• Open Access

Calcium and neurodegeneration

Authors


Mark P. Mattson, Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center 4F01, 5600 Nathan Shock Drive, Baltimore, MD 21224, USA. Tel: 410 558 8463; fax: 410 558 8465; e-mail: mattsonm@grc.nia.nih.gov

Summary

When properly controlled, Ca2+ fluxes across the plasma membrane and between intracellular compartments play critical roles in fundamental functions of neurons, including the regulation of neurite outgrowth and synaptogenesis, synaptic transmission and plasticity, and cell survival. During aging, and particularly in neurodegenerative disorders, cellular Ca2+-regulating systems are compromised resulting in synaptic dysfunction, impaired plasticity and neuronal degeneration. Oxidative stress, perturbed energy metabolism and aggregation of disease-related proteins (amyloid β-peptide, α-synuclein, huntingtin, etc.) adversely affect Ca2+ homeostasis by mechanisms that have been elucidated recently. Alterations of Ca2+-regulating proteins in the plasma membrane (ligand- and voltage-gated Ca2+ channels, ion-motive ATPases, and glucose and glutamate transporters), endoplasmic reticulum (presenilin-1, Herp, and ryanodine and inositol triphosphate receptors), and mitochondria (electron transport chain proteins, Bcl-2 family members, and uncoupling proteins) are implicated in age-related neuronal dysfunction and disease. The adverse effects of aging on neuronal Ca2+ regulation are subject to modification by genetic (mutations in presenilins, α-synuclein, huntingtin, or Cu/Zn-superoxide dismutase; apolipoprotein E isotype, etc.) and environmental (dietary energy intake, exercise, exposure to toxins, etc.) factors that may cause or affect the risk of neurodegenerative disease. A better understanding of the cellular and molecular mechanisms that promote or prevent disturbances in cellular Ca2+ homeostasis during aging may lead to novel approaches for therapeutic intervention in neurological disorders such as Alzheimer's and Parkinson's diseases and stroke.

Calcium and neuronal death: the basics

Neurons are excitable cells that rapidly transfer electrochemical signals in a highly controlled spatio-temporal manner. A major intracellular messenger that mediates many physiological responses of neurons to chemical and electrical stimulation is Ca2+ (Fig. 1). The influx of Ca2+ through voltage-dependent and ligand-gated channels in the plasma membrane is a critical signal for the release of neurotransmitters from presynaptic terminals and for responses of the postsynaptic neuron (Yuste et al., 2000; Burnashev & Rozov, 2005; Hartmann & Konnerth, 2005). Glutamate, the major excitatory neurotransmitter in the central nervous system (CNS), induces an increase in the concentration of cytoplasmic Ca2+ by directly activating α-amino-3-hydroxy-5-methylisoxazole-4-propionate acid (AMPA) and N-methyl-d-aspartate (NMDA) receptor channels and by indirectly activating voltage-dependent Ca2+ channels (VDCC). Accordingly, antagonists of AMPA and NMDA receptors, or VDCC, have been reported effective in protecting CNS neurons against glutamate-mediated neuronal death (excitotoxicity) (Weiss et al., 1990; Mattson, 2003). In addition, activation of metabotropic glutamate receptors coupled to the GTP-binding protein Gq11 stimulates the release of inositol triphosphate (IP3), which activates Ca2+ channels in the endoplasmic reticulum (ER). Ca2+ is removed from the cytoplasm by the activities of the plasma membrane Na+/Ca2+ exchanger, plasma membrane and ER Ca2+-ATPases, and Ca2+-binding proteins such as calbindin and parvalbumin (Mattson et al., 1989a,b, 1991; Hof et al., 1999). By buffering intracellular Ca2+ loads, Ca2+-binding proteins such as calbindin may serve as endogenous anti-excitotoxic proteins (Mattson et al., 1991). Ca2+ can also be transported into and released from mitochondria (Werth & Thayer, 1994; Duchen, 2000; Nicholls et al., 2003).

Figure 1.

Subcellular systems involved in the disruption of neuronal Ca2+ homeostasis in aging and neurodegenerative disorders. Oxidative stress resulting from the aging process and disease-specific mechanisms causes peroxidation of lipids in the plasma membrane that impairs the function of ion-motive ATPases and glucose transporter proteins. This promotes membrane depolarization and cellular energy depletion, which results in excessive Ca2+ influx through glutamate receptor channels (GRC) and voltage-dependent Ca2+ channels (VDCC), and accumulation of Ca2+ within the cell. Perturbed Ca2+ homeostasis may also result from endoplasmic reticulum stress and mitochondrial dysfunction. Abnormal aggregation of proteins such as Aβ and α-synuclein likely contribute to the damage and dysfunction of Ca2+-regulating systems. Impaired function of the proteasome may contribute to the accumulation of protein aggregates throughout the cell. Excessive amounts of Ca2+ within the neuron can cause dysfunction of a myriad of cellular processes, including mitochondrial oxidative phosphorylation, protein production, and proper folding in the endoplasmic reticulum, and transcriptional regulation in the nucleus. Perturbed Ca2+ homeostasis may first adversely affect synaptic plasticity, followed by degeneration and death of neurons. Calcium dependent proteases (CDPs).

