**6. Epigenetic alterations in AD**

While APP mutations account for a small part of EOAD, mutations on PSEN1/PSEN2 have been identified as critical genes in EOAD [63], which are shown to increase Aβ42/Aβ40 ratios and promote Aβ42 accumulation [64, 65]. After proteolytic cleavage of full-length presenilin to generate N-terminal and C-terminal fragments and assembly into γ-secretase complex, γsecretase is transported to cell surface where it acts on APP processing and cleavage. Both PSEN1 and 2 mutations increase formation of Aβ species and deposition of amyloid plaques [63, 66, 67]. PSEN1 mutations by altering APP γ-secretase cleavage site promote Aβ42 genera‐ tion [68]. PSEN2 mutations lead to AD with a slower progression than PSEN1 mutations [67]. Besides their role in APP processing, presenilins are involved in many other cellular functions such as Notch signaling and differentiation [69], calcium homeostasis [70], gene expression via interaction with transcriptional coactivators like CREB-binding protein [71]. It was reported that PSEN1 exhibited neuroprotective functions through ephrin-B [72], and that defects in these functions with genetic modifications are implicated in AD pathogenesis. In AD trans‐ genic animal models, APP mutations or in combination with presenilin 1 mutations induced Aβ plaque formation similarly to what were seen in AD human brains [73]. Interestingly, comparatively to sporadic AD cases, patients with PSEN1 mutations had more senile plaques and NFTs developed in their brains, suggesting that PSEN1 may enhance tau deposition as

Among all identified genetic factors involved in the disease, APOE gene has been extensively studied over the past decade or so. APOE is a major risk gene associated with AD, is located on chromosome 19 [75], and encodes for apolipoprotein E, a 34-kDa lipid binding protein involved in triglycerides and cholesterol transport [76–81]. Three ApoE isoforms which differ by single amino acid substitutions have been found in humans: ApoE2, ApoE3, and ApoE4. ApoE4 carriers present with high levels of cholesterol and LDL in the plasma, which predis‐ pose the carriers to cardiovascular disease and AD [82]. ApoE4 carriers contribute to about 50% of AD cases. While ApoE2 decreases the risk and delays the onset of AD [53], ApoE4 multiplies the risk of EOAD and late-onset AD (LOAD) by approximately 3-fold for hetero‐ zygous and 10-fold for homozygous carriers [62]. ApoE4 is shown to increase amyloid plaque

formation by altering Aβ clearance [83] and promoting fibrillary aggregations [84].

New methods of genetic mapping of single-nucleotide polymorphisms (SNP) provided new information regarding genes involved in increasing or decreasing the risk of AD. Besides wellestablished AD risk genes described above, nine additional AD-related genes including complement receptor 1 (CR1), bridging integrator 1 (BIN1), clusterin (CLU), phosphatidyli‐ nositol-binding clathrin assembly protein (PICALM), MS4A4/MS4A6E, CD2AP, CD33, EPHA1, and ATP-binding cassette transporter (ABCA7) have been unveiled by genome-wide

The CR1 gene polymorphism rs6656401 was the first found to be associated with AD in European population [85]. A study conducted on two Canadian cohorts further showed that

well [74].

228 Update on Dementia

**5. Genetic risk factors in sporadic AD**

association and sequencing studies [85, 86].

Epigenetic variabilities such as histone modifications, DNA methylation/demethylation, and microRNA regulation have been reported not only in the aging processes of different tissues but also in neurodegenerative diseases such as AD. These epigenetic changes might play a pathogenic role in disease mechanism [102–108].

#### **6.1. Histone modifications**

A few recent studies reported histone modifications in AD [105, 109, 110]. For example, histone acetylation such as H4 acetylation was decreased in APP/PS1 transgenic mice, which might be involved in cognitive deficits [109, 111, 112]. Another study reported increased H3 and H4 acetylation in the 3xTg-AD mouse model compared to wild-type mice [113]. Levels of phosphorylated histone proteins such as HDAC6 and H3S10 were found to be increased in AD brain regions and neurons [114, 115]. Finally, levels of methylation, acetylation, and phosphorylation of histone H3 were showed to be elevated in AD individual cortex [116].

