Parkinsonism (tremor, rigidity, and bradykinesia) is a clinical feature of α-synucleinopathies (PD, DLB, and MSA) and many other neurological diseases. Most differential diagnoses, such as MSA [6], tauopathies (progressive supranuclear palsy and corticobasal degeneration) [7], and vascular parkinsonism [8] usually show distinct clinical syndromes. Nonetheless, inaccurate clinical diagnosis of PD may be a common phenomenon, especially in non-specialized centres [9].

In the frame of parkinsonism, DLB and PD-D share neuropathological features of αSyn and LB as well as a variation of clinical features, including cognitive impairments and hallucination [10]. The DLB diagnosis is based on general consensus criteria when cognitive symptoms precede motor pathology by 1 year at least, whereas PD-D is diagnosed when motor impairment occurs at least 1 year before cognitive deficits [11]. Clinical features include the onset of hallucinations, which typically occur spontaneously in DLB, possibly related to temporal lobe dysfunctions, whereas in PD-D, they often appear after dopaminergic treatment [12]. Rigidity and bradykinesia are common parkinsonian features in DLB, whilst tremors at rest are less common. RBD is commonly associated with DLB and should precede cognitive impairment by several years [13]. Besides RBD, hyposmia, constipation and depression are considered to be prodromal symptoms that could help in the differential diagnosis of PD-D versus DLB [13,14,15]. DLB is a type of dementia that is characterized by a more rapid decline in cognitive abilities such as executive function, episodic memory, attention, and constructive abilities compared to PD-D (another type of dementia) [3]. These differences are more evident during early and mid-stages of the pathologies. Nevertheless, they tend to converge at later stages of the disease [16]. Magnetic resonance imaging studies showed greater grey matter loss in parietal, occipital and frontotemporal areas in DLB, which correlate with the visual-spatial hallucination [15]. The selective impairment of neurotransmitter systems also differs between DLB and PD-D. Indeed, whereas the striatal dopaminergic system [17] and the cholinergic system [18] in the basal forebrain and pedunculopontine nucleus are impaired in PD-D, the cortical serotoninergic systems are more impaired in DLB compared to PD [19]. Furthermore, higher tau pathology and amyloid-β deposition have been found in DLB compared to PD-D [20,21,22,23]. To sum up, clinical differences between PD-D and DLB are probably related to disease stage and the sequence of affected anatomic, biochemical or functional compounds of the CNS. In principle, the two diseases may completely mimic each other at some points of the clinical course.

Although the sum of clinical manifestations and the neurobiological changes may indicate differences among synucleinopathies, there is a high degree of overlap not only among synucleinopathies, but also among other neurodegenerative diseases that may appear as parkinsonism but do not show overt αSyn pathology, such as Alzheimer’s disease, corticobasal degeneration and progressive supranuclear palsy [24] (Fig. 1). Therefore, the definitive diagnosis still relies on the post-mortem observation of neuropathological changes associated with the presence of LBs and LNs throughout the nervous system. Detection of LBs and oligodendroglia cytoplasmic inclusions (GCIs) in MSA can be achieved by immunohistochemical staining (Fig. 2a-d). While LBs and GCIs are distinguishable, and LBs in PD and DLB may be observed in different brain regions, they are not discriminable from appearance and structure in the brain tissue (Fig. 2b-d). Control persons without αSyn pathology contain no LBs or GCI; they show neuromelanin-containing cells in the substantia nigra instead (Fig. 2a).

