Address correspondence and reprint requests to P. Hemachandra Reddy, PhD, Neurogenetics Laboratory, Neurological Sciences Institute, Oregon Health and Science University, 505 NW 185th Avenue, Beaverton, OR 97006, USA. E-mail: email@example.com
Silent information regulator 2, a member of NAD+-dependent histone deacetylase in yeast, and its homologs in mice and humans, participate in numerous important cell functions, including cell protection and cell cycle regulation. The sirtuin family members are highly conserved evolutionarily, and are predicted to have a role in cell survival. The science of sirtuins is an emerging field and is expected to contribute significantly to the role of sirtuins in healthy aging in humans. The role of sirtuins in neuronal protection has been studied in lower organisms, such as yeast, worms, flies and rodents. Both yeast Sir2 and mammalian sirtuin proteins are up-regulated under calorie-restricted and resveratrol treatments. Increased sirtuin expression protects cells from various insults. Caloric restriction and antioxidant treatments have shown useful effects in mouse models of aging and Alzheimer's disease (AD) and in limited human AD clinical trials. The role sirtuins may play in modifying and protecting neurons in patients with neurodegenerative diseases is still unknown. However, a recent report of Huntington's disease revealed that Sirtuin protects neurons in a Huntington's disease mouse model, suggesting that sirtuins may protect neurons in patients with neurodegenerative diseases, such as AD. In this review, we discuss the possible mechanisms of sirtuins involved in neuronal protection and the potential therapeutic value of sirtuins in healthy aging and AD.
Research on aging began several decades ago, as did research on Alzheimer's disease (AD), yet scientists are struggling to unravel a clear connection between aging and AD. Eat less, age well, and remember well appears to be the new mantra in cutting-edge research on aging and human health. Emerging evidence on the consequences of caloric restriction (CR) highlights the therapeutic value of CR in extending lifespan and in delaying or avoiding numerous age-related disorders, such as cardiovascular and neurological disorders (Anson et al. 2003; Bordone and Guarente 2005; Guarente and Picard 2005). CR triggers the over-expression of Sir2, a NAD-dependent class III histone deacetylase, which participates in transcriptional silencing of telomeres and cell-mating type loci, recombination at DNA, cell cycle regulation, and lifespan extension in yeasts (Guarente 2000; Blander and Guarente 2004; Lamming et al. 2004). Sir2 is also known to mediate the nutrient-sensing pathway of aging in nematodes and fruit flies (Parker et al. 2005; Tissenbaum and Guarente 2001; Wood et al. 2004). Seven Sir2 homologs have been identified in mice and humans [sirtuin (SIRT)1–7], and they offer promise for understanding the mechanisms of neuronal protection and aging (Frye 1999, 2000; Denu 2003; North and Verdin 2005; Porcu and Chiarugi 2005).
Even if CR proves useful, the public might not accept this type of therapy, feeling reluctant to give up a few favorite tastes. Many plant flavonoids – a class of naturally occurring plant compounds that function as antioxidants and that are found in, for example, purple grape juice and red wine – mimic CR-type effects, thus earning their name CR mimetics (Howitz et al. 2003; Lamming et al. 2004; Wood et al. 2004). One such proven CR mimetic is resveratrol, a flavonoid derived from red wine, which is implicated in the ‘French paradox’ that links wine drinking to longer lifespan and healthy aging. It is a matter of time before many more CR mimetics are found, as nearly 35 000 plant species offer more than 4000 flavonoids in different fruits and vegetables, not to mention other types of phytochemicals (Nijveldt et al. 2001; Duncan et al. 2003; Williams et al. 2004), and studies are under way to identify many more CR mimetics. If flavonoids are proven useful in extending healthy lifespans, it is not inconceivable that just the right dose of fruits, vegetables, or their extracts will be added to daily meals to promote healthy aging. In this review, we give an overview of the mechanisms of aging and AD, and discuss how mammalian sirtuins might trigger neuroprotective mechanisms in healthy aging and AD.
