To generate a mouse model of transient middle cerebral artery occlusion (MCAO), we induced focal cerebral ischemic stroke in 8-week-old male C57BL/6 mice (20–25 g) as described previously (12). In brief, the mice were anesthetized using sevoflurane, then an incision was made to allow separation of the cervical skin from the external carotid artery. Next, a nylon monofilament was inserted into the left internal carotid artery through the origin of the middle cerebral artery, where it remained for 1 h to induce cerebral ischemia. After the ischemic episode, mice were reperfused for 24 h. A decrease in cerebral blood flow >70% was considered a successfully induced MCAO model, and this was monitored using a laser Doppler flow meter (Perimed, Switzerland) and throughout cerebral reperfusion. The preparation for the sham surgery was only separated external carotid artery but no suture was placed. The number of the mice in the sham control group is 10, and in the SHS group is 20 in 24h and 10 in other time point groups, respectively, which are the ones that survived 24 h after transient MCAO.

TTC (Merck, Darmstadt, Germany) staining was used to observe cerebral infract volume in MCAO mice. After 24 hours of reperfusion, mice were euthanized with a narcotic overdose and the brain was completely stripped. After freezing in at −20°C for 20 min, brain tissue was cut into 2-mm-thick slices. The tissue slices were then incubated in a 2% TTC solution for 15-30 min at 37°C to identify the cerebral infract.

Neurological deficits in MCAO mice were assessed using the modified neurological severity score (mNSS) test, as described previously (12). The highest score for the mNSS test is 18, and higher scores indicate more severe neurological deficit.

Serum cTnI in peripheral blood samples was measured using a mouse TNNI3/cTn-I (Troponin I Type 3, Cardiac) ELISA kit (Elabscience, Wuhan, China). According to the manufacturer’s protocol, after 100 μL each of blank, standard and sample were pipetted into individual wells and incubated at 37°C for 90 min, 100 μL of biotinylated detection Ab working solution was added to each well and incubated at 37°C for 1 h. After 3 washes, 100 μL horseradish peroxidase (HRP) Conjugate Working Solution was added to each well followed by another incubation at 37°C for 30 min. Then samples were washed 5 times and 90 μL of Substrate Reagent was added to each well. After a 15-minincubation at 37°C, 50 μL of Stop Solution was added to each well and the optical density (OD value) of each well was measured using a microplate reader (Molecular Devices, USA) set to 450 nm.

After the mice were euthanized by excessive anesthesia, they were infused with 4% paraformaldehyde in the heart, then the hearts were collected and fixed in 4% paraformaldehyde. Subsequently, the hearts were dehydrated, paraffin embedded, cut into 3-mm-thick slices, and baked at 72°C for 1 h. Heart tissue slices were placed at room temperature for 10 minutes before dewaxing with water, and antigens were retrieved using EDTA (pH9.0) in a high-pressure chamber. Next, the tissue endogenous peroxidase was blocked with 3% hydrogen peroxide (ZSGB-Bio, Beijing, China) for 10 min at room temperature, followed by incubation with anti-CD45 antibody (Cell Signaling Technology (CST), MA, USA, 1:50 dilution) at 4°C overnight in the dark. The tissue was then incubated with goat anti-mouse Alexa Fluor 594 (Thermo Fisher, MA, USA, 1:500 dilution) at room temperature for 1 h and nuclei were stained with 4’,6-diamidino-2-phenylindole (DAPI) (Solarbio, Beijing, China). Tissue was mounted on slides with a coverslip and photographed using a laser-scanning confocal microscope (FluoView 1200, Olympus, Japan).

To analyze the pathophysiological process of cardiac damage induced by ischemic stroke, we downloaded an expression profiling dataset (GSE102558) from Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), which was contributed by R. Veltkamp and based on the GPL20710 platform of [MoGene-2_0-st] Affymetrix Mouse Gene 2.0 ST Array (mogene20st_Mm_ENTREZG_19) and analyzed the expression data of heart tissue 24 hours after MCAO or sham surgery.