#### **6.2. DNA methylation/demethylation**

Genes containing CpG islands are methylated in their promoter regions. Differences in methylation have been reported in APP, BACE, PS1, and APOE genes [105, 107]. For example, one study showed that methylation in APP promoter region was decreased in the brains of old AD patients compared to young [117]. Evidence suggests that hypo-methylated promoter region of APP gene was correlated with an increased Aβ production [118], which resulted in an increase of the genome-wide hypo-methylation, leading to upregulation of neuro-inflam‐ mation and apoptosis genes, subsequently applying a positive feedback control on Aβ production [105].

The changes in DNA methylation at PSEN1 and APOE promoter regions are variable based on results from different studies. PSEN1 promoter may be up- or downregulated by DNA methylation in AD [119, 120]. PSEN1 promoter hypo-methylation increased PSEN1 expression which resulted in an elevated Aβ production [121]. The APOE gene presents a duality in its structure; while its 5′-promoter CpG site is hypo-methylated, the 3′-CpG island is hypermethylated. Wang et al. suggested that aberrant epigenetic modifications in these CpG sites may contribute to LOAD [105, 118]. High levels of CLU (APOJ) gene, due to high methylation of CpG regions in the promoter of CLU, were observed in AD and might be associated with disease severity and clinical progression [122].

Tau promoter region was also found to be affected by methylation changes during AD [117]. For example, Aβ25–35 induced demethylation and increased tau phosphorylation and NFTs formation [123], which may be resulted from hypo-methylation of protein phosphatase 2A (PP2A) [102, 103, 124].

Recently, Sánchez-Mut et al. studied CpG 5′-region gene methylation patterns in different brain regions of AD mouse models and found hyper-methylation of three new target genes which could be involved in AD: thromboxane A2 receptor (TBXA2R), sorbin and SH3 domain containing 3 (SORBS3), and spectrin β4 (SPTBN4) [125]. Finally, genes involved in cell cycle and apoptosis were found to be modulated by DNA methylation and upregulated in AD neurons and aging AD brains [105].

#### **6.3. miRNAs regulation in AD**

MicroRNAs (miRNAs) are noncoding regulatory RNAs that are known to modulate ∼60% of genome via post-transcriptional gene silencing. The alterations in epigenetic modulations by miRNAs may promote abnormal expression of genes involved in AD [126, 127]. For example, Kumar et al. discovered a unique signature of seven circulating miRNAs in the plasma that could differentiate AD from non-AD individuals with >95% accuracy [128]. Similarly, another miRNA-based signature from blood samples has been reported, which allowed disease detection with 93% accuracy and 95% specificity [129]. It was also reported that four miRNAs (miRNA-9, miRNA-125b, miRNA-146a, miRNA-155) were involved in pathogenic signaling in AD brains and increased levels of these miRNAs were found in the CSF and brain samples of AD patients [130].

Within an exhausted list of miRNAs in AD pathogenesis, some directly regulate APP mRNA [105]. For example, miR-101 subexpression decreased APP levels and Aβ plaque formation in neurons [131]. Conversely, miR-16 over-expression may trigger an impaired APP expression [132]. miR-124 was reported to alter splicing of APP exons 7 and 8 in neurons [133], and to regulate BACE1 expression [112]. Over-expression of miR-29c, miR-298, miR-328, and miR-195 reduced BACE1 expression and thereby decreased Aβ generation [134–136].

Several miRNAs were found to regulate tau expression. For example, miR-132 was downre‐ gulated in some tauopathies [133]. miR-9, miR-124, and miR-15a were reported to be down‐ regulated in AD, affecting tau levels [78, 133]. The miR-15/ERK1 pathway that modulates tau hyper-phosphorylation was found to be downregulated in AD brains [78]. Altered levels of miR-26a in AD inhibited GSK-3β expression and thus affected production of NFT and Aβ in AD [137–139]. In a mouse model with impaired miRNA production, tau was highly phos‐ phorylated leading to NFT formation in mouse brains [140]. Finally, downregulation of miR-212 was involved in NFT density in AD [139, 141].
