Sirtuins in Alzheimer’s Disease: SIRT2-Related GenoPhenotypes and Implications for PharmacoEpiGenetics

Sirtuins (SIRT1-7) are NAD+-dependent protein deacetylases/ADP ribosyltransferases with important roles in chromatin silencing, cell cycle regulation, cellular differentiation, cellular stress response, metabolism and aging. Sirtuins are components of the epigenetic machinery, which is disturbed in Alzheimer’s disease (AD), contributing to AD pathogenesis. There is an association between the SIRT2-C/T genotype (rs10410544) (50.92%) and AD susceptibility in the APOEε4-negative population (SIRT2-C/C, 34.72%; SIRT2-T/T 14.36%). The integration of SIRT2 and APOE variants in bigenic clusters yields 18 haplotypes. The 5 most frequent bigenic genotypes in AD are 33CT (27.81%), 33CC (21.36%), 34CT (15.29%), 34CC (9.76%) and 33TT (7.18%). There is an accumulation of APOE-3/4 and APOE-4/4 carriers in SIRT2-T/T > SIRT2-C/T > SIRT2-C/C carriers, and also of SIRT2-T/T and SIRT2-C/T carriers in patients who harbor the APOE-4/4 genotype. SIRT2 variants influence biochemical, hematological, metabolic and cardiovascular phenotypes, and modestly affect the pharmacoepigenetic outcome in AD. SIRT2-C/T carriers are the best responders, SIRT2-T/T carriers show an intermediate pattern, and SIRT2-C/C carriers are the worst responders to a multifactorial treatment. In APOE-SIRT2 bigenic clusters, 33CC carriers respond better than 33TT and 34CT carriers, whereas 24CC and 44CC carriers behave as the worst responders. CYP2D6 extensive metabolizers (EM) are the best responders, poor metabolizers (PM) are the worst responders, and ultra-rapid metabolizers (UM) tend to be better responders that intermediate metabolizers (IM). In association with CYP2D6 genophenotypes, SIRT2-C/T-EMs are the best responders. Some Sirtuin modulators might be potential candidates for AD treatment.


Introduction
About 45-50 million people suffer from Alzheimer's disease (AD) (75 million in 2030; 145 million in 2050; 7.7 million new cases/year). The global economic cost of dementia is over US $604 billion, equivalent to 1% of the global gross domestic product. In terms of costs, AD accounts for $226 billion/year in the USA and €160 billion/year in Europe (>50% are costs of informal care, and 10-20% are costs of pharmacological treatment). It is estimated that in the USA alone, the direct cost of AD in people older than 65 years of age could be over $1.1 trillion in 2050 (from 2015 to 2050, the estimated medical costs would be about $20.8 trillion) [1]. Despite its relevance, paradoxically, no new drugs

Sirtuins
Sirtuins (Table 1) were discovered in yeast following the characterization of a yeast gene silencing modifier (Silent Information Modifier 2, SIR2) with a particular role in maintaining genomic stability. SIR2 homologs have been identified in different species. This category of protein deacetylases is important in the regulation of cell cycle progression, maintenance of genomic stability, and longevity. In yeast, SIR2 interacts with protein complexes that affect both replication and gene silencing. In metazoans, the largest SIR2 homolog, SIRT1, is implicated in epigenetic modifications, circadian signaling, DNA recombination and DNA repair. Mammalian SIRT1 participates in modulating DNA replication [20]. Sirtuins (Sirt1-Sirt7) are NAD + -dependent protein deacetylases/ADP ribosyltransferases, which play decisive roles in chromatin silencing, cell cycle regulation, cellular differentiation, cellular stress response, metabolism and aging [21]. Different sirtuins control similar cellular processes, suggesting a coordinated mode of action [22].

SIRT1
SIRT1 (10q21.3) is a NAD + -dependent histone deacetylase involved in transcription, DNA replication, and DNA repair, acting as a stress-response and chromatin-silencing factor [23]. SIRT1 interacts with SUV39H1 and NML in the energy-dependent nucleolar silencing complex (ENOSC), downregulating ribosomal RNA (rRNA) transcription during nutrient deprivation, reducing energy expenditure and improving cell survival [24]. Histones and proteins associated with the enhancement of mitochondrial function and antioxidant protection are currently SIRT1 substrates. Sir2 proteins (in yeast and mice) are NAD + -dependent histone deacetylases, with deacetylating activity on lysines 9 and 14 of histone H3 and lysine-16 of histone H4 [25]. SIRT1-related gene silencing results from deacetylation of histone tails, recruitment and deacetylation of histone H1, and spreading of hypomethylated H3-K79 activated by SIRT1-mediated heterochromatin formation [26].
Fluctuations in intracellular NAD + levels regulate SIRT1 activity. SIRT1 influences the nuclear organization of protein-bound NADH. Free and bound NADH are compartmentalized inside the nucleus, and its subnuclear distribution depends on SIRT1 [27]. In the liver, SIRT1 coordinates the circadian oscillation of clock-controlled genes, including genes that encode enzymes involved in metabolic pathways. G1/S progression is affected by the absence of SIRT1, as well as circadian gene expression, accompanied by lipid accumulation due to defective fatty acid beta-oxidation [28]. Several members of the Sir2 family can regulate life span in response to diet [29]. Hst2 is a Sir2 homolog that promoting the stability of repetitive ribosomal DNA is responsible for Sir2-independent life span extension. DNA stability is critical for yeast life span extension by calorie restriction. Sirtuins also affect the regulation of replicative aging by maintenance of intact telomeric chromatin. An age-related decrease of Sir2 protein is accompanied by an increase in histone H4 lysine-16 acetylation and loss of histones at subtelomeric regions in yeast cells, and this epigenetic change neuroprotection dependent on the neuronal population, and that SIRT1 and 3 decrease in parallel to AD progression, while expression of SIRT5 increases during the progression of AD [71]. Frontal cortex histone deacetylase (HDAC) and SIRT levels are altered during the clinical course of AD. HDAC1, HDAC3 and HDAC6 tend to increase in AD while SIRT1 decreases. HDAC1 levels negatively correlate with perceptual speed, while SIRT1 levels positively correlate with perceptual speed, episodic memory, global cognitive score, and Mini-Mental State Examination (MMSE). Furthermore, HDAC1 positively, while SIRT1 negatively correlate with cortical neurofibrillary tangle formation [72].
There is an age-related decline in the serum levels of SIRT1 in the population, and serum SIRT1 levels have been proposed as an early biomarker of AD [73]. A clear decline in SIRT1 levels is observed in patients with AD and mild cognitive impairment (MCI) as compared to healthy subjects [73,74]. SIRT1 is neuroprotective in AD. SIRT1 knockdown inhibits cell survival, proliferation, and functionality. These effects are associated with suppressed AKT activity, CREB activation and increased p53 expression [75]. Overexpression of SIRT1 preserves learning and memory in 10-month-old 3×Tg-AD mice and enhances cognitive performance in healthy non-transgenic mice. Novel pathways of SIRT1 neuroprotection may involve enhancement of cell proteostatic mechanisms and activation of neurotrophic factors [76].
SIRT1 is down-regulated in neurodegenerative disorders and shows a protective role in Parkinson's disease by reducing the formation of α-synuclein aggregates [77].
Many other factors cooperate with SIRT1 in the regulation of brain homeostasis. One example is FoxOs. The mammalian forkhead transcription factors of the O class (FoxOs) are present in brain centers associated with cognition (i.e., hippocampus, amygdala, nucleus accumbens). FoxOs may be required for memory formation and consolidation. FoxOs influence survival of CNS cells, pathways of apoptosis and autophagy, and stem cell proliferation and differentiation. FoxOs also interact with multiple cellular pathways (i.e., growth factors, Wnt signaling, Wnt1 inducible signaling pathway protein 1 (WISP1), silent mating type information regulation 2 homolog 1 (Saccharomyces cerevisiae), SIRT1) that retro-control FoxOs and determine the fate of cells involved in cognition and memory processes [80].
Humic acid (HA) is a potential pathogenic factor in vascular diseases and AD. HA contributes to Aβ-induced cytotoxicity mediated through the activation of endoplasmic reticulum stress by stimulating PERK and eIF2α phosphorylation together with mitochondrial dysfunction caused by down-regulation of the Sirt1/PGC1α pathway. Over-expression of Sirt1 reduces loss of cell viability by HA and Aβ [81].
The aspartyl protease β-site AβPP-cleaving enzyme 1 (BACE1) catalyzes the rate-limiting step in Aβ production in AD, and the adipocytokine leptin reduces Aβ production and decreases BACE1 activity. The transcription factor nuclear factor-kappa B (NF-κB) regulates BACE1 transcription and NF-κB activity is regulated by SIRT1. Leptin activates SIRT1. Leptin attenuates the activation and transcriptional activity of NF-κB by reducing the acetylation of the p65 subunit in a SIRT1-dependent manner [82]. SIRT1 activity in AD is reduced in parallel with the accumulation of hyperphosphorylated tau in the brain. The activation of SIRT1 with resveratrol reverses the ICV-STZ-induced decrease in SIRT1 activity and the increase in ERK1/2 and tau phosphorylation, as well as the cognitive impairment in experimental animals, where SIRT1 protects hippocampal neurons from tau hyperphosphorylation [83].
The BACE1 promoter contains multiple PPAR-RXR sites, and direct interactions among SIRT1-PPARγ-PGC-1 at these sites are enhanced with fasting. There is increased transcription of β-secretase/BACE1, the rate-limiting enzyme for Aβ generation, in eNOS-deficient mouse brains and after feeding a high-cholesterol diet. Modest fasting reduces BACE1 transcription in the brain, in parallel with elevated PGC-1 expression. The suppressive effect of PGC-1 is dependent on activated PPARγ, via SIRT1-mediated deacetylation in a ligand-independent manner [84].

