Effect of the micro-environment on α-synuclein conversion and implication in seeded conversion assays

Being a natively unfolded protein [18], α-synuclein aggregation is ruled by thermodynamics [19]. The first event in the assembly of α-synuclein is supposed to be dimerization [20], leading in turn to the generation of aggregation-competent oligomers, encompassing a wide variety of structures [21], and possessing their own kinetics of fibrillization and toxicity [22].

The conversion behaviour of α-synuclein, described as “prion-like”, has been reported by evidences of graft-to-host spreading of α-synuclein aggregates [23‐25]. Aggregated α-synuclein can be transmitted from neuron to neuron [26] either after neuronal death or by exocytosis-mediated pathways [26], while the uptake may occur be endocytic pathways [27], substantiating the neuron-to-neuron transmission and the predictable spreading observed in some forms of α-synucleinopathies. Recently, α-synuclein was shown to interact with high density lipoproteins [28] and the presence of α-synuclein, both in its native and misfolded conformation, was found in extracellular vesicles as well [29, 30], suggesting a potential novel route for the spreading of the pathology. Moreover, α-synuclein transmission along axonal connections was demonstrated ex vivo and in vivo in various animal models.

Luk and co-workers demonstrated that fibrillary α-synuclein is able to trigger the conversion of monomeric α-synuclein into aggregates closely reminiscent of LBs and neurites in a cell model overexpressing α-synuclein [31]. Fibrillary α-synuclein was further found to be transported anterogradely through the axon and taken up by the next neuron in mouse embryonic cultures [32], while in vivo studies conducted with rats showed the trans-neuronal spreading of fibrils from the olfactory bulb toward other non-olfactory regions [33]. The transport of fibrillary α-synuclein forms were shown to follow an anterograde pathway through the axon and passaged to the next neuron in mouse embryonic cultures. Moreover, in vivo studies conducted in rats demonstrated that α-synuclein fibrils spread from the olfactory bulb toward other non-olfactory regions [33]. Others showed that α-synuclein aggregates were inoculated in the medulla oblongata of rats, and the caudo-rostral spreading was followed by showing fibers and neuronal abnormalities in the rostral mesencephalon and in prosencephalic regions after 18 weeks [34]. Lastly, PD pathology was successfully passaged to non-human primates by administration of human PD-derived LBs.

The spreading of α-synuclein through different routes may highlight the differential affinity of different cell types toward α-synuclein binding and internalization, as the cellular milieu differs among populations of cells derived from different developmental lineages. Thus, the cellular environment may play an instructive role for the spreading of misfolded α-synuclein, affecting first highly vulnerable cell lineages.

In the next section, we will address the environmental factors that may affect the aggregation of α-synuclein, its seeding ability and thus the modes of spreading through various routes.

A plethora of environmental factors have been found to affect protein misfolding and aggregation [35]. Physical factors, such as molecular crowding and temperature, may affect protein aggregation by enhancing the entropy of the system. Similarly, chemical factors such as pH variations or the presence of other proteins or biomolecules, such as nucleic acids, phospholipids and proteoglycans, may affect the aggregation process. Moreover, post-translational modifications (PTMs) may also interfere with the aggregation process by affecting the surface charge or the topology of the protein. The effect of these conditions on α-synuclein aggregation is summarized in Table 1.

Additional, often overlooked factors in PD patients, which influence the misfolding and aggregation features of α-synuclein, are missense point mutations in the α-synuclein gene (SNCA) (e.g. A53T, A53E) as well as the potential increase of α-synuclein expression levels, induced by duplications or triplications of SNCA. The genetic A53T missense mutation causes conformational changes in α-synuclein with structural alterations of the β-sheet structures, thermodynamically more stable in aqueous solutions, exhibiting a significant influence on the aggregation process [36]. Experimentally, the integration of a genetic PD associated point mutation in recombinant α-synuclein substrate would accelerate the seeding response of oligomeric α-synuclein in seeded aggregation assays. However, it may also promote the self-aggregation of the substrate and the generation of false-positives, as we discuss in a dedicated section below.

