Sex as a biological variable. Our study examined male and female animals, and similar findings were reported for both sexes.

Antibodies and reagents. The following antibodies were used in this study: AT8 (Thermo Fisher Scientific, MN1020), AT100 (Thermo Fisher Scientific, MN1060), His (Proteintech, 66005-1-Ig), MAP2 (Proteintech, 17490-1-AP, and Invitrogen, 13-1500), P62 (Abcam, 56416), Ubiquitin (Cell Signaling Technology, 3936S), GFP (Proteintech, 66002-1-Ig), GAPDH (Proteintech, 60004-1-Ig), β-actin (Proteintech, 66009-1-Ig), tau (Thermo Fisher Scientific, MA5-12808), Iba1 (Wako, 019-19741), GFAP (Invitrogen, PA5-16291), Synapsin I (Proteintech, 20258-1-AP), Synaptophysin (Proteintech, 17785-1-AP), PSD95 (Proteintech, 20665-1-AP), Alexa Fluor 488–Goat anti-Mouse (Invitrogen, A11001), Alexa Fluor 488–Goat anti-Rabbit (Invitrogen, A11034), Alexa Fluor 594–Goat anti-Mouse (Invitrogen, A11005), Alexa Fluor 594–Goat anti-Rabbit (Invitrogen, A11012), DAPI (Biofroxx, EZ3412B205), HRP-conjugated anti-mouse IgG (Bio-Rad, 170-6516), and HRP-conjugated anti-rabbit IgG (Bio-Rad, 170-6515). The TUNEL in situ cell death detection kit was purchased from Roche (11684795910). DiI was purchased from Invitrogen (D-282).

Mice. Specific pathogen–free (SPF) tau P301S transgenic mice (line PS19) and WT C57BL/6J mice were obtained from The Jackson Laboratory (stock numbers: 008169 and 000664, respectively). The mice were kept under SPF conditions at a 12-hour light/12-hour dark cycle at 22°C and fed food and water ad libitum. Two-month-old mice were sex-matched and randomly assigned to each experimental group. The sample size was determined by Power and Precision (Biostat).

Purification and aggregation of recombinant human tau RD. Purification of recombinant human K18 was performed as previously described (48). In brief, His-tagged K18 sequence in the pRK172 expression vector was transformed into E. coli BL21 (DE3) cells (ABclonal, CD011). The cells were harvested and resuspended in cold buffer A (20 mM Tris, 500 mM NaCl, and 10 mM imidazole, pH 8.0). The mixture was sonicated by an ultrasonic cell disruptor and centrifuged for 20 minutes at 12,000 rpm at 4°C. Then, the supernatants were purified through affinity chromatography with Ni-NTA agarose beads (Cytiva, 17371201). The beads were washed with buffer B (20 mM Tris-HCl, 20 mM imidazole, 0.15 M NaCl, pH 8.0) and eluted by buffer C (20 mM Tris-HCl, 250 mM imidazole, 0.15 M NaCl, pH 8.0). The purity of the protein was confirmed through Coomassie Brilliant Blue staining (Biofroxx, 1912GR025). The endotoxin of the purification of recombinant human K18 was removed using the EtEraser HP kit (Bioendo, ER0015). The purified K18 monomers were lyophilized and resuspended in PBS with 12.5 μM low–molecular weight heparin and 2 mM DTT at a final concentration of 1 mg/mL and then fibrilized by agitation at 1,000 rpm at 37°C for 12 hours. To study the effect of PM2.5 on tau aggregation, PM2.5-tau PFFs were prepared by adding 0.01, 0.05, or 0.1 mg/mL PM2.5 at a final concentration into the tau fibrilization reaction. ThS assay was conducted to evaluate tau aggregation. In brief, PFFs were incubated with 20 μM ThS and read at excitation and emission wavelengths of 440 nm and 480 nm using a SpectraMax microplate reader (Molecular Devices). The monomers and PFFs were stored at –80°C until use.

TEM. The morphologies of PFFs were assessed by TEM. The tau PFFs or PM2.5-tau PFFs (1 mg/mL) were dropped onto the 200 mesh carbon-supported copper grids for 3–5 minutes. We removed excess liquid with filter paper, then negatively stained with 2% phosphotungstic acid for 2 minutes and dried at room temperature. The grid was viewed at 80 kV under a transmission electron microscope (HT7800/HT7700, Hitachi).

