Microglia, Calcification and Neurodegenerative Diseases

been detected in these brain regions, providing evidence of a local inflammatory reaction (Hirsch Hunot, 2009). The injection of lipopolysaccharide (a potent microglia activator) into the substantia nigra produces microglial activation and the death of dopaminergic cells. These findings support the hypothesis that microglial activation and neuroinflammation contribute to PD pathogenesis (Herrera et al., 2000). Neurodegenerative Diseases - Processes, Prevention, Protection and Monitoring focuses on biological mechanisms, prevention, neuroprotection and even monitoring of disease progression. This book emphasizes the general biological processes of neurodegeneration in different neurodegenerative diseases. Although the primary etiology for different neurodegenerative diseases is different, there is a high level of similarity in the disease processes. The first three sections introduce how toxic proteins, intracellular calcium and oxidative stress affect different biological signaling pathways or molecular machineries to inform neurons to undergo degeneration. A section discusses how neighboring glial cells modulate or promote neurodegeneration. In the next section an evaluation is given of how hormonal and metabolic control modulate disease progression, which is followed by a section exploring some preventive methods using natural products and new pharmacological targets. We also explore how medical devices facilitate patient monitoring. This book is suitable for different readers: college students can use it as a textbook; researchers in academic institutions and pharmaceutical companies can take it as updated research information; health care professionals can take it as a reference book, even patients' families, relatives and friends can take it as a good basis to understand neurodegenerative diseases.

Microglia, Calcification and Neurodegenerative Diseases 303 influenced by surrounding astrocytes and inflammatory T-cell subsets, which can affect their phagocytic capacity and antigen-presenting cell properties. All neurological disorders lead to activation of the microglia. Thus, microglial reaction represents the main mediator of the inflammatory process in neurodegenerative diseases, and microgliosis is directly related to the physiopathology of these. For example, microgliosis is associated with atypical and insoluble components caused by irregular protein folding and degradation pathways, altered subcellular localization, and the abnormal interactions with other cellular proteins found in AD, PD HD, Down syndrome and normal aging. Microgliosis is also associated with the formation of extracellular ionic precipitates, such as hydroxyapatites, which are frequently observed within the CNS areas involved in the disease (Rodriguez et al., 2009a, Saura et al., 1995, and is also present in encephalopathies caused by prions. This innate immune response is currently considered to be a potential pathogenic factor, since microglial reaction may engender neurodegenerative events, including amyloid-beta plaque formation, dystrophic neurite growth, and excessive tau phosphorylation.

Microglial reaction: Two sides of the same coin
In the healthy CNS, ramified resting microglia are active cells since they permanently scan their microenvironment (Wake et al., 2009). In response to any CNS injury or immunological stimuli, the microglia rapidly evolve from a surveillance state towards a more reactive one, through important phenotypical changes in response to activation signals released by the tissue (Schwartz et al., 2006). Microglia undergo a dramatic morphological transformation into amoeboid form and express an upregulated catalogue of molecules, such as CD14, major histocompatibility complex (MHC) molecules, chemokine receptors, CD11c, integrins, neurotrophins, and several other markers (Kettenmann et al., 2011). As such, reactive microglia can perform functions essential to neuron survival, such as phagocytosis to clear toxic and cellular debris, and innate immunity. Also involved in the release of trophic and anti-inflammatory factors, microglia facilitate repair through the guided migration of stem cells to the site of inflammation and injury. In contrast, once microglia become overactivated they can produce detrimental effects through excessive production of a large array of cytotoxic factors, such as NO, TNF-α, reactive oxygen species, and pro-inflammatory cytokines (Lull & Block, 2010, Milligan & Watkins, 2009. Currently, the conditions that determine whether microglial reaction will be detrimental or beneficial to neuronal survival are poorly understood. However, it is becoming more widely accepted that although microglial activation is necessary and crucial for host defense and neuron survival, microglial overactivation leads to deleterious consequences. Since every single microglial cell generates its own response to damage according to the nature and intensity of the signals released by the injured tissue, microglial cells do not constitute a homogenous cell population, but instead present a range of different phenotypes closely related to the evolution of the lesion process. In addition, some microglial cells become increasingly dysfunctional as they age, and may participate directly in the development of neurodegeneration (Block et al., 2007, Stoll et al., 2002. Microglia adopts a phenotype that mostly exacerbates tissue injury or promotes brain repair. Microglia can thus present two phenotypes, one of which is deleterious (also called M1 microglial phenotype) and the other benign (M2 microglial phenotype), depending on their intrinsic properties, interaction with the cellular microenvironment, and presence of pathogenic factors (Halliday & Stevens, 2011, Henkel et al., 2009. Therefore, controlling microglial cell activation and the acquisition of positive or negative phenotypes is of major therapeutic interest in all CNS disorders related to neuroinflammation.

Astrocyte-microglia interactions: Who's the bad guy after all?
