Reveal the correlation between hub hypoxia/immune-related genes and immunity and diagnosis, and the effect of SAP30 on cell apoptosis, ROS and MDA production in cerebral ischemic stroke

Abstract

Background: Cerebral ischemic stroke (CIS) is a common cerebrovascular disease. The purpose of this study was to investigate the potential mechanism of hypoxia and immune-related genes in CIS.

Methods: All data were downloaded from public databases. Hub mRNAs was identified by differential expression analysis, WGCNA analysis and machine learning. Hub mRNAs were used to construct the classification models. Pearson correlation analysis was used to analyze the correlation between hub mRNAs and immune cell infiltration. Finally, the SAP30 was selected for verification in HMC3 cells.

Results: The SVM, RF and DT classification models constructed based on 6 hub mRNAs had higher area under the curve values, which implied that these classification models had high diagnostic accuracy. Pearson correlation analysis found that Macrophage has the highest negative correlation with CCR7, while Neutrophil has the highest positive correlation with SLC2A3. Drug prediction found that ruxolitinib, methotrexate, resveratrol and resatorvid may play a role in disease treatment by targeting different hub mRNAs. Notably, inhibition of SAP30 expression can reduce the apoptosis of HMC3 cells and inhibit the production of ROS and MDA.

Conclusion: The identification of hub miRNAs and the construction of classification diagnosis models provide a theoretical basis for the diagnosis and management of CIS.

Keywords: SAP30; cerebral ischemic stroke; hypoxia; immune.

Figure 1

Volcanic maps of DEmRNAs. (A) Volcanic map of DEmRNAs in the GSE58294 dataset; (B) Volcanic map of DEmRNAs in the GSE16561 dataset.

Figure 2

Identification of hub modules and candidate hub mRNAs based on WGCNA. (A) Sample clustering dendrogram to detect outliers; (B) Sample clustering dendrogram and trait heatmap; (C) Scale-free fitting index and average connectivity for different soft threshold power (β); (D) mRNA is divided into different modules by hierarchical clustering, and different colors represent different modules; (E) Modules with dissimilarity <25% are merged; (F) Heatmap of correlation between ME and CIS; (G) Scatter plot of mRNAs in green module; (H) Scatter plot of mRNAs in red module.

Figure 3

Identification and functional enrichment analysis of intersection mRNAs. (A) Venn diagram of intersection of the DEmRNAs in the GSE58294 dataset, the DEmRNAs in the GSE16561 dataset, the candidate key mRNAs in the WGCNA and the set of IRGs and HRGs; (B) Venn diagram of the intersection mRNAs, IRGs and HRGs; (C) GO functional enrichment analysis of intersection mRNAs; (D) KEGG functional enrichment analysis of intersection mRNAs; (E) A PPI network was constructed based on STRING database to study the regulatory relationship between intersection mRNAs.

Figure 4

Identification of hub mRNAs and construction of SVM, RF and DT classification models. (A) LASSO regression analysis was performed on 26 intersection mRNAs; (B) Mean decreased accuracy sorting of SAP30, S100A12, CCR7, SLC2A3, JAK2, TLR4 and IL32; (C) Trend chart of AUC with the increase of DEmRNA quantity; (D) Correlation between SAP30, S100A12, CCR7, SLC2A3, JAK2 and TLR4. Red and blue represent positive and negative correlations, respectively. (E) Expression heatmap of SAP30, S100A12, CCR7, SLC2A3, JAK2 and TLR4; (F) ROC curve of SVM classification model in GSE58294 dataset; (G) ROC curve of RF classification model in GSE58294 dataset; (H) ROC curve of DT classification model in GSE58294 dataset.

Figure 5

ROC curve validation of SVM (A), RF (B) and DT (C) classification models in GSE16561 dataset.

Figure 6

Analysis of immune cell infiltration. (A) The level of immune cell infiltration was analyzed by ssGSEA method; (B) Stacked histogram of the proportion of each immune cell in the sample analyzed by CIBERSORT method; (C) Box diagram of the proportion of each immune cell in the sample analyzed by CIBERSORT method; (D) Correlation between hub mRNAs and immune cell infiltration. Red and blue represent positive and negative correlations, respectively. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001, Abbreviation: ns: no significant significance.

Figure 7

Construction of TFs regulatory network and ceRNA regulatory network. (A) TFs regulatory network; (B) Volcano map of DEmiRNAs in the GSE95204 dataset; (C) CeRNA regulatory network.

Figure 8

Drug prediction and molecular docking of hub mRNAs. (A) Drug prediction of hub mRNAs; (B) Molecular docking of methotrexate and S100A12; (C) Molecular docking of methotrexate and TLR4; (D) Molecular docking of resatorvid and TLR4; (E) Molecular docking of resveratrol and SLC2A3; (F) Molecular docking of ruxolitinib and JAK2.

Figure 9

The relative content of SAP30, ROS and MDA and the apoptosis rate in HMC3-OGD/R model group and HMC3 control group. (A) The relative expression level of SAP30 in HMC3-OGD/R model group and HMC3 control group was detected by real time-PCR; (B) Fluorescence value of ROS in HMC3-OGD/R model group and HMC3 control group; (C) Content of MDA in HMC3-OGD/R model group and HMC3 control group; (D) Apoptosis rate in HMC3 control group was detected by flow cytometry; (E) Apoptosis rate in HMC3-OGD/R model group was detected by flow cytometry; (F) Histogram of apoptosis rate in HMC3-OGD/R model group and HMC3 control group. ^****P < 0.0001; Abbreviations: HMC3-NC: HMC3 control group; HMC3-OGD/R HMC3-OGD/R model group. Q1-2 and Q1-4 quadrants represent late apoptotic cells and early apoptotic cells, respectively.

Figure 10

The relative content of SAP30, ROS and MDA and the apoptosis rate in HMC3-si-NC-OGD/R group and HMC3-si-SAP30-OGD/R group. (A) Real time-PCR was used to detect the expression of SAP30 to screen out effective interference targets in HMC3 cell; (B) The relative expression level of SAP30 in HMC3-si-NC-OGD/R group and HMC3-si-SAP30-OGD/R group; (C) Fluorescence value of ROS in HMC3-si-NC-OGD/R group and HMC3-si-SAP30-OGD/R group; (D) Content of MDA in HMC3-si-NC-OGD/R group and HMC3-si-SAP30-OGD/R group; (E) Apoptosis rate in HMC3-si-NC-OGD/R group was detected by flow cytometry; (F) Apoptosis rate in HMC3-si-SAP30-OGD/R group was detected by flow cytometry; (G) Histogram of apoptosis rate in HMC3-si-NC-OGD/R group and HMC3-si-SAP30-OGD/R group. ^*P < 0.05, ^**P < 0.01, ^***P <0.001, ^****P < 0.0001. Q1-2 and Q1-4 quadrants represent late apoptotic cells and early apoptotic cells, respectively.