Deep learning for Alzheimer's disease: Mapping large-scale histological tau protein for neuroimaging biomarker validation.

Abnormal tau inclusions are hallmarks of Alzheimer's disease and predictors of clinical decline. Several tau PET tracers are available for neurodegenerative disease research, opening avenues for molecular diagnosis in vivo. However, few have been approved for clinical use. Understanding the neurobiological basis of PET signal validation remains problematic because it requires a large-scale, voxel-to-voxel correlation between PET and (immuno) histological signals. Large dimensionality of whole human brains, tissue deformation impacting co-registration, and computing requirements to process terabytes of information preclude proper validation. We developed a computational pipeline to identify and segment particles of interest in billion-pixel digital pathology images to generate quantitative, 3D density maps. The proposed convolutional neural network for immunohistochemistry samples, IHCNet, is at the pipeline's core. We have successfully processed and immunostained over 500 slides from two whole human brains with three phospho-tau antibodies (AT100, AT8, and MC1), spanning several terabytes of images. Our artificial neural network estimated tau inclusion from brain images, which performs with ROC AUC of 0.87, 0.85, and 0.91 for AT100, AT8, and MC1, respectively. Introspection studies further assessed the ability of our trained model to learn tau-related features. We present an end-to-end pipeline to create terabytes-large 3D tau inclusion density maps co-registered to MRI as a means to facilitate validation of PET tracers.

Caspase-6-cleaved tau is relevant in Alzheimer's disease and marginal in four-repeat tauopathies: Diagnostic and therapeutic implications.

AIM: Tau truncation (tr-tau) by active caspase-6 (aCasp-6) generates tau fragments that may be toxic. Yet the relationship between aCasp-6, different forms of tr-tau and hyperphosphorylated tau (p-tau) accumulation in human brains with Alzheimer's disease (AD) and other tauopathies remains unclear. METHODS: We generated two neoepitope monoclonal antibodies against tr-tau sites (D402 and D13) targeted by aCasp-6. Then, we used five-plex immunofluorescence to quantify the neuronal and astroglial burden of aCasp-6, tr-tau, p-tau and their co-occurrence in healthy controls, AD and primary tauopathies. RESULTS: Casp-6 activation was strongest in AD and Pick's disease (PiD) but almost absent in 4-repeat (4R) tauopathies. In neurons, the tr-tau burden was much more abundant in AD and PiD than in 4R tauopathies and disproportionally higher when normalising by p-tau pathology. Tr-tau astrogliopathy was detected in low numbers in 4R tauopathies. Unexpectedly, about half of tr-tau positive neurons in AD and PiD lacked p-tau aggregates, a finding we confirmed using several p-tau antibodies. CONCLUSIONS: Early modulation of aCasp-6 to reduce tr-tau pathology is a promising therapeutic strategy for AD and PiD but is unlikely to benefit 4R tauopathies. The large percentage of tr-tau-positive neurons lacking p-tau suggests that many vulnerable neurons to tau pathology go undetected when using conventional p-tau antibodies. Therapeutic strategies against tr-tau pathology could be necessary to modulate the extent of tau abnormalities in AD. The disproportionally higher burden of tr-tau in AD and PiD supports the development of biofluid biomarkers against tr-tau to detect AD and PiD and differentiate them from 4R tauopathies at a patient level.

18F-flortaucipir PET to autopsy comparisons in Alzheimer's disease and other neurodegenerative diseases.

Few studies have evaluated the relationship between in vivo18F-flortaucipir PET and post-mortem pathology. We sought to compare antemortem 18F-flortaucipir PET to neuropathology in a consecutive series of patients with a broad spectrum of neurodegenerative conditions. Twenty patients were included [mean age at PET 61 years (range 34-76); eight female; median PET-to-autopsy interval of 30 months (range 4-59 months)]. Eight patients had primary Alzheimer's disease pathology, nine had non-Alzheimer tauopathies (progressive supranuclear palsy, corticobasal degeneration, argyrophilic grain disease, and frontotemporal lobar degeneration with MAPT mutations), and three had non-tau frontotemporal lobar degeneration. Using an inferior cerebellar grey matter reference, 80-100-min 18F-flortaucipir PET standardized uptake value ratio (SUVR) images were created. Mean SUVRs were calculated for progressive supranuclear palsy, corticobasal degeneration, and neurofibrillary tangle Braak stage regions of interest, and these values were compared to SUVRs derived from young, non-autopsy, cognitively normal controls used as a standard for tau negativity. W-score maps were generated to highlight areas of increased tracer retention compared to cognitively normal controls, adjusting for age as a covariate. Autopsies were performed blinded to PET results. There was excellent correspondence between areas of 18F-flortaucipir retention, on both SUVR images and W-score maps, and neurofibrillary tangle distribution in patients with primary Alzheimer's disease neuropathology. Patients with non-Alzheimer tauopathies and non-tau frontotemporal lobar degeneration showed a range of tracer retention that was less than Alzheimer's disease, though higher than age-matched, cognitively normal controls. Overall, binding across both tau-positive and tau-negative non-Alzheimer disorders did not reliably correspond with post-mortem tau pathology. 18F-flortaucipir SUVRs in subcortical regions were higher in autopsy-confirmed progressive supranuclear palsy and corticobasal degeneration than in controls, but were similar to values measured in Alzheimer's disease and tau-negative neurodegenerative pathologies. Quantification of 18F-flortaucipir SUVR images at Braak stage regions of interest reliably detected advanced Alzheimer's (Braak VI) pathology. However, patients with earlier Braak stages (Braak I-IV) did not show elevated tracer uptake in these regions compared to young, tau-negative controls. In summary, PET-to-autopsy comparisons confirm that 18F-flortaucipir PET is a reliable biomarker of advanced Braak tau pathology in Alzheimer's disease. The tracer cannot reliably differentiate non-Alzheimer tauopathies and may not detect early Braak stages of neurofibrillary tangle pathology.

