Alzheimer’s Disease (AD) is the most common form of dementia [1]. Patients diagnosed with AD experience disordered cognition and memory, as well as changes in behaviour and personality [2]. The vast majority of AD is diagnosed in patients aged over 65 years and classified as late onset (LOAD), with the remaining ~1% of cases termed early onset AD (EOAD) [3]. The molecular neuropathology of AD includes the presence of extracellular amyloid beta (Ab) plaques and proteinaceous lesions of aggregated neuronal tau protein across multiple brain regions [4,5]. Implicated regions include brainstem structures as the origin of progressive neurodegeneration [6,7], which then follows a predictable spreading pattern to the cholinergic basal forebrain [8], the (trans)entorhinal cortex, hippocampus, and the neocortex [9]. Sadly, there are currently no specific preventative measures or cures for AD.

The heritability of LOAD and EOAD is 60-80% and 90-100% respectively [10], and over 70 genomic loci have been associated with AD to-date [11]. These data support the presence of a strong genetic component in the development of AD (Figure 1). The rarest but most severe AD phenotypes are due to various autosomal dominant gene mutations in APP, PSEN1, and PSEN2; these lead to EOAD [12]. The most important genetic risk factor for developing LOAD is having the allelic variant e4 of APOE (APOE4), whereby homozygous (e4/e4) individuals have a 50-60% lifetime risk of developing LOAD [13].

Figure 1. Genes associated with increased risk for developing AD. Gene penetrance and frequency affect AD age at onset. APP & PSEN mutations are 100% penetrant but very rare. Microglial TREM2 mutations can be more common with mixed phenotypic effects. The APOE4 allelic variant is relatively common within the population and significantly increases AD risk. Various other GWAS hits are rare and have small effects. Some genes have multiple loci that are differentially associated with AD, e.g., TREM2, SORL1, ABCA7. Made with reference to three GWAS datasets [11,14,15].

ApoE was first associated with AD following its discovery in amyloid plaques [16]. Binding of ApoE4 to cell surface receptors induces microglial and astrocytic dysregulation, affecting cholesterol homeostasis and impeding Aβ clearance [17]. In human neurons, ApoE4 stimulates ERK1/2 MAP-kinase signalling through MAP3K12 and MAP2K7 [18]. MAP2K7 then phosphorylates cFos to induce APP expression and subsequent Aβ production and neurodegeneration, but also phosphorylates CREB, paradoxically stimulating synapse formation [18]. Introducing a conditional APP Swedish point mutation in human neurons, Zhou et al., [19] report increased total functional synapses upon expression; an effect that was abolished by inhibiting APP cleaving enzyme BACE1. Restoring Aβ to physiologic levels in neurons not expressing APP was synaptogenic and increased synaptic transmission [19]. These findings support a cautious approach in AD therapeutic strategies involving clearance of Aβ peptides and plaques.

In the noradrenergic (NA) neurons of the locus coeruleus (LC), the presence of ApoE4 increases tau neurotoxicity by selectively binding vesicular monoamine transporter 2, inhibiting vesicular uptake of neurotransmitter NA, leading to NA oxidation to DOPEGAL [20]. This toxic metabolite activates enzymatic cleavage of tau at site N368, triggering LC neurodegeneration, reducing hippocampal volume, and inducing cognitive dysfunction [20]. In a transgenic mouse model of tauopathy (P301S), human APOE4 knock-in (KI) mice (P301S/E4) show significantly exacerbated tau-mediated brain-wide neurodegeneration relative to E2/E3 KI and APOE knock-out (EKO) mice [21].

Wang et al. [22], showed that depletion of astrocytic ApoE4 rescues loss of synapses, reduces tau-mediated neurodegeneration and levels of hyperphosphorylated tau (pTau) in P301S mouse brain. Consistent with these findings, a binding partner of ApoE4, the astrocyte-secreted protein glypican-4 (Gpc4), drives tau hyperphosphorylation and propagation in vitro and in vivo [23]. Glpc4 enhances ApoE4-dependent trafficking of surface ApoE receptor LRP1 and tau propagation by internalising LRP1-ApoE4-Gpc4-pTau complexes [23].

