Abstract
Tauopathies are morphologically, biochemically, and clinically heterogeneous neurodegenerative diseases defined by the accumulation of abnormal tau proteins in the brain. There is no effective method to prevent and reverse the tauopathies, but this gloomy picture has been changed by recent research advances. Evidences from genetic studies, experimental animal models, and molecular and cell biology have shed light on the main mechanisms of the diseases. The development of radiology and biochemistry, especially the development of PET imaging, will provide important biomarkers for the clinical diagnosis and treatment. Given the central role of tau in tauopathies, many treatments have constantly emerged, including targeting phosphorylation, targeting aggregation, increasing microtubule stabilization, tau immunization, clearance of tau, anti-inflammatory treatment, and other therapeutics. There is still a long way to go before we obtain drug therapy targeted at multifactor mechanisms.
INTRODUCTION
Tauopathies are believed to be involved in dozens of neurodegenerative diseases, clinically manifesting with a wide spectrum of symptoms, including cognitive, behavioral, or motor impairments [1]. They are histopathologically defined by the filamentous inclusions of unusually phosphorylated tau protein in neurons and neuroglia [2]. Recently, mechanisms and therapeutic strategies of tauopathies become a hot spot in this field. The discoveries of biomarkers of tauopathies open up a new avenue for revealing the real roles of tau in mechanisms of neurodegenerative diseases and provide effective evidences for clinical treatments [3]. In this review, we will integrate and interpret recent advances that have provided new clues for mechanisms and therapeutics of tauopathies. Moreover, we also discuss the promising and novel therapeutic strategies for other neurodegenerative disorders with the lessons learned from the recent knowledge of tauopathies.
TAU ISOFORMS AND GENETICS
Tau is a member of microtubule-associated protein family which is concentrated in axons of central nervous system and peripheral nervous system, and plays a physiological role in dendrites [4]. Located on chromosome 17q21, a single gene containing 16 exons (E) encodes the human tau proteins. Six tau isoforms have been formed by alternative splicing of E2, 3, and 10 (Fig. 1A). The alternative splicing of E2 and 3 produces three tau isoforms-two N-terminal inserts, one insert or no insert, while the alternative splicing of E10 generates two tau isoforms-3R (three repeat domains) or 4R (four repeat domains) [5]. Except promoting the formation of microtubules and keeping their stability, tau also plays a key role in maintaining neuronal architecture and axonal transport and serves as an anchor for enzymes of cell signaling or other proteins. Meanwhile, it regulates cell activity and viability [6].
Tau six isoforms and mutation of MAPT in human brain. A) Six tau isoforms are produced via inserting a 29-aminoacid or 58-aminoacid encoded by E2 (light brown) and E3 (brownish yellow) in the N-terminal half and by the presence or absence of 31-aminoacid repeat domains encoded by E10 (brown) in carboxy-terminal half. The alternative splicing E2 and E3 produces three tau isoforms-2N (two N-terminal inserts), 1N (one insert) or 0N (no insert). The alternative splicing of E10 generates two tau isoforms-3R (three microtubule-binding repeat domains) or 4R (four microtubule-binding repeat domains). E1, 4, 5, 7, 9, 11, 12, and 13 (grey) constitute the basic component of tau protein, whereas E0, 4a, 6, 8, and 14 are not. B) Mutations of MAPT genes, mainly relation with frontotemporal dementia and parkinsonism linked to chromosome 17. 55 sites mutations in MAPT including 43 exon mutations and 15 non-coding-region mutations flanking E10 are revealed. MAPT, microtubule-associated protein tau.
A great majority of genetic studies put focus on the mutation of MAPT and related genes. Mutation of MAPT and related genes promote tau aggregation, reduce its affinity for microtubule, disturb the balance between 3R and 4R and regulate the splicing of mRNA [7], leading to tauopathies. Over 50 mutations on MAPT, which is located on 17q21-22, have been identified to date [4] (Fig. 1B). Mutations in E9-12 increase the mRNA splicing, leading to tau hyperphosphorylation, tangle formation, synapse loss, glial activation, neuronal loss, and memory impairment [7–9]. Missense mutations of MAPT jeopardize physiological functions of tau protein and facilitate the aggregation of microtubule. Nearly 80% of missense mutations are associated with frontotemporal dementia and other neurodegenerations including progressive supranuclear palsy (PSP), Pick’s disease (PID), and corticobasal degeneration (CBD) [10]. Mutations of R406W and A152T are found to be associated with both FTDP-17 and AD-like pathological process [11, 12]. A152T increases the levels of p-tau protein and decreases clearance of p-tau protein. A152T substitution increases the risk of neurodegeneration by increasing p-tau levels and enhancing network hyperexcitability, combined with other pathogenic factors [13]. Interestingly, the same mutation was linked with distinct syndromes in the same family; for example, the son has the clinical manifestation of CBS while the father has frontotemporal dementia (FTD) [4].
Two haplotypes, including H1 with a 900-kb inversion polymorphism and H2 with non-inversion polymorphism, also contribute to the neurodegenerative diseases. H1 causes PSP and Parkinson’s disease as well as CBD through H1c. H1c is located in one of intron 0 regulatory region of MAPT. Recent studies have shown that MAPT H1c haplotype (rs242557) is a genetic risk factor for CBD and PSP [14]. H2 confers protection by increasing E3 expression of MAPT in grey matter [15]. KANSL1 is a gene encoding chromatin modifier which affects the expression of gene via the lysine 16 acetylation of histone H4. Its haploinsufficiency causes 17q21.31 microdeletion syndrome [16]. Additionally, the mutations in 18q deletion syndrome and other 21 genes have been reported to be related to tauopathies. In the other 21 genes, some genes (ATP6AP2, SLC9A6, LRRK2, PRKN, SNCA, and CYP27A1) were detected to occur with tauopathy through some “specific” mutations. Others (including CLN6, CNBP, DMPK, ITM2B, NPC1, NPC2, PANK2, PRNP, PSEN2 and SLC17A5) were detected to form AD-like neurofibrillary tangles (NFTs) in tauopathy (Table 1) [17].
