The transcriptome characteristics of the postmortem parietal cortex were compared between individuals who died from sepsis and non-septic critically ill individuals in the GEO database GSE221379. The complete RNA datasets from the selected samples were downloaded for subsequent analysis; however, information regarding the age, etiology, medication treatment, and prognosis of each sample was not available. The raw data was processed using the “Deseq2” R package to identify the differentially expressed genes (DEGs) in the human brain between septic and non-septic critically ill patients.log2 (fold change). The “ggplot2” R package was utilized to create a volcano plot, which illustrates the results. Furthermore, the “pheatmap” R package was employed to perform a bidirectional hierarchical clustering analysis of the collected DEGs (Xiantao tool, www.xiantao.com) (12).

To annotate the potential functions of the EP-DEGs, DAVID Bioinformatics Resources v6.8 was employed for Gene Ontology (GO) mapping and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. The biological processes (BP), cellular components (CC), and molecular functions (MF) associated with each DEG were assessed. Additionally, all EP-DEGs were submitted to the DAVID website (version 6.8).The enriched GO terms were visualized using the GOplot package in the R programming language. Additionally, the GO plot package was utilized to analyze the overlap of genes involved in the identified GO terms. Furthermore, the pathways in which the DEGs were involved were predicted and mapped using the KEGG database in conjunction with DAVID (Xiantao tool, www.xiantao.com) (13–15).

The STRING database’s multi-protein online tool was used to predict PPI networks between two genomes, and Cytoscape 3.6.1 was employed to visualize these networks. To explore the biological processes associated with the gene cluster, the genes in the cluster with the highest score were imported into STRING for further analysis of the protein interaction network (Xiantao tool, www.xiantao.com) (16).

Throughout the MCAO procedure, the head temperature was maintained at 36° with the assistance of a heat lamp. After 60 minutes of MCAO, the intraluminal suture was removed to achieve brain reperfusion (17). “In brief, mice were anesthetized using a mixture of isoflurane, oxygen, and air. The right common carotid artery was carefully isolated and ligated with a medical suture. Additionally, the right internal carotid artery and external carotid artery were separated. A ligation was performed at a distance of 1 cm from the internal and external cervical vascular branches, specifically at the site of the external carotid artery. A small incision was made proximal to the heart, near the ligated area of the artery. Through this incision, a nylon monofilament coated with silicone and heat-blunted at its tip was carefully inserted into the right internal carotid artery. Laser Doppler flowmetry was utilized to continuously monitor the blood flow in the right brain artery. The filament was secured when its tip reached the origin of the middle cerebral artery, which was confirmed by a recognizable decrease (~30% of baseline) in cortical blood flow. After 1 hour of occlusion, the filament was removed to allow for reperfusion, and the blood flow was restored to more than 70% of the baseline level. For sham-operated mice, the same surgical procedures were performed, with the exception of filament insertion. Throughout the aforementioned procedures, the mice were maintained on a warming pad. A total of 66 mice were successfully induced with MCAO and included in further experiments. However, 15 mice died within 24 hours following reperfusion and were subsequently excluded from the study”. In order to investigate the role of Tregs in MCAO, the experimental groups were divided into four groups: Control group, MCAO group, Tregs Depletion +MCAO (DT+MCAO) group, and DT group.

Brain tissue sections (20 μm) were hydrated and deparaffinized, and antigen retrieval was performed before blocking with 10% bovine serum albumin for one hour at room temperature. Primary antibodies, including anti-2 (1:100, AiFang Biological, AF14722), anti-IL-10 (1:100, AiFang Biological, AF06580), anti-IL-2(1:100, AiFang Biological, AF02481), anti-IL-6 (1:100, AiFang Biological, AF06790), anti-FSP1 (1:100, AiFang Biological, AF14902), and anti-GPX4 (1:100, AiFang Biological, AF06580), were used to incubate the sections overnight at 4°. In the case of astrocyte cells, the cells were initially washed three times in PBS, fixed in 4% paraformaldehyde for 30 minutes, rinsed again in cold PBS, and then permeabilized for 30 minutes with 0.3% Triton X-100. Following the permeabilization step, the cells were blocked with 10% BSA for 1 hour before incubating with relevant primary and secondary antibodies. Observation under a fluorescence microscope (FV3000, Olympus) followed (18).

