Introduction

Stress granules (SGs) are membrane-free dynamic structures formed by the aggregation of mRNA and proteins, induced by stimuli such as oxidative stress, endoplasmic reticulum stress and hypoxia in eukaryotic cells1,2,3. SG formation and depolymerization are highly dynamic processes regulating the dynamic RNA transcription–protein translation balance and improving cell survival under adverse conditions4,5,6,7. Continuous formation of primary SGs (size, 600 nm) promotes the conversion of prion-like domains in SGs to stable amyloid fibers8,9. This irreversible change leads to the SG structure changing from a highly dynamic one into a stable solid one, altering RNA processing and protein synthesis7,10,11. However, in neural cells, 1-h H2O2 stimulation can transform primary SGs into mature SGs (size, 5 μm), accompanied by cell function stabilization. Thus, accelerating SG maturation may be critical for cellular structural and functional stability maintenance.

TRIM21 and EWSR1 are major regulators in SG formation and maturation12,13,14,15. Under arsenite-induced oxidative stress, SGs are highly enriched in TRIM2116. TRIM21 knockdown leads to prevention of K63-linked ubiquitination of G3BP1 (a core protein of SGs), along with SG formation; in contrast, TRIM21 overexpression results in G3BP1 ubiquitination and SG formation inhibition17. In cells, oxidative stress stimulates EWSR1 translocalization from the nucleus to SGs and facilitates alternative splicing and expression of relevant genes. EWSR1 mutation can reduce the mobility of SGs and affect their dynamics15. Therefore, TRIM21 and EWSR1 can regulate SG stability.

SGs are associated with inflammasome activation and cytoskeletal protein depolymerization18,19. G3BP1 mediates protein–RNA interactions in SGs through the NTF2L and RNA-binding domains20,21. In hypoxia-induced neural cells, G3BP1 can competitively inhibit the binding of NLRP3 inflammasomes to RNA helicase, inhibiting inflammasome aggregation and thereby regulating damaged neural cells22. Primary SGs transported via cytoskeletal proteins (microfilaments or microtubules) fuse and form into mature SGs23,24. However, when colchicine disrupts microtubules, SGs become smaller (size, < 600 nm) but their number increases, which leads to a blockade of SG aggregation; this inhibits SG maturation and function and accelerates cell damage24,25. Thus, SGs can inhibit inflammasome aggregation, and they can be regulated through cytoskeletal depolymerization.

Protein nanoparticles (PNs) originate from microfilament and microtubule depolymerization and inflammasomes26. According to electric double layer theory, protein nanoparticles can induce membrane potential changes through free cation adsorption and further activate Ca2+-related ion channels, then regulate changes in the cell osmotic potential—an important regulatory mechanism underlying biological osmotic pressure (OP)27. Moreover, cell tension, based on the cytoskeleton (i.e., microfilaments, microtubules, and intermediate fibers), osmolarity, and cytoskeletal tension, regulates cell volume changes28. Therefore, we hypothesized that SGs inhibit abnormal protein nanoparticles, including microfilaments, microtubules, and inflammasomes, thereby regulating cell tension, membrane potential, and osmolarity changes. Herein, we used a fluorescence resonance energy transfer (FRET)-based intermediate filament (IF) tension probe, the tension of which is positively correlated with transmembrane osmotic pressure in live cells29. Combined with electrohysiological techniques, we observed electrochemical and osmotic tension occurrence and recovery in inflammation-induced U87 cells. Thus, we elucidated that under the electrochemical–tension interactions, SGs formation regulate osmotic pressure and the membrane potential by regulating intracellular protein nanoparticles and ions, revealing newer functions of SGs from the biophysical perspective.

Materials and methods

Reagents

Lipopolysaccharide (LPS) and nigericin were purchased from MedChemExpress (Monmouth Junction, NJ, USA), H2O2 from Beyotime (Shanghai, China), and the Caspase-1 Activity Assay Kit (#C1101) from Beyotime Biotechnology (Shanghai, China). Furthermore, mouse anti–apoptosis-associated speck-like (ASC) antibodies (#SC-514414) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), rabbit anti-G3BP1 antibodies (#ZF-0311) from Origene (Shanghai, China), and rabbit anti-β-actin antibody (AP0060) from Bioworld Technology (Nanjing, China). Moreover, the vimentin (#31922), EWSR1-G511A (#50734), and EWSR1-WT (#50733) plasmids were obtained from Addgene (Shanghai, China) and the GFP-G3BP1 (#HG16688-ACG) plasmid from SinoBiological (Beijing, China). Finally, TRIM21-targeting small interfering RNA (siRNA; F: 5′-GCAGGAGUUGGCUGAGAAGTT-3′; R: 5′-CUUCUCAGCCAACUCCUGCTT-3′) and negative control (NC) siRNA were designed by and purchased from GenePharma (Shanghai, China).

1 L standard artificial cerebrospinal fluid (ACSF) Recipe: NaCl (Sodium Chloride): 126 mM (7.35 g); KCl (Potassium Chloride):3.0 mM (0.224 g); NaHCO3 (Sodium Bicarbonate): 26 mM (2.18 g); NaH2PO4 (Sodium Phosphate Monobasic): 1.25 mM (0.15 g); CaCl2 (Calcium Chloride): 2.0 mM (0.22 g); MgCl2 (Magnesium Chloride): 1.0 mM (0.095 g); Glucose: 10 mM (1.80 g). All components were dissolved in approximately 900 ml of sterile distilled water, adjusted to pH 7.4 using HCl/NaOH, supplemented with distilled water to a constant volume of 1 L, sterilized through a 0.22 μm filter membrane, and stored at 4 °C.

