What this is
- This research investigates α-synuclein pathology in -negative cases of LRRK2-related Parkinson's disease (PD).
- It compares six LRRK2 cases without Lewy bodies to five cases with disease and six healthy controls.
- The study employs a () to detect non-inclusion α-synuclein aggregates, revealing significant differences in pathology.
Essence
- Non-inclusion α-synuclein aggregates are prevalent in -negative LRRK2 cases, indicating that their lack of Lewy bodies does not reflect an absence of α-synuclein pathology. Instead, it suggests a deficiency in forming characteristic inclusions.
Key takeaways
- -negative LRRK2 cases show 2.5–10× more particulate signal than disease cases. This indicates a higher abundance of non-inclusion α-synuclein aggregates in LRRK2 cases.
- Particulate signal is prominent in pontocerebellar tracts and inferior olivary nuclei in LRRK2 cases, areas not typically affected in idiopathic disease. This suggests unique regional pathology in LRRK2 cases.
- Healthy controls exhibited low levels of particulate signal, with some cases showing higher levels, potentially indicating early-stage pathology. This raises questions about the progression of α-synuclein aggregates.
Caveats
- The small cohort size limits the generalizability of findings. Larger studies are needed to confirm the prevalence of non-inclusion α-synuclein aggregates in LRRK2 cases.
- Heterogeneity in clinical diagnoses among cases may obscure specific pathological features. Future research should focus on more homogeneous groups.
- The absence of complementary methods to confirm non-inclusion α-synuclein aggregates in LRRK2 cases is a limitation, necessitating further validation through additional techniques.
Definitions
- Lewy body: Abnormal aggregates of protein, primarily α-synuclein, found in neurons of patients with Lewy body diseases.
- Proximity ligation assay (PLA): A technique used to detect specific proteins in cells or tissues by amplifying signals from nearby binding events.
Simplified
Introduction
Lewy body diseases (LBDs), a collective term encompassing Parkinson’s disease (PD), PD with dementia (PDD), and dementia with Lewy bodies (DLB), are common and debilitating neurodegenerative diseases [37, 56]. They feature varying degrees of motor symptoms (resting tremor, rigidity, and bradykinesia) as well as non-motor dysfunctions, such as hyposmia, sleep disturbances, gut dysfunction, depression, and dementia [37, 70]. Neuropathologically, PD is defined by neuronal loss particularly in the substantia nigra pars compacta (SNpc) as well as characteristic proteinaceous inclusions in the neurons of multiple brain regions [18]. The inclusion pathology, encompassing Lewy bodies (LBs) in the soma and Lewy neurites (LNs) in neuronal processes, present in stereotypical patterns used to stage disease severity [4, 12]. Aggregated forms of the neuronal protein α-synuclein, phosphorylated on serine-129, are found in these inclusions in both PD and DLB [18, 49]. Though most PD cases are idiopathic, mutations in a number of genes can cause autosomal dominant and recessive forms of monogenic PD, which constitute approx. 10–15% of cases [8, 17, 37, 58, 59]. Of these, mutations in the leucine rich repeat kinase (LRRK2) have been particularly puzzling, as some LRRK2 patients are clinically diagnosed with PD but do not contain LBs upon autopsy [38, 59, 90]. This has led to discussion whether these LB-negative LRRK2-PD cases also represent a synucleinopathy or a completely different aetiology not associated with α-synuclein aggregation [14, 38, 50, 71, 78].
