Tau and MAPT genetics in tauopathies and synucleinopathies

Abstract

MAPT encodes the microtubule-associated protein tau, which is the main component of neurofibrillary tangles (NFTs) and found in other protein aggregates. These aggregates are among the pathological hallmarks of primary tauopathies such as frontotemporal dementia (FTD). Abnormal tau can also be observed in secondary tauopathies such as Alzheimer’s disease (AD) and synucleinopathies such as Parkinson’s disease (PD). On top of pathological findings, genetic data also links MAPT to these disorders. MAPT variations are a cause or risk factors for many tauopathies and synucleinopathies and are associated with certain clinical and pathological features in affected individuals. In addition to clinical, pathological, and genetic overlap, evidence also suggests that tau and alpha-synuclein may interact on the molecular level, and thus might collaborate in the neurodegenerative process. Understanding the role of MAPT variations in tauopathies and synucleinopathies is therefore essential to elucidate the role of tau in the pathogenesis and phenotype of those disorders, and ultimately to develop targeted therapies. In this review, we describe the role of MAPT genetic variations in tauopathies and synucleinopathies, several genotype-phenotype and pathological features, and discuss their implications for the classification and treatment of those disorders.

A pathological hallmark of several neurodegenerative disorders is accumulation of misfolded proteins aggregates. For instance, accumulation of misfolded microtubule-associated protein tau (MAPT) is the hallmark of tauopathies, which may be further classified as primary and secondary. In primary tauopathies, tau is the main aggregated protein, as it is the case in frontotemporal dementia (FTD), corticobasal degeneration (CBD), progressive supranuclear palsy (PSP), Pick’s disease (PiD) chronic traumatic encephalopathy (CTE), and primary age-related tauopathy (PART). In secondary tauopathies, other misfolded proteins are present alongside tau, such as amyloid beta (Aβ) in Alzheimer’s disease (AD)¹. Similarly, synucleinopathies are characterized by the accumulation of misfolded alpha-synuclein and include Parkinson’s disease (PD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA)^2–4. Interestingly, evidence suggests that tau and alpha-synuclein might collaborate in neurodegeneration by co-localizing and promoting misfolding and accumulation of one another^5–7. Tau pathology is also frequently observed in synucleinopathies and vice-versa⁵. However, a clear cause-effect between misfolded protein aggregates and neurodegeneration has not been demonstrated in most tauopathies and synucleinopathies, and it is possible that these aggregates are a byproduct of neurodegeneration, or that misfolded protein aggregation is a protective mechanism that becomes overwhelmed in individuals with disease⁸. Supporting this hypothesis, tau and alpha-synuclein pathology can also be seen in healthy individuals and are positively correlated with age⁸. For these reasons, elucidating the role of misfolded protein aggregation and understanding the role of tau in neurodegeneration is important for both tauopathies an synucleinopathies. While genetic variations in MAPT, the gene encoding tau, is a cause or risk factor for many tauopathies and synucleinopathies⁹, their role in the pathogenesis and phenotype of most of those disorders needs to be clarified. Herein, we describe the roles of MAPT variations in the phenotype and pathological features of tauopathies and synucleinopathies, and discuss the implications of these findings for disease classification and therapy.

MAPT and Tau: structure, function and interaction with alpha-synuclein

MAPT, located on chromosome 17q21.31, is composed of 15 exons and encodes the microtubule associated protein tau. The gene has two main haplotypes due to a 900 kb inversion on chromosome 17q21¹⁰. The H1 haplotype (direct orientation) is observed in all populations and has normal patterns of genetic variation, resulting in multiple different subhaplotypes¹¹. The H2 haplotype (inverse orientation) is almost exclusive to individuals of European ancestry and has a prevalence of about 20% in this population¹². Although the H2 haplotype sequence differs significantly from the H1 haplotype, sequence diversity and variability within the haplotype itself is limited¹⁰. Tau is particularly present in neurons of the central nervous system, especially in axons. Six known isoforms are expressed in the adult human brain (figure 1).

Figure 1.

Structure of the six tau isoforms expressed in the adult human brain

Human MAPT is located on chromosome 17q21.31 and has two main haplotypes: H1 (direct orientation) and H2 (inverse orientation). The gene contains 15 exons, including exon 0, which encodes the 5’ untranslated region (UTR). Tau isoforms can have zero, one, or two amino-terminal inserts based on alternative splicing of exons 2 and 3 (0N, 1N, 2N; shown in blue), The insertion of exon 3 is dependent on the presence of exon 2. Tau isoforms can also include three or four microtubule-binding repeats, depending on splicing of exon 10 (3R or 4R; shown in orange). Exon 4A corresponds to the alternative exon located between exons 4 and 5, and is expressed outside of the brain. Exon 6 is also expressed outside the brain, while exon 8 insertion has not been reported in humans.

Adapted from Refs 1 and 9.