DEGs were identified using the GEO2R (http://www.ncbi.nlm.nih.gov/geo/geo2r) online web tool provided by the GEO database. The parameters |log₂fold change|≥1 and adjusted P value <0.05 were used as cut-offs for defining DEGs. A heatmap of the DEGs was constructed using the R package, “ComplexHeatmap”.

The Metascape (http://metascape.org/gp/index.html) online web tool was used to conduct Gene Ontology (GO) enrichment analysis, which includes molecular function (MF), biological process (BP), and cellular component (CC), as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (13, 14). The data were analyzed using the R packages “clusterProfiler” and “GOplot” and visualized by the R package “ggplot2”. P<0.05 and false discovery rate (FDR) <0.25 were set as the cut-off values.

To evaluate infiltrated immune cells propotions in cardiac tissue, the online bioinformatics algorithms CIBERSORTx (https://cibersortx.stanford.edu/index.php) (15) and mice leukocyte gene signature matrix (16) were used. Correlation analysis and visualization of immune cells were conducted using the R package, “ggplot2”.

Inflammatory-related genes (IRGs), named “HALLMARK_INFLAMMATORY_RESPONSE” were obtained from the Molecular Signatures Database (MSigDB) (http://www.gsea-msigdb.org/) (17). The online tool Venn (http://DrawVennDiagram(ugent.be)) was used to identify differentially expressed IRGs (DIREGs) by identifying genes in the intersection of the Venn diagram. Correlations between DIREGs and visualization were analyzed using the R packages “igraph” and “ggraph”.

TRIzol Reagent (Thermo Fisher, MA, USA) and reverse transcriptase (TransGen, Beijing, China) were used to extract total RNA from mice hearts and reverse-transcribe RNA into cDNA, respectively. A SYBR Green qPCR kit (TransGen, Beijing, China) was used to amplify Aplnr, Ccrl2, Cdkn1a, Irak2, Serpine1 and Gapdh by RT-qPCR. The PCR system include 0.5 μL of cDNA, 0.1 μL each of upstream and downstream primers (100 μM), 7.5 μL of 2×SYBR Green qPCR reagent, and 6.8 μL of ddH₂O. Three-step PCR program was carried out as follows: preheating program is at 94°C for 3 min, amplification program include forty cycles of denaturation at 95°C for 15 s and annealing at 62°C, and elongation program is at 75°C for 3 min. The relative expression level of each gene was calculated using the 2^-ΔΔCT method. The primers used were listed in Table 1 .

Samples used for Western Blot analysis were prepared following methods in our previous study (12). Briefly, after grinding, a portion of the heart tissue powder was incubated in RIPA buffer with a combination of phosphatase and protease inhibitors to lyse cells, and abicinchoninic acid (BCA) kit (Solarbio, Beijing, China) was used to quantify protein concentration. Then, an equal volume of each sample was denatured for 5 min at 95°C in loading buffer, and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Gels were transferred to a polyvinylidene fluoride membrane (PVDF), which were then blocked for 2 h in 5% bovine serum albumin (BSA) and incubated overnight at 4°C with the following primary antibodies: anti-IRAK2 (Cell Signaling Technology (CST), MA, USA, 1:1000 dilution), anti-APLNR, anti-CDKN1A, anti-SERPINE1 or anti-GAPDH (Proteintech, Rosemont, USA, 1:500, 1:1000, 1:10000,1:10000 dilution respevtively). Membranes were then washed three times with phosphate-buffered saline containing Tween 20 (PBST), and incubated for an hour with horseradish peroxidase (HRP) conjugated secondary antibody (Proteintech, Rosemont, USA, 1:10000 dilution). Immunoblots were visualized using a gel imaging system (Fusion FX7, ViIbertLourmat, France).

All data were processed using SPSS 22.0 statistical software. Comparisons between two groups were performed using an independent samples t-test, and correlation analyses were performed using Pearson’s method. One-way analysis of variance (ANOVA) was used for multiple-group comparisons, and the least significant difference post-hoc test was used for intergroup comparisons. P values < 0.05 were considered statistically significant.