During normal physiological activity, the intracellular Ca2+ concentration increases only transiently (seconds to a few minutes) and has no adverse effects on the neurons. However, in pathological states, and more insidiously in normal aging, the ability of neurons to control Ca2+ fluxes and recover from a Ca2+ load is compromised. Oxidative stress is an important factor implicated in the disruption of neuronal Ca2+ homeostasis and neuronal death. Sources of oxidative stress include superoxide anion radical produced during oxidative phosphorylation; hydrogen peroxide, generated from superoxide in a reaction catalyzed by superoxide dismutases; hydroxyl radical, produced from hydrogen peroxide in a reaction catalyzed by Fe2+ and Cu+ and peroxynitrite, produced by the interaction of nitric oxide with superoxide.

Particularly disruptive to Ca2+ homeostasis is membrane-associated oxidative stress (MAOS), which typically involves lipid peroxidation caused by hydroxyl radical and peroxynitrite (Fig. 2). MAOS impairs the function of ion-motive ATPases (Na+/K+-ATPase and Ca2+-ATPase), and glutamate and glucose transporters, thereby promoting membrane depolarization and Ca2+ influx through NMDA receptors and VDCC (Mattson, 1998). Thus, oxidative stress renders neurons vulnerable to a form of Ca2+-mediated death called excitotoxicity in which glutamate receptors are overactivated even as the intracellular Ca2+ concentrations rise beyond tolerable levels (Arundine & Tymianski, 2003).

Figure 2.

Amyloid β-peptide induces membrane-associated oxidative stress, cellular Ca2+ overload and apoptosis. Upper panels: primary rat hippocampal neurons in culture were exposed to vehicle (left) or Aβ (right) for 24 h. Cells were then processed using a fluorescence-based method for visualization of sites of membrane lipid peroxidation. Neurons exposed to Aβ exhibited high levels of lipid peroxidation (yellow and red). Middle panels: cultured hippocampal neurons were exposed to Aβ for 24 h, an image showing intracellular free Ca2+ levels was acquired using fura-2 (left) and then the cells were immunostained with an Aβ antibody (right). The neuron with high amounts of Aβ immunoreactivity (pink) associated with its cell surface exhibited a very high concentration of intracellular Ca2+ (upper neuron). In contrast the neuron with little Aβ on its surface had a low intracellular Ca2+ concentration (lower neuron). Lower panels: cultured vascular endothelial cells were exposed to aggregated Aβ for 24 h, and were then stained with the DNA-binding dye bisbenzamide. Note that cells in contact with the Aβ exhibit apoptotic (condensed and fragmented) nuclear chromatin, whereas those not contacted by Aβ exhibit nuclei with normal appearing (diffuse and lightly stained) chromatin. These images were modified from Mark et al. (1997c), Mattson et al. (1993c), and Blanc et al. (1997).

An important role for Ca2+ release from the ER in excitotoxicity has been demonstrated in studies showing that blockers of the two different types of ER Ca2+ channels, IP3 receptors and ryanodine receptors, can protect neurons against excitotoxic injury (Frandsen & Schousboe, 1991; Mattson et al., 2000). In addition, mitochondria play important roles in the regulation of neuronal Ca2+ homeostasis (Werth & Thayer, 1994), and it has been shown that genetic and pharmacological manipulations that enhance mitochondrial Ca2+ sequestration can protect neurons against excitotoxicity (Keller et al., 1998; Duchen, 2000). Several proteins have been identified that are located in the ER or mitochondria that can protect neurons against Ca2+-mediated death, including glucose-regulated protein 78 (Yu et al., 1999), Bcl-2 (Guo et al., 1997) and Herp (Chan et al., 2004) in the ER, and manganese superoxide dismutase (Mn-SOD) (Keller et al., 1998), Bcl-2 (Murphy et al., 1996; Kruman & Mattson, 1999) and uncoupling proteins (Liu et al., 2006) in mitochondria.

How does Ca2+ kill neurons? Considerable evidence for several different, cross-amplifying, cascades has been obtained (Fig. 1). First, Ca2+ activates (either directly or indirectly) cysteine proteases called calpains and caspases that degrade a variety of substrates, including cytoskeletal proteins, membrane receptors and metabolic enzymes (Chan & Mattson, 1999; Nixon, 2003). Calpains may also play an important role in the triggering of apoptotic cascades by virtue of their ability to activate caspases (Leist et al., 1997; Stefanis, 2005). Second, Ca2+ induces oxidative stress (Lafon-Cazal et al., 1993; Mattson, 2003). This occurs through several different mechanisms, including activation of oxygenases such as those in the arachidonic acid metabolism cascade, perturbation of mitochondrial Ca2+ and energy metabolism, and induction of MAOS. The reactive oxygen species (ROS) generated in response to glutamate-induced Ca2+ influx include superoxide anion radical, hydrogen peroxide, hydroxyl radical, nitric oxide and peroxynitrite (Lipton et al., 1993; Mattson, 1998). Third, Ca2+ triggers apoptosis, a form of programed cell death (Ankarcrona et al., 1995). This might occur by Ca2+-mediated induction/activation of pro-apoptotic proteins such as Bax, Par-4, and p53 leading to mitochondrial membrane permeability changes, release of cytochrome c and caspase activation (Duan et al., 1999; Dargusch et al., 2001; Culmsee & Mattson, 2005).

