The Neurophysiology and Pathology of Brain Zinc

Our understanding of the roles played by zinc in the physiological and pathological functioning of the brain is rapidly expanding. The increased availability of genetically modified animal models, selective zinc-sensitive fluorescent probes, and novel chelators is producing a remarkable body of exciting new data that clearly establishes this metal ion as a key modulator of intracellular and intercellular neuronal signaling. In this Mini-Symposium, we will review and discuss the most recent findings that link zinc to synaptic function as well as the injurious effects of zinc dyshomeostasis within the context of neuronal death associated with major human neurological disorders, including stroke, epilepsy, and Alzheimer's disease.


Introduction
Brain zinc, in its free ionic form (Zn 2ϩ ), is present within synaptic vesicles at glutamatergic nerve terminals and is synaptically released during neuronal activity. Zn 2ϩ is also bound to metalloproteins and intracellularly mobilized upon oxidative stress. A growing and exciting body of evidence indicates that Zn 2ϩ plays a dynamic role in both the physiology and pathophysiology of brain function.
Synaptic activation releases vesicular Zn 2ϩ , bringing its concentrations in the synaptic cleft to transiently rise. The exact amount of such release is controversial, but many laboratories have indicated (with the limitation of the current imaging techniques) that transient Zn 2ϩ increases may reach 1-100 M (Vogt et al., 2000;Qian and Noebels, 2005;Frederickson et al., 2006). Others have indications for lower (submicromolar) concentrations (Komatsu et al., 2005). Fuelling the controversy is the fact that measurement of actual Zn 2ϩ levels within the synaptic cleft is technically challenging, given the short time in which the free ion is present in the synapse (Hurst et al., 2010). Some authors have alternatively suggested that the ion does not diffuse in the cleft and is actually only externalized following exocytosis. In this view, Zn 2ϩ remains bound to the plasma membrane, forming a "veneer" on presynaptic terminals (Kay, 2003;Kay and Tó th, 2008). While this is an intriguing hypothesis, the interpretation of these results needs to take into account the variability in preexisting vesicular Zn 2ϩ levels, as these are known to be affected by changes in previous synaptic activity (i.e., sensory experience), the animal age, and the methods used in the preparation of brain slices (Frederickson et al., 2006;Nakashima and Dyck, 2009). Future studies combining electrophysiology with state-of-the-art synaptic Zn 2ϩ imaging are likely to give a more accurate description of the precise dynamics and concentrations of the ion during activity-dependent synaptic activity.
Exogenously applied Zn 2ϩ profoundly affects the activity of glutamate, GABA A , and glycine ionotropic receptors. Extracellular Zn 2ϩ therefore is likely to be intimately linked to the balance of excitation and inhibition in the brain. Indeed following stimulation of Zn 2ϩ -containing fibers, endogenous Zn 2ϩ has been shown to block postsynaptic NMDA (Vogt et al., 2000;Molnár and Nadler, 2001a) and GABA A (Ruiz et al., 2004) receptors. However, the modulation of postsynaptic receptors by Zn 2ϩ is likely complex, as other investigators have failed to find effects of vesicular Zn 2ϩ on GABA A receptors (Molnár and Nadler, 2001b) and neuronal excitability (Lopantsev et al., 2003;Lavoie et al., 2007).
More recent findings also indicate that synaptically released Zn 2ϩ activates a specific metabotropic Zn 2ϩ -sensing receptor ( Fig. 1B) (Besser et al., 2009;Chorin et al., 2011). Zn 2ϩ can flux into neurons and be taken up in organelles such as mitochondria (Sensi et al., 2000;Caporale et al., 2009;Dittmer et al., 2009), and recent evidence in non-neuronal cells indicates that some uptake might occur in the endoplasmic reticulum and the Golgi apparatus (Qin et al., 2011) as well. Neurons also keep cytosolic [Zn 2ϩ] i levels very low by using several ZnT transporters (the SLC30 family), ZIP transporters (the SLC39 family), and Zn 2ϩ buffering metallothioneins (MTs) as well as other transporters such as a putative Na ϩ /Zn 2ϩ exchanger ).