In this study, we used an α-synuclein aggregate-specific proximity ligation assay (PLA) to study six LB-negative LRRK2 cases. Based on dual antibody recognition followed by a rolling circle signal amplification, α-synuclein PLA has become a predominant strategy for in situ detection of small, oligomeric α-synuclein aggregates, which are undetected by immunohistochemistry [36, 63, 75, 76]. We have previously shown that α-synuclein PLA strongly labels non-inclusion α-synuclein aggregates preceding regular Lewy pathology but has variable detection of LBs and LNs, depending on the antibody and PLA kit vendor (Duolink or Navinci) [36]. In the present study, we used the conformation-specific MJFR14-6–4-2 antibody in the Navinci PLA application that equally detects α-synuclein aggregate pathology whether organized into inclusions (i.e., LBs and LNs) or not (i.e., oligomeric) [36]. This allowed us to assess both the well-characterized Lewy pathology and the less studied non-inclusion (oligomeric) α-synuclein pathology in one tissue section. The PLA staining patterns from LB-negative LRRK2 cases were compared with five idiopathic LBD cases (Braak stage 6) and six healthy controls. We demonstrate that these LRRK2 cases consistently display α-synuclein PLA-positive non-inclusion pathology and typically at higher levels than regular LBD cases, while most controls only present with very little PLA signal. These results indicate that α-synuclein aggregation is also a characteristic of LB-negative LRRK2 cases. As such, LB-negative LRRK2 cases may be associated with an inability to form LBs and LNs, rather than a lack of α-synuclein aggregate pathology [15, 16].
Materials and methods
Human tissue
All non-LRRK2 cases had been verified to not contain mutations in LRRK2 or other PD-related genes by genotyping on Illumina’s NeuroX array for neurodegenerative diseases, encompassing more than 24,000 neurodegeneration-specific variants in addition to the standard Illumina exome content of approx. 240,000 variants [52, 53].
| Case ID | Clinical diagnosis | Sex | Age at death (y)a | Disease duration (y) | Postmortem delay (h) | Braak PD stage | LRRK2 mutation(s) | Thal phase (Aβ) | Braak tangle stage (tau) |
|---|---|---|---|---|---|---|---|---|---|
| LRRK2 1 | PD | M | 92 | 11 | 16 | LB-negative | M1646T | 1-2b | 3 |
| LRRK2 2 | Healthy control | F | 94 | – | 24 | LB-negative | N2081D | N/A | N/A |
| LRRK2 3 | Healthy control | F | 87 | – | 76 | LB-negative | M1646T | 5c | 2 |
| LRRK2 4 | PSP | F | 61 | 5 | 43 | LB-negative | N2081D | 0 | 0d |
| LRRK2 5 | PD | M | 84 | 7 | 15 | LB-negative | N2081D, M1646T, P1542S | 0b | 1b,e |
| LRRK2 6 | PD | F | 79 | 9 | 32 | LB-negative | G2019S, N2081D | 2 | 1 |
| LBD 1 | PD | M | 78 | 12 | 32 | 6 | – | 4 | 4 |
| LBD 2 | DLB + AD | M | 85 | N/A | 96 | 6 | – | 5 | 5 |
| LBD 3 | PD | M | 80 | 23 | 24 | 6 | – | N/A | 1 |
| LBD 4 | DLB + AD | F | 79 | N/A | 72 | 6 | – | 5 | 5–6 |
| LBD 5 | PDD | F | 81 | N/A | 48 | 6 | – | N/A | N/A |
| Control 1 | Healthy control | F | 88 | – | 100 | 0 | – | 1 | 1 |
| Control 2 | Healthy control | M | 82 | – | 45 | 0 | – | 2 | 2 |
| Control 3 | Healthy control | M | 78 | – | 93 | 0 | – | 3 | 2 |
| Control 4 | Healthy control | M | 70 | – | 43 | 0 | – | 3 | 2 |
| Control 5 | Healthy control | F | 89 | – | 24 | 0 | – | 1 | 2 |
| Control 6 | Healthy control | F | 89 | – | 30 | 0 | – | 0 | 2 |
Proximity ligation assay
Conformation-specific α-synuclein antibody MJFR14-6-4-2 (MJF-14; Abcam, #ab214033, 1 mg/mL) was conjugated to complementary NaveniLink proximity probes (Navinci, #NL.050) according to the manufacturer’s instructions. Briefly, 10 µL of Modifier was added to 100 µL antibody, before mixing with the lyophilized oligonucleotides (Navenibody 1 or Navenibody 2) and incubation overnight at room temperature (RT). 10 µL of Quencher N was then added and incubated for 15 min at RT, whereafter conjugated antibodies were stored at 4 °C.