Cell culture

The human glioblastoma U87 cells were procured from the American Animal Model Library and Culture Collection. U87 cells were cultured in Dulbecco’s Modified Eagle’s Medium (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (Gibco), 100 µg/mL penicillin, and 100 µg/mL streptomycin (Gibco) at 37 °C in a humidified atmosphere with 5% CO2.

Probe construction and transfection

A vimentin fluorescent tension probe was constructed using a cloning kit and restriction enzyme cloning technology, according to the FRET principles. We constructed fluorescent sensors with circularly permutated cpVenus and cpCerulean (cpVenus-7aa-cpCerulean; denoted as cpstFRET). In cpstFRET, cpCerulean (cyan) and cpVenus (yellow) were the donor and acceptor, respectively; they remained parallel under normal conditions. When the tension probe received an external force, the angle of cpstFRET changed, reducing FRET efficiency. The cpstFRET probe (1 µg) was transfected into cells by using 3 µL of Lipofectamine 2000 (Thermo Fisher, USA).

cpstFRET analysis

FRET efficiency was determined on the basis of the dipole angle between the donor/enhanced cyan fluorescent protein (eCFP) and the acceptor/enhanced yellow fluorescent protein (eYFP). The donor and acceptor were assessed using argon lasers at 458 and 514 nm, respectively. Donor and acceptor emission images were captured using a Dual View 2 splitter (MAG biosystems; BioVision Technologies, Exton, PA, USA). The FRET/acceptor emission ratio was calculated for each pixel in the clearest optical plane for each image field. The cells were applied using ImageJ with a binary mask. A donor mask was generated by applying a threshold to the donor image, and a similar mask was generated for the acceptor channel. The CFP/FRET ratio (intensity of the CFP channel divided by that of the FRET channel) was calculated as follows: E = eCFPdonor/eYFPacceptor. It is correlated negatively with FRET efficiency but positively with force. For final image generation, we applied Pseudocolor in ImageJ.

Cytoplasmic osmolarity and protein nanoparticle count rate measurement

Cell culture medium, HEPES (osmotic solution), and a trypsin solution were all calibrated to 300 ± 10 mOsm/kg. Human glioblastoma U87 cells were cultured in the calibrated medium in 90-mm dishes. When their density exceeded 95%, the cells were stimulated with drugs for a certain duration. Brain tissues were also processed separately.

The cells or brain tissues were then digested and suspended in calibrated HEPES and transferred to 1.5-mL microcentrifuge tubes, followed by centrifugation (13,000×g, 5 min, 4 °C), ultrasonication (75% amplitude, 5 times, 5 s; Sonics and Materials, Connecticut, CT, USA), and another centrifugation (13,000×g, 10 min, 4 °C). Then, 50 µL of the supernatant solution (i.e., cytoplasm) was added to 0.5-mL test tubes. Next, we measured the cytoplasmic osmolarity in an Osmomat 3000 Freezing Point Osmometer (Gonotec, Berlin, Germany) and 050 Membrane Osmometer (Gonotec), which were calibrated three times. Kilocycles per second (Kcps) of cytoplasmic nanoparticles were detected in a Nanosight NS300 (Malvern Instruments, Malvern, UK).

Intracellular calcium and chloride ion measurement

Enhanced N-[ethoxycarbonylmethyl]−6-methoxy-quinolinium bromide (MQAE; Beyotime, China) and fura-2-acetoxymethyl ester (Fura-2 AM; Abcam, Cambridge, MA, USA) were used to detect intracellular chloride ion (Cl−) and calcium ion (Ca2+) concentrations, respectively. Cells were allowed to grow with adhesion on confocal dishes and used for experiments when the cell density was 50–70%. U87 cells were incubated with 5 µM MQAE at 37 °C for 30 min and washed five times with Krebs-HEPES buffer. The cells were incubated with 2 µM Fura-2 AM at 37 °C for 30 min and then treated with Hanks’ balanced salt solution, followed by incubation for another 30 min. The fluorescence intensity was detected using a Thunder Imaging Systems fluorescence microscope (SP8; Leica, Wetzlar, Germany), and images were obtained every 60 s. The normalized values of fluorescence intensity for MQAE (F0/Ft) and Fura-2 AM (Ft/F0) were calculated based on ion fluorescence intensities right after stimulation (Ft) and before stimulation (F0) for 15 min.

Immunofluorescence staining

Cells were seeded in 90-mm dishes and grown to 70–80% confluence. Next, the medium was removed, and the cells were treated with different drugs for a specified duration. The cells were then fixed using 4% paraformaldehyde for 30 min, permeabilized using 2% Triton-100, blocked with 4% bovine serum albumin at room temperature for 1 h, and incubated with antibodies at 4 °C overnight. Next, the cells were washed with phosphate-buffered saline and then exposed to goat antimouse FITC secondary antibody (1:200), goat antimouse tetramethylrhodamine secondary antibody (1:200), and FITC-phalloidin, followed by incubation at room temperature for 2 h in the dark. Finally, 4′,6-diamidino-2-phenylindole was added to stain nuclei, and fluorescent cells were observed under a confocal laser scanning microscope (SP5; Leica).