On the dark side, Zn 2ϩ is also a potent neurotoxin (Fig. 2) involved in variety of conditions that have been associated with excitotoxicity, including ischemia, epilepsy, and brain trauma . Zn 2ϩ promotes both neuronal and glial death in vitro and in vivo . Landmark in vivo studies have shown that the transsynaptic movement of Zn 2ϩ , a process also called "Zn 2ϩ translocation," plays a role in neuronal death associated with transient global ischemia (TGI) (Tønder et al., 1990;Koh et al., 1996). However, more recent evidence indicates that Zn 2ϩ mobilization from intracellular pools is also a crucial contributor to neuronal injury (Aizenman et al., 2000;Lee et al., 2000Lee et al., , 2003Hwang et al., 2008).
Finally, Zn 2ϩ can play an important role in the development of Alzheimer's disease (AD). The cation is a key component of amyloid plaques, and it is now suggested that Zn 2ϩ deregulation in the brain facilitates the synaptic deficits and cognitive decline observed in AD, while restoring brain Zn 2ϩ homeostasis may represent an important and novel therapeutic avenue (Corona et al., 2010) (Fig. 3A).
In this Mini-Symposium, we review novel, exciting, and sometimes controversial, findings that substantiate a major role for Zn 2ϩ in the physiological and pathological functioning of the brain.

Zn 2؉ and synaptic function
Free or "chelatable" Zn 2ϩ is concentrated within synaptic vesicles at glutamatergic terminals through the activity of the specific transporter ZnT3 . The role of synaptic Zn 2ϩ in regulating plasticity has been addressed either using extracellular metal chelators or in animal models lacking synaptic Zn 2ϩ (ZnT3 KO and the Mocha mutation), but the precise roles this ion plays in synaptic function is still controversial. While some studies suggest that long-term potentiation (LTP) and synaptic excitability in CA3 hippocampal neurons are unaffected by Zn 2ϩ under physiological conditions (Vogt et al., 2000;Lopantsev et al., 2003), others reports have shown the opposite (Li et al., 2001;Huang et al., 2008). In CA3 hippocampal neurons, Zn 2ϩ can directly promote the transactivation of the BDNFrelated TrkB pathway (Huang et al., 2008); however, TrkB signaling can also be activated by extracellular Zn 2ϩ in a metalloproteinase-dependent manner by releasing pro-BDNF and converting it to mature BDNF (Hwang et al., 2005). Paralleling these in vitro Zn 2ϩ -mediated neurotrophic effects, chronic dietary treatment with Zn 2ϩ has been found to induce an increase of brain levels of BDNF (Nowak et al., 2004;Corona et al., 2010) (Fig. 3B). It should be noted that in CA3 neurons, BDNF has been shown to activate a Zn 2ϩindependent Ca 2ϩ current that is mediated by TRPC3 channels (Li et al., 2010). Zn 2ϩ also modulates LTP in the amygdala through regulation of feedforward GABAergic inhibition (Kodirov et al., 2006). A direct role for synaptic Zn 2ϩ in learning and memory has been only recently uncovered. ZnT3 KO mice exhibit clear cognitive deficits, but only in animals aged beyond 6 months (Adlard et al., 2010). In other studies, 3-to 4-month-old ZnT3 KO animals show impaired contextual discrimination, spatial working mem- Figure 1. Zn 2ϩ and synaptic function. A, A high-affinity "Zn 2ϩ sensor" in NMDA receptors. NMDAR subunits contain in their extracellular region a tandem of clamshell-like domains, the N-terminal domain (NTD) and the agonist-binding domain, which is directly connected to the transmembrane pore region. In the GluN2A subunit, the NTD forms a discrete high-affinity Zn 2ϩ binding site that underlies allosteric inhibition of NMDAR-mediated synaptic currents by nanomolar Zn 2ϩ concentrations. B, Synaptic Zn 2ϩ activates a specific mZnR. Synaptic Zn 2ϩ released from the mossy fibers activates metabotropic Ca 2ϩ release via the ZnR. The expression of GPR39 (left panel) and synaptic Zn 2ϩ -dependent Ca 2ϩ release (right panel) are eliminated in the CA3 pyramidal cell layer in GPR39 KO mice. The activity of the mZnR triggers phosphorylation of ERK1/2 and regulation of Cl Ϫ transport, which lead to increased inhibitory drive. ory, or learned fear extinction (Martel et al., 2010;Martel et al., 2011;Sindreu et al., 2011).
