/AKT pathway also regulates key proteins such as mTOR, FOXO and BAD, leading to multiple biological effects (Manning and Cantley 2007). Therefore, we also measured the levels of AKT and found that, as expected, LPS induced a high decrease in the levels of this protein, whereas metformin treatment increased it. PI3K phosphorylation has an important function as an upstream activator of Akt. PI3K/Akt pathway plays a crucial role in the regulation of inflammatory mediators, in inflammatory cell recruitment, and in immune cell function (Hawkins and Stephens 2015). In accordance with our results, Liu et al., (2018) (Guimei, et al. 2018) has shown that metformin is able to decrease levels of IL-6 and TNF-α and increase Akt/PI3K expression in High-Fat Diet/Streptozotocin-Induced Type 2 Diabetes in Rats. These authors suggested that Akt/PI3K pathway possibly regulate the effects of metformin on inflammatory cytokines. Moreover, Tang et al., (2017) (Tang, et al. 2017) showed that ! 127! metformin ameliorates sepsis-induced brain injury by inhibiting apoptosis, oxidative stress and neuroinflammation via the PI3K/Akt signaling pathway. In brief, metformin was able to prevent, totally or partially, most of the alterations induced by LPS. Thus, microglial activation (to both M1 and M2 phenotypes), expression of NADPH, production of ROS, activation of the inflammasome or production of pro-inflammatory cytokines induced by LPS were reduced or avoided by metformin. These anti-inflammatory actions of metformin seem to be mediated by the activation of the AMPK signaling, an energy sensing which mediates downregulation of the NF-κB pathway. Since metformin is an inhibitor of complex I of the mitochondrial respiratory chain, it increases the AMP/ATP or ADP/ATP ratios, which activates AMPK after binding of AMP or ADP (El-Mir, et al. 2000; Owen, et al. 2000) with the subsequent downregulation of the NF-κB pathway. The results discussed so far support the use of metformin to prevent the inflammation associated with PD. However, metformin has also shown negative aspects that must be carefully considered before its possible use in this disease. Thus, metformin does not reduce the initial damage induced by LPS to astrocytes, as demonstrated by the loss of GFAP staining at the injection site of the endotoxin. The loss of GFAP staining in the SN after the injection of LPS is characteristic in this model. This loss was avoided by the treatment with minocycline (Tomas-Camardiel, et al. 2004), but not with dexamethasone (Castaño, et al. 2002a) which seems to suggest the existence of more than one mechanism able to induce the disappearance of astrocytes after exposure to LPS. The early loss of astrocytes is followed by an astrogliosis that repopulates the area initially affected (Herrera, et al. 2000). This transitory disappearance of astrocytes alters the BBB in the SN, perhaps facilitating the arrival from the blood of elements that can contribute to damaging this structure. Furthermore, the loss of dopaminergic neurons from the SN is a key feature of PD. In previous works, our group demonstrated that the injection of LPS in this structure, besides provoking a ! 