Possible Role of Amyloid Cross-Seeding in Evolvability and Neurodegenerative Disease

Aggregation of multiple APs in neurodegenerative disorders

One of the major features in the pathogenesis of neurodegenerative diseases is the aggregation of multiple APs [11]. As an example, the co-presence of aggregated Aβ and tau is a defining pathological feature of AD. Furthermore, in dementia with Lewy bodies (DLB) as well as familial AD, αS aggregates frequently co-localize with those of Aβ [12, 13]. Moreover, in amyotrophic lateral sclerosis (ALS), SOD-1 (Cu/Zn superoxide dismutase) and TAR DNA-binding protein of 43 kDa (TDP-43) may also be co-aggregated [14]. Perhaps the most prominent disorder featuring aggregation of multiple APs is the ALS/parkinsonism-dementia complex (PDC) of Guam that is associated with aggregation of Aβ, tau, αS, and TDP-43 [15–17]. Because the etiology of ALS/PDC complex is unknown, several hypotheses have been provided, including cycad toxicity and flying fox consumption [18]. Parallel pathologic mechanisms may also apply to the ALS/PDC compex found in Kii peninsula in Japan [19]. Collectively, multiple combinations of APs may be general phenomena in neurodegeneration.

The CS of APs in experimental models of neurodegeneration

The mechanism by which the aggregation of multiple APs occurs in neurodegenerative diseases is obscure. In this regard, recent study shows that the CS of APs may play a major role in a wide variety of experimental models of neurodegeneration, including AD, PD, DLB, ALS and transmissible spongiform encephalopathy [20]. In AD, the CS of Aβ with tau might be caused by various mechanisms. Supporting this, Aβ directly associated with tau in vitro [21], while phosphorylation of tau by Aβ was essential to pathogenesis in both cell-based and animal studies in AD [22]. Similarly, aggregation of αS was increased by Aβ in vitro [23] as well as in transgenic mouse model of α-synucleinopathies [24]. Moreover, Aβ promotes the aggregation of other APs, such as prion protein (PrP) and TDP-43 [11, 25]. Collectively, Aβ may play a central role in the CS of APs, yet other APs may also fulfill a similar role in the CS of APs. For instance, stimulation of αS aggregation in the presence of tau and PrP in α-synucleinopathies may indicate the CS of APs other than Aβ [26, 27]. In the similar context, SOD-1 and TDP-43 may co-aggregate in ALS. In ALS/PDC complex, it is probable that environmental factors might cause neurotoxicity, ultimately promoting the CS of APs. Taken together, the CS of APs may underlie a broad range of neurodegenerative conditions, which prompts the question as to why the aggregation of multiple APs, apparently injurious to the brain, has emerged and persisted against the pressures of natural selection.

Bacteria

Specifically, bacterial APs are highly conserved, being involved in biofilm formation, which benefit bacteria during invasion, host adhesion, and resistance to destruction [29]. Curli, the best studied bacterial AP, is made by enteric bacteria such as Escherichia coli and Salmonella spp, and its key element, CsgA, has been found to contain amyloidogenic peptide repeat motifs shared by prions and αS that assemble into amyloid fibers. In E. coli, the polymerization of the major curli fiber subunit protein CsgA into an amyloid fiber depends on the minor curli subunit protein, CsgB. The outer membrane-localized CsgB protein shares approximately 30% sequence identity with the amyloid-forming protein CsgA, suggesting that CsgB might also have amyloidogenic properties. Also described, curli promotes the conversion of PAP_248–286 into the amyloid SEVI, exemplifying the CS of dissimilar amyloid sequences (Fig. 2a) [30]. Thus, the CS of CsgA and CsgB may play an important role in bacterial physiology.

