In this symposium titled ‘Human Neuropathology: Tau, a key protein in the human brain,’ Dr Corey T. McMillan started with discussion of TAR DNA-binding protein 43 (TDP-43) proteinopathy staging and subtyping then Dr Sònia Sirisi Dolcet presented data regarding how in Alzheimer’s disease (AD), occurrence of synaptic oligomeric tau may be an early event. Dr Jennifer S. Rabin then discussed how cognitive decline may arise from interactions between cerebral amyloid angiopathy (CAA) and parenchymal amyloid-beta (Aβ) and Dr Frederique J. Hart de Ruyter presented data on the assessment of retinal phosphorylated tau (p-tau) and Aβ in neurodegenerative diseases. Next, Dr Jamie M. Walker discussed the use of Nanostring Digital Spatial Profiling (DSP) to detect protein expression in the hippocampus in people with AD compared to those without and the symposium ended with Dr Brian D. Hitt showing data from his work on developing antibodies to detect tau seeding in in vitro cells and brain tissue.
TDP-43 proteinopathies: Subtyping and staging
TDP-43 proteinopathies have different neuropathological staging patterns that, according to Dr McMillan, do not completely account for mixed neuropathology.1,2 Subtype and Stage Inference (SuStaIn) is an algorithm to aid disease progression modelling that attempts to identify temporal and spatial TDP-43 spread utilising brain cross-sectional ‘snapshots’ from imaging data.3 Dr McMillan’s study compared SuStaIn staging utilising ordinal (0−3+) neuropathological ratings in 21 anatomic regions.
In amyotrophic lateral sclerosis (ALS), a pattern of staging was identified whereby pathology first develops in the spinal cord then spreads to the medulla and motor cortex then the frontal, angular and temporal cortices. There was some uncertainty in the model in early stages. In frontotemporal lobar degeneration (FTLD), there was more model certainty, especially in the middle stages of disease where there was a widespread pattern of progression. In limbic-predominant age-related TDP-43 encephalopathy (LATE) ±AD, in early stages there was more model certainty. Here, pathology was shown in the amygdala, which then spread through medial temporal lobes, then the anterior cingulate, orbitofrontal cortex then lateral temporal cortex. Both similarities and differences were found between FTLD and LATE.
The SuStaIn algorithm can help model TDP-43 proteinopathy staging
The SuStaIn outputs relative to neuropathological standards for staging showed generally good correlation except for ALS, where there was some uncertainty. Whereas with some standard staging methods around 12% of FTLD cases do not fit the pattern,1 with SuStaIn, the majority did fit. Additionally found were potential subtypes of ALS and FTLD and homogenous LATE progression.
Early synaptic oligomeric tau may occur in Alzheimer’s disease
P-tau accumulation has been shown both pre- and post-synaptically in people with AD.4 Dr Dolcet discussed how tau seeds generated by different phosphorylated oligomers may precede tangle formation5 and asked if synaptic oligomeric tau could be an early event in AD. Her work utilised array tomography on serial sections of superior temporal and primary visual areas in healthy controls (n=24; mean age 78.12) and people with sporadic AD (n=29; mean age 83.69, Braak stages V−VI).
Tau seeding can occur early in Alzheimer’s disease both synaptically and post-synaptically
A decrease in excitatory synaptic density was found in people with AD compared to controls, correlating with dystrophic neurites. P-tau and P22+ oligomeric tau was found to colocalise with synaptophysin in synapses. P22+ oligomeric tau forms were found in the synapse in AD cases including globular oligomers in dystrophic neurites and fibrils in the synapse and post-synaptic terminals. Oligomeric tau was also found in regions without AD-relevant pathology. In the temporal cortex there was a positive correlation between oligomeric tau and neurofibrillary tangles (NFTs); however, synaptic oligomeric tau was found to precede NFT presence. Though less dense post-synaptically, the presence of oligomeric tau in these regions suggested, said Dr Dolcet, that trans-synaptic tau seeding could be possible. She also concluded that pathological tau forms appearing early in AD may be a future therapeutic target.
A potential link between vascular burden and Alzheimer’s disease pathology
CAA occurrence, with a build-up of amyloid in the cerebrovascularture, increases the risk of AD.6 Dr Rabin’s study aimed to examine the relationship between CAA, tau deposition, cognitive decline and parenchymal Aβ burden and assess whether the association between CAA and cognitive decline is mediated by tau burden. The cohort included participants with no/mild CAA (n=1101; mean age at death 89.9; 39.1% with dementia) or moderate/severe CAA (n=621; mean age at death 90.5; 56.0% with dementia). Prior to death, cognition was assessed annually for a mean of 8.8 years.
