Gut microbiota-activated PLCγ2 monocytes drive atrial fibrillation

Abstract

Atrial fibrillation (AF) is the most common sustained arrhythmia, conferring risks of stroke and heart failure. Monocyte activation-associated inflammation is implicated in AF, yet the recruitment of monocytes and their role in atrial remodeling remain unclear. Here we show that monocytes from AF patients upregulate phospholipase C gamma 2 (PLCG2). In a combined AF male mouse model, phospho-PLCγ2⁺ monocytes are recruited to the atria, where they trigger endothelial‑to‑mesenchymal transition (EndMT) through the secreted phosphoprotein 1 (SPP1)-integrin α9β1 signaling axis. Monocyte-specific deletion of PLCG2 blocks atrial recruitment of these monocytes and reduces AF. Depletion of the gut microbiota with antibiotics reduces the number of phospho-PLCγ2⁺ monocytes in the atria, suppresses EndMT, and attenuates AF inducibility, and these effects are reversed by fecal microbiota transplantation (FMT). Mechanistically, gut microbiota-derived bacterial membrane vesicles activate spleen tyrosine kinase (Syk)/PLCG2 signaling in monocytes. These findings establish a gut microbiota-monocyte-EndMT axis, wherein monocyte-endocardial crosstalk drives AF.

Atrial fibrillation involves systemic inflammatory activation that remodel the atrial endocardium. Here, the authors show that gut microbiota activate PLCy2 positive monocytes, promoting endothelial-to-mesenchymal transition and atrial fibrillation through a microbiota–monocyte–endocardial axis.

Introduction

Atrial fibrillation (AF), the most prevalent sustained cardiac arrhythmia, arises from a multifactorial etiology encompassing electrophysiological disturbances, structural/electrical remodeling, and molecular mechanisms involving inflammatory cascades and oxidative stress^1,2. Established clinical risk factors, including advanced age, tobacco use, alcohol abuse, and comorbidities (e.g., heart failure, hypertension, diabetes mellitus), synergistically potentiate AF progression^3,4. Contemporary therapeutic paradigms emphasize dual approaches: pharmacological management and catheter-based interventions, both guided by rate and rhythm control strategies^5–7. Nevertheless, therapeutic efficacy demonstrates significant interindividual variability mediated by age-related pathophysiological changes, immune system modulation, and alterations in metabolic homeostasis^8,9. Crucially, pro-inflammatory mediators (e.g., NLRP3 inflammasome, NF-κB signaling pathway, TNF-α, IL-6, CRP) and reactive oxygen species synergistically drive cardiomyocyte injury through calcium dysregulation mechanisms^10,11. The pathological synergy promotes RyR2-mediated Ca²⁺ release, facilitating atrial ectopic depolarization and perpetuating arrhythmogenic substrates via three interconnected pathways: (1) enhancement of myocardial remodeling and fibrosis¹², (2) disruption of electrical conduction homogeneity¹³, and (3) generation of micro-reentry circuits¹³. While anti-inflammatory therapies demonstrate potential in reducing postoperative AF incidence through cytokine modulation, current clinical trials reveal limited therapeutic success¹⁴. The translational gap likely stems from the inherent complexity of inflammatory pathway interactions and current limitations in achieving pathway-specific therapeutic targeting¹⁴.

Immune cells constitute essential cellular components within cardiac tissue, particularly in the atrial chambers¹⁵. Activated monocytes and resident macrophages orchestrate atrial remodeling through dual mechanisms: initiating inflammatory cascades via cytokine secretion (e.g., IL-1β, IL-6, TNF-α) and directly participating in cellular clearance and scar formation¹⁶. The atrial macrophage population predominantly originates from circulating monocyte differentiation¹⁷. During acute injury phases or chronic cardiac pathology, these recruited monocytes differentiate into CCR2⁺ MHC-II^high macrophages that facilitate necrotic cell clearance while amplifying inflammation through pro-inflammatory mediators¹⁸. In contrast, CCR2^-MHC-II^low macrophage subsets emerge during later reparative stages, promoting extracellular matrix remodeling through matrix metalloproteinase (MMP) secretion and TGF-β-mediated fibroblast-to-myofibroblast transdifferentiation¹⁹. Notably, macrophages exhibit electrophysiological interactions with cardiomyocytes via connexin 43 (Cx43) gap junctions, synchronizing their membrane potentials with cardiac electrical activity²⁰. Experimental models demonstrate that Cx43 knockdown or macrophage depletion induces conduction abnormalities²¹, suggesting their potential role in arrhythmogenesis, though direct involvement in AF pathogenesis requires further validation. The inflammatory milieu of early atrial fibrillation features rapid monocyte recruitment and monocyte chemoattractant protein-1 (MCP-1) secretion, establishing a self-perpetuating cycle through adhesion molecule upregulation (e.g., ICAM-1/VCAM-1) and interleukin induction (e.g., IL-1, IL-6, TNF-α)²². These inflammatory mediators drive atrial electrical remodeling via distinct pathways: IL-6 suppresses connexin expression, while TNF-α impairs ion channel function²³. Concurrently, MCP-1/CCR2 signaling activates MCPIP transcription factors that promote cardiomyocyte apoptosis and structural reorganization²⁴. The resultant myocardial fibrosis creates both a substrate for arrhythmia maintenance and anatomical barriers disrupting normal conduction patterns, ultimately establishing the pathophysiological foundation for AF progression²⁵.

Endothelial-to-Mesenchymal Transition (EndMT) describes the phenotypic transformation of endothelial cells characterized by morphological, structural, and functional alterations under physiological or pathological stimuli²⁶. This process involves the acquisition of mesenchymal features, including SMA/TWIST expression²⁶. Experimental evidence implicates EndMT in murine myocardial fibrosis, with approximately 27–35% of cardiac fibroblasts originating from endothelial-derived mesenchymal cells²⁷. Notably, suppression of EndMT correlates with attenuated fibrotic progression²⁸. Pro-inflammatory cytokines can induce microvascular endothelial cells to adopt fibroblast-like properties, evidenced by α-SMA, calponin, and collagen I expression, confirming EndMT activation in inflammatory milieus²⁹. While endocardial EndMT and fibrosis are recognized hallmarks of AF, the mechanistic interplay between monocyte-derived signals and EndMT regulation remains undefined.

Metabolic dysregulation-associated conditions, including hypertension, hyperglycemia, and obesity, constitute established risk factors for AF pathogenesis¹³. Atrial tissues exhibit intrinsic metabolic vulnerability due to diminished metabolic reserves compared to ventricular myocardium, rendering them susceptible to stressors (e.g., hyperglycemia, pressure overload, aging)³⁰. This metabolic susceptibility predisposes to electromechanical remodeling and arrhythmogenic substrate formation³⁰. Emerging evidence highlights gut microbiota as a modulator of cardiac fibrosis via metabolic pathway regulation, with specific microbial shifts (e.g., Nitrosomonas depletion, Amygdalaceae enrichment) correlating with elevated IL-6, STAT3, and IL-17 levels that sustain low-grade inflammation and AF perpetuation^31,32. Zuo et al. reported that disordered gut microbiota and alterations in microbial metabolite profiles were significantly associated with atrial fibrillation³³. Nevertheless, the potential tripartite interaction between gut microbiota, monocyte signaling, and EndMT activation in AF remains unexplored. In the present study, we aimed to investigate whether gut microbiota‑derived signals promote monocyte recruitment and activation in the atrium and their role in driving EndMT and AF progression. We demonstrate that phospho-PLCγ2⁺ monocytes preferentially infiltrate atrial tissue in AF models, and that this infiltration depends on the gut microbiota. Mechanistically, PLCγ2 activation in monocytes triggers EndMT via the SPP1-integrin α9β1 signaling axis, ultimately establishing a pro‑arrhythmic substrate. Our findings identify the gut microbiota-monocyte-EndMT axis as a potential driver of AF progression.

Results

Single-cell profiling of AF atrial tissues

To characterize cellular and transcriptional diversity during AF, scRNA-seq was performed on atrial tissues from AF patients and controls (Fig. 1A). Initial quality control metrics, including mean reads per cell, median UMI counts, median genes per cell, and the percentage of mitochondrial reads, were comparable between the AF and control groups (Supplementary Fig. 1C–F), confirming data robustness for downstream analysis. A total of 31,334 high-quality single-cell transcriptomes were captured and clustered into 16 distinct populations using UMAP analysis (Fig. 1B). These 16 clusters were further classified into 10 major cell types based on canonical marker genes (Fig. 1B, C). A total of 36,601 genes were identified across the 16 clusters. Pseudotime trajectory analysis further suggested distinct cellular state dynamics between the groups (Supplementary Fig. 1B). Comparative analysis showed the proportions of cell clusters between control and AF groups, with an increase in the proportion of monocytes in AF group compared to controls (Supplementary Fig. 1A). We next examined the expression of monocyte-specific markers. Feature plot analysis revealed that FCN1, a highly specific human monocyte marker, was strongly and selectively expressed in this cluster (Fig. 1D). The same cluster also showed positive expression of the pan-myeloid markers LYZ (Supplementary Fig. 1H), S100A8 (Supplementary Fig. 1I), and S100A9 (Supplementary Fig. 1J). To distinguish monocyte subsets, we evaluated the expression of CD14 and FCGR3A (Fig. 1E, F). A dot plot of monocyte-specific markers confirmed the enrichment of these genes in the monocyte population (Supplementary Fig. 1G). To further investigate the potential cell-cell communication between monocytes and other cell types, we analyzed specific interaction markers (Fig. 1G). To further investigate the role of monocytes in AF, DEGs in monocytes from AF and control atrial tissues were analyzed. PLCG2 was identified as a highly expressed gene in monocytes from AF tissues (Fig. 1H). Heatmaps illustrate the increased expression of PLCG2 in monocytes of AF groups (Fig. 1I). Visualization via feature plots demonstrated an increase in PLCG2 gene expression at the transcriptional level in AF monocytes relative to control monocytes (Supplementary Fig. 1K, L). Since CD14 is a canonical surface marker of monocytes, we assessed the proportion of phospho-PLCγ2⁺CD14⁺ cells in atrial tissues. As expected, phospho-PLCγ2⁺CD14⁺ monocytes were significantly increased in AF patients compared to controls (Fig. 1J). To validate PLCG2 expression in monocytes, monocytes were isolated from AF and control atrial tissues. Quantitative analysis confirmed that PLCG2 expression was significantly upregulated in monocytes from AF tissues compared to controls (Fig. 1K).

Fig. 1. Single-cell transcriptional landscape of AF atrial tissue.

