SEARCH

SEARCH BY CITATION

Keywords:

  • amyloid;
  • amyloid cytotoxicity;
  • amyloid diseases;
  • amyloid oligomers;
  • cell membranes

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

A great deal must still be learnt on the structural features of amyloid assemblies, particularly prefibrillar aggregates, and the relationship of the latter with amyloid cytotoxicity. Presently, it is recognized that the population of unstable, heterogeneous amyloid oligomers and protofibrils is mainly responsible for amyloid cytotoxicity. Conversely, mature fibrils are considered stable, harmless reservoirs of molecular species devoid of toxicity in the polymerized state. This view has been modified by recent reports showing that mature fibrils grown at different conditions can display different structural features and stabilities, possibly leading them to undergo disassembly with the leak of toxic oligomers. Fibril polymorphism is paralleled by oligomer polymorphism and both can be traced back to amyloid growth from differently destabilized monomers with distinct structural features at differing conditions. Recent research has started to unravel oligomer structural and biophysical features and the relationship between the latter and oligomer cytotoxicity. These data have led to the proposal that, together, both oligomer and membrane physical features determine the extent of oligomer–membrane interaction with the resulting disruption of membrane integrity and cell impairment. Such a view can help to explain the variable vulnerability of different cell types to the same amyloids and the lack of relationship between amyloid load and the severity of clinical symptoms. It also stresses the importance, for cell/tissue impairment, of the presence, in tissue, in addition to toxic oligomers, of fibrils conformers of reduced stability as a possible source of toxic oligomers, whose leakage can be favoured upon interaction with suitable surfaces or by other environmental conditions.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

Amyloid diseases include over 25 familial, sporadic or transmissible pathological conditions, in some cases with remarkable clinical relevance due to their dramatic prevalence in the population (type 2 diabetes mellitus, Alzheimer’s and Parkinson’s diseases). These pathologies are characterized by the intracellular and/or extracellular presence, in specific tissues and organs, of fibrillar deposits grown from one of approximately 25 different peptides or proteins, either wild-type or carrying destabilizing mutations (reviewed in [1–3]). The deposited material is found in the extracellular matrix (amyloid) or inside the cells (inclusion bodies, aggresomes) where it impairs tissue physiology and cell viability (reviewed in [1,2]). Typically, the fibrils found in amyloid diseases are long, unbranched polymeric assemblies 2–10 nm wide, displaying a regular core structure that results from a couple of parallel or antiparallel β-sheets propagating along the fibril main axis and whose β-strands are perpendicular to the latter (the so-called cross-beta structure). Each of these peptides/proteins is characteristic of a specific disease or of a group of closely similar pathological conditions associated with peculiar clinical signs (reviewed in [2,3]). When the aggregated polypeptides are variants resulting from genetic mutations associated with early onset, familial forms, they display an increased propensity to misfold and to aggregate; in this regard, recent research has provided clues on the relationship between the amino acid sequence and both the aggregation propensity of any unfolded polypeptide chain [4] and the time of onset and severity of its clinical signs [5].

The presence, in selected tissues/cells, of fibrillar proteinaceous deposits is a shared hallmark of any peculiar amyloid condition; this finding was at the origin of the amyloid cascade hypothesis, proposed almost 20 years ago to explain the molecular basis of such pathological conditions. The hypothesis states the existence of a direct causative link between the presence of protein aggregates in tissues and the appearance of the clinical symptoms [6]. Although the structural and physical features of the aggregated species (mature fibrils and their precursors) and their biological and cellular effects are still under intense investigation, the amyloid hypothesis has gained solid support by a large number of biochemical and genetic studies [7,8].