Fig.2

Physiological roles of the CR of APs in microorganisms. a) Bacterial curli protein promotes the conversion of PAP_248–286 into the amyloid SEVI. Electron microscopic images of SEVI fibers formed in the absence of curli (left) and in the presence of 5 mol% CsgA (middle) and CsgB (right) fibers. Fibers were grown at a concentration of 440μM PAP248–286 at 37°C under 1400 rpm orbital shaking for 7 days. Bars = 500 nm. Quantification of the fibers are shown. b) Heterologous prion-forming proteins interact to cross-seed aggregation in Saccharomyces cerevisiae. Modified from Hartman et al. [30] (a) and Keefer et al. [31] (b) with permission.

Yeast

In addition, it was recently shown that the CS of yeast prions, including Rnq1 and Sup35, is a predominant mechanism leading to self-propagating and aggregation of the translation termination factor sup35 for formation of the yeast prion [PSI⁺], suggesting that CS of APs may be physiological in yeast (Fig. 2b) [31]. Collectively, it is predicted that the CS of APs might evolve as a physiological phenomena in the lower organisms.

Hormesis

A recent study suggests that APs are composed of structurally heterogeneous populations in part due to the ‘intrinsically disordered structure’ of APs that lack fixed or ordered three-dimensional structures [32, 33]. As such, it is conceivable that AP heterogeneity might correspond to diverse stressors and that APs might retain information regarding a variety of stressors through structural changes. Given that AP evolvability is beneficial, one might further predict that the CS of multiple APs may synergistically increase the heterogeneity of APs aggregates, allowing greater capacity to cope with increasingly diverse stressors (Fig. 3). Thus, the CS of APs may be physiologically regulated to acquire the resistance of stresses, namely hormesis.

Fig.3

Role of the CS of APs in evolvability and neurodegenerative diseases. The CS of multiple APs, such as Aβ, αS, tau, SOD-1, TDP-43, and PrP, may stimulate protein aggregation, resulting in increased evolvability for offspring to cope with forth-coming stressors. Instead, antagonistic pleiotropy in aging might promote neurodegeneration in parental brain. Therefore, the CS of multiple APs may be beneficial for offspring, but detrimental for parents.

Transgenerational transmission

Since the requirements for evolvability include not only hormesis, but also heredity to generate adaptive genetic diversity [34], it follows that transgenerational transmission is a critical component. In case of unicellular organisms, such as yeast and bacteria, protofibrillar APs may be easily transmitted from parental to progeny cells in concert with cell division. By contrast, the transgeneration of AP protofibrils in multicellular organisms may be more complex. In humans, protofibrillar APs may be transmitted to offspring via germ lines [6, 35]. As far we are aware, few reports demonstrate the presence of APs protofibrils in vivo, including germ cells, however, further investigations are warranted since amyloid fibrils are abundantly present in semen [36].

Because monomeric APs may be unstable due to its intrinsically disordered nature [37], it is possible that oligomers and protofibrils of APs might be more stable. Thus, it is interesting to determine whether the protofibrils composed of heterogeneous APs might be more stable compared to the homogeneous protofibrils of APs (Fig. 3). There is also great interest in the role of exosomal RNAs, including ncRNAs and miRNA, transmitted transgenerationally in a non-genetic manner through germ cells. [38]. Considering that prefibrillar Aβ aggregates preferentially bind to exosomes [39], it is possible that these biologic processes might in some way be relavant to the transgenerational transmission of Aβ prefibrils. Taken together, further studies may reveal that multiple APs may cooperate to increase the efficiency of evolvability, including the CS of APs.

Antagonistic pleiotropy

On the other hand, neurodegenerative diseases may manifest in parental brain through the antagonistic pleiotropy mechanism in aging [9, 40], whereby in neurodegeneration multiple APs might be more detrimental compared to singular AP. Although the neurotoxicity associated with the evolvability of APs protofibrils may be well regulated during the reproductive stage, the same might be reduced or absent in post-reproductive senescence, leading to neuropathological phenotypes characterized by mature fibrils and the CS of APs. Thus, the CS of APs is primarily beneficial in development/reproduction, but is detrimental in aging. This viewpoint might explain why the CS of APs, a deleterious pathological phenomenon during aging has not been selected out during the evolutionary process.

