Epigenetic modifications of chronic hypoxia-mediated neurodegeneration in Alzheimer’s disease

Chronic hypoxia is a reduction of oxygen supply, which is a direct consequence of hypoperfusion and restricts the function of organs, tissues or cells. Efficient oxygen delivery to brain tissues is essential for normal brain function and cells survival. Clinically, chronic hypoxia is a common pathophysiological event and contributes significantly to the progression of widespread diseases including fever, chronic obstructive pulmonary disease, obstructive sleep apnea syndrome, stroke, cancer, neurodegenerative disorders and etc.[16]. It has been shown that individuals who have suffered severe hypoxia or ischemia are more susceptible to developing AD[17‐19]. Researchers reported that Aβ deposited in the brain after severe head injury[12, 20], and cardiac arrest induced a time-dependent increase level of Aβ in the serum of heart attack patients[21]. A recent clinical prospective study by Yaffe and colleagues, demonstrated that older women with sleep-disordered breathing and hypoxia had a significantly increased risk of developing cognitive impairment and dementia[22]. Their findings suggested that chronic hypoxia but not sleep fragmentation or duration was associated with the higher risk of mild cognitive impairment or dementia. Another prospective study named Health, Aging, and Body Composition (Health ABC) that took a follow-up observation over the 11 years has shown that anemia was associated with an increased risk of developing dementia among older adults[23].

Consistent results have been published in transgenic mouse models or cell models showing the impact of chronic hypoxia on AD. It is found that hypoxia increased both mRNA and protein levels of APP in cells and elevated the level of Aβ in culture medium significantly[24]. Our previous study has reported that repeated hypoxia treatment could lead to more and larger senile plaque formations and more Aβ42 production in aged APP^Swe + PS1^A246E double transgenic mice[25]. Interestingly, similar neuropathology was observed in the study of prenatal hypoxia in AD development in adult mice[26]. It has been found that prenatal chronic intermittent hypoxia significantly contributed to memory and cognition deficit in adult mice[26]. Moreover, prenatal hypoxia increased senile plaque formation and Aβ production in adult mice[26]. Another study also showed that hypoxia treatment impaired spatial learning and memory of AD transgenic mice, and caused ER stress and neuronal apoptosis[27]. These effects were associated with abnormal calpain activation because suppression of m-calpain expression could attenuate the hypoxia-induced ER stress and apoptosis[27].

Abnormal cleavaged β-amyloid peptide is the main constituent of neuritic plaques, which derives from sequential cleavage of APP by β-secretase and γ-secretase[28]. In the brain, the clearance of Aβ is mainly degraded by neprilysin (NEP) and insulin-degrading enzyme (IDE)[29]. Several studies including ours have shown that chronic hypoxia may elevate the production of Aβ and reduce the degradation of Aβ[25, 26, 30‐33]. Chronic hypoxia in the human neuroblastoma SH-SY5Y cells caused reduced expression ADAM10, part of the α-secretases and increased the expression of β-site APP cleaving enzyme 1 (BACE1)[34, 35]. Our previous study has demonstrated that chronic hypoxia increased Aβ generation by altering β- and γ-cleavage of APP[25]. The ratio of C99/C83 was elevated by chronic hypoxia in the brain of AD transgenic mice and the expression level of APH-1a was enhanced under hypoxia condition, which in turn would lead to increase in Aβ production[25]. Others have also shown that hypoxia could up-regulate BACE1 and APH-1a at both transcriptional and translational levels in vitro and in vivo[30, 31, 33]. It is widely accepted that hypoxia-inducible factor-1 (HIF-1) is the master regulator of the cellular response to chronic hypoxia[36]. Further investigations have revealed a functional hypoxia responsive element in the BACE1 gene promoter, to which HIF-1α can bind and result in increased activity of β-secretase under hypoxia conditions[30, 33]. Similar HIF-1α binding site AP4 was identified in the promoter of APH-1a[31]. The binding of AP4 and HIF-1 to the promoter under hypoxia conditions may significantly affect the expression of APH-1a and lead to increased activity of γ-secretase[31]. All these findings suggest that hypoxia may increase the generation of Aβ in vivo and in vitro.

It is known that NEP plays a key role in the degradation of Aβ in the brain and it is one of the most important Aβ-degrading enzymes[29, 37]. Evidence showed that NEP mRNA, protein and activity levels were declined not only in AD but also in the normal aging in the brains[38‐42]. Fisk et al. demonstrated that hypoxia reduced NEP expression at the protein and mRNA levels as well as its activity[43, 44]. We previously tested the expression levels of NEP in the cortex of AD transgenic mice, and found that the protein level of NEP started to reduce at the age of 6-month old in the brain of AD mice in comparison with age-matched wide-type mice[26]. Interestingly, the results also documented that chronic hypoxia induced significantly down regulation of NEP protein level in both transgenic mice and wide-type mice[26]. The declined level of NEP would reduce the clearance of Aβ. Given to the aggravated pathogenesis of AD and the over-load of Aβ burden under hypoxia condition, the chronic hypoxia-induced down regulation of NEP was implied to be a significant event in AD.

