Examining the mechanisms that link β-amyloid and α-synuclein pathologies

β-amyloid (Aβ) and α-synuclein (α-syn) are aggregation-prone proteins typically associated with two distinct neurodegenerative disorders: Alzheimer's disease (AD) and Parkinson's disease. Yet α-syn was first found in association with AD plaques several years before being linked to Parkinson's disease or Lewy body formation. Nowadays, a large subset of AD patients (~50%) is well recognized to co-exhibit significant α-syn Lewy body pathology. Unfortunately, these AD Lewy body variant patients suffer from additional symptoms and an accelerated disease course. Basic research has begun to show that Aβ and α-syn may act synergistically to promote the aggregation and accumulation of each other. While the exact mechanisms by which these proteins interact remain unclear, growing evidence suggests that Aβ may drive α-syn pathology by impairing protein clearance, activating inflammation, enhancing phosphorylation, or directly promoting aggregation. This review examines the interactions between Aβ and α-syn and proposes potential mechanistic links between Aβ accumulation and α-syn pathogenesis.

show no evidence of concurrent Aβ deposition [23,24]. Aβ and α-syn thus appear to interact in a disease-specifi c and anatomical-specifi c manner. By examining the mecha nisms by which these proteins interact we will probably enhance our understanding of why Aβ and αsyn patho logies co-exist in many, but not all, AD patients.
In the present review we shall examine the evidence supporting a role for synergistic interactions between Aβ and α-syn in the development and progression of AD-LBV. We will also discuss putative mechanisms by which Aβ and α-syn could interact and infl uence disease progression.

Clinical prevalence of the Lewy body variant of Alzheimer's disease
Between 50 and 60% of AD patients exhibit signifi cant amounts of both Aβ and α-syn pathology at autopsy [3,6]. Th ese AD-LBV patients often present with a more aggressive form of dementia featuring a higher rate of cognitive decline and shortened survival versus pure AD [9,11,12,25]. After AD, DLB appears to be the second most common form of age-related dementia [3]. Clinically, pure DLB patients who lack Aβ pathology often exhibit diff erent cognitive defi cits compared with AD or other dementias. Th e core diagnostic criteria for DLB include fl uctuating cognition and attention, persistent visual halluci nations, and spontaneous Parkinsonian symptoms. DLB patients may also possess greater defi cits in working memory, attention, executive function, and visuospatial ability than AD patients [3,12]. Pure DLB cases are rela tively rare (~20%), however, and Aβ pathology is often also present [3,26,27]. In these cases of mixed pathology, clinical diag nosis can be more diffi cult as the cognitive decline more closely resembles the cognitive profi le of AD with the addition of some of the unique DLB-associated symptoms [25,26].
Interestingly, not all studies have reported such cog nitive diff erences between AD patients, AD-LBV patients, and DLB patients, perhaps as a result of varying methodology such as not including postmortem α-syn or Aβ histochemical assessments [12]. Additionally, given the diff ering cognitive profi les of AD, AD-LBV, and DLB, direct comparison of cognitive decline and the rate of decline can be diffi cult and dependent on the assessments utilized [11,25,26]. PDD patients, who can show very similar pathology and symptoms to DLB patients, can also sometimes exhibit both Aβ plaques and Lewy bodies. Although the onset of dementia in PDD patients is highly variable, it appears to be infl uenced by Aβ pathology [4,13]. Several disease states that involve α-syn pathology can thus also exhibit varying degrees of Aβ pathology.

Synergistic interactions between β-amyloid and α-synuclein
Given the intriguing overlap of Aβ and α-syn pathology that occurs in these various dementing disorders, researchers have begun to examine the interactions between these two proteins and pathologies. Both in vitro and in vivo experiments have started to identify potential mechanisms by which Aβ and α-syn may interact, providing critical results that promise to advance our understanding of these inter-related neurodegenerative diseases.

