Intracellular Responses of Glial Cells To Monomeric Tau Protein Are Closely Mediated By Heparan Sulfate Proteoglycans (HSPGs)

The conversion of soluble tau protein to insoluble, hyperphosphorylated neurofibrillary tangles is a major hallmark leading to neuronal death observed in neurodegenerative tauopathies. Recent work suggests that extracellular, soluble tau binds to negatively charged heparan sulfate proteoglycans (HSPGs) available on the cell surface. In addition, LRP1 has recently been recognized as a major tau receptor, mediating tau uptake and spread. We hypothesized based on this data that monomeric tau would be endocytosed in both an HSPG- and LRP-dependent manner, activating intracellular signaling pathways that would regulate cellular phenotypes. Using live-cell confocal microscopy and flow cytometry, we show that soluble 0N4R monomers were rapidly endocytosed by SH-SY5Y and C6 glioma cells, via actin-dependent macropinocytosis. We also demonstrated the crucial role of HSPGs and LRP1 in cellular endocytosis of monomeric tau by observing reduced tau uptake in C6 glial cells with genetic knockouts of xylosyltransferase-1 – a key enzyme in HSPG synthesis – and LRP1. An ERK1/2 inhibition experiment showed that inhibiting the MEK-ERK1/2 signaling pathway attenuated IL-6 and IL-1β gene expression but not TNF-α . An LRP1 knockdown experiment led to an attenuated propensity for tau uptake and further elevated IL-6 gene expression. Collectively, our data suggest that tau has multiple extracellular binding partners that mediate its internalization through distinct mechanisms. Additionally, this study demonstrates the important role of both HSPG and LRP1 in regulating cellular immune responses to tau protein monomer, which provides a novel target for alleviating the neuroinflammatory environment before the formation of neurofibrillary tangles. tangles; Xylt1, xylosyltransferase-1; FTD, frontotemporal dementia; CNS, central nervous system; Aβ, β-amyloid peptide; ISF, interstitial fluid; CSF, cerebrospinal fluid; HSPG, HS, heparan LRP1, lipoprotein receptor-related protein 1; MT,


Introduction
More than twenty-five neurological disorders feature filamentous aggregates of hyperphosphorylated tau protein in the brain and are collectively known as neurodegenerative tauopathies [1]. Progressive deposition of tau-containing neurofibrillary tangles (NFTs) throughout various brain regions is a well-characterized hallmark of tauopathies; however, the mechanism behind NFT formation and spread remains poorly understood. Tau binds and stabilizes axonal microtubules (MTs) but can be hyperphosphorylated and lose its affinity for MTs [2]. MAPT gene mutation has also been proposed to explain the impaired ability of tau to promote the assembly of tau and MTs [3], leaving it available to interact with other cellular components. Indeed, in the past few decades many additional binding partners, subcellular locations, and physiological roles of tau have been identified [4].
Increasing evidence suggests that tau can be physiologically released into the extracellular space, both as monomers and aggregate forms. Tau has been found in both the interstitial fluid (ISF) and cerebrospinal fluid (CSF) in both humans and wild-type rodents, suggesting that tau release can occur in the absence of neurodegeneration [5][6] [7][8] [9] [10]. Many species of tau have been characterized outside of cells including free monomer, C-terminal truncations, low molecular weight aggregates, fibrils, or membrane-encapsulated tau [11][12][13] [14]. Interestingly, endogenous tau was released in an activitydependent manner from primary cortical cultures [15]. Several cell culture studies show that tau monomer and oligomer in various phosphorylation states are constitutively released into the extracellular space [12] [16] [17]. Tau monomer in the ISF may eventually have pathogenic consequences as tau monomers are able to adopt both an inert and a seed-competent conformation, while some seed-competent monomers can form multiple unique pathological strains [18] [19]. Furthermore, full-length, soluble tau monomer added externally to culture media eventually forms aggregates inside cells [20]. Taken together, these findings suggest that cellular uptake of physiologically released tau monomer may represent an initial step in neurodegenerative tauopathies, where cellular uptake provides the initial conditions for aggregation.
Several groups have demonstrated that full-length monomer is readily internalized through clathrin-protein and could have increased pathogenic consequences on the cell by converting wild-type tau into a more pathological relevant tau protein seed, leading to the subsequent pathological transmission, as discussed in one of our previous papers [28]. This possibility also merits further investigation. Recently, the involvement of astrocytes in the progression of tau pathology has drawn much attention because of their close communication with neuronal cells [28][29] [30]. Astrocytes, like microglia, express genes involved in phagocytosis [31], eliminate synaptic debris [32], or protein aggregates as seen by clearance of β-amyloid peptide (Aβ) [33]. For instance, the enhanced cellular uptake of tau fibrils by astrocytes has been related to the intensive autophagy-lysosomal pathway as reported in several tau pathology mouse models. Moreover, astrocytes develop more neurotoxic features by transforming into reactive astrocytes, which are induced by activated microglia and neuroinflammation in various human neurodegenerative disorders at a more advanced stage [34]. However, the cellular responses of astroglial cells to tau monomers have been less investigated. Most importantly, astrocytic LRP1 has been implicated in Aβ cleavage during the progression of Alzheimer's Disease (AD), while the mechanism by which LRP1 regulates cellular uptake of tau monomers in glial cells requires further investigation.
In this study, cellular endocytosis of tau monomer was examined in several cell lines, as well as neuronal and glial cells. To investigate the molecular contributors to tau endocytosis and altered intracellular activities, we subsequently examined tau uptake propensity and the resulting cellular signaling pathways in HSPG and LRP1-deficient glial cells. Findings from this study provide novel information for the progression of tau protein-related pathology and shed light on exacerbated pro-inflammatory signaling pathways affected by tau propagation that may provide a therapeutic target.