To deplete regulatory T cells (Treg cells), diphtheria toxin (DT) was injected intraperitoneally (ip). At a dosage of 0.05 mg/g of body weight, an injection of the antibody CD25 was administered three days before subjecting the rats to a 60-minute transient middle cerebral artery occlusion (tMCAO). To ensure continuous depletion of Treg cells, this injection was repeated every three days until 14 days after the occurrence of the stroke., as previously described. Delayed injection of DT began 5 days after the tMCAO procedure, and then repeated every 3 days until the animals were sacrificed. To deplete microglia and macrophages, the mice (9-10 weeks old, weighing 25-30 g) were provided with PLX5622 in their diet (Research Diets) at a concentration of 1200 parts per million (PPM) or 1200 mg/kg of chow. This treatment was initiated 7 days before surgery and continued until the end of the experiment (19).

We employed a previously-described OGD/R procedure. The specific methodology used in this research study was based on the references provided below (20). In this study, astrocyte cultures were divided into four experimental groups: (1) the normal group; (2) the OGD/R group, where astrocytes were washed three times with PBS and cultured in DMEM (no glucose). The cells were then placed in an incubator with an atmosphere of 0.1% O2, 94.9% N2, and 5% CO2 (hypoxia) for 2 hours. After that, the cells were placed in DMEM and incubated in an atmosphere of 95% air and 5% CO2 for 48 hours; (3) the OGD/R+10μg/ml IL-10 group;,where astrocytes were washed with PBS three times and re-suspended in DMEM (no glucose) supplemented with IL-10 (10μg/mL) during OGD/R injury; (4) the OGD/R+10μg/ml IL-10 + 10μM GPX4-inhibitor group, where astrocytes were washed with PBS three times and re-suspended in DMEM (no glucose) supplemented with GPX4-inhibitor (RSL3 10 μM) during OGD/R injury.

ELISA was performed to validate the expression levels of IL-1β, NLRP3, FTL, and TNF-α in four groups. Tissue homogenate or cell supernatant were determined using commercial ELISA kits (IL-1β: Cat. #AF-02323M2 China; NLRP3: Cat. #AF-02936M2, China; TNF-α: Cat. #AF-02415M2, China; FTL: Cat. #JL31443, China.) according to the manufacturer’s instructions

Tregs were isolated from the spleen and lymph nodes of mice, as previously described. Tregs and glial cells were co-cultured using the following method. The culture medium was removed from the semi-confluent glial cell culture, and Treg cells (1.0×10⁶/well) were then added. The cells were cultured in the presence of 2-mercaptoethanol (×1000 dilution), anti-CD3 antibody (2 μg/mL), anti-CD28 antibody (2 μg/mL), for 3–6 days. Half of the medium was changed every 2 days, and anti-CD3 and anti-CD28 antibodies were added only on the first day. Various cytokines were added at the indicated concentrations described below. To prepare the conditioned medium (CM), semi-confluent astrocytes (2×10⁵/well) in 24-well dishes were washed three times using serum-free medium, cultured in DEME medium containing 10% FCS supplemented with 4 mM L-glutamine for 24 h, collected in sterile tubes, and stored at −25°C until further use. Subsequently, 0.5 mL of CM was added to the 0.5 mL Treg culture in 24-well dishes.

Statistical analyses were conducted using GraphPad Prism 9.0.The Student’s t-test and χ² test were used to analyze the normally distributed data, assessing the mean difference between two groups and testing the association between categorical variables, respectively. Non-normally distributed data was analyzed using the Kruskal-Wallis test, a nonparametric test that evaluates differences between three or more groups without assumptions of normality or homogeneity of variance. A significance threshold of p<0.05 was set. Normality and homogeneity of variance were tested, and appropriate adjustments for multiple comparisons were made using Bonferroni correction (21).

The preprocessing of the sample information primarily involves filtering and quality control procedures. The filtering process includes removing low-quality samples and eliminating outliers. Common metrics, such as sequencing depth and ribosomal RNA contamination, are assessed to perform quality control of the samples. Additionally, sample normalization is conducted to ensure the comparability of gene expression levels. The identification of DEGs is achieved through statistical analysis methods, such as the t-test or the DESeq2 package in R.Genes exhibiting significant differences in expression levels, based on specific criteria such as fold change and statistical significance, are classified as differentially expressed genes (DEGs).Subsequent downstream analysis involves further examination and prioritization of these DEGs. This identification process helps to identify key genes that are relevant to the research question of interest. The identified DEGs serve as potential targets for subsequent investigation and functional annotation ( Figures 1A–D ).