The importance of the self-assembly properties in the onset of neurodegenerative diseases have led to the investigation of the kinetic of the aggregation. During the onset of the pathology, this process is typically slow. However, it can be mimicked in vitro and, under appropriate conditions, it may occur much faster than during pathology. Independently from physical and chemical factors, which may influence the fibrillization process, the presence of a molecule which is prone to aggregation and can act as a nucleus or seed for the fibrillization, allows the reduction of the lag phase and avoids the possible denaturation of the monomers during the reaction [90, 91]. The growth of the α-synuclein fibrils follows a sigmoidal curve (Fig. 1). It displays a lag phase, during which the nucleation process subsequently, the exponential phase indicates the elongation of the fibrils, before a plateau phase can be observed once the monomers have been converted into β-sheet enriched species, which saturates the signal (Fig. 1). The seeded conversion assays have been applied for the detection of misfolded proteins for the diagnosis of prion diseases. In this frame, a fundamental innovation was the development of the protein misfolding cyclic amplification (PMCA) [93]. The aim of the study was to reproduce the conversion of PrP in vitro by seeding the reaction with minute amounts of the misfolded form of PrP. To this end, the aggregation of PrP^C from healthy brain homogenates was triggered by the addition of brain homogenates from scrapie-infected hamsters in a cyclic reaction, consisting in cycles of incubation and sonication. The former step allows the seed to contact and convert its native counterpart and to form growing oligomers, whilst the latter step breaks longer fibrils to increase the number of smaller seeding-competent aggregates, hastening the conversion process. The end-product was shown to be proteinase K (PK)-resistant by Western blotting. The PMCA was further automated [94‐96] to make it suitable for the analyses of biological fluids (blood, urine or the cerebrospinal fluid (CSF) [97, 98]) and, recently, its application was expanded to the detection in biofluids of other neurodegenerative-related molecules, such as the CSF from patients with α-synucleinopathies [99].

To overcome the inherent complications of the PMCA, such as poor reproducibility, the time-consuming need of PK-resistant PrP^res detection via Western blotting, the careful handling of the infectious end-product and the lack of consistency with different prion strains [100], other in vitro, cell-free, amplification assays for misfolded proteins have been developed. The substitution of sonication by shaking [101, 102] represented a major advancement. Indeed, by increasing the air-water interface, shaking is thought to cause the unfolding of PrP and to enhance the fragmentation into oligomers, thus favouring the seed-substrate interaction [103]. The assay, termed real-time quaking-induced conversion (RT-QuIC), was further refined by the real-time screening of the course of the reaction by adding the fluorescent dye Th-T (Fig. 1). In this assay, small amounts of misfolded PrP^Sc seeds recruit single recombinant PrP^C substrate molecules and induce their conversion by integrating them into a growing amyloid aggregate concomitant with a conformational change of the substrate into a seeding-competent state (Fig. 1) [92, 101]. The RT-QuIC consists of cycles of incubation steps, where the size of the amyloid product, PrP^Res, increases exponentially and vigorous shaking steps, which fragment PrP^Sc aggregates into smaller seeds promoting the conversion (Fig. 1). The PrP^Sc-seeded conversion reaction product, PrP^Res, is enriched in β-sheet secondary structures. The generation of these β-sheets can be monitored in real-time by using a temperature-controlled shaking fluorescence plate reader and by adding a fluorescence dye, Th-T, in the reaction mix.

In contrast to previous diagnostic assays (colorimetric ELISA, Western blot etc.) for analysis of surrogate biomarker proteins (such as tau and 14–3-3), whose regulation in CSF is based on neurodegeneration process in brain tissue, RT-QuIC analysis allows an ultrasensitive (femtomolar) detection of pathogenic proteins, exhibiting a specificity close to 100% [92]. Recently, the RT-QuIC was successfully applied for the template-induced seeding conversion of α-synuclein, which enables the detection of α-synuclein misfolded species reported by different groups, employing different protocols in regards of the physical conditions (e.g., shaking, temperature, chaotropic agents) and chemical micro-environment (e.g., pH, ionic strength, presence of detergents).