References: 1,9,13,198,227

The protein has four functional domains: an N-terminal projection domain, a proline-rich domain, a microtubule-binding domain (R), and a C-terminus¹³. The major role of tau in the central nervous system is to promote microtubules assembly, involved in neurite polarity, axon elongation, and axonal transport¹³. These functions are influenced by the exonic composition of MAPT isoforms: compared to their counterparts with three microtubule-binding domains (3R) (figure 1), 4R tau isoforms have greater affinity for microtubules and lead to increased microtubule assembly. Tau has other roles in addition to its association with microtubules, including regulation of neuronal plasticity, nucleolar organization, and DNA protection against oxidative stress¹. Tau can also undergo multiple posttranslational modifications (PTMs) that regulate its activity. The most notable of these is phosphorylation, which decreases the affinity of tau for tubulin and facilitates its dissociation from microtubules. Less is known about the influence of other PTMs and epigenetic mechanisms on MAPT, but they are also likely involved in tauopathies. For instance, tau acetylation prevents degradation of phosphorylated tau, and autopsy studies have shown increased levels of tau acetylation in different tauopathies^14–16.

Soluble tau can accumulate in the cytoplasm in the form of nonfibrillar aggregates known as pre-tangles, which eventually undergo conformational changes and form intracellular, insoluble, densely packed neurofibrillary tangles (NFTs)¹⁷. Accumulation of NFTs, and possibly to a greater extent their precursors, might lead to neuronal dysfunction and death by disrupting microtubule assembly, axonal transport and mitochondrial trafficking¹⁸.

Alpha-synuclein, encoded by SNCA, is the main component of proteins aggregates that are the pathological hallmark of synucleinopathies: Lewy bodies (LBs) in Parkinson’s disease (PD) and dementia with LBs (DLB), and glial cytoplasmic inclusions (GCIs) in multiple system atrophy (MSA). In addition to being involved in these three disorders, alpha-synuclein may interact with tau and might contribute to tauopathy via several mechanisms, including inhibition of tau-tubulin binding, tau hyperphosphorylation, initiation of tau polymerization, and promotion of tau aggregation⁵. Conversely, dysfunction of tau-mediated axonal transport may promote accumulation of tau and alpha-synuclein in neurons¹⁹. Alpha-synuclein and tau also co-localize in axons and Lewy bodies, and they synergistically promote the fibrillation and accumulation of each other in vitro and in vivo^6,7,20. There is some pathological overlap between tauopathies and synucleinopathies. LBs are observed in more than half of AD autopsies, and the majority of PD patients also have some AD pathology^21,22.

Tauopathies

Frontotemporal Dementia

FTD is a clinically and neuropathologically heterogeneous group of neurodegenerative disorders characterized by frontotemporal lobar degeneration (FTLD), executive dysfunction, behavioral abnormalities, personality changes, and progressive speech and language difficulties. Affected individuals might also develop motor symptoms such as parkinsonism with or without atypical features^23–26. The three main clinical subtypes of FTD are behavioral variant FTD (bvFTD), progressive non-fluent aphasia (PNFA), and semantic dementia (SD). Pathologically, FTLD is categorized based on the type of inclusion observed, the two most common being tau (FTDL-tau) and TAR DNA-binding protein 43 (TDP-43, FTDL-TDP)^23,27. The clinical and pathological features of FTD and its subtypes are described in table 1. Although clinicopathological correlations have been established, clinical presentation cannot reliably predict autopsy findings. Most cases of SD are associated with FTLD-TDP, while tau inclusions are seen in most cases of PNFA. Individuals with bvFTD can exhibit FTLD-tau, FTLD-TDP, or other types of pathology^28–30. Brain atrophy patterns seem to correlate more closely with clinical phenotype (Table 1)³¹.

Table 1.

