Spilling the T: T cells in tauopathy mechanisms, disease progression, and therapeutic horizons

Abstract

Primary and secondary tauopathies are major causes of dementia characterized by the abnormal accumulation of microtubule-associated protein tau. Tau aggregation initiates neuronal dysfunction and death, while despite the distinct and pathological features of individual tauopathies, accumulating evidence shows that these disorders converge on shared mechanistic pathways marked by chronic inflammation and progressive neuronal loss. The innate immune response has been extensively studied in this context, yet the contribution of the adaptive immune system, particularly T cells, has only recently gained attention. In this review, we examine emerging evidence for adaptive immune involvement in tauopathies, including mechanisms that drive T cell activation and infiltration into the central nervous system, their effector functions once within the parenchyma, and how these responses interact with innate immune signaling. We also highlight potential therapeutic strategies aimed at modulating adaptive immunity in primary and secondary tauopathies. Current findings suggest that tau accumulation and glial-driven neuroinflammation creates an environment that promotes the activation and recruitment of effector T cells into the tauopathy brain. Importantly, blocking or reshaping the T cell responses shows promise for mitigating neurodegeneration, highlighting the adaptive immune system as a potential therapeutic target.

Introduction

Tauopathies are a heterogeneous group of neurodegenerative diseases that are unified by the abnormal accumulation of microtubule-associated protein tau (MAPT)[1, 2]. Tau protein localizes primarily in neuronal axons, where it binds tubulin and promotes microtubule assembly and stability [3–5]. Beyond this structural role, tau has also been implicated in myelination, synaptic plasticity, and neurogenesis [2, 6, 7]. Tauopathies include the most prevalent neurodegenerative disorder, Alzheimer’s disease (AD), and less prevalent forms including frontotemporal lobar degeneration- tau (FTLD-tau), progressive supranuclear palsy (PSP), and Pick’s disease (PiD). Each disease is clinically, phenotypically, and morphologically distinct, yet recent studies have revealed converging mechanistic pathways that culminate in neurodegeneration and neuroinflammation. This shared convergence suggests immune involvement may be a unifying factor across distinct tauopathies. At present, disease modifying treatments remain limited, though emerging therapies in AD show early indications of slowing progression in select patient populations [8, 9]. While tau pathology is a well- recognized driver of neurodegeneration and subsequent neuroinflammation, less attention has been given to the role of adaptive immunity in shaping disease progression. Increasing evidence now points to a role for the adaptive immune system in neurodegeneration, raising key questions about the mechanisms and drivers of this process.

In this review, we summarize the current literature on the impact the adaptive immune system plays in tau-dependent neurodegeneration with a focus on effector T cells (Fig 1). Specifically, we will: 1) briefly describe the tau protein and neuropathological features of primary and secondary tauopathies; 2) outline potential mechanisms of tau-induced neurodegeneration and neuroinflammation; 3) discuss adaptive immune responses, specifically T cells, in tauopathy; 4) highlight potential therapeutic targets and approaches. Our understanding of adaptive immunity in tauopathy and neurodegeneration is rapidly evolving and may have important translational implications.

Fig. 1.

In tauopathies, tau hyperphosphorylation promotes aggregation in part by inhibiting tau degradation. Activated microglia and astrocytes release pro-inflammatory cytokines (IL-1β,TNF-α, IFN-αβ) and chemokines (CCL4,CXCL9, and CXCL10) generating signals that recruit effector T cells across a compromised BBB. Once within the parenchyma, infiltrating T cells secrete cytolytic granules and proinflammatory cytokines such as TNF-α, IFN-γ, IL-17, and GzmK. This further amplifies neuroinflammation and contributes to a neurotoxic milieu

Tau protein: Physiological and pathological characteristics

Tau is a highly conserved protein encoded by the MAPT gene, which is located on chromosome 17 in humans and on chromosome 11 in mice. In the human brain, six major isoforms are expressed, all spliced from the single MAPT gene [10]. Alternative splicing of exons 2 and 3 generates N-terminal inserts that produce isoforms with 0, 1, or 2 inserts, denoted as 0 N, 1 N, or 2 N [1, 7, 10]. Alternative splicing of exon 10 produces tau with either three (3R; lacking exon 10) or four (4R; exon 10 present) repeat domains. Isoform expression is developmentally regulated. In the fetal human brain, only 0N3R isoform is present but in the adult human brain all six isoforms are expressed at an approximate 1:1 ratio of 3R to 4R tau [11]. Mice also express 3R tau during neuronal development but adult mice almost exclusively express 4R tau in the brain [12]. In humans, shifts in the 3R:4R ratio are implicated as a potential driver of tauopathy. Each of these isoforms have specific anatomical locations and biochemical characteristics that may contribute to 3R and 4R disease specific molecular conformations [13, 14]. Mass spectrometry studies have shown that 4R isoforms significantly increase in 4R and 3R/4R tauopathies such as PSP, corticobasal degeneration (CBD), and AD while 4R isoform amount did not change in 3R tauopathy PiD or non-tauopathy FTLD-TDP [15].

