Monitoring α-synuclein aggregation

Abstract

Synucleinopathies, including Parkinson's disease (PD), dementia with Lewy Bodies (DLB), and multiple system atrophy (MSA), are characterized by the misfolding and subsequent aggregation of alpha-synuclein (α-syn) that accumulates in cytoplasmic inclusions bodies in the cells of affected brain regions. Since the seminal report of likely-aggregated α-syn presence within the Lewy bodies by Spillantini et al. in 1997, the keyword “synuclein aggregation” has appeared in over 6000 papers (Source: PubMed October 2022). Studying, observing, describing, and quantifying α-syn aggregation is therefore of paramount importance, whether it happens in tubo, in vitro, in post-mortem samples, or in vivo. The past few years have witnessed tremendous progress in understanding aggregation mechanisms and identifying various polymorphs. In this context of growing complexity, it is of utmost importance to understand what tools we possess, what exact information they provide, and in what context they may be applied. Nonetheless, it is also crucial to rationalize the relevance of the information and the limitations of these methods for gauging the final result. In this review, we present the main techniques that have shaped the current views about α-syn structure and dynamics, with particular emphasis on the recent breakthroughs that may change our understanding of synucleinopathies.

1. Introduction

1.1. The spectrum of synucleinopathies

Synucleinopathies, such as Parkinson's disease (PD), dementia with Lewy Bodies (DLB), and multiple system atrophy (MSA), are a group of neurodegenerative disorders of unknown aetiology with very diverse clinical and pathological presentations. However, these diseases share certain standard features, the most significant being the aggregation of alpha-synuclein (α-syn), present in cytoplasmic inclusions bodies in the cells of affected brain regions (Papp and Lantos, 1992; Spillantini et al., 1997). These cytoplasmic inclusion bodies are diverse in nature and cellular localization across these synucleinopathies. Lewy bodies (LBs) are found in neurons of PD and DLB patients' brains, while glial cytoplasmic inclusions (GCIs) are detected in oligodendrocytes of MSA patients' brains. LBs are eosinophilic inclusions comprising filamentous structures in a dense core surrounded by a peripheral halo (Wakabayashi et al., 2007) containing numerous membranous components and dysmorphic organelles (Shahmoradian et al., 2019). GCIs are packed triangle or sickle-shaped, filamentous structures (Wakabayashi and Takahashi, 2006).

Neuronal and oligodendroglial pathologies are far more diverse than such a simple description suggests both in their nature or form and the number of anatomical stigmas they present. Behind the established terms, LBs and GCIs, there is, in fact, a myriad of inclusions, both intracellular and extracellular, more or less mature and more or less toxic. Still, all of them are characterized by misfolded, aggregated α-syn.

The neuronal pathology, known as Lewy pathology, is the most studied in PD. Although this pathology has various origins, most research focuses on the characterization of LBs in the lower brainstem nuclei, where they are well structured, with a classic core surrounded by a halo. Cortical LBs cannot be considered “classical” and are less structured (Sakamoto et al., 2002), and, overall, much less studied. This morphological heterogeneity of the two types of LBs is explained by their enrichment in α-syn. Although aggregated filamentous α-syn is a central (not principal) constituent of LBs and GCIs (Laferrière et al., 2022; Mccormack et al., 2016; Spillantini et al., 1997; Wakabayashi and Takahashi, 2006), only around two-thirds of LBs in the Substantia Nigra are composed of α-syn; and they represent no more than one-third of cortical LBs. Finally, the glial inclusions seen in MSA are little studied (except, of course, by pathologists); hence, the status of α-syn aggregates is less evident in these aggregate subtypes.

1.2. Synuclein aggregation leads to synucleinopathies

The cornerstone of aggregate formation is the ability of α-syn, a so-called amyloid protein, to aggregate (Bourdenx et al., 2017). Fig. 1A shows a schematic summary of the aggregation pathway of this protein. α-syn exists in various conformational forms. Indeed, it is present in at least two structural isoforms: a soluble, natively unfolded monomer and a membrane-bound, helix-rich form. Some authors have suggested the existence of an alternative, tetrameric, aggregation-resistant organization of the protein (Bartels et al., 2011). Isoforms can undergo dramatic structural changes, resulting in rich sheet-like assemblies that are the precursors of more complex amyloid forms of the protein. From a myriad of in vitro studies, we can deduce that the behaviour of α-syn follows a dynamic equilibrium in which the monomer aggregates. First into several types of small oligomeric species that sheet interactions can stabilize, and then into higher molecular weight, insoluble protofibrils that can polymerize into amyloidogenic fibrils by reassembling those contained in LBs. However, the exact mechanism governing the fundamental conformational change from the standard α-syn monomer to the pathological, disease-associated form remains elusive.

Fig. 1.

α-syn aggregation (adapted from (Bezard and Dehay, 2022). A) Schematic summary of the aggregation pathway of α-syn. Different monomeric species exist in a continuous equilibrium. Aggregation of α-syn usually involves the formation of rich b-sheet-like assemblies, which precede the formation of more complex fibrillated forms of the protein found in Lewy Bodies. B) Energy diagram illustrating protein structural transition. Unlike classic proteins, where different conformational forms are associated with relatively deep energy pits (left), α-syn structure oscillates continuously between various conformations with similar energy levels (right). C) Under special conditions, α-syn can adopt extremely stable forms, termed amyloid structure, which is then very hard to reverse. D) α-syn aggregation process is a highly complex process. Nonetheless, it is generally accepted that amyloid fibril formation is a process that relies on several stages: primary nucleation, growth, and proliferation by elongation and secondary nucleation. These phases are commonly subdivided into three depending on fibril mass: elongation, growth, and stationary.

When applied to α-syn, the notion of folding, or so-called “pathological” conformation, appears restrictive. α-syn is one of the most intrinsically disordered proteins. Its structure oscillates continuously between various conformations, perhaps on a continuum, ranging from a fully unfolded to a fully folded protein (Robustelli et al., 2018). This shape transition can be illustrated by an energy diagram, representing the shape's stability as a function of the energy required to maintain it (Fig. 1B). Imagine that a ball is pushed through this diagram. It will eventually end up ‘locked’ in a pit, the most bottomless pit, for which the free energy will be the lowest, a state of ‘rest.’ This happens to a standard ‘structured’ protein; it takes on the form in which the energy to hold it is lowest, the ‘deep pit’ (Schor et al., 2016). However, under physiological conditions, the conformational energy landscape that characterizes α-syn has no natural “deep pit”; all the energy pits are more or less equal in depth. The protein thus continuously ‘vacillates’ between the different energy states (Fig. 1B).

Other external factors can create ‘energy pits’. These correspond to a β sheet-like organization of the protein (Raskatov and Teplow, 2017) (Fig. 1C): they can be caused by pH changes, post-translational modifications, cleavages, or amino acid substitutions in the protein due, for example, to point mutations (over eight are known to date). The effect of these point mutations has been extensively studied in tubo experiments. Their exact role, however, is not yet well understood. Most of them are located in the N-terminal domain, and, by themselves, do not induce important changes into the conformation of the monomeric form, although they affect α-syn interaction with the membrane (Fares et al., 2014). It is tempting to posit that these mutations increase the self-assembly propensity of α-syn. While some of them indeed do (E46K, H50Q, or A53T), some others (G51D or A53R) seem to produce the opposite effect, thereby preventing aggregation (Villar-Piqué et al., 2016). Interestingly, it has been suggested that the fibres formed by each mutant variant may present structural and biophysical differences that would support the pathological role of the mutant. (Boyer et al., 2020; Ruggeri et al., 2020).

In its β sheet form, α-syn can self-aggregate into large amyloid stacks. These are incredibly stable and correspond to a bottomless energy pit, making the protein pathological. Each folding state has a corresponding energy pit (Fig. 1C): the deeper the pit is, the more difficult it is to reverse the protein's shape. While partially folded states can be changed with minimal energy expenditure, amyloid fibrils produce a bottomless pit, which will require considerable energy expenditure to reverse (and thus return to the unfolded state of the protein).

The formation of the amyloid fibril is a process that relies on several stages: primary nucleation, growth, and proliferation by elongation and secondary nucleation (Perrett et al., 2014). Primary nucleation is critical. It creates the first fibrils. Secondary nucleation can be more problematic. It can lead to an exponential increase in protofibrils and then fibrils. Once the assembly of monomers has started, the amyloid structure acts like a “black hole.” It attracts more and more monomers into the initiated superstructure, making the system increasingly stable in terms of energy, which can concomitantly catalyze the formation of new oligomers (Fig. 1D). In this state, the cell subject to this process is exploited. It will produce more and more α-syn to feed this “monster”.

The final fine structure of amyloid fibres likely depends on several factors. First, the several possible conformations of the monomer β sheets, are the starting point of the process. Then, multiple factors such as pH or particular cellular or extra-cellular environment will affect the aggregation process per se, giving rise to different species of α-syn aggregates (polymorphs) with variable toxicity and manifestations (De Giorgi et al., 2020). These different α-syn fibrils, will spread differently and produce different pathophysiological responses, depending on their particular properties (Lau et al., 2020). This fiber heterogeneity could be at the origin of the different spectrum of synucleinopathies we observe, such as PD, DLB, or MSA, each reflecting the growth and spread of the different amyloid species that would be specific to a given disease manifestation (Peelaerts et al., 2015).

While these insoluble aggregates represent the most stable and thus “final” forms of α-syn aggregates, increasing experimental evidence points to smaller soluble species, usually called oligomers, as the main spreading toxic species. These smaller aggregates, precisely due to their instability and smaller size, would be responsible for cytotoxicity and disease expansion (Emin et al., 2022; Kulenkampff et al., 2021). Be as it may, understanding and monitoring the aggregation process at every level is critical in synucleinopathy research.

1.3. Monitoring aggregation is central to synucleinopathy research

Since the seminal report by Spillantini et al. in 1997 (Spillantini et al., 1997), the number of scientific reports featuring “synucleinopathy” as a keyword has reached 84,804 on PubMed. Among those, the keyword “synuclein aggregation” accounts for 6433 papers (October 2022). Given the above, one understands that studying, observing, describing, and quantifying α-syn aggregation is of paramount importance, whether it happens in tubo (Table 1), in vitro (Table 2), in post-mortem samples (Table 3), or in vivo (Table 4).

Table 1.

In tubo methods.

