Identification of Gene Expression Signatures for Phenotype-Specific Drug Targeting of Cardiac Fibrosis

Figures

• 

  Figure 1. Global gene expression changes in MF models. Unsupervised two-dimensional hierarchical clustering of samples based on gene expression level (A). Different colors in principal component analysis indicate different treatment models (B). Bars representing the number of up- and down-regulated genes in all treatment groups compared to control with adjusted cut-offs: FDR = 0.05, LFC = ±2 (C).

• 

  Figure 2. Heatmap of k-means clustering of genes with subsequent gene ontology (GO) enrichment analysis. Unsupervised gene clustering ordered genes into groups based on the expression pattern across all samples. 2000 top-ranked expressed genes were used for a GO enrichment analysis for each gene cluster. The figure shows the most associated GO cellular components, biological processes and molecular functions for each cluster of genes. Green and red colors denote highly and weakly expressed genes, respectively.

• 

  Figure 3. The heatmaps of differentially regulated genes in regulated pathways across all models of myocardial remodeling. The heat maps show the expression of genes compared to untreated control in the transforming growth factor beta (TGF-ß) signaling pathway (A), apelin signaling pathway (B), and hypoxia-inducible factor 1-alpha (HIF-1 alpha) signaling (C) in all treatment groups. Red = upregulation; green = downregulation; gray = non-significant changes. The treatment groups DOX (N = 5), MYO (N = 5), RemoLV (N = 3) and Hyper (N = 3) are relative to control animals (N = 3). (Significance for p < 0.05, moderated t-statistics adjusted for multiple testing).

• 

  Figure 4. The heat maps of DEGs in significantly regulated pathways in specific models of MF. The heat maps show the expression of genes compared to untreated control in the tumor necrosis factor-alpha (TNF-alpha) signaling pathway (A); adrenergic signaling pathway (B); and Forkhead box protein-O (Fox-O) signaling (C) in all treatment groups. Red = upregulation; green = downregulation; gray = non-significant changes. The treatment groups DOX (N = 5), MYO (N = 5), RemoLV (N = 3) and Hyper (N = 3) are relative to control animals (N = 3). (Significance for p < 0.05, moderated t-statistics adjusted for multiple testing).

• 

  Figure 5. Drug prediction based on the gene expression signature datasets. Bubble plot lists of selected significant chemical perturbagens (predicted drugs) based on the gene expression signature datasets of DOX (N = 5), MYO (N = 5), RemoLV (N = 3) and Hyper (N = 3) groups. Selected perturbagens reflect (Pearson’s correlation coefficient) the gene expression signatures of MF models. The size of the dot represents p-value. ARB, angiotensin II receptor blocker; ACE, angiotensin converting enzyme.

• 

  Figure 6. Study design and applied molecular biology methods. The study population comprises 5 groups of pigs. DOX animals (N = 5) received doxorubicin and MYO animal (N = 5) Myocet® infusions in doses equivalent to human treatment regimens. Catheter-based reperfused anterior wall myocardial infarction was induced by balloon inflation in mid-LAD, and induced adverse remodeling of the enlarged left ventricle (LV), resulting in volume-overload myocardial fibrosis of the posterior non-ischemic wall (RemoLV, N = 3). Artificial aortic isthmus stenosis was designed to study pressure overload-induced myocardial remodeling (Hyper, N = 3). Sham interventions served as control (Control, N = 3). After a pre-defined follow-up (FUP), the hearts were explanted and subjected to RNA expression analysis following RNA-seq bioinformatics pipeline and drug prediction model. Graphical elements in both panels were created with BioRender.com.