Comprehensive analysis of macrophage-associated inflammatory genes in AMI based on bulk combined with single-cell sequencing data

Data collection

Single-cell data were collected from the GSE163465 dataset in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). This dataset consists of one sample from mice in the AMI group and one sample from mice in the Sham group. Bulk data were collected from the GSE236374 and GSE183272 datasets in the GEO database. The GSE236374 dataset contains three healthy Sham samples and three samples with AMI. The GSE183272 dataset contains five samples with AMI and five healthy Sham samples. The GSE114695 dataset includes six healthy Sham samples and six samples with AMI. Healthy samples from the dataset and samples from day seven after AMI were used for this study.

Cell clustering analysis

Nineteen cell clusters were obtained after UMAP dimensionality reduction clustering of single-cell data (GSE163465) using the Seurat package (v 4.3.0.1) and cell-type annotation using the SingleR package (v 2.2.0).

Identification of AMI-associated differentially expressed genes (DEGs)

To identify AMI-associated DEGs, differential analysis was conducted on samples from the training dataset using the FindMarkers function in the Seurat software package. The established thresholds were set at p-value <0.05 and —log2FC—>0.5.

Analysis of intercellular communication

To understand the AMI network, the CellChat software package (v 1.6.1) was used to assess the strength and probability of cell–cell communication in AMI mice. Circle plots were used to show the number and strength of interactions between any two cell types within each cellular signaling pathway. Subsequently, the strength and contribution of interactions between dominant cells and other cells was further investigated.

Identification of DEGs and Co-DEGs

All the data from the GSE236374 dataset and the GSE183272 dataset were re-normalized using the sva package v3.52.0 for batch effect removal. The threshold of DEG screening was p-value <0.05 and —log2FC—>0.5. DEGs (n = 80) identified in the Sc-RNA dataset were overlapped with those (n = 1, 907) identified in the bulk-RNA dataset, forming the Co-DEG for subsequent analysis.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) and Wiki pathway enrichment analysis

Enrichment analysis of the Co-DEGs was performed using the DAVID database (https://david.ncifcrf.gov/) and the results obtained were visualized using the ggplot2 package (v 3.4.4), highlighting the pathways ranking in the top list. Likewise, for the top 3 Co-DEGs, all functionally enriched pathways of significantly expressed genes were determined using Wiki pathway databases.

Screening of key genes and immune infiltration analysis

The inflammatory response score for both the AMI and the Sham groups was calculated using the gene set variation analysis (GSVA) algorithm. The key genes related to inflammation-related pathway and macrophage M0 and M2 scores were selected for further analysis. The screened key genes underwent Pearson correlation analysis with genes in inflammatory response pathways to assess the extent of correlation between these key genes and genes associated with inflammation-related pathway.

Validation of the expression of the screened genes in the original datasets

The GSE163465 and GSE236374 datasets were used for gene expression validation. Differential expression of key genes in the AMI and the Sham groups was compared using rank sum test of the ggpubr package (v 0.6.0).

Dataset validation and single-gene Gene Set Enrichment Analysis (GSEA)

The GSE114695 dataset was used for gene expression validation, and GSEA of key genes was conducted using the clusterProfiler package (v 4.8.3; https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html).

Construction of AMI mouse model

Twelve clean-grade SD male mice, weighing 18–22 g each, were purchased from Hangzhou Medical College Laboratory Animal Center. All mice were reared adaptively for two weeks under pathogen-free condition with 20−24 °C and free access to food and water under natural light with day and night. Mice were randomly divided into Sham group (n = 6) and AMI group (n = 6). Mice were used for constructing AMI model by ligating the left anterior descending (LAD) coronary artery (Peng et al., 2023) and anesthetized with 5% isoflurane.The mice heart was exposed by creating a transverse chest incision, and the LAD branch was ligated using a puncture needle in the interventricular groove, approximately 2–3 cm below the left auricle. The myocardium displayed a change in color, becoming white or cyanotic after ligation. Subsequently, the heart was repositioned into the chest cavity, followed by closure of the chest and suturing of the skin. The Sham group mice underwent only thoracotomy without LAD ligation. After 1 week of LAD ligation in mice, the cardiac function in the left ventricle was assessed using M-mode echocardiography. The mice were anesthetized and secured in a supine position on a wooden board. The chest was fully exposed, and the chest fur removed with hair removal cream. An ultrasound instrument was adjusted to display an M-mode echocardiogram. Left ventricular end-diastole diameter (LVEDD), left ventricular end-systolic diameter (LVESD), left ventricular fraction shortening (LVFS), and left ventricular ejection fraction (LVEF) of the mice were collected for each group, which were used to assess the level of the cardiac function of the left ventricles of the mice. At the end of treatment, the mice were euthanized by 5% inhalation of isoflurane and cervical dislocation. The heart tissues of mice were isolated and frozen. All study procedures were approved by Institutional Animal Care and Use Committee of Zhejiang Center of Laboratory Animals (NO. ZJCLA-IACUC-20010546).

