Apolipoprotein E: Cholesterol metabolism and Alzheimer’s pathology

Introduction

Alzheimer’s disease (AD) is a devastating neurodegenerative disease associated with profound memory loss and cognitive dysfunction. More than a century after Alois Alzheimer described the first case of AD, according to Alzheimer’s Disease International, the disease now afflicts 50 million people worldwide. Since age is the greatest risk factor for AD, increasing life expectancy makes a dramatic contribution to these demographics. It was only in the late 1970 s, when Robert Katzman defined AD as one of the world’s greatest killers, that AD was recognized as an epidemiological disease (Katzman, 1976). Two different types of AD exist; the early onset (EOAD) and the late onset (LOAD) forms. By definition, patients below the age of 65 when diagnosed suffer from EOAD, and patients who develop symptoms after 65 years of age have LOAD. Katzman observed that the general decline in cognition and the progression of neurodegeneration followed a similar pattern in LOAD and EAOD. The more aggressive form of EOAD is rare, accounting for only 1–5 % of all AD cases, and is caused by de novo or familial genetic mutations. Affected genes encode the amyloid-β (Aβ) precursor protein (APP) or APP processing proteins, each of which trigger enhanced production of the Aβ-peptide that forms neurotoxic oligomers and ultimately aggregates in extracellular deposits, called plaques. In LOAD, mechanisms involved in reduced Aβ clearance, rather than overproduction of Aβ, are believed to be a major contributor to Aβ-toxicity and plaque deposition (Wildsmith et al., 2013). Importantly, the vast majority of AD cases are defined as LOAD and the most prevalent genetic risk factor for these cases is apolipoprotein E (ApoE) isoform ɛ4 (E4); 45–65 % of AD patients are E4 positive (Farrer et al., 1997). Each E4 allele decreases the age of AD-onset by approximately five years (Roses, 1994). In the past two decades, research into the role of ApoE in AD increased exponentially, driven especially by numerous failures of Aβ-targeting clinical trials (Panza et al., 2019). A complete understanding of the pathological mechanism of E4 in AD will provide an urgently needed alternative research strategy for the discovery of druggable targets.

In the following sections I introduce the three major molecular players in AD: amyloid-β, microtubule-binding protein tau, and ApoE. I then highlight some important mechanisms by which E4 contributes to AD.

Amyloid-β and hyperphosphorylated tau

In the late 19^th century, the German physician Alois Alzheimer was confronted with the 51-year old patient Auguste D. who suffered from profound memory loss, confusion, and irritability. Today, Auguste D. would have been diagnosed with EOAD. Following her death in 1906 at the age of 56, Alois Alzheimer examined her brain. Besides neuronal cell death and massive loss of neuronal tissue, he observed (1) abnormal deposits around neurons, which are today known to be plaque depositions of accumulated Aβ, and (2) fibrillary tangles inside neuronal cell bodies, caused by hyperphosphorylation and accumulation of tau. To this day, these occurrences are still known as the major pathological hallmarks of AD.

Amyloid-β, the major component of extracellular plaques, is a proteolytic fragment of the transmembrane protein APP. Aβ is highly prone to self-assembly and forms soluble oligomers and fibers, ultimately accumulating in solid extracellular deposits. It wasn’t until the end of the 1990 s that scientists discovered that the soluble Aβ-oligomers, rather than monomers or plaques, are neurotoxic (Arriagada et al., 1992). The famous “Nun study” demonstrated that even huge amounts of plaque deposits do not necessarily cause cognitive decline (Snowdon et al., 1997). Today it is considered that plaques entrap toxic material to protect the brain, explaining the failure of plaque-targeting drugs in clinical studies – solubilized plaques release toxic material.

Tau is a microtubule-binding and stabilizing protein primarily expressed in neurons. Hyperphosphorylation of tau results in microtubule-dissociation, translocation to the cell body and dendrites, and aggregation into neurofibrillary tangles (Wischik et al., 1996). Seeds of accumulated tau can be transmitted from one neuron to another, comparable to an infection. Braak and colleagues described tau seeding originating in the entorhinal cortex, the main interface between hippocampus and cortex. From there, seeding proceeds along axons of the perforant pathway to the hippocampus, a region critical for memory formation. In the final stages, tau seeds reach the neocortex, where long-term memories are stored (Braak et al., 1993). The spread of tau deposition matches the brain networks responsible for the cognitive functions that decline in AD. For instance, mild cognitive impairment is associated with neuronal death in the entorhinal cortex.

How do Aβ and tau act together? Soluble forms of Aβ accumulate into plaques and tau into tangles. George Bloom described Aβ as the trigger and tau as the bullet (Bloom, 2014). More specifically, upstream Aβ triggers the conversion of tau from a normal to a toxic state, which then enhances Aβ toxicity in a feedback loop that accelerates AD pathology. Due to the self-propagation of soluble Aβ and tau species, the disease spreads through the brain in a prion-like fashion. AD pathology is thought to start at least 20 years before symptoms arise.

