Identification and characterization of biomarkers associated with endoplasmic reticulum protein processing in cerebral ischemia-reperfusion injury

Data acquisition and pre-processing

The GSE163614 dataset was retrieved from the National Center for Biotechnology Information (NCBI) database. ENSRNOG identifiers were mapped to gene symbols utilizing the annotation file Rattus_norvegicus.mRatBN7.2.107.gtf. This dataset comprised three CI/reperfusion injured rats and three control rats (Yi et al., 2021).

Datasets GSE78731 (Oh et al., 2018) and GSE97537 (Zou et al., 2019) were also sourced from the Gene Expression Omnibus (GEO) within the NCBI database. Probe identifiers were transformed into gene symbols using the platform-specific file. Entries with a single probe mapping to multiple genes were excluded. When multiple probes were mapped to one gene, their data were averaged. From GSE78731, five Middle Cerebral Artery Occlusion (MCAO) and five control samples were extracted, and from GSE97537, seven MCAO and five sham samples were derived.

Differential gene expression analysis

Employing the ‘limma’ package in R, differential expression between the three CI/reperfusion injured rats and three controls in the GSE163614 dataset was conducted. DEGs were identified using a threshold of |log2(Fold Change)| > log2(1.2) and FDR < 0.05. This analysis was followed by the generation of volcano and heat maps.

Isolation of PER-related genes

PER-related pathways were sourced from the single-sample Gene-Set Enrichment Analysis (ssGSEA) official website. PER scores were computed via ssGSEA, and the t-test function was applied to compare the MCAO and sham groups. Associations between the PER score and DEGs were determined using R’s ‘cor’ function. Genes exhibiting |cor| > 0.7 were designated as PER-associated DEGs. These genes were further assessed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment via the clusterProfiler package.

Identification of central genes in PER-related genes

A protein-protein interaction (PPI) network of the PER-related DEGs was constructed using the STRING database. Essential interactions were discerned using a minimum score of 0.7. Visualization and degree analysis of the PPI were executed using Cytoscape.

Identification of diagnostically relevant per-associated genes

For datasets GSE78731 and GSE97537, crucial genes linked to PER were scrutinized. Receiver operating characteristic (ROC) curves were produced for these DEGs, and their diagnostic efficacy was evaluated based on the Area Under the Curve (AUC).

Correlation of key PER genes with hypoxia

To further explore the possible mechanisms of CI-reperfusion, the HIF-1 signalling pathway gene set was downloaded from the KEGG official website. HIF-1 signalling scores for MCAO and sham groups were determined using the ssGSEA approach. The correlation between key diagnostic genes and the HIF-1 pathway score was analyzed via the Spearman correlation method.

Cell cultivation

The PC12 cell line was sourced from the Cell Bank of the Shanghai Institute of Cell Biology, affiliated with the Chinese Academy of Sciences. The cells were cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Gibco, Grand Island, NY, USA) at 37 °C. Under normoxic conditions, PC12 cells were maintained in an atmosphere consisting of 95% air and 5% CO2. For simulating ischemic conditions, an oxygen and glucose deprivation/reperfusion (OGD/R) model was employed. Cells were exposed to glucose-free DMEM in a hypoxic environment of 95% N2 and 5% CO2 for a duration of 2 h. Post hypoxia, cells were reintroduced to regular glucose conditions and an oxygenated environment for a recovery period of 24 h.

qRT-PCR analysis

For assessing gene expression alterations, quantitative reverse transcription PCR (qRT-PCR) was utilized. Total RNA from PC12 cells was extracted and then reverse-transcribed using the Takara RNA PCR Kit (AMV) Ver 3.0 (TaKaRa Bio Inc., Shiga, Japan). The synthesized cDNA was subjected to PCR amplification using SYBR Green I as the fluorescent dye on an ABI PRISM 7300 real-time PCR system. The thermal cycling conditions were initiated with a denaturation step at 95 °C for 30 s, followed by 45 amplification cycles, each consisting of 95 °C for 10 s (denaturation) and 62 °C for 31 s (annealing/extension). To ensure reproducibility and accuracy, all experiments were performed in duplicate and repeated twice. The specific primer sequences used for each target gene are provided in Table 1. All primer sequences were validated by BLAST.