Technique	Provided Info	Advantages	Disadvantages	Refs.
Circular Dichroism	Secondary structure (and tertiary to some extent)	Simple technique. Compatibility with vesicular dispersions, real-time structural information (kinetics)	Very sensible to light scattering. Limited to studies of one protein. Limited information	Weinreb et al., 1996; Kelly et al., 2005
FTIR	Secondary and tertiary structure	Easy to use. Very sensitive to b-sheets conformations	In tubo studies of one protein. Limited information	Calero and Gasset, 2005
soluble NMR	Strong structural information on soluble α-syn species	Very high structural information. Key for naturally disordered proteins such as α-syn	Need a soluble pure and often isotope-labelled sample. Complex and expensive technique	Eliezer et al., 2001
Solid state NMR	Strong structural resolution on insoluble α-syn fibrils	No need for solubility. Great 3D resolution. Detailed short-range inter-atomic distance	Difficult sample preparation, spectra collection, and analysis. Expensive technique	Heise et al., 2005; Tuttle et al., 2016
“Classic” Diffraction techniques	Diffraction pattern. The tertiary structure of fibrils	Unambiguous structural information about tertiary structures	Limited resolution and utility with non-crystallo-genic samples.	Sunde et al., 1997; Serpell et al., 2000
Microcrystal Electron Diffraction	Diffraction patterns without the need for macroscopic crystallization	Overcome the inability of fibrils to form macro-crystals. Great resolution	Complex equipment is needed.	Shi et al., 2013; Rodriguez et al., 2015
“Classic” EM	Morphology of the fibrils and their size distribution	Relatively accessible. Versatile technique. Possibility to combine with immunochemistry	Limited resolution to resolve relevant structural polymorphisms	Spillantini et al., 1998; Lashuel, 2020
Cryo-EM	Very high resolution on α-syn fibrils	Impressive resolution capabilities. Resolve different fine structural properties of the fibres	Expensive and relatively complex technique. Not readily available	Guerrero-Ferreira et al., 2019; Schweighauser et al., 2020
AFM	Morphology, size distribution. and mechanical properties of fibres	Information about heights, periodicities, mechanical strength and. Kinetic information	Need for synthetic surfaces. May influence the native properties of fibres	Goldberg and Lansbury Jr, 2000; Watanabe-Nakayama et al., 2020
ThT fluorescence	Formation of b-sheet structures (amyloids). Fibril formation kinetics through fluorescence.	Great enhancement of fluorescence upon binding to b-sheets. Real-time monitoring of aggregate formation.	May affect the aggregation process Fluorescence intensity depends on the polymorphs. Not always sensible	Levine III, 1993; De Giorgi et al., 2020
Site-specific labeling	Aggregation monitoring through fluorescence imaging.	Small fluorophores attached to specific and non-perturbing regions. Region-specific information upon aggregation. Through FRET or ESIPT.	Needs a relatively complex chemical label of α-syn before its use.	Shimogawa and Petersson, 2021

Table 2.

Live cells methods.

Technique	Provided Info	Advantages	Disadvantages	Refs.
cell-NMR	α-syn conformation inside cells	Structural information in more relevant cellular environments	Need a soluble pure and isotope-labelled sample. Complex and expensive technique	Selenko and Wagner, 2007; Theillet et al., 2016
ThS	Fluorescence increases upon binding to b-sheet structures	Direct binding to amyloid fibres. Useful as a confirmatory technique for amyloid presence	High backgrounds. It may present specificity problems	Luk et al., 2009; Gaur et al., 2021
Site-specific labelling	Fluorescence imaging of pre-labelled seeds	Cell-permeable, and non-perturbing probes. Information on uptake and interaction with cells	Restricted to labelled pre-formed seeds	Haney et al., 2016; Jun et al., 2019
Protein fusion fluorescence	Direct fluorescence imaging of proteins fused to α-syn	The direct expression on living cells. Quantitative aggregation information through FRET	The bulky probe can alter the native conformation and aggregation process	McLean et al., 2001; Yamasaki et al., 2019
BiFC	Emission of fluorescence by two protein fragments when they are in close proximity	Spatial and temporal resolution for aggregation. Suitable for early oligomers	Bulky probes. Fluorescence potentially not arising from just dimerization	Outeiro et al., 2008, Bae et al., 2014

Table 3.

Post-mortem techniques.

Technique	Provided Info	Advantages	Disadvantages	Refs.
FTIR	Secondary and tertiary structure by IR absorption. Presence of amyloid fibrils	Presence of α-syn inclusion and information about its environment	Basic information. Mainly as an orthogonal technique	Araki et al., 2015; Shahmoradian et al., 2019
EM	Visualization of α-syn in its context with immuno-electron-dense tags	Visualization of α-syn in its context. Identity of intracellular inclusions	Complex sample preparation. Limited fine fibril resolution	Baba et al., 1998, Bourdenx et al., 2020
Immunohistochemistry	Visual localization of the desired antigen thanks to immunologic specificity.	Direct visualization of the antigen in its context.	Limited by Abs specificity to discern different aggregates	Altay et al., 2022
Western Blot	Separation by mass using electrophoresis combined with immunodetection.	Common method. Separation by mass of species present in a sample. Immunodetection by different Abs	Non-quantitative. Limited by antibody specificity. Mass by itself does not tell much about an aggregate	Peng et al., 2018
AS-PLA	Informs on aggregated protein through an amplification signal of two close-distance antibodies	Good sensitivity towards early aggregates. Recognition of early aggregates in previously not recognized regions	Cannot fundamentally between oligomers and other types of higher-order aggregates	Roberts et al., 2015; Bengoa-Vergniory et al., 2020
ELISA	Quantification of immobilized species by double antibody binding + enzymatic reaction	Fast and relatively easy technique. Quantitative. Powerful combined with previous separation methods	Used Abs are critical. Limited sensibility. Previous sample preparation is critical	Lassen et al., 2018
Purification methods	Obtention of protein fractions from tissue homogenates depending on the shape, size or density of the species.	Straightforward methods. Powerful information when combined with posterior characterization of fractions.	Need of posterior characterization of the fractions	Laferrière et al., 2019
Digestion methods	Informs about aggregate nature through protein solubility/insolubility and/or proteolytic profile	Evaluation of fibrilization nature on digestion-resistance terms.	Limited to “solvent resistance” information	Kushnirov et al., 2020

Table 4.

Animal models of α-syn aggregation.

Technique	Provided Info	Advantages	Disadvantages	Refs.
Protein construct fluorescence	In vivo or post-mortem fluorescence. Information about the aggregate formation and its localization	Generation of transgenic animals expressing the fused protein. Possibility of in vivo and post-mortem analysis	Bulky probe. Can alter the native conformation, its behaviour, and the aggregation process	Kaminski Schierle et al., 2011; Osterberg et al., 2015
BiFC	In vivo or post-mortem fluorescence about oligomer formation and propagation	Expression specific at different zones. Oligomerization-specific fluorescence	Bulky probes. Fluorescence potentially not arising from just pure dimerization	Dimant et al., 2013; Kiechle et al., 2019

In animal models.

The past few years have witnessed tremendous progress in understanding aggregation mechanisms and identifying various polymorphs. In this context of growing complexity, it is of utmost importance to understand what tools we possess, what information they provide, and what context they may be applied to. Nonetheless, it is also crucial to rationalize the relevance of the information and the limitations of these methods for gauging the final result. In this review, we present the main techniques that have shaped the current views about α-syn structure and dynamics, with particular emphasis on the recent breakthroughs that may change our understanding of synucleinopathies.

2. Semantic-driven review organization

Synucleinopathies research involves various testbeds, i.e. experimental modalities, and the aggregation monitoring methods may apply only to some but not all modalities. We, therefore, propose a review organized in ascending order of model complexity, from in tubo to in vivo approaches. Of course, some techniques have applications in several models and will appear repeatedly. Thus, the first part is dedicated to the various in tubo structural and kinetic techniques that have shaped our understanding of α-syn aggregation and the relations between fine protein structure and pathogenicity. Then, we move towards more complex models, i.e., in vitro cell cultures and, ultimately in vivo (in animals) or post-mortem tissues. Each of these modalities provides specific information and proposes specific challenges to track and monitor α-syn aggregation. Before diving into the presentation of the various techniques and due to the often-confusing terms found in the literature (Bezard, 2022), we would like to explicit the meaning of some terms in the present review:

In tubo suggests an experiment performed in a liquid environment with neither a living cell nor necessarily a physiological buffer.

In vitro reports experiments executed in the presence of living cells which could be cell lines, primary cultures, or differentiated pluripotent stem cells.

In vivo describes an experiment performed on a living organism. Such an organism could be a model organism - such as flies (Drosophila melanogaster), worms (Caenorhabditis elegans), fishes (zebrafish) - or mammals - such as mice, rats, and non-human primates. In most instances, in vivo will mean ‘in living mice”. When performed in another species, it will be indicated.

Post-mortem suggests an experiment performed on biological tissues (here, brain tissue in most cases). The tissue originates from human subjects, patients, or animals previously processed through an in vivo experiment. The nature of the tissue may vary from fresh-frozen (i.e., frozen after death – humans or euthanasia - animals) to fixed cryopreserved frozen (animals) or paraffin-embedded formalin-fixed tissues (most often typical appearance of human post-mortem tissues used in anatomo-pathology).

Recombinant means that the protein in question, here α-syn, has been produced in a bacteria genetically engineered to this end.

Oligomer: The scientific literature describes various early prefibrillar α-syn oligomeric species, coined “oligomers”. They differ from each other in structure, molecular weight, and morphology. Oligomers come in a wide range of molecular weights, with variable levels of secondary structure (ranging from mainly α-helical to β-sheet with different amounts of disordered regions) and exposed hydrophobic areas. Operationally, these oligomers are sometimes termed soluble as opposed to insoluble fibrils. This review uses this operational definition, although we reckon that a more accurate description would provide the sedimentation coefficient since all proteins can be pelleted down if centrifuged rapidly enough for a sufficiently long time.

Fibril: Apart from α-syn oligomers, other (late) conformers associated with synucleinopathies include protofibrils and amyloid fibrils. The fibrillar form of α-syn is mainly located within intracytoplasmic inclusion bodies, i.e., LBs and GCIs are not found free in cytosol and often exhibit a weight in megadaltons.

Polymorphs: During amyloid stacking, α-syn can adopt several distinct structural folds, resulting in fibril polymorphs. Together with their specific fold, their supramolecular properties are inherited and propagated as the polymorphs self-replicate by templated growth.

3. Structural and aggregation-monitoring techniques

3.1. Basic spectroscopic techniques

The process of α-syn aggregation has primarily been studied in cell-free systems. Although non-directly extrapolable, they have contributed much to our basic kinetic and aggregation-mechanisms-related knowledge. Spectroscopical biophysical techniques, such as Circular Dichroism and Fourier-Transformed Infra-Red, are relatively simple techniques that provide helpful essential structural pieces of information. They have been mainly used as complementary and orthogonal techniques.

3.1.1. Circular dichroism

Circular dichroism (CD) is a popular technique for characterizing and following α-syn fibrillization in cell-free systems. It informs about the secondary structure by measuring the difference in absorbance of right and left-circularly polarized light, usually reporting the ellipticity (θ) at different wavelengths. The most common variant of CD in protein study works in the far-UV (180–240 nm), where the signal depends on the environment of the peptide bond, directly related to the secondary structure of α-syn. By knowing the base spectra of the different α-syn forms, it is possible to quantify the relative concentration of each of the species at each time-point in the same sample based on the contribution of each of the species on the spectra (Calero and Gasset, 2005; Kelly et al., 2005). First structural studies on human α-syn used CD to infer the protein structural conformation, suggesting that the protein acquires a random coil structure in solution (Weinreb et al., 1996). Subsequent studies used different forms of α-syn to characterize its structural diversity under various conditions: Davidson et al. used recombinant α-syn to study the effects of lipid binding on the secondary structure of α-syn, observing a shift from unstructured to a highly enriched α -helical conformation (Davidson et al., 1998). Serpell et al. used a set of α-syn strains to study secondary structure upon fibril formation using diffraction techniques and CD. Authors showed a transition from disordered monomer to β-sheet structure upon aggregation in all α-syn species (Serpell et al., 2000). Since then, the CD has become a handy tool for studying and monitoring α-syn structure in in tubo studies (Burré et al., 2013; Rovere, 2019). In recent years, vibrational circular dichroism (VCD) has emerged as an alternative that provides more specific and extensive information about the structure of the protein. In contrast to classic CD, in VCD, the signal arises from vibrational transitions of chemical bonds in response to infrared radiation (Polyanichko et al., 2011). Applied to the study of α-syn, it offers a better sensitivity towards fibrils, providing insights into the size and the higher-order morphology of the fibrils (Van de Vondel et al., 2018) or the possibility of distinguishing fibrillar structures belonging to different levels of fibrillar maturation (Martial et al., 2019).

