I. INTRODUCTION

Top Previous Next References

Many persons live for nine or more decades and enjoy a well-functioning brain until the very end of life. We therefore know what the brain is capable of and, accordingly, a major goal of research in the area of the neurobiology of aging is to identify ways to facilitate successful brain aging in everyone. Studies of brains of the oldest old have provided evidence for stability as well as plasticity in successful brain aging (Fig. 1). In many brain regions, there is very little or no decrease in numbers of neurons, while in some brain regions neuronal loss may occur but may be compensated by expansion of dendritic arbors and increased synaptogenesis in the remaining neurons (19). It is thought that many neurons remain in the brain for a lifetime, although in some brain regions such as the olfactory bulb and dentate gyrus of the hippocampus, there may be a continuous replacement of neurons from a pool of progenitor (stem) cells (91, 291). This regenerative capacity of some brain regions may persist throughout life. Changes in the cellular structure of the brain and the functions of its neuronal circuits are controlled by an intricate array of intercellular signaling molecules and intracellular signal transduction pathways. Several such cellular signal transduction systems are altered during brain aging. Examples of widely used signaling mechanisms affected by aging include protein phosphorylation (alterations in kinases and phosphatases) (150), cellular calcium homeostasis (215), and gene transcription (180). Among neurotransmitter systems, dopaminergic signaling appears to be consistently altered during aging with a progressive decrease in signaling via the D2 subtype of receptor (303). In addition to signaling pathways, cellular systems that regulate protein folding (chaperone proteins) and degradation (proteasomal and lysosomal systems) are altered in brain cells during aging (158) (Fig. 2). These kinds of alterations that occur during normal aging may set the stage for catastrophic neurodegenerative disorders that may be triggered by particular genetic predispositions or environmental factors, while other age-related changes may represent adaptive protective responses to the aging process.

View larger version (48K): [in this window] [in a new window]   Fig. 1. During aging there is a progressive accumulation of damaged molecules and impaired energy metabolism in brain cells. Neurons and glial cells may adapt to the adversities of aging by increasing their ability to cope with stress, compensating for lost or damaged cells by producing new neurons and glia, and remodeling neuronal circuits. If adaptation is not successful, then molecular damage to neurons and inflammatory processes result in synaptic dysfunction and neuronal degeneration and death.

View larger version (63K): [in this window] [in a new window]   Fig. 2. Mechanisms involved in regulating protein turnover and their modification by cellular stress. Proteins damaged by oxidative stress or other modifications can be degraded by the proteasomal or lysosomal systems. Protein chaperones such as heat shock proteins (HSP40, HSP70, and HSP90), glucose-regulated proteins (GRP78 and GRP94), and ubiquitin play important roles in controlling protein folding and targeting proteins for proteolytic degradation. ER, endoplasmic reticulum.

Studies of embryonic and early postnatal development, and of synaptic plasticity in the young adult, have made a major contribution to our current understanding of the molecular and cellular mechanisms that determine whether brain aging occurs successfully or manifests dysfunction or disease. This is because the same intercellular signals and intracellular transduction pathways that regulate neurite outgrowth, synaptogenesis, and cell survival during development are also operative throughout life (213). Although new signaling systems continue to be discovered, the major classes of signaling molecules important in brain aging include neurotrophic factors, neurotransmitters, cytokines, and steroids. Neurotrophic factors such as neurotrophins [brain-derived neurotrophic factor (BDNF), nerve growth factor (NGF), neurotensin (NT)-3, and NT-4], fibroblast growth factors, glial-derived neurotrophic factor (GDNF), and ciliary neurotrophic factor (CNTF) have been shown to promote the survival of specific populations of brain neurons under experimental conditions relevant to brain aging and neurodegenerative disorders (232). In addition to their roles as mediators of synaptic transmission, neurotransmitters such as glutamate, acetylcholine, and dopamine also play important roles in regulating the formation of neuronal circuits during development and in influencing the neurodegenerative process in brain disorders of aging (215). Sex steroids (estrogen and testosterone) and stress steroids (glucocorticoids) have been shown to have quite striking effects on brain function and structure, and alterations in regulation of their production and signaling mechanisms have been reported to occur during aging (239). The status of such neurotransmitter, trophic factor, cytokine, and steroid hormone signaling systems is likely to have a major influence on the outcome of brain aging.

While the brain can age successfully, its cells may face considerable adversity during the journey (Fig. 1). Increased oxidative stress (oxyradical production) and accumulation of oxidatively damaged molecules (proteins, nucleic acids, and lipids) promote dysfunction of various metabolic and signaling pathways (178). Neurons may also face energy deficits as the result of alterations in the cerebral vasculature and in mitochondrial function (131). As in other organ systems, cells in the brain encounter a cumulative burden of oxidative and metabolic stress that may be a universal feature of the aging process. Each of the major classes of cellular molecules, including proteins, nucleic acids, and lipids, is oxidatively modified during brain aging. Protein modifications include carbonyl formation (34, 35, 74); covalent modification of cysteine, lysine, and histidine residues by the lipid peroxidation product 4-hydroxynonenal (261, 268, 266); nitration on tyrosine residues (326); and glycation (249). DNA and RNA bases are subject to oxidative modification, with a prominent example being the formation of 8-hydroxydeoxyguanosine (331). Double bonds in membrane lipids are oxidized resulting in the production of a variety of lipid peroxides and aldehydes (218). These modifications of proteins, nucleic acids, and lipids are greatly exacerbated in neurodegenerative disorders such as Alzheimer's disease (AD) and Parkinson's disease (PD) consistent with a major role for oxidative stress of aging in the pathogenesis of those disorders (207).

Analyses of experimental animal and cell culture models of age-related neurodegenerative disorders have provided insight into the mechanisms that result in increased oxidative stress and damage to proteins, nucleic acids, and membrane lipids (Fig. 3). The pathogenesis of AD involves altered proteolytic processing of the -amyloid precursor protein (APP) resulting in increased production of a long (42 amino acid) form of amyloid -peptide which self-aggregates and forms insoluble plaques in the brain (216). As amyloid -peptide aggregates, it generates reactive oxygen species that can induce membrane lipid peroxidation in neurons resulting in the impairment of the function of membrane ion-motive ATPases and glucose transporter proteins, which, in turn, disrupts cellular ion homeostasis and energy metabolism (216). These oxidative actions of amyloid -peptide can cause dysfunction of synapses and may render neurons vulnerable to excitotoxicity and apoptosis (80, 219). In PD, degeneration of dopaminergic neurons in the substantia nigra may be triggered by mitochondrial impairment and increased oxidative stress resulting from aging and exacerbated by environmental factors (147). In the common late-onset forms of AD and PD, the neurodegenerative cascade is most likely promoted by environmental factors (see sects. IV and V) that result in increased oxidative and metabolic stress. On the other hand, more rare inherited forms of these disorders in which disease onset occurs at an early age (30-60 years of age) are caused by specific mutations; for example, mutations in APP and presenilins that cause AD (122) and mutations in -synuclein and parkin that cause PD (165, 281). Some neurodegenerative disorders are purely genetic including Huntington's disease (HD) and related trinucleotide repeat disorders (389); such disorders may not, therefore, be considered as diseases of aging, although aging processes may affect the age of disease onset and clinical course.

