Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cell. Author manuscript; available in PMC 2011 July 23.
Published in final edited form as:
PMCID: PMC2911635

This article has been retractedRetraction in: Cell. 2015 March 17; 158(4): 959    See also: PMC Retraction Policy

SIRT1 Suppresses β-amyloid Production by Activating the α-Secretase Gene ADAM10


A hallmark of Alzheimer's disease (AD) is the accumulation of plaques of Aβ 1-40 and 1-42 peptides, which result from the sequential cleavage of APP by the β and γ-secretases. The production of Aβ peptides is avoided by alternate cleavage of APP by the α and γ-secretases. Here we show that production of β-amyloid and plaques in a mouse model of AD are reduced by overexpressing the NAD-dependent deacetylase SIRT1 in brain, and are increased by knocking out SIRT1 in brain. SIRT1 directly activates the transcription of the gene encoding the α-secretase, ADAM10. SIRT1 deacetylates and coactivates the retinoic acid receptor β, a known regulator of ADAM10 transcription. ADAM10 activation by SIRT1 also induces the Notch pathway, which is known to repair neuronal damage in the brain. Our findings indicate SIRT1 activation is a viable strategy to combat AD, and perhaps other neurodegenerative diseases.


Alzheimer's Disease (AD) is a neurodegenerative disorder affecting up to one third of individuals reaching the age of 80 (Tanzi and Bertram, 2005). It displays a stereotypical brain pathology; deposition of plaques of amyloid beta (Aβ) peptides and phosphorylation and tangles of the neurofibrillar protein tau (Hardy and Selkoe, 2002). Affected individuals suffer neurological damage resulting in memory loss, cognitive and functional decline and death.

Insight into the etiology of AD has come from studies of the rare familial form of early onset AD (Selkoe, 1997). Dominant mutations have been found in the gene encoding the neuronal membrane protein amyloid precursor protein (APP), which can be cleaved in two sequential steps by the β- and γ-secretases to generate Aβ 1-40 and 1-42 amyloid peptides (Tanzi and Bertram, 2005). A second class of dominant mutations giving rise to familial AD fall in genes presenilin 1 and 2 (PSEN1 and 2), which encode components of the γ-secretase (De Strooper, 2007). These findings suggest a pathway of AD in which sequential cleavage of APP by the β and γ-secretases leads to accumulating β-amyloid and the downstream pathologies of AD, such as Aβ plaques, tau tangles and neurodegeneration.

Interestingly, the production of Aβ peptides is avoided by an alternate APP cleavage pathway mediated by the α-secretase followed by the γ-secretase (Postina et al., 2004). Indeed, α-secretase cleavage of APP has been shown to be protective in AD models (Mattson, 1997; Kojro and Fahrenholz, 2005). The α– and γ–secretases also sequentially cleave the notch receptor to generate a notch intracellular domain (NICD) (Kojro and Fahrenholz, 2005; van Tetering et al., 2009). The NICD activates nuclear genes, such as HES1 and HES5 to facilitate neurogenesis during development and damage repair in adults (Kopan and Ilagan, 2009; Yoon and Gaiano, 2005; Costa et al. 2005). ADAM10 encodes the α-secretase (Saftig and Hartmann, 2005; Fahrenholz et al., 2008), and transcription of this gene is activated by the retinoic acid receptor (RAR) (Fahrenholz et al., 2008; Prinzen et al., 2005; Tippmann et al., 2009). Indeed, several studies suggest a link between retinoic acid signaling in the brain and AD (Goodman and Pardee, 2003; Corcoran et al., 2004; Goodman et al., 2006).

Sirtuins are NAD-dependent deacetylases that counter aging and have a wide spectrum of metabolic and stress-tolerance functions (Sinclair, 2005). Of the seven mammalian sirtuins, the SIR2 ortholog SIRT1 deacetylates numerous regulatory proteins, such as PGC-1α, p53, FOXO, HSF and HIF-2α to trigger resistance to metabolic, oxidative, heat and hypoxic stress (Guarente, 2009). SIRT1 has been directly implicated in neuronal protection against stress in cultured cells (Qin et al., 2006). In mice, SIRT1 has been shown to protect against neurodegeneration in the p25 over-expression model (Kim et al., 2007), as well as in Wallerian degeneration slow mice (Araki et al. 2004). More generally, SIRT1 mediates at least some of the effects of calorie restriction (Guarente, 2008), a diet reported to protect against models of neurodegenerative diseases, including AD (Patel et al., 2005).

Here, we investigate whether SIRT1 levels in brain affect the Aβ plaque formation, pathology and cognitive decline in Alzheimer's Disease mouse model (APPswe/PSEN1dE9 double transgenic). We show that SIRT1 can suppress AD in a mouse model for this disease. The induction of brain pathology and behavioral deficits was mitigated in AD mice over-expressing SIRT1 in brain, and exacerbated with SIRT1 knocked out in the brain. SIRT1 directly activates transcription of ADAM10, which encodes the α-secretase by deacetylating RARβ. SIRT1 appears to direct APP processing toward the α-secretase and away from the β-secretase, which results in a reduction in the production of toxic β-amyloid peptides. Furhermore, by activating ADAM10, which is also known to cleave the membrane-bound notch receptor thus liberating an intracellular domain that activates nuclear genes for neurogenesis, SIRT1 also activates notch pathway. As a result of this, SIRT1 helps to mitigate AD by increasing neurogenesis and neuroprotection which would be a second way besides suppressing β-amyloid production.


Effects of SIRT1 levels in a mouse model of AD

We wished to test whether the levels of SIRT1 in the brain could influence Aβ plaque formation and the accruing pathological and cognitive decline in a mouse model of AD. We thus employed a model in which two linked transgenes, encoding the human APPswe and PSEN1dE9 alleles drive Aβ plaque formation and learning and memory deficits (Jankowsky et al., 2004). Plaques are evident in this model at 3-5 months of age, and they progress in number and size as mice grow older. In order to test the effects of increasing SIRT1 in these AD mice, they were crossed to strains bearing a SIRT1 transgenic allele (Tg in figures) that results in over-expression of the protein in the brain by about 2-fold (Bordone et al., 2007). Neither the SIRT1 transgene nor the knock-out allele described below affected expression of the APPswe or PSEN1dE9 disease genes (not shown). In all experiments, we compare congenic C57BL/6 littermates by quantifying cortical plaque number in serial coronal cross sections.