ApoE-isoforms and lipid transport in the brain and periphery

Apolipoproteins transport lipids, such as cholesterol and fats, which make up the major components of the cell membrane, and deliver their cargo to cells by ligand-induced receptor endocytosis. For cellular uptake, ApoE binds to members of the low-density lipoprotein receptor (LDLR) related protein (LRP) family. ApoE is expressed in several tissues – liver hepatocytes are the main peripheral source, with the majority of ApoE in the brain being secreted by astrocytes. ApoE is the main apolipoprotein in the nervous system, where cholesterol plays an important role in membrane fluidity, vesicle formation, synaptogenesis, and repair. The human brain contains 25 % of the body’s total cholesterol, which it must produce locally due to the difficulty cholesterol molecules face in crossing the blood-brain barrier (BBB). Besides AD, ApoE plays a role in cardiovascular diseases. In fact, ApoE was first described in the 1970 s as an arginine-rich, blood-cholesterol clearing protein. The different ApoE isoforms were discovered by separation of serum proteins derived from hyperlipidemia patients on a pH gradient via isoelectric focusing. The numbering of these isoforms refers to their separation based on their isoelectric point (IEP), which describes the pH at which the charge of the protein is neutral (Ordovas et al., 1987; Shore and Shore, 1969). In the general population, E2, E3, and E4 are the major alleles and have approximate allele frequencies of 8 %, 78 %, and 14 %, respectively. After the discovery of cholesterol lowering statins, ApoE-research stagnated. In the 1990 s, ApoE gained new popularity following its detection in plaques in the brains of AD patients, and after the discovery that E4 dramatically increases the risk for AD, whereas E2 decreases the risk (Nagy et al., 1995; Strittmatter et al., 1993).

Evolutionarily, E4 is the oldest isoform and carries arginines at the amino acid positions 112 (Arg-112) and 158 (Arg-158). The most common allele, E3, evolved about 200,000 years ago via an arginine to cysteine substitution at position 112 (Arg112Cys). The youngest isoform, E2, evolved from E3 about 80,000 years ago via an Arg158Cys substitution (Huebbe and Rimbach, 2017). The two polymorphisms alter the molecular structure, lipidation, receptor binding, degradation, and toxicity of the protein. Overall, ApoE contributes to coronary artery disease, myocardial infarction, and AD in the same isoform-specific stepwise pattern – from highest to lowest contribution: E4 > E3 > E2. Until now, E4 has been shown to be the greatest genetic risk factor for AD and the ApoE gene ranks fifth among the most studied human genes (Dolgin, 2017). Interestingly, Alois Alzheimer himself described “adipose inclusions”, indicating a defect in lipid metabolism. To date, E4 has been described as contributing to AD in a multitude of different ways, including through peripheral and central pathways. In the following section I will focus on a selection of mechanisms by which E4 contributes to AD pathology.

Cholesterol metabolism and Aβ-clearance

In the periphery, ApoE is present in “good” high-density lipoprotein cholesterol (HDL), which is capable of removing lipids for degradation, but not in “bad” low-density lipoproteins (LDL). In contrast to E2 and E3, E4 is poorly lipidated, which leads to different HDL/LDL ratios in people according to their ApoE genotype (Bennet et al., 2007). Peripheral circulating HDL particles are capable of traversing the BBB via ApoA-1 mediated transcytosis, thus contributing to Aβ clearance (Dal Magro et al., 2019). Importantly, the capacity of ApoE isoforms to bind to Aβ in the brain correlates with their lipidation efficiency in forming HDL-like particles: E2 > E3 > E4 (Strittmatter et al., 1993). Studies on Aβ-overproducing AD mouse models suggest that the E4-genotype and ApoE deficiency promote Aβ pathology to a comparable extent (Bell et al., 2012; Liu et al., 2015). Overexpression of the primary ApoE lipidator, ABCA1, increased Aβ-clearance in an AD mouse model (Wahrle et al., 2008), suggesting ApoE lipidation as a potential drug target. AD-linked single nucleotide polymorphisms have been discovered in several genes encoding various apolipoproteins and their numerous receptors. Thus, apolipoprotein metabolism became a new focus in understanding the various mechanisms of Aβ clearance via microglia, astrocytes, and neurons in the brain, as well as endothelial cells and pericytes at the BBB (Pohlkamp et al., 2017). Whereas LDLR plays an important role in Aβ clearance from the brain across the BBB (Castellano et al., 2012), the function of LRP1 in Aβ metabolism seems to be more complicated and partially conflicting (for a review, see Shinohara et al., 2017).

Microglia are the resident immune cells in the brain and provide the most important mechanism for Aβ degradation. ApoE modulates their inflammatory response in an isoform-specific manner. Specific types of activated microglia are found around plaque deposits in AD brains. Microglia express the receptor TREM2 on their surface, which represents the second greatest genetic risk factor for LOAD, after ApoE. Interestingly, ApoE binds to TREM2 (Atagi et al., 2015). This interaction is potentially involved in a process that puts microglia in a state in which they phagocytose Aβ-particles (Shi and Holtzman, 2018). ApoE was also described as a checkpoint inhibitor of unresolvable inflammation in response to Aβ plaques (Yin et al., 2019). However, the precise mechanism – describing, for example, how E4 would alter this microglial response – is not understood. Recently it has been found that ApoE isoforms differentially regulate the transcriptome of brain cells, particularly those of microglia and astrocytes, with consequences for the expression of genes regulating inflammation and lipid metabolism (TCW et al., 2019). This and other recent studies stress that ApoE-isoform-specific functions in cholesterol metabolism are involved in AD pathology.