Clinical and pathological features of main tauopathies and synucleinopathies

Disease	Proteinopathy type	Motor features	Cognitive features	Protein accumulation pattern	Brain atrophy and neuronal degeneration pattern	References
FTD (FTLD-tau)	Primary tauopathy	Parkinsonism with symmetrical bradykinesia, postural rigidity, absence of resting tremor, and minimal or absent response to levodopa. Features can overlap with CBS or PSP.	bvFTD: personality changes, disinhibition, apathy, impulsivity, loss of empathy, hyperorality, compulsive behavior, repetitive movements, ritualistic behaviors, and stereotypy of speech. Deficits in executive tasks with relative sparing of episodic memory and visuospatial skills. PNFA: impaired speech production and non-fluent speech, impaired comprehension of complex sentences. Spared single word comprehension and object knowledge. SD: impaired word comprehension, loss of semantic memory, surface dyslexia or dysgraphia. Spared repetition and speech production.	FTLD-tau pathologic subtypes: AGD - accumulation of 4R tau-containing argyrophilic grains in the medial temporal lobe. GGT - 4R-tau containing globular glial inclusions and neuronal globular or tangle-like inclusions. Pathology can overlap with PSP and CBD.	bvFTD - Bilateral atrophy of the frontal lobes, insula, anterior cingulate, and anterior temporal lobes. PNFA - atrophy of the anterior perisylvian cortex, mostly in the dominant hemisphere. SD - asymmetrical atrophy of the dominant anterior inferior temporal lobe.	23,26,31–33,36,211–214
PSP	Primary tauopathy	Heterogeneous phenotype with many possible variants. Ocular motor dysfunction, dysarthria, postural instability, and bradykinesia / akinesia are common. Can also present as CBS.	Behavioral and/or language impairment similar to subtypes of FTD	Accumulation of 4R-tau NFTs in subcortical structures, especially the subthalamic nucleus, basal ganglia, and brainstem. The NFTs may be associated with 4R-tau positive tufted astrocytes, oligodendroglial coiled bodies, and neuropil threads.	Atrophy of the subthalamic nucleus and brainstem tegmentum, depigmentation of the substantia nigra.	33,79
CBD	Primary tauopathy	CBS: asymmetric levodopa-resistant parkinsonism, dystonia, myoclonus, orobuccal or limb apraxia, alien limb phenomena, cortical sensory deficit. Can present similarly to PD or PSP.	Can present similarly to FTD or AD-like dementia.	Accumulation of 4R-tau positive neuronal inclusions, white and grey matter threads, coiled bodies, and astrocytic plaques. NFTs and corticobasal bodies are seen.	Asymmetric focal cortical atrophy and depigmentation of the substantia nigra.	85,96
PiD	Primary tauopathy	Most commonly presents as bvFTD, but can also present as other types of FTD and CBS.		3R-tau Pick bodies, predominantly in the dentate gyrus and temporal lobes.	Cerebral atrophy predominant in frontal and temporal lobes.
CTE	Primary tauopathy	Usually late in disease: parkinsonism and, in some cases, motor neuron disease.	Irritability, aggressively, impulsivity, depressive symptoms, and memory impairment. Speech abnormalities and dementia can develop later.	Focal deposition of tau-positive NFTs, tangles, threads, and astrocytes that is most pronounced around cortical sulci and penetrating vessels. The medial temporal lobe, basal ganglia, diencephalon, and brainstem can also be involved, especially later in the disease. TDP-43, Aβ, and alpha-synuclein deposition can also be observed.	Generalized atrophy of the cerebral cortex, notably medial temporal lobes, diencephalon, and mamillary bodies.	104,218
PART	Primary tauopathy	Usually not observed.	Can be associated with memory loss in aging. Can also present with progressive decline in episodic and semantic memory, processing speed, and attention. Symptoms are usually milder and progress more slowly than in AD.	Accumulation of 3R and 4R-tau NFTs mostly in the medial temporal lobe, basal forebrain, brainstem, and olfactory bulb and cortex. The NFTs are identical to those of AD, but Aβ deposits are not seen. Symptom severity correlates with NFT burden.	Neocortical and medial temporal lobe atrophy might be observed in some individuals.	106,219,220
AD	Secondary tauopathy	Extrapyramidal symptoms or parkinsonism can be present late in disease but are not common nor predominant.	Short-term memory loss, visuospatial impairment, and executive dysfunction. Language and behavior impairment typically occur later. Can also develop personality changes, loss of empathy, and obsessive / compulsive behaviors.	NFTs and Aβ extracellular neuritic plaques. NFTs accumulate in a stereotypical manner: 1) entorhinal cortex and hippocampus 2) other limbic structures such as the amygdala and thalamus, 3) in neocortex. LBs are also seen in most cases, predominantly in the amygdala.	Predominant atrophy of medial, basal, lateral temporal, and medial parietal lobes.	208–210
PD	Synucleinopathy	Asymmetric and levodopa-responsive parkinsonism. Shuffling, festinating gait. Postural instability and autonomic dysfunction can occur later in disease.	Sleep dysfunction, depression, anxiety, apathy. Dementia occurs late in disease.	Accumulation of misfolded alpha-synuclein into LBs, predominantly in the substancia nigra. NFTs and Aβ plaques can also be identified.	Loss of dopaminergic neurons, predominantly in the ventrolateral area of the substantia nigra pars compacta.	2.135,215
DLB	Synucleinopathy	Parkinsonism (often less levodopa-responsive than PD), dysautonomia.	Dementia, fluctuating cognition, visual hallucinations, RBD, apathy, depression.	LBs predominantly in the brainstem. LBs can also accumulate in limbic structures and in the cerebral cortex, particularly the frontal and temporal lobes. Aβ plaques with or without NFTs can be seen.	Pallor of the substantia nigra, atrophy of the midbrain, hypothalamus, substantia innominate, lateral prefrontal cortex, and left premotor cortex.	3,170,216
MSA	Synucleinopathy	Levodopa-unresponsive parkinsonism (can be responsive to levodopa, especially early in disease), pyramidal signs, cerebellar abnormalities, and autonomic dysfunction. Predominant symptoms depend on subtype: MSA-P (parkinsonism) and MSA-C (cerebellar dysfunction).	Sleep dysfunction, depression, anxiety, attention deficit. Presence of dementia should prompt consideration of alternative diagnosis, such as DLB.	Presence of argyrophilic GCIs that are alpha-synuclein and tau-positive.	Neuronal loss and gliosis in the striatonigral and olivopontocerebellar systems.	4,179,217