Live Cell Confocal Microscopy
All cells were plated at 50,000 cells/well in Lab-Tek II confocal imaging chambers (Thermo-Fisher, cat# 155360) and cultured for 24-48 hours in incubators prior to assay. On the day of assay, cell chambers were transferred to a pre-warmed live-cell imaging chamber slide incubator on a Nikon A1 confocal microscope and allowed to equilibrate for approximately 10 min. at 37 °C and 5 % CO2. All data acquisition was performed with a 488 nM excitation laser with a 60x oil-objective attached to a heating element to prevent heat-sink from the imaging chamber during acquisition. Thirty seconds after acquisition began, a pre-warmed protein-media mixture was added to cells in a 2X concentration to ensure complete mixing.

Tau Internalization Assay and Flow cytometry
To test the tau uptake propensity by different cell types (HEK293, SH-SY5Y, C6 glioma), singlecell C6 subclone (CL5), LRP1 knockdown C6 cells, and primary rat astrocytes (RA) were seeded at 5 × 10 5 cells/well in a 24-well plate. After culturing overnight, cells were incubated with tau-AF488 at fixed incremental concentrations of tau proteins (0.25 μM, 0.5 μM, and 1.0 μM). After at least 30 mins., the cells were washed with PBS, trypsinized with 0.05% Trypsin-EDTA (Thermo Fisher Scientific, USA), and the fluorescence analyzed with an Accuri flow cytometer (MBIC, Mellon Institute). The relative tau uptake for each condition was calculated by subtracting the non-specific binding of the fluorophore to the cell membrane, which was conducted with identical experimental procedures but at 4 ℃ to limit internalization.

LRP1 Knockdown by Small Interfering RNA (siRNA)
Double-stranded, rat LRP1 specific siRNAs and non-targeting siRNA were synthesized by IDT.
The sequences of LRP1 specific siRNAs were as follows: Gene expression of LRP1 was analyzed by real-time PCR analysis 48 hrs. after transfection.

Statistical Analysis
Statistical significance was determined by unpaired t-tests, one-way analysis of variance (ANOVA), or two-way ANOVA when appropriate using GraphPad Prism 8 software. A five percent cutoff was applied to determine statistical significance, and a p-value of <0.05 was denoted with one asterisk (*).
In all cases, at least three independent biological replicates from different days were performed and the standard error of the mean is shown.

