Characterizing the Range of Tau Forms Linked to Different Brain Diseases

The aim of this project is to validate lab models of tau-related pathology as suitable for preclinical studies targeting different forms of tau.

Tau is best known as a central protein in Alzheimer’s disease, but it occurs in different forms that are associated with different diseases that can cause dementia. Lab models of disease must accurately mimic the disease presentation in order to produce accurate findings that translate to humans. Each tau-associated disease needs to be comprehensively characterized by the type of tau that is being misprocessed.

A screening method for evaluating the many available lab models of tau could ensure that the proper tool is being used for the intended target disease. Henry Pan, PhD, and his colleagues are developing a tool to achieve this purpose, able to rapidly screen for different types of tau in cells and animal models to confirm or isolate the target version of tau.

For their work, the researchers will use a freshly developed neural network that can classify proteins based on structure and features illustrating their interactions. Their plan is to enhance this method, known as EMBER, by adding information from fluorescent tags bound to tau. The fluorescent patterns will vary depending on which form of tau they bind.

Dr. Pan and his coworkers will then compare the “”fluorescent fingerprints”” of various tau proteins from different lab models with those from human brain tissues to assess their similarity. As a final step to confirm any similarities, the team will examine the tau variants at the near-atomic level using electron microscopy on flash-frozen samples.

Confirming the nature of tau strains from lab models is important for research into the causes of Alzheimer’s disease and for drug development studies. This work is expected to generate a tool that can rule in or rule out the validity of current lab models of tau protein variants and ensure that studies focus on models using variants that are most like those in humans.

In this proposal, we will systematically compare tau protein aggregate structures in a widely used mouse model and a newly developed rat model that better mimics human disease progression. Using a high-throughput fluorescence fingerprinting method, we will rapidly compare the rodent aggregate structures and determine if the structures match those found in the diseased patient brain extracts they were infected with. After finding the best matches, we can then validate this match between the rodent model and its human counterpart at a near-atomic resolution using cryogenic electron microscopy.

If there are cases where the fidelity of tau protein aggregate structure is preserved for a given disease and rodent model, then we will have discovered areas of focus for effective drug development. If fidelity is not preserved across all diseases and rodent model conditions, then we know that this widely accepted assumption that fidelity is preserved between human patients and rodent models in drug development is incorrect, and we will have a workflow that can effectively benchmark the development of better drug models since we would have proven that current drug models are ineffective.