Recent research has increasingly confirmed the view that the ability to misfold and to aggregate into amyloid assemblies can be considered a generic property of most, perhaps all, polypeptide chains under specific perturbing conditions rather than a peculiar behaviour of the few peptides found aggregated in amyloid diseases associated in some way to some peculiarities of their amino acid sequences (reviewed in [1,2]). Nevertheless, only a very limited number of proteins and peptides are found aggregated in specific diseases under physiological conditions. Such an apparent paradox can be explained by considering the highly co-operative character of protein folding, the structural adaptations set up by protein evolution to increase the resistance of natural proteins against their primordial tendency to aggregate (reviewed in [9]) and the evolution of highly effective molecular machineries aimed at performing the quality control of protein folding (reviewed in [10]). In spite of the existence of these (and several other) biological adaptations, any alteration of specific parameters affecting the conformational state of the precursor, monomeric variants that nucleate aggregate growth of a specific protein/peptide can trigger the build-up of aggregation-prone species. These alterations include the expression level (resulting from increased synthesis or reduced degradation), structural perturbations (arising from specific mutations, lack of ligands or chemical modifications) or any impairment of the quality control of protein folding in the cell; minor changes in the environmental conditions (macromolecular crowding, redox state, heat shock, presence of suitable surfaces, etc.) affecting the conformational state of a polypeptide chain can also contribute to trigger protein/peptide aggregation (reviewed in [1,2]).

Generally, in amyloid diseases a toxic gain of function is involved resulting from the appearance of an unnatural toxic fold in the aggregated proteins/peptides. However, the formerly accepted idea that such a toxic fold is present in mature fibrils that, accordingly, are mainly responsible for cell demise in the affected tissues, is questioned by an impressive body of experimental data. Presently, mature fibrils are considered harmless and substantially inert deposits of toxic precursors, and their growth is considered a passive cell defence mechanism. Such a view can help to explain the lack of direct correlation between the density of fibrillar plaques in the brains of Alzheimer’s disease patients and the severity of their clinical symptoms [11]. Actually, in some cases, amyloid fibrils can impair cell viability directly [12,13] or indirectly; in the latter case, they can trigger an inflammatory response in tissue with secretion of toxic cytokines by macrophages and glial cells or even leak toxic oligomers (see below). Moreover, in systemic amyloidoses, the huge amount of material that in some cases is deposited in the affected organs can hinder proper blood–tissue exchange of oxygen and nutrients with cell sufferance. Nevertheless, the notion remains generally accepted that, at least in most cases, the oligomeric assemblies transiently arising in the path of fibrillization of several peptides and proteins associated with amyloid disease are the main or even the sole cytotoxic species [14–18]. However, the different intermediate assemblies appearing in the fibrillization path can differ greatly in their ability to impair cell physiology and viability. For example, it is known that the neuroinflammatory response in the Alzheimer’s disease brain is more specifically associated to the presence of fibrillar Aβ [19], whereas small Aβ42 oligomers impair long-term potentiation in the brains of people with Alzheimer’s disease [14], raise endoplasmic reticulum stress [20] and eventually cause cell death following an aggregation state-specific uptake [21].

In spite of the growing knowledge of the effects of amyloids on cell biochemical and functional features, for many amyloid diseases no information is currently available on the identity of the supramolecular assemblies responsible for tissue damage in vivo and on the molecular mechanism(s) of cell impairment. In addition, a substantial lack of information still remains on the structural features of the toxic prefibrillar oligomeric assemblies, a study that is made more difficult by the heterogeneity and instability of these species. Accordingly, increasing efforts are dedicated to investigating the conformational features of the monomers in the soluble oligomeric fibril precursors, as well as the structural properties and the structure–toxicity relationship of the latter (see below). In general, the oligomeric prefibrillar assemblies are highly unstable, transient, flexible and dynamic structures still exposing patches of hydrophobic residues on their surface; this feature explains oligomer instability, their tendency to organize into more stable higher order assemblies and to interact with cell surfaces such as those exposed by other macromolecules and cell membranes, thus providing a general rationale of their toxic effects.

This review will focus on the importance of the biochemical and biophysical features of cell membranes and amyloid oligomers/fibrils, as well as of the polymorphism of the latter, in modulating the oligomer/fibril–membrane interaction, an event that, at least in many cases, is considered the first step of amyloid cytotoxicity. It will also describe our knowledge on the structural features and polymorphism of prefibrillar aggregates (notably the early oligomeric species) and the importance of the cell surfaces in promoting protein aggregation and aggregate toxicity. Finally, very recent data on the importance of the relative contributions of the biochemical and biophysical features of cell membranes and amyloid oligomers to the resulting toxic effects of the latter will be discussed.