Cognitive decline in Alzheimer’s disease may be exacerbated by both cerebral amyloid angiopathy and Aβ burden
A positive association was found between CAA and neuritic plaque burden, that decreased in the severe CAA groups, and between CAA and tau burden. When stratified by Aβ burden, there was no relationship with CAA in tau in low plaque cases, but there was in higher plaque cases. Correlations were found between CAA, Aβ burden and cognitive decline, which, said Dr Rabin, suggested the former two may promote the latter. In participants with a high plaque burden, the association between CAA and cognitive decline was postulated to be tau mediated. This may mean, discussed Dr Rabin, that CAA accelerates tau burden, that may in turn affect vasculature, necessitating both pathologies to be therapeutically targeted.
Could retinal p-tau and Aβ be utilised in neurodegenerative disease assessment?
Dr Ruyter discussed the need for easily accessible, patient friendly, low cost biomarkers in AD. Neurons in the retina can be directly visualised in a non-invasive way and his studies aimed to assess post-mortem Aβ presence and tau pathology in the retina of people with neurodegenerative diseases (n=46; mean age 71.0) or cognitively normal controls with (n=10; mean age 83.4) or without (n=6; mean age 56.5) Braak neuropathology. Retinas were digitally analysed to assess mean surface area with immunoreactivity to early and late stage disease p-tau antibodies.
Retinal p-tau may help predict brain p-tau burden
Only in three cases was Aβ detected in the retina, including one AD case, no plaques were observed. Depending on the epitope, p-tau showed a diffuse signal in AD cases compared to none in controls. There were significant differences in quantification of p-tau between AD cases (n=17) and controls and other tauopathies (n=9) and controls. P-tau was only found in the retina in the far and mid periphery regions, not the central region. It was also found that when p-tau in the retina was low, so was p-tau in the brain. However, this was not necessarily vice versa leading Dr Ruyter to conclude that retinal p-tau as a biomarker was specific but not sensitive.
Using spatial proteomics to ascertain differences in hippocampus neuropathology in Alzheimer's disease, primary age-related tauopathy and ‘resistant and resilient’ individuals
Dr Walker discussed Nanostring DSP. Here, protein expression with 73 UV-photocleavable oligo-conjugated antibodies (‘multiplex technology’) in hippocampal sections (CA1, CA2, entorhinal cortex) were examined with 18 regions of interest per slide including neurons with or without NFTs and the immediate neuronal microenvironment. Protein expression was examined in people with AD or primary age-related tauopathy (PART) compared to those without (‘resistant’ and ‘resilient’ cohorts).7
Protein expression levels differ between people with or without Alzheimer’s disease or primary age-related tauopathy
Dr Walker found a correlation between NFTs in neurons and increased expression of Ab-processing and tangle-located proteins, which, she postulated, “reinforces the assumption that tangle and plaque pathology may develop synergistically in AD.” Also shown was significantly higher expression in the ‘resilient’ and ‘resistant’ cases of proteins involved in misfolded protein degradation, amyloid processing, neuronal integrity maintenance and cholesterol removal. In PART and ‘resistant’ cases there were higher levels of microglia markers. Dr Walker concluded that ‘resistance’ and ‘resilience’ may be linked to higher expression of protective proteins and neuronal integrity.
Developing antibodies to detect tau seeding in Alzheimer’s disease and other tauopathies
Previous work shows that even tau monomers can be seeds, with a hairpin domain within the R2 repeat domain end terminal being key.8 Tau seeding has been detected by Dr Hitt using a biosensor Förster Resonance Energy Transfer (FRET) assay. Here, the repeat domain of tau is coupled to a biosensor such that a FRET signal is produced when tau seeds enter an in vitro cell and trigger aggregation. This can be visualised and can be quantified using flow cytometry. Dr Hitt aimed to make monoclonal antibodies to linear peptide antigens in the R1−R2 and R1−R3 transitions of the repeat domain, substituting in a fluorinated proline residue that favours seeding confirmations.
Following antibody generation, the most sensitive and specific forms were selected for detecting AD versus non-AD tauopathies, named MD2.2 and MD3.1. While in control brains there was no seeding activity, in AD brains, most of the seeding activity was found in the immunoprecipitate, showing the antibodies efficiently bind tau seeds. In Progressive Supranuclear Palsy (PSP) brains, this was similar but in Corticobasal degeneration (CBD) and Pick disease (PiD) brains, most of the seeding activity was left in the supernatent so the antibodies were not as sensitive in binding tau seeds.
Antibodies to tau seeds can reveal neuropatholgical differences between tauopathies
Tissue staining showed MD2.2/MD3.1 in the AD cases in perinuclear granular aggregates without neuritic fibrils, with a smaller amount shown in PSP cases and little staining in PiD or CBD cases. In conclusion, the antibodies MD2.2 and MD2.1 were shown to be novel anti-tau agents targeting seed-compentent but not inert tau, predominantly in AD and PSP cases. These may be useful as tools for studying tau seeding in brain tissue.