The classical view of astroglia, as simply presenting non-excitable support to neurons, has changed radically in recent years. Astrocytes are now seen as elements that generate various local signals, including glutamate, to communicate with neurons and that influence the tissue outcome during neurodegeneration (Allaman et al., 2011). There is increasing evidence in support of this active role of astrocytes, suggesting that atypical astrocyte activation or astroglial dysfunction constitute maladaptive responses to brain injury that may feed the ongoing pathologic process during neurodegeneration. For example, astrocyte dysfunction is a key factor in the pathogenesis of human neurological disorders (Seifert et al., 2006), and in the cognitive impairment of aged rats (Andrés et al., 2000). Furthermore, glutamate-induced chronic lesion in rat brain not only presents a lack of astrogliosis but also long term atrophy of astrocytes, suggesting a maladaptive response that may be a cause of the on-going pathologic process (Rodriguez et al., 2009a). Moreover, astrocytes influence microglial behaviour (figure 1). For instance, astrocytes play a critical role in the activation of microglia under infectious conditions (Ovanesov et al., 2008). In addition, astroglial chemokines are involved in microglia/macrophage activation in multiple sclerosis with MCP-1/CCL2 and IP-10/CXCL10 directing reactive gliosis (Tanuma et al., 2006). Therefore, it is reasonable to assume that astrocytic activity can be influenced by microglial activation. Although a clear account of this dynamic relationship has yet to be proposed, the astrocyte-microglia interplay may determine the phenotype that microglial cells adopt during neurodegeneration. Some findings have implicated astrocytes in chronic microgliosis, with a transition from an initial neuroprotective activity to a later cytotoxic one. TNF-α secretion is crucial for rapid autocrine microglial activation with both neuroprotective and cytotoxic effects, a process that is also fed by TNF-α released by reactive astrocytes (Suzumura et al., 2006). TNF-α actions leading to neuronal death or survival are dose dependent (Bernardino et al., 2008), since it can activate two specific receptors: TNFR1, with an intracellular death domain, and TNFR2, mainly involved in neuroprotection (Fontaine et al., 2002). Consequently, low concentrations of TNF-α would initially induce TNFR2-mediated neuroprotection, whereas a subsequent high concentration of TNF-α would be able to activate TNFR1 both in astrocytes and microglia and contribute to cell injury through the death domain of the receptor. Astroglial S100 is another of the factors that control microglial activity. Astrocytes release S100 constitutively (Van Eldik & Wainwright, 2003) and increase this release upon stimulation by several factors, including TNF-α (Edwards & Robinson, 2006). Under normal conditions, released S100 acts as a neurotrophic factor, countering the stimulatory effect of neurotoxins on (Reali et al., 2005) and stimulating astrocyte glutamate uptake (Tramontina et al., 2006). Released S100 modifies astrocytic, neuronal and microglial activities, depending on the extracellular concentration of the former and the expression of the specific receptor for advanced glycation end-products (RAGE). At micromolar concentrations, S100 upregulates IL-1 and TNF-α expression in activated microglia via RAGE, with the requirement of concurrent activation of NF-κB and AP-1 transcription factors (Bianchi et al., 2010). Furthermore, factors that modulate microglial reactivity, such as Ca 2+ concentration, Signalling systems in the microglia-astrocyte-neuron cross-talk. Astrocytes and microglia monitorize neuronal activity by sensing neurotransmitter release. In the same way, microglial cells have receptors to molecules released by astrocytes. Microglia integrate all these signals and release molecules that may modulate neuronal and astroglial activity. During neurodegeneration, changes in physiological parameters may trigger neuronal injury and/or microglial activation (molecules inside a red square) that modify those signal transduction systems. Aβ, amyloid beta; Ach, acetylcholine; Ado, adenosine; BDNF, Brainderived neurotrophic factor; GABA, γ-aminobutyric acid; GDNF, Glial cell-derived neurotrophic factor; Glc, glucose; Glu, glutamate; 5-HT, serotonin; VIP, Vasoactive intestinal peptide; IFNγ, interferon gamma; ROS, reactive oxygen species; Tau, taurine reactive oxygen species or TNF-α, also modify the RAGE response to S100 (Edwards & Robinson, 2006) in a microglia-astroglia cross-talk that integrates these signaling systems. In contrast, at high concentrations S100 binds the RAGE, which may mediate microglial activation during the course of brain damage (Bianchi et al., 2010). An increased release of S100 during neurodegeneration (Li et al., 2011) will enhance inflammatory cytokine production and potentiate the switch of microglia to chronic cytotoxic activity, feeding the neurotoxic process and leading to neurodegeneration.
Thus, CNS neurodegeneration involves chronic microgliosis with a putative transition from an initial neuroprotective activity to a later cytotoxic one, with S100 as a key modulator of microglial transition towards a cytotoxic response. This suggests a role for astrocytes in promoting the cytotoxic microglial phenotype through secretion of TNF-α, S100 and other signals (Donato et al., 2009, Suzumura et al., 2006. In this scenario, the interplay between trophic, neuroprotective, inflammatory and cytotoxic functions of both cell types during brain injury will determine the evolution of the neurodegenerative process. Precise control of these processes thus requires a dynamic view of their interactions so as to allow effective development of approaches to neuroprotection.

