Unveiling Pyroptosis-Related Hub Genes in Ischemic Stroke Provides Insights for Enhanced Risk Assessment

doi:10.21203/rs.3.rs-3865292/v1

Download PDF

Research Article

Unveiling Pyroptosis-Related Hub Genes in Ischemic Stroke Provides Insights for Enhanced Risk Assessment

https://doi.org/10.21203/rs.3.rs-3865292/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Stroke is the second-leading global cause of death. The immune storm triggered by ischemia-reperfusion injury after stroke is a crucial damaging factor. This study analyzed the expression of key pyroptosis genes in stroke and their correlation with immune infiltration.

Methods: Middle Cerebral Artery Occlusion datasets were obtained and pyroptosis-related genes were identified. Differential expression and functional analyses of pyroptosis-related genes were performed. Differences in functional enrichment between high-risk and low-risk groups were determined. After selecting pyroptosis-related genes with differential expression, a MCAO diagnostic model was constructed and validated. High and low-risk MCAO groups were constructed for expression and immune cell correlation analysis with pyroptosis-related hub genes. A regulatory network between pyroptosis-related hub genes and miRNA was built, and protein domains were predicted. The expression of key pyroptosis genes was validated in the MCAO rat model.

Results: Twenty-five pyroptosis genes showed differential expression, including four hub genes (WISP2, MELK, SDF2L1, and AURKB). The high- and low-risk groups showed significant expression differences for WISP2, MELK and SDF2L1. In immune infiltration analysis, 12 immune cells exhibited expression differences in MCAO samples. Further analysis demonstrated significant positive correlations between the pyroptosis-related hub gene SDF2L1 and immune cell-activated dendritic cells in the high-risk group and immune cell natural killer cells in the low-risk group.

Conclusion: This study identified four pyroptosis-related hub genes, with elevated WISP2, MELK, and SDF2L1 expression closely associated with the high-risk group. Analysis of inflammatory cell types in immune infiltration provides a theoretical basis for predicting ischemic stroke risk levels and treatment.

Stroke

Ischemia-reperfusion injury

Pyroptosis

Immune infiltration

Ischemic stroke (IS), characterized by rapid clinical signs of brain damage, stands out as a significant contributor to mortality and disability, notably prominent in China where the incidence surpasses 39.3%, particularly affecting males, with a prevalence exceeding 41% [1, 2]. Thrombolytic therapy is the main clinical treatment strategy for IS, aiming to restore blood supply to the ischemic area and protect neuron vitality and function [3]. Restoration of middle cerebral artery flow is an absolute requirement for cell survival in the ischemic zone. However, this restoration induces cell death through ischemia-reperfusion injury (IRI). IRI occurs due to interrupted cerebral blood flow or cerebral vasculature obstruction by a thrombus, depriving local brain tissues of oxygen and glucose, ultimately leading to cell death [4, 5]. IRI leads to severe damage to brain and nerve cells, complicating patient prognosis and causing functional impairments [6–8].

Pyroptosis is characterized by cell membrane pore formation, pro-inflammatory cytokine release, and cell lysis [9]. Recent studies indicate pyroptosis as a manifestation of the inflammatory response to IRI [6]. Understanding key genomic alterations in pyroptosis and its role in IRI is crucial for effective treatments. For instance, low-density lipoprotein receptor inhibits neuronal death after cerebral IRI by suppressing NLRP3-mediated neuronal pyroptosis [10]. Valproic acid, CHRFAM7A overexpression, and XBP-1 down-regulation also alleviate cerebral IRI by inhibiting the caspase-1-mediated canonical pathway of pyroptosis [11]. High-throughput analyses of genomic and transcriptomic data enhance our understanding of cellular dynamics for more effective clinical therapeutics. However, there is limited information on the transcriptomic dynamics of pyroptosis in IRI. Further exploration of pyroptosis's role in IRI and uncovering its molecular mechanisms could offer novel insights into IRI's pathogenesis.

In this study, we analyzed the gene expression dynamics of pyroptosis post-IRI from middle cerebral artery transcriptomes and unveiled the transcriptome signatures of pyroptosis and their associated regulatory mechanisms in IRI. Constructing the transcriptome signature of pyroptosis after IRI, a diagnostic model was developed for middle cerebral artery occlusion. Analysis of the immune infiltration underlying the IRI provides a theoretical basis for predicting ischemic stroke risk levels. Our data offer evidence for prognostic biomarkers of IRI and potential therapeutic targets for the middle cerebral artery.

Data Acquisition

We obtained Middle Cerebral Artery Occlusion (MCAO) related datasets (GSE97537, GSE61616, and GSE106680) from the GEO database (Table 1). Specifically, GSE97537 encompasses seven MCAO samples and five Control samples; GSE61616 includes five MCAO samples and five Control samples; GSE106680 comprises three MCAO samples and three Control samples. All MCAO and Control samples from these datasets were included in our study. A list of 338 Pyroptosis-Related Genes (PRGs) was compiled from the GeneCards [12] and Molecular Signatures Database (MSigDB) [13]. We employed the R package sva [14] to perform batch correction on GSE97537, GSE61616, and GSE106680, yielding the integrated GEO dataset (Combined Datasets). This combined dataset comprises 15 MCAO samples and 13 Control samples. Standardization, probe annotation, and normalization procedures were applied to the integrated GEO dataset using the R package limma [15].

Table 1

GEO Microarray Chip Information
	GSE97537	GSE61616	GSE106680
Platform	GPL1355	GPL1355	GPL15084
Species	Rattus norvegicus	Rattus norvegicus	Rattus norvegicus
Tissue	Brain	Brain	Brain
Samples in MCAO Group	7	5	3
Samples in Control Group	5	5	3
Reference	NA	PMID: 26305988	PMID: 29632486
MCAO, Middle Cerebral Artery Occlusion; GEO, Gene Expression Omnibus.

Differential Expression Analysis of PRGs

The samples in Combined Datasets can be categorized into MCAO and Control groups. The R package limma [15] was employed for differential analysis of genes in the MCAO and Control groups. Genes with logFC > 1.00 and adjusted p-value < 0.05 were identified as up-regulated Differential Expression Genes (DEGs), while those with logFC < -1.00 and adjusted p-value < 0.05 were considered down-regulated DEGs. The intersection of DEGs and PRGs was identified as Pyroptosis-Related Differentially Expressed Genes (PRDEGs). Volcano plots were generated by the R package ggplot2, heatmaps were generated by the R package pheatmap. Chromosome localization maps were generated using the R package RCircos [16].

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Enrichment Analysis

GO analysis [17] and KEGG [18] enrichment analysis was conducted with R package clusterProfiler [19], with selection criteria set at adjusted p-value < 0.05 and a False Discovery Rate (FDR) < 0.25.

Gene Set Enrichment Analysis (GSEA)

We performed GSEA on all genes in the integrated GEO dataset with the R package clusterProfiler with parameters set as follows: seed = 2020, permutations = 1000, minimum gene set size = 10, maximum gene set size = 500. The GSEA utilized the c2.cp.all.v2022.1.Hs.symbols.gmt All Canonical Pathways gene set from MSigDB [13]. The criteria for GSEA filtering were adjusted p-value < 0.05 and FDR < 0.25.