Abbreviations: FTD: frontotemporal dementia; FTLD: frontotemporal lobar degeneration; CBS: corticobasal syndrome; PSP: progressive supranuclear palsy; bvFTD: behavioral variant frontotemporal dementia; PNFA: progressive non-fluent aphasia; SD: semantic dementia; AGD: argyrophilic grain disease; GGT: globular glial tauopathy; CBD: corticobasal degeneration; NFTs: neurofibrillary tangles; PD: Parkinson’s disease; AD: Alzheimer’s disease; PiD: Pick’s disease; CTE: chronic traumatic encephalopathy; TDP-43: TAR DNA-binding protein-43; Aβ: amyloid beta; PART: primary age-related tauopathy; LBs: Lewy bodies; DLB: dementia with Lewy bodies; RBD: rapid eye movement sleep behavior disorder; MSA: multiple system atrophy; GCIs: glial cytoplasmic inclusions

FTLD-tau can be further divided based on the tau isoform contained in the inclusions. Pick’s disease (PiD) is associated with aggregation of 3R-tau, while 4R-tau is seen in argyrophilic grain disease (AGD) and globular glial tauopathy (GGT). Pathological patterns that can be seen in FTLD-tau can also include CBD and PSP (table 1)³². The pathology of the different subtypes of FTLD-tau and other tauopathies has been described in detail in a recent review³³.

More than 50 pathogenic MAPT mutations have been reported in FTLD, comprising 5-20% of familial cases depending on the population^34,35. The phenotype of affected individuals is highly variable, even in patients with the same variant from the same family. The different phenotypes described in FTLD-tau and corresponding MAPT mutations are shown in figure 2^9,25,36–63. Pathological features seen in MAPT-associated FTLD-tau are similar to the different FTLD-tau subtypes seen in sporadic cases and include PiD, AGD, GGT, CBD, and PSP^{32,37,64–66}. The vast majority of pathogenic mutations are either coding mutations affecting microtubule-binding repeats and their flanking region (exons 9-13, figure 2), reducing their ability to bind to microtubules, or variants in exon 10 and intron 10 that lead to differential splicing and a relative increase in 4R tau. Accordingly, mutations in exon 10 and intron 10 are mostly associated with predominantly 4R FTLD-tau (AGD, GGT, CBD, PSP), while most mutations in other exons are associated with predominant 3R pathology (PiD)³⁷. The p.V337M and p.R406W mutations, in exon 12 and 13 respectively, have been associated with mixed 3R and 4R isoforms and paired, straight, helical tau filaments³⁶. The molecular pathophysiology of specific MAPT mutations and the associated pathological and neuroimaging features has been comprehensively reviewed³⁶. Unlike other mutations, the p.A152T MAPT variant is not a disease-causing mutation but rather a risk factor for the development of FTD^67,68. The role of MAPT haplotypes in FTD is unclear. While some studies have found an association between FTD and the H1⁶⁹⁻⁷¹ or H2^72,73 haplotype, other studies, including a meta-analysis of the H2 haplotype, have found no such association^74–77. There are contradicting results regarding the effect of the H2 haplotype on AAO of FTD^{70,71,73,77,78}. Overall, it seems that these haplotypes do not have a major role in FTD, whereas rare MAPT mutations are important in FTD.

Figure 2.

MAPT mutations and associated clinical presentations in frontotemporal lobar degeneration

The vast majority of reported mutations are in exons 9-13 and in the intronic region following exon 10. The most commonly reported presentations are bvFTD, PSP, and unspecified types of FTD. The mutations associated with a bvFTD and unspecified FTD presentation are well distributed among exons 9-13, while all but 3 mutations presenting with a PSP phenotype are within exon 10 and the intronic region following it. The four mutations associated with a CBS presentation are located in exons 10 and 13. The two mutations associated with PNFA are in exon 12 and 13 and the two SD-associated mutation are located in exons 1 and 10. All the intronic mutations are associated with either bvFTD or an unspecified FTD subtype. PSP was the other clinical presentation observed in intronic mutations. The three most commonly reported mutations are N279K, P301L, and IVS10+16. Exons in orange encode the microtubule-binding repeats, while exons in blue encode the amino-terminal inserts.

References: 9,25,36–63,227

Progressive supranuclear palsy

PSP is traditionally considered an atypical parkinsonian syndrome with a heterogeneous phenotype⁷⁹. The main clinical and pathological features of PSP are detailed in table 1. Most identified pathogenic MAPT mutations (summarized in figure 2) are located in exon 10 and intron 10, which is consistent with the 4R-tau pathology seen in PSP. However, families with autosomal dominant PSP with MAPT mutations are rare^42,80,81, and the main role of MAPT in PSP is as the most important genetic risk factor rather than a monogenic cause. Numerous studies have shown that the MAPT H1c subhaplotype and the H2 haplotype are respectively associated with an increased and decreased PSP risk^75,82–85. The increased PSP risk associated with the H1c subhaplotype may be explained by increased levels of MAPT gene expression and proportion of 4R isoform^86–88. An increase in N-terminal exon-containing MAPT transcripts may explain the reduced risk seen with the H2 haplotype⁸⁹.