Enduring effects of excessive synaptic glutamate
Glutamate accounts for most of the excitatory synaptic activity in the CNS and has been implicated in learning, memory, synaptic plasticity and neurotrophic activity processes. Glutamate receptors have been classified into three groups: two ionotropic groups -Nmethyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA)-kainate receptors-; and one group of metabotropic receptors, which are coupled to G proteins. Although non-NMDA receptors are not initially permeable to Ca 2+ , glutamate release in the synaptic cleft increases post-synaptic and glial membrane permeability, leading to a transient increase in intracellular Ca 2+ concentration ([Ca 2+ ] i ) (Obrenovitch et al., 2000). Excessive activation of glutamate receptors can trigger neuronal death through a process characterized by chronic glutamate release and the consequent [Ca 2+ ] i dys-homeostasis in neurons and astrocytes (Arundine & Tymianski, 2004) and formation of [Ca 2+ ] i precipitates, the size and number of which depends on the CNS area involved and CNS maturation (Bernal et al., 2000b, Rodríguez et al., 2001. This process, defined as excitotoxicity, also involves cellular influxes of Na + and Cland efflux of K + , with ensuing cell swelling (Chen et al., 1998, Katayama et al., 1995. Because of the complexity and diversity of the processes taking place at the glutamatergic synapse, any disturbance at the pre-synaptic, postsynaptic, or astroglial level may trigger a chronic excitotoxic process. For example, ALS presents a loss of selectivity of ionotropic receptors (Obrenovitch et al., 2000) and deficiencies in glial re-uptake of glutamate (Liévens et al., 2000). These glutamate-related dysfunctions contribute to explaining phenomena such as the aging-associated hypoactivity of NMDA receptors observed in AD (Olney et al., 1997), and the specific AMPA-receptor increment detected in the hippocampus of aged, cognitive-impaired rats (Le Jeune et al., 1996). One of the consequences of excitotoxic-induced neuronal loss is the alteration of other neurotransmitter systems and neuromodulators. For example, long-term ibotenic-induced lesion in the basal forebrain of rat leads to a loss of cholinergic afferences and to decreased extracellular noradrenaline, glutamate, and taurine . This cortical reduction in glutamatergic transmission presents a temporal pattern which, together with the development of Ca 2+ precipitates and a decrease in the cholinergic and noradrenergic functions (Saura et al., 1995), mimics the neurochemical modifications described in AD. Similarly, one year after acute lesion, the cortical and hippocampal decrease in brainderived neurotrophic factor, fibroblastic growth factor and glucocorticoid receptor, and the increase in c-fos expression in the septal area were still significant (Boatell et al., 1992). Thus, www.intechopen.com excitotoxic lesions in basal forebrain can modify long-term cortical adaptative responses, and may modulate the expression of glutamate receptor. Some of these effects, such as the decrease in brain-derived neurotrophic factor and the increase in c-fos expression, also reflect the molecular alterations present in AD. Given these toxic effects, adaptations that act to control glutamatergic neurotransmission and Ca 2+ movements in the cell may be potentially protective. Over time, these defenses are developed to act at any time during an excitotoxic event, involve different cellular types such as neurons, astrocytes and microglia, and deal with the cellular and molecular mechanisms of glutamatergic neurotransmission. These mechanisms include defenses that: a) decrease neuronal excitability, b) decrease glutamate accumulation in the synapse, c) limit Ca 2+ mobilization in the postsynaptic neuron and protect against calcium-dependent degenerative effects, and d) enhance neuronal energy (Rodriguez et al., 2009b, Sapolsky, 2001. If the compensatory mechanisms are not sufficiently effective, the initial acute neuronal injury due to an increase in [Ca 2+ ] i leads, with time, to a chronic lesion. This secondary excitotoxicity appears in neurons after the massive entrance of Ca 2+ and Na + through ionotropic glutamate receptors, an entrance that is supplemented by Ca 2+ release from the endoplamic reticulum following activation of mGluRs. As a result, an excessive [Ca 2+ ] i increment occurs (Verkhratsky, 2007), which activates the mechanisms triggering neuronal death. Ca 2+ extrusion and buffering are activated when [Ca 2+ ] i increases (Mattson & Chan, 2003), with high expenditure of energy through Ca 2+ -ATPases. (Figure 4). Calcium homeostasis disturbances are present in all neurodegenerative disorders (Mattson, 2006). Dysregulation of Ca 2+ homeostasis alters the rapid and coherent activation of neurons, and is therefore ultimately responsible for many aspects of brain dysfunction and CNS diseases. For example, an increased rate of Ca 2+ -mediated apoptosis may cause neuronal death in the penumbra of cerebral ischemia, or may underlie the etiology of chronic neurodegenerative disorders such as PD and AD. Calcium precipitation that coincides with microglial activation, amyloid deposits and other ion accumulations in AD may thus be a key element of the neurodegenerative process (Ramonet et al., 2006).