128! strong pro-inflammatory response of the microglia, it induces the specific death of nigral dopaminergic neurons (Castano, et al. 1998; Herrera, et al. 2000) and that this harmful effect could be avoided with the use of several anti-inflammatory substances such as dexamethasone (Castaño, et al. 2002b) and of other substances not categorized as anti-inflammatory compounds nevertheless exerted an anti-inflammatory effect, such as minocycline (Tomás-Camardiel et al., 2004) or simvastatin (Hernandez-Romero, et al. 2008). Metformin, however, despite the aforementioned anti-inflammatory effects, fails to prevent the death of the dopaminergic neurons induced by LPS. Interestingly, the protective effect of dexamethasone against an inflammatory process induced in the SN is not general, but depends on the inflammatory agent used; thus, in spite of avoiding the dopaminergic damage induced by LPS, dexamethasone did not prevent the death of the dopaminergic neurons of the SN induced by the injection of thrombin into the striatum (Herrera, et al. 2008). In this case, dexamethasone not only failed to prevent the damage induced by thrombin but increased it by inducing the expression of the MAO B enzyme, which acts in the catabolism of DA producing DOPAC and H2O2. In consistence with the results presented here, in a previous work (Ismaiel, et al. 2016), it was shown that metformin, in spite of exhibiting strong in vivo anti-inflammatory properties, was unable to protect the damage induced by MPTP to the nigrostriatal dopaminergic system in mice; in fact, metformin reinforced the loss of DA and its metabolite DOPAC induced by MPTP. Other authors, however, have described a protective effect of metformin against MPTP on dopaminergic neurons in mice, which is supported by reduced brain oxidative stress, and TH-positive dopaminergic neurons (Lu, et al. 2016b; Patil, et al. 2014b). This discrepancy is possibly due to experimental differences in the dosage and MPTP exposure. Since MPTP and metformin are both inhibitors of complex I of the respiratory chain, they could be competing for their access to the complex and have some mutual influence on mitochondrial function. It can be postulated that in those previous studies metformin was not able to restore damaged neurons but reduced the harmful effects of ! 129! MPTP. In this sense, the use of metformin in other animal models of PD, such as the LPS model used in the present work, would allow a better study of its potential usefulness. If metformin is not able to avoid the dopaminergic damage induced by LPS in this experimental design, despite the anti-inflammatory effects demonstrated, it may be because it is exerting some harmful effect on the dopaminergic neurons independent of the inflammatory processes induced by the single injection of LPS. It is known that metformin is a weak inhibitor of complex I of the mitochondrial respiratory chain (El-Mir, et al. 2000; Owen, et al. 2000) and complex I deficiency has been long associated with mitochondrial dysfunction and PD risk (Schapira and Jenner 2011); Thus, despite its anti-inflammatory effects, metformin could be producing an alteration in the ability of mitochondria to produce the enormous amount of energy necessary for the functioning of the dopaminergic neurons, inducing a mitochondrial oxidant stress that could be fatal in the long term. The question that arises is why metformin should produce such a specific effect on a particular neuronal phenotype if it is administered orally. The key to explain this specific effect could be in the complex axonal arborization of these neurons (Hunn, et al. 2015). Thus, dopaminergic neurons seem to have a number of synapses and an unmyelinated axonal arborization that would greatly exceed those of other neuronal phenotypes (Pissadaki and Bolam 2013) making them very sensitive to any energy imbalance that they may suffer (Bolam and Pissadaki 2012; Surmeier, et