3.1.2. Fourier-transform infrared

Fourier-transform infrared (FTIR) is an absorption spectroscopy technique based on different vibration modes of hetero­atomic bonds in response to infrared radiation. FTIR spectra are conventionally represented in terms of absorbance intensity as a function of wave numbers in units of cm¹ (Calero and Gasset, 2005). The FTIR-amyloid study usually focuses on the 1700 to 1600 cm⁻¹ range, a region known as the amide band. In this region, most of the signal arises from stretching the amide carbonyl group, whose vibration frequency strongly depends on its particular electronic context. This gives a particular sensitivity towards β-sheet structures and their relative conformation, allowing to distinguish different β-sheet-containing species (Froula et al., 2019; Sarroukh et al., 2013; Zandomeneghi et al., 2004). FTIR has been routinely used as a complementary technique in saline solution for the study of the structural characteristics of different α-syn species (Biere et al., 2000), oligomeric α-syn species (Hong et al., 2008), and fibrillar α-syn polymorphs (Bourdenx et al., 2020; Bousset et al., 2013)).

More recently, FTIR has been applied to post-mortem human tissue samples. In 2015, Araki et al. used the more radiative-powerful synchrotron radiation FTIR to characterize the β-sheet content in LB of a PD patient. They observed a high content of β-sheet structures in the halo, similar to the β-sheet conformation of in tubo recombinant α-syn fibrils, while the core presented a substantial amount of proteins but with different secondary structures (Araki et al., 2015). In another relevant human-derived tissue study, Shahmoradian et al. used high-definition FTIR imaging with correlative immunofluorescence staining in LB of PD patients. Authors exploited that the presence of different biomolecules, such as lipids, modifies the absorption of infrared radiation to confirm the presence of lipids within the α-syn inclusions (Shahmoradian et al., 2019). Finally, our team leveraged FTIR differential β-sheet sensitivity while analysing patient-derived α-syn. We characterized the α-syn secondary structures in fractionated α-syn extracts from PD patients' mesencephalon, focusing on fractions containing the LBs and fractions supposedly containing soluble α-syn (Bourdenx et al., 2020). While the former exhibited secondary structures typical of large aggregates, the latter contained a minute proportion of smaller aggregates characterized by differential secondary β-sheet conformations unravelled by FTIR. Unexpectedly, we showed that, in nonhuman primates, a small amount of these singular α-syn aggregates was as toxic as larger amyloid fibrils present in the fractions containing the LBs (Bourdenx et al., 2020), thus reinforcing the need for (i) preclinical research in this species and (ii) in-depth characterization of secondary β-sheets structures, notably of patient-derived materials.

3.2. Advanced structural techniques

The insolubility, heterogeneity, and non-crystalline nature of amyloid fibrils have hampered a more detailed structural study with “classic” structural techniques such as X-ray diffraction or soluble nuclear magnetic resonance (NMR). However, more complex structural techniques are vital to understanding the role of species involved in synucleinopathies. Recent years have seen a revolution in structural biology techniques (Iadanza et al., 2018), mainly in cryogenic electron microscopy (cryo-EM) and solid-state NMR (ssNMR) fields, providing near-atomic-resolution structures of amyloid fibrils. They have contributed to the emerging theory of fibril polymorphism as a basis for clinical and pathological heterogeneities (Lau et al., 2020; Shahnawaz et al., 2020; Shrivastava et al., 2020; Tanaka and Komi, 2015). These techniques pave the way for a refined understanding of the different polymorph species, their involvement in diseases, and their interactions with other proteins and small molecules as part of the aggregation process.

3.2.1. Nuclear magnetic resonance

Nuclear magnetic resonance (NMR) is a powerful spectroscopic technique that has shed much information on α-syn atomic structure. It provides structural information based on the oscillatory response of nuclei with non-zero spins. When exposed to a magnetic field, spin transitions in these nuclei can be induced by resonant excitation through radiofrequency irradiation. The NMR transition frequencies are sensitive to several orientation-dependent anisotropic interactions, potentially providing vibrant three-dimensional information about covalent bonding, internuclear distances, and electric field gradients (Reif et al., 2021). This makes NMR one of the few techniques well-suited to structurally resolve intrinsically disordered proteins, such as α-syn. Besides, it provides information about specific local conformational structures that encode biological functions in health and disease (Salmon et al., 2010). In the early 2000s, soluble NMR significantly contributed to our comprehension of the intimate structural identity of α-syn in its different conformations. In 2001, Eliezer et al. used NMR to characterize the structure of a recombinant α-syn in a free solution and after binding to membranes (Eliezer et al., 2001). They showed that α-syn protein is mainly unfolded in solution but with the propensity to adopt an α-helical structure in the N-terminal while transitioning towards an α-helical state in the presence of lipidic membranes (Eliezer et al., 2001). Then, further studies shed more light on the structural dynamics of α-syn: Chandra and colleagues showed that the α -helical state upon membrane binding comprises, in fact, only the first 98 N-terminal residues, which are structured in two α -helical regions interrupted by a short break, while the 42C-terminal residues retain a random coil configuration (Chandra et al., 2003). Since then, soluble NMR has remained a powerful method for studying the α-syn monomeric conformation and monitoring its conformational changes (Bodner et al., 2009; Fernandez et al., 2004).

Soluble NMR is not limited to saline solution; the so-called in-cell-NMR is used to study α-syn in more relevant cellular environments. The fact that NMR proteins are radiolabelled can be exploited to create “NMR-visible” proteins in “NMR-invisible” intracellular environments (Selenko and Wagner, 2007). In a remarkable work, Theillet et al. used electroporation to introduce 15 N isotope-enriched α-syn into non-neuronal and neuronal cells. They showed that the dynamic and disordered nature of monomeric α-syn is fundamentally similar to that in tubo but presents amino-terminal acetylation and the overall adoption of more compact structures (Theillet et al., 2016). Unfortunately, NMR presents some limitations. Besides not being a readily available technique, it has a relatively low sensitivity. It requires the use of high-purity samples, usually isotopically enriched proteins, a characteristic that prevents, for instance, from studying native or amplified patient-derived extracts, primarily because of the reported (and concerning) amplification inability to recapitulate the seed structure (Lövestam et al., 2021).

The use of classic NMR is limited by the insolubility of many relevant α-syn aggregated forms. Solid-state NMR (ssNMR) represents a powerful approach to overcoming this limitation. This variant uses highly concentrated protein assemblies forming a solid to semi-solid sample. This sample is then studied under a so-called magic-angle spinning to average the spatial anisotropic contributions and take advantage of local order in fibrils (Habenstein and Loquet, 2016). This variant has been extensively used during the last two decades in the study of aggregated protein structures to overcome the solubility of these relevant species (Heise et al., 2005; Lv et al., 2012; Vilar et al., 2008).

Nonetheless, ssNMR is a complex technique, and characterization at the atomic scale of aggregated species has remained elusive until recently (Meier et al., 2017). Progress in sample preparation and more sophisticated ssNMR methods and analysis have enabled the determination of high-resolution 3D structures of amyloid fibrils. In 2016, Tuttle et al. published the first high-resolution 3D structure of a human α-syn pathogenic fibril showing a Greek-key topology and paving for a better understanding of these fibrils' behaviour (Tuttle et al., 2016). Despite its impressive resolution, ssNMR remains a complex and not always accessible technique. For example, to achieve this impressive resolution, authors had to use samples highly enriched with “NMR-visible isotopes”, collect 68 multidimensional spectra and perform complex computational analysis.

Despite cryo-EM increasing popularity for super-resolve fibril polymorphs structure in an arguably more straightforward manner than ssNMR, this technique is far from having said its last word in the field. For example, cryo-EM can poorly resolve sidechain contacts and their possible interactions between fibres, which likely play a role in cytotoxicity and transmission. Contrary to cryo-EM, the accumulation of short-range inter-atomic distance information inherent to ssNMR allows for a much better resolution and understanding of the role of these elements (Takamuku et al., 2022). Both techniques thus have a complementary role in the refined understanding of polymorph pathology and delicate structure influence.

3.2.2. Diffraction techniques

Diffraction techniques are based on irradiating a high fibrillar content protein sample with a beam of X-rays or electrons, followed by recording several diffraction patterns resulting from the scattering of incoming radiation. The fact that amyloid fibrils do not form homogeneous 3D crystals limits the resolution of diffraction techniques, but they still have been profusely used for studying the amyloid fibril structure. They provided strong evidence for the so-called cross-β structure (Serpell et al., 2000; Sunde et al., 1997). Since then, both X-ray diffraction and electron diffraction have been used as characterization or orthogonal techniques. The latter possesses the advantage that the region where diffraction has been collected can also be imaged, allowing the geometrical relationship between the sample and diffraction pattern (Morris and Serpell, 2012).

In 2013, Shi et al. introduced Microcrystal Electron Diffraction (MicroED) as a remarkable alternative to classic diffraction techniques. It uses tiny three-dimensional crystals in a transmission electron cryo-microscope, irradiating with a reduced electron beam. It provides very high resolutions while overcoming the lack of crystal formation needed for classic diffraction techniques (Shi et al., 2013). In 2015, Rodriguez et al. used MicroED to obtain a fully refined structure with 1.4 Å resolution of the “preNAC” and “NACore” of α-synuclein, a fragment known to be critical in protein fibrillization (Rodriguez et al., 2015).

3.2.3. Imaging techniques

Imaging techniques such as electron microscopy (EM) or atomic force microscopy (AFM) permit visualizing higher molecular weight oligomers, fibrils or aggregates for deciphering their differential morphologies. Even though they are not quantitative, they provide critical structural and morphological information.

3.2.3.1. Electron microscopy (EM) and cryo-electron microscopy (cryo-EM)

EM has been a game-changing technique that grounds our understanding of synucleinopathies (Baba et al., 1998; Goldsbury et al., 2011; Lashuel, 2020; Spillantini et al., 1998). The possibility of immuno-labelling α-syn with electron-dense tags provides information about protein identity, combined with super-resolution lengths and morphologies. Interestingly, EM provided the first hints on the fibrillar composition of LBs in the 60s (Duffy and Tennyson, 1965) while immunolabelled EM shed light on the α-syn composition of the fibrils (Baba et al., 1998; Spillantini et al., 1998). Since then, although it lacks the resolution for a more refined analysis, EM has become a standard tool to study structural properties or simply the presence of α-syn aggregates in in tubo (Shrivastava et al., 2020), in vitro (Mahul-Mellier et al., 2020; Volpicelli-Daley et al., 2011) and in post-mortem samples (Fig. 2A) (Baba et al., 1998; Recasens et al., 2014; Shahmoradian et al., 2019).

Fig. 2.