View larger version (26K): [in this window] [in a new window]   Fig. 3. Examples of sources of oxidative stress in neurons during aging and examples of molecules damaged by free radical-mediated processes. A major source of oxyradicals is mitochondria, in which superoxide anion radical (O·) is produced during oxidative phosphorylation. Superoxide is converted to hydrogen peroxide (H2O2) via the actions of mitochondrial manganese superoxide dismutase (Mn-SOD) and cytosolic Cu/Zn-SOD. Hydrogen peroxide is eliminated from cells by conversion to water in reactions catalyzed by catalase and glutathione peroxidases. However, hydrogen peroxide is an important source of hydroxyl radical (OH·) which is generated in the Fenton reaction which involves Fe2+ or Cu2+. Hydroxyl radical is a potent inducer of membrane lipid peroxidation. Oxyradicals can also be generated in response to calcium influx via the activation of nitric oxide (NO) synthase, resulting in NO production; NO can interact with superoxide to produce peroxynitrite. In addition, various oxygenases can be activated by calcium resulting in superoxide production. Oxyradicals (particularly hydroxyl, superoxide, and peroxynitrite) can damage proteins, lipids, and nucleic acids. Lipid peroxidation products such as 4-hydroxynonenal (HNE) can covalently modify proteins and impair their function. Arach acid, arachidonic acid; CaM, calmodulin; depol, depolarization; ER, endoplasmic reticulum; GSH, glutathione; GSHPx, glutathione peroxidase; GSHR, glutathione reductase; GSSG, reduced glutathione; LP, lipid peroxidation; PLA2, phospholipase A2; NOS, NO synthase; PS1, presenilin-1; PT, permeability transition pore; SOD1, Cu/Zn-superoxide dismutase; SOD2, manganese superoxide dismutase.

In the United States and other industrialized countries, life expectancy continues to increase, and therefore, more people will suffer from age-related neurodegenerative conditions. The negative impact of age-related neurodegenerative disorders on our societies is emphasized by the fact that more dollars are required to care for patients with AD, PD, and stroke than are spent on the combined care for patients with cardiovascular disease or cancer. Each neurodegenerative disorder is characterized by dysfunction and degeneration of specific populations of neurons in the brain (246). Neurons in brain regions involved in learning and memory processes, such as the hippocampus and cerebral cortex, are afflicted in AD. In PD, dopaminergic neurons in the substantia nigra degenerate resulting in motor dysfunction (147). A stroke occurs when a cerebral blood vessel becomes occluded or ruptures resulting in the degeneration of neurons in the brain tissue supplied by that vessel (65). Several genetic and environmental factors that may initiate the neurodegenerative process in AD, PD, and stroke have been identified, and this information has led to the development of valuable animal models of these disorders. Animal models of AD include transgenic mice overexpressing mutant forms of human APP (93, 133), transgenic and knockin mice expressing mutant forms of human presenilin-1 (PS1) (75, 108), and infusion of amyloid -peptide and excitotoxins into the brains of rats and mice (31, 97). Animal models of PD include administration of the toxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) to monkeys and mice resulting in selective degeneration of substantia nigra dopamine-producing neurons and associated motor dysfunction (71), and transgenic mice expressing mutant human -synuclein which exhibit degeneration of dopaminergic neurons and a behavioral phenotype with features similar to PD (210). Stroke models involve transient or permanent occlusion of the middle cerebral artery in rats and mice (65, 378). The mechanisms that result in neuronal dysfunction and/or death in these models are beginning to be understood, with increased oxidative stress, perturbed energy metabolism, altered calcium homeostasis, and activation of apoptotic cascades playing important roles in most cases (219). Data obtained using these various models have provided valuable insight into the cellular molecular mechansims of neurodegenerative disorders and will therefore be cited throughout the remainder of this article.

	II. ADAPTIVE CELLULAR AND MOLECULAR RESPONSES IN BRAIN AGING

Top Previous Next References

A goal of basic and clinical neuroscientists is to identify approaches for promoting the maintenance of cognitive, emotional, motor, and sensory functions throughout the life span.

This could be accomplished by avoiding genetic (through genetic counseling or germline gene therapy, for example) and environmental (dietary and behavioral factors, for example) factors that facilitate neuronal dysfunction and death, or by enhancing the ability of neurons to adapt to the aging process. Basic research is identifying cellular signaling mechanisms that promote cell survival, neurite growth, and/or synapse formation/plasticity; understanding these signaling pathways may reveal ways of promoting successful brain aging. Clinicians, geneticists, and epidemiologists should therefore work together to identify genetic and environmental factors that cause or affect risk of age-related neurological disorders.

Intercellular signaling mechanisms mediate the second-to-second functions of neuronal circuits as well as long-term changes in the biochemistry and structure of those circuits. Three major classes of intercellular signaling proteins that regulate neuronal survival and synaptic plasticity are neurotransmitters, neurotrophic factors, and hormones. Glutamate and GABA, the major excitatory and inhibitory neurotransmitters in the brain, play pivotal roles in regulating neuronal survival (231) and synaptic plasticity (18). By inducing the expression of neurotrophic factors such as BDNF, glutamate can promote neuronal survival (203). On the other hand, overactivation of glutamate receptors can cause neuronal death, particularly under conditions of increased levels of oxidative and metabolic stress, as occurs during aging and in age-related neurodegenerative disorders (224). By reducing neuronal excitability, GABA can protect neurons in experimental models of neurodegenerative disorders (231). Other neurotransmitters that can modify neuronal vulnerability in cell culture and animal models of neurodegenerative disorders include acetylcholine, dopamine, norepinephrine, and serotonin (89, 214, 288). During brain aging, these neurotransmitters may contribute to either degeneration or adaptive responses of neurons.

Neurotrophic factors promote the survival, outgrowth, and/or synaptogenesis of neurons. Examples of neurotrophic factors that have been shown to counteract untoward aspects of aging (oxidative stress and disturbed ion homeostasis, for example) include basic fibroblast growth factor (bFGF), NGF, BDNF, NT-4, transforming growth factor- (TGF-), tumor necrosis factor (TNF), and GDNF. These neurotrophic factors can protect one or more populations of brain neurons against excitotoxic, oxidative, and metabolic insults in models of stroke, AD, PD, and HD (232, 233). Signaling by cell adhesion proteins such as integrins may also play important roles in modulating neuronal survival (94). Growth factors, cytokines, and integrin ligands promote neuronal survival by inducing the expression of genes that encode proteins that suppress oxidative stress, stabilize cellular calcium homeostasis, and antagonize neurodegenerative biochemical cascades. Examples of three neuroprotective signaling cascades, activated by BDNF, bFGF, and the secreted form of APP are shown in Figure 4. bFGF, BDNF, TNF, and NGF can increase the production of one or more antioxidant enzymes [Cu/Zn-superoxide dismutase (SOD), Mn-SOD, glutathione peroxidase, and catalase] in hippocampal neurons (233, 229). NGF, BDNF, and TNF can induce expression of anti-apoptotic Bcl-2 family members (32) and inhibitor of apoptosis proteins (IAP) (364). Neurotrophic factors can also modulate the expression and/or activity of subunits of glutamate receptors and voltage-dependent calcium channels in ways that promote neuronal survival and synaptic plasticity (217, 225, 360). Kinases such as mitogen-activated protein (MAP) kinase and protein kinase C, and transcription factors such as NF-B and CREB, transduce the cell survival signals of neurotrophic factors and cytokines.