It is important to note that there are significant differences between human and mouse tau, which must be considered when interpreting findings from experimental tauopathy models. The mouse brain only expresses 4 tau isoforms, and 3R tau is restricted to the fetal stage and absent in adult tissue. This absence complicates modeling primary tauopathies, such as Pick’s disease, that are driven by 3R tau accumulation. Consequently, many experimental systems rely on knock-in or transgenic expression of human tau, often carrying disease-associated mutations. Human tau contains 11 additional amino acids in the N-terminal region compared to mouse tau [12]. While the proteins are highly homologous across the central and C-terminal domains, these N-terminal differences may affect interactions with binding partners and antibody recognition [7].

Tau is subject to several post-translational modifications (phosphorylation, acetylation, ubiquitination, etc.) that alter its solubility and microtubule binding affinity and can promote pathological protein aggregation[5, 7, 16]. One of the notable hallmarks of tauopathy is hyperphosphorylation of serine and threonine residues within the proline-rich and C-terminal domains [17–19]. Tau hyperphosphorylation can lead to tau aggregation through inhibiting tau degradation, driving mis-localization of tau from axons to somatodendritic compartments, or favoring the misfolding of tau into filaments [7, 20, 21].

Tauopathies can be broadly divided into two distinct categories: primary, in which tau pathology is the central driver of disease, and secondary, in which tau aggregation occurs alongside or downstream of another proteinopathy that is considered the central driver of disease. Primary tauopathies include PSP, PiD, or corticobasal degeneration (CBD). Primary tauopathies that fall under the FTD umbrella are collectively referred to as FTLD-tau [22]. Secondary tauopathies include Alzheimer’s disease (AD), where tau pathology develops subsequent to amyloid-β (Aβ) deposition [22]. Dementia is a major global heath burden, with the World Health Organization estimating 57 million cases worldwide and 10 million new cases annually, projected to reach 152 million by 2050 [23, 24]. Despite overlapping features, each disease has a distinct anatomical distribution of tau pathology, which aligns with its clinical presentation. Importantly, all share a common outcome: tau aggregation leads to neuronal dysfunction, neurodegeneration, and neuroinflammation. We refer the reader to several comprehensive reviews of neuropathology of tauopathies [2, 22, 25, 26] and the differing tau structures of each tauopathy [27]. Box 1.

Pathological characteristics of primary and secondary tauopathies. References: [22, 28–32]

Innate immunity in tauopathy

Neuroinflammation is a consistent and defining feature across primary and secondary tauopathies. Microglia and astrocytes, the resident innate immune cells of the brain, respond to tau pathology and associated central nervous system (CNS) damage through multiple mechanisms: releasing pro or anti-inflammatory mediators, phagocytosing debris or injured cells, or modulating neuronal activity [33–38]. Reactive microglia and astrocytes are found in brain regions that develop tau pathology, and their responses may contribute both to the clearance of pathological tau and to the release of pro-inflammatory mediators that exacerbate neuronal injury [34–36, 39, 40]. These glial-driven innate immune responses may promote the engagement of adaptive immunity, including infiltration of peripheral T cells [41]. Elevated levels of stress cytokines and chemokines, such as IL-1β and TNF-α, can inhibit neuronal repair, increase oxidative stress, and promote mitochondrial dysfunction contributing to neuronal death [39, 42, 43]. Prolonged release of these cytokines also contributes to the breakdown of the blood brain barrier (BBB) [41]. Cytokines and chemokines such as, CCL4 or CXCL10, which attract T cells, are elevated in AD and FTLD-tau [35]. At the same time, border- associated macrophages and T cells monitor the brain vasculature and meningeal compartments [44, 45]. Perivascular, meningeal, and choroid plexus macrophages are strategically localized at the interface between parenchyma and circulation, allowing immune surveillance and signaling across the entire CNS [38, 44–46].

Evidence from both human and mouse models demonstrate that regions of microglia activation and neuroinflammation often precede and predict areas vulnerable to future neurodegeneration. In PSP, for example, positron emission tomography (PET) imaging using the tracers ¹¹C-PK11195 (for microglia activation) and ¹⁸F-AV1451 (for tau burden) revealed widespread co-localization of microglial activation and tau aggregation [34]. This overlap is the most prominent in the insula, temporoparietal junction, brain stem, and cerebellum, and these imaging findings correlate with disease severity and mirror patterns observed in post-mortem FTLD-tau tissue [34]. Experimental evidence supports a causal link between microglial activation and tau pathology. Adoptive transfer of CX3CR1^−/− activated microglial into wild type mice induced endogenous tau hyperphosphorylation, demonstrating that activated microglia can promote pathological tau modifications [33]. Conversely, in tauopathy models, such as P301S mice expressing human APOE4, which develop neurodegeneration, pharmacological depletion of microglia using CSF1R inhibitor significantly reduced brain atrophy, accumulation of hyperphosphorylated tau, and prevented T cell infiltration[47, 48]. Together, these findings establish microglia as key drivers of tauopathy-associated neurodegeneration and as central mediators linking innate immune activation to downstream adaptive immune responses, including T cell recruitment into the diseased brain .

