Upregulated FoxO1 promotes arrhythmogenesis in mice with heart failure and preserved ejection fraction

Abstract

Myocardial fibrosis leads to cardiac dysfunction and arrhythmias in heart failure with preserved ejection fraction (HFpEF), but the underlying mechanisms remain poorly understood. Here, RNA sequencing identifies Forkhead Box1 (FoxO1) signaling as abnormal in male HFpEF hearts. Genetic suppression of FoxO1 alters the intercellular communication between cardiomyocytes and fibroblasts, alleviates abnormal diastolic relaxation, and reduces arrhythmias. Targeted downregulation of FoxO1 in activated fibroblasts reduces cardiac fibrosis, blunts arrhythmogenesis and improves diastolic function in HFpEF. These results not only implicate FoxO1 in arrhythmogenesis and lusitropy but also demonstrate that pro-fibrotic remodeling and cardiomyocyte-fibroblast communication can be corrected, constituting an alternative therapeutic strategy for HFpEF.

Fig. 1. Impaired electrical impulse propagation underlies arrhythmia in HFpEF.

a Experimental design. b Representative left ventricular M-mode echocardiographic images, pulse-wave Doppler showing E (early filling)– and A (atrial filling)–wave, and tissue Doppler showing E′. Quantification of ejection fraction (c, EF) and diastolic function (d, E/E’ ratio), n = 7 (Ctr) and n = 8 (HFpEF). Representative programmed electrical stimulation tracings and pacing protocol (e) and Incidence of pacing-induced ventricular tachycardia (VT, f). g Representative isochronal voltage maps. Action potential duration at 80% of repolarization (APD80, h) and conduction velocity (i) at different basic cycle lengths (BCL), n = 8 (Ctr) and n = 7 (HFpEF). j Representative isochronal voltage maps (*indicates the location of pacing electrode stimulating at 80-ms BCL). Conduction velocity (CV) in transverse (k) and longitudinal (l) directions, n = 5 (Ctr and HFpEF). m Representative picrosirius red images (scale bar = 100 µm) and fibrosis quantification (n), n = 7 (Ctr and HFpEF). o Quantification of hydroxyproline content, n = 6 (Ctr) and n = 7 (HFpEF). Data are expressed as mean ± SEM. Unpaired two-sided Student’s t-test (c–o). Two-sided Fisher’s exact test (f). Two-way repeated measures ANOVA followed by Bonferroni post hoc test (h, i). SR, sinus rhythm; ms, millisecond. Source data are provided as a Source Data file. Part of Fig. 1a was created in BioRender. Mesquita, T. (2024) https://BioRender.com/o74c855. Released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

Fig. 2. Transcriptomics of HFpEF hearts revealed a high convergence of metabolic signals toward FoxO1 activation.

a Heatmap of difference expressed genes in the heart. b Biologically relevant pathways based on Kyoto Encyclopedia of Genes and Genomes enrichment analysis. c Schematic illustration highlighting the multiple altered pathways in HFpEF hearts converge to Forkhead Box O signaling. d Representatives immunoblot images of major cardiac FoxO isoforms in the enriched nuclear fraction. e Protein quantification, n = 3 (Ctr) and n = 4 (HFpEF). f Relative expression of known downstream targets of FoxO1 and different fibroblast states in left ventricular HFpEF samples relative to controls based on RNA sequencing. g Experimental design of adult cardiac fibroblast isolation and mRNA quantification of fibrosis markers by qPCR (h), n = 10 (Ctr) and n = 9 (HFpEF). Data are expressed as mean ± SEM. Two-sided Fisher’s exact test (b). Unpaired two-sided Student’s t-test (e, h). Source data are provided as a Source Data file. Parts of Fig. 2g were created in BioRender. Mesquita, T. (2024) https://BioRender.com/e91m165. Released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

Fig. 3. Downregulation of FoxO1 enhances the propagation of electrical impulses and alleviates the arrhythmogenesis in HFpEF.

a Experimental design. b Cardiac transduction by AAV vectors showing paired brightfield (top) and epifluorescent (bottom) images of mCherry 4 weeks after AAV injection. Quantification of ejection fraction (c, EF) and diastolic function (d, E/E’ ratio), n = 10 (Ctr), n = 9 (HFpEF AAV9 scramble-sh) and n = 10 (HFpEF AAV9 FoxO1-sh). e Exercise capacity. Lung weight (f, LW) and heart weight (g, HW) normalized by the tibia length (TL). e–g n = 9 (Ctr), n = 9 (HFpEF AAV9 scramble-sh) and n = 10 (HFpEF AAV9 FoxO1-sh). Representative programmed electrical stimulation tracings and pacing protocol (h) and Incidence of pacing-induced ventricular tachycardia (VT, i). Action potential duration at 80% of repolarization (APD80, j) and conduction velocity (k) at different basic cycle lengths (BCL), n = 7 (HFpEF AAV9 scramble-sh) and n = 6 (HFpEF AAV9 FoxO1-sh). l Representative isochronal voltage maps (*indicates the location of pacing electrode stimulating at 80-ms BCL). Conduction velocity (CV) in transverse (m) and longitudinal (n) directions, n = 5 (HFpEF AAV9 scramble-sh and HFpEF AAV9 FoxO1-sh). o Representative picrosirius red images (scale bar = 100 µm) and fibrosis quantification (p), n = 7 (Ctr), n = 10 (HFpEF AAV9 scramble-sh) and n = 9 (HFpEF AAV9 FoxO1-sh). q Quantification of hydroxyproline content, n = 4 (Ctr, HFpEF AAV9 scramble-sh and HFpEF AAV9 FoxO1-sh). Data are expressed as mean ± SEM. Two-way repeated measures ANOVA followed by Bonferroni post hoc test (c–k). One-way ANOVA followed by Bonferroni post hoc test (e–q). Two-sided Fisher’s exact test (i). Unpaired two-sided Student’s t-test (m, n). Source data are provided as a Source Data file. Part of Fig. 3a was created in BioRender. Mesquita, T. (2024) https://BioRender.com/o74c855. Released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

