Tau’s function and dysfunction in the brain: when small changes have big consequences

With the increase of life expectancy and population growth, neurodegenerative diseases have risen too and are projected to be a major health public concern by 2050. Neurodegenerative diseases are characterized by the progressive decline of cognitive function leading to the subsequent loss of autonomy. Although the underlying causes of neurodegeneration are not well understood, aging is the main risk factor. Interestingly, more than 17 neurodegenerative diseases present Tau protein dysregulation as a hallmark of degeneration. Tau protein, which can be hyperphosphorylated, hyperacetylated, cleaved, alternative spliced, mutated, and form neurofibrillary tangles (NFT’s), has multiple cellular functions that could potentially be compromised and lead to neurodegeneration. Nevertheless, Tau’s biology remains largely unknown to this day despite its wide interest in the scientific community. In this perspective, we highlight some of the layers of regulation that make Tau’s universe so elusive and challenging for its apprehension and the ramifications of such complexity with emphasis at the nucleus.

Since its discovery in 1975, Tau protein has been extensively studied given its pivotal importance on Alzheimer’s disease. Initially, Tau was described for its role in stabilizing neuronal microtubules and maintenance of cytoskeleton. Since then, the knowledge about its functions has expanded to include many nuclear functions, for example, DNA protection under cellular stress, chromosome stability, heterochromatinization of rDNA, and regulation of transcription at the nucleolus, among others (Bukar Maina et al., 2016). How Tau regulates such a vast range of cellular processes is still unknown. Most studies on Tau’s biology have been correlative and descriptive, leaving a vacuum on the molecular mechanisms that govern Tau’s functions. Tau’s elusiveness can be greatly explained in its multifactorial regulatory complexity which involves:

Tau experiences alternative splicing: Tau mRNA undergoes alternative splicing giving rise to six protein variants. These variants differ from one another in the number of N domains at the projection domain (0, 1, and 2) which has been hypothesized to modulate its specific subcellular localization. The variants also differ in the number of microtubule binding repeats (that can include 3 or 4), which are responsible for the interaction with the microtubules, as well as critical for Tau’s interaction with DNA (Figure 1A). Interestingly, all six Tau isoforms are expressed in the brain. These variants can also be differentially expressed during development such as Tau 0N3R which is specifically expressed during neonatal development, or during diseases such as neurodegenerative diseases when Tau variants 0N3R and 0N4R become highly dysregulated. Nevertheless, both the role of these developmental variants and the reason they are re-expressed during disease have not yet been explored.

Figure 1.

Through the lenses of Tau’s complexity.

(A) Alternative splicing of the Tau gene gives rise to six known isoforms that are differentially expressed in the brain. Tau protein isoforms are structurally flexible and it is likely their structure is defined by their binding partners; thus, their flexibility allows for a wide range of interactions and conformations. Although Tau isoforms are similar to each other, they possess differences in their sequence which in turn leads to variability in their post-translational modification profiles (B) and cleavage patterns (C). These changes are not trivial, and their combined effect is likely to trigger major changes in localization, protein interactors, stability, and aggregation, among different byproducts. Such changes occur during DNA damage (D) where full-length Tau and its fragments increase in the nuclear compartment where they accumulate and may lead to nuclear dysregulation. Created with BioRender.com.

Tau is an intrinsically disordered protein: Tau belongs to the intrinsically disordered protein family and does not have a well-defined tertiary structure when free. However, when bound to its cellular interactors in a process known as “folding-upon binding mechanism”, it may acquire a wide variety of ordered states (Luo et al., 2014). Therefore, given that structure determines function, and that Tau’s tertiary conformation could be determined by its specific interactors, we should consider the phrase: “show me your friends and I’ll tell you who you are”. There could be a wide variety of unexplored Tau functions that come to be only within specific Tau interaction contexts. For example, Tau’s interaction with FKRBP5 could prevent its nuclear translocation as it does with Glucocorticoid receptor, thus affecting subcellular location.

Tau posttranslational modifications (PTM): Tau’s protein sequence is enriched with amino acids that can be modified. Tau has been known to be phosphorylated, acetylated, methylated, ubiquitinylated, SUMOylatied, and N-glycosylated, among others. These modifications are highly regulated and can modulate the affinity by which Tau binds to microtubules and other interactors, its stability, subcellular localization, turnover rate, aggregation capacity, and neurofibrillary tangle formation once they occur (Kaluski et al., 2017; Alquezar et al., 2020; Portillo et al., 2021; Figure 1B). Interestingly, a recent study reported that most of Tau PTM’s found in neurodegenerative disease are present in healthy brains as well, but at different levels (Morris et al., 2015). Since some of Tau’s modifiers are present only in certain sub-cellular locations, the type of modification(s) will also differ depending on the cellular compartments in which Tau is located (Ulrich et al., 2018; Portillo et al., 2021). Moreover, some Tau residues can be post-translationally modified by more than one PTM, suggesting some “competition” for these sites and their functions (Morris et al., 2015; Alquezar et al., 2020).

Finally, Tau PTM hierarchy is not well understood, nor which modifications prime and prevent others from occurring, or their space-temporal occurrence. For example, we showed a switch in nuclear location due to acetylation at residue 174. Interestingly, Tau acetyl-mimic mutant (Tau-K174Q) leads to a recapitulation of its nuclear localization, further affecting ribosomal gene expression and rates of protein translation. In contrast, the unmodifiable one (arginine mutant, Tau-K174R) cannot enter the nucleus or affect these genes.

