The chemicals used were hemin (Sigma, St. Louis, MO, United States). Hemin was dissolved in ammonia water and pH value was corrected to pH 7.4. Hemin was diluted with DMEM or Neurobasal medium (NB, Thermo Fisher Scientific, Waltham, MA, United States).

CTX-TNA2, an astrocyte cell line isolated from the cortex of 1-day old rats, was purchased from Bioresource Collection and Research Center (Taiwan) and maintained in Dulbecco’s Modified Eagle Medium (Sigma) supplemented with 10% (v/v) fetal bovine serum (Cytiva, Marlborough, MA, United States) and 1% penicillin-streptomycin-amphotericin B (Cytiva) in an incubator under 5% CO2 at 37°C. Prior to the hypoxic preconditioning, CTX-TNA2 cells were seeded at a density of 1 × 10⁶ cells in a 10-cm culture dish and were incubated in 5 mL NB without exosomes. Hypoxic treatment was conducted by incubating cells in an incubator with 1% oxygen. Two hypoxic preconditionings were designed as follows. One hypoxic treatment was to expose CTX-TNA2 cells to 2-h hypoxia followed by 6-h normoxia. The other was to expose CTX-TNA2 cells to 12-h hypoxia followed by 24-h normoxia.

At the end of hypoxia-renormoxia treatments, exosomes obtained from the culture medium were isolated using ExoQuick (System Biosciences, Palo Alto, CA, United States). In brief, the culture medium from 10-cm culture dishes was collected and centrifuged at 2600 × g for 30 min at 4°C. The supernatant was transferred to a new centrifuge tube. Five milliliters of culture medium mixed with 1 mL of ExoQuick and the tube was placed in a refrigerator at 4°C overnight. The supernatant was collected and centrifuged at 1,500 × g for 30 min at 4°C. Again, the supernatant was collected and centrifuged at 1,500 × g for 5 min. After removing the remaining supernatant, the exosome pellet was re-dissolved in PBS at 4°C and stored at 4°C. Approximately, 5 × 10⁸ exosome particles obtained from 1 × 10⁶ cells in a 10-cm culture dish were used for further experiments as 1-fold (1x) of exosomes.

Nanoparticle Tracking Analysis (NTA) was employed to characterize exosomes, including size distribution and exosome particle concentration by NanoSight NS300 (Malvern Panalytical, Malvern, UK). This technique was capable of detecting vesicles with sizes between 0.05 and 1 μm. At the same time, zeta potential of exosomes was analyzed using the Nano ZS (Malvern Panalytical).

An aliquot of exosomes was applied onto an electron microscopy grid and incubated for 5 min and fixed with 2% paraformaldehyde for 5 min. The samples were counterstained with 1% uranyl acetate for 1 min, air-dried overnight, and observed using a Transmission Electron Microscope JEM-1400Plus (JEOL Ltd., Tokyo, Japan).

MicroRNA(miRNA) libraries were prepared using the QIAseq miRNA UDI Library Kit (Qiagen, Venlo, Netherlands) and evaluated for size distribution using the Agilent 4150 TapeStation with HS D1000 ScreenTape (Agilent, Santa Clara, CA, United States). Sequencing was conducted with single end reads (100 bp) at a depth exceeding 10 million reads per sample. miRNA expression levels were quantified using the Qiagen GeneGlobe platform (Qiagen) with UMI-based correction.

Pregnant female Sprague-Dawley (SD) rats were supplied by BioLASCO Taiwan Co., Ltd. (Yilan, Taiwan). All animals (one rat/individually ventilated cage) were housed in an air-conditioned room (22 ± 2°C) on a 12 hr-light/dark cycle (07:00–19:00 h light) and had free access to food and water. To prepare primary cultured cortical neurons, embryonic day 17 fetal rat brains were obtained from pregnant female SD rats of 17-day gestation which were sacrificed by an overdose of Zoletil^® (Virbac, Taipei, Taiwan) to minimize pain or discomfort used. The use of animals and all experiments conducted were under approved protocols from the Institutional Animal Care and Use Committee (IACUC) of Taipei Veterans General Hospital, Taipei, Taiwan. The approval number is IACUC2022-235. In addition, the authors complied with the ARRIVE guidelines. All experiments were performed in accordance with relevant guidelines and regulations.

Cerebral cortices of fetal rats were isolated and dissociated mechanically. The dissociated cells were suspended in the Basal Medium Eagle medium (BME, Thermo Fisher Scientific) containing 20% fetal bovine serum, and were seeded onto a 35-mm culture dish (IWAKI, Tokyo, Japan) with a density of 5 × 10⁶ cells per dish. Afterwards, cells were maintained with serum-free NB medium supplemented with B27 (Thermo Fisher Scientific) in the incubator with 5% CO₂ at 37°C. Hemin-induced neurotoxicity was established by exposing primary cultured cortical neurons with hemin (30 μM).