To investigate molecular changes leading to SHS, MCAO was induced in mice for 1 h followed by reperfusion for 24 h. After the reperfusion period, we determined infarct volumes and neurological deficit scores. MCAO-treated mice had a significantly higher cerebral infarct volume and neurological deficit scores compared with sham-treated mice ( Figures 1A, B ). To examine whether MCAO induced cardiac damage, we next examined the level of serum cTnI, a marker of acute myocardial injury, which showed that MCAO-treated mice had significantly higher serum cTnI compared with sham-treated mice ( Figure 1C ). Furthermore, IF analysis revealed significantly increased leukocyte infiltration in the myocardial tissue of MCAO-treated mice compared with sham-treated mice ( Figures 1D, E ). Collectively, these data indicate that MCAO was successfully induced in mice and that MCAO-treated mice suffered acute cardiac dysfunction and inflammation.

To identify pathways that may contribute to SHS, we next evaluated gene expression profiles from a publicly available database of RNA expression profiling data of myocardial tissue from MCAO- and sham-treated mice after 24 h of perfusion. Applying the standard criteria of |log₂fold change| ≥1 and adjusted P value < 0.05, we identified 138 DEGs, including 91 upregulated genes and 47 downregulated genes in MCAO-treated mice versus sham-treated mice ( Figures 2A, B ). The box diagram indicates that the data has been standardized and is suitable for analysis ( Figure 2C ). The top 40 DEGs are shown in Figure 2D .

To understand the biological relevance of the DEGs for the development of SHS, we first analyzed the DEGs using the Metascape online tool (http://metascape.org). This showed that the DEGs were predominantly enriched in the regulation of endopeptidase activity, response to hormone, extracellular matrix organization, positive regulation of cell death, and entrainment of circadian clock by photoperiod ( Figures 3A, B ). The enriched ontology clusters colored by p-value is shown in Figure 3C . We then performed GO analysis on the DEGs and visualized the results using chord diagrams, which not only display the item, but also display the molecules contained within the item. In agreement with the MetaScape analysis, the BP changes of DEGs were significantly enriched in regulation of peptidase activity (GO:0052548), which includes the genes Fas, Nr4a1, Serpina3n, Serpine1, and others ( Figure 3D ). The response to oxidative stress (GO:0006979), negative regulation of immune system process (GO:0002683), lymphocyte proliferation (GO:0046651), and regulation of inflammatory response (GO:0050727) were also significantly enriched. The CC changes of DEGs were most enriched in collagen-containing extracellular matrix, membrane raft, membrane microdomain, apical part of cell, and mitochondrial inner membrane ( Figure 3E ). The MFs that were most enriched among the DEGs were enzyme inhibitor activity, endopeptidase regulator activity, peptidase regulator activity, endopeptidase inhibitor activity, and peptidase inhibitor activity ( Figure 3F ). Finally, we analyzed the DEGs using KEGG, which showed that the most enriched pathways were autophagy–animal (mmu04140), p53 signaling pathway (mmu04115), platinum drug resistance (mmu01524), biosynthesis of unsaturated fatty acids (mmu01040), and circadian rhythm (mmu04710) ( Figure 3G ). The terms of each category were listed in Table 2 .

Based on our experimental findings and the gene expression analysis, we next evaluated the immune cell infiltrate in cardiac tissues of MCAO- and sham-treated mice. We found that the proportion of resting natural killer (NK) cells was significantly lower in MCAO-treated compared with sham-treated mice (P=0.0263), whereas the proportion of activated NK cells was significantly higher compared with the sham-treated group (P=0.0043). Also, the proportion of B memory cells, M0 macrophages, and CD4⁺ naive T cells were significantly increased and the proportion of naïve B cells, CD4⁺ follicular T cells, and immature dendritic cells (DCs) were significantly decreased in MCAO-treated compared with sham-treated mice ( Figures 4A, B ). Correlation analysis of 25 types of immune cells determined that the strongest positive correlation was between activated DC cells and CD8⁺ naive T cells (r=1, P=0) and that the strongest negative correlation was between memory B cells and naive B cells (r=−0.85, P<0.001) ( Figure 4C ).

