Suckling mice aged 24–48 h were used to extract primary cells. Cerebral cortex tissues were stripped and minced, followed by digestion with trypsin. The digestion was terminated by the addition of 20% FBS in the DMEM medium. Cells were cultured in a flask kept inside an incubator maintained at 37°C and 5% CO₂. After 48 h, the primary astrocytes were digested with 0.125% Trypsin EGTA. The digestion was terminated by the addition of 20% FBS-DMEM medium. The cells were centrifuged and resuspended, followed by their inoculation in a new culture flask to allow for the growth of adherent cells for further experiments. Astrocyte U87 cells were cultured at 37°C, in 5% CO₂ with Dulbecco’s Modified Eagle’s Medium (DMEM; Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS; Gibco), 100 μg/mL penicillin, and 100 μg/mL streptomycin (Gibco).

The ischemia reperfusion model of astrocyte was cultured in vitro by oxygen glucose deprivation/reperfusion method. The cells were cultured in sugar-free DMEM medium in an incubator maintained at 37°C in an atmosphere of 95% N₂ and 5% CO₂ for 4 h. This was followed by further culture of the cells in normal DMEM medium for the subsequent experiments.

The FRET-based tension probes of glial fibrillary acidic protein (GFAP)-cpst-GFAP (GcpG) were designed and constructed as described in previous reports (Zhang et al., 2019; Zheng et al., 2022a). The integrity of all expression constructs was confirmed by DNA sequencing. Plasmids encoding the FRET sensor were transfected into cells with FuGENE® 6 transfection reagent (Roche Diagnostics, Basel, Switzerland) and Opti-MEM™ media (Invitrogen, Carlsbad, CA, United States), according to manufacturers’ instructions. The samples were analyzed using the following excitation lasers and emission filters: Cyan-458-nm laser and 447/60 bandpass filter; Yellow-514-nm laser and 580/23 bandpass filter. Cells were sorted with the simultaneous detection of cyan and yellow emission wavelengths.

The efficiency of FRET is determined by the distance and the dipole angular orientation between the donor/CFP and the acceptor/YFP. The cells were imaged using a confocal microscope (SP5; Leica, Wetzlar, Germany) equipped with a ×63 oil-immersion objective lens. The donor and acceptor were observed under 458 nm and 514 nm argon lasers, respectively. CFP/FRET ratios were calculated using the equation 1/E = Cerulean (donor)/Venus (acceptor) (Hu et al., 2022).

Astrocyte U87 cells were cultivated in 90-mm Petri dishes. When the cell proliferation reached about 95%, the medium was discarded, the cells were washed twice with isotonic Hepes buffer, and appropriate experimental stimulation was added. After 24 h, the drug-containing culture medium was aspirated, digested by typsin and washed, and the cell suspension was collected into 1.5 mL microcentrifuge tube. Centrifuge at 1,000 g at 4°C for 5 min and discard supernatant. Following ultrasonification (75% amplitude, five times, 5 s) (Sonics and Materials, Connecticut, CT, United States) and ultracentrifugation (20,000 g, 1 h, room temperature), 50 µL of the supernatant solution (cytoplasm) was transferred into 0.5-mL test tubes. The cytoplasmic OP was measured using the Osmomat 3000 Freezing Point Osmometer (Gonotec, Berlin, Germany) The diluted supernatant was used to assess the count rate (kilocycles per second, kcps) of cytoplasmic nanoparticles (NanoSight NS300; Malvern Instruments, Malvern, United Kingdom).

Calcium, chloridion and sodium fluorescent imaging was performed using the dyes Fluo-4 a.m. (excitation at 494 nm and emission at 516 nm, purchased from Abcam, Cambridge, England), N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide (MQAE, excitation at 355 nm and emission at 460 nm, purchased from Beyotime, Shanghai, China), and Enhanced NaTrium Green (ENG)-2 a.m. (excitation at 488–525 nm and emission at 545 nm, purchased from New Research Biosciences, Xian, China), respectively. After the specific treatment, the cell cultures were examined by confocal microscopy to confirm comparable fluorescence intensities.

Cell cytoplasm supernatants were obtained and processed according to the instructions of the human diglyceride (DAG) and cAMP ELISA kit provided by Jiangsu Enzyme Immunoassay Industry Co., Ltd. and Nanjing Yifeixue Biotechnology Co., Ltd, respectively.