SIRT2
SIRT2 is a highly conserved lysine deacetylase involved in aging, energy production, and lifespan extension. It has been interpreted that SIRT2 might promote neurodegeneration, because high levels of SIRT2 are present in AD, Parkinson's disease and other neurodegenerative disorders; however, in SH-SY5Y cells, elevated SIRT2 protects cells from rotenone or diquat-induced cell death and enzymatic inhibition of SIRT2 enhances cell death. SIRT2 protection is mediated, in part, through elevated SOD2 expression. SIRT2 reduces the formation of α-synuclein aggregates in Parkinson's disease. Some studies suggest that SIRT2 is necessary for protection against oxidative stress and that higher SIRT2 activity in neurodegeneration may be a compensatory mechanism to combat neuronal stress [89].
There is an association between human SIRT2 SNP rs10410544 C/T and AD susceptibility in the APOEε4-negative population [90,91]. When compared with the C allele, the T allele of rs10410544 shows a 1.709-fold risk for developing late-onset AD [90]. The SIRT2 SNP is associated with human AD risk in comparative models. The European population shows an increased risk of AD and association in the APOE ε4-negative population [91,92]. The SIRT2 rs10410544 SNP has also been associated with depression in European (Greek and Italian) AD cases in whom no association was found with AD [93]. In this study, the SIRT2-T/T genotype was associated with protection against depression.
α-Synuclein is acetylated on lysines 6 and 10 and these residues are deacetylated by sirtuin 2. Mutants blocking acetylation exacerbate α-synuclein toxicity in the substantia nigra. This suggests that sirtuin 2 might be a therapeutic option in some synucleinopathies [94].
Mitochondrial dysfunction is likely to be involved in AD pathogenesis. Mitochondria may lead to a dysfunction in autophagy/mitophagy due to the overactivation of SIRT2, which regulates microtubule network acetylation. Increased SIRT2 levels and decreased acetylation of Lys40 of tubulin are present in AD cells. SIRT2 loss of function achieved with AK1 (a specific SIRT2 inhibitor) or by SIRT2 knockout recovers microtubule stabilization and improves autophagy, favoring cell survival through the elimination of toxic Aβ oligomers [95]. SIRT2 inhibition in AD with small molecules (AGK-2, AK-7) reduces Aβ production and soluble β-AβPP, with an increase in soluble α-AβPP protein, and improves cognitive performance [96]. Whole brain radiotherapy (WBRT) produces unwanted sequelae, albeit via unknown mechanisms. In these cases, it appears that SIRT2 is linked to neurodegeneration. Canonical pathways for Huntington's, Parkinson's, and Alzheimer's diseases are acutely affected by brain radiation within 72 h of treatment. Loss of Sirt2 preferentially affects both Huntington's and Parkinson's pathways. Long-term radiation effects are found to be associated with altered levels of neurodegeneration-related proteins, identified as Mapt, Mog, Snap25, and Dnm1 [97]. Sirtuin inhibitors exert a neuroprotective effect in experimental models of Parkinson's disease [98,99] and Huntington's disease [100].

