Microarray Dataset Analysis
Initially, relevant microarray data associated with ischemic stroke were retrieved from the GEO database (https://www.ncbi.nlm.nih.gov/geo/), using the keywords "ischemic stroke," and limiting the study type to "Expression profiling by array" for the species "Homo sapiens." Subsequently, inclusion and exclusion criteria were established to filter the candidate datasets. The selected microarray datasets were required to meet the following conditions: (I) inclusion of both ischemic stroke patients and a control group of healthy individuals; (II) samples derived from brain tissue or peripheral blood; (III) a minimum total sample size of 20 cases per dataset; (IV) availability and usability of the dataset's annotation platform information for accurate and smooth conversion to corresponding gene names. Datasets needing to meet these criteria were excluded from the study.
Transcriptomic Data
Following our search methodology, three datasets were chosen for subsequent analysis: GSE16561, GSE37587, and GSE162955, with sample types and control conditions detailed in Table 1. GSE16561 and GSE37587 involve peripheral blood transcriptomic sequencing data related to ischemic brain injury, whereas GSE162955 contains transcriptomic sequencing data of brain necrotic areas and normal brain tissue from the contralateral side of patients who died due to ischemic brain injury (35).
Table 1
Candidate datasets for standard transcriptomic sequencing.
GSE ID | Sample Type | Comparison and Sample Size | Platform ID |
GSE16561 | Peripheral Blood | Ischemic Brain Injury (39) vs. Normal (24) | GPL6883 |
GSE37587 | Peripheral Blood | Ischemic Brain Injury 0–24 hours (34) vs. 24–48 hours (34) | GPL6883 |
GSE162955 | Brain Tissue | Ischemic Brain Injury Necrotic Tissue (6) vs. Contralateral Healthy Brain Tissue (6) | GPL17586 |
Given the significance of inflammatory responses following ischemic brain injury, our initial focus was on the peripheral blood transcriptome of ischemic stroke patients compared to a healthy control group. Subsequently, to discern the time-dependent nature of ischemic brain injury treatment, transcriptomes from acute (0–24 hours) and subacute (24–48 hours) phases were compared to elucidate characteristic changes in the early stages of ischemic brain injury. This comparison aims to guide clinical diagnosis and treatment. Thirdly, to gain a deeper understanding of the inflammatory response in brain tissue post-ischemic injury, we included transcriptomic sequencing data from both necrotic brain areas and healthy regions of patients who succumbed to ischemic brain injury. This approach was intended to clarify the transcriptomic characteristics exhibited by brain tissue in lethal ischemic brain injury.
The comprehensive analysis of these three datasets offers insights into the transcriptomic changes during the onset and progression of ischemic brain injury, particularly regarding immune-inflammatory responses. This contributes to a better understanding the inflammation associated with ischemic brain injury.
Gene Differential Expression Analysis
After the search and selection process was completed, the gene expression matrices and corresponding annotation platform files for the target datasets were downloaded using the GEOquery package in R software (Version R 4.1.0, https://cran.r-project.org/). The 'exprs' function was utilized to extract gene expression matrices and the 'pData' function for extracting clinical features of samples matching the expression matrices. Following acquiring the target datasets' expression matrices, the limma package in R was used for normalization and log2 transformation. Differential gene expression analysis between samples from ischemic brain injury patients and normal healthy controls was then conducted using the limma package. Genes meeting the criteria of |log fold change (FC)| > 1 and P < 0.05 were identified as differentially expressed genes (DEGs) (36).
Immune Infiltration Analysis
CIBERSORT, an RNA deconvolution algorithm using support vector regression, estimates the abundance of immune cells. The leukocyte signature matrix (LM22), containing 547 genes to differentiate 22 human hematopoietic cell phenotypes, including 7 T cell types, naive and memory B cells, plasma cells, NK cells, and myeloid subgroups, is commonly used. This matrix, highly sensitive and specific to human immune cell phenotypes, determines the diversity and overall profile of tumor-infiltrating immune cells. By applying the CIBERSORT method to ischemic brain injury-related datasets from the GEO database, we compared the immune cell types and levels of infiltration among regular patients and ischemic brain injury patients across different periods. The results are presented in stacked bar graphs and violin plots (P < 0.05 indicates statistical significance) (37).