We observed a marked reduction of plaques in 5-month old AD transgenic mice over-expressing SIRT1 (AD-Tg) compared to AD control mice (Figure 1A). Wild type mice without the AD disease genes (wt) were free of plaques. In order to inactivate SIRT1 in the brain, we used brain-specific knock out mice (BSKO) (Cohen et al., 2009) which lack full length SIRT1 in all neurons and glia (Figure 1B), but show normal adult brain morphology (Figure S1A). To test the effects of SIRT1 inactivation, BSKO mice were crossed to AD mice. Remarkably, AD-BSKO mice all died between 3 and 5 months of age, while AD mice or BSKO mice themselves lived well past 1.5 years of age (Figure 1C and data not shown). While we have not pinpointed the cause of death in AD-BSKO mice, this synthetic lethality is typically indicative of a strong genetic interaction between the genes involved, further linking SIRT1 to AD pathogenesis. Consistent with this, AD-BSKO mice displayed a marked increase in plaques in 3-month old mice compared to AD control mice (Figure 1D).

Figure 1
SIRT1 levels regulate the pathology in a mouse Alzheimer's Disease model

In order to monitor other features of brain pathology, we assayed for gliosis by immunological quantitation of the glial marker of inflammation, glial fibrillary acidic protein (GFAP) (Panter et al., 1985). A progressive increase in cortical GFAP staining was observed in AD control mice, and this was markedly reduced in AD-Tg mice and exacerbated in AD-BSKO mice (Figure 1E). In addition, phosphorylation of tau at Ser 199 and Ser 399 (Johansson et al., 2006) was evident in AD mice and suppressed in AD-Tg mice (Figures S1B and S1C).

AD mice show age-dependent decline in learning and memory (Reiserer et al., 2006), which can be measured with behavioral tests. In the fear-conditioning test (Bryan et al., 2009), mice were trained on day 1 by repeated exposure to a tone followed by a foot shock (Figure 2A). Memory of this training is evinced on day 2 by a freezing behavior when exposed to the tone without the shock. AD control mice showed a progressive decline in performance in this test at 8 and 11 months of age compared to 4 and 6 months of age (Figure 2B). This decline was clearly suppressed in older, age-matched AD-Tg mice. Wild type and SIRT1 transgenic mice without the AD genes showed roughly normal learning at all ages (Figure 2B). Conversely, AD-BSKO mice were precociously defective in this test at 4 months of age, while BSKO mice without the AD disease genes performed similar to wild type mice (Figure 2C).

Figure 2
SIRT1 improves the behavioral phenotype of Alzheimer's Disease mouse model

In the Morris water maze test, mice are trained to use visual cues to find a submerged platform in an opaque pool (Bryan et al., 2009). By determining the time required to find the platform (escape latency) as a function of days of training, we observed a marked decline in performance in 8 and 10 month old AD mice compared to 4 month olds (Figure 2D). This decline was significantly mitigated in AD-Tg mice. In mice without the AD genes, the SIRT1 transgene had no effect at any ages (Figure 2E), indicating that the suppression of decline in AD-Tg mice is not due to an inherent increase in learning and memory in SIRT1 over-expressing mice. As in the fear conditioning test, AD-BSKO mice displayed a precocious defect at 4 months of age (Figure 2F). We observed no differences in swim speed in any mice tested (Figure S1D).

SIRT1 activates α-secretase and reduces production of β-amyloid

Since Aβ plaques were reduced in SIRT1 transgenic mice, we quantitated Aβ levels in brains of AD mice overexpressing or lacking SIRT1 by ELISA. Strikingly, Aβ 1-42 levels were substantially reduced in AD-Tg mice and enhanced in AD-BSKO mice (Figure 3A). SIRT1 affected Aβ 1-40 in a like manner (Figure S2A). The reduction in Aβ levels in SIRT1 over-expressing mice was largest in 4-month old mice and grew smaller, but was still evident, in older mice. A second assay using immunoblotting also showed similar effects on Aβ levels by SIRT1 (Figure S2B).

Figure 3
SIRT1 reduces Aβ levels and increases α-secretase activity

A reduction of Aβ levels by SIRT1 could reflect a reduced production of the amyloid peptides. To address the potential of mice to produce Aβ, we prepared brain extracts and assayed the α and β-secretases (Kojro and Fahrenholz, 2005; Fahrenholz et al., 2008; Rockenstein et al., 2005). Remarkably, AD-Tg mice showed a doubling of α-secretase activity in 2-, 4-, and 6-month old mice compared to AD-controls (Figure 3B). Conversely, there was a small but significant reduction in BSKO mice with or without the AD genes.

In addition, we observed a small decrease in β-secretase activity (Figure 3C) and BACE1 (β-secretase) RNA (Figure S2C) in SIRT1 over-expressing mice at all ages. However, since β-secretase was induced by the progression of disease in AD mice (Fukumoto et al., 2004), it is likely that the small decrease in this activity in AD-Tg mice is a secondary consequence of a slower disease onset and progression. Consistent with this idea, β-secretase activity (Figure 3C) and BACE RNA levels (Figure S2C) were higher in AD-BSKO mice compared to AD mice. Examination of individual γ-secretase subunits pen2 (Figure S2D), Apha1a, Aph1b (Figure S2E) and presenilin 1 (Figure S2F) showed no differences in mice with altered levels of SIRT1.