Relevant examples of α-syn structural imaging techniques. A) Illustrative image of LB from the human brainstem using EM. Here, the LBs present the characteristic dense core and radiating filaments immunolabelled against human α-syn and visualized by the immunoperoxidase method (Baba et al., 1998). B) Cryo-EM maps of filaments purified from the putamen from 4 different MSA cases. Cryo-EM allows almost atomic resolution of fibrils, revealing relevant polymorphisms (Schweighauser et al., 2020). C) Cryo-ET imaging of α-syn aggregates seeded by recombinant PFFs in primary mouse neurons in a 1.4 nm thick tomographic slice. Red arrowheads mark fibrils. Right top magnified view of fibrils (red arrowheads). Right top magnified view of neurofilaments (blue arrowheads) (Trinkaus et al., 2021). D) AFM image of an in tubo-formed helical α-syn fibril. The AFM analysis provided the height profile and periodicity of the fibril (Kumar et al., 2017). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Remarkably, these last years have seen an impressive increase in the resolution capabilities of cryo-EM and image processing techniques, achieving resolutions close to the atomic scale (Kühlbrandt, 2014). Several cryo-EM studies have focused on studying the distinct structures of tubo-produced recombinant α-syn polymorphs, achieving resolutions ranging from 3 to 3.7 Å. These studies have used either full-length α-syn (Guerrero-Ferreira et al., 2019; Li et al., 2018a; Li et al., 2018b) or modified or mutant α-syn (Boyer et al., 2019; Guerrero-Ferreira et al., 2018; Sun et al., 2020; Zhao et al., 2020) to shed light on the relation between the different fine atomic structure of the polymorphs and its fibril-forming mechanisms, seeding capacity and cytotoxicity.

Although of outstanding scientific, technical, and ultimately clinical interests, the question of to what extent these in tubo-formed fibrils resemble those formed in animal, or human brains has yet to be solved. A similar question about Tau filaments suggested that fibrils purified from human post-mortem brains differed from those generated in tubo (Zhang et al., 2019). Burger et al. adopted a different approach. They used homogenates from PD and MSA brains to seed recombinant monomeric α-syn using a protein misfolding cyclic amplification (PMCA) assay (see Section 6.1). The authors resolved the resulting structures using cryo-EM, reporting different polymorphic structures for PD and MSA. (Burger et al., 2021). Although they showed that the fine structure of PMCA-generated fibrils is different from those generated purely from recombinant seeds, the question of to what extent the structure of these fibrils totally reproduces the characteristics and fine structure of in vivo-formed fibrils remains open. Fan et al. used a similar experimental framework to cryo-resolve fibrils amplified using PMCA from the cerebrospinal fluid (CSF) of prodromal and late symptomatic stage post-mortem PD patients. The prodromal sample was obtained from an initially healthy patient who converted to PD 7 years later. Cryo-EM studies showed a modest yet present structural polymorphism between early (prodromal) and late-stage (clinical) disease fibrils. Authors next used these polymorphs to seed the formation of recombinant α-syn fibrils, observing that the resulting recombinant fibrils seeded by the PD late-stage fibrils presented a greater capacity to seed endogenous α-syn aggregation, measured by quantification of phosphorylated α-syn-immunopositive aggregates in primary rat cortical neurons (Fan et al., 2021).

Interestingly, a 2020 study managed to atomically resolve fibrils extracted directly from fresh-frozen brain regions of several individuals with MSA and DLB using sarkosyl solubilization (Schweighauser et al., 2020) (Fig. 2B). Authors found two different types of fibrils in MSA brains, together with non-proteinaceous molecules. Furthermore, they observed that these fibrils were distinct from those found in the brains of DLB patients and those formed in tubo using pathological seeds and recombinant α-syn. Recently, the same group managed to extract and cryo-resolve fibrils from the brain of PD (one individual), PDD, and DLB patients' brains. Authors found all these Lewy pathology fibrils are made of a single proto-filament, which they termed Lewy fold, overall different from the previously reported MSA folds (Yang et al., 2022). These findings are significant in two ways. On the one hand, they once again highlighted the hypothesized relationship between specific fibril polymorphism and specific fibril toxicity (i.e., one polymorph would give rise to one “disease” or pathology phenotype). On the other hand, the authors showed how the fine structure from fibrils purified from the brain presents different structures than those grown in tubo using recombinant α-syn seeded with previously extracted fibrils (Lövestam et al., 2021).

Trinkaus et al. adopted a slightly different strategy in a competitive field of research. They used cryo-EM to image neuronal α-syn inclusions directly in primary mouse cortical neurons preserved by vitrification. Although this study does not achieve the structural richness of fibril-purified studies, it offers a unique high-resolution view of aggregates inside cells. Authors provoked aggregation by culturing the cells either with α-syn preformed fibrils (PFFs) or aggregates purified from MSA patients. Although with some differences, they obtained α-syn fibrils interspersed by numerous cellular organelles (Fig. 2C). Next, they used gold-labelled PFFs to study aggregate seeded growth in a neuroblastoma cell line (SH-SY5Y cells), observing a unidirectional fibril growth seeded mainly by the small fibrils (Trinkaus et al., 2021).

3.2.3.2. Atomic force microscopy (AFM)

AFM is based on scanning the sample with a tip probe, creating a topography image of the proteins, with information about the growth, directionality, and morphology of individual fibrils (Fig. 2D) (Goldsbury et al., 1999). In the field, it has been used to image and study the structural properties and the heterogeneity of α-syn individual oligomers and fibrils formed under different conditions providing both qualitative and quantitative data (e.g., heights and periodicities) (Ding et al., 2002; Goldberg and Lansbury Jr, 2000; Kumar et al., 2017; Li et al., 2018b; Sidhu et al., 2018). It has also been used to image the continuous formation of fibrillar species (in situ AFM), although it is likely that the electrical or chemical properties of the selected substrate might bias the assembly process (Jeong et al., 2013; Moores et al., 2011). Last years have seen the use of High Speed-AFM, allowing the visualization of structural dynamics of individual α-syn species with a much-improved time resolution. It permits imaging of the conformation dynamics of a monomeric and dimeric α-syn (Zhang et al., 2018b) and the fibril growth (Watanabe-Nakayama et al., 2020). The effect of fluorophore upon aggregation pathway is a problem that is usually ignored in fluorescence microscopy. In an exciting approach, Jadavi et al. used AFM-STED correlative microscopy to study fluorophores' effect on in vitro α-syn fibrilization. (Jadavi et al., 2022).

4. Fluorescence-based methods

Fluorescent labelling methods are valuable tools to track α-syn aggregation due to their hypothesized sensitivity towards different α-syn species and their potential ability to provide aggregation information in various models. Choosing a correct fluorescent probe for tracking aggregation is paramount to obtaining meaningful results. Beyond the classic requirements when working with biological samples, such as wavelength, fluorescence intensity, photo-stability, and bio-stability, an ideal aggregation-tracking fluorescent probe should undergo a significant fluorescence change upon binding to a specific structure or upon a change in its microenvironment (Gao et al., 2021; Haney et al., 2016; Shimogawa and Petersson, 2021). Moreover, it is crucial to understand and consider whether (and if yes, how) probe binding alters the functionality of α-syn or modifies its aggregation kinetics or final structure. While it is always a good policy to validate labelling in structure-sensitive experiments like CD, EM, or cell-based fibril seeding assays (Mučibabić et al., 2016), these techniques may not be accurate enough to detect more subtle structural changes.

4.1. Amyloid dyes

The so-called amyloid dyes include, notably, the symmetrical sulfonated azo dye Congo Red (CR) and the benzothiazole dyes thioflavin S (ThS) and thioflavin T (ThT). This family of organic molecules is classically used for studying amyloids due to their affinity and optical changes upon binding to the β-sheet structures present in amyloid structures (Reinke and Gestwicki, 2011). The binding sites of ThT or CR upon the β-sheet structure are reportedly different (Zhang et al., 2018a), explaining why specific samples would produce signals upon CR presence but not with ThT (De Giorgi et al., 2020). The practical outcome is that one can theoretically develop novel dye analogues that can be used to spectroscopically quantify the amount of amyloid richness and variety in a sample.

When bound to β-sheet structures, the optical changes of these compounds depend on different mechanisms. In the case of Congo Red binding, changes can be explained as aromatic rings become torsionally restricted upon binding. Although herein included in the fluorescence section, and indeed fluorescent (Yakupova et al., 2019), Congo-Red stained amyloids have been usually detected upon examination with polarized light in what has been traditionally called apple-green birefringence. (Housmans, 2021). CR has been central in early amyloidogenic research to study aggregation in vitro (Conway et al., 2000) and stain post-mortem tissue (Thiruchelvam et al., 2004). However, its use has been limited more recently, as studies show a problematic nonspecific binding or a likely interference with processes of protein misfolding and aggregation (Frid et al., 2007).

Regarding the thioflavin family, ThS can be used in cellular inclusions (Lázaro et al., 2016; Luk et al., 2009) and post-mortem fractions (Recasens et al., 2014) to confirm the presence of β-sheet structures in an easy way. However, the fact that it does not undergo a shift in the excitation or emission spectra upon binding results in high background fluorescence, making it suboptimal to be used in quantitative measurements (LeVine III, 1999).

ThT deserves special attention due to its widespread use and general reliance as a probe for α-syn fibrillation to the present day. In this case, fluorescence increase upon binding is thought to be due to the loss of rotational freedom between the benzothiazole and aniline rings, hence the nickname of rotor dyes for this family. Briefly, in the unbound state, the ultrafast twisting dynamics around such C—C bond are thought to cause rapid self-quenching of the excited state, resulting in low emission. However, the rotational freedom is restricted upon binding to fibrils, and the excited state is readily populated (Biancalana and Koide, 2010; Kuznetsova et al., 2016). As it binds to β-sheet structures, it theoretically enhances its fluorescence linearly with amyloid fibril mass. Although its use in brain sections is highly limited due to its weak fluorescence emission and potential background interference from other cellular components (Gao et al., 2021; Gaur et al., 2021), ThT fluorescence has become the most common proxy for aggregate formation or presence (Conway et al., 2000; Levine III, 1993).

This reliance on the increase in ThT signal as the aggregation probe cornerstone presents many limitations, even leading to false assumptions. The question of ThT impact upon aggregation itself is seldom considered. Some studies pointed out that ThT interactions with α-syn disordered monomer may promote amyloid aggregation (Coelho-Cerqueira et al., 2014), while others claimed that the effect of ThT on aggregation at moderate concentrations is negligible (Xue et al., 2017). Kumar et al., however, found that the sole presence of ThT during fibrilization modulated the fibrillar conformation, balancing the final fibril population towards different structures. Besides, these ThT-formed structures presented higher in vitro toxicities, reducing the viability of mammalian HEK-293 T cells (Kumar et al., 2017).

Far more concerning is the dogmatic reliance on the ThT signal for the selection/production of recombinant α-syn PFFs. However, polymorph diversity may reflect the pathological diversity (and hence possibly the clinical spectrum) (Lau et al., 2020; Shahnawaz et al., 2020). Sidhu et al. showed that in vitro formed β-sheet structural polymorphs present different ThT binding sites, directly affecting fluorescence intensities (Sidhu et al., 2018). Furthermore, Strohäker et al. observed that PMCA-amplified fibrils seeded by PD or MSA brain extracts present different morphological features supporting their variable ThT fluorescence (Strohäker et al., 2019). Using a similar approach, Shahnawaz et al. observed that MSA-seeded fibrils exhibit significantly lower ThT fluorescence than those amplified from PD samples (Shahnawaz et al., 2020).

Further exploring this puzzling question, De Giorgi and colleagues demonstrated how, while launching PFF formation from monomeric α-syn, a variety of α-syn fibrils are stochastically produced. Some newly generated PFFs are ThT-positive while others are ThT-negative despite exhibiting a clear β-sheet structure demonstrated using notably ssNMR (De Giorgi et al., 2020). The emergence of such ThT-negative fibril polymorphs has been largely experimentally ignored so far or mistaken for fibrillization inhibitions/failures. They present a specific atomic organization and show an exacerbated propensity towards self-replication in cortical neuron primary culture and living mice. Indeed their injection into the substantia nigra pars compacta of 8-week-old wild-type mice induces the formation of phosphorylated S129-positive α-syn aggregates that spreads towards the dorsal striatum, the nucleus accumbens, and the insular cortex (De Giorgi et al., 2020).