Neurodegeneration as a result of disturbances in calcium homeostasis
Regulation of intraneuronal Ca 2+ movements is a key element to ensure adequate cellular response at all times and produce a physiological effect. Ca 2+ participates in most cellular functions, including membrane excitability and secretion, energy production, synaptic transmission, gene regulation and plasticity, cell proliferation and cell death (Arundine & Tymianski, 2003). As a consequence, cell function requires tight control of Ca 2+ homeostasis between extra-and intracellular compartments at all times, including the production of sufficient energy from glycolysis to maintain different gradient concentrations such as a 1/10,000 inside-outside cell concentration (Verkhratsky, 2005). Most calcium within cells is sequestered in the mitochondria and the endoplasmic reticulum. Intracellular free calcium concentrations fluctuate widely, from roughly 100 nM to over 1 μM, due to release from cellular stores or influx from extracellular fluid. These fluctuations are integral to the role of calcium, and unless compensated for by some other mechanism, any dysregulation will have severe cellular consequences, the effects of which will be propagated to the surrounding cells, namely, neurons and glial cells (Beck et al., 2004). For example, activation of glutamate receptors by excitatory signals leads to a massive increase in cytoplasm Ca 2+ levels, which in turn activates a cascade of events to produce a neuronal response. A return to basal activity requires a considerable expenditure of energy to bring Ca 2+ back to initial levels. Any failure in these multiple, coordinated steps will alter neuronal signalling and interfere with neuronal network functions (Rodriguez et al., 2009b). Reduction of [Ca 2+ ] i involves a high mitochondrial intake of Ca 2+ that may lead to loss of the mitochondrial membrane potential and the production of reactive oxygen species, thereby decreasing cellular respiratory capacity and ATP formation from ADP and Pi (Chan et al., 2007a). As a result, there is an acceleration of anaerobic glycolysis with a net lactate production that contributes to tissue acidification and progression of damage. Disturbances in Ca 2+ homeostasis in astrocytes reduce neuronal support, in particular through alteration of the glutamate/glutamine cycle and a reduction of glucose delivery to neurons (Pellerin et al., 2007, Ramonet et al., 2004 (Figure 4). Consequently, astrocyte dysfunction can lead to increased synaptic glutamate levels and glutamate receptor overactivation, combined with reduced neuronal energy, resulting in neuronal damage. Under these conditions of high Ca 2+ and Pi and low ATP, formation of hydroxyapatite precipitates to reduce Ca 2+ cytoplasm activity at low energy costs can occur in neurons and astrocytes (Rodriguez et al., 2000). This new step in calcium homeostasis temporarily helps the cell to resist excessive stimulatory signals and return to basal activity by dissolving paracrystal elements. However, in most cases, the hydroxyapatite crystals show progressive growth and participate in cell death and CNS damage.
Depending on the importance of the damage, microglial activation may take place, initially expressing neuroprotective signals to help avoid further neuronal death, but then changing progressively to an inflammatory phenotype (Graeber, 2010). Glycolysis, highly stimulated in microglia to ensure and maintain their activated stage, reduces neuronal glucose and oxygen availability. Consequently, microglial participation in neuronal death includes not only neuroinflammation, but also reduction in neuronal energy availability (Allaman et al., 2011). If, as asserted by Gyuri Buzsaki in Rhythms of the Brain (Buzsáki, 2006), "Brains are foretelling devices and their predictive powers emerge from the various rhythms they perpetually generate", any significant alteration of the neuronal rhythms caused by an abnormally high Ca 2+ concentration in neurons or glia should alter the brain's ability to pause, adapt and learn, and can lead to disease. This is the case of neurodegenerative diseases, which exhibit diverse clinical and neuropathological phenotypes but share the common feature of progressively reduced cell function and survival within the nervous system, leading to neurological disability and often death. As such, several different CNS disorders can be induced following the same injury, due to the multi-directional interactions between the neurons, glial cells, extracellular matrix, endothelia and host immune cells that regulate tissue homeostasis and orchestrate neuroinflammation and degeneration. Furthermore, the characteristics of each neuronal population and network, the different gliopathic changes occurring between CNS areas, the various microglia phenotypes and the abundance and distribution of glutamate receptor subtypes and of Ca 2+ -binding proteins, all participate directly in the properties of the neurodegenerative parameters (Graeber & Streit, 2010, Rodriguez et al., 2004, Rodriguez et al., 2009a) that will determine the dynamics and progression of the disease in the specific affected areas. For example, PD and AD are both regarded as diseases that are initiated by neuronal death to which the immune system responds, as evidenced by astroglial and microglial activation with pathogenic consequences (Agostinho et al., 2010, Halliday & Stevens, 2011. Similarly, multiple sclerosis is typically considered a neuroinflammatory disorder, but one in which neuronal injury plays an active role in regulating neuroinflammation, as has been recently reported (Haider et al., 2011). In all of these diseases, dysregulation of Ca 2+ homeostasis has been considered a pathophysiological factor linked to neuronal degeneration, and the formation of intracellular Ca 2+ deposits with different characteristics as regards size and distribution -reflecting differential CNS area vulnerability -has frequently been reported (Hashimoto et al., 2003, Ramonet et al., 2006.

