Tau Oligomers Induce Brain Endothelial Cell Hyperpermeability and Increase NLRP3 Inflammasome Signaling and MMP-9 Activity

Abstract

Objective:

Microvascular hyperpermeability and blood-brain barrier (BBB) dysfunction is a key consideration in neurological disorders, particularly tauopathies, a group of neurodegenerative disorders driven by misfolded and aggregated tau protein. Tau pathology has been shown to activate microglial NLRP3 inflammasome, an innate immune system sensor that responds to changes in the microenvironment, including cellular stress. Increases in reactive oxygen species (ROS), for example, activate NLRP3 inflammasome signaling, which provides a platform for the maturation of caspase-1 enzyme. Mature caspase-1 can cleave and release pro-inflammatory IL-1β cytokine. Both NLRP3 inflammasome and IL-1β may activate downstream MMP-9 enzyme, a known inducer of endothelial cell barrier hyperpermeability. Endothelial cells make up the innermost layer of the BBB and as such, largely govern BBB structural and functional integrity. Whether tau can activate NLRP3 inflammasome signaling in cerebral endothelial cells is unknown. The objective of this study is therefore to understand the role of tau, in various states of aggregation, on endothelial cell permeability and to investigate if NLRP3 inflammasome signaling occurs in this context.

Methods:

Human brain microvascular endothelial cells (HBMECs) were grown as a monolayer in Transwell inserts and exposed to various tau polymorphs, including tau monomers, tau oligomers (oTau), and tau fibrils (fTau). Barrier permeability was measured using FITC-dextran fluorescent tracer (10 kDa) and Trans-Endothelial Electrical Resistance (TEER). Relative changes in gene expression were measured using RT-qPCR and normalized to GAPDH. Levels of NLRP3 sensor protein and IL-1β were measured by ELISA. Relative activity of caspase-1 and MMP-9 enzymes were calculated using fluorometry. Cell viability was reported using calcein AM, a measure of cell membrane integrity, and by measuring the redox potential of XTT. ROS formation, apoptosis, and necrosis were determined using a commercially available kits. An NLRP3 inflammasome inhibitor, MCC950, was applied in blocking studies prior to tau treatments.

Results:

Tau oligomers, but not monomers or fibrils, induced endothelial cell hyperpermeability in a dose-independent manner. At concentrations that compromised barrier function (100 nM; overnight), oTau did not alter cell viability and did not increase apoptosis or necrosis. Tau oligomers increased the formation of ROS and increased levels of both NLPR3 sensor protein and IL-1β cytokine. Enzymatic activity of caspase-1 and MMP-9 also increased in response to oTau, without changes in gene expression. These alterations were attenuated when NLRP3 inflammasome signaling was inhibited via MCC950, strongly suggesting that oTau activates NLRP3 inflammasome signaling in cerebral endothelial cells.

Conclusions:

Our study is the first to document a role for NLRP3 inflammasome signaling in cerebral endothelial cells following exposure to tau oligomers. Given the importance of endothelial cell functioning in BBB integrity, these data are significant in that they demonstrate a key role for endothelial cell signaling in tau pathogenicity and propose a mechanism by which tauopathies compromise BBB functional integrity. Taken together, these findings warrant future investigation into the therapeutic potential of NLRP3 inflammasome inhibition to ameliorate tauopathy related barrier breach.

Introduction:

The blood-brain barrier (BBB) is a critical physiological structure with highly limited permeability that allows for the maintenance of central nervous system (CNS) homeostasis and acts as an interface between the central and the peripheral circulations (1). Cerebral endothelial cells (CECs) make up the innermost lining of the BBB and are largely responsible for its distinctive functionality, allowing for the selective exchange of solutes between the blood and the brain (1, 2). CECs are sealed together by tight junction proteins, and disruption to CECs due to mechanical, inflammatory, or protein aggregate-related damage may result in barrier hyperpermeability (1, 2). BBB dysfunction plays a role in, and serves as a biomarker for, several neurological disorders (1, 3-5)

Tau is an endogenous, microtubule-stabilizing protein highly expressed in the brain (6, 7). Tauopathies are neurodegenerative diseases, including Alzheimer’s disease (AD), that are driven by the accumulation of misfolded and aggregated tau (7). Tau pathology occurs when physiological tau becomes hyperphosphorylated (pTau) (8), causing microtubule detachment that results in conformational changes promoting tau aggregation, which ultimately interferes with normal cellular and physiological functioning (6-9). The heterogeneity of aggregation structures is responsible for the various conformational states of tau, which are known as tau polymorphs (10). Tau polymorphs include pTau, tau oligomers (oTau), and tau fibrils (fTau), among others (8, 10-12). Subtle differences across these polymorphs are hypothesized to be responsible for the phenotypic and mechanistic differences across tauopathies (10, 13, 14). While pathogenic tau is known to play a role in mediating BBB dysregulation (15-17), the precise effects of tau polymorphs on the BBB are unclear. Importantly, identifying which tau polymorphs alter CEC functionality and the mechanisms governing such changes is a priority for tauopathy researchers.

