MIF is a Good Diagnostic Indicator for IAs and is Associated with Macrophage M1 Polarization
The mRNA sequencing data quality from dataset GSE54083 for intracranial aneurysms (IAs) and control samples was found to be satisfactory, with no significant outliers, and clear differences between the two groups (Figure 2A-C). Therefore, all samples were included in the study. Subsequently, we observed that the expression level of MIF in IAs was significantly higher than that in normal arterial samples (Figure 2D). The diagnostic ROC curve for MIF (AUC=0.980, CI=0.925-1.000) and the diagnostic PR curve, which showed a precision rate greater than 0.8, indicated its high diagnostic value for IAs (Figure 2E-F). Subsequently, we conducted an immune infiltration analysis using the Cibersort algorithm. First, the heatmap and bar chart results supported higher levels of M1-like macrophage infiltration in IAs compared to control samples (Figure 2G-H). Secondly, we found that the expression level of MIF was positively correlated with the degree of M1-like macrophage infiltration (r=0.783, p=0.008, Figure 2I). Similar results were obtained in the validation dataset GSE75436 (Supplementary Material 1). Therefore, MIF may play an important role in the pathological process of IAs and may be involved in the phenotypic polarization process of macrophages.
Levels of MIF and CD74 Methylation Are Downregulated
Upon quality assessment of the gene methylation data from both intracranial aneurysms (IAs) and normal arterial samples in the dataset GSE75434, no aberrant samples were detected. Principal Component Analysis (PCA) revealed a significant distinction between the two groupings without conspicuous correlation, thereby validating the biological grouping's relevance (Figure 3A-C). Figure 3D provides an overview of the dataset's probe detection, encompassing chromosomal loci and gene locus information. Subsequently, we conducted a differential methylation site analysis using the CHAMP pipeline. Between the two groups, there were marked differences in the methylation levels of MIF and CD74 (Figure 3E-F). As a result, the downregulation of MIF and CD74 gene methylation levels in intracranial aneurysms may contribute to the elevated mRNA expression of MIF and CD74 in IAs.
Single-Cell Data Quality Control and Cell Type Identification
After filtering out unsuitable cells and genes, a total of 9,419 cells and 17,821 genes were included. The "Harmony" package was used to integrate samples and eliminate batch effects, which demonstrated good integration (Figure 4A), with details on quality control provided in Supplementary Material 2. Setting the resolution to 0.4, an unsupervised clustering algorithm classified the 9,419 cells into 19 distinct subgroups (clusters 0-18, Figure 4B). Based on cell markers [28], manual cell annotation categorized the cells into ten defined lineages: vascular smooth muscle cells (VSMCs), endothelial cells, fibroblasts, pericytes, and immune cells (Figure 4C). The marker genes for each cell cluster are identifiable in Figure 4D. Subsequently, we examined the expression patterns of molecules related to the MIF pathway. Mif was highly expressed in VSMCs and macrophages, Pi3kr1 was moderately expressed in both VSMCs and macrophages, while Sirt1 was most abundantly expressed in VSMCs, followed by macrophages. Conversely, Nfkb1 was highly expressed in macrophages (Figure 4E). This suggests that the MIF signaling pathway may potentially mediate the interaction between VSMCs and macrophages within IAs tissue and may be particularly active.
Identification and functional analysis of VSMCs cells
We extracted the single-cell data of the VSMCs for a secondary sample integration, which showed good integration results (Figure 5A); then we performed dimensionality reduction again and clustered the 2,500 cells into 8 cell subgroups (clusters 0-7, Figure 5B). Cell identification was conducted based on the markers of VSMCs, and we discovered that the majority were secretory VSMCs (ACTA2 + MYH11 + COL1A1 + COL1A2+) followed by contractile VSMCs, while other phenotypes of VSMCs were sparse (Figures 5C-D). Subsequently, we observed a significant increase in the proportion of secretory VSMCs in IAs, while contractile VSMCs were markedly reduced (Figure 5E). Notably, there were considerable differences in the expression levels of Sirt1 and Mif among secretory VSMCs between different samples (Figure 5F). At the level of differential gene expression, secretory VSMCs expressed Tns1 and Itga9 highly and Anxa1 and S100a10 lowly (Figure 5G). Further, GO and KEGG enrichment analyses of differentially expressed genes mainly included terms like macrophage activation, positive regulation of macroautophagy, positive regulation of vascular associated smooth muscle cell migration, and the PI3K-Akt signaling pathway (Figure 5H). These findings suggest that secretory VSMCs may play an important role in the pathogenesis of IAs, especially in interaction with macrophages.
Macrophage Identification and Functional Analysis
We extracted the single-cell data of macrophages for secondary sample integration, which showed good integration results (Figure 6A); then we performed dimensionality reduction again and clustered the 2,414 cells into 10 cell subgroups (clusters 0-9, Figure 6B). Cell identification was conducted based on macrophage markers, distinguishing between M1-like and M2-like macrophages (Figure 6C). Figure 6D shows the expression level of macrophage markers in each cell subgroup. Notably, we found that the proportion of M1-like macrophages was significantly increased in the IAs tissues, and the expression levels of Pik3r1 and Nfkb1 in M1-like macrophages were also noticeably higher than in the control group (Figures 6E-F). Subsequent differential gene expression analysis of M1-like macrophages between different samples revealed high expression of Oasl1 and Cxcl10, while Snrpf and Pim3 were expressed at low levels (Figure 6G). A global GSEA analysis indicated a strong pro-inflammatory capability of M1-like macrophages (Figure 6H), which reconfirmed the important role M1-like macrophages play in the onset and progression of IAs.
Through Spearman correlation analysis, we observed a positive correlation between secretory VSMCs and M1-like macrophages (r=0.68, p<0.05, Figure 6I). Inference of intracellular signaling pathway activity revealed significant activity of the TGFb, P53, and hypoxia pathways in secretory VSMCs, while the NFkB and TNFa signaling pathways were upregulated in M1-like macrophages (Figure 6J). This suggests that the interaction between secretory VSMCs and M1-like macrophages may be involved in the occurrence and progression of IAs.
Cell communication analysis shows a significant enhancement of the MIF axis in IAs
We conducted a comprehensive analysis of cell communication using single-cell data. First, we observed frequent cellular communication between various cell types, especially between VSMCs, fibroblasts, and macrophages (Figure 7A). Secondly, we investigated the signaling strength of different receptor-ligand interactions for MIF, where the MIF and CD74/CD44 receptor-ligand relationship dominated (Figure 7B). Chord diagrams revealed the MIF axis as an important receptor-ligand interaction between VSMCs and M1-like macrophages (Figure 7C).
Next, we extracted the single-cell data for secretory VSMCs and M1-like macrophages for group-specific cell communication analysis. First, in IAs, the interaction between secretory VSMCs and M1-like macrophages was significantly increased and enhanced compared with control group samples (Figure 7D). Secondly, through the identification of conserved and specific signaling pathways, the study further found that the signaling strength of the MIF axis between secretory VSMCs and M1-like macrophages was significantly intensified in IAs (Figure 7E). Finally, we explored the expression of the main components of the MIF axis in different cell types and found that the signaling strength of the MIF-CD74/CD44 receptor-ligand was moderate (Figures 7F-G).