Cell autonomous microglia defects in a stem cell model of frontotemporal dementia tau

Abstract

Neuronal dysfunction has been extensively studied as a central feature of neurodegenerative tauopathies. However, across neurodegenerative diseases, there is strong evidence for active involvement of immune cells like microglia in driving disease pathophysiology. Here, we demonstrate that MAPT mRNA and tau protein are expressed in microglia in human brains and in human induced pluripotent stem cell (iPSC)-derived microglia like cells (iMGLs). Using iMGLs harboring the MAPT IVS10 + 16 mutation and isogenic controls, we demonstrate that a tau mutation is sufficient to alter microglial transcriptional states. We discovered that MAPT IVS10 + 16 microglia exhibit cytoskeletal abnormalities, stalled phagocytosis, disrupted TREM2/TYROBP networks, and altered metabolism. Additionally, we found that secretory factors from MAPT IVS10 + 16 iMGLs impact neuronal health, reducing synaptic density in neurons. Key features observed in vitro were recapitulated in human brain tissue and cerebrospinal fluid from MAPT mutations carriers. Together, our findings that MAPT IVS10 + 16 drives cell-intrinsic dysfunction in microglia that impacts neuronal health has major implications for development of therapeutics for tauopathies.

Introduction

Frontotemporal lobar degeneration (FTLD) encompasses a range of heterogeneous neurodegenerative disorders. Familial FTLD with tau inclusions (FTLD-tau) is characterized by dominantly inherited mutations in the microtubule-associated protein tau (MAPT) gene, while common variants in MAPT can also play a role in sporadic forms of FTLD-tau [1, 2]. FTLD-tau is defined by neuropathological hallmarks such as atrophy of the frontal and temporal lobes, abnormal tau protein inclusions, gliosis, and neuronal loss, as well as clinical hallmarks such as behavioral/language difficulties and cognitive dysfunction [3]. In the brain, MAPT is alternatively spliced to produce 6 major isoforms that differ based on N-terminal insertions and the number of repeats in the microtubule binding domain (3 repeats (3R); 4 repeats (4R)). A subset of MAPT mutations (e.g. MAPT IVS10 + 16) alter splicing and shift tau isoforms (4R > 3R). These mutations mimic a subset of sporadic tauopathies in which 4R tau protein accumulates in brains (e.g., progressive supranuclear palsy and corticobasal degeneration) [4].

Tau is a neuronal protein highly concentrated in axons, with minimal presence in other cellular compartments or non-neuronal cells [5, 6]. Nevertheless, tau-positive inclusions have been documented in astrocytes and oligodendrocytes in human neuropathological studies, including in patients carrying the MAPT IVS10 + 16 mutation [7–11]. Interestingly, microglial activation has been observed in presymptomatic MAPT mutation carriers, even in the absence of protein aggregation and atrophy in the frontotemporal regions [12, 13]. Furthermore, increased microglial activation has been proposed as a potential biomarker for FTLD linked to intronic mutations in MAPT, correlating with a more rapid cognitive decline [14–16]. Despite the evidence of tau inclusions in non-neuronal cells, little work has been done to evaluate MAPT mRNA and tau protein expression in glia [17–19]. This may be due to the prevailing viewpoint that glial inclusions arise from glial cells internalizing tau ladened neurons or capturing tau protein released by neurons [10, 20, 21]. Microglia, the resident immune cells of the brain, may also play a critical role in tauopathy pathophysiology [22–25]. Microglia can engulf neurons and synapses burdened with tau [26–28]. Impaired microglia-mediated clearance of phagocytosed tau can lead to propagation of tau pathology through secretion and promote a vicious cycle that may drive pathological decline [10, 25, 29, 30]. However, studies directly investigating the potential role of microglia as cell autonomous molecular drivers of tauopathy have been limited. Most studies of the contribution of microglia to tauopathy use animal models that express the human mutant tau transgene in neurons (and not microglia), which does not fully recapitulate human tau expression patterns [31, 32]. Thus, cell-intrinsic dysfunction of microglia in tauopathies remains largely unknown [33].

To begin to understand the contribution of tau to microglia dysfunction in tauopathies, we demonstrated that tau mRNA and protein are expressed in microglia isolated from human brains and in human induced pluripotent stem cell (iPSC)-derived microglia-like cells (iMGLs). We then leveraged iMGLs from MAPT IVS10 + 16 mutation carriers and isogenic controls to discover broad molecular and cellular defects in MAPT IVS10 + 16 iMGLs including dysregulated microglial states, altered cytoskeleton organization, reduced phagocytosis of human myelin and recombinant human tau preformed fibrils, and altered TREM2/TYROBP signaling. MAPT IVS10 + 16 shifts the secretome of iMGLs, which leads to elongated dendrites with reduced synapse density in human neurons. Many of the features we report in vitro are maintained in brains and cerebrospinal fluid (CSF) from MAPT mutation carriers. Our findings reveal that cell-intrinsic dysfunction in microglia may contribute to neuronal dysfunction and FTLD-tau pathophysiology.

Results

FTLD-causing MAPT IVS10 + 16 mutation causes a shift in microglia transcriptional states

Microglial activation is a key feature of FTLD-tau pathology [16, 34–36]; however, the cell intrinsic effects of microglia in FTLD-tau pathophysiology remain poorly understood. Tau protein was originally identified as a neuronal microtubule-associated protein specifically localized and enriched in the neuronal axons in the central nervous system (CNS) [5]. More recently, MAPT mRNA expression and tau protein levels have been reported in human iPSC-derived astrocytes [37], suggesting that tau may be expressed more broadly in CNS cells than previously thought. To determine whether tau is expressed in microglia, we analyzed publicly available data from a previously reported cohort of 10 human brains in which microglia were isolated and profiled by bulk RNA sequencing [38, 39] (Fig. 1A; Supplemental Table 1). Briefly, microglia were obtained from human brains (gray matter) using mechanical homogenization followed by myelin depletion. Cells were then isolated by FACS using CD11b and CD45 (see Methods for additional details). These isolated microglia were free of macrophages (high expressing CD45 cells), infiltrating monocytes, and other contaminating lymphocytes based on flow cytometry and transcriptomic readouts as previously reported [38, 39]. We found that microglia from all donors expressed MAPT (Fig. 1B), and the overall levels of MAPT were robust amongst all protein coding genes captured in the samples (Fig. 1C). Thus, we demonstrate that human microglia express MAPT mRNA in vivo.

Fig. 1. iPSC-derived microglia expressing MAPT IVS10 + 16 reveal shifts in microglia states compared with isogenic controls.

A–C Microglia isolated from human brains and analyzed by bulk RNA sequencing were analyzed for MAPT expression (n = 10) as previously reported in Olah et al [38, 39]. A Workflow for microglia isolation. B MAPT expression (TPM). Each dot represents microglia sampled from an individual brain. C MAPT expression, indicated by the red dashed line, relative to the 19,182 protein coding genes in microglia isolated from human brains. D Schematic representing how iPSC-derived human microglia-like cells (iMGLs) were generated from hematopoietic progenitor cells (HPCs) differentiated from two independent donors of MAPT IVS10 + 16 carriers: FTD1 and FTD2 and their isogenic controls (4 lines total; Supplemental Table 1). E Dotplots of CD45 and CD11b expression reveals that >95% of the iMGLs are double positive for CD45 and CD11b across donors and genotypes. F Immunocytochemistry of iMGLs for TMEM119. Scale bar, 12 µm. G–J Whole cell lysates were isolated by detergent extraction and analyzed by immunoblotting for total tau (Tau5), P2RY12 (microglial homeostatic marker), and GAPDH (loading control) (G) and by immunoprecipitation (Tau1/HJ8.5/HJ8.7 antibodies) and mass spectrometry (H–J). H. 3R and 4R specific peptides monitored by mass spectrometry plotted as a ratio for iMGLs. I Relative expression of tau protein in iPSC-derived neurons (iNeurons) compared with iMGLs in MAPT IVS10 + 16 and isogenic controls (WT). Expression of Tau-212 peptide relative to internal standard (15 N). J Peptides monitored by mass spectrometry are plotted relative to the mid-domain peptide at amino acid 181-190 (TPSL). Dashed lines, MAPT IVS10 + 16. Solid lines, MAPT WT (isogenic). Gray line/circle, iPSC-derived neurons from MAPT WT, isogenic to mutant line. Black line/circle, iPSC-derived neurons from MAPT IVS10 + 16 mutation carrier. Light orange, iMGLs from MAPT WT, isogenic to mutant line. Dark orange, iMGLs from MAPT IVS10 + 16 mutation carrier. K RNAseq and differential gene expression were performed on iMGLs from MAPT IVS10 + 16 and isogenic controls. Volcano plot. Red dots, FDR-BH ≤ 0.05. Blue dots, p-value ≤ 0.05. Gray dots, not statistically significant. L Percentage of genes defining specific microglia functions/states: chemokines (n = 12 genes), phagocytosis (n = 14 genes), DAM (n = 17 genes), LAM (n = 12 genes), and homeostatic (n = 10 genes) were explored in the differential gene expression. Red, significantly upregulated. Blue, significantly downregulated. White, not statistically significant. FDR-BH ≤ 0.05. DAM, disease associated microglia. LAM, lipid-associated myeloid cells.