Inflammasomes are a family of intracellular multiprotein complexes that monitor and respond to cellular stressors (18). Of these, NOD-like receptor family pyrin domain containing-3 (NLRP3) inflammasome is a major contributor of neurological diseases, including tauopathies (19). NLRP3 inflammasome is comprised of three components: the NLRP-3 sensor protein, the ASC adaptor, and the caspase-1 effector protease (20). These components exist in the cellular cytoplasm as distinct and inactive components, but may become activated in the event of several cellular stress mechanisms, including reactive oxygen species (ROS) formation (21). Upon activation, the components of NLRP3 come together to provide a platform for caspase-1 maturation, allowing for caspase-1 mediated cleavage and the release of interleukin (IL)-1β (20, 21). In addition to activating an inflammatory response (20), IL-1β has been shown to activate matrix metalloproteinase (MMP)-9 enzyme both directly (22-24) and indirectly via IL-1β (25, 26). Activated MMP-9 is a known inducer of CEC dysregulation through tight junction protein degradation and CEC structural reorganization (27, 28). While NLRP3 inflammasome activity has been reported in the context of inflammation (18, 20), neurodegeneration (19, 29-31), and BBB dysregulation (32), its role in tau-mediated CEC hyperpermeability has not yet been shown.

The objectives of this study were to (1) characterize the functional effect of tau polymorphs, including tau monomers, oTau, and fTau, on cerebral endothelial cell functioning; and (2) investigate the role of NLRP3 inflammasome signaling in CEC barrier permeability. Figure 1 depicts the in vitro methods and quantifiable outcomes used to test our hypothesis.

Figure 1. A schematic overview of study methods and results.

Human microvascular endothelial cells (HBMECs) were exposed to tau oligomers (oTau) (100 nM; overnight) and used for subsequent experimentation. Experimental methods include functional measures of relative permeability using TEER and FITC-dextran fluorescent assay. Biochemical assays include cell stress detection, RT-qPCR for mRNA expression, ELISA, and fluorometric assay for determination of protein levels and enzyme activity, respectively. Measurable outcomes include changes in HBMEC barrier permeability, changes in relative ROS generation, and changes in the NLRP3 inflammasome signaling pathway, including capsase-1 effector and mediators IL-1β and MMP-9.

We hypothesized that tau aggregates may activate NLRP3 inflammasome signaling in CECs, resulting in microvascular hyperpermeability.

Materials and Methods:

Reagents

Recombinant human tau proteins were purchased as monomeric human Tau-441 (2N4R) (ACRO Biosystems), human Tau-441 (2N4R) Wild-Type Oligomers (StressMarq Biosciences Inc.), and human Tau-441 (2N4R) Pre-Formed Fibrils (StressMarq Biosciences Inc.). Proteins were prepared per manufacturer’s instructions. Briefly, proteins were reconstituted in 1x phosphate buffer (pH=7.4), aliquoted, and used promptly.

Tau Treatment Protocol

Tau treatment protocols were designed to simulate pathologically relevant concentrations, as reported in the literature. Monomeric Tau was tested at 1 mg/mL and 5 mg/mL (33, 34). Oligomeric tau (oTau) was tested at 10 nM, 100 nM, and 250 nM (35, 36). Fibrillar tau (fTau) was tested using 10 nM and 25 nM (34, 37). All tau treatments were incubated overnight at 37 °C to simulate acute exposure. Following initial dose-testing and cell viability testing, oTau was given at a concentration of 100 nM. This dose was chosen based on the finding that (1) oTau induces hyperpermeability at this dose; and (2) that cells remain viable at this dose. When a change was observed compared to control conditions, experiments were repeated with NLRP3 inflammasome inhibition using pharmacological blocking agent, MCC950 (1 μM; 2 hrs) (Millipore, Sigma-Aldrich) to explore a role for NLRP3 inflammasome-dependent signaling (38, 39). Hydrogen peroxide (H₂0₂), as an inducer of barrier hyperpermeability (100 uM; 2 hrs), is used as a positive control (40).