Tau cleavage: Adding to the layers of complexity of Tau is the formation of its fragments. Tau can be targeted by a wide variety of proteases. For example, Tau has recognition motifs in its sequence for ADAM10, caspase 1, 2, 3, 6, 7, and 8, as well as calpain 1 and 2 (Boyarko and Hook, 2021; Figure 1C). In addition to this, it has been reported that Tau has intrinsic self-proteolytic activity coupled to autoacetylation. Tau truncation is a vital aspect of its biology to be understood since specific fragments form aggregates or NFT’s, while concomitantly, they interact and sequester other molecules thus rendering them incapable of functioning (Lester et al., 2021). Remarkably, Tau protein contains most of the sequences that regulate its clearance at the microtubule binding domain, therefore, the lifetime of their byproducts is uncoupled when it is cleaved.

Stress response sensor? It is likely that the outcome of the integration of these layers of regulation leads to the generation of a myriad of molecules that share fragments of Tau’s sequence but may behave as different molecules since their PTM profile, binding partners, subcellular localization, and the segments of Tau sequence may radically diverge from others. Such processes seem to be related to cellular stress. Indeed, there is a growing body of evidence that hints to Tau as a neuronal stress sensor, thus making it vital in neuronal homeostasis compared with the role it was previously thought to have played as a bystander. Among these stresses, Tau has been found to respond to oxidative stress, heat shock, glucose deprivation, loss of proteostasis, and genotoxic stress (Sultan et al., 2011; Ulrich et al., 2018; Asada-Utsugi et al., 2022). However, how these stresses affect Tau’s functions requires further research.

Nuclear roles of Tau: One of the main mechanisms of the cell to cope with stress is to induce changes in gene expression, which requires the transduction of signals to the nucleus. Tau translocates to the nucleus under various stress signals such as heat shock, oxidative stress, and irradiation (Sultan et al., 2011; Portillo et al., 2021). Among its functions in the nuclear compartment is heterochromatin structure (Frost et al., 2014) and can bind and protect DNA from damage (although the mechanism is unknown). When DNA is damaged, Tau 174Ac can induce changes in the nucleolus size and transcription of ribosomal genes and other pathways, helping the cell cope with the genotoxic stress.

One of the leading causes of aging and age-related diseases is DNA damage. It can trigger striking effects that include changes in Tau’s PTM profile, the generation of a myriad of Tau fragments, changes in subcellular localization, and stability. Interestingly, this phenomenon was recapitulated by cells susceptible to DNA damage accumulation, such as SIRT6 deficient cells. These changes in stability, fragmentation, and nuclear localization were in part dependent on the acetylation at residue 174 carried out by CREB-binding protein. This single PTM dramatically changed the levels, location, and function of Tau, thus further highlighting the effect a single PTM can have on a protein (Figure 1D; Portillo et al., 2021).

Recently, Tau’s nuclear function has been shown to have an important role in driving neurodegenerative diseases. Given that the integration of cellular signals occurs in the nucleus, and that Tau’s levels correlate with histone modifications such as H3K9ac (a mark known to be at transcriptionally active genes), it is critical to elucidate both the molecular mechanisms that drive Tau to the nucleus and its multiplicity of consequences when in the nucleus. One example of our poor understanding of Tau’s role as a driver of neurodegenerative diseases is that hyperphosphorylated Tau is a main marker of neurodegeneration, but is this modification pathological per se? It is likely that phosphorylation protects the cell by means of preventing Tau’s nuclear translocation (hyperphosphorylated Tau in the Tau1 epitope cannot enter the nucleus) where Tau’s accumulation can lead to both DNA and nuclear structural damage. This challenges the hypothesis that phosphorylation is a deleterious modification and the usage of Tau’s kinase inhibitors as a viable therapeutic option. Interestingly, NFT’s are formed by phosphorylated Tau fragments that contain the Microtubule binding repeats, also necessary to bind DNA in the nucleus. This in turn suggests the possibility of NFT’s as being a coping mechanism by which the cell prevents Tau fragments with DNA binding properties from entering the nucleus where, if they do, they can lead to both gene and architectural dysregulation (Figure 1D). Moreover, the fate of Tau fragments that do not form part of NFT’s, and their contribution to pathology such as Alzheimer’s disease, remains mostly unknown.

In brief, decoupling of Tau functions through its expression, splicing, PTMs, and fragmentation could be a powerful driver of neurodegeneration, therefore it is imperative to understand its contribution at a particular and combined level to better understand its effect, such as that observed in Tauopathies.

In conclusion, many unanswered questions concerning Tau biology need to be addressed. Of critical importance is the detailed role of Tau in the nucleus and its different compartments which have the potential to explain a wide variety of phenotypes in Tauopathies. Moreover, the main differences in Tau isoforms, fragments, and PTM code should be delineated since it is likely that their differences go beyond the well-described roles in the cytoplasm.

We thank Rebecca Dryer (Medical School for International Health, Ben Gurion University of the Negev, Be’er Sheva, Israel) and Christine Osborn (Department of Molecular and Cell Biology, University of Cape Town, Rondebosch, South Africa) for their constructive criticism and proofreading of this manuscript.

This work was supported by The David and Inez Myers Foundation, Beachwood, OH, USA (to DT).