The PLA staining was conducted using NaveniBright HRP kits using DAB as the chromogen (Navinci, #NB.MR.HRP.100) and counterstained with a haematoxylin-based nuclear dye. FFPE sections were deparaffinized and rehydrated in decreasing alcohol series, before antigen retrieval of microwaving for 2 × 5 min in sodium citrate (pH 6, DAKO, #S1699). Endogenous peroxidase activity was quenched in 0.3% hydrogen peroxide in PBS, after which samples were blocked in Navinci blocking buffer with supplement 1 for 1 h at 37 °C. PLA-conjugated MJF-14 was diluted 1:10,000 in antibody diluent with supplement 2 and incubated with the samples overnight at 4 °C. After washing off unbound antibody, samples were incubated in freshly prepared Reaction 1 (1 × Buffer 1 diluted in nuclease-free water and supplemented with Enzyme 1) for 30 min at 37 °C. Subsequently, Reaction 2 was similarly prepared (1 × Buffer 2 diluted in nuclease-free water and supplemented with Enzyme 2) and applied for 1 h at 37 °C. Sections were then incubated in HRP detection solution for 30 min at RT, before signal development for 5 min at RT. Finally, sections were briefly counterstained in Navinci Nuclear stain and staining developed in tap water, before a quick dehydration in 100% isopropanol and mounting with VectaMount Express Mounting Medium (VectorLabs, #H-5700) [22, 24–27].
Immunohistochemistry
Immunohistochemistry was performed as previously described for the Syn-O4 antibody [1, 83]. In brief, FFPE sections were deparaffinized and rehydrated as for PLA, followed by antigen retrieval for 15 min in 80% formic acid at RT. Endogenous peroxidase activity was quenched in 3% hydrogen peroxide in PBS, after which samples were blocked in 10% foetal bovine serum. Mouse monoclonal antibody Syn-O4 against aggregated α-synuclein [83] was diluted 1:5,000 and incubated on the sections overnight at 4 °C. Sections were incubated with HRP-conjugated anti-mouse secondary antibody as part of the REAL EnVision detection system (DAKO, #K5007) for 1 h at RT. Signal was visualized with 3,3′-diaminobenzidine (DAB) and sections were counterstained with haematoxylin. Finally, sections were dehydrated in increasing alcohol concentrations, cleared in xylene, and mounted with DPX mounting medium (Sigma, #06522) [40, 48].
Imaging and image analysis
Tissue sections were imaged at X20 magnification on an Olympus VS120 slide scanner and stored in the vsi file format. Digitalized images were opened in QuPath [3], stain vectors estimated using the auto function, and the entire tissue regions were outlined by thresholding on averaged channel values. Specific regions of interest were annotated according to The Brain Atlas and specific articles pertaining individual subregions as indicated in Suppl. Figure 3 [2, 29, 72, 86, 89].
5 total slide scans were selected to optimize parameters for detection of signals for quantitative image analysis. These cases (LBD 1 medulla, LBD 1 pons, LRRK2 6 midbrain, LRRK2 1 posterior hippocampus, and LRRK2 4 amygdala) encompassed different composition and densities of the various signals found across the cohort. Signals were divided into particulate PLA signal (not associated with inclusions) and Lewy-like PLA signal (signal found in structures resembling LBs and LNs) and quantified as area coverage (%) to allow a direct comparison between the two measures. Any neuromelanin in the various regions of interest was manually outlined and excluded from the analysis to avoid the algorithm mistaking it for PLA signal. Then, PLA signals were detected in a three-step protocol. First, all PLA signal was defined using a simple thresholder on the DAB channel with Gaussian pre-filtering with a smoothing sigma of 0.5 and a threshold of 0.1. Next, Lewy-like PLA signal was determined using a simple thresholder from the average channel intensity with Gaussian pre-filtering of 1.0 and a threshold of 80, where anything below the threshold corresponded to Lewy-like PLA. Finally, particulate PLA was defined as signal positive on the DAB thresholder but negative on Lewy pathology thresholder. The image analysis strategy is illustrated in Suppl. Figure 1.