The above findings suggest that SIRT1 may suppress AD by directing APP processing toward the α-secretase. To obtain further evidence for this hypothesis, we quantitated two of the products of α-secretase cleavage. First, α-secretase generates a large α-fragment ectodomain of APP (Figure 3D), which migrates slightly faster than APP in a gel (Lesne et al., 2006; Vieira et al., 2009). The ratio of α-fragment to APP was clearly increased in AD-Tg mice and decreased in AD-BSKO mice (Figure 3E). We have also determined the levels of α-fragment by using an α-fragment-specific antibody (APPalpha antibody), and observed that the levels of α-fragment in AD-Tg mice is increased compared to AD mice (Figure S2G). Second, the other product of APP cleavage by α-secretase is a fragment termed α-CTF (Lesne et al., 2006; Vieira et al., 2009) (Figure 3D), and again, levels of this protein were increased in AD-Tg mice (Figure 3F and 3H, performed by using two different antibodies). We have also analyzed the products of APP cleavage by beta secretase (Figure 3D), β-fragment and β-CTF. Both were decreased in AD-Tg mice compared to AD mice (Figure 3G and H). Together, all of the above findings indicate that α-secretase is activated by SIRT1.

SIRT1 activates ADAM10 transcription and notch pathway

We next determined RNA levels in the brain of ADAM10, which encodes the α-secretase (Saftig and Hartmann, 2005; Fahrenholz et al., 2008) and found a significant increase in SIRT1 over-expressing mice with or without the AD disease genes (Figure 4A). Conversely, AD-BSKO mice showed a significant decrease in ADAM10 RNA. We also observed elevated levels of ADAM10 and its precursor ADAM10-P in SIRT1 over-expressing mice with (Figure 4B, left panel and Figure S3A) or without (Figure 4B, middle panel) the AD genes, and decreased levels in AD-BSKO mice (Figure 4B, right panel). The ADAM10 homologs ADAM 9 and ADAM 17 (Saftig and Hartmann, 2005) did not show clear changes in SIRT1 over-expressing or BSKO mice with or without the AD disease genes (Figure S3B).

Figure 4
SIRT1 increases α-secretase levels by activating ADAM10

What might be the biological relevance of SIRT1 activation of ADAM10? This protease is known to cleave the notch receptor to release a notch intracellular domain (NICD), which activates transcription of genes involved in neurogenesis and patterning during embryonic development and the maintenance of brain function in adults (Yoon and Gaiano, 2005; Costa et al. 2005). To determine whether SIRT1 regulates the notch pathway in adults, we first quantitated levels of the NICD in brain extracts. SIRT1 Tg mice with or without the AD disease genes showed a significant elevation in NICD levels in AD-Tg mice compared to AD controls (Figure 4C). Moreover, transcription of the notch nuclear target genes HES1 and HES5 (Kopan and Ilagan, 2009; Yoon and Gaiano, 2005) were also significantly elevated in AD-Tg mice compared to AD mice (Figure 4D).

Activation of ADAM10 depends on deacetylase activity of SIRT1 and retinoic acid

The above findings suggest that SIRT1 is a positive activator of ADAM10 transcription. Consistent with this surmise, chromatin immuno-precipitation assays revealed direct interaction of SIRT1 with a specific region of the ADAM10 promoter -467 to -30 but not -2100 to -1900 (Figure 5A and S4A). This gene-proximal region of ADAM10 is where the retinoic acid receptor (RAR) is known to bind and activate transcription (Prinzen et al., 2005). Cellular effects of retinoic acid (RA) are mediated by the retinoic acid receptors (RARs) and retinoic X receptors (RXRs), which form heterodimers and bind to RAREs (retinoic acid response element) to activate gene expression (Mangelsdorf et al., 1995).

Figure 5
SIRT1 suppression of Aβ production is mediated by ADAM10

We next investigated the mechanism of ADAM10 activation by SIRT1 in cultured cells. SIRT1 or the catalytically inactive and dominant negative mutant His363-Tyr (SIRT1HY) were expressed in mouse embryo fibroblasts (MEFs) from SIRT1+/+ and SIRT1-/- embryos (Figure S4B) and endogenous ADAM10 RNA (Figure 5B) or NICD levels (Figure 5C) were quantitated. Expression of SIRT1 but not SIRT1HY increased levels of ADAM10 RNA and NICD protein in SIRT1-/- MEFs, indicating that the deacetylase activity of SIRT1 is required for ADAM10 activation. Further, addition of RA also increased ADAM10 RNA and NICD protein in SIRT1+/+ but not SIRT1-/- MEFs, and this activation was further enhanced by SIRT1 and blocked by SIRT1-HY. These findings suggest that the RAR is the target of SIRT1 activation of ADAM10. To bolster this claim, we also expressed SIRT1 or SIRT1-HY in N2A neuroblastoma cells (Figure S4C). Expression of SIRT1 but not SIRT1-HY activated ADAM10 RNA (Figure S4D) and NICD protein levels (Figure S4E) in N2A cells. Further, addition of RA also increased ADAM10 RNA and NICD protein, and this activation was further enhanced by SIRT1 and blocked by SIRT1-HY as in MEF cells.

SIRT1 deacetylates Retinoic Acid Receptor β

SIRT1 is known to deacetylate numerous mammalian transcription factors (Donmez and Guarente, 2010). To further investigate the relationship between SIRT1 and the RAR, we studied a luciferase reporter driven by the RARE in MEFs (Figure 6A). Luciferase levels were activated by SIRT1 over-expression and reduced in SIRT1-/- MEFs. Moreover, addition of retinoic acid stimulated luciferase while preserving the differences in SIRT1 over-expressing and SIRT1-/- MEFs. In contrast, the RAR antagonist LE 135 (Li et al., 1999) blocked activity in all cases. These findings clearly indicate that SIRT1 coactivates the RAR. Next, we carried out immunoprecipitation of endogenous SIRT1 on antibody-coated beads (Figure 6B). We observed co-immunoprecipitation of endogenous RARβ, the major isoform in the brain (Mangelsdorf and Evans, 1992; Dolle et al., 1990). As a specifity control, we checked LXR levels (another nuclear receptor) in the immunoprecipitates and observed none (Figure S5A). Finally, we found that acetylation of endogenous RARβ was enhanced in SIRT1-/- cells and reduced in SIRT1 over-expressing cells, further indicating that SIRT1 co-activates RARβ by deacetylating it. (Figure 6C). Also, acetylation of endogenous RARβ was unaffected by expressing catalytic mutant of SIRT1, SIRT1-HY in SIRT1-/- cells, indicating that catalytic domain of SIRT1 is required for deacetylation. We did not observe acetylation of endogenous RARα when identical immunoprecipitation experiments were performed using RARα antibody (Figure S5B), indicating the effects of SIRT1 are specific for RARβ.