Aging, CR and sirtuins
Healthy aging is a universal and natural phenomenon that is not affected by age-related diseases. Several hypotheses have been proposed to explain the biological mechanisms of healthy aging and extended lifespan in numerous organisms, from bacteria to humans. The best known of these hypotheses involve the reduced insulin/insulin growth factor type 1 (IGF-I), CR, longer telomeres, and decreased production of reactive oxygen species (ROS) in mitochondria (Dillin et al. 2002; Hekimi and Guarente 2003; Tatar et al. 2003; Balaban et al. 2005; Kenyon 2005). Despite what we know about healthy aging and ROS in particular, anti-aging drugs are still not available. CR without malnutrition and a diet that includes antioxidants, fruits and vegetables seem to offer hope for delaying unhealthy disorders that accompany aging but, as yet, no such clinical treatment has been developed (Hadley et al. 2005). Medicines that treat osteoporosis and cardiovascular, neurological and other age-related disorders extend healthy aging indirectly, but it is unclear if they also affect aging mechanisms.
The molecular effects of reduced insulin/IGF-I (Holzenberger et al. 2003) as well as CR (Rodgers et al. 2005) include the reduction of mitochondrial ROS, a slight increase in respiration rates, and an increase in the uncoupling of the electron transport chain (ETC), all of which have been found to lead to cell protection and extended lifespan (Bordone and Guarente 2005). In short, considerable cross-talk exists among CR, ROS and the ETC of mitochondria that control cell death and the aging process. The precise mechanism(s) involved in each age-related disorder are still not clearly understood. However, it has been proposed that when a cell loses its potential to protect itself from various age-related problems, disease-specific insults increase in the cell and ultimately lead to histopathological changes and the development of disease symptoms. The best example in this scenario is late-onset AD.
Table 1 summarizes the biological functions of Sir2 homologs in mouse and humans that are currently under intense investigation. Seven mammalian sirtuins have been found to be expressed ubiquitously across different types of tissues and have been categorized under four classes of proteins: SIRT1–3 (class I), SIRT4 (class II), SIRT5 (class III), and SIRT6–7 (class IV) (Frye 1999, 2000; North et al. 2003; Shi et al. 2005). Molecular phylogenetic analyses of 60 sirtuins from prokaryotes and eukaryotes suggest that the biological structure of sirtuins has been conserved evolutionarily and that sirtuins participate in important cellular functions (Frye 1999, 2000).
Table 1. Summary of mammalian sirtuins, their localization, interactions and cellular functions
Repression; reduced DNA damage; increased cell survival
SIRT1 deacetylates histone and non-histone proteins
Of the seven sirtuins discovered so far (see Table 1), SIRT1 is closely related to Sir2 and is implicated in the lifespan extension of lower organisms, such as yeast. Functions of SIRT1 are rapidly being identified in the current literature. SIRT1 interacts and deacetylates histones H1-Lys26, H3-Lys9, Lys14 and H4-Lys16, and reduces methylation of histone H3-Lys79 (Imai et al. 2000; Vaquero et al. 2004). SIRT1 also deacetylates and represses the activities of p53 (Vaziri et al. 2001), forkhead transcription factor (FOXO)3a, FOXO1 and FOXO4 (Motta et al. 2004; Yang et al. 2005) (Table 1). Under oxidative stress, SIRT1 increases the ability of FOXO3 to induce the cell cycle or to withstand oxidative stress, and SIRT1 reduces the apoptotic ability of FOXO3 (Brunet et al. 2004). The over-expression of SIRT1 in clonal PC12 cells has been found to reduce the cellular consumption of O2 by about 25%, and to deacetylate and interact with peroxisome proliferator-activated receptor-gamma co-activator-1α (PGC-1α), a master regulator of the cellular gluconeogenic pathway (Nemoto et al. 2005). SIRT1 physically associates with the human Hairy-related basic helix-loop-helix (bHLH) repressor proteins, Hairy/enhancer-of-split related with C-terminal WRPW motif (HES1) and Hairy/enhancer-of-split related with C-terminal YXXW motif (HEY2) both in vitro and in vivo (Takata and Ishikawa 2003). In studies of SIRT1-deficient mice, the mice were small and showed developmental defects in the retina and heart, and the SIRT1-deficient cells of these mice showed p53 hyperacetylation following DNA damage (Cheng et al. 2003). Mice with Sir2a or SIRT1 null alleles were also smaller and died early during the postnatal period, suggesting that the SIRT1 protein may be essential for embryogenesis and reproduction (McBurney et al. 2003).