al. 2017), regardless of how small it might be. Thus, treatment with metformin, despite the benefit of its anti-inflammatory action, could be producing an alteration in the ability of mitochondria to produce the enormous amount of energy necessary for the functioning of these neurons, inducing a mitochondrial oxidant stress that could be fatal in the long term. Another important factor that should be taken into consideration is the dose being used of metformin. Isoda et al. 2006 showed that the ability of metformin to inhibit proinflammatory responses in human vascular wall cells depends largely on the dose being used. ! 130! Should we use metformin for the prevention and treatment of Parkinson’s disease? Over the last 40 years, metformin has revolutionized the treatment of type-2 diabetes worldwide and is still the most influential oral antidiabetic drug today. The key role of insulin in nutrient sensing which balances growth and proliferation with life extending conservation makes metformin an interesting drug. Type-2 diabetes is a disease with high prevalence nowadays and seems to be a risk factor for some neurodegenerative diseases such as AD (Cooper, et al. 2015; Monette, et al. 2014; Palta, et al. 2014; Vincent and Hall 2015) and PD (D'Amelio, et al. 2009; Driver, et al. 2008; Gang, et al. 2007; Miyake, et al. 2010; Schernhammer, et al. 2011; Simon, et al. 2007; Xu, et al. 2011). Until now, only a few studies have assessed the effect of metformin on this neurodegenerative disease and the results are conflicting. While some of these studies have shown beneficial effects on AD (Allard, et al. 2016; Lennox, et al. 2014; Pintana, et al. 2012), and PD (Katila, et al. 2017; Lu, et al. 2016a; Patil, et al. 2014b) others, including ours, show harmful effects or no effect (Chen, et al. 2016a; Ismaiel, et al. 2016; Kuan, et al. 2017; Thangthaeng, et al. 2017). Trials using metformin to treat or protect against neurodegenerative diseases in humans and animals have also produced mostly conflicting results. Therefore, the data shows either positive, no or even detrimental effects of metformin on neurodegenerative processes in cell cultures, animals and humans. These contradictory results may be due to differences in the experimental design and to the multiple biological pathways in which metformin exerts influence. Precisely because of this complexity, it is possible that this compound cannot be used as a common treatment for different neurodegenerative disorders. In fact, it is possible that the patient's risk profile determines the type of effect produced by metformin (Wang, et al. 2017). Therefore, it will be necessary to assess very carefully the use of metformin for the treatment of diabetes in patients who may suffer from Alzheimer's or Parkinson's diseases. If metformin was to be used in the treatment of this type of disease, it would first be necessary to study in depth the most appropriate dosage in relation to the type of neurodegenerative ! 131! process that takes place. One of the major contributions of metformin is that it can help to understand the complex molecular mechanisms that underlie neurodegenerative processes, which would allow applying more specific treatments. ! 132! VI. CONCLUSION ! 133! 1- Mteformin treatment is able to decrease the immune associated alterations in in vitro and in in vivo PD model based on the intranigral injection of LPS. 2- Metformin attenuates the oxidative stress accompanying neuroinflammation in PD. 3- Inflammasome activity is reduced upon metformin treatment, which suggests that it could be a proposed route for metformin anti-inflammatory activity. 