Assay of cytosolic ROS production

The intracellular ROS was determined by measuring the oxidative conversion of cell permeable DCFH-DA (10 µmol/L) to fluorescent dichlorofluorescein. The fluorescence was detected by a microsmolaritylate reader (Thermo Varioskan LUX, MA, USA) at an excitation wavelength 488 nm and an emission wavelength 525 nm.

Caspase-1 activity assay

Cells were seeded into a six-well plate, treated with different drugs for a specified duration, suspended in lysis buffer, and placed on ice for 15 min for digestion. The suspension was centrifuged at 16,000–20,000×g at 4 °C for 10–15 min. The supernatant was then added to 0.5-mL test tubes, and caspase-1 activity was measured using a Caspase-1 Inflammasome Assay kit (Beyotime), according to the manufacturer’s instructions.

RNA extraction and quantitative polymerase chain reaction

RNA was isolated from U87 cells by using Trizol Total (Invitrogen, CA, USA) and transcribed into cDNA by using a PrimeScript RT Reagent Kit (Vazyme, Nanjing, China) with gDNA Eraser. The quality and quantity of the isolated RNA were determined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). Quantitative polymerase chain reaction (qPCR) was performed on the Stratagene Mx3000P system (Agilent Technologies) by using a SYBR Green PCR Core Reagents kit (Yeasen, Shanghai, China), according to the manufacturer’s instructions. The relative mRNA expression of IL-1β (forward primer: 5′-AAAGCTTGGTGATGTCTGGT-3′; reverse primer: 5′-GGACATGGAGAACACCACTT-3′) and IL-18 (forward primer: 5′-TCGCAGATGGCTCTTTGCCT-3′; reverse primer: 5′-GAGGCCGATTTCCTTGGTCT-3′) were measured and standardized to that of β-actin (forward primer: 5′-GCAAGTGCTTCTAGGCGGAC-3′; reverse primer: 5′-AAGAAAGGGTGTAAAACGCAGC-3′). The following PCR amplification steps were used: enzyme activation at 95 °C for 5 min, followed by 40 cycles of denaturing at 95 °C for 10 s and annealing and extending at 60 °C for 30 s. Finally, the 2−ΔΔCt method was used for relative quantification of target gene expression.

Electrophysiological analysis

The patch clamp technique was used to record the membrane potentials of the U87 cells. The cells were plated onto 9-mm glass coverslips at a suitable density, mounted in a chamber (Warner Instruments, Hamden, CT, USA), placed onto the stage of an inverted microscope, and perfused with the specified solution. Next, measurements were obtained at room temperature in whole-cell configuration. In particular, the cells were perfused with a solution containing 144 mM NaCl, 4.3 mM KCl, 2.5 mM CaCl2, 1.1 mM MgCl2, 10 mM glucose, and 10 mM HEPES (pH 7.3); moreover, mannitol was used to adjust the osmolality to 300–310 mOsm/kg. Ca2+, K+, and albumin concentrations were adjusted according to the experimental requirements. Patch electrodes were pulled from borosilicate capillary glass (outer diameter = 1.5 mm; inner diameter = 0.86 mm; Sutter Instruments, USA) with a vertical two-stage pipette puller (P-1000; Narishige). Measurements were performed using glass pipettes (between 3 and 6 MΩ) filled with an intracellular solution comprising 140 mM KCl, 2 mM MgCl2, 2 mM Na2ATP, 1 mM CaCl2, 5 ethylene glycol tetraacetic acid, and 5 mM glucose (pH 7.3, 290–300 mOsm/kg). All measurements were performed on a Multiclamp 700B amplifier at a data-sampling rate of 2 kHz with a Digidata 1550B Digitizer controlled using PClamp (version 10.6; Molecular Devices, Sunnyvale, CA, USA). Membrane potentials were recorded in gap-free mode using a current clamp; membrane resistance Rm > 1GΩ before membrane rupture, access resistance Ra < 25MΩ and hold current <−100pA after membrane rupture.

Animals and brain slice culture

Sprague Dawley rats (age = 10 days) were purchased from Shanghai Sippe-Bk Lab Animal Co Ltd. (Shanghai, China; SCXK(Hu) 2018-0006), and housed at 23 to 25 °C with a 12-hour light/dark cycle and allowed free access to water and standard laboratory diets. The animal experiment protocol and animal manipulations was approved by the Experimental Animal Ethical Committee of Nanjing University of Chinese Medicine, China (Approval No. 202201A037), and in accordance with the ARRIVE guidelines. All methods were performed in accordance with the relevant guidelines and regulations.

The rats were deeply anesthetized with ketamine/xylazine (60/5 mg/kg, i.p.) and euthanized by decapitation. The brain was rapidly removed and sectioned into 350-µm-thick slices, cultured in artificial cerebrospinal fluid at 37 °C under 1% O2 and 5% CO2, and stimulated with 100 µg/ml LPS for 12 h. Next, the sections were exposed to 200 or 500 µmol/L H2O2 for 1 h. The sections were divided into the control, LPS, and LPS + H2O2 groups. Finally, all brain slices were frozen for subsequent experimentations.

Hematoxylin and Eosin staining of rat brain slices in situ

Rat brain tissues were fixed in 4% paraformaldehyde at 4 °C for 24 h, dehydrated in an ethanol concentration series, and embedded in paraffin. Next, the paraffin-embedded brain tissues were sliced into 4-µm-thick consecutive sections on a microtome. Finally, the sections were dewaxed, dehydrated, and stained with hematoxylin and eosin (HE). The images of the cortical and hippocampal areas were acquired under a microscope at 200× magnification. The scores are based on the morphological structure of brain cells and the surrounding conditions of the cells. 0 points, normal; 1 point, a small amount of cell damage, no vacuoles around the cells; 2 points, obvious cell damage, with a small number of vacuoles around the cells; 3 points, obvious cell damage, with little inflammatory cell infiltration and obvious vacuoles around the cells; 4 points, a large number of cells damaged, with much inflammatory cell infiltration and a large number of vacuoles observed around the cells.