To understand the microglia cell intrinsic contributions of a mutation in the MAPT gene that causes autosomal dominant FTLD-tau, we leveraged human iPSCs from MAPT mutation carriers [40]. iPSCs from two independent donors who are heterozygous carriers of the MAPT IVS10 + 16 mutation (MAPT IVS10 + 16/WT; FTD1 and FTD2) were compared with CRISPR-corrected isogenic controls in which the mutant allele was corrected to WT (MAPT WT/WT; Supplemental Table 1; Supplemental Fig. 1; Fig. 1D). iPSCs were differentiated into iMGLs using a growth factor-based approach through hematopoietic progenitor cells (HPCs) as previously described [41, 42] (Fig. 1D; see Materials and Methods). Using flow cytometry, we demonstrated that iPSC lines, regardless of mutation status, share a similar propensity to form CD45 and CD11b double positive iMGLs (Fig. 1E). iMGLs were also similar in their expression of the microglia marker TMEM119 and Iba1 (Fig. 1F; Supplemental Fig. 2A). Thus, the MAPT IVS10 + 16 mutation does not lead to overt defects in the capacity to form mature microglia from iPSCs.

In neurons, MAPT IVS10 + 16 alters tau splicing, leading to elevated 4R tau protein expression [43]. To determine whether iMGLs express tau and the degree to which MAPT IVS10 + 16 alters tau splicing in iMGLs, we measured MAPT mRNA and tau protein. We found that MAPT mRNA is expressed in iMGLs (Supplemental Table 2), and tau protein is detectable by immunoblotting (Fig. 1G). Total MAPT mRNA and tau protein levels were similar between MAPT IVS10 + 16 and isogenic control iMGLs (Fig. 1G; Supplemental Table 2). One may question whether the source of the tau we detect is neuronal in origin. Unlike neurons and astrocytes which are derived from ectodermal neural progenitor lineages, microglia are derived from the mesoderm. Thus, this iMGL protocol does not produce contaminating cells of neuronal or astroglia fate. Using immunocytochemistry and transcriptomics, we demonstrate that iMGLs are free of the neuronal markers (Supplemental Fig. 2A, B). Thus, iMGLs are the source of the tau we detect. To more quantitatively measure tau isoforms, we used immunoprecipitation and mass spectrometry (Fig. 1H–J). MAPT IVS10 + 16 iMGLs produce more 4R tau isoforms compared to WT, consistent with the biological effect of the mutation and prior reports in human iPSC-derived neurons (Fig. 1H) [43]. To better understand how tau protein expression in iMGLs compares to iPSC-derived neurons, we measured a mid-domain peptide, Tau-212, which serves as surrogate for total tau (Fig. 1I). iPSCs from the same donors were differentiated into either iMGLs or iPSC-derived neurons. We found that iPSC-derived neurons express approximately 200-fold more tau than iMGLs (Fig. 1I). To determine whether tau isoform expression profiles differ between iMGLs and iPSC-derived neurons, we mapped tau peptide expression from N- to C-terminus (Fig. 1J). We discovered that tau isoform expression is similar to patterns we have previously described in human iPSC-derived neurons [44], the predominant tau isoform expressed is 3R0N (Fig. 1J). Thus, we demonstrate that iMGLs express MAPT mRNA and tau protein, and MAPT IVS10 + 16 alters tau splicing in iMGLs.

Microglia are heterogenous cells that can exist in diverse, transitory states that are influenced by genetics, environment, development, infection, and disease [45, 46]. To begin to understand the extent to which the MAPT IVS10 + 16 mutation impacts microglial transcriptional states, we analyzed mutant and isogenic control iMGLs by RNA sequencing (Fig. 1K; Supplemental Fig. 1; Supplemental Table 2). Principal component analyses (PCA) revealed that the presence of the MAPT IVS10 + 16 mutation was sufficient to drive global transcriptional signatures (Supplemental Fig. 1E). To determine whether the MAPT IVS10 + 16 driven transcriptional changes translate to shifts in microglia states, we plotted differential gene expression within each of the key microglial states [38, 47–51]: chemokine; phagocytosis; disease associated microglia (DAM); lipid associated microglia (LAM); and homeostatic microglia (Fig. 1L; Supplemental Table 3). Chemokines, which play a critical role in chemoattraction, were broadly significantly upregulated in MAPT IVS10 + 16 iMGLs. A number of genes involved in phagocytosis, DAM, and LAM were significantly dysregulated in mutant iMGLs, while homeostatic genes were largely unchanged (Fig. 1L; Supplemental Table 3).

The cytoskeletal network is altered in iMGLs expressing MAPT IVS10 + 16

As a microtubule-associated protein, tau has been well characterized for its mechanistic roles in polymerization, stabilization of microtubules, and suppression of microtubule dynamics [52]. Tau may also serve as a molecular linker that enables interaction between the actin and microtubule cytoskeletal networks to support cellular migration, morphological changes to stimuli, and a host of other cellular functions [53, 54]. Tau mutations and post-translational modifications have been shown to lead to mis-regulation of microtubule interaction and function [55–57]. Transcriptomic data revealed a significant upregulation of genes involved in semaphorin signaling and downregulation of genes belonging to the Rho family of GTPases in MAPT IVS10 + 16 iMGLs (Fig. 2A), both of which have been reported to act as molecular switches in their regulation of actin filament and microtubule networks [58–60]. Using immunocytochemistry, we demonstrate that tau co-localizes with α-tubulin-positive microtubules in MAPT IVS10 + 16 and WT iMGLs (Supplemental Fig. 3A, B). Thus, MAPT IVS10 + 16 does not significantly alter tau’s association with microtubules. To next determine whether microtubule and actin cytoskeletal networks are altered in mutant microglia, we performed immunocytochemistry to visualize and characterize the α-tubulin (Fig. 2B, C) and filamentous forms of actin (F-actin) (Fig. 2D–F) in MAPT IVS10 + 16 and WT iMGLs. In MAPT WT iMGLs, α-tubulin exhibits a radial array pattern originating from the microtubule organizing center (MTOC) (Fig. 2B, C; Supplemental Fig. 2C, D). However, in the presence of the mutant allele, α-tubulin distribution is shifted to the periphery (Fig. 2B-C; Supplemental Fig. 2C, D). Semaphorin signaling and the RhoH GTPase cycle contribute to the disassembly of F-actin [61, 62]. Consistent with the dysregulation of these pathways in mutant iMGLs, we observed a significant reduction of F-actin (phalloidin) in MAPT IVS10 + 16 iMGLs in comparison to the isogenic WT iMGLs based on immunocytochemistry (Fig. 2D, E) and flow cytometry (Fig. 2F, G). Thus, MAPT IVS10 + 16 is sufficient to alter actin and tubulin cytoskeletal networks in microglia.

Fig. 2. MAPT IVS10 + 16 alters the cytoskeleton in iMGLs.

A Pathway analysis of genes differentially expressed between MAPT IVS10 + 16 and isogenic controls shows an enrichment of genes regulating the cellular cytoskeletal network. Red bars, significantly upregulated genes. Blue bars, significantly downregulated genes. Significance defined as FDR-BH ≤ 0.05. Three biological replicates were included for each donor. B Immunocytochemistry for α-tubulin. Representative images from a Nikon spatial array confocal images acquired at 60x. Scale bar, 12 µm. C Quantification of mean fluorescence intensity (MFI) of α-tubulin staining across cellular area using the straight-line tool in ImageJ. MFI was plotted relative to distance across the cell (normalized and expressed as a percentage). D Immunocytochemistry for filamentous actin (F-actin; phalloidin; green) and nucleus (NucBlue). Scale bar, 12 µm. E Quantification of mean fluorescence intensity of phalloidin within each cell. ****, p < 0.0001. F Total intracellular levels of F-actin were quantified by flow cytometry using Phalloidin-488 represented as dot plots reveal the percentage of F-actin positive cells within CD45 + CD11b+ iMGLs. Representative dot plots from FTD1 donor pair. G Quantification of phalloidin (MFI). MFI normalized to WT. Results represent three independent experiments. **, p < 0.005; ***, p < 0.0005; p-values generated by one-way ANOVA with Sidak’s multiple comparisons test in (E) and (G).