The synaptic density was also detected by TEM. Mice were perfused with 2% glutaraldehyde. Then, the left hippocampal tissues of mice were postfixed with 1% OsO4 for 2 hours at room temperature. After that, the tissues were dehydrated and embedded using standard procedures. Ultrathin sections (90 nm) were stained with 2% uranyl acetate and 2.6% lead citrate, then viewed at 100 kV. Synapses from the DG region were identified by the presence of synaptic vesicles and postsynaptic densities.

PK digestion. The tau PFFs and PM2.5-tau PFFs were incubated with PK at a final concentration of 4 μg/mL for various times at 37°C. The interaction was terminated by the addition of protease inhibitor cocktail (Roche, 4693132001). Then the samples were boiled for 10 minutes in 1× SDS loading buffer. The samples were separated on a 10% SDS-PAGE and stained with Coomassie Brilliant Blue solution. The images were analyzed using ImageJ software (NIH). The quantitative measurements were performed of the levels of all remaining protein fragments with PK digestion relative to total proteins without PK digestion by analyzing the gray value of the whole lane.

Collection of PM2.5. Referring to previous studies (49, 50), the sampling site was located near a main road in the central urban area of Wuhan, which is a typical urban outdoor point with intensive human activities and traffic flow. A flow air particle sampler (Wuhan Tianhong TH-150C) was set on the rooftop of a building that was 20 m above the ground. The sampler removed PM10 based on the weight of the particles in the cutter and gathered the PM2.5 on a 90 mm diameter glass fiber filter by working at a constant flow rate of 400 L/min. After filtering with 12 gauze layers and lyophilizing, the PM2.5 was resuspended in PBS at a final concentration of 2 mg/mL and then stored at –80°C until use.

Transfection of cells with PFFs. Tau-HEK293 cells that stably express GFP-tagged tau RD in HEK293 cells (ATCC, CRL-1573) were cultured in DMEM. The cells were seeded into 6-well plates with coverslips (Thermo Fisher Scientific, 12-545-80) precoated with polyethyleneimine and cultured to a density of approximately 60%. Then, different PFFs were transduced into the cells. For transduction, PBS or different PFFs (4 μg), 96 μL opti-MEM (Gibco, 31985070), and 4 μL Lipofectamine 2000 (Invitrogen, 11668019) were mixed in a tube. After incubation for 20 minutes, the mixtures were directly added to the culture medium.

Immunofluorescence staining. For immunostaining, cells were fixed with 4% paraformaldehyde (PFA) solution with 1% Triton X-100 for 10 minutes at room temperature to get rid of soluble tau. The cell slides were then blocked with 5% BSA for 30 minutes. After incubation with primary antibodies overnight at 4°C, the coverslips were incubated in the dark with fluorescence secondary antibody solution at room temperature for 2 hours followed by incubation with DAPI (0.1 μg/mL) solution for 5 minutes. Between each step, the coverslips were washed 3 times with PBS solution. Finally, the coverslips were sealed onto slides using antifade mounting medium (Invitrogen, P36941).

Primary neuronal culture and treatment. The isolated primary neurons are cortical neurons prepared from embryos of pregnant tau P301S mice (Jackson Laboratory, stock number 008169) as previously described (51). To evaluate the effects of PM2.5 on tau aggregation, PM2.5 was added directly into the medium of neurons cultured in vitro for 1 week. Tau PFFs were added to the medium 24 hours later. The neurons were harvested for Western blotting or fixed with 4% PFA for immunofluorescence staining. To investigate the seeding activity of pure tau PFFs and PM2.5-tau PFFs in neurons, PBS or different PFFs were added directly into the medium. The phosphorylation and aggregation of tau were detected by Western blot and immunofluorescence staining.

Immunohistochemistry staining of brain sections. The mouse brain tissues were fixed with 4% PFA and embedded in paraffin, cut at a thickness of 4 μm, and attached to glass slides. Then the tissue sections were soaked in xylene 3 times for 15 minutes each and rehydrated with decreasing concentrations of ethanol in water. After soaking in tap water for 5 minutes, the tissue sections were placed in citrate buffer (0.1 M sodium citrate, pH 6.0) at 94°C for 20 minutes. The sections were washed 3 times with PBS for 5 minutes each time after the citrate buffer was cooled to room temperature.

