Neuropathological hallmarks in the post-mortem retina of neurodegenerative diseases

Abstract

The retina is increasingly recognised as a potential source of biomarkers for neurodegenerative diseases. Hallmark protein aggregates in the retinal neuronal tissue could be imaged through light non-invasively. Post-mortem studies have already shown the presence of specific hallmark proteins in Alzheimer's disease, primary tauopathies, synucleinopathies and frontotemporal lobar degeneration. This study aims to assess proteinopathy in a post-mortem cohort with different neurodegenerative diseases and assess the presence of the primary pathology in the retina. Post-mortem eyes were collected in collaboration with the Netherlands Brain Bank from donors with Alzheimer's disease (n = 17), primary tauopathies (n = 8), synucleinopathies (n = 27), frontotemporal lobar degeneration (n = 8), mixed pathology (n = 11), other neurodegenerative diseases (n = 6), and cognitively normal controls (n = 25). Multiple cross sections of the retina and optic nerve tissue were immunostained using antibodies against pTau Ser202/Thr205 (AT8), amyloid-beta (4G8), alpha-synuclein (LB509), pTDP-43 Ser409/410 and p62-lck ligand (p62) and were assessed for the presence of aggregates and inclusions. pTau pathology was observed as a diffuse signal in Alzheimer's disease, primary tauopathies and controls with Alzheimer's disease neuropathological changes. Amyloid-beta was observed in the vessel wall and as cytoplasmic granular deposits in all groups. Alpha-synuclein pathology was observed as Lewy neurites in the retina in synucleinopathies associated with Lewy pathology and as oligodendroglial cytoplasmic inclusions in the optic nerve in multiple system atrophy. Anti-pTDP-43 generally showed typical neuronal cytoplasmic inclusion bodies in cases with frontotemporal lobar degeneration with TDP-43 and also in cases with later stages of limbic-associated TDP-43 encephalopathy. P62 showed inclusion bodies similar to those seen with anti-pTDP-43. Furthermore, pTau and alpha-synuclein pathology were significantly associated with increasing Braak stages for neurofibrillary tangles and Lewy bodies, respectively. Mixed pathology cases in this cohort consisted of cases (n = 6) with high Braak LB stages (> 4) and low or moderate AD pathology, high AD pathology (n = 1, Braak NFT 6, Thal phase 5) with moderate LB pathology, or a combination of low/moderate scores for different pathology scores in the brain (n = 4). There were no cases with advanced co-pathologies. In seven cases with Braak LB ≥ 4, LB pathology was observed in the retina, while tau pathology in the retina in the mixed pathology group (n = 11) could not be observed. From this study, we conclude that the retina reflects the presence of the major hallmark proteins associated with neurodegenerative diseases. Although low or moderate levels of copathology were found in the brains of most cases, the retina primarily manifested protein aggregates associated with the main neurodegenerative disease. These findings indicate that with appropriate retinal imaging techniques, retinal biomarkers have the potential to become highly accurate indicators for diagnosing the major neurodegenerative diseases of the brain.

Keywords: Alpha-synuclein; Amyloid beta; Eye, retina; Immunohistochemistry; Neurodegeneration; Neuropathology; TDP-43; Tau.

Fig. 1

Localization and morphology of pTau, Aβ, αSyn, pTDP-43 and p62 in the retina. The medial frontal gyrus of cases with AD and FTLD-TDP was used as a positive control brain tissue for a pTau Ser202/Thr205 (AT8), c Aβ (4G8), k pTDP-43 and m p62. Mesencephalon tissue of a PD case was used as a positive control for g αSyn (KM51). b pTau was observed as a diffuse signal in the inner and outer plexiform layers of the retina, shown here in an AD case. Aβ was observed as d aggregates in the vessel wall in the inner retinal layers, e globular deposits in the photoreceptor layer and f cytoplasmic granular deposits in the ganglion cells. Anti-αSyn revealed Lewy neurites in h the inner plexiform layer of the retina and i the optic nerve, with the latter white matter tissue also revealing j oligodendroglial cytoplasmic inclusions with anti-SOX-10 labelling oligodendrocyte nuclei (Liquid Permanent Red). l pTDP-43 showed typical neuronal cytoplasmic inclusions in the outer plexiform layer. n Immunopositive cytoplasmic inclusions were observed with p62 showing overlap with the inclusions observed with anti-pTDP-43. o Furthermore, p62 showed intranuclear inclusions in the inner nuclear layer, shown here in a case with neuronal intranuclear inclusion disease. Immunostaining is shown with DAB (brown), and nuclei are counterstained with haematoxylin (blue). In j, nuclei are immunolabelled and stained with Liquid Permanent Red