A scRNA-seq was performed on atrial tissues from AF patients and controls. UMAP plots show the distribution of cells, with colors representing different specimens. B UMAP plots visualize cell types, with colors indicating distinct cell populations. C Dotplot showing marker genes of cell subtypes. D Feature plot showing FCN1 gene expression levels across all cells. E Feature plot showing CD14 gene expression levels across all cells. F Feature plot showing FCGR3A gene expression levels across all cells. G Dot plots illustrating specific cell-cell communication pathways between monocyte and other cells. H Volcano plots highlight DEGs in monocytes, with PLCG2 annotated. Colors indicate statistical significance based on the following criteria: grey (non-significant), green (significant Log2FC only), blue (significant q-value only), and red (significant for both Log2FC and q-value). I Heatmaps illustrate expression of PLCG2 in monocytes of AF and control groups. J Immunofluorescence staining of phospho-PLCγ2 (red) and CD14 (green) in atrial tissues, with DAPI (blue) staining nuclei. White arrows indicate CD14⁺phospho-PLCγ2⁺ cells, scale bar: 50 µm. K Monocytes isolated from AF and control tissues were analyzed for PLCG2 expression using violin plots (n = 15 control atria, n = 15 AF atria). Data are expressed as the mean ± SD. p values were determined by a two-tailed t-test (K). The source data are provided as a Source data file.

The PLCG2-diacylglycerol (DAG)-protein kinase C (PKC) pathway was activated in monocytes from combined AF (cAF) mice

To investigate the molecular mechanisms underlying AF, a cAF mouse model was established by feeding mice a high-fat diet (HFD) for 4 months followed by angiotensin II (Ang II) infusion for 1 month (Supplementary Fig. 2A), as previously described^17,34. Sham-operated normotensive mice served as controls. Representative atrial electrogram recordings from sham and cAF mice are shown in Supplementary Fig. 2B. The inducibility rate and duration of AF were significantly increased in cAF mice compared to sham controls (Supplementary Fig. 2C, D). To investigate the role of PLCG2 in monocytes during AF, monocytes were isolated from atrial tissues of cAF and sham mice. The number of phospho-PLCγ2 positive cells was significantly higher in cAF mice compared to sham controls (Supplementary Fig. 2E, F). To evaluate the function of monocytes, we determined the migration of monocytes in cAF mice (Supplementary Fig. 2G). After stimulation with 5 ng/mL MCP-1, the number of migrated monocytes was significantly increased in cAF mice compared to sham controls (Supplementary Fig. 2H). Since the DAG-PKC and Ca²⁺ signaling pathways are downstream of PLCG2 activation³⁵, we measured and inositol trisphosphate (IP3) and DAG levels in monocytes. Both IP3 and DAG concentrations were significantly elevated in monocytes from cAF mice compared to sham controls (Supplementary Fig. 2I, J). Western blot analysis revealed increased protein levels of phospho-PLCγ2, PKCδ, and calmodulin (CaM) in monocytes from cAF mice (Supplementary Fig. 2K). Additionally, intracellular Ca²⁺ levels were assessed in monocytes using the fluorescent indicator Fluo-3 AM. Quantitative analysis revealed a significantly higher Fluo-3 AM fluorescence intensity in monocytes isolated from cAF mice compared to those from sham controls, indicating elevated cytosolic Ca²⁺ concentrations in the cAF condition (Supplementary Fig. 2L, M). To quantitatively analyze intracellular calcium [Ca²⁺]ᵢ dynamics, THP-1 cells were treated with shPLCG2 or shCtrl and then stimulated with 1 µM thapsigargin (TG) in a calcium-free extracellular medium. Fluorescence values (F) were converted to the relative change ratio ΔF/F₀, where ΔF/F₀ = (F-F₀)/F₀. As shown in Supplementary Fig. 3A, B, control cells displayed a characteristic slow rise in [Ca²⁺]ᵢ, reaching a plateau of ΔF/F₀ = 2.285 ± 0.15, which reflects the steady-state balance between TG-induced passive ER leak and cellular calcium clearance mechanisms. In contrast, shPLCG2-treated THP-1 cells exhibited a significantly attenuated response, reaching a plateau of only ΔF/F₀ = 1.049 ± 0.07 (Supplementary Fig. 3A, B). We next investigated how PLCG2-dependent store content translates to amplified calcium influx under conditions permissive for store-operated calcium entry (SOCE). THP-1 cells overexpressing PLCG2 were stimulated with 10 µM cyclopiazonic acid (CPA) in the presence of a 10 mM calcium concentration with or without a specific Orai1 channel inhibitor GSK-7975A. PLCG2 overexpressing THP-1 cells exhibited a dramatically potentiated response, with the increased peak and sustained plateau (Supplementary Fig. 3C). GSK-7975A had no significant effect on the initial CPA-induced calcium peak (Supplementary Fig. 3C). However, GSK-7975A treatment caused the [Ca²⁺]ᵢ signal to rapidly decay after the peak, abolishing the high sustained plateau (Supplementary Fig. 3C). These data demonstrate that PLCG2 activation in monocytes potentiates Orai1-mediated SOCE to drive prolonged cytosolic calcium elevation.

Phospho-PLCγ2⁺ monocytes were recruited to the atrium in cAF mice

To investigate the role of PLCG2 in monocyte mobilization during AF, we analyzed phospho-PLCγ2⁺ monocytes in the blood of cAF mice. CD45⁺ cells were gated and analyzed in the blood of cAF mice (Fig. 2A). Monocytes were identified as CD45⁺CD11b⁺Ly6G⁻ cells, and their proportion was quantified (Fig. 2B). Subsequently, Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ monocytes were analyzed within the CD45⁺CD11b⁺Ly6G⁻ gate (Fig. 2C). The proportion of CD45⁺ CD11b⁺Ly6G⁻ cells was significantly increased in the blood of cAF mice compared to sham controls (Fig. 2D). Similarly, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes were significantly higher in cAF mice (Fig. 2E, F). Notably, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hiphospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes were significantly increased in the blood of cAF mice compared to controls (Fig. 2G, H). To further investigate the role of phospho-PLCγ2 in the atrium of cAF mice, we analyzed phospho-PLCγ2⁺ monocytes in atrial tissues. CD45⁺ cells were gated and analyzed in the atrium of cAF mice (Fig. 2I). The proportion of CD45⁺CD11b⁺Ly6G⁻ cells was quantified (Fig. 2J). Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ monocytes were analyzed within the CD45⁺CD11b⁺Ly6G⁻ gate (Fig. 2K). The proportion of CD45⁺CD11b⁺Ly6G⁻ cells was significantly increased in the atrium of cAF mice compared to controls (Fig. 2L). Interestingly, while the proportion of CD45⁺CD11b⁺ Ly6G⁻Ly6C^hi monocytes remained unchanged, the proportion of CD45⁺CD11b⁺Ly6G⁻ Ly6C^lo monocytes was significantly increased in cAF mice (Fig. 2M, N). Moreover, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hiphospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes were significantly increased in the atrium of cAF mice compared to controls (Fig. 2O, P). These results suggest that CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes and PLCγ2⁺ monocytes are recruited and colonize the atrium in cAF mice.

Fig. 2. phospho-PLCγ2+ monocytes were recruited in atrium of male cAF mice.

A The illustration gate of the CD45⁺ cells in blood of male cAF mice. B The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells in blood of male cAF mice. C The illustration gate of the Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ cells within CD45⁺ CD11b⁺Ly6G⁻ population. D Percentage of CD45⁺CD11b⁺Ly6G⁻ cells in blood of male cAF mice (n = 5 mice). E Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi monocytes in blood of male cAF mice (n = 5 mice). F Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in blood of male cAF mice (n = 5 mice). G Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^hiphospho-PLCγ2⁺ monocytes in blood of male cAF mice (n = 5 mice). H Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo phospho-PLCγ2⁺ monocytes in blood of male cAF mice (n = 5 mice). I The illustration gate of the CD45⁺ cells in atrium of male cAF mice. J The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells in atrium of male cAF mice. K The illustration gate of the Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ cells within CD45⁺CD11b⁺ Ly6G⁻ population. L Percentage of CD45⁺CD11b⁺Ly6G⁻ cells in atrium of male cAF mice (n = 5 mice). M Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi monocytes in atrium of male cAF mice (n = 5 mice). N Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in atrium of male cAF mice (n = 5 mice). O Percentage of CD45⁺CD11b⁺Ly6G^-Ly6C^hiphospho-PLCγ2⁺ monocytes in atrium of male cAF mice (n = 5 mice). P Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo phospho-PLCγ2⁺ monocytes in atrium of male cAF mice (n = 5 mice). Data are expressed as the mean ± SD. p values were determined by a two-tailed t-test (D–H, L–P). The source data are provided as a Source data file.

Phospho-PLCγ2⁺ monocytes promoted EndMT through the SPP1-α9β1 signaling pathway in cAF mice

Based on the cell-cell interaction analysis, we identified the SPP1-α9β1 signaling axis as a key pathway between monocytes and endothelial cells (Fig. 1C), suggesting its critical role in the function of phospho-PLCγ2⁺ monocytes in cAF. Endothelial cells, which form a single layer of squamous cells, are a key component of the endocardial surface in the heart³⁶. Endothelial dysfunction has been implicated in promoting atrial arrhythmic substrates, inducing thromboembolism, and predicting AF recurrence after cardioversion or ablation therapy³⁷. Given that endothelial dysfunction is implicated in atrial arrhythmogenesis, we investigated the occurrence of EndMT in the atrial endocardium of cAF mice. Immunofluorescence staining revealed increased levels of SMA (Fig. 3A) and TWIST (Fig. 3C) in the endocardial endothelial cells of cAF mice compared to controls. Quantification confirmed a significantly higher SMA/CD31 intensity ratio (Fig. 3B) and TWIST/CD31 intensity ratio (Fig. 3D) in cAF mice. Western blot analysis further demonstrated an increased expression levels of key EndMT-related proteins, including SMA, TWIST, Snail, Slug, and Col1a1, whereas a slightly decreased expression levels of VE-cadherin in the atrial tissue of cAF mice (Fig. 3E). Consistent with the protein data, qPCR analysis confirmed that the increased mRNA levels of Acta2, TWIST, Col1a1, Snail, and Slug in the atrial tissue of cAF mice, whereas VE-cadherin mRNA levels showed no significant change (Fig. 3H–M). Furthermore, masson’s trichrome staining indicated enhanced collagen deposition in the atrial tissue of cAF mice (Fig. 3F, G). These results collectively indicate enhanced EndMT in the atrial tissue of cAF mice. We next utilized an in vitro co-culture system to establish the role of PLCγ2⁺ monocytes in EndMT. Monocytes were transfected with PLCγ2-overexpressing vectors (overexpression validated in Supplementary Fig. 4B) and co-cultured with atrial endothelial cells. Western blot analysis of endothelial cell lysates revealed that PLCγ2-overexpressing monocytes robustly induced the expression of EndMT-related proteins (Supplementary Fig. 4A). To functionally validate the role of the SPP1-α9β1 axis in monocyte-endothelial cell interaction, we generated monocyte-specific integrin α9β1 knockout (M-α9β1^−/−) mice, with knockout efficiency confirmed by loss of α9 protein in isolated monocytes (Supplementary Fig. 4C). Crucially, the pro-EndMT effect was attenuated when PLCγ2-overexpressing monocytes from M-α9β1^−/− mice were used in the co-culture assay (Supplementary Fig. 4D). Similarly, knockdown of SPP1 in monocytes (validated in Supplementary Fig. 4E) abolished the ability of PLCγ2-overexpressing monocytes to trigger EndMT in endothelial cells (Supplementary Fig. 4H). These data genetically demonstrate that the SPP1-α9β1 axis is essential for phospho-PLCγ2⁺ monocytes to induce EndMT. In addition, collagen deposition was significantly reduced in the atrial tissue of M-α9β1^−/− cAF mice compared to WT cAF mice (Supplementary Fig. 4F, G). These results demonstrate that the reduction in EndMT observed in M-α9β1^−/− cAF mice is associated with a concomitant decrease in atrial fibrosis. Immunohistochemical staining for CD11b and CD31 in atrial tissues showed that the number of CD31⁺CD11b⁺ double-positive cells, indicating direct proximity or interaction, was significantly reduced in M-α9β1^−/− cAF mice compared to WT cAF mice (Supplementary Fig. 4I, J).