Calcium content assay

Calcium content was determined by a calcium assay kit (ab102505; Abcam, Cambridge, UK), according to the manufacturer’s indications. In brief, brain tissues were stimulated with LPS and H2O2 for 2 h and then extracted using a lysis reagent. The tissue extracts were incubated with 90 µL of chromogenic reagent and 60 µL of calcium assay buffer for 10 min in the dark. Finally, the optical density of the chromogen was measured at 575 nm on a microplate reader (model 680; Bio-Rad, Hercules, CA, USA), according to the manufacturer’s instructions.

Western blotting

Brain tissue and cell proteins were extracted using a RIPA lysis buffer. The lysates were centrifuged (12000 g/min) for 10 min, and the supernatants were collected. The protein concentration was measured by a BCA kit. After quantification, the proteins were resolved through 10–12% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes. The PVDF membranes were blocked with 5% nonfat dry milk in tris-buffered saline with Tween 20, and incubated with primary antibodies (ASC, G3BP1 and GAPDH) at 4◦C overnight. After washing with TBST, the membrane was incubated with HRP-conjugated secondary antibody for 2 h at room temperature. The relative expression level of the target protein was normalized to that of β-actin. The antibody reactivity was then detected using an ECL kit (Yeasen, Shanghai, China) and quantified using Image J.

Liquid chromatography–tandem mass spectrometry analysis and proteomic data processing

Cultured U87 cells were washed twice with an isotonic HEPES buffer and then subject to either no treatment (control group) or cotreatment with cytochalasin B (Cyto B; a microfilament depolymerizer) and Noc (a microtubule depolymerizer) for 30 min (Cyto B + Noc group). The treated cells were obtained through digestion, followed by centrifugation at 14,000×g at 4 °C, ultrasonication (Sonics and Materials, Newtown, CT, USA), and re-centrifugation at 14,000×g at 4 °C, to obtain cytoplasmic supernatants.

Next, we determined total protein concentrations in the supernatants by using a Nanodrop Microvolume Spectrophotometer. For all subsequent steps, we used 100-µg aliquots of the samples. Each sample was mixed with DTT and alkylated with IAA, then combined with urea. Filter-Aided Sample Preparation filters were activated using NH4HCO3 and centrifuged. All peptides were eluted via the membrane and dried in a vacuum centrifuge.

Next, peptides were resolved in the C18 reverse phase column (Thermo Fisher Scientific EASY-nLC) and analyzed on a Q Exactive Plus mass spectrometer (Thermo Fisher Scientific) with higher-energy collisional induced dissociation fragmentation. The mass spectrometer was operated in top 20 data-dependent mode with automated switching between mass spectrometry (MS) and tandem MS (MS/MS). Raw data were processed using MaxQuant and mapped onto the UniProtKB/Swiss-Prot database. All spectra were searched against the UniProt human database. Target decoy analysis was performed by searching a reverse database, with an overall false discovery rate of 1% at the protein and peptide levels. Label-free quantification was performed using the LFQ feature included in MaxQuant, according to default LFQ parameters.

Finally, the acquired data were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses30,31,32. All pathways were ranked based on their p values, and those with p ≤ 0.05 were considered significantly enriched.

Statistical analysis

All data are expressed as means ± standard deviations (SDs) of at least three independent experiments. Statistical analysis was performed using GraphPad Prism (version 8.0). One-way analysis of variance followed by the Tukey test was used to compare the statistical significance of differences between the samples and their respective controls. A p value of < 0.05 was considered to indicate statistical significance.

Results

Stress granules assembly regulate protein nanoparticles-related osmotic pressure and membrane potential

To assess the effects of SGs on osmotic pressure, which is closely associated with cytoskeletal tension, we designed a vimentin tension probe based on FRET and evaluated the IF tension, which indicates transmembrane osmotic pressure changes33,34. In addition, we stimulated U87 cells with 100 ng/mL LPS for 4 h and then 0.1 mmol/L H2O2 for 30 min and found that H2O2 stimuli could increase ROS contents to cause oxidative stress (Fig. S1A), and LPS and H2O2 co-stimuli induced the formation of SGs, mainly consisting of G3BP1 in U87 cells (Fig. S1B). Moreover, SGs inhibited ASC inflammasome, IL-1β, IL-18 and caspase-1 activation in LPS and nigericin–stimulated U87 cells (Fig. S1B–E). Next, we analyzed the effects of SGs on the osmotic pressure, Ca2+ signaling, and membrane potential. The EWSR1 mutant EWSR1-G511A is a SG inhibitor, which influences SG dynamics, whereas TRIM21 promotes SG production by inhibiting SG autophagy17. Moreover, EWSR1-G511A inhibited SG formation and increased IL-1β and IL-18 gene expression in U87 cells (Fig. S2A-B, Fig. S3A-B), and TRIM21 knockdown using TRIM21 siRNA promoted SG production, alleviating inflammatory stress (Fig. S3G–J). Therefore, the effects of SGs on the protein nanoparticle-related osmotic pressure (PN-OP) were evaluated using EWSR1-G511A and TRIM21 siRNA. The result demonstrated that SGs reduced cell tension, CFP/FRET ratio, intracellular protein nanoparticle number, and cytoplasmic osmolarity in H2O2 and LPS co-stimulated U87 cells (Fig. 1A–D). Furthermore, EWSR1-G511A cell tension, CFP/FRET ratio, intracellular protein nanoparticle number, and cytoplasmic osmolarity (p < 0.001 or p < 0.0001), whereas TRIM21 knockdown reduced them (p < 0.001 or p < 0.0001). Thus, SGs may increase PN-OP significantly. We also noted that cytoplasmic Ca2+and Cl− contents and membrane potential difference in EWSR1-G511A-stimulated U87 cells were higher than those in U87 cells stimulated with LPS and H2O2, and K+ contents were significantly decreased in U87 cells stimulated with LPS and H2O2 (p < 0.05, p < 0.01, or p < 0.001; Fig. 1E–H); nevertheless, TRIM21 siRNA reduced cytoplasmic Ca2+ and Cl− contents and membrane potential in EWSR1-G511A-stimulated U87 cells, and increased K+ contents (p < 0.05 or p < 0.01, Fig. 1E-H, Fig. S4). These results confirmed that SG formation is crucial for the regulation of PN-OP and membrane potential alterations.