PPI Network Construction
Data from monocyte Clusters 1, 5, and 7 were extracted for differential gene expression analysis. Comparisons were made between the sham group (control) and Days 2 (D02) and 14 (D14) post-ischemic brain injury, intersecting the differential expressed genes from the three groups. The STRING database (https://string-db.org/) was used for protein-protein interaction (PPI) analysis, and the network was visualized and optimized using Cytoscape 3.10.0 software (38).
Single-Cell Sequencing Data Acquisition in Ischemic Brain Injury Mouse Model
Single-cell sequencing data from a mouse model of ischemic brain injury were retrieved and downloaded from the GEO database. The retrieval and processing of the data are outlined as follows: Firstly, using "ischemic stroke" as a keyword and limiting the study species to "Mus musculus," ischemic brain injury-related single-cell sequencing data were searched in the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The candidate mouse model single-cell sequencing datasets included the sham group (control), peripheral blood, and brain tissue cells from Day 2 and Day 14 post-injury. Expression matrices were generated using the official 10 × Genomics software, Cellranger. The primary criteria for cell selection were: the number of genes expressed (nFeature) in a single cell should be less than 200, and nFeature_RNA should be greater than 5000. Cells with mitochondrial gene proportion (percent. mt) exceeding 20% were filtered out. Subsequently, data quality was inferred using analysis methods based on unique molecular identifiers (UMIs) and gene correlations (39).
Construction of Single-Cell Sequencing Analysis Subjects
Upon identifying the target datasets, we downloaded the corresponding single-cell sequencing matrix data and constructed single-cell analysis matrix objects using the Seurat package in R software (Version R 4.1.0, https://cran.r-project.org/). Focusing on cell types defined by the original authors of the datasets, we extracted T cells, neutrophils, and monocytes for subsequent analysis. Considering the time-dependent nature of inflammatory responses associated with cerebral infarction and the dynamic changes in immune cells, we employed trajectory analysis to explore the trends in T cells, neutrophils, and monocytes across different tissues post-brain injury. This trajectory analysis used the Monocle package in R (39).
KEGG Functional Enrichment Analysis
Using cell states derived from trajectory analysis (i.e., dynamic changes in immune cells) as cell grouping markers, we examined the biological processes and molecular functions exhibited by immune cells at different stages. In this instance, we did not opt for analysis based on KEGG signaling pathways. On the one hand, the analysis focused solely on specific cell subgroups, emphasizing the dynamic changes of these cells across different stages. Therefore, KEGG signaling pathways might not accurately reflect the functional characteristics of these subgroups over time. On the other hand, the differential genes between clusters within this subgroup were not pronounced, limiting the ability of KEGG signaling pathways to represent the dynamic changes of subgroup cells over time. Hence, the enrichment analysis primarily showcased the current cell subgroup's biological processes and molecular functions, presented in the form of bar graphs (40).
Animal Study Ethical Statement
This experimental protocol and animal usage plan adhered to the regulations and guidelines of our institution's Animal Ethics Committee and received the necessary approvals. All experiments were conducted to minimize animal suffering and discomfort and to reduce the number of animals used as much as possible. The animals' care, handling, and experimental procedures strictly followed internationally recognized animal welfare standards and complied with the ARRIVE guidelines (Kilkenny, 2010; McGrath, 2015). The animal experiments adhered to national laws, regulations, and standards concerning experimental animals, including the "Regulations for the Administration of Affairs Concerning Experimental Animals" and the "Guidelines for Ethical Review of Animal Welfare." Proper care was provided to all animals, and at the end of the experiments, all animals were euthanized using cervical dislocation following anesthesia with 150 mg/kg isoflurane (PHR2874, Sigma-Aldrich, USA).
