Glial contributions to neurodegeneration in tauopathies

Tau pathobiology

The MAPT gene on chromosome 17q21.31 encodes the tau protein which can be alternatively spliced into six distinct isoforms [8, 9]. These isoforms differ by inclusion of either three or four conserved repetitive domains (termed R) as well as the addition of one or two amino-terminal inserts [8]. Aggregates composed of both 3R and 4R tau are seen in disease states, though curiously several tauopathies including PSP, CBD, FTDP-17, and AGD exclusively feature 4R tau deposits. 4R tau does display a higher propensity for aggregation [10] which has led some to postulate that it is more pathogenic [11–13]. A recent study reported significantly higher levels of aggregated hyperphosphorylated tau (p-tau) and less-soluble tau species after using anti-sense oligonucleotides to increase the ratio of 4R:3R tau in mice expressing human tau under the endogenous promoter [12]. While this data supports 4R tau pathogenicity, other factors still contribute to 3R tau aggregation which is seen exclusively in PiD and in the mixed 3R and 4R tau inclusions in AD and CTE [1, 2].

In a healthy brain, tau is predominantly localized in mature neuronal axons and primarily functions to promote microtubule assembly and stability as well as vesicle and organelle transport along microtubules [14–16]. Phosphorylation of serine and threonine residues flanking the microtubule binding domain of tau regulate its interactions with tubulin and influence its conformational state [2, 17]. Therefore, inappropriate phosphorylation of tau at these regions can lower its affinity for tubulin and inhibit its ability to promote microtubule assembly [15]. Free tau species are vulnerable to hyperphosphorylation, which can leave the intrinsically disordered protein more prone to forming β–sheet conformations that promote aggregation into filamentous neurofibrillary tangles (NFTs) that fill neuronal soma and dense neuropil threads (NTs) that line neuronal processes. In addition to phosphorylation, tau can undergo a variety of other post-translational modifications such as acetylation [18–20], glycosylation [21, 22], methylation [23, 24], nitration [25, 26], O-glycosylation [27–29], polyisomerisation [30, 31], SUMOylation [32, 33], truncation [34–37], and ubiquitination [38–40]. These modifications alter tau structure, function and cellular localization which influence its pathophysiology [2, 14].

Seminal studies by Braak and Braak first described a spatial and temporal pattern in the appearance of tangles in AD patient brains that follow neuronal networks and correlate with cognitive decline. In the AD pattern, NFTs first appear in the transentorhinal region and progress along anatomical pathways to the hippocampus and eventually the neocortex [41, 42]. An analogous pattern has been recapitulated in two independent mouse models, rTgECtau mice, where mutant tau was exclusively expressed in the entorhinal cortex and neurons containing aggregates but lacking tau mRNA were found downstream in the dentate gyrus and hippocampus [43, 44]. Similar temporal progression of tau pathology is observed in AGD, though the brain regions involved differ [45]. Likewise, the spatial distribution of tangles is distinct in other tauopathies [1, 2] indicating additional mechanisms involving the vulnerability of certain neuronal populations contribute to disease.

The formation of NFTs was once solely attributed to inherent susceptibility of individual neurons to the disease process. While still a contentious topic [46], there is now substantial evidence that also supports the idea that propagation of pathological tau species occurs between cells [47–50]. Tau, not inside a membrane compartment, is readily detected in the conditioned media of cultured neurons [51–57] and in the interstitial fluid (ISF) of the brain under normal conditions [58–61]. The mechanism of tau release is still unclear, though reports have linked it to synaptic activity [54, 59]. It has also been found in exosomes [7, 55, 62]. Once released tau may be taken up by cells via macropinocytosis as well as potentially other mechanisms [56, 62, 63]. How tau escapes endosomal compartments once internalized is unknown, however cell culture studies have demonstrated that misfolded tau aggregates can mediate a templated misfolding or “seeding” of normal, monomeric tau to induce intracellular tau aggregation [57, 64–66]. Indeed, high molecular weight species of tau isolated from the ISF or cerebral spinal fluid (CSF) of transgenic mice or AD patients has been shown to seed intracellular tau both in vitro and in vivo [67, 68]. This emerging data suggests tau secretion may be a physiological process that is hijacked in disease states. In vivo tau spreading models further support this potential mechanism of tau propagation in that injection of recombinant tau fibrils or brain lysate containing tau aggregates into the brains of wildtype or young transgenic mice can induce robust pathology at the site of injection and in anatomically connected regions [69–73]. The induction of tau pathology in mice that do not otherwise develop tau inclusions supports the concept of seeding and the propagation of tau aggregates to neuronal populations anatomically connected to the site of injection supports a non-cell autonomous mechanism for disease progression. However, whether the spread of pathological tau species is necessary or sufficient for tauopathy and neurodegeneration in humans remains to be proven.