Gene Set Variation Analysis (GSVA)

Utilizing c2.all.v7.5.1.symbols.gmt gene sets from MSigDB [13], GSVA [20] was performed on all genes in the Combined Datasets for MCAO vs. Control and High Risk vs. Low Risk groups. Selection criteria for GSVA were set at |logFC| > 1.30 and adjusted p-value < 0.05 for MCAO vs. Control, and |logFC| > 1.00 and p-value < 0.05 for High Risk vs. Low Risk.

Construction of Middle Cerebral Artery Occlusion Diagnostic Model

To construct a diagnostic model for MCAO using the Combined Datasets, logistic regression analysis was applied to PRDEGs. PRDEGs were selected based on adjusted p-value < 0.10, and a logistic regression model was built. Subsequently, a Support Vector Machine (SVM) [21] model was constructed using the genes from the logistic regression model with the highest accuracy and lowest error rate. Finally, LASSO (Least Absolute Shrinkage and Selection Operator) regression analysis using the R package glmnet [22] was performed on the PRDEGs from the SVM model, with set.seed(100) as a parameter. LASSO regression reduces overfitting and improves model generalization by adding a penalty term (lambda × absolute value of the slope) in linear regression. The results were visualized through diagnostic model plots and variable trajectory plots. The outcome of the LASSO regression analysis represented the MCAO diagnostic model, with the included PRDEGs identified as hub genes. A nomogram [23] was generated using the R package rms.

Validation of Middle Cerebral Artery Occlusion Diagnostic Model

Receiver Operating Characteristic Curve (ROC Curve) [24] analysis was conducted using the R package pROC to assess the diagnostic performance of Pyroptosis-Related hub Genes (PRhubGs) in the integrated GEO dataset. The Area Under the Curve (AUC) of the ROC curve was calculated, with AUC values between 0.5 and 1 indicating varying levels of accuracy. Higher AUC values, approaching 1, indicate improved diagnostic performance. Specifically, AUC values in the range of 0.5 to 0.7 denote lower accuracy, 0.7 to 0.9 indicate moderate accuracy, and values above 0.9 suggest high accuracy. Additionally, single-sample Gene-Set Enrichment Analysis (ssGSEA) was employed using the R package GSVA to quantify the Pyroptosis Score (PScore) based on the expression matrix of PRhubGs. Group comparison plots, ROC curves, and AUC calculations were generated using R packages ggplot2 and pROC, respectively.

Calibration curves were drawn to assess the accuracy and discrimination of the MCAO diagnostic model based on PScore. Decision Curve Analysis (DCA) [25] was performed using the R package ggDCA. The DCA curves were plotted based on PScore from the integrated GEO dataset.

Construction of High and Low-Risk Groups in Middle Cerebral Artery Occlusion

Using the median expression value of RiskScore from the MCAO diagnostic model, samples were categorized into High Risk and Low Risk groups, calculated as:

$$RiskScore=WISP2\times 1.6643+MELK\times 0.4164+SDF2L1\times 1.8724+AURKB\times 0.8093$$

Differential analysis, with a threshold of |logFC| > 1.00 and adjusted p-value < 0.05, was performed on genes in these groups using the R package limma [15]. Up-regulated and down-regulated genes were visualized using volcano plots generated by the R package ggplot2, and heatmap plots were created using the R package pheatmap.

Expression Differences and Correlation Analysis

To explore the expression differences of PRhubGs in the MCAO and Control groups of the integrated GEO dataset, group comparison plots were generated based on the expression levels of PRhubGs. Spearman correlation analysis was performed on the expression levels of PRhubGs in the integrated GEO dataset, and the results were visualized using the R package pheatmap. Similarly, to investigate the expression differences of PRhubGs in the High Risk and Low Risk groups of MCAO samples, group comparison plots were created based on the expression levels of PRhubGs. Spearman correlation analysis was conducted on the expression levels of PRhubGs in the MCAO samples, and the results were visualized using the R package pheatmap.

Immune Infiltration Analysis

CIBERSORT [26] was employed to estimate the composition and abundance of immune cells in the transcriptome expression matrix. Using LM22 feature gene matrix and filtering for immune cell enrichment scores greater than zero, the specific results of the immune cell infiltration matrix were obtained. Subsequently, group comparison plots illustrating the expression differences of LM22 immune cells in the MCAO and Control groups of the integrated GEO dataset were generated using the R package ggplot2. Additionally, correlation heatmaps were created using the R package pheatmap to visualize the correlation analysis results between LM22 immune cells, PRhubGs, and immune cells.

ssGSEA [27] was used to quantify the relative abundance of immune cell infiltration. After labeling various infiltrating immune cell types, ssGSEA was applied to calculate enrichment scores, representing the relative abundance of immune cell infiltration in each sample. Samples with a p-value < 0.05 were filtered, resulting in the immune cell infiltration matrix. Lastly, correlation heatmaps were separately generated using the R package pheatmap to display the correlation analysis results between immune cells, PRhubGs, and immune cells in the High and Low-Risk groups of MCAO.

Regulatory Network Construction and Protein Domain Prediction

Transcription factors (TFs) control gene expression by interacting with PRhubGs in the post-transcriptional stage. Using the ChIPBase database, TFs regulating PRhubGs were identified and visualized using Cytoscape to construct the mRNA-TF Regulatory Network. The relationship between PRhubGs and miRNAs was analyzed using the TarBase database, and the mRNA-miRNA Regulatory Network was visualized with Cytoscape. Furthermore, potential direct and indirect drug targets of PRhubGs were predicted using the Comparative Toxicogenomics Database (CTD). The interaction between PRhubGs and drugs was visualized in the mRNA-Drug Regulatory Network using Cytoscape. Finally, the SMART database was employed for predicting protein domains, critical functional units with specific spatial structures. The protein domains of PRhubGs were queried using the SMART database and visualized for analysis.

Statistical Analysis

The analyses were conducted using R software (Version 4.2.0). Continuous variables are presented as mean ± standard deviation. The Wilcoxon Rank Sum Test was applied for group comparisons. Unless otherwise specified, Spearman correlation analysis was used to calculate correlation coefficients between different molecules, and results were considered significant if adjusted p-value < 0.05. p-values were adjusted with Benjamini-Hochberg (BH) correction method.

Analysis Flow Chart and Data Preprocessing

We obtained MCAO datasets (GSE97537, GSE61616, GSE106680, n = 28) from GEO databases. PRGs sourced from GeneCards and the Molecular Signatures Database were subjected to differential expression analysis, followed by GO and KEGG functional assessments. Functional enrichment differences between high-risk and low-risk MCAO groups were calculated using GSEA and GSVA. LASSO regression identified and validated PRDEGs for constructing the MCAO diagnostic model. High and low-risk MCAO groups were formed for expression and immune cell correlation analysis with pyroptosis-related hub genes. A regulatory network between pyroptosis-related hub genes and miRNA was established, and protein domain prediction was conducted. The expression of key pyroptosis genes in the MCAO rat model was validated using qRT-PCR (Fig. 1).

Integration of the Middle Cerebral Artery Occlusion Datasets

The MCAO datasets were involved in initial batch effect removal and formed a unified GEO dataset termed "Combined Datasets." The PCA graphs revealed a significant reduction in batch effects among the samples within the MCAO dataset following the batch effect removal procedure (Fig. 2A-D).