Statistical analysis
To compare levels of particulate and Lewy-like PLA between groups, we ran univariate analyses adjusting for age and sex using SPSS v.29.0.1.0 (IBM). Other potential covariates (post-mortem delay, Braak tau stage, and Aβ Thal phase) were assessed for their influence on the model and generally not included. Data were tested for assumptions of normal distribution of residuals using Shapiro–Wilk’s test, linear relation of covariates to the dependent variable, homogeneity of regression slopes, homoscedasticity, and homogeneity of variances using Levene’s test. In cases where assumptions of normality of residuals and/or homogeneity of variances were violated, the dependent variable was log-transformed, and data were re-tested to assure that they now fulfilled the assumptions. Correction for multiple comparisons in the analyses was done using Bonferroni’s post-test. P values under 0.05 were considered significant (*p < 0.05, **p < 0.01, ***p < 0.001), though values close to significance were also indicated, as the small group sizes limit the chance of reaching statistical significance. All graphs were compiled using GraphPad Prism 10, and data were plotted as mean ± SEM unless otherwise mentioned [60, 65, 67].
Results
The study included six LRRK2 cases, all of which were LB-negative on standard neuropathological assessments. Three of the cases had a clinical diagnosis of PD, one a diagnosis of progressive supranuclear palsy (PSP, neuropathologically confirmed), and two cases had no clinical diagnoses of neurodegenerative disease. These were compared with five idiopathic LBD cases, all of which were Braak stage 6, and six healthy controls, as based on standard clinicopathological assessments. None of the idiopathic LBD cases or the controls harboured any mutations in LRRK2 or other LBD-related genes. Clinical, demographic, genetic, and neuropathological information on the cases is summarized in Table 1.
Regions of interest and PLA terminology.Five brain regions (medulla, pons, midbrain, posterior hippocampus, and amygdala) were included for each case to allow comparative quantitative analyses of PLA signals. See Suppl. Figure 3 for delineation of the listed subregions in three of the cases included in this study.The PLA staining yielded two distinct patterns of signal: the particulate PLA not associated with inclusions (top panels) with increasing signal density from left to right and a Lewy-like PLA staining (bottom panel) with strong morphological similarity to LBs and LNs as stained by standard IHC. Presumed LBs are indicated by arrowheads, while examples of putative LNs are highlighted with arrows. Scale bar = 20 µm (applies to all images). Particulate and Lewy-like PLA were analysed separately in all sections and reported as area covered (%) a b
LRRK2 cases show abundant particulate PLA, while both particulate and Lewy-like PLA is found in LBD cases
In addition to its presence in the LRRK2 cases, some particulate PLA signal was also found in LBD cases and two of controls (Fig. 2a–b). In general, the LB-negative LRRK2 cases contained 2.5–10 times more particulate PLA signal than the LBD cases when analysing the entire tissue sections, though group differences only reached statistical significance for a couple of regions, when correcting for multiple comparisons with Bonferroni (Fig. 2b). This indicates that non-inclusion α-synuclein aggregates are more abundant in the LB-negative LRRK2 cases than LBD cases or controls, though larger groups are needed to determine more precise effect sizes and achieve statistical significance in individual brain regions [88].
Furthermore, even in the LBD cases, the particulate PLA signal was much more abundant than the Lewy-like PLA signal, in line with previously published studies using α-synuclein PLA in LBD cases [36, 75, 76]. This was also apparent when comparing the MJF-14 PLA with Syn-O4 IHC, which showed that especially particulate PLA signal outside the cell bodies, but also non-inclusion accumulations in the neuronal somas, were preferentially detected with PLA but not with IHC (Suppl. Figure 2a). Of note, the ratio between particulate and Lewy-like PLA increased with more rostral sections, with an average area coverage from 41-fold higher in medulla to more than 100-fold higher in posterior hippocampus and amygdala sections (Fig. 2b–c, Suppl. Table 1). Based on the limited cohort in this study, we could not conclude whether this difference in signal ratio was related to region-specific propensities for forming LBs (brainstem versus limbic/cortical), a rostral progression of disease, or other yet unknown causes.