Fig. 3. phospho-PLCγ2⁺ monocytes promoted EndMT through the SPP1-α9β1 signaling pathway in male cAF mice.

A Immunohistochemical staining of SMA (red) and CD31 (green) in the atrial endocardium of male cAF mice, scale bar: 50 µm. B Quantitation of SMA/CD31 fluorescence intensity ratio (n = 5 mice). C Immunohistochemical staining of TWIST (red) and CD31 (green) in the atrial endocardium of male cAF mice, scale bar: 50 µm. D Quantitation of TWIST/CD31 fluorescence intensity ratio (n = 5 mice). E Western blot analysis of EndMT-related protein expression levels in the atrial tissue of male cAF mice. F Masson’s trichrome staining in the atrial tissue of male cAF mice. Blue indicates collagen deposition, scale bar: 25 μm. G Quantification of collagen deposition in the atrial tissue of male cAF mice (n = 5 mice). H qPCR analysis of TWIST mRNA levels in the atrial tissue of male cAF mice (n = 5 mice). I qPCR analysis of Acta2 mRNA levels in the atrial tissue of male cAF mice (n = 5 mice). J qPCR analysis of VE-cadherin mRNA levels in the atrial tissue of male cAF mice (n = 5 mice). K qPCR analysis of Col1a1 mRNA levels in the atrial tissue of male cAF mice (n = 5 mice). L qPCR analysis of Snail mRNA levels in the atrial tissue of male cAF mice (n = 5 mice). M qPCR analysis of Slug mRNA levels in the atrial tissue of male cAF mice (n = 5 mice). Data are expressed as the mean ± SD. p values were determined by a two-tailed t-test (B, D, G–M). n = 3 independent experiments (E). The source data are provided as a Source data file.

Monocyte-specific knockout of PLCG2 attenuated AF inducibility and EndMT in cAF mice

To investigate the role of PLCG2 in monocytes during AF, we generated monocyte-specific PLCG2 knockout (M-PLCG2^−/−) mice and subjected them to sham or cAF conditions. Representative atrial electrogram recordings from M-PLCG2^−/− sham and cAF mice are shown in Fig. 4A. The AF inducibility rate was significantly increased in cAF mice compared to sham controls but was reduced in M-PLCG2^−/− cAF mice (Fig. 4B). Similarly, the duration of AF episodes was significantly shorter in M-PLCG2^−/− cAF mice compared to wild-type (WT) cAF mice (Fig. 4C). To assess the functional impact of PLCG2 knockout, monocytes were isolated from M-PLCG2^−/− sham and cAF mice. The number of phospho-PLCγ2 positive cells was significantly reduced in both M-PLCG2^−/−sham and M-PLCG2^−/− cAF mice compared to their WT counterparts (Fig. 4D, E). To investigate downstream signaling, IP3 and DAG levels were measured in monocytes. Both IP3 and DAG concentrations were significantly lower in M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 4F, G). Western blot analysis confirmed reduced expression of phospho-PLCγ2, PKCδ, and CaM in monocytes from M-PLCG2^−/− cAF mice (Fig. 4H). To evaluate the function of monocytes, monocyte migration was assessed using a transwell assay (Fig. 4I). After stimulation with 5 ng/mL MCP-1, the number of migrated monocytes was significantly reduced in M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 4I, J). Additionally, assessment of intracellular Ca²⁺ levels using the fluorescent indicator Fluo-3 AM revealed that monocytes from M-PLCG2^−/− cAF mice exhibited significantly lower Fluo-3 AM fluorescence intensity compared to those from WT cAF mice, indicating that PLCG2 deletion attenuated the pathological elevation of cytosolic Ca²⁺ in monocytes (Fig. 4K, L). To evaluate the impact of monocyte-specific PLCG2 knockout on endothelial cell function, EndMT markers were analyzed in the atrial endocardium. Immunofluorescence staining showed reduced levels of SMA in the atrial endocardium of M-PLCG2^−/− cAF mice (Supplementary Fig. 5A). The SMA/CD31 intensity ratio was significantly lower in M-PLCG2^−/− cAF mice compared to WT cAF mice (Supplementary Fig. 5B). Similarly, TWIST expression was reduced in the atrial endocardium of M-PLCG2^−/− cAF mice (Supplementary Fig. 5C), and the TWIST/CD31 intensity ratio was significantly lower compared to WT cAF mice (Supplementary Fig. 5D). These results demonstrate that monocyte-specific PLCG2 knockout attenuates AF inducibility and EndMT in cAF mice.

Fig. 4. Monocyte specific knockout of PLCG2 attenuated AF inducibility in male cAF mice.

A Representative atrial electrogram recordings from male M-PLCG2^−/− sham and cAF mice. Solid underlining represents burst pacing, and dashed underlining indicates AF. B AF induction success rate (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). C Total AF duration in male cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). Data points represent the cumulative duration of all AF episodes ≥250 ms for each mouse. D Monocytes were isolated from atrial tissues of male M-PLCG2^−/− sham and cAF mice. Immunofluorescence staining of phospho-PLCγ2 in monocytes of male M-PLCG2^−/− sham and cAF mice, scale bar: 50 µM. E Quantification of phospho-PLCγ2 monocytes (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). F, G Intracellular DAG and IP3 concentrations in monocytes of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). H Western blot analysis of of phospho-PLCγ2, PKCδ, CaM in monocytes of male M-PLCG2^−/− sham and cAF mice. I Representative fluorescence images of migrated monocytes from male M-PLCG2^−/− sham and cAF mice toward MCP-1 (5 ng/mL), scale bar: 50 µM. J Quantification of migrated monocytes (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). K Measurement of intracellular Ca²⁺ levels in monocytes from atrial tissues of male M-PLCG2^−/− sham and cAF mice using the fluorescent indicator Fluo-3 AM, scale bar: 50 µM. L Quantification of the Fluo-3 AM fluorescence intensity in monocytes (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). Data are expressed as the mean ± SD. p values were determined by one-way ANOVA with Tukey’s test (C, E–G, J, L). n = 3 independent experiments (H). The source data are provided as a Source data file.

Monocyte-specific knockout of PLCG2 suppressed the recruitment and colonization of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in the atrium of cAF mice

To investigate the role of PLCG2 in monocyte mobilization during AF, we analyzed phospho-PLCγ2⁺ monocytes in the blood of M-PLCG2^−/− sham and cAF mice. Monocytes were gated as CD45⁺CD11b⁺Ly6G⁻ cells, and their proportion was quantified (Fig. 5A). The proportion of CD45⁺CD11b⁺Ly6G⁻ monocytes was significantly reduced in the blood of M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 5B). Within this population, Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lo phospho-PLCγ2⁺ monocytes were further analyzed (Fig. 5C). The proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi and CD45⁺CD11b⁺Ly6G^-Ly6C^lo monocytes were significantly decreased in the blood of M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 5E, F). Notably, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi phospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes were also significantly reduced in the blood of M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 5D, G). To further investigate the role of PLCG2 in atrial monocyte colonization, we analyzed phospho-PLCγ2⁺ monocytes in the atrium of M-PLCG2^−/− sham and cAF mice. Monocytes were gated as CD45⁺CD11b⁺Ly6G⁻, and their proportion was quantified (Fig. 5J). The proportion of CD45⁺CD11b⁺Ly6G^- monocytes was significantly reduced in the atrium of M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 5K). Within this population, Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ monocytes were further analyzed (Fig. 5L). The proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes was significantly decreased in the atrium of M-PLCG2^−/− cAF mice compared to WT cAF mice (Fig. 5H), while the proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi monocytes showed no significant change (Supplementary Fig. 3D). As expected, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi phospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes were both significantly reduced in M-PLCG2^−/− cAF mice (Fig. 5I, M). These results demonstrate that monocyte-specific knockout of PLCG2 suppresses the recruitment and colonization of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in the atrium of cAF mice.

Fig. 5. Monocyte specific knockout of PLCG2 suppressed colonization of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in atrium of male cAF mice.

A The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells in blood of male M-PLCG2^−/− sham and cAF mice. B Percentage of CD45⁺CD11b⁺Ly6G⁻ cells in blood of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). C The illustration gate of the Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lo phospho-PLCγ2⁺ cells within CD45⁺CD11b⁺Ly6G⁻ population. D Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hiphospho-PLCγ2⁺ monocytes in blood of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). E Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi monocytes in blood of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). F Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in blood of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). G Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes in blood of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). H Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in atrium of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). I Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes in atrium of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). J The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells in atrium of male M-PLCG2^−/− sham and cAF mice. K Percentage of CD45⁺CD11b⁺Ly6G⁻ cells in atrium of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). L The illustration gate of the Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ cells within CD45⁺CD11b⁺Ly6G⁻ population. M Percentage of CD45⁺CD11b⁺Ly6G⁻ Ly6C^hiphospho-PLCγ2⁺ monocytes in atrium of male M-PLCG2^−/− sham and cAF mice (n = 6 mice for Sham+WT group, n = 7 mice for remaining each group). Data are expressed as the mean ± SD. p values were determined by one-way ANOVA with Tukey’s test (B, D–I, K, M). The source data are provided as a Source data file.

