In tauopathies, scientists tend to focus on neuronal tau. They track its pathology and blame it for dysfunctional microglia. Now, researchers led by Celeste Karch at Washington University, St. Louis, claim that microglia express their own tau—and that doing so does them in. In a preprint uploaded to medRxiv on May 16, they reported that iPSC-derived microglia with the IVS10+16 tau variant were sluggish at taking up myelin and tau fibrils, and at TREM2 signaling. To boot, they secreted neurotoxic molecules. Microglia isolated from human tauopathy brains also expressed tau, and their transcriptomes and proteomes suggested pathways gone awry, just as in the iMGLs.

  • Microglia isolated from human brain express tau.
  • Induced microglia carrying the IVS10+16 tau variant make little TREM2.
  • They hardly phagocytose tau fibrils and release neurotoxic molecules.

“These data support the idea that MAPT IVS10+16 mutation may alter microglia to be less active and impaired for phagocytosis,” wrote Tsuneya Ikezu of the Mayo Clinic in Jacksonville, Florida (comment below).

Kristine Freude of University of Copenhagen, Denmark, found the evidence compelling. “This … is significant as it corroborates the limited existing studies that report tau in glial cells,” she wrote (comment below).

While highly expressed in neurons, the tau gene tends to lie fallow in other brain cells. In tauopathies, however, tau inclusions form within astrocytes and oligodendrocytes, though these likely arise from phagocytosis of tau released from neurons (Forrest et al., 2022; reviewed by Kovacs et al., 2016; Mothes et al., 2023; Ferrer et al., 2019). Microglia can adopt an inflammatory posture before tangles can even be detected in tauopathies, but again, this has been attributed to neuronal tau (Bevan-Jones et al., 2019; Bolós et al., 2015). 

Karch and colleagues had another thought. What if microglia messed themselves up by expressing their own pathogenic tau? To see if this is possible, first author Abhirami Iyer analyzed previously published transcriptomic data on microglia isolated from cortical tissue of four people who had had Alzheimer’s disease and six controls from the ROSMAP cohort (Dec 2020 news). Iyer found that microglia from all donors expressed tau, though at levels much lower than what is seen in neurons.

Tau Tempers TREM2. Wild type, induced microglia-like cells (left) produce much more TREM2 (cyan) than iMGLs expressing the IVS10+16 tau mutation (right). The latter have much less TREM2 on the cell surface (not shown). [Courtesy of Iyer et al., medRxiv, 2024.]

Nevertheless, the scientists studied the potential effects of this tau by creating induced microglia-like cells, i.e., iMGLs, from iPSCs of two people who each carried one copy of the MAPT IVS10+16 variant, which causes autosomal-dominant frontotemporal dementia. Like microglia in vivo, iMGLs expressed tau mRNA and protein. The IVS10+16 iMGLs made as much tau as isogenic control cells, though they favored four-repeat over three-repeat tau, in keeping with the variant’s ability to alter tau splicing in neurons (Capano et al., 2022). 

RNA-Seq revealed that the transcriptional states had shifted in IVS10+16 iMGLs. They mobilized chemokine genes at the expense of genes needed for phagocytosis, or of genes from disease-associated and lipid-associated microglia of AD and other neurodegenerative diseases express. As a result, the tau variant iMGLs engulfed about one-third less fibrillar tau and 23 percent less myelin than did control cells, highlighting their listless phagocytosis.

The downregulation of DAM and LAM genes hinted that the IVS10+16 iMGLs might be less reactive than control cells. They had little TREM2 mRNA and almost none of the receptor protein. A DAM gene, TREM2 spurs microglia to engulf myelin debris as well as surround and compact amyloid plaques (see image above).

Totally Tubulin. Rather than branching out from the nucleus, as in wild-type iMGLs (left), α-tubulin (black) gathers around the cell periphery in IVS10+16 iMGLs (right). [Courtesy of Iyer et al., medRxiv, 2024.]

Given that tau binds microtubules, the scientists next looked at the iMGL cytoskeleton. The mutant tau cells had quieted genes regulating actin, a microfilament building block, and had only half as much actin as isogenic controls. The microtubule protein α-tubulin was distributed abnormally within the induced microglia, hanging around the furthest reaches of the cytosol rather than near the nucleus (image at right). The authors believe wonky cytoskeletons explain the phagocytosis defect, since that process requires the cytoskeleton to wrap around fragments of myelin, tau, and other debris.

To learn if IVS10+16 tau affected the microglial secretome, Iyer measured levels of 1,360 proteins in the iMGLs culture media. Eighty-nine were differently expressed in IVS10+16 versus control iMGLs, with most being overproduced (image below). The scientists think these 89 make for a toxic brew, since medium from tau-variant iMGLs stunted synapses when added to iPSC-derived neurons in culture.