Figure 6
SIRT1 deacetylates RARβ

SIRT1 suppresses Aβ production because it activates ADAM10

Does the increase in ADAM10 in SIRT1 overexpressing mice account for the reduction in β-amyloid production? To address this question, we assayed Aβ1-42 produced by N2A neuroblastoma cells expressing the APPswe and PSEN1dE9 transgenes (Borchelt et al., 1996). Over-expression of SIRT1 in these cells substantially increased ADAM10 (Figure 7B) and reduced Aβ levels (Figure 7A) compared to control cells, analogous to what we observed in mice. Conversely, silencing ADAM10 with 3 different shRNAs alone (Figure 7B) increased Aβ (Figure 7A). Of critical importance, in cells over-expressing SIRT1, ADAM10-shRNAs reduced ADAM10 to levels of control cells (Figure 7B) and also prevented the suppression in Aβ production by this sirtuin (Figure 7A). We also employed scrambled shRNA as control which did not have any silencing effect on ADAM10 (Figures S6A, S6B, S6C). This finding indicates SIRT1 over-expression reduces Aβ production by increasing ADAM10. Finally, as in the mice, the reduction of SIRT1 levels (by 3 different shRNAs 1, 2, and 3) increased Aβ production in these cells (Figures 7A, 7B and S6D). As a test for the specificity of these ADAM10 shRNAs, we showed that their effects were rescued by expressing ADAM10 mutant genes that were resistant to shRNA 1 or an siRNA to the ADAM10 3′ UTR (Figures S6E and S7A).

Figure 7
SIRT1 suppresses Aβ production by activating ADAM10

Correspondingly, NICD levels (Figure 7C) and expression of the notch targets HES1 (Figure 7D) and HES5 (Figure S7B) were increased by SIRT1 over-expression and decreased by shRNA for SIRT1 or ADAM10 in these cells. Most importantly, shRNA for ADAM10 reversed the increase in notch activity in SIRT1 over-expressing cells (Figures 7C and 7D). The above analysis draws a causal link between SIRT1 expression, ADAM10 activation and Aβ suppression. A recent report showed that SIRT1 could affect α-secretase activity in cultured cells by inhibiting expression of the rho-associated protein kinase ROCK1 (Qin et al., 2006). However, we observed no differences in ROCK1 levels in mice with altered levels of SIRT1 (Figure S7C).

In order to show that RARβ activation by SIRT1 suppresses Aβ production, we treated N2A neuroblastoma cells expressing the APPswe and PSEN1dE9 transgenes with RA and RAR antagonist LE135. RA decreased Aβ production thereby augmenting the SIRT1 effect, while LE135 increased Aβ production thereby blocking the effect of SIRT1 (Figure 7E). These data show that RARβ activation by SIRT1 affects Aβ production.


Our findings clearly indicate that SIRT1 can suppress AD in a mouse model for this disease. The induction of brain pathology and behavioral deficits in doubly transgenic APPswe/PSEN1dE9 mice was mitigated in mice over-expressing SIRT1 in the brain, and exacerbated in mice with SIRT1 knocked out in the brain. SIRT1 appears to direct APP processing toward the α-secretase and away from the β-secretase, which results in a reduction in the production of toxic β-amyloid peptides (Figure 6D). Indeed, SIRT1 directly activates transcription of ADAM10, which encodes the α-secretase. ADAM10 is also known to initiate activation of the notch pathway by cleaving the membrane-bound notch receptor thus liberating an intracellular domain that activates nuclear genes for neurogenesis (Costa et al., 2005; Hartmann et al., 2001). Studies in neuroblastoma cells show that the suppression of Aβ production by SIRT1 is prevented by normalizing ADAM10 levels with specific shRNAs. Thus we conclude that the activation of ADAM10 by SIRT1 is mechanistically linked to Aβ suppression.

The mechanism of activation of ADAM10 by SIRT1 appears to be deacetylation of the RARβ, which is known to activate ADAM10 transcription. SIRT1 but not a catalytically inactive mutant activated ADAM10 transcription and the notch pathway in neuroblastoma cells and MEFs. Moreover, SIRT1 coactivated RAR in reporter assays, bound to the receptor at endogenous levels, and determined its acetylation level. SIRT1 has also been shown to deacetylate and regulate other members of the nuclear receptor family, such as LXR, PPARγ, PPARα and the androgen receptor.

We can not be certain that the reduction in Aβ production in SIRT1 over-expressing mice is the sole mechanism conferring protection against memory decline in AD mice. First, activation of notch or perhaps other pathways by the increase in ADAM10 may also contribute to AD protection, since this pathway mediates repair in response to neuronal damage in the brain (Vieira et al., 2009; Mangelsdorf and Evans, 1995). Second, SIRT1 may also increase stress tolerance in neurons of the AD brain. Third, RAR targets other than ADAM10 may play some role in protection, since RAR signaling is known to be disrupted in the AD brain (Goodman and Pardee, 2003; Corcoran et al., 2004; Goodman, 2006). However, we can be confident that the increase in ADAM10 and resulting reduction in Aβ will benefit AD mice. Indeed, ADAM10 has been shown to prevent amyloid plaque formation and hippocampal behavioral defects in AD mice (Postina et al., 2004). Further, ADAM10 mutations have been associated with a reduction in α-secretase in familial late-onset AD (Kim et al., 2009).