Gut microbiota are required for phospho-PLCγ2⁺ monocyte-mediated regulation of AF inducibility and EndMT in cAF mice

PLCG2-mediated DAG lipid metabolism is influenced by intestinal microbiota³⁸, and gut microbiota have been implicated in the regulation of cardiovascular diseases, including AF³⁹. To investigate the role of gut microbiota in the colonization of phospho-PLCγ2⁺ monocytes during AF, we depleted gut microbiota using antibiotics in sham and cAF mice. Representative atrial electrogram recordings from antibiotics-treated sham and cAF mice are shown in Fig. 6A. The AF inducibility rate was significantly reduced in antibiotics-treated cAF mice compared to controls (Fig. 6B). Similarly, the duration of AF episodes was significantly shorter in antibiotics-treated cAF mice (Fig. 6C). To assess the functional impact of gut microbiota depletion, monocytes were isolated from antibiotics-treated sham and cAF mice. The number of phospho-PLCγ2 positive cells was significantly reduced in antibiotics-treated cAF mice compared to controls (Fig. 6D, E). To explore downstream signaling, IP3 and DAG levels were measured in monocytes. Both IP3 and DAG concentrations were significantly lower in antibiotics-treated cAF mice compared to controls (Fig. 6F, G). To evaluate the function of monocytes, monocyte migration was assessed using a transwell assay (Fig. 6H). After stimulation with 5 ng/mL MCP-1, the number of migrated monocytes was significantly reduced in antibiotics-treated cAF mice compared to controls (Fig. 6H, I). Additionally, evaluation of intracellular Ca²⁺ levels via Fluo-3 AM fluorescence showed that monocytes from antibiotics-treated cAF mice had significantly lower Fluo-3 AM intensity compared to those from cAF controls, suggesting that antibiotic intervention mitigated the aberrant calcium elevation in monocytes associated with cAF (Fig. 6J, K). To evaluate the impact of gut microbiota depletion on endothelial cell function, EndMT markers were analyzed in the atrial endocardium. Immunofluorescence staining showed reduced levels of SMA in the atrial endocardium of antibiotics-treated cAF mice (Supplementary Fig. 6A). The SMA/CD31 intensity ratio was significantly lower in antibiotics-treated cAF mice compared to controls (Supplementary Fig. 6B). Similarly, TWIST expression was reduced in the atrial endocardium of antibiotics-treated cAF mice (Supplementary Fig. 6C), and the TWIST/CD31 intensity ratio was significantly lower compared to controls (Supplementary Fig. 6D). To distinguish the specific contribution of gut microbiota depletion from potential pleiotropic antibiotic effects, we performed fecal microbiota transplantation (FMT) and post-antibiotic recovery studies. qPCR analysis confirmed that the combined antibiotic regimen effectively depleted fecal bacterial load, as evidenced by a dramatic reduction in 16S rDNA levels (Supplementary Fig. 3I, K, Day 0). Notably, FMT from cAF donor mice into antibiotic-pretreated recipients successfully restored gut bacteria (Supplementary Fig. 3I). Crucially, this restoration of the microbiota completely reversed the protective effect of antibiotics, as AF inducibility in the FMT group was significantly higher than in mice receiving antibiotics alone (Supplementary Fig. 3J). To further corroborate this finding, we tracked the natural recovery of the gut microbiota and AF susceptibility following antibiotic withdrawal. qPCR analysis revealed a time-dependent rebound in fecal 16S rDNA levels over 28 days after cessation of antibiotic treatment (Supplementary Fig. 3K). In parallel, the protective effect against AF induction gradually diminished, with AF inducibility returning to cAF levels by day 28 (Supplementary Fig. 3L). These results demonstrate that gut microbiota are essential for phospho-PLCγ2⁺ monocyte-mediated regulation of AF inducibility and EndMT in cAF mice.

Fig. 6. Gut microbiota are required for phospho-PLCγ2⁺ monocyte-mediated regulation of AF inducibility in male cAF mice.

A Representative atrial electrogram recordings in antibiotics-treated and/or male cAF mice. Solid underlining represents burst pacing, and dashed underlining indicates AF. B AF induction success rate (n = 7 mice). C Total AF duration in male cAF mice (n = 7 mice). Data points represent the cumulative duration of all AF episodes ≥250 ms for each mouse. D Monocytes were isolated from atrial tissues of antibiotics-treated and/or male cAF mice. Immunofluorescence staining of phospho-PLCγ2 in monocytes, scale bar: 50 µM. E Quantification of phospho-PLCγ2 monocytes (n = 7 mice). F, G Intracellular DAG and IP3 concentrations in monocytes of antibiotics-treated and/or male cAF mice (n = 7 mice). H Representative fluorescence images of migrated monocytes from antibiotics-treated and/or male cAF mice toward MCP-1 (5 ng/mL), scale bar: 50 µM. I Quantification of migrated monocytes (n = 7 mice). J Measurement of intracellular Ca²⁺ levels in monocytes from antibiotics-treated and/or male cAF mice using the fluorescent indicator Fluo-3 AM, scale bar: 50 µM. K Quantification of the Fluo-3 AM fluorescence intensity in monocytes (n = 7 mice). Data are expressed as the mean ± SD. p values were determined by one-way ANOVA with Tukey’s test (C, E–G, I, K). The source data are provided as a Source data file.

Gut microbiota promoted recruitment and colonization of phospho-PLCγ2⁺ monocytes in the atrium of cAF mice

To investigate the role of gut microbiota in monocyte mobilization during AF, we analyzed phospho-PLCγ2⁺ monocytes in the blood of antibiotics-treated sham and cAF mice. The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells was analyzed to determine the proportion of monocytes in the blood of antibiotics-treated sham and cAF mice (Fig. 7A). The proportion of CD45⁺CD11b⁺Ly6G⁻ monocytes was significantly reduced in the blood of antibiotics-treated cAF mice compared to controls (Fig. 7B). Within this population, Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ monocytes were further analyzed (Fig. 7C). The proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes were significantly decreased in the blood of antibiotics-treated cAF mice compared to controls (Fig. 7E, F). Notably, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi phospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo phospho-PLCγ2⁺ monocytes were also significantly reduced in antibiotics-treated cAF mice (Fig. 7D, G). To further investigate the role of gut microbiota in atrial monocyte colonization, we analyzed phospho-PLCγ2⁺ monocytes in the atrium of antibiotics-treated sham and cAF mice. The illustration gate of the CD45⁺CD11b⁺Ly6G^- cells was analyzed to determine the proportion of monocytes in the atrium of antibiotics-treated sham and cAF mice (Fig. 7J). The proportion of CD45⁺CD11b⁺Ly6G⁻ monocytes was significantly reduced in the atrium of antibiotics-treated cAF mice compared to controls (Fig. 7K). Within this population, Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ monocytes were further analyzed (Fig. 7L). The proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes was significantly decreased in the atrium of antibiotics-treated cAF mice compared to controls (Fig. 7H), while the proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi monocytes showed a slight reduction (Supplementary Fig. 3E). Interestingly, the proportions of CD45⁺CD11b⁺Ly6G⁻Ly6C^hiphospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo phospho-PLCγ2⁺ monocytes were both significantly reduced in antibiotics-treated cAF mice (Fig. 7I, M). These results demonstrate that gut microbiota promote recruitment and colonization of phospho-PLCγ2⁺ monocytes in the atrium of cAF mice.

Fig. 7. Gut microbiota promoted recruitment and colonization of phospho-PLCγ2⁺ monocytes in the atrium of male cAF mice.

A The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells in blood of antibiotics-treated and/or male cAF mice. B Proportion of CD45⁺CD11b⁺Ly6G⁻ cells in blood of antibiotics-treated and/or male cAF mice (n = 7 mice). C The illustration gate of the Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ cells within CD45⁺CD11b⁺Ly6G^-population. D Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi phospho-PLCγ2⁺ monocytes in blood of antibiotics-treated and/or male cAF mice (n = 7 mice). E Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^hi monocytes in blood of antibiotics-treated and/or male cAF mice (n = 7 mice). F Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in blood of antibiotics and/or male cAF mice (n = 7 mice). G Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes in blood of antibiotics and/or male cAF mice (n = 7 mice). H Percentage of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in atrium of antibiotics-treated and/or male cAF mice (n = 7 mice). I Proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lo phospho-PLCγ2⁺ monocytes in atrium of antibiotics-treated and/or male cAF mice (n = 7 mice). J The illustration gate of the CD45⁺CD11b⁺Ly6G⁻ cells in atrium of antibiotics-treated and/or male cAF mice. K Percentage of CD45⁺CD11b⁺Ly6G^- cells in atrium of antibiotics-treated and/or male cAF mice (n = 7 mice). L The illustration gate of the Ly6C^hiphospho-PLCγ2⁺ and Ly6C^lophospho-PLCγ2⁺ cells within CD45⁺CD11b⁺ Ly6G⁻ population. M Percentage of CD45⁺CD11b⁺Ly6G^-Ly6C^hiphospho-PLCγ2⁺ monocytes in atrium of antibiotics-treated and/or male cAF mice (n = 7 mice). Data are expressed as the mean ± SD. p values were determined by one-way ANOVA with Tukey’s test (B, D–I, K, M). The source data are provided as a Source data file.