Genetic mutation with tau pathology
CLINICOPATHOLOGICAL FEATURES OF THE TAUOPATHIES
Tauopathies are morphologically, biochemically, and clinically heterogeneous neurodegeneration defined by the accumulation of abnormal tau proteins in the brain. The neuropathological phenotypes distinguish themselves from each other according to different tau morphological features (pretangles, NFTs, threads, Pick bodies, dystrophic neuritis, grains, and spherical cytoplasmic inclusions) and tau isoforms (3R and 4R) [18].
3R and 4R tauopathies
Alzheimer’s disease
Alzheimer’s disease (AD) is the most prevalent neurodegenerative disease associated with age-related dementia. It is characterized by both Aβ extracellular plaques and intracellular neurofibrillary tangles made up of 3R and 4R tau [19, 20]. Pretangles are cytoplasmic tau immunoreactivity in neurons without apparent formation of fibrillary structures. In Alzheimer’s disease, such tau deposition is considered to represent a premature state prior to fibril formation (AD-pretangles), later to form neurofibrillary tangles and finally ghost tangles. This morphological evolution from pretangles to ghost tangles is parallel with their profile shift from four repeat (4R) tau-positive pretangles to three repeat (3R) tau-positive ghost tangles with both positive neurofibrillary tangles in between. This complementary shift of tau profile from 4R to 3R suggests that these tau epitopes are represented interchangeably along tangle evolution. The intracellular neurofibrillary tangles formed by abnormal tau proteins cause deficits through a loss-of-function mechanism in AD [21]. Clinically, AD presents with typical and atypical manifestations including memory and learning impairment, executive dysfunction, aphasia, visual, practical problems, and mental symptoms in the late of AD [22].
Primary age-related tauopathy
Primary age-related tauopathy (PART) replaces these names (“tangle-predominant senile dementia”, “tangle-only dementia”, “preferential development of NFT without senile plaques”, and “senile dementia of the neurofibrillary tangle type”) to describe the neurofibrillary tangle pathology that is observed in the brains of aged individuals [23]. In comparison with AD, PART has an association with MAPT tau H1 haplotype while having no genetic risk associated with apolipoprotein gene (ɛ4) [24]. However, neurofibrillary tangles in PART are made up of 3R and 4R tau as those in AD. Hence, it is difficult to distinguish AD and PART. The relation between PART and AD is still not clear.
Chronic traumatic encephalopathy
Chronic traumatic encephalopathy (CTE) is a progressive neurodegenerative disease caused by repetitive head injuries [25]. The symptoms will be present after a long period of latency. The periods for different individuals vary from several years to several decades. It classically presents with mood, cognitive, and behavioral symptoms, including depression, aggression, impulsivity, suicidal ideation, apathy, short-term memory loss, and executive dysfunction [26]. The concept was first introduced in 1927 [26], and the symptoms of CTE was reported by Harrison Martland in 1928 [27]. Subsequently, hyperphosphorylated tau was found in small cerebral vessels, cerebral sulci, periventricular areas, subpial regions, subcortical structures, and brainstem nuclei and it also had an immune response for both 3R and 4R tau [26]. Besides, the astrocytic tau pathology was observed in periventricular and subpial areas, and astrocytic tangles were formed in CTE [25].
4R tauopathies
Corticobasal degeneration
Corticobasal degeneration (CBD) is related with huge burden of tau inclusions in white matter, basal ganglia and brainstem. It classically presents with motor features (myoclonus, dystonia, gait abnormalities, bradykinesia, limb and axial rigidity) and higher cortical features (cortical sensory loss, aphasia, apraxia, cognitive impairment, alien limb, and behavioral changes) [28]. Some patients showed a frontotemporal pattern of cognitive deficits (such as personality change, a progressive non-fluent aphasia and impaired awareness of deficit) and supranuclear gaze palsy [29]. Besides these clinical symptoms, behavioral changes and psychiatric symptoms without motor symptoms were also found in CBD, but less severe degenerative changes were found in the substantia nigra and subthalamic nucleus of these patients [30]. CBS, CBD-bvFTD, PSP syndrome, and progressive non-fluent aphasia (PNFA) were 4 clinical subtypes of CBD [31]. Patients with CBD and Richardson syndrome have greater atrophy in anterior corpus callosum, more severe neuronal loss in substantia nigra and less neuronal loss in the subthalamic nucleus, when compared with PSP [32].
Progressive supranuclear palsy
PSP refers to Richardson’s syndrome or PSP syndrome [33]. In 1964, Richardson’s syndrome (PSP-RS) was first described by Steele, Richardson, and Olszewski [34]. Early postural instability/falls, supranuclear gaze palsy, variable cognitive impairments, apraxia of speech and pseudobulbar affect were also observed in patients [19]. Abnormal accumulation of 4R-tau proteins was seen in PSP patients, and they were present in different neuroanatomical patterns to produce variable clinical features [33]. PSP variants are well-recognized in recent years and include PSP-parkinsonism (PSP-P), PSP-CBS, pure akinesia with gait freezing (PAGF), PSP-bvFTD, and PNFA [35]. The cortical tau pathology was more significant in PSP-CBS, PSP-bvFTD, and PNFA than in PSP-P and PAGF. PSP-P and PAGF are brainstem-predominant PSP syndromes. Compared with PSP-RS, they showed more severe degeneration and less cortical tau pathology in substantia nigra, globus pallidus, and subthalamic nucleus [36].
Argyrophilic grain disease
Argyrophilic grain disease (AGD) is described in amygdale, entorhinal cortex, hippocampus, and neighboring temporal cortex in patients with adult onset dementia and is characterized by spindle-shaped “grains” [37]. In peri-amygdaloid white matter and hippocampus, AGD presents with oligodendroglial coiled bodies and pretangles [18]. Besides, tau pathology was found in different regions in different stages. The changes in AGD are dependent on tau neurofibrillary tangles, which are biochemically difficult to distinguish from AD. Cognitive decline, episodic memory loss, dementia, personality changes, aphasia, emotional and mood imbalance (irritability, delusions, amnesia, apathy, irritability, dysphoria and agitation), prominent abnormal behavior, and aggression have been noted in the patients of AGD, but no specific clinical syndrome was found to diagnose this disease [38].