Cells were inoculated into a glass plate with a diameter of 35 mm. The cells adhered to the wall and were allowed to grow until they reached 60%–70% confluence. They were subsequently treated with drugs and maintained in an incubator at 37°C in an atmosphere of 95% O₂ and 5% CO₂. After 12 h, the cells were analyzed using ×3 objective 60D tomographic scanning microscope (Nanolive SA, Tolochenaz, Switzerland) and STEVE software (Nanolive SA, Tolochenaz, Switzerland) to determine the changes in their morphology.

Whole-cell patch recordings were conducted as previously described (Pankonien et al., 2022). Briefly, the cells were transferred to a submersion-type recording chamber and perfused (1∼2 mL/min) with an extracellular solution composed of 140 mM NaCl, 4 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose, and 10 mM HEPES, with the pH adjusted to 7.4 using NaOH. The intracellular solution for membrane potential recordings was composed of 130 mM K-gluconate, 10 mM NaCl, 2 mM MgCl₂, 2 mM CaCl₂, 10 mM EGTA, 2 mM Na₂ATP, pH adjusted to 7.2 with KOH. The average resistance of the recording pipettes was 2∼6 MΩ, when filled with internal solution. Recordings were performed using a MultiClamp 700B amplifier and a Digidata 1550B interface (Molecular Devices, CA, United States) at room temperature (24∼25°C).

Whole-cell voltage clamp recordings were performed on astrocyte cells. Cells were spread at an appropriate density on polylysine-coated cell slides, placed in a bath on the stage of an inverted microscope, and perfused with the desired solution. The electro physiological results were performed using a Multiclamp 700B amplifier at a data-sampling rate of 2 kHz with a Digidata 1550B Digitizer controlled using PClamp 10.6 software (Molecular Devices, Sunnyvale, CA, United States). A glass electrode (outer diameter 1.5 mm; internal diameter 0.86 mm; World Precision Instruments, Sarasota, FL, United States) using a drawing instrument (P-1000; Narishige Scientific Instrument Lab., Tokyo, Japan) were drawn with resistance values of 2–5 MΩ after filling the inner fluid, and all experiments were performed at room temperature (23°C–25°C). Adjustments for capacitance compensation and series resistance compensation were made before recording. Series resistance was routinely compensated by 60%–80%. If the recording channel was unstable during the experiment (the series resistance increased by more than 15% at the end of the experiment), the cell was discarded from the data analysis. The Current trace diagram was performed in Graphpad Prism 7.

Rats were anesthetized using an intraperitoneal injection of 2% sodium pentobarbital (30 mg/kg). Body temperature was monitored and maintained at 36.5°C–37.5°C. Briefly, after making an incision in the midline skin, the right common carotid artery, external carotid artery (ECA), and internal carotid artery (ICA) were exposed. Subsequently, a 6–0 nylon monofilament with a rounded tip was inserted into the right ICA through the broken end of the ECA, in order to block the origin of the middle cerebral artery (MCA). Cerebral ischemia through the intraluminal suture was maintained for 60 min. It was followed by the removal of the monofilament and reperfusion. After 24 h of middle cerebral artery occlusion (MCAO), the brains were obtained by decapitation. This experiment was approved by the ethics committee of the Affiliated Huaian First People’s Hospital of Nanjing Medical University.

The infarct volume was evaluated at 24 h after MCAO using 2, 3, 5-triphenyltetrazolium chloride (TTC) staining. The volumes of contralateral and non-infarcted tissue of the ipsilateral hemispheres were outlined and measured with ImageJ software (NIH). The formula used for calculating infract volume is given below: Infract volume = (volume of contralateral—volume of non-ischemic ipsilateral)/2*(volume of contralateral) *100% (Zhou et al., 2021). The brains were harvested from different groups of rats and weighed (wet weight). They were dried for 2 days in a 70°C incubator, which was followed by the determination of their dry weight. The water content was calculated using the formula: Water content = (wet weight – dry weight)/wet weight * 100%.

The blood obtained from different groups was placed at room temperature for 1 h, followed by centrifugation at 3,000 rpm/min for 10 min in order to obtain serum. The levels of glutamic-pyruvic transaminase (ALT), glutamic oxalacetic transaminase (AST), alkaline phosphatase (ALP), blood urea nitrogen (BUN) and creatinine (CRE) in serum were measured using specific kits (Nanjing Jiancheng Bioengineering Institute, China) in accordance with the manufacturer’s instructions.

Data analysis was conducted using the statistical program Graph Pad Prism 7. More than three biological replicates were presented as mean ± SEM. Unpaired t-test (two-tailed p-value) was used to determine the statistical significance (p < 0.05 was considered to be statistically significant).