Isolation of Bone Marrow-Derived Neutrophils
Three male C57BL/6 mice (4–5 weeks old, 18–22 g) were procured from Hunan SJA Laboratory Animal Co., Ltd. (Changsha, China, https://www.lascn.net/SupplyDemand/Site/Index.aspx?id=6#). They were acclimatized for one week in a constant humidity (45%-50%) and temperature (25–27℃) SPF-grade animal room with a 12-hour light/dark cycle. The mice were fasted for 12 hours before administration but otherwise had ad libitum access to food and water. Our institution's Animal Ethics Committee approved all animal experimental procedures. For neutrophil isolation, mouse bone marrow was rinsed and filtered through a 200-mesh screen, then centrifuged at 400 g for 5 minutes and resuspended in 1 ml HBSS solution (14065056, Gibco, USA). Percoll density gradients (76%, 62%, 52%; P8370, Shanghai Yuning Bio-Tech Co., Ltd., Shanghai, China) were layered in a 15 ml centrifuge tube, overlaid with the cell suspension, and centrifuged at 1060 g for 30 minutes. Cells between the 76%-62% interface were collected into a new tube, washed and resuspended in 2 ml HBSS solution, then layered over 2 ml lymphocyte separation medium and centrifuged at 1600 g for 30 minutes. The intermediate cell layer was collected to obtain a neutrophil suspension, which was then washed with RPMI 1640 medium (11875119, Gibco, USA). The isolated neutrophils were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (10099141C, Gibco, USA) .
In Vitro Cell Culture
Mouse brain microvascular endothelial cells bEnd.3 (CL-0598) were obtained from Wuhan Procell Life Science & Technology Co., Ltd. (Hubei, China). All cells were tested for mycoplasma contamination before experiments. Cells were cultured in DMEM medium (PM150210, Wuhan Procell Life Science & Technology Co., Ltd., Hubei, China) supplemented with 10% FBS (164210, Wuhan Procell Life Science & Technology Co., Ltd., Hubei, China), 100 U/ml penicillin, and 100 U/ml streptomycin (PB180120, Wuhan Procell Life Science & Technology Co., Ltd., Hubei, China). Cultures were maintained at 37℃ in a 5% CO2 incubator, with media changed every three days. Upon reaching 80% confluency, cells were digested with trypsin (P4201, Beyotime, Beijing, China) for one minute at 37℃ and subsequently passaged.
Lentiviral Infection
A lentiviral interference vector, pSIH1-H1-copGFP (shRNA, catalog number: SI501A-1), was purchased from System Biosciences, USA, to construct a lentivirus-based Hspa8 gene interference vector. Lentiviral particles were packaged in HEK-293T cells (iCell-h237, iCell Bioscience Inc., Shanghai, China) using a lentiviral packaging kit (A35684CN, Invitrogen, USA), with a titer of 1×108 TU/ml. Target cells were incubated with the viral mixture at 40% concentration for 8 hours, followed by a medium exchange with DMEM containing 10% FBS. Cells were selected with 5µg/mL puromycin (A1113803, Thermo Fisher Scientific, China) for 4 weeks before experiments. The lentiviral sequences are listed in Table S1, with silencing efficiency validated in neutrophils (41).
RT-qPCR
Total RNA was extracted from tissues using Trizol (16096020, Invitrogen, USA). Reverse transcription was carried out to obtain cDNA for mRNA detection using a reverse transcription kit (RR047A, Takara, Japan). Quantitative real-time PCR (qRT-PCR) was performed following the TaqMan Gene Expression Assays protocol (Applied Biosystems, Foster City, CA, USA). GAPDH was used as the internal reference gene. The PCR program was as follows: initial denaturation at 95°C for 10 minutes, followed by 35 cycles of denaturation at 95°C for 15 seconds, annealing at 60°C for 30 seconds, and extension at 72°C for 45 seconds. All qRT-PCR reactions were performed in triplicates. Primer sequences are provided in Table S2, and the relative expression of the target gene in the experimental group compared to the control group was calculated using the 2−ΔΔCt method, where ΔΔCt = ΔCt experimental group - ΔCt control group, and ΔCt = Ct target gene - Ct internal reference gene. Ct represents the cycle number at which the real-time fluorescence intensity reaches the set threshold, indicating logarithmic growth of amplification at this point (42). The experiment was repeated three times for validation.