Cell Culture

Primary cultures of human brain microvascular endothelial cells (HBMECs) isolated from a male healthy brain cortex without positive selection were purchased from Cell Systems (ACBRI 376, NC2380990) in December of 2024. HBMECs are antibody-free and authenticated for the research conducted in this current study, including BBB modeling, neurovascular research, permeability assays, gene and mRNA expression studies, cytokine signaling, and inflammation and vascular activation assays, among other applications. HBMECs tested negative for bacteria, fungus, and mycoplasma with negative serology and PCR for HIV, HBV, and HCV. According to the manufacturer’s datasheet, no previous contamination has been reported and no reports of misidentification have been described.

For experimentation, HBMEcs were cultured in Complete Classic Medium supplemented with Culture Boost, per manufacturer’s instructions (Cell Systems). HBMECs were inoculated onto polystyrene petri dishes coated with 50 ug/mL bovine fibronectin diluted in phosphate-buffered saline. Cells were incubated at 37 °C with 5% CO₂. Media was changed every 48 hours until reaching confluence. Adherent cells were detached with 0.05% trypsin/EDTA (ThermoFisher) for subsequent experiments. HBMEC passages 4-12 were used for all experiments.

Monolayer Permeability

Two permeability assays were used to quantify the relative permeability across the HBMEC monolayer in several experimental conditions: (1) a FITC-dextran fluorescent tracer assay (41) and (2) TEER assay (42). Both protocols begin by seeding HBMECs on bovine-fibronectin coated Transwell inserts (Corning) for 72 hours. One hour prior to experimentation, Complete Classic Medium is replaced with Opti-MEM reduced serum media (ThermoFisher) and incubated for 1 hour at 37°C to reduce interreference. Experimental treatments are given per the design protocol described above. Following overnight incubation at 37°C, permeability is measured using both below measurements.

FITC-Dextran Fluorescent Assay:

Fluorescein isothiocyanate-labeled (FITC)-dextran, a 10 kDa fluorescent tracer is added to the apical (upper) chamber of the bi-chamber Transwell insert and allowed to incubate for 30 minutes at 37°C. Following this 30-minute incubation period, 100 μL of solution from the basal (bottom) chamber is removed and added to a clear-bottom black 96-well plate. Given the relatively large size of the FITC-dextran molecule against the relatively small (4 μm) pore size of the inserts, permeability across the monolayer is expected to be relatively limited. Hyperpermeability is reported as a relative increase in FITC-dextran leakage from the apical to the basal chamber compared to control, as measured by fluorescent intensity reading on the SYNERGY/HT microplate reader at excitation/emission of 450/680. Permeability is reported as Dextran Relative Fluorescent Units (RFU).

TEER:

The Trans-Endothelial Electrical Resistance (TEER) assay is used to validate the results of the Dextran fluorescent permeability assay. An endothelial ohm meter TEER –EVOM2 (World Precision Instruments) is used to measure the value of electrical conductance in the apical chamber of the Transwell inserts. Increased permeability results in a decrease in electrical resistance in the apical chamber. TEER is measured by loading the individual inserts on the Endohm chamber and placing the electrode in the well to create a current density flow across the membrane. TEER values are determined by subtracting the EVOM2 value recorded in wells with cells against the EVOM2 value recorded in wells without cells. The electrodes are dipped in saline between each reading and gently dried with water absorbing paper.

In all permeability experiments, groups have a sample size of 3-6 wells per condition, and all experiments are performed thrice at minimum. Representative data from one experiment are reported when findings from all experiments are statistically consistent.

Real-Time Quantitative Polymerase Chain Reaction (RT-qPCR)

HBMECs are seeded onto 6-well plates until 80% confluent. Complete Classic Medium is replaced with reduced serum media (Opti-MEM, ThermoFisher) and incubated for 1 hour at 37°C before experimental treatments are given to reduce interreference. Cells are treated with the appropriate protocol and then harvested using RNA isolation lysis buffer supplied by the QUIAGEN Mini Kit (Quiagen). The same kit is used to perform RNA extraction, per manufacturer's instructions. RNA concentration is determined using Thermos Scientific NanoDrop One/OneC. cDNA is obtained via the iScript cDNA synthesis kit (Bio Rad). Real-time (RT) quantitative (q) PCR is performed using iQ^™ SYBR Green Supermix (Bio Rad). Primers were designed and purchased from Invitrogen (ThermosFisher). A list of primer sequences may be found in the Supplemental Data as Table S1 and include GAPDH, NLRP3, IL-1β, CASP-1, and MMP-9. The cDNA is loaded into the Bio-Rad CFX Opus 96 machine for 10 minutes at 95 °C, followed by 40 cycles of denaturation for 15 seconds at 95°C and 1 minute annealing at 60°C. Cq values are normalized to GAPDH and delta-delta-Cq method is used to calculate fold change. Data are expressed as relative fold change.

