2. The Role of Apolipoprotein E in Alzheimer’s Disease Pathogenesis
For sporadic AD, the
APOE 4 allele is the most important genetic risk factor, as well as for the earlier stages of cognitive decline represented by mild cognitive impairment (MCI)
[21], but its expression is poorly understood. Astrocytes and activated microglia produced the major amount of ApoE in the brain. Having one
APOE ε4 allele conducts to a 4-fold increased risk of developing AD, while having two
APOE ε4 alleles conducts to a 12-fold increased risk, in comparison with the
APOE ε3-carriers. Conversely, the uncommon heterozygous carriers of the
APOE ε2 allele have an AD risk 40% lower and the homozygous carriers have a further reduced risk
[22]. In brain,
APOE ε4-carriers with normal cognition displayed higher Aβ and tau burden than
APOE ε3-carriers; conversely,
APOE ε2-carriers had reduced global Aβ burden, without differences in regional tau burden or accumulation over time
[23]. The contribution in AD pathogenesis from
APOE involves not only Aβ aggregation and its clearance, but also tau-mediated neurodegeneration
[24], microglia impairment
[25][26], astrocyte reactivity
[27], and blood–brain barrier disruption
[28][29].
The three ApoE isoforms bind and transport Aβ peptides with differential affinity during AD pathogenesis
[30][31], being highest for ApoE4, intermediate for ApoE3, and lowest for ApoE2
[32][33]. Therefore, their effects are also different concerning Aβ aggregation and clearance, but not Aβ production
[34][35]. ApoE also can affect tau-mediated neurodegeneration and tauopathy by modulating microglial responses to Aβ plaque pathology
[36][37][38]. Thus, different ApoE isoforms may be associated with an increased or reduced AD risk
[31][32], based on different combined effects of ApoE isoforms on both deposits of Aβ and neurofibrillary tangles
[39].
APOE and its ε2/ε3/ε4 alleles have been connected by several genetic studies to different disorders and physiological conditions. Epigenetic alterations could explain the association between
APOE and its associated diseases, considering that disease-associated genetic signals may also reflect a site sequence architecture for epigenetic codes
[12].
3. Apolipoprotein E, Human Longevity, and Alzheimer’s Disease
A genetic association of
APOE with both human longevity and AD was found, but the mechanistic contribution of
APOE in aging and long life is largely under investigation.
APOE pleiotropic roles may be explained by its exceptional epigenetic properties. In the AD brain, these epigenetic changes could contribute to neural cell dysfunction. Additionally, DNA methylation modifications have been found on specific genes associated with AD pathology such as
APOE. In the AD brain, it was shown that
APOE CpG islands were differentially methylated in an
APOE-genotype and tissue-specific way
[40]. In the
APOE CpG islands, AD typically showed a lower level of DNA methylation occurring in brain regions affected by AD pathophysiology (highest levels in the cerebellum, with moderate levels in the hippocampus and the lowest levels in the frontal lobe). However, in the
APOE CpG islands, there was a complex interplay among the presence of the
APOE ε4 allele, AD status, and DNA methylation levels. AD-specific methylation differences were mainly attributed to the
APOE ε3/ε4 heterozygous subjects
[40]. Allele variations in the major
APOE CpG islands of targeted replacement (TR) mice expressing human
APOE may affect its methylation in the brain
[41]. Epigenetic changes may link modified gene expression with environmental stimuli such as dietary patterns and physical exercise. In animal models,
APOE alleles may have alterations in epigenetic regulation responding to external stimuli reported in studies on
APOE TR mice
[42].
In the strategy for replacing mouse ApoE in the
APOE TR models, the differences between human and mouse
APOE gene clusters should be considered, the complexity of transcriptional control of human
APOE, and the structure of the targeting construct
[43]. Moreover, lifestyle factors like education, smoking, alcohol consumption, and physical activity may weaken genetic risk in the process of age-related cognitive decline and dementia. At this regard, twelve potentially modifiable risk factors have been investigated to prevent or delay up to 40% the risk for different types of dementias including AD, i.e., lower educational level, hypertension, hearing loss, smoking, obesity, depressive syndromes, physical inactivity, diabetes, low social contact, immoderate alcohol consumption, air pollution, and traumatic brain injury
[44]. The complex interactions among lifestyle, genetics, and age-related cognitive decline may encourage behaviors maintaining cognitive health in older age, including dietary habits
[45]. At this regard, ApoE may be important for the pathophysiology of lipid metabolism
[46] and central nervous system (CNS), although the role in healthy aging and longevity has seen its value grow
[47][48][49].