Aging and perturbed neuronal calcium homeostasis

As in most other tissues, the normal aging process in the nervous system is associated with increased amounts of oxidative stress (Floyd & Hensley, 2002; Poon et al., 2004; Hyun et al., 2006) and perturbed cellular energy metabolism involving impaired mitochondrial function (Calabrese et al., 2001; Drew & Leeuwenburgh, 2004). Measurements of oxidative stress and mitochondrial function in brain cells of animals of different ages have demonstrated increases in oxidative modifications of proteins, lipids and DNA, and reduced mitochondrial function (activity levels of electron transport chain enzymes) during aging (Floyd & Hensley, 2002). Data from studies of patients, and animal and cell culture models, have established pivotal roles for oxidative stress and impaired energy metabolism in the pathogenesis of Alzheimer's disease (AD), Parkinson's disease (PD), Huntington's disease (HD), stroke, and related neurodegenerative disorders (for a review, see Gardian & Vecsei, 2004; Mattson, 2004; Przedborski & Ischiropoulos, 2005; Saito et al., 2005; Abou-Sleiman et al., 2006; Mattson & Magnus, 2006). The specific molecular alterations that result in oxidative stress may differ among disorders; for example, amyloid β-peptide (Aβ) in AD (Chan et al., 2002), mitochondrial alterations in PD (Greenamyre et al., 1999), mutant huntingtin in HD (Brustovetsky et al., 2005), and ischemia in stroke (Kristian & Siesjo, 1998). However, the reactive oxygen (and nitrogen) species produced, and their consequences for neuronal function and viability, are similar among disorders. Superoxide, hydrogen peroxide, hydroxyl radical, nitric oxide and peroxynitrite have each been implicated in all of the major neurodegenerative disorders. ROS can directly damage proteins and nucleic acids, which could perturb neuronal Ca2+ homeostasis. Examples include peroxynitrite-mediated modification of VDCC and NMDA receptors (Leist et al., 1997; Ohkuma et al., 2001), and oxidative damage to mitochondrial DNA (Lee et al., 2002a).

Neurons expend relatively large amounts of energy to maintain ion gradients across membranes; Na+/K+- and Ca2+-ATPases in the plasma membrane and Ca2+-ATPases in the ER membrane are major ATP sinks. During normal aging of the CNS, there is a decrease in the efficiency of mitochondria in generating ATP (Navarro, 2004). Together with oxidative stress and activation of glutamate receptors, prolonged membrane depolarization can deplete ATP from neurons, thereby compromising ion-motive ATPase activities and promoting Ca2+ overload (Collins et al., 1989). The latter mechanisms are perhaps best typified in ischemic brain injury where cellular Ca2+ overload is believed to play a key role in the death of neurons caused by a stroke. Accordingly, glutamate receptor antagonists, VDCC blockers, and intracellular Ca2+ chelators can protect neurons in experimental models of stroke (Weiss et al., 1990; Turski et al., 1991; Tymianski et al., 1993).

As mitochondria are the major source of both ATP and ROS in neurons, they represent an organelle that plays a pivotal role in determining whether or not neurons succumb to injury or disease during aging. The complex processes involved in mitochondrial energy, ROS and Ca2+ metabolism have been investigated (Mattson & Kroemer, 2003; Polster & Fiskum, 2004). Several mitochondrial proteins that may have particularly important roles in mediating or protecting against Ca2+-mediated neuronal death include pro-apoptotic (Bax and Bad) and antiapoptotic (Bcl-2 and Bcl-xL) proteins that modify mitochondrial membrane permeability (Beal, 1998a; Murphy & Fiskum, 1999; Springer et al., 2000; Camandola et al., 2005), and uncoupling proteins that decrease mitochondrial ROS production (Maragos & Korde, 2004; Liu et al., 2006). Mitochondrial alterations, including membrane depolarization and permeability, Ca2+ overload, and release of apoptotic proteins such as cytochrome c and apoptosis-inducing factor have been implicated in the pathogenesis of AD, PD, HD, amyotrophic lateral sclerosis (ALS), and stroke (Dawson & Dawson, 2003; Mattson, 2004).

Alterations in the ER have been associated with age-related dysfunction and degeneration of neurons. ER stress, characterized by the accumulation of misfolded proteins and uncontrolled Ca2+ release, is implicated in the dysfunction and death of neurons in a range of age-related neurodegenerative conditions (Lindholm et al., 2006). As with other aspects of neuronal Ca2+ dysregulation, oxidative stress and metabolic impairment promote ER dysfunction. In regards to specific neurodegenerative conditions, ER dysfunction has clearly been shown to contribute to the death of neurons following a stroke (Paschen, 2003).

