Integrative Transcriptomic-Histological Analysis in Dilated Cardiomyopathy Unveils FGFR1 Inhibition as Anti-Cardiac Fibrotic and Cardioprotective Therapy

Figures

• 

  Figure 1. Transcriptomic Analysis and Network Modeling in DCM and Controls (A) Heatmap of Pearson’s correlation coefficients (upper values) and corresponding P values (lower values) illustrating the module-trait relationships between gene coexpression modules (y-axis) and clinical traits (x-axis). (B) A display of the 30 foremost genes based on intramodular connectivity within the purple module. Genes highlighted in yellow represent the top 5 hub genes, and circle size correlates with the intramodular connectivity. (HSPA4: 9.93, RAB15: 8.93, ANKRD1: 7.91, DNAJB5: 7.79, and FSTL3: 7.20). BNP = brain natriuretic peptide; DCM = dilated cardiomyopathy; LVDd = left ventricular diastolic diameter; LVDs = left ventricular systolic diameter; LVEF = left ventricular ejection fraction.

• 

  Figure 2. Comprehensive Analysis of Fibrosis and Gene Expression in DCM (A) HALO AI workflow. (B) Representative images of an Elastica Masson-stained heart biopsy sample (left) and postclassification by HALO AI (right). Classified regions: "fibrotic area" (green), "cardiac muscle area" (red), "pericardial tissue area" (dark blue), "blood area" (yellow), and "empty area" (light blue). (C) Percentage distribution of fibrotic area in dilated cardiomyopathy (DCM) samples. N = 58. (D) Linear regression shows the relationship between fibrosis rate (%fib) and fibrosis score (median expression of 114 fibrosis-related genes) with 95% CIs (dotted lines). (E) Scatterplot of gene expression vs correlation with fibrotic area. The x-axis (log scale) shows average gene expression; the y-axis shows correlation coefficients. The top 10 genes with average expression >300 are highlighted in red. (F) Linear regression of fibrosis rate (%fib) vs MMP2 expression with 95% CIs (dotted lines). (G) Bar plot comparing late gadolinium enhancement (LGE)+ and LGE− groups. ∗∗P < 0.01. n = 42. Error bars: SD. Analyses: Pearson's correlation (D, E, F); unpaired Student's t-test (G). Scale bar: 100 μm.

• 

  Figure 3. Validation of Fibrosis-Related Biomarkers Using iPSC-Derived Cardiac Fibrosis Model (A) Protocol to generate in vitro cardiac fibrotic organoid models. (B) Representative immunostaining of cardiac troponin T (cTNT), FN1, GFP, and 4′,6-diamidino-2-phenylindole (DAPI) in healthy and fibrotic cardiac organoids. 3-dimensional reconstruction images are shown. GFP staining specifically highlights epicardium-derived cells. Scale bar, 50 μm. (C) Quantification of the mean FN intensity in healthy and fibrotic organoids (n = 12). (D) Quantitative reverse transcription polymerase chain reaction expression analyses of MMP2 in healthy and fibrotic organoids (n = 12). (E) Quantitative reverse transcription polymerase chain reaction expression analyses of MMP2 in various fibrosis-related conditions (n = 3). (F) MMP2 ELISA analysis in the healthy and fibrotic organoids (n = 5). (G) The expression of FN1 in organoids treated with various inhibitors (1 μmol/L). Relative expression compared with healthy control (= 1.0) is shown. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 by 1-way analysis of variance with Tukey’s post hoc test in E and G and by unpaired Student's t-test in C to F. All error bars represent…

• 

  Figure 4. Inhibition of FGFR1 Signaling Pathway Attenuates Cardiac Fibrosis (A) Representative immunostaining of pFGFR1 and FGFR1 in biopsy samples with larger and smaller fibrotic areas. Scale bar: 100 μm. (B) Quantification of the pFGFR1/FGFR1 ratio in immunostained samples of the indicated conditions (n = 3). (C) Representative coimmunofluorescence of human cardiac biopsy sections showing pFGFR1 (red) with (top) cardiac troponin T (cTNT) (green) or (bottom) vimentin (VIM) (green) and DAPI (blue). Scale bar: 100 μm. (D) Protocol to test the effects of FGFR1 inhibition by AZD4547. (E) Quantitative reverse transcription polymerase chain reaction expression analyses of indicated genes in the DMSO- and AZD4547-treated fibrotic organoids (n = 6). (F) Representative immunostaining of cTNT (white), FN1 (red), GFP (green), and DAPI (blue) in the DMSO- and AZD4547-treated fibrotic organoids. GFP staining specifically highlights epicardium-derived cells. Three-dimensional reconstruction images are shown. Scale bar: 50 μm. (G and H) Quantification of the mean FN1 (G) and COL1A1 (H) immunofluorescence intensity in the indicated conditions (FN1, n = 19 for healthy; n = 32 for fibrotic organoids…