Detection of Enzyme Activity

Caspase-1:

The activity of caspase-1 enzyme, also known as Interleukin (IL)-1β converting enzyme, is measured using Caspase-1 Assay (Fluorometric) assay kit (Abcam). HBMECs are seeded on a black with clear-bottom 96-well plate and treated per experimental protocol. Following the treatment incubation period, cells are lysed using Lysis Buffer IV. Reaction Buffer I and YVAD-AFC substrate is added to wells and incubated at 37 °C for 2 hours, protected from light. Cleavage of YVAD-AFC substrate reveals a yellow fluorescence that may be measured by microplate reader with excitation/emission of 400/505 nm. Absence of cleavage by caspase-1 produces a blue emission. Relative change in enzyme activity is calculated against non-induced control cells and data are reported as Relative Fluorescent Units (RFU).

MMP-9:

Activity of Matrix Metalloproteinase (MMP)-9, also known as Gelatinase-B and/or collagenase-IV, is measured using SensoLyte^® 520 MMP-Fluorometric Assay Kit (AnaSpec) to determine a potential mechanism by which endothelial cell barrier hyperpermeability may occur. Manufacturer’s protocol is followed. Briefly, cells are seeded on a black with clear-bottom 96-well plate and treated with the appropriate experimental conditions. A Förster Resonance Energy Transfer (FRET) peptide substrate is added and the activity of MMP-9 enzyme is recorded by microplate reader at excitation/emission of 520 nm where fluorescent intensity is directly proportional to enzyme activity. Data are reported as Relative Fluorescent Units (RFU) compared to untreated and non-induced control cells.

Calcein AM Cell Viability

EZViable^™ Calcein AM Fluorometric Assay Kit (Biovision) is used to quantify the number of viable cells as defined by an impenetrable plasma membrane. Briefly, the hydrophobic and non-fluorescent calcein AM crosses the cell membrane and is readily cleaved by intracellular esterases in metabolically active cells. Cleavage reveals a strongly fluorescent signal that is maintained in the cell cytoplasm and measured at an excitation/emission of 490/515 nm read on a microplate reader. Results are reported as Calcein AM Relative Fluorescent Units (RFU) and are directly proportional to viable cells.

XTT Cell Viability

CyQUANT XTT Cell Viability Assay (ThermoFisher) is used to validate the results of the calcein AM cell viability results. Viable cells are defined as metabolically active cells, as determined by the redox potential to reduce water-soluble XTT to a water-soluble formazan. Cells were seeded according to protocol on a 96-well plate. The XTT Reagent and Electron Coupling Reagent were mixed as a working solution and immediately added to cells, incubated for four hours at 37 degrees. Absorbance was read using a microplate reader at 450 nm (XTT-specific absorbance) and 660 nm (background absorbance). Data represent the delta value and are reported as XTT Relative Fluorescent Units (RFU), which are directly proportionable to viable cells.

Apoptosis/Necrosis Detection

To identify the number of apoptotic and necrotic (late-stage apoptosis) cells, a commercially available Apoptosis/Necrosis Assay (Abcam) is used. Cells are grown in a clear 96-well plate until desired confluency is reached. Treatments are given as defined in the experimental protocol. Cells are washed and resuspended in Assay Buffer. Apopxin Green Indicator, a fluorescent sensor of phosphatidyl serine (PS) is used for the detection of cells undergoing early-stage apoptosis; 7-aminoactinomycin D (7-AAD) is used to detect late-stage cell death or necrosis. Cells are read on a microplate reader at excitation/emission of 490/520 nm (apoptosis) and 550/650 nm (necrosis), Data are reported as Relative Fluorescent Units (RFU) as compared to untreated control cells.

ROS Detection

The production of reactive oxygen species (ROS) is measured using the ROS/Superoxide Detection Assay Kit (cell-based) (Abcam). The ROS inhibitor N-acetyl-cysteine (NAC) (5 mM) is added to control cells and incubated at room temperature for 30 minutes to provide a negative control, per manufacturer’s protocol. The ROS detection mix is added and incubated at room temperature for 60 minutes. Without washing, samples are analyzed in a microplate reader with an excitation/emission of 590/525. Data are reported as Relative Amount of ROS compared to untreated control cells.