0N4R tau is rapidly endocytosed by immortalized cell lines via actin-dependent macropinocytosis pathway
We employed both live-cell confocal microscopy and flow cytometry to give insights into the kinetics of 0N4R tau endocytosis in three immortalized cell lines (Fig. 1). Purified 0N4R tau protein monomer was labeled with Alexa Fluor 488 dye (tau-488) as described in materials and methods, and quantification of live-cell confocal microscopy as number of vesicles per cell showed that rat C6 glioma cells endocytosed tau-488 monomer around three to four times as rapidly as HEK293 cells and SH-SY5Y cells ( Fig. 1A and 1B). This result was corroborated by quantitative flow cytometry analysis, as 99% of C6 glioma cells were positive for tau after 15 min. compared with 75.2% for SH-SY5Y cells and just 6.8% for HEK293 cells for 1 µM tau addition (Fig. 1C-E). Importantly, free Alexa Fluor 488 label and bovine serum albumin (BSA; 1 μM) labeled with Alexa Fluor 488 did not enter cells over the imaging periods used (Supplemental Fig. S1), suggesting that bulk fluid-phase endocytosis was not responsible for the rapid uptake of 0N4R tau that was observed. Because tau probably exists in lower concentrations in the ISF as compared with the cytosol in living systems [10], we also tested uptake of 25 nM tau-488 for all three cell types ( Fig. 1C-E). We observed a decrease in tau-488 fluorescence in SH-SY5Y and HEK293 cells over the time periods tested; however, no statistically significant difference was observed in C6 glioma cells for the 25 nM versus the 1 µM tau addition (Fig. 1E).
To confirm that this process occurred via cellular endocytosis, we pre-incubated HEK293 cells on ice for thirty minutes prior to tau addition and found that tau-488 monomer did not enter cells at temperatures non-permissive for endocytosis, as determined by confocal microscopy ( Fig. 2A and 2B).
Furthermore, flow cytometry confirmed that HEK293 cells preincubated at 4 °C endocytosed approximately seven times less tau-488 than cells incubated at 37 °C (Fig. 2C). To understand the molecular mechanism responsible for tau endocytosis, SH-SY5Y cells -an effective model of neurodegeneration [42] -were preincubated with cytochalasin-D as a potent actin-polymerization and macropinocytosis inhibitor.
Cytochalasin-D addition reduced tau uptake significantly over the entire observation time (Fig. 2D). Tau pathology is not solely restricted to neuronal cells, as tau inclusions have been found in astrocytes, oligodendrocytes, and even microglia [43] [44]; therefore, we also sought to validate a similar tau uptake mechanism in an immortalized glial cell line, the rat C6 glioma line. Diminished tau uptake was also observed in C6 glial cells (Fig. 2D) with the addition of cytochalasin-D. Taken together, these data suggest that 0N4R tau monomer is readily endocytosed by all three immortalized cell lines, largely by actinmediated macropinocytosis.