General structural features of amyloid oligomers

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

Prefibrillar assemblies of amyloid type occur in vitro and/or in vivo in the aggregation path of many proteins and peptides associated with amyloid disease, including Aβ peptides, α-synuclein, prion proteins, β2 microglobulin, amylin, transthyretin, serum amyloid A, and others (reviewed in [22]). Similar prefibrillar entities have also been imaged in the aggregation path of several proteins not associated with amyloid disease [17,23]. Although most of these assemblies are considered intermediates in the path of fibrillization, some of them, such as the frequently imaged small annular oligomers (‘doughnuts’), could be ‘dead end’ products of the process or, at any rate, structurally and functionally distinct types of amyloid oligomer. Soluble oligomers of several peptides and proteins have also been repeatedly detected in, and purified from, cultured cells and tissues where the monomeric precursors are expressed [14,24–26] reinforcing the idea that these species are really present in vivo and are directly associated with cell/tissue impairment.

Considerable information has recently been gained on the hierarchical growth of amyloid fibrils from structurally more simple precursors through a number of steps, as well as on the structural features of the ordered β-sheet-rich core of amyloid fibrils and the supramolecular organization of the latter [27–29]; yet, a severe lack of knowledge on the overall structural features of fibril precursors still remains. In general, the formation of oligomeric aggregation nuclei (nucleation) is considered a key event in the onset of protein aggregation and often the rate-limiting step of fibril growth [30]; it also explains the delay time of polymer appearance that is recorded in in vitro protein aggregation experiments. However, at variance with protein folding, where in-depth investigations carried out in the last decade have provided significant information on the structural features of folding intermediates and transition states, only reduced knowledge is presently available on both the energetically favourable conformational states an aggregating polypeptide chain can reach; the conformational properties of the oligomeric assemblies arising in the path of aggregation are also substantially unknown. In this regard, a quantitative kinetic model for the aggregation of β-lactoglobulin into amyloid using a simulation exploiting a combination of atomic force microscopy, particle size distribution, 1-anilino-8-naphthalene sulfonate, thioflavin T and dynamic light scattering measurements has very recently been reported, together with a proposed aggregation energy landscape for this protein [31]. Actually, some of the energy minima occurring in the protein aggregation energy landscape are expected to be scarcely defined due to the broad heterogeneity and instability of rapidly interconverting oligomeric states endowed with similar free energies. Conversely, the energy minima of the more structurally defined and stable higher order polymers (protofilaments and mature fibrils) can be much more easily identifiable, even though the same protein can organize into oligomers with different conformations growing into fibrils with differing morphologies and structural features under variable solution conditions both in vitro and in tissue [32,33].

As indicated above, amyloid fibril growth appears as a hierarchical process starting with misfolded/unfolded monomers. The structural features and molecular dynamics of the latter, the extent of their exposed hydrophobic surface and the environmental conditions determine, to a large extent, monomer arrangement into loose, transient, unstable, globular or tubular particles typically 2.5–5.0 nm in diameter, referred to as oligomers [34,35]. These globular entities frequently associate into beaded chains, small annular rings (‘doughnuts’ or ‘pores’), curvy protofibrils, large closed rings and/or ribbons. Some of these species can be off-pathway intermediates, whereas others often appear to be the precursors of longer ribbons or protofilaments, eventually organizing into mature fibrils (reviewed in [22]). In general, it is accepted that in most cases the transiently formed oligomeric species are richer in β-structure than the original monomers, but poorer in respect to the mature fibrils, expose hydrophobic clusters and result from the association of a number of partially folded monomers, as has been shown for α-synuclein [34] and other systems. Many in vitro studies have imaged the morphological modifications of amyloid assemblies during their growth from different disease-related and disease-unrelated peptides and proteins [32] and a theoretical frame for these modifications has been provided by molecular dynamics simulation studies [35].