Exploring the richness of rotor dyes, De Giorgi et al. showed that fluorescent compounds with a slightly different chemical structure but sharing with ThT its trimethine cyanine skeleton could detect these ThT-negative polymorphs. Notably, SYBR Green, the well-known DNA intercalating dye, could sense the amyloid structure of these ThT-negative polymorphs but appeared to be much less capable of doing so for ThT-positive ones, giving an almost mirror image of the ThT readings (De Giorgi et al., 2020). Thus, as with prion strains and luminescent conjugated polymers (LCPs) (Sigurdson et al., 2007), multiple external probing using these different molecules, now known under the term “fibrilloscope”, could discriminate the α-syn polymorph identities revealed by ssNMR. These data critically highlight the risk of overemphasizing one readout (here ThT positivity) over others at the expense of the description of the richness of the pathology and, hence, of the narrowness of the “therapeutic” tools designed for a single category of α-syn polymorphs. ThT fluorescence alone should not be taken anymore as a proxy of fibrillization.

Research towards new amyloid rotor dyes continues and produces new entities. Needham et al. synthesized a ThT-derived dye, termed ThX, with improved optical properties and higher binding affinity to amylogenic structures. This new probe presents much better sensitivities to earlier pathological amyloid species (Needham et al., 2020). In the same way, Gaur and colleagues developed an amyloid probe, termed RB1, with the ability to selectively stain α-syn aggregates in the cytoplasm of living cells (Gaur et al., 2021). Both, however, would require validation of previously identified ThT-negative strains to demonstrate a clear added value compared to ThT itself.

4.2. Fusion proteins

Beyond amyloid dyes, fluorescence tagging of α-syn monomers, oligomers or fibrils represents a powerful approach for the dynamic structural study of α-syn aggregation.

The aggregate formation is highly affected by the complex cellular environment, but it has primarily been studied in cell-free systems. Cellular models of protein aggregation and neurodegeneration represent the next complexity step to understanding the processes that govern pathological aggregation and propagation. The simplest and most used way to fluorescently label α-syn for live cell study is to express it fused to a fluorescent protein. Integrating the construct into the genome makes it possible to express the fluorescent protein-α-syn construct directly in the live cells model (Falkenburger et al., 2016; McLean et al., 2001; Outeiro and Lindquist, 2003). In the most straightforward approach, the direct visualization of aggregate formation is possible after the expression of the construct, such as α-syn-GFP or α-syn-mCherry. These free constructs are visualized as an even distribution of fluorescence through the cell or around the membrane, while aggregated inclusions appear as bright cytoplasmic spots (Opazo et al., 2008; Osterberg et al., 2015; Perrino et al., 2019). While using fused fluorescent proteins such as GFP or mCherry represents a straightforward method, it presents several problems. The most obvious is the size of these reporters, which are often larger than α-syn. They can potentially disrupt or affect the α-syn folding/misfolding properties, dynamics or toxicity (Mučibabić et al., 2016). Moreover, the expression of fused fluorescent proteins is generally restricted to the C- or N-terminus, which limits and may interfere with the study of aggregation or protein dynamics (Haney et al., 2016; Jun et al., 2019).

In addition to direct visualization, biophysical techniques such as Förster Resonance Energy Transfer (FRET) allow extracting quantitative information upon aggregation, permitting converting a change in fluorescence into an intramolecular distance. Briefly, FRET occurs between two sufficiently close fluorescent molecules when the energy of one fluorophore, the donor, flows to the other, the acceptor, by non-radiative transfer (Fig. 3A). The subsequent FRET signal strongly depends on the distance and the angle between the two fluorophores (Piston and Kremers, 2007). CFP and YFP are probably the most common FRET pairs used in the study of α-syn dynamics, either inter- or intra-molecularly. In the first case, singly- (such as αS-CFP and αS-YFP) or double-tagged (YFP-αS-CFP) α-syn fusion proteins are used as aggregation biosensors and to reveal specific α-syn conformations and aggregations states (Miraglia et al., 2020; Nakamura et al., 2008; Svanbergsson et al., 2021; Yamasaki et al., 2019). In the second, intramolecular FRET (Single-Molecule FRET) uses double-labelled α-syn (such as YFP-αS-CFP) to obtain information about early minor conformations that may initiate protein pathologic aggregation (Moosa et al., 2019). Alternative approaches use polarized light and measure fluorescence anisotropy due to energy migration FRET to distinguish between different aggregated states of α-syn (van Ham et al., 2010).

Fig. 3.

Schematic depiction of FRET and BiFC methods. A) Single-tagged (αS-CFP and αS-YFP) and double-tagged (YFP-αS-CFP, depicted in the image) α-syn can be used as FRET probes. FRET occurs between two sufficiently close fluorescent molecules by non-radiative transfer of energy from one fluorophore, the donor (CFP), flows to the other, the acceptor (YFP). Dimerization can be quantified by the amount of FRET fluorescence emitted by the acceptor protein. B) BiFC is based on the expression of α-syn fused to a fluorescent protein fragment. Fluorescence is only emitted when the two halves are brought into close proximity by protein dimerization, enabling the formation of the full-length fluorescent protein.

4.3. Bimolecular fluorescence complementation (BiFC)

Protein complementation assays (PCA) are valuable methods to study α-syn aggregation in living systems, especially when approaching the formation of early oligomers. These methods are based on the expression of non-bioluminescent/non-fluorescent protein fragments fused on the N- or C-terminal of α-syn. These fragments only emit fluorescence when the two halves are brought into close proximity, i.e. protein dimerization, to enable the formation of the full-length fluorescent protein (Fig. 3B). An early variant of this method uses bioluminescent protein, such as Gaussia Luciferase, to obtain a read-out indicating the formation of α-syn oligomeric species in different cell fractions (Danzer et al., 2011). Bimolecular fluorescence complementation (BiFC) represents a more exciting approach as it is based on the fusion of a fragment of a fluorescent reporter protein (usually GFP or YFP/Venus) to the protein terminus and thus allows also to direct image and monitor the dynamics of α-syn in in vitro cultures or animal models with sub-cellular accuracy (Goncalves et al., 2010). Since Outeiro and colleagues first implemented the BiFC assay to α-syn, the α-syn-BiFC system has been extensively used to study different aspects of α-syn behaviour in living cells, especially regarding early oligomerization events. In this seminal study (Outeiro et al., 2008), authors co-expressed both fragments of GFP-α-syn in H4 neuroglioma cells. In their first paper, the authors found that fluorescence arises only from both α-syn-fused fragments. Besides, the presence of the BiFC system stabilized the oligomers increasing their in vitro toxicity studied by the measure of adenylate kinase released from damaged cells. Studying the release and uptake of α-syn by cells is an interesting application of BiFC in cellular models. By co-culturing cells expressing each of the fragments individually, combined with blotting and immunocytochemistry, several studies have investigated how α-syn propagates between cells (Bae et al., 2018; Bae et al., 2014; Lee et al., 2021). BiFC readout can also be used for the search for different molecular and genetic regulators of α-syn oligomerization (Gonçalves et al., 2016; Lázaro et al., 2016) and pharmacological modifiers of α-syn release and uptake (Dominguez-Meijide et al., 2020).

However, as Frey et al. noted in a recent review (Frey et al., 2021), despite the widespread use of BiFC in recent years, these studies systematically lacked validation and proper characterization regarding several vital aspects. The authors found a marked differential expression of the two fragments, potentially leading to non-specific interactions. Furthermore, they found that Vn-α-syn shows a higher tendency to form oligomeric and higher-order aggregates. They suggest this may lead to fluorescence arising from α-syn dimerization but also from the association or entrapment of α-syn-Vc in the Vn-α-syn aggregates.

In our opinion, these results do not invalidate BiFC but rather define the limits of use. We agree and support the general idea of the importance of well-characterizing and validating our techniques before jumping to conclusions. As discussed, the α-syn aggregation process highly depends on particular conditions, including adding any bulky reporter. Considering how these reporters may modify α-syn aggregation natural dynamics is of utmost importance.

4.4. Fluorogenic biarsenical compounds

Fluorogenic biarsenical compounds (like FlAsH or ReAsH) represent a relatively small tag alternative compared with fluorescent proteins. It is based on the expression of a recombinant α-syn, bearing a tetra-cysteine tagged to the C-terminal region. This tag acts as a selective target for a fluorogenic biarsenical compound, which is later conjugated, either by making react the tag with the previously purified protein or by treating the transfected cells with the biarsenical compound (Roberti et al., 2007; Roberti et al., 2011). The addition of protein tagged with a second biarsenical compound acting as an acceptor (such as ReAsH) can be used to report intermolecular distances in associated and aggregated forms of the protein through FRET interaction. This method has been applied to study the aggregation of α-syn both in in tubo experiments (Dhavale et al., 2017) and living cells (Roberti et al., 2011). Although arsenal compounds represent less bulky alternatives than fused proteins to monitor the formation of different α-syn species, the labelling process for working with live cells is not simple, and it still constitutes a modification of the protein sequence.

4.5. Site-specific fluorescent probes

The selective introduction of one or multiple site-specific small synthetic fluorescent probes in α-syn represents an exciting and relatively new approach with promising application to study structural dynamics in tubo, and to a lesser extent, at least for the moment, in living cells. Besides potentially overcoming the problems of more bulky probes, they can provide rich region-specific information upon aggregation in real-time through fluorescence emission (Shimogawa and Petersson, 2021). A relatively common approach for α-syn-specific labelling is the introduction of a cysteine residue by site-directed mutagenesis, further labelled via fluorophore-attached maleimides (Cys-Mal). Although this method is limited to single-label sequences, it can be combined with native chemical ligation to obtain multi-label sequences (Haney et al., 2016). Alternatively, unnatural amino acid (UAA) mutagenesis represents a versatile and elegant approach to be used alone or in combination with other labelling methods. It is based on the genetic incorporation of UAAs at desired positions of the protein, which act as handles for a fluorophore, later incorporated via click chemistry (Haney et al., 2016; Saal et al., 2018) (for a review of labelling strategies see Shimogawa and Petersson, 2021).

Site-specific labelling with a fluorophore can be exploited to study detailed structural changes using distance-dependent fluorescent interactions such as FRET. Daniels et al. used a set of α-syn proteins with FRET pairs introduced via UAAs to measure distance changes between specific protein regions in response to cyclized nordihydroguaiaretic acid (NDGA) (Daniels et al., 2019). Depending on the case and needs, several of these methods can be combined to obtain site-specific, multiply labelled protein. Such an approach can be capitalized to create single-molecule FRET (smFRET) probes, which can detect conformational subpopulations and dynamics upon aggregation. In an interesting example, Pan et al. used native ligation to join the αS1–55 fragment containing Cys-Mal labelling with the αS56–140 fragment, labelled by UUA. Then, they used smFRET to evaluate the conformational changes in pY39 α-syn that contributed to the aggregation of α-syn (Pan et al., 2020). Remarkably, Lee et al. succeed in the entirely site-specific labelling of α-syn with three different colour fluorophores, combining native chemical ligation with region-specific cysteine protection of protein fragments. This methodology should expand the utility of smFRET, enabling simultaneous measurement of multiple distances and the study of more complex dynamic interactions (Lee et al., 2018).