Among the targets of synaptic Zn 2ϩ , NMDA receptors (NMDARs) are rather unique as they display very high sensitivity to extracellular Zn 2ϩ (Paoletti et al., 1997;Traynelis et al., 1998). At low nanomolar concentrations, Zn 2ϩ allosterically inhibits the activity of NMDARs containing the GluN2A subunit, a subunit that is widespread in the adult CNS. The cation acts on a discrete Zn 2ϩ -binding site located in the large bilobate N-terminal domain of the GluN2A subunit ( Fig. 1 A) (Paoletti et al., 2000). The Zn 2ϩ -GluN2A interaction likely mediates tonic inhibition of NMDARs by ambient Zn 2ϩ levels. At higher Zn 2ϩ concentrations (micromolar range), such as may occur during phasic synaptic release, Zn 2ϩ binds to the N-terminal domain of the GluN2B subunit, thereby inhibiting GluN2B-containing receptors (Rachline et al., 2005). The in vivo relevance of the highaffinity Zn 2ϩ inhibition of NMDARs has been recently addressed using a knock-in (KI) mouse line in which the GluN2A Zn 2ϩ site has been specifically eliminated. GluN2A-KI mice display a pronounced pain phenotype, showing both hypersensitivity to acute thermal and chemical nociception and enhanced allodynia in models of inflammatory and neuropathic pain (Nozaki et al., 2011). Moreover, in the KI animals, analgesia produced by exogenous Zn 2ϩ administrations is completely suppressed, revealing an essential role of the Zn 2ϩ -GluN2A interaction in the painrelieving effects of the cation (Nozaki et al., 2011).
Synaptic Zn 2ϩ also interacts with a selective metabotropic receptor, the mZnR, which has been recently identified as the previously orphan G-protein-linked receptor GPR39 ( Fig. 1 B) (Besser et al., 2009). Zn 2ϩ released from mossy fibers directly and specifically activates the mZnR in CA3 hippocampal neurons. The mZnR response is mediated by a G q -coupled pathway that triggers IP3 production, followed by Ca 2ϩ release from thapsigargin-sensitive stores. Subsequent activation of MAPK and CAMKII-dependent pathways has also been demonstrated. Employment of mice lacking GPR39 revealed that the ZnRdependent Ca 2ϩ response requires this protein. The mZnR is likely to modulate neuronal excitability as its activation enhances KCC2 function and Cl Ϫ efflux in postsynaptic neurons, thereby inducing a pronounced hyperpolarizing shift in the GABA A reversal potential (Chorin et al., 2011). Notably, this effect is absent in GPR39 KO or ZnT3 KO mice lacking mZnR or synaptic Zn 2ϩ , respectively. Thus, synaptic Zn 2ϩ -dependent enhancement of KCC2 function, via the mZnR, alters the Cl Ϫ gradient and may potentiate GABA A receptor-mediated inhibition. Synaptic Zn 2ϩ deficiency, induced by decreased dietary intake, acute chelation, or genetic manipulations, leads to enhanced susceptibility to seizures (Fukahori and Itoh, 1990;Cole et al., 2000;Blasco-Ibáñez et al., 2004). Based on these results and the documented role of KCC2 in epilepsy, we suggest that the mZnR may provide a crucial link between activity-dependent release of synaptic Zn 2ϩ and modulation of neuronal inhibition.

Zn 2؉ -dependent injury: a death by multiple cuts
Zn 2ϩ promotes neuronal death by affecting multiple systems. Zn 2ϩ can induce profound mitochondrial dysfunction by being sequestered in that organelle (for review, see Sensi et al., 2009). Zn 2ϩ can trigger mitochondrial depolarization and the generation of reactive oxygen species (ROS) (Sensi et al., 1999(Sensi et al., , 2000Dineley et al., 2005). These events, once investigated only in the neuronal somata, are now also identified in dendrites where they show a specific temporal pattern that differs from that found in the soma. This phenomenon may serve as primum movens for neuronal deafferentation and subsequent death (Medvedeva et al., 2009;Frazzini et al., 2011) (Fig. 2 A).
