The Effects of APP, Presenilin and ApoE4
In rare cases (< 5%), Alzheimer's disease (AD) is caused by genetic mutations in one of three genes: APP, PSEN1 and PSEN2, encoding amyloid precursor protein (APP), presenilin 1 (PS1) and presenilin 2 (PS2), respectively. Individuals with these inherited mutations may develop AD early in life; often in their 50s for APP mutations, and 40s for presenilin mutations (Selkoe, 2001). This type of AD is called familial AD or early-onset AD. In most cases, AD is not associated with these mutations and usually begins after 60. This type of AD is referred to as sporadic AD or late-onset AD.
Apolipoprotein E (ApoE) is the principal cholesterol carrier in the brain. It can have three major variants: ApoE2, ApoE3 and ApoE4. ApoE4 has been found to increase the risk for AD. How ApoE4 and the mutations in APP and presenilin predispose to AD is a subject of intensive research. The Amyloid Cascade Hypothesis (Hardy and Higgins, 1992; Selkoe and Hardy, 2016) centers on their relationship with the amyloid beta peptide (Aβ). This hypothesis appears to gain support from earlier experimental findings. However, recent studies have revealed new roles of APP, presenilin and ApoE4. This paper will focus on Aβ-independent mechanisms.
Amyloid Precursor Protein
AD patients with APP mutations are very rare. Only about two dozen cases have been reported (Selkoe, 2001). These mutations are located around the cleavage sites for α-, β-, and γ-secretases (Paper 7), suggesting that the APP fragments resulting from proteolytic processing at these sites may play a role in AD. It has been found that most of the mutations lead to higher Aβ42 level (Weggen and Beher, 2012). This finding serves as important basis for the Amyloid Cascade Hypothesis.
The pathogenic APP mutations also change the level of APP intracellular domain (AICD) (Kim et al., 2003; Wiley et al., 2005; Weggen and Beher, 2012). As discussed in Paper 7, AICD may enhance excitability via Ankyrin-G and GSK-3. In addition, AICD regulates the expression of APP (von Rotz et al., 2004), which has been demonstrated to play a crucial role in synaptic functions (Klevanski et al., 2015; Ludewig and Korte, 2017). Therefore, either higher or lower level of AICD may have detrimental effects on synaptic functions.
Down syndrome is a genetic disorder caused by the presence of three copies of chromosome 21 where the APP gene is located. The extra copy of APP gene will produce larger amount of APP, and through proteolytic processing, higher level of Aβ. Down patients are prone to develop AD. This fact has inspired the Amyloid Cascade Hypothesis (Selkoe and Hardy, 2016). However, the proteolytic processing of extra APP also results in elevated AICD which may cause AD-like features (Ghosal et al., 2009), as well as seizures (Vogt et al., 2011). Incidentally, Down syndrome is known to associate with seizures (Lott et al., 2012).
Presenilin is the catalytic subunit of the γ-secretase complex involved in the production of Aβ and AICD from APP (Paper 7). Earlier studies found that the pathogenic mutations in presenilin increased Aβ (Selkoe, 2001). However, in a recent study which analyzed the effects of 138 pathogenic mutations on the catalytic ability of presenilin to produce Aβ, only 13 mutations increased Aβ while all others decreased both Aβ42 and Aβ40. Furthermore, the Aβ42/Aβ40 ratios show no statistically significant correlation with the mean age at onset for the corresponding mutations (Sun et al., 2017). This raises the possibility that the presenilin mutations may predispose to AD via a mechanism independent of γ-secretase function.
In a seminal work, Cheung et al. (2008) demonstrated that presenilin mutants could bind with the inositol trisphosphate receptor (IP3R), thereby enhancing Ca2+ release from the endoplasmic reticulum. This discovery was substantiated by a follow-up study (Cheung et al., 2010), and corroborated by the finding that reduced IP3R expression rescued aberrant hippocampal long-term potentiation in PS1M146V knock-in mice (Shilling et al., 2014). Hence, presenilin mutations may exacerbate AD by elevating cytosolic Ca2+ concentration. This mechanism supports the hypothesis that AD is fundamentally caused by Ca2+ overload (Paper 6).
The ability of ApoE4 to clear Aβ from the brain appears to be less than other isoforms (Castellano et al., 2011). This report was used to support the Amyloid Cascade Hypothesis. On the other hand, several lines of evidence suggest that ApoE4 may reduce the level of brain-derived neurotrophic factor (BDNF), thereby enhancing excitability and Ca2+ overload.
The BDNF Cascade Hypothesis
According to the BDNF Cascade Hypothesis, neurodegenerative disorders are fundamentally caused by Ca2+ overload, which is initiated by BDNF deficiency, leading to neuronal hyperexcitability. Various neurodegenerative disorders differentiate at the neurons that exhibit hyperexcitability and Ca2+ overload. AD begins in the entorhinal cortex (EC) for two possible reasons. (1) The stellate cells of EC layer II are intrinsically very excitable. They regularly display subthreshold oscillations such that minimal elevation (1-3 mV) of membrane potential is sufficient to generate spikes (Alonso and Klink, 1993). (2) The expression of Tau proteins is almost twice as great in EC than elsewhere in the brain (Shukla and Bridges, 1999).
Each component in the main pathogenic cascade may be influenced by other factors. For instance, ApoE4 and Aβ oligomers may reduce BDNF level, AICD may enhance excitability and presenilin mutations can aggravate Ca2+ overload. The Ca2+ overload could lead to oxidative stress by disrupting mitochondrial functions (Peng and Jou, 2010). Numerous studies have demonstrated that oxidative stress promotes the production of Aβ (Oda et al., 2010; Zhao and Zhao, 2013; Arimon et al., 2015), which may aggregate to form Aβ oligomers and plaques.
From Figure 1, we see that there is a loop from BDNF deficiency to the production of Aβ oligomers which in turn can reduce BDNF level (Peng et al., 2009; Xia et al., 2017). Therefore, on the basis of this cascade alone, AD might originate from Aβ. However, by considering other experimental findings, this possibility can be ruled out. AD is known to begin in EC, but postmortem studies have shown that Aβ deposits appear first in neocortex (Thal et al., 2002). Recent PET studies on living people also suggest that Tau tangles, but not amyloid-β plaques, correlate with cognition and clinical symptoms (Tosun et al., 2017). In accelerated-senescence nontransgenic rats, named OXYS rats, Aβ deposits occur later than synapse loss, neuronal death, mitochondrial abnormalities, and Tau hyperphosphorylation (Stefanova et al., 2015). Moreover, Aβ oligomers are not the only factor that can cause BDNF deficiency. Other abnormalities may also reduce BDNF level, such as glucocorticoid elevation (Suri and Vaidya, 2013; Wosiski-Kuhn et al., 2014), estrogen deficiency (Carbone and Handa, 2013) and melatonin deficiency (Imbesi et al., 2008; Zhang et al., 2013; Rudnitskaya et al., 2015). Hence, drugs targeting Aβ alone will not halt disease progression via other pathogenic pathways. This explains why "clinical trials with Aβ therapies, including immunotherapy, have thus far failed to show any reduction in neurofibrillary pathology or improvement in cognitive performance of patients with AD" (Dai et al., 2017).
Author: Frank Lee