C6 glioma cells efficiently endocytose 0N4R monomer in an HSPG-dependent manner
Because tau endocytosis in both monomeric and oligomeric states has previously been shown to depend on the presence of heparan sulfate proteoglycans (HSPGs) [26] [45]. Heparin, a glycosaminoglycan, has been used to competitively inhibit binding of 2N4R tau to HSPGs, as well as other entities in the extracellular space such as viruses [46]. To relate tau endocytosis to HSPGs on the cell surface, tau uptake was examined in CHO and CHO745 -a CHO-derived cell line deficient in xylosyltransferase, thus incapable of synthesizing all glycosaminoglycans (GAGs) [47]. Quantitation of live-cell imaging showed that CHO745 cells rapidly endocytosed tau monomer but did so at a significantly slower rate than wildtype CHO cells (Fig. 3A). To further our understanding, heparan sulfate (HS) protein expression on the cell surface was quantified by immunolabeling and flow cytometry analysis (Fig. 3B). The relative expression levels are consistent with the differential endocytosis rate we observed (Fig. 1), where C6 cells had the highest expression of HS as well as the highest endocytosis rate across all the cell lines we tested, and as expected, CHO showed higher levels of HS relative to the CHO745, consistent with observed tau uptake.
To validate our hypothesis about the importance of HSPGs as key cell surface receptors for tau, we then constructed an HSPG-deficient C6 cell line by genetic knock-down of xylosyltransferase (Xylt1) using the clustered regularly interspaced short palindromic repeat (CRISPR)-Cas9 system [48]. This approach was informed by the previous approach in CHO cells that showed that Xylt1 deficiency would prevent glycosaminoglycan production and directly disrupt HSPG biosynthesis [47], Single guide RNA (sgRNA) was designed to target Exon7 of Xylt1, within the xylosyltransferase domain (Fig. 4A). C6 wild-type cells were transfected with a sequence-verified CRISPR/Cas9-sgRNA construct according to this design. We observed a higher level of GFP expression for transfected cells, while cells lacking the template showed no fluorescence (Fig. 4B). The top 3% of the GFP-positive cell population was subjected to FACS-based single-cell selection (see materials and methods), and two clones were characterized following fragment length analysis -clone 3 (CL3), and clone 5 (CL5). As Xylt1 catalyzes the first step in the biosynthesis of HSPG, the levels of heparan sulfate (HS) were quantified by immunostaining to validate the knockdown.
Flow cytometry analysis of these stained populations showed that CL5 had about half the HS expression compared with the non-edited C6 parental cells, while CL3 showed a similar level of HS expression compared with the parental cells (Fig. 4C). To confirm the HS expression data, mRNA levels of Xylt1 were quantified, with Xylt1 gene expression of CL5 observed to be about two-fold lower than the non-edited parental C6 (Fig. 4D). Western blot analysis confirmed the absence of Xylt1 expression in CL5, compared with non-edited parental C6 (Fig. 4E). These data confirmed the successful creation of a Xylt1-knockdown C6 cell line that will be referred to as CL5, by using the CRISPR/Cas9 system, which was used for subsequent studies.
Uptake of tau-488 by CL5 was then compared with parental C6 cells (Fig. 4F). Tau-488 was rapidly and efficiently internalized by C6 cells in a concentration-dependent manner, while the HSPG deficiency in CL5 significantly decreased the cellular uptake of tau at every concentration measured during a thirtyminute incubation. These data indicate that HSPGs indeed play a significant role in the uptake of 0N4R tau monomers by glial cells.

Intracellular responses stimulated by tau internalization is HSPG-dependent
Stancu and colleagues [49] showed previously that aggregated tau was capable of activating the NLRP3 inflammasome, increasing secretion of proinflammatory cytokines. Uptake of fibrillar Aβ also induced the assembly of the NLRP3 inflammasome, which led to caspase 1-dependent release of proinflammatory cytokines such as IL-1β and IL-18 [50]. To identify the cellular events affected by tau monomer endocytosis in C6 glial cells, we next evaluated mRNA transcriptional levels of the proinflammatory cytokines, IL-6 and TNF-α. Wild type 0N4R tau addition significantly upregulated pro-inflammatory gene expression of IL-6 and TNF-α (Fig. 5Ai), suggesting a direct role in activating a pathological cascade in glial cells.
As previous studies demonstrated that the activation of the MEK/ERK pathway was associated with early tau deposition in neurons and glial cells in tauopathies [51], we recognized that ERK1/2 phosphorylation might lead to cytokine transcription. Internalization of Aβ during the progression of neurofibrillary degeneration in Alzheimer's Disease (AD) has also led to MEK/ERK pathway activation [52]. To explore the potential mechanism through which tau induced immune responses, the phosphorylation of ERK1/2 was further examined for different experimental conditions [53]. Intriguingly, the phosphorylation of ERK1/2 increased in glial cells following tau monomer addition as compared with control cells, as determined by Western analysis (Fig. 5B).
To further understand whether cytokine addition was affected by physiological tau differently from pathological tau, several tau variants were examined. A triple mutant VPR (V337M, P301L, R406W) associated with frontotemporal dementia (FTD) that has been shown to hyperphosphorylate and have increased aggregation propensity [54][55] and a highly phosphorylated tau protein (ptau) that is expected to drive aggregation [56] were created as described in materials and methods, and purified as for wild-type tau. Interestingly, the highest ERK1/2 phosphorylation was concurrent with the highest expression of immune genes in the C6 cells incubated with VPR or ptau ( Fig. 5Ai and 5B). Addition of the MEK-ERK1/2 inhibitor U0126 abolished the increased levels of IL-1β and IL-6 and enhanced TNF-α mRNA levels that were induced by exogenous VPR addition (Fig. 5Aii). Taken together, we conclude that tau monomer addition induced gene expression changes of pro-inflammatory cytokines, which are regulated by ERK1/2 activation.
HSPGs have been implicated in cellular signaling from a variety of ligands, such as the Wnt, Hedgehog, transforming growth factor-β, and fibroblast growth factor during development [57]. To investigate the relevance of HSPGs and intracellular signaling activation by tau addition, we further evaluated ERK1/2 phosphorylation and mRNA transcriptional levels of the proinflammatory cytokines IL-6 and TNF-α in the HSPG-deficient C6-derived cell clone CL5. Intriguingly, CL5 cells displayed minimal activation of pro-inflammatory genes upon tau monomer addition compared to parental C6 (Fig. 5Aiii) and the phosphorylation of ERK1/2 was decreased (Fig. 5C). These data suggest that HSPGs regulate the proinflammatory gene expression of glial cells in response to monomeric tau via ERK1/2 signaling activation.