Although tau is predominantly produced by neurons in the brain, it is expressed at low levels in oligodendrocytes and astrocytes and tau pathology is prevalent in these cells across tauopathies [74]. Tau accumulates to form fine, branching coiled bodies and argyrophilic threads that line myelinating processes in oligodendrocytes [75, 76]. Astrocytic tau pathology occurs in several tauopathies but can appear differently. Diffuse granular p-tau clustered around a nucleus of dense tangles illustrates tufted astrocytes specific to PSP. Alternatively, circular tau puncta localized to distal processes compose astrocytic plaques in CBD while ramified bushy astrocytes are typical to AGD. Thorn-shaped astrocytes feature perinuclear tau deposits and are relatively more common as they are observed in PSP, AGD, PiD, AD and in the brains of the cognitively normal elderly [74, 77, 78]. The diversity of astrocytic tau pathology and the implications of each subtype are still largely unknown as is the percentage of tau in glial inclusions that are derived from astrocytes and oligodendrocytes versus neurons. Interestingly, glial fibrillary acidic protein (GFAP), an astrocyte-specific marker commonly up-regulated in activated states, is redistributed differently per each astrocytic phenotype [79]. In addition, reactive gliosis correlates more closely with thorn-shaped astrocytes as opposed to tufted astrocytes [77, 78] suggesting the first may be a common pathological response while the latter independent of the reactive gliotic process [74]. Furthermore, tau lesions impact glial functions leading to an array of deleterious consequences both within the glia themselves as well as non-cell autonomous effects on neuronal health. Tau inclusions have also been reported in microglia [80–82] despite a lack of tau expression, providing further evidence that pathological tau may also be transferring between cells in the brain.

Risk factors for tauopathies implicate a role for gliosis and neuroinflammation

Reactive gliosis and neuroinflammation were historically considered secondary events in tauopathies and other neurodegenerative diseases. Since the start of the twenty-first century however, accumulating evidence has suggested that aberrant activation of microglia and astrocytes drives chronic neuroinflammation which negatively impacts disease progression. Genetic studies have further implicated roles for the innate immune system in neurodegenerative diseases, particularly AD.

Whole exome sequencing studies have identified numerous gene variants that influence risk for developing AD with varying degrees. Notably, variants of TREM2, an immunoglobulin-like cell-surface receptor primarily expressed on microglia in the brain, were recently found to confer a 2 to 4-fold increased risk for AD [83]. Exactly how TREM2 variants confer AD risk is still under investigation, but current studies indicate it may be due to a loss-of-function in lipid sensing, microglia proliferation and or microglial response to Aβ plaques [84]. However, TREM2’s effect on AD risk is still second to the greatest risk factor for late-onset sporadic AD, apolipoprotein E (ApoE). In the brain, ApoE is predominantly secreted by glial cells and functions as a major transporter of lipoproteins between cells in the brain. Of the three ApoE alleles, ε2, ε3, and ε4, the ApoEε4 allele is associated with a 4–12-fold increased risk based on allele dosage [85, 86]. ApoEε4 is largely thought to influence AD pathogenesis by decreasing Aβ turnover and clearance as well as by directly influencing Aβ aggregation [87]. Additionally, ApoEε4 has been found to have reduced ability to suppress inflammatory stimuli and higher densities of NFTs have been reported in ApoEε4 carriers [88]. Interestingly, the ApoEε4 genotype has also been found to be over-represented in FTD [89, 90] including correlating with increased brain atrophy in patients [91]. Therefore, one possibility is that ApoEε4may increase neuroinflammation which may enhance tau pathology and/or neurodegeneration independent of its influences on Aβ. The exact contributions of ApoE and TREM2 on tau pathogenesis remains unclear and should be more thoroughly assessed in future research. Other gene variants associated with influencing AD risk that impact microglia function and complement signaling include CD33, CR1, ABCA7, SHIP1, BIN1, CD2AP, CLU, HLA-DRB5/DRB1, INPP5D, SORL1, EPHA1, PLD3, PICALM, and MS4A [2, 92, 93]. While in some cases protein products of these genes have been found to influence Aβ accumulation and structure such as CD33 and CLU [94, 95], additional studies are needed to understand their implications in primary tauopathies.