SIRT3
Mammalian SIRT3-5 are active in mitochondria where several clusters of protein substrates for SIRT3 have been identified. SIRT3 is the main mitochondrial Sirtuin involved in protecting stress-induced mitochondrial integrity and energy metabolism. SIRT3 is involved in the pathogenesis of some neurodegenerative diseases such as AD, amyotrophic lateral sclerosis, Parkinson's disease and Huntington's disease [101]. Mitochondrial dysfunction has been closely linked to the pathogenesis of AD [102]. Loss of SIRT3 accelerates neurodegeneration in brains challenged with excitotoxicity [103]. The increase in mitochondrial ROS increases Sirt3 expression in primary hippocampal cultures, where SIRT3 over-expression exerts a neuroprotective effect [102]. SIRT3 mRNA and protein levels are decreased in AD cerebral cortex and in the cortex of APP/PS1 double transgenic mice [104], and Ac-p53 K320 is increased in AD mitochondria. SIRT3 prevents p53-induced mitochondrial dysfunction and neuronal damage in a deacetylase activity-dependent manner. Mito-p53 reduces mitochondria DNA-encoded ND2 and ND4 gene expression with the consequent increase in reactive oxygen species (ROS) and reduced mitochondrial oxygen consumption. The expression of ND2 and ND4 is decreased in AD. SIRT3 restores ND2 and ND4 expression and improves mitochondrial oxygen consumption by repressing mito-p53 activity. SIRT3 dysfunction may lead to p53-mediated neuronal and mitochondrial damage in AD [105].
SIRT3 activates protein substrates involved in the production and detoxification of ROS (SOD2, catalase) and enzymes of the lipid beta-oxidation pathway. Microglia are the prime cellular source of ROS in the CNS. Sirtuin 3 is implicated in regulating cellular ROS levels. Sirt3 reduces cellular ROS by deacetylating forkhead box O 3a (Foxo3a), a transcription factor which transactivates antioxidant genes, catalase (CAT) and manganese superoxide dismutase (MnSOD). Sirt3 is localized in the ameboid microglial cells of the corpus callosum (CC) of the early postnatal rat brain and diminishes in the ramified microglial cells in the CC of the adult rat brain. Knockdown of SIRT3 in microglia leads to an increase in the cellular and mitochondrial ROS and a decrease in the expression of antioxidant MnSOD, reflecting a role for Sirt3 in ROS regulation in microglia. Conversely, SIRT3 overexpression increases CAT and MnSOD expression, and this effect is accompanied by an increase in the expression and nuclear translocation of Foxo3a, suggesting that Sirt3 regulates ROS by inducing the expression of antioxidants via activation of Foxo3a [106].
Aβ1-42 and SKI II induce free radical formation, disturb the balance between pro-and anti-apoptotic proteins and evoke cell death. Aβ1-42 increases the level of mitochondrial proteins (apoptosis-inducing factor AIF, Sirt3, Sirt4, Sirt5). p53 protein is essential at early stages of Aβ1-42 toxicity. After prolonged exposure to Aβ1-42, the activation of caspases, MEK/ERK, and alterations in mitochondrial permeability transition pores are additional factors contributing to cell death. Sphingosine-1-phosphate (S1P), Sirt activators and antioxidants (resveratrol, quercetin) enhance viability of cells under the toxic effects of Aβ1-42 [107].
Pituitary adenylate cyclase activating polypeptide (PACAP) is a neurotrophin with neuroprotective effects in AD. PACAP and SIRT3 expression is reduced in AD and in 3×TG mouse brains, inversely correlating with Aβ and tau protein levels. Treatment with PACAP protects neurons against Aβ toxicity. PACAP stimulates mitochondrial Sirt3 production. Knocking down Sirt3 abolishes the neuroprotective effects of PACAP, and this effect can be reversed by over-expressing Sirt3 [108].

SIRT6
SIRT6 is involved in telomere maintenance, DNA repair, genome integrity, energy metabolism, and inflammation, contributing to life span regulation. SIRT6 is deficient in AD patients [109]. SIRT6 promotes DNA repair, an activity that declines with age with the consequent accumulation of DNA damage. SIRT6 regulates Tau protein stability and phosphorylation through increased activation of the kinase GSK3α/β [110]. SIRT6 protein expression levels are reduced in AD brains. Aβ42 decreases SIRT6 expression, and Aβ42-induced DNA damage is prevented by the overexpression of SIRT6 in hippocampal neurons. A negative correlation between Aβ42-induced DNA damage and p53 levels is currently being seen, and upregulation of p53 with Nutlin-3 prevents SIRT6 reduction and DNA damage induced by Aβ42. p53-dependent SIRT6 expression protects cells from Aβ42-induced DNA damage [111].

APOE-Related Phenotypes
Multiple studies demonstrate the powerful influence of APOE genotypes on the AD phenotype. From these studies, several conclusions can be drawn: (i) the age-at-onset is 5-10 years earlier in 80% of APOE-4/4 carriers; (ii) serum ApoE levels are lowest in APOE-4/4 carriers, intermediate in APOE-3/3 and APOE-3/4, and highest in APOE-2/3 and APOE-2/4 carriers; (iii) cholesterol levels are higher in patients harboring the APOE-4/4 genotype than in carriers of other genotypes; (iv) HDL-cholesterol levels tend to be lower in APOE-3 homozygotes than in APOE-4 allele carriers; (v) LDL-cholesterol levels are higher in APOE-4/4 carriers with an APOE genotype-related pattern similar to total cholesterol; (vi) serum triglycerides tend to show the lowest levels in APOE-4/4 carriers (vii) nitric oxide levels tend to show reduced values in APOE-4/4 carriers (viii) serum and CSF Aβ levels show differential patterns in APOE-4/4 carriers as compared with carriers of other genotypes (APOE-3/3, APOE-3/4); (ix) blood histamine levels are dramatically reduced in APOE-4/4 carriers; (x) brain atrophy and AD neuropathology are markedly increased in APOE-4/4 > APOE-3/4 > APOE-3/3; (xi) brain mapping activity shows increased slow wave activity in APOE-4/4 from early stages of the disease; (xii) brain hemodynamics (reduced brain blood flow velocity, increased pulsatility and resistance indices) is significantly worse in APOE-4 carriers than in APOE-3 carriers; brain hypoperfusion and neocortical oxygenation as assessed with optical topography mapping is also more deficient in APOE-4 carriers; (xiii) lymphocyte apoptosis is enhanced in APOE-4 carriers; (xiv) cognitive deterioration is faster in APOE-4/4 patients than in carriers of other APOE genotypes; (xv) some metabolic and hematological deficiencies (iron, ferritin, folic acid, vitamin B12) accumulate more in APOE-4 carriers than in APOE-3 carriers; (xvi) some behavioral disturbances, alterations in circadian rhythm patterns, and mood disorders are slightly more frequent in APOE-4 carriers; (xvii) aortic and systemic atherosclerosis is also more frequent in APOE-4 carriers and the size of atheroma plaques in the aorta wall tends to be almost two-fold higher in APOE-4/4 carriers; (xviii) liver metabolism and transaminase activity also differ in APOE-4/4 with respect to other genotypes; (xix) hypertension and other cardiovascular risk factors also tend to accumulate in carriers of the APOE-4 allele; and (xx) APOE-4/4 carriers are the poorest responders to conventional drugs. All these phenotypic features clearly illustrate the biological disadvantage of APOE-4 homozygotes and the potential consequences that these patients may experience when they receive pharmacological treatment for AD and/or concomitant pathologies [2][3][4]6,[112][113][114][115][116][117][118][119][120][121][122][123][124].