4- Although metformin treatment is able to reduce the pro and anti-inflammatory forms of activated microglia, it could not protect against LPS induced damage to the astrocytes. 5- It was not able to protect against the death of DAergic neurons of the SN. On the contrary, Metformin exacerbated the damage. 6- We can conclude that the use of metformin treatment in diabetic patients should be done with caution due to the increased susceptibility of developing PD. 7- Metformin can be considered a risk factor for developing PD in diabetic patients. ! 140! Calcutt, N. A., M. E. Cooper, T. S. Kern, and A. M. Schmidt (2009). Therapies for hyperglycaemiainduced diabetic complications: from animal models to clinical trials. Nat Rev Drug Discov. 8(5):417-29. Cameron, A. R., V. L. Morrison, D. Levin, M. Mohan, C. Forteath, C. Beall, A. D. McNeilly, D. J. Balfour, T. Savinko, A. K. Wong, B. Viollet, K. Sakamoto, S. C. Fagerholm, M. Foretz, C. C. Lang, and G. Rena (2016). Anti-Inflammatory Effects of Metformin Irrespective of Diabetes Status. Circ Res. 119(5):652-65. Carling, D. (2004). The AMP-activated protein kinase cascade--a unifying system for energy control. Trends Biochem Sci. 29(1):18-24. Carreno-Muller, E., A. J. Herrera, R. M. de Pablos, M. Tomas-Camardiel, J. L. Venero, J. Cano, and A. Machado (2003). Thrombin induces in vivo degeneration of nigral dopaminergic neurones along with the activation of microglia. J Neurochem. 84(5):1201-14. Castaño, Herrera A. J., Cano J., and Machado A. (2002a). The degenerative effect of a single intranigral injection of LPS on the dopaminergic system is prevented by dexamethasone, and not mimicked by rh-TNF-α, IL-1β and IFN-γ. Journal of Neurochemistry. 81(1):150-157. Castano, A., A. J. Herrera, J. Cano, and A. Machado (1998). Lipopolysaccharide intranigral injection induces inflammatory reaction and damage in nigrostriatal dopaminergic system. J Neurochem. 70(4):1584-92. Castaño, A., A. J. Herrera, J. Cano, and A. Machado (2002b). The degenerative effect of a single intranigral injection of LPS on the dopaminergic system is prevented by dexamethasone, and not mimicked by rh-TNF-α, IL-1β and IFN-γ. Journal of Neurochemistry. 81(1):150-157. Cattaneo, F., G. Guerra, M. Parisi, M. De Marinis, D. Tafuri, M. Cinelli, and R. Ammendola (2014). Cell-surface receptors transactivation mediated by g protein-coupled receptors. Int J Mol Sci. 15(11):19700-28. ! 141! Cayero-Otero, Maria Dolores, Ana M. Espinosa-Oliva, Antonio J. Herrera, Irene Garcia- Dominguez, Mercedes Fernandez-Arevalo, Lucia Martin-Banderas, and Rocio M. de Pablos (2018). Potential Use Of Nanomedicine For The Anti-Inflammatory Treatment Of Neurodegenerative Diseases. Current pharmaceutical design. Chang, Z. L., A. Novotney, and T. Suzuki (1990). Phospholipase C and A2 in tumoricidal activation of murine macrophage-like cell lines. FASEB J. 4:1753. Chaudhuri, K. R., and A. H. Schapira (2009). Non-motor symptoms of Parkinson's disease: dopaminergic pathophysiology and treatment. Lancet Neurol. 8(5):464-74. Chen, Fang, Rong Rong Dong, Kai Long Zhong, Arijit Ghosh, Su Su Tang, Yan Long, Mei Hu, Ming Xing Miao, Jian Min Liao, Hong Bing Sun, Ling Yi Kong, and Hao Hong (2016a). Antidiabetic drugs restore abnormal transport of amyloid-β across the blood–brain barrier and memory impairment in db/db mice. Neuropharmacology. 