Fig. 1
figure 1

SGs regulate protein nanoparticle-related osmotic pressure in U87 cells stimulated with LPS and H2O2. All U87 cells were incubated with 100 ng/mL LPS for 4 h and then with 0.1 mmol/L H2O2 for 30 min in the presence or absence of EWSR1-G511A or TRIM21 siRNA. (A) Images taken during a 30-min time lapse of FRET analysis (calibration bar: 0.1–1.5; scale bar: 10 μm). (B) Standardized CFP/FRET ratio of vimentin at minute 30. (C) Cytoplasmic osmolarity determined on a freezing-point osmometer. (D) protein nanoparticle count rate. (E-G) cytoplasmic Ca2+ (E), Cl− (F) and K+(G). (H) Membrane potential variation. Ctrl, control. Data represent mean ± SD of n = 3. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Full size image

Stress granule formation effectively inhibits inflammasome to maintain PN-OP and membrane potential stability

Stress granule generation can inhibit inflammasomes in U87 cells. We previously demonstrated that inflammasome and depolymerized cytoskeleton are important sources of cytoplasmic protein nanoparticles, which are closely related to osmotic pressure and membrane potential27,35,36. TRIM21 downregulation can not only promote SG production but also influence inflammation37,38. To assess whether inflammation is involved in the SG-based regulation of osmotic pressure, we measured cell tension, intracellular osmolarity, and membrane potential. As illustrated in Fig. 2A–D, TRIM21 siRNA significantly reduced the CFP/FRET ratio, intracellular protein nanoparticle number, and cytoplasmic osmolarity in inflamed U87 cells stimulated with LPS and nigericin (p < 0.05 or p < 0.0001). Moreover, Fluo4, MQAE and EPG-4 were used to measure Ca2+, Cl− and K+ contents respectively, and that TRIM21 siRNA was found to directly reduced Ca2+ and Cl− contents, increased K+ contents and blocked membrane potential alterations in inflammation-related U87 cells (p < 0.01; Fig. 2E–H, Fig.S5). Taken together, these results exhibited that SGs regulate PN-OP and membrane potential by inhibiting inflammation in U87 cells.

Fig. 2
figure 2

SGs affect protein nanoparticle-related osmotic pressure in inflammasomes in U87 cells stimulated with LPS and nigericin. All U87 cells were incubated with 100 ng/mL LPS for 4 h and then with 20 µmol/L nigericin for 30 min in the presence or absence of EWSR1-G511A or TRIM21 siRNA. (A) Images taken during a 30-min time lapse of FRET analysis (calibration bar: 0.1–1.5; scale bar: 10 μm). (B) Standardized CFP/FRET ratio of vimentin at minute 30. (C) Cytoplasmic osmolarity determined on a freezing-point osmometer. (D) protein nanoparticle count rate. (E-G) cytoplasmic Ca2+ (E), Cl− (F) and K+(G). (H) Membrane potential variation. Ctrl, control; Nig, nigericin. Data are expressed as means ± SD(n = 3). *p < 0.05, **p < 0.0,.****p < 0.0001.

Full size image

To verify the regulatory effects of SGs on inflammation-related osmotic pressure further, we assessed how EWSR1-G511A, a SG inhibitor, affects PN-OP and inflammation indicators. The results demonstrated that EWSR1-G511A upregulated CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number (p < 0.05 or p < 0.0001; Fig. 3A–C) and increased IL-1β and IL-18 gene expression in U87 cells under oxidative stress and inflammation stimulated by LPS, nigericin, and H2O2 (Fig. S3C–D). However, EWSR1-G511A was observed to not affect osmolarity indicators (Fig. 2) or IL-1β or IL-18 gene expression in inflamed U87 cells stimulated by LPS and nigericin stimuli (Fig. S3E-F). Thus, SGs may inhibit osmolarity changes by blocking protein nanoparticle generation within inflammasomes without direct regulation of inflammation in U87 cells.