Despite the clear association between the MAPT H1c subhaplotype and risk of PSP, only few studies have investigated correlations between MAPT variations and clinicopathological phenotype. No difference in age at onset (AAO), motor and cognitive symptoms, and survival between heterozygous and homozygous H1 PSP patients have been reported⁹⁰. In addition, similar rates of cognitive impairment and no significant difference in tau burden and Braak NFT stage were found when comparing H1/H1 and H1/H2 carriers with PSP⁹¹. This lack of significant impact of tau haplotypes on PSP pathology was also previously noted^92,93. Interestingly, another study analyzed cognitive performance in 305 PSP patients, and found that carriers of the H1c subhaplotype had better performance in terms of executive function, attention, and general cognitive function⁹⁴. The H1c subhaplotype has also been associated with a more severe pathological phenotype⁹⁵. Compared to FTD, in PSP the H1 subhaplotypes and the H2 haplotypes seem to have a more prominent role.

Corticobasal degeneration

CBD is a pathological entity that most commonly presents clinically as corticobasal syndrome (CBS), a form of atypical parkinsonism. The pathological features and possible clinical phenotypes of CBD are described in Table 1⁹⁶. CBD and PSP show some MAPT-related genetic overlap: MAPT mutations are a rare cause of CBD, and the H1 MAPT haplotype and H1c subhaplotype are well-established risk factors for CBD; the H2 haplotype is also associated with reduced risk^{11,75,97–100}. Genotype-phenotype correlations studies of MAPT in CBD are even more limited than they are for PSP. The only such study found no correlation between tau pathology burden and MAPT haplotype in 36 H1/H1 and 9 non-H1/H1 CBD autopsy cases⁹³. There are major limiting factors for determining the influence of MAPT on the clinical and pathological characteristics of CBD, including the significant clinicopathological overlap with other tauopathies, the rarity of CBD (and even lower prevalence of non-H1/H1 patients), and necessity of autopsy for a definitive diagnosis.

Pick’s disease

PiD is a unique pathological entity, as it is the only pure 3R-tau tauopathy¹⁰¹. Its main pathological features and the associated clinical presentations are described in table 1. Factors that contribute to the 3R-tau nature of Pick bodies include the disease-specific patterns of tau filament phosphorylation and folding seen in PiD^41,102. MAPT mutations associated with PiD tend to be localized outside of exon and intron 10, as these mutations typically lead to a relative increase and 4R-tau and are associated with corresponding 4R-tau pathologies³⁷. Some pathological features seen in PiD are associated with specific mutations. For instance, the MAPT p.K257T and p.S320F mutations are associated with Pick bodies and axonal inclusions, while the p.L266V and p.L315R mutations are also associated with glial inclusions. In contrast to 4R-tauopathies, the MAPT H1 haplotype does not seem to be associated with PiD¹⁰³. This lack of association might explain some of the conflicting results seen in association studies of MAPT haplotypes in FTD, as the different studies may have different proportions of patients with PiD pathology. This emphasizes the importance of pathological confirmation of disease, or preferably, identification of reliable biomarkers that can distinguish between the pathologies. This may also indicates that PiD pathology might be associated with distinct diseases mechanisms and will have different responses to treatment. This distinction is therefore also important for future therapy development and participation in clinical trials.

Chronic traumatic encephalopathy

CTE is a pathologically defined neurodegenerative disease that usually begins to develop about 10 years after repetitive traumatic brain injury¹⁰⁴. Its clinical and pathological features are described in table 1. Although it is considered a tauopathy, TDP-43, Aβ, and alpha-synuclein deposition can also be observed¹⁰⁴. Genetic studies in CTE are limited, and no association with MAPT haplotypes or variants have been reported so far^104,105. Further studies are needed to elucidate the genetics of CTE.

Primary age-related tauopathy

PART is a neuropathological pattern commonly observed in older individuals¹⁰⁶. Its pathology and associated clinical features are described in table 1. Despite the fact that research on PART has increased in the last years, it remains an understudied entity. The H1 MAPT haplotype was overrepresented in two small studies that included a total of 30 patients^107,108, which are too underpowered to determine whether this association is true. Next steps in the study of PART involve a better characterization of the clinical spectrum associated with the pathological features, and genetic association studies that will evaluate the impact of MAPT on its incidence, pathology, and symptomatology.

Alzheimer’s disease

AD is the most prevalent neurodegenerative disorder and the most common cause of dementia¹⁰⁹. The pathology and symptomatology of the disease is described in table 1. It is considered a secondary tauopathy: in addition to intracellular NFTs, AD is also characterized by aggregation of Aβ in the form of extracellular neuritic plaques. The two proteins may act synergistically in a positive feedback loop to promote their aggregation: Aβ leads to tau hyperphosphorylation, while extracellular secreted tau leads to increased level of Aβ^110–112. A recent study has shown that the biochemical characteristics of tau vary among individuals with AD, and that features that promote tau seeding, such as high levels of high-molecular weight, hyperphosphorylated, oligomeric tau are associated with increased tau burden and a more severe clinical phenotype¹¹³. Early case-control genetic studies of MAPT in AD have been inconsistent in the identification of risk variants^87,114–120. More recent studies, including meta-analyses, have identified an association between the H1 haplotype (particularly the H1c subhaplotype) and increased risk of AD and an association with reduced risk with the H2 haplotype^{75,88,89,121–123}. In addition, the A allele of the rs393152 single nucleotide polymorphism (SNP), which is located within the extended MAPT locus, was found to be significantly associated with increased risk of AD in a meta-analysis⁸⁸.