Age and Sex
In our sample, females represent 57.55% and males 42.45% of the total. This female:male ratio is similar in all SIRT2 and APOE genotypes; however, the age at onset of the disease shows interesting differences, especially related to APOE genotypes. SIRT2 variants do not influence the age at onset in AD, except in the case of SIRT2-T/T males, who show a tendency to develop the disease at an earlier age than carriers of the other SIRT2 genotypes ( Figure 6). Among SIRT2-C/C carriers, females represent 57.56% of the sample (age: 71.55 ± 8.51 years, range: 51-73 years) and males 42.44% (age: 71.11 ± 9.43 years, range: 50-94 years). SIRT2-C/T females (59.13%; age: 71.63 ± 9.56 years, range: 50-94 years) and males (40.87%; age: 71.23 ± 9.44 years, range: 51-97 years) exhibit a similar age at onset; and SIR2-T/T males (40.08%; age: 69.84 ± 8.21 years, range: 52-84 years) tend to show an earlier age at onset than females (51.92%; age: 71.50 ± 9.61 years, range: 51-98 years) ( Figure 6).
In the case of APOE, there is a clear influence of the APOE-4 allele on the age at onset, with APOE-4 carriers (especially patients harboring the APOE-2/4 and APOE-4/4 genotypes) showing an earlier age at onset than their counterparts ( Figure 7).
APOE-SIRT2 bigenic haplotypes show significant differences in age at onset, with particular relevance in 23CC vs.

Age and Sex
In our sample, females represent 57.55% and males 42.45% of the total. This female:male ratio is similar in all SIRT2 and APOE genotypes; however, the age at onset of the disease shows interesting differences, especially related to APOE genotypes. SIRT2 variants do not influence the age at onset in AD, except in the case of SIRT2-T/T males, who show a tendency to develop the disease at an earlier age than carriers of the other SIRT2 genotypes ( Figure 6). Among SIRT2-C/C carriers, females represent 57.56% of the sample (age: 71.55 ± 8.51 years, range: 51-73 years) and males 42.44% (age: 71.11 ± 9.43 years, range: 50-94 years). SIRT2-C/T females (59.13%; age: 71.63 ± 9.56 years, range: 50-94 years) and males (40.87%; age: 71.23 ± 9.44 years, range: 51-97 years) exhibit a similar age at onset; and SIR2-T/T males (40.08%; age: 69.84 ± 8.21 years, range: 52-84 years) tend to show an earlier age at onset than females (51.92%; age: 71.50 ± 9.61 years, range: 51-98 years) ( Figure 6). In

Lipid Metabolism and BMI
Total cholesterol levels are significantly higher in SIRT2-C/T carriers (p = 0.05 vs SIRT2-C/C). Other parameters associated with lipid metabolism are similar among carriers of SIRT2 variants ( Table 2). Body Mass Index (BMI) tends to be higher in SIRT2-C/C (28.06 ± 4.31 kg/m 2 ) than in  Table  2). Table 2. SIRT2-related phenotypes in patients with Alzheimer's disease.

Lipid Metabolism and BMI
Total cholesterol levels are significantly higher in SIRT2-C/T carriers (p = 0.05 vs SIRT2-C/C). Other parameters associated with lipid metabolism are similar among carriers of SIRT2 variants ( Table 2) Table  2). Table 2. SIRT2-related phenotypes in patients with Alzheimer's disease.

Biochemical and Metabolic Parameters
Most biochemical parameters do not show any significant difference among SIRT2 variants, except cholesterol (p < 0.05 Table 2).

Pharmacogenetics and Pharmacoepigenetics
The genes involved in the pharmacogenomic response to drugs fall into five major categories: (i) genes associated with disease pathogenesis; (ii) genes associated with the mechanism of action of drugs (enzymes, receptors, transmitters, messengers, components of the epigenetic machinery); (iii) genes associated with drug metabolism (phase I-II reaction enzymes); (iv) genes associated with drug transporters; and (v) pleiotropic genes involved in multifaceted cascades and metabolic networks [2,19,[125][126][127]. All these genes are subjected to the epigenetic machinery for the specific regulation of their expression in physiological and pathological conditions [128][129][130]. Epigenetic regulation is responsible for the tissue-specific expression of genes involved in pharmacogenetic processes; consequently, epigenetics plays a key role in drug efficacy and safety and in the development of drug resistance. Epigenetic changes affect cytochrome P450 enzyme expression, major transporter function, and nuclear receptor interactions [129][130][131][132].
In the case of AD, most studies coincide in that the APOE and CYP2D6 genes are the most influential genes for the pharmacogenetic outcome, representing pathogenic (APOE) and metabolic (CYPD2) genes associated with the therapeutic response to conventional treatments [

Pharmacogenetics and Pharmacoepigenetics
The genes involved in the pharmacogenomic response to drugs fall into five major categories: (i) genes associated with disease pathogenesis; (ii) genes associated with the mechanism of action of drugs (enzymes, receptors, transmitters, messengers, components of the epigenetic machinery); (iii) genes associated with drug metabolism (phase I-II reaction enzymes); (iv) genes associated with drug transporters; and (v) pleiotropic genes involved in multifaceted cascades and metabolic networks [2,19,[125][126][127]. All these genes are subjected to the epigenetic machinery for the specific regulation of their expression in physiological and pathological conditions [128][129][130]. Epigenetic regulation is responsible for the tissue-specific expression of genes involved in pharmacogenetic processes; consequently, epigenetics plays a key role in drug efficacy and safety and in the development of drug resistance. Epigenetic changes affect cytochrome P450 enzyme expression, major transporter function, and nuclear receptor interactions [129][130][131][132].
In the case of AD, most studies coincide in that the APOE and CYP2D6 genes are the most influential genes for the pharmacogenetic outcome, representing pathogenic (APOE) and metabolic (CYPD2) genes associated with the therapeutic response to conventional treatments [2][3][4]6,7,19,125,127,133].