Furthermore, Zn 2ϩ induces a multiconductance cation channel activity in the inner mitochondrial membrane that is consistent with the activation of the mitochondrial permeability transition pore (mPTP; Jiang et al., 2001;Sensi et al., 2003;Gazaryan et al., 2007), thereby leading to release of pro-apoptotic mitochondrial proteins such as cytochrome C (Cyt-C) and apoptosis-inducing factor (AIF) (Jiang et al., 2001). Zn 2ϩ can also profoundly affect mitochondrial trafficking and morphology (Malaiyandi et al., 2005).
Zn 2ϩ deregulation favors neuronal death by increasing cytosolic oxidative stress through PKC (protein kinase C)-dependent activation of NADPH oxidase Noh et al., 1999) as well as by activating the neuronal isoform of nitric oxide synthase Kim and Koh, 2002). Moreover, Zn 2ϩ triggers a lethal depletion of neuronal ATP by inhibiting the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) enzyme, a process mediated by the reduction of cytosolic NAD ϩ (Sheline et al., 2000) and reverted by pyruvate (Sheline et al., 2000). Evidence also indicates that Zn 2ϩ -mediated NAD ϩ depletion requires the activation of the sirtuin pathway as SIRT proteins are NAD ϩ -catabolic protein deacetylases and sirtuin inhibitors are neuroprotective against both acute and chronic Zn 2ϩ -dependent toxicity, while, on the contrary, sirtuin activators promote NAD ϩ depletion and neuronal death (Cai et al., 2006). Zn 2ϩ also triggers autophagic neuronal death (Hwang et al., 2008). Lysosomes are the organelles in which autophagic degradation occurs. Lysosomal hydrolases released into the cytosol promote cell death by breaking down cellular components as well as by activating death inducers such as BID through a process termed lysosomal membrane permeabilization (LMP). Recently, evidence has indicated that LMP is a key contributing mechanism in oxidative-and Zn 2ϩ -induced hippocampal neuronal death (Hwang et al., 2008). Following exposure to H 2 O 2 or toxic levels of Zn 2ϩ , Zn 2ϩ rapidly accumulates in lysosomes and Zn 2ϩladen lysosomes undergo membrane disintegration releasing the toxic enzyme cathepsin. Exposure to the cell-permeable Zn 2ϩ chelator tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) not only completely blocks lysosomal Zn 2ϩ rises but also inhibits LMP. Thus, Zn 2ϩ deregulation may function as a linker between oxidative stress and LMP; however, what favors Zn 2ϩ accumulation in lysosomes under oxidative conditions is still completely unknown. Zn 2ϩ may enter lysosomes through Zn 2ϩ transporters, Zn 2ϩ -permeable channels, or exchangers. Alternatively, Zn 2ϩ may be released inside lysosomes from proteins, but the precise inter-organelle dynamics of Zn 2ϩ inside cells warrant further investigations.
Interestingly, Zn 2ϩ deprivation can also be a trigger for neuronal death and a converging set of evidence indicates that, most likely, neurons possess a finely tuned "Zn 2ϩ set-point." In this respect, even though chelation of deregulated [Zn 2ϩ ] i is neuroprotective, excessive depletion of [Zn 2ϩ ] i by high-affinity cellpermeable Zn 2ϩ chelators can be lethal . Zn 2ϩ depletion can enhance endonuclease activity  as well as counteract neuronal apoptosis by inhibiting Bax and Bak activation (Ganju and Eastman, 2003). In addition Zn 2ϩ chelation by TPEN promotes neuronal apoptosis by inducing caspase-11 and caspase-3 activation . TPEN-induced Zn 2ϩ depletion also favors a "dying-back" pattern of axon and dendrite degeneration due to Zn 2ϩ -mediated ATP depletion and autophagy (Yang et al., 2007).