101:123-136. Chen, W. W., X. Zhang, and W. J. Huang (2016b). Role of neuroinflammation in neurodegenerative diseases (Review). Mol Med Rep. 13(4):3391-6. Chen, Xu, Xuan Li, Wenyan Zhang, Jie He, Bo Xu, Bin Lei, Zhenhua Wang, Courtney Cates, Thomas Rousselle, and Ji Li (2018a). Activation of AMPK inhibits inflammatory response during hypoxia and reoxygenation through modulating JNK-mediated NF-κB pathway. Metabolism: clinical and experimental. 83:256-270. Chen, Y. C., C. H. Kuo, Y. M. Tsai, Y. C. Lin, H. P. Hsiao, B. H. Chen, Y. T. Chen, S. L. Wang, and C. H. Hung (2018b). Suppressive effects of metformin on T-helper 1-related chemokines expression in the human monocytic leukemia cell line THP-1. Endocr Res.1-7. Chersi, F., M. Mirolli, G. Pezzulo, and G. Baldassarre (2013). A spiking neuron model of the cortico-basal ganglia circuits for goal-directed and habitual action learning. Neural Netw. 41:212-24. ! 142! Chiba, K., A. J. Trevor, and N. Castagnoli, Jr. (1985). Active uptake of MPP+, a metabolite of MPTP, by brain synaptosomes. Biochem Biophys Res Commun. 128(3):1228-32. Chung, M. M., C. J. Nicol, Y. C. Cheng, K. H. Lin, Y. L. Chen, D. Pei, C. H. Lin, Y. N. Shih, C. H. Yen, S. J. Chen, R. N. Huang, and M. C. Chiang (2017). Metformin activation of AMPK suppresses AGE-induced inflammatory response in hNSCs. Exp Cell Res. 352(1):75-83. Clark, I. E., M. W. Dodson, C. Jiang, J. H. Cao, J. R. Huh, J. H. Seol, S. J. Yoo, B. A. Hay, and M. Guo (2006). Drosophila pink1 is required for mitochondrial function and interacts genetically with parkin. Nature. 441(7097):1162-6. Cohen, G., and R. E. Heikkila (1974). The generation of hydrogen peroxide, superoxide radical, and hydroxyl radical by 6-hydroxydopamine, dialuric acid, and related cytotoxic agents. J Biol Chem. 249(8):2447-52. Conductier, G., N. Blondeau, A. Guyon, J. L. Nahon, and C. Rovere (2010). The role of monocyte chemoattractant protein MCP1/CCL2 in neuroinflammatory diseases. J Neuroimmunol. 224(1- 2):93-100. Cooper, Claudia, Andrew Sommerlad, Constantine G. Lyketsos, and Gill Livingston (2015). Modifiable Predictors of Dementia in Mild Cognitive Impairment: A Systematic Review and Meta-Analysis. American Journal of Psychiatry. 172(4):323-334. D'Amelio, Marco, Paolo Ragonese, Graziella Callari, Norma Di Benedetto, Barbara Palmeri, Valeria Terruso, Giuseppe Salemi, Giorgia Famoso, Paolo Aridon, and Giovanni Savettieri (2009). Diabetes preceding Parkinson's disease onset. A case–control study. Parkinsonism & Related Disorders. 15(9):660-664. D'Ardenne, K., N. Eshel, J. Luka, A. Lenartowicz, L. E. Nystrom, and J. D. Cohen (2012). Role of prefrontal cortex and the midbrain dopamine system in working memory updating. Proc Natl Acad Sci U S A. 109(49):19900-9. ! 143! D., Breder Christopher, Tsujimoto Masafumi, Terano Yoshitake, Scott Donald W., and Saper Clifford B. (1993). Distribution and characterization of tumor necrosis factor!α!like immunoreactivity in the murine central nervous system. Journal of Comparative Neurology. 337(4):543-567. Dahlstroem, A., and K. Fuxe (1964). Evidence for the Existence of Monoamine-Containing Neurons in the Central Nervous System. I. Demonstration of Monoamines in the Cell Bodies of Brain Stem Neurons. Acta Physiol Scand Suppl.SUPPL 232:1-55. Dahlström, A., and K. Fuxe (1964). Localization of monoamines in the lower brain stem. Experientia. 20(7):398-9. Damier, P., E. C. Hirsch, Y. Agid, and A. M. Graybiel (1999). The substantia nigra of the human brain. I. Nigrosomes and the nigral matrix, a compartmental organization based on calbindin D(28K) immunohistochemistry. Brain. 