SIRT1 in disease pathologies
Recent molecular studies suggest that SIRT1 expression is implicated in disease pathomechanisms (Langley et al. 2002; Yeung et al. 2004; Pagans et al. 2005). SIRT1 in promyelocytic leukemia protein (PML) nuclear bodies deacetylates and negatively regulates PML-induced p53 transactivation and prevents premature cellular senescence (Langley et al. 2002). In human non-small-cell lung cancer cells, SIRT1 physically interacts with the Rel/p65 subunit of nuclear factor (NF)-kappa B, deacetylates it at Lys310, and inhibits NF-kappa B transcription (Yeung et al. 2004) (Table 1). NF-kappa B up-regulates genes essential for cell survival. Treatment with resveratrol was found to potentiate SIRT1 proteins in the promoter region of inhibitor of apoptosis 2 (cIAP-2), which correlated with a loss of NF-kappa B expression and tumour necrosis factor-α-induced apoptosis (Yeung et al. 2004). SIRT1 was also found to interact with and deacetylate the human immunodeficiency virus (HIV) Tat protein and to act as a transcriptional co-activator during Tat transactivation (Pagans et al. 2005). In addition, SIRT1 was found to interact directly with and to repress B cell leukemia 11A (BCL11A) protein (Senawong et al. 2005). The over-expression of Ezh2, a histone-lysine methyltransferase, in a cell culture model for prostate cancer promotes the formation of polycomb repressive complex (PRC4) that contains SIRT1 and the PRC2 isoform, Eed2 (Kuzmichev et al. 2005). The four and a half lens fiber-cell intrinsic membrane (LIM2) protein in prostate cancer cells enhances the interaction of FOXO1 and SIRT1 (Yang et al. 2005). These studies strongly implicate SIRT1 functions in important pathological pathways that are, as yet, to be understood.
CR and plant flavonoids trigger SIRT1 expression
Nutritional stress in PC12 cells and in mice induced a FOXO3a-dependent increase in SIRT1 expression through the interaction of SIRT1 with p53 (Nemoto et al. 2004). CR in Friends virus B (FVB) mice activated SIRT1 in white adipose tissue, and SIRT1 in turn triggered fat mobilization, enhanced lipolysis of triglycerides, released free fatty acid, and inhibited the peroxisome proliferator-activated receptor-gamma (Picard et al. 2004). In the liver of fasted C57BL/6 mice, the levels of SIRT1, PGC-1α, phosphoenol-pyruvate kinase, pyruvate and NAD+ increased and the levels of lactate decreased, suggesting that SIRT1 controls the regulation of gluconeogenic and glycolytic genes (Rodgers et al. 2005). CR and IGF-I/insulin mechanisms in the liver tissues of growth hormone receptor knockout (GHRKO) mice and bovine growth hormone transgenic (bGH Tg) mice were distinct but overlapped, suggesting a major role for the Akt/FOXO1 pathway in the regulation of aging (Al-Regaiey et al. 2005). Resveratrol treatment also increased cell and animal survival by stimulating SIRT1-dependent deacetylation of p53 (Howitz et al. 2003). CR and resveratrol in Fisher 344 rats increased SIRT1 expression, deacetylated Ku70, and suppressed Ku70-Bax-mediated apoptosis (Cohen et al. 2004). Thus, CR and resveratrol treatments increased SIRT1 expression, deacetylase and several transcription proteins, and protected cells from apoptotic death.
