Effects of Phenylacetylglutamine on the Susceptibility of Atrial Fibrillation in Overpressure-Induced HF Mice

Abstract

Phenylacetylglutamine (PAGln), a gut metabolite is substantially elevated in heart failure (HF). The increase of PAGln in plasma is associated with atrial fibrillation (AF), and contributes to AF pathogenesis. However, the role of PAGln in AF with HF remains uncertain. Therefore, this study aimed to determine the effect of PAGln on AF after HF. Thoracic aortic coarctation (TAC) created overpressure-induced HF mice for 4 weeks. Histopathology, biochemical, echocardiographic for assessment of cardiac function, and electrophysiological examination of several electrophysiological indexes (ERP, SNRT, and the occurrence rate of AF) were performed at the end of the HF mice model. We found that plasma PAGln levels were significantly elevated in PAGln-treated HF mice and that PAGln aggravated maladaptive structural remodeling and electrical remodeling, which aggravated the vulnerability of AF, shortened the ERP duration, prolonged the SNRT, increased the occurrence rate of AF in HF mice. Mechanistically, PAGln exacerbated ROS accumulation and increased the levels of phosphorylated PLB and CAMK II. Overall, PAGln played a vital role in promoting the occurrence of AF in HF mice by activating the CAMK II signaling pathway.

Keywords:

• Phenylacetylglutamine
• CAMK II
• atrial fibrillation
• heart failure

Introduction

Atrial fibrillation (AF) is the most common arrhythmia in patients with heart failure (HF), and its prevalence tends to increase with the severity of HF, reaching rates of up to 50%.¹ Patients who have both HF and AF generally have worse prognoses than those with normal sinus rhythms.² The coexistence of AF and HF can exacerbate each other through various mechanisms, including cardiac remodeling and impairing left ventricular function due to rapid heart rates.^3,4 Thus, it is imperative to investigate the risk factors and underlying mechanisms that lead to AF after HF, and identifying treatments to mitigate these effects could have significant clinical benefits.

HF is a significant health concern, and identifying treatments that can mitigate its effects is crucial. The gut microbiota, which is a rich microbial community residing in the gastrointestinal tract that is often referred to as the “second genome,” plays a pivotal role in maintaining normal physiological functions and contributing to various diseases.^5,6 However, the precise mechanisms through which imbalances in the gut microbial community lead to diseases remain largely unclear. Recent research has highlighted the role of gut microbiota-derived metabolites, including phenylacetylglutamine (PAGln),⁷ short-chain fatty acids,⁸ bile acids (BAs),⁹ catechin, and trimethylamine N-oxide (TMAO),¹⁰ in the development of cardiovascular diseases (CVDs). These metabolites are promising targets for preventing and treating CVDs.^11,12 In particular, PAGln has garnered increased amounts of attention as a key player in CVDs. Studies have demonstrated that PAGln is associated with CVD and is linked to major adverse cardiovascular events, including myocardial infarction, stroke, and death.¹³ PAGln can influence CVD-related phenotypes and thrombosis by interacting with adrenergic receptors such as α2A, α2B and β2.¹⁴ In addition, recent investigations have suggested that the PAGln has prognostic value for HF.¹⁵ Furthermore, increased PAGln levels in the blood have been associated with atrial AF and contribute to its pathogenesis.¹⁶ Despite these findings, the precise mechanisms by which PAGln affects AF after HF remain unclear. Determining the role of PAGln in HF is essential for developing preventive and therapeutic strategies for AF following HF. Moreover, identifying the molecular players involved in gut microbiome-related diseases could pave the way for effective treatments for various lifestyle-related conditions, including obesity, diabetes, and CVDs. Therefore, comprehensive research on the impact of PAGln on HF is crucial for providing further evidence for the prevention and treatment of AF after HF.

Results

PAGln exacerbated cardiac dysfunction and enlargement of the left atrium in overpressure-induced HF mice

To explore the effect of PAGln on AF after HF, we conducted animal studies, and schematic diagram of the experimental groupings and procedure was shown in Figure 1A. The Clostridium sporogenes is one of the most important gut microbes that produces PAA (phenylacetic acid), which is followed by host generation of PAGln. So, we measured the relative abundance of Clostridium sporogenes in the stool between the Sham and HF group. Consistent with the previous study, we found that the relative abundance of Clostridium sporogenes was significantly increased in HF mice (Figure 1B). Plasma PAGln levels were measured by mass spectrometry as described previously. The concentration of PAGln was significantly increased in HF group than in Sham group, and PAGln intervention further increased the plasma concentration of PAGln (Figure 1C). After 4 weeks of PAGln intervention, we measured cardiac function and left atrial diameter (LAD) by echocardiography to confirm the correlation between PAGln and atrial function in HF mice. The LAD was significantly more enlarged in the HF group than in the Sham group, and PAGln intervention further aggravated the enlargement of the LAD (Figure 1D; Table 1). The left ventricular ejection fraction (LVEF) was significantly decreased in HF group than in Sham group, and PAGln intervention further aggravated the decrease of the LVEF (Figure 1E). We also measured plasma NT-proBNP concentrations and the mRNA levels of ANP and BNP in atrial tissue to further confirm the effect of PAGln on atrial function. PAGln intervention aggravated the increase of the plasma NT-proBNP concentrations in HF mice (Figure 1F) and aggravated the increase in the mRNA expression of ANP and BNP in HF mice atrial tissue (Figure 1G).