In vitro examinations
Using cell-free assays, researchers fi rst began to explore potential direct interactions between Aβ and α-syn. For example, incubation of recombinant human α-synuclein (hSYN) with Aβ 42 promoted and increased the formation of high-molecular-weight hSYN oligomers [19]. Interestingly, while Aβ 42 induced α-syn oligomer formation, coincubation with the less pathogenic Aβ 40 did not. Virtually identical results were observed in a cell culture model where extracellular Aβ 42 , but not Aβ 40 , promoted the formation of intracellular α-syn aggregates [19]. Expanding upon these results, Aβ 40 and Aβ 42 were both shown to directly interact with α-syn in vitro [20]. However, α-syn appears to induce a greater structural change in Aβ 42 . While Aβ 40 remains soluble in solution following α-syn co-incubation, Aβ 42 instead forms oligomers and insoluble precipitates [20]. Th e implication that α-syn may preferentially interact with Aβ 42 is important given the known toxic and aggregate-prone properties of Aβ 42 relative to other Aβ isoforms [28,29].
Continuing to implicate Aβ 42 as a critical component in the interactions between Aβ and α-syn is the result that a mutation in presenilin 1, which increased Aβ 42 , also enhanced the pathogenic phosphorylation and aggregation of α-syn in both patients and cells [21]. Th is result not only provides further support for the idea that Aβ 42 plays a key role in the aggregation of α-syn, but also suggests a possible mechanism; the Aβ 42 -induced phosphory lation of α-syn. In the normal brain, about 4% of αsyn is phosphorylated at serine 129 (pS129-syn). In contrast, up to 90% of α-syn is phosphorylated at this site in synucleinopathies such as DLB, suggesting an important role in pathogenesis [30,31]. Indeed, phosphorylation of α-syn at serine 129 (Ser129) can promote fi bril formation in vitro [30]. Aβ 42 -induced phosphorylation of α-syn therefore provides an intriguing mechanism by which Aβ may enhance α-syn pathology.
Th ere is considerable in vitro evidence that α-syn can also interact with tau, the other major pathological protein in AD. Whereas α-syn has been shown to selfpolymerize in vitro [32], tau instead requires cofactors to polymerize [33]. Interestingly, α-syn can itself serve as a cofactor to promote tau polymerization and both proteins co-localize within inclusion bodies [17,18,34]. Recent evidence has shown that co-transfection of α-syn with tau induces insoluble, cytotoxic, α-syn aggregate formation [34]. Likewise, cellular seeding with α-syn fi brils induces the formation of cytotoxic neurofi brillary tangle-like inclusions [35]. Extracellular seeding of α-syn fi brils can also promote recruitment of soluble α-syn into insoluble Lewy body-like inclusion bodies [36].

Transgenic models that combine β-amyloid and α-synuclein pathology
Mouse models of overlapping Aβ and α-syn pathology lend further support to the theory that Aβ and α-syn interact synergistically to create a more severe disease course. Double-transgenic mice expressing human amyloid precursor protein (APP) and wild-type hSYN develop motor defi cits at 6 months compared with 12 months in single-transgenic hSYN mice [19]. Th ese human APP/ hSYN mice also develop spatial memory defi cits and increased numbers of Lewy body-like inclu sions [19]. Th ese double-transgenic mice therefore provide a useful model for examining the potential inter actions between Aβ and α-syn.
To model all three of the pathologies that co-exist within AD-LBV patients, Clinton and colleagues crossed 3xTg-AD transgenic mice with a mutant α-synuclein transgenic line [22]. Th e 3xTg-AD model develops Aβ plaque and neurofi brillary tangle pathology via co-expression of mutant APP, mutant presenilin-1, and mutant tau. By adding a mutant α-syn (A53T) transgene to the mix, this model (hereafter referred to as AD-LBV mice) successfully recapitulated all three major AD-LBV pathologies [22]. Interestingly, AD-LBV mice exhibit accelerated cognitive dysfunction versus 3xTg-AD or α-syn lines, suggesting that this complex model mimics an important feature of AD-LBV. Similar to the single-transgenic α-syn mouse, the AD-LBV mice develop Lewy body-like inclusions. However, AD-LBV mice show increased levels of insoluble α-syn, pS129-syn, and Lewy body pathology at much earlier ages than single-transgenic α-syn mice. Two other pathological results of interest are that AD-LBV mice develop increased levels of insoluble Aβ 42 tau at younger ages than 3xTg-AD mice. Although this model uses mutant transgenes, the results nevertheless provide important additional evidence that Aβ, tau, and α-syn can interact synergistically to acceler ate pathogenesis and cognitive decline.
While much of the in vivo evidence linking Aβ and αsyn comes from the use of transgenic mice that express mutant genes, the study of these models has yielded invaluable additions to our knowledge of neuro degenerative disease [1,18,37]. Importantly, similar results have been found in models regardless of whether mutant [22] or wild-type [19] α-syn transgenes were utilized. Interestingly, investigations of familial AD presenilin 1, presenilin 2, and APP mutation carriers also reveal increased development of Lewy body pathology [38][39][40]. Both mouse models and human cases thus suggest that diseaseassociated APP and presenilin mutations can enhance the pathological accumulation of wild-type α-syn.

