Tau Strains Could Help Explain Differences Between Neurodegenerative Diseases

I think it's strong. Over the last two decades, there's been a growing body of work that's come from across the global community arguing for the existence of tau strains. Major progress from the Laboratory of Molecular Biology at Cambridge has now validated it at the structural level.

This concept of strains, adopted from the original prion disease hypothesis, states that structure dictates disease phenotypes, kinetics, cell tropism, and brain-region spread in diseases like Creutzfeldt-Jakob and other classic prion diseases—and the field has borrowed this “playbook” from the original prion researchers.

The evidence is strong, but my take on it is that we're not done yet; I don't think we have the full picture. I don't think it'll be as clean as one structure, one strain per one disease; we're just at the tip of the iceberg.

Methods like cryo-electron microscopy (cryo-EM) and the conventional ways of producing samples that allow you to study the structures of tau filaments at the atomic level—they only see what they can see, and that's not a dig at cryo-EM; it's just the limitation of that method. We use it in my lab, too.

As resolution improves, as our methods improve, and as we bridge different disciplines together, we may unlock the rest of the iceberg that's underwater. We'll see that there are perhaps still clusters of strains that exist between diseases and may be shared between certain diseases.

I don't think it'll be this perfect pillar of separation that the field thinks of, based on the cryo-EM data we have. 

This comes back to the idea that tau doesn't exist in a vacuum or void. It's normally functioning, even though I don't think we still have a full handle on what it's doing physiologically.

There's a good consensus on the things that we think it's doing, but the point is, it's a complex protein. It's expressed by six alternative splice isoforms with upwards of hundreds of post-translational modifications, and it’s an ensemble of those proteins (“proteoforms”) doing something physiologically, and ultimately, in these different diseases, coalescing together in different combinations.

“Tau, in itself, is incredibly dynamic. It's this intrinsically disordered protein, and it really doesn't want to fold by itself.” — Dr. Carlo Condello.

You do these experiments in vitro of the many protein-aggregating diseases; you express a protein recombinantly in a cell-free system in a tube, with something like α‐synuclein and amyloid-β, and these will catch fire and aggregate. If you turn away for a second, it's done, right? How does it want to do that?

I think we're now appreciating from the physical chemistry and protein biophysics side that there are certain cofactors that induce it or coerce it to want to misfold.

These tricks have been worked on in the past 15–20 years, and we know that there are these polyanions, charged molecules, that interact with tau, and it requires these negatively charged small molecules to aggregate.

What we know from cellular neuroscience is that the different cell types in the brain are dramatically different in their constituents and the functions they perform. So, we know there's probably a good chance that the cocktail of molecules that are in an astrocyte, oligodendrocyte, or a neuron is vastly different.

There is a neuron-centric idea of neurodegenerative diseases; yes, tau is highly expressed in neurons and has a physiological role in neurons, but I think the field often underappreciates that oligodendrocytes highly express tau as well—it's hugely important in how oligodendrocytes form myelin sheaths.

So, it's huge for oligodendrocyte biology, and astrocytes also express it to a lesser extent. We know less about tau’s role there, but neuropathologists can see tau accumulate in postmortem human brains in all three of these cell types in primary tauopathies.

We know from neuroscience and genomics that these cells are vastly different in what they express and likely what they harbor in terms of small molecules.

So, I think it's both the complexity of the cellular environments and the fact that tau itself is so complex. It requires both coming together to understand the biology.

We have at least two structurally-validated bona fide Alzheimer’s disease positron emission tomography (PET) tracers that are designed against tau. However, when they are applied to PSP or CBD, it's a mess; there's off-target binding.

I think as the field appreciates that this is actually a problem that we have to address, we'll see more of an earnest effort to develop small molecules—certainly for PET tracers, and that’s what my lab is doing in collaboration with colleagues in the IND at UCSF.

I've had the experience of requesting human postmortem samples from non-Alzheimer’s tauopathies, and so often is the case that I see patients were clinically diagnosed with one thing, but then neuropathologically switched and diagnosed with another.

You hardly see that for Alzheimer's anymore; the biomarkers and the neurologists have a very clear Alzheimer's phenotype in most typical cases, but everything else is a gray zone.

If we could have an in vivo diagnosis of PSP vs CBD or an atypical Alzheimer's variant, that would be huge for the field.

Being able to use tracers, especially strain-specific tracers, could not only help a patient be treated more appropriately, but it could also help determine whether they have a comorbid tau neuropathology, which could alter how a neurologist or clinical researcher recruits them into a clinical trial. 

I want to put front and center the idea that these proteinopathies don't manifest in a void; they manifest in one of the most complex organs of our body—and the organ that we really know the least about.

In order to understand tau biology and its drugability, we need to understand the basic mechanisms by which these different cells express and use tau physiologically, and then what goes awry.

We don't know which comes first. We know all these cell types are involved in forming tau, but we only have the snapshot at the end.

Who is actually the culprit? Is it glial cells or neuronal cells? And is it a process that we can ultimately learn about mechanistically and drug very early, or create biomarkers?

Upstream of tau, can we understand the mechanistic biology in these different cell types, and does that matter in the disease?

If we appreciate that the tau conformers are different in disease, let's zoom out and go backward. How do we understand the mechanisms by which we arrived at these different tau strains and different cell types?

If we can understand those basic mechanisms, it could open up a whole new chapter for how we drug, not just tau itself, but how we identify a specific kinase signal transduction pathway that's only in oligodendrocytes that gives rise to that first biochemical change in tau that then kicks off a PSP-specific disease biology well before neurons.