Noxious Secretome? IVS10+16 iMGLs overproduced most secretory proteins (red), but held others back (blue). [Courtesy of Iyer et al., medRxiv, 2024.]

All told, the findings indicate that mutant tau has profound cell-autonomous effects on induced microglia. Would that play out in vivo? Many scientists do not believe induced microglia fully recapitulate the properties of microglia in their normal environment. In fact, Kiran Bhaskar, Jonathan Hulse, and Karthikeyan Tangavelou, all at the University of New Mexico in Albuquerque, wondered if the iMGLs turned on neuronal genes, including tau, because they were in culture (comment below).

To look for in vivo evidence of cell-autonomous effects, Iyer and colleagues combed through published bulk RNA-Seq data on middle temporal gyrus tissue from three controls and two people carrying MAPT IVS10+16. They found 100 genes whose up- or downregulation in carriers mimicked that seen in iMGLs (Minaya et al., 2023).  These included immune activation and cytoskeletal organization genes. Likewise, CSF proteomics data from 37 people with various tau mutations were compatible with an uptick in microglial adaptive immunity pathways, and a failing of extracellular matrix pathways, just as in the iMGLs.

To Karch, this was profound. “I was blown away by the degree to which the proteins that changed in people with tauopathies matched those that changed in the media of the iMGLs,” she told Alzforum. To her mind, these findings highlight how accurately iMGLs reflect what is going on in people.

Karch wants to understand how these findings pertain to other tau mutations, so she is repeating the experiments using the P301L, R406W, and V337M tau mutations.—Chelsea Weidman Burke


  1. Microglia in the human brain are not known to develop tau pathology but do play a significant role in propagation of tau pathology in animal models of tauopathy (Asai et al., 2015; Shi et al., 2019; Wang et al., 2022), in corticobasal syndrome (Palleis et al., 2024), and across the FTD spectrum (Bevan-Jones et al., 2020). The functional role of endogenously expressed tau in microglia is underexplored. This study, although limited to one particular FTLD-MAPT mutation, IVS10+16, is a well-controlled investigation of how this mutation impacts iMGL biology in comparison to isogenic iMGL controls.

    Interestingly, the authors detected 3R tau in these MAPT mutant iMGLs, and more 4R tau than in control iMGLs, which is consistant with reports on this mutation (Hutton et al., 1998). MAPT IVS10+16 iMGLs show upregulation of chemokines, downregulation of DAM and LAM genes, reduction of phagocytosis, TREM2 signaling, and energy metabolism.

    These findings were validated by bulk RNA-Seq of human brain tissues isolated from MAPT IVS10+16 mutation carriers and control cases without neuropathological change. Publicly available proteomic datasets from the CSF samples of MAPT carriers and controls show enrichment of protein modules involved in extracellular matrix, complement, adaptive immunity, autophagy, and synapse assembly in symptomatic cases, whereas integrin signaling is more enriched in presymptomatic cases.

    Finally, to determine the biological effect of iMGL on neurons, conditioned media from MAPT IVS10+16 or control iMGLs were applied to iNeurons, which showed reduced synaptic density and increased dendritic length.

    These data support the idea that the MAPT IVS10+16 mutation may alter microglia to be less active and impaired for phagocytosis and energy metabolism. This is consistent with a recent study showing little activation of microglia in FTLD-tau brain (Hartnell et al., 2024). However, the study is inconclusive on whether this is due to the endogenous expression of tau, since they have not tested the effect of silencing MAPT expression on these MAPT IVS10+16 iMGLs.

    It will also be of interest to learn if misfolded tau is found to accumulate in these MAPT mutant microglia. 


    . Depletion of microglia and inhibition of exosome synthesis halt tau propagation. Nat Neurosci. 2015 Nov;18(11):1584-93. Epub 2015 Oct 5 PubMed.

    . Microglia drive APOE-dependent neurodegeneration in a tauopathy mouse model. J Exp Med. 2019 Nov 4;216(11):2546-2561. Epub 2019 Oct 10 PubMed.

    . Microglial NF-κB drives tau spreading and toxicity in a mouse model of tauopathy. Nat Commun. 2022 Apr 12;13(1):1969. PubMed.

    . Association of Neurofilament Light Chain, [18F]PI-2620 Tau-PET, TSPO-PET, and Clinical Progression in Patients With β-Amyloid-Negative CBS. Neurology. 2024 Jan 9;102(1):e207901. Epub 2023 Dec 14 PubMed.

    . Neuroinflammation and protein aggregation co-localize across the frontotemporal dementia spectrum. Brain. 2020 Mar 1;143(3):1010-1026. PubMed.

    . Association of missense and 5'-splice-site mutations in tau with the inherited dementia FTDP-17. Nature. 1998 Jun 18;393(6686):702-5. PubMed.

