Multi-microRNA Diagnostic Panel for Heart Failure with Preserved Ejection Fraction in Preclinical and Clinical Settings

Figures

• 

  Figure 1. DSS animal model developed diastolic dysfunction and resembles the HFpEF features. (A) Experimental design. (B) Mean blood pressure (MBP) and (C) left ventricle end diastolic pressure (LVEDP) assessed by intracardiac haemodynamic measurement at the end of 28 weeks. (D) Diastolic function indicated by E/e′ and (E) systolic function indicated by ejection fraction (EF) assessed by echocardiography. (F) Pulmonary congestion indicated by lung wet/dry ratio. (G) Plasma analysis of NT‐proBNP by ELISA. (H) Representative images of four chamber view (top panel), Haematoxylin and Eosin staining (HES) (second panel), Picrosirius Red (Sirius Red, third panel) and Wheat Germ Agglutinin (WGA) stained histological left ventricular sections from DSS rats subjected to low (LS) or high (HS) salt diet. (I) Left ventricular weight adjusted to tibia length. (J) Quantification of collagen deposition from Picrosirius Red stained sections in DSS rats subjected to LS or HS diet. (K) Quantification of cardiomyocyte cross‐sectional area from WGA‐stained sections in DSS rats subjected to LS (number of hearts = 3, number of cardiomyocytes = 60 per each heart) or HS (number of hearts = 3, number of…

• 

  Figure 2. qPCR‐based expression profiling of circulatory miRNAs in hypertension‐induced diastolic dysfunction preclinical model. (A) Flow chart of circulatory miRNAs screening. (B) Volcano plot of dysregulated circulatory miRNAs in hypertension‐induced diastolic dysfunction preclinical model. Horizontal dashed line indicates the adjusted P value < 0.05. (C) Expression heat map of dysregulated circulatory miRNAs (miR‐wise z‐score of 2ΔCq). According to guide in the right, red colour represents higher plasma level of circulatory miRNAs than mean and blue colour represents lower plasma level of circulatory miRNAs than mean. (D–G) Fold change of rno‐miR‐21‐5p (D), rno‐miR‐140‐3p (E), rno‐let‐7b‐5p (F) and rno‐let‐7e‐5p (G). The numerical data in Parts (D–G) are expressed as mean ± SD.

• 

  Figure 3. Correlations of dysregulated circulatory miRNAs with cardiac function, structure and stress marker as well as pathway enrichment analysis. Chord diagrams represent the significant correlations of rno‐miR‐21‐5p (A), rno‐miR‐140‐3p (B), rno‐let‐7b‐5p (C) and rno‐let‐7e‐5p (D) using R 2 derived from spearman correlation analysis. The stronger is the correlation, the wider is the base of each cord toward the cardiac function, structure and stress marker. (E,F) KEGG (E) and Reactome (F) pathway enrichment analysis of predicted target genes of dysregulated circulatory miRNAs in hypertension‐induced diastolic dysfunction preclinical model.

• 

  Figure 4. Diagnostic performance of panels of identified miRNAs in preclinical and clinical settings. (A) Schematic representations of implemented stepwise algorithm. In the first step, four circulatory miRNAs (miRNA‐21‐5p, miRNA‐140‐3p, let‐7b‐5p and let‐7e‐5p) have been identified by differential expression analysis in hypertension‐induced diastolic dysfunction preclinical model. In the second step, the diagnostic performances of 15 different combinations from four identified miRNAs have been analysed using d‐score derived from linear discriminant analysis (LDA). From this step, receiver operating characteristic (ROC) curve analysis showed an AUC of above 0.9 in 8 miRNA combinations out of 15. Forest plot shows the diagnostic performance of all 15 combinations of miRNAs in discrimination of animals with diastolic dysfunction (N = 11) from control (N = 15). Dashed horizontal line shows the cut‐off of AUC 0.9. In the third step, eight identified miRNA combinations in step2 were tested in clinical setting after LDA using a publicly available dataset (GSE53437) via ROC curve and logistic regression analysis. (B) Forest plot shows the ROC curve analysis of eight miRNA combinations in…

• 

  Figure 5. Gender‐specific expression pattern of identified miRNAs in clinical setting. (A) Gender‐specific expression patterns of four identified miRNAs have been tested in a clinical setting using a publicly available dataset (GSE104150). (B) Receiver operating characteristic (ROC) curve analysis of the combination of four identified miRNAs regardless of gender showed an area under the curve (AUC) of 0.8 in discrimination of heart failure patients from control. The numerical data in (A) are expressed in mean ± SD.