Cerebral ischemic injury induces extracellular glutamate accumulation, further facilitating intercellular accumulation of cations and neurocyte swelling (Mahmoud et al., 2019; van Putten et al., 2021). We exposed primary astrocytes and the astrocyte cell line to medium containing glutamate (2 mM). Our previous research demonstrated that glutamate stimulated MF and MT depolymerization in astrocytes, resulting in the production of actin and tubulin monomers. These monomers function as intracellular PNs and promote hyper-osmosis and astrocyte swelling (Zheng et al., 2021). We hypothesize that protein nanoparticle-induced osmotic pressure (PN-OP) plays a significant role in astrocyte edema. In this study, we found that glutamate induced an increase in the levels of NLRP3 and ASC. MCC950 (NLRP3 specific inhibitor) or Z-VAD-FMK (caspase-1 pan inhibitor) effectively rescued the glutamate-induced ASC and NLRP3 in the primary astrocytes (Figure 1A). Interestingly, MF stabilizer jasplakinolide (JK) and MT stabilizer taxol (TAX) co-treatments also reduced ASC and NLRP3, suggesting that MF and MT stabilization relieved the NLRP3 inflammatory effects in response to glutamate stimuli. In addition, MCC950, Z-VAD-FMK, JK-TAX treatments significantly decreased glutamate-induced astrocyte swelling, as shown by 3D cell imaging (Figure 1B). Thus, we postulate that NLRP3 inflammasome is another intracellular PNs involved in astrocytic OP regulation. To investigate this theory, a glial fibrillary acidic protein (GFAP) FRET-based tension probe was applied in primary astrocytes to monitor IF tension in real-time, as it has been shown to be an effective probe in evaluating the transmembrane OP effects in living cells (Qian et al., 2022). MCC950, Z-VAD-FMK, and JK-TAX treatments significantly reduced glutamate-induced IF tension (Figure 1C). In parallel, these inhibitors decreased the cytoplasmic OP and the amount of intracellular PNs, as detected by the freezing point osmometer and nanoparticle tracking analyzer, respectively (Figures 1D,E). Meanwhile, the fluorescent imaging results showed that NLRP3 and MF/MT inhibition decreased the sodium and chloride ion levels in primary astrocytes (Figures 1F,G), which are closely associated with intracellular osmosis. Similar results were also observed in the astrocyte cell line U87 (Supplementary Figure S1). Taken together, the cell volume, IF tension, cytoplasmic OP and the intracellular ion levels showed variations which were consistent with regard to the inhibition of PNs. These findings suggest that NLRP3 inflammasome production and MF/MT depolymerization are the two main sources of intracellular PNs, which upregulate cell osmosis and promote astrocyte swelling.

According to the Donnan effect, intracellular PNs can absorb cations and promote the re-arrangement of free ions near the plasma membrane, leading to changes in membrane potential (Nguyen and Kurtz, 2006; Lang et al., 2014; Yu et al., 2021). Membrane potential was measured by the current clamp of primary astrocytes and showed depolarization corresponding to the glutamate stimuli. However, a decrease in the intracellular PNs through the MF/MT stabilization and NLRP3 inhibition resulted in a significant recovery of the depolarized astrocytes (Figure 2A).

Astrocytes respond to external stimuli by activating cell surface receptors that increase the intracellular levels of secondary messengers, including diacylglycerol (DAG), cyclic adenosine phosphate (cAMP) and Ca²⁺. Increased Ca²⁺ concentration in cells can activate adenylate cyclase 3, leading to the enhanced synthesis of intracellular cAMP. We performed the DAG and cAMP ELISA assays, and the Flou-4 a.m. Ca²⁺ fluorescent imaging in primary astrocytes. Glutamate stimulated an increase in the levels of the second messengers, including diacylglycerol (DAG), cAMP and Ca²⁺. However, when PN production was blocked by the NLRP3 inhibitor MCC950, caspase-1 inhibitor Z-VAD-FMK, or MF/MT stabilizers JK/TAX, it also caused a decrease in the levels of second messengers (Figures 2B–D). These results indicate that intracellular PNs can trigger changes in the membrane potential and induce second messengers during astrocyte swelling.