In addition to genetic predisposition, there is evidence that environmental factors that promote neuroinflammation contribute to tau pathogenesis. It is widely recognized that traumatic brain injury (TBI) predisposes individuals to dementia, particularly AD. Increasing evidence further indicates that repetitive mild TBI, with or without concussions, can have long-term consequences leading to tauopathy and neurodegeneration as seen in CTE [96]. Neuroinflammation may be a significant contributor to secondary cell death immediately following moderate to severe TBI and inflammatory effects have been shown to persist up to 17 years post-injury [97, 98]. Similarly, mild TBIs instigate reactive gliosis and prime microglia to over-react to future insults. It is hypothesized that while acute gliosis is arguably protective following TBI, repetitive insults provoke microglia and astrocytes to release markedly higher levels of proinflammatory molecules that can affect neuronal homeostasis and regulate tau release and aggregation [99]. Additionally, environmental toxins and viral infections have all been shown activate gliosis and impact tau pathophysiology [100–102]. Altogether, genetic and environmental risk factors for AD and other tauopathies implicate that glial cells and chronic inflammation may have a more active role in the degenerative process than previously thought. In AD, Aβ plaque deposition may initially provoke gliosis [4, 103] while repetitive mild TBIs have been shown to prime microglia and lead to exacerbated inflammatory responses that are speculated to contribute to the development of CTE [99]. The hypothesis that chronic neuroinflammation plays a causal role in neurodegeneration is rapidly changing the way the field approaches disease research.

Microgliosis in tauopathy

Microglia are the resident immune cells in the brain and have a nuanced role in neuroprotection and maintenance of homeostasis. Yet under pathological conditions microglia become activated and transform into a ramified, branched state. These cells have the capacity to migrate, proliferate and efficiently phagocytose pathogens and cellular debris, including protein aggregates [104]. Furthermore, activated microglia may release a host of proinflammatory cytokines including interleukin (IL) 1-beta (IL-1β), tumor necrosis factor alpha (TNF-α), IL-6, IL-18 and interferon gamma (IFN-γ) as well as produce nitric oxide, reactive oxygen species and many others associated with a neurodegenerative phenotype [105]. Microglia can also take on a state that is believed to promote tissue remodeling and repair through release of anti-inflammatory cytokines like IL-4, transforming growth factor beta (TGF- β), YM1, arginase 1, and IL-10. Both phenotypes have been reported in disease states making their roles in neurodegeneration unclear [105]. For instance, one study found significantly higher levels of IL-1β and TGF- β in the frontal cortex of AD patient brains corresponding with the presence of ramified, activated microglia and increased levels in TNF-α and IL-6. Conversely in PSP, only IL-1β was significantly increased in the substantia nigra and subthalamic nucleus [92, 106]. This highlights several inherent differences between AD and other tauopathies. First, the spatial distribution of neuroinflammation is dependent on the deposition of protein aggregates specific to each disease [107]. Second, increases in TGF- β and other cytokines associated with remodeling and repair are often reported in response to plaque deposition and are hypothesized to be protective against Aβ pathology [108, 109]. This indicates neuroinflammation may arise differently in AD due to Aβ deposition as opposed to pure tauopathies like PSP and CBD where only tau deposits are seen. In contrast, expression of IL-1β, TNF-α and IL-6 all feed into a cascade that leads to increases in tau hyperphosphorylation, reduction in synapse markers, and neuronal loss [110]. Limited data is available regarding cytokine transcript levels in human tissue of other rarer tauopathies, however neuropathologists note morphologically activated glial cells routinely accompany tau deposition [74]. It is possible that both inflammatory and repair associated microglia coexist in disease states as they try to combat the accumulation of misfolded proteins while also attempting to counteract neuroinflammation. Meanwhile, experimental evidence does demonstrate that dysregulation of proinflammatory molecules is detrimental for tau pathology.

