Inhibition of Soluble Epoxide Hydrolase Confers Neuroprotection and Restores Microglial Homeostasis in a Tauopathy Mouse Model

TL;DR

Blocking an enzyme called sEH protects the brain and reduces harmful inflammation in a mouse model of tau-related neurodegeneration.

Contribution

This study shows that sEH inhibition or EET augmentation benefits tauopathies by targeting both neurons and microglia.

Findings

sEH inhibition improved cognition, reduced Tau pathology, and rescued neuronal loss in a mouse model.

TPPU treatment upregulated genes related to actin cytoskeleton and excitatory synapses in neurons.

sEH inhibition reversed harmful microglia states and improved Tau clearance in primary cultures.

Abstract

The epoxyeicosatrienoic acids (EETs) are derivatives of the arachidonic acid metabolism with anti-inflammatory activities. However, their efficacy is limited due to the rapid hydrolasis by the soluble epoxide hydrolase (sEH). Accordingly, inhibition of sEH has been shown to stabilize the EETs and dampen neuroinflammation in Aβ mouse models of Alzheimer’s disease (AD). However, the role of the sEH-EET signaling pathway in other cell types of the CNS and in other neurodegenerative conditions are less understood.

Here we examined the mechanisms and the functional role of the sEH-EET axis in tauopathy by treating the PS19 mice with a small molecule sEH inhibitor TPPU and by crossing the PS19 mice with Ephx2 (gene encoding sEH) knockout mice, followed by single-nucleus RNA-sequencing (snRNA-seq), biochemical and immunohistochemical characterization, and behavioral analysis. We also tested the effect of the sEH-EET pathway in primary microglia cultures and human induced pluripotent stem cell (iPSC)-derived neurons that develop seeding-induced Tau inclusions.

We show that sEH inhibition improved cognitive function, rescued neuronal cell loss, and reduced Tau pathology and microglia reactivity. snRNA-seq revealed that TPPU treatment resulted in the upregulation of actin cytoskeleton and excitatory synaptic pathway genes. Treating the human iPSC-derived neurons with TPPU led to enhanced synaptic density without affecting Tau accumulation, indicating a cell-autonomous effect of sEH blockade in neuroprotection. Further, sEH inhibition reversed disease-associated and interferon-response microglia states in PS19 mice and EET supplementation enhanced Tau phagocytosis and clearance in primary microglia cultures.

These findings demonstrate that sEH blockade or EET augmentation confer therapeutic benefit against neurodegenerative tauopathies through parallel targeting of neuronal and microglial pathways.

Figures

A](#F1)), which assesses the hippocampus dependent long-term recognition memory by measuring the percent time spent exploring a novel object. The four groups did not exhibit object bias during the training phase (Figure S3E). However, compared with the vehicle treated WT (WT_Veh) mice, the PS19_Veh group displayed no preference to the novel object ([Figure 1A](#F1)). TPPU treatment did not exert any appreciable effect in WT mice (WT_TPPU) similar to our previous study [[23](#R23)], but significantly improved the NOR score in PS19 mice (PS19_TPPU, [Figure 1A](#F1)). We further performed fear co

E](#F2) and [2F](#F2)), consistent with the anti-inflammatory activity of sEH blockade reported before [[20](#R20)]. Examination of microglia morphology using IMARIS imaging documented reduced filament length and branches and increased cell volume and surface area in PS19 microglia, which were normalized by TPPU treatment ([Figure 2G](#F2) and [2H](#F2)). Reduced Tau pathology and neuroinflammation was also observed by genetic knockout of *Ephx2* on PS19 background (Figure S4) [[26](#R26)].

D](#F3)). Consistent with the immunostaining results, TPPU treatment resulted in robust reductions of microglia population and restoration of DG cells.

A](#F4) and Figure S6A). We found no appreciable differences between WT_Veh and WT_TPPU microglia, with both consisted mostly of HM ([Figures 4B](#F4) and [4C](#F4)). In PS19 microglia, the HM cluster was drastically diminished while the transitional, DAM and IFN clusters were greatly expanded. Pseudotime trajectory analysis based on RNA velocity indicated that DAM was derived from transitional microglia while IFN may be directly converted from HM ([Figures 4D](#F4)). TPPU treatment led to substantial reductions of DAM and IFN clusters and partial replenishment of the HM pool while the transit

A](#F5)). Regardless, GO term analysis documented that, while comparison of PS19_Veh vs WT_Veh failed to identify enriched pathways, TPPU treatment of PS19 mice resulted in the upregulation of multiple neuronal and synapse pathway genes compared to vehicle treated group ([Figure 5B](#F5)). These included genes related to glutamatergic synapse function, axon guidance and actin cytoskeleton organization ([Figure 5C](#F5)). Of note, numerous postsynaptic genes, including *Dlg4*, *Grik3*, *Glk4*, *Shank1*, *Shank2* and *Shank3*, were downregulated in PS19_Veh group, which were rescued by TPPU trea

Full text

Background

Alzheimer’s Disease (AD) is a progressive neurodegenerative disorder manifested by the deposition of beta amyloid (Aβ) plaques, the accumulation of neurofibrillary tangles (NFTs) composed of aggregated Tau protein, and extensive neurodegeneration. These pathological events are accompanied by the hyperactivation of microglia and astrocytes and chronic neuroinflammation [1]. Although anti-Aβ antibodies have shown clinical efficacy in prodromal AD [2, 3], it is known that NFTs correlates more closely with cognitive decline than Aβ plaques [4]. Therefore, anti-Aβ therapies may not be effective in late stages when Tau pathology and neuroinflammation ensue. In addition, NFTs is a pathological feature of a group of diseases combinedly termed tauopathies and can cause neurodegeneration in and of itself [5, 6]. Thus, there is strong premise for developing therapies targeting Tau and neuroinflammatory pathways for AD and broader tauopathy conditions [1, 7, 8].