The possibility of neuronal injury triggered by Zn 2ϩ depletion should be considered in light of the fact that Zn 2ϩ -chelating strategies have been proposed as therapeutic measures in the aftermath of an ischemic insult as well as for the treatment of AD (Corona et al., 2011). The idea of a Zn 2ϩ set-point is further substantiated within the context of AD as recent findings indicate that, in an AD mouse model, dietary supplementation of the cation starting as early as 1-month-old animals largely prevents the development of age-dependent mitochondrial dysfunction and hippocampal-dependent cognitive deficits and induces a potent increase in BDNF levels ( Fig. 3B) (Corona et al., 2010).

Zn 2؉ dyshomeostasis and ischemic neuronal injury
A series of recent studies has promoted a reevaluation of the "calcium-centric" hypothesis that has dominated the field of ischemic neuronal death in the past three decades (Choi, 1988). Real-time "single-cell" ionic imaging techniques have in fact shown that Ca 2ϩ works in synergy with Zn 2ϩ to promote ischemic death. Ca 2ϩ imaging in acute hippocampal slices undergoing oxygen and glucose deprivation (OGD; an in vitro model of brain ischemia) using a low-affinity Ca 2ϩ -sensitive probe has shown that the OGD-driven increase in the probe fluorescence is substantially stunted (by 70%) by TPEN, indicating that, at least in some models, ischemia promotes a parallel, and possibly interdependent, surge of [Ca 2ϩ ] i and [Zn 2ϩ ] i (Stork and Li, 2006).
Indeed, analyzing the ionic changes of CA1 pyramidal neurons exposed to OGD, a more recent study has dissected the interplay between the two cations. The study showed that, within few minutes after OGD induction, neurons undergo Ca 2ϩ deregulation and irreversible alteration of plasma membrane permeability (Medvedeva et al., 2009). Surprisingly, both processes are Figure 3. Zn 2ϩ in Alzheimer's disease. A, Zn 2ϩ released during neurotransmission is trapped by amyloid, depriving targets essential for LTP. In addition, the Zn 2ϩ transfers inappropriately 4 to APP and inhibits its ferroxidase activity and ability to facilitate iron release from neurons, leading to pro-oxidant intraneuronal iron accumulation as a downstream consequence of extracellular Zn 2ϩ accumulation. B, Zn 2ϩ supplementation is beneficial in an animal model of AD. 3xTg-AD mice chronically fed (11-13 months) with water containing 30 ppm of ZnSO 4 are protected from the appearance at 12-14 months of age of hippocampus-dependent memory deficits (as assessed with the Morris water maze test). Mice were tested when the platform was removed 1.5 h (a, left panel; to investigate short-term memory) and 24 h (a, right panel; to investigate long-term memory) after the last training trial. Zn 2ϩ -fed 3xTg-AD mice exhibited a marked recovery in their long-term memory as indicated by the decreased time (latency) they used to reach the point where the platform used to be. b, Zn 2ϩ supplementation promotes metalloproteinase (MMPs) activation in 3xTg-AD mice as shown by gelatin zymography indicating a significant increase of MMP-2 and MMP-9 induction in 3xTg-AD mice brains. c, BDNF immunoblotting reveals that Zn 2ϩ -fed 3xTg-AD mice showed a fourfold increase in BDNF levels compared to untreated mice (modified from Corona et al., 2010). Error bars indicate mean values Ϯ SEM. * indicates p Ͻ 0.05 in a and c and p Ͻ 0.01 in b. preceded by elevations in [Zn 2ϩ ] i and associated with mitochondrial Zn 2ϩ uptake as well as with depolarization and [Zn 2ϩ ] i chelation with TPEN results in delaying the phenomena (Medvedeva et al., 2009). A novel link explaining the synergistic deregulation of the two cations is offered by the downstream effects produced by glutamate and Ca 2ϩ influx on the acidification of the neuronal cytosol. Acidosis is a potent trigger for Zn 2ϩ mobilization from MTs and a recent study indicates that glutamate/ Ca 2ϩ influx leads to acidification of the neuronal cytosol, and that this is key to promote neuronal [Zn 2ϩ ] i rises (Kiedrowski, 2011). Synaptically released Zn 2ϩ can promote ischemic neuronal death (especially in TGI) by entering postsynaptic neurons through routes that are used by Ca 2ϩ , such as NMDARs and voltage-sensitive Ca 2ϩ channels (VSCC), however most Zn 2ϩ preferentially fluxes through Ca 2ϩ -and Zn 2ϩ -permeable GluA2lacking AMPA receptors (Ca/ARs; for review, see Sensi et al., 2009). Ca/ARs are highly expressed and dynamically upregulated after TGI (Pellegrini-Giampietro et al., 1997) on postsynaptic membranes in the dendritic tree of TGI-vulnerable neurons and their pharmacological inhibition prevents Zn 2ϩ influx and is highly neuroprotective in brain slices undergoing OGD or in animals exposed to TGI (Yin et al., 2002;Noh et al., 2005).