Heparin attenuates tau uptake and mitigates immune response to tau in primary astrocytes
In order to ensure the observations were not limited to immortalized cell lines, experiments in primary astrocytes were performed to determine the effects of heparin in regulating intracellular activities affected by tau endocytosis. The propensity for tau uptake was examined in rat primary astrocytes by fluorescent imaging or flow cytometry to visualize and quantify tau internalization. Monomeric VPR-488 was rapidly internalized by the primary astrocyte (Fig. 6A), and the amount of internalized tau increased over time from 1 hr. to 6 hrs., while heparin addition attenuated the tau uptake, indicating tau uptake was partially mediated by HSPGs. Phase-contrast images were overlaid with fluorescent images to visualize the spatial localization of tau, where tau protein was observed to be localized primarily within the cell body ( Fig. 6A). Additionally, the nuclear stain Hoechst and an astrocyte-specific marker, GFAP, were co-imaged with VPR-488 to determine the spatial distribution of tau relative to the cell nucleus (Fig. 6B). Tau protein was evenly distributed outside of the cell nucleus, with negligible overlap with intracellular GFAP protein.
In order to quantify this observation, monomeric VPR-488 uptake was determined in primary astrocytes by flow cytometry (Fig. 6C and D), showing that all cells appear to take up tau following a 6-hour incubation.
Heparin addition led to a 50% decrease in tau uptake compared to the control group without heparin addition. A similar effect of heparin reduction of wild-type tau and ptau internalized by astrocytes was observed (Supplemental Fig. S2), suggesting heparin competes with tau for binding with HSPGs at the cell surface, regardless of physiological tau state. Further, the presence of heparin also ameliorated the immune activation of astrocytes upon the addition of unlabeled monomeric VPR by driving the IL-6 and IL-1β gene expression back to baseline levels, though TNFα levels remained elevated (Fig. 6E).

LRP1 receptor plays a role in regulating intracellular responses of glial cells to tau protein
LRP1 has been recently identified as a major regulator of tau internalization, and tau transmission in neurons [58]. Ligand binding to LRP1 also activates cytoplasmic signaling pathways including several important intracellular mediators such as JNK and ERK and has been shown to mediate the immune responses of microglia to extracellular signals [59]. To identify potential interactions between HSPGs and LRP1 upon tau addition, we measured LRP1 expression in both C6 cells and CL5 cells. The addition of VPR protein downregulated LRP1 levels in C6 cells while LRP1 levels in CL5 cells remained at similar levels as the parental C6 cells (Fig. 7A). While the cellular mechanism underlying the interplay between HSPG and LRP1 expression needs further investigation, these data clearly indicate that HSPG levels mediate LRP1 expression in the presence of tau.
To validate our observation and identify the role of LRP1 in regulating intracellular responses to monomeric tau, LRP1 gene expression was reduced in C6 glial cells using specific LRP1-siRNAs. LRP1 was successfully knocked down by 20%-70% using three different LRP1-siRNA (Fig. 7B). LRP1-siRNA3 knockdown cells and control cells were incubated with 1 μM VPR-488 at 37 °C for 30 minutes and then analyzed by fluorescent microscopy (Fig. 7Ci). Tau uptake was slightly decreased in LRP1 knockdown cells, which was subsequently quantified as a 30% reduction by flow cytometry (Fig. 7Cii). LRP1-siRNA3 knockdown exacerbated IL-6 gene expression increases as compared to negative control cells in response to tau (Fig. 7D), suggesting an intermediate regulatory role of LRP1 in response to inflammatory stimuli.
Interestingly, ERK appeared unresponsive in LRP1 knockdown cells (Fig. 7E), implying that other possible as yet unknown signaling pathways, regulated by LRP1, may be also responsible for the activation of immune pathways following tau uptake.