As outlined above, the mechanistic picture of how misfolded monomeric peptides/proteins self-organize into oligomeric assemblies proposed by several authors still awaits definitive experimental confirmations. One of the most widely accepted models, the experimentally and theoretically supported ‘nucleated conformational conversion’ model [36,37], proposes that a group of unstable misfolded monomers in solution coalesce, generating relatively disordered, highly unstable ‘molten’ oligomers. Subsequently, in such a generic ‘two step’ mechanism, the monomers undergo extensive structural reorganization in the oligomers; the latter become increasingly structured, growing into higher-order assemblies with an increasingly more compact hydrophobic core, and eventually into mature fibrils [37–39]. According to this scenario, intermolecular hydrophobic interactions, together with protein concentration, medium temperature and polarity, must be among the major determinants of the rate and the extent of the hydrophobic collapse of the misfolded monomers into different types of oligomer. On the other hand, when the hydrophobicity of the misfolded monomer is low, the coalescence step can be skipped and the monomers establish predominantly intermolecular secondary interactions organizing directly into ordered oligomers [36]. Such a ‘one step’ mechanism can be at the basis of protein aggregation from natively folded proteins (reviewed in [39]), in which the exposure of hydrophobic residues is probably modest. According to this view, the choice between the two-step or one-step mechanism of oligomer growth depends on the balance between the rapidly forming intermolecular hydrophobic interactions and the slower exchange of the directional hydrogen bonds into the assembling oligomers [38]. These ideas provide a theoretical frame to the widely accepted view that aggregation is a generic property of virtually every polypeptide chain and that amyloids grown from structurally different peptides and proteins share a basic structure (see below).

The importance of the relative contributions of the hydrophobic forces and the hydrogen bonds also explains the finding that different aggregates (fibrillar or nonfibrillar) with similar β-sheet content, yet with differing biophysical and/or morphological features, can be grown from the same protein/peptide at different solution conditions favouring or disfavouring hydrogen bonding [40,41]; it also provides a molecular explanation of amyloid oligomer polymorphism, whose importance is increasingly recognized. In the steric zipper model of amyloid fibril formation and structure, the existence of distinct crystalline polymorphs of the same zipper-forming segments found in different aggregating peptides/proteins provides a possible molecular basis of amyloid polymorphism [42]. Amyloid polymorphism can rationalize several observations, including prion, and other protein, polymorphism propagation [31,43] (reviewed in [44]), the variable cytotoxicity [21,45,46] and the structural heterogeneity [47,48] of amyloids grown from the same peptides and proteins at different environmental conditions, and the appearance of in-path and off-path fibril intermediates [34,48].

As pointed out above, the heterogeneity, remarkable instability and intrinsically disordered nature of amyloid fibril precursors make it challenging to get solid data on the conformational features of these species. The study of the conformational features of amyloid assemblies has greatly benefited from the development of antibodies capable of specifically recognizing shared features of amyloid oligomers grown from different peptides and proteins. Various reports have described the generation of antibodies able to specifically recognize amyloid oligomers grown from differing peptides and proteins but not mature fibrils [18] or to cross-react with both amyloid pores [49] and with pores produced by pore-forming proteins [50]. Another recent report described an Ig directed against Aβ amyloid fibrils able to recognize both a shared, sequence-independent, epitope present in amyloid fibrils grown from different peptides and a type of soluble Aβ oligomer (fibrillar oligomer), but not a different type of apparently similar oligomer (prefibrillar oligomer) [50]. These data indicate the existence of structurally distinct families of amyloid oligomers with similar appearance, yet with different structural features and potentially different cytotoxicities. In the case of the Aβ peptides, it has therefore been proposed that at least two alternative aggregation nuclei for Aβ amyloid fibrils must exist: one type evolving into mature fibrils only after extensive structural reorganization and another type, possibly the true fibril precursor, growing into aggregates with increasing sizes by monomer addition [51]. These data could match the recently reported structural differences between Aβ fibrils grown from Aβ peptides in vitro or in tissue [30]. In conclusion, the use of conformational antibodies specific to different amyloid assemblies has allowed these distinct entities to be distinguished on the basis of their different structural features (reviewed in [52]); it has also provided an experimental basis for amyloid polymorphism investigation, showing that the same protein/peptide at the beginning of the aggregation can organize into oligomeric and prefibrillar assemblies with clearly different structural properties [51].