For immunohistochemistry staining, the tissue sections were incubated with quenching buffer (3% hydrogen peroxide solution) for 10 minutes followed by blocking with 5% BSA for 30 minutes. After incubating in the primary antibody solution overnight at 4°C, the tissue sections were rinsed in PBS 3 times and incubated with HRP-conjugated secondary antibody solution at 37°C for 1 hour. Next, the tissue sections were incubated with DAB solution, then stained in hematoxylin solution. After dehydrating in 75%, 95%, and 100% ethanol for 5 minutes each and immersing in xylene for 10 minutes, the slides were mounted with a cover glass by neutral balsam.

Sequential protein extraction. The cells were lysed on ice in TX-100 buffer before centrifugation at 100,000g for 1 hour at 4°C. Then the supernatants containing the TX-soluble fraction were collected and quantified using BCA assay (Invitrogen). The pellets were washed and sonicated twice with TX-100 buffer followed by centrifugation at 100,000g for 30 minutes at 4°C each time. The final pellets were resuspended and sonicated with SDS buffer (2% SDS in TBS with protease and phosphatase inhibitors). All samples were normalized by the BCA protein quantitation kit, analyzed by SDS-PAGE, and detected by Western blot analysis.

Sedimentation analysis. The PBS solution containing 0.05% Triton X-100 with protease and phosphatase inhibitors was used to lyse the cells. The cell lysate was sequentially centrifuged at 500g and 1,000g for 5 minutes. Ten percent of the supernatant was set as the total fraction. Then 90% of the supernatant was centrifuged at 100,000g at 4°C for 1 hour. The supernatant was collected as the supernatant fraction, and the pellet was washed and sonicated twice with PBS followed by centrifugation at 100,000g at 4°C for 30 minutes. Then the pellet was resuspended and sonicated in RIPA buffer containing 2% SDS with protease and phosphatase inhibitors, which was the pellet fraction. All the samples were normalized by BCA protein quantitation and analyzed by Western blot.

Intranasal instillation of PM2.5. According to the bulletin from the Wuhan Ecological Environment Bureau, the maximum concentration of PM2.5 in Wuhan in 2021 was 215 μg/m3. It has been reported that the conversion factor is 12.3 for converting mouse dose (mg/kg) to human equivalent dose (mg/kg), based on body surface area (52). An individual with 60 kg will breathe in approximately 20 m3 of air containing 4,300 μg of PM2.5, resulting in a dose of 71.667 μg/kg/d in humans and an equivalent dose of 881.504 μg/kg/d in mice based on body surface area. Thus, 36 μg PM2.5 in 18 μL PBS was administered to mice by intranasal instillation every 2 days for 4 months, 9 μL in each nostril. To avoid the solution being ingested, mice were placed in the supine position, their heads were manually held at an angle of 70°–90°, and the solution was dripped several times so that PM2.5 would remain in the nasal cavity rather than descending into the pharynx.

Stereotaxic injection. Stereotaxic injection of the mouse hippocampus was performed as previously described (53). There are no experimental data to support any lateralization effects in this model. Five micrograms of PFFs (5 μL, 1 mg/mL) were injected into the left hippocampus (anteroposterior –2.5 mm, mediolateral +2.0 mm, dorsoventral –1.7 mm) of 2-month-old tau P301S mice at a rate of 0.1 μL/min using a 10 μL Hamilton syringe. The control mice were injected with 5 μL of sterile PBS. After injection, the syringe was maintained for 5 minutes for complete absorption of the PBS or PFFs. Postoperative care was provided until the mice recovered from anesthesia.

Morris water maze. Three months after stereotaxic injection, 5-month-old mice were tested by Morris water maze, which is a reliable tool for assessing spatial learning and memory but not sensitive to spatial working memory assessment (54). During the training phase, mice were trained with extra maze cues. Each subject was tested 4 times per day to reach the platform within 60 seconds. If the mouse failed to find the platform within 60 seconds, it was manually guided to the platform and allowed to remain there for 15 seconds. On the first day, visual training (a visible flag was placed on the platform only on the first day) was performed. Then, each mouse was tested for 5 consecutive days without a visible flag. The probe trial was performed after the training test, in which the platform was removed and the escape latency and the time spent in the platform quadrant were recorded and analyzed by ANY-Maze software.