ApoE4 accelerates tauopathy

Stressed neurons express ApoE, and the E4 isoform in particular undergoes enhanced proteolysis to neurotoxic fragments that stimulate tau hyperphosphorylation under these conditions (Brecht et al., 2004). Mutations in tau leading to hyperphosphorylation cause Frontotemporal Dementia (FTD) with tauopathy. In a tau-mutant FTD mouse model, ApoE-deficiency had a protective effect, whereas the E4-genotype accelerated neurodegeneration, neuroinflammation, and tau propagation (Shi et al., 2017). In agreement with this, in human FTD patients with tau mutations, the E4-genotype decreased the age of disease onset (Koriath et al., 2019). Additionally, tau pathology is strongly associated with chronic inflammatory processes, particularly activation of microglia involving ApoE and TREM2 (Keren-Shaul et al., 2017; Krasemann et al., 2017; Shi and Holtzman, 2018). However, research into the effect of ApoE on tauopathy is at an early stage.

ApoE4 causes an endosomal traffic jam in neurons

The different amounts of positively charged arginines in ApoE isoforms affect their net charge, and thereby their IEP follows the order E2 (5.9) < E3 (6.2) < E4 (6.4). At its IEP, a protein is uncharged, becomes hydrophobic, and self-assembles. After cellular uptake, ApoE enters an intravesicular (endosomal) sorting machinery in which the endosomal lumen undergoes gradual acidification. Luminal pH is critical for endosomal function. ApoE binds to its receptor via the interaction of domains that are oppositely charged. In early endosomes, increasing amounts of protons intervene in the binding of ligand to receptor at pH 6.4, causing dissociation, which is required for re-expression of the receptor at the surface and ligand re-secretion (Van der Horst et al., 2009). Further acidification in late endosomes and lysosomes assists in the sorting and degradation of biomolecules. ApoE containing endosomes have the propensity to convert into recycling endosomes that stay at the periphery (Heeren et al., 2006) and do not experience further acidification. Notably, E4 has the most basic IEP (6.4) of the three isoforms. Recent data indicate that the congruence of the IEP of E4 and the early endosomal pH causes E4 to accumulate, leading to its intracellular entrapment, along with its receptor (ApoER2/LRP8) and glutamate receptors relevant for synaptic function and plasticity. Endosomal acidification attenuates E4 mediated defects in synaptic plasticity, thus endosomal pH provides a novel drug target (Xian et al., 2018).

ApoE4 and HSV-1 as partners in crime

Recently, the link between the very prevalent herpes simplex virus 1 (HSV-1) infection and AD has become one of huge interest. As early as 1995 it was reported that E4-frequency is increased not only among AD patients but also HSV-1 infected individuals suffering from cold sores. Moreover, the combination of HSV-1 infection and the E4-genotype has been suggested to cause AD, whereas either of these features alone has not (Lin et al., 1996). More recent research indicates that E4 facilitates HSV-1 endocytosis and infection in the brain to a higher degree than E3 does (Burgos et al., 2006). Interestingly, cell membrane cholesterol plays a key role in HSV-1 entry, infection, replication, and cell-to-cell spread (Wudiri and Nicola, 2017).

Besides that, heparan sulfate proteoglycans (HPSGs), polysaccharides that decorate cell surface and secretion proteins, serve as receptors for HSV-1 particles. Moreover, HSPGs are receptors for ApoE, and recombinant ApoE fragments have been used to reduce viral infection, including that caused by HSV-1 (Dobson et al., 2006; Tudorache et al., 2017). HSPGs also bind Aβ and tau, and are enriched in plaques and neurofibrillary tangles (Holmes et al., 2013; Zhang et al., 2014). Recent findings credit Aβ with potent antimicrobial properties as it may entrap pathogens like HSV-1 in plaques (Eimer et al., 2018). HSPGs have also been implicated in the propagation of tau species from neuron to neuron (Katsinelos et al., 2018). Antiviral drugs for herpes have been shown to reduce Aβ aggregation and tau hyperphosphorylation in vitro. In line with this, treatment of humans with the respective drugs is associated with a decrease in the incidence of AD (Qin and Li, 2019).

Taken together, the AD risk factor ApoE4 is linked to the accumulation of both Aβ and tau, the major pathological hallmarks of AD, as well as inflammatory responses in neurodegeneration. Numerous findings indicate that the basic function of ApoE as a lipid transporter is responsible for this. Accordingly, E4 lipidation and improving its trafficking through the endosomal system attenuates AD-relevant impairments. Thus, improving the main functions of E4 is currently the focus of different drug development strategies for AD prevention and treatment in E4 carriers.