Calculation of myocardial infarct area in mice by triphenyltetrazolium chloride (TTC) staining

Surviving mice were euthanized through cervical dislocation 1 week after successful AMI modeling. Following anesthesia for blood sampling from the abdominal aorta, fresh hearts were removed, thoroughly washed with saline, and then placed in a refrigerator at −20 °C for about 20 min. Afterward, the hearts were removed from the refrigerator, sliced into 2–3 mm-thick slices and placed in 2% TTC staining solution prewarmed at 37 °C. After 15 min of staining, it was observed that the infarcted area of the mouse heart appeared white, whereas the non-infarcted area was red. Image J (v 1.51 J8) was used for analysis. The percentage of the white infarcted area to the whole section area was used as an index.

Detection of related gene expression in mouse cardiac macrophages by quantitative reverse transcription polymerase chain reaction (qRT-PCR)

The method for isolating mouse macrophages involved excising the mouse hearts, washing the hearts with heated Hanks Balanced Salt Solution (HBSS) to remove blood, preparing cell suspensions through chopping, homogenizing and digesting, discarding the supernatant after centrifugation, and lysing erythrocytes with ACK lysis buffer (A10492-01; Gibco). After washing, the cells were resuspended in 2 ml HBSS and incubated with 50 ul anti-CD45 (mouse) magnetic beads (Miltenyi 130-052-301) for 30 min. Subsequently, cells were passed through the MACS magnetic column with magnets in place. The column was rinsed with buffer, and labeled cells were eluted for RNA collection (Carlson et al., 2017).

Transfection efficiency was determined using qRT-PCR (Zhang et al., 2020). RNA from mouse cardiac macrophages in both the AMI and the Sham groups were extracted using Trizol solution (15596026; Thermo Fisher Scientific, USA) and reverse transcribed into cDNA (RR037Q; Takara, Beijing, China). The expression levels of Saa3, Ms4a4c, Fcgr4, and Acp5 were detected using qPCR (CN830S, Takara) with the following reaction conditions were 95 °C for 10 min, 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min, for a total of 40 cycles. The expression levels of Saa3, Ms4a4c, Fcgr4, and Acp5 were calculated using the 2^−ΔΔct method. GAPDH was used as the internal marker gene. The primer sequences were designed as follows (Table 1).

Statistical analysis

Statistical analysis was conducted using the R program (v 4.0.2; R Core Team, 2020). Data for different genes in the two groups were compared using t-test with GraphPad Prism (v 10.1.0) software, with P < 0.05 considered statistically significant.

Analysis of scRNA-seq data

Single-cell data from the GSE163465 dataset were subjected to Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction, revealing 19 different clusters in the AMI group, including B cells, basophils, and macrophages (Figs. 1A, 1B). Analysis of the percentage of each cell type in the AMI group compared to the Sham group indicated obvious differences in the percentage of different clusters between the two groups (Fig. 1C, Fig. 1D). DEG analysis on the GSE163465 dataset revealed a total of 80 DEGs between the AMI and the Sham groups, as shown in the volcano plots. (Screening criteria: threshold adjP<0.05, and —log₂FC—>0.5) (Fig. 1E).

Analysis of intercellular communication in AMI

The CellChat package was used to analyze scRNA-Seq data for understanding potential intercellular communication in AMI mice from the GSE163465 dataset (Fig. 2A, Fig. 2B). Results indicated macrophages as the dominant contributors to cellular communication, closely interacting with many cells,

The CCL signaling pathway was shown to be closely involved in the interaction between macrophages and monocytes (Fig. 2B). Further investigation of the CCL signaling pathway was conducted with CellChat for better undrstanding cell-to-cell communications (Fig. 2C). Ccl6-Ccr2 and Ccl5-Ccr5 were identified as the main contributor of Ligand-Receptor (L −R) for the CCL signaling pathway. (Fig. 2D). In addition, analyzing the probability of communication for ligand–receptor pairs sent by macrophages to other cell types showed a substantial increase in Ccl 6-Ccr 2 from macrophages to monocytes, and in Ccl 4-Ccr 5 and Ccl 3-Ccr 5 from neutrophils to macrophages (Figs. 2E, 2F). The expression of chemokines and chemokine receptors on the major cell types was further investigated, revealing that the chemokines ccl6 and ccl9 and the chemokine receptor ccr5 can function in macrophages (Fig. 2G).