α-Synucleinopathies can be divided into different stages, recapitulating the disease progression useful for sample characterization [25,26]. According to the LB distribution and the progression of pathology, DLB patients can be grouped according to their main pathologies in brainstem, limbic and neocortical. Lewy pathology can be assessed by different types of αSyn antibody [27], which is a major advantage allowing a better neuropathological characterization compared to previously used methods that rely on haematoxylin-eosin or ubiquitin staining [28]. A staging system for αSyn pathology in sporadic PD has been reported by Braak et al. [29]. It is useful to predict the routes of neuronal damage according to the selective susceptibility of neurons and to define the disease progression. According to Braak et al., LB pathology originates from outside the CNS and follows a retrograde pathway through the vagus nerve, because αSyn aggregates had been found also in the peripheral autonomic nervous system [29⇓-31], supporting the body-first model (pathogenic αSyn originating from the enteric nervous system and not from the CNS) [32]. Since the risk of developing dementia in the DLB has been associated with body-first symptoms, whereas brain-first symptoms are associated with PD, which shows a slower progression towards dementia [32], the spreading pathway is supposed to have an influence on the type of -α-synucleinopathy.

Braak stage I is defined by the presence of αSyn inclusions in the dorsal IX/X motor nucleus in the medulla oblongata, in the anterior olfactory nucleus as well as in the intermediate reticular zone. The LNs are found in a higher number than LBs. Braak stage II is characterized by the location of LBs and LNs in the reticular formation and in the caudal raphe nuclei. LBs and LNs are also found at early Braak stages (I-II) in incidental DLB cases (iDLB) in the substantia nigra, with no relationship between αSyn burden and neuronal loss [33]. Braak stage III is characterized by lesions in the magnocellular nuclei of the basal forebrain and in the substantia nigra, while stage IV by evident degeneration and depigmentation of the pars compacta of the substantia nigra. LNs can be detected in the amygdala and in the hippocampus, while LBs are detectable in the anteromedial temporal mesocortex. In Braak stage V, the LB pathology spreads into the neocortex, whilst the burden of lesions increases in the brainstem, olfactory bulb and substantia nigra. The advanced stage VI exhibits LBs and LNs throughout the whole neocortex.

Even though prodromal symptoms such as autonomic dysfunction can be identified by neuropathological diagnostic, the Braak stage classification is restricted to PD, and not applicable to DLB or amygdala-predominant Lewy pathology [34].

A similar classification system for LB pathology has been put forward by the DLB consortium for Lewy pathology, in which PD-D and DLB are grouped together [35,36], including amygdala-dominant cases. The observation of αSyn types or strains exhibiting different features of spreading and influencing the susceptibility of neuronal subtypes, has been proposed to explain differences among α-synucleinopathies.

The spreading of αSyn inclusions is supposed to proceed in the caudo-rostral direction through neuronal transfer [65]. Evidence of neuronal spreading of αSyn pathology was first reported in patients with neuronal graft transplant, who developed Lewy pathology years after the surgical procedure [66,67,68] akin to the prion-like transmission [69]. Misfolded αSyn may either be actively released [70] by neurons through exocytosis [71] or following their death, spread to neighbouring neurons, where it is internalized following the endocytic pathway [72]. Passive diffusion and transfer through tunnelling nanotubes have been proposed as well [73,74]. Spreading along neurites has been shown in different cell models. For instance, in mouse embryonic cultures, fibrillary αSyn was found to be transported in an anterograde direction and taken up by neighbour neurons [75]. Studies in rats showed the neuron-to-neuron spreading of pathogenic αSyn from the olfactory bulb toward non-olfactory areas [76]. Upon injection of pathogenic αSyn in the medulla oblongata of rats, caudal-to-rostral spreading of αSyn was observed, supporting the spreading along interconnected neurons [65]. A similar work was conducted by administering LBs from PD to non-human primates, which later developed PD pathology [77].

Together, these lines of evidence corroborate the central role of αSyn spreading in disease development and progression. Like most neurodegenerative diseases [78,79], synucleinopathies are being recognized as protein-misfolding diseases [80], defined by the template-induced conversion process of one or more proteins. As such, αSyn has been shown with different structures which may lead to different clinical pictures and therefore different diseases. In the next section, we will focus on the impact that different conformations of αSyn may have on the phenotypic variability observed in synucleinopathies.

