Characterization of Tau Pathology Heterogeneity Across the Alzheimer’s Disease Spectrum

Neurofibrillary tau tangles (NFTs) are a hallmark pathology of Alzheimer’s disease (AD) that forms inside neurons. Although misfolded tau is the primary pathology in a number of diseases (“primary tauopathies”), in Alzheimer’s disease it develops only after the development of amyloid beta pathology, and its structure and location in the brain are disease specific. In AD, tau pathology first emerges in the brain’s temporal lobes and then spreads in a very consistent pattern that in most patients correlates well with clinical symptoms. This predictable propagation is described by Braak stages, named for the scientist who first characterized them from autopsied brains. Scientists hypothesize that the pattern of NFT appearance and accumulation may reflect propagation via networked neurons, which can have axons that extend from one brain region to another.

The initial development of AD tau PET brain imaging enabled scientists to confirm the close temporal relationship among Braak stages of tau pathology, neurodegeneration, and clinical symptoms. Improvements in PET imaging resolution and tracers have further revealed deeper nuances in the spread of NFTs than captured by Braak staging. For example, Drs. Hansson and Mattsson-Carlgren and others have identified four distinct patterns of AD tau deposition (‘AD tau PET subtypes’). They propose that knowing where pathological AD tau accumulates in the brain—and why it deposits in those locations—could eventually lead to better individualized treatments.

Drs. Hansson and Mattsson-Carlgren are truly at the forefront of PET imaging to study tau propagation through the Swedish BioFINDER-2 project. BioFINDER-2 is among the largest AD studies worldwide, enrolling over 1,500 participants since 2017. The study features repeated MRI and PET imaging with comprehensive analyses of genetic, cognitive, CSF, and blood plasma data collected every two years. It also employs the latest generation AD tau PET tracer, which is more sensitive to early AD tau deposition. In their last funding cycle, Drs. Hansson and Mattsson-Carlgren used the BioFINDER-2 dataset to test their hypothesis that there are four AD tau PET subtypes. Each subtype has distinct genetic associations and changes in fluid biomarkers and brain structure and function. They found that the brain regions with the highest burden of AD tau also had the most significant loss of brain tissue (atrophy). They also found distinct patterns of brain activity (detected with functional MRI) specific to each subtype. Most APOE4 carriers aligned with one subtype, suggesting that genetic variations can indeed influence the path of AD tau progression in the brain. Finally, and crucially, they found that the different subtypes had different clinical paths and timelines, establishing the critical relevance of their findings.

Building upon their findings from their prior funding cycle, the team now seeks to further characterize the molecular and genetic factors that are associated with distinct AD tau PET subtypes. Notably, they will also extend their previous analyses to participants at earlier disease stages. They have three experimental aims. In the first aim, they are using state-of-the-art MRI techniques to determine if AD tau deposition and spread are driven by changes in brain microstructure or function. They predict that brain regions physically or functionally connected to regions with high tau levels will be more likely to accumulate tau over time. The second aim focuses on deep characterization of the genetic and molecular signatures unique to each subtype. They are looking for gene variants that distinguish the AD tau PET subtypes (beyond APOE4). They are also analyzing and comparing the levels of 3,000 proteins in cerebrospinal fluid (CSF) (‘proteomics’) as a first step to developing biomarkers specific to each subtype. Since PET imaging is expensive and tau PET imaging continues to be primarily unavailable in clinical settings, biomarkers that can be measured in CSF or blood are important for clinical translation. In the third aim, they are determining if AD tau PET subtypes emerge at early disease stages. This effort involves expanding tau PET subtyping to preclinical AD patients (who may still have low levels of tau) and creating a platform for others to use in clinical research settings. The team is taking advantage of a wealth of biological and imaging data that was collected from patients across the full disease spectrum in the negative ‘A4’ clinical trial (Anti-Amyloid Treatment in Asymptomatic Alzheimer’s) of an early anti-amyloid immunotherapy, solanezumab.

In the first year of funding, the group has made substantial progress toward understanding the biological drivers of tau pathology and developing clinically accessible tools to monitor disease progression. Using multimodal MRI, tau-PET, genetics, proteomics, and machine learning approaches in the BioFINDER-2 cohort, the team has demonstrated that individual differences in brain microstructure and functional connectivity shape where tau accumulates and how it spreads over time. They have also shown that tau production and tau propagation are genetically distinct processes, identifying several genome-wide significant loci associated with these mechanisms. In parallel, the investigators have established a robust quality-controlled proteomics platform spanning more than 8,000 CSF and plasma samples and developed machine learning models capable of generating synthetic tau-PET scans from MRI, plasma biomarkers, and clinical data, providing a promising and scalable alternative when tau-PET imaging is unavailable.

Building on these findings, the team now plans to prospectively evaluate the clinical utility of synthetic tau-PET in diagnostic decision-making while extending their multimodal analyses to better define the molecular mechanisms underlying distinct tau-PET subtypes. Ongoing work will integrate high-dimensional proteomics, genetics, whole-genome sequencing, and mechanistic modeling to identify subtype-specific biomarkers and biological pathways driving tau accumulation and disease progression. The investigators are also developing a comprehensive genetic-proteomics database to enable biomarker discovery, causal inference, and validation across independent Alzheimer’s disease cohorts, with the long-term goal of improving early diagnosis, prognosis, and precision therapeutic strategies.