In the lipid metabolism pathophysiology, ApoE may be related with normal/pathological aging, while its function in CNS pathophysiology needs further clarification
[50]. In fact, in the CNS, there was about a quarter of total body cholesterol that may exert a significant impact on synaptic plasticity
[51]. With advancing age, cholesterol metabolism may modify, and its related brain changes may be associated with the pathophysiology of AD
[51]. So, in longevity and healthy aging, lipid and cholesterol maintenance are a critical factor also from an interventional point of view. Dietary interventions may manage the detrimental effects of the
APOE ε4 allele
[52], with a Mediterranean dietary pattern potentially including higher n-3 polyunsaturated fatty acid (PUFA) intakes
[53][54].
Studies on longevity and healthy aging are related because subjects who live longer tend to be healthier for a greater part of their lives
[55]. Healthy aging can be described as achieving older age maintaining intact cognition and/or mobility and without disabilities or multimorbidity. This last can be defined as the coexistence of two or more chronic diseases in the same subjects
[56]. The detrimental effects of the
APOE ε4 allele on longevity could influence the probability of a long human lifespan
[55]. The
APOE ε2 allele has a greater frequency in long-lived individuals than the ε4 allele
[57]. Thus, the main longevity factor is the
APOE ε3/ε3 genotype. The greater frequency of the ε3 allele in older individuals and their offspring than in controls derives from the higher amount of the homozygous
APOE ε3/ε3 genotype in comparison with the ε2/ε3 or ε3/ε4 genotypes
[58].
4. Specific Epigenetic Modifications of Apolipoprotein E in Alzheimer’s Disease
In response to environmental stimuli, epigenetic marks and signals may enable temporal combination of regulatory events through mechanisms including DNA methylation, histone modification/chromatin conformation, and noncoding microRNAs (miRNAs). In the
APOE gene, several studies investigating DNA methylation suggested an age-dependent flow and
APOE DNA methylation specific for brain area. The impact of aging on
APOE methylation is based on the general link between DNA methylation and longevity, although current studies on the association among aging and
APOE methylation patterns are limited and with little sample size as later described. The
APOE genomic sequence is approximately 4 kb in size (chromosome19:45408714-45412650, hg19) including its promoter. This region encompasses 172 CpG dinucleotides
[59]. At the
APOE locus, three functional SNPs may modify DNA methylation, and of these, rs405509 is in the promoter region, while the other two SNPs (rs429358 and rs7412), which define the
APOE e2/e3/e4 isoforms, are within exon 4 (
Figure 1). Genetic variability in the
APOE neural expression may contribute to the risk of AD considering that relative
APOE ε4 mRNA expression is higher in AD patient than in healthy controls
[60]. Moreover, together with the qualitative effect on the AD risk of the
APOE e2/e3/e4 polymorphisms, functional
APOE promoter mutations may determine quantitative variation of expression of these alleles that is a fundamental determinant of AD occurrence. In the late 1990s and early 2000s, polymorphic sites in the first intron and the proximal promoter the of
APOE gene cluster (−1019 to +407) affecting
APOE expression were identified, showing the deleterious effect of the Th1/E47cs T allele and the protective effect of the −491 T allele
[61][62][63][64][65] (
Table 1). Notably, these polymorphisms have been related with a differential AD risk
[66][67][68].
Figure 1. Apolipoprotein E (APOE) gene pleiotropic roles may be explained by its exceptional epigenetic properties. The APOE ε2/ε3/ε4 alleles are produced by two cytosine-phosphate-guanine (CpG)-altering SNPs (rs429358 and rs7412) in the core region of the APOE CpG islands. APOE ε4 carriers have the greatest number of CpG dinucleotide sites, while APOE ε2 carriers have the smallest number, so methylation levels for most CpG sites are in the order of APOE ε4 carriers > APOE ε3/ε3 > APOE ε2 carriers. In the promoter region, there is another SNP, rs405509. Furthermore, other epigenetics mechanisms are linked to the processing of primary miR-650 to mature miR-650 is mis regulated. miR-650 can significantly reduce the expression of APOE.
In AD, the association between these polymorphic sites and the variability of sequence in the proximal promoter with ApoE protein levels are not clearly understood. In fact, among different studies, findings on the levels of expression of
APOE RNA and the relationship with the ApoE levels varied. In human AD postmortem brain, there was elevated methylation in frontal lobe of a 5′-C-phosphate-G-3′ (CpG) island overlapping with exon four and downstream
[69].