View larger version (21K): [in this window] [in a new window]   Fig. 4. Examples of cellular signaling pathways that modulate neuronal plasticity and survival during aging. Neurotrophic factors such as brain-derived neurotrophic factor (BDNF) activate membrane receptor tyrosine kinases that initiate kinase signaling cascades that ultimately regulate the expression of genes that encode proteins which enhance neuronal survival and plasticity. Such gene targets include those encoding proteins that suppress apoptotic cascades, reduce oxidative stress, and stabilize cellular calcium homeostasis. AKT/PKB, Akt kinase; CaMKIV, calcium/calmodulin-dependent protein kinase IV; CREB, cAMP response element binding protein; IRS1, insulin receptor substrate-1; MAPK, mitogen-activated protein kinase; MEK, MAP kinase kinase; PI3K, phosphatidylinositol 3-kinase; PKG, protein kinase G; R, receptor; sAPP, secreted form of amyloid precursor protein; SHC, src homology domain-containing adaptor protein; trkB, high-affinity BDNF receptor.

Another type of adaptive response that may protect neurons against the adversities of aging and disease is a stress response that involves protein chaperones that exhibit neuroprotective properties. Examples of such stress proteins include heat shock proteins (e.g., HSP-70, HSP-90, and HSP-60) and glucose-regulated proteins (e.g., GRP-78 and GRP-94). These chaperone proteins interact with many different proteins in cells and function to ensure their proper folding, on the one hand, and degradation of damaged proteins, on the other hand (86, 96). They may also interact with, and modify the function of, apoptotic proteins including caspases (14, 293). Levels of some of these chaperone proteins may be increased during aging as a protective response (180, 182). Cell culture and in vivo studies have shown that HSP-70 and GRP-78 can protect neurons against injury and death in experimental models of neurodegenerative disorders (197, 380). Interestingly, caloric restriction, a dietary manipulation that increases life span and brain "health span" (the time window of life during which the brain maintains a level of function that permits a productive life-style), can increase the expression of chaperone proteins in the brains of rats and mice (see sect. IV).

Synapses are sites where the actions of neurotrophic factors, stress proteins, and anti-apoptotic Bcl-2 and IAP family members may play particularly important roles in preserving the integrity and function of neuronal circuits (115, 228). The impact of aging is likely to be most severe in synapses, because these compartments are sites of repetitive calcium influx and oxyradical production; it is therefore of great importance to understand how genes and environment affect synaptic homeostasis (see sect. VI).

Research performed in many different laboratories during the past 10 years has revealed that regeneration/compensation can occur in the adult brain and that populations of stem cells or neural progenitor cells (NPC) may play a role in this process by dividing and then differentiating into neurons or glia (91). Various neurotrophic factors and cytokines may promote neurogenesis, neurite outgrowth, and synaptic recovery after brain injury (146, 232). Damage to axons and dendrites can result in regrowth of those processes; however, in contrast to the peripheral nervous system, the brain contains a number of inhibitory signals that may prevent successful reinnervation of target cells (312). If synaptogenesis does occur, it may or may not replace lost function depending on the degree of specificity of neuronal connections in the circuits involved. For example, reorganization of the morphology of hippocampal neurites and synapses after stress-induced damage correlates with behavioral improvement (333). On the other hand, many brain functions rely on "memories" based on the past history of synaptic activity, and such memories are unlikely to be restored by new synapse formation.

Stem cell biology is a rapidly growing area in the fields of neuroscience research and aging. Embryonic stem cells have received considerable attention because of their ability to form any type of cell in the body including neurons (99). They are therefore a potential cell source for replacement of neurons lost in neurodegenerative disorders. Two major populations of pluripotent NPC are present in the adult brain, one in the subventricular zone and the other in the subgranular layer of the dentate gyrus of the hippocampus (91). These NPC cells can give rise to either neurons or astrocytes, and there is increasing evidence that some of the progeny of the NPC survive and become functional, although many may undergo programmed cell death (Fig. 5). Newly generated cells in the brain can be identified by giving animals the thymidine analog bromodeoxyuridine (BrdU); the phenotype of their differentiated progeny can then be determined by double-labeling using antibodies against neuronal (e.g., neural cell adhesion molecule or ₃-tubulin) or astrocyte [glial fibrillary acidic protein (GFAP)] markers. Several signals that control the proliferation, differentiation, and survival of NPC have been identified (91, 291). bFGF and epidermal growth factor (EGF) can maintain NPC in a proliferative state, whereas BDNF and NT-3 can promote their differentiation and/or survival, and bone morphogenetic protein can induce NPC to become astrocytes (91, 241). Additional signals that control NPC cell fate include Notch (344), Numb (40), neurogenin (339), and the secreted form of APP (255). Brain injury is a potent stimulus for neurogenesis (194), and this effect is likely mediated by the trophic factors and cytokines induced by cell injury (232).

View larger version (20K): [in this window] [in a new window]   Fig. 5. Regulation of neurogenesis and gliogenesis. Neural stem cells capable of producing neurons and astrocytes are maintained in a self-replicating state by cytokines and neurotrophic factors such as leukemia inhibitory factor (LIF) and epidermal growth factor (EGF). Under the appropriate conditions, the stem cells can form neuron- or glia-restricted progenitor cells which, in turn, can cease dividing and differentiate into neurons or glia. New neurons may integrate into neuronal circuits or may die, and glia may also live or die. BDNF, brain-derived neurotrophic factor; bFGF, basic fibroblast growth factor; BMP, bone morphogenic protein; CNTF, ciliary neurotrophic factor; HB-EGF, heparin-binding epidermal growth factor; IL-6, interleukin-6; NGF, nerve growth factor; NT-3, neurotrophin-3; NT-4, neurotrophin-4; TGF-, transforming growth factor-; PDGF, platelet-derived growth factor; ROS, reactive oxygen species; sAPP, secreted form of amyloid precursor protein.

