PFKM-Driven Lactate Overproduction Promotes Atrial Fibrillation via Triggering Cardiac Fibroblasts Histone Lactylation

Abstract

Increasing evidence has clarified that atrial fibrillation (AF) is associated with enhanced glycolysis, leading to lactate accumulation. However, whether glycolysis promotes AF remains unknown, as does whether histone lactylation plays a role in its pathogenesis. In the study, spontaneous AF mice are established to monitor AF susceptibility and atrial substrates at different ages (3, 5, 7 months), indicating that enhanced glycolysis acts as a promoter during AF development by inducing atrial fibrosis. The promoting effect of glycolysis on AF and the pivotal enzyme in driving glycolysis are confirmed by treatment with glycolysis inhibitor 2-deoxyglucose (2-DG) and adeno-associated virus-mediated atrial PFKM expression. Furthermore, lactate stimulates primary mouse cardiac fibroblast (CF) activation. Mechanistically, the observations indicated that atrial lactate accumulation promotes global lactylation and H3K18 lactylation in atrial fibroblasts. P300-mediated H3K18 lactylation up-regulates TGF-β1 transcription, leading to activation of CF, and thereby contributing to atrial fibrosis. The results reveal a novel role of the metabolic-epigenetic axis in AF pathogenesis, which raises the possibility of potential therapeutic strategies targeting AF.

Keywords: H3K18la; PFKM; atrial fibrillation; glycolysis; histone lactylation.

Figure 1

Glycolysis is associated with enhanced AF in atrial pacing rabbits, CREM mice and human. A) The results of KEGG pathway analysis from GEO database (GSE128188) with atrium in people with AF and sinus rhythm (n = 5). B) Schematic diagram of atrial tissue obtained from AF patients, rapid atrial pacing rabbits and CREM mice for the detection of glycolytic pathway and lactate content. C) Representative bands and quantification of the protein levels of HK, GLUT1, PFKM, PKM and LDHA in atrial tissue with AF and sinus rhythm (n = 6). D) The atrial concentration of lactate (n = 11). E) Representative bands and quantification of the protein levels of GLUT1, HK, PFKM, PKM and LDHA in atrial tissue from rabbits in control group and rapid atrial pacing group (n = 6). F) The atrial concentration of lactate (n = 6). G) Schematic diagram showing dynamic monitoring of the pathological mechanism of AF in CREM mice at different months of age. H) Representative examples of AF induction attempt in CREM mice and AF inducibility (n = 6). I) Representative bands and quantification of the protein levels of GLUT1, HK, PFKM, PKM, and LDHA in atria of 7‐month CREM mice (n = 6). J) The atrial concentration of lactate (n = 6). K) Representative bands and quantification of the protein levels of HK, GLUT1, PFKM, PKM, and LDHA in atria of CREM mice in 5 months (n = 6). L) The atrial concentration of lactate (n = 6). SR = sinus rhythm; AF = atrial fibrillation; Con = control; Pacing = atrial rapid pacing. The data are given as mean ± SEM and compared by Student's t test (C, D, E, F, I, J, K, and L) and one‐way ANOVA (H). The diagram was created in https://BioRender.com.

Figure 2

PFKM knockdown alleviates glycolysis and decreases AF susceptibility in CREM mice. A) The atrial concentration of lactate (n = 6). B) Representative bands and quantification of the protein levels of HK, PFKM, PKM, and LDHA in atrial tissue 3‐month control and 3‐month CREM mice (n = 6). C) The atrial concentration of PFKM in two groups (n = 8). D) qPCR validation of the relative abundance of PFKM in 3‐month, 5‐month, and 7‐month CREM mice left atrial tissue (n = 6). (E) Representative images of atrial Immunohistochemistry staining of mice. Scale bar = 50um F) The correlation between expression of PFKM and atrial lactate levels (n = 20). G) Schematic diagram showing the establishment of the PFKM‐knockout model in CREM mice. H) Representative bands and quantification of the protein levels of PFKM in two groups (n = 6). I) The atrial concentration of lactate (n = 6). J) Representative examples of AF induction attempt in the CREM‐Con and in the CREM + AAV‐PFKM group. K) AF inducibility. L) Representative echocardiography and images of LA dimensions in the CREM‐Con and CREM+AAV‐PFKM group. M) LA diameters in the two groups (n = 10). N) Representative images of atrial Masson staining and the collagen volume fraction in atria of two groups. (n = 6). Scale bar = 50um O) qPCR validation of the relative abundance of Col3, α‐SMA, and TGF‐β1 in atrial tissue with two groups (n = 6). P) Representative bands and quantification of the protein levels of Col3, α‐SMA and TGF‐β1 in atrial tissue with two groups (n = 6). AF = atrial fibrillation; Con = control; LA = left atrium. The data are given as mean ± SEM and compared by Student's t test (A, B, C, H, I, M, N, O, and P), one‐way ANOVA (D). AF inducibility (K) was compared by Fisher exact test. The diagram was created in https://BioRender.com.

Figure 3

PFKM‐mediated activation of glycolysis is the primary driver of AF. A) Schematic diagram showing the establishment of atrial overexpression PFKM model in mice. Mice were injected with adeno‐associated virus through the tail vein, and AF susceptibility was tested 28 days later. The control group was treated with placebo. B) Representative bands and quantification of the protein levels of HK, PFKM, PKM and LDHA in two groups (n = 6). C) The atrial concentration of lactate (n = 6). D,E) Representative examples of AF induction attempts and AF inducibility in two groups (n = 10). F) Representative echocardiography and images of LA dimensions in the AAV‐Con and AAV‐OE‐PFKM group. G) LA diameters in the two groups (n = 6). H) Representative images of atrial Masson's staining and the collagen volume fraction in atria of two groups (n = 6). Scale bar = 50um I) qPCR validation of the relative abundance of Col3, α‐SMA, and TGF‐β1 in atrial tissue with two groups (n = 6). J) Representative bands and quantification of the protein levels of Col3, α‐SMA and TGF‐β1 in atrial tissue with two groups (n = 6). GAPDH was used for normalization. The data are given as mean ± SEM and compared by Student's t test (B,C,G, H, I and J). AF inducibility (E) was compared by the Fisher exact test. The diagram was created in https://BioRender.com.