TGI-related apoptosis is also modulated by Zn 2ϩ as the cation induces mPTP opening in isolated postischemic mitochondria extracted immediately after TGI and these intramitochondrial Zn 2ϩ increases are linked to increased proteolytic cleavage of BCL-xL and the accumulation of the pro-apoptotic byproduct, deltaN-BCL-xL (Bonanni et al., 2006). Furthermore, the extracellular Zn 2ϩ chelator, clioquinol (CQ), is neuroprotective and decreases the expression levels of caspase-3 and -9 and AIF in the hippocampus of CQ-treated gerbils undergoing ischemia .
Neurons can be also killed by intraneuronal mobilization of the metal. Studies using ZnT3KO mice have, in fact, shown that glutamate-driven [Zn 2ϩ ] i accumulation can result from Zn 2ϩ released from sources such as MTs, mitochondria, and lysosomes (for review, see Sensi et al., 2009).
MTs are key players in excitotoxic and ischemic injury as they release Zn 2ϩ upon oxidative stress, a phenomenon occurring in neurons and glia (Aizenman et al., 2000;Malaiyandi et al., 2001Malaiyandi et al., , 2004 that can lead to both caspase-dependent and caspaseindependent forms of cell death (Aizenman et al., 2000;McLaughlin et al., 2001;Du et al., 2002). In the case of caspasedependent cell death, the liberated Zn 2ϩ triggers a signaling cascade that creates a permissive environment for the effective activation of proteases and nucleases. K ϩ is a key modulator of this process as cells undergoing caspase-mediated death develop an early, robust drop of intracellular K ϩ levels (Yu et al., 1997;Hughes and Cidlowski, 1999) reaching, in some cases, a final concentration of 50 mM . This loss of [K ϩ ] i favors the activation of caspases, including caspase 3, while the process and the subsequent neuronal death is inhibited when [K ϩ ] i is maintained at physiological levels Hughes et al., 1997;Yu et al., 1997). In cortical and midbrain neurons, K ϩ efflux is facilitated by a dramatic enhancement of delayed rectifier, voltage-activated K ϩ currents mediated by Kv2.1-encoded channels (Fig. 2 B) (McLaughlin et al., 2001;Pal et al., 2003;Redman et al., 2006). This K ϩ current surge results from the following series of events: (1) Zn 2ϩ -dependent activation of Src kinase in parallel to Zn 2ϩ -induced inhibition of protein tyrosine phosphatase (PTP), ensuring phosphorylation of Kv2.1 N-terminal tyrosine residue Y124 (Redman et al., 2009); (2) Zn 2ϩ -triggered activation of p38 MAPK via the MAP3K ap-optosis signaling kinase-1 (ASK-1; McLaughlin et al., 2001;Aras and Aizenman, 2005), resulting in the phosphorylation of Kv2.1 C-terminal serine residue S800 (Redman et al., 2007(Redman et al., , 2009, and (3) exocytotic, SNARE-dependent membrane insertion of the Y124/S800 dual-phosphorylated Kv2.1 channel, leading to enhanced K ϩ current densities (Pal et al., 2003(Pal et al., , 2006. Preventing the rise of intracellular Zn 2ϩ (McLaughlin et al., 2001), blocking any of the signaling steps along the pathway (McLaughlin et al., 2001;Pal et al., 2004Pal et al., , 2006Aras and Aizenman, 2011), or interfering with the functional expression or membrane insertion of Kv2.1 (Pal et al., 2003(Pal et al., , 2006 is sufficient to prevent neuronal death following oxidative, nitrosative, or chemical injury. Indeed, targeting the K ϩ current surge-signaling pathway might provide novel therapeutic strategies in neuroprotection (Aras and Aizenman, 2011).