Western Blot
Cell pellets were resuspended in 200 µL RIPA lysis buffer (P0013B, Beyotime, Beijing, China) in 1.5 mL Eppendorf tubes. After homogenization and incubation on ice for 10 minutes, the samples were centrifuged at 3000 g and 4℃ for 10 minutes. The supernatant was transferred to a pre-cooled Eppendorf tube for protein quantification using the BCA protein assay kit (A53226, Thermo Fisher Scientific, Rockford, IL, USA). Post-BCA protein concentration measurement, samples were equalized with loading buffer and boiled (99℃, 10 min) before storage at -20℃. Gel concentration was chosen based on the target protein molecular weight, with a loading volume of 20 µL per well. PVDF membranes were activated with methanol and transferred using a BioRad electrophoresis system, with voltage or current fixed according to protein molecular weight. Membranes were blocked with 5% BSA at room temperature for 1 hour, followed by overnight incubation at 4℃ with primary antibodies against Hspa8 (ab51052, 1:1000, Abcam, UK) and GAPDH (ab181602, 1:1000, Abcam, UK). The next day, membranes were brought to room temperature on a shaker for 1 hour and washed three times with TBST (containing 0.1% Tween 20). HRP-conjugated goat anti-rabbit IgG (1:20000, ab205718, Abcam, UK) was applied for 1 hour at room temperature, followed by three TBST washes. The detection solution (P0019, Beyotime, Beijing, China) was mixed in a 1:1 ratio and applied uniformly to the membrane. Protein bands were visualized using the gel imaging system and quantified using AlphaView SA software (Version 3.4.0), with protein expression represented as the ratio of target protein band intensity to GAPDH band intensity (43). Each experiment was replicated three times.
CCK-8 Assay
Cells in the logarithmic growth phase were seeded at a density of 5×104 cells per well in a 96-well plate and cultured overnight. Cell proliferation viability was measured using the CCK-8 kit (E606335, Sangon Biotech). After 24 hours of culture, 10 µL of CCK-8 solution was added to each well. The plates were then incubated at 37℃ in a humidified incubator for 1 hour before measuring absorbance at 450 nm using an Epoch microplate spectrophotometer (Bio-Tek, USA) (44). Each group had three replicate wells, and each experiment was repeated thrice.
Transwell Assay for In Vitro Blood-Brain Barrier Model
BEnd.3 cells were seeded in the upper chamber of a Transwell to form a monolayer, serving as a model for migration studies and representing the blood-brain barrier. Neutrophils (1 × 106) were added to the upper chamber, with FBS-free medium in the lower chamber. After 3 hours, the lower chamber was stained with DAPI. The permeability rate was calculated as the percentage of stained area relative to the total field of view.
Detection of ROS
Cells from each group (1×104) were seeded into 96-well plates. After 24 hours, cells were washed with PBS and incubated with PBS containing DCFH-DA (10 µM, for total ROS detection) for 45 minutes. Post-incubation, cells were washed and examined using a fluorescence microscope (AF6000, Leica, Germany) (44).
Construction of Ischemic Brain Injury Mouse Model
A total of 32 male C57BL/6 mice (4–5 weeks old, 18–22 g) were divided randomly into four groups, obtained from Hunan Slick Jinda Laboratory Animal Co., Ltd. Mice were acclimated for one week in an SPF facility with constant humidity (45%-50%) and temperature (25–27℃), on a 12-hour light/dark cycle. Food was withheld for 12 hours before administration, with free access to water. Our Institutional Animal Care and Use Committee approved all animal experiments. The I/R injury model was established using the MCAO method. Mice were anesthetized with 3% isoflurane (792632, Sigma Aldrich, St. Louis, MO, USA) and placed on a stainless steel surgical table. After making a midline neck incision, the left external carotid artery (ECA) and left pterygopalatine artery were isolated and ligated. The internal carotid artery (ICA) was clamped at the peripheral bifurcation with the pterygopalatine artery. A blunted 6 − 0 nylon suture (0.2–0.22 mm tip diameter) was advanced through the ICA to the bifurcation of the ICA and ECA. The nylon suture was tied to the ECA with a 6 − 0 suture. After cutting the ECA and rotating the nylon suture, it was advanced until resistance was felt, positioning the tip at 6 mm and 9 mm from the bifurcation of the pterygopalatine artery and the ICA-ECA, respectively. The nylon suture was removed after 1 hour to restore blood flow, with the entire procedure maintained at 37.0 ± 0.5℃. Mice were euthanized 24 hours post-reperfusion. Sham-operated mice underwent artery isolation without ligation (45, 46, 47, 48).
Animal grouping mice were divided into four groups: Sham (8 mice), MCAO (8 mice), MCAO + sh-NC (8 mice), and MCAO + sh-Hspa8 (8 mice).