Another promising alternative is using environment-sensitive probes relying upon excited-state intramolecular proton-transfer (ESIPT). These ESIPT fluorophores owe their interest to their environment-sensitive characteristics and substantial Stokes shifts. Variation between different α-syn states will cause different charge distributions, hydrogen-bond formations and dipole modifications, leading to different fluorometric responses of the fluorophore (Gao et al., 2021; Sedgwick et al., 2018). Fauerbach et al. used a mutant 18Cys α-syn conjugated with an ESIPT-based probe to monitor aggregation processes occurring during the early lag time inaccessible by ThT (Fauerbach and Jovin, 2018). Kucherak et al. used a newly synthesized ESIPT probe to label a cysteine mutant of α-syn, discriminating unstructured, membrane-bound, and fibrillar states of α-syn (Kucherak et al., 2018).

The necessity of exogenously labelling the α-syn protein sequence is the main limitation of these approaches and the reason why site-specific labelling literature refers mainly to cell-free assays, i.e., in tubo experiments. The very few examples of such α-syn labelling applied to living cells have been limited to tracking the fate of α-syn PFFs formed by mixtures of non-labelled and labelled α-syn. Haney et al. incubated hippocampal neurons with PFFs containing a 25% Cys-Mal-labelled α-syn, demonstrating that labelled-α-syn in position 114 presents a similar ability to seed aggregation and toxicity in living cells as wild-type α-syn (Haney et al., 2016). Jun et al. used a small photo-convertible molecule as fluorophore introduced by UAA to study the internalization and subsequent endo-lysosomal trafficking of α-syn (Jun et al., 2019). Zhang et al. used cysteine mutant-tagged α-syn at position 90 to generate fibrils of different lengths that were then applied to SH-SY5Y cells. The authors monitored cell-surface interactions and uptake, reporting that uptake and toxicity are inversely dependent on the PFF average size (Zhang et al., 2020).

5. Immuno-labelling reporting of α-syn aggregation

5.1. Antibodies against α-syn

The present section does not aim to in-depth review the pros and cons of each antibody but to introduce the reader to the variety of antibodies against α-syn for later discussing the techniques for which immunolabelling is required. Indeed, immunolabelling methods, combined with previous purification steps, are critical in understanding the role of α-syn aggregation in synucleinopathies (Spillantini et al., 1998). Shortly after the discovery of the pivotal role of α-syn, several antibodies were raised against different regions of native α-syn -including the NAC region, the N-terminal region, and the C-terminal region (Culvenor et al., 1999; Dickson et al., 1999). The first α-syn antibodies were used in clinical post-mortem studies. Even before the pathological role of α-syn was unravelled in PD, Ueda et al. discovered a 140 amino acids-long protein in amyloid preparation purified from the cortex of AD patients (Ueda et al., 1993). They developed, without knowing, the first α-syn antibodies against the 61–95 amino acids fragment (now known as the NAC fragment). Shortly after the role of α-syn became apparent in PD and DLB, Baba et al. raised a monoclonal antibody (mAB), LB509, against highly purified LBs from DLB cortices. Using this antibody, they confirmed previous work showing the presence of α-syn in LB and highlighted the presence of high-molecular species of putatively aggregated α-syn in LB from human samples using an immunoblotting experiment. LB509 is now a routine Ab in anatomo-pathology. The parallel development of post-translational modification-specific antibodies was of tremendous importance to better understand the influence of those modifications in pathogenesis and identify them as disease signatures. Among the historically first targeted modifications, we can find nitration (the mAb nSyn 8, 14 and 24) (Giasson et al., 2000) or oxidisation (mAb Syn 303, Syn 505, and Syn 514) (Duda et al., 2002). α-syn phosphorylation at serine S129 (PS129) (Fujiwara et al., 2002) deserves a specific highlight as it has developed almost into a synonym of α-syn pathology for most investigators. Posterior research has witnessed the development of many antibodies targeting different epitopes of α-syn or specific disease-associated α-syn post-translational modifications, including phosphorylation, ubiquitination, nitration and C- and N-terminal truncations (Vaikath et al., 2019).

Immunolabelling is undoubtedly an enormously powerful tool in synucleinopathy research, but the results' interpretation should be cautiously conducted. Some of these antibodies, especially those targeting phosphorylated α-syn at serine 129 (pS129-α-syn) or ubiquitinated α-syn, have been developed into recognized and widely used markers of α-syn pathology. However, assuming that phosphorylation equals pathological, i.e., aggregated, α-syn, despite not all pathological α-syn being phosphorylated and not all phosphorylated α-syn being pathological (Bezard, 2022), may have dramatic scientific and (ultimately) therapeutic consequences. In the same way, plenty of known (plus the yet unknown) modifications found in anatomopathological landmarks may interfere with the immunoreactivity of these α-syn antibodies, making interpreting the immunolabelling even more difficult. Likely, the currently used antibodies do not capture the full spectrum of aggregated and not aggregated α-syn pathological populations. Besides, there is a fundamental problem of aiming to the primary structure of α-syn when studying aggregated protein, especially considering the now better-understood heterogeneity underlying α-syn pathology. Indeed, different stages and modalities of disease likely possess different biochemical and structural antibody epitopes. Whether the currently-used antibodies can distinguish the entire proteome spectra underlying different pathology remains fundamentally relevant (Altay et al., 2022).

Accumulated evidence points towards a possible disease-relevant role for α-syn oligomers (Alam et al., 2019; Winner et al., 2011), calling for new tools going beyond primary structure and targeting particular conformations. Unfortunately, intermediate oligomers are unstable, variable and highly heterogeneous (Kulenkampff et al., 2021; Luth et al., 2015). Insight into the biochemical and structural properties of native oligomeric α-syn is therefore challenging to obtain. These factors make generating high-quality and high-specificity conformational antibodies challenging, which is nonetheless critical for the robust interpretation of results (Desai et al., 2021).

Despite this, the last years have seen the development of different antibodies raised against particular protein structural motifs, termed conformation-specific antibodies (Brännström et al., 2014; Covell et al., 2017; Vaikath et al., 2015; van Diggelen et al., 2019). These tools represent an exciting development for studying disease-relevant native oligomers in fixed brain tissue of animals (Duffy et al., 2018) and lysates from human LB (Vaikath et al., 2019). Two recent studies have used conformational antibodies and super-resolution microscopy to study the uptake by neuronal cells of pathologically relevant α-syn oligomers and fibrils (Bigi et al., 2021) and the uptake and secretion of seeds by SH-SY5Y cells (Sang et al., 2021). In an exciting approach, Gibbs et al. used computational modelling to predict conformational epitopes likely to be exposed only on oligomers but not on monomers or insoluble aggregates (Gibbs et al., 2022). They generated monoclonal Ab (mAb) against these epitopes. These mAb present low reactivities with monomers and insoluble fibrils, including post-mortem LBs. Besides, adding the mAb to cultures of primary rat dopaminergic neurons exposed to α-syn oligomers resulted in a higher survival rate, measured by posterior quantitation of TH-positive neurons. Also, they reduced the uptake of PFFs in a cellular seeding assay in rat hippocampal neurons, in which neurons with aggregates of human α-syn were quantified after 14 days of culturing (Gibbs et al., 2022).

The lack of head-to-head comparison through a series of validating steps makes the comparison between various antibodies elusive. Kumar et al. accepted the challenge and performed a systematic antibody characterization through a validation pipeline which allowed a systematic investigation of the specificity of α-syn antibodies using well-defined and well-characterized preparations of various α-syn species, including monomers, fibrils, and different oligomer preparations that were characterized by distinct morphological, chemical and secondary structure properties (Kumar et al., 2020). Testing 18 antibodies, of which 16 were previously reported as conformation- or oligomer-specific antibodies, Kumar et al. convincingly showed that i) none of the antibodies tested were specific for one particular type of α-syn species, including monomers, oligomers or fibrils; ii) all antibodies which were reported to be oligomer-specific also recognized fibrillar α-syn; and iii) a few antibodies showed high specificity for oligomers and fibrils but did not bind to monomers (Kumar et al., 2020). The authors concluded that the great majority of supposedly α-syn aggregate-specific antibodies did not differentiate between oligomers and fibrils, thus highlighting the importance of exercising caution when interpreting results obtained using these antibodies and the need for complete characterization and validation of antibodies before their use in mechanistic studies (Kumar et al., 2020).

5.2. Biochemical methods

The combination of biochemical methods and immunoassays provides precious information about the identity of α-syn species, notably regarding the purification and characterization of those extracted from fresh-tissue samples (Fig. 4). Techniques such as western blotting and dot blotting will not be explicitly explained here, as they are performed routinely in most laboratories. They represent techniques of enormous importance and are usually the final step of the myriad of techniques presented in this section. It is nonetheless worth mentioning filter retardation assay as a potentially powerful technique, as it could allow the capture of aggregated over specific sizes due to specific membrane properties such as acetate membrane (i.e., independently of the supposed conformer specificity of an Ab). The selection of the membrane material and properties is critical in dot blot experiments. Subsequent Dot Blotting with the desired Abs allows for the relative quantification of these species (Bourdenx et al., 2020; Recasens et al., 2014).

Fig. 4.

Biochemical analysis of biological tissues. The combination of purification methods, and characterization techniques, notable immunoassays, provide precious information about the identity of the different α-syn species present in samples extracted from fresh-tissue samples. Characterization of the different fractions may involve, for example, EM (Laferrière et al., 2019), filter retardation assay (Bourdenx et al., 2020), or digestion methods + western blotting (Peng et al., 2018).

One of the most popular methods to study α-syn fibrils derived from ex vivo samples is to evaluate the resistance of aggregates against different protein-digesting chemicals. While relative protein solubility is usually used to purify samples, it also holds potential as an analytical tool. Amyloid fibrils are characterized by their cross-beta fold structure, an incredibly stable fold that confers resistance against solvent agents. α-syn aggregates are diverse and complex, presenting different dynamic regions, with usually a single part of the protomer having the β-fold structure. At the same time, smaller oligomers are usually entirely soluble. Consequently, different polymorphs are expected to fragment differently towards specific solving agents. The evaluation of its resistance and the so-called proteolytic profile towards these agents permits us to infer useful analytical information about the type of aggregates we may be dealing with.

The most well-known example of these agents is Proteinase K (PK). The digestion resistance to this endo-protease is usually cited as one of the main properties of amyloid fibrils. Contrary to other endo-proteases, PK presents a low amino acid specificity, which leads to almost-complete hydrolysis of “non-PK-protected” regions (Kushnirov et al., 2020). Besides PK, further fractionation characterization can be done using different mixtures of protein dissolvent such as urea, SDS, Triton, or trypsin. The subsequent analysis of digestion patterns with Blotting assays, such as western blot or dot blot, permits to evaluate of the differences between different types of samples in digestion-resistance terms (Bourdenx et al., 2020; De Giorgi et al., 2020; Laferrière et al., 2019; Peng et al., 2018). In an elegant approach, Landureu et al. used PK and thermolysin to achieve limited proteolysis. Then, immunoblotting with several Abs directed to different known epitopes was performed to investigate the exposition of different regions. Authors applied it to three previously known α-syn fibril polymorphs displaying different binding properties, confirming a different exposure of α-syn N-terminal domain (Landureau et al., 2021).