In addition to microglia, reactive astrocytes are also evident in tauopathy. Some forms of tauopathy feature tau aggregation within astrocytes. In PSP, tau inclusions accumulate within astrocytes, which adopt a tufted morphology, most prominently in the motor cortex and corpus striatum [28]. In CBD, hyperphosphorylated tau aggregates within astrocytic processes to form astrocytic plaques [22, 25, 49]. Comparative analyses of preclinical CBD and end-stage CBD brains show that astrocytic plaques are the earliest and most prominent lesion type, preceding overt neuronal tau aggregation [50]. Astrocytic expressions of inflammatory mediators are an important feature of both primary tauopathy and AD. Early astrocyte activation may contribute to BBB disruption and the release of inflammatory cytokines such as CCL2, CCL5, and CXCL10, all of which are elevated in the CSF, plasma, and post-mortem brain tissue of individuals with FTLD-tau and AD [51, 52]. Upregulation of these cytokines could establish a chemotactic gradient that promotes T cell migration across the BBB [53]. Specifically, CCL2 is known to recruit T cells via CCR2 signaling, facilitating transmigration from the circulation into the brain [54]. CCL5 binds to CCR5, enhancing adhesion and chemotaxis of activated and memory T cells, while CXCL10 binds to CXCR3 on effector T cells, increasing their migratory efficiency [55, 56]. Collectively, these mechanisms position astrocytes as key regulators that prime the CNS microenvironment for adaptive immune cell infiltration in tauopathies.

Adaptive immune system and the CNS

The adaptive immune system is composed of two principal cell types, T and B cells, that recognize discrete antigens and generate long-lived protection against future encounters. B cells mediate humoral immunity through antibody production and can directly recognize intact antigens via the B cell receptor (BCR) [57, 58]. BCR antigen engagement initiates a signaling cascade that drives B cell activation and differentiation. In addition to antibody secretion, B cells function in conjunction with T cells to shape immune responses. Although generally less efficient than antigen presenting cells (APCs), B cells can present antigen to CD8⁺ and CD4⁺ T cells. This process has been shown to be highly efficient when the antigen recognized by the BCR is presented to T cells with shared antigen specificity. B cells express co-stimulatory molecules that promote pro-inflammatory T cell activation and secrete pro and anti-inflammatory cytokines [59–61]. B cells have been identified within the meninges, choroid plexus, and perivascular spaces in patients with AD [62, 63]. However, it remains unclear whether these observations reflect disease-specific mechanisms or secondary responses to neuroinflammation, and current data supporting a pathogenic role of B cells remains limited. Further investigations into B cell function in neurodegenerative disease is warranted given their integral role in adaptive immunity and potential relevance to tauopathies. While B cells constitute a major arm of the adaptive immune system, the remainder of this review will focus on T cells and their effector functions. Readers are referred to comprehensive reviews for detailed discussions of B cell development, function, and potential roles in tauopathies [58, 64].

T cells mediate cellular immunity and recognize peptide antigens presented on major histocompatibility complex (MHC) molecules on antigen-presenting cells [65, 66]. There are two main subsets of conventional T cells: CD4⁺ helper T cells and CD8⁺ cytotoxic T cells. Each expresses a unique T cell receptor (TCR) composed of α and β glycoprotein chains, though a minor population express γ and δ chains [65]. T cells undergo rigorous developmental selection within the thymus to ensure self-tolerance that seeks to eliminate self-reactive conventional T cells and promote the development of self-reactive immunosuppressive regulatory T cells (Tregs). The result is a mature repertoire of T cells that are both self-restricted and self-tolerant yet capable of recognizing foreign antigens[65, 67, 68].

Once mature, naïve T cells circulate through peripheral lymphoid organs awaiting antigen encounter. Antigen recognition and T cell activation occur when dendritic cells (DCs), after detecting pathogens via pathogen recognition receptors (PRRs), migrate to draining lymph nodes carrying processed antigen. Within the lymph node, DCs present antigenic peptides to naïve T cells via MHC molecules (MHCI to CD8⁺ T cells and MHCII to CD4⁺ T cells) and secrete chemokines that attract and position T cells for engagement. T cell activation requires three signals. First, TCR recognition of the peptide-MHC complex. Second, co-stimulatory interactions between T cells and APCs, such as CD40 ligand on the T cell binding to CD40 receptor on the APC. Third, inflammatory cytokines, such as IL-12 or Type I IFNs, promote optimal T cell expansion. These signals initiate intracellular signaling cascades and metabolic reprogramming, leading to clonal expansion and differentiation into effector T cells [69–71]. Activated CD4⁺ helper T cells secrete cytokines including IL-2 and IL-12 that promote CD8⁺ T cell proliferation, survival, and transcriptional programming [65]. Cytotoxic CD8⁺ T cells induce apoptosis in target cells through death receptor ligands and the release of cytolytic granules containing perforin and granzymes [72]. Effector T cells subsequently migrate along chemokine gradients to sites of infection or inflammation, where they execute pathogen clearance, or in pathological contexts, contribute to tissue injury. Multiple layers of regulation exist to maintain immune balance and prevent autoimmunity. Regulatory T cells (Tregs), inhibitory receptor signaling (e.g. PD-1, CTLA-4), and cytokine-mediated feedback loops serve to restrain excessive activation and maintain tolerance [68, 73, 74]. Among these mechanisms, Tregs are particularly critical, as the primary function of this cell population is to sustain immune homeostasis and balance immune responses. CD4⁺ Tregs express TCRs that recognize antigen presented on MHCII molecules and are characterized by high expression of FOXP3 and CD25 [75]. Tregs suppress immune activation through multiple mechanisms. Firstly, Tregs engage APCs, limiting physical access of effector T cells to antigen-MHC complexes. In addition, Tregs express the inhibitory receptor CTLA-4, which binds CD80/CD86 on APCs with high affinity, thereby reducing co-stimulatory signaling required for T cell activation [76]. Activated Tregs also secrete anti-inflammatory cytokines including IL-10 and TGFβ to alter the local inflammatory milieu [77]. Finally, Tregs express a high-affinity IL-2 receptor subunit (IL2RA or CD25) that limits accessibility of IL2 for CD8 + T cells that express a lower affinity IL-2 receptor [78]. Collectively, these suppressive mechanisms highlight the therapeutic potential of targeting T cell regulatory pathways. When regulatory processes fail, uncontrolled adaptive immune activation can promote autoimmune pathology and exacerbate neuroinflammatory conditions, including those observed in tauopathies [25, 73].