APOE has a well-defined CpG island external to the promoter region and overlapping with the
APOE 3′-exon. In the human genome, these 3′-CpG islands are very rare, representing < 1% of the total CpG islands, and are also conserved in other mammals
[70][71]. However, the
APOE CpG island methylation level relates to the level of expression of four known
APOE transcripts. The majority of the total
APOE mRNA, with higher expression in the AD frontal lobe than in the frontal lobe of control subjects, is constituted by circular RNAs, mRNAs, and truncated
APOE transcripts. The findings of several studies suggested several changes in the epigenome and the regulatory role of epigenomic elements related to the risk or clinical presentation of several neurological diseases, although the exact clinical significance of these signatures in the quantities of RNA and methylation level of CGI in the
APOE 3′-exon was still unclear
[69] (
Table 1).
At the level of the individual CpG site, epigenetic regulation was shown by up/down patterns in the methylation profiles between samples and tissues. Significant differences in the global methylation levels among several regions of the brain were discovered across postmortem brain tissues. In brain regions primary affected by AD such as frontal lobe, temporal lobe, and hippocampus, methylation levels were lower. Conversely, in the cerebellum, a region apparently without profound pathological alterations in AD but with recent important findings, the highest methylation levels, suggesting a correlation between the methylation levels of the
APOE CpG islands and the vulnerability of brain regions in AD patients
[40]. In fact, age- and AD-related alterations in several cerebellar subregions may also impact numerous functional domains, especially those affecting cognitive processing
[72].
Genetic variants, which consist of CpG-altering SNP, can modify DNA methylation levels. These genetic variations may act like regulatory elements connecting genetic changes not only with the protein isoforms, but also with epigenetic variability
[73] (
Table 1). As previously described, the
APOE ε2/ε3/ε4 alleles are produced by two CpG-altering SNPs (rs429358 and rs7412) residing in the core region of the
APOE CpG islands. The
APOE ε4 allele, if compared with ε2 or ε3 alleles, adds one more CpG, further saturating a small 12 bp region with 4 CpG sites. On the contrary, the
APOE ε2 allele eliminates 1 CpG and opens a 33-bp CpG-free region. Consequently, these two SNPs may alter the regional CpG burden and probably influence global DNA methylation of the CpG islands (
Figure 1). These CpG load changes might change the binding profiles of methyl-CpG binding domain proteins, associated to methylated DNA through their exclusive amino acid patterns
[74].
Furthermore, within the
APOE CpG islands, there is evidence of indirect indicators of protein binding which consist of histone marks and a DNase I hypersensitivity cluster. These findings suggested that the
APOE CpG islands and exon 4 may be a site for chromatin remodeling and protein binding. Considering that environmental stimuli could influence DNA methylation gradually with aging, the differences in
APOE CpG island methylation between healthy subjects and patients with AD increased with age
[75]. Taken together, different methylation scenarios may be represented by the inheritance of different ε2/ε3/ε4 alleles in the
APOE CpG islands, which could accumulate or change continuously with age, also modified by environmental factors. Recent results showed that methylation levels for most CpG sites may be in the order of
APOE ε4-carriers (greatest number of CpG sites) >
APOE ε3/ε3-carriers >
APOE ε2-carriers (smallest number of CpG sites)
[76] (
Table 1). These changes could potentially alter protein binding, with some consequences on biological systems, even affecting the pathophysiological processes of multiple diseases and plasma lipids levels.
APOE methylation could partially mediate the effects of age on plasma lipid (
Figure 2).
Table 1. Overview of studies illustrating epigenetic signatures of apolipoprotein E gene (APOE) in aging and Alzheimer’s disease (AD).
Figure 2. Despite the high lifetime risk linked to the presence of the APOE ε3/ε4 and APOE ε4/ε4 genotypes (the greatest risk factor for developing Alzheimer’s disease, AD), stochastic factors (such as environment, diet, physical exercise, and aging), may play a significant role as epigenetic modifiers, influencing the imbalance among the different ApoE isoforms. The role of APOE ε4 allele in AD pathogenesis involves not only amyloid-β aggregation and clearance, but also tau-mediated neurodegeneration, microglia dysfunction, astrocyte reactivity, and blood–brain barrier disruption. These changes may only occur in brain regions profoundly affected by AD pathophysiology (highest levels in the cerebellum, with moderate levels in the hippocampus and the lowest levels in the frontal lobe).
In the epigenetic scenario, miRNAs are known to be small non-coding RNAs with a length of ~22 nucleotides. They are also implicated in AD, as shown by the altered expression of miRNA 650 (miR-650) in AD brains
[78]. Bioinformatic analysis showed that miR-650 may target the expression of three AD-related components:
APOE, presenilin 1 (PSEN1), and cyclin-dependent kinase 5 (CDK5), with recent findings confirming that miR-650 may reduce in vitro the expression of
APOE, PSEN1, and CDK5
[78].