Profound degeneration of wake-promoting neurons in Alzheimer's disease.

INTRODUCTION: Sleep-wake disturbances are a common and early feature in Alzheimer's disease (AD). The impact of early tau pathology in wake-promoting neurons (WPNs) remains unclear. METHODS: We performed stereology in postmortem brains from AD individuals and healthy controls to identify quantitative differences in morphological metrics in WPNs. Progressive supranuclear palsy (PSP) and corticobasal degeneration were included as disease-specific controls. RESULTS: The three nuclei studied accumulate considerable amounts of tau inclusions and showed a decrease in neurotransmitter-synthetizing neurons in AD, PSP, and corticobasal degeneration. However, substantial neuronal loss was exclusively found in AD. DISCUSSION: WPNs are extremely vulnerable to AD but not to 4 repeat tauopathies. Considering that WPNs are involved early in AD, such degeneration should be included in the models explaining sleep-wake disturbances in AD and considered when designing a clinical intervention. Sparing of WPNs in PSP, a condition featuring hyperinsomnia, suggest that interventions to suppress the arousal system may benefit patients with PSP.

Spatial genomics of AAVs reveals mechanism of transcriptional crosstalk that enables targeted delivery of large genetic cargo.

Integrating cell type-specific regulatory elements (e.g. enhancers) with recombinant adeno-associated viruses (AAVs) can provide broad and efficient genetic access to specific cell types. However, the packaging capacity of AAVs restricts the size of both the enhancers and the cargo that can be delivered. Transcriptional crosstalk offers a novel paradigm for cell type-specific expression of large cargo, by separating distally-acting regulatory elements into a second AAV genome. Here, we identify and profile transcriptional crosstalk in AAV genomes carrying 11 different enhancers active in mouse brain. To understand transcriptional crosstalk, we develop spatial genomics methods to identify and localize AAV genomes and their concatemeric forms in cultured cells and in tissue. Using these methods, we construct detailed views of the dynamics of AAV transduction and demonstrate that transcriptional crosstalk is dependent upon concatemer formation. Finally, we leverage transcriptional crosstalk to drive expression of a large Cas9 cargo in a cell type-specific manner with systemically-administered engineered AAVs and demonstrate AAV-delivered, minimally-invasive, cell type-specific gene editing in wildtype animals that recapitulates known disease phenotypes.

Molecular characterization of selectively vulnerable neurons in Alzheimer's disease.

Alzheimer's disease (AD) is characterized by the selective vulnerability of specific neuronal populations, the molecular signatures of which are largely unknown. To identify and characterize selectively vulnerable neuronal populations, we used single-nucleus RNA sequencing to profile the caudal entorhinal cortex and the superior frontal gyrus-brain regions where neurofibrillary inclusions and neuronal loss occur early and late in AD, respectively-from postmortem brains spanning the progression of AD-type tau neurofibrillary pathology. We identified RORB as a marker of selectively vulnerable excitatory neurons in the entorhinal cortex and subsequently validated their depletion and selective susceptibility to neurofibrillary inclusions during disease progression using quantitative neuropathological methods. We also discovered an astrocyte subpopulation, likely representing reactive astrocytes, characterized by decreased expression of genes involved in homeostatic functions. Our characterization of selectively vulnerable neurons in AD paves the way for future mechanistic studies of selective vulnerability and potential therapeutic strategies for enhancing neuronal resilience.

Selective Vulnerability of Brainstem Nuclei in Distinct Tauopathies: A Postmortem Study.

The brainstem nuclei of the reticular formation (RF) are critical for regulating homeostasis, behavior, and cognition. RF degenerates in tauopathies including Alzheimer disease (AD), progressive supranuclear palsy (PSP), and corticobasal degeneration (CBD). Although the burden of phopho-tau inclusion is high across these diseases, suggesting a similar vulnerability pattern, a distinct RF-associated clinical phenotype in these diseases indicates the opposite. To compare patterns of RF selective vulnerability to tauopathies, we analyzed 5 RF nuclei in tissue from 14 AD, 14 CBD, 10 PSP, and 3 control cases. Multidimensional quantitative analysis unraveled discernable differences on how these nuclei are vulnerable to AD, CBD, and PSP. For instance, PSP and CBD accrued more tau inclusions than AD in locus coeruleus, suggesting a lower vulnerability to AD. However, locus coeruleus neuronal loss in AD was so extreme that few neurons remained to develop aggregates. Likewise, tau burden in gigantocellular nucleus was low in AD and high in PSP, but few GABAergic neurons were present in AD. This challenges the hypothesis that gigantocellular nucleus neuronal loss underlies REM behavioral disorders because REM behavioral disorders rarely manifests in AD. This study provides foundation for characterizing the clinical consequences of RF degeneration in tauopathies and guiding customized treatment.