Acute amnestic encephalopathy in amyloid-β oligomer–injected mice is due to their widespread diffusion in vivo

Abstract

Amyloid-β (Aβ) oligomers are the suspected culprit as initiators of Alzheimer's disease (AD). However, their diffusion in the brain remains unknown. Here, we studied Aβ oligomers' dissemination and evaluated their in vivo toxicity. Wild-type mice were injected with 50 pmol of synthetic Aβ oligomers (of different size) in the hippocampus. Oligomers diffused largely in the brain as soon as 1 hour and up to 7 days after injection. A transient encephalopathy with memory impairment was induced by this unique injection. The immunoreactivity of the postsynaptic marker PSD95 was diffusely decreased. Similar results (both on memory and PSD95 immunoreactivity) were obtained with delipidated and high molecular weight oligomers (>50 kDa) but not with smaller assemblies. Tau hyperphosphorylation was observed in the oligomer-injected brains. Finally, fos immunostaining was increased in Aβ-derived diffusible ligands–injected mice, suggesting neuronal hyperactivity. Rapid and widespread diffusion of Aβ oligomers was demonstrated in vivo and associated with decreased synaptic markers and memory deficits which gives new insight to the pathogenicity of Aβ.

Introduction

Diagnosis of Alzheimer's disease (AD) relies on 2 hallmark lesions: extracellular deposits of beta-amyloid peptides (Aβ deposits) and intraneuronal aggregates of hyperphosphorylated tau protein (neurofibrillary tangles) (Duyckaerts et al., 2009). A current physiopathological hypothesis states that the overproduction or insufficient clearance of Aβ is the event that induces a cascade of secondary brain alterations (synaptic damage, inflammation, tau hyperphosphorylation, and neuronal death) leading to dementia (Hardy and Higgins, 1992, Hardy and Selkoe, 2002,). This sequence of events has, however, never been reproduced by injecting fibrillar or monomeric Aβ peptides in the brain of animals (Games et al., 1992); moreover, the failure of recent therapeutic trials in which a reduction of the amyloid load did not modify the progression of the symptoms does not support the amyloid cascade hypothesis (Doody et al., 2014, Gilman et al., 2005, Holmes et al., 2008, Salloway et al., 2014). One hypothesis to explain these failures could be that the toxic Aβ species have not been specifically targeted (Goure et al., 2014). Experimental data from the last decade suggest indeed a revision of the model. Numerous reports show that the toxic pool of Aβ is not made of extracellular aggregates but of soluble oligomeric molecular assemblies (Selkoe, 2008). The failure of the cascade hypothesis model could hence be related to its emphasis on the concentration of Aβ rather than on the aggregation state of the peptide; hence, fibrils of Aβ are more toxic than monomers but less than oligomers in vitro (Chromy et al., 2003). Oligomers are aggregates of Aβ that remain soluble and show an exquisite synaptotoxicity. Solubility-diffusibility and/or active transport are required to explain the spreading of Aβ toxicity but have not yet been thoroughly analyzed in vivo.

Synthetic Aβ oligomers such as Aβ-derived diffusible ligands (ADDLs) have been shown to be 10 times more neurotoxic in vitro than fibrillar Aβ and 50 times more toxic than monomeric Aβ (Lambert et al., 1998, Stine et al., 2003). ADDLs colocalize with PSD95, a marker of postsynaptic densities (Lacor et al., 2004). They alter the shape, composition, and density of synapses (Lacor et al., 2007). They modify mGluR5 receptor clustering and lateral diffusion at the cell membrane triggering an intracellular calcium influx and causing drastic synaptic changes (Renner et al., 2010). Only few studies have investigated the effects of in vivo administration of Aβ oligomers. Injections of different types of Aβ oligomers in the lateral ventricles acutely induce marked memory impairments (Balducci et al., 2010, Cleary et al., 2005, Garcia et al., 2010, Youssef et al., 2008). Unilateral injection restricted to the hippocampus is sufficient to induce a general deficit in memory (Moon et al., 2011, Pearson-Leary and McNay, 2012) while a localized lesion of the hippocampus should not cause such impairments (Moser et al., 1995). A diffusion of the Aβ oligomers or a widespread reaction to its local injection could explain the severity of the symptoms. It has been hypothesized that induced memory impairments are tightly associated with the synaptotoxic effects of oligomers (Selkoe, 2008). Widespread synaptic dysfunction is nonetheless required to promote behavioral effects. This implies large diffusion of Aβ oligomers through the brain parenchyma but in vivo evidence is currently lacking.

In the present study, Aβ oligomers' diffusion and toxicity were studied after injection of various types of ADDLs in wild-type mice. The consequences of a unique injection in the hippocampal formation (hippocampus or adjacent rhinal cortex) were examined. As lipids have been suggested to act as cofactors for Aβ-induced toxicity (Hong et al., 2014, Martins et al., 2008), both regular (synthesized in lipid containing medium) and delipidated ADDLs were tested for their toxicity. The diffusion of Aβ species in brain parenchyma was assessed by different methods (histological mapping and imaging mass spectrometry–IMS [Chaurand, 2012, Cornett et al., 2007]). Local and distant synaptotoxic effects were quantified as well as other neuropathological alteration (tau phosphorylation, changes in evoked neuronal activity). Consequences of ADDLs injection on learning and memory tasks were also analyzed.