Tau hyperphosphorylation represents one of the most critical pathological hallmarks in Alzheimer's disease (AD) and related tauopathies, including progressive supranuclear palsy (PSP), corticobasal degeneration (CBD), and frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17). This process involves the excessive addition of phosphate groups to the tau protein, fundamentally altering its physiological function and leading to the formation of neurofibrillary tangles (NFTs) that disrupt neuronal integrity and function.
The tau protein, encoded by the MAPT (Microtubule-Associated Protein Tau) gene located on chromosome 17q21.31, plays a fundamental role in maintaining neuronal cytoskeletal stability and axonal transport. Under normal physiological conditions, tau functions as a microtubule-stabilizing protein, binding to tubulin and promoting microtubule assembly and stability essential for axonal transport of organelles, , and signaling molecules [1]. [@crews2009]
In Alzheimer's disease and other tauopathies, tau undergoes extensive post-translational modifications, with hyperphosphorylation being the most prominent and consequential change. This pathological modification reduces tau's affinity for microtubules, promotes its aggregation into paired helical filaments (PHFs), and leads to the formation of NFTs within neurons. The progression of tau pathology follows a characteristic pattern that closely correlates with clinical disease progression, beginning in the entorhinal cortex and spreading throughout the hippocampal formation and cortical regions [2]. [@tsai2004]
The human tau protein exists as six isoforms generated by alternative splicing of the MAPT gene. These isoforms differ by the inclusion of zero, one, or two inserts of 29 amino acids in the microtubule-binding repeat domain (R1-R4) and the presence of three or four repeat domains. The isoforms with three repeats (3R-tau) and four repeats (4R-tau) are expressed in different proportions in the adult human brain, with 4R-tau being more effective at stabilizing microtubules [3]. [@patrick1999]
The tau protein possesses an N-terminal projection domain that projects away from the microtubule surface and interacts with neuronal membranes, and a C-terminal microtubule-binding domain containing the repeat regions. This structural organization allows tau to span multiple tubulin dimers along the microtubule lattice, creating stable cross-bridges that maintain axonal integrity [4]. [@barrett2000]
The six tau isoforms range from 352 to 441 amino acids in length, with molecular weights between 45-65 kDa. The N-terminal region contains two acidic regions that provide a negative charge, while the C-terminal region contains the repeat domains that mediate microtubule binding. The balance between these domains determines tau's functional state and its propensity for pathological aggregation [5]. [@liu2008]
Tau contains approximately 85 potential phosphorylation sites, including 45 serine, 35 threonine, and 5 tyrosine residues. Under normal physiological conditions, approximately 2-3 moles of phosphate per mole of tau are present, distributed across specific regulatory sites. The phosphorylation state of tau dynamically regulates its microtubule-binding affinity, allowing rapid cytoskeletal remodeling in response to cellular signals [6]. [@sontag2014]
Key phosphorylation sites include: [@mandelkow2019]
The coordinated activity of protein kinases and phosphatases maintains tau in a state that optimizes its physiological function while allowing dynamic regulation [7]. [@zhou2018]
The hyperphosphorylation of tau in AD results from the dysregulation of several protein kinases that normally regulate tau phosphorylation. Three kinase families have been identified as the primary contributors to pathological tau phosphorylation: glycogen synthase kinase-3β (GSK3β), cyclin-dependent kinase 5 (CDK5), and calcium/calmodulin-dependent kinase II (CaMKII) [8]. [@barghorn2000]
GSK3β is a serine/threonine kinase that plays a central role in tau hyperphosphorylation. The enzyme can phosphorylate tau at multiple sites, including Ser9, Ser396, and Thr404. GSK3β activity is increased in AD brains through multiple , including increased expression, reduced inhibitory phosphorylation at Ser9, and altered subcellular localization. In AD brains, GSK3β colocalizes with NFT in neurons, suggesting an active role in tau pathology [9]. [@wischik1988]
GSK3β exists in two isoforms (α and β) with distinct subcellular localizations and functions. GSK3β is constitutively active but can be inhibited through phosphorylation at Ser9 by Akt and other kinases. In AD, the balance shifts toward active GSK3β due to: [@cisewski2015]
The activation of GSK3β in AD results from multiple factors, including neuroinflammation, oxidative stress, and impaired glucose metabolism. Amyloid-β oligomers can activate GSK3β through various signaling pathways, creating a potential vicious cycle between amyloid and tau pathologies. Additionally, mitochondrial dysfunction in AD leads to increased GSK3β activity through energy deprivation pathways [11]. [@braak1991]
CDK5, in complex with its regulatory subunit p35 or p39, phosphorylates tau at multiple sites including Ser202, Thr205, and Ser396. CDK5 activity is high in neurons during development but declines in the adult brain under normal conditions. In AD, CDK5 activity increases due to the upregulation of p35 and the calpain-mediated generation of the truncated p25 fragment, which stabilizes CDK5 and redirects it to abnormal substrates including tau [12]. [@braak2011]
The p25/CDK5 complex exhibits altered substrate specificity and longer half-life compared to the physiological p35/CDK5 complex. This dysregulated kinase activity contributes to pathological tau phosphorylation at multiple epitopes that are characteristic of AD tau pathology. Neurotoxic insults that raise intracellular calcium levels can activate calpain, leading to p35 cleavage and p25 generation [13]. [@fitzsimmons2020]
