Diagnostic potential of energy metabolism-related genes in heart failure with preserved ejection fraction

Abstract

Background: Heart failure with preserved ejection fraction (HFpEF) is associated with changes in cardiac metabolism that affect energy supply in the heart. However, there is limited research on energy metabolism-related genes (EMRGs) in HFpEF.

Methods: The HFpEF mouse dataset (GSE180065, containing heart tissues from 10 HFpEF and five control samples) was sourced from the Gene Expression Omnibus database. Gene expression profiles in HFpEF and control groups were compared to identify differentially expressed EMRGs (DE-EMRGs), and the diagnostic biomarkers with diagnostic value were screened using machine learning algorithms. Meanwhile, we constructed a biomarker-based nomogram model for its predictive power, and functionality of diagnostic biomarkers were conducted using single-gene gene set enrichment analysis, drug prediction, and regulatory network analysis. Additionally, consensus clustering analysis based on the expression of diagnostic biomarkers was utilized to identify differential HFpEF-related genes (HFpEF-RGs). Immune microenvironment analysis in HFpEF and subtypes were performed for analyzing correlations between immune cells and diagnostic biomarkers as well as HFpEF-RGs. Finally, qRT-PCR analysis on the HFpEF mouse model was used to validate the expression levels of diagnostic biomarkers.

Results: We selected 5 biomarkers (Chrna2, Gnb3, Gng7, Ddit4l, and Prss55) that showed excellent diagnostic performance. The nomogram model we constructed demonstrated high predictive power. Single-gene gene set enrichment analysis revealed enrichment in aerobic respiration and energy derivation. Further, various miRNAs and TFs were predicted by Gng7, such as Gng7-mmu-miR-6921-5p, ETS1-Gng7. A lot of potential therapeutic targets were predicted as well. Consensus clustering identified two distinct subtypes of HFpEF. Functional enrichment analysis highlighted the involvement of DEGs-cluster in protein amino acid modification and so on. Additionally, we identified five HFpEF-RGs (Kcnt1, Acot1, Kcnc4, Scn3a, and Gpam). Immune analysis revealed correlations between Macrophage M2, T cell CD4+ Th1 and diagnostic biomarkers, as well as an association between Macrophage and HFpEF-RGs. We further validated the expression trends of the selected biomarkers through experimental validation.

Conclusion: Our study identified 5 diagnostic biomarkers and provided insights into the prediction and treatment of HFpEF through drug predictions and network analysis. These findings contribute to a better understanding of HFpEF and may guide future research and therapy development.

Keywords: diagnostic biomarkers; energy metabolism-related genes; heart failure with preserved ejection fraction; immunoscape; machine learning; targeted drug prediction.

Figure 1

Identification of DE-EMRGs. (A) Volcano plot of DEGs. Red and blue dots indicate upregulated and down-regulated genes, respectively, and gray is the non-significant gene. (B) The intersection of DEGs with EMRGs. DE-EMRGs, differentially expressed energy metabolism-related genes; DEGs, differentially expressed genes; EMRGs, energy metabolism-related genes.

Figure 2

Identification of key DE-EMRGs using machine learning algorithms. (A) LASSO logistic regression algorithm used to screen key genes. (B) The importance of the variables was ranked by RMSE loss after permutations, with higher values indicating that the variable contributes greater to the accuracy of the model. (C) SVM-RFE algorithm to screen diagnostic biomarkers. (D) Venn diagram demonstrates the intersection of diagnostic biomarkers obtained by the three algorithms. DE-EMRGs, differentially expressed energy metabolism-related genes; LASSO, least absolute shrinkage and selection operator; RMSE, root mean square error; SVM-RFE, support vector machine-recursive feature elimination.