Rapid HSPG-dependent endocytosis of tau monomers
Extracellular tau is present in brain cerebrospinal fluid (CSF) and interstitial fluid (ISF) of AD patients either as free tau or within vesicles [16]. The majority of extracellular tau consists of soluble oligomers and monomers, while a minority of tau species are in a truncated form from cleavage by various pro-inflammatory cytokines released by reactive astrocytes and microglia in AD brains [60]. The cellular internalization of extracellular tau is a major part of neurodegenerative disease progression, which is still under investigation [61]. The cellular pathways of tau internalization are highly dependent on the overall size of tau species, ranging from receptor-mediated endocytosis to actin-dependent macropinocytosis [25] [45]. Importantly, HSPGs, highly expressed on the cell surface, have been identified as an important cell-surface endocytosis receptor of tau in various studies [45] [62]. Ranch and coworkers reported that knockouts of extension enzymes of the HSPG biosynthetic pathway, such as extension enzymes exostosin 1 (EXT1), exostosin 2 (EXT2), and exostosin-like 3 (EXTL3) sufficiently reduced the uptake of tau oligomers in HEK293 cells [26]. Primary neurons and mouse models with heparin pretreatment showed diminished tau aggregate uptake, suggesting HSPGs are required for tau aggregate internalization in vivo [45]. We have observed a similar effect of HSPGs on regulating, in particular, monomeric tau protein, using a more CNS relevant cell line. In our study, reduced tau uptake was observed in both CHO745 and Xyltknockout C6 glial cell lines and indicated the role of HSPGs in tau uptake.
The interaction between tau monomers and HSPGs has also been characterized at the structural level characterized by surface plasmon resonance and NMR mapping. Several studies suggest that tau-

HSPG interactions are regulated by specific sulfation patterns of the HS chain, including 6-O-sulfation [26],
3-O-sulfation [63], and N-sulfation [64]. Interestingly, tau also contains a heparin-binding domain; residues in both the proline-rich region (PRR) and the second repeat motif (R2) have been shown to be specifically recognized by negatively charged sulfo groups on HS [63]. This suggests a complex regulation of tau interactions with HSPG and heparin that merit more attention in further studies.

Cellular events regulated by cell surface heparan sulfate proteoglycans
Cell surface HSPGs have been implicated in a variety of receptor signaling pathways via the binding of extracellular molecules and entities such as viruses [62]. HSPGs act as internalizing receptors of macromolecular cargo such as FGF2, lipoproteins, and ApoE in a receptor-independent mechanism via the clathrin-coated pit endocytosis pathway. There are two major classes of cell surface HSPGs, which are characterized by the type of core protein -the glypican (GPCs) family of glycosylphosphatidylinositol (GPI)-linked proteins and the syndecan family of transmembrane proteins [57]. Most relevant to our study, the binding of ligands to cell surface syndecan HSPG activates ERK1/2 signaling to facilitate actindependent endocytosis [62].
As a co-receptor, HSPGs activates receptor signaling following binding between cytokines and their corresponding receptors, such as the binding between FGF and FGFR and IL-6 and IL-6R [57]. For instance, a recent study demonstrated the inflammatory effects of low molecular weight heparin to attenuate IL-6 release in patients with COVID-19 by competing with HSPGs for the binding of IL-6 [65]. Moreover, the transcriptional level of IL-6 was significantly decreased after the addition of a heparinase enzyme inhibitor. The underlying mechanism behind these effects was that cell surface HSPGs regulate IL-6 ligandreceptor interaction and IL-6 secretion via facilitating the activation of JAK associated with the IL-6 receptor. Downstream ERK 1/2 signaling has also been found to mediate the transcriptional activation of IL-6 [66]. An ERK1/2 inhibitor reduced the transcriptional level of IL-6, suggesting the ERK activation plays a role in transcription and subsequent IL-6 secretion. Here, no ERK1/2 signaling activation was observed in HSPG-deficient cells. Taken together, these data suggest that HSPG interactions play a critical role in regulating inflammatory gene expression in glial cells following the addition of exogenous monomeric tau via intracellular ERK1/2 signaling activation (Fig. 8). However, further study is needed to identify the primary contributor to ERK1/2 phosphorylation, HSPG-dependent endocytosis of tau or positive-feedback regulation of IL-6 secretion, or both.