APOE-and TOMM40-Related Therapeutic Response to Multifactorial Treatments
Different studies document the impact of APOE genotypes on AD therapeutics [2][3][4][5]7,19,121,122,[124][125][126][127]133]. We have performed prospective and retrospective studies in which it was clearly demonstrated that APOE-4 carriers are the worst responders to conventional treatments [2,3,6,118,120,127,134]. The TOMM40 locus is located near to and in linkage disequilibrium with the APOE locus on 19q13.2. The TOMM40 gene encodes an outer mitochondrial membrane translocase involved in the transport of amyloid-β and other proteins into mitochondria, and a poly T repeat in an intronic polymorphism (rs10524523) (intron 6) in the TOMM40 gene has been implicated in AD [135][136][137][138][139][140]. Different variants in the APOE-TOMM40 region influence disease risk, age at onset of AD [135][136][137][138][139][140][141], cognitive aging [142] and pathological cognitive decline [143]. The intronic poly T (rs10524523) affects expression of the APOE and TOMM40 genes in the brains of patients with late-onset AD (LOAD) [144]. The expression of both genes is increased with disease. The 523 locus may contribute to LOAD susceptibility by modulating the expression of TOMM40 and/or APOE transcription [144]. The TOMM40 gene rs10524523 ("523") variable-length poly T repeat polymorphism is associated to a certain extent with similar AD phenotypes as those reported for APOE, such as brain white matter changes [145,146] or different biomarkers [147][148][149].
With this therapeutic strategy, AD patients respond with a significant cognitive improvement during the first 9 months (Figure 13), and a progressive decline is observed thereafter, as with many other conventional treatments. This indicates that current treatments only provide a transient benefit, but they do not protect against progressive neuronal death once the neurodegenerative process is activated decades before the onset of the disease. This response is highly influenced by the baseline MMSE Score at the time of diagnosis and the starting point of treatment, and also by the genetic background of the patients, with APOE-3/3 carriers behaving as the best responders and APOE-4 carriers being the worst responders ( Figure 14).
Patients with different SIRT2 genotypes respond similarly during the first 3 months of treatment, with a significant improvement, and only SIRT2-C/T carriers maintain cognitive improvement over baseline levels for one year. Globally, SIRT2-C/T carriers are the best responders, SIRT2-T/T carriers show an intermediate pattern, and SIRT2-C/C carriers are the worst responders ( Figure 15). other conventional treatments. This indicates that current treatments only provide a transient benefit, but they do not protect against progressive neuronal death once the neurodegenerative process is activated decades before the onset of the disease. This response is highly influenced by the baseline MMSE Score at the time of diagnosis and the starting point of treatment, and also by the genetic background of the patients, with APOE-3/3 carriers behaving as the best responders and APOE-4 carriers being the worst responders ( Figure 14). Patients with different SIRT2 genotypes respond similarly during the first 3 months of treatment, with a significant improvement, and only SIRT2-C/T carriers maintain cognitive improvement over baseline levels for one year. Globally, SIRT2-C/T carriers are the best responders, SIRT2-T/T carriers show an intermediate pattern, and SIRT2-C/C carriers are the worst responders ( Figure 15).   other conventional treatments. This indicates that current treatments only provide a transient benefit, but they do not protect against progressive neuronal death once the neurodegenerative process is activated decades before the onset of the disease. This response is highly influenced by the baseline MMSE Score at the time of diagnosis and the starting point of treatment, and also by the genetic background of the patients, with APOE-3/3 carriers behaving as the best responders and APOE-4 carriers being the worst responders ( Figure 14). Patients with different SIRT2 genotypes respond similarly during the first 3 months of treatment, with a significant improvement, and only SIRT2-C/T carriers maintain cognitive improvement over baseline levels for one year. Globally, SIRT2-C/T carriers are the best responders, SIRT2-T/T carriers show an intermediate pattern, and SIRT2-C/C carriers are the worst responders ( Figure 15).

APOE-SIRT2 Bigenic Genotype-Related Cognitive Response to Treatment
The study of APOE-SIRT2 bigenic clusters revealed important differences in cognitive performance at diagnosis that influence the therapeutic response to multifactorial treatments. Significant differences in cognition at baseline levels were found between the following genotypes:  (Figure 16). This heterogeneity at baseline levels is determinant, together with the genomic background of each patient, for the pharmacogenetic outcome. 24CT carriers show improvement only for the first month (p < 0.05); 33CC carriers at 6-9 months (p < 0.05); and 33TT and 34CT at 3 months (p < 0.05). According to these bigenic clusters, 33CC carriers are better responders than 33TT and 34CT carriers, and 24CC and 44CC are the worst responders ( Figure 16).

APOE-SIRT2 Bigenic Genotype-Related Cognitive Response to Treatment
The study of APOE-SIRT2 bigenic clusters revealed important differences in cognitive performance at diagnosis that influence the therapeutic response to multifactorial treatments. Significant differences in cognition at baseline levels were found between the following genotypes:  (Figure 16). This heterogeneity at baseline levels is determinant, together with the genomic background of each patient, for the pharmacogenetic outcome. 24CT carriers show improvement only for the first month (p < 0.05); 33CC carriers at 6-9 months (p < 0.05); and 33TT and 34CT at 3 months (p < 0.05). According to these bigenic clusters, 33CC carriers are better responders than 33TT and 34CT carriers, and 24CC and 44CC are the worst responders ( Figure 16).

APOE-SIRT2 Bigenic Genotype-Related Cognitive Response to Treatment
The study of APOE-SIRT2 bigenic clusters revealed important differences in cognitive performance at diagnosis that influence the therapeutic response to multifactorial treatments. Significant differences in cognition at baseline levels were found between the following genotypes:  (Figure 16). This heterogeneity at baseline levels is determinant, together with the genomic background of each patient, for the pharmacogenetic outcome. 24CT carriers show improvement only for the first month (p < 0.05); 33CC carriers at 6-9 months (p < 0.05); and 33TT and 34CT at 3 months (p < 0.05). According to these bigenic clusters, 33CC carriers are better responders than 33TT and 34CT carriers, and 24CC and 44CC are the worst responders ( Figure 16).

CYP2D6-Related Therapeutic Response to Multifactorial Treatments
CYP2D6 genophenotypes are highly influential in the response to cholinesterase inhibitors and other medications in AD [2,7,19,118,127,133,150]. In our sample, the distribution and frequency of CYP2D6 genophenotypes was as follows: Extensive metabolizers (EM), 59.46%; intermediate metabolizers (IM), 20.06%; poor metabolizers (PM), 5.36%; and ultra-rapid metabolizers (UM), 6.12% (Figure 17). Significant differences were found in cognitive performance at diagnosis between EMs and PMs (p = 0.01), IMs vs. PMs (p = 0.02), and PMs vs. UMs (p = 0.004). The lowest MMSE Scores were detected in PMs ( Figure 17). There is an accumulation of APOE-3/4 and APOE-4/4 carriers in PMs, and APOE-4/4 carriers are over-represented in UMs (Figure 18). In APOE-CYP2D6 bigenic genophenotypes, over 50% of the cases among APOE-4/4 carriers are IMs, PMs and UMs, and 100% of APOE-2/4 cases are EMs (Figure 19). The concentration of IMs, PMs and UMs in APOE-4/4 carriers may justify, in part, the poor cognitive performance of APOE-4/4 carriers in response to conventional treatments. According to the metabolizing condition of the patients, EMs are the best responders, PMs are the worst responders, and UMs tend to be better responders than IMs (Figure 20). CYP2D6 genophenotypes are highly influential in the response to cholinesterase inhibitors and other medications in AD [2,7,19,118,127,133,150]. In our sample, the distribution and frequency of CYP2D6 genophenotypes was as follows: Extensive metabolizers (EM), 59.46%; intermediate metabolizers (IM), 20.06%; poor metabolizers (PM), 5.36%; and ultra-rapid metabolizers (UM), 6.12% ( Figure 17). Significant differences were found in cognitive performance at diagnosis between EMs and PMs (p = 0.01), IMs vs PMs (p = 0.02), and PMs vs UMs (p = 0.004). The lowest MMSE Scores were detected in PMs ( Figure 17). There is an accumulation of APOE-3/4 and APOE-4/4 carriers in PMs, and APOE-4/4 carriers are over-represented in UMs (Figure 18). In APOE-CYP2D6 bigenic genophenotypes, over 50% of the cases among APOE-4/4 carriers are IMs, PMs and UMs, and 100% of APOE-2/4 cases are EMs (Figure 19). The concentration of IMs, PMs and UMs in APOE-4/4 carriers may justify, in part, the poor cognitive performance of APOE-4/4 carriers in response to conventional treatments. According to the metabolizing condition of the patients, EMs are the best responders, PMs are the worst responders, and UMs tend to be better responders than IMs (Figure 20).   CYP2D6 genophenotypes are highly influential in the response to cholinesterase inhibitors and other medications in AD [2,7,19,118,127,133,150]. In our sample, the distribution and frequency of CYP2D6 genophenotypes was as follows: Extensive metabolizers (EM), 59.46%; intermediate metabolizers (IM), 20.06%; poor metabolizers (PM), 5.36%; and ultra-rapid metabolizers (UM), 6.12% ( Figure 17). Significant differences were found in cognitive performance at diagnosis between EMs and PMs (p = 0.01), IMs vs PMs (p = 0.02), and PMs vs UMs (p = 0.004). The lowest MMSE Scores were detected in PMs ( Figure 17). There is an accumulation of APOE-3/4 and APOE-4/4 carriers in PMs, and APOE-4/4 carriers are over-represented in UMs ( Figure 18). In APOE-CYP2D6 bigenic genophenotypes, over 50% of the cases among APOE-4/4 carriers are IMs, PMs and UMs, and 100% of APOE-2/4 cases are EMs (Figure 19). The concentration of IMs, PMs and UMs in APOE-4/4 carriers may justify, in part, the poor cognitive performance of APOE-4/4 carriers in response to conventional treatments. According to the metabolizing condition of the patients, EMs are the best responders, PMs are the worst responders, and UMs tend to be better responders than IMs (Figure 20).