3R tauopathies
Pick’s disease
PID is first used to describe atrophy of the frontotemporal lobes in patients with behavioral disturbances and progressive language disorder, now used to describe the disease with tau-positive intraneuronal “Pick body” inclusions [39]. Clinically, it presents with the disorder of social comportment, executive impairments, PPA, visuospatial deficits and akinetic-rigid syndrome with executive. The clinical symptoms of PID were associated with bvFTD, PPA, and CBS [40]. The autopsy found other tauopathies including CBD, PSP, AGD, AD and TDP-43 proteinopathies in bvFTD patients [41]. The poor specificity of bvFTD for PID neuropathology necessitated the recent clinical research criteria for bvFTD in order to more accurately predict PID. The studies of PPA variants showed various results. Semantic variant PPA (svPPA) and PNFA were present in different tauopathies. The former is due to TDP-43 proteinopathies and the latter is due to PSP and CBD. The “Pick body” inclusions were mainly made up of 3R tau and distributed in hippocampus, dentate gyrus, and cortical areas [42]. The pathogenesis of mutations in tau proteins has been mentioned in FTDP-17. Recent studies have developed a novel 3R mutant tau transgenic model for better understanding the natural history, progression and therapy of PiD [43]. Moreover, a novel missense mutation in E12 of MAPT (p.Q336H) has been suggested to cause PiD, possibly by increasing 3R tau isoforms aggregation [44].
BIOMARKERS OF TAUOPATHIES
A biomarker is a parameter which reflects the progress of disease and the effects of treatment. Recently, neuroimaging biomarkers of tauopathies will provide strong underpinnings for prevention, diagnosis, and therapeutic targeting.
Biomarkers of tau in neuroimaging
Positron emission tomography (PET) may be one of the noninvasive imaging technologies to measure tauopathies and associated neurodegeneration in the brain. PET imaging of tau pathology relies on two major radioactive tracers: 18F-labeled (t1/2 =109.8 min) tracers and 11C-labeled (t1/2 = 20.3 min) tracers [45]. With the development of the PET tracers, many tracers were reported to be applied to the detection of tau pathology including 18F-FDDNP, 18F-THK523, 18F-THK5117, 18F-THK5105, 18F-THK5351, 18F-AV-1451(T807), 18F-T808, 11C-PBB3, 11C-THK951, lansoprazole, RO6931643, RO6924963, and RO6958948 [3, 46–52]. The characteristics of these tracers were described in Table 2. A phase 0 trial of 18F-FDDNP to evaluate tau values in Parkinson’s disease dementia has been completed, but no results have been posted (https://clinicaltrials.gov/ct2/show/NCT02243982). 13 clinical trials for 18F-AV-1451 have been recruiting patients. Two trials of 18F-AV-1451 have been completed, but no results are reported (https://clinicaltrials.gov/ct2/show/NCT02079766 and https://clinicaltrials.gov/ct2/show/NCT01992380). In the living human brain, 18F-AV1451 makes it possible for the regional distribution of tau pathology visually [53]. A phase II trial for18F-AV1451 to measure the distribution of tau proteins in living human brain is ongoing (https://clinicaltrials.gov/ct2/show/NCT02191267). At the same time, 18F-T807 is also in phase II process as a diagnostic PET imaging tracer for AD and neurodegeneration characterized by tau pathology [54]. Besides, 18FMNI-798 was used to evaluate tau protein burden in AD in a phase I and II trial (https://clinicaltrials.gov/ct2/show/NCT02640092]). RO6931643, RO6924963, and RO6958948 as the new tracers for PET are being tested in humans, and a phase I trial has been completed to assess the ability to detect tauopathy (https://www.clinicaltrials.gov/ct2/show/NCT02187627).
Tracers of PET imaging for tau
18F-FDDNP: 2-(1-(6-[(2-[18F] fluoroethyl) (methyl) amino]-2- naphthyl) ethylidene) malononitrile; 18F-THK523:2-(4-aminophenyl)-6-(2-[18F] fluoroethoxy) quinoline; 18F-THK5105:6-[(3-[18F] fluoro-2-hydroxy) propoxy]-2-(4-dimethylaminophenyl) quinolone; 18F-THK5351:6-[(3-18F -2-hydroxy) propoxy]-2-(4-methylaminopyridyl) quinoline; 18F-T807: (E)-4-(2-(6-(2-(2-(2-[18F] fluoroethoxy) ethoxy) ethoxy) pyridin-3-yl) vinyl)-N-methyl benzenamine; 18F-T808:2-(4-(2-[18F] fluoroethyl) piperidin-1-yl) benzo (4, 5) imidazo (1, 2-a) pyrimidine; 11C-PBB3: a tau PET tracer constructed around a phenyl/pyridinyl-butadienyl-benzothiazole/ benzothiazolium (PBB) scaffold; NFTs, neurofibrillary tangles; PHF, paired helical filaments; AD, Alzheimer’s disease; CBD, Corticobasal degeneration; PSP, Progressive supranuclear palsy.
Biomarkers of tau in cerebrospinal fluid
CSF tau species have been considered as the potential biomarkers for tauopathies, which are useful to diagnose and identify different forms of neurodegeneration. Several researchers found that CSF t-tau and p-tau have been considered as biomarkers for AD, and the concentration of t-tau in AD has a two- to three-fold increase than in controls, while the p-tau in CSF shows a specificity of 80–100% [55]. Tau phosphorylation sites found in CSF up to now are Ser199, Thr231, Thr 181, Ser 396/404, and Ser235 [55]. As for p-tau181, it has low specificity and sensitivity to distinguish AD and non-AD cases [56], and this ability increases after binding to Aβ1 - 42 and t-tau [57]. The diagnostic power of Aβ1 - 42/p-tau181 is higher than Aβ1 - 42 and t-tau in the discrimination between AD and FTD. When focusing on the discrimination between AD and CJD, p-tau181P/T-tau shows significantly higher diagnostic power than Aβ1 - 42, t-tau, and Aβ1 - 42/p-tau181P.