Y-maze test. The Y-maze test can better assess changes in spatial working memory, which is mainly regulated by the hippocampus and the prefrontal cortex (55). Y-maze is easily disturbed by the environment, but it makes up for the deficiency of the Morris water maze test. Each arm was 40 cm long, 12 cm high, 3 cm wide at the bottom, and 10 cm wide at the top. The arms converged in an equilateral triangular central area that was 4 cm at its longest axis. The 3 arms were randomly designated. Start arm: The mouse started to explore when the arms were always open. Novel arm: The arm was blocked during the first trial but opened during the second trial. The test consisted of 2 trials separated by an intertrial interval (ITI) to assess spatial recognition memory. The first trial (training) lasted for 5 minutes, and the mice were allowed to explore only 2 arms (start arm and another arm) in the maze, with the third arm (novel arm) being blocked. The series of arm entries was recorded visually. After a 1-hour ITI, the second trial (retention) was conducted, during which all 3 arms were accessible, and novelty versus familiarity was analyzed. In the second trial, the mice were placed back in the maze in the same starting arm, with free access to all 3 arms for 5 minutes. Recordings were taken and analyzed using the ANY-Maze software. Then the number of entries and time spent in each arm were analyzed.

Electrophysiology. Electrophysiology of hippocampal neurons was performed using a whole-cell patch-clamp as described previously (56). Briefly, the murine brains were placed in precooled, oxygenated cutting solution (2.5 mM KCl, 25 mM d-glucose, 25 mM NaHCO3, 1.26 mM NaH2PO4, 0.5 mM CaCl2, 7.2 mM MgCl2, 3.1 mM Na-pyruvate, 97 mM choline chloride, and 11.35 mM ascorbic acid, pH = 7.3–7.4) and cut along the sagittal direction at 350 μm thickness by using a Leica VT1200S vibratome. Then, the brain slices containing the dorsal hippocampus were incubated with oxygenated artificial cerebrospinal fluid (118 mM NaCl, 2.5 mM KCl, 26 mM NaHCO3, 1 mM NaH2PO4, 2 mM CaCl2, 2 mM MgCl2 and 22 mM glucose, pH = 7.3–7.4) for 1 hour at room temperature. The recording microelectrodes were made with borosilicate glass (Warner Instruments, Inc.) by using a horizontal drawing puller (Sutter Instrument) to form a high-resistance seal (GΩ) between the recording microelectrodes and the cell membrane, containing an intracellular solution with a resistance of 6 MΩ (20 mM KCl, 100 mM CsCH3SO3, 4 mM Mg-ATP, 0.3 mM Tris-GTP, 10 mM HEPES, 7 mM Tris2-Phosphocreatine, 3 mM QX-314, pH = 7.3–7.4, osmolarity = 298 mOsm). The fEPSPs in CA1 neurons were recorded by stimulating the Schaeffer fibers from CA3. Three trains of high-frequency stimulation (100 Hz, 1-second duration) were applied to induce LTP. Electrical signals were recorded by an Axon Instruments clamp 700B amplifier at 25 kHz and digitalized with a Digidata 1,440 digitizer (Molecular Devices) controlled by Clampex 10.1 (Molecular Devices).

Golgi staining. Golgi staining of the hippocampus was performed using a Golgi staining kit (FD Neuro Technologies, Inc., PK-401). In brief, the brain tissue was soaked in the silver impregnation solution for 2 weeks in the dark and then incubated with stock solution for 3–7 days. The samples were cut at 60 μm and attached on gelatin-coated glass slides followed by incubation with staining solution for 10 minutes. After rinsing in double-distilled water, the sections were dehydrated through 95% and 100% ethanol, cleared in xylene, and mounted with mounting medium. Images of dendritic spines were taken at 100× original magnification using an Axioplan (ZEISS) microscope. The spine density was counted in all clearly evaluable areas of 50~100 μm of secondary dendrites from each imaged DG neuron.

Statistics. Data were expressed as means ± SEM and analyzed with GraphPad Prism (version 8.0). Two-group comparisons were performed with 2-tailed Student’s t tests. Comparisons among 3 groups or more than 3 groups were performed with 1-way ANOVA followed by Tukey’s multiple comparisons for post hoc tests. For the time course studies, the level of significance was determined using the 2-way ANOVA and Holm-Šídák multiple comparisons test. Differences with P values less than 0.05 were considered significant.

Study approval. All experiments involving animal use were approved by the Laboratory Animal Welfare Ethical Committee of Renmin Hospital of Wuhan University (No. 20210402).

Data availability. Supporting data values for each figure are made available in the Supporting Data Values file.