• 

  Figure 5. Transcriptomic and Functional Analysis of Gene Response to AZD4547 (A) Heatmap representation of fibrosis-related gene expression in control organoids (healthy; n = 3, shown in yellow), fibrotic organoids (fibrosis; n = 3, shown in blue), and fibrotic organoids treated with AZD4547 (fibrosis+AZD4547; n = 3, shown in red). Expression values have been log2-transformed for visualization. (B) Volcano plot comparing gene expression between fibrosis and fibrosis+AZD4547. Red and blue dots represent significantly up- and down-regulated genes by fibrosis+AZD4547, respectively, and gray dots signify genes with no significant change. The x-axis represents Log2 fold change, and the y-axis represents -log10 (P) – adjusted P values. (C, D) GO enrichment analysis, emphasizing biological processes, based on differentially expressed genes in fibrosis+AZD4547. Circle size signifies the gene ratio, which indicates the proportion of genes in a specified pathway relative to all differentially expressed genes. The color gradient indicates Fisher exact P value, adjusted by the Benjamini-Hochberg method for multiple comparisons: (C) for genes down-regulated by AZD4547 and (D) for genes up-regulated by…

• 

  Figure 6. AZD4547 Treatment Enhances Cardiac Function in Mouse Cardiac Injury Model (A) Study design: Adult C57BL/6 mice were treated with saline or Ang II/PE via an osmotic minipump for 35 days. AZD4547 treatment commenced the day after implantation with daily oral gavage. (B) M-mode echocardiographic images showing LVDd and LVDs. (C) Echocardiographic analysis of fractional shortening (FS) and left ventricular ejection fraction (LVEF), showing improvement in AZD4547-treated mice. n = 6, n = 5, n = 6 biologically independent mice per group. (D) Picrosirius red staining of cardiac sections showing collagen (pink). Reduced fibrosis was observed in PE/Ang II + AZD4547 compared to PE/Ang II + saline. Insets: magnified LV myocardium view. Scale bar: 100 μm. (E) Quantitative fibrosis analysis (% of the ventricular area) from D. n = 3 biologically independent mice per condition. Statistical analysis: 1-way analysis of variance with Tukey’s post hoc test. Error bars denote the SEM. Significance: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Abbreviations as in Figure 1.

• 

  Figure 7. Cellular and Molecular Changes in Left Ventricles After PE/AngII and AZD4547 Treatment (A) UMAP plot showing scRNA-seq data integration from left ventricles of control, NS, and AZD treatment, 7 days post-pump implantation. (B) Reclustering of fibroblasts reveals 8 subclusters (0-7) in a UMAP plot. (C) Bar plot showing the percentage distribution of fibroblast subclusters by treatment origin. (D) Violin plots of fibrosis-related gene expression (Col1a1, Col1a2, Mmp2) across fibroblast subclusters. (E) UMAP plot of reclustered cardiomyocytes reveals 10 subclusters (0-9). (F) Bar plot showing the percentage distribution of cardiomyocyte subclusters by treatment origin. (G) Violin plots show log-normalized expression of Mmp2 (left) and Fgfr1 (right) across 7 major cardiac cell types in control, NS, and AZD treatment groups. Statistical analysis: 1-way analysis of variance with Tukey’s post hoc test. Error bars represent SEM. Significance: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

• 

  Figure 8. AZD4547 Modulates Ligand Receptor-Mediated Intercellular Communication and Target Gene Expression in PE/Ang II–Treated Hearts (A) Network diagrams of total predicted ligand-receptor interactions among major cardiac cell types—fibroblasts (FB), endothelial cells (EC), ventricular cardiomyocytes (CM), and macrophages (Macro)—under PE/Ang II + Normal saline (NS) and PE/Ang II + AZD4547 (AZD) conditions. Edge thickness is proportional to the number of inferred interactions between each pair of cell types. Edge thickness is proportional to the number of inferred interactions between each pair of cell types. (B) Relative information flow for the top 10 signaling ligand-receptor pairs in (left) the FGF pathway acting on cardiomyocytes (CM), and (right) the NPR1 pathway acting on fibroblasts (FB). Red bars denote NS, and teal bars AZD4547. Ligand names in bold indicate those most reduced by AZD4547. (C) FGF signaling networks plotted as circular chord diagrams for NS (left) and AZD (right). Chord width reflects interaction strength from the source (ligand-producing) to the target (receptor-expressing) cell types. (D) NPR1 signaling networks, displayed as in (C), highlight altered…