Apart from the above-described immunological tools, recently introduced biophysical techniques, such as single-molecule spectroscopy (notably fluorescence), small angle X-ray scattering in solution and solid-state NMR are also of great value to obtain information on the structural features and polymorphisms of amyloid oligomers, fibril aggregation nuclei and other assemblies populating the path of amyloid fibril growth (reviewed in [53,54]). Chemical cross-linking of oligomers followed by separation and analysis of the different species so-stabilized is another promising approach [55], similarly molecular dynamics simulation, a useful computational tool to address amyloid conformational variability [56].

In conclusion, the data provided by different experimental approaches indicate that the various types of prefibrillar aggregate resulting from unfolding/misfolding of various proteins and peptides share some basic structural features that differ from those displayed by the folded monomers or their fibrillar counterparts. However, these features can vary considerably in different oligomers, in terms of compactness of the hydrophobic core and, hence, of the accessibility to water of hydrophobic patches normally sequestered into the core of the monomeric protein, thus explaining the variable intrinsic instability of these assemblies. This view also provides clues on the molecular basis of oligomer/fibril polymorphism and on the inherent tendency of amyloid oligomers, and other unstable prefibrillar aggregates, to interact inappropriately with cellular components, notably membranes and, accordingly, on their toxic potential.

Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

Proteins are synthesized, fold and perform their biological function in the very complex and crowded intra/extracellular environment in close contact with other molecules and biological surfaces, such as membranes and macromolecular assemblies. Such conditions can profoundly influence the structural features of proteins, favouring compact states over extended ones [57]. In addition, peptides and proteins can interact with, and be actively recruited by, inorganic, synthetic, or biological surfaces, in particular membranes [58], but also macromolecules [59–62], thus modifying their conformational states, resulting in non-native, aggregation-prone conformations. Such a view has led to the proposal that surfaces can catalyse amyloid aggregate nucleation and growth by a mechanism substantially different from that observed in the bulk solution (reviewed in [63,64]). Surfaces, apart from populating aggregation-prone conformers possibly different from those arising in solution, can also strongly increase the local concentration of the latter, thus speeding up aggregate nucleation, possibly accompanied by structural alterations of membrane integrity [65]. These considerations account for the increasing interest in the investigation of the physicochemical features of protein interaction with, and aggregation on, artificial or natural surfaces, even in relation to the structure and lipid composition of the latter.

Apart from favouring protein misfolding and aggregation, synthetic and natural phospholipid bilayers can also be key interaction sites with prefibrillar aggregates (reviewed in [63]). In general, protein aggregation onto, or aggregate interaction with, cell membranes results in lipid extraction, loss of membrane integrity, derangement of selective permeability and impairment of specific membrane-bound protein function (reviewed in [63]). Such an effect is not exclusively observed with amyloid aggregates; in fact, a similar behaviour has recently been reported for nonamyloid, native-like oligomers and fibrils of the yeast prion Ure2 [66].

Lipid composition can affect the ability of phospholipid membranes to recruit and unfold polypeptide monomers and to nucleate and recruit amyloid oligomers. This is confirmed by the key well-known role of anionic surfaces (anionic phospholipid-rich liposomes, glycosaminoglycans, nucleic acids, natural membranes) both as triggers of protein/peptide fibrillization and as potent inductors of β-sheet structures (reviewed in [63]). These data agree with the selective antitumour specificity of endostatin [67], other antitumour protein aggregates, and a number of antimicrobial peptides possibly targeting the presence of phosphatidylserine at the surface of cancer cells and tumour vascular endothelial cells; accordingly, they have led to the proposal that a shared fold in amyloid aggregates grown from different peptides and proteins could recognize, although with different efficiency, anionic sites in cell membranes [68].