Other mammalian sirtuins
Relatively very little is known about the remaining six mammalian sirtuins (SIRT2–7). SIRT2 was found to be an NAD+-dependent α-tubulin deacetylase that co-localizes with the cytoplasmic tubulin network (North et al. 2003), increases expression during the M phase of mitotic cell cycle, controls mitotic exit (Dryden et al. 2003), and also interacts with the homeobox transcription factor (HOXA10) (Bae et al. 2004) (Table 1). SIRT3 was found to localize to the inner membrane and matrix of mitochondria and showed NAD+-dependent protein deacetylation in mitochondria (Onyango et al. 2002; Schwer et al. 2002). In the brown adipose tissue of C57BL/6 mice, CR enhanced the expression of SIRT3, PGC-1α, uncoupling protein (UCP)1, Cytochrome Oxidase II (COX II), COX IV, and ATP synthase. Furthermore, SIRT3 stimulated cyclic adenosine monophosphate (cAMP) response element binding protein (CREB) phosphorylation, reduced ROS production, and increased the respiration rates of the mice (Shi et al. 2005), strongly indicating a role of SIRT3 in mitochondrial function. Although expressed ubiquitously, SIRT4, SIRT5, SIRT6 and SIRT7 failed to deacetylate 3H-labeled acetylated histone H4 (North et al. 2003), suggesting that these proteins may have unknown cellular functions. In a recent study, SIRT6 localized to the nucleus and failed to show deacetylation activity, but did show mono-ADP-ribosyltransferase activity in vitro (Liszt et al. 2005).
It has been proposed that in late-onset AD, ROS may activate β-secretase of the APP molecule and generate Aβ peptides (Tamagno et al. 2002, 2003, 2005). In studies of aging and APP transgenic mice, chronic ROS exposure was found to result in oxidative damage to mitochondrial and cellular proteins, lipids and nucleic acids, resulting in a shut-down of mitochondrial energy production (Reddy and Beal 2005). Recent molecular, cellular and animal model studies of familial AD have revealed that Aβ enters mitochondria and interacts with an Aβ-induced alcohol dehydrogenase protein, disrupts the ETC, generates ROS and inhibits cellular ATP (Lustbader et al. 2004; Reddy and Beal 2005). These results suggest that age-dependent interactions of Aβ with mitochondrial proteins cause mitochondrial dysfunction in AD (Anandatheerthavarada et al. 2003; Lustbader et al. 2004; Reddy and Beal 2005). Overall, in both early-onset familial and late-onset sporadic AD, aging is a major contributing factor for disease development and progression.
How can sirtuins protect AD neurons?
In mouse models of AD, CR has been found to diminish AD symptoms (Mattson et al. 2003; Patel et al. 2005) and, in a primate model, CR increased neurotrophic factors and attenuated behavioral deficits (Maswood et al. 2004), suggesting a potential therapeutic value of CR for patients with AD. However, further investigations are required to evaluate whether CR-induced sirtuins may have any role in diminishing AD symptoms. Resveratrol was found to extend the lifespan of mice through the over-expression of Sir2 or SIRT1 in a yeast diet (Bordone and Guarente 2005) and to be epidemiologically linked to longer life in humans (Luchsinger and Mayeux 2004; Panza et al. 2004). It was also found to protect cells against Aβ-induced ROS production and DNA damage in vitro (Jang and Surh 2003; Russo et al. 2003; Savaskan et al. 2003). In addition, in a rat model of sporadic AD, resveratrol was found to prevent cognitive impairment induced by intracerebroventricular streptozotocin, which may result from its antioxidant effects (Sharma and Gupta 2002). However, it is still not clear how the expression of intracellular sirtuins, triggered by CR or CR mimetics, could mitigate any intracellular accumulation of Aβ.