Mutations that severely reduce notch signaling result in embryonic lethality, because of a deficit in brain development (Yoon and Gaiano, 2005). Brain-specific SIRT1 KO mice activate cre in early embryos, yet are born in expected numbers and do not display any detectable defects in adult brain morphology or in performing behavioral tasks. It is possible that the reduction of notch signaling in the absence of SIRT1 is not sufficiently severe to cause dysfunction during embryonic development. Alternatively, the activation of notch by SIRT1 may have evolved for maintenance of the adult brain, rather than development.

SIRT1 has been implicated in protection against metabolic syndrome and diabetes, and sirtuin activators may offer promising new treatments for these increasingly common disorders (Milne et al., 2007). Recent studies also implicate SIRT1 in tumor protection, at least in some mouse models (Firestein et al., 2008). Our findings indicate that another important disease of aging, Alzheimer's Disease, is also mitigated by genetic activation of SIRT1. It may therefore be critically important to develop sirtuin activators tailored to cross the blood brain barrier to treat neurodegenerative diseases. The broad efficacy of SIRT1 activation in treating diseases of aging may derive from evolutionary selection for tissue maintenance to allow for delayed reproduction in the face of dietary or other stressors. The challenge is to harness this protection pharmacologically in a developed world, in which many of these stressors no longer exist.

Experimental Procedures

Mouse strains

All mice used were in congenic C57Bl/6 except SIRT1-/- mice for CHIP assay. Double transgenic (APPswe,PSENdE9) Alzheimer's Disease mouse model was purchased from Jackson Laboratory. SIRT1 transgenic mice have been described previously (Bordone et al., 2007). SIRT1 brain-specific knockout mice were generated by crossing a SIRT1 allele containing a floxed exon 4 (Cheng et al., 2003) with Cre-expressing mice driven by the brain-specific nestin promoter (Cohen et al., 2007). All mice were housed at controlled temperature (25°C) and 12:12 h light-dark cycle.


Mice were perfused with 4 % paraformaldehyde, cryoprotected, sectioned 40 μm-thick and collected at 150 μm intervals. 6 sections per brain were analyzed and consecutive sections of cortex were used for Aβ and GFAP staining. Vectastain kit (Vector laboratories) were used to perform Aβ and GFAP staining according to manufacturer's directions. Thionin staining was performed as described (Pozniak et al., 2002). For Aβ and GFAP stainings, 6E10 (Covance) and GFAP (Abcam) antibodies were used, respectively. The plaque numbers, the sizes of the plaques and GFAP staining were quantified using NIH ImageJ program.

Western blotting and Immunoprecipitation

Mouse brains were homogenized in RIPA buffer (50 mM Tris-HCl pH: 8.0, 1mM EDTA, 0.1% SDS, 150 mM NaCl, 1% NP40, 0.1 % Sodium-deoxycholate) including Complete Protease Inhibitor mixture (Roche), centrifuged, 100 μg of the supernatant was loaded onto 4-15 % gradient SDS-PAGE gels and immunoblotted with anti-SIRT1 (Upstate), 6E10, APPα, APPβ, APP C-terminal (Covance, Figure 3F), APP C-terminal (Santa Cruz, Figure 3H), ADAM10, ADAM10-Prodomain, ADAM9, ADAM17, Aph1a, Aph1b, Pen2, Rock1, NICD, p-Tau Ser399, RARα, RARβ, RXR, LXR (Abcam), Presenilin1 (Cell Signalling), Ac-K (Immuno-chem) antibodies. In order to separate the APPα fragment from APP (Figure 3E), 10-20 % SDS-PAGE gels were used and run for 2.5 hours. For western blotting using cells, cells were harvested and extracted in RIPA buffer as explained above. Proteosome Inhibitors were used when analyzing NICD. Western blotting experiments were performed with at least three mice from each genotype and age and representative data are shown.

The immunoprecipitations were carried out by using Pierce Direct IP Kit (Thermo Scientific). Immobilizing the antibody covalently to agarose beads, this method results in purified antigen free from antibody contamination To show the endogenous interaction between SIRT1 and RARβ, anti-SIRT1 antibody or normal rabbit serum (NRS) was couple to beads and then wild type (wt) and SIRT1-/- (SIRT1-/-) MEFs were incubated with the beads. To determine RARβ acetylation, anti-RARβ antibody or NRS was coupled to beads and then, wild type (wt), SIRT1-/- (SIRT1-/-), SIRT1-/- overexpressing SIRT1-HY and SIRT1-overexpressing (mSIRT1) MEFs were incubated with the beads. The eluate was blotted with anti-SIRT1, anti-RARβ and anti-Ac-K antibodies.

Fear conditioning

10-15 mice from each indicated genotype and age were trained and tested in a fear conditioning device containing one observation chamber (30 × 24 × 21 cm; MED-Associates, Inc.). Fear to the tone was assessed by measuring freezing behavior. The output from video camera mounted in the chamber was fed into a video processor, and the mice were videotaped throughout each of the sessions. Freezing behavior, defined as the absence of all movements except for the breathing, was scored by the computer program. Freezing was quantified by computing the percentage of observations in which the mouse had been scored as freezing during test. The experimenter was blind to the genotypes and ages of the mice.

Morris Water Maze

Morris water maze testing procedure has been described previously (Bryan et al., 2009; Cao et al., 2007). 6-12 mice from each indicated genotype and age were trained to find the visible platform three trials a day for the first day and tested to find the hidden platform for 7 consecutive days. In each trial, the mice were allowed to swim until it found the hidden platform, or until 2 min had elapsed and, at which point the mouse was guided to the platform. The mouse was then allowed to sit on the platform for 10 s before being picked up. During the test days, the platform was hidden 1cm beneath the water. The escape latency was recorded by a video camera. The experimenter was blind to the genotypes and ages of the mice. The swim speed of each mouse was calculated by video tracking system and MATLAB software.