Arachidonic acid is a polyunsaturated fatty acid released from the membrane phospholipids, which can be further metabolized to epoxyeicosatrienoic acids (EETs) by the cytochrome P450s (CYPs) [9]. The EETs have been demonstrated to possess inflammation resolving and anti-inflammatory properties [10]. However, their efficacy is limited due to the rapid hydrolysis by the soluble epoxide hydrolase (sEH). sEH, encoded by the EPHX2 gene, is widely expressed in both the peripheral tissues and the central nervous system (CNS) [11][12]. Elevated sEH levels have been implicated in neurovascular and neurological conditions such as vascular cognitive impairment [13], depression [14], schizophrenia [15], and Parkinson’s disease [16]. Multiple lines of evidence support a direct involvement of sEH in AD. Specifically, EPHX2 is located within the PTK2B-CLU AD risk locus in chromosome 8 where a disease-associated polymorphism has been reported to influence its expression [17, 18]. Proteome-wide association studies also revealed a potential link of EPHX2/sEH with AD [19, 20]. As well, targeted lipidomic analyses of brain lysates, plasma and cerebrospinal fluid samples identified dysregulation of CYP and sEH metabolic pathways in AD [21, 22]. Consistent with the human data, sEH inhibition by genetic deletion or pharmacological blockade has been shown to augment the EET levels and reduce Aβ pathology in mice [23–25]. While these effects have been largely attributed to the immune modulatory function of the EETs on glial cells, particularly microglia, whether the sEH-EET axis acts on other cell types such as neurons are not well-understood. In addition, a possible role of sEH inhibition in Tau pathogenesis has not been reported.

Here we treated the PS19 Tau P301S transgenic mice that develop age-dependent pathological Tau accumulation and neuronal cell loss with a selective sEH inhibitor, 1-trifluoromethoxyphenyl-3-(1-propionylpiperidin-4-yl) urea (TPPU) [11, 14]. We also crossed the Ephx2 null mice with the PS19 mice [26]. We show that both genetic ablation and pharmacological inhibition of sEH reduced microglia reactivity, attenuated Tau pathology and rescued neurodegeneration and cognitive impairment. Single nucleus RNA-sequencing (snRNA-seq) revealed microglia and dentate granule cells as the major cell types altered in PS19 mice, which were rescued by TPPU treatment. Using primary microglia cultures and human induced pluripotent stem cell (iPSC)-derived neurons [27], we demonstrate a direct effect of neuronal sEH inhibition in promoting synaptic health while augmenting the EET improved microglia phagocytosis and Tau clearance.

Methods

Study design

The goal of this study was to determine the role of the sEH-EET epoxy lipid signaling pathway in diseases of tauopathy including AD, using a small molecule sEH inhibitor TPPU or by genetic ablation of Ephx2 in PS19 mice. Additionally, we treated the iPSC neurons or primary microglia cultures with TPPU or EET, respectively, to evaluate the mechanisms. The animals, plates, and slides were randomized prior to the experimentation and were blinded to the experimenters. The minimum number of mice for all experiments was at least six per group based our previous studies [23, 28]. The exact number of mice are listed in the figure legends and marked on the bar graph figures. All in vitro experiments were performed at least twice, each with at least three technical repeats. Additional experimental details are described in this section and in figure legends.

Human subjects

Postmortem brain tissues were provided by the University of Pennsylvania Center for Neurodegenerative Disease Research. Informed consent was obtained from all subjects. The demographic data can be found in Supplemental Table 1. The influence of gender was not analyzed due to small sample size.

Mice and treatment

The PS19 mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Ephx2 null mice were provided by B. Hammock [26]. They were housed in a pathogen free mouse facility with ad libitum access to food and water on a 12 hr light/dark cycle. Male and female mice at approximately equal ratio were used for all experiments. All procedures were performed in accordance with NIH guidelines and approval of the Baylor College of Medicine Institutional Animal Care and Use Committee (IACUC).

Behavioral analysis

The sEH inhibitor TPPU was synthesized as described [11], and dissolved in 10% v/v aqueous polyethylene glycol 400 (PEG400, ThermoFisher). Wild-type and PS19 mice were treated with either vehicle (1% PEG400) or TPPU at 3 mg/kg via drinking water starting at 6–6.5 months and continuously for a total of 12 weeks. Every 2 weeks, mice were supplied with a new water bottle containing fresh TPPU or vehicle solution.

Behavioral assays were performed blind to genotype and treatment. Prior to each assay, mice were habituated in the test room (150 lux, 60 dB white noise) for at least 30 min. At least one day was given between different assays for mice to recover.

Novel object recognition was performed in a Plexiglass arene (22 cm by 44 cm). In the training day, mice were allowed to freely explore the arena with two identical LEGO objects for 10 min. Twenty-four hours post training, the mice underwent testing, in which the mice were placed in the same arena with one object previously explored during the training phase and one novel object differing in color and shape but sharing a common size and volume. The animals were allowed to freely explore the objects for 10 min. The time spent exploring each object was measured by ANYmaze software. Exploration of an object was defined by the distance of nose toward the object at less than or equal to 2 cm. The exploration preference was calculated by percent time spent on the novel object.