Possible connection between SIRT1 and AD
A recent study has highlighted the neuroprotective roles of SIRT1 in Huntington's disease (Parker et al. 2005; Sinclair 2005). Resveratrol-induced SIRT1 in neurons from HdhQ111 knock-in mice and Sir2 in the neurons of polyQ mutant TgCaenorhabditis elegans (both models for Huntington's disease) rescued neuronal dysfunction caused by polyQ toxicity (Parker et al. 2005). Similar to Huntingdon's disease neurons, AD neurons may be rescued by the over-expression of SIRT1 (via CR or resveratrol treatment). The precise connection between the over-expression of SIRT1 and protection of neurons from AD patients is not clearly understood, but SIRT1 is undergoing intense investigation in many laboratories.
Insulin/IGF-1 is known to protect neurons against AD by facilitating the clearance of Aβ from the brain or preventing tau hyperphosphorylation (Gasparini and Xu 2003). Lifespan extension studies have found that a decrease in insulin/IGF-1 levels increases SIRT1 expression and that SIRT1 is a primary causal factor for longevity (Cohen et al. 2004). One explanation for this apparent paradox between lifespan extension and neuronal protection is that insulin/IGF-1 is an upstream signal, and its impact on downstream signals might be context dependent (Tang 2005a). Therefore, although SIRT1 in liver, kidney and adipose tissues may contribute to lifespan extension, its role in the CNS may be different. In addition, SIRT1 has been found to be triggered by CR, flavonoids and non-steroidal anti-inflammatory drug treatments. This altered SIRT1 expression suppresses GTPase Rho and Rho-associated kinase, and might promote non-amyloidogenic or non-pathogenic pathways in APP processing (Tang 2005b).
The current literature suggests that the intracellular accumulation of soluble Aβ causes axonopathy and transport deficits about a year before the formation of Aβ deposits (Stokin et al. 2005) and that such deficits may ultimately lead to cognitive deficits in early-onset AD (Billings et al. 2005). Intracellular Aβ 42 was found to stimulate the over-expression of the tumor suppressor and transcription factor p53 and to cause apoptosis in AD neurons (Culmsee and Mattson 2005; Ohyagi et al. 2005).
Figure 1 illustrates the role of SIRT1, SIRT2 and SIRT3 in neuronal protection in AD. We propose two related roles of SIRT1 in the nucleus of AD-affected neurons. First, SIRT1 in the nucleus of AD-affected neurons may deacetylate and repress p53 activity of the neurons and prevent the apoptotic death of these neurons. Second, SIRT1 may deacetylate and suppress apoptotic activities of FOXO proteins and promote neuronal survival, thus providing a novel therapeutic option. FOXOs share functional similarities and cross-talk considerably with p53 (You and Mak 2005). In motoneurons, FOXO3a induces neuronal death through the Fas pathway, in cooperation with c-Jun N-terminal kinase (JNK) (Barthelemy et al. 2004). FOXO proteins directly induce bim gene expression and cause apoptosis in sympathetic neurons (Gilley et al. 2003). Although more definitive research results are lacking, these studies raise the possibility that FOXOs might be involved, either directly or in cooperation with p53, in contributing to neuronal death in AD.
SIRT1 therapeutics comes with some caveats. Because SIRT1 localizes predominantly to the nucleus, SIRT1 is expected to neutralize p53, FOXOs and other pathological protein molecules in the nucleus, but not those proteins localized to other organelles. p53 is also known to localize in cytoplasm and mitochondria (see Table 1 and Fig. 1). If p53 in cytoplasm and mitochondria is contributing to sporadic AD, the dominant form of AD, then SIRT1-based drugs may be not as effective in treating the disease.