The CNS is now recognized as an active participant in systemic immune surveillance rather than an isolated immune-privileged organ. T cells routinely patrol CNS interfaces, including the meninges, perivascular space, and choroid plexus, where they interact with local antigen-presenting and stromal cells [79]. Meningeal lymphatic vessels absorb cerebrospinal fluid (CSF) from the adjacent subarachnoid space via arachnoid cuff exit points and drain into the deep cervical lymph nodes (dCLNs) where CNS-derived antigens are presented to the adaptive immune system [80–82]. This coordinated drainage network ensures that CNS antigens are sampled by peripheral immune cells, allowing for continuous immune education and rapid detection of infection or injury. CNS antigens that drain through the lymphatic pathways can reach the dCLNs and potentially activate CNS-specific T cells. Evidence of immune crosstalk between the CNS and periphery is apparent from assessment of cytokines in different tauopathies. Serum cytokine profiling from patients with FTLD, PSP, and CBD showed significant elevations in pro-inflammatory cytokines TNF-α, TNF-R1, M-CSF, IL-17, IL-2, IP10 and IL-6 compared to healthy controls. When combined with PET imaging for microglia activation, individuals with high pro-inflammatory cytokine scores exhibited greater neuroinflammation in the frontal lobe and brainstem and lower survival rates [83]. These inflammatory signatures were also present in the CSF from FTLD-tau and FTLD-TDP43 cases, providing disease-specific profiles that could distinguish these disorders from AD [84]. Among these cytokines, IL-17, which is produced primarily by TH17 CD8⁺ T cells, and γδ T cells, was consistently elevated and known to amplify local inflammation and recruit neutrophils. These analytes were predictive and could determine disease progression [84]. Meningeal γδ T cells have recently been found to have a major function within the CNS, regulating short term memory formation and behavior via release of IL-17 [85, 86]. Neutralizing IL-17 at later disease stages in the 3xTg mouse, which develop amyloid-β plaques and tau accumulation, delayed short-term memory deficits [87].

Furthermore, proteomic analysis of CSF from pathologically confirmed PSP, non-PSP FTLD, and healthy controls identified CTLA-4 levels inversely correlated with PSP rating scale (PSPRS) scores, suggesting that stronger T cell inhibitory signaling may confer milder disease severity [88]. CTLA-4 is a negative regulator of T cell activation that binds to CD80/CD86 to block co-stimulation, and higher levels of CTLA-4 expression may reflect a compensatory anti-inflammatory response [88]. Importantly, distinct innate immune shifts are observed in the peripheral blood of AD, PSP, CBD, and FTD patients, characterized by pro-inflammatory immune profiles compared to healthy controls [89]. These profiles displayed a transition towards pro-inflammatory monocyte activation and myeloid phenotypes [89]. Together, these observations support the view that adaptive immune activation in tauopathies is not a primary trigger but rather a secondary response to chronic innate immune dysregulation and tau-driven neuroinflammation.

Human studies demonstrate that T cells are present in the CNS under homeostatic conditions. CSF from healthy individuals contains predominantly CD4⁺ memory-like T cells (CD4⁺CD45RA⁻CD27⁺CD69⁺) measured by flow cytometry, which are thought to monitor the CNS without penetrating the parenchyma [90, 91]. Histochemical analyses of non-inflamed post-mortem brain tissue have identified CD8⁺ lymphocytes localized to the perivascular space between blood vessels and the glia limitans [92, 93]. In a comparative study analyzing freshly isolated T cells from peripheral blood and the corpus callosum, (CC) immediately after death, a local population of differentiated CD8⁺ and to a lesser extent CD4⁺ effector T cells (CD27⁻CD45RA^±) was detected with very few naïve CD27⁺ T cells [93]. Despite expressing effector-associated markers, these T cells showed limited cytolytic enzyme expression. Both CD4⁺ and CD8⁺ T cells expressed CD127, supporting IL-7 driven homeostatic maintenance, and few co-expressed CD38 or HLA-DR, indicating minimal recent antigenic stimulation. These T cells also expressed CXCR1 and CXCR3 chemokine receptors, as well as, tissue residency markers CD69 and CD103, while lacking CCR7 and showing downregulation of S1PR1, consistent with long term CNS residency and restricted recirculation [92, 93]. Collectively, these characteristics suggest that homeostatic T cells contribute to barrier maintenance, cytokine balance, and rapid immune readiness within the CNS without provoking damaging inflammation.