Involvement of perturbed calcium homeostasis in neurodegenerative disorders

Studies of patients, and animal and cell culture models, have provided a wealth of data supporting the involvement of alterations in Ca2+ regulation in the pathogenesis of stroke and chronic neurodegenerative disorders. Stroke is a major cause of disability and death in elderly populations worldwide (Ingall, 2004). Evidence that perturbed cellular Ca2+ homeostasis is pivotal to the death of neurons following a stroke is strong. Neurons in the affected brain regions exhibit increased activities of Ca2+-dependent proteases, and inhibitors of calpains protect neurons against ischemic injury in animal models of stroke (Hong et al., 1994). Studies of animal and cell culture models have clearly demonstrated that neuronal Ca2+ overload occurs in neurons subjected to ischemia (Kristian & Siesjo, 1998). Activation of glutamate receptors and VDCC and impairment of Na+/Ca2+ exchangers and ion-motive ATPases contribute to neuronal Ca2+ overload and cell death after a stroke (Kristian & Siesjo, 1998; Nellgard & Wieloch, 1992).

Analyses of brain tissue from patients with a neurodegenerative disease have revealed evidence that alterations in cellular Ca2+ homeostasis contribute to the neurodegenerative process. For example, amounts of free and protein-bound Ca2+, and the activity of Ca2+-dependent proteases, are increased in neurons containing neurofibrillary tangles as compared with tangle-free neurons in brain tissue from AD patients (Murray et al., 1992; Nixon, 2003). Studies of experimental models of AD have shown that overactivation of glutamate receptors can induce changes in the cytoskeleton of neurons similar to those seen in neurofibrillary tangles (Mattson, 1990; Stein-Behrens et al., 1994). In AD, the proteolytic processing of the β-amyloid precursor protein (APP) is altered; cleavage by β- and/or γ-secretases is increased resulting in increased production of neurotoxic Aβ (Mattson, 2004). In the most common late-onset form of AD, it is likely that age-related increases in oxidative stress, metabolic impairment, and inflammatory processes contribute to increased Aβ production.

Αmyloid β-peptide disrupts neuronal Ca2+ homeostasis by generating ROS (hydrogen peroxide and hydroxyl radical) in a process catalyzed by Cu+ and Fe2+ (Hensley et al., 1994). Aβ accumulates on cell membranes and the ROS generated by Aβ oligomers induce MAOS, which can impair the function of membrane ion-motive ATPases (Na+/K+- and Ca2+-ATPases), and glucose and glutamate transporters. The latter actions of Aβ cause an elevation of basal intracellular Ca2+ levels and sensitize neurons to excitotoxicity and apoptosis (Fig. 2) (Mattson et al., 1992; Mark et al., 1995, 1997a,b; Keller et al., 1997). Aβ induces lipid peroxidation, which disrupts Ca2+ homeostasis by a mechanism involving the production of 4-hydroxy-2,3-nonenal (HNE), an aldehyde that can covalently modify proteins on cysteine, lysine, and histidine residues. HNE has been shown to modify and impair the function of ion-motive ATPases, the neuronal glucose transporter GLUT3 and the astrocyte glutamate transporter GLT-1 (Keller et al., 1997; Mark et al., 1997a,b; Blanc et al., 1998). Increased levels of HNE have been detected in association with degenerating neurons in tissue samples from patients with AD (Cutler et al., 2004) and even in the cerebrospinal fluid of AD patients (Lovell et al., 1997). In addition to inducing oxidative stress, Aβ oligomers may disrupt Ca2+ homeostasis by forming Ca2+-conducting pores in cell membranes (Kawahara & Kuroda, 2000).

Analyses of CNS tissues from patients with PD, HD, and ALS suggest a role for cellular Ca2+ overload in the death of the vulnerable neurons in these disorders. Excitotoxicity and excessive Ca2+-mediated nitric oxide production are believed to contribute to the death of dopaminergic neurons in PD (Beal, 1998b). Dopaminergic neurons expressing relatively high levels of the Ca2+-binding proteins calbindin and calretinin appear to be resistant to degeneration in PD (Yamada et al., 1990; Mouatt-Prigent et al., 1994). The medium spiny neurons that succumb in HD have low levels of Bcl-2 and parvalbumin and are highly vulnerable to excitotoxicity compared to other populations of striatal neurons (Liang et al., 2005). Medium spiny neurons from transgenic mice expressing mutant huntingtin exhibit enhanced mitochondrial Ca2+ overload when exposed to glutamate compared to medium spiny neurons from control mice (Tang et al., 2005). Evidence that perturbed Ca2+ homeostasis contributes to the demise of motor neurons in ALS includes laser-activated microprobe mass analysis of spinal cord tissue sections from ALS and control subjects documented a significant increase in the Ca2+ content of motor neurons in the ALS patients (Kasarskis et al., 1995); motor neurons resistant to ALS express high levels of Ca2+-binding proteins (Alexianu et al., 1994); motor neurons may die by an excitotoxic mechanism, and the only drug proven effective in slowing the progression of ALS (riluzole) targets excitotoxicity (Van Damme et al., 2005). In addition, MAOS occurs in neurons affected in PD, HD and ALS (Yoritaka et al., 1996; Beal, 1998b; Pedersen et al., 1998; Smith et al., 1998; Cutler et al., 2002) and, through the mechanisms described above, may disrupt neuronal Ca2+ homeostasis in each disorder.