Cholesterol and gangliosides could also play a role in modulating protein aggregation at, and aggregate interaction with, cell membranes. Actually, the biochemical and biophysical features of the cell membrane can affect the conformation, distribution and proteolytic processing of membrane proteins involved in neurodegenerative conditions such as Alzheimer’s disease or prion disease; in addition, the protein/peptide interaction, with the cell surface, particularly with areas rich in membrane cholesterol and gangliosides, such as lipid rafts, is considered an important requirement for cytotoxicity (reviewed in [63,69,70]). In this regard, it has been reported that a loss of cholesterol in neuronal membranes enhances amyloid peptide generation and that the interaction of prefibrillar aggregates with the cell membrane, resulting in cytotoxicity, is impaired when the membrane is enriched in cholesterol (reviewed in [69]). Overall, the data presently available support the idea that membrane cholesterol and gangliosides can modulate conformational changes and aggregation of specific protein/peptides [71], even though it may affect protein/peptide oligomerization into amyloids in several ways with opposite effects (reviewed in [69]). Moreover, a higher membrane rigidity following increased cholesterol can be protective against any perturbation of membrane integrity and cell demise following aggregate growth at, or interaction with, the cell membrane [72,73].

Amyloid fibrils can be a source of toxic species in tissue

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

As reported above, amyloid fibrils are presently considered far less toxic to cells than amyloid oligomers and other prefibrillar aggregates, even though, in some cases, direct fibril cytotoxicity [12,13], or cytotoxicity associated with fibril assembly and growth on lipid membranes [74], has been described. The toxicity of fibrillar material can also result from fibril breakage with the generation of shorter fibrils [75] or leakage of toxic species [76]; the latter phenomenon has also been observed in vivo [25]. Previous data on the effects on 4′-iodo-4′-deoxydoxorubicin [77], and more recent findings on the solubilization of both transthyretin fibrils by doxycycline [78] and β2-microglobulin fibrils into toxic oligomers by a number of tetracyclines (S. Giorgetti, S. Raimondi, A. Relini, M. Bucciantini, K. Pagano, A. Corazza, M.C. Mimmi, M. Salmona, P. Mangione, F. Fogolari, L. Colombo, L. Marchese, A. Gliozzi, M. Stefani, G. Esposito, M. Stoppini & V. Bellotti, unpublished results) further reinforce the idea that amyloid fibrils, under a panel of different conditions, can be disassembled or, at least, can leak toxic oligomers.

Overall, these studies highlight the importance of amyloid fibrils in amyloid cytotoxicity and synaptotoxicity, not as directly responsible for the toxic insult, apart from the inflammatory reaction they can trigger in tissue, but possibly as providers of toxic species. This view implies that amyloid plaques in tissue could not be considered to always be protective by recruiting toxic species arising from peptide/protein misfolding; rather, they can also be potential sources of toxicity. Actually, recent data suggest that fibrils of the same protein/peptide grown from structurally different oligomers or deposited under differing conditions and, hence, with variable structures and stabilities, can display different cytotoxicity or ability to leak toxic oligomers. These data concern monomer/oligomer recycling within amyloid fibril both in vitro [79] and in vivo [80], the different stabilities of fibrils grown from the same monomer at different conditions [81] and the leakage of toxic oligomers around fibrillar deposits in vivo [25]. Accordingly, spontaneous fibril decomposition or fragmentation in tissue can be seen as a possible key determinant of amyloid cytotoxicity.

Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

In spite of the increasing awareness that mature amyloid fibrils can display a toxic potential, both in tissue and to cultured cells, amyloid oligomers, either grown from monomers or leaked from mature fibrils, are still considered the amyloid entities endowed with the highest direct cytotoxicity. Therefore, the ability of differing conditions to generate structurally different oligomers with variable cytotoxicity appears of particular importance for the role of these species as key players of cell/tissue impairment. In this regard, oligomer (and mature fibril) heterogeneity and polymorphism are undoubtedly key issues. As reported above, many data indicate that different conditions can affect variously the fold of the same peptide/protein, resulting in different ways the same peptide/protein misfolds, entering into aggregation paths where oligomers with different conformational features eventually growing into polymorphic mature fibrils are generated [51,82]. The same can be true for the aggregation, under similar conditions, of protein/peptide variants carrying specific amino acid substitutions or chemical modifications altering the physicochemical properties of their polypeptide chains with respect to those displayed by the unmodified counterparts [83]. Specific conformational features of the monomer can also determine the way it aggregates under suitable conditions and the toxicity of the aggregates themselves; besides being associated with the presence of specific amino acid substitutions, in some cases these conformational peculiarities can be much more subtle, as it has been shown for the Aβ42 peptide [84].

As indicated above, a number of recently reported studies have sought to assess the structure–toxicity relationship of amyloid oligomers and fibrils indirectly, for example by establishing a relationship between their stability and the ability to impair cell viability [85]. In the case of the Aβ peptides, the most studied model, many data support the idea that Aβ interaction with the cell membrane implies peptide conformational changes and is important in determining Aβ toxicity [86]; it could therefore be proposed that the toxicity of any Aβ species can be, at least in part, in relation to its ability to change structure on, or within, the cell membrane and, hence, ultimately, to its flexibility and relative stability. Similar conclusions have been drawn in a recent research on the relationship between the differing amyloid conformations at two different temperatures of the peptide encoded by the huntingtin exon-1 with an expanded polyglutamine stretch and the relative toxicity of either conformation to cultured cells and in animal tissue [87]. These data suggest that the same protein in different brain areas experiences varying conditions that modulate its stability, leading it to aggregate into fibrils with differing physical and biological properties. Finally, it has been reported that soluble 30–50 nm-sized annular α-synuclein oligomers are released by mild detergent treatment from glial inclusions purified from multiple system atrophy brain tissue; the authors showed that in such α-synucleinopathy the pathological aggregates can be a source of annular α-synuclein species at variance with recombinant α-synuclein, whose aggregates yielded only spherical oligomers after detergent treatment [88].

Our recent results on HypF-N, a bacterial protein not associated with any amyloid disease, further confirm and extend the generality of these considerations, providing clues on the structural features of amyloid oligomers and their relationship to toxicity. At different destabilizing conditions, HypF-N misfolds and generates two types of amyloid oligomer morphologically similar but with varying stabilities and structural features. These oligomers have been carefully characterized in terms of burial of hydrophobic residues and, as a consequence, density of packing and flexibility, as well as the extent of hydrophobic exposed area. The less stable oligomers grew into mature fibrils, whereas the more stable ones eventually assembled into stable curvy protofibrils with no further evolution. The two types of oligomer displayed differing cytotoxicities and abilities to interact with, to permeabilize and to cross the plasma membrane of exposed cells; the less compact, less stable and hydrophobic assemblies being substantially more toxic [43] (Fig. 1). These data establish a direct link between general structural features of prefibrillar oligomers of a protein and their ability to grow into distinct, stable amyloid assemblies, providing clues on the relationship between oligomer conformational features and their ability to stick to, disassemble and permeabilize the cell membrane and to impair cell viability. Taken together, these results can tentatively be generalized, proposing that the conformational, structural (exposure of hydrophobic surface and flexibility) and stability features of amyloid assemblies can remarkably affect the way and the extent the latter interact with the cell membrane, thus being a major determinant of cytotoxicity.

image

Figure 1.  The structural and biophysical features of amyloid oligomers dictate the extent of their interaction with the plasma membrane of exposed cells and of the resulting disassembly of membrane architecture and permeabilization. Highly flexible, less compact oligomers with extended exposed hydrophobic surface area interact in depth with the lipid bilayer. The interaction probably results in phospholipid pull-out with considerable membrane disassembly and nonspecific permeabilization. The latter is considered one of the key features of the impairment of viability in cells exposed to toxic amyloid aggregates.