Fig. 4. scRNA-seq revealed transcriptome signatures of fibroblasts in HFpEF hearts.

a Heatmap of difference expressed genes in the heart. b, c Gene set enrichment analysis. d Schematic depicting the experimental design of scRNA-seq. e Uniform manifold approximation and projection (UMAP) embedding of all cells and from all conditions. f Violin plots showing the expression of known markers of activated fibroblasts (FB) in all identified clusters. g Expression profiles and the number of cells expressing markers of FB activation within activated and quiescent FBs clusters. h Gene ontology terms in activated FBs clusters. i The relative number of cells within activated and quiescent FBs clusters. j Pseudo-bulk expression profiles and the number of cells expressing markers of FB activation in FBs cells. Kolmogorov-Smirnov test (b, c) and two-sided Fisher’s exact test (h). Source data are provided as a Source Data file. Parts of Fig. 4 d were created in BioRender. Mesquita, T. (2024) https://BioRender.com/x03s068. Released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

Fig. 5. Downregulation of FoxO1 alters intercellular communication between cardiomyocytes and fibroblasts in HFpEF hearts.

a Gene set enrichment analysis plots within all FB clusters showing upregulation of the extracellular matrix organization term in scramble-treated HFpEF compared to control (top) and downregulation of this gene set term by the treatment with AAV9-FoxO1-sh (bottom); NES normalized enrichment score. b Representative immunofluorescence images for periostin and quantification (c), n = 4 (Ctr, HFpEF AAV9 scramble-sh and HFpEF AAV9 FoxO1-sh). d Representative immunofluorescence images for connexin 43 (Cx43), vimentin, and alpha sarcomeric actinin (αSA). e Relationship analysis between vimentin area and Cx43 fragmentation per field of view (scale bar = 25 µm), n = 4 (Ctr 11 random fields, HFpEF AAV9 scramble-sh 19 random fields and HFpEF AAV9 FoxO1-sh 15 random fields). Dashed line: one phase exponential fit. Data are expressed as mean ± SEM. Kolmogorov-Smirnov test (a) and one-way ANOVA followed by Bonferroni post hoc test (c). Source data are provided as a Source Data file.

Fig. 6. Nuclear translocation of FoxO1 is necessary to activate fibroblast reprogramming.

a Experimental design of adenoviral overexpression of GFP and flag-fused mutant FoxO1 (FoxO1^AAA) in fibroblasts from neonatal hearts. b Representative images showing the expression of GFP and Flag (fused with FoxO1, scale bar  = 100 µm) and quantification of subcellular localization (c). d mRNA expression of FBs activation markers, n = 6 (Ad-GFP and Ad-FoxO1^AAA). e mRNA expression at indicated times after actinomycin D treatment, presented as percent of transcript levels relative to time point zero, n = 3 (Ad-GFP and Ad-FoxO1^AAA). Data are expressed as mean ± SEM. Unpaired two-sided Student’s t-test (d). Two-way repeated measures ANOVA followed by Bonferroni post hoc test (e). mRNA levels were fitted using one-phase exponential decay. a.u., arbitrary unit. Source data are provided as a Source Data file. Parts of (a) were created in BioRender. Mesquita, T. (2024) https://BioRender.com/a30n975. Released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

Fig. 7. Targeted downregulation of FoxO1 in activated fibroblasts alleviates fibrotic remodeling and arrhythmias in HFpEF.

a Experimental design. b Representative left ventricular M-mode echocardiographic images, pulse-wave Doppler showing E (early filling)– and A (atrial filling)–wave, and tissue Doppler showing E′. Quantification of ejection fraction (c, EF) and diastolic function (d, E/E’ ratio). e Exercise capacity. Lung weight (f, LW) and heart weight (g, HW) normalized by the tibia length (TL). c–g, n = 8 (HFpEF AAV9-GFP and HFpEF AAV9-Cre). Representative programmed electrical stimulation tracings and pacing protocol (h) and incidence of pacing-induced ventricular tachycardia (VT, i). j Representative picrosirius red images (scale bar = 200 µm) and fibrosis quantification (k), n = 5 (HFpEF AAV9-GFP and HFpEF AAV9-Cre). l Schematic illustration of our working hypothesis. Data are expressed as mean ± SEM. Unpaired two-sided Student’s t-test (c–g, k). Two-sided Fisher’s exact test (i). Source data are provided as a Source Data file. Parts of (a and l) were created in BioRender. Mesquita, T. (2024) https://BioRender.com/p96k811 and https://BioRender.com/f20a900. Released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.