The proteolytic cleavage of APP by β-/g-secretases leads to the generation of 40- and 42-amino acid-long Aβ peptides (Aβ₄₀, Aβ₄₂), of which Aβ₄₂ is less common but more toxic than Aβ₄₀ [24]. According to the amyloid cascade hypothesis the cause of AD is excessive levels of extracellular Aβ₄₂ [25]. Factors contributing to this include missense mutations in APP and PSEN1/2 genes which increase production of Aβ₄₂, and/or APOE4 inheritance and neuroinflammation leading to failures of Aβ₄₂ clearance [24]. These Aβ₄₂-driven effects purportedly culminate in dysregulated kinase-phosphatase signalling, inducing tau hyperphosphorylation, propagation, and deposition of aggregated tau, leading to neuronal cell death [26,27].

Protein changes in AD include metabolism/processing of Aβ and pTau, and accumulation of plaques and tangles respectively. These changes can be targeted and used as biomarkers, detected by combining imaging techniques (magnetic resonance imaging, MRI; positron emission tomography, PET) and fluid (blood; cerebrospinal fluid, CSF) analysis techniques. For instance, using a radioligand with affinity for neurofibrillary tangle tau and the spatial topography information conferred by PET, Therriault et al. [28], were able to apply Braak staging to living humans. Late tau-PET-based Braak stages (IV-VI) correlated with higher Aβ₄₂/Aβ₄₀ ratios in CSF. Elevated concentrations of CSF pTau₁₈₁, pTau₂₁₇, pTau₂₃₁, pTau₂₃₅, and plasma pTau₂₃₁ were detectable from PET-based Braak stage II compared to stage 0 [28].

Blood-based biomarkers, including soluble Aβ, total/phosphorylated tau, and inflammatory factors, have attracted research interest, as blood collection and biomarker detection are relatively non-invasive and economical. Despite this gain in popularity, blood-based testing has limitations. Minute biomarker concentrations with indistinguishable contributions from peripheral and central nervous systems, coupled to an inherently complex matrix of proteins in human blood can interfere with reliable protein identification and quantification.

Regarding treatment of AD, as of 26/12/2022 (index date for this editorial) there are 235 interventional clinical trials for AD, whereby the vast majority (197; 83.8%) involve drug therapy (Table 1).

While most participants are enrolled in drug-based clinical trials, a sizeable minority (~20%) are undergoing behavioral training, using biological interventions, or various devices or diets. This diversity of approaches offers some hope and promise to those suffering from AD-related symptoms globally.

As post-mortem brain tissue analyses reveal, the presence of accumulated plaques is not sufficient to predict LOAD. Conversely, while neural tissue patterning of tau spreading and aggregation strongly correlates with AD neurodegeneration, tau aggregation may manifest at a late and irreversible stage in the disease process.

A consistent presentation among patients diagnosed with LOAD is the rate of disease progression and age-at-onset of disease. That is, a pattern of neurodegeneration emerges in the population after a (somewhat) predictable period of environmental exposure (i.e., living to beyond the age of 70 years). Within a shared environment, some people never develop AD and die beyond the age of 90 from other causes; others carry one or two copies of APOE4 and die of LOAD, while others are heterozygous for alleles with highly penetrant, disease-associated mutations and develop EOAD. Assuming there are no overt yet-unknown environmental causes of AD, the in-between cases – as per the described extremes – must depend on genome sequence identity coupled to gene expression: a variable resistance to the persistent environmental onslaught that is life.

So which events might signify impending neurodegeneration? One’s attention might focus on two potential causes of AD, which are not mutually exclusive: (i) Specific molecular cell surface contacts of Aβ plaques (rather than mere existence of plaques); (ii) Specific molecular interactions of toxic Aβ₄₂ oligomers in equilibrium with Aβ plaques. The interactions in (i) and (ii) include pathologic stabilisation of integral membrane proteins and dysregulated activation of various intracellular signalling cascades. Consequent effects include excitotoxicity, and dysregulated proteostasis and post-translational modifications, which lead to death of cells containing hyperphosphorylated and aggregated tau protein.