Fig. 3
figure 3

Mature SGs stabilize protein nanoparticle-related osmotic pressure in U87 cells. All U87 cells were incubated with 100 ng/mL LPS for 4 h and then with 0.1 mmol/L H2O2 and 20 µmol/L nigericin for 30 min in the presence or absence of EWSR1-G511A or TRIM21 siRNA. (A) Standardized CFP/FRET ratio of vimentin at minute 60 (n ≥ 3). (B) Cytoplasmic osmolarity determined on a freezing-point osmometer. (C) protein nanoparticle count rate. Ctrl, control; Nig, nigericin. Data are expressed as means ± SD (n = 3). *p < 0.05, ****p < 0.0001.

Full size image

Mature stress granule formation is beneficial to protein nanoparticle-related osmotic pressure recovery

SGs form in two stages: formation and maturation39. In the formation stage, under stimulation, the disordered domain of proteins such as G3BP1 is driven by phase separation, enabling the aggregation of neighboring stress particle components, followed by the formation of primary SGs (size < 600 nm). Next, in the maturation stage, several primary SGs fuse with each other through cytoskeletal transport, forming mature SGs40. To confirm the effects of SG formation and maturation on osmolarity, we examined LPS-stimulated U87 cells treated with H2O2 over different periods and noted that H2O2 continuously increased CFP/FRET ratio, and cytoplasmic osmolarity over 15–45 min of treatment (Fig. 4A-B); thus, SG formation may affect osmolarity alterations. H2O2 also increased intracellular protein nanoparticle number (Fig. 4C). SGs can gradually mature after 60 min of H2O2 stimulation in inflamed cells41. We noted that 60 min of H2O2 stimulation significantly normalized cell tension, CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number in LPS-treated U87 cells. Thus, SG formation may regulate PN-OP changes, and mature SGs may be beneficial to PN-OP recovery. Moreover, PN-OP indicators tended to remain stable 1 h after H2O2 withdrawal in LPS-stimulated U87 cells, suggesting that SG depolymerization does not affect PN-OP stability.

Fig. 4
figure 4

Mature SG formation is beneficial to protein nanoparticle-osmotic pressure recovery in U87 cells. All U87 cells were stimulated with 100 ng/mL LPS for 4 h, followed by incubation with 0.1 mmol/L H2O2 for 0, 15, 30, 45, and 60 min and then by stimulant removal and further incubation for 60 min (i.e., until 120 min after initial stimulation). (A) Standardized CFP/FRET ratio of vimentin at minutes 0, 15, 30, 45, 60, and 120 (n ≥ 3). (B) Cytoplasmic osmolarity determined on a freezing-point osmometer. (C) protein nanoparticle count rate. Ctrl, control. Data are expressed as means ± SD (n = 3). ****p < 0.0001.

Full size image

EWSR1-G511A prevents protein nanoparticle-related osmotic pressure recovery by inhibiting SG formation and depolymerization

Because EWSR1-G511A inhibited SG formation, we next assessed the effects of EWSR1-G511A-inhibited SG formation on the PN-OP regulation process. In LPS-stimulated U87 cells treated with H2O2 for different periods, it was continuously increased CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number over 15–45 posttreatment min, with all the values normalizing after 60 min of H2O2 stimulation. However, EWSR1-WT could not delay or accelerate PN-OP changes in U87 cells (p > 0.5, Fig. 5A-C), whereas EWSR1-G511A delayed the recovery of cell tension (p < 0.05), CFP/FRET ratio (p < 0.05), cytoplasmic osmolarity (p < 0.001) and intracellular protein nanoparticle number (p < 0.0001) after 60 min of H2O2 stimulation (Fig. 5D-F). Thus, EWSR1 mutation may inhibit SG formation, resulting in incomplete SG maturation after 1 h of H2O2 stimulation, affecting PN-OP stability.

Fig. 5
figure 5

EWSR1 mutation prevents protein nanoparticle-related osmotic pressure recovery by inhibiting SG formation and depolymerization in U87 cells. All U87 cells were stimulated with 100 ng/mL LPS for 4 h, followed by incubation with 0.1 mmol/L H2O2 for 0, 15, 30, 45, and 60 min and then by stimulant removal and further incubation for 60 min (i.e., until 120 min after initial stimulation) in the presence or absence of EWSR1-G511A. (A and D) Standardized CFP/FRET ratio of vimentin at minutes 0, 15, 30, 45, 60, and 120 (n ≥ 3). (B and E) Cytoplasmic osmolarity determined on a freezing-point osmometer. (C and F) protein nanoparticle count rate. Ctrl, control. Data are expressed as means ± SD (n = 3). ***p < 0.001.****p < 0.0001.

Full size image

In cells cultured under normal conditions, intracellular SGs produced by stress stimuli can be completely degraded within 1 h42. Similar regulatory mechanisms were observed in U87 cells stimulated with LPS and H2O2. As shown in Fig. 5A-C, CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number tended to remain stable after 45 min of H2O2 withdrawal in LPS-stimulated U87 cells. In contrast, EWSR1-G511A inhibited PN-OP recovery, indicated by significant increases in cell tension, CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number after 1 h of H2O2 withdrawal, suggesting that EWSR1-G511A delays SG depolymerization, preventing PN-OP restoration.