Fig. 2

Prevalence of retinal proteinopathies within the different clinicopathological subgroups. The sunburst graphs show the presence (brown) and absence (blue) of proteinopathy and the prevalence in the different subgroups. a pTau was present in 39% of the cases and was present in all typical AD cases and most primary tauopathy cases (7/8). Furthermore, pTau was observed in cases with controls with neuropathologic changes (controls +), synucleinopathies and FTLD-TDP-43, all of which showed Braak stages I–III for neurofibrillary tangles. b Aβ was present in 26% of all cases and showed variable presence in all groups except controls without neuropathological changes (controls −) and atypical AD. c αSyn pathology was present in 30% of all cases and primarily seen in cases with synucleinopathies. Other cases that showed αSyn pathology were synucleinopathies with mixed (tau) pathology (7/11) and one FTLD-TDP-43 case. d pTDP-43 pathology was observed in 21% of the cases and was also variably observed in different groups, although the highest incidence was found for FTLD-TDP-43 cases (7/8). e p62 generally showed immunopositivity in all groups, in 56% of all cases, and seemed to be specifically associated with FTLD-TPD-43, showing the highest incidence here (7/8). AD Alzheimer’s disease; controls −/ + controls without/with neuropathological changes; FTLD-TDP frontotemporal lobar degeneration-TAR DNA-binding protein; OND other neurodegenerative diseases

Fig. 3

Relation between retinal protein aggregates and neuropathological staging systems. Shown is the percentage of cases with retinal proteinopathy (brown indicates presence, green indicates absence) in the retina nerve associated with neuropathological staging systems. a pTau in de retina significantly correlated with Braak NFT stages I–II and V–VI. b pTau in the retina showed a positive trend in the association with Thal amyloid phase. c No significant association was found between Aβ in the retina and Aβ in the brain. d αSyn pathology in the retina significantly correlated with increasing Braak LB stages. e pTDP-43 in the retina was significantly associated with LATE stages 2 and 3. NFT neurofibrillary tangles; LB Lewy bodies; LATE limbic predominant age-related TDP-43 encephalopathy

Fig. 4

Proteinopathy heterogeneity in the brain is larger than in the retina. The upset plots show the number of cases on the y-axis that have a specific combination of proteins depicted on the x-axis. The Venn diagrams represent the number of cases with one specific protein (in colours). The overlapping circles in the Venn diagrams show less co-occurrence of proteins in the retina (a) than in the brain (b). In the retina, pTau (n = 19) and αSyn (n = 13) are most often observed solely, whereas, in the brain, most cases show copathology of pTau, Aβ and αSyn without pTDP-43 (n = 25). In the retina, only one case showed the presence of all four proteins. The sizes of the circles reflect the number of cases in which a certain protein is observed. In the retina, pTau (green) and αSyn pathology (blue) are most often observed. In brain tissue, the most common pathological proteins observed are pTau (brown) and Aβ (green). pTDP-43 (pink) was shown to be the least prevalent in the retina and brain

Fig. 5

Proteinopathy distribution across different cliniconeuropathological groups. The spider plot illustrates the spread of the different pathological proteins within the retina across the major clinicopathological groups. Data are shown in percentages of cases in which the proteins were observed. Each step towards the outside of the plot represents 10% with a maximum of 100%. The legend shows the cliniconeuropathological groups depicted in different colours. Controls with pathology (orange) showed co-presence of p62 (50%), Aβ (35%), pTau (50%) and pTDP-43 (11%). In typical AD (light grey), retinal pTau is observed in 100% of the cases and co-occurs with p62 in 55% and with Aβ in 25% of these cases. Atypical AD cases (purple) showed p62 positivity (50%), together with some pTDP-43 inclusions (25%). The mixed pathology group cases (blue) showed retinal αSyn pathology in 64%, p62 positivity in 45%, Aβ in 27% and pTDP-43 in 18%. Primary tauopathies (green) showed pTau in nearly 90%, occurring mainly with Aβ (38%) and p62 (30%). Synucleinopathies (dark grey) showed mostly αSyn pathology in the retina (90%), and also showed a prevalence of p62 (55%), pTau (33%), Aβ (26%) and pTDP-43 (19%). FTLD-TDP (brown) showed evenly co-occurrence of pTDP-43 and p62 with 88% and in some cases co-occurrence with pTau and Aβ (25%)