Particulate PLA, Lewy-like PLA, and LB densities in LRRK2 LB-negative cases, regular LBD, and non-neurodegenerative controls.Overview of PLA staining appearance in medulla (around DMV), pons (around LC), midbrain (around SNpc), posterior hippocampus (around CA2), and amygdaloid complex (in the amygdala). Scale bar = 50 µm (applies to all images).Quantitative analysis performed in the entire tissue sections showing particulate PLA area coverage (%,) and Lewy-like PLA area coverage (%,) for the three groups. All graphs show mean ± SEM with data points for individual cases. Univariate analyses adjusting for age and sex followed by Bonferroni’s multiple comparison test. *< 0.05, **< 0.01, ***< 0.001.cornu ammonis 2 of the hippocampus,dorsal motor nucleus of the vagus,locus coeruleus, SNpc (substantia nigra pars compacta) a b–c b c p p p CA2 DMV LC
LRRK2 cases display varying amounts of particulate PLA signal in nuclei and regions typically affected in PD but also in unusual regions
In addition to the prominent brainstem pathology, five out of six LRRK2 cases displayed high levels of particulate PLA in hippocampus and entorhinal cortex, and all cases in amygdala-associated cortex (Fig. 3, Suppl. Figures 3–4). In the hippocampus, particulate PLA signal was found in all hippocampal subregions (dentate gyrus as well as CA1-3), though the specific distribution varied between cases. In entorhinal and amygdala-associated cortex, particulate PLA signal was present in both the superficial (layers 1–3) and deep cortical layers (layer 5–6), without any obvious difference in signal density. This was observed for both the LRRK2 cases and the LBD cases, though LBs were selectively accumulated in the deep cortical layers, in line with previous reports from us and others [11, 30, 36].
Healthy control cases generally displayed low particulate PLA signal, though two cases showed some abundance around specific neuronal populations in the DMV, SNpc, and amygdala (see Controls 2 and 6, Fig. 3 and semi-quantitative overview in Fig. 6).
PLA signal in specific brain regions from selected cases. Specific brain regions were identified from the medulla (dorsal motor nucleus of the vagus, reticular formation, and inferior olivary nucleus), pons (locus coeruleus and raphe nucleus), midbrain (substantia nigra pars compacta, SNpc, and ventral tegmental area, VTA), posterior hippocampus (CA3, CA2, and entorhinal cortex) and amygdala. Three cases were displayed from each group, encompassing most of the variation in signal densities in each group. Scale bar = 10 µm (applies to all images). Note that LRRK2 4 was almost entirely depigmented and had substantial neuronal loss in several regions, incl. SNpc. See also semi-quantitative overview particulate versus Lewy-like PLA in all cases in Fig. 6
Prominent PLA signal in the inferior olivary nucleus of LRRK2 cases.LRRK2 cases displayed particulate PLA signal in the inferior olivary nucleus, with signal predominantly present in the neuropil surrounding the neuronal cell bodies. Note that LRRK2 4, which contained no PLA signal in the medulla, was left out of the figure.Only LBD 1 of the LBD cases displayed PLA signal in the inferior olivary nucleus. Note that in this case, signal is not confined to the neuropil but also found in the neuronal cell bodies. Scale bars = 50 µm. See also semi-quantitative overview particulate versus Lewy-like PLA in all cases in Fig. a b 6
PLA in pontine nuclei, transverse and longitudinal fibres of the basilar pons.Particulate and Lewy-like PLA signal was assessed in six ROIs in the basilar pons of all cases as indicated in the overviews from one LRRK2 case (left) and one LBD case (right). Magnified panels show transverse fibres (outlined in blue) and longitudinal fibres (outlined in yellow), while in-between areas correspond to pontine nuclei. Scale bars = 200 µm.LRRK2 cases displayed dense particulate PLA signal in the transverse (pontocerebellar) fibres and to some degree in the pontine nuclei. Note that LRRK2 4, which was negative in the pons, was left out of the figure. Scale bar = 20 µm (applies to all images).LBD cases displayed much less signal in pontine nuclei and transverse fibres than the LRRK2 cases. One case (LBD 5) presented with some Lewy-like PLA signal in the longitudinal fibre bundles (arrows). Scale bar = 20 µm (applies to all images). See also semi-quantitative overview particulate versus Lewy-like PLA in all cases in Fig.and quantifications in Suppl. Figure 4 a b c 6