122 ( Pt 8):1421-36. Danner, H., and C. Pfister (1982). [7 types of neurons in the substantia nigra of the rat. Golgi rapidimpregnation study]. J Hirnforsch. 23(5):553-6. Date, I., M. F. Notter, S. Y. Felten, and D. L. Felten (1991). Stereotaxic injection of GD1a ganglioside induces limited recovery of striatal dopaminergic system in MPTP-treated aging mice. J Neurosci Res. 28(4):525-30. Dawson, T. M., H. S. Ko, and V. L. Dawson (2010). Genetic animal models of Parkinson's disease. Neuron. 66(5):646-61. de la Monte, S. M., L. Longato, M. Tong, and J. R. Wands (2009). Insulin resistance and neurodegeneration: roles of obesity, type 2 diabetes mellitus and non-alcoholic steatohepatitis. Curr Opin Investig Drugs. 10(10):1049-60. ! 144! de Rijk, M. C., W. A. Rocca, D. W. Anderson, M. O. Melcon, M. M. Breteler, and D. M. Maraganore (1997). A population perspective on diagnostic criteria for Parkinson's disease. Neurology. 48(5):1277-81. DeLong, M. R. (1990). Primate models of movement disorders of basal ganglia origin. Trends Neurosci. 13(7):281-5. Detaille, D., B. Guigas, C. Chauvin, C. Batandier, E. Fontaine, N. Wiernsperger, and X. Leverve (2005). Metformin prevents high-glucose-induced endothelial cell death through a mitochondrial permeability transition-dependent process. Diabetes. 54(7):2179-87. Diamond, C. E., H. J. Khameneh, D. Brough, and A. Mortellaro (2015). Novel perspectives on noncanonical inflammasome activation. Immunotargets Ther. 4:131-41. Dinarello, C. A. (2007a). A signal for the caspase-1 inflammasome free of TLR. Immunity. 26(4):383-5. Dinarello, Charles A. (2007b). A signal for the caspase-1 inflammasome free of TLR. Immunity. 26(4):383-385. Dinsmore, C. J., and P. Soriano (2018). MAPK and PI3K signaling: At the crossroads of neural crest development. Dev Biol. Dohi, Kenji, Hirokazu Ohtaki, Tomoya Nakamachi, Sachiko Yofu, Kazue Satoh, Kazuyuki Miyamoto, Dandan Song, Shohko Tsunawaki, Seiji Shioda, and Tohru Aruga (2010). Gp91phox (NOX2) in classically activated microglia exacerbates traumatic brain injury. Journal of Neuroinflammation. 7(1):41. Dong, Qiaoyun, Yanyong Wang, Ping Gu, Rusheng Shao, Li Zhao, Xiqi Liu, Zhanqiang Wang, and Mingwei Wang (2015). The Neuroprotective Mechanism of Low-Frequency rTMS on Nigral Dopaminergic Neurons of Parkinson’s Disease Model Mice. Parkinson’s Disease. 2015:8. ! 145! Downey, J. S., and J. Han (1998). Cellular activation mechanisms in septic shock. Front Biosci. 3:d468-76. Driver, J. A., T. Kurth, J. E. Buring, J. M. Gaziano, and G. Logroscino (2008). Parkinson disease and risk of mortality. A prospective comorbidity-matched cohort study. 70(16 Part 2):1423- 1430. Drougard, A., T. Duparc, X. Brenachot, L. Carneiro, A. Gouaze, A. Fournel, L. Geurts, T. Cadoudal, A. C. Prats, L. Penicaud, D. Vieau, J. Lesage, C. Leloup, A. Benani, P. D. Cani, P. Valet, and C. Knauf (2014). Hypothalamic apelin/reactive oxygen species signaling controls hepatic glucose metabolism in the onset of diabetes. Antioxid Redox Signal. 20(4):557-73. El-Mir, M. Y., D. Detaille, R. Villanueva G, M. Delgado-Esteban, B. Guigas, S. Attia, E. Fontaine, A. Almeida, and X. Leverve (2008). Neuroprotective role of antidiabetic drug metformin against apoptotic cell death in primary cortical neurons. J Mol Neurosci. 34(1):77-87. El-Mir, M. Y., V. Nogueira, E. Fontaine, N. Averet, M. Rigoulet, and X. Leverve (2000). Dimethylbiguanide inhibits cell respiration via an indirect effect targeted on the respiratory chain complex I. J Biol Chem. 275(1):223-8. EPDA (2018). European Parkinson's Disease Association. https://epda.eu.com/, accessed on: 07/06/2018. Fahn, S., and D. Sulzer (2004). Neurodegeneration and neuroprotection in Parkinson disease. NeuroRx. 