Determination of Protein Levels

NLRP3 levels:

A calorimetric sandwich enzyme linked immunosorbent assay (ELISA) was used to quantify levels of NLRP3 sensor protein (Abcam). Cells were grown in a 6-well plate until confluent and treated per experimental design. Since NLRP3 is an intracellular protein, cells were harvested and cell lysate was used to determine protein levels. When ready for harvesting, cells were washed with phosphate buffered saline before 1 mL extraction buffer was applied. Cells were scraped and collected, vortexed, and incubated on ice for 30 minutes. Following incubation, cells were centrifuged at 15,000 x g at 4° Celsius for 15 minutes. The supernatant was collected and protein levels were determined per manufacturer’s protocol using the Human NLRP3 ELISA Kit (Abcam). Absorbance was read at 450 nm on a Molecular Devices microplate reader with SoftMax Pro 7.1 software.

IL-1β levels:

Levels of IL-1β were measured using a colorimetric sandwich ELISA (Abcam). Cells were cultured in a 6-well plate until confluent and treated per experimental design. Since IL-1β is expected to be secreted from cells following NLRP3 inflammasome activation, the cell culture supernatant, rather than cell lysate, was collected and vortexed before incubating on ice for 30 minutes. Following incubation, the cells were centrifuged for at 15,000 x g at 4° Celsius for 15 minutes. The protocol for Human IL-1β ELISA Kit (Abcam) was followed. Absorbance was measured at 450 nm using a Molecular Devices microplate reader on SoftMax Pro 7.1 software.

Statistical Analysis

GraphPad Prism version 10.6.01 is used to analyze and report results. All experiments follow the workflow: (1) generate descriptive statistics to define the mean, medium, range, max, min, standard deviation, standard error of the mean, and upper and low confidence intervals (95% and 5%, respectively); (2) test for Gaussian distribution using the Shapiro-Wilk Test for normality at p<0.05; (3) perform the appropriate statistical test based off the previously defined parameters. In the case of two groups, a student’s T test is used to compare the mean values between groups. Under conditions of normally distributed data, a one-way parametric (ordinary) analysis of variance (ANOVA) is used to test the difference in mean values. Tukey’s post hoc test then used to determine which groups differ from one another. In the case of non-parametric (non-normal distribution) data, the Kruskal-Wallis test followed by Dunn’s post hoc analysis is used to compare rank mean values. Statistical significance is defined at p < 0.05.

Figures and Graphs

All schematic figures are created using BioRender.com. All graphs are generated using GraphPad Prism. Mean values are presented within all bar graphs and “ns” denotes no statistically significant difference between groups; “*” denotes statistical significance at p<0.05; “**” denotes statistical significance at p<0.01; “***” denotes statistical significance at p<0.001. Sample sizes (n=) are reported as the number of wells used in one individual experiment and are listed above each graph. The range of datapoint within one experiment may be visualized as the number of dots within the bar graphs. All experiments are performed at minimum thrice.

Results:

Tau oligomers induce microvascular hyperpermeability

Tau oligomers, when incubated overnight with HBMECs, induce microvascular hyperpermeability compared to control as measured by FITC-dextran fluorescent permeability assay (Figure 2b) and TEER (Figure 2c). Barrier permeability is significantly increased at 10 nM, 100 nM, and 250 nM concentrations. Hyperpermeability was attenuated with inhibition of NLRP3 inflammasome using MCC950. Microvascular permeability was not significantly altered when HBMECs were exposed to tau monomers (Figure 2a) or tau fibrils (Figure 2d, 2e).

Figure 2. Tau polyforms differentially alter endothelial cell permeability. Hyperpermeability is dependent on NLRP3 inflammasome signaling.

Commercially purchased tau protein monomers, oligomers (oTau), and fibrils (fTau) were incubated with HBMECs overnight at differing concentrations. (a) Tau monomers had no significant effect on endothelial cell permeability at 1 μg/mL and 5 μg/mL as determined by FITC-dextran fluorescent tracer assay. (b) FITC-dextran fluorometric assay shows oTau significantly induced hyperpermeability at 10 nM, 100 nM, and 250 nM with no dose-dependent effects; blocking of NLRP3 inflammasome attenuated increases in permeability. (c) Results from TEER support the finding that hyperpermeability is induced by oTau, however, MCC950 did not significantly affect permeability. (d) fTau did not significantly alter permeability at 5 nM and 25 nM according to FITC-dextran fluorescent or (e) TEER measurements. H₂0₂ (100 μM; 2 hrs) is used as positive control. RFU = relative fluorescent units. FITC = fluorescein isothiocyanate-labeled. TEER = trans-endothelial electrical resistance. One-way ANOVA with Tukey’s test was used to determine mean differences between groups.