Recently, studies have begun using PET to further examine neuroinflammation in the neurodegenerative process. The most popular tracers investigated to date bind to translocator protein (TPSO) which is expressed on activated microglia, astrocytes, and other infiltrating immune cells in the brain. TPSO signal has been shown to increase with microglia activation in tauopathies including AD, PSP, PiD, and FTDP-17 [111–113] as well as several other neurodegenerative diseases and injury models such as other frontotemporal dementias (FTDs), Parkinson’s disease, stroke and TBI [114]. Interestingly, despite many reports characterizing microglial activation in response to Aβ plaque deposition, Maeda et al. observed stronger correlation of TPSO tracer ¹¹C–AC-5216 binding in response to NFTs and NTs in AD and non-AD tauopathy brains. These findings were further investigated in PS19 human-tau transgenic mice, which carry a P301S mutation in the human MAPT gene that is causative for FTDP-17 [115]. The authors observed that radiotracer accumulation in the hippocampus and entorhinal cortex strongly correlated to tau deposition. Importantly, TPSO signaling preceded thioflavin-S-positive tangles and MRI-measureable regional atrophy in the mice implying TPSO PET may be more adept at detecting neuroinflammation induced by tau aggregation [111]. Other PET tracers are also being further developed for imaging of neuroinflammation. Future studies are needed to understand how signal changes spatially and temporally relate to pathology progression in human AD and other primary tauopathies. However, these data do support that tau-induced gliosis is an intrinsic process in tauopathies and that further research should aim to elucidate potential causal relationships between microglia activation, tauopathy progression, and neurodegeneration.

Microglia may contribute to tau spreading

Many groups have reported tau spreading between synaptically connected neuronal populations [43, 44, 64, 69–71], however lateral propagation of tau aggregates has also been described suggesting that multiple mechanisms of spreading may co-exist. In the rTgECtau mice that selectively express human tau in the entorhinal cortex, tau inclusions were seen to propagate to the dentate gyrus and hippocampus by 18 months of age but cortical neurons lacking tau expression outside of the entorhinal cortex did not have tau deposits. Yet by 24 months, tau aggregates were noted in non-synaptically connected neurons as well as glial cells that surrounded degenerating axon terminals [43]. Similarly, Braak staging has also described tau pathology in unconnected brain regions in AD patients past Braak stage III [41] and glial tau pathology is widely observed across tauopathies [74]. Glial cells are potentially affected by tau pathology as neuronal axons and dendrites degenerate and release toxic, aggregated tau species. Additionally, recent work suggests that microglia play a plausible role in the synaptic and non-synaptic spread of tau pathology.

Microglia readily take up both soluble and insoluble forms of tau [7, 56, 82, 116]. Once engulfed, tau is either degraded [116] or re-released in exocytosing microvesicles called exosomes [7, 55]. Interestingly, increased levels of exosome-associated tau have been found by some groups in the CSF and blood of individuals with AD and FTDs [55, 117]. This led one group to hypothesize that microglia actively contribute to tau propagation by phagocytosing and exocytosing tau protein [7]. They observed a significant reduction in pathologically phosphorylated tau staining by AT8 following pharmacological depletion of microglia in two different tauopathy mouse models. Furthermore, microglial ablation rescued neuronal excitability deficits and resulted in significantly lower levels of proinflammatory cytokines. The authors went on to show that microglia rapidly phagocytose tau and secrete it in exosomes. Inhibition of exosome synthesis reduced tau secretion from microglia in vivo and impeded the development of tau pathology. They concluded that microglia play a significant role in non-synaptic tau propagation and neurotoxicity. While intriguing, this study only examined p-tau species, not fibrillar aggregates, and the number of animals used in each experiment was small. In addition, reducing microgliosis and thereby the levels of proinflammatory cytokines may have also altered the progression of tau pathology independent of exosome synthesis. Further work is needed to truly understand the contribution of microglia-derived exosomes in the spread of tauopathy.