Abnormal intracellular and extracellular accumulations of aggregated proteins are conspicuous features of neurodegenerative disorders that are strongly implicated in the disease process (Mattson & Sherman, 2003). Chronic inhibition of mitochondrial complex I results in aggregation of α-synuclein, which is associated with apoptosis of neural cells (Sherer et al., 2002). Transgenic mice expressing PD-causing mutations of α-synuclein exhibit intracellular aggregates of α-synuclein and neuronal degeneration in several brain regions (Lee et al., 2002c). Similarly, when mutant α-synuclein is expressed in cells of the substantia nigra of rats, dopaminergic neurons exhibited α-synuclein inclusions, neuritic degeneration and cell death (Lo Bionco et al., 2002). Mutations in α-synuclein and huntingtin may promote cell death through protein aggregation and impaired clearance by the proteasome (Moulder et al., 1999; Ostrerova-Golts et al., 2000; Bence et al., 2001). In contrast to Aβ, where the mechanism by which the aggregating peptide disrupts Ca2+ homeostasis is beginning to be elucidated, little is known about if and how mutations in α-synuclein, parkin, DJ-1, and huntingtin affect neuronal Ca2+-regulating systems. However, induction of oxidative stress and disruption of membrane integrity may be involved (Lee et al., 2002d; Kegel et al., 2005; Firdaus et al., 2006; Furukawa et al., 2006).

Genetic factors, Ca2+ dysregulation and neurodegeneration

The work of molecular geneticists has been invaluable in our efforts to understand the mechanisms responsible for neuronal dysfunction and death in age-related neurodegenerative disorders. Data obtained from studies of cultured cells and mice expressing disease-causing mutations support roles for perturbed neuronal Ca2+ homeostasis in the demise of neurons (Fig. 3). Some cases of inherited early onset AD are caused by mutations in APP, presenilin-1 or presenilin-2. APP mutations result in increased production of neurotoxic forms of Aβ that, as described above, disrupt neuronal Ca2+ homeostasis by inducing MAOS. Presenilin mutations that cause AD result in an abnormality in ER Ca2+ regulation characterized by an increased capacity of the ER to release Ca2+ in response to IP3 and Ca2+ influx through plasma membrane channels (Ca2+-induced Ca2+ release) (Guo et al., 1997, 1999). Stutzmann et al. (2006) provided convincing evidence that presenilin-1 mutations cause an increase in the level of ryanodine receptors and Ca2+ release through those channels in response to Ca2+ influx or release from IP3 stores. The consequences of the latter abnormality include increased vulnerability of neurons to excitotoxic, metabolic and oxidative insults (Guo et al., 1997, 1999; Mattson et al., 2000). Tau mutations that cause frontotemporal lobe dementia have been reported to disrupt neuronal Ca2+ homeostasis by enhancing Ca2+ influx through VDCC (Furukawa et al., 2003).

Figure 3.

Examples of genetic and environmental factors that have been shown to either disrupt (left) or stabilize (right) neuronal calcium homeostasis in the context of aging and neurodegenerative disorders. Modified from Mattson & Magnus (2006).

Individuals with the E4 isoform of apolipoprotein E are at increased risk of AD compared to those with E2 or E3 isoforms (Mahley et al., 2006). The mechanism by which apolipoprotein Es modify the neurodegenerative process in AD may involve, in part, effects on neuronal Ca2+ homeostasis. The apoE4 isoform (which increases the risk of AD) may amplify Ca2+ signaling directly (Hartmann et al., 1994; Tolar et al., 1999). It has also been suggested that apoE4 is less effective than E2 and E3 in scavenging 4-hydroxynonenal, suggesting that individuals with apoE4 are compromised in their ability to protect cells against age- and disease-related increases in membrane-associated oxidative stress (Pedersen et al., 2000).

Mutations in several different genes have been linked to early-onset inherited PD; they include α-synuclein, Parkin, and DJ-1 (Moore et al., 2005). PD α-synuclein mutations have been shown to perturb Ca2+ homeostasis in cultured neurons by increasing membrane ion permeability (Volles & Lansbury, 2002; Furukawa et al., 2006). Parkin has been shown to interact with the synaptic vesicle-associated protein synaptotagmin XI, suggesting a potential mechanism whereby Parkin mutations impair synaptic signaling (Huynh et al., 2003).