Microvascular hyperpermeability is not due to cell inviability, apoptosis, or necrosis

The microvascular hyperpermeability that occurs when HBMECs are exposed to 100 nM of tau oligomers is not a result of inviable cells as measured by calcein AM (Figure 3a), a measure of cell membrane integrity or XTT (Figure 3b), a measure of cellular redox potential. At concentrations that induce barrier hyperpermeability (e.g. 100 nM), tau oligomers do not significantly increase apoptosis (Figure 3c) or necrosis (Figure 3d) as measured by phosphatidylserine (PS) and 7-AAD fluorescent sensors, respectively. These data suggest that oTau alters HBMEC permeability via other mechanisms.

Figure 3. At concentrations that induce hyperpermeability, tau oligomers do not alter cell viability, increase apoptosis, or necrosis.

oTau (100 nM; overnight) did not significantly alter the cell viability of HBMECs, as indicated by (a) calcein AM, a measure of cell membrane integrity. (b) XTT assay, a measure of cellular redox potential, validates calcein AM findings. Cell viability significantly decreased following 250 nM oTau exposure. Tau oligomers do not increase apoptosis (c) or necrosis (d) at concentrations less than 500 nM, as indicated by phosphatidylserine (PS) or 7-AAD fluorescent signaling, respectively. H₂0₂ (100 μM; 2 hrs) is used as a positive control. XTT = (2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide); PS = phosphatidylserine; 7-AAD = 7-amino-actinomycin D. RFU = Relative Fluorescent Units. Student’s T- test was used to determine differences in calcein AM RFU. One-way ANOVA with Tukey’s test was used to determine mean value differences in experiments measuring XTT, PS, and 7-AAD RFUs.

Tau oligomers generate reactive oxygen species (ROS) and increase activity of NLRP3 inflammasome signaling

Exposure of HBMECs to tau oligomers at 100 nM increases the generation of ROS formation compared to untreated control conditions. Increases in ROS are mitigated with direct inhibition of NLRP3 inflammasome (Figure 4a). oTau increases levels of NLRP3 sensor protein (Figure 4c) 118.70%, from 2.31 ng/mL to 5.03 ng/mL. Increased protein levels are not due to increased transcriptional activity as oTau had no effect on mRNA expression (Figure 4b). Tau oligomers also increased the enzymatic activity of the NLRP3 effector protein, caspase-1 (Figure 4e) with no significant effect on mRNA expression (Figure 4d). oTau did upregulate both mRNA expression (Figure 4f) and protein levels (Figure 4g) of IL-1β, a downstream effector of NLRP3 inflammasome. The relative activity of MMP-9 enzyme increased 32.81% (Figure 4i) in response to oTau, with no change in mRNA expression (Figure 4h). Blocking studies indicate that inhibition of the NLRP3 inflammasome with MCC950 reduces observed changes in NLRP3 (Figure 4f) and IL-1β levels (Figure 4h) as well as in caspase-1 (Figure 4g) and MMP-9 activity (Figure 4i).

Figure 4. Tau oligomers increase ROS and NLRP3 inflammasome signaling.

oTau (100 nM; overnight) (a) increased ROS formation relative to control as measured by Oxidative Stress Detection Reagent. Inhibition of NLRP3 inflammasome with MCC95 significantly attenuated ROS increase. N-acetyl-cysteine (NAC), an ROS inhibitor, is used as a negative control (5 mM; 30 mins). oTau did not increase mRNA expression of (b) NLRP3 inflammasome or (d) caspase-1 enzyme, however increases in (c) NLRP3 inflammasome levels and (e) caspase-1 activity did significantly increase, as measured by ELISA and fluorometric assay. Increases were mitigated with MCC950. Tau oligomers increased both (f) mRNA and (g) levels of IL-1β in cell culture supernatant and the latter was attenuated by MCC950. While tau oligomers did not alter (h) mRNA expression of MMP-9, (i) increased MMP-9 activity was measured and modified by MCC950. NLRP3 inflammasome blocking studies performed with MCC950 (1 μM; 2 hrs). Changes in gene expression are calculated using delta-delta-Cq method and mRNA expression is reported as normalized relative fold change. Students T-test was used to determine differences between NLRP3, caspase-1, IL-1β, and MMP-9 mRNA expression, respectively. One-way ANOVA with Tukey’s test was used to determine mean values between groups in all other experiments.