Astrogliosis in tauopathy

Astrocytes are the most abundant cell type in the brain and are instrumental in supporting neuronal health and function. However, astrogliosis can also be a major contributor to chronic neuroinflammation that diminishes neuronal integrity [118]. Many of the proinflammatory cytokines secreted by microglia can also be synthesized and secreted by astrocytes. Furthermore, signals secreted by microglia, such as IL-1β, TNF-α, IL-6, and C1q have been shown to co-activate astrocytes leading to neuronal dysfunction and death [107]. For instance, reduction of astrocyte-derived cytokine S100β was reported in a study that inhibited IL-1R signaling in mice [119]. The authors demonstrated that IL-1β stimulated S100β secretion that activated GSK-3β in neurons to reduce neuronal β-catenin signaling which has been implicated in tau phosphorylation [120]. Yet, blocking IL-1R restored β-catenin levels by inhibiting GSK-3β. This demonstrates that IL-1β has indirect effects on neurons via affecting cross-talk with astrocytes. Interestingly, IL-1β has also been shown to impact inflammatory responses of astrocytes by binding to and stabilizing IL-6 and COX-2 mRNA, which was shown to be dependent on PKC kinase [121]. Recent studies have further demonstrated how inflammatory signaling can regulate toxic gain-of-function and loss-of-function in astrocytes. These studies illustrated activated microglia induce what was termed an “A1” astrocytic subtype by secretion of IL-1α, TNF, and C1q. A1 astrocytes lost their ability to promote neuronal survival, growth, synaptogenesis, and phagocytosis and were also highly toxic to neurons, though the exact mechanism of toxicity is still unknown [122, 123]. Importantly, this group has demonstrated that A1 astrocytes are upregulated in AD and other neurodegenerative disorders, though additional tauopathies were not tested. This study highlighted the array of consequences that microglial-derived cytokines can have on astrocyte gene expression and function which can ultimately impact neuronal integrity.

As previously described in this review, astrocytic tau pathology is common across many tauopathies. Why and how tau accumulates in astrocytes however is still not understood and very little research has attempted to investigate these questions. Tau accumulation in astrocytes has been reported in some murine tauopathy models [124, 125], however there has been no characterization beyond the initial lesions. One study did attempt to investigate consequences of astrocytic tauopathy in mice by placing the wild-type and a mutant P301L human tau transgene under control of an astrocyte specific promoter [126]. These mice, particularly those expressing P301L tau, displayed age-dependent phosphorylation, fibrillization and asymmetric accumulation of tau in astrocytes beginning at 12 months and progressing up to 24 months of age. The astrocytic pathology in the mice resembled both tufted astrocytes, characteristic of PSP, and astrocytic plaques that are more commonly seen in CBD. In addition to tau deposition, astrocytes displayed a redistribution of the cytoskeletal protein GFAP characteristic to human neuropathology in tauopathies [79]. Tau accumulation and GFAP displacement was followed by induction of low molecular weight heat shock proteins as well as mild disruption of the blood brain barrier evidenced by higher levels of immunoglobulin (IgG) and albumin in brain regions with robust pathology. Finally, neurofilament staining revealed ballooned neurons and axonal degeneration in areas with abundant tau pathology [126]. Likewise, expressing tau in either glia or neurons in a Drosophila model was neurotoxic and co-expression in both cell types synergistically enhanced cell death [74, 127]. While these are admittedly artificial models of astrocytic tauopathy, they demonstrate that tau accumulation in astrocytes is sufficient to cause neuronal degeneration.

Secreted proinflammatory factors

Neuroinflammation in tauopathies: Cause or effect?

Gliosis and neuroinflammation are prevalent in tauopathy patient brains [4, 5, 106] and recapitulated across many animal models [127, 170–174]. Furthermore, microgliosis, astrogliosis and inflammatory markers like IL-β, TNF-α and IL-6 have been shown increase in response to tau pathology [171]. However, it is still a matter of debate whether aberrant neuroinflammation causes tau pathophysiology or if glial cells respond first to tau toxicity. Yoshiyama et al. has provided the most compelling evidence to date that microgliosis can precede tau tangle formation and is capable of driving neurodegeneration. In their initial paper describing the PS19 tau transgenic mouse, the authors were surprised by the striking increase in CD11b immunoreactivity in 3-month-old animals, prior to the accumulation of tau deposits [115]. Additional radiograms utilizing [³H]DAA1106 clearly demonstrated an age-dependent microglial-activation in the hippocampus, amygdala and entorhinal cortex. Moreover, CA3 neurons in the hippocampus of 4-month-old mice were immunoreactive for IL-β and COX-2. To further test the hypothesis that microgliosis was capable of driving tauopathy, the immunosuppressant FK506 was given to the mice beginning at 2 months of age. Not only did treatment significantly reduce tau pathology and brain atrophy, it dramatically increased the life-span of the mice. While these data offer a mechanistic link between aberrant microglial activity and tauopathy progression, more recent studies have revealed earlier forms of tau aggregation in PS19 tau transgenic mice using a cellular FRET-based biosensor assay that utilizes recombinant repeat domain tau (RD-tau) fused to either yellow or cyan fluorescent protein. In the presence of tau seeds, the RD-tau aggregates and FRET signal can be measured by flow cytometry [66]. This assay has led to new insights regarding tau toxicity and disease progression. For instance, it has revealed tau seeding activity in PS19 mice as early as 1.5–2 months of age [66]. Therefore, it is possible that tau seeds invoke early microglial activation, which in turn accelerate tau pathology and neurodegeneration. It also raises interesting questions regarding microglial activation and tau seeding, especially considering the dramatic effects of FK506 treatment. Do activated microglia physically contribute to tau seeding or spread? Do proinflammatory molecules activate pathways that encourage tauopathy development? Is gliosis required for neurodegeneration in tauopathies or does it exacerbate it? Future studies should investigate the link between microgliosis, neuroinflammation and tau seeding as well as consider the possible effects of tau strains which have been shown to have different degrees of seeding activity and provoke unique microglial phenotypes [175].

