It is clear from above that many cell processes are disrupted in AD and they are initiated by different events. Consequently, any hypothesis that explains AD has to incorporate the idea of multiple primary lesions to a possible single mode of pathogenesis. The Amyloid Cascade Hypothesis was proposed in 1991 by John Hardy and David Allsop (reviewed by Tanzi and Bertram 2005 ). This hypothesis suggested that the mismetabolism of APP was the initiating event in AD pathogenesis, subsequently leading to the aggregation of Aβ, specifically Aβ42. Formation of neuritic plaques would instigate further pathological events, including the formation of NFTs, NTs and disruption of synaptic connections, which would lead to a reduction in neurotransmitters, death of tangle-bearing neurons and dementia (Hardy and Allsop 1991 ).
Since Aβ was first identified in 1984 as the main component of amyloid plaques (Glenner and Wong 1984 ), evidence has accumulated to suggest that this peptide is indeed the primary neuropathological insult in AD. Central to this hypothesis is the observation that the vast majority of mutations causing familial EOAD increase the ratio of fibrillogenic A42 (Tanzi and Bertram 2005 ). In addition, transgenic mice models expressing pathogenic mutations of APP (Hsiao et al 1996 ) and PS1 have increased levels of Aβ and amyloid plaques. Furthermore, individuals with trisomy 21 (Down’s syndrome, DS) have 3 copies of APP and develop advanced AD usually within the fourth decade of life.
Consistent with this hypothesis, large amounts of Aβ is neurotoxic to cells (Goodman and Mattson 1994 ). Aβ may exert its neurotoxic effects in a variety of ways, including disruption of mitochondrial function via binding of the Aβ-binding alcohol dehydrogenase (ABAD) protein (Lustbader et al 2004 ), induction of apoptotic genes through inhibition of Wnt (reviewed by Caricasole et al 2003 ) and insulin signalling (Xie et al 2002 ), formation of ion channels (reviewed by Kagan et al 2002 ) triggering loss of calcium homeostasis (Goodman and Mattson 1994), stimulation of the JNK/SAPK pathway (Kim et al 2003 ) or activation of microglia cells leading to the expression of pro-inflammatory genes, an increase in reactive oxygen species, and eventual neuronal toxicity and death (reviewed by Bamberger and Landreth 2001 ; Dugue et al 2003).
In normal metabolism, Aβ levels are tightly regulated by amyloid degrading-enzymes, which include the insulin-degrading enzyme (IDE) and neprilysin. These proteases degrade specific types of Aβ: neprilysin and IDE degrade Aβ monomers, while other proteases cleave the fibrillogenic and oligomeric types. In addition to degradation, extracellular Aβ can be cleared and exported, via binding of A2M and APOE to LRP, and subsequent degradation via the lysosomal pathway. The Aβ peptide can also be removed from the brain and into the periphery by LRP mediated endocytosis, where it is then degraded at peripheral sites. Endosomal dysfunction is one of the earliest known disease-specific pathologies of sporadic AD and abnormally enlarged early endosomes coincide with early rises in soluble Aβ that precedes deposition of Aβ (Cataldo et al 2004 ).
Aβ42 accumulates in the brains of sporadic AD patients and the only known risk factor for LOAD is APOE ε4, which as discussed above, affects Aβ aggregation. In addition, genetic studies of LOAD have observed positive associations with other genes involved in APP processing, degradation and clearance - A2M, BACE1, IDE, LRP1, NCT, NEP and PLAU - although the results have not been consistent (reviewed by Tanzi and Bertram 2005).
Taken together, these data indicate that although the mechanism for increased Aβ differs between familial and sporadic AD, increased levels of Aβ can initiate this devastating disease (reviewed by Hardy 1997 ). Furthermore, Aβ deposition may take decades to develop and may be influenced by lifelong patterns of Aβ clearance (reviewed by Friedland 2002 ).