Analysis of Differentially Expressed Genes Associated with Cell Pyroptosis

The Combined Datasets were stratified into MCAO and Control groups. Differential expression analysis unveiled a total of 675 DEGs. A total of 507 genes exhibited upregulated expression, while 168 showed downregulated expression (Fig. 3A). We intersected all DEGs and PRGs and determined a total of 25 PRDEGs (Fig. 3B), including PLIN2, MCM2, HSPB1, CDT1, MCM3, CKS2, CKLF, MCM6, MCM5, MELK, CDCA3, CDCA7, SDF2L1, CCNA2, UHRF1, PBK, TOP2A, LOX, PRC1, AKR1B10, WISP2, FAM111A, AURKB, BUB1B, and TTK.

Subsequently, the 25 PRDEGs among different sample groups in the Combined Datasets showed obvious differential expression (Fig. 3C). The 25 PRDEGs were distributed on each chromosome (Fig. 3D). Notably, the majority of the PRDEGs were located on chromosome 9, including CKS2, MELK and PLIN2.

GO and KEGG Enrichment Analyses

GO and KEGG enrichment analyses were conducted to further elucidate the biological relevance of the 25 PRDEGs in the context of MCAO (Table 2 and Fig. 4A). The analysis revealed that the PRDEGs were predominantly enriched in biological processes (BP) such as double-strand break repair via break-induced replication, DNA replication initiation, negative regulation of chromosome organization, DNA unwinding involved in DNA replication, and DNA-templated DNA replication. Regarding cellular components (CC), enrichment was observed in chromosomal regions, CMG complex, DNA replication preinitiation complex, chromosome centromeric regions, and nuclear chromosomes. In terms of molecular functions (MF), the PRDEGs were associated with single-stranded DNA helicase activity, ATP-dependent activity acting on DNA, single-stranded DNA binding, 3'-5' DNA helicase activity, and DNA helicase activity. Additionally, significant enrichment was identified in biological pathways such as DNA replication and Cell cycle. Furthermore, network diagrams were generated to depict the relationships among BP, CC, MF, and pathways (Fig. 4B-E). Finally, the enriched KEGG pathways, particularly DNA replication (Fig. 5A) and Cell cycle (Fig. 5B), were visualized using the Pathview R package.

Table 2

Results of GO and KEGG Enrichment Analysis for PRDEGs
ONTOLOGY	ID	Description	GeneRatio	BgRatio	pvalue	p.adjust	qvalue
BP	GO:0000727	double-strand break repair via break-induced replication	4/24	12/18800	1.00E-09	2.27E-07	1.38E-07
BP	GO:0006270	DNA replication initiation	5/24	38/18800	1.06E-09	2.27E-07	1.38E-07
BP	GO:2001251	negative regulation of chromosome organization	6/24	89/18800	1.19E-09	2.27E-07	1.38E-07
BP	GO:0006268	DNA unwinding involved in DNA replication	4/24	22/18800	1.47E-08	2.10E-06	1.28E-06
BP	GO:0006261	DNA-templated DNA replication	6/24	159/18800	3.95E-08	3.99E-06	2.43E-06
CC	GO:0098687	chromosomal region	9/24	366/19594	2.56E-10	1.87E-08	9.98E-09
CC	GO:0071162	CMG complex	4/24	11/19594	5.68E-10	2.07E-08	1.10E-08
CC	GO:0031261	DNA replication preinitiation complex	4/24	12/19594	8.51E-10	2.07E-08	1.10E-08
CC	GO:0000775	chromosome, centromeric region	5/24	227/19594	7.09E-06	1.06E-04	5.64E-05
CC	GO:0000228	nuclear chromosome	5/24	228/19594	7.25E-06	1.06E-04	5.64E-05
MF	GO:0017116	single-stranded DNA helicase activity	4/24	23/18410	1.93E-08	1.45E-06	7.94E-07
MF	GO:0008094	ATP-dependent activity, acting on DNA	5/24	103/18410	1.94E-07	7.28E-06	3.99E-06
MF	GO:0003697	single-stranded DNA binding	5/24	120/18410	4.16E-07	1.04E-05	5.70E-06
MF	GO:0043138	3'-5' DNA helicase activity	3/24	14/18410	7.02E-07	1.32E-05	7.20E-06
MF	GO:0003678	DNA helicase activity	4/24	72/18410	2.15E-06	3.23E-05	1.77E-05
KEGG	hsa04110	Cell cycle	7/12	126/8164	1.31E-10	3.02E-09	2.35E-09
KEGG	hsa03030	DNA replication	4/12	36/8164	1.54E-07	1.77E-06	1.38E-06
GO, Gene Ontology; BP, Biological Process; CC, Cellular Component; MF, Molecular Function; KEGG, Kyoto Encyclopedia of Genes and Genomes; PRDEGs, Pyroptosis-Related Differentially Expressed Genes.

GSEA of the Integrated GEO Dataset

GSEA was employed to elucidate the impact of gene expression levels on MCAO within the Integrated GEO Dataset. The analysis aimed to explore the connection between the expression of all genes in the Integrated GEO Dataset and the biological processes involved, cellular components affected, and molecular functions exhibited (Fig. 6A). Detailed results are presented in Table 3. The GSEA revealed significant enrichment of all genes in the Integrated GEO Dataset in biologically relevant functions and signaling pathways, notably in the IL18 Signaling Pathway (Fig. 6B), Inflammatory Response Pathway (Fig. 6C), Il1 and Megakaryocytes in Obesity (Fig. 6D), and Oxidative Damage Response (Fig. 6E).

Table 3

Results of GSEA for Combined Datasets
ID	setSize	enrichmentScore	NES	pvalue	p.adjust	qvalues
REACTOME_NEUTROPHIL_DEGRANULATION	341	7.14E-01	2.81E + 00	1.00E-10	6.71E-09	4.37E-09
WP_TYROBP_CAUSAL_NETWORK_IN_MICROGLIA	51	8.50E-01	2.65E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_ANTIGEN_PROCESSING_CROSS_PRESENTATION	81	7.79E-01	2.63E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_EXTRACELLULAR_MATRIX_ORGANIZATION	210	6.89E-01	2.62E + 00	1.00E-10	6.71E-09	4.37E-09
WP_BURN_WOUND_HEALING	74	7.76E-01	2.59E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_IMMUNOREGULATORY_INTERACTIONS_BETWEEN_A_LYMPHOID_AND_A_NON_LYMPHOID_CELL	56	8.15E-01	2.59E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_SIGNALING_BY_INTERLEUKINS	343	6.54E-01	2.57E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_CYTOKINE_SIGNALING_IN_IMMUNE_SYSTEM	497	6.37E-01	2.56E + 00	1.00E-10	6.71E-09	4.37E-09
KEGG_HEMATOPOIETIC_CELL_LINEAGE	61	7.88E-01	2.54E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_INTERLEUKIN_4_AND_INTERLEUKIN_13_SIGNALING	86	7.45E-01	2.54E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_INTEGRIN_CELL_SURFACE_INTERACTIONS	61	7.86E-01	2.53E + 00	1.00E-10	6.71E-09	4.37E-09
WP_MICROGLIA_PATHOGEN_PHAGOCYTOSIS_PATHWAY	34	8.62E-01	2.50E + 00	1.00E-10	6.71E-09	4.37E-09
KEGG_COMPLEMENT_AND_COAGULATION_CASCADES	55	7.85E-01	2.49E + 00	1.00E-10	6.71E-09	4.37E-09
REACTOME_INTERLEUKIN_10_SIGNALING	38	8.44E-01	2.48E + 00	1.00E-10	6.71E-09	4.37E-09
PID_INTEGRIN1_PATHWAY	50	8.00E-01	2.48E + 00	1.00E-10	6.71E-09	4.37E-09
KEGG_LEISHMANIA_INFECTION	51	7.94E-01	2.48E + 00	1.00E-10	6.71E-09	4.37E-09
WP_IL18_SIGNALING_PATHWAY	203	6.26E-01	2.37E + 00	1.00E-10	6.71E-09	4.37E-09
WP_INFLAMMATORY_RESPONSE_PATHWAY	27	8.21E-01	2.31E + 00	2.46E-07	6.03E-06	3.93E-06
WP_IL1_AND_MEGAKARYOCYTES_IN_OBESITY	20	8.59E-01	2.26E + 00	2.76E-07	6.56E-06	4.28E-06
WP_OXIDATIVE_DAMAGE_RESPONSE	29	7.98E-01	2.25E + 00	3.89E-07	8.79E-06	5.73E-06
GSEA, Gene Set Enrichment Analysis.