Epigenetic Drugs
Epigenetic drugs are chemicals or bioproducts that target regulatory components of the epigenetic machinery [18]. Epigenetic drugs reverse epigenetic changes in gene expression and might open future avenues for the treatment of major problems of health [11,19,138,[150][151][152][153][154][155][156][157][158][159][160][161]. Within this growing category of drugs, several inhibitors of histone deacetylation and DNA methylation have been approved by the US FDA for hematological malignancies, and some epigenetic drugs are being evaluated in clinical trials for the treatment of several diseases [18].

Epigenetic Drugs
Epigenetic drugs are chemicals or bioproducts that target regulatory components of the epigenetic machinery [18]. Epigenetic drugs reverse epigenetic changes in gene expression and might open future avenues for the treatment of major problems of health [11,19,138,[150][151][152][153][154][155][156][157][158][159][160][161]. Within this growing category of drugs, several inhibitors of histone deacetylation and DNA methylation have been approved by the US FDA for hematological malignancies, and some epigenetic drugs are being evaluated in clinical trials for the treatment of several diseases [18].

Epigenetic Drugs
Epigenetic drugs are chemicals or bioproducts that target regulatory components of the epigenetic machinery [18]. Epigenetic drugs reverse epigenetic changes in gene expression and might open future avenues for the treatment of major problems of health [11,19,138,[150][151][152][153][154][155][156][157][158][159][160][161]. Within this growing category of drugs, several inhibitors of histone deacetylation and DNA methylation have been approved by the US FDA for hematological malignancies, and some epigenetic drugs are being evaluated in clinical trials for the treatment of several diseases [18].
Several epigenetic drugs have been unsuccessfully tested in AD models, and none have passed preclinical or early phase clinical trials [10,11]. Hypermethylation of the SIRT1 gene and demethylation of the β-amyloid precursor protein (APP) gene are common findings in AD. However, the expression of SIRT1 is decreased, while that of APP is increased in AD. The treatment of human neuroblastoma SK-N-SH cells with the epigenetic drugs, the DNA methylation inhibitor 5-aza-2 -deoxycytidine (DAC) and the histone deacetylase inhibitor trichostatin A (TSA), in the presence of Aβ25-35, showed that DAC and TSA have different effects on the expression of SIRT1 and APP under amyloid toxicity. The MAPT (Microtubule-associated protein τ), PSEN1 (presenilin 1), PSEN2 (presenilin 2), and APOE genes are up-regulated by Aβ25-35, but they do not respond to DAC and/or TSA [156].   Table 3. Cont.

Categories Drugs
Other compounds

Sirtuin Modulators
The mammalian sirtuins (SIRT1-7) are NAD + -dependent lysine deacylases with central effects in cell survival, inflammation, energy metabolism, cancer, aging, cardiovascular disorders and neurodegeneration. Consequently, members of this family of enzymes represent promising pharmaceutical targets for the treatment of age-related neurodegenerative disorders and cancer. A series of sirtuin modulators have been discovered and characterized during the past decades [18,157]. SIRT1-activating compounds of different pharmacological categories (Tables 3 and 4), provide health benefits in animal models. Compared with natural products, the synthetic sirtuin modulators exhibit greater potency, solubility, and target selectivity, together with higher toxicity as well [18,157]. Despite promising considerations on sirtuins as potential therapeutic targets for AD [158,159], no breakthroughs have been reported and few epigenetic drugs are in clinical trials for AD or other neurodegenerative disorders [10,11,18]. In the following paragraphs, some examples of sirtuin modulators (activators and inhibitors) are shown (Tables 3 and 4). neurodegenerative disorders [10,11,18]. In the following paragraphs, some examples of sirtuin modulators (activators and inhibitors) are shown (Tables 3-4).     About 6-10% of AD patients are deficient in folate and over 40% of the cases suffer cardiovascular disorders or diseases which represent vascular risk factors. Folic acid is cardio-and neuro-protective in early-stage AD in transgenic mice. Folic acid treatment restores SIRT1 expression, which is suppressed in 3×Tg mice, through enhanced AMPK expression [160]. About 6-10% of AD patients are deficient in folate and over 40% of the cases suffer cardiovascular disorders or diseases which represent vascular risk factors. Folic acid is cardio-and neuro-protective in early-stage AD in transgenic mice. Folic acid treatment restores SIRT1 expression, which is suppressed in 3×Tg mice, through enhanced AMPK expression [160]. About 6-10% of AD patients are deficient in folate and over 40% of the cases suffer cardiovascular disorders or diseases which represent vascular risk factors. Folic acid is cardio-and neuro-protective in early-stage AD in transgenic mice. Folic acid treatment restores SIRT1 expression, which is suppressed in 3×Tg mice, through enhanced AMPK expression [160].  About 6-10% of AD patients are deficient in folate and over 40% of the cases suffer cardiovascular disorders or diseases which represent vascular risk factors. Folic acid is cardio-and neuro-protective in early-stage AD in transgenic mice. Folic acid treatment restores SIRT1 expression, which is suppressed in 3×Tg mice, through enhanced AMPK expression [160]. About 6-10% of AD patients are deficient in folate and over 40% of the cases suffer cardiovascular disorders or diseases which represent vascular risk factors. Folic acid is cardio-and neuro-protective in early-stage AD in transgenic mice. Folic acid treatment restores SIRT1 expression, which is suppressed in 3×Tg mice, through enhanced AMPK expression [160].
Food-derived polyphenols protect against age-related diseases, such as atherosclerosis, cardiovascular disease, cancer, arthritis, cataracts, osteoporosis, diabetes, hypertension and AD. Resveratrol and pterostilbene are polyphenols with anti-aging effects on oxidative damage, inflammation, telomere attrition and cell senescence [162].
Resveratrol is a potent activator of SIRT1, mimicking caloric restriction to prevent aging-related disorders. A randomized, double-blind, placebo-controlled, phase II trial of resveratrol in mild-to-moderate AD cases revealed that resveratrol crosses the blood-brain barrier and modulates the CNS immune response [163].
Resveratrol delays axonal degeneration. The effect of resveratrol on Wallerian degeneration is lost when SIRT1 is inhibited. Knocking out Deleted in Breast Cancer-1 (DBC1), an endogenous SIRT1 inhibitor, restores the neuroprotective effect of resveratrol. It appears that resveratrol protects against Wallerian degeneration by promoting the dissociation of SIRT1 and DBC1 in cultured ganglia [168].