Besides AD, a reduced ratio of CSF p/t-tau has been proved to have 82.4% positive predictive value to detect TDP pathology in FTLD [58]. The levels of the 20–22 kDa NH2-truncated form of human tau in CSF could be rarely discovered in non-demented cases, which could distinguish between cognitive impairment patients and cognitively healthy subjects influenced by heterogeneous neurological disorders with a higher sensitivity of 85% and a low specificity of 65% [59]. It is not a biomarker for the diagnosis of cognitive impairment.
Biomarkers of tau in peripheral blood
Platelets are not only the main reserve of the amyloid precursor protein, but include several tau molecules. High molecular weight tau proteins were detected in platelets of peripheral blood in AD [60]. Tau fragments, tau-A and tau-C (a caspase-3 cleaved tau fragment) are detected in serum, which serve as the novel biomarker in serum for the neurodegeneration [61–63]. Recent studies have shown that plasma tau is not an AD biomarker in individual people, and it just partly reflects AD pathology [64]. The levels of tau proteins were significantly increased in serumas indicators for severe TBI and Olympic boxing [65]. Minor axonal damage induced by punches causes the increase of plasma tau. The plasma tau will significantly increase after a bout in 1–6 days [65]. Circulating miRNAs also provide the evidence for AD. The serum levels of cell-free miR-125b in serum decrease inAD [66].
MECHANISMS OF TAUOPATHIES
Tau hyperphosphorylation
Phosphorylation plays the pivotal role in physiological and pathological function of tau. In the longest tau isoform, there are 85 potential sites of phosphorylation–45 serines, 35 threonines and 5 tyrosines [67]. Among them almost 40 sites become abnormal phosphorylated in tauopathies [68]. Phosphorylation sites in tau are in the C-terminal tail region and proline-rich region [69]. Many factors such as accelerated aging, traumatic brain injury (TBI), LRRK2, hypothermia, β-N-methylamino-L-alanine (BMAA), ischemia, oxidative stress, methanol, and Cytoplasmic FMR1 interacting protein 2 (CYFIP2) induce hyperphosphorylation of tau [70–75]. Among them, accelerated aging contributes to an acceleration of tau hyperphosphorylation, leading to the development of behavioral changes which influence tau pathogenesis [75]. Recent studies indicated that tau phosphorylation is a risk factor associated with the extent of cognitive impairment due to TBI. Meanwhile, elevation of GSK-3β kinase activity and reduction in PP2A phosphatase activity have been suggested as, respectively, positive and negative factors associated with the severity of TBI. Altogether, these results suggest a role of tau phosphorylation pathway for the alterations of injured neurons in mediating cognitive damages of TBI [76]. Tau phosphorylation is exquisitely sensitive to temperature, increasing by 80% for each degree below 37°C, due to exponential decrease in PP2A activity during direct hypothermia, or anesthesia-induced hypothermia [77]. BMAA leads to tau hyperphosphorylation by activating metabotropic glutamate receptor 5 (mGluR5). Then, mGluR5 results in the release of PP2Ac and its phosphorylation at Tyr307 to inhibit PP2A [71]. Transient brain ischemia has been shown to induce tau hyperphosphorylation. Recent studies confirmed that tau protein was dephosphorylated during brain ischemia. In addition, the activity of GSK-3β increased and the activity of PP2A decreased. After reperfusion, tau protein was hyperphosphorylated; the activity of GSK-3β decreased; the activity of PP2A remained low. Importantly, the interaction of tau with GSK-3β and PP2A was altered during ischemia and reperfusion [78, 79]. Methanol not only leads to acute neurological sequelae, including blindness and death, but low concentrations of methanol also lead to hyperphosphorylation of tau and apoptosis of hippocampus neurons [74]. The lack of CYFIP2 increases the expression of alpha-calcium/calmodulin-dependent kinase II in order to exacerbate tau hyperphosphorylation [80].
MicroRNAs (miRNAs) participate in the progress of tau phosphorylation, such as miR-922, miR-125b, and miR-132. MiR-922 induces tau phosphorylation through reducing the expression of ubiquitin carboxy-terminal hydrolase L1 (UCHL1), and affects tau phosphorylation through influencing the numbers of NFTS [81]. MiR-125b promotes tau hyperphosphorylation by its overexpression. When miR-125b was injected into mice, cdk5, p35, and p44/42-MAPK signaling were upregulated and the anti-apoptotic factor Bcl-W and the phosphatases DUSP6 and PPP1CA were suppressed, leading to tau phosphorylation [82]. MiR-132 loss promotes tau pathology through regulating the expression of inositol 1,4,5-trisphosphate 3-kinase B (ITPKB), which regulates tau phosphorylation [83]. In the Drosophila tauopathy model, proto-oncogenes suppresses tau-mediated neurodegenerative diseases by reducing abnormal tau hyperphosphorylation [84].
Tau aggregation
Tau is a highly soluble unfolded protein and not aggregated with each other in physiological conditions and concentrations. Tau aggregation is not only the result of the phosphorylation, but also can be affected by other factors. Negatively charged molecules are the inducers of non-phosphorylated tau aggregation [4], including heparin, anionic micelles, RNA, synthetic particles, fatty acids, fatty acid-like molecules (such as arachidonic acid), and lipid vesicles. Heparin induces the polymerization of tau fragments by binding to the second and third microtubule-binding repeats [85], while arachidonic acid is more effective than heparin in promoting aggregation of full-length tau when the number of K18, a tau fragment identified as one of the four repeats of MBD, which is united to the lipid bilayer, oversteps a critical surface density[86, 87].