ELISA and secretase activity assays

Aβ 1-42 and Aβ 1-40 levels of brain tissue were quantitated by using Aβ 1-42 and Aβ 1-40 ELISA kits (Biosource), respectively. Phosphorylation levels of pTau Ser199 was quantitated by using ELISA (Sigma). α- and β-secretase activities were quantitated by using α- and β-secretase activity assay kits (R&D Systems), respectively. Aβ 1-42 levels of the conditioned media from APPswe/PSEN1dE9 cells were measured by using Aβ 1-42 ELISA kit (Biosource). All ELISA and secretase activity assays were performed according to the manufacturer's directions.

RNA isolation and analysis

Total RNA from mouse brains was isolated by using Trizol (Qiagen). For real time q-PCR analysis, cDNA was synthesized from total RNA by SuperScript III reverse transcriptase (Invitrogen) with random primers. The cDNA was then subjected to PCR analysis with gene specific primers in the presence of CYBR green (Bio-Rad). Relative abundance of mRNA was obtained by normalization to 18S levels.

Chromatin Immunoprecipitation Analysis (CHIP)

CHIP analysis from mouse brains was performed using Magna Chip kit (Millipore) according to manufacturer's directions by using SIRT1 antibody (Upstate) and IgG. The primer sequences for ADAM10 promoter are as the following. Primer 1 forward: AGAAGCCGAAAGCCCTTCT, primer 1 reverse: GAGCGCTCTTTTCCCTGTG, primer 2 forward: AGGCCAATCCCTGCTC TC, primer 2 reverse: GACGGCACCCAATACACTC primer 3 forward: TATCGCGGC TAAAATCATGG, primer 3 reverse: CTACCCCGAAGCTGTCAAGA, primer 4 forward: GGCCTCGCGAGTGTATTG primer 4 reverse: CTTCCCTGCCCTCGCT CT Upstream primer 1 forward: CAAACCCCAGGCTCTGTTTA Upstream primer 1 reverse: CAGGACCTCTGGAAG AGCAG Upstream primer 2 forward: CCAGCAACC ACATGGTGA Upstream primer 2 reverse: TTGATCTTTTGGGT TTGTTTCTC Upstream primer 3 forward: GAGAAACAAACCCAAAAGATCAA Upstream primer 3 reverse: CCCTCTCCCAGGGTTCTCT

Cells and transfection

The plasmids pBABE-mSIRT1 and SIRT1+/+ and SIRT1-/-MEFs have been described (Rockenstein et al., 2005). N2A cells were from ATCC. N2A cells stably overexpressing APPswe and PSEN1dE9 were kindly provided by Hyo-Jin Park and Seong Hun Kim (University of Florida). The plasmids expressing mSIRT1 and mSIRT1-HY were purchased from Addgene. Three SIRT1-shRNA plasmids and three ADAM10-shRNA plasmids were purchased from Open Biosystems. Transfections were performed by using Effectene transfection reagent (Qiagen). ADAM10-shRNA sequences used were CCAGGAGAGTCTAAGAACTTA (shRNA 1), CAGCTCTA TATCCAGACAGAT (shRNA2), GAGTTATCAAATGGGACACAT (shRNA3) or siRNA (GCACAAAGUCUUAGAAUAUUU) for the 3′ UTR. ADAM10-ORF clone was purchased from Open Biosystems and siRNA against ADAM10 3′ UTR was purchased from Thermo Scientific. siRNA and ADAM10-ORF plasmid were cotransfected with DharmaFECT Duo transfection reagent from Thermo Scientific. Via PCR, silent mutations were introduced within the ADAM10 protein coding region (wt shRNA1 target: CCA GGA GAG TCT AAG AAC TTA, mutant shRNA1 target: CCC GGT GAA AGC AAA AAT CTC) to render the gene resistant to shRNA 1. Cells were treated with Retinoic Acid (Sigma) and LE 135 (Axon Medchem) of final concentration 1 μM and 5 μM, respectively as described (Li et al., 1999; Dolle et al., 1990).

Luciferase assay

The plasmid expressing RARE-luciferase was purchased from Addgene. Cells were cotransfected with RARE-luciferase reporter and pRL-TK (Renilla luciferase; Promega), 24 hr later treated with RA and LE-135 for additional 48 hrs. Luciferase activity was then measured by using the Dual-Luciferase Reporter Assay System (Promega). The final GL (firefly luciferase) activity was normalized to RL (Renilla luciferase) activity. The experiments were performed in triplicates and were repeated three times.

Statistical Analysis

The analysis was performed using Student's t test, and significant differences are demonstrated by single asterisk (*) indicating p < 0.01. Error bars in figures represent SEM.