GSVA of the Integrated GEO Dataset

GSVA was conducted to investigate the differential enrichment of the c2.all.v7.5.1.symbols.gmt gene set between the MCAO and Control groups within the Combined Dataset. Detailed information is provided in Table 4 and Fig. 7A. The GSVA results indicated that 12 pathways exhibited statistically significant differences (p-value < 0.05) between the MCAO and Control groups. These pathways include macrophage markers, ZERBINI response to SULINDAC DN, CLEC7A inflammasome pathway, TESAR ALK targets human ES 4D DN, TESAR ALK targets human ES 5D DN, RUNX3 regulates immune response and cell migration, CD22 mediated BCR regulation, metal sequestration by antimicrobial proteins, GALI TP53 targets apoptotic UP, RUNX1 regulates transcription of genes, SMIRNOV circulating endotheliocytes in cancer DN, and activation of C3 and C5. The GSVA scores of the 12 pathways also show obvious differences between the MCAO and Control groups (Fig. 7B).

Table 4

Results of GSVA for Combined Datasets
ID	logFC	AveExpr	p value	adj.p value
WP_MACROPHAGE_MARKERS	1.35E + 00	3.22E-02	1.99E-25	6.33E-22
ZERBINI_RESPONSE_TO_SULINDAC_DN	1.37E + 00	1.47E-02	1.19E-24	1.90E-21
REACTOME_CLEC7A_INFLAMMASOME_PATHWAY	1.45E + 00	4.26E-02	9.84E-24	8.91E-21
TESAR_ALK_TARGETS_HUMAN_ES_4D_DN	1.58E + 00	4.31E-02	2.18E-22	6.59E-20
TESAR_ALK_TARGETS_HUMAN_ES_5D_DN	1.58E + 00	4.31E-02	2.18E-22	6.59E-20
REACTOME_RUNX3_REGULATES_IMMUNE_RESPONSE_AND_CELL_MIGRATION	1.41E + 00	2.87E-02	2.60E-22	7.32E-20
REACTOME_CD22_MEDIATED_BCR_REGULATION	1.34E + 00	7.56E-03	5.07E-22	1.11E-19
REACTOME_METAL_SEQUESTRATION_BY_ANTIMICROBIAL_PROTEINS	1.50E + 00	4.69E-02	1.54E-21	2.76E-19
GALI_TP53_TARGETS_APOPTOTIC_UP	1.36E + 00	1.90E-03	9.94E-18	3.31E-16
REACTOME_RUNX1_REGULATES_TRANSCRIPTION_OF_GENES	1.32E + 00	7.23E-03	5.54E-15	1.15E-13
SMIRNOV_CIRCULATING_ENDOTHELIOCYTES_IN_CANCER_DN	1.30E + 00	-1.78E-02	1.96E-13	2.54E-12
REACTOME_ACTIVATION_OF_C3_AND_C5	1.35E + 00	-1.26E-02	2.31E-13	2.96E-12
GSVA, Gene Set Variation Analysis.

MCAO Diagnostic Model

To ascertain the diagnostic value of the 25 PRDEGs in MCAO, a logistic regression model was constructed based on these genes (Fig. 8A). We identified 14 statistically significant PRDEGs (p-value < 0.05), including AURKB, BUB1B, CCNA2, CDCA3, CDCA7, CKLF, FAM111A, MELK, PBK, PRC1, SDF2L1, TOP2A, TTK, and WISP2.

Subsequently, an SVM model was developed using the 14 PRDEGs and SVM algorithm, revealing optimal performance with a gene count of 4 (Fig. 8B-C). These 4 PRDEGs—WISP2, MELK, SDF2L1 and AURKB—were identified as key contributors to the SVM model. A LASSO regression analysis based on these genes yielded the final MCAO diagnostic model (Fig. 8D-E).

To further validate the diagnostic model's utility, a Nomogram (Fig. 7F) was constructed based on the cell death-related hub genes (hub genes). Results demonstrated that the expression level of WISP2 significantly contributed to the diagnostic model's efficacy, surpassing other variables, while MELK exhibited lower utility compared to other variables.

Validation of the MCAO Diagnostic Model

To validate the diagnostic model for MCAO, ROC curves were generated using the expression levels of the cell death-related hub genes (hub genes) WISP2, MELK, SDF2L1, and AURKB in the Combined Datasets. The ROC curves indicated high accuracy (AUC > 0.9) in distinguishing between different groups (Fig. 9A-D).

Utilizing the ssGSEA algorithm, PScores were calculated based on the expression of the four hub genes across all samples in the Combined Datasets. The PScore between the MCAO and Control groups has statistically significant variations (p-value < 0.001) (Fig. 9E). Furthermore, ROC curves based on PScore confirmed their high accuracy (AUC > 0.9) in discriminating between different groups (Fig. 9F).

The accuracy and discriminatory power of the MCAO diagnostic model were evaluated through Calibration analysis, presenting a Calibration Curve that slightly deviated from the ideal diagonal line but demonstrated a close fit (Fig. 9G). Additionally, DCA based on PScore from the Combined Datasets assessed the clinical utility of the MCAO diagnostic model, revealing a stable net benefit within a certain range and overall favorable performance (Fig. 9H).

GSEA of the High and Low-Risk Groups

The MCAO samples were stratified into High-Risk and Low-Risk groups based on their RiskScores, calculated using the formula:

$$RiskScore=WISP2\times 1.6643+MELK\times 0.4164+SDF2L1\times 1.8724+AURKB\times 0.8093$$

Differential expression analysis identified 18 DEGs in MCAO samples (adjusted p-value < 0.05). Among these, 12 genes were upregulated (logFC > 1.00, adjusted p-value < 0.05) and 6 genes were downregulated (logFC < -1.00, adjusted p-value < 0.05) (Fig. 10A-B). GSEA was performed to understand the impact of gene expression in MCAO samples on the condition. Notably, all genes in MCAO samples were significantly enriched in processes such as Apoptosis (Fig. 10D), Il1 and Megakaryocytes in Obesity (Fig. 10E), IL18 Signaling Pathway (Fig. 10F), and Photodynamic Therapy-induced Nfkb Survival Signaling (Fig. 10C, G, Table 5).