However, although some studies have observed a strong correlation between dementia and senile plaques (Cummings et al 1996 ), others have identified only a weak correlation (Dickson et al 1995 ). In contrast, strong correlations have been observed with the number of tangles (Arriagada et al 1992 ) and studies examining the temporal relationship between plaques and tangles observe that tangles appear first (Schonheit et al 2004 ). Nonetheless, increasingly, attention has turned towards soluble, oligomeric and even intracellular Aβ42, rather than insoluble forms of Aβ (reviewed by Bloom et al 2005 ). Soluble Aβ levels correlate more strongly with dementia severity (McLean et al 1999 ), indeed although plaques have been observed in the brain of a non-demented individual, there were no amyloid-β diffusible ligands (ADDLs) in the CSF (Georganopoulou et al 2004 ). Furthermore, treatment with insulin improves cognition and decreases intracellular Aβ levels (Gasparini et al 2001 ) and a recent study observed that neuronal cells not only accumulated intraneuronal Aβ42, but that the earliest Aβ42 immunoreactive SPs developed along the projections and at terminals of early Aβ42 accumulating neurons (Gouras et al 2000 ). Transgenic animal studies have provided further evidence supporting this (reviewed by Gotz et al 2004 ). Soluble Aβ levels alter hippocampal synaptic efficiency in rats (Walsh et al 2002 ), while increased expression of Aβ42 in Drosophila induces formation of diffuse amyloid deposits, age-dependent learning defects and extensive neurodegeneration (Iijima et al 2004 ). In addition, cognitive defects in mice are caused by small oligomeric Aβ assemblies (Westerman et al 2002 Billings et al 2005 ), before Aβ deposition (Arendash et al 2004 ). Compelling evidence for the amyloid cascade hypothesis was provided with the observation that Aβ immunotherapy decreased Aβ levels in mice and this rescued cognitive deficits (Oddo et al 2004 Billings et al 2005 ). Furthermore, amyloid-containing brain regions have decreased expression of genes important in long term potentiation (LTP) and memory consolidation (Dickey et al 2003 ).
One tenet of this hypothesis is that plaques precede tangles. Observations from autopsied AD brains support this, as some AD patients have many neocortical plaques but no or few tangles. In addition, mutations in tau cause FTDP-17, which is characterised by the severe deposition of tau in NFTs, but no amyloid deposition (reviewed by Hardy and Selkoe 2002 Furthermore, recent transgenic mouse studies have shown that Aβ influences formation of tau pathology (reviewed by Bloom et al 2005; Lewis et al 2001 Gotz et al 2001 Oddo et al 2003 ) and age of disease onset (Oddo et al 2003). Aβ can influence tau pathology by affecting tau-kinases, through inhibition of insulin and wnt signalling, which negatively regulate GSK3β activity (reviewed by Caricasole et al 2003 Gasparini et al 2001 ), through influencing p53 levels, which activates GSK3 (reviewed by Jope and Johnson 2004 ) and through activation of Cdk5 (Cruz and Tsai 2004 ). Both AD pathologies can further be related by deficiencies in axonal transport. As this influences plasticity, a dysfunction in axonal transport may contribute to pathological development. APP is a potential receptor for the kinesin light chain (KLC) (Kamal et al 2000 ), while tau inhibits kinesin dependent transport (Stamer et al 2002 ). Transport deficiencies leads to the selective increase in Aβ42 peptides (Stokin et al 2005 ) and elevated levels of tau are observed in degenerating neurons (Stamer et al 2002). Although recent work disputes a transport role for APP (Lazarov et al 2005 ), in vivo data from D. melanogaster provides further support (Gunawardena and Goldstein 2001 ). However, an alternative mechanism could involve dysregulation of GSK3β. KLCs are in vivo substrates for GSK3. Phosphorylation of KLCs reduces the levels of kinesin-1 mediated fast axonal transport (Morfini et al 2002 ), as observed in PS1 mutant cells (Pigino et al 2003). Either mechanism is consistent with the amyloid cascade hypothesis and would disrupt axonal transport. Initial disruption may lead to an autocatalytic spiral where axonal blockages and Aβ generation are mutually stimulatory (Stokin et al 2005). Subsequently, tau hyperphosphorylation may be the response of the cell to try and restore normal axonal transport (Stamer et al 2002).
The spread of AD pathology can be mediated by soluble extracellular Aβ that induces apoptosis and tau hyperphosphorylation in surrounding cells. Furthermore, extracellular Aβ can increase de novo production of intracellular species (Yang et al 1999 ), and amyloid deposits may be reservoirs of Aβ peptides that can diffuse and cause tau pathology over the course of many years (reviewed by Bloom et al 2005).
In conclusion, the majority of the evidence at present suggests that intracellular Aβ is the initiating lesion in the development of Alzheimer’s disease. Dysregulation of APP processing leads to increased levels of Aβ which induces cell death processes and disturbed axonal transport, leading to defects in cognition, synaptic plasticity and development of tau pathology. In addition, these effects can be created by disturbances in diverse signalling pathways such as Notch, Wnt and Insulin, and therefore the pathogenesis of AD is likely to be extremely complex and with many contributing factors that enhance pathology.
Click to read this topic
6/6/2012 2:43:17 PM - 220.127.116.11