Also, of particular relevance in this section, some physicochemical methods permit the separation of different protein fractions found in ex vivo homogenates. For example, sucrose gradient purification allows post-mortem homogenates content based on mass separation. For that, samples are placed on the top of a sucrose step gradient of different concentrations, thus different densities. Ultra-centrifugation of this sucrose gradient at controlled times and speed permits the obtention of various fractions, which can then be studied separately (Emin et al., 2022). In an influential work, Recasens et al. used a sucrose gradient to separate a PD patient mesencephalon homogenate into 20 fractions. Posterior analysis of the fractions with confocal and electronic microscopy, combined with filter retardation assay, permitted the identification of LB-enriched fractions containing pathological α-syn. Most interestingly, they demonstrated the capacity of these purified fractions to induce nigrostriatal neurodegeneration in monkeys and rodents (Recasens et al., 2014). Building on this technique, Bourdenx et al. showed that amyloids with a specific cross-beta structure actually “hide” into the fraction containing supposedly mostly soluble α-syn. They demonstrated that these peculiar conformers could also induce degeneration and pathological changes throughout the brains of primates but not of rodents, thereby proving the need for involving different species to assess the toxicity of α-syn-β-amyloids (Bourdenx et al., 2020). Using only rodents would artificially focus on large aggregates while involving primates widening the spectrum of toxic species (Bourdenx et al., 2020).

Similarly, density floatation gradient and sedimentation assays are practical separation methods after solubilization in solvents such as sarkosyl. In the first case, separation is achieved by exploiting the fact that bigger aggregates are less dense than smaller ones and monomers. This density-based separation is achieved by loading the samples in gradients of iodixanol. Upon centrifugation, the different species move until they reach an equilibrium where their density is the same as the surrounding gradient medium. Again, subsequent blotting with particular Abs allows for relative quantification of the different fragments, such as total α-syn, pS129 α-syn or specific conformation (Bourdenx et al., 2020). Complementary sedimentation assays take advantage of that different species, notably depending on the shape, size and density, tend to form sediments faster (Laferrière et al., 2019). After the first solubilization step, the different entities can be fractionated depending on their sedimentation velocity upon ultracentrifugation, allowing them to isolate particular species or generate sedimentation profiles. Laferriere et al. showed the differential content in aggregates between PD and MSA brains using such sedimentation assay. Applying this to a transgenic mouse model of MSA, they showed that the heavy S129-phospho-α-syn load in these mice was made of soluble α-syn while PD and MSA brains exhibited aggregated and phosphorylated α-syn (Laferrière et al., 2020).

5.3. Enzyme-linked immunosorbent assay (ELISA)

ELISA represents a widely used qualitative method for studying α-syn aggregation as it possesses broad applicability ranging from in tubo to ex vivo tissue or patients' samples. There exist several modalities of ELISA, but, simply put, it is based on the immobilization of the antigen to the surface, either directly or using the first Ab, and the addition of an enzyme-conjugated detection Ab that provides the signal for the quantification of the antigen. ELISA is a relatively simple procedure. It is usually applied to test large pools of samples or purified fractions. The technique's success depends directly on the specificity of the Ab for the desired species. Earlier modalities in the aggregated α-syn study used two monoclonal Ab with identical epitope specificity, so the signal only arises if there is more than one binding site in the structure (El-Agnaf et al., 2006; Kulenkampff et al., 2021). Later studies have used ELISA to quantify the soluble aggregates, i.e., oligomers, using conformational Abs. These studies used Abs raised against epitopes missing on natively unfolded α-syn. Paleologou et al. used a polyclonal rabbit Ab, termed

FILA-1, detecting elevated levels of these soluble aggregates in DLB brain lysates (Paleologou et al., 2009) (but see review by (Kumar et al., 2020)). In this line, Lassen et al. developed a sandwich-type α-syn monoclonal rabbit Ab, MJF14–6–4-2, and demonstrated its ability to detect oligomers in detergent extracts of the whole brain from Thy-1 promoter overexpressing α-syn mice (Lassen et al., 2018). Although promising as a fast and easy diagnosis tool for human-derived samples, these ELISA methods are limited because conformational antibodies are imperfect tools (Kumar et al., 2020). While ELISA usually represents a “routine method”, we believe it becomes precious if it is preceded by fractionation and purification steps earlier mentioned. A careful selection of Abs might then lead to, for instance, quantifying the total α-syn present in patients' brain-derived extracts carefully fractionated using either of the above-described techniques.

We should also highlight the single molecule array (SIMOA), an emergent ELISA variant, which uses femtoliter-sized chambers with Abs immobilized on a paramagnetic bed. In the end, each bed captures either one immunocomplex or none. Fluorescence is then electronically measured, giving this technique a much higher sensibility than ELISA (Kulenkampff et al., 2021). SIMOA has recently been applied to α-syn research by Kumar et al. to assess the immunoreactivity of conformation-specific antibodies (Kumar et al., 2020) and Ren et al. to search for potential biomarkers in PD patients' plasma (Ren et al., 2022).

5.4. Single protein immuno-fluorescent imaging methods

These methods represent unique approaches to image and potentially characterize individual protein aggregates in different systems, notably from human-derived samples. The ability to image individual aggregates provides potentially higher sensibilities than “bulk methods” such as ELISA (Kulenkampff et al., 2021).

sFIDA technology (surface-based fluorescence intensity distribution analysis) represents a promising approach towards the sensitive detection of aggregates, which is especially interesting for the design of biosensors for diagnostics. It is based on the immobilization of Ab directed against linear epitopes of α-syn, such as Syn211, on the glass surface of a microtiter plate and the subsequent treatment with the same Ab functionalized with a fluorophore. Insensitivity towards monomers is achieved as the first immobilized Ab masks the protein epitope. The second Ab, functionalized with a fluorophore, thus permits the detection and counting of single particles of (in theory) all aggregated α-syn subtypes. As a drawback, the technique cannot yet discriminate the size or subtype of aggregate, as aggregates are typically smaller than the resolution limit. Blömeke et al. recently demonstrated the applied sFIDA technology to CSF samples of PD and DLB patients, finding significantly higher levels of α-syn aggregates compared with controls (Blömeke et al., 2022). In a similar approach as sFIDA, the single-molecule pull-down (SiMPull) assay uses a classic immuno-precipitation to immobilize aggregates and a pair of primary-secondary Abs to detect at single protein resolution. The method was applied by Je et al. to the study of α-syn oligomers. Although this method does not directly discriminate between monomers and aggregates, the different intensity distributions can be used to identify the species (Je et al., 2017).

5.5. Alpha-synuclein proximity ligation assay (AS-PLA)

α-syn proximity ligation assay (AS-PLA) represents a suitable method to detect and study the smaller α-syn aggregates, usually oligomers in literature, that other techniques may fail to detect, thanks to its potential higher sensibility. It has particular relevance when it comes to exploring post-mortem tissue surfaces. In some cases, the liability and heterogenicity of early aggregates limit the use of traditional immunohistochemistry with conformational antibodies (Behere et al., 2021). Others might claim that antigen preservation is not the issue but rather the limited knowledge of the conformers in the tissue.

Simply put, PLA methods detect protein-protein multimerization through an amplification reaction mediated by an oligomer chain. This oligomer chain will only bind the two protein–binding antibodies if they are sufficiently close to each other (<40 nm), i.e. possibly aggregated (Sekiya et al., 2022; Söderberg et al., 2006). Although this technique has been profusely used in the study of so-called early oligomers, it is crucial not to mistake its ability to detect smaller species with selectivity. PLA will generate, in principle, signal against all kinds of aggregated or non-aggregated species provided (i) two monomeric proteins are close enough and (ii) the epitope for the Ab is exposed. While some individual fibres may present lower signals with PLA, as epitopes are not close enough (Roberts et al., 2015), this does not mean PLA can universally differentiate between oligomers and other higher aggregated forms of a-syn.

PLA was first applied to an α-syn aggregation study by Roberts et al. by using an α-syn Ab (syn211) displaying blocking activity on the epitope (Roberts et al., 2015). This blocking prevents the liaison of more than one Ab per monomer, making the essay only sensible towards multimeric α-syn species. The authors first demonstrated the technique's sensitivity to detect in tubo recombinant α-syn-formed oligomers preferentially. Then, they applied the method in fixed post-mortem PD brain tissue, observing previously undetected diffuse deposits of α-syn, often in mildly affected regions (Roberts et al., 2015). Over the past few years, AS-PLA became a popular tool to study what authors usually termed early α-syn oligomers. It has been applied to study oligomer distribution in the midbrain of young mice (Behere et al., 2021) or the brain samples of MSA and PD patients (Sekiya et al., 2019). Bengoa-Vergniory and colleagues used AS-PLA in several models and situations to evaluate the presence of oligomers upon treatment with the molecular tweezer CLR01: PD patients' brain extracts, human dopaminergic cultures over-expressing α-syn, and human α-syn-overexpressing mouse model tissue (Bengoa-Vergniory et al., 2020). Sekiya et al. used PLA and classic histochemistry against pS129-α-syn to compare the distribution of early aggregates, i.e. oligomers, with Lewy-related pathology (LRP) distribution. The authors used diverse formalin-fixed paraffin-embedded post-mortem brain samples from neuropathologically-confirmed cases of PD and control patients. First, the authors found the distribution of both oligomers and LRP variables between patients, confirming the heterogenicity of what we collectively call PD. Second, the authors also found a broad distribution of oligomers, which do not necessarily correlate with the presence of LRP. Nonetheless, oligomer presence, for example, in the hippocampus correlated with observed symptoms of dementia, providing another hypothesis for why not all patients fit the Braak staging scheme (Sekiya et al., 2022).

Outside the central nervous system, the potential higher sensibility of AS-PLA has also been used to search for early α-syn oligomeric biomarkers such as in gastro-intestinal biopsy tissue (Ruffmann et al., 2018) or skin biopsies of PD and MSA patients (Mazzetti et al., 2020; Vacchi et al., 2021). Besides using PLA as a marker for α-syn aggregation, it can also be applied to study the colocalization and interaction of α-syn with other proteins in mechanisms possibly relevant to α-syn aggregation, such as the SNARE proteins (Almandoz-Gil et al., 2018), α-tubulin in the murine and human brain (Amadeo et al., 2021) and α-syn with retrograde and anterograde transport proteins in dopaminergic cultures (Bengoa-Vergniory et al., 2020). Interestingly, PLA can also detect modified proteins, such as p-α-syn, by using one Ab targeted to the protein and one targeted to the modification (Arlinghaus et al., 2021).

5.6. Quantification of immunohistochemical reaction

Although immunocytochemical and immunohistochemical studies are the cornerstone of the many post-mortem studies performed on tissue originating from animal models or deceased patients, there is some vagueness in the definition of the measured endpoints. The above discussion gives a sense of the hot debates about the specificity and reliability of the various antibodies. We would like to add to this debate the issue of quantification. To illustrate the lack of generalization in the assessment procedures, allow us to take the example of α-syn phosphorylation at serine 129 (pS129-α-syn) (Fujiwara et al., 2002). This original elegant demonstration suggested excessive phosphorylation of α-syn in synucleinopathy brains (including those in transgenic animals), as well as the underlying alterations in the activities of various kinases and phosphatases and the conformation of α-syn, may contribute to abnormalities of α-syn. Since this publication, in the overwhelming majority of post-mortem and experimental studies, the assessment of α-syn phosphorylation was taken as an indirect but supposedly reliable marker of aggregation without stating the endpoint's proxy nature. However, many papers have been satisfied with even very diffuse pSer129 positivity to qualify experimental histological images as synucleinopathy without considering the presence of actual subcellular inclusions in the measured/quantified signal. In other words, the immunohistochemical signal is often measured through optical density (without defining the experimental process for avoiding background variations) and measured as such or as an assessment of surface staining (Teil et al., 2022). A punctated signal could also be measured either quantitatively (very rarely) (Peng et al., 2018) or semi-quantitatively using a pathology-like rating scale- the actual meaning being more or less defined) (Henrich et al., 2020; Luk et al., 2012). Another option is to bypass the issue of α-syn quantification and use p-syn staining to identify cells that would then be characterized for the expression of other markers (Chu et al., 2019; Henrich et al., 2020). Surprisingly, very few papers report the counting of carefully defined aggregates (here, aggregate means an intracytoplasmic object more extensive than a simple punctum) (De Giorgi et al., 2020; Schmidhuber et al., 2022). We recently elaborated on the need to disclose the analysis of experimental synucleinopathies better, highlighting that the semantic and conceptual efforts should be made by scientists working with human tissues and experimental/animal material (Bezard, 2022).