Several studies have described increased T cell abundance in the brains of people with primary and secondary tauopathy [94–98]. Increased CD4⁺ and CD8 T⁺ cells were described in individuals with FTLD-tau, particularly those with PiD. Notably, the abundance of CD8⁺ T cells was positively correlated with p-tau staining in FTLD-tau, suggesting a possible link between tauopathy and T cells [94]. Increased CD8⁺ T cells were also noted in the substantia nigra of people with PSP compared to people with PD or without neurological disease [98]. CD8⁺ T cells were evident in individuals with chronic traumatic encephalopathy (CTE) in proximity to p-tau and microglia [99, 100]. Increases in brain and leptomeningeal localized CD8⁺ T cells have been described in people with AD [101–103]. Notably the abundance of parenchymal T cells appears to increase with increased Braak stage and p-tau, similar to the positive correlation of CD8⁺ T cells and p-tau described in FTLD-tau [48, 94]. In advanced AD, hippocampal tissue contains the highest numbers of CD3⁺ T cells, 90% of which are CD8⁺ T cells, and extravascular T cells correlate with tau pathology, but not Aβ burden, reinforcing tau as a primary driver of T cell infiltration [97]. Similar to observations in humans, several mouse models of tauopathy exhibit increased T cell abundance in the CNS. Increased T cells were reported in P301L-tau/PS2APP mice, Thy-tau22 mice, PS19 mice which express human P301S-tau, and 3xTg mice, particularly in brain regions that had high levels of p-tau [48, 104, 105]. It is important to note that the presence of T cells in postmortem tissue does not directly establish causality; however, when considered alongside consistent peripheral immune alterations observed across multiple tauopathies, these findings suggest effector T cells may functionally contribute to progressing neuronal damage and promoting neuroinflammatory pathways. While emerging spatial transcriptomics analyses of AD and tauopathy brains have largely focused on innate immune changes, specifically microglia activation, these datasets have revealed upregulation in T cell associated signaling pathways, indicating the utility for spatially resolved approaches to assess T cell activation states within diseased brains [106–108].

Sequencing and flow cytometry analysis of T cells in the CSF and blood have also identified disease-associated changes in T cells. The majority of immunocytes in the CSF are T cells and in contrast to the brain parenchyma there are more CD4⁺ than CD8⁺ T cells. One recent study found that CD4⁺ T cells in the CSF and from peripheral blood mononuclear cells (PBMCs) from individuals with non-autosomal dominant, early-onset AD exhibited increased interferon-signaling associated gene expression (101, 109). Mass spectrometry analysis of peripheral blood from AD patients revealed a distinct immune signature characterized by increased CD8⁺ T effector memory cells expressing CD45RA⁺T_EMRA[103]. Notably, higher T_EMRA frequencies were negatively correlated with cognitive performance, suggesting a link between peripheral T cell activation and cognitive decline. CD8⁺ T cells in the CSF largely consist of effector memory (TEM) cells, including CD45RA⁺ T_EMRA cells [101, 110]. CD8⁺, and to a lesser extent CD4⁺, T cells in the CSF exhibit increased clonal expansion in people with AD compared to healthy controls, potentially indicative of an antigen-driven immune response. Clonally expanded CD8⁺ T cells in the CSF of individuals with AD exhibited increased expression of particular granzymes (GZMA, GZMH, and GZMK) and NKG7 suggesting increased cytotoxic function. Clonally expanded CD8⁺ T cells in elderly individuals with cognitive impairment also exhibited increased expression of CXCR6, a homing receptor that can mediate T cell entry into the brain [109]. Notably, non-classical monocyte expression of CXCL16, the ligand for CXCR6, is also increased in the CSF of individuals with cognitive impairment [109].

The cognate antigen(s) that could potentially lead to clonal expansion of T cells in human tauopathy and tauopathy mouse models are unclear. Sequence analysis suggested that clonal T cells in the CSF of human AD patients may recognize Epstein-Barr virus antigens, although the relevance of this observation is not certain [103]. Other studies have identified autoreactive T cells against tau peptides in human PBMCs. Recent studies examining the HLA peptidome, the repertoire of peptides presented on MHC molecules, provide new insights [111, 112]. Specific HLA class II alleles (HLA-DRB104 and HLA-DQB1802) bind strongly to post-translationally modified tau peptides, particularly acetylated PHF6 sequences, and are associated with reduced neurofibrillary tangle burden and slower clinical progression [113, 114]. These alleles may facilitate more efficient presentation of tau-derived antigens, promoting deletion of autoreactive T cells or induction of regulatory T cell responses that temper neuroinflammation.