1(1):139-54. Fallon, J. H., and K. B. Seroogy (1985). The distribution and some connections of cholecystokinin neurons in the rat brain. Ann N Y Acad Sci. 448:121-32. Fearnley, J. M., and A. J. Lees (1991). Ageing and Parkinson's disease: substantia nigra regional selectivity. Brain. 114 ( Pt 5):2283-301. ! 146! Fink, J. S., and G. P. Smith (1980). Mesolimbicocortical dopamine terminal fields are necessary for normal locomotor and investigatory exploration in rats. Brain Res. 199(2):359-84. Fleming, I. N., C. M. Elliott, and J. H. Exton (1996). Differential translocation of rho family GTPases by lysophosphatidic acid, endothelin-1, and platelet-derived growth factor. J Biol Chem. 271(51):33067-73. Foretz, M., and B. Viollet (2015). Therapy: Metformin takes a new route to clinical efficacy. Nat Rev Endocrinol. 11(7):390-2. Forno, L.S. (1982). Pathology of Parkinson’s disease. In Movement Disorders. C.D. Marsden and S. Fahn, eds. Pp. 25-40. London: Butterworth Scientific. Frey, E. A., D. S. Miller, T. G. Jahr, A. Sundan, V. Bazil, T. Espevik, B. B. Finlay, and S. D. Wright (1992). Soluble CD14 participates in the response of cells to lipopolysaccharide. J Exp Med. 176(6):1665-71. Fukata, M., and M. T. Abreu (2012). Mucosal Bacterial Recognition and Signaling Systems in the Intestine. In Physiology of the Gastrointestinal Tract. 5th edition. J. Leonard, ed. Pp. 1135– 1145. London: Elsevier. Galdieri, L., H. Gatla, I. Vancurova, and A. Vancura (2016). Activation of AMP-activated Protein Kinase by Metformin Induces Protein Acetylation in Prostate and Ovarian Cancer Cells. J Biol Chem. 291(48):25154-25166. Gang, Hu, Bidel Siamak, Jousilahti Pekka, Antikainen Riitta, and Tuomilehto Jaakko (2007). Coffee and tea consumption and the risk of Parkinson's disease. Movement Disorders. 22(15):2242-2248. Gao, H. M., J. Jiang, B. Wilson, W. Zhang, J. S. Hong, and B. Liu (2002). Microglial activationmediated delayed and progressive degeneration of rat nigral dopaminergic neurons: relevance to Parkinson's disease. J Neurochem. 81(6):1285-97. ! 147! Gao, H. M., H. Zhou, and J. S. Hong (2012). NADPH oxidases: novel therapeutic targets for neurodegenerative diseases. Trends Pharmacol Sci. 33(6):295-303. Gerfen, C. R. (1992). The neostriatal mosaic: multiple levels of compartmental organization. Trends Neurosci. 15(4):133-9. Gerfen, C. R., and D. J. Surmeier (2011). Modulation of striatal projection systems by dopamine. Annu Rev Neurosci. 34:441-66. Ghosh, S., M. J. May, and E. B. Kopp (1998). NF-kappa B and Rel proteins: evolutionarily conserved mediators of immune responses. Annu Rev Immunol. 16:225-60. Gitler, A. D., P. Dhillon, and J. Shorter (2017). Neurodegenerative disease: models, mechanisms, and a new hope. Dis Model Mech. 10(5):499-502. Goedert, M., M. G. Spillantini, K. Del Tredici, and H. Braak (2013). 100 years of Lewy pathology. Nat Rev Neurol. 9(1):13-24. Goldberg, M. S., and P. T. Lansbury, Jr. (2000). Is there a cause-and-effect relationship between alpha-synuclein fibrillization and Parkinson's disease? Nat Cell Biol. 2(7):E115-9. Graham, D. G. (1978). Oxidative pathways for catecholamines in the genesis of neuromelanin and cytotoxic quinones. Mol Pharmacol. 14(4):633-43. Grahn, J. A., J. A. Parkinson, and A. M. Owen (2008). The cognitive functions of the caudate nucleus. Prog Neurobiol. 