Graphical summary of particulate and Lewy-like PLA across all cases and nuclei/regions examined. Based on the quantifications of area coverage, semi-quantitative scales were made for particulate PLA () and Lewy-like PLA (). Note that the same colour inanddenotes a 300-fold difference in area %. LBD cases stand out with a high density of Lewy-like PLA across the regions examined, while LRRK2 cases display the highest particulate PLA density. N/A, not available a b a b
Neuronal loss may explain some of the variation in PLA signal density between cases
Cell loss and extra-neuronal neuromelanin in the locus coeruleus of LBD cases. Locus coeruleus displayed varying degrees of loss of its noradrenergic, neuromelanin-containing neurons in the LBD cases, as illustrated with a progressive fall in the density of melanized neurons (left to right). Extra-neuronal neuromelanin was apparent in some cases (arrowheads in LBD 2 and 5), evincing neuronal loss. PLA signal density was also lower in LBD cases with extensive neuronal loss (arrows indicate examples of Lewy-like PLA). Scale bar = 50 µm
Discussion
In recent years, several studies have demonstrated previously undetected α-synuclein aggregate pathology not organised into LBs in PD, DLB, and multiple system atrophy (MSA) using α-synuclein PLA [36, 63, 75, 76]. With their signal amplification creating a high signal-to-noise ratio, the PLAs are capable of detecting aggregate species with low epitope density, which enables detection of, e.g., α-synuclein oligomers. Oligomers and other small α-synuclein aggregates are generally not detected by IHC, as their epitope density is so low that—even if the antibodies label them—they are indistinguishable from physiological α-synuclein species or deemed as background [5]. Some studies have used proteinase K pretreatment to degrade physiological α-synuclein before IHC, demonstrating abundant synaptic α-synuclein aggregates that appear comparable to PLA-detected non-inclusion aggregates [41, 54, 73, 80]. Nevertheless, the proteinase K pretreatment also affects tissue integrity and is likely to degrade α-synuclein oligomers with a low beta-sheet content [5]. How the α-synuclein PLA signal corresponds to these proteinase K resistant synaptic aggregates and whether oligomers with low beta-sheet content contribute to the α-synuclein PLA signal is currently unknown.
In the present study, we used the Navinci version of the recently published MJF-14 PLA [36], which labels both inclusions of Lewy-type morphology as well as the non-inclusion aggregates. Though the exact nature of the α-synuclein species detected by the PLA is not yet determined, the use of the aggregate-specific MJF-14 antibody should ensure that only pathological α-synuclein species are detected [7, 36, 66]. With this assay, we present the first evidence of a broad accumulation of α-synuclein aggregates not organised into inclusions in multiple brain regions of LRRK2 cases without LBs. Of the six cases examined, five of them presented with significant particulate PLA signal in all the five brain regions examined, but the exact distribution and density differed between the various cases and brain regions investigated. Though most of the LRRK2 mutations in our cases (Table 1) are not traditionally listed as pathogenic/PD-related, there is evidence suggesting a link between PD and mutations in both M1646T and N2081D [13, 42, 64, 79]. Some studies have also shown increased LRRK2 kinase activity with M1646T and N2081D mutations compared to wildtype LRRK2, though to a lower degree than with G2019S LRRK2 [35, 39, 62]. Lastly, M1646T and N2081D mutations in LRRK2 were quite prevalent among PD-diagnosed LRRK2 cases with Lewy pathology in the Parkinson’s UK brain bank—more so than G2019S or R1441C/G/H mutations. With the low penetrance of LRRK2 mutations for development of PD (estimated 25–40% for G2019S LRRK2 [13]), it is clear that there is still much to uncover regarding the pathogenicity of LRRK2 mutations and their interaction with other modulators of disease penetrance.