CNS calcification
In immature human CNS, Ca 2+ -mediated excitotoxicity is associated with a calcification process that directly correlates with neuronal loss and the extent of injury. Revealed by Alizarin red staining and appearing in TEM and X-ray microanalyses of animal neurodegeneration models, small and large intracellular Ca 2+ precipitates indicate the formation of a paracrystalline structure of hydroxyapatites localized within neurons and astrocytes (figure 2). Glutamate analog microinjection in rat CNS leads to an intracellular Ca 2+ precipitation similar to brain calcification in humans (Ramonet et al., 2006(Ramonet et al., , 2002. As these Ca 2+ deposits can be observed in several areas of rat brain after microinjection of different excitotoxins (Bernal et al., 2000b, Rodriguez et al., 2000, Saura et al., 1995, their formation does not depend on the glutamate receptor subtype initially stimulated. However, their size, number and distribution vary with both the activated receptor and the CNS area. For example, sensitivity to AMPA-induced calcification decreased from the globus pallidus, cerebral cortex, hippocampus, medial septum, to retina (Rodriguez et al., 2000). Moreover, in medial septum, the degeneration associated with microinjection of ibotenic and quisqualic acids was characterized by significant atrophy and no calcification (Mahy et al., 1996, Saura et al., 1995. In similar conditions, AMPA microinjection resulted in similar atrophy and Ca 2+ deposits at the injection site (Rodriguez et al., 2009a). Ca 2+ deposits do not occur in all cells that degenerate in response to excitotoxins. For example, in the basal forebrain and medial septum, the calcification observed in GABAergic cells was not detected in cholinergic neurons. The former, together with astrocytes, seem to participate actively in the calcification process . Differences in the neuronal phenotype of Ca 2+ buffering and extrusion systems, specific energy needs, expression of the glutamate subtype receptor and different astroglial populations, should explain this variability. The ultrastructural study of tissue affected by excitotoxicity has also contributed to our understanding of calcification. Ca 2+ deposits within hypertrophied astrocytes have been characterized in the basal forebrain and hippocampus which ranged from 0.5 to 10 μm in diameter and were formed by numerous, small, needle-shaped crystals associated with cellular organelles, such as microtubules, cisternae, vesicles or mitochondria, with no signs of neurodegeneration (figure 2). Larger inclusions were surrounded by reactive microglia, a finding that was also observed in tissue after specific localization by in vitro autoradiography (Bernal et al., 2000b, Petegnief et al., 1999. X-ray microanalysis has shown an electron-diffraction ring pattern characteristic of a crystalline structure similar to apatites (Kim, 1995), and a Ca/P ratio of 1.3±0.2 of cytoplasmic deposits ( Figure 2). This ratio, lower than the theoretical apatite value of 1.67, is also typical of biological crystals, which do not present an ideal organization (Rodriguez et al., 2000). As biological hydroxyapatites, these deposits are similar to those observed in several human peripheral nervous system tissues (Kodaka et al., 1994).

Fig. 2. Characterization of calcium deposits induced by ibotenic acid in the rat brain.
Microphotographs of a) Nissl stained section of a rat hippocampus 15 days after the injection and b) Alizarin red stained section of the same hippocampus showing calcium deposits associated with the lesion. c-d) calcium deposits showed different sizes in the rat globus pallidus 2 months after injection. e) Isolectin B4 histochemistry (brown staining) counterstained with alizarin red showing the microglial reaction (arrowhead) associated with calcium deposits. f) Hypertrophic astrocyte with an intracytoplasmic calcium deposit by TEM. g) Detail of the ultrastructure of calcium deposit within an astrocyte. Note the normal appearance of the surrounding mitochondria (arrowhead). X-ray image h) and spectrum analysis i) of one calcium deposit in a non-osmificated sample with a calculated Ca/P ratio of 1.3. False colour X-ray image mapping j) and distribution plots k) of Ca and P of the same deposit. Bars; a-b, 1 mm; c, 100 μm; d-e, 20 μm; f, 0.6 μm, g, 0.2 μm; h, 10 μm www.intechopen.com Experimental models of bone formation (i.e. hydroxyapatite formation in vitro) (Andre-Frei et al., 2000) have shown that rather than Ca 2+ , a minimal amount of phosphorus, as inorganic phosphate, is crucial for crystal nucleation in a collagen matrix. Similarly, organic phosphate residues of the phosphoproteins also play a direct and significant role in the process of in vitro nucleation of apatite by bone collagen, whereas collagen itself does not promote the precipitation of Ca 2+ or phosphate (Andre-Frei et al., 2000). Therefore, excitotoxicity-induced calcification in the rat brain depends on an increase in intracellular inorganic phosphate (i.e. ATP depletion) and, most importantly, on the degree of protein phosphorylation. Thus, the Ca 2+ -binding-protein-dependent kinases and activity of the neurotrophic factor ultimately determine calcification. In aqueous solutions, hydroxyapatite crystallization takes place in two sequential steps (Barat et al., 2011): in the first, crystal nucleation occurs spontaneously with subsequent growth to some nanometers, when phosphate and Ca 2+ ions reach a certain concentration; in the second step, an accretion process of these nanocrystals on a proteinic net takes place until reaching a maximum size of 20 micrometers. While the first process facilitates resolubilization of the