There is currently no definitive evidence that monogenic forms of AD are caused by MAPT mutations ^124–127. Table 2 summarizes the genetic findings related to MAPT in AD. Genotype-phenotype studies of MAPT in AD are limited: non-replicated studies have shown an association between the H1 haplotype and a higher prevalence of symptoms in individuals who met AD pathological criteria at autopsy¹²⁸, a decreased NFT and Aβ plaque burden¹²⁹, and a faster rate of cognitive decline in affected individuals¹³⁰. Studies correlating cerebrospinal fluid levels of tau and MAPT mutations have shown conflicting results^131–134.

Table 2.

Reported MAPT genetic findings in tauopathies and synucleinopathies

Disease	Causative variants	Risk variants and haplotypes	Phenotype-genotype correlations	References
FTD	Numerous causative variants reported; see figure 1.	p.A152T is a risk variant. Role of MAPT haplotype is unclear.	Earlier age of onset, higher prevalence of movement disorder, and lower prevalence of language disorder compared to other genetic causes of FTD. Earlier AAO compared to sporadic cases of FTD. H2: possibly earlier AAO. See figure 1 for phenotypes associated with reported causative variants.	66–68,70–73,75–78,194,221
PSP	Numerous causative variants reported; see figure 1.	H1c: increased risk. H2: decreased risk. H1d, H1g, and H1o: increased risk (one large study). H1b and H1q: increased risk (one small study each).	H1c: better cognitive performance (one study), more severe pathological phenotype (one study). H1: earlier AAO (one study).	82–84,93–95,99,193,222
CBD	Numerous causative variants reported; see figure 1.	H1c: increased risk. H2: decreased risk.	No correlation between tau pathology burden and MAPT haplotype (one study).	99
PiD	Numerous causative variants reported; see figure 1.	MAPT haplotype are not associated with PiD risk.	See FTD.	41,43,50
CTE	None.	No significant association found (one study)	None reported.	105
PART	None.	H1 overrepresented in two small studies.	None reported.	107,108
AD	p.R406W mutation and 17q21.31 microduplication have shown AD-like phenotype but lack Aβ pathology.	H1 and H1c: increased risk. H2: decreased risk. Unclear association with pA152T variant rs393152: increased risk with A allele	H1: faster cognitive decline, decreased NFTs and Aβ burden, lower prevalence of symptoms among those that met AD pathological criteria (one study each). Conflicting results for genotype associated with CSF tau levels.	68,69,88,123,125–134,190,192
PD	None.	H1: increased risk. H2: decreased risk.	H1: decreased cognitive performance, increased risk of PD dementia (mixed results), non-tremor dominant phenotype, hallucinations. H1p: PD dementia (one study). H1h: non-tremor dominant PD (one study). Intronic MAPT SNPs (rs2435207 and rs11079727): later onset of motor symptoms in LRRK2-associated PD. Mixed genotype-pathology findings.	88,138.148–162,168,169,195,224–226
DLB	None.	p.A152T, H1/H1 diplotype, and H1g subhaplotype are potential risk factors (mixed results).	H1: Increased alpha-synuclein deposition, particularly in brainstem regions (one small cohort).	171–173,177,178
MSA	None.	H1, H1j and H1x: increased risk. H1e and H2: reduced risk. (one study each)	No association between genotype and AAO (one study).	180,181,186

Abbreviations: FTD: frontotemporal dementia; AAO: age at onset; FTLD: frontotemporal lobar degeneration; PSP: progressive supranuclear palsy; CBD: corticobasal degeneration; PiD: Pick’s disease; CTE: chronic traumatic encephalopathy; PART: primary age-related tauopathy; AD: Alzheimer’s disease; NFTs: neurofibrillary tangles; Aβ: amyloid beta; CSF: cerebrospinal fluid; PD: Parkinson’s disease; DLB: dementia with Lewy bodies; MSA: multiple system atrophy

Despite the fact that NFTs are a pathological hallmark of the disease, the influence of MAPT variations seems to be minor in AD. An issue that can complicate genetic studies is the significant clinical overlap between AD and other tauopathies and the lack of neuropathological confirmation in many cases. Further research is required to clarify the impact of MAPT variation on the clinicopathological phenotype of AD.

Synucleinopathies

Although this group of diseases is defined by the presence of alpha-synuclein accumulation, tau pathology is quite common, and a role for MAPT genetic variants have been demonstrated with different degrees of certainty.

Parkinson Disease

PD is the most common synucleinopathy¹³⁵. The main clinical manifestations and pathological findings of PD are described in table 1. Despite being considered a synucleinopathy, NFTs (and Aβ plaques) can also be found in PD, with a higher tau pathology burden in those who also developed dementia¹³⁶. The MAPT H1 haplotype is consistently identified as one of the strongest genetic risk factors for PD, while the H2 haplotype is associated with reduced risk^75,137–141. Subhaplotype analysis have yielded inconsistent results between different populations^142–145 or no association with PD^137,146,147. Some studies have suggested genotype-phenotype correlations, yet many of those were not replicated, and therefore need to be considered cautiously. For example, several studies suggested that MAPT variants affect cognition in PD, demonstrating an association of the H1 haplotype with decreased cognitive performance^148,149 and increased risk of dementia^150–153. However, these results were not replicated in other studies^154–160. Other features that have been reported to be more common in H1/H1 PD patients include a non-tremor dominant phenotype^161,162 and hallucinations¹⁵⁵. Although one study reported an association between the H2 haplotype-tagging SNP rs8070723 and an older PD AAO¹⁶³, other studies found no such association^164–167. Reports of correlations between MAPT variations and PD pathological findings have shown inconsistent results including increased LB burden associated with the H1¹⁶⁸ or H2 haplotype¹⁶⁹, decreased NFTs in patients with the H1 haplotype¹²⁹, or no association between MAPT genotype and pathology¹⁵⁹.