TRIM21 downregulation accelerates SG maturation and induces protein nanoparticle-related osmotic pressure recovery

H2O2 stimulation induced CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number to reach their maximum values at 45 posttreatment min and then began to decrease until PN-OP normalized at 1 h in LPS-stimulated U87 cells. Moreover, PN-OP indicators tended to be stable after 1 h of H2O2 withdrawal (Fig. 6A-C). TRIM21 knockdown accelerates SG production, enhancing SG functions37,38. Thus, we examined the effects of TRIM21 knockdown on PN-OP. Compared with NC siRNA (Fig. 6A-C), TRIM21 siRNA induced CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number to reach their maximum values at 30 min and normalized at 45 min. However, CFP/FRET ratio, cytoplasmic osmolarity, and intracellular protein nanoparticle number did not change significantly after 1 h of H2O2 withdrawal in U87 cells (Fig. 6D-F). These results indicated that TRIM21 knockdown may accelerate SG formation, resulting in a rapid increase and recovery of PN-OP. TRIM21 siRNA also induced membrane potential stabilization in U87 cells (Fig. 1G). Taken together, these results suggested that TRIM21 knockdown promotes SG maturation and induces PN-OP recovery and membrane potential stability, thus protecting the U87 cells.

Fig. 6
figure 6

TRIM21 downregulation accelerates SG maturation inducing protein nanoparticle-related osmotic pressure recovery in U87 cells. All U87 cells were stimulated with 100 ng/mL LPS for 4 h, followed by incubation with 0.1 mmol/L H2O2 for 0, 15, 30, 45, and 60 min and then by stimulant removal and further incubation for 60 min (i.e., until 120 min after initial stimulation) in the presence or absence of TRIM21 siRNA. (A and D) Standardized CFP/FRET ratio of vimentin at minutes 0, 15, 30, 45, 60, and 120 (n ≥ 3). (B and E) Cytoplasmic osmolarity determined on a freezing-point osmometer. (C and F) protein nanoparticle count rate. Ctrl, control. Data are expressed as means ± SD (n = 3). *p < 0.05, ****p < 0.0001.

Full size image

Cytoskeleton depolymerization–induced protein nanoparticles generation blocks SG formation and promotes PN-OP alterations

Cytoskeleton depolymerization leads to the production of many protein nanoparticles, followed by an increase in cell tension and osmolarity43. This, we exposed LPS-stimulated U87 cells to Cyto B and Noc and observed that cell tension, CFP/FRET ratio, cytoplasmic osmolarity, intracellular protein nanoparticle number, and Ca2+ content increased significantly (p < 0.001 or p < 0.0001; Fig. 7A-E). SG generation also increased cell tension, osmolarity, and protein nanoparticles (Figs. 1 and 2). H2O2 treatment also directly impaired PN-OP stability, manifested as increases in cell tension, CFP/FRET ratio, cytoplasmic osmolarity, intracellular protein nanoparticle number, and Ca2+ content (p < 0.05, p < 0.01 or p < 0.0001; Fig. 7A–E). To confirm the effects of cytoskeleton depolymerization on SGs, we exposed U87 cells stimulated with LPS and H2O2 to Cyto B and Noc and noted persistent increases in PN-OP and Ca2+ signaling indicators (p < 0.01 or p < 0.0001; Fig. 7A-E). Thus, cytoskeleton depolymerization may lead to an increase in protein nanoparticles, which blocked SG assembly, leading to the formation of primary SGs sized < 600 nm, inhibiting SG maturation and exacerbating PN-OP alterations.

Fig. 7
figure 7

Protein nanoparticle generation induced by cytoskeleton depolymerization blocks SG formation in U87 cells. All U87 cells were incubated with 0.1 mmol/L H2O2, 10 µmol/L Cyto B, and 100 µmol/L Noc for 15 min. (A) Images taken during a 15-min time lapse of FRET analysis (calibration bar: 0.1–1.5; scale bar: 10 μm). (B) Standardized CFP/FRET ratio of vimentin at minute 15 (n = 3). (C) Cytoplasmic osmolarity determined on a freezing-point osmometer. (D) protein nanoparticle count rate. (E) Normalized fluorescence intensity of Ca2+. Ctrl, control. Data are expressed as means ± SD of n = 3. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Full size image

Because cytoskeletal depolymerization can induce protein nanoparticle generation in U87 cells, we performed proteomic analysis based on liquid chromatography (LC)–MS to identify changes in cytoplasmic protein nanoparticles in Cyto B + Noc-stimulated cells. In total, 80 and 357 protein types were upregulated and downregulated in the Cyto B + Noc group compared with the control group, respectively (Fig. S6A), suggesting that cytoskeletal polymerization facilitates changes in various protein nanoparticle indicators. Furthermore, to identify the pathways altered in the Cyto B + Noc group, we performed subcellular localization, GO, and KEGG pathway enrichment analyses. The categories related to the cytoplasm, endoplasmic reticulum, mitochondria, and nucleus were significantly enriched in the Cyto B + Noc group (Fig. S6B), suggesting that protein nanoparticles demonstrate cytoplasmic translocation. In the GO enrichment analysis, the cellular component, biological process, and molecular function terms related to cytoplasmic ribonucleoprotein granule, structural constituent of cytoskeleton, protein–RNA complex assembly, and microtubule plus-end binding were enriched (Fig. S6D). The KEGG pathway enrichment analysis revealed that significant differentially expressed proteins (q < 0.05) between the control and Cyto B + Noc groups were mainly enriched in apoptosis pathways (mainly the mTOR and cGMP-PKG pathways) and metabolic pathways (mainly the tricarboxylic acid cycle, glyoxylate and dicarboxylate metabolism, and lipoic acid metabolism). Cyto B + Noc-induced cytoskeletal polymerization potentially affected Ca2+ flow and PN-OP by regulating the cAMP and calcium pathways (Fig. S6C). Therefore, protein nanoparticle production elicited by cytoskeletal depolymerization may affect various intracellular osmolarity- and Ca2+-related signaling pathways and influence electrochemical–tension interactions related to PN-OP in U87 cells.