crystal, the second produces a stable precipitate and requires a catalytic agent. These two mechanisms may help explain the size differences we found between several areas of the CNS. For example, the large insoluble Ca 2+ precipitates (mean size 20 μm) found after AMPA microinjection in globus pallidus (Petegnief et al., 1999) fit well with the second step theory, whereas the small deposits (mean size lower than 3 μm) obtained in hippocampus after the same procedure (Rodriguez et al., 2004) may reflect the lack of a catalytic agent for accretion, or an equilibrium between crystal formation and solubilization. Furthermore, blockade of glial glutamate uptake in rat striatum (Liévens et al., 2000) produced a spherical lesion with a central necrotic core surrounded by a penumbra zone similar to that caused by focal ischemia. Three days after treatment, an astroglial reaction and small Ca 2+ deposits (mean diameter < 1 μm) were observed in the penumbra area. Eleven days later, these deposits had disappeared, the penumbra zone had recovered from injury and the necrotic area was partially repaired (Liévens et al., 2000). In this scenario, compensatory mechanisms helped normalize Ca 2+ homeostasis and avoid further neuronal death. The tissue recovered the ability to use extrusion mechanisms, and re-solubilization of Ca 2+ precipitates took place. When Ca 2+ deposits are localized extracellularly due to cell death, a microglial reaction is activated for their phagocytic removal. This microglial reaction also participates in the neuronal death seen in chronic neurodegenerative processes, but is dissociated from astrogliosis. In some animal models of neurodegeneration, a recovery has been observed associated with the disappearance of Ca 2+ deposits. In other excitotoxic rodent models, the on-going neurodegenerative process increased with time and the Ca 2+ deposits remained present, associated with microglial reaction.

The calcification process and ageing
In the neonate mammalian brain, considered more resistant to hypoxia-ischemia than adult CNS, dysregulation of Ca 2+ homeostasis together with lactate acidosis are considered the main factors causing neuronal death. As premature-neonates are more resistant to hypoxiaischemia than term neonates, we studied the relationship between differences in human brain vulnerability to hypoxia-ischemia during the perinatal period and brain calcification in the basal ganglia, cerebral cortex, and hippocampus (Rodríguez et al., 2001). The number and size of the observed non-arteriosclerotic calcifications were area-specific and increased in term neonates ( Figure 3). The basal ganglia presented the highest degree of calcification and the hippocampus the lowest, mainly in the CA1 subfield. In all cases, neuronal damage was associated with astroglial reaction and Ca 2+ precipitates, with microglial reaction absent in the hippocampus. These data are consistent with those obtained following long-term excitotoxic lesions in adult rat brain and support the involvement of excitotoxic processes in hypoxia-ischemia damage. A comparison between lifespan and degree of calcification (Figure 3) demonstrated that in all cases, highest calcification occurred within two months of hypoxia-ischemia, and that semi-calcification time was very short (less than 10 days). Independent of subjective measurements, this last parameter suggests that calcification depends on the degree of brain differentiation and initial cerebral injury, but not on the time-course of the lesion. Moreover, the mechanisms leading to Ca 2+ precipitation seem to be similar for all brain areas. If this is true, neurons of each CNS structure degenerate through a common mechanism, which is linked to disturbances in Ca 2+ homeostasis. As each area of the brain participates in specific physiological functions, the resultant pathology will depend on the specific neuronal death of the area affected. Aging increases neuronal vulnerability to toxic compounds, including drugs that impair energy metabolism and induce secondary excitotoxic processes (Brouillet et al., 1993). However, a decreased susceptibility of aged rats to excitotoxins such as quinolinic or kainic acids has been reported (Kesslak et al., 1995). AMPA-induced Ca 2+ deposits in rat hippocampus are age-dependent, since young rats (3 months old) present greater areas of calcification than middle-aged ones (15 months old) (Bernal et al., 2000a). In this study, glial reaction, -aminobutiric acid (GABA)-uptake activity and immunostaining of Ca 2+ binding proteins showed the same response. Therefore, the vulnerability of hippocampal neurons to AMPA-induced neurodegeneration decreases with age between 3 and 15 months. Similar results have been found in other brain areas, such as the striatum and the nucleus basalis magnocellularis. This reduced vulnerability may be related to several factors: for example, age-associated variations in the relative abundance of glutamate receptors and pre-synaptic alterations of glutamate release may explain, at least in part, an increased resistance to excitotoxicity in the hippocampus (Mullany et al., 1996, Nicolle et al., 1996. This effect is compatible with the increased vulnerability to excitotoxicity observed in the oldest animals (Brouillet et al., 1993), since some of the factors responsible for injury resistance may follow a biphasic pattern, with a progressive increase until reaching maturity followed by a subsequent decrease (Coleman et al., 1990). Many authors have also described biphasic variations in several parameters during aging, with an opposite tendency before and after middle age (Villa et al., 1994). We observed a biphasic variation in