Bacterial MVs activated Syk signaling to promote AF inducibility in cAF mice

To investigate the role of gut microbiota in priming phospho-PLCγ2⁺ monocyte recruitment and colonization during AF, we analyzed fecal components in antibiotics-treated sham and cAF mice. qPCR quantification of bacterial 16S rDNA confirmed significant depletion of gut bacteria in antibiotics-treated groups (Fig. 8A), which was accompanied by marked alterations in phylum-level community composition (Fig. 8B). Syk, a key upstream activator of PLCγ2 signaling³⁵, exhibited increased phosphorylation (p-Syk) in monocytes from cAF mice compared to controls. Notably, antibiotic treatment reduced p-Syk levels in monocytes from cAF mice relative to the untreated cAF group (Fig. 8C). To investigate microbiota-derived mechanisms, we isolated bacterial MVs from cAF mouse feces. TEM confirmed MVs as spherical nanoparticles (Fig. 8D), with an average size of 95.6 nm (Supplementary Fig. 3F). Next, bacterial MVs were labeled with a fluorescent lipophilic dye DiO to track the internalization of MVs in THP-1 cells⁴⁰. DiO-labeled bacterial MVs were internalized by THP-1 cells within 12 h (Fig. 8E). To test Syk’s role in microbiota-mediated PLCγ2 activation, THP-1 cells were treated with 100 μg/mL MVs ± the Syk inhibitor R406 (1 or 10 μM). MV treatment increased p-Syk and phospho-PLCγ2 levels, which were attenuated by R406 in a dose-dependent manner (Fig. 8F). Downstream signaling analysis revealed that R406 (1/10 μM) significantly reduced MV-induced IP3 and DAG production in THP-1 cells compared to MV treatment alone (Fig. 8G, and Supplementary Fig. 3G). Transwell migration assays demonstrated that R406 (1/10 μM) suppressed MV-enhanced THP-1 cell migration toward MCP-1 (5 ng/mL; Fig. 8H, I). Additionally, R406 (1 or 10 μM) significantly attenuated the MV-induced elevation of intracellular Ca²⁺ in THP-1 cells, as quantified by Fluo-3 AM fluorescence intensity (Fig. 8J, K). To validate Syk’s role in vivo, monocyte-specific Syk knockout (M-Syk^−/−) mice were generated and confirmed by loss of Syk protein in isolated monocytes (Supplementary Fig. 3H). M-Syk^−/− cAF mice exhibited reduced AF inducibility and shorter AF duration compared to WT cAF mice (Fig. 8L–N). These results suggest that bacterial MVs activate Syk signaling to promote AF inducibility in cAF mice. To further delineate the specific roles of PLCG2, integrin α9β1, and Syk signaling in phospho-PLCγ2⁺ monocyte activation, we compared both the proportion and absolute number of activated phospho-PLCγ2⁺ monocytes in monocyte-specific knockout mice (M-PLCG2^−/−, M-α9β1^−/−, and M-Syk^−/−) subjected to a high-fat diet combined with angiotensin II infusion. The proportion and absolute number of phospho-PLCγ2⁺ cells were both significantly reduced in CD45⁺CD11b⁺Ly6G^-Ly6C^hi and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in M-PLCG2^−/− and M-Syk^−/− mice compared to controls (Supplementary Fig. 7G–J). In contrast, deficiency of integrin α9β1 did not significantly alter the proportion or absolute number of phospho-PLCγ2⁺ monocytes in either subset compared to controls (Supplementary Fig. 7G–J). To differentiate between monocytes and macrophages, we assessed the surface expression of CCR2 and F4/80 by flow cytometry in both Ly6C^hi and Ly6C^lo populations that were positive for phospho-PLCγ2. Flow cytometric analysis revealed that a proportion of CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes showed positive expression for F4/80 in the atrium of cAF mice (Supplementary Fig. 7A–C). Interestingly, a substantial subset of CD45⁺CD11b⁺Ly6G⁻ Ly6C^hiphospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo phospho-PLCγ2⁺ monocytes expressed the chemokine receptor CCR2 in the atrium of cAF mice (Supplementary Fig. 7D–F).

Fig. 8. Bacterial MVs primed Syk signaling to promote AF inducibility in male cAF mice.

A qPCR quantification of bacterial 16S rDNA in feces of antibiotics-treated and/or male cAF mice (n = 7 mice). B Abundance of bacterial phyla in feces of antibiotics-treated and/or male cAF mice. C Western blot analysis of of p-Syk and Syk in monocytes of antibiotics-treated and/or male cAF mice. D TEM of bacterial MVs, scale bars, 100 nm. E Internalization of DiO-labeled MVs (green) in THP-1 cells at 0 h and 12 h post-treatment, scale bar: 100 µM. F Protein levels of p-Syk, Syk, phospho-PLCγ2, and PLCγ2 in THP-1 cells treated with 100 µg/mL MVs ± R406 (1 or 10 µM). G Intracellular DAG concentration in THP-1 cells (n = 4 independent experiments). H Representative fluorescence images of migrated THP-1 cells toward MCP-1 (5 ng/mL) after MV/R406 treatment, scale bar: 50 µM. I Quantification of migrated THP-1 cells (n = 4 independent experiments). J Measurement of intracellular Ca²⁺ levels in THP-1 cells after treatment with 100 µg/mL MVs ± R406 (1 or 10 µM) using the fluorescent indicator Fluo-3 AM, scale bar: 50 µM. K Fluo-3-AM intensity quantification (n = 4 independent experiments). L Representative atrial electrogram recordings in WT and M-Syk^−/− male cAF mice. Solid underlining represents burst pacing, and dashed underlining indicates AF. M AF induction rate in WT and M-Syk^−/− male cAF mice (n = 6 mice). N Total AF duration in male cAF mice (n = 6 mice). Data points represent the cumulative duration of all AF episodes ≥250 ms for each mouse. Data are expressed as the mean ± SD. p values were determined by a two-tailed t-test (N) and one-way ANOVA with Tukey’s test (A, G, I, K). n = 3 independent experiments (C–F). The source data are provided as a Source data file.

Discussion

The pathogenesis of AF involves a complex interplay of electrophysiological abnormalities, pathophysiological remodeling, and genetic predisposition⁴¹. Emerging evidence highlights inflammatory cascades as key drivers of AF initiation and progression²². Systemic inflammatory activation recruits and activates monocytes, triggering the release of pro-inflammatory cytokines and profibrotic mediators⁴². This inflammatory milieu promotes cardiomyocyte apoptosis and extracellular matrix remodeling, ultimately creating arrhythmogenic substrates through ion channel dysfunction and structural disruption of atrial tissue⁴³. However, the mechanisms underlying monocyte differentiation and colonization in atrial tissue during AF remain poorly understood. To characterize cellular diversity during AF, scRNA-seq was performed on atrial tissues from AF patients and healthy controls. Eleven major cell types were identified, with monocytes showing a significant increase in AF tissues. Differential gene expression analysis revealed elevated PLCG2 expression in monocytes from AF tissues. Immunofluorescence staining confirmed increased phospho-PLCγ2 levels in monocytes isolated from AF atrial tissues compared to controls.

PLCG2, located on human chromosome 16, contains a PH domain, X/Y domain, C2 domain, and SH2 domain, the latter of which acts as an autoinhibitory region⁴⁴. PLCγ2 activation occurs through three primary mechanisms: tyrosine phosphorylation, interaction with Rac GTPase, and SH2 domain-mediated conformational changes⁴⁴. Activated PLCγ2 hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP2) to DAG and IP3, both of which function as secondary messengers⁴⁵. IP3 induces calcium release from the endoplasmic reticulum⁴⁶, while DAG activates PKC to propagate downstream signaling⁴⁷. To investigate the PLCG2-DAG-PKC pathway in monocytes during AF, we established a cAF mouse model by feeding an HFD for 4 months followed by Ang II infusion for 1 month^17,34. Phospho-PLCγ2⁺ monocytes were significantly increased in cAF mice compared to controls. IP3 and DAG levels were elevated, and protein expression of phospho-PLCγ2, PKCδ, and CaM was enhanced in cAF monocytes. Calcium levels and monocyte migration toward MCP-1 (5 ng/mL) were also significantly increased in cAF mice, indicating activation of the PLCG2-DAG-PKC pathway. To further define the role of PLCG2, we generated monocyte-specific PLCG2 knockout (M-PLCG2^−/−) mice. IP3 and DAG levels were significantly reduced in M-PLCG2^−/− cAF monocytes compared to WT controls. Protein expression of phospho-PLCγ2, PKCδ, and CaM was suppressed, and calcium levels and monocyte migration were significantly decreased in M-PLCG2^−/− cAF mice. These results demonstrate that monocyte-specific PLCG2 knockout inhibits the DAG-PKC pathway and monocyte migration in cAF mice.

Monocytes are heterogeneous immune cells that circulate in the blood and mediate innate immune responses to inflammation⁴¹. In mice, monocytes are classified into two distinct populations based on Ly6C expression: Ly6C^hi (pro-inflammatory) and Ly6C^lo (patrolling) subsets⁴⁸. Ly6C^hi monocytes exhibit CCR2hiCX3CR1lo expression, enabling antigen transport to lymph nodes and differentiation into macrophages or dendritic cells (DCs) at inflammatory sites⁴⁹. Conversely, Ly6C^lo monocytes (CX3CR1^hiCCR2^lo) crawl along the vascular endothelium to surveil the vasculature and participate in tissue repair⁵⁰. While Ly6C^hi monocyte activation has been observed in the blood during AF progression¹⁷, the dynamics of atrial monocyte populations remain unclear. In cAF mice, we found increased proportions of both CD45⁺CD11b⁺Ly6G⁻Ly6C^hi and CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes in the blood. However, only CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes were significantly elevated in the atrium, suggesting their predominant role in atrial activation and differentiation during cAF. We hypothesize that Ly6C^hi monocytes primarily drive early inflammatory responses, whereas Ly6C^lo monocytes mediate vascular endothelial cell-dependent myocardial remodeling in late-stage AF. Notably, phospho-PLCγ2⁺ monocytes (CD45⁺CD11b⁺Ly6G⁻Ly6C^hi/Ly6C^lo) were significantly increased in both blood and atrial tissues of cAF mice, indicating their recruitment and colonization. Monocyte-specific PLCG2 knockout (M-PLCG2^−/−) selectively reduced atrial CD45⁺CD11b⁺Ly6G⁻Ly6C^lo monocytes, with only a marginal decrease in Ly6C^hi monocytes. These results demonstrate that PLCG2 predominantly regulates Ly6C^lo monocyte recruitment and colonization in AF pathogenesis.

AF is characterized by disordered electrical activity in the atria, leading to ineffective myocardial contraction, thrombus formation, and impaired blood flow from the atria to the ventricles⁵¹. Endothelial dysfunction is a key contributor to thrombogenesis in AF⁵². Morphologically, endocardial endothelial cells form a thin, transparent layer covering the myocardium. Pathological mechanical stress during cardiovascular disease can induce cytoskeletal reorganization, altering endothelial function and promoting endothelial dysfunction⁵³. Endothelial cells produce molecules that inhibit platelet reactivity, regulate fibrinolysis, and stimulate antithrombotic pathways⁵⁴. Dysfunctional endothelial cells reduce the production of these protective molecules, resulting in a prothrombotic state and increased secretion of pro-inflammatory cytokines⁵⁵. EndMT, a specialized form of epithelial-mesenchymal transition (EMT), plays critical roles in cardiovascular disease development^56,57. EndMT has been implicated in endocardial fibrosis during AF^58,59. To investigate endothelial cell involvement in cAF, we analyzed EndMT markers in the atrial endocardium of cAF mice. Immunofluorescence staining revealed increased levels of SMA and TWIST in the atrial endocardium of cAF mice, indicating EndMT activation. To assess the impact of monocyte-specific PLCG2 knockout on endothelial function, we analyzed EndMT markers in M-PLCG2^−/− cAF mice. Notably, SMA and CD31 levels were significantly reduced in the atrial endocardium of M-PLCG2^−/− cAF mice compared to WT controls, demonstrating that monocyte-specific PLCG2 knockout suppresses EndMT in cAF.