Curiously, several of the LRRK2 cases did not show particularly dense PLA staining in the brainstem regions such as DMV, LC, and SNpc. Instead, we detected a noteworthy affection of inferior olivary and pontine nuclei as well as pontocerebellar fibres, raising the speculation whether this is a general feature of LRRK2 cases. These regions are generally implicated in MSA rather than the LBDs, though sparse Lewy pathology has been reported in some PD and DLB cases [74, 85]. If this observation is shown to be a common characteristic in LRRK2-mutant cases, it may suggest a slight cerebellar dysfunction in these patients, which should then be investigated clinically. In general, the LRRK2 cases presented with a higher degree of particulate PLA signal than both the regular LBD cases and the non-neurodegenerative controls, though one LBD case did show particulate PLA signal at levels similar to most of the LRRK2 cases [44].
The striking prevalence of particulate PLA in the LRRK2 cases suggests that their lack of LBs does not reflect an absence of disease-associated α-synuclein aggregates but rather an inability to form the characteristic aggregate-containing inclusions. Though knowledge of the mechanisms involved in LB formation is limited, LBs are clearly complex structures, and their formation appears to be a specific cellular response, which might indeed be perturbed in some cases [20, 43, 46, 55, 77, 81, 84]. LRRK2 has mechanistically been associated with vesicular trafficking and autophagy, and mutation-associated perturbations of proteostasis could thus represent a link to the absence of LBs [28, 45, 69]. However, since the absence of LBs is not linked to any specific LRRK2-mutation and varies even within the same family, LRRK2 is not the sole player in the perturbation of LB formation [38, 59, 90]. Consequently, further studies to determine whether such non-inclusion α-synuclein aggregate pathology is a general feature of PD patients with mutations in LRRK2 or in Parkin (PRKN), who also frequently present without LBs, are strongly motivated [19].
Curiously, a recent study using α-synuclein seed amplification assay (SAA) on cerebrospinal fluid (CSF) from more than 1100 participants in the Parkinson’s Progression Marker Initiative (PPMI) cohort showed that SAA-positivity was significantly lower among LRRK2 PD than idiopathic PD patients [78]. At the time of the study, neuropathological examination was only available for a small subset of cases, including one SAA-negative LRRK2 carrier, who did not contain Lewy pathology. In contrast, all cases with a positive SAA showed typical Lewy pathology upon neuropathological examination [78]. At present, the correlation between SAA-positivity and presence of α-synuclein PLA signal has not been studied, though such studies would be highly valuable. A recent study, however, did examine the relation between CSF SAA-positivity and regional presence of Lewy pathology, demonstrating significantly lower CSF SAA-positivity in cases with early (Braak stages 1–3) or focal pathology (e.g., amygdala-predominant) [6]. Moreover, they found SAA-positivity in brain homogenates from multiple brain regions without accompanying Lewy pathology, indicating a seeding activity of smaller or less organized α-synuclein aggregates [6]. As such, the lack of CSF SAA-positivity in some of the LRRK2 cases in the PPMI cohort does not necessarily reflect an absence of α-synuclein aggregates but could also be caused by predominantly focal pathology, which is not reflected in the cerebrospinal fluid [6]. Alternatively, pathology consisting of non-inclusion (presumably oligomeric) α-synuclein may have a lower seeding efficiency than mature α-synuclein fibrils in the SAA, in line with evidence from experimental models [47, 57]. Additional SAA analyses on homogenates from various brain regions could thus be an important next step to assess the presence of seeding-competent α-synuclein aggregates, even in the absence of Lewy pathology. Combined with α-synuclein PLA staining in post-mortem cases from the upcoming follow-up studies on the PPMI cohort, this will enhance our understanding of the LRRK2-related PD neuropathology.