Differential Tau-HSPG interaction defined by specific tau variants
Previous work describes a 'paperclip-like' conformation for tau in solution, where the N-and Cterminal ends fold over in proximity to the center of the repeat domains [67]. Site-specific phosphorylation and mutations have been found to directly impact the conformation of monomeric tau and affect the stability of a folded conformation, contributing to the propensity for tau protein to aggregate [68]. A recent study reported that FTD-associated mutations enhanced tau oligomer internalization in SH-SY5Y cells, presenting in vitro evidence that mutation-induced conformational changes may lead to changes in tau binding to HSPGs on the cell surface [69]. Indeed, site-specific modifications of tau monomer may cause conformational changes to tau structure that facilitate conversion of inert wild-type tau into seed-competent tau states that have a high propensity to form aggregates and filaments [18]. In our study, no significant difference was observed in monomeric tau uptake between different protein variants within 30 minutes.
Nevertheless, differential pro-inflammatory gene expression and ERK1/2 phosphorylation was observed with these variants. This result implies that differential tau-HSPG interaction may be responsible for intracellular change; further investigation of tau protein variant binding to heparan sulfate proteoglycans may uncover differential affinities. Alternatively, changes to the tau variants following uptake -such as aggregation or phosphorylation -could alter intracellular pathways. To our knowledge, our study is the first that finds different tau variants' behavior in activating downstream cellular pathways in a CNS-relevant cellular system, which argues that the tau-HSPG interaction may be also epitope and tau variant-dependent.

Crosstalk between LRP1 and HSPGs in regulating intracellular responses of glial cells to tau monomer
Apart from HSPG-dependent endocytosis, tau internalization is also regulated by specific receptormediated endocytosis, as suggested by several previous reports [25] [58]. LRP1 has recently been recognized as a master regulator of tau uptake; LRP1 knockdown abolished uptake of various forms of tau including monomers, oligomers, and fibrils in H4 neuroglioma cells [58]. Our study showed consistent results as observed in neuronal cells, demonstrating the involvement of LRP1 in tau endocytosis by glial cells. In addition, the cooperative interplay between HSPGs and LRP1 has been found to regulate neuronal Aβ uptake [70]. Interestingly, LRP1 has been shown to facilitate the endocytosis and lysosomal degradation of the HSPG-hedgehog complex [71], suggesting a possible LRP1-based regulation of HSPG-dependent endocytosis.
In our study, downregulation of LRP1 mRNA was observed in the glial cells upon monomeric tau addition, while LRP1 remained at a similar level in HSPG-deficient cells compared to control cells upon tau addition. Further, LRP1 knockdown exaggerated IL-6 mRNA expression, while no ERK activation was observed. Thus, we suggest that IL-6 changes likely were altered via the JNK pathway (Fig. 8).
Alternatively, upregulated IL-6 expression may be the outcome of cooperation between LRP1 and HSPG via its binding to monomeric tau. Further studies are needed to identify the intracellular signaling pathways involved in the immune response of glial cells to tau, with a special focus on the MAPK and JNK pathways, which have been previously identified as the key protein kinases regulated by Aβ-LRP1 interaction in microglial cells [59].