Considerable interest in fundamental mechanisms of cellular aging has arisen from studies of telomerase, a reverse transcriptase that adds a six-base DNA repeat onto the ends of chromosomes (telomeres) and thereby maintains their integrity during successive rounds of cell division (221). Expression of the catalytic subunit of telomerase (TERT) and telomerase activity are associated with cell immortalization and cancer and are absent from most somatic cells in the adult, suggesting an important role in the aging process. Indeed, expression of TERT in normal fibroblasts makes them immortal (without transforming them) (24). Telomerase is present in brain cells during development, where it is thought to play a role in the maintenance of NPC in a proliferative state, and in promoting survival of NPC and their neuronal and glial progeny (87, 166). Telomerase is also present in NPC in the adult brain where its expression may be influenced by brain injury and other environmental factors. Recent studies suggest that TERT expression can be induced by BDNF and sAPP (W. Fu and M. P. Mattson, unpublished data). It has been reported that TERT can prevent apoptosis of cultured neurons in experimental models relevant to AD and stroke (87, 387), suggesting that if TERT expression could be induced in neurons or NPC in the adult brain, it may increase the resistance of neurons to age-related neurodegenerative disorders (221). DNA damage may trigger neuronal death in age-related neurodegenerative processes, and telomerase may suppress such DNA damage responses (198); nuclear localization and reverse transcriptase activity appear to be critical for the antiapoptotic function of TERT (P. Zhang and M. P. Mattson, unpublished data).

The possibility that the aging process impairs neurogenesis is suggested by studies in which BrdU-labeled cells were quantified in the brains of middle-aged and old rats (174). This adverse effect of aging on neurogenesis may be counteracted by many of the same environmental conditions that promote successful brain aging. Indeed, dietary restriction can increase neurogenesis (183). Interestingly, neurogenesis can also be increased by environmental enrichment (164, 254) and physical exercise (351). Because no specific molecular markers of NPC have been established, and because NPC cannot be labeled by the usual BrdU method in clinical studies of humans, it is not known whether abnormalities in neurogenesis contribute to the pathogenesis of age-related neurodegenerative disorders. However, recent studies of experimental models of AD have shown that amyloid -peptide can impair neurogenesis (124). Both the proliferation and survival of NPC in the dentate gyrus of the hippocampus are reduced in APP mutant mice. Infusion of amyloid -peptide into the lateral ventricle of adult mice impairs neurogenesis of NPC in the subventricular region. Moreover, exposure of cultured human NPC to amyloid -peptide impairs their proliferation and differentiation and can induce apoptosis (124). These experimental findings suggest that adverse effects of amyloid -peptide on NPC may contribute to depletion of neurons and cognitive impairment in AD. Although it is not known whether a failure of neurogenesis contributes to the pathogenesis of PD, several studies in rodents, nonhuman primates, and humans suggest that functional recovery can occur after transplantation of NPC or mobilization of endogenous NPC (21, 323). The development of methods for identifying NPC and their recent progeny in post mortem brain tissue sections from human patients would greatly facilitate our understanding of the relative contributions of neuronal degeneration and impaired neurogenesis to neurodegenerative disorders.

The implications of neurogenesis for facilitating successful brain aging and treating age-related neurodegenerative disorders are quite profound. It may be possible to mobilize endogenous NPC in the brain or to introduce exogenous NPC to replace dysfunctional or dead neurons and glia. As described above, three different behavioral modifications have been shown to enhance neurogenesis (dietary restriction and increased intellectual and physical activities). In addition, pharmacological approaches for mobilizing NPC are being developed. For example, it has been shown that antidepressant drugs such as serotonin transport inhibitors can increase production of BDNF and stimulate neurogenesis (5, 201). There have already been reports of improved functional outcome following NPC transplantation in models of traumatic central nervous system injury (238), demyelinating disorders (372), and PD (21). Particularly intriguing are reports suggesting that stem cells in other organs of the body, including bone marrow, are capable of forming neurons and glial cells (242). A major clinical hurdle in interindividual transplantations is immune attack on the transplanted cells; this could be avoided by using a person's own stem cells for transplantation. Embyronic stem cells may also prove valuable in treating various neurodegenerative disorders because of their multipotent properties and their reduced reactivity toward immune cells.

	III. GENETIC FACTORS IN BRAIN AGING AND NEURODEGENERATIVE DISORDERS

Top Previous Next References

The probability of living a long life with preservation of a high level of brain function is strongly influenced by the genes one inherits. Accordingly, genetic factors also play important roles in determining one's risk of age-related neurodegenerative disorders. In this section we review the evidence for the involvement of specific genes in the determination of life span and risk of neurological disorders of aging.

There is ample evidence that life span (52, 193, 274), intelligence (57, 240), and risk of various neurological disorders (51, 337, 355) are determined, in part, by heritable factors. However, the specific genes involved and their mechanisms of action are largely unknown. One gene that appears to have an influence on aging in general, and on risk of age-related neurodegenerative disorders, is apolipoprotein E (327). Three alleles of apolipoprotein E encode proteins that differ in two amino acids; E2 contains a cysteine in each position, E3 contains a cysteine in one of the positions, and E4 does not contain a cysteine in either position. Individuals with an E4 allele have a reduced life span (126) and are at increased risk of AD (157). The mechanism whereby E4 may accelerate brain aging has been suggested to involve a decreased antioxidant and neuroprotective properties of this isoform (Fig. 6). The cysteine residues in E2 and E3 may bind to and thereby detoxify 4-hydroxynonenal, a cytotoxic product of lipid peroxidation (266).

View larger version (37K): [in this window] [in a new window]   Fig. 6. Possible mechanisms of action of different apolipoprotein E isoforms in promoting or preventing age-related pathological changes in the brain and periphery. In the brain, apolipoprotein E is produced mainly by astrocytes. Apolipoprotein E can promote neuronal survival and outgrowth and may play important roles in adaptive responses to aging and brain injury. The beneficial effects of apolipoprotein E may involve an antioxidant function. The E2 and E3 isoforms are more effective than the E4 isoform in their antioxidant and biological activities. Mechanisms of apolipoprotein production and metabolism in the periphery are shown in the bottom panel. Individuals with the E4 isoform are prone to atherosclerosis, which may be due to a diminished antioxidant activity of this isoform resulting in enhanced damage to vascular endothelial cells.

Other genes linked to life span that may influence brain aging are those encoding growth hormone (10) and major histocompatability complex (MHC) proteins (36). Growth hormone levels decrease with aging, and this change is ameliorated by caloric restriction. The density of microvessels in the brains of rodents decreases during aging, and this can be reversed by treating the animals with growth hormone, which may increase production of insulin-like growth factor I (IGF-I) in the brain (332). The latter studies further showed that IGF-I can reverse age-related impairment in learning and memory. IGF-I can also protect neurons against injury and death in experimental models of AD and related neurodegenerative disorders (46, 386). MHC genes are expressed in neurons (58). Mice that are genetically deficient for class I MHC proteins exhibit incomplete synaptogenesis in the developing visual system and enhanced long-term potentiation of synaptic transmission in the hippocampus (135), suggesting important roles for MHC proteins in learning and memory.

Inherited variability of mitochondrial genes encoding proteins involved in oxidative phosphorylation and other aspects of mitochondrial function may also contribute to aging (60) and the pathogenesis of neurodegenerative disorders (258). Mitochondrial DNA damage has been shown to increase in brain cells during aging (121), and it has been proposed that accumulation of mitochondrial DNA mutations is a major factor in the aging process itself (260).