5. Epigenetics of Apolipoprotein E and Cognitive Function: Contrasting Evidence in Alzheimer’s Disease
Several lifestyle and environmental stimuli could explain the effects of
APOE genotype on AD and cognitive functioning, such as exercise
[79], education
[80], and vitamin D status
[81] (
Figure 2). Vitamin D is often referred to as a neurosteroid with neuroprotective and anti-inflammatory properties. Although the interplay between
APOE genotype and vitamin D metabolism or transport in the nervous system is yet to be established, it is known that
APOE contributes to the transport of lipid-soluble vitamins in the circulation and influences several immunological, inflammatory, and neurodegenerative processes
[82]. Furthermore,
APOE polymorphism and dietary responsiveness to fat-soluble vitamins, flavonoids, and n-3 PUFA were described, indicating
APOE ε3 as a more flexible and responsive genotype than
APOE ε4
[82]. Among implications for the development and progression of AD, vitamin D supplementation may be another potential strategy to consider for the
APOE ε4 allele-carriers. Some reports showed that higher vitamin D concentrations in
APOE ε4 homozygous carriers allow them to perform better at memory scores
[83]. Then, compared to the
APOE ε3/ε3-carriers, the
APOE ε4-carriers showed earlier onset of cognitive impairment in AD. However, after the disease onset, the effect of the
APOE genotype on the progression of cognitive impairment remained debated
[84].
For this reason, epigenetic modifications of
APOE, such as DNA methylation, may have a key role in maintaining intact cognitive function in older age. Growing DNA methylation levels at the
APOE promoter region were found on postmortem prefrontal cortex samples of sporadic AD individuals using MALDI-TOF mass spectrometry and lymphocytes
[77]. A notably age-specific epigenetic drift was identified, supporting a potential role of epigenetic effects in AD development. These results may partly explain the differences between
APOE ε4 carriers and noncarriers in the benefits triggered by long-term exercise that might depend, at least partially, on mechanisms of metabolic response of prefrontal cortex to physical activity
[79].
Numerous studies have indagated the relationship between
APOE DNA methylation and AD or MCI
[85][86][87]. Instead, the association between
APOE DNA methylation and cognitive function in healthy subjects without cognitive impairment was evaluated by two studies, with controversial findings
[88][89]. Liu and colleagues found an inverse association between DNA methylation in the
APOE gene region and delayed recall capacity among 289 older African American people with a mean age of 67 years during normal cognitive aging
[88]. Conversely, the other study, conducted in a large European cohort, observed no association between general cognitive functioning and
APOE DNA methylation
[89].
Many reports have suggested that neuroinflammation may have a key role in AD pathogenesis
[90]. Dietary habits are known to influence systemic inflammation, neuroinflammation, and inflammaging
[91]. A recent study conducted in a cohort of racially diverse middle-aged people (
n = 411), pursued to identify DNA methylation sites associated with cognitive function in the genomic region of
APOE. Regarding the inflammatory potential of the diet, among the dietary inflammatory index, cognitive performance, and the methylation level of several CpG sites have been detected significant relationships
[92].
However, studies are contrasting in this regard, and whether epigenetic biomarkers could be used for predicting AD is still unclear. In the
APOE gene, DNA methylation at two CpG sites (3/13) that are known to show age-dependent changes was related with the total cholesterol and high-density lipoprotein cholesterol ratio, but not with cognitive status, family history of AD, or the risk of cardiovascular disease in a blood-based DNA methylation study of 5828 people from the Generation Scotland cohort
[89]. These findings supported that there is no evidence yet for considering
APOE methylation as a biomarker for predicting AD or cardiovascular disease, although
APOE methylation was associated with the blood levels of cholesterol
[89].
Some limitations could affect specific methodologies used in the studies cited for assessing DNA methylation of
APOE. Overall, at the transcriptional level, all major cell types have AD pathology, and single cell-level resolution may be critical; moreover, changes in gene expression, including directionality, can be conditional on cell type. The number of significant differentially expressed genes for non-neuronal populations were substantially smaller, likely reflecting reduced power in lower-abundance cell types
[93]. Given that astrocytes are the primary producers of brain ApoE, alterations of epigenetically regulated
APOE expression in glia may explain a significant part of the genetic AD risk linked to this gene
[94]. Furthermore, although several imputation methods exist, a major deficiency lies in the inability to cope with large datasets, such as DNA methylation chips. Therefore, specific methods for imputing missing methylation data are needed
[95].