Relationships between systemic sclerosis and atherosclerosis: screening for mitochondria-related biomarkers

Abstract

Background: Patients with systemic sclerosis (SSc) are known to have higher incidence of atherosclerosis (AS). Mitochondrial injuries in SSc can cause endothelial dysfunction, leading to AS; thus, mitochondria appear to be hubs linking SSc to AS. This study aimed to identify the mitochondria-related biomarkers of SSc and AS.

Methods: We identified common differentially expressed genes (DEGs) in the SSc (GSE58095) and AS (GSE100927) datasets of the Gene Expression Omnibus (GEO) database. Considering the intersection between genes with identical expression trends and mitochondrial genes, we used the least absolute shrinkage and selection operator (LASSO) as well as random forest (RF) algorithms to identify four mitochondria-related hub genes. Diagnostic nomograms were then constructed to predict the likelihood of SSc and AS. Next, we used the CIBERSORT algorithm to evaluate immune infiltration in both disorders, predicted the transcription factors for the hub genes, and validated these genes for the two datasets.

Results: A total of 112 genes and 13 mitochondria-related genes were identified; these genes were then significantly enriched for macrophage differentiation, collagen-containing extracellular matrix, collagen binding, antigen processing and presentation, leukocyte transendothelial migration, and apoptosis. Four mitochondria-related hub DEGs (IFI6, FSCN1, GAL, and SGCA) were also identified. The nomograms showed good diagnostic values for GSE58095 (area under the curve (AUC) = 0.903) and GSE100927 (AUC = 0.904). Further, memory B cells, γδT cells, M0 macrophages, and activated mast cells were significantly higher in AS, while the resting memory CD4⁺ T cells were lower and M1 macrophages were higher in SSc; all of these were closely linked to multiple immune cells. Gene set enrichment analysis (GSEA) showed that IFI6 and FSCN1 were involved in immune-related pathways in both AS and SSc; GAL and SGCA are related to mitochondrial metabolism pathways in both SSc and AS. Twenty transcription factors (TFs) were predicted, where two TFs, namely BRCA1 and PPARγ, were highly expressed in both SSc and AS.

Conclusion: Four mitochondria-related biomarkers were identified in both SSc and AS, which have high diagnostic value and are associated with immune cell infiltration in both disorders. Hence, this study provides new insights into the pathological mechanisms underlying SSc and AS. The specific roles and action mechanisms of these genes require further clinical validation in SSc patients with AS.

Keywords: atherosclerosis; immune infiltration; machine learning; mitochondria-related genes; systemic sclerosis.

FIGURE 1

Flowchart of the study design.

FIGURE 2

Identification of differentially expressed genes (DEGs). DEG heatmaps and volcano plots for the (A, B) SSc and (C, D) AS datasets.

FIGURE 3

Identification of mitochondria-related DEGs and functional enrichment between SSc and AS. (A) Venn diagram of the DEGs; (B) Venn diagram of the mitochondria-related DEGs; (C, D) GO and KEGG enrichment analyses of DEGs with the same expression trends; (E, F) GO and KEGG enrichment analyses of the mitochondria-related DEGs.

FIGURE 4

Screening of mitochondria-related hub genes and evaluation of their diagnostic values: (A) LASSO analysis for screening mitochondria-related hub genes in GSE58095; (B) identification of mitochondria-related hub genes according to the importance of variables by random forest (RF) analysis of GSE58095; (C) LASSO analysis for screening mitochondria-related hub genes in GSE100927; (D) identification of mitochondria-related hub genes according to the importance of variables by RF analysis of GSE100927; (E) Venn diagram of the four common mitochondria-related hub genes between the two datasets; (F) receiver operating characteristic (ROC) curves of the four hub genes to assess their diagnostic values in the GSE58095 and GSE100927 datasets.

FIGURE 5

Expression levels of the four hub genes in the GSE58095 and GSE100927 datasets (A) Expression level of IFI6 in GSE58095; (B) Expression level of FSCN1 in GSE58095; (C) Expression level of GAL in GSE58095; (D) Expression level of SGCA in GSE58095; (E) Expression level of IFI6 in GSE100917; (F) Expression level of FSCN1 in GSE100927; (G) Expression level of GAL in GSE100927; (H) Expression level of SGCA in GSE100927. *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 6

Development of the diagnostic nomogram model: (A) nomogram predicting the probability of SSc; (B) calibration curves of the SSc risk models; (C) ROC curve of the SSc risk model; (D) nomogram predicting the probability of AS; (E) calibration curves of the AS risk models; (F) ROC curve of the AS risk model.

FIGURE 7

Immune cell infiltration analyses in SSc and AS: (A) boxplot showing the comparison of 22 kinds of immune cells between SSc and the control group; (B) boxplot showing the comparison of 22 kinds of immune cells between AS and the control group; (C) heatmap representing the associations of the differentially infiltrated immune cells with the four hub genes in SSc for the threshold of p < 0.05; *p < 0.05; **p < 0.01; ***p < 0.001; (D) heatmap representing the associations of the differentially infiltrated immune cells with the four hub genes in AS for the threshold of p < 0.05; *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 8

Gene set enrichment analysis (GSEA) for the four mitochondria-related hub genes in (A–D) SSc and (E–H) AS.