Next, we evaluated which inflammatory response genes may contribute to SHS. We identified 5 IRGs among our DEGs (DEIRGs): Aplnr, Ccrl2, Irak2, Cdkn1a, and Serpine1 ( Table 3 ). All of the DEIRGs except Aplnr were significantly upregulated in MCAO myocardium compared with sham myocardium. Correlation analysis showed that Ccrl2 and Irak2 were highly positively correlated (r=0.902), whereas Aplnr and Cdkn1a were highly negatively correlated (r=−0.860) ( Figure 5 ).

We used Pearson’s rank correlation analysis to explore the association between DEIRGs and the immune infiltrate in cardiac tissue of MCAO-treated mice. Aplnr expression was positively correlated with the proportion of infiltrating resting NK cells and negatively correlated with the proportion of infiltrating activated NK cells ( Figure 6A ). Ccrl2 expression was positively correlated with the proportion of infiltrating eosinophil cells and negatively correlated with the proportion of infiltrating plasma cells ( Figure 6B ). Cdkn1a, Irak2, and Serpine1 expression were positively correlated with the proportion of infiltrating activated NK cells and negatively correlated with the proportion of infiltrating resting NK, plasma cells and immature DC respectively ( Figures 6C–E ).

To validate the DEIRGs identified and further characterize their expression patterns during cardiac dysfunction, we performed RT-qPCR-based analysis of Aplnr, Ccrl2, Cdkn1a, Irak2, and Serpine1 transcription levels at 1, 3, 6, 12 and 24 h after MCAO. The results showed that these five genes exhibited significant changes in expression in heart tissue following MCAO. It begins within 1 h after MCAO, continuously changing over the 24 h observation period.

Among them, Aplnr expression significantly decreased in MCAO-treated mice to 50% that of the sham group by 1 h after MCAO and continued to decrease thereafter. By 24 h, Aplnr transcription was approximately 30% of that detected in the sham mice ( Figure 7A ). By contrast, Ccrl2 increased in the early stage of reperfusion, reaching >2-fold that of sham mice by 1 h post-MCAO, and peaking at 3.5-fold higher levels at 3 h, before decreasing to similar levels as sham controls by 24 h post-MCAO ( Figure 7B ). Cdkn1a and Irak2 underwent similar increases, up to 5- and 2.7-fold of their expression compared to sham mice within 1 h after MCAO, respectively, and steadily increased until 12 h after MCAO. Notably, average Cdkn1a expression in the MCAO group reached approximately 25-fold its levels in comparison to the sham group, whereas Irak2 transcript levels peaked at >5-fold that of the sham group. Although gradually decreasing by 24 h post-MCAO, transcript levels of Cdkn1a and Irak2 were greater than 4.7 and 2.7 times, respectively, in MCAO mice compared to sham mice at the final time point ( Figures 7C, D ). However, Serpine1 followed a different trend, decreasing to approximately 70% and 50% of its levels in comparison to sham mice at 1 h and 3 h, respectively. Serpine1 transcription then increased, reaching approximately 1.2-, 2.2- and 2.4-fold that of its expression in the sham group at 6 h, 12 h, and 24 h after MCAO, respectively ( Figure 7E ).

These results verified that the expression of these inflammation-related genes indeed fluctuates significantly after MCAO, suggesting that MCAO-induced ischemic stroke may share a close relationship with cardiac injury. Although Ccrl2 expression in MCAO mice did not significantly differ from its expression in comparison to the sham group at 24 h, the other four genes all had significantly different transcription levels from that of the sham group, leading us to classify them as target genes.

To further determine whether Aplnr, Cdkn1a, Irak2, and Serpine1 could serve as candidate target genes, we examined their protein levels in heart tissue at 24 h post-reperfusion in MCAO- and sham-treated mice. The results showed that APLNR protein levels were significantly lower, while CDKN1A, IRAK2 and SERPINE1 accumulation was significantly higher in MCAO-treated heart tissue compared to that of sham controls ( Figure 8 ).