Although the vast majority of cases of AD are sporadic with no clear pattern of inheritance and a late age of onset (70s and 80s), some cases are inherited in an autosomal dominant manner with complete penetrance and an early age of onset (40s and 50s). The first gene linked to familial AD is located on chromosome 21 and encodes the APP, the source of the 40- to 42-amino acid amyloid -peptide (A) that forms insoluble amyloid plaques in the brains of all AD patients (122). Several different disease-causing mutations in APP have been reported, all of which are located within or adjacent to the A sequence, and all of which increase production of A-(142). The most intensely studied APP mutations are the "Swedish" mutation (a 2-amino acid substitution adjacent to the NH₂ terminus of A; Ref. 177) and the "London" mutation (a missense mutation adjacent to the COOH terminus of A; Ref. 42). In addition, several AD kindreds have been identified in which mutations within the A sequence are pathogenic (253). Two other genes linked to early-onset familial AD are those encoding PS1 (chromosome 14) and PS2 (chromosome 1); more than 70 different PS1 mutations (all except one are missense mutations), and 2 PS6 mutations have been reported (122). PS1 and PS2 are structurally similar integral membrane proteins with eight transmembrane domains and are localized mainly in the endoplasmic reticulum (ER). The presenilin-1 mutations tend to cluster in the cytoplasmic loop region and in or near transmembrane domain 2. The identification of the mutations in the APP and presenilin genes has led directly to experiments that have revealed, at least in part, how they cause AD (Fig. 7; Ref. 226).

View larger version (56K): [in this window] [in a new window]   Fig. 7. Mechanisms underlying the pathogenic actions of mutations in amyloid precursor protein (APP) and presenilins. Mutations in APP, as well as presenilin mutations, shift the proteolytic processing of APP such that more A is produced and less sAPP is produced. Presenilins play an essential role in -secretase cleavage of APP and may also facilitate Notch cleavage and release of the transcription-regulating Notch COOH-terminal domain (NICD). Presenilin mutations have a major impact on endoplasmic reticulum (ER) calcium signaling, effectively increasing the pools of ryanodine- and inositol trisphosphate-sensitive stores. The normal functions of APP and presenilins, and the consequences of Alzheimer's disease-linked mutations in these proteins, may be particularly important in synapses.

A well-documented abnormality that results from APP mutations, as well as presenilin mutations, is increased production of A [particularly A-(142)] and decreased production of sAPP (6, 122, 216). A can impair synaptic function and can render neurons vulnerable to excitotoxicity and apoptosis by the following mechanism. During the process of self-aggregation, A generates reactive oxygen species (hydrogen peroxide and hydroxyl radical) by a mechanism that may involve metal-catalyzed oxidation of methionine (128, 134, 353). When this process occurs in the immediate vicinity of cell membranes, lipid peroxidation is initiated (204, 205). In neurons, A-induced lipid peroxidation impairs the function of ion-motive ATPases (sodium and calcium pump proteins), glucose transporter proteins (204-206), and GTP-binding proteins (163). A can also induce oxidative stress in astrocytes resulting in impaired glutamate transport (23). By this mechanism, A disrupts neurotransmitter signaling, destabilizes cellular calcium homeostasis, and renders neurons vulnerable to excitotoxicity and apoptosis (204, 227). Oxidative stress induced by A may be particularly detrimental for neuronal function and survival when it occurs in synapses (162).

Exposure of cultured neurons to A can trigger programmed cell death which manifests characteristic mitochondrial membrane alterations and release of cytochrome c and caspase activation. A stimulates the production of apoptotic proteins including Par-4, Bax, and the tumor suppressor protein p53 (55, 80, 109, 111, 262). Analyses of post mortem brain tissue from AD patients reveals evidence for neuronal apoptosis including upregulation of Par-4 (109) and caspase activation (38). Agents that stabilize mitochondrial function and caspase inhibitors protect neurons against A-induced death (109, 111, 234).

In addition to increasing production of A, APP and presenilin mutations decrease the production of sAPP (6). A decrease in sAPP levels may contribute to the pathogenesis of AD, because sAPP normally functions in modulating synaptic plasticity (learning and memory) and in promoting survival of neurons (90, 142, 230). The mechanism whereby sAPP promotes neuronal survival and synaptic plasticity involves activation of potassium channels and of the transcription factor NF-B; these actions of sAPP stabilize cellular Ca²⁺ homeostasis and suppress oxyradical production (225).

AD patients typically exhibit emotional disturbances and abnormal stress responses involving increased glucocorticoid production that likely result from pathological changes in brain regions that control such behaviors and stress responses including limbic structures such as the amygdala and hippocampus and the frontal cortex (300). Studies of APP mutant transgenic mice suggest that such abnormal stress responses are the result of increased levels of A in these brain regions. APP mutant mice exhibit an age-dependent increase in sensitivity to physiological stressors, which is associated with abnormalities in hypothalamic-pituitary-adrenal function and dysregulation of blood glucose levels (267). Two related neuropeptides that may play a role in the alterations in affective behaviors in AD patients and APP mutant mice are corticotropin-releasing hormone (CRH) and urocortin. CRH is expressed in brain regions prone to degeneration in AD, and CRH can protect cultured hippocampal and cortical neurons against cell death caused by A (271). Urocortin and urocortin II are CRH-related neuropeptides that act on CRH receptors expressed by neurons in the hippocampus and functionally related brain regions. Urocortin can protect hippocampal neurons against excitotoxic and oxidative injury by activating the type I CRH receptor and a signaling pathway involving cAMP-dependent protein kinase, protein kinase C, and MAP kinase (272). These findings suggest roles for CRH and urocortin in antagonizing the neurodegenerative process in AD. It remains to be determined whether pharmacological manipulations of CRH/urocortin signaling will benefit AD patients.

Two major consequences of presenilin mutations are perturbed cellular calcium homeostasis (226) and altered APP processing (122) (Fig. 6). At this point in time, it remains unclear which defect is a primary consequence of the mutations and which is secondary. PS1 and PS2 mutations increase the vulnerability of cultured cells to apoptosis and excitotoxicity (108, 110, 112, 367). Hippocampal neurons in PS1 mutant knockin mice are more vulnerable to excitotoxicity and apoptosis (108). Perturbed calcium regulation in the ER is central to the cell death-promoting effects of PS1 mutations (Fig. 7). The abnormality involves an increased pool of ER calcium resulting in increased calcium responses when cells are challenged with glutamate or agonists that stimulate calcium release from the ER (39). Abnormal ER calcium signaling caused by presenilin mutations may promote altered capacitative calcium influx through voltage-dependent channels in the plasma membrane (186, 375). Presenilin mutations result in altered APP processing, and there is evidence that presenilins are critical for -secretase activity (369) and Notch cleavage (63). Altered APP processing may not account for the spectrum of effects of PS mutations. Instead, altered APP and Notch processing caused by PS mutations may result from altered calcium-mediated regulation of the enzyme activities that mediate cleavage of the two proteins (191, 276, 284, 316, 361). Indeed, one or more secretase activities are sensitive to calcium (44). Notch signaling may promote neuronal survival by enhancing cellular calcium homeostasis, an action antagonized by a protein called Numb; alterations in Notch and Numb functions may play roles in the pathogenesis of AD (40).