86(3):141-55. Graybiel, A. M. (2000). The basal ganglia. Curr Biol. 10(14):R509-11. Grillner, S., and B. Robertson (2015). The basal ganglia downstream control of brainstem motor centres—an evolutionarily conserved strategy. Curr Opin Neurobiol. 33:47-52. ! 148! Grofová, I. (1975). The identification of striatal and pallidal neurons projecting to substantia nigra. An experimental study by means of retrograde axonal transport of horseradish peroxidase. Brain Res. 91(2):286-91. Grofova, I., and E. Rinvik (1970). An experimental electron microscopic study on the striatonigral projection in the cat. Exp Brain Res. 11(3):249-62. Gruzman, A., G. Babai, and S. Sasson (2009). Adenosine Monophosphate-Activated Protein Kinase (AMPK) as a New Target for Antidiabetic Drugs: A Review on Metabolic, Pharmacological and Chemical Considerations. Rev Diabet Stud. 6(1):13-36. Gu, J., S. Ye, S. Wang, W. Sun, and Y. Hu (2014). Metformin inhibits nuclear factor-kappaB activation and inflammatory cytokines expression induced by high glucose via adenosine monophosphate-activated protein kinase activation in rat glomerular mesangial cells in vitro. Chin Med J (Engl). 127(9):1755-60. Gu, M., J. M. Cooper, J. W. Taanman, and A. H. Schapira (1998). Mitochondrial DNA transmission of the mitochondrial defect in Parkinson's disease. Ann Neurol. 44(2):177-86. Gubellini, P., and P. Kachidian (2015). Animal models of Parkinson's disease: An updated overview. Rev Neurol (Paris). 171(11):750-61. Guimei, Liu, Bei Jia, Liang Li, Yu Guoyong, Li Lu, and Li Quanhong (2018). Stachyose Improves Inflammation through Modulating Gut Microbiota of High!Fat Diet/Streptozotocin!Induced Type 2 Diabetes in Rats. Molecular Nutrition & Food Research. 62(6):1700954. Gulley, R. L., and M. Smithberg (1971). Synapses in the rat substantia nigra. Tissue Cell. 3(4):691- 700. Gulley, R. L., and R. L. Wood (1971). The fine structure of the neurons in the rat substantia nigra. Tissue Cell. 3(4):675-90. ! 149! Gunton, J. E., P. J. Delhanty, S. Takahashi, and R. C. Baxter (2003). Metformin rapidly increases insulin receptor activation in human liver and signals preferentially through insulin-receptor substrate-2. J Clin Endocrinol Metab. 88(3):1323-32. Haber, S. N. (1986). Neurotransmitters in the human and nonhuman primate basal ganglia. Hum Neurobiol. 5(3):159-68. Haeusgen, W., R. Boehm, Y. Zhao, T. Herdegen, and V. Waetzig (2009). Specific activities of individual c-Jun N-terminal kinases in the brain. Neuroscience. 161(4):951-9. Hailman, E., H. S. Lichenstein, M. M. Wurfel, D. S. Miller, D. A. Johnson, M. Kelley, L. A. Busse, M. M. Zukowski, and S. D. Wright (1994). Lipopolysaccharide (LPS)-binding protein accelerates the binding of LPS to CD14. J Exp Med. 179(1):269-77. Han, C., M. W. Rice, and D. Cai (2016). Neuroinflammatory and autonomic mechanisms in diabetes and hypertension. Am J Physiol Endocrinol Metab. 311(1):E32-41. Han, J., Y. Li, X. Liu, T. Zhou, H. Sun, P. Edwards, H. Gao, F. S. Yu, and X. Qiao (2018). Metformin suppresses retinal angiogenesis and inflammation in vitro and in vivo. PLoS One. 13(3):e0193031. Harijith, A., D. L. Ebenezer, and V. Natarajan (2014). Reactive oxygen species at the crossroads of inflammasome and inflammation. Front Physiol. 5:352. Harvey, B. K., Y. Wang, and B. J. Hoffer (2008). Transgenic rodent models of Parkinson's disease. Acta Neurochir Suppl. 101:89-92. Hattori, Y., K. Hattori, and T. Hayashi (2015). Pleiotropic benefits of metformin: macrophage targeting its anti-inflammatory mechanisms. Diabetes. 64(6):1907-9. Hawkes, C. H., K. Del Tredici, and H. Braak (2010). A timeline for Parkinson's disease. Parkinsonism Relat Disord. 16(2):79-84.