Huntington's disease is caused by trinucleotide (CAG) repeat expansions in the huntingtin gene resulting in polyglutamine repeat expansions in the huntingtin protein. Studies of cells from HD patients, and of cultured cells and transgenic mice expressing mutant huntingtin, support a role for perturbed Ca2+ homeostasis in the degeneration of neurons that occurs in HD (Bezprozvanny & Hayden, 2004). For example, Ca2+ regulation is abnormal in mitochondria from HD patient lymphoblasts and from brains of HD mice. Mutant huntingtin causes a potentiation of NMDA receptor activity in medium spiny striatal neurons from HD mice. In addition, mutant huntingtin binds to IP3 receptors and enhances their opening to release Ca2+ from the ER.

The selective vulnerability of spinal cord motor neurons in ALS is believed to result, in part, from their large size and their particular complement of AMPA receptors and VDCC (von Lewinski & Keller, 2005). The work of Simpson et al. (2002) have provided considerable evidence that cellular Ca2+ overload is of central importance in the death of motor neuron sin ALS. Mutations in Cu/Zn-SOD are responsible for some cases of inherited ALS (Bruijn et al., 2004). When expressed in cultured neurons and transgenic mice, the ALS Cu/Zn-SOD mutations render neurons vulnerable to oxidative stress and excitotoxicity (Kruman et al., 1999). ALS Cu/Zn-SOD mutations can impair synaptic glucose and glutamate transport and render neurons vulnerable to ischemic injury (Guo et al., 2000). Overexpression of the Ca2+-binding protein parvalbumin protects motor neurons and delays disease onset in Cu/Zn-SOD mutant mice (Beers et al., 2001). Calcium influx through AMPA receptor channels has been implicated in motor neuron damage in ALS, and may damage the neurons by causing the misfolding and aggregation of Cu/Zn-SOD (Tateno et al., 2004).

Diet, lifestyle and neuronal vulnerability

It has become apparent in recent years that the risk of developing a neurodegenerative disease as one ages can be modified depending on certain environmental factors including the quantity and molecular composition of food consumed, exercise (of the body and the brain), and exposure to environmental hazards (Fig. 3). In many cases, the same factors that increase the risk of type 2 diabetes and cardiovascular disease also increase the risk of neurodegenerative disorders. This is most apparent in stroke; risk is increased by diets high in calories and saturated fats, and a sedentary lifestyle (Ingall, 2004). However, overeating and lack of exercise may also be risk factors for AD and PD (Mayeux, 2003). In addition, cognitively challenging lifestyles may protect neurons against AD (Stern, 2006). Animal studies suggest cause–effect relationships between dietary and lifestyle factors and neuronal vulnerability to disease (Martin et al., 2006). For example, rats maintained on an intermittent fasting regimen exhibit reduced brain damage and improved functional outcome in a stroke model (Yu & Mattson, 1999). In addition, dietary energy restriction protects hippocampal neurons and improves cognitive function in animal models relevant to AD (Bruce-Keller et al., 1999; Zhu et al., 1999; Duan et al., 2001; Patel et al., 2005; Wang et al., 2005a). Dietary energy restriction has also been shown to protect dopaminergic neurons in mouse (Duan & Mattson, 1999) and monkey (Maswood et al., 2004) models of PD, and striatal and cortical neurons in an HD model (Duan et al., 2003). On the other hand, dietary energy restriction was detrimental (Pedersen & Mattson, 1999), whereas a high energy diet was beneficial (Dupuis et al., 2004) in mouse models of ALS.

Exercise and cognitive stimulation were reported to have beneficial effects on disease pathogenesis and functional outcome in some, but not all, animal models of neurodegenerative disorders. Studies of AD mouse models have suggested that environmental enrichment (cognitive stimulation) can reduce (Lazarov et al., 2005), exacerbate (Jankowsky et al., 2003), or have no effect on (Arendash et al., 2004) Aβ accumulation. However, environmental enrichment did improve cognitive function in multiple models (Arendash et al., 2004; Jankowsky et al., 2005), suggesting the possibility that cognitive stimulation may act downstream of Aβ production. In other studies, it was shown that voluntary exercise decreases Aβ accumulation in a mouse model of AD (Adlard et al., 2005). In regards to other neurodegenerative disorders, exercise and/or environmental enrichment have been shown to improve the outcome in animal models of stroke, PD and HD (Young et al., 1999; Endres et al., 2003; Spires et al., 2004).

Direct evidence that environmental factors impact on neurodegenerative processes by modifying neuronal Ca2+ homeostasis is lacking. However, inference and indirect evidence supports mechanism(s) involving Ca2+ regulation. Perhaps the strongest evidence comes from the literature concerning the effects dietary energy restriction and excess on neuronal vulnerability. As described above, dietary energy restriction protects neurons against damage in models in which Ca2+ plays a major role in the cell death process, including severe epileptic seizures (Bruce-Keller et al., 1999) and stroke (Yu & Mattson, 1999). Conversely, diabetes renders neurons vulnerable to Ca2+-mediated cell death (Li et al., 1998; Vannucci et al., 2001). Exercise and environmental enrichment have also been shown to modify the vulnerability of neurons to excitotoxicity and related forms of Ca2+-mediated death. In most cases, the effects of exercise and enrichment are beneficial (Endres et al., 2003; Gobbo & O'Mara, 2005), although detrimental effects have also been reported in some models (Ramsden et al., 2003).