    . Glial reactivity and T cell infiltration in frontotemporal lobar degeneration with tau pathology. Brain. 2024 Feb 1;147(2):590-606. PubMed.

  2. Iyer et al. present compelling evidence of MAPT expression and presence of endogenous tau in microglia (iMGLs), utilizing both human brain samples and induced microglia-like cells derived from induced pluripotent stem cells (iPSCs). This finding is significant as it corroborates the limited existing studies that report tau in glial cells.

    Furthermore, this study is notable because most data on the cellular localization of tau and its phagocytosis derive from transgenic mouse models, which exclusively express human tau in neurons. The primary observational results indicate that, in iMGLs derived from iPSCs carrying the MAPT IVS10+16 mutation, TREM2 expression is reduced, resulting in defects in phagocytosis, cytoskeletal organization, endolysosomal function, and metabolic processes in these microglia. Future research should explore the interaction between tau and TREM2 and how MAPT mutations lead to TREM2 downregulation and subsequent cellular phenotypes. 

  3. Microtubule-associated protein tau traditionally has been viewed solely as a neuronal protein, pivotal in the progression of neurodegenerative conditions such as Alzheimer’s disease and primary tauopathies. Consequently, our focus has predominantly centered on studying the pathological accumulation of tau in neurons, with any observed changes in microglia being perceived as a downstream response to this neuronal pathology. However, this narrative has never explained why microglial dysregulation often precedes the pathological accumulation of neuronal tau in human disease.

    Here, the research team led by Dr. Celeste Karch at Washington University in St. Louis, Missouri, illuminated a crucial insight into this dilemma: Microglia also express tau protein if its genome carries the MAPT IVS10+16 mutation, which is known to increase 4R tau splice variant expression. The authors show that tau expression in microglia seems to affect normal microglial function and triggers significant transcriptional alterations in important pathways governing microglial function.

    While data from FTD iMGLs show tau expression compared to CRISPR-corrected “control” iMGLs, it is important to consider a few points. First, earlier studies have shown that microglia can cross-seed tau via exosomes, which are known to contain mRNAs (derived from neurons). Therefore, it is important to determine the identity/origin of tau mRNA, especially in microglia isolated for in vivo experiments. Second, we sometimes see that cell confluency may result in the de novo expression of certain neuronal proteins during the maturation steps of deriving iMGLs. Therefore, consideration of aberrant transcription/translational factors driving microglial tau expression may be important. Finally, extensive investigations are warranted to delineate the role of microglia tau “physiological” function (in mutant carriers) thoroughly and to develop a mouse model of (MAPT IVS10+16) FTDs to validate the observed phenotype. It would also be important to determine whether microglial tau expression is specific to this specific splice-site mutation, or whether other intronic/exonic tau mutants/ haplotypes also display such a phenomenon. Nonetheless, this study may have profound implications for the neuroimmune changes in tauopathies.

  4. Our current understanding is that tau, encoded by the MAPT gene, is highly enriched in neuronal axons and present at minimal levels in non-neuronal cells. However, in the brains of individuals with FTLD-tau, those inclusions have been detected in glial cells, and they drive glial activation and dysfunction. This underscores the role of tau-inclusion-bearing glia in neurodegenerative diseases (Ezerskiy et al., 2022; Chung et al., 2021).

    Glial cells, especially microglia, are active phagocytes. A critical question is whether glial tau inclusions result from phagocytosed tau released from tau-bearing neurons or from tau expressed intrinsically by glia. By leveraging human iPSC-derived microglia-like cells (iMGLs), the current study demonstrated that microglia express tau mRNA and protein isoforms.

    Furthermore, it reports that if the primary tauopathy MAPT IVS10-16 mutation is expressed by microglia, then it influences their transcriptomic states and alters cell functions in a cell-autonomous manner. Additionally, human iMGLs bearing this mutation regulated neuronal synapses. This pioneering study uncovered cell-autonomous effects of microglia-expressed tau and illuminated the contribution of intrinsically developed tau inclusions in microglia to tau pathogenesis.

    This intriguing research leaves several issues to be addressed by the field:

    1. Tau expression in microglia: Emerging studies support cell-autonomous tau expression in glia, predominantly in astrocytes and oligodendrocytes. Evidence of tau inclusions in microglia is still scarce. However, the current study provides convincing evidence that microglia, whether isolated from the human brain or differentiated from iPSCs (hMGLs), express tau at baseline levels. Does microglial tau expression change with aging or under various tauopathy conditions, including Alzheimer's disease?
    2. The nature of microglial tau: Characterization of microglial tau is needed at the basic cell biology and biochemistry levels, including post-translational modifications, solubility/aggregation, and cellular location. Are these tau species phosphorylated or acetylated at similar sites as neuronal tau? It would be informative to profile the solubility of microglial tau, particularly microglia bearing FTD mutations, compared to neuronal tau. Do these tau species aggregate or form oligomers? Where are these tau species located in microglia, and are they secreted? How do mutations affect these processes? Moreover, Li Gan’s group generated a tau interactome landscape in human iPSC-derived neurons and uncovered tau binding partners involved in diverse cellular processes, including synaptic activity and mitochondrial function (Tracy et al., 2022). It would be intriguing to examine tau interactors in microglia and compare with neuronal tau partners.
    3. Contribution of microglial tau to neuroinflammation: The current study aligns with a recent case report by Richard Bevan-Jones and colleagues, which describes microglial activation in frontotemporal regions lacking tau aggregation or atrophy in a presymptomatic carrier of the IVS10+16 mutation (Bevan-Jones et al. 2019). It is possible that microglial tau with FTD mutations, such as IVS10+16, plays an early role in microglial activation in frontotemporal dementia before neuronal tau inclusions develop. It remains to be investigated how microglia respond to immunogenic tau released from neurons at a more advanced disease stage and how FTD mutations affect this.
    4. Mechanisms of microglial activation: We and others have shown that extracellular tau fibrils activate microglia and induce inflammatory signaling pathways such as TREM2-TYROBP, NFκB, and cGAS-STING, both in vitro and in vivo (Wang et al., 2022; Udeochu et al., 2023; Jin et al., 2021). One underlying mechanism is that, upon entering microglia, neuronal tau triggers the release of mitochondrial DNA, which is sensed by cGAS, leading to the activation of the STING-IFN response. Previous studies have observed diverse cellular and pathological heterogeneity in primary tauopathies, suggesting different mechanisms are involved (Chung et al., 2021). How intrinsically expressed tau activates microglia remains to be determined.


    . In vivo evidence for pre-symptomatic neuroinflammation in a MAPT mutation carrier. Ann Clin Transl Neurol. 2019 Feb;6(2):373-378. Epub 2019 Jan 2 PubMed.

    . Cellular and pathological heterogeneity of primary tauopathies. Mol Neurodegener. 2021 Aug 23;16(1):57. PubMed.

    . Astrocytic 4R tau expression drives astrocyte reactivity and dysfunction. JCI Insight. 2022 Jan 11;7(1) PubMed.

    . Tau activates microglia via the PQBP1-cGAS-STING pathway to promote brain inflammation. Nat Commun. 2021 Nov 15;12(1):6565. PubMed.

    . Tau interactome maps synaptic and mitochondrial processes associated with neurodegeneration. Cell. 2022 Feb 17;185(4):712-728.e14. Epub 2022 Jan 20 PubMed.

    . Tau activation of microglial cGAS-IFN reduces MEF2C-mediated cognitive resilience. Nat Neurosci. 2023 May;26(5):737-750. Epub 2023 Apr 24 PubMed.

    . Microglial NF-κB drives tau spreading and toxicity in a mouse model of tauopathy. Nat Commun. 2022 Apr 12;13(1):1969. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Most Detailed Look Yet at Activation States of Human Microglia

Mutations Citations

  1. MAPT IVS10+16 C>T
  2. MAPT P301L
  3. MAPT R406W
  4. MAPT V337M

Paper Citations

  1. . Association of glial tau pathology and LATE-NC in the ageing brain. Neurobiol Aging. 2022 Nov;119:77-88. Epub 2022 Jul 31 PubMed.
  2. . Aging-related tau astrogliopathy (ARTAG): harmonized evaluation strategy. Acta Neuropathol. 2016 Jan;131(1):87-102. Epub 2015 Dec 10 PubMed.
  3. . Astrocytic uptake of neuronal corpses promotes cell-to-cell spreading of tau pathology. Acta Neuropathol Commun. 2023 Jun 17;11(1):97. PubMed.
  4. . Involvement of Oligodendrocytes in Tau Seeding and Spreading in Tauopathies. Front Aging Neurosci. 2019;11:112. Epub 2019 May 28 PubMed.
  5. . In vivo evidence for pre-symptomatic neuroinflammation in a MAPT mutation carrier. Ann Clin Transl Neurol. 2019 Feb;6(2):373-378. Epub 2019 Jan 2 PubMed.
  6. . Direct Evidence of Internalization of Tau by Microglia In Vitro and In Vivo. J Alzheimers Dis. 2015;50(1):77-87. PubMed.
  7. . Recapitulation of endogenous 4R tau expression and formation of insoluble tau in directly reprogrammed human neurons. Cell Stem Cell. 2022 Jun 2;29(6):918-932.e8. PubMed.
  8. . Conserved gene signatures shared among MAPT mutations reveal defects in calcium signaling. Front Mol Biosci. 2023;10:1051494. Epub 2023 Feb 9 PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Cell autonomous microglia defects in a stem cell model of frontotemporal dementia. 2024 May 16 10.1101/2024.05.15.24307444 (version 1) medRxiv.