Corresponding to changes in the membrane potential and the levels of various second messengers, we predict the activation of voltage-dependent ion channels. Ca²⁺-activated TRPM4 channel has been reported to predominantly control the massive Na⁺ influx involved in edema progression (Stokum et al., 2018). Furthermore, Ca²⁺-activated TMEM16A (also known as anoctamin-1) function as Cl⁻ channels, and open in response to cerebral infarction (Liu et al., 2019). To investigate the involvement of these ion channels in astrocyte osmosis, primary astrocytes induced by glutamate and the U87 cell line were individually exposed to glyburide (Glyb, TRPM4 inhibitor) and niclosamide (Niclo, TMEM16A inhibitor). Results showed that the two inhibitors effectively attenuated the osmotic effects, including IF tension (Figure 3A), cytoplasmic OP (Figure 3B), and the Na⁺ and Cl⁻ levels in astrocytes (Figures 3C,D, Supplementary Figure S2).

Another second messenger of cAMP promotes voltage-dependent activation of hyperpolarization and cyclic nucleotide gated (HCN) cation channel involved in the Na⁺/K⁺ electrochemical gradient flow (Benarroch, 2013). Cystic fibrosis transmembrane conductance regulator (CFTR) is a chloride ion channel, and its activity is enhanced by an increase in the cAMP levels (Kunzelmann and Mehta, 2013b; Weidenfeld and Kuebler, 2017). We found that artemisinin (Arte, HCN inhibitor) and CFTRinh-172 (CFTR inhibitor) caused a significant suppression of glutamate-induced IF tension, cytoplasmic OP, and the Cl⁻ and Na⁺ content in both primary astrocytes and the U87 cells (Figure 3 and Supplementary Figure S2). Further, Arte could effectively reduce the HCN current (Supplementary Figure S3). Taken together, the glutamate stimuli simultaneously induced various ion channels, which were dependent on both membrane potential and second messengers. These ion channels could federatively promote the ion influx-mediated astrocyte swelling.

As intracellular PNs and various ion channels regulate astrocyte swelling, we next screened for potential compounds against edema. Shingshot (SSH) inhibitor Sennoside A (SenA) can inhibit cofilin dephosphorylation to stabilize MF (Lee et al., 2017). Taxol is widely used as a MT stabilizer (Yang and Horwitz, 2017). Nimodipine is a voltage-gated L-type calcium channel inhibitor that blocks Ca²⁺ influx to maintain intracellular calcium homeostasis (Carlson et al., 2020). Glyburide (Glyb) is SUR1-TRPM4 channel inhibitor (King et al., 2018; Nakayama et al., 2018), while Niclosamide (Niclo) and Benzbromarone (Bbr) function as TMEM16A inhibitors (Miner et al., 2019; Dwivedi et al., 2023). Artemisinin (Arte) and Ivabradine are specific blockers of HCN channel (Ide et al., 2019), and CFTRinh-172 is a specific inhibitor of CFTR channel (Melis et al., 2014). The glutamate-induced astrocyte U87 cells and oxygen-glucose deprivation (OGD)-induced primary astrocytes were individually treated with the above-mentioned inhibitors. All of them attenuated the glutamate-induced effects on IF tension and cytoplasmic OP; with SenA, Glyb, Niclo, Arte and CFTRinh-172 showing high efficiencies (Figure 4).

To investigate whether these MF stabilizer and ion channel inhibitors display a combined effect against glutamate-induced primary astrocyte edema, different compound designs were tested for their osmotic effects. Glyb is widely used as a drug against edema; however, Glyb treatment alone failed to ensure complete recovery of the glutamate-induced IF tension (Figure 5A). Combinations of two compounds (SenA and Glyb) and three compounds (SenA, Glyb and Niclo) showed mild effects. Combinations of four and five compounds, containing Glyb/SenA/Niclo/Arte and Glyb/SenA/Niclo/Arte/CFTRinh-172, showed the highest efficacy on the IF tension (Figure 5A). However, including CFTRinh-172 in the combination dramatically reduced cell viability (Figure 5C), and had little impact on cytoplasmic OP (Figure 5B), thereby suggesting that SenA/Glyb/Niclo/Arte is the optimal combination for the treatment of cytotoxic edema.

Cerebral ischemia induces extracellular glutamate accumulation, ion disequilibrium and the resultant edema. The MCAO rat model was used to further confirm the efficacy of the above combination compounds in vivo. The MCAO group showed a significant increase in the volume of cerebral infarction and water content in brain tissues, while the combination compounds treatment attenuated injuries (Figures 6A–C). In addition, the serum biochemical analysis, which included ALT, AST, ALP, BUN and CRE, and hematoxylin and eosin (HE) staining of tissue section (heart, liver, spleen, lung and kidney), showed no difference among the control, MCAO and drug combination-treated MCAO groups (Figures 6D,E). This confirmed the safety of the drug combination as well as its effective inhibition of cerebral edema.