MAPT IVS10 + 16 iMGLs exhibit impaired phagocytosis of myelin and tau preformed fibrils

As the CNS resident immune cells, microglia perform constant surveillance functions [63]. Microglia are required to adapt their cell morphology for rapid migration through narrow spaces in the parenchyma and to phagocytose debris or foreign cargo in their microenvironment [64]. Efficient microglial phagocytosis relies on downstream signaling cascades that mediate regulation of F-actin networks, which mobilizes around the phagocytic cup to enable target cargo engulfment [65]. Consistent with our observation that mutant iMGLs exhibited reduced F-actin levels and an altered microtubule network, transcriptomic analyses revealed a significant downregulation of genes involved in phagocytosis and amoeboid-type cell migration, including MERTK (Fig. 3A; Supplemental Table 2). MAPT IVS10 + 16 iMGLs also exhibited a significant upregulation of genes involved in the negative regulation of phagocytosis, suppression of phagosomal maturation, and the prevention of phagosomal-lysosomal fusion (Fig. 3A).

Fig. 3. MAPT IVS10 + 16 iMGL exhibit reduced phagocytosis of myelin and tau fibrils.

A Pathway analysis of genes differentially expressed between MAPT IVS10 + 16 and isogenic controls shows an enrichment of genes regulating endocytosis and phagocytosis. Red bars, significantly upregulated genes. Blue bars, significantly downregulated genes. Significance defined as FDR-BH ≤ 0.05. Three biological replicates were included for each donor. B Phagocytic uptake of human myelin (20 µg/ml, C-E) or recombinant human tau preformed fibrils, Tau-PFF (500 nM, F-H) analyzed by IncuCyte live cell imaging over 24 h. C Representative images 12 h post myelin treatment. Scale bar, 100 µm. D Representative quantification of myelin area over time. E Quantification of myelin area at 24 h post-treatment. ****, p < 0.0001. F Representative images 12 h post Tau-PFF treatment. Scale bar, 100 µm. G Representative quantification of Tau-PFF area over time. H Quantification of Tau-PFF area at 12 h post-treatment. *, p < 0.05; **, p < 0.005. Data in C-H is representative of 8 replicates per donor line per treatment group. I, J. Tau-PFF uptake was measured by flow cytometry. I Histogram of Tau-PFF mean fluorescence intensity (MFI) within CD45 + CD11b+ cells 24 h post-treatment. J Quantification of Tau-PFF MFI. Data is representative of three biological replicates per line. ***, p < 0.005; ****, p < 0.0001. p-values by two-ANOVA with Sidak’s multiple comparisons test in (E), (H) and (J).

Studies investigating the role of microglia in tauopathies have shown that microglia phagocytose extracellular tau or neurons laden with tau aggregates [10, 26]. When microglia fail to degrade engulfed tau, they become hypophagocytic, senescence-like, and release tau seeds [29, 30, 66, 67]. To date, how the expression of mutant tau in microglia can impact these functions has not been well explored. To investigate the direct cell autonomous impact of MAPT IVS10 + 16 on microglial phagocytosis, we challenged iMGLs with physiological (myelin), microbe mimicking (zymosan A bioparticles), and disease-relevant (tau preformed fibrils (PFF)) substrates and performed hourly live cell imaging for 24 h (Fig. 3B). MAPT IVS10 + 16 iMGLs phagocytosed significantly less myelin than isogenic controls (Fig. 3C–E); however, there was no difference in zymosan A uptake (Supplemental Fig. 4). Tau PFFs were taken up at a significantly reduced rate in the MAPT IVS10 + 16 iMGLs compared with isogenic controls (Fig. 3F–H). We further validated the defects in tau uptake in MAPT IVS10 + 16 iMGLs by measuring Tau-PFF fluorescence 24 h after its addition using flow cytometry (Fig. 3I, J). Thus, MAPT IVS10 + 16 drives phagocytosis deficits in iMGLs that are substrate specific to dying neurons and tau aggregates.

TREM2 signaling and metabolic networks are altered in MAPT IVS10 + 16 iMGLs

TREM2 is a microglia-enriched membrane protein that binds to adapter proteins and signaling mediators like DAP12 (also known as TYROBP) or DAP10 [68]. Transcriptomic analyses revealed a significant downregulation of TREM2 and the TREM2-associated TYROBP network in MAPT IVS10 + 16 iMGLs (Fig. 4A). Interestingly, we observed a loss of TREM2 staining by immunocytochemistry (Fig. 4B) and a significant reduction of TREM2 at the cell membrane by flow cytometry (Fig. 4C, D) in mutant iMGLs. Once at the cell surface, TREM2 is cleaved and released as a soluble form (sTREM2) [69–71]. Increased sTREM2 levels have been equated with increased microglial activity and cognitive resilience [72–74]. In MAPT IVS10 + 16 iMGLs, sTREM2 levels were significantly reduced compared with isogenic controls (Fig. 4E).

Fig. 4. MAPT IVS10 + 16 disrupts TREM2 signaling in iMGLs.

A Pathway analysis of genes differentially expressed between MAPT IVS10 + 16 and isogenic controls shows an enrichment of genes involved in TREM2 signaling. Red bars, significantly upregulated genes. Blue bars, significantly downregulated genes. Significance defined as FDR-BH ≤ 0.05. Three biological replicates were included for each donor. B Immunocytochemistry for TREM2 (cyan) and nucleus (NucBlue) Scale bar represents 12 µm. C, D TREM2 on the cell membrane (mTREM2) was measured by flow cytometry. C Histogram overlay of mTREM2 MFI. D mTREM2 MFI quantification in both donor pairs. *, p < 0.05. E Soluble TREM2 (sTREM2) levels measured by ELISA in media from both donor pairs. **, p < 0.05; p-values were calculated using an unpaired t-test with Welch’s correction in (D) and (E). F TREM2 signaling network genes are differentially expressed in MAPT IVS10 + 16 iMGLs. Logfold change plotted for each gene. G TREM2 mediates pathways in microglia that regulate actin/cytoskeletal organization, proliferation and survival, cytokines, autophagy, and metabolism. Those genes involved in these pathways that were significantly differentially expressed FDR-BH < 0.05 are annotated in red (upregulated in mutant iMGL) or blue (downregulated in mutant iMGL).

TREM2-DAP12/10 interactions initiate a signaling cascade in microglia that can result in integrin activation, cytoskeleton rearrangement, MAPK signaling, mTOR signaling, calcium mobilization, and energy metabolism [68, 75]. Consistent with the downregulation of TREM2 protein, we observed a broad effect on genes and pathways regulated by TREM2 (Fig. 4F, G; Supplemental Table 2). MAPT IVS10 + 16 iMGLs exhibited altered gene expression of TREM2-related pathways associated with proinflammatory cytokines, transcriptional regulation of DAM genes, actin cytoskeleton rearrangement, and phagocytosis (Fig. 4F, G). We also observed a significant reduction in MTOR and other endolysosomal genes (Fig. 4F, G; Supplemental Fig. 5A). MTOR is an important regulator of cellular metabolism in microglia [76–78]. Consistent with these findings, we identified a broad number of genes enriched in glycolysis that were altered in MAPT IVS10 + 16 iMGLs (Supplemental Fig. 5B, C). MAPT IVS10 + 16 iMGLs exhibited a significant increase in genes involved in mitochondrial oxidative phosphorylation, the mechanism by which cells translate nutrients into energy (Supplemental Fig. 5B, C). Together, these findings suggest that the MAPT IVS10 + 16 mutation impacts TREM2 expression and signaling that may ultimately contribute to broad defects in phagocytosis, cytoskeletal organization, endolysosomal function, and metabolism functions in microglia.

