Soluble amyloid-β oligomers as synaptotoxins leading to cognitive impairment in Alzheimer's disease

Abstract

Alzheimer's disease (AD) is the most common form of dementia in the elderly, and affects millions of people worldwide. As the number of AD cases continues to increase in both developed and developing countries, finding therapies that effectively halt or reverse disease progression constitutes a major research and public health challenge. Since the identification of the amyloid-β peptide (Aβ) as the major component of the amyloid plaques that are characteristically found in AD brains, a major effort has aimed to determine whether and how Aβ leads to memory loss and cognitive impairment. A large body of evidence accumulated in the past 15 years supports a pivotal role of soluble Aβ oligomers (AβOs) in synapse failure and neuronal dysfunction in AD. Nonetheless, a number of basic questions, including the exact molecular composition of the synaptotoxic oligomers, the identity of the receptor(s) to which they bind, and the signaling pathways that ultimately lead to synapse failure, remain to be definitively answered. Here, we discuss recent advances that have illuminated our understanding of the chemical nature of the toxic species and the deleterious impact they have on synapses, and have culminated in the proposal of an Aβ oligomer hypothesis for Alzheimer's pathogenesis. We also highlight outstanding questions and challenges in AD research that should be addressed to allow translation of research findings into effective AD therapies.

Keywords: Alzheimer’s disease; amyloid-β oligomers; memory loss; neuronal dysfunction; synapse failure.

Figure 1

Synapse under siege. A simplified scheme illustrating some of the mechanisms by which AβOs (AβOs; represented as red asterisks) impact synapse plasticity and function. At the pre-synaptic terminal, AβOs inhibit microtubule-based fast axonal transport (FAT) in a tau-independent manner (Decker et al., ; Ramser et al., ; Gan and Silverman, ; Takach et al., 2015), resulting in reduced or interrupted transport of various cargoes, including brain-derived neurotrophic factor (BDNF), to synapses. FAT inhibition may involve physical or functional interaction of AβOs with axonal voltage-gated calcium channels (VGCC), calcium influx into the axon and activation of calcineurin (CaN). Deregulated calcium levels also triggers increased pre-synaptic glutamate release (Brito-Moreira et al., 2011). Release of the NMDA receptor (NMDAR) co-agonist, D-serine, is also increased in neurons exposed to AβOs (Brito-Moreira et al., ; Madeira et al., 2015), resulting in elevated basal excitatory tonus. At the dendritic spine, AβOs aberrantly activate NMDARs, triggering increased calcium influx and increased production of reactive oxygen species (ROS; De Felice et al., ; Decker et al., ; Saraiva et al., ; Paula-Lima et al., 2011), likely mediated by mitochondrial dysfunction. NMDAR-mediated calcium influx causes ryanodyne receptor-mediated calcium release from endoplasmic reticulum (ER) stores (Paula-Lima et al., 2011). Elevated calcium levels at the spine activate CaN to dephosphorylate NMDARs and AMPARs at specific phosphoepitopes, causing their removal from synapses and internalization (Snyder et al., ; Hsieh et al., ; Jürgensen et al., 2011). CaN activation has also been implicated in Aβ-induced spine loss (Wu et al., ; not represented in the scheme for simplicity). Calcium-dependent activation of calcium/calmodulin-dependent kinase II (CaMKII) and casein kinase II (CKII) also appears to mediate removal of NMDARs from synapses (De Felice et al., ; not represented in the current scheme). AβOs block insulin signaling by instigating removal of insulin receptors (IRs) from the neuronal plasma membrane (De Felice et al., ; not represented here), inhibiting IR autophosphorylation (Zhao et al., ; not represented), and by inhibiting IR-mediated tyrosine phosphorylation of the insulin receptor substrate-1 (IRS-1; Bomfim et al., ; not represented). AβOs instigate brain microgliosis and increased microglial production/secretion of TNF-α (Ledo et al., ; unpublished results). Aberrant TNF-α signaling activates cell stress kinases (c-Jun N-terminal kinase, JNK; IκB kinase, IKK; double-stranded DNA-dependent protein kinase, PKR; Bomfim et al., ; Lourenco et al., 2013), resulting in serine phosphorylation and further inhibition of IRS-1. PKR also phosphorylates eukaryotic initiation factor 2 α (eIF2α; Lourenco et al., 2013). eIF2α is further phosphorylated by PKR-related ER resident kinase (PERK), a kinase that executes one of the three branches of the unfolded protein response (UPR) that is activated by ER stress in neurons exposed to AβOs (Lourenco et al., ; Ma et al., 2013). Insulin signaling inhibition and eIF2α-P lead to blockade of protein synthesis, essential for synaptic plasticity (Buffington et al., 2014). Additional mechanisms (not included here) thought to further contribute to synapse failure include deregulation of the actin cytoskeleton at spines (Lacor et al., ; Davis et al., 2011), loss of pre- and post-synaptic proteins (e.g., synaptophysin, PSD-95; Roselli et al., ; Sebollela et al., ; Figueiredo et al., 2013), Fyn kinase-mediated surface removal of NMDARs and spine loss (Um et al., 2012), among others.