Download figure to PowerPoint

Conversely, as pointed out above, the structural and biochemical features of the cell membranes also play an important role in modulating the cytotoxic potential of amyloid oligomers. Actually, a large number of studies have appeared in the last 10 years that have provided convincing data indicating that the ability of amyloids to grow on, to interact with, and to permeabilize cell membranes depends strictly on the biophysical features of the latter, including the curvature, compactness, rigidity and density of charge, primarily associated to membrane lipid composition (see above). This view is further supported by our preliminary data obtained using cultured cells either with normal, enriched or depleted in membrane cholesterol exposed to the two types of HypF-N oligomer, showing that the level of cytotoxicity of the same oligomer is modulated by the lipid composition and biophysical features of the cell plasma membrane (C. Cecchi et al., unpublished data). These and previously published results [71] can lead to the proposal that the notion of oligomer cytotoxicity must be considered relative, being the net result of the interplay between oligomer and membrane structural features determining the way and the extent of oligomer–membrane interaction and the severity of the resulting membrane structural perturbations. Apart from establishing a more complex link between oligomer/membrane structural features and the resulting cytotoxicity, these data also provide a rationale explaining the different vulnerability to the same amyloids of different cell types, either cultured or in tissue [89,90].

Concluding remarks

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References

The data reported in the last few years have made it increasingly evident that amyloid fibrils and their precursors grown from structurally different monomers of the same peptide/protein or from monomers misfolded at different environmental conditions can display differing structural features. Such a conformational polymorphism appears to be of great importance in determining the cytotoxic potential of amyloids. On this line of evidence, it is increasingly being recognized that amyloid fibrils, previously considered as harmless reservoirs of toxic oligomers can indeed be a source of cytotoxic species to exposed cells, depending on their structural and stability properties and the conditions of the environment where they grow and are deposited. Actually, increasing information supports the idea that mature fibrils can indeed be a source of toxic oligomers following their fragmentation by thermal motion or by interaction with disassembling surfaces or molecules; conversely, it can also be considered that the increased number of free ends in fragmented fibrils can recruit more efficiently misfolded monomers, hindering their assembly into toxic oligomers.

This view still assigns to the oligomers the role of direct cytotoxic entities; however, also in this case, structural polymorphism can play a key role in modulating the ability of these assemblies to interact with cell membranes and, hence, their toxic effects to exposed cells. Unfortunately, at variance with mature fibrils, obtaining information on the structural features of fibril precursors, in particular the early oligomeric assemblies, is made challenging by the heterogeneous, transient, unstable and highly flexible nature of the latter. Nevertheless, new experimental approaches are starting to unravel the structural features of these assemblies, providing information on the structure–toxicity link. Finally, the importance of the biophysical features of synthetic or biological membranes in determining the way amyloids can grow on, or interact with, them must also be taken into consideration. Overall, all these points can be considered as details of a highly incomplete, yet remarkably complex picture, where amyloid polymorphism and the cell membrane biochemical and biophysical features both contribute, although differently in differing systems, to the overall cytotoxicity of these assemblies. Such a view leads to the consideration of amyloid oligomer cytotoxicity as a relative concept rather than an inherent property. Apart from reconciling the amyloid fibril toxicity versus amyloid fibril safety alternative views, these data provide new clues to explain the molecular determinants of sporadic amyloid diseases, the variable susceptibility of different cell types to amyloid cytotoxicity and the lack of any direct relationship between amyloid load and the severity of the clinical signs. We are still at the beginning in the path to unravel the oligomer mystery. However, the present intense research warrants that in the near future we will be able to obtain solid knowledge in this field that will be of great value for rational drug design against oligomer cytotoxicity.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. General structural features of amyloid oligomers
  5. Cell membranes can promote protein misfolding and aggregation and are key targets of amyloid cytotoxicity
  6. Amyloid fibrils can be a source of toxic species in tissue
  7. Both oligomer/fibril and membrane biochemical and physicochemical features can modulate their reciprocal interaction and the resulting toxic effects
  8. Concluding remarks
  9. Acknowledgements
  10. References