Fairfoul and colleagues detected by RT-QuIC the seeding activity of misfolded α-synuclein derived from both CSF and brain homogenates [104]. The reaction was promoted by the addition of zirconia/silica beads to achieve sensitivity of 95% and specificity of 92–100%, respectively for DLB and PD diagnosis. Further, the authors reported the detection of prodromal stages of DLB, showing positive RT-QuIC results from cases affected by rapid eye movement (REM) behaviour sleep disorder, a hallmark of the prodromal stage of DLB. A biochemical and biophysical characterization of the RT-QuIC products from RT-QuIC reactions seeded with DLB derived brain homogenates was conducted by Sano and co-workers [105]. They obtained an accuracy of 100% specificity, 92% sensitivity for DLB and 95% sensitivity for PD. The tested α-synuclein was either wild-type, phosphorylated at residue 129 (pSer129) (commonly observed in patients with α-synucleinopathies) or non-phosphorylable α-synuclein, containing a point mutation (S129A). The seeding activity of S129A oligomers was higher than that of wild-type oligomers regardless of phosphorylation state. In a second RT-QuIC passage (that is, using the RT-QuIC products as seed for a second run), the highest seeding activity was observed when oligomeric species were produced by S129A α-synuclein in the presence of ATP and casein kinase (CK, that phosphorylates α-synuclein at residue S129), whereas wild-type α-synuclein in presence of either CK alone or both CK and ATP, showed a lower aggregation signal. These results point toward a possible protective effect of pSer129, which may generate inert fibrils faster hindering the aggregation. However, the presence of ATP alone was not investigated, as the presence of the nucleoside may promote the aggregation regardless the primary sequence of the substrate.

Another group, Groveman et al. [106], reported an aggregation of monomeric α-synuclein in reactions seeded with CSF or brain homogenate from DLB and PD patients. 100% of specificity and 93% of sensitivity for α-synucleinopathies (DLB and PD grouped together) was obtained by using as substrate α-synuclein owing a point mutation, K23Q, which confers a lower propensity to self-aggregation.

Manne and co-workers [107] achieved α-synuclein aggregation in the RT-QuIC by using as seed either CSF or brain homogenates in combination of SDS and chaotropic agents (zirconia/silica beads, similarly to Fairfoul’s protocol) to enhance the conversion of monomeric α-synuclein in RT-QuIC reactions, seeded with DLB and PD brain homogenates, with PD CSF and with a subset of progressive supranuclear palsy (PSP, a neurodegenerative disease characterized by tau aggregates). Moreover, researchers induced α-synuclein aggregation in reactions seeded with protein extracts from a microglial cell culture infected with pre-formed α-synuclein fibrils. The observations indicate that the seed for the RT-QuIC can also derive from a cell system.

Candelise and co-workers recently published a study in which positive RT-QuIC signals could be obtained specifically from DLB-brain seeded reactions [108]. In order to avoid possible contamination from macromolecules that may skew the aggregation process, a new pre-analytical fractionation protocol for brain homogenates was established. Under avoidance of bead addition (artificial inducer of α-synuclein aggregation), the α-synuclein RT-QuIC was successfully applied for the specific discrimination of DLB patients (positive response) from PD (negative response on control level). Biochemical analyzes confirmed the generation of a fibrillary, proteinase K (PK)-resistant seeding competent α-synuclein species only in DLB-brain seeded RT-QuIC reactions, pointing toward the existence of different α-synuclein strains in DLB and PD cases. Importantly, the RT-QuIC results were neither influenced by the brain region analyzed nor by the co-pathology or other epidemiological factors, such as age, sex, severity of the disease, and age of onset [108]. Controversial findings regarding a lack of α-synuclein seeding in PD patients might be explained by the addition of beads and the pre-analytical treatment (fractionation protocol) [104, 106]. Further factors, such as the kind of purification method of recombinant α-synuclein substrate or the presence of additional compounds (e.g. SDS, metals) may also have an impact on the seeding response.

The standard purification protocol includes a HPLC/FPLC step [105]. This step is particularly important for removing nucleic acids, as the presence of DNA/RNA influences the aggregation kinetics. Moreover, α-synuclein is prone to bind metal ions, specially copper (e.g. His50) which accelerate its aggregation. Therefore, standard protocols to purify recombinant α-synuclein from E. coli include chelator compounds, commonly EDTA, in the purification buffers. Finally, there are different purified α-synuclein species generated during the purification process (oligomeric and monomeric). Specifically the oligomeric species (removable by an additional filtration step) is prone to aggregation which results in a higher seeding propensity in the RT-QuIC as well as in self-aggregation of the substrate.