It has been indicated that environmental factors and epigenetic mechanisms are likely to contribute to the etiology of LOAD[8‐10]. Epigenetic mechanisms modify heritable and non-heritable traits without altering the underlying DNA code, mediated through the reversible modifications of DNA and histones[45]. Epigenetic processes play an important role in normal physical functions of cells and the body, so aberrant epigenetic modification are hypothesized to contribute to a majority of pathologies[46]. Abnormal PS1 methylation patterns have previously been associated with hypomethylation in promoter[47]. DNA demethylation has a strong correlation with transcriptional activation[48]. Hypomethylation in promoter CpG islands of other AD-associated genes such as APP and BACE1 has also been reported, which in turn may lead to abnormal up regulation of these genes and over-production of Aβ[49, 50]. In the study of epigenetic differences in monozygotic twins discordant for AD, a significantly reduction of DNA methylation was observed in the temporal cortex neuronal nuclei of AD twin[51]. It was shown that hypoxia can reduce global DNA methylation in cancer cell lines in vivo and in vitro[52]. Importantly, hypoxia could cause long-lasting change in DNA methylation change in promoter regions, some of which could be highly correlated with transcriptional modulation in a number of genes involved in neural growth and development[53]. Chen et al. found that Aβ could reduce global DNA methylation and increase NEP promoter methylation and further suppress the NEP expression in mRNA and protein levels[54]. However, the methylation status of the NEP promoters did not regulate its expression in vitro[55], and chronic hypoxia did not affect the methylation patterns of NEP gene promoters in mouse primary cortical and hippocampal neurons[32]. Interestingly, chronic hypoxia could cause significant down regulation of NEP by up-regulating G9a histone methyltransferase and histone deacetylase 1 (HDAC1), which resulted in increased expression of H3K9me2 and decreased expression of H3-Ace respectively[32]. In addition, methylation inhibitor 5-Aza, HDAC inhibitor valproic acid VA, and siRNA-mediated knockdown of G9a or HDAC1 could reverse the expression of NEP[32]. Others also reported that the NEP promoter could be repressed by HDACs and the expression of NEP was repressed in neuronal cells via the competitive binding of HDACs to its promoter[55]. All these findings suggest that chronic hypoxia aggravates AD by epigenetic modulations, mainly focusing on DNA methylaion and histones modifications, which are of tremendous importance to the onset and progression of the disease (Figure 1).

Currently, the treatment of AD focuses mainly on improving symptoms, targeting cholinergic and glutamatergic neuron transmission, with no therapeutics that can cure the disease[56]. Researchers around the world have worked hard to search for novel therapeutics for AD, which include the anti-Aβ (drugs to reduce its production, prevent its aggregation and promote its clearance), anti-tau (drugs to prevent its aggregation or phosphorylation), neurotrophins, and others[56]. Epigenetic regulation is a strong candidate therapeutic, since epigenetic modifications are reversible while genetic mutations are not. Drugs which can modulate DNA methylation and remodel the structure of chromatin through post-translational modifications of histones are promising potential candidates[57‐60]. Recently, several drugs have been reported to be effective on AD mice models by modulating DNA methylation and histone acetylation separately, although none have yet entered clinical development[58, 59, 61, 62]. Sodium valproic acid is one of these disease-modifying drugs, which may not only attenuate AD pathogenesis in transgenic AD mice, but also show effect in hypoxia-induced AD in the model. Valproic acid is a widely used antiepileptic drug, which has recently been found to have neuroprotection[63, 64], and histone deacetylase inhibitory property may offer potential therapeutic option for AD[65]. It has been reported that valproic acid could inhibit Aβ production, reduce neuritic plaque formation, and improve behavioral deficits in AD mice[66]. Further study has demonstrated that valproic acid could attenuate the prenatal hypoxia-induced AD neuropathology, improved learning deficits and decrease Aβ42 levels in AD transgenic mice by up regulation of NEP[67]. Moreover, valproic acid could also restore memory deficit in adult offspring caused by prenatal hypoxia via elevating NEP expression and activity[68]. Hence, epigenetic drugs that increase histone acetylation by inhibiting HDAC could be useful in the disease treatment. Strategies targeting Aβ burden by epigenetic regulations, may represent new opportunities for therapeutic intervention to AD.