Supplementary Material



We are grateful to Hyo-Jin Park and Seong Hun Kim for N2A APPswe/PSEN1dE9 cells. We thank Dennis Selkoe for helpful discussions. We also thank Takashi Nakagawa, Hung Chun Chang and Petra Simic for comments on the manuscript. This work is supported by an American Parkinson Disease Association post-doctoral fellowship to G.D. and grants from the NIH and a gift from the Paul F. Glenn Foundation to L.G.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Araki T, Sasaki Y, Milbrandt J. Increased nuclear NAD biosynthesis and SIRT1 activation prevent axonal degeneration. Science. 2004;305(5686):1010–3. [PubMed]
  • Borchelt DR, Thinakaran G, Eckman CB, Lee MK, Davenport F, Ratovitsky T, Prada CM, Kim G, Seekins S, Yager D, Slunt HH, Wang R, Seeger M, Levey AI, Gandy SE, Copeland NG, Jenkins NA, Price DL, Younkin SG, Sisodia SS. Familial Alzheimer's Disease-linked Presenilin 1 variants elevate Aβ 1-42/1-40 ratio in vitro and in vivo. Neuron. 1996;17:1005–1013. [PubMed]
  • Bordone L, Cohen D, Robinson A, Motta MC, van Veen E, Czopik A, Steele AD, Crowe H, Marmor S, Luo J, Gu W, Guarente L. SIRT1 transgenic mice show phenotypes resembling calorie restriction. Aging Cell. 2007;6:759–67. [PubMed]
  • Bryan KJ. Methods of Behavior Analysis in Neuroscience. CRC Press; 2009. Transgenic mouse models of Alzheimer's disease: behavioral testing and considerations. [PubMed]
  • Cheng HL, Mostoslavsky R, Saito S, Manis JP, Gu Y, Patel P, Bronson R, Appella E, Alt FW, Chua KF. Developmental defects and p53 hyperacetylation in Sir2 homolog (SIRT1)-deficient mice. PNAS. 2003;100:10794–9. [PubMed]
  • Cao D, Lu H, Lewis TL, Li L. Intake of sucrose-sweetened water induces insulin resistance and exacerbates memory deficits and amyloidosis in a transgenic mouse model of Alzheimer Disease. J Biol Chem. 2007;282:36275–36282. [PubMed]
  • Cohen DE, Supinski AM, Bonkowski MS, Donmez G, Guarente LP. Neuronal SIRT1 regulates endocrine and behavioral responses to calorie restriction. Genes Dev. 2009;23:2812–2817. [PubMed]
  • Costa RM, Drew J, Silva AJ. Notch to remember. Trends in Neuroscience. 2005;28:429–435. [PubMed]
  • Corcoran JPT, So PL, Maden M. Disruption of the retinoid signaling pathway causes a deposition of amyloid β in the adult rat brain. European J Neuroscience. 2004;20:896–902. [PubMed]
  • De Strooper B. Loss-of-function presenilin mutations in Alzheimer disease. Talking Point on the role of presenilin mutations in Alzheimer disease. EMBO Rep. 2007;8(2):141–6. [PubMed]
  • Dollé P, Ruberte E, Leroy P, Morriss-Kay G, Chambon P. Retinoic acid receptors and cellular retinoid binding proteins 1. A systemic study of their differential pattern of transcription during mouse organogenesis. Development. 1990;110:1133–1151. [PubMed]
  • Donmez G, Guarente L. Aging and disease: connections to sirtuins. Aging Cell. 2010;9:285–290. [PubMed]
  • Fahrenholz F. Up-regulation of the α-secretase pathway. In: Fisher A, editor. Advances in Alzheimer's and Parkinson's disease. Chapter 38. Springer; 2008. pp. 369–374.
  • Firestein R, Blander G, Michan S, Oberdoerffer P, Ogino S, Campbell J, Bhimavarapu A, Luikenhuis S, de Cabo R, Fuchs C, Hahn WC, Guarente LP, Sinclair DA. The SIRT1 deacetylase suppresses intestinal tumorigenesis and colon cancer growth. Plos One. 2008;3:e2020. [PMC free article] [PubMed]
  • Fukumoto H, Rosene DL, Moss MB, Raju S, Hyman BT, Irizarry MC. β-secretase activity increases with aging in human, monkey and mouse brain. Am J Pathol. 2004;164:719–725. [PubMed]
  • Goodman AB, Pardee AB. Evidence for defective retinoid transport and function in late onset Alzheimer's disease. PNAS. 2003;100:2901–205. [PubMed]
  • Goodman AB. Retinoid receptors, transporters, and metabolizers as therapeutic targets in late onset Alzheimer disease. J Cell Physiology. 2006;209:598–603. [PubMed]
  • Guarente L. Hypoxic hookup. Science. 2009;324(5932):1281–2. [PubMed]
  • Guarente L. Mitochondria--a nexus for aging, calorie restriction and sirtuins? Cell. 2008;132:171–6. [PMC free article] [PubMed]
  • Hardy J, Selkoe DJ. The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science. 2002;297(5580):353–6. [PubMed]
  • Hartmann D, Tournoy J, Saftig P, Annaert W, De Strooper B. Implication of APP secretases in notch signaling. J Mol Neurosci. 2001;17:171–81. [PubMed]
  • Jankowsky JL, Fadale DJ, Anderson J, Xu GM, Gonzales V, Jenkins NA, Copeland NG, Lee MK, Younkin LH, Wagner SL, Younkin SG, Borchelt DR. Mutant presenilins specifically elevate the levels of the 42 residue β-amyloid peptide in vivo: evidence for augmentation of a 42-specific γ-secretase. Hum Mol Genet. 2004;13:159–170. [PubMed]
  • Johansson S, Jämsä A, Vasänge M, Winblad B, Luthman J, Cowburn RF. Increased tau phosphorylation at the Ser396 epitope after amyloid beta-exposure in organotypic cultures. Neuroreport. 2006;17:907–911. [PubMed]
  • Kim D, Nguyen MD, Dobbin MM, Fischer A, Sananbenesi F, Rodgers JT, Delalle I, Baur JA, Sui G, Armour SM, Puigserver P, Sinclair DA, Tsai LH. SIRT1 deacetylase protects against neurodegeneration in models for Alzheimer's disease and amyothrophic lateral sclerosis. EMBO J. 2007;26(13):3169–79. [PubMed]