The role of the MAPT H1 haplotype as a risk factor for PD is well-established. Larger, and preferably longitudinal studies, are required to properly examine whether MAPT variants have true effects on the clinical course of PD.

Dementia with Lewy bodies

DLB is one of the most common causes of dementia¹⁷⁰. Table 1 describes the main clinical and pathological findings associated with the disease. Genetic association studies investigating MAPT as a potential risk factor for DLB have found overrepresentation of the p.A152T¹⁷¹ variant, H1/H1 diplotype¹⁷², and H1g subhaplotype¹⁷³ in affected individuals. However, these results have not been reported in three large studies^174–176. In an autopsy cohort of 22 DLB patients, an increased alpha-synuclein deposition the brains of H1/H1 patient was noted, particularly in brainstem regions^177,178. These results have yet to be replicated in other cohorts. Despite the fact that tau pathology is not uncommon in DLB, current evidence on a possible influence of MAPT variation on disease risk or phenotype is limited and suggest a minor effect at best. Further studies are needed to clarify these potential associations and the role of MAPT in DLB.

Multiple system atrophy

MSA, the third major synucleinopathy, is a rare neurodegenerative disorder that can have a broad range of manifestations¹⁷⁹. The clinical and pathological findings of the disease are described in table 1. Few studies on the role of MAPT variation in MSA have been published. A positive association between the H1 haplotype and MSA has been reported¹⁸⁰ , while another study¹⁸¹ suggested that the H1j and H1x subhaplotype are risk factors for MSA. The H1e and H2 haplotypes were associated with reduced risk. However, these studies included only 59 and 213 MSA cases, respectively. Other studies did not identify associations between MAPT and MSA, although they only genotyped a limited number of variants^182–185. In another study, no difference in haplotype distribution when comparing 22 early-onset (defined by the authors as ≤ 55 years) and 18 late-onset MSA cases was found¹⁸⁶. Similar to DLB, current evidence, although limited, suggest that MAPT variations are either not significant or have a minor role in MSA.

Overlap between tauopathies and synucleinopathies

Tauopathies and synucleinopathies share various clinical, pathological, and genetic characteristics. Cognitive impairment is a feature of all the reviewed disorders (except MSA), and parkinsonism is also common in PD, DLB, MSA, PSP, and CBD. In fact, the four latter diseases are commonly considered as “Parkinson-plus” syndromes¹⁸⁷. Parkinsonism can also be seen in FTD, CTE and late-stage AD (table 1). Tau pathology may be present in all of these disorders, either as a primary or secondary finding in tauopathies, and also as a component of other inclusions such as Lewy bodies and glial cytoplasmic inclusions in synucleinopathies (table 1). In addition, the MAPT H1 haplotype is associated with increased risk of PSP, CBD, AD, PD, and potentially FTD, PART, DLB, and MSA (table 2). This overlap may suggest possible interactions between alpha-synuclein and tau, and/or shared mechanisms of neurodegeneration. Another hypothesis is that different MAPT variants, and/or tau or alpha-synuclein accumulation make neurons more vulnerable in general, and neurodegeneration occurs due to additional factors. Indeed, different alterations in MAPT could contribute to the variation in observed phenotypes. For instance, the A allele of the H1-tagging rs242557 SNP is associated with increased risk of PSP and CBD but not PD^139,188. If tau and alpha-synuclein are themselves pathogenic, it is possible that their co-occurrence in tauopathies and synucleinopathies is the result of a positive feedback mechanism, where the predominant protein promotes its own accumulation by enhancing misfolding and aggregation of its counterpart. Further studies are required to determine the causes and implications of the potential interactions between tauopathies and synucleinopathies, notably because current evidence is insufficient to determine whether misfolded alpha-synuclein and tau aggregates are pathogenic, protective, both (at different disease phases) or by-products of neurodegeneration⁸. Better understanding of these interactions could have important implications for future research and therapeutics.

Discussion – The role of MAPT in the spectrum of neurodegeneration

The current review highlights the potential roles of MAPT variation and the tau protein in tauopathies and synucleinopathies. These diseases vary in terms of age of onset, clinical manifestations, pathology, and pattern of inheritance (table 1). They also show some overlap. This overlap is even more prominent in FTD, PSP, and CBD, where the symptomatology of one disease can be associated with the pathology of another. In a recent study, 310 patients with a syndrome likely to be caused by frontotemporal lobar degeneration (bvFTD, PNFA, SD, PSP, CBD) were evaluated, and it was found that 62% of patients met the diagnostic criteria for more than one disorder. The same study also showed increased syndromic overlap in 46 patients at follow-up and a continuous spectrum of association between brain atrophy and clinical manifestations¹⁸⁹. Genetic evidence also reinforces the concept of a spectrum, as the H1 haplotype is associated with PSP, CBD, and potentially FTD, and the fact that the same MAPT mutations might cause different clinical syndromes (figure 2). PSP and CBD also share several genetic risk factors outside of MAPT^85,100.