Mutations in genes encoding the proteins -synuclein and Parkin can cause early-onset familial PD; -synuclein mutations are inherited in an autosomal dominant manner, while mutations in parkin are inherited in an autosomal recessive manner (280). The -synuclein gene is located on chromosome 4, and the Parkin gene is located on chromosome 6. -Synuclein is a vesicle-associated protein, and Parkin is a cytoplasmic protein. -Synuclein is axonally transported and associates with vesicles in presynaptic terminals, suggesting a role in regulation of vesicle trafficking (148). Parkin is expressed primarily in neurons where it is localized at particularly high levels in neurites (137). Studies of the pathogenic actions of three missense mutations in -synuclein (A53T, A30P, and G209A) have provided new insight into the events that lead to the dysfunction and degeneration of dopaminergic neurons in PD (Fig. 8). Expression of -synuclein mutations in cultured cells increases their vulnerability to oxidative stress and apoptosis (342). Overexpression of wild-type or mutant -synuclein induces apoptosis in cultured neurons (305); PC12 cells overexpressing mutant -synuclein exhibit decreased proteasome activity and increased vulnerability to mitochondrial dysfunction and apoptosis (343). Masliah et al. (210) reported evidence for loss of dopaminergic neurons and Lewy body-like cytoplasmic inclusions in -synuclein mutant mice. However, another line of mice expressing mutant -synuclein driven by a tyrosine hydroxylase promoter did not exhibit pathology in the substantia nigra (211). -Synuclein forms aggregates that may exhibit toxic properties similar to those of A including production of reactive oxygen species (347) and increased membrane ion permeability (357). Collectively, these findings suggest that -synuclein mutations may promote neuronal degeneration by causing abnormalities in protein degradation and oxidative stress. Some data point to a loss of function, as opposed to a gain of function in the pathogenic action of -synclein mutations. Thus -synuclein knockout mice exhibit a defect in dopamine release (1), and overexpression of wild-type (but not mutant) -synuclein protects cultured neural cells against apoptosis (56).

View larger version (65K): [in this window] [in a new window]   Fig. 8. Mechanisms of neuronal degeneration in familial Parkinson's disease (PD). In sporadic PD, alterations in dopamine metabolism and/or exposure to environmental toxins such as MPTP and rotenone can induce oxidative stress in dopaminergic neurons resulting in their dysfunction and death. Familial PD can be caused by mutations in -synuclein or Parkin. Mutations in -synuclein may cause excessive protein aggregation that triggers apoptosis and/or impair presynaptic function. Parkin mutations may interfere with protein degradation by the proteasome. In both sporadic and familial PD, dopaminergic neurons are subjected to increased oxidative stress that may trigger apoptosis.

Parkin is a ubiquitin-protein ligase that presumably functions in protein degradation; Parkin mutations result in loss of the ubiquitin-protein ligase activity (320). It was recently reported that parkin can ubiquitinate -synuclein (321), suggesting a link between impaired proteasomal degradation of -synuclein and the neurodegenerative process. These observations strongly suggest that a defect in protein degradation is central to the pathogenesis of PD and further suggest strong mechanistic interactions between oxidative stress and protein degradation in neurodegenerative disorders in general, since oxidative damage to proteins often makes them targets for ubiquitination and proteasomal degradation. In addition to disease-causing mutations, risk of PD may be influenced by genetic polymorphisms, a possibility that is currently being investigated by several laboratories (43, 356).

The genetics of HD are seemingly straightforward; all cases of this disease are believed to be caused by the expansion of trinucleotide (CAG) repeats in the huntingtin gene (located on chromosome 4) resulting in long stretches of polyglutamine repeats in the huntingtin protein (76, 199). However, the consquences of the huntingtin mutation may be subject to modification by aging, as suggested by evidence for variability in age of disease onset and progression of the clinical phenotype. HD patients manifest progressive motor dysfunction characterized by involuntary body movements due to degeneration of neurons in the basal ganglia, principally the caudate and putamen (3). As the disease progresses, the neurodegenerative process may spread to regions of the cerebral cortex, thalamus, and cerebellum resulting in cognitive dysfunction and emotional disturbances. Recently, a closely related familial HD-like disorder was described that appears to result from a trinucleotide expansion in a yet-to-be identified gene (202).

The alterations caused by polyglutamine expansions in huntingtin that result in neuronal death are beginning to be revealed (Fig. 9). Overexpression of mutant human huntingtin in cultured cells and transgenic mice can induce spontaneous cell death (apoptosis) and can increase the vulnerability of neurons to excitotoxicity (130, 192, 294, 346). Several different behavioral abnormalities have been described in mice expressing mutant huntingtin including motor deficits and cognitive dysfunction (250, 311). The reason that trinucleotide expansions in huntingtin promote degeneration of striatal neurons in HD is unclear. Mutant huntingtin self-aggregates resulting in the formation of inclusions in the nucleus and cytoplasm (120, 311). Neurons in mice expressing mutant huntingtin exhibit increased caspase activation, and administration of caspase inhibitors to the mice can suppress the neurodegenerative process (257), suggesting that mutant huntingtin triggers programmed cell death. Cells expressing mutant huntingtin exhibit altered proteasomal function that may trigger apoptosis (145). Mutant huntingtin may trigger apoptosis by weakening the interaction of huntingtin with huntingtin interacting protein-1 (Hip-1), thereby allowing Hip-1 to interact with a protein called Hippi that then recruits caspase-8 and thereby initiates the cell death cascade (223).

View larger version (58K): [in this window] [in a new window]   Fig. 9. Pathogenic mechanisms of mutant huntingtin. Huntington's disease is caused by polyglutamine expansions in the huntingtin protein. Mutant huntingtin may self-aggregate and trigger activation of caspases. Data also suggest that mutant huntingtin can cause a depletion of BDNF production by suppressing transcriptional activity.

Interestingly, adverse effects of mutant huntingtin on cell function are not limited to the nervous system because abnormalities in adipocytes have been described that may be related to the well-known alterations in energy metabolism in HD patients (82). The ability of dietary supplementation with creatine to delay motor symptoms and increase survival in huntingtin mutant mice (7) is consistent with impaired cellular energy metabolism in the neurodegenerative process. Finally, it was recently reported that levels of BDNF are decreased in HD (390), suggesting that decreased neurotrophic support may be a consequence of huntingtin mutations that promotes the degeneration of striatal neurons.

In each of the three neurodegenerative disorders just described (AD, PD, and HD), there is considerable evidence suggesting that abnormalities in mitochondrial function contribute to the disease process. In each disorder, alterations in activities of enzymes involved in oxidative phosphorylation have been demonstrated including a decrease in activity of the -ketoglutarate dehydrogenase complex in AD (97a) and a defect in complex I in PD (306b). Such abnormalities in mitochondrial energy metabolism may precede and contribute to the increased oxidative stress and perturbations in neuronal calcium homeostasis that occurs in each disorder. Interestingly, the metabolic deficits, and oxidative stress and calcium dysregulation, may not be limited to the brain cells affected in the disorders, as they have been documented in peripheral cells including fibroblasts and lymphocytes (95a, 306b). Based on these and additional data, Blass (23a) has introduced the concept of a "mitochondrial spiral," in which metabolic deficits result in oxyradical production and calcium dysregulation, to explain the pivotal role of mitochondrial alterations in neurodegenerative disorders. Both genetic and environmental factors may promote such neurodegenerative mitochondrial spirals (97a).