Figure 4

Lactate induces cardiac fibroblasts activation. A) Schematic representation of atrial‐derived lactate concentrations treated for cardiac fibroblasts and cardiac myocytes. B) Representative optical microscope image showed the morphology of cardiac fibroblasts and cardiac myocytes treated with control and lactate. Scale Bar = 1000um. C) The cell viability of CM (n = 6). D) Representative images of Tunel⁺ staining of CM (n = 5). Scale Bar = 20um. E) Representative bands and quantification of the protein levels of Bax and Bcl2 in CM (n = 6). F) The cell viability of CF in two groups (n = 6). G) Representative images of α‐SMA staining of CF (n = 3). Scale Bar = 100um. H) Representative bands and quantification of the protein levels of Col3, α‐SMA and TGF‐β1 in CF (n = 6). Con = control; lac = lactate. The data are given as mean ± SEM and compared by Student's t test (C, E, F, G, and H). The diagram was created in https://BioRender.com.

Figure 5

Enhanced glycolysis promotes cardiac fibroblasts histone lactylation in the atrium. A) Representative bands and quantification of the protein levels of Pan Kla (n = 7) and H3K18la (n = 6) in atria with SR and AF people. B) Representative bands and quantification of the protein levels of Pan Kla and H3K18la in atria with Con and rapid atrial pacing rabbit (n = 6). C) Representative bands and quantification of the protein levels of Pan Kla and H3K18la in atria with Con and 7‐month CREM mice (n = 6). D) Representative Pan Kla immunohistochemistry staining of Con and 7‐month CREM mice groups (n = 6). Scale bar = 100um. E) Representative H3K18 lactylation immunohistochemistry staining of Con and 7‐month CREM mice groups (n = 6). Scale bar = 20um. F) Immunofluorescence co‐staining for α‐SMA (green) with Pan Kla (red) in the atria with control and h CREM group. Scale bar = 40um. G) Immunofluorescence co‐staining for α‐SMA (green) with H3K18 la (red) in the atria with control and 7‐month CREM group. Scale bar = 50um. H) Representative bands and quantification of the protein levels of Pan Kla and H3K18la in CF isolated from SR and AF individuals (n = 3). I) Representative bands and quantification of the protein levels of Pan Kla and H3K18la in CF isolated from Con and 7‐month CREM mice (n = 6). The data are given as mean ± SEM and compared by Student's t test (A, B, C, D, E, H, and I) or Wilcoxon test (A, C, and E).

Figure 6

H3K18la regulates TGF‐β1 transcription to activate CF. A) Heatmaps for H3K18la binding peaks in control and lactate‐treated CF from mice. Color depth indicates the relative number of reads; genes with similar distribution patterns are clustered together through a clustering algorithm to show the binding trends of lactylation modifications on all genes. B) Volcano plot of differentially expressed genes in two groups, red represents up‐regulated genes and blue represents down‐regulated genes (n = 3). C) The top 2 enriched de novo motifs of the upregulated and downregulated genes with differential H3K18la modification. D) Bubble chart showing the top gene ontology terms of upregulated genes with increased H3K18la modification. E) Bioinformatics analysis filtered TGF‐β1 as downstream targets of H3K18la. Normalized read densities for H3K18la and RNA‐Seq at the TGF‐β1 gene. F) Gene expression analysis by RT‐qPCR and H3K18la occupancy analysis by ChIP‐qPCR in control and lactate treated CF (n = 6). G) Gene expression analysis by RT‐qPCR and H3K18la occupancy analysis by ChIP‐qPCR in CF isolated from the atrium in control and 7‐month CREM mice (n = 6). H) Schematic diagram about cells divided into control group, lactate group, and lactate +si TGF‐β1 group to observe the activation of CF induced by lactate. I) Representative bands and quantification of protein levels of Col3, α‐SMA and TGF‐β1 in CF (n = 6). J) Representative images of α‐SMA staining of CF. Scale bar = 100 µm. The data are given as mean ± SEM and compared by Student's t test (F and G), or one‐way ANOVA (I). The diagram was created in https://BioRender.com.

Figure 7

P300 serves as a writer for H3K18la to mediate TGF‐β1 transcription. A–C) ClusPro and ZDOCK 3.0 were performed to analyze the binding mode between the peptide and mucus P300(A)/MOF(B)/GCN5(C). D) Co‐immunoprecipitation (CoIP) analysis for H3K18la in the presence of P300, MOF, and GCN5 was conducted on cell lysates from CF treated with lactate. The lysates were immunoprecipitated using anti‐P300/MOF/GCN5 antibodies, followed by western blotting to assess H3K18la levels (n = 3). E) Immunofluorescence co‐staining for H3K18la (green) with P300/MOF/GCN5 (red) in CF treated with lactate. Scale bar = 20um. F) qPCR validation of the relative abundance of α‐SMA and TGF‐β1 in CF (n = 6). G) Representative images of α‐SMA staining of CF. Scale bar = 100 µm. H–I) Representative bands and quantification of the protein levels of Col3, α‐SMA, TGF‐β1, and H3K18la in CF, followed by lactate stimulation for 24 h (n = 6). J) Detection of H3K18la occupancy rates (n = 6). The data are given as mean ± SEM and compared by one way ANOVA (F, H, I, and J).