Pterostilbene
Pterostilbene, a resveratrol derivative, shows neuroprotective effects in age-related disorders and AD models. Pterostilbene diet affects markers of cellular stress, inflammation, and AD pathology, with upregulation of peroxisome proliferator-activated receptor (PPAR) alpha expression and no effect on SIRT1 levels [169].

Curcumin
Curcumin, extracted from the yellow pigments of turmeric (Curcuma longa), shows antioxidant, anti-apoptotic and neuroprotective effects. Curcumin prevents Aβ25-35-induced cell toxicity in cultured cortical neurons, improves mitochondrial membrane potential (∆Ψm), decreases ROS generation and inhibits apoptotic cell death. Curcumin also activates the expression of SIRT1 with a subsequent decrease in the expression of Bax in the presence of Aβ25-35. The protective effects of curcumin can be blocked by SIRT1 siRNA [170].
Curcumin exerts a neuroprotective effect against the toxicity induced by acrolein. Curcumin restores the expression of γ-glutamylcysteine synthetase, reactive oxygen species, and reactive nitrogen species levels and has no effect on glutathione (GSH) and protein carbonyls. Acrolein activates Nrf2, NF-κB, and Sirt1, and these in vitro effects can be modulated by curcumin. Acrolein also induces a decrease in pAkt, which is counteracted by curcumin [171].

Nicotinamide Riboside
Defective cellular bioenergetics and DNA repair contribute to AD pathogenesis. Cellular NAD + depletion upstream of neuroinflammation, pTau, DNA damage, synaptic dysfunction, and neuronal degeneration may be pathogenic in AD. Treatment with nicotinamide riboside (NR) lessens pTau pathology in transgenic models with no effect on Aβ accumulation. NR-treated 3× TgAD/Polβ+/mice exhibit reduced DNA damage, neuroinflammation, and apoptosis of hippocampal neurons and increased activity of brain SIRT3 [172].

Oleuropein Aglycone
Oleuropein aglycone (OLE) is a polyphenol present in extra virgin olive oil. OLE is able to induce autophagy, a process by which aggregated proteins and damaged organelles are eliminated through lysosomal digestion. Autophagy is defective in AD and OLE is able to decrease aggregated proteins and improves cognition by modulating several pathways including the AMPK/mTOR axis and the activation of autophagy gene expression mediated by sirtuins and histone acetylation or EB transcription factor [173].
Poly(ADP-ribose) polymerase-1 (PARP1) activation contributes to Aβ-induced neurotoxic events in AD. OLE treatment in TgCRND8 mice restores PARP1 activation and the levels of its product, PAR, to control values. PARP1 activation and PAR formation upon exposure to N-methyl-N -nitro-N-nitrosoguanidine (MNNG) are abolished by pretreatment with either OLE or PARP inhibitors. OLE-induced reduction of PARP1 activation is paralleled by overexpression of SIRT1, and by a decrease in NF-κB and the pro-apoptotic marker p53 [174].

Honokiol
Honokiol (poly phenolic lignan from Magnolia grandiflora) is a SIRT3 activator with antioxidant activity. Honokiol enhances SIRT3 expression, reduces reactive oxygen species generation and lipid peroxidation, enhances antioxidant activity and mitochondrial function reducing Aβ and sAPPβ levels in transgenic models. Honokiol increases the expression of AMPK, CREB, and PGC-1α, inhibiting β-secretase activity and consequently leading to reduced Aβ levels [175].

Flavonoids
Flavonoids are nutraceuticals with potential beneficial effects in AD, aging and age-related inflammatory disorders. Flavonoids can reduce extracellular amyloid deposits and neurofibrillary tangles by mediating amyloid precursor protein (APP) processing, Aβ accumulation and tau pathology. The antioxidant and anti-inflammatory effects of flavonoids as well as their modulatory effects on sirtuins and telomeres also contribute to ameliorating neurodegeneration. Some flavonoids can inhibit poly (ADP-ribose) polymerases (PARPs) and cyclic ADP-ribose (cADP) synthases (CD38 and CD157), elevate intracellular nicotinamide adenine dinucleotide (NAD + ) levels and activate NAD + -dependent sirtuin-mediated signaling pathways [176].

Tripeptides
SIRT1 attenuates the amyloidogenic processing of APP in AD pathology. A CWR tripeptide has been characterized as a potential SIRT1 activator with capacity for enhancing SIRT1 activity. This tripeptide decreases the acetylation of p53 in IMR32 neuroblastoma cells and protects cells against Aβ toxicity [178].

Ampelopsin (Dihydromyricetin)
Ampelopsin is a natural flavonoid from the Chinese herb Ampelopsis grossedentata, with pleiotropic effects including anti-inflammatory, anti-oxidative and anti-cancer functions. Studies in a rat model with D-gal-induced brain aging revealed that expression of miR-34a can be suppressed with ampelopsin. The up-regulation of miR-34a is associated with aging-related diseases. Ampelopsin activates autophagy through up-regulation of SIRT1 and down-regulation of mTOR signaling pathways in connection with down-regulation of miR-34a [179].

Cystatin C
Cystatin C (CysC) is a natural cysteine protease inhibitor that reduces Aβ40 secretion in human brain microvascular endothelial cells. The CysC-induced Aβ40 reduction is caused by degradation of β-secretase BACE1 through the ubiquitin/proteasome pathway. The α-secretase ADAM10, which is transcriptionally upregulated in response to CysC, is required for the CysC-induced sAPPα secretion. Knockdown of SIRT1 abolishes CysC-triggered ADAM10 upregulation and sAPPα production. CysC can direct amyloidogenic APP processing to the non-amyloidogenic pathway, mediated by proteasomal degradation of BACE1 and SIRT1-mediated ADAM10 upregulation [180].