  • Kim M, Suh J, Romano D, Truong MH, Mullin K, Hooli B, Norton D, Tesco G, Elliott K, Wagner SL, Moir RD, Becker KD, Tanzi RE. Potential late-onset Alzheimer's disease-associated mutations in the ADAM10 gene attenuate α-secretase activity. Hum Mol Genet. 2009;19:3987–3996. [PMC free article] [PubMed]
  • Kojro E, Fahrenholz F. The non-amyloidogenic pathway: structure and function of α-secretases. Subcell Biochem. 2005;38:105–127. [PubMed]
  • Kopan R, Ilagan XG. The canonical notch signaling pathway: unfolding the activation mechanism. Cell. 2009;137:216–33. [PMC free article] [PubMed]
  • Lesné S, Koh MT, Kotilinek L, Kayed R, Glabe CG, Yang A, Gallagher M, Ashe KH. A specific amyloid-β protein assembly in the brain impairs memory. Nature. 2006;440:352–357. [PubMed]
  • Mangelsdorf DJ, Evans RM. The RXR heterodimers and orphan receptors. Cell. 1995;83:841–850. [PubMed]
  • Mangelsdorf DJ, Evans RM. Retinoid receptors as transcription factors. Cold Spring Harbor Press; 1992.
  • Mattson MP. Cellular actions of beta-amyloid precursor protein and its soluble and fibrillogenic derivatives. Pysiol Rev. 1997;77(4):1081–132. [PubMed]
  • McBurney MW, Yang X, Jardine K, Hixon M, Boekelheide K, Webb JR, Lansdorp PM, Lemieux M. The mammalian Sir2a protein has a role in embryogenesis and gametogenesis. Mol Cell Biol. 2003;23:38–54. [PMC free article] [PubMed]
  • Milne JC, Lambert PD, Schenk S, Carney DP, Smith JJ, Gagne DJ, Jin L, Boss O, Perni RB, Vu CB, Bemis JE, Xie R, Disch JS, Ng PY, Nunes JJ, Lynch AV, Yang H, Galonek H, Israelian K, Choy W, Iffland A, Lavu S, Medvedik O, Sinclair DA, Olefsky JM, Jirousek MR, Elliott PJ, Westphal CH. Small molecule activators of SIRT1 as therapeutics for the treatment of type 2 diabetes. Nature. 2007;450:712–6. [PMC free article] [PubMed]
  • Panter SS, McSwigan JD, Sheppard JR, Emory CR, Frey WH., 2nd Glial fibrillary acidic protein and Alzheimer's disease. Neurochem Res. 1985;10:1567–76. [PubMed]
  • Patel NV, Gordon MN, Connor KE, Good RA, Engelman RW, Mason J, Morgan DG, Morgan TE, Finch CE. Caloric restriction attenuates Aβ-deposition in Alzheimer transgenic models. Neurobiol Aging. 2005;26:995–1000. [PubMed]
  • Postina R, Schroeder A, Dewachter I, Bohl J, Schmitt U, Kojro E, Prinzen C, Endres K, Hiemke C, Blessing M, Flamez P, Dequenne A, Godaux E, van Leuven F, Fahrenholz F. A disintegrin-metalloproteinase prevents amyloid plaque formation and hippocampal defects in an Alzheimer disease mouse model. J Clin Invest. 2004;113(10):1384–7. [PMC free article] [PubMed]
  • Pozniak CD, Barnabé-Heider F, Rymar VV, Lee AF, Sadikot AF, Miller FD. p73 is required for survival and maintenance of CNS neurons. J Neurosci. 2002;22:9800–9. [PubMed]
  • Prinzen C, Müller U, Endres K, Fahrenholz F, Postina R. Genomic structure and functional characterization of the human ADAM10 promoter. FASEB Journal. 2005;19:1522–4. [PubMed]
  • Qin W, Yang T, Ho L, Zhao Z, Wang J, Chen L, Zhao W, Thiyagarajan M, MacGrogan D, Rodgers JT, Puigserver P, Sadoshima J, Deng H, Pedrini S, Gandy S, Sauve AA, Pasinetti GM. Neuronal SIRT1 activation as a novel mechanism underlying prevention of Alzheimer Disease amyloid neuropathology by calorie restriction. J Biol Chem. 2006;281:21745–21754. [PubMed]
  • Ramsey KM, Mills KF, Satoh A, Imai S. Age-associated loss if SIRT1-mediated enhancement of glucose-stimulated insulin secretion in beta cell-specific SIRT1-overexpressing (BESTO) mice. Aging Cell. 2008;7:78–88. [PMC free article] [PubMed]
  • Reiserer RS, Harrison FE, Syverud DC, McDonald MP. Impaired spatial learning in the APPswe, PSEN1dE9 bigenic mouse model of Alzheimer's disease. Genes Brain Behav. 2006;6:54–65. [PubMed]
  • Rockenstein E, Mante M, Alford M, Adame A, Crews L, Hashimoto M, Esposito L, Mucke L, Masliah E. High beta-secretase activity elicits neurodegeneration in transgenic mice despite reductions in amyloid-beta levels: implications for the treatment of Alzheimer disease. J Biol Chem. 2005;280(38):32957–67. [PubMed]
  • Saftig P, Hartmann D. ADAM10: A major membrane protein ectodomain sheddase involved in regulated intramembrane proteolysis. In: Hooper NM, Lendeckel U, editors. The ADAM family of proteases. 2005. pp. 85–121.
  • Selkoe DJ. Alzheimer's disease: genotypes, phenotypes and treatments. Science. 1997;275:630–631. [PubMed]
  • Sinclair DA. Toward a unified theory of caloric restriction and longevity regulation. Mech Aging Dev. 2005;126:987–1002. [PubMed]
  • Tanzi RE, Bertram L. Twenty years of the Alzheimer's disease amyloid hypothesis: a genetic perspective. Cell. 2005;120(4):545–55. [PubMed]
  • Tippmann F, Hundt J, Schneider A, Endres K, Fahrenholz F. Up-regulation of the α-secretase ADAM10 by retinoic acid receptors and acitretin. FASEB Journal. 2009;23(6):1643–54. [PubMed]
  • van Tetering G, van Diest P, Verlaan I, van der Wall E, Kopan R, Vooijs M. The metalloprotease ADAM10 is required for Notch1 S2 cleavage. J Biol Chem. 2009;284:31018–31027. [PMC free article] [PubMed]
  • Vieira SI, Rebelo S, Domingues SC, da Cruz e Silva EF, da Cruz e Silva OA. S655 phosphorylation enhances APP secretory traffic. Mol Cell Biochem. 2009;9:84–7. [PubMed]
  • Yoon K, Gaiano N. Notch signaling in the mammalian central nervous system: insights from mouse mutants. Nature Neuroscience. 2005;8:709–715. [PubMed]