In order to gain a thorough understanding of the role of the 31 differentially expressed genes (DEGs) identified in the brain tissue dataset, functional and pathway enrichment analyses were conducted using DAVID. As depicted in Figure 2A , these genes exhibit significant enrichment in biological processes such as zinc ion binding, calcium-dependent protein binding, and cytokine activity. Additionally, they are found to be primarily enriched in the inflammatory response, innate immune response, and neutrophil chemotaxis. The enrichment analysis also revealed that all of these genes are enriched in the extracellular space. Furthermore, they are enriched in various KEGG pathways, using the cnetplot function in the ClusterProfiler package. Notably, IL6 and CXCL3,were found to be enriched in at least two BPs, CCs, and MFs ( Figures 2A–G ).

Through the application of LogFC in conjunction with KEGG analysis, it was discovered that IL-6 remained involved in the aforementioned processes. To investigate the interactions among the 26 DEGs, we constructed a protein-protein interaction (PPI) network using the STRING database. The PPI network was visualized using the Cytoscape application ( Figure 3A ).The PPI network consisted of 26 nodes, representing molecular pathways and processes associated with the 26 core DEGs. Working modules were developed using the MCODE plugin for Cytoscape. The results revealed a module consisting of 9 genes ( Figures 3B–E ).

To evaluate the potential impact of Tregs in MCAO mice, we conducted an immunofluorescence assay to measure the protein expression of two activation of astrocytes markers (IL-2, IL-6) in the brain tissues of each group. Immunofluorescence examination revealed a significant induction in the expressions of IL-2 and IL-6 in the brain of MCAO mice when compared to the control group, which was then promoted in the MCAO +DT group, as demonstrated in Figures 4A–Z . We also examined the expression of inflammatory factors IL-1β, NLRP3, and TNF-α and found that DT could reduce the elevated levels of inflammatory factors caused by MCAO, please refer to supple Figure 4 for the results.

In MCAO mice, there was a significant increase in the production of LC3 in the astrocytes, as illustrated in Figures 5A–J . The electron microscope results showed that the mitochondria in the experimental group exhibited swelling, disappearance of mitochondrial ridges ( Figures 5K, L ), and an increase in autophagosomes ( Figures 5O, P ). However, after applying DT intervention, the mitochondrial swelling could be alleviated and mitochondrial autophagy could be inhibited ( Figures 5M, N, Q, R ).

The MCAO group had significantly increased fluorescence intensity levels of FSP1 and decreased in GXP4 compared to the control group ( Figures 6A–L ). The MCAO reduction in fluorescence intensity levels of GXP4 and induction in FSP1 was partially attenuated by DT treatment ( Figures 6M–Z ). We also examined the expression FTL and found that DT could reduce the elevated levels of FTL by MCAO, please refer to supple Figure 6 for the results.

In vitro models demonstrated that OGD induces astrocytes activation. Following OGD treatment, the fluorescence intensity levels of IL-2 ( Figures 7E–H ) and IL-6 ( Figures 7Q–T ) significantly increased compared to the baseline levels ( Figures 7A–D, M–P ). The fluorescence intensity levels were comparable in the control group and the OGD+IL-10 group ( Figures 7I–L, U–Z ).We also examined the expression of inflammatory factors IL-1β, NLRP3, and TNF-α and found that IL-10 could reduce the elevated levels of inflammatory factors caused by OGD, please refer to supple Figure 7 for the results.

In vitro experiments showed that co-culturing tregs with astrocytes ( Figures 8A–H ) resulted in alleviation of neuroinflammation induced by OGD. Additionally, OGD resulted in a significant decrease in GPX4 protein expression levels ( Figure 8J ), whereas DT treatment increased GPX4 expression levels in OGD-treated pipes ( Figure 8K ). Moreover, the addition of IL-10Ra, an IL-10 inhibitor, weakened the effect of OGD ( Figure 8L ), while clearing tregs in vitro weakened the effect of IL-10Ra ( Figures 8M–Q ).