Acetylation of tau prevents degradation of phosphorylated tau, leading to tau aggregation [88]. In an animal model of tauopathy, the insoluble tau fraction was observed after the acetylation, suggesting that acetylation promoted soluble tau to insoluble tau in order to affect the process of aggregation [89]. CREB binding protein (CBP) and p300 acetylate different residues of tau, changing tau’s intrinsic propensity to aggregate [90]. In addition to acetyltransferase, acetylation also regulates the process of tau aggregation through other ways. At KXGS motifs, acetylation and phosphorylation act in a competitive relationship and are regulated by HDAC6 [90]. The inhibitors of HDAC6 facilitate tau acetylation and prevent tau phosphorylation at these motifs to interfere with tau aggregation propensity. In addition, HDAC6 could also impact on turnover of tau through regulating the acetylation of its substrate Hsp90, suggesting combination treatment of HDAC6 and Hsp90 inhibitors as a potent therapeutic treatment for diseases with accumulation of tau [91]. Recently, ac-K174 has a critical role in tau accumulation and toxicity. Besides, the acetyl-mimic mutant K174Q inhibits tau degradation and promotes tau aggregation [92]. Recent studies suggest that stress granules (SG) are a form of pathology associated with neurodegenerative diseases. SG proteins, such as TIA-1 and TTP, bind to phosphorylated tau, and that binding increases with the severity of the disease. Meanwhile, SG proteins might contribute to generating tau pathology, perhaps by creating local environments with high concentrations of phospho-tau [93]. Latest studies suggest that the pathological tau aggregation may be initiated through the regulated self-assembly of core SG-associated RNA binding proteins, such as TIA-1 [94].
Additionally, truncation at both ends (N-terminal or C-terminal) of the tau molecule could confer a toxic gain of toxic function. The in vivo model of tauopathy based on the truncated tau protein clearly shows that truncation is a “productive” modification that can initiate tau neurofibrillary degeneration [95]. The truncated forms of tau protein are especially found in PHFs, suggesting that tau truncation might contribute to tau aggregation [96]. Recent studies suggest that there is a complex interaction between phosphorylated and truncated tau species [97]. Moreover, tau truncation could modulate tau spreading. Finally, changes in tau 3R/4R ratio may result in differences in microtubule stability.
Neurotoxicity of tau oligomers
As mentioned above, hyperphosphorylation of tau consists of neurofibrillary tangles (NFT). Before the formation of NFT, tau forms two intermediate forms of NFT: tau oligomers, not discovered by atomic force microscopy (AFM), and granular tau oligomers, detected by AFM [1]. Tau oligomers play a crucial role in neurodegeneration, synaptic dysfunction and mitochondrial loss. Tau oligomers impair memory by disrupting anterograde memory storage when researchers inject the monomers of full-length recombinant p-tau-441 or fibrils into the C57BL/6 wild mice hippocampus [98]. Recent animal studies found that mice expressing TauC3 (a caspase-cleaved form of tau) showed memory and learning impairment and the decrease of synaptic number at early stage accompanied with tau oligomer formation [99].
Tau oligomers cause synaptic dysfunction through affecting the density of presynaptic and neuronal trafficking and also cause mitochondrial dysfunction by activating caspase-9 [98]. Intracellular p-tau accumulation causes mitophagy deficits, which inserts into the mitochondrial membrane to activate the membrane potential and to cause the decrease of PTEN-induced kinase 1/Parkin. PTEN-induced kinase 1/Parkin reduces p-tau-induced impairment [100]. Except these, miR-219 targeting tau synthesis directly causes the toxicity of tau [101]. Recently, a new mechanism of tau toxicity is found to contribute to glial activity impairment in AD and other tauopathies, which induces microtubule bundling and activate the Rho-GTPase-ROCK pathway [102].
Mechanism of autophagy-lysosome system in tauopathies
Autophagy is a catabolic system which can degrade abnormal proteins, lipids and organelles through lysosomes. It plays an important role in cell homeostasis and keeps the metabolic balance between synthesis and degradation and subsequent turnover of cytoplasmic materials in a stressful environment. Autophagy is classified into three types: microautophagy, chaperone-mediated autophagy (CMA) and macroautophagy [103, 104]. Many factors inhibit autophagy which leads to tau accumulation in neurodegenerative diseases, including mTOR, p53 tumor suppressor, cAMP, NH4Cl, cathepsin inhibitors,3-methyladenine, chloroquine, Bcl-2, Bcl-XL, and increased concentration of Ca2 + in cytoplasm which comes from the endoplasmic reticulum [105–108]. The mTOR is a primordial inhibitory signal which participates in the initial process of signal transduction. When growth factors such as insulin-like growth factor bind to insulin-like growth factor receptors (IGF1R), class I PT3K pathway, which catalyzes the conversion of PIP2 to PIP3, is activated. PIP3 recruits AKt and DDK1 to membrane, allowing the latter one to phosphorylate AKt. The active AKt inhibits the activity of the TSC1/TSC2 complex, whose activity can lead to an overall inhibition of mTORC1 signaling. By contrast, intracellular signals include nutrient starvation, low energy levels (increase of the AMP/ATP ratio), hypoxia, and DNA damage inhibit mTOR activity [109]. CMA (chaperone-mediated autophagy) may generate tau fragment aggregation [110]. There are two CMA-targeting motifs in MBDs of tau, 336QVEVK340 and 347KDRVQ351, which participate in the transportation of tau to lysosomal membranes. The overexpression of neuronal PAS domain protein 4 promotes the clearance of pathological tau by inducing autophagy [111] and provides a novel therapeutic approach to tauopathies.
Mechanism of ubiquitin-proteasome system in tauopathies
Like autophagy-lysosome system, the ubiquitin-proteasome system (UPS) is a primary mechanism of organism that mediates the degradation of abnormal or misfolded proteins and short-live proteins. The system contains two parts–ubiquitination and proteasome, which interact with each other to mediate the degradation of abnormal tau [112]. In the brain in AD, polyubiquitin chain of PHF-tau is found, which is attached to others through four lysine residues (Lys63, Lys48, Lys11, and Lys6) [106]. The affinity between tau and microtubule is reduced when tau is modified with ubiquitination. Meanwhile, researchers found that E3 ubiquitin ligase activity played a role in the ubiquitination of tau [113]. E3 ubiquitin ligase determines the specificity of interaction between ubiquitin and substrates.