Intracerebroventricular Injection of Lentivirus
Lentiviruses used in the study were purchased from Shanghai Ji Kai Gene Chemical Technology Co., Ltd. One hour post-ischemic brain injury model establishment, mice were anesthetized with 2.5% isoflurane and fixed on a Stoelting animal stereotaxic instrument. Using the bregma as a reference point, injections were made 1.5 mm lateral (right) and 0.9 mm posterior, at a depth of 3.5 mm into the right ventricle using a 10 µL Hamilton syringe (Microliter No.701; Hamilton Company, Switzerland) at a rate of 0.5 µL/min. The needle was left in place for 5 minutes post-injection to prevent backflow, then slowly withdrawn. Bone wax was used to seal the bone hole, followed by scalp suturing. Mice were kept warm on an electric heating pad until recovery from anesthesia (45, 49).
Neurological Function Scoring
Neurological function was assessed 24 hours post-surgery using the modified Garcia neurological behavior scoring system, which includes six parts: spontaneous activity, spontaneous limb movement, forelimb stretching activity, climbing response, bilateral trunk touch response, and vibrissae response. Scores from each part were summed to reflect the level of neurological function, ranging from a minimum of 3 to a maximum of 18, with lower scores indicating more severe neurological deficits. Scoring was conducted independently by two observers unrelated to the experiment, using the average of their scores.
The modified Bederson scoring system was used: 0, no deficit; 1, forelimb flexion; 2, unilateral circling when tail lifted; 3, spontaneous unilateral circling; 4, longitudinal rolling when tail lifted; 5, spontaneous longitudinal rolling. Scores for each mouse were determined by averaging 5–10 tests (each lasting 15 seconds). Higher scores indicate more severe neurological deficits, with scoring conducted independently by two observers unrelated to the experiment, using their average scores (50, 51, 52, 53).
Brain Tissue TTC Staining
Post-euthanasia, mouse brain tissues were collected and sliced into 3 mm sections, then incubated in a 2% 2,3,5-triphenyl tetrazolium chloride (TTC) solution at 37°C in the dark for 30 minutes. Stained tissue sections were photographed using a microscope (Zeiss, Germany), with red indicating normal brain tissue and white indicating infarcted tissue. Infarct areas were analyzed using ImageJ software, calculated as the contralateral area minus the ipsilateral non-infarcted area. Total infarct volume was calculated by multiplying each section's infarct areas and multiplying by the section thickness (53, 54).
Immunofluorescence Staining
Brain tissue cryosections from each group were air-dried at room temperature, fixed in 4% paraformaldehyde for 10 minutes, washed with PBS, and incubated with 0.2% TritonX-100 in PBS for about 10 minutes. After another PBS wash, sections were blocked with 5% goat serum at room temperature for 30 minutes, followed by overnight incubation at 4℃ with primary antibody against CD11b (ab8878, 1:100, Abcam, UK). The next day, sections were washed with PBS and incubated with fluorescent secondary antibody (1:300) at room temperature in the dark for 1 hour. After a final PBS wash, sections were mounted with a DAPI-containing mounting medium. Observations were made under a fluorescence microscope (AF6000, Leica, Germany). Negative controls were incubated with PBS instead of the primary antibody (55).
Statistical Analysis
In this study, microarray data retrieval was first conducted using the GEO database, and candidate datasets were selected based on predefined inclusion and exclusion criteria. Upon obtaining the microarray data, we used the GEOquery package in R software to download the gene expression matrix and related annotation platform files. The exprs function was utilized to extract the gene expression matrix, and the pData function to obtain clinical features matching the expression matrix. The limma package was then employed for normalization and log2 transformation of the candidate datasets. Differential gene expression analysis between brain infarction patients and normal healthy controls was performed using the limma package, with genes identified as differentially expressed if they met the |log fold change (FC)| > 1 and P < 0.05 criteria. Additionally, the CIBERSORT method was used to assess RNA reverse convolution, employing support vector regression to estimate the abundance of immune cells. We first retrieved and downloaded relevant data from the GEO database for single-cell sequencing data. After selection and quality control, we constructed single-cell analysis matrix objects in R using the Seurat package. Trajectory analysis was performed using the monocle R package to understand the dynamics of immune cells. Finally, we conducted a KEGG functional enrichment analysis, focusing on cell subgroups' biological processes and molecular functions at different stages, and presented the results graphically (56).