MAPT IVS10 + 16 iMGL secretome alters neuronal synapses

Here we show that MAPT IVS10 + 16 is sufficient to broadly disrupt microglial transcriptional states and key microglial functions. We identified genes and pathways associated with the adaptive immune response, the innate immune function, metabolism, and lysosomal functions that are dysregulated in mutant microglia (Supplemental Fig. 5D). To determine the extent to which the genes we detected in vitro are reflected in the pathologic processes in FTLD-tau patients, we compared the iMGL transcriptomic data with transcriptomic data previously generated from the middle temporal gyrus of MAPT IVS10 + 16 mutation carriers and neuropathology free controls (Supplemental Table 1) [79]. We identified 100 genes that were significant differentially expressed in the same direction in mutant iMGLs and FTLD-tau brains (Supplemental Table 4). These genes were enriched in pathways associated with adaptive immune activation, NF-κB, inflammasome activation, endolysosomal transport, cell adhesion molecules, and cytoskeletal organization (Fig. 5A). Interestingly, we also observed a number of the commonly dysregulated genes in pathways involving synaptic assembly and signaling, synaptic vesicle membrane organization, regulated exocytosis, and presynaptic endocytosis (Fig. 5A).

Fig. 5. MAPT IVS10 + 16 iMGLs produce secretory factors that alter synapse density.

A Genes shared between MAPT IVS10 + 16 iMGLs and brains of patients with FTLD-tau with MAPT mutations (n = 100 genes) were enriched in pathways involved in synaptic function and immune response. Genes defined by BH-FDR ≤ 0.05 in each dataset. B, C Secretory proteins were monitored using OLINK. B Volcano plot showing secretory factors altered in comparisons of media from MAPT IVS10 + 16 iMGLs and controls. Three technical replicates were included for each donor pair. Red dots, significantly upregulated (p < 0.05). Blue dots, significantly downregulated (p ≤ 0.05). Gray dots, not significant. C Pathway analysis of differentially expressed secretory factors (p < 0.05). Red, upregulated genes. Blue, downregulated genes. D Proteins in cerebrospinal fluid (CSF) were measured by SomaScan proteomics platform in MAPT mutation carriers compared with normal controls. Network analyses revealed several modules that were significantly enriched in CSF from MAPT mutation carriers compared with controls. Sx, symptomatic. PreSx, pre-symptomatic. E Diagram of paradigm. F Media from control neurons, WT iMGL, or MAPT IVS10 + 16 iMGL was added to iPSC-derived neurons for 7 days. Immunocytochemistry was performed for dendritic marker (MAP2; green), pan-synaptic marker (Synapsin; red) and nucleus (NucBlue). Scale bar represents 50 µm. G Quantification of synaptic density/mm². Data was normalized to neurons cultured in neuronal media. **, p < 0.05; ****, p-value < 0.0001; one-way ANOVA with Sidak’s multiple comparisons test. H Immunocytochemistry was performed for dendritic marker (MAP2; green) and nucleus (NucBlue). Scale bar represents 50 µm. I Quantification of dendritic length (μm). Data was normalized to neurons cultured in neuronal media. *, p < 0.05; ****, p-value < 0.0001; one-way ANOVA with Sidak’s multiple comparisons test.

Microglia secrete many soluble factors such as chemokines, cytokines, and neurotrophic factors that are involved in neuronal homeostasis, synaptic pruning and plasticity, immune responses, and tissue repair in the CNS [80]. Because of a crucial role for microglia-derived secretory factors in their communication with neurons and other CNS cell types, we next investigated the impact of the MAPT IVS10 + 16 mutation on the microglial secretome. Using a proximity extension assay proteomics platform (OLINK), we assessed the levels of 1360 secretory proteins after QC [81]. Eighty-nine proteins were differentially expressed in MAPT IVS10 + 16 iMGL media compared with isogenic controls (p-value < 0.05, Fig. 5B; Supplemental Table 5). The majority of differentially expressed secretory factors (n = 73) were significantly upregulated in MAPT IVS10 + 16 iMGL media and enriched in pathways associated with NF-κB signaling, T cell immune response, type II interferon production, and complement activation (Fig. 5C). The 16 significantly downregulated secretory factors were enriched in pathways associated with lipid particles, cytokine production, tubulin deacetylation, and secretory granule vesicles (Fig. 5C). To determine whether these pathways mimic changes occurring in cerebrospinal fluid (CSF) from MAPT patients, we examined publicly available aptamer-based proteomics data (SomaScan; 4,138 proteins) from 37 MAPT carriers and 39 controls [82]. Network analyses identified pathways enriched in the extracellular matrix, complement, adaptive immunity, autophagy, and synapse function, among others (Fig. 5D). These pathways are consistent with our molecular and cellular findings in mutant iMGLs in vitro and in the pathways shared in FTLD-tau brains.

Secretory factors derived from microglia conditioned medium have been shown to promote neuronal survival and maturation [83, 84] by regulating synaptic responses, density, and network connectivity in primary neurons cultured in microglia conditioned medium [85–87]. To determine whether media from MAPT IVS10 + 16 iMGL is sufficient to alter healthy neurons, we cultured iPSC-derived neurons driven by dox-inducible Ngn2, which allowed us to generate a homogenous population of excitatory, glutamatergic neurons. After 7 days in culture, iPSC-derived neurons were cultured in naïve neuronal media or media isolated from MAPT IVS10 + 16 or isogenic WT iMGLs for an additional 7 days (Fig. 5E). Neuronal health and synapses were evaluated by immunocytochemistry for MAP2, a neuronal dendritic marker, and synapsin, a synaptic vesicle marker (Fig. 5F). Quantification of synapse density/mm² revealed a significant reduction in synapse density in neurons cultured with MAPT IVS10 + 16 iMGL media compared with neurons cultured with WT iMGL media (Fig. 5F, G). We also observed a decrease in synapse density when comparing the naïve neuronal media-treated cells with the WT iMGL media-treated cells (Fig. 5F, G). Neurons cultured with iMGL media, regardless of mutation status, had significantly longer dendrites (Fig. 5H, I). MAPT IVS10 + 16 iMGL media produced neurons with the longest dendrites (Fig. 5H, I). This is consistent with prior work demonstrating a role of microglia secretory factors in synapse maintenance [83, 84]. Thus, these findings reveal common dysregulated gene/protein expression in MAPT IVS10 + 16 iMGLs and FTLD-patient tissues (brain and CSF) and implicate microglia-derived soluble factors as potential contributors to neurodegeneration in FTLD-tau.

Discussion

Microglia have traditionally been viewed as cells that respond and react to increasing tau aggregates and neuronal damage in tauopathies. Here, we report endogenous expression of MAPT mRNA (which encodes the tau protein) in microglia from human brains. Additionally, we shed light on the cell intrinsic contributions of microglia expressing a FTLD-causing MAPT mutation to disease pathogenesis. We developed a stem-cell-based model to uncover the cell-autonomous molecular effects of MAPT mutations specifically in human microglia. MAPT IVS10 + 16 leads to a shift in microglial transcriptional states. We demonstrate that MAPT IVS10 + 16 causes alterations in cytoskeletal arrangement, defects in phagocytosis of myelin and tau PFFs, and reduced TREM2 signaling. Conditioned media from MAPT IVS10 + 16 iMGLs is sufficient to alter neuronal synapse density and dendritic length. Together, these findings have important implications for our understanding of the early involvement of microglia in the pathogenesis of tauopathies.

In this study, we report that microglia sorted by flow cytometry from a previously analyzed cohort of 10 human brains express MAPT mRNA [38, 39]. While MAPT expression varies among these 10 donors, we found no correlation between this variability and the clinicopathologic status of the brain donors. We also discovered MAPT mRNA and tau protein are expressed in human iMGLs. Notably, mass spectrometry analysis from a pure population of mature iMGLs revealed full length tau protein, consistent with prior reports in iPSC-derived neurons and astrocytes [37, 43]. In tauopathies, microglia have been shown to play an important role in disease progression. Microglia acquire a hypofunctional, senescence-like and paracrine-mediated state upon phagocytosis of neurons burdened with insoluble tau aggregates. Microglia may go on to serve as vectors of tau spread via exosomes due to an inability to break down disease-related tau [10, 22, 23, 25, 29, 66, 67, 88–90]. Additionally, microglia-derived secretory factors and their proinflammatory state may initiate tau aggregation in neurons in tauopathy mouse models [88, 89, 91–93]. However, these studies have relied on mouse models that do not express mutant human tau in microglia, and thus, our knowledge of the role of mutant tau within microglia as a driver of disease and the presence of cell-autonomous defects in microglia has remained limited until now.