Implications of efficient monomeric tau endocytosis for tau pathogenesis
Extensive studies suggest that the progression of tau pathologies including FTD, Pick disease, progressive supranuclear palsy and corticobasal degeneration, rely on the effective spreading of pathogenic tau, known as 'seeds', along both synaptic and non-synaptic pathways [61] [72]. This propagation of tau pathology involves the transfer of pathogenic tau seeds from 'recipient cell' to 'donor cell' and templatedseeding of aggregation via recruitment of normal tau and formation of new tau seeds in donor cells [73].
Several studies have shown that soluble tau oligomers may act as the toxic seeds inducing endogenous tau misfolding [74] and transneuronal propagation [15]. However, monomeric tau internalization has been less well investigated as it has been considered less pathogenic. Given the abundance of monomeric tau in neurons, as well as the presence of extracellular tau in the ISF and CSF, monomeric tau may serve a role in the pathogenesis of tauopathies under as yet unknown conditions. Even though the templating-seeding properties of tau seeds mostly rely on the recruitment of endogenous tau protein [61], rapid endocytosis of monomeric tau protein has provided higher levels of intracellular tau protein available for seeding in both neuronal cells and astrocytes. It remains more plausible that this seeding occurs under pathological conditions where tau seeds facilitate intracellular tau aggregate formation [30] [75]. Phosphorylated tau protein may have a higher seeding propensity since phosphorylated tau was more sensitive to heparininduced aggregation compared to wild-type protein based on in vitro study [76].
Our results show more abundant and rapid accumulation of tau protein in astrocytic-like cells compared to neuronal cells, which is consistent with prior work demonstrating astrocytes exhibit greater avidity for oligomeric tau than neurons [77]. Astrocytes are enriched in phagocytosis pathways, suggesting the natural cellular mechanisms of enhanced uptake increased tau internalization by astrocytes as compared to neurons. For instance, astrogliosis has been observed concurrent with the upregulation of TFEB (transcriptional factor EB) -a key regulator of phagocytosis in astrocytes -within a range of clinical neurodegenerative diseases [78]. Astrocytic internalization of tau disrupted intracellular Ca 2+ signaling, leading to a significant reduction in the levels of released gliotransmitters and synaptic vesicles, in an ATPdependent manner [77]. Informed by these previous studies, rapid endocytosis of monomeric tau by astrocytes observed in our study would likely alter astrocytic neurotransmission, causing synaptic degeneration. We suggest that the seeding propensity for aggregation and potential alterations in neurotransmission following uptake of monomeric tau may have an overlapping role in the progression of tau pathogenesis. Future studies will focus on further investigating the functional alterations in CNS cell lines following monomeric tau internalization.

Conclusions
Our study shows that monomeric tau is rapidly endocytosed by multiple CNS    shows quantitative decrease in tau uptake following knockdown. Gray filled curve shows distribution with vehicle only. Rat C6 glioma cells were transiently transfected with either non-targeting siRNA (NC) or LRP1-specific siRNA3 for 48, followed by incubation with 1 μM AF488VPR or vehicle for 24 hrs. D. RT-qPCR analysis was performed to detect the expression level of IL-6 in C6 cells post-transfection as in (C), with β-Actin was used as an internal control. Open bars, NC-siRNA tranfected; Gray shaded bars, LRP1-siRNA3 (n=2, * indicates p<0.05 between different conditions) E. Cell lysates were then analyzed by immunoblotting (i.) against p-ERK1/2, total ERK (t-ERK), and β-actin, followed by quantitation (ii.) of p-ERK/β-actin ratios from a representative western blot experiment. * indicates p<0.05 between different test conditions.

Fig. 8 Schematic illustration of the intracellular responses affected by monomeric tau endocytosis.
Monomeric tau is rapidly internalized by both neuronal and glial cells, mediated by the actin-dependent macropinocytosis pathway. In addition, an important role of HSPG has been demonstrated in regulating inflammatory gene expression, including IL-6 and IL-1β, partially via ERK1/2 activation following tau uptake. LRP1 modulates IL-6 responses to tau endocytosis, and this interaction is HSPG-dependent but not ERK-dependent.