The fear conditioning protocol involved a training phase, a contextual test, and a cued test as previously described [29]. During the training phase, the mice were placed in the training chamber and allowed to freely explore the environment for 2 min, then an 80-dB and 5 kHz tone was presented (auditory conditioned stimulus (CS)) for 30 seconds, immediately followed by a foot shock (0.8 mA) for 2 sec, with the conditioning pattern repeated once. The animals were returned to their original housing cages afterwards. The contextual test was performed 24 hours later by returning the mice to the same training chamber for 5 minutes with no presentations of shock or CS. One hour later, the cued fear test was performed by placing the mice to a cued chamber consisting of a different geometric shape, flooring, light brightness and scent. After 3 minutes in the chamber, the auditory stimulus was presented for 3 minutes. The percentage of time freezing in each trial was recorded by FreezeFrame4 (Actimetrics, Coulbourn Instruments) software.

After behavioral assays and a total of 12 weeks of treatment with TPPU or vehicle, mice were sacrificed. The brains were perfused with 0.01M PBS (pH 7.4), microdissected, and frozen or fixed in 4% paraformaldehyde (PFA) for further analysis.

Unbiased stereology

The unbiased stereology was performed in brain slices from 9.5–10-month-old mice as previously described [30]. Briefly, brain sections were immunofluorescence stained using the anti-NeuN antibody. The number of neurons in selected region was counted from six slides of sagittal sections (30 μm thickness, 300 μm apart). Unbiased stereology was performed using optical dissector method on a Zeiss microscope with motorized stage. Data acquisition and analysis was performed using Stereo Investigator software by MBF Bioscience. The selected region of counting was outlined under 10× objective lens, NeuN-positive neurons were counted in counting frame area of 2500 μm^2^ with sampling grid area of 4900 μm^2^ under 40× objective lens and 15 optical counting dissectors.

Single nucleus RNA-sequencing and data analysis

The snRNA-seq was performed as described [31]. Specifically, the 9.5-month-old WT_Veh, WT_TPPU, PS19_Veh and PS19_TPPU mice (2M+2F/group) were perfused transcardially with cold saline. Brain hippocampi were dissected and dissociated, nuclei were stained by Hoechst-33342 and collected via the SONY SH800 FACS sorter. For each 10x Genomics run, 100k–400k nuclei were collected. 16k nuclei for each channel were loaded into the 10x controller. snRNA-seq library was performed by the 10x Genomics 3’ v3.1 kits. All PCR reactions were performed using the Biorad C1000 Touch Thermal cycler with 96-Deep Well Reaction Module. 13 cycles were used for cDNA amplification and 16 cycles were used for sample index PCR. As per 10x protocol, 1:10 dilutions of amplified cDNA and final libraries were evaluated on a bioanalyzer. The male and female samples were pooled sequenced using 100 cycle run kits and the Single Index configuration. Gene expression libraries were sequenced using paired end 150bp reads to an average 60–70% sequencing saturation on an Illumina Novaseq.

Raw sequencing files were demultiplexed and gene expression libraries were processed using bcl2fastq. After mapping sequencing reads to mm10 reference genome by cellranger-7.1.0, the ambient RNA for each single cell was estimated and removed by Cellbender [32]. Next, doublet cells were identified using Scrublet from the filtered feature barcode matrices produced by Cellbender [33]. Cells were scored as candidate doublets by Scrublet and removed if their doublet score exceeded 0.25. Finally, remaining cells were filtered to have less than 5% of their UMI’s mapped to mitochondrial genes and to express greater than 200 genes by Scanpy [34].

Single-cell analysis was conducted by following the scanpy pipeline (v1.9.5). Single cells were batch corrected and projected into a low-dimensional representation by single cell Variational Inference [35]. Major cell types and subclusters were marked and identified using well-known marker genes. Spliced and unspliced transcript counts were determined using the run10x command from velocyto and normalized and smoothed across 15 nearest neighbors using the filter_and_normalize and moments functions from the scVelo [36, 37]. Global scVelo velocity, transition matrix estimates and streamline visualizations were then calculated using the velocity and velocity_embedding_stream functions from the scVelo package with Dynamical Modeling. The normalized spliced and unspliced transcript counts act as the input matrix for the pseudo-time trajectory inference and the visualization of pseudotime UMAP. Differential gene expression analysis was performed between different groups and subtypes by using the Wilcoxon test implemented in the “FindMarkers” function, only genes that passed FDR <0.05, average log2 fold change >0.1, and % non-zero expressing cells >10% were retained as differentially expressed genes.