monoamine oxidase B (MAO-B) during aging in most human brain areas: up to the age of 50-60 years old, MAO-B levels remain constant, but start to increase thereafter (Saura et al., 1997). This finding may be due to the presence of MAO-B rich reactive astrocytes in response to neuronal degeneration. A similar increase in plaque-associated astrocytes has been found in patients with AD (Saura et al., 1994). As MAO-B activity is associated with reactive oxygen species production, astrocytes may contribute to the age-associated decline of neurological functions. The evidence that an increase in AMPA receptor correlates negatively with MAO-B in age-associated learning-impaired rats also suggests that a gliopathic reaction may be involved in neuronal dysfunction (Andrés et al., 2000). Fig. 3. Calcification depends on the brain area but also on the glutamate receptor involved. a) AMPA induces small calcium deposits in the rat hippocampus, affecting mainly the CA1 radiatum and lacunosum moleculare subfields. b) AMPA microinjection in the globus pallidus induces larger calcium deposits. c) NMDA microinjection in the hippocampus induces the formation of large calcium deposits located mainly in the pyramidal CA1 and granual dentate gyrus. d) The plots show comparison of the AMPA dose-response study in the hippocampus and the globus pallidus. e-f) Correlation plots of the hypoxia ischemiainduced calcification in the basal ganglia e), cerebral cortex f) and hippocampus g) of premature and term neonates. Calcification was calculated in a representative area (1 mm 2 ) and the lifespan corresponds to the time of injury in days. k, days to reach half of the maximal calcified area. Bars, 300 μm Thus, the correlation between the calcification process, neuronal loss and the extent of CNS injury disappears with aging, but differences in CNS area vulnerability to calcification are maintained. The components that underlie the specific vulnerability of each brain area are thus already expressed in human neonates. The permanent area differences are associated with significant variations in the response to specific Ca 2+ channel blockers such as nimodipine and TMB-8 (Bernal et al., 2009, Petegnief et al., 2004, and illustrate the functional diversity of each area and the difficulty encountered in ensuring the efficacy of such types of treatment. Similar results concerning differences between calcium precipitates and brain area susceptibility have been observed in congenital toxoplasmosis (Safadi et al., 2003, Surendrababu et al., 2006. Cerebral calcification has been described in 65% of these patients, with calcified foci distributed predominantly in the cortex in the form of tiny flecks, and as linear streaks in the basal ganglia. The calcification process can thus be considered a new stage in cytoplasmic calcium homeostasis taking place in a diversity of CNS injuries to reduce calcium signalling at no energy cost. When located extracellularly due to cell-death, these precipitates activate a permanent microglial reaction aimed at their removal but which rapidly turns into chronic damage and aggravation of neurodegeneration.

Uncoupling of retaliatory systems and energy availability
The balance between retaliatory system actions and energy metabolism constitutes a fine equilibrium in physiological conditions, but it can be disrupted by glutamate-mediated neuronal injury to then participate in the evoked neurodegenerative process. For example, AMPA-microinjection in medial septum induces a progressive cholinergic and GABAergic loss associated with a long-term decline of the hippocampal functions and decreased glutamatergic activity (Rodriguez et al., 2005). Other effects of this lesion imply modifications of adenosine and taurine transmissions, glutamate recycling and glucose metabolism (Ramonet et al., 2004, Rodriguez et al., 2005. Over time, adenosine replaces GABA functions to avoid further excitotoxic damage when cholinergic and GABAergic processes are compromised. Long-term septal lesion-induced neuronal loss in the hippocampus is apoptotic, with enhancement of neuronal glycolysis. Together with the cleavage of caspase 3, a glutamateglutamine cycle displacement towards glutamine production reduces glutamate synthesis (Ramonet et al., 2004). In addition, synaptic glutamine is decreased, probably through expulsion to vessels, where it exerts a vasodilatory effect through NO synthesis inhibition (Mates et al., 2002). In this scenario, the reduction in glutamate signaling and increased neuronal energy metabolism both reflect a neurodegenerative process with deficient adaptation of the retaliatory systems and a chronic energy requirement to execute the apoptotic program. This chronic energy requirement induces mitochondrial damage, in turn leading to acidosis in cells and the extracellular space (Hertz, 2008). Mitochondrial damage forces the cell to shift from an aerobic to an anaerobic metabolism, and as a result lactate is produced with the formation of two ATPs and the release of two protons. After trauma and ischemia, extracellular lactate increases dramatically and pH decreases. To ensure neuronal viability during and even after human hypoxia, glial glucose is only oxidized to lactate, which is rapidly transported into neurons for its complete oxidation (Sibson et al., 1998). In parallel, a Ca 2+ influx causes rapid cytoplasmic acidification (Verkhratsky, 2007) through: a) activation of membrane Na + /H + exchanger to restore the Na + gradient, and b) the Ca 2+ -dependent displacement of protons bound to cytoplasmic anions (Arundine & Tymianski, 2004). Furthermore, H + also appears during some chemical reactions, such as phospholipid hydrolysis.