Some inhibitors of proteasome, lactacystin, epoxomicin, MG132, and leupeptin put the degradation of abnormal tau off. TNF-α/NFκB reduces 26S proteasome and then decreases the activity of proteasome by S5b/PSMD5 pathway, which plays a vital role in UPS-associated tauopathies [114]. Not only in AD, but in transient cerebral ischemia, proteasome dysfunction causes E10 exclusion of tau and a decrease of Tra2β, leading to the imbalance of 4R/3R and neurodegeneration [115]. UBB + 1 interacted with polyubiquitin chains to inhibit the degradation of polyubiquitinated substrate through 26S proteasome, a dose-dependent inhibitor of proteasome [116]. Maintained stability and activity of UPS is a potential therapy for neurodegenerative diseases and tauopathies.
Propagation of tauopathies
Preclinical phase of the disease, mild cognitive impairment and severe dementia are three stages of clinical symptoms of AD. Each stage has two sequential neuropathological stages according to Braak stages (stage I-II, stages III–IV, stages V–VI) and approximate three stages described by Delacourte (Delacourte stages 1–3, Delacourte stages 4–6 and Delacourte stages 7–10) [3, 117]. The six or ten sequential neuropathological stages reveal the propagation of tau pathology, but the precise process of tau propagation is still a mystery.
Release of tau
The appearance of extracellular tau tangles is identified at stage III, and how tau is released is crucial for us to understand the process of extracellular tau tangle formation. Extracellular tau is not only from the dying neurons but also from other unconventional pathways [118]. Calcium affects the release of endogenous tau by an unconventional pathway [119]. However, the concentration of calcium also influences the process of endogenous tau secretion. The release of tau is increased when the agonist S-AMPA increases the activity of glutamate receptors [120]. The pre-synaptic compartment consists of abundant tau, released tau, and tau fragments via the pattern of potassium-induced depolarization [121, 122].
Endocytosis spreading pathway
Extracellular tau not only can aggregate tangles in interstitial fluid and CSF, but also can be internalized via endocytosis. Uptake of tau fibrils occurs through adsorptive endocytosis, which is time- and temperature-dependent [123]. Another endocytosis may occur when extracellular tau interacts with receptors, which is activated by intracellular calcium via M1/M3 muscarinic receptor stimulation [124]. Meanwhile, the interaction between tau and muscarinic receptors generates toxic influences and leads to the release of tau into the extracellular space and neuron death [125]. Other mechanisms have been identified, including the entry of non-fibrillar and soluble tau aggregates via dynamin-driven endocytosis [126] and proteoglycan-mediated macropinocytosis. Except abnormal tau, PBS-soluble phosphorylated high-molecular-weight tau takes part in tau propagation and is present in the form of endogenous tau [127].
Trans-synaptic spreading pathway
Synapses play a crucial role in the spreading of tauopathy. Studies of human postmortem tissue roughly described the anatomical distribution of the tauopathies, and demonstrated that these pathology areas were linked by synapses [117]. Furthermore, in vivo studies revealed that tau pathology was propagated to other areas associated with synapses from the initial location [128]. How the precise process of propagation happened is still uncertain, except trans-synaptic and endocytosis spreading. Tau may be transferred through tunneling nanotubes between cells and membrane penetration [129]. Besides, new research in animal models showed that microglia and exosomes promoted tau propagation in the mammalian brain and revealed that microglia could transfer tau to other neurons [130]. Recently, an animal model showed that amyloid promoted tau propagation and increased the toxicity of tau [131].
POTENTIAL HYPOTHESIS MECHANISM
Although above-mentioned mechanisms can cause tauopathies in their own ways, many evidences from animal’s experiments, preclinical models and postmortem research show that the mechanisms interact with each other to cause tauopathy (Fig. 2). Abnormal tau proteins are mainly degraded through two systems–UPS and autophagy-lysosome system. Tau proteins induce neuron death and neurodegeneration by a series of processes including hyperphosphorylation, aggregation, oligomers, and then the formation of the NFTs. The NFTs are made up of paired helical filaments (PHFs) and straight filaments, which are composed of abnormal tau proteins. Subsequently, abnormal tau in extracellular space propagates via endocytosis and trans-synaptic pathway and causes the damage of neighboring neurons, leading to the tauopathies.
Mechanisms of tauopathies. In the brain, the mutations of genetic (➀) and the disorder of splicing (➁) induce the production of abnormal tau proteins. Abnormal tau proteins detach from microtubule, leading to the instability of microtubule and the damage of axon transport. Meanwhile, the imbalance of protein kinases and phosphatases and other factors can induce tau hyperphosphorylation (➂). Abnormal tau proteins and hyperphosphorylated tau aggregate with each other, non-phosphorylated tau also aggregate after interacting with negatively charged molecules (➃). The primary form of aggregation is tau oligomers and cause mitochondrial (➄), then tau oligomers polymerize to form NFTs (➅) and cause tauopathies. The UPS (➆) and autophagy-lysosome system (➇), mediated abnormal tau degradation, are blocked by the inhibitors, causing the increase of abnormal tau levels. Intracellular tau secretes to extracellular space via targeting to special receptors (a), exosomes (b), and with the help of microglia (c). These abnormal tau pass to next neuron through endocytosis (f), trans-synaptic pathway (d), or exosomes (e), resulting in new pathology. MTs, microtubules; NFTs, neurofibrillary tangles; DUBs, deubiquitinating enzymes; 3MA, 3-methyladenine; CMA, chaperone-mediated autophagy; UPS, ubiquitin proteasome system.
EXPERIMENTAL ANIMAL MODELS OF TAUOPATHIES
Experimental animal models of tauopathies serve as efficient systems for interpreting the role of abnormal tau in the neurodegeneration and the experiments of treatment. Some crucial aspects of human tauopathies have been recapitulated in modified animals including Caenorhabditis elegans, Drosophila, and especially mouse models (Table 3) (see review in [132]). Caenorhabditis elegans and Drosophila are two easy tools for genetic studies. The Caenorhabditis elegans models have revealed tau toxicity, axonopathy, motor dysfunction, and early death [133]. This model is also used to study drug screening [134]. The Drosophila models are used to identify several pathological pathways or susceptibility genes. Besides, these models profit from low redundancy of the genome and low consumption [135].