Tau preformed fibrils (PFF) preparation

Human truncated tau containing four microtubule-binding repeats (4R) with a 5′ Myc tag was cloned into the pRK172 plasmid (a generous gift from Virginia Lee) and expressed in BL21(DE3) cells. Recombinant tau was purified as described previously [38] with minor modifications. Briefly, protein expression was induced with 1 mM IPTG for 2 hours. The cell pellets were resuspended on ice in 40 mL of buffer containing 750 mM NaCl, 20 mM NaF, 100 mM MES (adjusted to pH 7.0 with 5 N KOH), 10 mM EGTA, and 5 mM MgSO4, supplemented with protease inhibitors (Recom ProteaseArrest^™^ Protease Inhibitor Cocktail; G-Biosciences), PMSF (0.1 mM), and lysozyme (1 mg/mL). After incubation for 20 minutes, DTT (5 mM) and 10% Triton X-100 (4 mL) were added. The mixture was then boiled in water for 20 minutes with loosely capped tubes and subsequently cooled on ice for 10 minutes. After ultracentrifugation at 186,000 × g for 30 minutes at 4°C, the supernatant was collected and subjected to dialysis against FPLC buffer (20 mM MES, pH 6.8, 10 mM NaCl, 1 mM EGTA, 1 mM MgSO4, 5 mM DTT), supplemented with 0.1 mM PMSF. The dialyzed sample was then purified using ion exchange chromatography on a 5 mL GE HiTrap SP HP column. Fractions containing tau were identified by Coomassie-stained SDS-PAGE and further purified using a Pierce High-Capacity Endotoxin Removal Spin Column (ThermoFisher). The protein was concentrated using an Amicon Ultra-15 centrifugal filter (3-kDa MWCO) and buffer-exchanged to PBS. The protein concentration was determined using a BCA assay, and tau monomers were snap-frozen and stored at −80°C.

Tau fibrillization was initiated by incubating tau monomer (120 μM in PBS) with 10 μM DTT, 40 μM heparin, and 0.1% sodium azide in a shaker at 37°C and 200 rpm for 3 days. Fibrils were pelleted by ultracentrifugation at 130,000 rpm at 4°C and resuspended in PBS. The fibril suspension was sonicated using a probe sonicator at 40% amplitude (30 seconds on/off cycles for 10 minutes). The concentration of PFF was determined using a BCA assay.

For fluorescent labeling, 170 μL of tau fibrils (100 μM) were mixed with 1 μL of Alexa Fluor 488 (10 mg/ml), 20 μL of 1 M sodium bicarbonate, and 9 μL of PBS. The reaction was incubated at room temperature in the dark for 1 hour. After labeling, the solution was centrifuged to pellet large aggregates. The supernatant, containing smaller fibrils, was passed through Zeba^™^ Spin Desalting Columns (7K MWCO; ThermoFisher) to remove excess dye. The pellet was washed twice with PBS to eliminate unbound dye. Finally, the washed pellet was combined with the column-treated supernatant and sonicated to produce smaller, uniform PFFs for downstream applications.

Human iPSC-derived neuronal cultures and treatments

The method for culturing isogenic human WT and P301S 4R-tau iPSC lines and differentiation to i^3^N neurons were adapted from [27]. Specifically, cells were maintained in Matrigel-coated 6 cm petri dishes (Millipore Sigma) using mTeSR Plus medium and passaged with Accutase (Stem Cell Technologies). For pre-differentiation, iPSCs were plated at 1.5 × 10^6^ cells per well in Matrigel-coated 6-well plates (Millipore Sigma) and cultured in Knockout DMEM/F-12 medium (ThermoFisher) supplemented with 2 μg/mL doxycycline (Sigma-Aldrich), 1X N2 supplement (ThermoFisher), 1X Non-essential amino acids (ThermoFisher), 10 ng/mL BDNF (PeproTech), 10 ng/mL neurotrophin-3 (PeproTech), 1 μg/mL mouse laminin (Millipore Sigma), and 2 μM ROCK inhibitor (Tocris). The medium was replaced 24 hours later, omitting ROCK inhibitor, and pre-differentiation was maintained for 3 days (day −3 to day 0). On day 0, pre-differentiated precursor cells were dissociated with Accutase and re-plated onto poly-D-lysine (PDL) and laminin-coated coverslips at a density of 5 × 10^4^ cells per coverslip in 24-well plates. Neuronal cultures were established in a 1:1 mixture of Neurobasal-A (ThermoFisher) and DMEM/F-12 media, supplemented with B27 and N2 (ThermoFisher), 0.5X GlutaMAX (ThermoFisher), 1X non-essential amino acids, 10 ng/mL BDNF, 10 ng/mL NT-3, 1 μg/mL mouse laminin, and 2 μg/mL doxycycline. On day 7, doxycycline was removed, and half of the medium was replaced every other day up to a total of 6 additional weeks (day 49). Cells were treated with either vehicle (DMSO) or TPPU (2.5 μM) along with the half medium change. The half of the medium was replaced every other day with fresh i^3^N maturation medium containing the respective treatments (vehicle or TPPU) to maintain a final concentration of TPPU at 2.5 μM until the cells were collected. Tau PFF was added to P301S i^3^Ns at 1.5mg/mL final concentration once on day 7, either with vehicle (DMSO) or TPPU (2.5 μM). Subsequent medium changes were performed every other day with fresh i^3^N maturation medium containing only the respective treatments (vehicle or TPPU) to maintain a final concentration of TPPU at 2.5 μM until the cells were collected.

Primary microglia cultures and treatment

Primary microglia cultures were prepared as described previously [39]. In brief, mouse cortices and hippocampi were isolated from postnatal day 1 to day 3 pups. Tissue was digested with 2.5% trypsin at 37°C for 15 min, centrifuged, triturated, and resuspended in DMEM medium with 10% FBS. Cells were plated onto PDL–coated T-75 flasks to generate mixed glial cultures. These mixed glia cultures were allowed to grow for 2 weeks, with refeeding every other day. The mixed glia cultures were subjected to separation via CD11b microbeads selection according to the manufacturer’s instructions (Miltenyi Biotec). Enriched microglia were plated in 24 well PDL-coated plates in DMEM with 10% FBS and 1% Pen/Strep and supplemented with 10 ng/mL M-CSF. Microglial cells were allowed to rest for 24–48 hours before treatments.