ALS involves degeneration of spinal cord motor neurons resulting in progressive paralysis and death (196, 301). Most cases of ALS are sporadic, but some result from mutations in the gene encoding the antioxidant enzyme Cu/Zn-SOD, which is located on chromosome 21 (62, 302). Transgenic mice expressing mutant Cu/Zn-SOD exhibit progressive motor neuron degeneration and a clinical phenotype remarkably similar to ALS patients (118, 368). Lipid peroxidation is increased in spinal cord motor neurons of ALS patients and transgenic mice (268, 269), and administration of vitamin E to Cu/Zn-SOD mutant mice delays disease onset (117), suggesting an important role for lipid peroxidation in the neurodegenerative process. In addition, creatine delayed the neurodegenerative process in a mouse model of ALS (168). ALS Cu/Zn-SOD mutations cause impairments in synaptic glucose and glutamate transport (114) and increase the vulnerability of motor neurons to excitotoxic injury by increasing oxidative stress and perturbing cellular calcium homeostasis (172, 185).

The remarkably large size of axons of motorneurons and their correspondingly high density of neurofilaments has led to the suggestion that impaired axonal transport plays a role in the pathogenesis of ALS. Studies of axonal transport in Cu/Zn-SOD mutant mice and of mice lacking or overexpressing neurofilament proteins support a role for impaired axonal transport in ALS (365, 383). Apoptosis of motor neurons in ALS is suggested by studies showing that levels of the proapoptotic protein Par-4 are increased in spinal cord motor neurons of ALS patients and Cu/Zn-SOD mutant mice (269), and levels of caspase activation are also increased in spinal cord tissue from Cu/Zn-SOD mutant mice (264). In addition, caspase inhibitors (190) and the antiapoptotic protein Bcl-2 (358) can protect motor neurons in Cu/Zn-SOD mutant mice. Moreover, neurotrophic factors that can prevent apoptosis of motor neurons in culture can also prevent motor neuron loss and disease progression in Cu/Zn-SOD mutant mice (243).

Genetic risk factors for stroke, an age-related neurological disorder, overlap with risk factors for coronary artery disease due to shared mechanisms of atherosclerosis and blood clot formation in each disorder. Familial hypercholesterolemia caused by mutations in the low-density lipoprotein receptor results in early-onset atherosclerotic vascular disease (334). Apolipoprotein E genotype affects risk of atherosclerosis (286). Polymorphisms in the apolipoprotein(s) gene and the gene for methylene tetrahydrofolate reductase have been shown to be prominent risk factors for atherosclerotic vascular disease (252). Polymorphisms in the fibrinogen and factor XIII genes have been linked to stroke (37), suggesting a contribution of genetic differences in regulation of clot formation to disease risk. An inherited stroke syndrome called CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) was recently linked to mutations in the Notch-3 gene on chromosome 19 (59), and the cellular and molecular alterations caused by these mutations are currently being investigated.

Collectively, the data that have accumulated in studies of the age-related neurodegenerative disorders described above suggest that each disorder shares abnormalities that contribute to neuronal dysfunction and death. The alterations include increased oxidative stress, dysregulation of protein trafficking and processing, metabolic impairment, and disruption of cellular calcium homeostasis. Genetic factors that cause, or increase risk of, a disorder do so by impacting directly or indirectly one or more of the cellular systems involved in oxyradical metabolism, protein processing, energy metabolism, and calcium homeostasis. Section IV describes a rapidly growing body of evidence demonstrating that age-related neurodegenerative cascades can be influenced by environmental factors.

	IV. DIETARY FACTORS IN BRAIN AGING AND NEURODEGENERATIVE DISORDERS

Top Previous Next References

There are numerous dietary factors that have been reported to affect brain physiology in ways that could, in theory, modify brain aging and the pathogenesis of neurodegenerative disorders (221a). These range from amino acids such as tryptophan (16) to caffeine and related stimulants (85) to omega-3 fatty acids (348). Many such dietary factors have been shown to affect mood or cognition. However, this review focuses on a more limited number of dietary factors for which considerable supportive experimental, clinical, and epidemiological data have accrued to justify the development of recommendations to the general public vis-à-vis risk reduction for neurodegenerative disorders. These factors include caloric intake, folic acid, and antioxidants.

The mean and maximum life spans of many different organisms including yeast, roundworms, rodents, and monkeys can be increased by up to 50% simply by reducing their food intake (9, 176, 362). The incidence of age-related cancers, cardiovascular disease, and deficits in immune function is decreased in rodents maintained on such dietary restriction (DR) feeding regimens (362). Data from clinical and epidemiological studies in humans support the antiaging and disease prevention effects of DR. Thus a low-calorie diet decreases the risk of the most prominent age-related diseases in humans including cardiovascular disease, diabetes, and cancers (29, 179, 187). Recent findings reviewed in this section strongly suggest that DR can delay age-related functional deficits in the brain and may reduce the risk of major neurodegenerative disorders including AD, PD, and HD. DR may also increase resistance of neurons to acute insults such as stroke and severe epileptic seizures.

Biochemical and molecular analyses of the brains of old rats and mice that had been maintained on calorie-restricted diets reveal a retardation of changes that occur during aging of animals fed ad libitum including increases in levels of GFAP and oxidative damage to proteins and DNA (74, 245). Gene array analysis of the expression levels of thousands of genes in the brains of young rats and old rats that had been maintained on control or restricted diets revealed changes in gene expression in brain cells during aging and showed that DR can suppress many of those changes (180). Age-related changes in the expression of genes that encode proteins involved in innate immune responses, oxidative stress, and energy metabolism are counteracted by DR. This retardation of brain aging at the molecular level may underlie the preservation of brain function during aging in animals maintained on DR. For example, DR attenuates age-related deficits in learning and memory ability and motor function in rodents (140, 336). Studies of human populations suggest that DR can promote successful brain aging in humans. For example, epidemiological data suggest that the risk of developing AD, PD, and stroke is lower in individuals with a low calorie intake (30, 195, 235).