Table 5

Results of GSEA for Risk Group
ID	setSize	enrichmentScore	NES	pvalue	p.adjust	qvalues
WP_BURN_WOUND_HEALING	74	7.62E-01	2.75E + 00	1.00E-10	1.34E-08	9.46E-09
PID_INTEGRIN1_PATHWAY	50	7.71E-01	2.55E + 00	1.00E-10	1.34E-08	9.46E-09
REACTOME_NEUTROPHIL_DEGRANULATION	341	5.75E-01	2.51E + 00	1.00E-10	1.34E-08	9.46E-09
REACTOME_INTERLEUKIN_10_SIGNALING	38	8.12E-01	2.50E + 00	1.00E-10	1.34E-08	9.46E-09
REACTOME_INTEGRIN_CELL_SURFACE_INTERACTIONS	61	7.21E-01	2.50E + 00	1.00E-10	1.34E-08	9.46E-09
REACTOME_EXTRACELLULAR_MATRIX_ORGANIZATION	210	5.95E-01	2.45E + 00	1.00E-10	1.34E-08	9.46E-09
WP_TYROBP_CAUSAL_NETWORK_IN_MICROGLIA	51	7.05E-01	2.36E + 00	1.06E-08	7.78E-07	5.48E-07
WP_OVERVIEW_OF_PROINFLAMMATORY_AND_PROFIBROTIC_MEDIATORS	67	6.70E-01	2.35E + 00	2.84E-09	2.72E-07	1.92E-07
REACTOME_DEGRADATION_OF_THE_EXTRACELLULAR_MATRIX	94	6.35E-01	2.35E + 00	1.00E-10	1.34E-08	9.46E-09
REACTOME_ANTIGEN_PROCESSING_CROSS_PRESENTATION	81	6.45E-01	2.33E + 00	1.87E-09	2.03E-07	1.43E-07
WP_RETINOBLASTOMA_GENE_IN_CANCER	66	6.61E-01	2.31E + 00	1.17E-08	8.37E-07	5.90E-07
REACTOME_INTERLEUKIN_4_AND_INTERLEUKIN_13_SIGNALING	86	6.30E-01	2.30E + 00	1.45E-09	1.65E-07	1.16E-07
WP_MICROGLIA_PATHOGEN_PHAGOCYTOSIS_PATHWAY	34	7.43E-01	2.30E + 00	3.63E-07	1.80E-05	1.27E-05
REACTOME_DNA_STRAND_ELONGATION	29	7.61E-01	2.29E + 00	9.35E-07	4.02E-05	2.84E-05
PID_INTEGRIN3_PATHWAY	29	7.57E-01	2.28E + 00	1.34E-06	4.90E-05	3.46E-05
REACTOME_COLLAGEN_FORMATION	55	6.73E-01	2.28E + 00	1.87E-07	9.93E-06	7.00E-06
WP_APOPTOSIS	65	6.10E-01	2.13E + 00	2.24E-06	6.54E-05	4.62E-05
WP_IL1_AND_MEGAKARYOCYTES_IN_OBESITY	20	7.63E-01	2.12E + 00	7.31E-05	1.09E-03	7.68E-04
WP_IL18_SIGNALING_PATHWAY	203	5.13E-01	2.11E + 00	1.00E-10	1.34E-08	9.46E-09
WP_PHOTODYNAMIC_THERAPYINDUCED_NFKB_SURVIVAL_SIGNALING	29	6.78E-01	2.05E + 00	1.02E-04	1.40E-03	9.85E-04
GSEA, Gene Set Enrichment Analysis.

GSVA of the High and Low-Risk Groups

To explore the differences in the c2.all.v7.5.1.symbols.gmt gene set between the High-Risk and Low-Risk groups in MCAO samples, GSVA was conducted on all genes from MCAO samples (Table 6).

Table 6

Results of GSVA for Risk Group
ID	logFC	AveExpr	p value	adj.p value
REACTOME_ATTACHMENT_AND_ENTRY	-1.16E + 00	-2.95E-02	6.00E-07	1.09E-03
MYLLYKANGAS_AMPLIFICATION_HOT_SPOT_21	-1.15E + 00	-5.44E-02	8.54E-07	1.09E-03
LAU_APOPTOSIS_CDKN2A_DN	-1.09E + 00	-3.06E-02	8.58E-07	1.09E-03
REACTOME_RUNX1_REGULATES_TRANSCRIPTION_OF_GENES	-1.10E + 00	-7.92E-02	5.21E-05	5.52E-03
REACTOME_NECTIN_NECL_TRANS_HETERODIMERIZATION	1.02E + 00	4.04E-02	8.22E-04	1.81E-02
GSVA, Gene Set Variation Analysis

GSVA revealed five pathways that displayed statistically significant differences (p-value < 0.05) between the High-Risk and Low-Risk groups, namely: attachment and entry, MYLLYKANGAS amplification hot spot 21, LAU apoptosis CDKN2A DN, RUNX1 regulates transcription of genes, and NECTIN NECL trans heterodimerization (Fig. 11A-B).

Correlation and Expression Analysis of Cell Death-Related Hub Genes

The four hub genes, including WISP2, MELK, SDF2L1, and AURKB showed significantly differential expression between the MCAO and Control groups (p-value < 0.001, Fig. 12A). Besides, these hub genes also showed a significant differential expression (p-value < 0.05) between the High-Risk and Low-Risk groups. Specifically, WISP2 showed highly significant expression differences (p-value < 0.001), while MELK and SDF2L1 demonstrated substantial statistical significance (p-value < 0.01). The four hub genes exhibited a significant positive correlation among the cell death-related hub genes in both the Integrated GEO Dataset and MCAO samples (Fig. 12C-D).

Integrated GEO Dataset Immunoinfiltration Analysis

The CIBERSORT algorithm was applied to compute the correlation between 22 immune cell types and the samples from the MCAO and Control groups. The immunoinfiltration analysis revealed the proportions of immune cells in the Integrated GEO Dataset (Fig. 13A), demonstrating positive enrichment scores for 19 immune cell types (Fig. 13B). A total of 12 immune cell types showed statistically significant expression differences (p-value < 0.05) between the MCAO and Control groups, including memory B cells, CD8 T cells, CD4 naive T cells, regulatory T cells (Tregs), monocytes, M0 macrophages, M1 macrophages, resting dendritic cells, resting mast cells, activated mast cells, eosinophils, and neutrophils (Fig. 13B). Notably, positive correlations were observed between CD4 naive T cells and memory B cells (r value = 0.68), while negative correlations were identified between resting dendritic cells and memory B cells, as well as neutrophils and Tregs (r value = -0.65) (Fig. 13C). Furthermore, a significant positive correlation was shown between the cell death-related hub gene WISP2 and the immune cell type Dendritic cells resting, and a notable negative correlation was shown between MELK and the immune cell type Macrophages M1 (Fig. 13D).