The major transgenic mouse models of tauopathies
ND, not determined; Ref, reference; MoPrP, mouse prion protein promoter; CaMK-II, Ca 2 + /calmodulin kinase II; NFTs, neurofibrillary tangles; PHF1, paired helical filaments 1; 2′3′-CNP, 2′,3′-cyclic nucleotide 3′-phosphodiesterase; GFAP, glial fibrillary acidic protein; PDGFβ, platelet-derived growth factorβ.
The first transgenic mouse models of human tauopathies were produced by only expressing specific human tau isoforms (4R/2N) [136]. These models exhibited hyperphosphorylated tau and pre-tangle tauopathy, and NFTs were not observed in these models. To observe the NFTs, novel models were generated [132]. Meanwhile, the relationship between neuronal loss and neurofibrillary pathology was also found. Not only did the transgenic (tg) mice models of P301L demonstrate the process of tauopathies in neuronal and glial, but G272V, N279K, P301S, V337M, and R406W were created to explore the function of tau in tauopathies [132]. These models exhibited different aspects and initiation time of tauopathy. Subsequently, the double mutant (K257T and P301L, G272V and P301S) and triple mutant (G272V, P301L, and R406W) have been reported to test pathogenic hypotheses of tauopathy and therapeutic strategies [137, 138]. In the future work, more available models for human tauopathies should be produced.
TARGETING TAU THERAPEUTICS IN TAUOPATHIES
Most tau-targeted therapies (Fig. 3) are growing at a rapid pace due to the discovery of the vital role of tau in neurodegeneration. Some drugs used in the treatment of tauopathies have been put into clinical phase II or III trials (Table 4), and at the same time new drugs continue to be discovered.
Targeting tau-based therapeutics for tauopathies. The therapeutics of tauopathies consist of targeting phosphorylation, targeting aggregation, increasing microtubule stabilization, tau immunization, clearance of tau, anti-inflammatory treatment, and other therapeutics. MT, microtubule; PelA, peloruside A; NAP, daventide; NSAID, non-steroidal anti-inflammatory drugs.
Drugs of tauopathies for clinical trials
LMTX, leuo-methylthioninium; AD, Alzheimer’s disease; CBD, Corticobasal degeneration; HD, Huntington’s disease; PDD, Parkinson’s disease dementia; PSP, Progressive supranuclear palsy; FTD, Frontotemporal dementia; FTDP-17, Frontotemporal dementia and parkinsonism linked to chromosome 17; NA, not available; bvFTD, Behavioral variant frontotemporal dementia. *It includes preclinical phase.
Targeting phosphorylation in tauopathies
As a hallmark in AD and other neurodegenerative diseases, abnormal tau phosphorylation has the indelible effects on the process of pathology and diseases. The protein kinases provide a vital target site for tau phosphorylation, especially glycogen synthase kinase 3 (GSK-3). A number of GSK-3 kinase inhibitors, such as lithium, valproate, escitalopram, and tideglusib, decrease tau phosphorylation and alleviate the process of diseases [139–142]. Currently, some of these drugs have been put into the clinical trials to assess the pharmacokinetics and discharge excretions. The information is provided in Table 4. No clinical benefits have been found in a phase II clinical trials for valproate in AD [143]. Escitalopram reduces the levels of tau hyperphosphorylation, which are induced by Aβ1 - 42 through the pathway of Akt/GSK-3β [144]. Other tau protein kinases are inhibited for prevention of tau phosphorylation, but no tau protein kinases inhibitors except GSK-3 have been applied to clinical trials until now. To develop other kinases inhibitors and to apply those to clinical trials provide a new direction for the treatment of tauopathy, but the safety and efficacy of drugs should be taken seriously.
The activation of PP2A also supplies a direction for reducing tau phosphorylation. Sodium selenate, phosphotyrosyl phosphatase activator, memantine and metformin reduce tau phosphorylation through elevating the activity of PP2A [145–147]. Memantine has been applied in clinic to attenuate the symptoms of AD. Currently, folic acid reduces tau phosphorylation via inhibiting PP2A demethylation reactions [148]. The methylation of DNA and protein plays a key role in regulating the activity of phosphatase and the inhibition of methylation will be of tremendous value in the treatment of tauopathies.
Targeting aggregation in tauopathies
To inhibit or reverse tau aggregation becomes one of the great therapeutic strategies for neurodegeneration. Some small molecule inhibitors are demonstrated to inhibit and disassemble tau aggregation, such as rhodanines, phenylthiazol-hydrazides, anthraquinones, benzothiazoles, phenothiazines, and so on [139, 149]. Methylene blue is testified to decrease the levels of tau in ex vivo slice cultures from transgenic mice of tauopathy [150]. The positive results have been found in a phase II trial of methylene blue [151]. However, the National Toxicology Program used rodents to show that long-term usage of methylene blue resulted in some malignancies and lymphomas within the intestines, so leuco-methylthioninium (LMTX) is developed in AD and FTD which are more bioavailable and less toxic than MTC [152]. At present, LMTX is being applied in several phase III studies in AD and bvFTD, the clinical effects will be achieved in early 2016 [151]. Recent animal studies found that anle138b reduced tau aggregation, ameliorated disease symptoms, improved cognition, and increased survival time of tau transgenic PS19 mice [153].
Natural products have also shown the function of anti-aggregation, such as oleuropein aglycone, hydroxytyrosol, oleuropein, and grape seed polyphenolic [154, 155]. Recently, O-GlcNAc, one modification of tau, is demonstrated to reduce the aggregation propensity of tau [156]. The above-mentioned inhibitors can prevent tau aggregation or disassemble tau aggregation, but we still need much work to do before these strategies are applied in clinic.
Increasing microtubule stabilization in tauopathies
Hyperphosphorylated and polymeric forms of tau result in a reduction of microtubule stability, and this reveals that keeping the stability of microtubule is one therapeutic strategy for tauopathies. Paclitaxel, peloruside A, Epothilone D, NAP (davuentide), BMS-241027 and TPI-287 improve axonal function, MT density, and cognitive performance [157–159]. Paclitaxel is not appropriate for human tauopathies treatment because of the poor blood–brain barrier permeability, while Epothilone D permeates the blood–brain barrier. The compound makes some improvements in a phase II clinical trial and has no serious effects on patients with mild to moderate AD [160].