Beneficial effects of DR have been demonstrated in several of the animal models just described. The two DR protocols most commonly used in such studies are every-other-day feeding (the animals must go a whole day without food and then eat ad libitum on the next day) and paired feeding (the restricted animals are given food pellets that contain 30-40% fewer calories than the pellets given to the control animals). Both of these feeding regimens increase the life spans of rats and mice by 30-40% (103, 362). Rats maintained on DR for 2-4 mo exhibit increased resistance of hippocampal neurons to kainate-induced degeneration in a model relevant to the pathogenesis of AD and epilepsy; kainate selectively damages hippocampal pyramidal neurons and there is a profound deficit in learning and memory (31). When rats were maintained for several months on DR, damage to hippocampal neurons was decreased, and learning and memory were preserved compared with rats fed ad libitum (31). Studies of PS1 mutant knockin mice showed that the PS1 mutations increase the vulnerability of hippocampal and cortical neurons to excitotoxicity and apoptosis by a mechanism involving enhanced calcium release from the ER (108, 111). When PS1 mutant knockin mice were maintained on DR, they exhibited increased resistance of hippocampal CA1 and CA3 neurons to excitotoxic injury compared with mice fed ad libitum (388). Lipid peroxidation in the hippocampus after kainate administration was decreased in DR PS1 mutant mice, suggesting that suppression of oxidative stress is one mechanism whereby DR protects neurons (388).

DR has beneficial effects in animal models of PD and HD. The vulnerability of midbrain dopaminergic neurons to MPTP toxicity was decreased in mice maintained on dietary restriction with more dopaminergic neurons surviving exposure to MPTP; the motor function of the mice was also improved in the restricted mice (69). Administration of the succinate dehydrogenase inhibitor (mitochondrial toxin) 3-nitropropionic acid (3-NP) to rats and mice results in selective degeneration of striatal neurons and motor impairment, a model of HD. When rats were maintained on DR for several months before administration of 3-NP, more striatal neurons survived exposure to 3-NP, and their motor function was improved (31). The ability of DR to improve outcome after a stroke was demonstrated in a rat model in which the middle cerebral artery was transiently occluded resulting in damage to the cerebral cortex and striatum supplied by that artery and unilateral motor dysfunction. When rats were maintained on DR for several months and then subjected to a stroke, they exhibited reduced brain damage and improved behavioral outcome (380). The neuroprotective effects of DR in animal models of several different neurodegenerative disorders suggest that low-calorie diets may prove beneficial in reducing the incidence and/or severity of the corresponding human neurodegenerative disorders.

Although the findings described in the preceding paragraphs document quite striking neuroprotective effects of DR, this dietary manipulation has not proven beneficial in all animal models of neurodegenerative disorders. For example, when one line of APP mutant transgenic mice was maintained on an every-other-day feeding regimen or was fed ad libitum, they died within 2-3 wk (267). Additional analyses in the latter study showed that the APP mutant mice were hypersensitive to the stress associated with fasting for an entire day and became severely hypoglycemic during the days they were without food. The APP mutant mice exhibited abnormalities in the regulation of the stress-responsive hypothalamic-pituitary-adrenal axis including altered glucocorticoid and blood glucose regulation in response to restraint stress. However, when every-other-day feeding was begun in APP mutant mice that were less than 3 mo of age, they survived, suggesting a role for age-dependent amyloid deposition in the aberrant stress response. Transgenic ALS mice expressing the G93A Cu/Zn-SOD mutation did not benefit from DR. When they were maintained on an every-other-day feeding regimen, the age of disease onset was unchanged (270). Moreover, once the ALS mice on DR became symptomatic, the disease progressed more rapidly and they died sooner than did ALS mice that were fed ad libitum. These findings are interesting in that they suggest that the pathogenic mechanism of action of the Cu/Zn-SOD mutation is not subject to modification by DR and/or that DR does not exert the same kind of neuroprotective action on spinal cord motor neurons that it exerts on neurons in the brain.

There is emerging evidence from studies of human populations that is consistent with the possibility that DR can reduce the risk of human neurodegenerative disorders. The following epidemiological data suggest that individuals with a low calorie intake may have reduced risk for AD and PD. There is a strong correlation between per capita food consumption and risk for AD (105). For example, the reported incidence of AD in China and Japan is approximately one-half that in the United States and Western Europe, and this is correlated with a lower calorie intake in China and Japan (1,600-2,000 calories/day) compared with the United States and Western Europe (2,500-3,000 calories/day). Overeating is also a major risk factor for stroke (30). Although there are caveats with the latter observations (for example, per capita food consumption is a very poor measure of energy intake, and disease diagnosis may differ among the countries), they are consistent with a protective effect of low-calorie diets against age-related neurodegenerative disorders. More convincing evidence that DR can protect against neurodegenerative disorders comes from population-based case-control studies by Mayeux and colleagues who found that individuals with the lowest daily calorie intakes had the lowest risk of AD (235) and PD (195). Interestingly, the risk of PD and AD was more strongly correlated with calorie intake than with weight or body mass index. More recently, Hendrie et al. (127) reported findings from a population-based longitudinal prospective study which indicate that the incidence of AD increases among individuals living in industrialized countries compared with genetically similar individuals that live in nonindustrialized countries. Although the environmental factors that increase risk of AD in industrialized countries are not known, one clear difference between the two environments is calorie intake, which is much higher in industrialized countries. Together, the epidemiological and experimental data provide strong evidence that DR can reduce risk of AD, PD, and stroke, three of the most devastating neurodegenerative conditions in the elderly.

Because dietary restriction increases life span and reduces risk of many different age-related diseases including cardiovascular disease, diabetes, and cancers, it might be expected that it modifies shared biochemical cascades that lead to cell dysfunction and disease. In the case of neurodegenerative disorders, it is clear that while different genetic and environmental factors may initiate the neurodegenerative process in different disorders, a shared biochemical cascade ensues. Increased oxidative stress, perturbed cellular calcium homeostasis, and impaired energy metabolism occur in every neurodegenerative disorder studied to date (216, 219). These alterations render neurons vulnerable to apoptosis, a biochemical cascade of molecular interactions involving proteins such as Par-4, Bcl-2 family members, and caspases (219).

It has been shown that DR can stabilize mitochondrial function and reduce oxidative stress in brain cells of rodents (113), and this may increase the resistance of neurons to many different types of genetic and environmental factors. DR can induce the expression of genes that encode proteins that promote neuronal survival and plasticity (Fig. 10). For example, levels of heat shock protein-70 (HSP-70) and glucose-regulated protein-78 (GRP-78) are increased in cortical, striatal, and hippocampal neurons of rats and mice maintained for several months on a dietary restriction feeding regimen (69, 380). HSP-70 and GRP-78 can protect neurons against excitotoxic and oxidative injury (197, 379), suggesting that their increased levels contribute to the neuroprotective effect of dietary restriction.

View larger version (21K): [in this window] [in a new window]   Fig. 10. Model of the mechanisms whereby dietary restriction, intellectual activity, and exercise promote neuronal survival and plasticity. Dietary restriction, activity in neuronal circuits, and physical exercise each induces a mild cellular stress response, as a result of energetic factors (reduced glucose availability in dietary restriction and increased energy demand during intellectual and physical activity, for example). Neurons respond to these stresses by activating signaling pathways that induce the expression of genes encoding proteins, such as growth factors and protein chaperones, that promote neuronal survival and plasticity (neurogenesis, neurite outgrowth, and synaptic plasticity).

DR can induce the expression of several different neurotrophic factors in brain cells. Levels of BDNF are increased in neurons in the cerebral cortex, hippocampus, and striatum of rats and mice maintained on dietary restriction (6