Screening of genes associated with CI/reperfusion injury and investigation of their relationship with PER-associated pathways

The main technical roadmap of this study is shown in Fig. S1. Differentially expressed genes (DEGs) from three rats with CI/reperfusion injury and three control rats in the GSE163614 dataset were analyzed. Both volcano (Fig. 1A) and heat maps were generated. Using the anno_GO_keywords function, an enrichment analysis word cloud (Fig. 1B) was produced, revealing 651 upregulated genes (red) and 446 downregulated genes (blue). In addition, the top 10 genes with the most significant differences in expression up- and down-regulation were summarised in Table S1, respectively. The ssGSEA analysis indicated a stronger association of CI/reperfusion injury with PER in the MCAO group, as evidenced by a higher PER score (Fig. 1C).

Functional enrichment analysis of PER-associated gene pathways

DEGs were correlated with the PER score, and gene sets related to a |cor| > 0.7 threshold were isolated. This resulted in 648 genes positively correlated with the PER score and 446 negatively correlated genes (Table S2). Subsequent GO and KEGG enrichment analyses (using an FDR < 0.05 threshold) identified 1056 biological processes (BP), 132 cellular components (CC), 42 molecular functions (MF), and 73 KEGG signalling pathways. The top 10 enriched items are displayed as a bar graph in Fig. 2, the more the colour tends to be red, the smaller the p.adjust value, indicating a higher degree of enrichment. Importantly, the PER-associated genes showed significant involvement in neurotraumatic processes, such as leukocyte migration, wound healing, and angiogenesis regulation (Table S3).

Identification of central genes in PER-associated gene set

The Protein-Protein Interaction (PPI) network of the PER-associated genes was built via the STRING database (Mering et al., 2003) to predict potential biological functions among proteins and further explored using Cytoscape (Fig. 3A), followed by a degree analysis (Fig. 3B). The 20 most pivotal nodes, both positively and negatively correlated, were highlighted as key genes (Table 2). A node with a higher degree value suggests its centrality within the network.

Diagnostic potential of key PER-associated genes

For diagnostic relevance, we examined the key PER genes across the independent datasets GSE78731 and GSE97537. Seventeen of the top 20 genes showcased significant degree correlations (Figs. 4A and 4B, Table 2): ITGB1, RPS5, RPS3A, RPS27, RPL11, RPL5, RPS27L, FN1, PTPRC, RPLP0, STAT3, RPL6, RPL10A, RPL26, RPL27, RPSA, RPS15, RPS24, RPS4X, TP53, DLG4, PRKCB, GRIN2A, GNAO1, THY1, NTRK2, PRKCA, GRIA3, PLCB1, AGT, CAMK2A, DLG1, SHC3, HSPA4L, PAK1, ADCY5, APOE, NR3C1, ADRB1, RGS8. ROC curves of these 17 genes evaluated their diagnostic efficacy (Figs. 4C and 4D), with each gene demonstrating AUCs exceeding 0.7, suggesting their diagnostic promise. Further validation in the GSE163614 dataset supported these findings (Fig. 5).

Evaluating the association of diagnostic key PER genes with hypoxia

Decreased cerebral blood flow can instigate hypoxia, precipitating ischemic neuronal injury and favouring anaerobic tissue metabolism (Kalogeris et al., 2012). To contextualize hypoxia, the HIF-1 signalling pathway was extracted from KEGG. Its scores across groups were computed using ssGSEA, revealing an elevated score in the MCAO group (Fig. 6A). Correlation analyses between the 17 diagnostic genes and the HIF-1 signalling pathway score identified six genes—PRKCB, ADCY5, CAMK2, PLCB1, NTRK2, and DLG4—that displayed significant associations with hypoxia (Fig. 6B).

Investigating the interplay between key hypoxia-linked genes and apoptosis

Exacerbated endoplasmic reticulum stress (ERS) activation during CI/reperfusion can stimulate apoptosis via pathways such as CHOP, Caspase12, and JNK (Wang et al., 2022). We probed the connections between crucial hypoxia-associated genes and apoptosis. The MCAO group displayed a marked under-expression of these genes compared to the Sham group (Fig. 7). Subsequent analyses showed an elevated apoptosis score in the MCAO group (Fig. 8A). Furthermore, ADCY5, CAMK2A, PLCB1, NTRK2, and DLG4 showed significant negative correlations with apoptosis score (Fig. 8B). These genes appear to be instrumental in hypoxia-induced apoptosis during CI/reperfusion injury. PCR-based validation further highlighted their reduced expression in cells subjected to glucose deprivation/reperfusion (OGD/R), which suggests their consistent downregulation during such metabolic stress (Fig. 9).