The generation of p25 from p35 represents a critical pathological switch that activates CDK5 toward abnormal substrates. The p25 fragment lacks the membrane-targeting domain of p35, leading to diffuse cytoplasmic localization of CDK5 and enhanced phosphorylation of soluble substrates including tau. This mechanism provides a link between excitotoxicity, calcium dysregulation, and tau pathology [14]. [@fitzsimmons2020a]
CaMKII is abundant in neuronal synapses and can phosphorylate tau at multiple sites upon calcium influx. The activation of NMDA-type glutamate receptors leads to calcium influx and subsequent CaMKII activation, potentially linking excitatory neurotransmission to tau phosphorylation. CaMKII-mediated tau phosphorylation may contribute to activity-dependent modulation of microtubule dynamics [15]. [@mandelkow2003]
The imbalance between kinase and phosphatase activities in AD heavily favors phosphorylation due to the reduced activity of protein phosphatases. Protein phosphatase 2A (PP2A) accounts for approximately 70% of tau phosphatase activity in the brain. PP2A activity is significantly reduced in AD brains due to multiple including [16]: [@bierer1995]
The PP2A holoenzyme consists of a catalytic subunit (PP2Ac), a scaffolding subunit (PP2Aa), and a regulatory subunit. Multiple regulatory isoforms determine substrate specificity and subcellular localization. The downregulation of specific PP2A isoforms in AD contributes to the selective vulnerability of certain neuronal populations [17]. [@mandelkow2012a]
The reduction in PP2A activity creates a permissive environment for tau hyperphosphorylation by shifting the kinase/phosphatase balance toward increased phosphorylation. Restoring PP2A activity has been proposed as a therapeutic strategy to counteract tau hyperphosphorylation, and several compounds that activate PP2A have shown promise in preclinical models [18]. [@stamer2002]
Other kinases implicated in tau hyperphosphorylation include: [@arendt2009]
The combined dysregulation of multiple kinases creates a synergistic effect on tau pathology, with different kinases contributing to phosphorylation at distinct epitopes [19]. [@frost2009]
The transition from hyperphosphorylated tau to filamentous aggregates involves a conformational change in the tau protein. Hyperphosphorylation at multiple sites reduces the net negative charge of tau, diminishing its electrostatic repulsion and allowing inter-molecular interactions. Additionally, phosphorylation induces a change in tau's secondary structure from random coil to β-sheet conformation, facilitating the formation of ordered aggregates [20]. [@gomezisla1997]
The phosphorylated tau assemble into paired helical filaments (PHFs) and straight filaments (SFs) that constitute the core of NFTs. PHFs are approximately 10 nm in diameter with a characteristic 80 nm periodicity. The filament core contains the microtubule-binding repeat domains of tau, which form the stacked β-sheet structure that drives filament formation [21]. [@williams1996]
The aggregation process follows a nucleation-dependent mechanism: [@braak2011a]
Soluble tau oligomers may represent the most toxic species, with filament formation representing a protective mechanism that sequesters toxic intermediates [22]. [@arriagada1992]
NFT formation follows a sequential process: [@selkoe2016]
The progression of NFT pathology follows a predictable pattern in AD, moving from the entorhinal cortex (Braak stages I-II) to the hippocampal formation (Braak stages III-IV) and finally to the neocortex (Braak stages V-VI) [23]. [@avila2018]
The regional distribution of NFTs correlates strongly with cognitive impairment, with NFT burden in the entorhinal cortex and hippocampus being the best neuropathological correlate of memory dysfunction. This anatomical pattern has led to the hypothesis that tau pathology propagates along neuronal circuits in a prion-like manner [24]. [@hooper2008a]
Recent research has revealed that tau filaments from different tauopathies exhibit distinct structural characteristics, suggesting the existence of "strains" that may determine the clinical phenotype. The identification of distinct tau filament folds in AD, PSP, CBD, and chronic traumatic encephalopathy (CTE) has important implications for understanding disease-specific and developing targeted therapies [25]. [@sontag2012]
Cryo-electron microscopy studies have revealed that: [@wischik2015]
These structural differences may explain the distinct clinical presentations of different tauopathies and have implications for biomarker development and therapeutic targeting [26]. [@pedersen2015]
Hyperphosphorylated tau has markedly reduced affinity for microtubules, leading to microtubule destabilization and impaired axonal transport. This loss of function contributes to synaptic dysfunction and neuronal degeneration by disrupting the delivery of essential components to synapses and the removal of metabolic waste products [27]. [@blennow2019]
The microtubule-binding affinity of tau is regulated by phosphorylation in a site-specific manner. Phosphorylation at Ser262 in the repeat domain dramatically reduces microtubule binding, while phosphorylation at other sites has more modest effects. The combination of multiple phosphorylations creates a synergistic reduction in microtubule binding [28]. [@schott2022]
The disruption of microtubule integrity by hyperphosphorylated tau impairs axonal transport, leading to accumulation of organelles and in axons and axonal swellings. This transport deficit contributes to synaptic loss, energy deprivation, and the propagation of pathology to connected neurons through prion-like [29]. [@janelidze2020]
Axonal transport depends on microtubule-based motor kinesin and dynein that move cargo along microtubules. The loss of microtubule stability and direct interference by hyperphosphorylated tau with motor function disrupts this essential cellular process, contributing to the dying-back pattern of neurodegeneration [30]. [@hall2012]
Tau pathology directly contributes to synaptic dysfunction through multiple . Hyperphosphorylated tau can: [@yoshiyama2007]
Synaptic loss, rather than neuronal loss, correlates most strongly with cognitive decline in AD, making tau-induced synaptic dysfunction a critical therapeutic target [31].