Zn 2ϩ can also affect Cl Ϫ homeostasis as a recent study reported that [Zn 2ϩ ] i rises inhibit the activity of the K ϩ /Cl Ϫ cotransporter-2 (KCC2), the major Cl Ϫ outward transporter in neurons and thereby a key determinant of GABAergic neurotransmission . OGD-triggered [Zn 2ϩ ] i rises are followed by a profound KCC2 inhibition and a depolarizing shift in the GABA A reversal potential, a process reversed by intraneuronal Zn 2ϩ chelation. This process again indicates that [Zn 2ϩ ] i dyshomeostasis is an early and critical component of ischemic injury.
Adding a new angle to Zn 2ϩ -dependent ischemic neuronal loss, evidence indicates that the cation also promotes injury by inhibiting the ubiquitin-proteasome system (Chen et al., 2009). Recent evidence in brain slices undergoing OGD also indicates that Zn 2ϩ favors ischemic spreading depression, a wave of neuronal and glial depolarization that is thought to be a contributing factor in the enlargement of the infarct area (Carter et al., 2011).
Parenchymal acidosis is also a key modulator of Zn 2ϩ dyshomeostasis upon cerebral ischemia. Ischemic acidosis can increase Zn 2ϩ influx through VSCC and Ca/ARs and promote Zn 2ϩ release from MTs, thereby favoring an overall neurotoxic increase in [Zn 2ϩ ] i levels (Jiang et al., 2000;Sensi et al., 2003;Frazzini et al., 2007). As protons also block NMDARs, ischemic acidosis can therefore serve as a switch to decrease NMDAR-mediated neuronal death while potentiating injury triggered by the activation of VSCC and AMPARs. Data from cultured neurons indicate that, in fact, AMPAR activation promotes ROS-mediated [Zn 2ϩ ] i rises that are enhanced by mild acidosis (Frazzini et al., 2007). Interestingly, Zn 2ϩ can itself disrupt the neuronal acid-base equilibrium by blocking the Na ϩ /H ϩ exchanger, thereby creating a feedforward loop as the cation promotes intracellular acidification and also delays recovery from intracellular acidification (Dineley et al., 2002).
Zn 2؉ in Alzheimer's disease A␤ accumulation in the neocortex in AD is pathognomonic of AD, yet the mere production of this ubiquitously expressed 39 -43 residue peptide does not offer explanations for why amyloid only forms in the neocortex, why mice and rats do not develop amyloid pathology with age, or why women and APP transgenic mice have accelerated amyloid formation. The exceptional colocalization of A␤ and Zn 2ϩ in the glutamatergic synapses of the neocortex offers plausible explanations (Fig. 3A). Zn 2ϩ induces the rapid, but reversible, aggregation of A␤ into amyloid precipitates (Bush et al., 1994;Cherny et al., 1999), the pathological hallmark of AD. The rat/mouse A␤ possesses three amino acid substitutions that attenuate the interaction of Zn 2ϩ and prevent Zn 2ϩ -induced precipitation (Bush et al., 1994). As described earlier, Zn 2ϩ is released in a dissociable form by glutamatergic fibers in the cortex and hippocampus, and ZnT3 loads Zn 2ϩ into these synaptic vesicles. The distribution of ZnT3 expression closely approximates with the anatomical sites of A␤ deposition. ZnT3 is not appreciably expressed outside of the brain, and therefore the synaptic release of Zn 2ϩ in the neocortex is a cogent explanation for why A␤, which is released in the same vicinity, is liable to precipitate only in the brain. While several reports have found Zn 2ϩ to be enriched in extracellular amyloid deposits (Lovell et al., 1998;Lee et al., 1999;Miller et al., 2006;Adlard et al., 2008), this represents only a small fraction of the total cortical volume, and the tissue total Zn 2ϩ concentrations only rise during advanced pathology (Religa et al., 2006).