How might diet and lifestyle affect neuronal Ca2+ homeostasis? One possibility is that the environmental factors modify neurotrophic factor signaling, which, in turn, affects Ca2+ homeostasis. For example, dietary energy restriction (Lee et al., 2002a,b; Maswood et al., 2004) and exercise (Smith & Zigmond, 2003; Vaynman & Gomez-Pinilla, 2006) can increase brain-derived neurotrophic factor (BDNF) and glial cell line-derived neurotrophic factor (GDNF) signaling. Conversely, BDNF levels were reported to be decreased in the sciatic nerve of diabetic rats (Rodriguez-Pena et al., 1995), and both BDNF and GDNF can ameliorate adverse effects of diabetes on the nervous system (Nitta et al., 2002; Anitha et al., 2006). Both BDNF and GDNF have been shown capable of protecting neurons against Ca2+-mediated death (Cheng & Mattson, 1994; Perez-Navarro et al., 1996; Tomac et al., 2002). Additional mechanisms by which dietary factors, exercise and cognitive stimulation may affect neuronal Ca2+ homeostasis are by modifying the expression of protein chaperones (heat-shock proteins, glucose-regulated protein 78, etc.) and mitochondrial uncoupling proteins (Lee et al., 1999; Yu & Mattson, 1999; Liu et al., 2006).

Hormesis and the stabilization of neuronal Ca2+ homeostasis

Hormesis is a general term describing any process in which exposure of a cell or organism to a moderate level of stress (oxidative, metabolic, thermal, etc.) results in adaptive changes in the cell/organisms that confer resistance to more severe stress that would otherwise cause damage, disease and/or death (Arumugam et al., 2006; Mattson & Cheng, 2006). The heat-shock response (Giffard et al., 2004), ischemic preconditioning (Kirino, 2002) and electroconvulsive shock therapy (Altar et al., 2004) are examples of hormesis. In the following paragraphs, I describe examples of how neuronal Ca2+ homeostasis can be enhanced through hormesis-based mechanisms; a schematic representation of this concept is shown in Fig. 4.

Figure 4.

Hormesis-based Ca2+ stabilization mechanisms. Environmental factors such as dietary energy restriction, exercise, and cognitive stimulation impose a mild and beneficial stress on neurons. Within the neurons, this stress manifests as reduced energy availability or increased energy demand, increased oxyradical production, and/or increased ionic fluxes. The latter types of stress result in the activation of kinases (RTK, receptor tyrosine kinases; MAPK, mitogen-activated protein kinases; Akt, xxxxxxxx; PKA, cyclic AMP-dependent protein kinase; PKC, protein kinase C). Kinases activate (or inhibit) transcription factors (Nrf-2, nuclear factor E2-related factor 2; NF-κB, nuclear factor kappa of B cells; CREB, cyclic AMP response element binding protein). Such transcription factors modify the expression of various genes that mediate adaptive responses of the neuron to the stress, including antioxidant enzymes (AOE), protein chaperones, and neurotrophic factors. Among the proteins whose function (through phosphorylation by kinases) and levels (through transcriptional regulation) affected by mild stress are those involved in regulation of cellular Ca2+ homeostasis, including glutamate receptor channel proteins (GRC), voltage-dependent Ca2+ channels (VDCC), and Ca2+-binding proteins.

Because of the potential for Ca2+ to damage and kill neurons, several different mechanisms have evolved to prevent Ca2+ overload. Not surprisingly, Ca2+ itself activates hormetic pathways. One Ca2+-stabilizing neuroprotective mechanism involves activation of the transcription factor CREB (cyclic AMP response element binding protein) by Ca2+ and calmodulin. CREB induces the expression of genes that encode proteins that promote neuronal survival and plasticity, including BDNF (Soriano et al., 2006). Nuclear factor-κB (NF-κB) is a transcription factor activated in response to Ca2+ release from the ER; NF-κB has been shown to stabilize Ca2+ homeostasis by modifying the expression of glutamate receptor subunits (Furukawa & Mattson, 1998) and down-regulating the expression of IP3 receptor channels (Camandola et al., 2005). NF-κB is also activated by some cytokines and neurotrophic factors that are released from neurons and glial cells when they are under stress; examples include tumor necrosis factor (Barger et al., 1995; Mattson et al., 1997), BDNF (Cheng & Mattson, 1994), nerve growth factor (Carter et al., 1996) and a secreted form of APP (Barger & Mattson, 1996). Several different neurotrophic factors induced by mild hormetic stress have been shown to protect neurons against excitotoxicity and other forms of Ca2+-mediated death, in part, by modifying the expression of Ca2+-regulating proteins, including AMPA and NMDA receptor subunits, antioxidant enzymes, and Ca2+-binding proteins (Cheng & Mattson, 1991, 1992; Mattson et al., 1993b, 1995; Cheng et al., 1995; Fawcett et al., 2000).