To investigate the relationship between monocytes and EndMT during AF, scRNA-seq was employed to analyze potential cell-cell communication. CellPhoneDB analysis identified specific signaling markers between monocytes and endothelial cells. We identified the SPP1-α9β1 signaling axis as a key interaction, critical for phospho-PLCγ2⁺ monocyte function in cAF mice. Previous studies demonstrated that SPP1⁺ macrophages contribute to cardiac microenvironment formation during AF¹⁷. Integrin α9β1, a cell adhesion molecule, promotes vascular endothelial cell proliferation and differentiation⁶⁰ and regulates adhesion, migration, and angiogenesis via direct binding to VEGFA^61,62. To validate the SPP1-α9β1 axis, we generated monocyte-specific integrin α9β1 knockout (M-α9β1^−/−) mice. Immunohistochemical analysis of CD11b (monocyte marker) and CD31 (endothelial marker) revealed significantly fewer CD31⁺CD11b⁺ double-positive cells in M-α9β1^−/− cAF mice compared to WT controls. These results indicate that phospho-PLCγ2⁺ monocytes induce endothelial dysfunction through the SPP1-α9β1 signaling pathway in cAF mice.

Obesity is a well-established risk factor for cardiovascular diseases⁶³. Studies in mice have shown that HFD feeding transiently reduces peripheral blood monocyte counts, with a marked decline in Ly6C^lo monocytes and a biphasic response in Ly6C^hi monocytes (initial decrease followed by rebound)^64,65. In HFD-induced obese mice, elevated Ly6C^hi inflammatory monocytes in circulation correlate with adipose tissue macrophage infiltration⁶⁶. In our cAF mouse model, HFD and Ang II were used to induce obesity and hypertension, respectively^17,34. While both Ly6C^hi and Ly6C^lo monocyte proportions increased in peripheral blood, only Ly6C^lo monocytes were significantly elevated in atrial tissue of cAF mice, suggesting tissue-specific recruitment mechanisms. Phenotypic analysis precisely defines the atrial phospho-PLCγ2⁺ cell population. The high CCR2 expression across subsets unequivocally identifies them as monocyte-derived, recruited cells. The comparatively lower frequency of F4/80⁺ cells within the Ly6C^lo subset indicates that only a portion of these recruited cells have acquired a classic tissue-resident macrophage marker during our observation window. This pattern highlights that our model primarily captures the critical phase of monocyte recruitment, a dominant CCR2-driven process, rather than the function of resident macrophages. Our genetic targeting strategy is thus ideally suited to dissect the role of this recruited monocyte-derived continuum in AF remodeling.

Gut microbiota and their metabolites are pivotal regulators of cardiovascular disease pathogenesis⁶⁷. Through metabolic reprogramming and immune modulation⁶⁸, gut microbiota dynamically adapt to dietary changes and intestinal environmental shifts, functioning as a “metabolic filter” that converts nutrients into bioactive compounds such as trimethylamine N-oxide (TMAO), short-chain fatty acids (SCFAs), and secondary bile acids⁶⁹. These metabolites are critically involved in cardiovascular disease initiation and progression. Gut microbiota contribute to inflammation, cardiac fibrosis, and myocardial remodeling during AF by modulating metabolic pathways⁷⁰. Dysbiosis of gut microbiota has been identified in AF patients⁷¹. Yu et al.⁷² demonstrated that low-dose trimethylamine N-oxide (TMAO) treatment enhances atrial autonomic ganglion activity, upregulates structural remodeling factors (e.g., c-Fos), and increases serum levels of pro-inflammatory cytokines (e.g., IL-6 and TNF-α). Obesity is closely associated with gut microbiota alterations. Ley et al.⁷³ reported an increased Firmicutes/Bacteroidetes ratio in obese individuals, characterized by elevated Firmicutes and reduced Bacteroidetes abundance. Turnbaugh et al.⁷⁴ further demonstrated that fecal microbiota transplantation (FMT) from obese rats to germ-free mice recapitulates obesity-related metabolic phenotypes, whereas FMT from lean mice does not induce such changes. To investigate the role of gut microbiota in phospho-PLCγ2⁺ monocyte dynamics during AF, we depleted gut microbiota using antibiotics in cAF mice. Antibiotic-treated cAF mice exhibited significantly reduced AF inducibility and shorter AF duration compared to controls. Immunofluorescence analysis revealed suppressed expression of EndMT markers (SMA and CD31) in the atrial endocardium of antibiotic-treated cAF mice. Flow cytometry showed significant reductions in CD45⁺CD11b⁺Ly6G^-Ly6C^hi phospho-PLCγ2⁺ and CD45⁺CD11b⁺Ly6G⁻Ly6C^lophospho-PLCγ2⁺ monocytes in the atrium of antibiotics-treated cAF mice compared to controls. These findings indicate that gut microbiota prime the recruitment and colonization of phospho-PLCγ2⁺ monocytes in the atrium, driving AF progression.

Previous studies have detected bacterial DNA within atherosclerotic plaques, with the identified bacterial taxa matching those found in the gut microbiota of the same individuals⁷⁵. Bacterial MVs, which mimic the regulatory functions of gut bacteria in metabolism and immune responses⁷⁶, were isolated from fecal samples of cAF mice to investigate their impact on phospho-PLCγ2⁺ monocytes during AF. DiO-labeled bacterial MVs were internalized by THP-1 cells within 12 h, activating the PLCG2-DAG-PKC signaling pathway. Syk, a non-receptor tyrosine kinase (NRTK), serves as a critical signaling adaptor in immune cell activation⁷⁷. In monocytes, Syk and PLCγ2 are activated downstream of Fcγ receptor crosslinking, regulating inflammatory cytokine secretion and phagocytosis through downstream transcription factors³⁵. Additionally, Toll-like receptor 2 (TLR2) binding to Gram-positive bacterial peptidoglycan and TLR4 binding to Gram-negative bacterial lipopolysaccharide both induce PLCγ2 phosphorylation, promoting cytokine secretion⁷⁸. To investigate Syk’s role in gut microbiota-mediated PLCγ2 activation, THP-1 cells were treated with 100 μg/mL bacterial MVs ± the Syk inhibitor R406. R406 treatment suppressed the PLCG2-DAG-PKC signaling pathway in bacterial MV-treated THP-1 cells. To further validate Syk’s role in AF, monocyte-specific Syk knockout (M-Syk^−/−) mice were generated. M-Syk^−/− cAF mice exhibited significantly reduced AF inducibility and shorter AF duration compared to WT cAF mice, demonstrating that bacterial MVs prime Syk signaling to promote AF. A limitation of our study is the exclusive use of male mice. We chose this approach to minimize the confounding effects of the estrous cycle on monocyte function, thus reducing experimental variability. Despite we used a broad-spectrum antibiotic cocktail to deplete gut microbiota, we cannot completely exclude the possibility that the observed effects on PLCγ2⁺ monocytes and atrial fibrillation susceptibility are partially attributable to pleiotropic effects of the antibiotics themselves, rather than solely to microbiota depletion. Future studies employing germ-free mice or FMT will be essential to definitively establish a causal role for specific microbial taxa in modulating the PLCγ2 pathway and AF risk.

In summary, scRNA-seq suggested upregulated phospho-PLCγ2 expression in monocytes from the atrium of AF patients. In a cAF mouse model, phospho-PLCγ2⁺ monocytes were recruited to the atrium, accompanied by activation of the DAG-PKC signaling pathway. Further analysis demonstrated that phospho-PLCγ2⁺ monocytes promote EndMT in the atrial endocardium through the SPP1-α9β1 signaling axis, ultimately driving AF pathogenesis. Monocyte-specific knockout of PLCG2 suppressed the recruitment and differentiation of Ly6C^lo monocytes in the atrium of cAF mice. Antibiotic treatment alleviated AF by inhibiting the bacterial MV-initiated Syk/PLCG2 signaling pathway in monocytes. These findings establish a pathway in which phospho-PLCγ2⁺ monocytes infiltrate atrial tissue and drive endocardial EndMT via the SPP1-α9β1 axis, promoting AF progression.

Methods

Clinical samples

Human left atrial appendage samples from patients with persistent AF (n = 15) and individuals with sinus rhythm (SR, n = 15) undergoing cardiac surgery were collected from the First Affiliated Hospital of Bengbu Medical University between August 2020 and October 2023. The diagnosis of AF was conducted based on the 2016 European Society of Cardiology Guidelines³⁴, and patients with hypertension in AF patients were included. Exclusion criteria included hematological system diseases, thyroid diseases or any other immune diseases. The normal controls were individuals with SR and without any family history of AF or conditions meeting the exclusion criteria. The clinical information of participants was provided in Supplementary Data 1. Atrial samples were stored at −80 °C until analysis. All participants provided written informed consent prior to participation. The study protocol conformed to the ethical guidelines of the Declaration of Helsinki and was approved by the Institutional Ethics Committees of Clinical Ethics of First Affiliated Hospital of Bengbu Medical University (No. 2019KY023).

Animals

Six- to ten-week-old male C57BL/6 mice were used in the study. Monocyte-specific knockout of PLCG2 (M-PLCG2^−/−), integrin α9β1 (Itga9, M-α9β^−/−, and Syk (M-Syk^−/−) was achieved using the Cre-loxP system. The Lyz2-Cre (LysM-Cre) transgenic mouse line (B6.129P2-Lyz2^tm1(cre)Ifo/J) was used as the Cre driver. All floxed mice were on a C57BL/6 background. PLCG2^flox/flox mice were described previously⁷⁹. To generate monocyte-specific PLCG2-deficient mice (LysM-Cre: PLCG2^flox/flox, M-PLCG2^−/−), PLCG2^flox/flox mice, in which exons 3–5 are flanked by loxP sites, were crossed with LysM-Cre transgenic mice (The Jackson Laboratory). To generate monocyte-specific integrin α9β1-deficient mice (LysM-Cre: Itga9^flox/flox, M-α9β1^−/−), Itga9^flox/flox mice^80,81, in which exon 8 is flanked by loxP sites, were crossed with LysM-Cre transgenic mice (The Jackson Laboratory). To generate monocyte-specific Syk-deficient mice (LysM-Cre:Syk^flox/flox, M-Syk^−/−)⁸², Syk^flox/flox mice (Jackson Laboratory) mice, in which exon 1 is flanked by loxP sites, were crossed with LysM-Cre transgenic mice (Jackson Laboratory). To generate monocyte-specific knockout mice, heterozygous Lyz2-Cre mice were crossed with homozygous floxed mice for the target gene (PLCG2, Itga9, or Syk). Offspring obtained from mating PLCG2^flox/flox, Itga9^flox/flox, or Syk^flox/flox with Lyz2-Cre mice were genotyped by PCR analysis of tail DNA. The primer sequences were listed in the Supplementary Table 1. All mice were grouped and housed under specific pathogen-free conditions with a 12/12 h light/dark cycle, an ambient temperature of 22–24 °C, and relative humidity of 40–60%, with free access to food and water. All experimental procedures were performed in accordance with the guidelines of Institutional Animal Care and Use Committee of Zhejiang University. Mice were monitored daily throughout the experimental period. Predefined humane endpoints included rapid body weight loss (>15–20% of initial weight), anorexia≥24 h, impaired mobility, or severe organ system dysfunction (respiratory, circulatory, gastrointestinal, urogenital, musculoskeletal, or neurological). For atrial tissue harvest, animals were euthanized by 20% CO₂ inhalation followed by cervical dislocation.