The cells were treated with vehicle (DMSO) or11,12- EET (10 μM) for the next 18 hours. For Tau fibril phagocytosis assay, fluorescently labelled (Alexa-488) Tau fibrils at 2.5 μg/ml were added to the culture for an additional 6 hours. The Tau containing media was removed and replaced with 0.01% Trypsin for 5 minutes to eliminate Tau attached to the cell surface. Thereafter the cells were washed with PBS. One set (0 hr) was fixed in 4% PFA for Tau phagocytosis, the other set was incubated in fresh media for another 24 hours to measure Tau clearance. Images were acquired by the Leica STELLAS confocal microscope using 40X objective.

Western blotting

Cells or brain tissues were collected and resuspended in modified radioimmunoprecipitation (RIPA) assay buffer containing protease and phosphatase inhibitor mixture. Cell suspensions were sonicated after resuspension, whereas mouse brain tissues were homogenized, sonicated, and then centrifuged at 14,000 × g for 45 min at 4°C, as described previously [23]. Briefly, protein concentrations were estimated using a BCA kit (Thermo Fisher). Lysates were separated on 7.5%–15% SDS-polyacrylamide electrophoresis gels (Bio-Rad). After the separation, proteins were transferred to a nitrocellulose membrane, and nonspecific binding sites were blocked by treating with either Odyssey blocking buffer (LI-COR) or TBS with 5% bovine serum albumin (BSA) followed by antibody incubation. The bands were quantified using Fiji ImageJ software and normalized to the loading control (GAPDH). The primary antibodies used are described in Table S3

Immunofluorescence staining

Mice were anesthetized and perfused transcardially with PBS and fixed in 4% PFA overnight at 4 °C, dehydrated with 30% sucrose in PBS and serially sectioned at 30 μm on a sliding microtome (Leica). Free floating sections were blocked with 10% donkey serum and 3% BSA in PBS with 0.3% Triton X-100 for 1 hour at room temperature. Sections were then incubated with primary antibodies (see Table S3 for details) in blocking solution overnight at 4 °C. On the next day, sections were washed and incubated with secondary antibodies and DAPI in blocking buffer for 1 h at room temperature. After washing with 0.1% Triton X-100 in PBS, sections were mounted with ProLong Gold Anti-fade reagent (Invitrogen).

Primary cultures or i^3^Ns grown on coverslips were fixed in 4% PFA for 20 minutes, permeabilized with 0.3% Triton X-100 in PBS for 5 minutes at room temperature. Coverslips were then incubated in blocking buffer (5% normal donkey serum, 0.01% Triton X-100 in PBS) for 1 hour at room temperature. After blocking, coverslips were incubated with primary antibodies (see Table S3 for details) overnight in blocking buffer at 4 °C. Coverslips were then washed in PBS followed by incubation with secondary antibodies and DAPI in blocking buffer for 1–1.5 hours at room temperature. After brief washing with PBS, coverslips were then mounted on slide glasses with ProLong Gold antifade reagent.

Image quantification

Area fluorescence in the hippocampus was averaged across at least 4 consistently represented sections (300 μm apart). Images were obtained with a Leica confocal microscope and analyzed by Fiji ImageJ. Background fluorescence was subtracted by the software before quantification.

In vivo synaptic colocalization quantification was performed as previously described [40]. Images were obtained by the Leica confocal microscope, using a 63X oil lens with 6X digital zoom. Z-stacks of 10 μm thickness were obtained with a 0.2 μm step size. Z-stacks with pre- and postsynaptic puncta were analyzed using the Spots feature of IMARIS. Spots were generated automatically with manual adjustment for accurate puncta representation for each channel separately, and total number of spots were recorded for each channel, and analyzed by the Co-localize Spots MATLAB (MathWorks) plugin. Pre- and postsynaptic puncta were defined as colocalized if their centers were within 200 nm.

For in vitro synaptic colocalization quantification, images were captured by the Leica confocal microscope using a 63X oil lens with 2.5X digital zoom. Synapses were identified as puncta positive for the presynaptic marker SYN1 and the postsynaptic marker PSD95, in proximity to the dendritic marker MAP2. Dendritic length and the number of synaptic puncta were measured using IMARIS software (Bitplane), employing the filament feature for MAP2+ dendrites and the spot feature for SYN1 and PSD95 puncta. Colocalization was quantified by counting the number of PSD95 puncta colocalized with SYN1 puncta within a 1 μm distance [41]. Synaptic density was determined by using the Coloc-channel in IMARIS software and calculating the number of voxels or volumetric pixels in SYN1, PSD95, or their colocalized puncta divided by the length of the dendrite.

For microglia morphology quantification, Iba1-positive microglia were imaged by confocal microscopy using a 63X oil lens with 4X digital zoom to generate Z-stacks of the tissue thickness (~30 μm) with a step-size of 0.2 μm. Images were analyzed using IMARIS software, the filament function was used to generate filaments for individual cell in the images and microglia processes were automatically rendered based on the Iba1 signal.

Statistical analysis

All data were analyzed with GraphPad Prism 10.4.1 and presented as mean ± SEM (*p < 0.05, ** p <0.01, *** p< 0.001 and **** p< 0.0001). For simple comparisons, Welch’s or student’s t-test were used. For multiple comparisons, one- or two-way ANOVA with Tukey’s multiple comparison test were used. All samples or animals were included in the statistical analysis unless otherwise specified.