The hormesis mechanisms described in the previous paragraph typically function over time periods of hours to days. However, more rapid neuroprotective responses to Ca2+ influx have been described. For example, phosphorylation of glutamate receptor channels and VDCC by the Ca2+-dependent kinases has been shown to modulate Ca2+ influx (Bartschat & Rhodes, 1995; Lu et al., 2000; Leonard et al., 2002). Another rapid response mechanism involves activation of the Ca2+-responsive actin-severing protein gelsolin (Furukawa et al., 1997). Actin filaments enhance opening of NMDA receptor channels and VDCC by reducing channel rundown. Sustained Ca2+ influx results in depolymerization of actin filaments, resulting in channel rundown and a reduction in Ca2+ influx. Actin filaments may also regulate Ca2+ release from the ER (Wang et al., 2002).

Other hormesis pathways that are not directly activated by Ca2+ may also stabilize neuronal Ca2+ homeostasis. It is well established that exposure of neurons to a mild metabolic stress, such as occurs during a brief ischemia, can protect them against excitotoxicity and other Ca2+-mediated neurodegenerative processes. This type of metabolic hormesis is mediated by transcription factors such as hypoxia inducible factor 1 and production of cytoprotective proteins such as basic fibroblast growth factor and erythropoietin (Grimm et al., 2002; Liu et al., 2005). Exposure of neurons to subtoxic levels of noxious chemicals is increasingly recognized as a means of inducing hormesis. This has been well established for mitochondrial toxins such as cyanide (Jensen et al., 2002), but may also be a major mode of action of many of the health-promoting phytochemicals present in vegetables and fruits (Mattson & Cheng, 2006). Examples include sulforaphane (present at high levels in broccoli) (Gao & Talalay, 2004), curcumin from tumeric root (Wang et al., 2005b), and resveratrol from red grapes (Parker et al., 2005). Pathways that may mediate the beneficial effects of these phytochemicals in neurons include Nrf-2 – ARE (antioxidant response element pathway) and histone deacetylases called sirtuins.

Targeting Ca2+ regulation to prevent and treat neurodegenerative disease

The findings described above and elsewhere (Mattson, 2003; Braunewell, 2005) suggest that therapeutic approaches that stabilize neuronal Ca2+ homeostasis may be capable of retarding or preventing neuronal degeneration in acute and chronic neurodegenerative conditions. Although many different drugs that target Ca2+ influx have demonstrated efficacy in animal models of stroke, AD and other neurodegenerative disorders, very few have been successful in clinical trials. However, significant beneficial effects of the L-type Ca2+ channel blocker nimodipine (Tollefson, 1990) and the NMDA open channel blocker memantine (Bullock, 2006) in AD patients, and of the glutamate-modulating agent riluzole in ALS patients (Miller et al., 2003) have shown that Ca2+-regulating systems are viable targets. However, drugs that suppress Ca2+ influx may compromise the normal functions of neurons, which is a particularly important concern in neurodegenerative disorders that require long-term treatment including AD, PD and ALS. Therapeutic strategies aimed at enhancing endogenous Ca2+ removal and buffering mechanisms would be expected to have fewer side-effects. In this regard, activation of neurotrophic factor-signaling pathways, as described above, might prove beneficial.

Perhaps the most exciting development in therapeutics for neurodegenerative disorders has been the possible use of immunization. Data obtained from studies of mouse models of AD, and of AD patients, suggest that Aβ aggregation and accumulation in the brain can be modified by the immune system. Immunization of APP mutant mice with human Aβ results in reduction in the accumulation of Aβ in, and removal of aggregated Aβ from, the brain (Gelinas et al., 2004). In an initial clinical Aβ vaccination trial, a few of the patients developed severe CNS inflammation, but some of the patients who tolerated the treatment appeared to benefit as indicated by magnetic resonance imaging analyses and cognitive testing (Fox et al., 2005; Gilman et al., 2005). Other studies have shown that, depending upon the particular Aβ antibody, and presumably the epitopes with which it interacts, Aβ antibodies may either inhibit peptide aggregation and toxicity, or they may exacerbate the toxicity of the peptide (Mattson & Chan, 2003). Accordingly, clinical trials are in progress in which AD patients are treated with well-characterized Aβ antibodies (passive immunization), an approach that is expected to have fewer side-effects and more consistent therapeutic efficacy.

As with other major age-related diseases, the burden of neurodegenerative disorders may best be lessened through the implementation of preventative changes in diet and lifestyle. As described above, a reduction in dietary energy intake, exercise and cognitive stimulation may each enhance the ability of neurons to control Ca2+ fluxes during aging. A better understanding of how this is accomplished at the cellular and molecular levels may identify additional targets and strategies for neuroprotective therapeutic interventions.

Acknowledgments

This work was supported by the National Institute on Aging Intramural Research Program of the National Institutes of Health.

Ancillary