6. Amplification techniques

At this point, we would like to briefly introduce the amplification techniques. Protein misfolding cyclic amplification (PMCA) and real-time quaking-induced conversion (RT-QuIC) are not stricto sensu approaches for monitoring aggregation but amplifying certain polymorphs of interest. Initially developed to detect different prion strains with high sensitivity, they hold great potential in synucleinopathy diagnosis and research (Fairfoul et al., 2016)(Poggiolini et al., 2022). They are based on the ability of trace amounts of misfolded protein to “trap” monomeric protein into its potential energy pit and grow. A combination of aggregate-growing/aggregate-breaking cycles permits to increase in the template units that can induce further aggregation while, in principle, preserving its ultra-structure. Then, classical laboratory techniques such as western blot or real-time fluorescence should allow signal quantification, while cryo-EM can provide ultra-structural information. (Brandel et al., 2019; Burger et al., 2021).

6.1. Protein misfolding cyclic amplification (PMCA)

PMCA has proved helpful in replicating fibrils from patients' tissue extracts keeping their conformation (Jung et al., 2017). Several studies have recently investigated its potential as a diagnostic tool. Among those, we chose to highlight Shahnawaz et al., who investigated its ability to detect oligomers in the CSF of PD and MSA patients (Shahnawaz et al., 2020), and Fenyi et al., who obtained positive results for detecting and amplifying aggregates in gastrointestinal biopsies using PMCA (Fenyi et al., 2019). Shahnawaz et al. showed very convincingly that the α-syn-PMCA assay can discriminate between CSF samples of PD and MSA patients, with an overall sensitivity of 95.4%. The obtained aggregates were further distinguished using a plethora of techniques, supporting PMCA data. Further, Shahnawaz et al. found that the properties of CSF-originating aggregates were like those of brain-amplified aggregates (Shahnawaz et al., 2020). This result validates the hypothesis of a continuum of pathology between the brain and the body fluids that can be accessed in living patients. It also strengthens the hypothesis of differential conformational strains of PD and MSA α-syn aggregates, paving the way for developing a biochemical assay capable of discriminating between PD and MSA at an early disease stage.

6.2. Real-time quaking-induced conversion (RT-QuIC)

RT-QuIC was born as a variant of PMCA by introducing some modifications in the protocol. Importantly, the real-time aggregation is now followed by ThT fluorescence (hence the name “real-time”) (Atarashi et al., 2011) (of note the need for using complementary dyes because of the ThT fluorescence “bias” and the existence of ThT-negative aggregates possibly as toxic as the positive ones). Several studies have confirmed the α-syn RT-QuIC assay to have high sensitivity and specificity for PD. In 2017, Fairfoul et al. developed an RT-QuIC-based assay to detect aggregated α-syn in DLB and PD cerebrospinal fluid with sensitivities over 90% (Fairfoul et al., 2016). Manne et al., focused on the submandibular glands as more accessible peripheral tissue, finding also positive results (Manne et al., 2020). More recently, Vascellari et al. reliably found misfolded α-syn in the CSF-isolated samples from rapid-eye-movement sleep behaviour disorder patients (Vascellari et al., 2022). Poggiolini et al. contributed to the most comprehensive study to date. Several important features were shown, thanks to the convincing power of their study. Despite a superior sensitivity towards PD (89%), no quantitative parameters correlated with PD severity. Conversely, although sensitivity was lesser for MSA (75%), the quantitative parameters correlated with MSA disease progression. Conversion of RBD patients into PD was somewhat predictable from early RT-QuIC. However, the percentage of converters (31%) was probably too low for concluding the predictivity (9 out of 14 converters showed baseline RT-QuIC positivity) (Poggiolini et al., 2022). Limitations were acknowledged by the authors and the next steps were clearly defined notably regarding the putative usefulness of α-syn RT-QuIC as an early biomarker detecting synucleinopathy in idiopathic RBD patients prior to conversion. This is a result awaited by the entire research community to better understand the relationship between a-syn RT-QuIC signature and the progression from prodromal to different synucleinopathies.

7. Facing the challenge of in vivo reporting of aggregation

Despite the complexity of α-syn aggregation dynamics, the current knowledge comes primarily from cell-free studies and, to a lesser extent, from cellular models. Given the importance of the molecular environment in which α-syn aggregation occurs (Lövestam et al., 2021), animal models designed to monitor aggregation or aggregated α-syn represent a crucial step in our understanding. On the one hand, they allow the validation of in vitro results and the detection of the differences between the processes occurring in the models mentioned above and the in vivo ones. Moreover, they provide a platform for reliably testing protein aggregation upon pharmacological challenges or genetic modifications.

The expression of recombinant α-syn as a fluorescent fusion protein in live organisms represents a straightforward method of studying aggregation in vivo. In a first approach, Kaminski Schierle et al. used an α-syn–YFP transgenic C.elegans worms. The authors used FRET between YFP (donor) and the forming aggregate (acceptor) to study aggregation. To do so, they measured the excited state lifetime of YFP using time-correlated single-photon counting (TCSPC) fluorescence lifetime microscopy. Interestingly, they found that α-syn–YFP forms different species with different associated toxicities in vivo than in vitro, highlighting the importance of the model system over α-syn aggregates structure (Kaminski Schierle et al., 2011).

In an exciting work, Osterberg et al. used previously developed α-syn-GFP transgenic mice for in vivo multiphoton imaging (Unni et al., 2010) to study the seeded aggregation of α-syn in cortical neurons of living mice (Osterberg et al., 2015). The authors showed that the injection of recombinant PFFs in the primary sensory cortex seeded aggregation of transgenically expressed- and endogenous α-syn in cortical neurons of living mice. The in vivo system allowed the imaging of the same neuron over months, observing stage-like maturation of the inclusions and, finally, selective cell death of inclusion-bearing neurons. Besides, the in vivo use of FRAP allowed them to obtain α-syn mobility measured in different conditions and maturation states, detecting slower diffusion kinetics within inclusions. In a recent paper, Schaser et al. developed a transgenic mouse model expressing the human A53T α-syn-GFP fused protein (Schaser et al., 2020). Using a similar experimental approach to Osterberg et al. (Osterberg et al., 2015), Schaser et al. observed the initial formation of aggregates in the axons and a time-dependent retrograde spread of aggregated α-syn in neurons, spreading later to astrocytes (Schaser et al., 2020). Besides mice models, the same group developed a zebrafish model expressing human GFP-α-syn in neurons. FRAP mobility experiments showed that human α-syn in zebrafish behaves similarly to that in mammals. The authors also performed pharmacological and genetic manipulations in the model, concluding that presynaptic aggregation of α-syn cannot be modulated by phosphorylation at serine-129 on its own (Weston et al., 2021).

The development of animal models based on previously described α-syn BiFC has become another valuable tool to monitor α-syn aggregation in living models as it permits detecting the fluorescence directly in vivo or fixed tissue. C. elegans models have been developed to study α-syn transmission in the context of ageing. By expressing each of the α-syn BiFC fragments in different but connected anatomical parts (Kim et al., 2016) or specific neuronal sub-populations (Tyson et al., 2017), the authors monitored the BiFC fluorescence in worms at different ages. They confirmed that α-syn could propagate and aggregate in this system. Prasad et al. recently developed a transgenic α-syn-BiFC Drosophila melanogaster as a more straightforward in vivo test bench amenable to genetic and pharmacologic manipulations. The analysis of different soluble fractions of brain homogenates permitted the assessment of the influence of a set of suspected risk factors or mutations (Prasad et al., 2019).

In 2013, Dimant et al. first developed a BiFC rat model. By co-injecting AAVs encoding human α-syn fused to the N- or C-terminal half of VenusYFP, the authors were able to detect fluorescence and to image α-syn aggregates in fixed sections of the nigrostriatal pathway and in vivo in the cortex, using confocal and two-photon microscopy, respectively (Dimant et al., 2013). More recently, diverse α-syn-BiFC mice models have appeared based on transgenic or AAV approaches. Similarly to Dimant et al., Cai et al. developed a BiFC-based mouse model based on the AAVs co-injection and characterized it at several levels. At four weeks post nigral injection of both AVVs, authors observed venus-YFP colocalizing with TH immunofluorescence in the SN. SN cells were also positive for human α-syn immunohistochemistry. This AAV-mediated human α-syn overexpression was related to ipsilateral nigral neurodegeneration, measured by TH-positive cell counting. YFP was observed in the striatum, supporting an anterograde spread of aggregated ⍺-syn. To explore the nature of aggregates, brain sections were treated with PK showing a significant reduction in fluorescence. The authors concluded that most of this fluorescence was thus caused by non-fibrillar oligomers. Finally, the analysis of Triton-X100- and SDS-treated midbrain homogenates through immunoblotting using an anti-human α-syn Ab shows myriad aggregated α-syn species in the ventral midbrain and the striatum eight weeks post-injection. (Cai et al., 2018). Kiechle et al., on the other hand, have generated two transgenic mouse models, expressing either split Gaussia luciferase or split Venus Yellow Fluorescent Protein (YFP) for aggregate detection within the brain and with subcellular resolution. The authors were able to quantify α-synuclein oligomerization in vivo in the pre-synapse. They confirmed the age-dependency on aggregation and the cell-to-cell transfer of de novo generated α-syn aggregates (Kiechle et al., 2019).

8. Concluding remarks

Researchers have access to an extensive panel of tools for monitoring α-syn aggregation. Although biophysical, non-Ab-based techniques are sound and give rise to unquestionable results in tubo or in vitro experiments, the challenge we collectively face is with the in vivo monitoring of such aggregation. A subsidiary question deals with our ability to produce and use extremely well-characterized antibodies for which we should not claim too rapidly about specificity towards particular conformations. The past few years have witnessed a structural biology revolution demonstrating the possible relationship between exquisite structural details, motifs, and differential pathologies. These results highlight how understanding the particular fibrillization conditions may lead to specific α-syn aggregated species (so-called polymorphs) and thus different pathological manifestations. Such basic science knowledge should now be transferred to in vivo experiments comparing the pathological outcome of exposure to in-depth characterized purely recombinant polymorphs, recombinant-made PMCA- or RT-QuiC-amplified patient-derived seeds, or crude patient-derived extracts. Combining single particle tracking with refined Abs and spatial omics approaches is likely to be the key to unlocking the disease pathology complexity (within a disease) and the diseases' specific polymorphs (if they exist). The in vivo understanding of these conditions leading to specific conformers, and thus specific diseases, is of paramount importance both for advancing the knowledge and for, at last, developing therapeutic approaches for the unrelenting synucleinopathies.

Funding

This review has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (Grant agreement No. #951294 to EB). JEP is a Ph.D. fellow in that ERC program. This study received financial support from the French government in the framework of the University of Bordeaux's IdEx “Investments for the Future” program/GPR BRAIN_2030. The opinions expressed and arguments employed herein do not necessarily reflect the official views of these funding bodies.

Declaration of Competing Interest

EB is a director and shareholder of Motac Neuroscience Ltd. The other authors declare no conflict of interest.

Data availability

No data was used for the research described in the article.