Genetic ablation of ZnT3 abolishes interstitial (Lee et al., 2002) and vessel-wall (Friedlich et al., 2004) amyloid pathology in transgenic mice overexpressing human A␤. The increase in the levels of soluble A␤ in the brains of the APP transgenic ϫ ZnT3 KO mice (Lee et al., 2002) confirmed that Zn 2ϩ holds the amyloid mass in a dissociable equilibrium (Huang et al., 1997). Ablation of ZnT3 also abolished the difference in genders for this mouse model in amyloid burden. Female mice have greater levels of dissociable Zn 2ϩ in this system (Lee et al., 2002), and ovariectomy raises hippocampal synaptic vesicle Zn 2ϩ levels further, whereas estrogen replacement opposed this rise .
As mentioned above, Zn 2ϩ may be a key modulator of synaptic activity and substrate for LTP. This may explain why ZnT3 KO mice develop a cognitive and memory loss by the age of 6 months, becoming a phenocopy for the cognitive loss seen in the AD model transgenic A␤ overexpressors (Adlard et al., 2010). Therefore, by trapping extracellular Zn 2ϩ , amyloid pathology may deprive these targets of physiological Zn 2ϩ and so contribute to downstream cognitive loss through a variety of mechanisms. At the same time, Zn 2ϩ flux through the NMDAR promotes the attachment of A␤ oligomers to the NR2B subunit, which may also impair LTP, but can be reversed by treatment with the Zn 2ϩ ionophore, CQ (Deshpande et al., 2009). This ionophoric mechanism that liberates Zn 2ϩ from A␤ oligomers, returning Zn 2ϩ to the relatively deficient neighboring cells, may explain the rapid benefits of PBT2 (an analog of CQ) on cognition and neurite outgrowth in AD animal and cell culture models (Adlard et al., 2008(Adlard et al., , 2011, as well as the rapid efficacy of the drug candidate in a phase 2 clinical trial of AD patients (Lannfelt et al., 2008;Faux et al., 2010).
The trapping of Zn 2ϩ by extracellular amyloid also impacts upon neuronal iron homeostasis. The amyloid protein precursor (APP) is a ferroxidase that catalytically loads Fe 3ϩ into transferrin, and is required for optimal iron export from neurons (Duce et al., 2010). Brain neuronal iron levels are increased in APP knock-out mice, as well as in AD, which provokes oxidative damage (Smith and Goldin, 1997;Duce et al., 2010). APP ferroxidase activity is 75% decreased in AD cortical tissue, caused by dissociation of Zn 2ϩ from amyloid, and not caused by a decrease in APP levels (Duce et al., 2010). Abnormal iron homeostasis can also have broad sequelae on heme synthesis, and is another of the downstream ramifications of Zn 2ϩ trapping by amyloid.
One major question to be answered is why extracellular Zn 2ϩ begins to react with soluble A␤ with advanced aging. Extracellular A␤ concentrations are elevated in uncommon familial AD mutations, but there is no evidence of an elevation with age in sporadic cases. The prediction is that extracellular Zn 2ϩ levels may rise with age. Zn 2ϩ coreleased with glutamate in the synapse must be, like glutamate, taken back into the cells by a very rapid transport with a pattern of Zn 2ϩ levels in the synaptic cleft that is likely to not be steady, but rather rapidly sinusoidal. There is no evidence for increased synaptic Zn 2ϩ in AD, but it is possible that Zn 2ϩ reuptake, which is energy dependent, may be fatigued with aging. Recent data have implicated the presenilins (PSs), whose mutations cause familial AD, in Zn 2ϩ uptake (Greenough et al., 2011). Together, these data indicate that PS may be able to influence A␤ aggregation by metal ion clearance in the extra-neuronal vicinity, which is currently being studied further.

Conclusions
Critical new findings have begun to uncover the many physiological roles for synaptically released Zn 2ϩ , as well as for intracellularly mobilized Zn 2ϩ , acting as an important player in the modulation of neuronal excitability and survival. New territories, however, need to be explored. For instance, the physiopathological activity of Zn 2ϩ in glial cells and how this is factored within the context of neuron-glia interaction requires further investigation. A more detailed road map of the regulatory processes that affect Zn 2ϩ homeostasis and Zn 2ϩ -dependent signaling is also needed. All these steps are crucial to find better pharmacological tools able to modulate cellular Zn 2ϩ . These drugs are urgently needed as they are likely to have an important impact in the management of major neurological conditions like AD, epilepsy, and stroke.