Although several studies have found increased T cells in the CNS associated with tauopathy, the functional role of T cells in tauopathy and neurodegeneration is still unclear. Immunodepletion of T cells using an anti-CD3 antibody protected against memory deficits in Thy-tau22 mice [104]. Similarly, combined anti-CD4 and anti-CD8 antibody treatment conferred protection against memory deficits, reduced brain atrophy, and reduced neuroinflammation in P301S mice that expressed human APOE4 [48]. Conversely, genetic knockout of CD8⁺ T cells increased p-tau accumulation, reactive astrocytes, highly activated microglia, and neuronal injury in P301S mice [115]. These discordant findings provide valuable insights into the context dependent role of effector T cells in tauopathy and emphasize the critical importance of therapeutic timing. Notably, key differences between the mouse models may influence disease outcome. In the case of immunodepleting, the mice were hemizygous for a P301S transgene regulated by the prion protein promoter whereas in the genetic knockout of CD8 the mice expressed two copies of a P301S transgene regulated by a Thy1.2 promoter, raising the possibility that transgene dosage or neuronal expression patterns influenced the degenerative process [115]. While both models exhibit rapid and aggressive pathology, the presence of human APOE4 in the antibody depletion paradigm further exacerbates neuroinflammation and neurodegeneration, potentially amplifying the impact of peripheral immune modulation. A second major distinction is the timing of immune manipulation. In the CD8 germline knock out, CD8⁺ T cells were absent from birth, whereas the combined anti-CD4/anti-CD8 antibody treatment was initiated at 6 months of age, coinciding with the onset of tau pathology. Germline deletion of CD8⁺ T cells is known to induce compensatory immune functions to replace the loss of critical effector mechanisms [116]. In this context, early loss of CD8⁺ T cells was associated with increased tau accumulation, astrocytic and microglia reactivity and NfL levels indicating a potentially protective function for CD8⁺ T cells during early life or presymptomatic tauopathy [115]. In contrast, immunodepleting primarily targets circulating and infiltrating T cells, while preserving tissue resident memory T cells (TRM), which are relatively resistant to systemic antibody depletion. Together, these findings support a working model in which CD8⁺ T cells play a protective role early in life by maintaining immune homeostasis or limiting the spread of tau pathology, whereas infiltrating effector CD8⁺ T cells become detrimental following the establishment of neuroinflammation and neurodegeneration. Accordingly, therapeutic strategies aimed at modulating effector T cell activity at onset of disease, rather than during early immune development, may represent a more effective approach for targeting the adaptive immune system in tauopathies.

Using scRNAseq and multidimensional flow cytometry, a consistent signature of CD8⁺ T cells comprises clonally expanded T cells in the CNS of both mice and humans. These cells express immune checkpoint genes (TIGIT, PD-1, LAG3), transcription factor profiles associated with T cell exhaustion (Tox⁺, Eomes⁺), and high levels of Gzmk [103, 117–119]. This profile is reminiscent of CD8⁺ T cells that increase with aging in mice and humans. Unlike Gzmb, Gzmk does not cleave caspases in target cells to elicit apoptosis, and the functional role of Gzmk in the immune response is still not well understood. Previous studies found that Gzmk increased IFNγ-induced secretion of inflammatory cytokines by fibroblasts and augmented expression of genes in the senescence-associated secretory phenotype, potentially via activation of protease-activated receptor-1 (PAR1) [120, 121]. Gzmk-mediated PAR1 activation may also play a role in T cell diapedesis of endothelial cells and inducing Ca²⁺ dyshomeostasis and tau phosphorylation in neurons [25, 122, 123]. These findings could suggest Gzmk-dependent inflammation and neuronal perturbations could promote tauopathy and neurodegeneration. Another study found punctate Gzmk staining on microglia in P301S mice [115]. Although the functional relevance of Gzmk deposition is not clear, loss of CD8⁺ T cells resulted in increased dysfunctional microglia, suggesting a possible role for Gzmk in modulating microglial function [115]. Finally, two recent studies identified a novel role for Gzmk in facilitating complement activation via either C3 cleavage or cleavage of C4 and C2 to form a C4bC2a convertase [, 124]. Knockout of Gzmk reduced lung inflammation in an ovalbumin airway challenge mouse model and inflammation in the imiquimod-induced dermatitis [125, 126].

In addition to Gzmk, other potential effector functions for how T cells could influence tauopathy and neurodegeneration include secretion of proinflammatory cytokines and direct cytotoxic functions by CD8⁺ T cells. T cells are major producers of IFNγ, which is elevated in brain tissue of P301S mice [41, 48]. IFNγ can induce tau phosphorylation and IFNGRs are upregulated in astrocytes in neurodegenerative diseases [127–130]. Infiltration of T cells into a neuroimmune axis model of AD resulted in increased IFNγ signaling and neuronal toxicity and P301S mice exhibit upregulation of MHC-II molecules in microglia, likely indicative of IFNγ signaling [131]. Administration of an antibody to neutralize IFNγ reduced neurodegeneration, CD11c⁺ microgliosis, and tau phosphorylation in P301S mice expressing APOE3 [48]. Therefore, T cell-derived IFNγ could drive toxic neuroinflammation and tau phosphorylation. Little work has been published on the potential role of direct CD8⁺ T cell killing of cells in tauopathy. However, cytolytic genes such as Gzmb and Prf1 are expressed in a subset of CD8 T cells in the CNS raising the possibility that direct cytotoxic killing could factor into the neurodegenerative process.