Figure 3

Diagnostic value of key DE-EMRGs in HFpEF. (A–E) The ROC curves of the key DE-EMRGs. (F) ROC curves of the diagnostic model in the GSE180065 dataset. (G) Nomogram for HFpEF samples. (H) Calibration curve to assess the predictive power of the nomogram. DE-EMRGs, differentially expressed energy metabolism-related genes; HFpEF, heart failure with preserved ejection fraction; ROC, receiver operating characteristic.

Figure 4

Single gene GSEA-GO for the 5 key DE-EMRGs. Enrichment in GO collection by Chrna2 (A), Ddit4l (B), Gng7 (C), Prss55 (D), and Gnb3 (E). Each line represents one gene set with unique color. Gene sets were considered significant only when |NES| > 1, P < 0.05, and q < 0.05. Only several leading gene sets (Top 10) were displayed in the plot. GSEA, gene set enrichment analysis; GO, Gene Ontology; DE-EMRGs, differentially expressed energy metabolism-related genes; NES, normalized enrichment score.

Figure 5

Targeted drugs and regulatory networks for key DE-EMRGs. (A) DrugBank-based drug-key DE-EMRGs interaction network. (B) Integrated miRNA-key DE-EMRGs and key DE-EMRGs-TFs interaction networks for the 5 biomarkers. Blue squares represent nine hub genes. Yellow circles represent TFs that have connectivity with biomarkers. Pink diamonds represent miRNAs associated with biomarkers. DE-EMRGs, differentially expressed energy metabolism-related genes; miRNA, microRNA; TFs, transcription factor.

Figure 6

Identification of DEGs among biomarker-based subtypes. (A) Heatmap depicts consensus clustering solution (k = 2) for 5 biomarkers in 10 HFpEF samples; (B) Delta area curve of consensus clustering indicates the relative change in area under the CDF curve for k = 2 to 6. (C) Volcano plot of DEGs between Cluster1 and Cluster2. Red and blue dots indicate upregulated and down-regulated genes, respectively, and gray is the non-significant gene. DEGs, differentially expressed genes; HFpEF, heart failure with preserved ejection fraction; CDF, cumulative distribution function.

Figure 7

The association of biomarkers with immune microenvironment. (A) Immune cell infiltration between two groups by XCELL algorithms and only statistically significant ones are shown. (B) Scatter plots show the correlation of biomarkers with the infiltration of Macrophage M2, T cell CD4+ Th1, and B cell. *P < 0.05; **P < 0.01.

Figure 8

Assessment of cardiac function in the HFpEF mouse model. Echocardiography was performed at 15 weeks after a combination of a HFD (60% kilocalories from fat) and L-NAME (0.5 g l−1 in drinking water) or a standard (chow) diet in mice. (A) Representative recordings of echocardiographic images of the LV. (B) IVSs, IVSd, LVPWs, LVPWd, E/A, and LVEF were measured by echocardiography. *P < 0.05, **P < 0.01, ***P < 0.001 vs. corresponding control group. ns represents no significance. Data are means ± SEM. HFpEF, heart failure with preserved ejection fraction; HFD, high-fat diet; L-NAME, Nω-nitro-L-arginine methyl ester; LV, left ventricle; IVSs, inter-ventricular septum thickness end systolic; IVSd, inter-ventricular septum thickness end diastolic; LVPWs, left ventricular systolic posterior wall thickness; LVPWd, left ventricular posterior wall thickness end-diastolic; E/A, the early (E) wave peak velocity, representing the passive filling, to the late (A) wave peak velocity ratio, representing the active filling due to the atrial contraction; LVEF, left ventricular ejection fraction.

Figure 9

RNA expression of the 5 biomarkers was measured in HFpEF and control samples. RNA expression of Gng7 (A), Prss55 (B), Chrna2 (C), Gnb3 (D), and Ddit4l (E) were measured in blood samples using qRT-PCR. P-values were calculated using a two-sided unpaired Student’s t-test. *P < 0.05; ****P < 0.0001; ns represents no significance. HFpEF, heart failure with preserved ejection fraction; qRT-PCR, quantitative reverse transcription polymerase chain reaction.