Immunization strategies for tauopathies
Active immunization
Active immunization of tau pathology decreases the levels of aggregated tau through the functions of tau antigen. The phosphoepitopes of tau are demonstrated to reduce the levels of accumulated tau and prevent the process of tau related pathology via inducing the active immunization, such as a PHF1 phosphoepitope in P301L transgenic mouse, phosphorylated epitope S422(PS422), pS396/pS404, pS231, pS212/pS214, pS202/pT205, and AADvac1 [161–164]. A phase I trial is underway for pS396/pS404 by AC Immune and Janssen and other information has been reported about this trial [165]. Tau antigen not only induces active immunization, but carries a risk of causing autoimmunity and related tauopathy.
Passive immunization
Not only active immunization, but passive immunization has an effect on the treatment of tau pathology. These antibodies consist of anti-tau antibodies, AT8, three monoclonal antibodies of tau, MC31, DA31, HJ9.3, HJ9.4, PHF1, PHF13, PHF6, anti-pSer413 antibody, MAb86 and C2N-8E12 [166–174]. Except these, a novel antibody-DC8E8 is found to distinguish mis-disordered tau with physiological tau, and it inhibits the interaction between pathological tau, which reduces tau oligomers and neurofibrillary pathologies in animal models [175]. Recently, a new therapeutic strategy for passive immunization is found to prevent phospho-tau pathology by using encapsulated cell implants to transmit recombinant anti-amyloid-β antibodies to protect against the misfolded proteins [176]. Studies on tau antibodies still need to explore therapeutic development.
Clearance of tau in tauopathies
Targeting autophagy-lysosome system
Autophagy-lysosomal system degrades the levels of abnormal tau. Therefore, enhancing the activity of autophagy and inhibiting the degradation of lysosome will be efficient to promote the degradation of tau and alleviate neurodegeneration. Rapamycin and temsirolimus elevate the activity of autophagy and reduce abnormal tau in the mouse model of tauopathy via targeting the mTOR pathway [177, 178]. Besides the mTOR pathway, other drugs also induce autophagy via targeting the process of autophagy-lysosome formation, such as trehalose, methylene blue, lithium, sodium valproic, carbamazepine, and sulforaphane [150, 179–183]. The development of autophagy-lysosome activators that can elevate clearance of pathological tau without negative effects will be of immense therapeutic value for tauopathies and other proteinopathies.
Targeting ubiquitin-proteasome system
It becomes difficult to clear aggregated proteins after the dysfunction of UPS. The enhancement of ubiquitin-proteasome activity promotes the process of tau degradation. Rolipram promotes proteasome activity. IU1, a deubiquitinating enzyme related to proteasome, puts proteasomal degradation off [184, 185]. CHIP, carboxy-terminus of heat shock protein 70-interacting protein, interacts with phosphorylated tau firstly. Then it transfers phosphorylated tau to proteasome when heat shock protein 90 captures the abnormal tau [186]. Besides, cAMP promotes degradation of misfolded proteins by increasing the activity of 26S proteasome through PKA and Rpn6 [187]. These two pathways interact with 19s subunit to affect the activity of proteasome. Maybe, increasing the levels of cAMP will be a useful way of treating tauopathies.
Anti-inflammatory treatments in tauopathies
Neuroinflammation plays a vital role in the process of AD and other related tauopathies. Microglial activation promotes the expression of pro-inflammatory cytokines and leads to tau hyperphosphorylation and neuronal death [188]. In tauP301L mice deficient in CX3CR1 receptor, microglia are over-expressed and tau is hyperphosphorylated [189], thus leading to the aggregation of tau and related tauopathies through two pathways. One is related to the activity of p38 MAP kinase [190], another is related to the increased NRF2-ARE (nuclear factor (erythroid-derived 2)-like 2-antioxidant response element) activity to inhibit the overactivation of microglia [189]. So, targeting CX3CL1-CX3CR1 and/or neuronal IL-1 receptor-p38 and NRF2-ARE signaling pathway is a valid therapeutic strategy for tauopathies. Recent studies find that SCM-198 (Leonurine) inhibits the overactivation of microglia via NF-κB and c-Jun N-terminal kinase pathways [191]. Besides, immunosuppressant FK506 reduces tau aggregation and the expression of microglia. The immunophilin FKBP52 (a member of FK506-binding protein family) directly interacts with hyperphosphorylated tau to inhibit the aggregation of tau [192].
CONCLUSION AND OUTLOOKS
Tauopathies are morphologically, biochemically, and clinically heterogeneous neurodegenerative diseases defined by the accumulation of abnormal tau proteins in the brain. Although the development of various studies, medicine, and technology has helped interpret many vital aspects of the tauopathies, more effective treatments for tauopathies still need to be explored. Efforts to discover imaging markers with the characteristics of lower toxicity, higher kinetics, and greater permeability across the blood brain barrier should be continued and combined with more efficient approaches to be applied in clinical trials. Given the central role of tau in tauopathies, many treatments have constantly emerged, including targeting phosphorylation, targeting aggregation, increasing microtubule stabilization, tau immunization, clearance of tau, anti-inflammatory treatment, and other therapeutics. It still has a long way to go before we obtain a drug therapy targeting multifactor mechanisms.
Footnotes
ACKNOWLEDGMENTS
This work was supported by grants from the National Key R&D Program of China (2016YFC1305803), the National Natural Science Foundation of China (81471309, 81371406, 81571245, and 81501103), the Shandong Provincial Outstanding Medical Academic Professional Program, Taishan Scholars Program of Shandong Province (ts201511109, tsqn20161078, and tsqn20161079), Qingdao Key Health Discipline Development Fund, Qingdao Outstanding Health Professional Development Fund, and Shandong Provincial Collaborative Innovation Center for Neurodegenerative Disorders, Research Award Fund for Outstanding Young and Middle-aged Scientists of Shandong Province (BS2015SW006), Projects of medical and health technology development program in Shandong province (2015WSA02054).