Therapeutic horizons for tauopathies

Collectively, the data presented suggest that the adaptive immune system contributes to tauopathy disease progression, positioning it as a potential therapeutic target. Although most efforts have been centered on AD, this is an expanding field with increased interest in other tauopathies. Current literature remains limited and largely derived from preclinical mouse models, but several studies highlight the therapeutic potential of modulation T cell responses.

Experimental depletion of total T cells (CD3⁺) or combination of CD4⁺ and CD8⁺ T cells has demonstrated neuroprotective effects in tauopathy models, including reduced brain atrophy, decreased glial activation, diminished tau aggregation, and improved performance in memory based behavioral assays [104, 132]. Curiously, treatment of P301S mice with fingolimod, an S1PR modulator that sequesters lymphocytes into lymph nodes, increased CD8⁺ T cells in the brain and exacerbated neurodegeneration [133]. This raises the question of whether peripheral T cell infiltration or expansion of tissue resident T cells is responsible for the increased abundance of T cells and associated damage in this model. These findings suggest that broad suppression of T cells may be insufficient and highlight the importance of selectively modulating T cell subsets and functional states. Although antibody-based therapies exhibit limited BBB penetration, accumulating evidence indicates that peripheral immunodepleting to prevent T cell infiltration is more therapeutically beneficial than targeting the relatively sparse T cell population that reside in the brain.

Beyond depletion strategies, efforts have been turned toward steering the adaptive immune response away from pro-inflammatory effector phenotypes and toward regulatory, anti-inflammatory states. [134, 135]. As previously alluded to, targeting regulatory T cells to harness immune suppressive functions has emerged as an active therapeutic strategy for Alzheimer’s Disease. Prior studies have reported significantly reduced number and impaired suppressive capacity of Tregs in AD patients, characterized by reduced capacity of Tregs to suppress pro-inflammatory responses following ex vivo stimulation, suggesting a breakdown in immune regulation during disease progression [136, 137].

Preclinical studies aimed at enhancing Treg activity either through adoptive transfer or cytokine mediated expansion using low dose IL-2 administration, have demonstrated protective effects in the APPPS1 amyloid models as well as the 3xTg, which exhibits both amyloid and tau pathology. In these models, Treg expansion improved cognitive function, reduced pro-inflammatory cytokines, reduced Aβ plaque deposition following adoptive transfer, as well as reduced plaque- associated microglia following IL-2 treatment [136, 137]. Notably, both studies found depletion of Tregs was detrimental leading to exacerbated memory impairment. Recently, a phase II clinical trial (NCT01988506) was completed investigating the efficacy of low dose IL-2 treatment for mild to moderate AD. Low-dose IL-2 preferentially expands Tregs while minimizing the activation of effector T cells or natural killer cells, owing to the constitutively high expression of CD25 (IL-2Rα) on Tregs which forms a high affinity IL-2 receptor capable of responding to low IL-2 concentrations [138, 139]. Treatment was shown to be well tolerated with no adverse side effects. Results showed a significant increase in Treg populations and suppressive functions as well as reduced circulating inflammatory mediators including CCL2, CCL11, and IL-15 and increased levels of IL-4 and CCL13 [140]. Most notably, IL-2 treated patients exhibited stabilization of NfL alongside a trend toward improved cognitive performance from baseline. These findings support the need for disease specific immune and neurodegenerative biomarkers to identify patients most likely to benefit from immune-modulating therapies and monitor target engagement. Similarly, immune checkpoint-based strategies, including PD-1 pathway modulation, have been explored in amyloid-β models [141, 142]. Although checkpoint modulation, including CTLA-4 inhibition, has not yet been extensively tested in tauopathy models, the effects on immune regulation in other disease context warrants research in tau-driven neurodegeneration.

Complementary strategies aimed at expanding Tregs without systemic immune suppression are also under investigation. Another on-going clinical trial aimed at enhancing Tregs, involves intranasal administration of anti-CD3 (NCT06489548). Preclinical studies of autoimmune disease models have shown that mucosal delivery of anti-CD3 promotes Treg expansion and ameliorated disease severity [143, 144]. In the context of AD model 3xTg, intranasal anti-CD3 treatment significantly improves memory function and modulated microglia activation profiles, characterized by reductions in key microglia neurodegenerative phenotypes (MGnD), such as Trem2, Apoe, and Clec7a gene expression (143). This approach utilizes a distinct immune regulatory mechanism. Delivery of anti-CD3 via the nasal mucosa induces antigen-specific immune tolerance through interactions with the mucosal dendritic cells, leading to selective activation and expansion of Tregs without triggering systemic T cell depletion [145, 146]. In contrast, systemic administration of anti-CD3 results in broad immunosuppression through widespread T cell depletion, substantially increasing susceptibility to infection and malignancy, which are particularly pronounced in elderly populations. Together, these studies underscore the premise of immune modulation that are T cell specific rather than broadly suppressing adaptive immune responses. This distinction reflects a broader paradox in tauopathies, wherein specific T cell populations or functional states exert either a protective effect, such as having homestatic roles early in disease, or have a pathogenic effect, such as infiltrating T cells late in the disease to amplify neuroinflammation.