We show that the MAPT IVS10 + 16 mutation induces changes in the cytoskeleton of microglia, characterized by the redistribution of the tubulin network towards the cell periphery and a decrease in F-actin levels. Tau is associated with the regulation of cytoskeletal organization and dynamics, acting as a connector between the actin and tubulin cytoskeleton during processes like axon outgrowth and growth cone turning in cortical neurons [52, 53, 55, 94]. Microtubule remodeling is critical for microglial activation [95, 96]. Interestingly, COS-7 fibroblasts expressing a phosphomimetic form of tau (Y18E) also undergo microtubule redistribution, affecting endosomal and lysosomal motility [97]. Actin dynamics are crucial for microglial functions such as phagocytosis [98, 99], suggesting that the reduced phagocytic capacity of MAPT IVS10 + 16 iMGLs could be linked to lower F-actin levels in these cells. Thus, future studies are required to elucidate the mechanisms by which tau influences cytoskeletal network remodeling in microglia and its impact on microglial phagocytosis.

Our discovery that MAPT IVS10 + 16 leads to downregulation of the TREM2-TYROBP signaling network and a decrease in TREM2 expression at the gene, membrane and secreted protein level suggests tau impacts a major microglia signaling molecule implicated in FTLD-tau and secondary tauopathies such as Alzheimer’s disease (AD). TREM2 signaling has been shown to play a key role in connecting several microglial functions such as cell survival, proliferation, phagocytosis, endolysosomal pathways, metabolism, and inflammation [68, 75, 100]. Homozygous mutations in the gene encoding TREM2, an activating receptor exclusively expressed in microglia in the CNS, are clinically associated with FTLD-like syndrome [101]. TREM2 plays a crucial role in microglial phagocytic function [23, 98, 100]. TREM2 missense mutations linked to FTLD and FTLD-like syndrome lead to impaired TREM2 trafficking to the cell surface and reduced microglial phagocytosis, highlighting a direct association between mature membrane-bound TREM2 levels and microglial phagocytic activity [102]. Studies leveraging TREM2 loss-of-function mutations in mouse models of tauopathy point to a role for TREM2 signaling in the regulation of amyloid accumulation, tau pathology, and neuroinflammation in AD. [23, 103]. However, whether TREM2 plays a neuroprotective or neurotoxic role in tau pathology may depend on the disease stage in animal models [104–107].

Among the many roles of TREM2 in microglia, mounting evidence points to a role in regulating the endolysosomal network and sustaining microglial metabolic balance through mTOR signaling [77, 78]. Previously, we have shown that TREM2 loss-of-function mutation in iMGLs leads to lysosomal dysregulation [77]. Consistent with our prior work, transcriptomics analyses in MAPT IVS10 + 16 iMGLs revealed a downregulation of genes involved in the late endosomal/lysosomal processes. Interestingly, we also observed downregulation of MTOR, a sensor of cellular energy and homeostasis, in MAPT IVS10 + 16 iMGLs. TREM2-dependent mTOR signaling is crucial to sustain biosynthetic and energy metabolism in microglia [78] and consistent with this, we observed a dysregulation of genes involved in glycolysis and a selective upregulation of genes involved in mitochondrial oxidative phosphorylation in MAPT IVS10 + 16 iMGLs. Maintenance of lysosomal homeostasis is crucial to controlling microglial mTOR levels and initiating immune responses to tau pathology in the tau-P301S tauopathy mouse model [108]. Microglia activation also induces mitochondrial alterations and initiates a metabolic shift from oxidative phosphorylation to glycolysis [109]. iMGLs carrying hypofunctional TREM2 risk variants (e.g. R47H) exhibit deficiencies in their ability to transition from oxidative phosphorylation (OXPHOS) to glycolysis, along with reduced mitochondrial respiration rates [110]. Our findings point to a role for MAPT IVS10 + 16 in suppression of TREM2 and associated gene networks that regulate cell survival, proliferation, phagocytosis, endolysosomal pathways, metabolism, and inflammation. Further studies are needed to understand the molecular mechanisms involved and to explore whether activating TREM2 signaling will rescue the observed defects in phagocytosis, cytoskeletal network, and metabolism in the MAPT IVS10 + 16 iMGLs.

Here, we describe broad alterations in the MAPT IVS10 + 16 iMGL secretome. Both actin and microtubule dynamics are involved in microglial secretory function [95, 99]. MAPT IVS10 + 16 iMGLs exhibit less secreted lipoprotein lipase (LPL), a protein associated with microglial phagocytosis [111]. Knockdown of microglia specific Lpl in a mouse model of obesity, resulted in overall decreased microglial reactivity, phagocytic ability, mitochondrial disturbances, and an inability to transition to an adaptive immunometabolic state [112]. Microglia and complement signaling play crucial roles in synapse elimination during developmental pruning and disease, sometimes involving aberrant phagocytosis of synapses and viable neurons [26, 113–115]. Interestingly, we found that neurons treated with MAPT IVS10 + 16 conditioned microglial media had fewer synapses. This could be driven by several secretory factors that are altered in the MAPT IVS10 + 16 microglia. MASP1, an activator of the lectin pathway of complement proteins, is upregulated in MAPT IVS10 + 16 iMGLs secretome. MASP1 has been implicated in loss of presynaptic inputs in axotomized spinal motoneurons following peripheral nerve injury [116]. While complement-mediated synapse elimination has been observed primarily through microglial contact-dependent phagocytosis of synaptic terminals [117], several studies have demonstrated that microglia can also modulate neuronal synaptic density and network connectivity through soluble secretory factors [85–87]. Microglial activation and synaptic loss occur prior to the accumulations of fibrillary tau tangles in the P301S tauopathy mouse model [118]. Microglia-derived secretory factors can trigger tau pathology in neurons [30, 92]. Thus, our finding that synaptic density is altered when neurons are exposed to MAPT IVS10 + 16 conditioned media suggests a pivotal role for microglia in tauopathy pathophysiology. Further work is required to understand whether tau is required for these effects and whether the secreted proteins that are changed due to MAPT IVS10 + 16 may serve as potential new biomarkers of microglial dysfunction in tauopathies.

In this study, we have developed a stem-cell-based model to uncover the cell-autonomous molecular effects of MAPT mutations specifically in human microglia. iPSC-based microglial modeling is a powerful cellular platform that captures endogenous human tau expression to explore the molecular, cellular, and disease-relevant phenotypes resulting from FTLD-causing MAPT genetic mutations. By studying these changes in human microglia, we can better identify their contributions to tauopathies and other neurodegenerative diseases.

Experimental procedures

Patient consent

The informed consent was approved by the Washington University School of Medicine Institutional Review Board and Ethics Committee (IRB 201104178 and 201306108). The University of California San Francisco Institutional Review Board approved the operating protocols of the UCSF Neurodegenerative Disease Brain Bank (from which brain tissues were obtained). Participants or their surrogates provided informed consent for autopsy, in keeping with the guidelines put forth in the Declaration of Helsinki, by signing the hospital’s autopsy form. If the participant had not provided future consent before death, the DPOA or next of kin provided it after death. All data were analyzed anonymously.

Induced pluripotent stem cell (iPSC)

Human iPSC lines were generated using non-integrating Sendai virus in dermal fibroblasts as previously described [40, 79]. MAPT IVS10 + 16 mutation carrying GIH36C2 (FTD1) and GIH178C1 (FTD2), and their corresponding CRISPR-Cas9 corrected isogenic control lines, GIH36C2Δ1D01 and GIH178C1Δ2A10, were used in this study and donor-specific information are provided in Supplemental Table 1. Human iPSCs were cultured in mTesR1 on Matrigel-coated tissue culture-treated plates using Accutase™ (ThermoFisher Scientific, Cat#: NC9464543) for dissociation during routine passaging. Each line was analyzed for pluripotency markers by immunocytochemistry (ICC) and quantitative PCR (qPCR); for spontaneous differentiation into the three germ layers by ICC and qPCR; and for chromosomal abnormalities by karyotyping. Mutation status was confirmed in iPSC by Sanger sequencing (Supplemental Fig. 1). Predicted off-target editing events were evaluated using Sanger sequencing and confirmed to be similar to the unedited donor sequence (Supplemental Fig. 6A). Allelic drop out in the CRIPSR-corrected, isogenic MAPT WT/WT controls was evaluated by quantitative PCR of gDNA (qgPCR) using Kasp genotyping as previously described [119, 120]. Isogenic lines were found to carry two WT alleles, similar to an unrelated WT donor (Supplemental Fig. 6B). Karyotyping and Sanger sequencing to verify mutation status were performed every 20 passages. iPSCs were maintained with less than 5% spontaneous differentiation.

