Systems biology analysis reveals miRNA–mRNA networks associated with cardiomyocyte responses to SARS-CoV-2 infection

Fig. 1. Histopathological Features of Myocardial Injury in COVID-19. Representative H&E-stained sections of cardiac tissue from COVID-19 patients demonstrating hallmark lesions. The squares are emphasized and shown at a higher magnification in the right panel: (1) Necrosis due to attack by cytotoxic lymphocytes; (2) Immune infiltrate/ lymphocyte attacking; (3) Fibers of different sizes (thick, wide, irregular); (4) Degenerative perinuclear edema; (5) Lymphocyte attacking/ hyperchromatic nucleus; (*) Myocardial interstitial fibrosis with dense ECM accumulation.

Fig. 2. Transcriptomic Landscape of SARS-CoV-2-Infected Cardiomyocytes. A Summary of RNA-seq datasets, showing cell types, viral load (MOI), exposure duration, and total number of DEGs. B Heatmap of enriched Gene Ontology (GO) terms across conditions. Only GO terms with log₂(adjusted p-value) > 2 in at least four models are displayed. Hierarchical clustering was performed using Euclidean distance. C Bar plot quantifying DEGs involved in immune response, muscle changes, ECM remodeling, and coagulation per dataset. D Circos plot linking overlapping genes across functional GO categories; ribbon thickness denotes gene overlap. E Density plots of log₂(fold change) values for DEGs per GO category, visualized using cubic spline regression.

Fig. 3. Cardiomyocyte Transcriptional Programs Associated with Extracellular Matrix Remodeling and Myocardial Disorganization. A Dot plot of GO enrichment for immune, ECM, muscle, and coagulation-related terms; dot size = gene count, color = adjusted p-value. B Gene interaction network showing selected genes expressed in cardiomyocytes and their connectivity across functional categories (e.g., CAV1, CXCR4). C GO enrichment for each cluster: Cluster 1 = collagen organization; Cluster 2 = matrix degradation and integrin pathways. D H&E-stained section and binarized ECM/nuclei images used for fractal analysis. ECM (D = 1.1992) and nuclear (D = 1.5035) patterns indicate structural disarray and cellular disorganization, respectively. The ECM-related signatures identified in cardiomyocytes are interpreted as transcriptional activation of pro-fibrotic signaling programs, which may indirectly contribute to myocardial fibrosis through intercellular communication.

Fig. 4. MicroRNA Regulatory Networks in SARS-CoV-2-Infected Cardiomyocytes. A Bubble heatmap of top 40 predicted miRNAs targeting DEGs; size = significance (–log10 p-value), color = number of targets. Hierarchical clustering was performed using Euclidean distance considering rows and columns. B–D miRNA–target interaction networks for: B immune regulation; C ECM remodeling; D muscle contraction. Notable hubs: miR-21-5p, miR-145-5p (immune); miR-29 family (ECM); miR-1-3p, miR-199a-5p (muscle). Networks illustrate post-transcriptional regulation of cardiac pathophysiology. E Molecular network illustrating miRNAs (grey triangles) predicted to regulate genes (colored circles) implicated in cardiomyocyte alterations. Node size reflects the degree of connectivity (number of edges). Edges represent documented or predicted miRNA–mRNA interactions. Gene nodes are color-coded according to functional categories: immune response (purple), muscle contraction (pink), ECM remodeling (blue), and coagulation (green).

Fig. 5. Circulating miRNA Profiles Reveal COVID-19 Severity Signatures. A Workflow of circulating miRNA profiling by TLDA in plasma samples from healthy donors (n = 4), non-ICU COVID-19 patients (n = 4), and ICU COVID-19 patients (n = 4). B Principal component analysis (PCA) showing separation among the three groups, with ICU samples forming a distinct cluster. C Volcano plots showing differentially expressed circulating miRNAs between the study groups. D Venn diagram illustrating the overlap between the Top 40 candidate miRNAs predicted from cardiomyocyte transcriptomic meta-analysis and the 694-miRNA TLDA panel. Of the 40 predicted miRNAs, 34 were represented in the TLDA array, and 11 (32.4%) were significantly dysregulated in plasma samples from COVID-19 patients. The remaining predicted miRNAs were either not significantly altered in plasma or were not represented in the TLDA panel. E Functional classification of target genes regulated by the 11 shared miRNAs, grouped according to immune response, extracellular matrix remodeling, coagulation, and cardiac muscle-related processes.

Fig. 6. Model of COVID-19-Induced Myocardial Fibrosis. Illustrative model of five sequential stages: A Homeostasis: Healthy myocardium. B Inflammation: Immune infiltration and cytokine storm; regulated by miR-145-5p, miR-126-3p, miR-21-5p, miR-454. C ECM Remodeling: Myofibroblast differentiation and ECM deposition; regulated by miR-199a-5p and miR-29 family. D Interstitial Fibrosis: ECM excess and apoptosis. E Fibrotic Scar: Persistent myofibroblast activity leads to scar formation and loss of elasticity; ECM-regulatory miRNAs remain elevated. Original artwork created for this study by the authors.