Induction of a combined AF (cAF) model

A cAF mice model was established by treatment of high-fat diet (HFD) for 4 months, treatment of angiotensin (Ang) II for 1 month, and induction of AF^17,34. This model mimics clinical AF risk factors, including hypertension and obesity. Briefly, mice were fed an HFD for 4 months to induce obesity. Subsequently, Ang II (2.0 mg/kg/day) or saline was subcutaneously infused via osmotic minipumps for 1 month³⁴. Blood pressure was measured using a tail-cuff system (Softron, Tokyo, Japan). Finally, mice were anesthetized with 1% pentobarbital sodium and AF was induced by intracardiac pacing using a Millar 1.1F octapolar electrophysiology catheter (Scisense, USA) inserted into the right atrium and ventricle. Provocative testing consisted of a programmed electrical stimulation protocol followed by burst pacing. The programmed stimulation involved delivering double extrastimuli (S1-S2-S3) at S1 cycle lengths of 120 and 100 ms, with progressive shortening of the S2 and S3 intervals down to 10 ms. Subsequently, rapid burst pacing was administered at cycle lengths gradually accelerated from 50 ms to 10 ms, in bursts lasting 3 and 6 s. AF inducibility was defined as the induction of at least one AF episode lasting ≥1 s during provocative pacing. The total AF duration was defined as the cumulative duration of all AF episodes lasting ≥250 ms recorded in each study.

To determine the status of monocytes in cAF mice, atrial and blood samples were collected from two groups: sham control group (n = 5) and cAF group (n = 5). To investigate the role of PLCG2 in cAF mice, atrial and blood samples were obtained from four groups: sham+WT group (n = 6), sham+M-PLCG2^−/− group (n = 7), cAF+WT group (n = 7), cAF+M-PLCG2^−/− group (n = 7). To investigate the role of integrin α9β1 in EndMT of cAF mice, cAF mice were divided into two groups: cAF+WT group (n = 5), cAF+M-α9β1^−/− group (n = 5). To evaluate the role of gut microbiota in monocyte differentiation and colonization in cAF mice, atrial and blood samples were obtained from four groups: sham+water group (n = 7), sham+antibiotics group (n = 7), cAF+water group (n = 7), cAF+antibiotics group (n = 7). To investigate the role of Syk in cAF mice, cAF mice were divided into two groups: cAF+WT group (n = 6), cAF+M-Syk^−/− group (n = 6). To deplete gut microbiota, the cAF mice were orally administered a combined antibiotic cocktail (1 g/L ampicillin, 1 g/L neomycin and 0.5 g/L vancomycin) in drinking water for 2 weeks⁸³. Control mice received drinking water. Baseline characteristics of experimental animals were listed in the Supplementary Data 2.

Histology and immunohistochemistry

Atrial tissues were fixed in 4% paraformaldehyde (PFA; Beyotime, Shanghai, China). Tissue sections were blocked with 5% goat serum for 30 min and incubated with primary antibodies at 4 °C, including anti-phospho-PLCγ2 (Tyr759, Clone PA5-105761, 1:200; Invitrogen, MA, USA), anti-CD14 (1:200; Invitrogen, MA, USA), anti-CD31 (1:100; Abcam, Cambridge, UK), anti-SMA (1:200; Abcam, Cambridge, UK), anti-TWIST (1:200; Invitrogen, MA, USA) and anti-CD11b (1:100; Abcam, Cambridge, UK). Next, tissues were incubated with Alexa Fluor 488- or 647- conjugated secondary antibodies. Nuclei were counterstained with 4,6-diamidino-2-phenylindole (DAPI; Beyotime, Shanghai, China). Images were captured using a Zeiss fluorescence microscope (ZEISS, Oberkochen, Germany). Quantification of SMA and TWIST expression was performed in five random fields using ImageJ software.

Isolation of monocytes

Monocytes were isolated using the EasySepTM Mouse Monocyte Isolation Kit (Stemcell Technologies, Vancouver, Canada)⁸⁴. Briefly, single-cell suspensions were prepared from blood and atrial tissues by immunomagnetic negative selection. Blood samples were lysed by incubation with ammonium chloride solution on ice for 15 min. After discarding the supernatant, cells were resuspended at a concentration of 1 × 10⁸ nucleated cells/mL. For monocyte isolation from atrial tissue, the atrium was dissected and processed to generate single-cell suspensions^17,85. Tissues from mice were minced and enzymatically digested using a solution containing 450 μ/ml of collagenase I, 125 μ/ml of collagenase XI, 60 μ/ml of DNase I, and 60 μ/ml of hyaluronidase (Sigma-Aldrich, St. Louis, MO) at 37 °C for 30 min with gentle agitation. The cell suspension was passed through a 70-μm strainer to obtain single-cell suspensions. In the EasySep™ negative selection procedure, unwanted cells are labeled with antibody complexes and magnetic particles. For each 1 mL of sample, cells were incubated with cocktail (100 μL of Component A + 100 μL of Component B). The unwanted cells (T cells, B cells, NK cells, granulocytes, etc.) are targeted for removal. To assess the purity of the isolated monocytes, cells were analyzed by flow cytometry using the following antibody panel: anti-CD45 (Clone 30-F11), anti-CD11b (Clone M1/70), anti-Ly6G (Clone 1A8), and anti-Ly6C (Clone AL-21). For subsequent Western blot analysis, monocytes isolated from the pooled atrial tissues of 5–7 mice were lysed in RIPA buffer. The total protein concentration was determined by BCA assay, yielding approximately 5–10 μg of protein per mouse.

Isolation and culture of primary mouse atrial endothelial cells

Mouse atrial endothelial cells were isolated⁸⁶, and single-cell suspensions were prepared from atrial tissue using the Neonatal Heart Dissociation Kit (Miltenyi Biotec). The cell suspension was filtered through a 70-μm strainer and centrifuged. Endothelial cells were positively selected using CD31 MicroBeads (Miltenyi Biotec) according to the manufacturer’s instructions. Purified endothelial cells were cultured in Endothelial Cell Growth Medium (PromoCell) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37 °C in a 5% CO₂ incubator. Cells at passages 2–4 were used for experiments.

Monocyte-endothelial cell co-culture

For co-culture experiments, endothelial cells were seeded in 6-well plates and grown to 80–90% confluence. Transfected monocytes were then added to the endothelial cell culture at a ratio of 5:1 (monocytes: endothelial cells) in fresh medium. The cells were co-cultured for 72 h. After co-culture, monocytes were carefully removed by gentle washing, and endothelial cells were lysed for subsequent Western blot.

Single-cell RNA sequencing (scRNA-seq)

scRNA-seq was performed on left atrial appendage tissues obtained from two AF patients and two SR controls. Individual patient characteristics for scRNA-seq have been provided in the Supplementary Data 3. scRNA-seq libraries were constructed using the DNBelab Single-Cell Kit (BGI, Shenzhen, China) according to the manufacturer’s instructions. Sequencing was performed on the DNBelab C4 platform. Briefly, single-cell suspensions were processed for droplet generation, library preparation, bead collection, reverse transcription, and cDNA amplification to generate barcoded libraries. Indexed scRNA-seq libraries were constructed following the manufacturer’s protocol. The sequencing libraries were quantified using the Qubit^TM ssDNA Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA). Gene expression data were normalized and analyzed using the Seurat package. Principal component analysis (PCA) was performed, followed by dimensionality reduction using uniform manifold approximation and projection (UMAP). Cell clustering was performed using the “FindClusters” function, and the results were visualized in two dimensions using UMAP. Marker genes for each cluster were identified using Seurat’s “FindAllMarkers” function. Differentially expressed genes (DEGs) between cell subpopulations were identified using Seurat’s ‘FindMarkers’ function with the Wilcoxon rank-sum test under default parameters.

Flow cytometry

Single-cell suspensions were prepared from atrial tissues and blood (as detailed in the ‘Isolation of monocytes’ section)^17,84. Single-cell suspensions were obtained by passing the digested tissue through a 70 μm nylon mesh strainer. Erythrocytes were lysed using ammonium chloride buffer (BD Biosciences, Mountain View, CA, USA), and the remaining cells were blocked with an Fc Receptor Blocker (ABACE Biology, Beijing, China). The following antibodies were used to determine the percentage of monocytes and PLCG2⁺ monocytes: CD45 (Clone 30-F11), CD11b (Clone M1/70), Ly6G (Clone 1A8), Ly6C (Clone AL-21), CCR2 (Clone 475301), F4/80 (Clone T45-2342) (BD Biosciences, Mountain View, CA, USA), and phospho-PLCγ2 (Tyr759, Clone 4NPRN4) (eBioscience, San Diego, CA, USA). Cell suspensions were centrifuged to remove the supernatant and resuspended in PBS. A live/dead dye (Invitrogen, Waltham, MA, USA) was included to exclude non-viable cells. Surface staining was performed in the dark at 4 °C for 30 min. After centrifugation (5 min), cells were fixed, permeabilized, and incubated with the phospho-PLCγ2 antibody. Cells were washed, and flow cytometry was performed using a FACS Aria II instrument (BD Biosciences, San Jose, CA, USA). Data were analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA).

Plasmid transfection and shRNA knockdown

For PLCG2 overexpression, monocytes were transfected with 2 μg of mouse PLCG2 overexpression plasmid (OriGene) or empty vector control using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. After 48 h, transfection efficiency was validated by Western blotting. For SPP1 and PLCG2 knockdown, monocytes were transfected with 50 nM of SPP1 or PLCG2-specific shRNA or scrambled control shRNA (GenePharma) using Lipofectamine RNAiMAX (Invitrogen). The sequence for SPP1 shRNA (shSPP1) and PLCG2 shRNA (shPLCG2) was listed in Supplementary Table 2. Knockdown efficiency was confirmed by Western blotting after 48 h.