Role of the CS of APs to promote evolvability in non-neuronal tissues

Given that type 2 diabetes mellitus (T2DM) and neurodegenerative disorders, such as AD and PD, are analogous in terms of amyloidosis associated with stressors [41], there might be common mechanisms underlying these two diseases. Similar to AD, the aggregation of amylin in pancreatic β cells during T2DM pathogenesis could represent antagonistic pleiotropy of amylin evolvability. Hypothetically, gestational DM might promote transmission of the amylin protofibrils to deliver information about maternal stressors, which if deficient, might promote the risk of type 1 DM in offspring. T2DM, however, might manifest later through antagonistic pleiotropy in parental aging. Furthermore, it is interesting to note that amylin and αS co-aggregate in pancreatic β cells, suggesting that the CS of these APs might be pathologically important [42].

Experimental approach

The effect of the CS of multiple APs on evolvability would be assessed experimentally using transgenic (tg) mice model of neurodegenerative diseases. For instance, it was shown that tau and αS synergistically promoted fibrillization each other both in vitro and in the bigenic (αS/tau) mice expressing wild-type human αS plus P301L mutant tau [43]. According to our EVH, it is anticipated that the offspring born from the tg mice with administration to neurotoxins, such as 1-methyl–4-phenyl-1,2,3,6-tetrahydropyridine and 6-hydroxydopamine, might be more resistant compared to those born from the same parent without treatment. Such effects might be more stronger if the parents are the bigenic (αS/tau) mice compared to the single tg mice. Essentially similar results could be obtained using tg mice for αS and amyloid precursor protein [24]. In both cases, it should be considered that the fertility of the bigeic mice might be compromised compared to those of the single tg mice.

Aβ aggregation by viral infection

Notably, it was recently shown that the seeding of Aβ aggregation was stimulated by infection of Herpes simplex virus type 1 (HSV1) in cellular and tg mice models of AD (Fig. 4a) [45]. Mechanistically, it was suggested that Aβ aggregates entrap virus at the cell membrane and protect the brain against viral infection [45]. Based on a series of experimental results, the authors proposed the “antimicrobial protection model” (APM) of AD [46], suggesting that Aβ fibrillization may drive neuroinflammatory pathways that help fight the infection and clear Aβ/pathogen deposits.

Fig.4

Seeding of Aβ fibrils by HSV1 infection. a) Ultrstractural analysis of immunoelectron micrography shows Aβ fibrillization seeded by HSV1 in cell culture leading to virus capture and entrapment. Please see the experimental conditions in Eimer et al. (2018) [45]. Reprinted from Eimer et al. (2018) [45] with permission. b) Schematic of Aβ evolvability of and disease manifestation related to HSV1 infection. Aβ evolvability might be an epigenetic phenomenon transmitted transgenerationally to confer resistance against the HSV1 infection in offspring during reproduction, which may be beneficial in evolution. However, evolvability might lead to T2DM during parental aging through the antagonistic pleiotropy mechanism. The Aβ evolvability is increased by various causes, such as the CS of APs, may result in an efficient delivery of information of stresses associated with the HSV1 infection for offspring, but increased frequency of AD in parents aging. The CS of APs and the antagonistic pleiotropy may be targets of therapy strategy (Tx).

Given that amyloid-fibril structures are shared among APs, it is likely that APs other than Aβ may similarly entrap viral particles and neutralize viral infection. Indeed, many studies recently have described an association between virus infection and neurodegenerative as another important common feature of these disorders, not only HSV1 in AD, but also H1N1 influenza virus in PD and retroviruses in ALS [47]. As Aβ is protective against HSV1, it is possible that αS and TDP-43 might also be protective against some other viruses. In particular, it is of interest to determine whether the CS of APs might further enhance protection against various viral infections. It would be great interest to experimentally validate such hypotheses using tg mouse models for the various APs.