The sequence of occurrence of symptoms significantly differs between PD and DLB [34], which may reflect differences in the route of spreading of αSyn between the two pathologies [15]. Similarly, neuronal loss diverges between diseases as well as among neuronal areas and subpopulations involved (e.g., glutamatergic neurons in the hippocampus and dentate gyrus [81]). As neuronal populations are widely different, it is suggestive to speculate that each group of neurons displays a specific vulnerability towards different conformations of the same pathogenic seed. Selective vulnerability would account for the different brain regions affected in similar synucleinopathies, also supporting the differential spreading and staging of the diseases. Specifically, cortical neurons may be more sensitive to a DLB-derived strain, whereas substantia nigra areas may be more affected by a PD-derived seed, as reflected by the prevalence of cognitive symptoms observed in DLB compared to the prevalence of motor disorders found in PD. In a recently published paper, Candelise et al. brought evidence of a different seeding activity between PD and DLB (discussed in detail in the next section). They assayed both frontal cortex (FCx)- and substantia nigra pars compacta (SNc)-derived samples and compared the results from different brain regions of the same disease (PD or DLB) by two-way ANOVA [45], before and after the RT-QuIC aggregation assay. The analysis of the total content of αSyn before RT-QuIC showed no significant differences among brain regions of the same pathology. However, the seeding amplification assay showed a significant increase of the signal derived from FCx of DLB patients when compared to reactions seeded with the substantia nigra of the same disease. This result indicates that FCx-derived seeds could promote αSyn aggregation more effciently compared to SNc-derived seeds, supporting the selective vulnerability of different brain areas toward αSyn. FCx neurons, mainly glutamatergic and GABAergic neurons, could be more sensitive toward DLB strains of αSyn, whereas cholinergic and dopaminergic SNc neurons would be less affected by the same αSyn strain, but still be capable of seeding the aggregation assay. This hypothesis is in line with DLB clinical manifestation, showing cortical function impairment prior to the motor symptoms dependent on the nigrostriatal pathway.

Proteinaceous strains are not only critical for the understanding of pathology spreading and the selective routes affected in various forms of synucleinopathies, but may also serve as a biomarker for early and selective detection of the pathology. They are emerging as a powerful diagnostic tool, due to the aggregation techniques that have been developed in the last decade. To this end, promising results have recently been put forward on the ability of aggregation assays to detect prodromal forms of synucleinopathies such as iRBD and PAF [51,63]. Likewise, the recent successful amplification of αSyn seeds derived from skin biopsies of synucleinopathy patients [54,56,57,58,82], from saliva [83] and from olfactory swab [84] may offer the possibility of an accurate, ante-mortem diagnosis of synucleinopathies before the onset of symptoms.

Different folding conformations of pathogenic seeds (referred to as strains) have already been suggested to explain the clinical heterogeneity of neurodegenerative diseases. Different strains of a particular misfolded protein possess distinct biochemical and pathological conformational properties, which are maintained after serial transmissions [85,86]. Previous studies reported a strain-like behaviour of αSyn regarding the transmissibility after inoculations in rodents [87,88,89]. This observation could explain the heterogenic phenotypes in α-synucleinopathies, based on the differential vulnerability of neuronal subset of brain regions toward conformers of the same protein. The prion-like behaviour of αSyn has been reported for MSA αSyn strain [88]. Evidence for the passage of αSyn aggregates from graft to host further corroborates the prion-like nature of αSyn [66,67,68]. The misfolded species may fold in different conformations and each of which may possess its own spreading characteristics [89,90].

A possibility to discriminate between different types of misfolded αSyn in α-synucleinopathies is the analysis of the seeding conversion behaviour of a misfolded αSyn seed either by RT-QuIC or by PMCA.

Candelise et al. recently showed that brain-derived fractions from PD and DLB patients subjected as seeds for RT-QuIC assay, exhibit different signal responses in the RT-QuIC assay [45]. Although DLB and PD are considered variants of the same disease spectrum [2,3,10], a positive RT-QuIC reaction was only obtained from DLB cases but not from PD, suggesting either the existence of different types of αSyn seeds in DLB and PD brains [45] or differential environmental factors influencing seeding conversion activity of a misfolded seed [42]. Although Western blot analysis did not show difference in the total amounts of isolated αSyn seeds between groups, the different intensities of seeding activity might also be affected by minuscule higher amounts of pathogenic αSyn seeds in the DLB cohort compared to PD.