Discussion:

BBB compromise is increasingly recognized as a major component of neurological disorders, including tauopathies, neurodegenerative diseases driven by pathological tau protein. While misfolded and aggregated tau is known to play a role in neuroinflammation, neurodegeneration, and BBB dysfunction, the effects of tau polymorphs on microvascular permeability is not clearly defined, particularly at the level of cerebral endothelial cells. Our findings demonstrate that tau aggregates, specifically tau oligomers, induce NLRP3 inflammasome-dependent CEC barrier hyperpermeability. Conversely, Tau monomers have no effect on barrier permeability, and tau fibrils do not acutely alter CEC permeability in a statistically significant manner.

These data are consistent with other studies pointing to extracellular tau oligomers as drivers of pathology (10, 14, 43-46). Hossen et al. (43) show that oligomeric tau decreases the expression of tight junction proteins through the generation of oxidative stress and RhoA/ROCK signaling. RhoA/ROCK signaling has been shown to regulate NLRP3 inflammasome activation in vitro and in vivo (47). Hossen et al. furthermore show evidence of oTau mediated hyperpermeability through the downregulation of key tight junction proteins that are known to mediate barrier integrity. The Lasagna-Reeves Lab provides evidence for tau oligomers as early neurodegenerative agents (44, 45) in both animal models and human studies, and proposes targeting tau oligomers to ameliorate the most prevalent tauopathy, Alzheimer’s disease. Castillo-Carranza (48) uses a mouse model to show that reductions in tau oligomers reverses tauopathy phenotype, and that phenotypic changes occur without altering the levels of tau monomers or tau fibrillary tangles (48). Fox et al. (49) show that tau oligomers, not neurofibrillary tangles, disrupt neural system integration in mice. Taken together, these data provide strong evidence for the toxicity of tau oligomers over tau fibrils, neurofibrillary tangles, and other tau polymorphs.

Contrary to our findings, however, other groups have reported that exposure to tau fibrils may induce barrier hyperpermeability. Guzman-Hernandez and Fossati (2023) use an Electrical Cell Impedance Sensing-Zθ (ECIS) system to measure changes in TEER following CEC exposure to protofibrillar tau aggregates. The authors show progressive loss of barrier integrity following a 48 hour exposure period to 5 nM and 25 nM of tau protofibrils. The researchers also report decreases in tight junction protein expression occurring at 25 nM fTau concentrations. While our experiments did not support these findings, we did observe trends towards hyperpermeability in the context of fTau exposure, albeit these changes were not statistically significant. These discordant findings may be due to several reasons. Firstly, our experimental design sought to simulate acute exposure to tau aggregates and therefore, our incubation period was 24 hours rather than 48 hours. It is possible and perhaps likely that longer exposure times results in significant barrier damage. Secondarily, our model leveraged newly available recombinant tau fibrils, which ensures standardization of fTau exposure across all experiments. Before recombinant tau aggregates were commercially available, in vitro models of tau pathology used aggregation protocols to create tau aggregates from tau monomers. One commonly used protocol for tau aggregation is through the addition of a polyanionic cofactor, such as heparin or arachidonic acid (36, 50, 51). The translational utility of this method however, has recently been questioned after important differences have been established between aggregates made from acidification protocols and the tau aggregates isolated from human patients diagnosed with tauopathy (52, 53). Zhang et al (53) use cryo-electron microscopy (EM) to highlight some of these differences (53). As the impact of tau polymorphs on the pathophysiology of BBB functioning is still being untangled, researchers should interpret findings cautiously and a commitment to methodological transparency must be prioritized.

Our studies further contribute to the study of aberrant microvascular functioning in the context of tauopathy by identifying NLRP3/MMP-9 as a mechanism by which CEC hyperpermeability occurs. Previously published data using animal studies (31, 54, 55), gene expression analysis (22, 29, 55, 56), and in vitro modeling (56, 57) have consistently demonstrated an important role for NLRP3 inflammasome signaling in tauopathy and neurodegenerative disease more broadly. Primarily, these studies have focused on immune cell activation that results in barrier disruption and/or neuronal damage (32) or on CECs via other signaling pathways (47). Our studies, however, show that NLRP3 inflammasome signaling occurs directly in CECs, independent of immune cell activity. Importantly, we observed increases in NLRP3 inflammasome levels and MMP-9 enzyme activity without corresponding increases in gene expression. These findings are consistent with what is currently reported in the literature and supports previous observations that NLRP3 inflammasome components and downstream effectors exist in their inactive forms intracellularly. Cellular stressors, including ROS formation, may incite intracellular changes that increase signaling of the NLRP3 inflammasome pathway, ultimately resulting in increased IL-1β and MMP-9 activity. Our lab has previously shown that increased activity of NLRP3 inflammasome signaling in CECs by other inducers of neuroinflammation and neurodegeneration (e.g. extracellular ATP) results in MMP-9 mediated disruption of tight junction proteins, including zonula occludens (ZO)-1 and claudin-5 (38). Both ZO-1 and claudin-5 are key junctional proteins that mediate barrier integrity and our previous published data indicate that MMP-9 mediated ZO-1 and claudin-5 disruption are linked to CEC barrier hyperpermeability. These data are supported by Hossen et al. (47) who, as referenced above, show evidence of oTau mediated barrier hyperpermeability occurring in the context of decreased tight junction proteins expression, as well as several others who have shown that CEC permeability occurs due to tight junction protein changes (2, 25, 27). Together, these data provide evidence that NLRP3 inflammasome and MMP-9 are key effectors of barrier disruption following exposure to tau oligomers.