Salidroside
Salidroside (Rhodioloside) is a glucoside of tyrosol present in the vegetal Rhodiola rosea. Together with rosavin, salidroside might be responsible for the antidepressant and anxiolytic effects of this plant. Memory performance and neuroinflammation in D-galactose (D-gal)-induced sub-acute aging models show deterioration associated with activated nuclear factor kappa B (NF-κB) p65/RelA and down-regulation of SIRT1 expression in the hippocampus. Treatment with Salidroside ameliorates D-gal-induced memory deficits and inflammatory mediators including TNF-α and IL-1β. Salidroside also inhibits the NF-κB signaling pathway via up-regulation of SIRT1 [184].

CDP-Choline
CDP-Choline (Citicoline) is a choline donor and an intermediate of DNA metabolism. This old compound, developed in the 1970s, has been used as a neuroprotectant in some European countries and Japan for decades. Several studies demonstrated its utility in AD incorporated to multifactorial interventions [2,3,120,127,134,150]. Citicoline potentiates neuroplasticity and is a natural precursor of phospholipid synthesis. In addition to its conventional properties, citicoline increases SIRT1 expression [185].

Melatonin
Melatonin is a pleiotropic endogenous substance with antioxidant, neuroprotectant, antiexcitotoxic and immunomodulatory effects. Melatonin and its kynuramine metabolites are important for the attenuation of inflammatory responses and progression of neuroinflammation. Sirtuins influence circadian oscillators which are under the control of melatonin [188].

S-Linolenoyl Glutathione
Glutathione (GSH) is the most abundant endogenous free radical scavenger in mammalian cells. A series of novel S-acyl-GSH derivatives are capable of preventing amyloid oxidative stress and cholinergic dysfunction in AD models. The longevity of the wild-type N2 Caenorhabditis elegans strain is enhanced by dietary supplementation with linolenoyl-SG (lin-SG) thioester with respect to the ethyl ester of GSH, linolenic acid, or vitamin E. Life-span extension is mediated by the upregulation of Sir-2.1, a NAD-dependent histone deacetylase ortholog of mammalian SIRT1. Lin-SG-mediated overexpression of Sir-2.1 appears to be related to the Daf-16 (FoxO) pathway [189].

Taurine
Taurine is a naturally occurring β-amino acid in the brain with neuroprotective effects. Taurine attenuates Aβ1-42-induced neuronal death and intracellular Ca 2+ and ROS generation. SIRT1 expression is recovered by taurine in Aβ1-42-treated SK-N-SH cells, suggesting that taurine prevents Aβ1-42-induced mitochondrial dysfunction by activation of SIRT1 [190].

Sulfobenzoic Acid Derivative AK1
Sirtuin 2 inhibition may be a neuroprotective strategy in some neurodegenerative disorders. The SIRT2 inhibitor AK1 provides some neuroprotection in the hippocampus of rTg4510 mice (a model of the tauopathic frontotemporal dementia, characterized by the formation of tau-containing neurofibrillary aggregates and neuronal loss) [192].

Phytic Acid
Phytic acid (inositol hexakisphosphate) is a phytochemical found in food grains and is a key signaling molecule in mammalian cells. Phytic acid provides protection against amyloid precursor protein-C-terminal fragment-induced cytotoxicity by attenuating levels of increased intracellular calcium, hydrogen peroxide, superoxide, and Aβ oligomers, and moderately upregulates the expression of autophagy proteins. Phytic acid increases brain levels of cytochrome oxidase and decreases lipid peroxidation. In Tg2576 mice, phytic acid exerts a modest effect on the expression of AβPP trafficking-associated protein AP180, autophagy-associated proteins (beclin-1, LC3B), sirtuin 1, the ratio of phosphorylated AMP-activated protein kinase (PAMPK) to AMPK, soluble Aβ1-40, and insoluble Aβ1-42 [193].

Gamma Secretase Inhibitors
Gamma-secretase is an intramembrane-cleaving protease responsible for the abnormal proteolytic cleavage of APP and the production of neurotoxic Aβ peptides implicated in the pathogenesis of AD [2]. Most gamma-secretase inhibitors have failed in AD due to toxicity and/or inefficacy. 2-Hydroxy naphthyl derivatives are a subclass of NAD + analog inhibitors of sirtuin 2, with gamma-secretase inhibitory activity. 2-Hydroxy-1-naphthaldehyde is the minimal pharmacophore for gamma-secretase inhibition. A GXG signature nucleotide-binding site (NBS) shared by the gamma-secretase subunit presenilin-1 C-terminal fragment (PS1-CTF), SIRT2, and Janus kinase 3 (JAK3) is the target protein determinant of inhibition [194].

Donepezil
Donepezil is the most prescribed drug worldwide for the treatment of AD [195]. In addition to its anticholinesterase activity, donepezil increases SIRT1 activity and inhibits the generation of reactive oxygen species [196].
A therapeutic intervention with a multifactorial treatment in AD demonstrates some benefit in terms of cognitive improvement for the first 3-9 months of treatment, depending upon the pharmacogenetic profile of each patient (Figures 13-16). SIRT2-C/T carriers are the best responders, SIRT2-T/T carriers show an intermediate response, and SIRT2-C/C carriers are the worst responders to treatment ( Figure 15). In APOE-SIRT2 bigenic clusters, 33CC carriers respond better than 33TT and 34CT carriers, whereas 24CC and 44CC carriers are poor responders (Figure 16). SIRT2 also interacts with CYP2D6 and this interaction contributes to modulate the pharmacogenetic response to conventional treatments. The frequencies of CYP2D6 genophenotypes in AD are as follows: Extensive metabolizers (EM), 59.46%; intermediate metabolizers (IM), 20.06%; poor metabolizers (PM), 5.36%; and ultra-rapid metabolizers (UM), 6.12% (Figure 17). CYP2D6-EMs are the best responders, PMs are the worst responders, and UMs tend to be better responders than IMs ( Figure 20). There is an accumulation of APOE-3/4 and APOE-4/4 genotypes in CYP2D6-PMs and UMs ( Figure 18). In association with CYP2D6 genophenotypes, SIRT2-C/T-EMs are the best responders ( Figure 22). A major conclusion from the results obtained in the present study would be that the influence of SIRT2 in AD pathogenesis and in AD-related genophenotypes is very mild; however, the interaction of SIRT2 variants with other genes (i.e., APOE, CYP2D6) may be relevant, affecting age at onset, clinical course, rate of cognitive decline, and pharmacoepigenetic outcome. In this context, if the direct or indirect role of sirtuins in AD pathogenesis can be confirmed and their neuroprotective effects clearly demonstrated, it would be likely that some Sirtuin modulators might become potential candidates for AD treatment in the future.
Author Contributions: R.C. designed the study, performed the clinical studies, analyzed the data and wrote the paper; J.C.C. performed genetic analysis of APOE and CYP2D6 variants; N.C. and P.C. contributed in clinical trials, database searching, data collection and references; A.G.K., A.V.V. and D.G. performed genetic analysis of SIRT2 variants and contributed to the written manuscript; and L.C. was responsible for biochemical studies in clinical trials.
Funding: This research received no external funding.