Connection between SIRT2 and AD
In studies of SIRT2 knockdown mice, SIRT2 was found to deacetylate Lys40 of α-tubulin and co-localize with the cytoplasmic tubulin network, and the mice showed hyper-acetylated tubulin (North et al. 2003). Interestingly, hyperphosphorylation of the tubulin-associated protein tau, the hallmark pathology of AD, is associated with reduced acetylation of cytoskeletal proteins (for review, see Mattson 2003). It is well known that acetylation increases the stability of microtubules, and any reduction in acetylation may damage neuronal function (Morales and Fifkova 1991). If SIRT2 deacetylates α-tubulin, then how can SIRT2 be useful in stabilizing the cytoskeletal network? One possibility is that SIRT2 may interact with and modulate other cytoskeletal proteins that participate in axonal transport deficits or that are expressed in related neurons exhibiting AD symptoms. A recent study showed that SIRT2 was expressed in mouse olfactory sensory neurons (Yu et al. 2005). Because tubulin is enriched in olfactory sensory neurons and cilia are hyper-acetylated (Poole et al. 2001), there may be maximal activity of SIRT2 in these cilia. Studies have shown that olfactory sensory functions are severely impaired in patients with AD (Getchell et al. 2003). Thus, we speculate that SIRT2 might provide neuroprotective functions in olfactory sensory neurons that have been affected by AD pathology. In addition, a recent study by de la Monte and Wands (2004) has found that CNS neurons with abundant insulin/IGF-1 receptors may be vulnerable to the adverse effects of an AD-associated neuronal thread protein that accumulates in cortical neurons and co-localizes with the tau cytoskeleton. This suggests that, through unknown mechanism(s), SIRT2, which is localized to cytoplasm, may also control the levels of insulin/IGF-1 in the cytoskeleton of AD brains and protect them from the ill-effects of over-expressing insulin/IGF-1 receptors. Finally, findings from de la Monte and Wands (2004) also suggest that SIRT2 may deacetylate p53 in the cytoplasm. These functions of SIRT2 are illustrated in Fig. 1.
Connection between SIRT3 and AD
Recent literature suggests that three independent lines of evidence link mitochondria to aging and AD pathology. First, both familial and sporadic AD pathologies converge in mitochondria, where there is a reduction in mitochondrial ROS levels that substantially increases the lifespan of mice (Schriner et al. 2005). From Schriner et al. (2005), it could be inferred that the increased murine lifespan may be associated with over-expressed catalase activity. A second line of evidence to support this possibility came from two recent studies that focused on the indirect role of SIRT3 in human aging. The survivorship function of the G477T marker in elderly subjects shows that SIRT3 or a SIRT3-linked gene may be related to human longevity (Rose et al. 2003). In an investigation of 945 humans from 20 to 106 years old, variable number tandem repeat (VNTR) polymorphism in intron 5 of SIRT3 was associated with survival at the oldest ages (Bellizzi et al. 2005). Finally, in brown adipose tissue of obese mice, over-expressed SIRT3 triggered a slight up-regulation of UCP1 in the mitochondrial inner membrane, which led to decreased energy expenditure (Shi et al. 2005). Interestingly, the protein UCP4 is expressed only in the inner membrane of brain mitochondria where SIRT3 predominantly localizes (Fig. 1). We propose that SIRT3 may deacetylate p53 and other pathological proteins in mitochondria. Furthermore, as shown in Fig. 1, we propose that the interaction of SIRT3 with UCP4 and the deacetylation activity of SIRT3 might provide critical functions to mitochondria and may attenuate ROS levels or increase O2 consumption, ultimately leading to neuronal protection and healthy longevity.
Sirtuins are emerging as a focus of research in studies of both healthy aging and age-related diseases. Sirtuins have become promising engines of lifespan extension based on findings of their roles in lower organisms, but the implications of these findings for higher organisms, such as monkeys and humans, are still unclear. Findings from lower organisms appear to suggest that sirtuins can help increase healthy lifespan in humans and delay or even stop age-related illnesses, such as AD. Recent research on CR in the aging of rodent models of AD suggests that CR activates sirtuins, decreases ROS, increases O2 consumption and boosts cellular functions. The precise mechanistic connection between the over-expression of sirtuins and extended healthy aging or delaying age-related diseases in humans has yet to be established. Further research on sirtuins, to address issues such the effects of CR and CR mimetics on the expression of sirtuins, may eventually lead to therapeutic strategies for healthy aging and a delay in disease progression in patients with AD.
Authors thank Sandra Oster, PhD, for critical reading of the manuscript. This research was supported, in part, by the American Federation for Aging Research and National Institutes of Health grant AG22643.