Results

sEH inhibition rescues neuronal and synapse loss and cognitive impairment in PS19 mice

We reported that the sEH levels were elevated Aβ mouse models and AD brain samples [23]. To assess whether it is also the case in tauopathy diseases, we performed Western blot analysis of total and phospho-Tau levels in the hippocampus of PS19 transgenic mice and medial frontal cortices of human tauopathy patients including corticobasal degeneration (CBD), Picks disease, and progressive supranuclear palsy (PSP) (Supplementary Figure S1). While the sEH levels were comparable between WT and PS19 mice at 6 months, it was significantly higher in PS19 mice at 9 months when frank Tau pathology develops (Fig. S1, A-D). Upregulation of sEH expression was also observed in postmortem tauopathy patient samples compared to non-cognitively impaired controls (Figures S1E and S1F). These results implicate a dysregulation of the sEH pathway by Tau pathology and provide the rationale for testing the therapeutic effect of sEH inhibition in tauopathy conditions.

Accordingly, we administered the sEH inhibitor TPPU to PS19 mice of both sexes at 3 mg/kg via drinking water starting at 6–6.5 months of age when neuropathological changes start to manifest and continuously for 3 months [23]. Vehicle or TPPU-treated littermate wild-type (WT) mice were used as controls. We chose 9.5–10 months for analysis because Tau pathology, neuronal loss, immune system changes and behavioral deficits can be readily detected in PS19 animals [28]. Analysis of water consumption and body weight in WT and PS19 mice showed no appreciable differences between vehicle and TPPU treated groups (Figure S2), indicating that TPPU does not exert adverse reactions. Likewise, open field test revealed no group differences in locomotor activities or movement behavior (Figure S3, A-D).

We next evaluated the effect of TPPU in cognition using the novel object recognition (NOR) test (Figure 1A), which assesses the hippocampus dependent long-term recognition memory by measuring the percent time spent exploring a novel object. The four groups did not exhibit object bias during the training phase (Figure S3E). However, compared with the vehicle treated WT (WT_Veh) mice, the PS19_Veh group displayed no preference to the novel object (Figure 1A). TPPU treatment did not exert any appreciable effect in WT mice (WT_TPPU) similar to our previous study [23], but significantly improved the NOR score in PS19 mice (PS19_TPPU, Figure 1A). We further performed fear conditioning to test the hippocampal dependent (contextual) and independent (cued) associative learning (Figure 1B). All groups showed similar freezing percentage during the conditioning phase (Figure S3F). In the context test, the PS19_Veh group exhibited significantly decreased freezing percentage compared to WT_Veh or WT_TPPU groups, suggesting an impaired contextual memory. TPPU treatment of PS19 mice resulted in a close to significant increase (p = 0.071) in freezing frequency compared to the PS19_Veh group. The percentage of freezing in the post-cue test was indistinguishable between groups. Thus, TPPU treatment ameliorated the impaired hippocampus-dependent memory in PS19 mice.

To investigate the underlying mechanism of improved behavior by TPPU, we measured the NeuN-positive neurons in hippocampal area CA1 and dentate gyrus (DG) using unbiased stereology [30]. The results showed substantially reduced numbers in PS19_Veh group compared to WT_Veh and WT_TPPU controls and partial but significant rescue by TPPU treatment (Figure 1C and 1D). Similar results were obtained by co-immunostaining of pre- and post-synaptic proteins Bassoon and Homer1, respectively, and quantifying the number of pre- and post-synaptic as well as co-localized puncta (Figure 1E and 1F). Overall, these results demonstrate neuroprotective and synaptic promoting properties of TPPU, providing basis for its rescue of behavioral deficits in PS19 mice.

TPPU treatment ameliorates Tau pathology and neuroinflammation in PS19 mice

Having demonstrated a beneficial role of TPPU in reversing neuronal phenotypes of PS19 mice, we asked whether sEH inhibition could influence Tau pathology. Immunostaining using AT8 and MC1 antibodies (Figures 2A and 2B) and Western blotting using PHF1 and AT8 (Figures 2C and 2D) both showed significant reductions in TPPU treated PS19 mice.

The development of Tau pathology is accompanied by increased reactive gliosis and neuroinflammation. Indeed, immunostaining using anti-Iba1, -GFAP and -COX2 antibodies showed higher microglia and astrocyte immunoreactivities and proinflammatory marker COX2 levels in PS19 mice, all of which were significantly reduced by TPPU treatment (Figure 2E and 2F), consistent with the anti-inflammatory activity of sEH blockade reported before [20]. Examination of microglia morphology using IMARIS imaging documented reduced filament length and branches and increased cell volume and surface area in PS19 microglia, which were normalized by TPPU treatment (Figure 2G and 2H). Reduced Tau pathology and neuroinflammation was also observed by genetic knockout of Ephx2 on PS19 background (Figure S4) [26].

Reversal of microglial states in TPPU treated PS19 mice

We next sought to interrogate the cell types contributing to TPPU-mediated changes in PS19 mice by conducting single-nucleus RNA-sequencing (snRNA-seq) of hippocampal samples collected from 9.5-month-old WT and PS19 mice treated with vehicle or TPPU. Nuclei were isolated by fluorescence activated cell sorting (FACS). After stringent quality control including doublet removal, batch effect correction and normalization (Figure S5), we obtained a total of 84,226 high-quality single cell transcriptomes (Supplementary Table 2), which were annotated into 8 major cell types based on the expression of well-known cell-type-specific markers (Figures 3, A–C). Consistent with our previous study [31], cell type composition analysis revealed expanded microglia and loss of DG cells in PS19_Veh group compared to WT_Veh controls (Figure 3D). Consistent with the immunostaining results, TPPU treatment resulted in robust reductions of microglia population and restoration of DG cells.