Generation of hematopoietic progenitor cells from iPSCs

Human iPSCs were differentiated into microglia using a two-step protocol modified from previous reports. To generate hematopoietic progenitor cells (HPCs), a STEMdiff Hematopoietic kit (STEMCELL Technologies, Cat#: 05310) was used following manufacturer’s instructions as previously reported [41, 42, 77]. Briefly, iPSCs were detached with ReLeSR™ (STEMCELL Technologies, Cat#: 05872) and passaged in mTeSR1 supplemented with Rock inhibitor, Y-27632 Dihydrochloride (STEMCELL Technologies, Cat #72302) to achieve a density of 45–80 aggregates/well. On day 0, cells were transferred to Medium A from the STEMdiff Hematopoietic Kit to pattern iPSCs towards mesoderm. On day 3, mesodermal cells were exposed to Medium B to further promote differentiation into HPCs, and cells remained in Medium B for 10 additional days. After 12 days in culture, HPCs were collected for fluorescence activated cell sorting (FACS) to enrich for CD43 + CD34+ HPCs.

FACS for HPC generation and banking

For sorting of CD43 + CD34+ HPCs, HPCs were collected from the floating population with a serological pipette and from the adherent population after incubation with Accutase at 37 °C for 20 mins. Cells were filtered through 70 μm filters to remove large clumps, then washed with FACS buffer (2% FBS in 1x Dulbecco’s Phosphate Buffered Saline (DPBS), modified without calcium chloride and magnesium chloride, sterile-filtered, Sigma-Aldrich D8537; 300 g at 4 °C for 5 mins), then immunolabeled in the dark for 20 mins at 4 °C in FACS buffer containing the following antibodies: CD43-APC (Biolegend, Cat#: 343206, Clone: 10G7), CD34-FITC (Biolegend, Cat#: 343504, Clone: 581), CD45-Alexa 700 (Biolegend Cat#: 304024; Clone: H130). After immunolabeling, HPCs were washed once with FACS buffer and then resuspended in ~300 µL FACS buffer and filtered through 5 mL polystyrene round-bottom tube with cell-strainer cap (Falcon, Cat#: 352235) and sorted with a Becton Dickinson FACSAria II cell sorter. All steps were performed on ice and using pre-chilled reagents and a refrigerated centrifuge. After sorting, HPCs were spun down at 300 g at 4 °C for 10 mins and the supernatant was aspirated. Cell counts were determined using Invitrogen Countess 3 automated cell counter and the cell pellet was then frozen in 1 mL CryoStor® CS10 (STEMCELL Technologies, Cat#: 07930) for every 1 × 10⁶ HPCs and stored in liquid nitrogen until use.

Differentiation of HPCs into microglia-like cells (iMGLs)

For iMGL differentiation, HPCs were thawed into 5% FBS in DMEM/F12 and centrifuged at 300 g for 6 mins at room temperature (RT). The resulting cell pellet was resuspended in iPSC-microglia medium comprised of DMEM/F12 (Sigma-Aldrich, cat#: D8437), 2X insulin-transferrin-selenite (Gibco, cat#: 41400-045), B-27™ Supplement (50X, 1: 25, Life Technologies, cat#: 17504-044), N-2 Supplement (100X, 1:200, ThermoFisher Scientific, cat#: 17502001), glutamax (1X, ThermoFisher Scientific, cat#: 35050061), 1X non-essential amino acids (Thermo Scientific, cat#: 11140050), 400 mM monothioglycerol (Sigma Aldrich, cat#: M1753), 5 µg/mL human insulin (recombinant, Sigma Aldrich, cat#: I2643), and freshly supplemented maintenance cytokines (100 ng/mL IL-34 [PeproTech, Cat#: 200-34], 50 ng/mL TGFβ1 [PeproTech, Cat#: 100-21], and 25 ng/mL M-CSF [Peprotech, Cat #:300-25]). Cells were then maintained on Matrigel coated 6-well plates. On days 6 and 12 post-HPC generation, iMGLs were split based on cell density per well. Floating and adherent cells from each well were evenly transferred from three wells into one new well to refresh Matrigel and promote proliferation during this phase. On day 25 post-HPC generation, iMGLs were replated onto Matrigel-coated tissue culture in microglia media supplemented with maturation cytokines (100 ng/mL CD200 [Novoprotein, Cat#: C311] and 100 ng/mL CX3CL1 [Peprotech, Cat#: 300-31]) [77]. All experiments described in this study were performed in mature iMGLs harvested on day 28 post-HPC generation.

Assessment of iMGL cell surface marker expression by flow cytometry

To evaluate cell surface markers, iMGLs (28 days post-HPC generation; 100,000 cells) were treated with Zombie Aqua™ dye (Fixable Viability kit, BioLegend, Cat# 423102) for 20 mins in PBS (1:1000) followed by a wash with FACS buffer. iMGLs were incubated for 10 min on ice with anti-CD16/CD32 to block Fc receptors (1:50; Miltenyi Biotec, Cat #:120-000-442). Then, iMGLs were stained with CD11b-PE/Cy7 (Cat#: 101216; Clone: M1/70), CD45-Alexa Fluor 700 (Cat#: 304024; Clone: H130), TREM2-APC (R&D, FAB17291A) for 20 min at 4 °C in the dark followed by washing with FACS buffer. Cell surface markers were acquired on a BD X-20 analyzer. The resulting data were analyzed with FlowJo software (FlowJo 10.8.1).

Immunoblotting

iMGLs from two independent rounds of differentiation were lysed in 1x RIPA lysis buffer (150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris pH-7.4) containing protease inhibitor cocktail (1:500, Sigma, Cat#: P8340) and phosphatase inhibitor (PhosSTOP™, Roche, Cat#: 04906837001) and incubated on ice for 20 mins. Lysates were then centrifuged at 14000 g for 20 min at 4 °C, and the supernatant was isolated for subsequent analyses. To evaluate total protein concentration, a BCA assay (Pierce™ BCA Protein Assay Kits ThermoScientific, Cat#: 23225) was performed. SDS-PAGE was performed using 20 µg of iMGL cell lysates mixed with 4 × Laemmli sample buffer (Bio-Rad, Cat#: 161-0747) and 10% β-mercaptoethanol and heated at 95 °C for 10 min. A AnykD™ Criterion™ TGX™ Precast Midi Protein Gel (BioRad, Cat#5671123) was used. Proteins were transferred to PVDF membrane and blocked for 1 hr at RT in 5% milk in phosphate buffered saline with 0.1% Tween 20 (PBS-T). Membranes were probed with the mouse anti-Tau5 (a generous gift from Binder lab), P2RY12 (1:100, Sigma-Aldrich, Cat# HPA014518), and GAPDH (1:500, Thermo Fisher Scientific, Cat# MA5-15738, RRID: AB_10977387) overnight at 4 °C. Membranes were subsequently washed and incubated in affiniPure Goat anti-mouse HRP (1:2000, Jackson Immuno Research Labs, Cat# 115-035-174, RRID: AB_2338512) for 1 hr at RT, washed, and developed using SuperSignal West Pico PLUS (ThermoScientific, Cat#: 34580) on a Bio-Rad Chemidoc Imaging System.

Immunoprecipitation and mass spectrometry analyses

iMGLs and iPSC-derived neurons were homogenized in 300 µL 50 mM PBS containing 1x Protease inhibitors (Roche) for 10 mins. Lysates were sonicated on ice and centrifugated at 10,000 x g for 20 mins at 4 °C to isolate cell lysate. Total protein concentration was determined using the BCA assay method. Tau was then immunoprecipitated using a mixture of Tau1, HJ8.5, and HJ8.7 antibodies, following a previously described protocol with adjustments [121–123]. For each sample, 2.5 ng of 15 N labeled recombinant 2N4R tau isoform (0.5 ng/µL, a gift from Dr. Guy Lippens, France) was added to 30 µg total protein. Soluble tau was immunoprecipitated in a solution containing detergent (1% NP-40), 5 mM guanidine hydrochloride, and protease inhibitors (Roche), using a tau antibody-conjugated sepharose bead cocktail (50% slurry containing 3 µg antibody/mg beads). The immunoprecipitation was conducted by rotating the samples overnight at 4 °C. After incubation, sepharose beads were washed with 25 mM TEABC buffer. Immobilized proteins were digested on the beads overnight (16–18 h) using 0.4 µg trypsin at 37 °C. The digested peptides were diluted with 300 µL of a 3% oxidation solution (Formic acid + H₂O₂) and incubated overnight at 4 °C. The digested and oxidized samples (soluble tau) were desalted using the Oasis µElution HLB plate (Waters^TM Corp) following the manufacturer’s protocol. The eluent was lyophilized and reconstituted with 25 µL of 2% ACN, 0.1% FA in water prior to mass spectrometry analysis on Vanquish Neo UHPLC (Thermo Scientific) coupled to Orbitrap Exploris 480. Trypsinized tau peptides were quantified by comparing with the corresponding isotopomer signals from the 15 N internal standard using Skyline software (version 20.2, MacCoss Lab, Department of Genome Sciences, University of Washington).