Functional enrichment analysis of intersecting Co-DEGs

In the analysis of T bulk sequencing data from the GSE236374 and GSE183272 datasets, DEGs between the AMI and the Sham groups were identified. With thresholds of p-value <0.05 and —log2FC—>0.5, a total of 1,907 DEGs were discovered, of which 33 overlapped (Fig. 3A). Venn diagrams comparing DEGs from bulk RNA-seq data and scRNA-seq data revealed 33 Co-DEGs (Fig. 3B). Heatmap visualization of the Co-DEGs demonstrated significant differences in their distribution between the AMI and the Sham groups (Fig. 3C). GO and KEGG analyses were respectively performed on the Co-DEGs. GO analysis showed their main involvement in biological processes (BPs) such as regulation of locomotion and regulation of immune system process (Fig. 4A), cellular components (CCs) including fibrinogen complex, plasma lipoprotein particles (Fig. 4B), and molecular functions (MFs) relating to extracellular matrix binding and extracellular matrix structural composition (Fig. 4C). KEGG enrichment analysis demonstrated their involvement in pathways such as amoebiasis, AGE-RAGE signaling pathway in diabetic complications, etc (Fig. 4D). Wiki pathway analysis showed predominant enrichment of the,Co-DEGs in the inflammatory response pathway, spinal cord injury, and adhesion foci: PI3K-Akt-mTOR signaling pathway (Fig. 4E).

Screening for macrophage-associated key genes

Inflammatory pathway-related genes were significantly enriched in AMI compared to the Sham group (Fig. 5A). Taking the inflammation-related pathway and macrophage M0 and M2 score as conditions, Saa3, Ms4a4c, Acp5, and Fcgr4 were screened out as key genes (Fig. 5B). The relationship between these 4 genes and specific inflammatory response pathway-related genes was displayed (Fig. 5C).

Immune cell correlation analysis of macrophage-associated key genes

Immune infiltration analysis was performed on the four identified key genes (Saa3, Fcgr4, Acp5, and Ms4a4c). All of them were all positively correlated with M0 macrophages and negatively correlated with M2 macrophages, NK cell stimulation, and others (Figs. 6A–6D).

Validation of the expression of the four screened genes in different datasets

The analysis of the GSE163465 dataset revealed high expression of the key genes (Saa3, Ms4a4c, Fcgr4, and Acp5) in macrophages and monocytes in AMI (Figs. 7A–7D). Additionally, these same genes exhibited significantly high expression in AMI compared to the Sham group in the bulk RNA-seq dataset (Figs. 7E–7H). The validation results from the scRNA-seq dataset exhibited a consistent pattern with the bulk RNA-seq dataset.

Data set validation and GSEA

Analysis results from the GSE114695 dataset demonstrated a notable increase in the expression of all the four key genes in the AMI group compared to the Sham group (Figs. 8A–8D). GSEA results revealed that Saa3 was enriched in the cytokine-cytokine receptor interaction pathway (Fig. 8E), while Ms4a4c, Acp5, and Fcgr4 were significantly enriched in the Salmonella infection pathway (Figs. 8F–8H).

Cardiac function assessment of mice within the constructed AMI model

After 1 week of LAD coronary artery ligation in mice, the cardiac function in the left ventricle was assessed using M-mode echocardiography. Compared with the Sham group, mice in the AMI group developed cardiac systolic and diastolic dysfunction, as indicated by a significant decrease in ejection fraction (EF), fractional shortening (FS), along with an increase in LVESD (Fig. 9).

Assessment of infarct size in AMI mice

After 1 week of successful AMI modeling, the infarct area in mouse hearts was measured. The AMI group exhibited increased and much larger infarct area compared with the Sham group (Fig. 10A), confirming the successful AMI modeling.

Expression validation of 4 key genes by qRT-PCR

Validation of the expression of Saa3, Ms4a4c, Fcgr4, and Acp5 was conducted using qRT-PCR. The results revealed a significant up-regulation of in cardiac macrophages in the AMI group compared to the Sham group, while no significant difference was observed in the expression of MS4A4C between the Sham and AMI groups (Figs. 11A–11D).