Implications for therapies

Despite the significant clinical and economic burden tauopathies place on society, there are currently no treatments capable of curing or even slowing disease progression. The pursuit for tau-based therapies has rapidly expanded over the past ten years and today drug discovery efforts are fervently ongoing. Drug development is currently investigating tau immunotherapies, small molecule inhibitors, and microtubule stabilizers [176–179]. There have been many preclinical studies published in these areas and some agents have just started progressing through clinical trials.

While drug discovery is an active area of research in the dementia field, it is also important to have a full understanding of the mechanisms underlying disease. Initial immunization studies for Aβ were halted due to severe neuroinflammatory adverse events, some of which resulted in death. Additionally, many of the Aβ monoclonal antibodies have led to side effects, such as microhemorrages and brain edema, which are attributed to microglia-induced damaged to cerebral vasculature as they respond to antibodies coating cerebral amyloid angiopathy [180]. Therefore, careful consideration of the neuro-immune system should be taken in approaching tau immunotherapy. Some studies have reported that microglia can mediate tau clearance and this is enhanced with tau monoclonal antibody treatment [116, 181]. Antibodies typically stimulate antigen phagocytosis in an Fc-dependent (fragment crystallizable domain) manner. However, binding of Fc receptors is also known to activate microglia and stimulate release of proinflammatory molecules which can exacerbate the degenerative process [182]. Alternatively, it has been shown that antibodies that block tau seeding activity can prevent the formation of tangles and brain atrophy in mice [183]. It is possible that simply binding extracellular tau may prevent propagation of pathological tau seeds and slow disease progression. This type of mechanism would not require Fc effector function for tau antibodies. A recent study reported that an Fc-effectorless tau antibody reduced p-tau in vivo akin to the full-effector antibody with the same epitope specificity and affinity [184]. In addition, the effectorless tau antibody did not stimulate microglial release of IL-1β, TNF-α or IL-6 in vitro, which in the case of the full Fc effector containing antibody led to neuronal toxicity. A recent paper also showed that single chain fragment variables (scFvs) derived from an anti-tau antibody decreased p-tau accumulation in the brain of PS19 tau transgenic mice indicating that microglial activation via the Fc domain of an antibody is not required for the protective effect of such a treatment [185]. Additional work is needed to stringently test if Fc-effectorless tau antibodies can prevent formation of aggregated tau and neurodegeneration in vivo as well as characterize the effects of inflammation stimulated by tau immunotherapies.

There is also renewed interest in targeting inflammatory pathways since the discovery of TREM2-mediated risk for AD. In the past, clinical trials with various NSAIDs and glucocortoroids failed to rescue cognitive deficits in AD patients or prevent disease progression despite promising data from preclinical animal studies [186]. However, it is possible that more targeted therapies or starting earlier in the disease process will have positive effects. Current data indicates that inflammation is initially stimulated by Aβ in AD and that chronic gliosis influences tau pathogenesis. If this is true, it is possible that targeted therapies that interrupt neuroinflammation may be utilized after Aβ accumulation begins to delay or prevent tauopathy in AD. In the context of primary tauopathies, targeting specific proinflammatory molecules or pathways may alter the progression of disease and symptoms.