Further analysis of microglia population identified four subclusters: homeostatic (HM), transitional, disease-associated (DAM), and interferon-responsive (IFN) (Figure 4A and Figure S6A). We found no appreciable differences between WT_Veh and WT_TPPU microglia, with both consisted mostly of HM (Figures 4B and 4C). In PS19 microglia, the HM cluster was drastically diminished while the transitional, DAM and IFN clusters were greatly expanded. Pseudotime trajectory analysis based on RNA velocity indicated that DAM was derived from transitional microglia while IFN may be directly converted from HM (Figures 4D). TPPU treatment led to substantial reductions of DAM and IFN clusters and partial replenishment of the HM pool while the transitional microglia population remained largely the same. Pathway analysis of differentially regulated genes (DEGs) showed that, compared to HM, the DAM was enriched in innate immune responses, endocytosis, and lysosomal pathways (Figures S6C and S6D), while the anti-viral defense and cytokine stimulus pathway genes were upregulated in the IFN subtype (Figures S6E and S6F). Both the DAM (Apoe, Axl, Cd9, Csf1, Ctsb, Ctsd, and Mitf) and IFN (Ifi204, Ifi207 and Stat1) genes were upregulated in PS19_Veh microglia and suppressed by TPPU treatment (Figures 4E and 4G). KEGG enrichment analysis identified lysosome, antigen processing and presentation, and cholesterol and lipid metabolism as top enriched pathways upregulated in PS19_Veh microglia, which were downregulated by TPPU (Figures 4F and 4H). These results demonstrate a broad reversal of microglia states by sEH inhibition.

TPPU promotes glutamatergic synapses in a cell-autonomous manner

Transcriptomic analysis of dentate granule cells across groups revealed limited DEGs (Figure 5A). Regardless, GO term analysis documented that, while comparison of PS19_Veh vs WT_Veh failed to identify enriched pathways, TPPU treatment of PS19 mice resulted in the upregulation of multiple neuronal and synapse pathway genes compared to vehicle treated group (Figure 5B). These included genes related to glutamatergic synapse function, axon guidance and actin cytoskeleton organization (Figure 5C). Of note, numerous postsynaptic genes, including Dlg4, Grik3, Glk4, Shank1, Shank2 and Shank3, were downregulated in PS19_Veh group, which were rescued by TPPU treatment (Figure 5C). Western blot analysis of selected presynaptic and postsynaptic proteins from the brain lysates also revealed reduced synaptic protein levels in PS19_Veh group and significant elevation of postsynaptic proteins by TPPU treatment (Figures 5D and 5E). These results indicated a potentially preferential effect of TPPU on postsynaptic compartment.

The improved neuronal survival and function in TPPU treated PS19 mice may be mediated by a direct effect of sEH inhibition in neurons or indirectly through immune pathway modulation. To distinguish these possibilities, we differentiated human iPSCs expressing isogenic WT or P301S mutant 4-repeat (4R) Tau to excitatory neurons (i^3^Ns) [27], and treated the cultures with either DMSO (vehicle) or 2.5 μM TPPU (Figure 6). Immunostaining of seven-week-old WT and P301S i^3^Ns with synapsin-1 (SYN1) and PSD95, alongside MAP2 to mark the dendrites, revealed a significant reduction in SYN1, PSD95 and colocalized SYN1/PSD95 puncta in P301S i^3^Ns compared to WT. Treatment with TPPU significantly restored the number of total SYN1, PSD95 and their colocalized puncta, as well as the number of voxels per 10 μm of MAP2-positive dendrites, in P301S i^3^Ns. We then seeded the P301S i^3^Ns with K18 preformed fibrils (PFF) at day 7 of differentiation to induce Tau aggregation [27, 31], and treated the cultures with vehicle or 2.5 μM TPPU continuously for a total of seven weeks. Immunostaining indicated that TPPU treatment increased both the total number and voxels per 10 μm of MAP2-positive dendrites of SYN1, PSD95 and their colocalized puncta or voxels in P301S i^3^Ns seeded with PFF (Figures 7A and 7B). Intriguingly, neither the total Tau nor AT8-positive Tau were affected by TPPU treatment (Figures 7C and 7D). Immunostaining staining also showed comparable levels of MC1 and AT8 immunoreactivities (Figures 7E–H). Thus, sEH inhibition conferred a synaptoprotective effect in mutant Tau i^3^Ns under both aggregated and non-aggregated conditions independent of Tau pathology.

Improved microglia phagocytosis and clearance by EET

Tau is known to be secreted from neurons and Tau pathology can transmit in a prion-like spreading manner [42–45]. As professional phagocytes, microglia actively take up and degrade extracellular materials, including Tau. Since sEH inhibition did not have an appreciable effect on Tau accumulation in iPSC-derived neurons, we wondered whether the reduced Tau pathology by TPPU treatment in PS19 mice was due to enhanced Tau uptake and clearance by microglia. We have shown that microglia do not express sEH but addition of exogenous 11,12-EET (herein referred to as EET) can suppress microglia inflammation [23]. We thus tested whether EET may modulate Tau phagocytosis and clearance in primary microglia cultures by first treating the cultures with vehicle (DMSO) or 10 μM EET for 18 hours, followed by adding the fluorescently labelled Tau PFF (2.5 μg/ml) for 6 hours and quantifying the intracellular fluorescence either at time 0 (phagocytosis) or 24 hours later (clearance). The results showed that addition of EET effectively increased both the uptake and clearance of Tau PFF (Figure 8). Therefore, the reduced Tau pathology by TPPU treatment in vivo may be attributed by improved microglia phagocytosis and Tau clearance through increased EET levels.