Comparison of APM with other hypotheses

There are at the present, at least three hypotheses which are relevant to the aggregation of APs, namely APM, EVH and ACH (Table 1). The ACH focuses on neurodegeneration, a pathological phenomenon in aging, whereas the EVH covers evolvability, a physiological phenomenon mainly in the developmental/reproduction stage in offspring/parents, as well as aging by antagonistic pleiotropy in parent. In contrast to these two hypotheses, the APM addresses a singular factor, namely viral infection, which may occur during any stage of life.

Because aging-associated neurodegeneration and viral encephalitis are commonly associated with various stresses, including neurotoxins, inflammation, oxidative stress and hyperthermia, both the APM and the ACH may correspond to hormesis in the EVH, but fails to encompass the relevance to transgenerational transmission from parents to offspring of the information regarding stressors (Fig. 4b). Although the APM is physiological, whereas the ACH is pathological, both are similar in that Aβ is dominant compared to other APs, and in that both hypotheses are based on experimental results (Table 1). On the other hand, the EVH presumes the presence of the APs protofibrils, but is attractive due to a quality of inheritance of APs (Table 1). At this point, the three hypotheses have both advantages and disadvantages, and the biological significance of the CS of APs may depend on all these hypotheses.

Current strategy: Anti-aggregation

In the past decade or more, the ACH has been the predominant theory in the field of AD, which also has strongly influenced views on other related neurodegenerative conditions. As a result, the current experimental therapeutic paradigm is directed at mitigating protein aggregation to reduce the neurotoxicity of AP aggregates. Based on this, clinical trials using various therapeutic modalities, including antioxidants, anti-inflammatory agents and immunotherapies, have been evaluated extensively and/or are still in progress [48–50]. Despite numerous studies, none has proven efficacious, failing to meet established target endpoints. To be effective, it is highly possible that therapeutic interventions must be initiated early in disease pathogenesis. As a result, a phase 3 study of Aβ immunotherapy was conducted in dominantly inherited AD to assess the effect of therapeutic intervention in the pre-symptomatic stage [51]. Recently, however, it was revealed that the trial in early-onset AD was unsuccessful and thus terminated [52]. As for the failure of such clinical trials, at least two possibilities should be considered. First, immunotherapy against Aβ was not sufficiently efficacious. More specifically, although it was previously shown that there were significant decreases in amyloid plaque load in autopsy brain from Aβ immunotherapy patients, it is possible that residual neurotoxic APs protofibrils were still present. Indeed, neuropathologic examination of participants in previous clinical trials (e.g., AN1792: an active Aβ42 immunization by Elan Pharmaceuticals) may support of this notion [53]. Despite efficient removal of senile plaques in AD brain, most patients had progressed to severe dementia, possibly due to continued tau propagation [53], suggesting that the cross-seeding of Aβ with tau might be important in the pathogenic mechanisms of AD and the simultaneous targeting of Aβ and tau might be therapeutically more effective in AD. Since the current therapy strategy, especially immunotherapy, is directed at a single molecule among various APs, to improve efficacy, it may be necessary to consider collectively, protein aggregation involving combinations of multiple APs (Fig. 4b).

Antagonistic pleiotropy as a therapy target

Alternatively, the protein aggregation of Aβ might be unrelated to neurotoxicity during neurodegeneration, and in this case, it may be necessary to devise a novel therapy strategy. Presuming that neurodegeneration might be a manifestation of evolvability through antagonistic pleiotropy in aging, it follows that an attractive alternate therapy strategy might focus on the antagonistic pleiotropy mechanism (Fig. 4b). In this regard, a recent study suggests that a 2q22 region corresponding to the TGFβ/activin receptor-signaling pathways might be linked to the risk of major human diseases, including neurodegenerative disorders [54]. It is therefore possible that targeting and modifying the TGFβ/activin receptor-signaling pathways could be therapeutically beneficial for neurodegeneration [37]. This might be assessed by applying various treatment strategies such as receptor antagonists and antisense RNA, to tg mouse models of neurogeneration.