In this context, it has been reported that PD-seeded reactions may show positive seeding responses as demonstrated by several groups [43,47] employing other protocols, indicating the presence of seeding-competent αSyn in PD-derived material (Table 1).

The observed αSyn aggregation from two control cases that later on developed PD, strongly suggests the possibility of applying this detection system to prodromal stages of the pathology. Positive RT-QuIC results from prodromal cases of synucleinopathies, such as iRBD or PAF, have recently been reported in two studies involving large patient cohorts [51], further corroborating the ability of αSyn-seeded RT-QuIC to detect an early stage of the pathological aggregation.

Recently, van der Perren and co-workers reported differences in PMCA-amplified seeds derived from PD, DLB and MSA brain homogenates [91]. Interestingly, MSA strains show similarities with PD strains. However, MSA strains are more potent in provoking motoric deficits, nigrostriatal neurodegeneration, αSyn spreading, and inflammation, indicating a more aggressive nature of these strains. In contrast, DLB strains display more modest neuropathological features in experimental transmission experiments [91].

Semi-quantitative parameters of the RT-QuIC may indicate potential strain distinctions, as different conformations of the same pathogenic seed could show differences in promoting the aggregation. A potential strain-typing scheme for the RT-QuIC response curve may apply semi-quantitative seeding parameters as published before [41]. For example, there are five theoretical RT-QuIC outcomes that may help in the discrimination of the strains. In the first scenario, the presence of signal indicates the ongoing conversion from a pathogenic seed scenario (Fig. 3a). This is the case of a case-control study, in which the material derived from the healthy patient is compared to a diseased one or when the signal could be detected only in one specific group of α-synucleinopathies (e.g., in DLB but not in PD), indicating a putative strain difference underlying the two pathologies [45].

If two pathogenic seeds are compared, various results could be observed. A seed may be faster in triggering the aggregation of its substrate, resulting in a shorter lag-phase (Fig. 3b). A similar result could also be obtained by the same seed applied at different concentrations, highlighting the importance of quantification of the seeds. Although having similar maximum intensities, the two putative strains may be distinguished both by their lag-phase and by area under the curve. In a different case (Fig. 3c), the lag-phases may be identical but the strain may be identified by the lower maximum of fluorescence and hence a lower area under the curve. This may reflect a lower ability of a seed to bind the substrate, resulting in a weaker aggregation. This may be the case observed in Groveman’s work [46], in which PD-seeded reactions showed a reduced maximal fluorescence signal during the run (I_Max) compared to DLB-seeded reactions. This result may also be obtained by a strain with a less avid binding to the detecting dye, as reported by Ferreira and co-workers [92]. The most complex scenario (Fig. 3d) involves seeds with similar latency and avidity for the substrate (lag-phase and maximum intensity, respectively) but different exponential phases. A steeper curve would indicate a stronger ability to convert the substrate once the reaction is triggered. Here, although the maximum intensity and the lag-phases are similar, the two strains may be discriminated by the area under the curve values, with the steeper signal resulting in a larger area. Lastly, a seed may promote conversion faster than another but displaying a lower I_Max (Fig. 3e). This case resembles the observation reported by Shahnawaz and co-workers [52], with MSA-derived seeds showing a faster but less intense ThT-based PMCA signal compared to PD-derived seeds.

All autopsies had been performed either as clinical autopsies with first-line relatives, next of kin or their legally authorized representatives giving informed consent or as legal autopsies on behalf of investigating authorities. The use of specimens obtained at autopsies for research upon anonymization is in accordance with local ethical standards and regulations at the University Medical Center Hamburg-Eppendorf.

Not applicable.

All authors declare no biomedical financial interests or potential conflicts of interest.