The toxic effects of tau extend beyond the neuron in which it is expressed. Tau can be released into the extracellular space and taken up by neighboring neurons, where it can template the aggregation of endogenous tau. This trans-synaptic spread provides a mechanism for the progression of pathology through connected brain regions [32].
Certain neuronal populations exhibit particular vulnerability to tau pathology. The cholinergic neurons of the basal forebrain, layer II entorhinal cortex neurons, and pyramidal neurons of the hippocampus show early and severe involvement in AD. The reasons for this selective vulnerability include [33]:
The selective vulnerability of specific neuronal populations may determine the clinical presentation of different tauopathies. For example, the relative sparing of hippocampal neurons in PSP compared to AD explains the prominent memory impairment in AD versus the early gait and oculomotor deficits in PSP [34].
The progression of tau pathology in AD follows a predictable pattern that was systematically characterized by Heiko and Eva Braak. This staging system remains the gold standard for neuropathological assessment of AD and provides a framework for understanding disease progression [35]:
The Braak stage correlates with cognitive impairment, though significant variability exists between individuals. Some patients with high Braak scores maintain relatively normal cognition, possibly due to cognitive reserve [36].
The relationship between amyloid-β (Aβ) and tau pathologies remains an active area of investigation. Aβ may promote tau pathology through:
The "Amyloid Cascade Hypothesis" originally proposed that Aβ triggers downstream tau pathology, though current models acknowledge more complex interactions between these two hallmark pathologies [37].
Given the central role of kinases in tau hyperphosphorylation, several kinase inhibitors have been investigated as potential therapies:
However, kinase inhibitor development has faced challenges due to toxicity, lack of brain penetration, and limited efficacy in clinical trials. The broad physiological functions of these kinases create on-target toxicity that limits therapeutic windows [38].
Lithium, a mood stabilizer used clinically for decades, inhibits GSK3β and has shown some promise in AD trials, though results have been mixed. More selective GSK3β inhibitors have been developed, including:
The challenge of achieving sufficient brain penetration while avoiding toxicity remains a significant hurdle [39].
Activating protein phosphatases, particularly PP2A, represents an alternative approach to reducing tau hyperphosphorylation. The PP2A activator fingolimod (FTY720) has shown promise in preclinical models of tauopathy. Other PP2A activators under investigation include [40]:
Small molecules designed to prevent tau aggregation include:
Methylene blue reached Phase III trials but showed limited efficacy. Newer compounds with improved pharmacokinetic properties continue to be developed [41].
Active and passive immunization strategies targeting tau have shown promise in preclinical models:
Several anti-tau antibodies have reached clinical trials, including:
The challenge of achieving sufficient antibody penetration into the brain remains a significant limitation [42].
Cerebrospinal fluid (CSF) for tau pathology include:
The combination of decreased Aβ42 and elevated p-tau181 in CSF provides high diagnostic accuracy for AD. P-tau181 appears to be the most specific marker for AD-type tau pathology, distinguishing AD from other neurodegenerative conditions [43].
PET ligands that bind to tau aggregates enable in vivo visualization of tau pathology:
Tau PET provides information about regional tau burden that correlates with cognitive impairment and allows monitoring of disease progression and treatment responses [44].
Emerging blood-based show promise for detecting tau pathology:
These blood tests offer significant advantages for screening and longitudinal monitoring compared to CSF and PET approaches [45].
Transgenic mouse models expressing mutant human tau have provided crucial insights into tau pathology:
These models have been instrumental in testing therapeutic interventions and understanding disease [46]. Key findings from animal models include:
Non-human primate models and iPSC-derived neurons provide additional platforms for studying human-specific aspects of tau pathology [47].