Stress granules elicit neuroprotection by preventing inflammation and changing osmolarity in rat brain tissues

We evaluated the effects of SGs in H2O2-and LPS-stimulated rat brain slices. Compared with the LPS group, the H2O2-induced LPS group demonstrated significant decreases in rat brain osmolarity, protein nanoparticle, and Ca2+ contents (Fig. 8A–C), suggesting that SGs positively affect changes in intracellular osmolarity and Ca2+ signaling regulation in the brain. HE staining revealed that LPS caused neural cell damage and inflammation, then increased HE score. However, H2O2 could decrease HE score in LPS-induce brain tissues (Fig. 8D-E). Western blotting results confirmed that ASC expression increased in the LPS group, however, 500 µmol/L H2O2 induced increase in G3BP1 expression and inhibit ASC expression (Fig. 8F). Thus, SGs may inhibit inflammation and regulate Ca2+ signaling and intracellular osmolarity, protecting brain tissues from LPS-induced inflammation.

Fig. 8
figure 8

SG formation limited neural damage in LPS-stimulated rat brain tissues. (A) osmolarity value determined on a freezing-point osmometer. (B and C) protein nanoparticle count rate (B) and Ca2+ content (C) in brain tissues. (D-E) Representative images of HE-stained brain tissues and HE score (scale bars: 100 μm). (F) G3BP1 and ASC expression. Ctrl, control. Data are expressed as means ± SD (n = 3). ***p < 0.001, ****p < 0.0001.

Full size image

Discussion

Stress granules, membrane-less condensates of RNA and RNA-binding proteins, form in response to stress stimuli3. Recent studies have suggested that SGs are associated with various nerve diseases and thus have physiological importance44,45. SG formation can specifically inhibit NLRP3 inflammasome activation and maintain cell survival under stress18. The current results suggested that SG generation is accompanied by changes in intracellular protein nanoparticles and that inhibition of SG maturation increases protein nanoparticle number; thus, SGs may somewhat stabilize intracellular protein nanoparticle number. Inflammasomes such as NLRP3 inflammasomes are involved in mass protein nanoparticle production; moreover, protein nanoparticles are a crucial factor regulating transmembrane osmotic pressure46. Therefore, SGs can maintain the stability of intracellular osmotic pressure, Ca2+ signaling, and membrane potential via protein nanoparticle production inhibition, protecting cellular structure and function in inflammation-mediated U87 cells.

According to electric double layer theory, protein nanoparticles, but not protein microparticles, may be involved in the regulation of free ions via cation adsorption, resulting in the formation of charge gradients between the inner and outer sides of the cell membrane47. Intracellular protein nanoparticles can be effectively regulated through cytoskeletal depolymerization34. Primary SGs fuse with each other via cytoskeletal proteins (i.e., microfilaments or microtubules) to form mature SGs. Thus, cytoskeletal depolymerization is involved in SG regulation25. The current result demonstrated that cytoskeletal depolymerization–induced protein nanoparticles production and Ca2+ signaling may delay SG generation and maturation. These results clarified a novel function of SGs facilitating membrane potential and intracellular protein nanoparticle-related osmotic pressure regulation.

Stress granules homeostasis is modulated by TRIM21 and EWSR1, accelerating or inhibiting SG formation and maturation48. TRIM21 and EWSR1 are also involved in electrochemical–tension interactions, manifesting as intracellular protein nanoparticle and Ca2+ signaling regulation and playing an antagonistic role against PN-OP via the regulation of SG maturation and formation. EWSR1 reduces SG formation, increasing intracellular abnormal protein nanoparticle number, indirectly upregulating inflammatory responses and tension effects of PN-OP, and facilitating inflammation-mediated cellular edema. TRIM21 knockdown promotes early SG maturation, further inducing rapid PN-OP recovery and Ca2+ signaling and membrane potential stability, thereby protecting cells. Thus, SG formation and maturation may be involved in PN-OP stabilization and intracellular electromechanical recovery.

Studies have suggested that SG maturation can block mass production of protein nanoparticles, such as G3BP1 and DDX3X, eliciting two varied supramolecular structure assembly modes with NLRP3 in response to oxidative stimuli and then blocking IL-1β and IL-18 release and neuroinflammatory responses49,50. However, protein nanoparticle production, such as that via cytoskeletal depolymerization, can block SG transport and delay SG maturation. Thus, intracellular SGs and protein nanoparticles can regulate each other, exerting crucial effects on cellular electromechanical homeostasis. Therefore, as cytoplasmic protein aggregation particles, SGs are involved in chemical signal regulation; moreover, they elicit biophysical roles in the regulation of membrane potential and tension effects of intracellular osmolarity via their interaction with protein nanoparticles.

Conclusion

This study revealed a novel biophysical function of SGs: SGs suppress abnormal protein nanoparticle generation and Ca2+ signaling and restore membrane potential and transmembrane osmotic pressure in H2O2- and LPS- co-induced U87 cells. Moreover, EWSR1 expression and TRIM21 suppression elicit electromechanical activity, which enables the formation and early maturation of SGs; these SGs then stabilize protein nanoparticle-related osmotic pressure and membrane potential. However, protein nanoparticle production restrains SG maturation. Therefore, SG maturation–protein nanoparticle production interactions are vital in the induction of electromechanical activities through the regulation of Ca2+ signaling, membrane potential, and tension effect of PN-OP, which was novel mechanism involving electrochemical–tension interactions.