Western blotting

Cells were collected and lysed to extract proteins, which were separated by 10% SDS-PAGE and transferred onto polyvinylidene difluoride (PVDF) membranes (Invitrogen, Waltham, MA, USA). Membranes were blocked with 5% skimmed milk in TBST for 1 h at room temperature. Primary antibodies, including anti-CaM (Santa Cruz Biotechnology, Dallas, TX, USA), anti-phospho-PLCγ2 (Tyr759; Invitrogen, Waltham, MA, USA), anti-PKCδ, anti-p-Syk, anti-Syk, anti-PLCγ2, anti-SMA, anti-TWIST, anti-Snail, anti-Slug, VE-cadherin (Cell Signaling Technology, Danvers, MA, USA), and anti-α9, anti-Col1a1, anti-SPP1, anti-β-actin (Abcam, Cambridge, UK), were incubated with the membranes at 4 °C overnight. After washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (Cell Signaling Technology, Danvers, MA, USA) for 1 h at room temperature. Protein bands were visualized using a Bio-Rad imaging system (Bio-Rad, Hercules, CA, USA).

16S rRNA profiling of gut microbiota

Feces were collected from antibiotic-treated cAF mice. 16S rRNA sequencing libraries were constructed at BGI (Shenzhen, China). Paired-end reads were merged, and quality control was performed to remove low-quality sequences and chimeras, resulting in high-quality clean data. The Divisive Amplicon Denoising Algorithm (DADA2) was used to improve data accuracy and taxonomic resolution by removing sequences with high similarity and chimeras. Amplicon sequence variants (ASVs) were generated to replace traditional operational taxonomic units (OTUs), and a final ASV feature table and representative sequences were obtained. Following classification, the data were converted into OTU/ASV for further analysis.

Quantitative real-time PCR (qPCR)

Total RNA was extracted from monocytes using TRIzol reagent (Invitrogen, Waltham, MA, USA). One microgram of RNA was reverse-transcribed into cDNA using a Reverse Transcription Kit (Qiagen, Hilden, Germany). Bacterial DNA was extracted from fecal samples of antibiotic-treated and/or cAF mice. DNA concentration was quantified using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). qPCR was performed using an ABI PRISM 7500 real-time PCR system with the miScript SYBR Green PCR Kit (Qiagen, Hilden, Germany). Standard curves were generated to calculate bacterial concentrations in each sample. The qPCR cycling conditions were as follows: 50 °C for 2 min, 95 °C for 2 min, 40 cycles of 95 °C for 15 s, 60 °C for 15 s and 72 °C for 1 min. Primer sequences for genes and bacterial-specific primers were listed in the Supplementary Table 3.

Intracellular Ca²⁺ detection

Monocytes were cultured in 24-well plates. Intracellular Ca²⁺ levels were measured using the Fluo-3 AM fluorescent probe (Thermo Fisher Scientific, Waltham, MA, USA). Monocytes were incubated with 5 μM Fluo-3 AM in the dark for 45 min at room temperature. Fluorescence images were acquired using an Olympus fluorescence microscope (Olympus, Tokyo, Japan). To measure calcium release from intracellular stores independently of extracellular influx, shPLCG2 or shCtrl-treated THP-1 cells were perfused with a calcium-free HBSS supplemented with 0.5 mM EGTA. After recording a stable baseline for 30 s, 1 µM thapsigargin (TG) was applied to inhibit the endoplasmic reticulum Ca²⁺-ATPase pump and induce passive leak from the ER. To evaluate SOCE, 10 µM CPA was applied to deplete ER stores after baseline recording. The extracellular solution was rapidly switched to an HBSS containing 10 mM CaCl₂ to permit calcium influx. THP-1 cells were overexpressing PLCG2 and pre-treated with the specific Orai1 channel inhibitor GSK-7975A (10 µM) for 30 min prior to the start of imaging, and the inhibitor was maintained throughout the experiment. Fluorescence intensity was quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Enzyme-linked immunosorbent assay (ELISA)

The levels of DAG and IP3 in monocytes were measured using commercial ELISA kits for DAG (ELK Biotechnology, Wuhan, China) and IP3 (Abcam, Cambridge, UK). Samples were added to 96-well plates and incubated with biotin-conjugated antibodies specific to DAG or IP3. After washing, avidin-conjugated horseradish peroxidase (HRP) was added to each well and incubated. The reaction was developed using TMB substrate solution and terminated by adding sulfuric acid. Absorbance was measured at 450 nm using a microplate reader (ELX800, BioTek, Winooski, VT, USA). DAG and IP3 concentrations were determined by comparing sample optical density (OD) values to a standard curve.

Migration of monocytes

For migration assays, 1 × 10⁵ monocytes were seeded in FluoroBlok cell culture inserts (Corning, Corning, NY, USA). Monocyte chemoattractant protein-1 (MCP-1; 5 ng/mL, BioLegend, San Diego, CA, USA) was added to the lower chamber, and cells were incubated for 90 min at 37 °C. Migrated monocytes on the lower side of the membrane were fixed with 4% PFA (Beyotime, Shanghai, China) and stained with DAPI (Beyotime, Shanghai, China). Fluorescence images were acquired from five random fields using a Zeiss fluorescence microscope (ZEISS, Oberkochen, Germany). Migrated cell numbers were quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Immunofluorescence

For immunostaining, 5 × 10³ monocytes were seeded on coverslips overnight. Cells were permeabilized with 0.1% Triton X-100 (Biosharp, Hefei, Anhui, China) and fixed with 4% PFA for 20 min at 4 °C. After blocking with 5% goat serum (Beyotime, Shanghai, China) for 1 h at 37 °C, cells were incubated with anti-phospho-PLCγ2 (Tyr759) antibodies (1:200; Invitrogen, Waltham, MA, USA) at 4 °C overnight. Cells were then incubated with Alexa Fluor 647-conjugated secondary antibodies (Thermo Fisher Scientific, Waltham, MA, USA) for 1 h at room temperature. Fluorescence images were acquired from five random fields using a Zeiss fluorescence microscope. Cell numbers were quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Masson’s trichrome staining

Atrial tissue samples were fixed in 10% neutral buffered formalin for 24 h at room temperature, followed by routine dehydration, paraffin embedding, and sectioning at 4-μm thickness. For collagen deposition analysis, Masson’s trichrome staining was performed according to the manufacturer’s protocol (Sigma-Aldrich). In brief, deparaffinized and rehydrated sections were sequentially stained in Weigert’s iron hematoxylin working solution for 10 min, Biebrich scarlet-acid fuchsin solution for 5 min, and aniline blue solution for 5 min. Sections were then rinsed, differentiated in 1% acetic acid, dehydrated, cleared, and mounted with a resinous medium. The collagen volume fraction (CVF) was quantified using ImageJ software (NIH).

Isolation and identification of membrane vesicles (MVs)

Fecal samples were collected from cAF mice. One gram of feces was resuspended in 10 mL of filtered phosphate-buffered saline (PBS). The suspension was centrifuged at 4000 × g for 10 min at 4 °C, followed by a second centrifugation at 6000 × g for 20 min at 4 °C. The supernatant from each step was collected for subsequent centrifugation. The supernatants were sequentially filtered through 0.45 μm and 0.2 μm filters. Filtered supernatants were ultracentrifuged at 100,000 × g for 60 min at 4 °C. The resulting pellet was resuspended and loaded onto qEV10/70 nm columns (IZON, Oxford, UK) to isolate membrane vesicles (MVs). The diameter of the MVs were analyzed using a nanoparticle tracking analysis system (Malvern Instruments, Malvern, UK). MVs were visualized using transmission electron microscopy (TEM; JEOL, Tokyo, Japan) as previously described⁴⁰.

Uptake of membrane vesicles (MVs) by THP-1 cells

MVs were fluorescently labeled with 10 μM 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO) for 20 min at 37 °C. THP-1 cells were incubated with 100 μg/mL DiO-labeled MVs for 12 h. After washing, cells were fixed with 4% PFA (Beyotime, Shanghai, China) for 20 min at 4 °C and stained with DAPI (Beyotime, Shanghai, China) for 10 min at 37 °C. Fluorescence images were acquired using a Zeiss fluorescence microscope (ZEISS, Oberkochen, Germany). THP-1 cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS). To investigate the Syk/PLCG2 signaling pathway in vitro, THP-1 cells were treated with 100 μg/mL MVs and/or the Syk inhibitor R406 (1 μM or 10 μM) for 48 h.

Fecal microbiota transplantation (FMT) and recovery studies

cAF mice received a combined antibiotic cocktail (1 g/L ampicillin, 1 g/L neomycin and 0.5 g/L vancomycin) delivered in drinking water for 2 weeks. Fresh fecal pellets were collected from cAF donor mice, homogenized in sterile anaerobic PBS (100 mg/mL), and centrifuged briefly to remove large particulate matter. Recipient mice, following the 2-week antibiotic pretreatment, received 200 µL of the clarified fecal supernatant via oral gavage every other day for a total of 3 administrations over 1 week. Control antibiotic-treated mice received sterile PBS gavage. For the recovery study after antibiotic withdrawal: A separate cohort of cAF mice received the 4-week antibiotic regimen. Following cessation of treatment (Day 0), mice were returned to regular drinking water. Fecal samples and AF inducibility assessments were performed at baseline (Day 0) and on Days 3, 7, 14, and 28 post-antibiotic withdrawal.

Statistic analysis

Data were analyzed using GraphPad Prism 10.0 (GraphPad Software, San Diego, CA, USA). For comparisons between two groups, Student’s t-test was used. For comparisons involving more than two groups, one-way ANOVA followed by Tukey’s test was applied. A p-value < 0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

P.K., D.W., L.M., B.T., and W.L. conceived the research, designed the study. P.K., D.W., P.T., J.M., B.T., W.D., Y.Z., Q.C., Y.T., Z.C., M.X., Z.S., W.W., and H.W. performed experiments; P.K., D.W., and P.T. analyzed data; P.K., D.W., and L.M. interpreted data and wrote the manuscript.

Peer review

Peer review information

Nature Communications thanks Takuo Emoto, Robert Koeth and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the grants from the National Natural Science Foundation of China (82470415), Anhui Province Outstanding Youth Project (2022AH030141), Anhui Province excellent scientific research and innovation team (2022AH010083), the Natural Science Research Project of Anhui Educational Committee (202304295107020087), and the Clinical Translation Project of Anhui Province (202427b10020068).

Data availability

The data for scRNA-seq have been deposited in the Gene Expression Omnibus (GEO) database under accession code GSE322768. The 16S rRNA sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1509956. All data in this study are provided in the supplementary data and source data. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77219-3.