The role of MAPT in neurodegenerative disorders extends beyond genetics and involves epigenetics and PTMs of tau. Tau hypomethylation was reported in AD¹⁹⁰ (although not seen in two prior studies^191,192), PSP^95,193,194, and in the putamen of PD patients¹⁹⁵. Tau hyperphosphorylation leads to its aggregation and is the most studied PTM, but other mechanisms are also likely involved in the pathogenesis of tauopathies and related disorders. Tau acetylation prevents degradation of phosphorylated tau, and post-mortem studies have shown increased levels of tau acetylation in AD, CBD, PSP, and other tauopathies^14–16. Using cryo-electron microscopy, two recent studies have shown that tau filaments have disease-specific structures and folding patterns that are influenced by distinct PTM signatures that involve a combination of acetylation, ubiquitination, trimethylation, and phosphorylation^196,197. Another potential factor that explains the clinical spectrum seen in tauopathies is that the proportion of different tau isoforms varies between different brain regions and between healthy and diseased individuals¹⁹⁸.

The role of tau in neurodegeneration, in addition to its microtubule-related functions, may also involve other pathophysiologic pathways such as oxidative stress and mitochondrial dysfunction^1,18. This may partially explain the failure of tau-targeting medications in clinical trials of AD¹⁹⁹ and other tauopathies²⁰⁰. This would also provide insight on why the risk of developing PD, and potentially other synucleinopathies, is influenced by MAPT variations, beyond the co-localization and synergistic fibrillation of tau and alpha-synuclein. Another factor to consider is that the 17q21.31 region, where MAPT is located, contains other candidate genes for neurodegeneration, such as NSF and KANSL1, which could potentially contribute to the overrepresentation of the H1 haplotype in several neurodegenerative disorders^201–203. NSF is involved in synaptic vesicle exocytosis and is a substrate of the PD-associated gene LRRK2²⁰⁴. Recently, de novo heterozygous NSF variants were reported in two unrelated children with early infantile epileptic encephalopathy²⁰⁵. KANSL1 is involved in chromatin modification and neuronal process, and KANSL1 haploinsufficiency has been shown to be sufficient to cause 17q21.31 syndrome, a rare multi-system disorder that is notably associated with intellectual disability, hypotonia, and epilepsy²⁰⁶.

Beyond its role as a risk factor, MAPT seems to be a disease modifier, as numerous correlations between genotype and motor, non-motor, and pathological features of different diseases have been reported (table 2). However, many of these associations need to be more clearly defined, as they were either observed in a small number of studies or not replicated. Better insight into these genotype-phenotype associations may lead to a better understanding of mechanisms of different symptoms, and to patient stratification for prognosis, treatment, and inclusion in clinical trials.

Concluding remarks

Despite the unclear mechanisms by which tau accumulation is associated with neurodegeneration, further understanding of the genetics MAPT is essential to improve disease classification and genotype-phenotype correlations of tauopathies and synucleinopathies. Furthermore, classification systems incorporating genetic information might better represent distinct pathogenic mechanisms and therefore better guide the development of new therapies and inclusion criteria in future clinical trials. In addition to genetics, should these criteria be based on specific symptoms, clinical diagnosis, genotype, brain atrophy patterns on imaging, potential biomarkers, or a combination of multiple factors? Currently, there is no definitive answer to this question, although it is clear that the current classification system, which is mainly based on clinical symptoms, has several limitations that might impede the implementation of targeted therapies. For instance, patients with similar clinical presentations of frontotemporal dementia may have 3R and 4R tauopathies and require different treatments, as these pathologies might be associated with different underlying mechanisms. This distinction might also expand to 4R tauopathy subtypes, as the different protein inclusions observed in AGD, GGT, PSP, and CBD could reflect distinct pathogenic mechanisms responding differently to therapy. In that case, more reliable pathological markers would have to be identified, since clinical manifestations and genotypes can only hint at a specific type of pathology but not reliably predict it. Whether tau-targeting drugs will prove to be beneficial in neurodegenerative diseases is still to be determined, since thus far no tau-targeting therapy has been successful in clinical trials^199,200. However, this does not exclude MAPT and tau as therapeutic targets of interest. For instance, MAPT mutations and haplotype leading to increased tau expression could be targeted with antisense nucleotides (ASO), which could decrease tau expression via several mechanisms including splicing modulation, translational arrest, and targeted degradation²⁰⁷. It is also possible that only certain genetic subtypes of tauopathies (and potentially synucleinopathies) could respond to tau therapy. In any case, further understanding of the role of MAPT in neurodegenerative disorders will be crucial, and might lead to a better understanding of neurodegeneration, the mechanisms of specific diseases, the development of therapies that either target tau or biologically related mechanisms, and the selection of specific individuals who may benefit from such treatments.