Whether directly targeting T cells will yield clinical benefit in tauopathies remains uncertain; however, the growing evidence linking adaptive immune activation with disease severity provides strong rationale for continued investigation. In parallel, additional clinical efforts have sought to focus on attenuating neuroinflammation more broadly. For instance, Neflamapimod, a p38α kinase inhibitor originally developed for rheumatoid arthritis, suppresses IL-1β and TNF-α production by peripheral immune cells. Neflamapimod has advanced to phase II clinical trials for AD, where it improved episodic memory and preserved synaptic function in a small, randomized study. These effects are hypothesized to be due to inhibition of NF-κB signaling downstream of p38α MAPK leading to reduced microglia activation and improved synaptic function and integrity. Additional investigations have explored the utility of approved anti-inflammatory biologics to reduced inflammation such as TNFα inhibitor etanercept which showed promise in the P301S tauopathy mouse model by reducing microglia activation, attenuated neuronal loss, and reduced tau phosphorylation [147]. However, in a phase II randomized placebo-controlled double blinded trial involving patients with mild to moderate AD (NCT01068353), etanercept treatment failed to significantly improved cognition, behavior or alter inflammatory profiles [148]. These divergent outcomes highlight key limitations in translating preclinical findings to human therapeutic studies.

While mice and humans share many homologous immune mechanisms, there are substantial differences in the innate and adaptive immune systems that complicate direct translation. Mice exhibit higher portions of circulating lymphocytes whereas humans have a greater abundance of neutrophils. Additional distinctions include differences in T cell signaling pathways, such as all mouse T cells express CD28 whereas 80% of human T cells express CD28, and antigen presentation where activated human T cells express MHCII molecule, a function absent in murine T cells [149]. Moreover, disease timescales differ markedly; tauopathy progression in mouse models is genetically accelerated over months, whereas human disease progresses over decades, likely shaping distinct immune responses. Collectively, these limitations emphasize the need for careful interpretation of preclinical immunomodulatory studies and stress the importance of therapeutic strategies that account for disease stage, immune context, and species-specific differences. Taken together, emerging evidence supports modulation of the adaptive immune response as a promising avenue to influence disease progression across tauopathies.

Over the past decade, the concept of neuroinflammation has been recognized as playing an active and dynamic role in tauopathy pathogenesis (Fig 1). Evidence from both human and animal studies converge on a model in which tau aggregation triggers innate immune activation, primarily through microglia and astrocyte activation, which in turn remodels the CNS immune microenvironment. The resulting release of pro-inflammatory cytokines and chemokines disrupts the BBB, and recruitment of adaptive immune cells collectively creates a self-perpetuating cycle of neurodegeneration. Within this immune response pathway, T cells emerge as infiltrating functional participants capable of modulating disease progression through cytotoxic, inflammatory, and regulatory actions. Despite growing histological, transcriptomic, and functional evidence implicating T cells in tauopathy, several critical questions remain unresolved. It is still unclear whether T cell infiltration reflects an antigen-specific and disease-specific response to tau pathology or instead represents a secondary consequence of chronic neuroinflammation. Resolving this distinction will be essential for therapeutic design, as antigen driven responses may require targeted immune modulation, whereas secondary infiltration would necessitate earlier intervention to prevent neuroinflammatory escalation. In addition, although effector T cells are detected within the human CNS in tauopathies, whether these cells exert direct cytotoxic activity remains uncertain. Available evidence suggests a more senescent or exhausted phenotype in human tissues, in contrast to the more cytotoxic and inflammatory T cell response observed in mouse models, underscoring the important species-specific difference that complicates translation. Addressing these questions will require integrated approaches combining single-cell, spatial, and longitudinal immune profiling to delineate the origins, activation states, and functional roles of T cell responses with the CNS and peripheral compartments. Ultimately, therapeutic targeting of the adaptive immune system holds promise for slowing disease progression as well as yielding biomarkers for early detection and treatment monitoring.

Abbreviations

MAPT

Microtubule associated protein tau

AD

Alzheimer’s disease

FTLD-tau

Frontotemporal lobar degeneration-tau

PSP

Progressive supranuclear palsy

PiD

Pick’s disease

CBD

Corticobasal degeneration

CTE

Chronic traumatic encephalopathy

AGD

Argyrophilic grain disease

GGT

Globular glial tauopathy

Aβ

Amyloid-β

PET

Positron emission tomography

CNS

Central nervous system

BBB

Blood brain barrier

BCR

B cell receptor

MHC

Major histocompatibility complex

TCR

T cell receptor

Treg

Regulatory T cells

DCs

Dendritic cells

APC

Antigen presenting cells

PRR

Pathogen recognizing receptor

IFN

Interferon

CSF

Cerebrospinal fluid

dCLNs

Deep cervical lymph nodes

PSPRS

Progressive supranuclear palsy rating scale

TEM

Effector memory T cells

PBMCs

Peripheral blood mononuclear cells

scRNAseq

Single-cell RNA sequencing

TRM

Resident memory T cells

MGnD

Microglia neurodegenerative phenotypes

Authors’ contributions

Authors’ contributions: Both authors contributed equally.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

N/A.

Consent for publication

N/A

Competing interests

The authors declare no competing interests.