Tau peptides as prepared above were loaded directly onto an HSS T3 75 μm × 100 μm, 1.8 μm C18 column (Waters) at a temperature set to 65 °C using a Vanquish Neo UHPLC (Thermo Fischer Scientific, San Jose, CA, USA). Peptides were separated with a flow rate of 0.4 µL/min using buffer A (0.1% FA in water) and buffer B (0.1% FA in acetonitrile). Tau peptides eluted from the column with a gradient of 4–8% buffer B for 10 min, followed by 8–20% buffer B for another 8 min, ramping up to 95% buffer B in the next 2 min, and column cleaning for an additional 2 min. The Thermo Orbitrap Exploris 480 was equipped with a Nanospray Flex electrospray ion source (Thermo Fisher Scientific, San Jose, CA, USA) and operated in positive ion mode. Peptide ions sprayed from a 10 μm SilicaTip emitter (New Objective, Woburn, MA, USA) into the ion source (spray voltage = 2200 V) were targeted and isolated in the quadrupole. Isolated ions were fragmented by HCD, and ion fragments were detected in the Orbitrap (resolutions of 15000, 30000, and 60000 for N-terminal, mid-domain, and MTBR tau peptides, mass range of 150–1,500 m/z).

RNA-sequencing

RNA was extracted from iMGLs using RNeasy Mini Kit (Qiagen, Cat#: 74104). Total RNA integrity was determined using Agilent Bioanalyzer or 4200 Tapestation. Library preparation was performed using Ribo-ZERO kits (Illumina-EpiCentre). RNA was sequenced on an Illumina NovaSeq-6000 using paired end reads extending 150 bases. RNA-seq reads were then aligned and quantitated to the Ensembl release 101 primary assembly with an Illumina DRAGEN Bio-IT on-premises server running version 3.9.3-8 software.

Principal component and differential expression analyses

Transcriptome analyses were based on 12,272 protein-coding genes. These genes were expressed in ≥50% of the biological replicates analyzed in this research and with one or more counts per million (CPM). Principal component analyses (PCA) were calculated using a subset of the 500 most variable genes based on regularized-logarithm transformation (rlog) counts. PCA and Volcano plots were created using ggplot2 [124]. Differential gene expression was performed using the DESeq2 for R package 4.3.1 [125]. Unadjusted p-value (≤0.05) and Benjamini-Hochberg adjusted p-value (FDR ≤ 0.05) corrections were applied. Genes passing the FDR ≤ 0.05 threshold were analyzed by EnrichR and ToppGene [126, 127].

Human brain RNA-sequencing dataset

The bulk RNA-seq dataset for FACS-isolated microglia from ROSMAP was downloaded from the AMP-AD Knowledge Portal and processed as previously described [38]. Briefly, microglia were collected from postmortem human brains [38]. Gray matter was mechanically homogenized followed by myelin depletion with antibody-based beads [38]. The resulting fraction was then enriched for CD11b positive cells using antibody-based beads, and cells were isolated by FACS using CD11b and CD45 [38]. Library construction was performed using SmartSeq-2 [128]. These isolated microglia were demonstrated to be free of CD45 high macrophages, infiltrating monocytes, and other potentially contaminating cells from the blood. RNA-seq of the FACS-isolated microglia demonstrated high expression of typical microglial genes (CX3CR1, TREM2, P2RY12, GPR34, TMEM119) and an absence of markers for other blood cells (GYPA, ITGA2B, GP1BA, SELP, CD3D, CD3E, CD3G, NCAM1, CD19, MS4A1, SDC1, CCR2, ENPP3, CD34, MCAM). Ten samples, AD (N = 4) and healthy control (N = 6), were included in the analysis. The sample size, brain region, age, sex, postmortem interval (PMI), and disease status are summarized in Supplemental Table 1.

To determine whether the differentially expressed genes in the iMGLs capture molecular processes that occur in primary tauopathies, we analyzed gene expression in human brain RNA-seq datasets (Supplemental Table 1) from the middle temporal gyrus from MAPT IVS10 + 16 mutation carriers (2 samples) and healthy controls (3 samples) [79]. Differential expression analyses were performed including sex, age-at-death, RNA integrity number (RIN), and brain tissue source as covariates.

Immunocytochemistry

iMGL (50,000 cells per well) were seeded onto 8-well chamber slides (Millicell EZ SLIDE, Cat#: PEZGS0816) on day 25 post HPC generation in iMGL basal medium with maintenance and maturation cytokines. iMGLs were fed on alternate days until 28 days post HPC generation. Cells were washed one time with 1x DPBS and fixed with paraformaldehyde (4% w/v) for 20 min at RT followed by three washes with 1x DPBS. Cells were treated with blocking buffer (1x DPBS with 0.3% Triton X-100 and 3% BSA) for 1 hr at RT followed by permeabilization in 0.5% Triton X-100 in 1x DPBS for 15 mins. Primary antibodies (see below) were then added at specified dilutions in blocking buffer and placed at 4 °C overnight. Cells were then washed 3 times with 1x DPBS for 5 min each wash then stained with donkey Alexa Fluor conjugated secondary antibodies (Invitrogen,1:500) in blocking buffer for 2 h at RT in the dark. After secondary antibody staining, cells were washed three times with 1x DPBS PBS and cover slipped with Fluoromount-G™ (ThermoFisher).

Primary antibodies were used at the following dilutions in this study: rabbit anti-TMEM119 (1:100, Abcam, Cat#: ab185333), goat-anti IBA1 (1:500, Abcam, Cat#: ab5076), goat anti-TREM2 (1:100, R&D, Cat#: AF1828), rabbit α-tubulin (1:100, Abcam, Cat#: ab18251), mouse anti-NeuN (1:1000, Millipore Sigma, Cat#: MAB377), and Alexa Fluor™ Plus 647 Phalloidin (1:400, Invitrogen, Cat#: A30107). Nuclei were stained with NucBlue™ Fixed Cell ReadyProbes™ Reagent (DAPI; ThermoFisher Scientific, Cat# R37606). Images were acquired in the Nikon AX-R with NSPARC (Nikon’s Spatial Array Confocal detector), a laser scanning confocal microscope. Images were acquired using a 60x, 4-7x zoomed Z-stack images were acquired in the AX NSPARC mode. Image analyses were performed in Fiji. Representative images shown in figures are projected into max intensity z plane and scale bar is as indicated in respective figure legends. The number of cells analyzed for each staining are as indicated in figure legends. Six to eight confocal images acquired at 60x across two independent wells of a 8-well chamber slide were analyzed for quantification of α-tubulin quantification shown in Fig. 2C. Three to four independent wells with five to thirty cells per well were analyzed for F-actin quantification shown in Fig. 2E.

Phagocytosis assays

iMGLs (50,000 cells) were seeded into Matrigel-coated flat bottom 96-well plate on day 25 post-HPC generation in iMGL basal medium with maintenance and maturation cytokines and fed on alternate days until day 28. On day 28 post-HPC generation, iMGLs were treated with phagocytic substrates such as human myelin (20 µg/ml, prepared as previously reported in [77]), tau preformed fibrils (Human Recombinant Tau-441 (2N4R) P301S mutant protein pre-formed fibrils ATTO 488, TAU-PFF, StressqMarq, Cat# SPR-329-A488, 500 nM), or Alexa Fluor™ 488 conjugated Zymosan A (S. cerevisiae) BioParticles™ (ThermoFisher Scientific, Cat#Z23373, 0.5 µg/ml). Tau-PFF was prepared for treatment by sonication in water bath at 20 A for 3 cycles of 30 s on and 30 s off. Imaging was performed hourly for 24 h following substrate addition using an Incucyte S3 Live Cell Analysis Instrument (Sartorius). Eight biological replicates per donor line per treatment group were used in each experiment. Data analyses were performed using the IncuCyte S3 software v2021A and bar graphs to represent specific time-points were generated in GraphPad Prism 10 (version 10.0.3).