Potential mechanisms that link β-amyloid and α-synuclein pathology
While the interactions between AΒ and α-syn have been well documented, far less is known about the mechanisms by which these proteins exert their eff ects on one another. Research aimed at elucidating these mechanisms will no doubt provide critical knowledge about the progression of AD, AD-LBV, DLB, and PDD, and thus may also yield new therapeutic targets.

Phosphorylation of α-synuclein and tau
One mechanism that may underlie the eff ects of Aβ on α-syn is by indirectly infl uencing the phosphorylation state of α-syn. As previously mentioned, Ser129 α-syn phosphorylation is a pathogenic change observed in DLB and other synucleino pathies, and interactions with Aβ and tau can enhance phosphorylation at this residue in vitro and in vivo [22,30]. Several kinases have been identifi ed that may mediate Ser129 phosphorylation, but the most impor tant appear to be casein kinase 2 and polo-like kinase 2 (PLK2) [30,[41][42][43][44]. PLK2 was recently shown to phosphorylate α-syn at Ser129 with greater effi ciency than casein kinase 2 [44]. Interestingly, PLK2 expression is elevated in AD and DLB patient neurons. Phos phory lation of Ser129 was also recently detected in synaptic-enriched fractions from AD patients [45]. Upregulation of PLK2 in AD and DLB could thus potentially mediate the increased phos phory lation of αsyn observed in these patients. Increased phosphorylation of Ser129 has in turn been shown to increase the propensity of α-syn to form aggregates [30]. In contrast, one recent study found that increasing pS129 through various mechanisms, including increased PLK2 expres sion, did not alter the aggregation state [46]. While the evidence is clear that pS129-syn is associated with pathogenic changes, further research is needed to clarify the functional eff ects of Ser129 phosphorylation and the potential role of pS129-syn in Aβ and synuclein interactions.
Th e hyperphosphorylation of tau and its aggregation into neurofi brillary tangles and dystrophic neurites ( Figure 1) is a hallmark of AD. A number of studies have shown that Aβ can modulate tau phosphorylation and aggregation [47][48][49]. α-syn can also infl uence tau pathology [50,51]. Th e interactions between α-syn and tau appear bidirectional, however, as tau can also induce synuclein aggregation and phosphorylation [34]. In fact, tau overexpression can induce PLK2 expression, providing a potential mechanism for this eff ect [52]. Aβ could therefore possibly drive synuclein pathology indirectly by fi rst enhancing tau pathogenesis (Figure 2).

β-amyloid-induced infl ammation
Infl ammation is a critical component of AD [53] and also contributes to the pathogenesis of Lewy body disorders [54,55]. Although there is currently limited evidence connecting Aβ-induced infl ammation with α-syn aggre gation, we speculate that the eff ects of Aβ on infl amma tory processes could indirectly drive the phosphorylation and aggregation of α-syn. A growing body of evidence suggests that Aβ can indeed infl uence tau pathology via this kind of mechanism. For example, Aβ-induced release of proinfl am matory cytokines can in turn activate kinases such as cyclin-dependent kinase 5 that promote tau phos phorylation [37,56,57]. Interestingly, cyclin-depen dent kinase 5 has also been implicated in Lewy body formation -this same kinase may therefore infl uence α-syn aggregation [58]. In further support of this hypo the sis, age-related changes in microglial activation and cyto kine release can enhance nitric oxide production, increas ing α-syn nitration [59]. Nitration and oxidation of α-syn can in turn accelerate α-syn aggre gation [60]. Th e relationship between α-syn and infl am mation appears to be reciprocal, as α-syn can itself can drive astrocytic and microglia activation [61,62]. Notably, one recent report showed that tau overexpression can also drive infl ammation and enhance α-syn accumulation and phosphorylation [52].
Clearly a great deal more work is needed to determine whether infl ammation truly infl uences the interactions between Aβ and α-syn. However, infl ammatory-mediated changes in cytokine expression and kinase activation probably infl uence α-syn in much the same way as they modulate tau.