A modified sulforhodamine B (SRB) assay was employed to measure cell viability. At the end of experiment, cells in 96-well plates were washed with phosphate-buffered saline (PBS) and 10% trichloroacetic acid (TCA, Merck, Boston, MA, United States) and were then incubated at 4°C for 1h. Afterwards, TCA solution was removed, and the cells were washed with double distilled H₂O. Cells were then incubated with SRB solution (0.4% in 1% acetic acid, Merck) for 10 min. Afterwards, SRB solution was removed, and the cells were washed with 1% acetic acid. After removing acetic acid, the sample was air-dried and 20 mM unbuffered Tris base was used to dissolve the resulting formazan product. The absorption was measured by an ELISA reader (TECAN Sunrise, Männedorf, United States) at 540 nm with a reference wavelength of 690 nm.

At the end of 16-h treatments, the cells were collected, washed with phosphate buffered saline (PBS), and lysed in radioimmunoprecipitation assay (RIPA, Cell Signaling Tech. Beverly, MA, United States) lysis buffer containing 20 mM Tris HCl, 150 mM NaCl, 1% (v/v) NP-40, 1% (w/v) sodium deoxycholate, 1 mM ethylenediaminetetraacetates (EDTA), 0.1% (w/v) sodium dodecyl sulfate polyacrylamide (SDS) and 0.01% (w/v) sodium azide (pH 7.5) for 20 min on ice. Lysates were then centrifuged at 13800 × g for 10 min, and the protein concentrations of supernatants were determined by Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). Protein samples (30 μg) were run on 12–13.5% SDS-polyacrylamide gel electrophoresis and then transferred onto a polyvinylidene difluoride (PVDF, Bio-Rad, Hercules, CA, United States) at 100 V for 120 min. Blots were probed with primary antibodies including antibodies against GAP43 (Cell Signaling Technology), iNOS, COX-2, HO-1 (StressGen, Victoria, CA, United States), GPX4 and active-caspase 3 (Cell Signaling Technology) overnight at 4°C. After incubation of primary antibodies, the membrane was washed and incubated with a secondary antibody for 1 h at room temperature. The secondary antibodies were horseradish peroxidase-conjugated secondary IgG (Chemicon, Temecula, CA, United States). The immunoreaction was visualized using Amersham Enhanced Chemiluminescence (Amersham Pharmacia Biotech, Piscataway, NJ, United States). After this measurement, the bound primary and secondary antibodies were stripped by incubating the membrane in stripping buffer (100 mM 2-mercaptoethanol, 2% SDS) at 50°C for 5 min. The membrane was reprobed with a primary antibody against β-actin (Millipore, Billerica, MA, United States).

At the end of treatments, the primary cultured cortical neurons were fixed with 4% paraformaldehyde (Merck). Cells were then washed with 0.1 M PBS, incubated with 0.3% Triton X-100 (Sigma) and 1% goat serum (GS; Jackson ImmunoResearch, West Grove, PA, United States), and blocked with 3% GS for 60 min. Next, cells were processed for immunostaining using monoclonal antibody specific for rat Neu N (Cell Signaling Technology) in 1% GS-PBS at 4°C for 24 h. The cells were then incubated in fluorescein conjugated-IgG (FITC) (Jackson ImmunoResearch) for 1 h at room temperature, mounted in glycerol (Merck) and exosome stained with PKH26 (Sigma). Controls consisted of omissions of primary antibodies. The sections were visualized by a fluorescence confocal microscope (Olympus FluoView, Norfolk, VA, United States).

All data are expressed as the mean ± S.E.M. The results were analyzed by one-way analysis of variance (one-way ANOVA) followed by the LSD test (cell viability and Western blot assay) or Tukey multiple comparison (NTA) as post hoc methods. The significance level was set at p < 0.05.