Immunoinfiltration Analysis of the High and Low-Risk Groups

The ssGSEA algorithm was applied to calculate the immune infiltration abundance of 28 immune cell types in the high and low-risk groups (p-value < 0.05; Fig. 14A). Twelve immune cell types exhibited significant differences (p-value < 0.05), including activated CD4 T cells, central memory CD4 T cells, effector memory CD4 T cells, effector memory CD8 T cells, gamma delta T cells, regulatory T cells, type 1 T helper cells, activated dendritic cells, immature dendritic cells, myeloid-derived suppressor cells (MDSCs), natural killer cells, and plasmacytoid dendritic cells. Correlation heatmaps demonstrated the relationships among the infiltration abundances of the 12 immune cell types in different groups (Fig. 14B-C). Positive correlations were observed in both the high and low-risk groups for all immune cell types except the effector memory CD8 T cells, with the low-risk group showing more negative correlations.

Lastly, correlation dot plots showcased the associations between cell death-related hub genes and the immune infiltration abundances in the high and low-risk groups of the MCAO dataset (Fig. 14D-E). In the high-risk group, a significant positive correlation was observed between the immune cell Activated dendritic cell and the hub gene SDF2L1. In the low-risk group, a notable positive correlation was observed between the immune cell Natural killer cell and the hub gene SDF2L1.

Regulatory Network Construction and Protein Domain Prediction

To elucidate the regulatory landscape, we initially obtained TF binding to cell death-related hub genes from the ChIPBase database, forming the mRNA-TF regulatory network (Fig. 15A). The network comprised two cell death-related hub genes and 53 TFs (Table S2). Utilizing the TarBase database, we identified miRNAs associated with cell death-related hub genes, constructing the mRNA-miRNA regulatory network. Visualization of this network through Cytoscape (Fig. 15B) revealed four hub genes and 41 miRNAs involved (Table S3). Exploration of the CTD database facilitated the acquisition of drugs linked to cell death-related hub genes, leading to the construction of the mRNA-Drug regulatory network. Cytoscape visualization (Fig. 15C) showcased two hub genes and 18 associated drugs (Table S4). To gain insights into the protein architecture of the four cell death-related hub genes, including WISP2, MELK, SDF2L1 and AURKB, we leveraged the SMART database for protein domain prediction (Fig. 15D-G). Notably, WISP2 exhibited domains such as IB, VWC, and TSP1; MELK contained the S_TKc domain; SDF2L1 featured the MIR domain; and AURKB included the S_TKc domain.

IS, characterized by rapid clinical signs of brain damage, ranks as the second leading global cause of death and the third in terms of disability [2]. The current clinical approach, centered on thrombolytic therapy for IS, faces a formidable challenge due to the lack of precisely defined targets and treatment mechanisms [1]. Pyroptosis, marked by cell membrane pore formation and pro-inflammatory cytokine release, emerges as an important player in the inflammatory response to IRI. The exploration of the intrinsic relationship between pyroptosis and IRI emerges as a crucial research avenue [6]. The urgency lies in unraveling this complex web of interactions to not only comprehend the pathogenesis of IS comprehensively but also to pave the way for targeted therapeutic strategies. This study, grounded in advanced RNA transcriptome analyses, seeks to bridge the gap in our understanding by elucidating the connections between pyroptosis, IRI, and the dynamic expression of RNA.

The comparison of MCAO patients and controls reveals a notable enrichment of genes in immune response pathways like IL1 and megakaryocytes signaling, IL18 signaling pathway, and Nfkb survival signaling (Fig. 6). This emphasizes the crucial involvement of immune cells in the complex molecular events following IRI [28]. Besides, our comparative analysis unveiled the double-strand break repair pathway, DNA replication pathway and cell cycle pathway have been extensively documented in the MCAO samples. The results showcase the consistency between our findings and previous research on the unexpected effects of DNA damage on IRI [9, 29]. Additionally, our study introduces innovative perspectives by uncovering the involvement of CD22-mediated BCR regulation and TP53 targets apoptotic in the IRI-induced pyroptosis process. These novel pathways offer insights into potential mechanisms underlying pyroptosis post-IRI.

Our study identified 25 key genes associated with pyroptosis in the context of IRI. Employing logistic modeling and SVM analyses, we strategically focused on four hub genes—WISP2, MELK, SDF2L1, and AURKB—as central players orchestrating the expression landscape of MCAO. WISP2, an adipokine known for its role in pro-inflammatory responses, emerged as a central player in modulating the inflammatory cascade in pyroptosis (30). Importantly, its anti-inflammatory properties align seamlessly with pathways governing immune responses and tissue repair (31). This dual role positions WISP2 as a critical regulator in the intricate network of pyroptotic events, making it a noteworthy focus of our study. MELK, a key regulator of the cell cycle, was identified as a significant influencer of cell survival dynamics in the context of pyroptosis (32). Its role in shaping cell cycle regulation connects with broader pathways governing cellular fate after IRI (33). This connection highlights MELK as a key player not only in pyroptotic processes but also in the broader landscape of cellular responses to ischemic insult, further emphasizing its relevance in our study. The study detailed the analysis of WISP2 and MELK in pyroptotic mechanisms. The consistency of our findings with existing literature further supports the robustness and reliability of our study, contributing valuable insights to the field and paving the way for potential therapeutic interventions in IRI-induced pyroptosis.

SDF2L1, linked to endoplasmic reticulum stress, influences cellular responses to ischemic insult (34). AURKB is a crucial regulator of cell division that influence cellular responses to pyroptosis (35–36). The identification of SDF2L1 and AURKB as key regulators enriches our understanding of their specific roles in IRI-induced pyroptosis. Their association with distinct biological pathways underscores their significance as potential therapeutic targets, emphasizing their intricate contributions to molecular events governing pyroptosis post-IRI injury. Upon validating the roles of SDF2L1 and AURKB in the MCAO rat model, our study makes a significant contribution by explicitly establishing their connections with IRI. This marks a pivotal revelation, as our research is the first to uncover the intricate relationship between these two genes and IRI. The identification of SDF2L1 and AURKB as contributors to the IRI landscape underscores their importance in understanding the pathophysiology of IRI-induced pyroptosis. This discovery holds particular value as it represents a novel finding and enhances the broader understanding of the molecular events underlying IRI.

Leukostasis, infiltration of neutrophils and macrophages, and activation of complements and microglial cells occur in patients and animal models with IRI [9]. In this study, the rising level of inflammatory cytokines and chemokines, including NF-kB pathway, IL18 pathway, and IL1 pathway (Fig. 7), were detected in MCAO, indicating a critical role of inflammation in IRI. WISP2, potentially pro-inflammatory, might influence pyroptosis post-IRI considering its link to the immune microenvironment [30]. Moreover, the infiltration of dendritic cells and natural killer cells were positively correlated with SDF2L1 expression in high and low IRI risk groups, respectively. The result implicates the role of endoplasmic reticulum stress responses in shaping the immune microenvironment during pyroptotic events. These findings collectively emphasize the multifaceted nature of the immune response in the context of IRI, implicating various cellular components and signaling pathways.