Impaired protein degradation
Another common mechanism thought to underlie many neurodegenerative disorders is dysfunction in protein clearance mechanisms. Indeed, impairments in both the ubiquitin-proteasome system and the autophagylysosome pathway occur in AD and Parkinson's disease, and both pathways are important in Aβ and α-syn degradation [63][64][65][66][67]. Soluble oligomeric Aβ, in addition to aggregated α-syn, impairs the normal function of the proteasome [63,68]. Th e ubiquitin-proteasome system is also critical in the degradation of tau, and the E3 ligase (C-terminus Hsp70 interacting protein) targets both tau and α-syn for degradation [69][70][71]. Interestingly, proteasomal impair ment caused by one pathogenic protein may in turn reduce degradation of other pathogenic proteins. For example, Aβ-induced proteasome dysfunction increases the accumulation of tau [64,68].
Both α-syn and Aβ are also degraded by autophagy. Pathogenic interactions between α-syn and Aβ could therefore infl uence the function of this critical pathway. For example, a subset of neurons with increased levels of α-syn has been shown to recruit the autophagy pathway to compensate for impaired ubiquitin-proteasome system function. An increased burden on lysosomal degradation could thus drive dysfunction in vulnerable neuronal populations [63,66]. Uptake of Aβ 42 was also shown to induce lyso somal leakage, providing another possible mechanism for both direct and indirect interactions between Aβ 42 and α-syn [19,63]. Interestingly, activation of autophagy by over expression of Beclin-1 can reduce Figure 2. Potential mechanisms linking β-amyloid and α-synuclein pathology. Studies support several putative mechanisms by which β-amyloid (Aβ) and α-synuclein (α-syn) may interact to enhanced pathology and cognitive decline. Such mechanisms include (left to right): chronic infl ammation and microglial activation induced by both Aβ and α-syn; direct interactions and hybrid oligomerization of Aβ and α-syn; Aβ-induced kinase activation and α-syn phosphorylation; impairment of proteasome and autophagy degradation pathways; and Aβ-induced phosphorylation of tau leading to tau-mediated enhancement of α-syn aggregation. CK-2, casein kinase 2; PLK-2, polo-like kinase 2; PHF, paired helical fi laments; NFT, neurofi brillary tangle; p-Tau, phosphorylated tau; pS129, phosphorylated at serine 129; p-syn, phosphorylated α-synuclein; UPS, ubiquitinproteasome system.
Finally, there is the possibility of disruption of cyto plasmic protease activity. For example, the serine protease neurosin (kallikrein-6) has been shown to degrade α-syn and to prevent its polymerization [63]. Intriguingly, neurosin is dysregulated in Parkinson's disease and decreased in the brains of AD patients, providing another possible mechanism by which impaired protein clearance could drive synuclein pathology [63]. Impaired protein degradation clearly plays a substantial role in many neurodegenerative disorders. Th e combined actions of Aβ and α-syn on the ubiquitin-proteasome system and autophagy-lysosome systems provide a poten tial mechanism to explain the acceleration of pathology and cognitive decline in patients with overlapping pathologies. Dysfunction in the lysosomal system may also facilitate the direct interaction between Aβ and αsyn in neuronal subpopulations where Aβ and α-syn coexist [19].

Direct interactions between β-amyloid, α-synuclein, and tau
Aβ and α-syn do not normally exist in the same subcellular compartment in healthy cells, thus limiting their potential for direct interaction [19]. In pathological states, however, the localization of many proteins including Aβ and α-syn can be altered. For example, Aβ and αsyn have both been detected within mitochondria [74,75]. Likewise, both proteins can accumulate within lysosomes and autophagasomes [76,77]. Direct interactions between these proteins could thus potentially occur within damaged or diseased cells. To date, most of the evidence supporting direct inter actions between Aβ and α-syn comes from in vitro experiments. For example, cell-free studies show that α-syn can promote confor mational changes in Aβ that are detected by NMR spectroscopy [20]. Aβ and α-syn can also form complexes and can co-immunoprecipitate from AD-LBV patient brains and transgenic models, providing some in vivo evidence for direct interactions [78]. Th is same study provided evidence that these two proteins can form hybrid porelike oligomers that increase calcium infl ux. Tau can also enhance α-syn aggregation and toxicity [34], and both proteins can co-localize within AD-LBV patient neurons, dystrophic neurites, and Lewy bodies [17,18,79]. If direct interactions between Aβ and α-syn do indeed play a role in AD-LBV pathogenesis, it will be important to understand why these interactions occur only in some patients and brain regions but not in others.

Conclusions
Th e co-existence of Aβ and α-syn pathologies in dementia patients clearly does not simply represent two concurrent yet independent disease states. Evidence suggests instead that Aβ and α-syn may interact synergistically to enhance each others' aggregation and accelerate cognitive decline. Th e mechanisms by which these two aggregation-prone proteins interact remain unclear. However, growing evidence suggests that Aβ may infl uence α-syn pathology by modulating protein clearance, driving infl ammation, activating kinases, or directly altering α-syn aggregation. While a great deal of work is needed to confi rm and clarify these putative mechanisms, the prevalence of combined AD and LB disease clearly justifi es the need.

Competing interests
The authors declare that they have no competing interests.