Memory loss caused by β-amyloid protein is rescued by a β₃-adrenoceptor agonist

Abstract

Accumulation of the neurotoxic β-amyloid protein (Aβ) in the brain is a key step in the pathogenesis of Alzheimer's disease (AD). Although transgenic mouse models of AD have been developed, there is a clear need for a validated animal model of Aβ-induced amnesia which can be used for toxicity testing and drug development. Intracranial injections of Aβ_1–42 impaired memory in a single trial discriminative avoidance learning task in chicks. Memory inhibition was closely associated with the state of aggregation of the Aβ peptide, and a scrambled-sequence of Aβ_1–42 peptide failed to impair memory. Aβ had little effect on labile (short-term and intermediate) memory, but blocked consolidation of memory into long-term storage mimicking the type of anterograde amnesia that occurs in early AD. Since noradrenaline exerts a modulatory influence on labile memory in the chick, we examined the effects of two β-adrenoceptor (AR) agonists on Aβ-induced amnesia. A β₃-AR agonist (CL316243), but not a β₂-AR agonist, rescued Aβ-induced memory loss, suggesting the need for further studies on the role of β₃-ARs in AD.

Introduction

Alzheimer's disease (AD) is the most common cause of dementia in the elderly. Typically the disease is characterized in its earliest stages by memory problems. Initially patients may have trouble remembering recent events, but as the disease spreads, older and more established memories are lost. Wandering and disorientation occur and as the disease progresses, the symptoms worsen (Storey et al., 2001).

It is now recognized that AD is caused by a build-up of neurotoxic Aβ in the brain (Small et al., 2001, Walsh and Selkoe, 2004). Aβ peptides, particularly the longer species such as Aβ_1–42, are considered to be the major culprits in disease pathogenesis. Aβ_1–42 aggregates more readily than the more commonly produced Aβ_1–40 (Jarrett and Lansbury, 1993). Aβ_1–42 can aggregate to form oligomers, protofibrils that ultimately lead to the formation of amyloid plaques. However, recent studies suggest that it is the oligomeric or low molecular weight protofibrillar Aβ species, rather than the amyloid plaques, that are the most neurotoxic (Klein et al., 2001, Walsh and Selkoe, 2004).

While much of the focus in the field of AD has been on the chronic lesions that characterize the disease (amyloid plaques, neurofibrillary tangles, gliosis, cell death), it is increasingly recognized that Aβ may exert effects that are acute and independent of long-term chronic neurodegeneration (Palop et al., 2006). For example, oligomeric Aβ can rapidly alter calcium homeostasis and disrupt long-term potentiation (LTP) (Klein et al., 2001, Walsh and Selkoe, 2004). Thus increasingly it is thought that the acute effects of Aβ on synaptic plasticity may be a more important contributor to cognitive decline than the neuropathological features such as amyloid plaques, neurofibrillary tangles and cell loss (Palop et al., 2006).

The development of therapeutic drugs to treat AD must rely heavily upon models that can be used to test Aβ toxicity. Early biological studies used assays of cell death or mitochondrial activity (Koh et al., 1990, Shearman et al., 1994) to evaluate therapeutic potential. However, the relevance of these assays for therapy is unclear, as agents that block neurotoxicity in cell culture have not yet been found to have efficacy in clinical trials. For example, although antioxidants block the neurotoxic effect of Aβ in cell culture (Subramaniam et al., 1998), they do not have a clear effect in clinical trials (Boothby and Doering, 2005, Tabet et al., 2000). Ultimately, a validated in vivo model of Aβ toxicity is needed. While several APP transgenic mouse models of AD have been developed (German and Eisch, 2004, van Dooren et al., 2005), these animal models also have their problems. For example, although behavioural deficits occur in APP transgenic mice (Morgan, 2003), these deficits are subtle and can be difficult to quantify.

Several studies have examined the effect of exogenous Aβ peptides on memory. Direct injection of Aβ peptides into the brain has been reported to block memory in rodents (Townsend et al., 2006a, Townsend et al., 2006b, Townsend et al., 2007), but behavioural assays in mammals can be difficult to carry out. In contrast, chicks provide a relatively inexpensive and rapid method for the assessment of memory. The cortical and pallial regions in the mammal and the bird derive from the same pallial regions in the embryo (Jarvis et al., 2005). Brain regions involved in memory processing and memory mechanisms in birds are the same as those in mammals, for example, avian NMDA receptors play an important role in memory and synaptic plasticity as in mammals (Gibbs et al., 2008, Rickard et al., 1994) and protein synthesis is involved in both mammalian and avian memory (Izquierdo et al., 2006, Matthies, 1989, Rose, 2000). Mileusnic et al. (2007) and Mileusnic et al. (2004) previously reported that intracranial injection of Aβ into the chick blocks memory in a passive avoidance task. However, these studies lacked a suitable non-amyloidogenic peptide control needed to assess the validity of the memory effects and it is unclear in these studies whether effects on chick memory were caused by specific Aβ toxicity or whether a non-specific action was involved.

The development of a rapid, accurate and well-validated model of Aβ-induced cognitive dysfunction would be a major benefit for drug screening. While much attention has been focussed on cholinergic drugs, it is increasingly recognized that many other neurotransmitter systems are involved in the cognitive decline that occurs in AD. Noradrenaline is of particular interest as it plays an important role in modulating memory. In non-human primates and rats, working memory, a specialized form of short-term memory and the consolidation of long-term memory are regulated by noradrenaline (Berridge and Waterhouse, 2003, Ramos and Arnsten, 2007). In chicks, short-term memory (Gibbs and Summers, 2005) and the consolidation of intermediate into long-term memory triggered 30 min after training are also regulated by noradrenaline (Gibbs and Summers, 2002a, Gibbs and Summers, 2002b). In AD patients, noradrenaline levels are substantially reduced, suggesting a role for noradrenaline in the cognitive decline. The locus coeruleus, a structure that innervates the prefrontal cortex and the major source of noradrenaline in the forebrain, is affected in AD brains (Hertz, 1989, Szot et al., 2006).

In the present study, we have used chicks and a single trial discriminative avoidance task to examine the effects of Aβ on memory. We show that this model of Aβ-induced amnesia is a rapid, reproducible and highly quantifiable method for assessing Aβ neurotoxicity. We show that the effect of Aβ is specific, as a scrambled-sequence Aβ peptide does not block memory, and that there is a correlation between the state of Aβ aggregation and its effect on memory.

As noradrenaline has been implicated in memory processing, we have used our model (Gibbs and Summers, 2002b, Gibbs, 2008) to examine the effect of adrenoceptor agonists on Aβ-induced amnesia in chicks. We report that a β₃-adrenoceptor (AR) agonist (but not a β₂-AR agonist) blocks the amnestic effects of Aβ_1–42 peptides. As β₃-AR agonists have not yet been tested for their efficacy in clinical trials of AD, our studies suggest that further studies on the role of the β₃-AR in AD are needed.