Mechanisms of the inhibitory effects of amyloid β-protein on synaptic plasticity

Abstract

Alzheimer's disease can be considered a protein misfolding disease. In particular, inappropriate processing of a proteolytic fragment of amyloid precursor protein, amyloid β-protein (Aβ), in early stages of Alzheimer's disease may lead to stabilization of small oligomers that are highly mobile and have a potential to be extremely toxic assemblies. Recently, the importance of such soluble species of Aβ in triggering synaptic dysfunction, long before neuronal loss occurs, has become apparent. Animal models have revealed that plasticity of hippocampal excitatory synaptic transmission is relatively selectively vulnerable to Aβ both in vitro and in vivo. This review focuses on the mechanisms of Aβ inhibition of long-term potentiation at synapses in the rodent hippocampus from two complimentary perspectives. Firstly, we examine evidence that the synaptic activity of this peptide resides primarily in oligomeric rather than monomeric or fibrillar Aβ species. Secondly, the importance of different oxidative/nitrosative stress-linked cascades including JNK, p38 MAPK and NADPH oxidase/iNOS-generated reactive oxygen/nitrogen free radicals in mediating the inhibition of LTP by Aβ is emphasised. These mechanistic studies provide a plausible explanation for the sensitivity of hippocampus-dependent memory to impairment in the early preclinical stages of Alzheimer's disease.

Introduction

There is now extensive evidence that many neurodegenerative disorders, including Alzheimer's disease (AD), may form a subgroup of protein misfolding diseases (Dobson, 2002, Caughey and Lansbury, 2003). In particular, in the early stages of AD small oligomers of a proteolytic fragment of amyloid precursor protein (APP), amyloid β-protein (Aβ), may be highly mobile and have a potential to be extremely toxic assemblies (Klein et al., 2001, Selkoe, 2002). Increased production or reduced clearance of such small diffusible Aβ species may be responsible for neuronal dysfunction long before structural evidence of neurodegeneration is apparent in AD. Here we discuss our recent studies into how Aβ disrupts synaptic plasticity in the rodent hippocampus.

Why study hippocampal synaptic plasticity in models of early AD? Early AD, and age-related Mild Cognitive Impairment (MCI) which is a very strong predictor of clinical AD, are largely diseases of the medial temporal lobe, especially the hippocampus (Laakso, 2002). Hence most patients with early AD and by definition all people diagnosed with MCI have relatively selective deficits in hippocampal-dependent memory. There is now extensive evidence that the mechanisms underlying synaptic plasticity including long-term potentiation (LTP) mediate such memory (Morris et al., 2003). Therefore, we believe that studies of synaptic plasticity including LTP provide a key physiological readout of how memory mechanisms may be disrupted in animal models of early AD.

What causes AD? The Aβ cascade hypothesis of AD proposes that the increased levels of the 42-residue form of Aβ, Aβ42 triggers a sequence of events that lead to clinical dementia (Hardy and Selkoe, 2002, Selkoe and Schenk, 2003). A very small proportion of AD patients have a familial form of the disease where there is a mutation in certain genes, including APP and presenilins. APP is a presynaptic terminal protein with a large extracellular domain. Patients with early onset familial APP-linked AD often have autosomal dominant mutations in the APP gene in the regions coding for the amino acids near where APP is cleaved to form Aβ. Intramembrane cleavage of APP is carried out by an enzyme complex, called γ-secretase, that includes presenilin. AD-related missense mutations in APP and presenilins usually result in a relative increase in the production of Aβ42. This Aβ species is more toxic than the normal product, Aβ40. It also has a much greater tendency to oligomerize into large assemblies.

Perhaps the strongest clinical evidence supporting the Aβ hypothesis comes from a clinical trial of immunization against Aβ in AD. Based on very encouraging preclinical studies of transgenic APP-linked animal models of familial AD (Schenk et al., 1999, Janus et al., 2000, Morgan et al., 2000), patients with moderate to severe sporadic AD were immunized with Aβ42 (Hock et al., 2003, Orgogozo et al., 2003, Ferrer et al., 2004). Although the clinical trial had to be stopped because ∼6% of patients developed a life-threatening non-infectious meningoencephalitis, a recent report of thirty of these patients showed a strong relationship between the presence of anti-Aβ antibody titres and a lack of further worsening of symptoms and performance of a hippocampus-dependent task over a one-year study period (Hock et al., 2003).

Many different conformations of Aβ are present in the brains of patients with AD. Originally most research focused on the hallmark neuropathological plaques, which are primarily composed of insoluble fibrillar Aβ. However, correlations of the amount of fibrillar Aβ in the brain measured post mortem with indices of clinical dementia severity are poor (Terry, 2000). In contrast, more recent research has found that brain levels of soluble non-fibrillar Aβ correlate much more closely with clinical dementia ratings (Lue et al., 1999, McLean et al., 1999). There are several known species of soluble Aβ that can form assemblies of increasing size, including monomeric, oligomeric and protofibrillar Aβ. Our studies examined the role of the smaller ones in the synaptic actions of Aβ, focussing in particular on monomeric and low-n oligomeric assemblies, species that would be expected to predominate at the earlier stages of disease progression.