To delineate the effect of hypoxic preconditionings on exosome secretion, CTX-TNA2 astrocyte cells were treated with 3, 6, and 12 h hypoxia. SRB assay showed that hypoxia (1% oxygen) for 3, 6, 12 h did not cause significant cell loss of CTX-TNA2 cells (Figure 1A). Therefore, 2-h hypoxia followed by 6-h normoxia (2H/6R) and 12-h hypoxia followed by 24-h normoxia (12H/24R) were designed to treat CTX-TNA2 cells to harvest exosomes (Figure 1B). Western blot assay showed that compared with the control, Aix and CD 63 levels were elevated in the 2H/6R exosomes and 12H/24R exosomes (Figure 1C). TEM analysis showed that control exosomes, 2H/6R exosomes and 12H/24R exosomes all exhibited a spherical morphology with an intact lipid bilayer membrane (Figure 1D). Furthermore, the NTA assay showed that the sizes of 2H/6R exosomes and 12H/24R exosomes were slightly larger but not statistically different from the control exosomes, ranging from 89.8 ± 9.9 nm in the control exosomes to 99.4 ± 12.7 mm in 2H/6R exosomes and 105.1 ± 22.3 nm in the 12H/24R exosomes (n = 3/group) (Table 1). At the same time, the concentrations of exosomes particles were not statistically different among 3 exosomes (Table 1). Furthermore, no statistical differences were observed in zeta potentials of 3 exosomes (Table 1).

To delineate the effects of 2H/6R exosomes and 12H/24R exosomes, a hemin-induced neurotoxicity model was established in primary cultured cortical neurons. The SRB assay showed that 16-h incubation of hemin (30 μM) increased neuronal death (Figure 2A). Co-incubation with 2H/6R exosomes concentration-dependently attenuated hemin-induced neuronal death. In contrast, 12H/24R exosomes did not significantly alter hemin-induced neuronal death (Figure 2A). At the same time, Western blot assay showed that hemin reduced GAP43 levels (a biomarker of neurite outgrowth). Co-incubation of 2H/6R exosomes, but not 12H/24R exosomes, attenuated hemin-induced reduction in GAP 43 levels, indicating that 2H/6R exosomes may attenuate hemin-induced reduction of neurite outgrowth (Figures 2B,C). Moreover, the immunofluorescent staining data showed focal bead-like swellings, neuritic beading and discontinuities of neurites in hemin-treated primary cultured cortical neurons after 1-h incubation of hemin, indicating hemin-induced impairments in neurite outgrowth (Figure 3). Co-incubation of 2H/6R exosomes significantly attenuated hemin-induced neurite impairment (Figure 3). At the same time, intensive PKH26 fluorescence (a fluorescent dye specific for exosomes) was detected in the cytosol of primary cultured cortical neurons treated with hemin plus 2H/6R exosomes, suggesting that exosomes were successfully endocytosed by primary cultured cortical neurons.

To further investigate the differential effects of 2H/6R exosomes and 12H/24R exosomes, iNOS and COX-2 (two proinflammatory biomarkers) were measured in the primary cultured cortical neurons. Western blot assay demonstrated hemin (30 μM)-induced neuroinflammation by increasing iNOS (Figures 4A,B) and COX-2 (Figures 4C,D) in hemin-treated primary cultured cortical neurons. Co-incubation with 2H/6R exosomes significantly prevented hemin-induced elevations in iNOS (Figure 4A) and COX-2 (Figure 4C). In contrast, 12H/24R exosomes further increased hemin-elevated iNOS (Figure 4B) and COX-2 (Figure 4D). These data indicated opposite effects of 2H/6R exosomes and 12H/24R exosomes on hemin-induced neuroinflammation. At the same time, we investigated the effects of exosomes on HO-1 expression (an enzyme catalyzing hemin). Western blot assay showed that co-incubation of H/6R exosomes but not 12H/24R significantly attenuated hemin-induced elevation in HO-1 expression (Figure 5), suggesting that 2H/6R exosomes can reduce hemin-induced HO-1 expression.

The neurotoxic mechanisms underlying hemin-induced neuronal death were investigated by measuring glutathione hydroperoxidase 4 (GPX4, a biomarker of ferroptosis) and active caspase 3 (a biomarker of apoptosis). Western blot assay demonstrated that hemin (30 μM) reduced GPX4 (Figures 6A,B) and increased active caspase 3 (Figures 6C,D). Co-incubation with 2H/6R exosomes attenuated hemin-induced reduction in GPX4 (Figure 6A) and elevation in active caspase 3 (Figure 6C). In contrast, 12H/24R exosomes did not alter hemin-induced reduction in GPX4 (Figure 6B) and elevation in active caspase 3 (Figure 6D), indicating indicate that 2H/6R exosomes appear to reduce hemin-induced ferroptosis and apoptosis.

So far, we have demonstrated differential effects of 2H/6R exosomes and 12H/24R exosomes on hemin-induced neurotoxicity; 2H/6R exosomes, but not 12H/24R exosomes, possessed a neuroprotective activity. The cargos, especially miRNA in both exosomes were analyzed using RNA Sequencing Analysis. We found significant changes in the miRNA expression profiles, including the expression levels of let-7c-5p, miR-191a-5p, and miR-709 in 2H/6R exosomes (Figure 7).