Our study not only revealed transcriptome signatures linked to pyroptosis in IRI-induced injury but also crafted a diagnostic model for MCAO. This model, built on four pivotal pyroptosis-associated genes—WISP2, MELK, SDF2L1, and AURKB—showed a strong discriminatory ability to distinguish MCAO cases from controls. Using logistic regression, we pinpointed these genes as crucial contributors to the predictive accuracy of our model. Integrating these genes into a diagnostic tool facilitates efficient risk stratification, aiding clinicians in identifying individuals at higher risk of adverse outcomes post-MCAO. Categorizing MCAO patients into high and low-risk groups using our diagnostic model revealed significant changes in key biological pathways, including attachment and entry, apoptosis CDKN2A downregulation, RUNX1 regulates transcription, and Nectin/Necl trans heterodimerization, seen in both groups. These findings suggest a potential link between risk stratification via our diagnostic model and dysregulation of these pathways. The observed pathway variations in high and low-risk groups emphasize the clinical relevance of our diagnostic model. While promising, future studies should validate its clinical utility in larger patient cohorts and explore potential therapeutic interventions based on the identified risk groups.

While our study provides insights into the transcriptome dynamics associated with pyroptosis in IRI-induced injury, several limitations should be acknowledged. Firstly, three public RNA transcriptome datasets of patients with MCAO were collected from public datasets. More research efforts should prioritize the collection of larger and more varied datasets to enhance the robustness and generalizability of our findings. Furthermore, the current study primarily focused on gene expression changes. The biological regulatory mechanisms of pyroptosis such as epigenetics and post-translational modifications could also participate in the IRI process. Future investigations should delve into the epigenetic landscape and protein-level alterations to unravel the complete molecular picture associated with pyroptosis in IRI induced injury.

In conclusion, our analysis depicted the transcription dynamics of IS pathogenesis and explored the potential molecular targets for IRI. By unveiling essential insights underlying MCAO and constructing a diagnostic model, we identified the specific mechanisms and molecular targets for IRI, especially those associated with DNA replication and immune response pathways, which warrant meticulous validation in subsequent research endeavors. These findings contribute to the ongoing efforts in refining IS therapeutic strategies, offering a foundation for more targeted and effective interventions.

Funding

This work was supported by grants from the Special Research Start-up Fund for Doctors of Yijishan Hospital, Wannan Medical College (KY288306152022), Anhui Provincial Key Clinical Specialty Recommendation Project Funding, the Research Foundation of Yi Ji Shan Hospital, Wannan Medical College (YR202003) and Colleges and Universities Natural & Science Fund of Anhui (KJ2021A0832), and the young and middle-aged scientific research funding of Wannan Medical College (No. WK2021F29).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Author contributions

WY and LC conceived and designed the research, conducted bioinformatic analyses, and drafted the manuscript. LB and YF provided insightful guidance and revised the manuscript. MJ administered the project, YQ revised the manuscript and provided support. All authors contributed to the article and read and approved the final version of the manuscript.

Data availability statement

The datasets (GSE97537, GSE61616 and GSE106680) analyzed for this study can be found in the Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) database.