Multiple studies examining LRRK2 carriers (see Suppl. Table 2 for comprehensive list) have shown that LB-negative cases display Alzheimer’s disease (AD) or PSP-like tau pathology or show unspecific nigral cell loss without any underlying pathology [9, 32, 68, 82, 87, 90]. However, other pleomorphic pathologies also include MSA and argyrophilic grain disease [31, 61]. Except for the PSP-diagnosed LRRK2 4, which was confirmed by neuropathology, tau pathology was not prominent among the LRRK2 cases in this cohort, and we found no correlation of particulate PLA with either Braak tangle stage or Aβ Thal phase. The sparse tau pathology in our LRRK2 cases is somewhat contrasting with the findings from Henderson et al. who examined a cohort of 11 LRRK2 cases, of which four were LB-negative [32]. These cases, mostly with G2019S mutation, all contained AD-like tau pathology, though levels in LB-negative cases tended to be lower than in the LB-positive cases [32]. Whether the divergent neuropathological findings are related to specific LRRK2 mutations is yet to be determined in larger cohort studies, as is the exact relationship between particulate PLA and other protein pathologies.
For most of the healthy control cases, stained in parallel with LRRK2 and regular LBD cases, only very low levels of particulate PLA were detected. None of the healthy controls showed any LBs and only had limited, if any, other neurodegenerative proteinaceous inclusions, such as Aβ plaques and tau tangles. Two of the controls (Control 2 and 6), however, did present with higher levels of particulate PLA, which were closer to the levels seen in some of the LBD and LRRK2 cases. This was seen particularly in the amygdala and SNpc but not in DMV or LC, which are usually thought to be affected early on with Lewy-like pathology. Previous studies using α-synuclein PLA have shown that it appears to detect earlier aggregate species than the Lewy pathology [36, 63, 76]. These two controls could thus represent early asymptomatic stages of LBD, perhaps of the brain-first subtype [10, 33, 34], that would have become symptomatic, had they lived longer. As we did have two other controls (Control 4 and 5) that were virtually void of any PLA signal in the study, it is unlikely that the α-synuclein PLA accumulation is a normal, disease-unrelated process. Whether there is a certain threshold above which the PLA-positive species become problematic and how they progress in the brain is currently unknown and requires further studies in larger cohorts of early-stage LBD cases and aged controls.
The main limitation of this study pertains to the relatively small cohort size, which relates to the scarcity of LB-negative LRRK2 cases available for post-mortem examination. As such, larger studies are needed to determine whether the non-inclusion α-synuclein aggregate pathology is a general feature of LRRK2 cases and how its extent and distribution may correlate with other protein pathologies. Additionally, the inclusion of cases with different clinical diagnoses (within both LRRK2 and LBD groups) could obscure features that are specific for one clinical diagnosis but were not apparent in our study due to group heterogeneity. The study is also limited by the lack of complementary methods to confirm the presence of non-inclusion α-synuclein aggregates in the LRRK2 cases without Lewy pathology. One strategy for future studies could be to perform SAA on brain homogenates from regions with and without particulate PLA signal, which would inform on the presence of seeding-competent α-synuclein species in these regions. Nevertheless, we here provide compelling evidence that Lewy pathology is not representative of all α-synuclein aggregates in the brain.
In conclusion, we demonstrate substantial levels of α-synuclein aggregates, but no Lewy-like inclusions, in LB-negative LRRK2 cases. The particulate PLA affects both classical PD-related brain regions in the LRRK2 cases, but also other regions not usually implicated in PD, particularly the inferior olivary nucleus and the pontocerebellar pathways. These results motivate further studies on not only LRRK2 cases but also PRKN PD cases, which also frequently present without LBs upon autopsy. Furthermore, we need to also examine larger aged cohorts to understand how this particulate PLA signal progresses in the brain and when it becomes associated to clinical symptoms.
Supplementary Information
Below is the link to the electronic supplementary material. Supplementary file1 (PDF 14378 KB)