Discussion

Here we investigated the role of the sEH-EET axis in Tau pathogenesis. We show that sEH blockade by pharmacological inhibition or genetic ablation reversed heightened microglia states and rescued neuronal and synapse loss in PS19 mice. This was accompanied by reduced Tau pathology and improved cognition. Mechanistically, we reveal a parallel effect of the sEH-EET pathway in neurons and microglia where sEH inhibition promotes neuronal and synaptic health in a cell-autonomous manner whereas supplementing EET improves microglia phagocytosis and clearance.

Elevated sEH levels have been observed in multiple neurological disorders including depression, Parkinson’s disease and AD and blocking of sEH activity has been shown to provide therapeutic benefit in related disease models [14, 16, 23, 25, 46–48]. The protective effects have been largely attributed to augmented circulating EETs resulting from sEH inhibition, which act on microglia to suppress neuroinflammation. While this is also the case for tauopathy conditions, we present evidence that sEH exhibits an independent function in neurons. Specifically, our snRNA-seq revealed that multiple genes related to actin cytoskeleton organization, axon guidance and glutamatergic synapses were down regulated in DG neurons of PS19 mice. This is consistent with the fact Tau is normally localized to axons but redistributes to dendrites and postsynaptic compartment under pathological conditions, where it influences actin cytoskeleton, spine density and synaptic function through interacting with NMDAR, PSD95 and other synaptic proteins [49, 50]. Therefore, the rescue of postsynaptic protein expression by TPPU treatment provides the basis for the improved synaptic and cognitive function in PS19 mice. Importantly, using iPSC-derived neurons, we demonstrate that TPPU treatment increased synaptic density in P301S neurons with or without Tau inclusions, demonstrating a neuroprotective and synaptic promoting function of sEH inhibition through a cell-autonomous mechanism. The primary pathway leading to the changes of multiple neuronal pathway genes by TPPU warrants additional investigation.

Our results show that long-term TPPU treatment in mice leads to reduced Tau pathology. However, TPPU does not affect Tau inclusions in cultured iPSC neurons. Besides the intraneuronal pathology, Tau is known to be secreted from neurons and undergo cell-to-cell transfer or prion-like spreading [42–45]. Microglia has been shown to uptake extracellular Tau and mediates its clearance [51]. The augmented microglial uptake of Tau fibrils by EET supports the idea that reduced Tau pathology by TPPU in vivo is due to the elevated microglia uptake and clearance by EET. However, we cannot rule out the possibility that TPPU can directly modulate Tau pathology in neurons and the negative result observed in iPSC neurons is due to the differences between the in vivo and in vitro systems or the treatment regimes.

An anti-inflammatory role of EET downstream of sEH inhibition has been well-established. Our snRNA-seq analysis revealed that this is likely due to the prominent reductions of DAM and IFN microglial subtypes in TPPU-treated PS19 mice. The fact that Mitf and Stat1, which have been reported as key components for the DAM and IFN subtypes [52, 53], respectively, were increased in PS19 and reduced by TPPU supports the notion that downregulation of the MITF and STAT1 pathways drive the anti-inflammatory activities of EET. That the HM is mostly exhausted in PS19 microglia and replenished by TPPU suggests that sEH inhibition facilitates microglia state transition and rejuvenates the chronically activated microglia induced by pathological Tau.

Our work revealed that the sEH-EET pathway exerts parallel effects on neurons and microglia: sEH inhibition promotes synaptic protein expression and synaptic density without affecting Tau pathology in a cell-autonomous manner, while the EET non-cell-autonomously enhances Tau phagocytosis and clearance by microglia. Nevertheless, sEH is known to be highly expressed in the vasculature where it mediates vascular inflammation and barrier function [10, 13, 54]. Therefore, it remains possible that the overall therapeutic effect of TPPU is due to its inhibition in vascular endothelia and possibly other cell types besides neurons and microglia. Interestingly, Wu et al. recently reported a liver-brain interaction in which hepatic sEH regulates cerebral Aβ pathogenesis [25]. Thus, although our earlier work showed that TPPU enters the brain where its levels correlate with EETs [23], we cannot exclude the possibility that TPPU could also exert its effect through inhibition of liver sEH.

Conclusions

Several classes of sEH inhibitors have been developed [55]. Overall, they are well-tolerated in preclinical studies. Among these, TPPU is widely used as a tool compound because of its superior potency, specificity, and pharmacokinetics [11, 56, 57]. A TPPU derivative EC2056 is being tested in humans and the results so far revealed no adverse effects (ClinicalTrials.Org, NCT04908995) [58]. Multiple studies have demonstrated a dysregulation of the sEH metabolic pathway in AD and a beneficial role of sEH inhibition in Aβ mouse models. We report here that it also ameliorates Tau pathology through parallel modulation of neuronal and immune systems. These features, combined with the availability of a positron emission tomography (PET) tracer for sEH imaging in the human brain, allowing for biomarker and target engagement assessment [59], make sEH an attractive therapeutic target for AD, tauopathy diseases, and possibly other neurodegenerative conditions.