Functional relevance of the calcification process
The massive astroglial production of lactate to help compensate for the neuronal energy depletion caused by excitotoxicity is a key factor in brain calcification (Figure 4). pH reduction associated with increased lactate concentration facilitates solubility of Ca 2+ and the formation of H 2 PO 4 -, HPO 4 2-and PO 4 3-ions from inorganic phosphate (Rodriguez et al., 2000) and phosphorylated proteins. Because of the very high Ca 2+ / H 2 PO 4 -, HPO 4 2-, PO 4 3affinity, apatite nucleation may occur, with the subsequent growth of crystalline formations together with neurodegeneration. In this case, calcification of each lesioned area will also depend on phosphate availability and the differential capacity of glial cells to release lactate during degeneration. Wide variations have been described in the extent of calcification in pathological cases (Ramonet et al., 2002, Rodríguez et al., 2001 and animal species (Ramonet et al., 2006) However, the homogeneous morphology of these deposits suggests common synaptic processes (Ramonet et al., 2006) where variability depends on cellular type (astrocyte or neuron), glutamatergic activity, and energy availability (Figure 4). These factors modify Ca 2+ homeostasis and may trigger cellular calcification through a common mechanism (Ramonet et al., 2006). Thus, hydroxyapatite formation, with the subsequent reduction of free Ca 2+ ions, may take place as an alternative homeostatic step to reduce excitotoxicity (Ramonet et al., 2006, Rodriguez et al., 2000, and a number of findings lend support to this interpretation. For example, it has been observed that mitochondria close to Ca 2+ deposits appear normal at electron microscopy level , Rodriguez et al., 2000, despite Fig. 4. Excitotoxicity modifies cell calcium homeostasis in the brain. Drawing of the excitotoxic process induced by glutamate, with the intercellular precipitation of calcium as part of the calcium homeostasis. The metabolic pathway of lactate with the communication between endothelial, astroglial, microglial and neuronal compartments is included in the diagram the fact that mitochondrial dysfunction constitutes a primary event in NMDA-induced degeneration (Schinder et al., 1996). This hypothesis is also consistent with the finding that neurons undergoing prolonged stimulation of NMDA receptors can survive in the presence of [Ca 2+ ] i chelators. Very high levels of cytoplasmic Ca 2+ are not necessarily neurotoxic, and an effective uptake of this element into mitochondria is required to trigger NMDA-receptorstimulated neuronal death (Stout et al., 1998). Moreover, in rat globus pallidus, an AMPAdose-response study has shown a dose-dependent increase in calcification, which was not accompanied by an increase in astrogliosis (Petegnief et al., 1999). In the hippocampus, AMPA induced a calcified area larger than the injured area. In this same structure, the selective adenosine-A2a-receptor antagonist 8-(3-chlorostyryl)-caffeine increased NMDAinduced neuronal loss while calcification was decreased (Robledo et al., 1999). All these data indicate that Ca 2+ precipitation does not necessarily reflect neuronal death and that, as proposed for retinal excitotoxic damage (Chen et al., 1999), in addition to Ca 2+ other factors such as Na + and Clinflux, cell swelling and acidosis induce excitotoxic neuronal damage.

Conclusions
Neurodegenerative disorders are characterized by the appearance of distinct neurodegenerative parameters that determine the induction of a chronic process with underlying glutamate-mediated excitotoxicity and Ca 2+ dys-homeostasis. At tissue level, the pathogenesis of each disorder depends on the neuronal type involved, synaptic density, glial interactions, and vicinity of vascularization. For each neuron, astrocyte and microglial type, the group of glutamate and cytokine receptors, the Ca 2+ binding protein content, protein phosphorylation levels, and all elements that participate in energy needs and glucose availability will constitute the factors involved in the appearance of the lesion. In this scenario, CNS calcification can be considered one of the few common mechanisms already available at an early age to help buffer disturbances in Ca 2+ homeostasis at no energy cost. Over time, CNS maturation includes a massive increase in synaptic connections, the organization of inhibitory systems and greater cellular complexity. As it becomes more sophisticated, the CNS relies on a greater diversity of mechanisms to prevent CNS injury. This would explain why calcification is observed in neurodegenerative diseases, but does not correlate with CNS damage.