While we provided strong evidence for the hypothesis that tau oligomers increase NLRP3 inflammasome signaling in CECs, limitations to our findings exist. While the use of recombinant tau oligomers allows for a standardized methodology, there are limitations to its applicability in physiological models. As already mentioned, tau is a highly dynamic molecule that may exist in transient states, varying across microenvironments. Future studies should continue to investigate the role of tau polymorphs on models of BBB structure and function to clearly define causal mechanisms. Secondly, our experimental design studies the acute effect of tau polymorphs on CECs using a 24 hour exposure period. Longer exposure times may alter outcomes and may be used to shed light on sub-acute and chronic exposures with relevance to chronic neurodegenerative processes and may provide rationale for further investigation of fTau and mechanisms governing aberrant CEC functioning. Thirdly, our model utilizes a monolayer model of the BBB to allow for detailed investigation into the role of CECs in response to tauopathy. While advantageous in that this model provides an isolated understanding of CEC-specific mechanisms, the converse is that this model lacks insight on signaling interactions and cross-talk among neighboring BBB cells. Future studies can incorporate a more complex co-culturing modeling of the BBB to understand how interacting cells influence endothelial cell signaling to investigate the potential for additive or synergistic effects. Fourthly, while our studies identify the generation of ROS formation as one possible regulator for NLRP3-dependent changes, further investigation into the upstream mediators of NLRP3 inflammasome are warranted to elucidate the full mechanistic pathway in CECs. Finally, our studies identify the canonical pathway for NLRP3 inflammasome activity as a mechanism for endothelial cell barrier hyperpermeability. This hypothesis was formed based off the existing literature, which currently favors NLRP3 inflammasome as a mediator for neuroinflammatory and neurodegenerative disease. It is likely, however, that multiple interacting mechanisms are involved in CEC barrier dysfunction and these remain to be clarified. Many outstanding questions on tau pathology and microvascular dysfunction persist. Our study adds value to the field by identifying NLRP3 inflammasome signaling as a mechanism for CEC hyperpermeability following tau oligomer exposure.

In conclusion, our findings demonstrate that oligomeric tau induces cerebral endothelial cell hyperpermeability while increasing NLRP3 inflammasome signaling and MMP-9 enzymatic activity. These data position CEC as a target of tau pathology and posit inhibition of NLRP3 inflammasome signaling as a potential therapeutic strategy in mitigating BBB dysfunction.

Perspectives:

This pre-clinical study characterizes the effect of tau protein polymorphs, including tau monomers, tau oligomers, and tau fibrils, on human microvascular endothelial cell (HBMEC) permeability to elucidate the effects of tauopathy on BBB functionality at the level of cerebral endothelial cells (CECs). CECs exposed to tau oligomers, but not tau monomers or tau fibrils, induce barrier hyperpermeability and increase activity of the NLRP3 inflammasome signaling pathway with subsequent increases in pro-inflammatory IL-1β and corresponding increases in the known CEC barrier disruptor, MMP-9 enzyme. This study demonstrates a role for NLRP3 inflammasome signaling in barrier dysfunction and positions NLRP3 inflammasome as a potential therapeutic target for ameliorating tau protein pathology.

Supplementary Material

Supplemental Data

Table 1. Primer sequences used in RT-qPCR experiments.

List of Abbreviations

BBB

Blood Brain Barrier

CEC

cerebral endothelial cell

HBMEC

human brain microvascular endothelial cells

oTau

oligomeric tau or tau oligomers

fTau

fibrillar tau or tau fibrils

NLRP3

NOD-like receptor pyrin domain containing-3

MMP-9

Matrix Metalloproteinase-9

IL-1β

interleukin-1 beta

MCC950

a direct NLRP3 inflammasome inhibitor (molecular formula: C₂₀H₂₄N₂O₅S)

Data availability:

All data are available in the main text.