Ethics approval: Not applicable

Informed consent to participate: Not applicable

Consent to publish: Not applicable

Feigin VL, Brainin M, Norrving B, Martins S, Sacco RL, Hacke W, Fisher M, Pandian J, Lindsay P (2022) World Stroke Organization (WSO): Global Stroke Fact Sheet 2022. Int J Stroke (2022) 17:18–29. https://doi.org/10.1177/17474930211065917
Wu S, Wu B, Liu M, Chen Z, Wang W, Anderson CS, Sandercock P, Wang Y, Huang Y, Cui L, Pu C, Jia J, Zhang T, Liu X, Zhang S, Xie P, Fan D, Ji X, Wong KL, Wang L; China Stroke Study Collaboration (2019) Stroke in China: advances and challenges in epidemiology, prevention, and management. Lancet Neurol 18:394–405. https://doi.org/10.1016/S1474-4422(18)30500-3
Sporns PB, Fullerton HJ, Lee S, Kim H, Lo WD, Mackay MT, Wildgruber M (2022) Childhood stroke. Nat Rev Dis Prim 8:12. https://doi.org/10.1038/s41572-022-00337-x
Liao Y, Cheng J, Kong X, Li S, Li X, Zhang M, Zhang H, Yang T, Dong Y, Li J, Xu Y, Yuan Z (2020) HDAC3 inhibition ameliorates ischemia/reperfusion-induced brain injury by regulating the microglial cGAS-STING pathway. Theranostics 10:9644–9662. https://doi.org/10.7150/thno.47651
Tao T, Liu M, Chen M, Luo Y, Wang C, Xu T, Jiang Y, Guo Y, Zhang JH (2020) Natural medicine in neuroprotection for ischemic stroke: Challenges and prospective. Pharmacol Ther 216:107695. https://doi.org/10.1016/j.pharmthera.2020.107695
Zheng Y, Xu X, Chi F, Cong N (2022) Pyroptosis: a newly discovered therapeutic target for ischemia-reperfusion injury. Biomolecules 12:1625. https://doi.org/10.3390/biom12111625
Siesjö BK (1992) Pathophysiology and treatment of focal cerebral ischemia. Part I: Pathophysiology. J Neurosurg 77:169–184. https://doi.org/10.3171/jns.1992.77.2.0169
Siesjö BK (1992) Pathophysiology and treatment of focal cerebral ischemia. Part II: Mechanisms of damage and treatment. J Neurosurg 77:337–354. https://doi.org/10.3171/jns.1992.77.3.0337
Eltzschig HK, Eckle T (2011) Ischemia and reperfusion—from mechanism to translation. Nat Med 17:1391–1401. https://doi.org/10.1038/nm.2507
Sun R, Peng M, Xu P, Huang F, Xie Y, Li J, Hong Y, Guo H, Liu Q, Zhu W (2020) Low-density lipoprotein receptor (LDLR) regulates NLRP3-mediated neuronal pyroptosis following cerebral ischemia/reperfusion injury. J Neuroinflammation 17:330. https://doi.org/10.1186/s12974-020-01988-x
Cao X, Wang Y, Gao L (2021) CHRFAM7A overexpression attenuates cerebral ischemia-reperfusion injury via inhibiting microglia pyroptosis mediated by the NLRP3/Caspase-1 pathway. Inflammation 44:1023–1034. https://doi.org/10.1007/s10753-020-01398-4
Stelzer G, Rosen N, Plaschkes I, Zimmerman S, Twik M, Fishilevich S, Stein TI, Nudel R, Lieder I, Mazor Y, Kaplan S, Dahary D, Warshawsky D, Guan-Golan Y, Kohn A, Rappaport N, Safran M, Lancet D (2016) The GeneCards Suite: From Gene Data Mining to Disease Genome Sequence Analyses. Curr Protoc Bioinformatics 54: 1.30.1–1.30.33 https://doi.org/10.1002/cpbi.5
Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP (2011) Molecular signatures database (MSigDB) 3.0. Bioinformatics 27:1739–1740. https://doi.org/10.1093/bioinformatics/btr260
Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD (2012) The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics 28:882–883. https://doi.org/10.1093/bioinformatics/bts034
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK (2015) limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43:e47–e47. https://doi.org/10.1093/nar/gkv007
Zhang H, Meltzer P, Davis S (2013) RCircos: an R package for Circos 2D track plots. BMC Bioinformatics 14:244. https://doi.org/10.1186/1471-2105-14-244
Mi H, Muruganujan A, Ebert D, Huang X, Thomas PD (2019) PANTHER version 14: more genomes, a new PANTHER GO-slim and improvements in enrichment analysis tools. Nucleic Acids Res 47:D419–D426. https://doi.org/10.1093/nar/gky1038
Ogata H, Goto S, Sato K, Fujibuchi W, Bono H, Kanehisa M (1999) KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res 27:29–34. https://doi.org/10.1093/nar/27.1.29
Yu G, Wang L-G, Han Y, He Q-Y (2012) clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters. Omi A J Integr Biol 16:284–287. https://doi.org/10.1089/omi.2011.0118
Hänzelmann S, Castelo R, Guinney J (2013) GSVA: gene set variation analysis for microarray and RNA-Seq data. BMC Bioinformatics 14:7. https://doi.org/10.1186/1471-2105-14-7
Sanz H, Valim C, Vegas E, Oller JM, Reverter (2018) SVM-RFE: selection and visualization of the most relevant features through non-linear kernels. BMC Bioinformatics 19:432. https://doi.org/10.1186/s12859-018-2451-4
Engebretsen S, Bohlin J. Statistical predictions with glmnet (2019) Clin Epigenetics 11:123. https://doi.org/10.1186/s13148-019-0730-1
Wu J, Zhang H, Li L, Hu M, Chen L, Xu B, Song Q (2020) A nomogram for predicting overall survival in patients with low‐grade endometrial stromal sarcoma: A population‐based analysis. Cancer Commun 40:301–312. https://doi.org/10.1002/cac2.12067
Park SH, Goo JM, Jo C-H (2004) Receiver operating characteristic (ROC) curve: practical review for radiologists. Korean J Radiol 5:11. https://doi.org/10.3348/kjr.2004.5.1.11
Van Calster B, Wynants L, Verbeek JFM, Verbakel JY, Christodoulou E, Vickers AJ, Roobol MJ, Steyerberg EW (2018) Reporting and interpreting decision curve analysis: a guide for investigators. Eur Urol 74:796–804. https://doi.org/10.1016/j.eururo.2018.08.038
Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA (2015) Robust enumeration of cell subsets from tissue expression profiles. Nat Methods 12:453–457. https://doi.org/10.1038/nmeth.3337
Xiao B, Liu L, Li A, Xiang C, Wang P, Li H, Xiao T (2020) Identification and verification of immune-related gene prognostic signature based on ssGSEA for osteosarcoma. Front Oncol 10:607622 https://doi.org/10.3389/fonc.2020.607622
Carroll MC, Holers VM (2005) “Innate Autoimmunity.” p. 137–157 https://doi.org/10.1016/S0065-2776(04)86004-8
Pressly JD, Park F (2017) DNA repair in ischemic acute kidney injury. Am J Physiol Physiol 312:F551–F555. https://doi.org/10.1152/ajprenal.00492.2016
Ruiz-Fernández C, González-Rodríguez M, Abella V, Francisco V, Cordero-Barreal A, Ait Eldjoudi D, Farrag Y, Pino J, Conde-Aranda J, González-Gay MÁ, Mera A, Mobasheri A, García-Caballero L, Gándara-Cortés M, Lago F, Scotece M, Gualillo O (2022) WISP-2 modulates the induction of inflammatory mediators and cartilage catabolism in chondrocytes. Lab Investig 102:989–999. https://doi.org/10.1038/s41374-022-00793-9
Tong Y, Ding X-B, Chen Z-X, Jin S-Q, Zhao X, Wang X, Mei S-Y, Jiang X, Wang L, Li Q (2016) WISP1 mediates hepatic warm ischemia reperfusion injury via TLR4 signaling in mice. Sci Rep 6:20141. https://doi.org/10.1038/srep20141
Tang Q, Li W, Zheng X, Ren L, Liu J, Li S, Wang J, Du G (2020) MELK is an oncogenic kinase essential for metastasis, mitotic progression, and programmed death in lung carcinoma. Signal Transduct Target Ther 5:279. https://doi.org/10.1038/s41392-020-00288-3
Speers C, Tsimelzon A, Sexton K, Herrick AM, Gutierrez C, Culhane A, Quackenbush J, Hilsenbeck S, Chang J, Brown P (2009) Identification of novel kinase targets for the treatment of estrogen receptor-negative breast cancer. Clin Cancer Res 15:6327–6340. https://doi.org/10.1158/1078-0432.CCR-09-1107
Sasako T, Ohsugi M, Kubota N, Itoh S, Okazaki Y, Terai A, Kubota T, Yamashita S, Nakatsukasa K, Kamura T, Iwayama K, Tokuyama K, Kiyonari H, Furuta Y, Shibahara J, Fukayama M, Enooku K, Okushin K, Tsutsumi T, Tateishi R, Tobe K, Asahara H, Koike K, Kadowaki T, Ueki K (2019) Hepatic Sdf2l1 controls feeding-induced ER stress and regulates metabolism. Nat Commun 10:947. https://doi.org/10.1038/s41467-019-08591-6
Shaalan AK, Teshima THN, Tucker AS, Proctor GB (2021) Inhibition of Aurora Kinase B activity disrupts development and differentiation of salivary glands. Cell Death Discov 7:16. https://doi.org/10.1038/s41420-020-00393-w
Zhu Q, Ding L, Zi Z, Gao S, Wang C, Wang Y, Zhu C, Yuan Z, Wei F, Cai Q (2019) Viral-mediated AURKB cleavage promotes cell segregation and tumorigenesis. Cell Rep 26:3657-3671.e5. https://doi.org/10.1016/j.celrep.2019.02.106
Grünberg JR, Elvin J, Paul A, Hedjazifar S, Hammarstedt A, Smith U (2018) CCN5/WISP2 and metabolic diseases. J Cell Commun Signal 12:309–318. https://doi.org/10.1007/s12079-017-0437-z

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Unveiling Pyroptosis-Related Hub Genes in Ischemic Stroke Provides Insights for Enhanced Risk Assessment

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Data Acquisition

Differential Expression Analysis of PRGs

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Enrichment Analysis

Gene Set Enrichment Analysis (GSEA)

Gene Set Variation Analysis (GSVA)

Construction of Middle Cerebral Artery Occlusion Diagnostic Model

Validation of Middle Cerebral Artery Occlusion Diagnostic Model

Construction of High and Low-Risk Groups in Middle Cerebral Artery Occlusion

Expression Differences and Correlation Analysis

Immune Infiltration Analysis

Regulatory Network Construction and Protein Domain Prediction

Statistical Analysis

Results

Analysis Flow Chart and Data Preprocessing

Integration of the Middle Cerebral Artery Occlusion Datasets

Analysis of Differentially Expressed Genes Associated with Cell Pyroptosis

GO and KEGG Enrichment Analyses

GSEA of the Integrated GEO Dataset

GSVA of the Integrated GEO Dataset

MCAO Diagnostic Model

Validation of the MCAO Diagnostic Model

GSEA of the High and Low-Risk Groups

GSVA of the High and Low-Risk Groups

Correlation and Expression Analysis of Cell Death-Related Hub Genes

Integrated GEO Dataset Immunoinfiltration Analysis

Immunoinfiltration Analysis of the High and Low-Risk Groups

Regulatory Network Construction and Protein Domain Prediction

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1