Garrido and co-workers [109] applied the protocol originally proposed by Fairfoul [104] and used in other studies [110, 111] to assess the presence of α-synuclein aggregation competent species in the CSF derived from α-synucleinopathies patients carrying a genetic PD mutation in the gene coding for leucine -rich repeat kinase 2 (LRRK2). By testing 10 idiopathic PD cases, they found a specificity of 90% and a specificity of 80%. Samples derived from LRRK-PD subjects showed a specificity of 40% (6 out of 15), while 3 of 16 (18.8%) of non-manifesting LRRK2 carriers showed a positive RT-QuIC signal response.

Finally, De Luca et al [112] performed a cross-sectional observational study, carried out by seeding the RT-QuIC reaction with brain material derived from cases affected by PSP, frontotemporal dementia (FTD) and corticobasal degeneration (CBD), PD and MSA. Serial steps of high speed centrifugation were conducted to enrich the seeding-competent α-synuclein species and diluted to a working concentration of 10^− 3. Moreover, they obtained 47 samples of olfactory mucosa (OM) from PD, MSA, PSP and CBD cases. Researchers were able to detect seeding activity from reactions seeded with brain homogenates from PD and MSA patients, while no signal could be detected from reactions seeded with PSP, CBD, FTD or controls. When testing the OM samples, they detected seeding activity in 10 out of 18 PD cases (55.5% sensitivity) and in 9 out of 11 MSA cases (81.8% sensitivity). Thus, the overall sensitivity of PD and MSA grouped cases was 65.5%, whereas negative controls from PSP and CBD gave positive results respectively in 1 out of 6 and 2 out of 12 cases (overall specificity 84.3%). Both brain -and OM seeded reactions showed a lag phase of approximately 40 h. The group further analyzed the RT-QuIC fibrils by transmission electron microscopy, considering the length, the shape and the number of fibrils in the RT-QuIC end-products seeded with OM derived material. They found a statistically significant difference between PD derived fibrils and MSA derived fibrils, with the former owing a shorter over-twist distance compared to the latter. These fibrils were different from those derived from the positive RT-QuIC assays seeded with PSP and CBD material (respectively, PSP derived were longer than MSA, whilst CBD were the shortest produced during the RT-QuIC).

Thus, outside of the great improvement in the usage of a seed derived from a source other than CSF (which dramatically improves the well-being of a patient, compared to the painful procedure of the lumbar puncture), they also produced evidences of a difference in the seeding activity of PD derived α-synuclein compared to MSA derived α-synuclein, indicating a putative strain difference between both pathologies, in line with previous studies reporting putative strain differences among synucleinopathies [113‐115].

Even though several studies (mentioned before) observed a seeding activity of pathogenic α-synuclein, their protocols were different for chemical and physical conditions and, more importantly, for the starting seed and the kind of purified α-synuclein substrate used for the reactions. Most of the groups used human full length α-synuclein from different companies, and different kinds of beads were added in the reaction to accelerate the conversion. In addition, the purified α-synuclein substrate can already be partially aggregated (preformed fibrils) or genetic (PD mutations were included in the substrate sequence), enhancing the seeding propensity in the RT-QuIC reaction. The presence of the positively charged histidine moiety (required for the purification), may further promote the self-aggregation properties of α-synuclein. Moreover, the addition of further compounds such as SDS can change the seeding kinetic and the sensitivity of the α-synuclein RT-QuIC. Differences among protocols and their clinical accuracy (specificity, sensitivity) are summarized in table form (Table 2).

It is therefore apparent that multiple experimental settings may be exploited for the detection of α-synuclein seeding activity in the RT-QuIC. A different “micro-environment” in the reactions, different physical conditions for the aggregation, a different source for the seed, the usage of different substrates or more probably the combinations of these factors, may lead to the detection of α-synuclein seeding activity from different α-synucleinopathies.

Upon protein aggregation, a population of conformational variants could be produced, which compete for the same source (i.e., the native version of the misfolded protein) as a function of the environmental conditions. The different conformations that arise from the same misfolded protein have been referred as strains.

The term “strain” was first introduced to discriminate the pathological phenotypes associated with the intracerebral inoculation of the prion protein in goats [116‐118]. Upon misfolding, a single unit may accommodate into different conformations, giving rise to different pathological frames. Strains hence refer to separate conformations of the same protein. Since their inception in the prion field, a variety of strains has been characterized [119, 120]. The conformation may be altered by the genetic background of the host and the source of the pathogenic protein, as well as post-translational modifications, chemical and physical conditions of the micro-environment in which the process takes place, leading to a heterogeneous strain population.