Tau uptake - flow cytometry

On day 25 post-HPC generation, 100,000 iMGLs were seeded in Matrigel-coated 24-well tissue culture plates in iMGL basal medium with maintenance and maturation cytokines and fed on alternate days until day 28. iMGLs were either treated with PBS vehicle or 50 nM TAU-PFF (sonicated as described above) on day 28 for 24 h. Cells were then harvested on ice from the 24-well plates, washed thrice with 1x DPBS, and replated into U-bottom 96-well plate for flow cytometry staining. iMGLs were first treated with Zombie Aqua™ dye (Fixable Viability kit, BioLegend, Cat# 423102) for 20 mins in PBS (1:1000) followed by washing with FACS buffer and centrifugation at 500 g for 5 mins. iMGLs were then processed and stained with CD45-Alexa Fluor 700, CD11b-PE/Cy7 for defining iMGLs by flow cytometry. iMGLs were then run on a BD X-20 analyzer and acquired for the stained markers, and TAU-PFF ATTO 488 staining was quantified in CD45 + CD11b+ iMGLs. The resulting data were analyzed with FlowJo software from BD (FlowJo 10.8.1). Bar graphs for quantitative representation of flow cytometry data were generated in GraphPad Prism 10 (version 10.0.3).

sTREM2 ELISA

On day 25 post-HPC generation, iMGLs were seeded into 6-well plates (500,000 cells/well) in iMGL basal medium with maintenance and maturation cytokines and fed on alternate days until reaching maturation on day 28. On day 28 post-HPC generation, supernatants from iMGL culture were harvested, protease and phosphatase inhibitors were added. Samples were analyzed for sTREM2 using a commercially available Human TREM2 ELISA Kit (Abcam, Cat#: ab224881) following manufacturer’s instructions. Briefly, samples were diluted 1:5 in sample diluent and 50 µl samples were added to each well of pre-coated microplate strips. Three biological replicates were run along with technical duplicates. Plates were read using 450 nm using a BioTek Microplate spectrophotometer.

Proteomics

Media from iMGL cultures were analyzed by OLINK Explore 3072 platform, a high-throughput, multiplex protein immunoassay-based proximity extension assay including a protein library of 2944 protein assays [81]. iMGL basal medium with maintenance and maturation cytokines served as negative controls for assessing background level of the proteins measured. Briefly, target proteins are bound by unique pairs of oligonucleotide-labelled antibodies. When in close proximity to one another, the oligonucleotides hybridize to form a PCR target sequence. The resulting DNA amplicon is quantified on an Illumina NovaSeq platform using Next Generation Sequencing. Protein levels derived from cycle threshold values are directly proportional to the concentration of target protein in the sample and are reported on a log2 scale as normalized protein expression (NPX). The technical QC was passed for 2902 of the 2944 assays. Using our standard proteomics data analyses pipeline [81], low detection proteins were filtered out. Proteins with detectability in all of the mutant samples, all of the WT samples, or in all samples were included in the analyses (n = 1490). Proteins were excluded if the average coefficient of variation (CV) was above 20% within the technical triplicates (n = 1360 protein assays). We averaged the triplicate measurements for each sample and compared the expression of secretory factors between mutation carriers vs isogenic control WT iMGL cultures by paired t-tests. Significant secretory factors (p-value ≤ 0.05) were categorized into those upregulated or downregulated and functional annotations of these factors into pathways were derived from EnrichR gene list enrichment analysis tools [126].

iMGL conditioned media treatment on iPSC-neurons

Human iPSCs from a normal, healthy control donor were engineered to introduce a dox regulatable form of Ngn2 (F12468.13.CRISPRiΔ1E11.NgN2Δ3B09) in the AAV safe harbor locus as previously described [129]. Engineered iPSCs were characterized for pluripotency markers and stable karyotype. To differentiate iPSCs into cortical neurons, iPSCs were plated at a density of 65,000 cells per well in neural induction media (StemCell Technologies, Cat #05839) in a 96-well v-bottom plate to form neural aggregates. After 5 days, cells were transferred into culture plates. The resulting neural rosettes were isolated by enzymatic selection (Neural Rosette Selection Reagent; StemCell Technologies, Cat #05832) and cultured as neural progenitor cells (NPCs). NPCs were differentiated into neurons upon 1 µg/mL doxycycline induction (doxycycline hyclate, Sigma, cat#: D5207-10G) in BrainPhys™ Neuronal Media (STEMCELL Technologies, cat#: 05790) supplemented with BDNF (20 ng/ml, PeproTech, 450-02-100 µg), GDNF (20 ng/ml, PeproTech, 450-10-100ug), cAMP (0.5 mM, Sigma, cat#: D0627), glutamax (1X, ThermoFisher Scientific, cat#: 35050061), B-27™ Supplement (50X, 1: 100, Life Technologies, cat#: 17504-044), N-2 Supplement (100X, 1:200, ThermoFisher Scientific, cat#: 17502001), and pen/strep (ThermoFisher Scientific, 15140-122). Neurons were then replated on coverslips in 24 well plate tissue culture dishes, coated with Poly-l-ornithine (0.01%, R&D Systems, cat#: 3436-100-01; coating overnight) and laminin I (10 µg/ml, R&D Systems, cat#: 3400-010-03; coating for 2 h) at 5 × 10⁴ cells/well. On day 7, neuron media was removed and replaced with naïve neuronal conditioned media or conditioned media from isogenic MAPT WT or MAPT IVS10 + 16 iMGL cultures (400 µl/well). Conditioned media was harvested from day 28 mature iMGL cultures and centrifuged at 300 g for 5 mins to remove cells and was aliquoted and stored in −80 °C. One day prior to day 7 of neuron culture, required volume of iMGL conditioned media was thawed overnight at 4 °C. On day 7, media was prewarmed to RT and then added to neuronal cultures. Neurons were cultured for an additional 7 days with media changes every 2 days.

On day 14, neuronal cells were first fixed with 4% paraformaldehyde followed by three washes with 1X DPBS and permeabilized with 0.1%Triton X-100 and then incubated for 1 hr in blocking buffer (1X DPBS with 0.1% saponin, 1% BSA, and 5% FBS). Neurons were then incubated in blocking buffer containing primary antibody (see below) overnight at 4 °C. After washing 3 times with 1x DPBS, secondary antibodies were added in blocking buffer for 1 hr at RT in the dark. Antibodies used for immunocytochemistry of neurons included synapsin-1(Millipore, cat #: AB1543P, 1:500), MAP2 (Abcam cat #: ab5392, 1:500), and goat Alexa Fluor conjugated secondary antibodies (Invitrogen,1:1000). Coverslips were mounted on slides with Fluoromount-G™ (Thermo Fisher Scientific, 00-4958-02). Images were acquired with an LSM980 confocal system (Zeiss) with Airyscan and a 20x or 63x oil-immersion objective lenses.

Synapse density shown in Fig. 5G was quantified as the number of synapsin puncta per mm² of field area (number of cells analyzed for Synapsin-1 puncta density per image field; neuron media, n = 103; WT iMGL media, n = 72; IVS10 + 16 iMGL media, n = 87). Morphometric analyses were performed on 40x confocal images, 3D reconstructed in Imaris software, and semi-automatically analyzed for mean dendritic length (number of neurites analyzed; neuron media, n = 172; WT iMGL media, n = 158; IVS10 + 16 iMGL media, n = 105, respectively).

Disclosures

C.S. may receive income based on technology licensed by Washington University to C2N Diagnostics.

Author contributions

Designed experiments: AKI, LV, CMK. Performed and analyzed experiments: AKI, LV, FSM, MM, MA, RKK, AR, SY, EPD, KMS, AV, GG, SML, GH, KJM, JM, CET, CS, CMK. Provided experimental tools: ALN, SS, WWS, LTG, ST. Provided funding: CMK. Wrote the manuscript: AKI, CMK. Revised and approved manuscript: AKI, LV, FSM, MM, MA, RKK, AR, SY, EPD, KMS, AV, GG, SML, GH, KJM, JM, ALN, SS, WWS, LTG, ST, CET, CS, CMK.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41380-025-03073-2.