Figure 1. Plasma PAGln concentrations are significantly increased in overpressure-induced HF mice and, PAGln aggravates atrial inflammation infiltration in overpressure-induced HF mice. (A) Schematic diagram of the experimental groupings and procedure. (B) The relative quantity of Clostridium sporogenes in stool in the Sham and HF groups (n = 8). (C) Plasma levels of PAGln in were significantly increased in the four groups (n = 8). (D) The left atrial diameter (LAD) was significantly enlarged in overpressure-induced HF mice, and PAGln intervention further promoted the enlargement of LAD (n = 8). (E) The left ventricular ejection fraction (LVEF) was significantly decreased in overpressure-induced HF mice, and the intervention with PAGln exacerbated cardiac dysfunction (n = 8). (F) The concentration of NT-proBNP in plasma is significantly increased in overpressure-induced HF mice, and PAGln intervention further promotes the increase of NT-proBNP (n = 8). (G) The mRNA expression of ANP and BNP was significantly increased in the atrial tissue for overpressure-induced HF mice, and PAGln intervention further promoted the expression of ANP and BNP (n = 8). (H) The images of HE stain for left atrium in four groups mice, arrow indicated atrial inflammation infiltration in overpressure-induced HF mice (I) The mRNA expression of the inflammation marker, IL-1β, IL-6, TNF-α was significantly increased in overpressure-induced HF mice, and PAGln intervention further promoted the expression of inflammation marker (n = 8). Data are expressed as mean ± SD, one-way ANOVA with Welch correction (unequal variance). *P < 0.05; ns, not statistically significant.

Table 1 The data of cardiac function indexes were measured by transthoracic echocardiography

Group	Sham (n = 8)	Sham + PAGln (n = 8)	HF (n = 8)	HF + PAGln (n = 8)
LA (mm)	1.62 ± 0.041	1.61 ± 0.053	1.96 ± 0.081*	2.16 ± 0.064*^,**
LVEDV (cm^3)	0.076 ± 0.033	0.078 ± 0.041	0.116 ± 0.034*	0.128 ± 0.032*
LVESV (cm^3)	0.014 ± 0.007	0.015 ± 0.008	0.051 ± 0.017*	0.055 ± 0.019*
EF (%)	82.26 ± 6.175	79.21 ± 4.149	58.79 ± 2.750*	51.71 ± 4.221*^,**
FS (%)	44.73 ± 7.804	45.24 ± 6.308	28.18 ± 3.294*	24.34 ± 3.596*

* P < 0.05 vs Sham group.

** P < 0.05 vs HF group.

PAGln exacerbated structural remodeling of the atrium in poverpressure-induced HF mice

To evaluate the effect of PAGln to atrial structural remodeling, we performed pathological staining for atrial tissue, such as H&E staining, Masson’s trichrome staining, and wheat germ agglutinin (WGA) staining. H&E staining showed that PAGln intervention aggravated disordered atrial myocyte arrange and the increase of inflammatory cell infiltration (Figure 1H). We also measured mRNA expression of IL-1β, IL-6 and TNF-α in atrial tissue (Figure 1I), and the results were consistent with H&E staining. The mRNA expression of IL-1β, IL-6, TNF-α was significantly increased in overpressure-induced HF mice, and PAGln further promoted the expression of inflammation marker.

Furthermore, Masson’s trichrome staining revealed that the proportion of atrial fibrosis was larger in the HF + PAGln group than in the HF group (Figure 2A). We then quantified the atrial fibrosis areas (Figure 2B) and measured the mRNA expression level of CTGF, Col1A1 and Col3A1 in atrial tissue (Figure 2C). The results showed that PAGln aggravated atrial fibrosis and the increase in the expression of fibrosis markers (CTGF, Col1A1 and Col3A1) in atrial tissue. Cardiomyocyte hypertrophy is also an important pathological change in overpressure-induced HF mice. WGA staining revealed that atrial cardiomyocyte hypertrophy was exacerbated in the HF + PAGln group than in the HF group (Figure 2D). We then quantified the atrial cardiomyocyte area, and the results showed that PAGln aggravated the increase in myocyte area in HF mice, but not in Sham mice (Figure 2E).

Figure 2. PAGln intervention aggravate atrial cardiomyocyte fibrosis and cardiomyocyte hypertrophy. (A) The images of Masson’s trichrome staining of the left atrium in four groups mice. (B) Statistical analysis of myocardial fibrosis area in four groups mice (n = 8). (C) the mRNA expression of CTGF, Collagen I, Collagen III was significantly increased in overpressure-induced HF mice, and PAGln intervention further promoted the expression of fibrosis marker (n = 8). (D) The images of wheat germ agglutinin (WGA) stain in the atrial myocyte area, and (E) the statistical analysis of the atrial myocyte area in overpressure-induced HF mice (n = 8). Data are expressed as mean ± SD, one-way ANOVA with Welch correction (unequal variance). *P < 0.05; ns, not statistically significant.

PAGln exacerbated electrical remodeling of the atrium in overpressure-induced HF mice

After exploring the effect of PAGln on atrial structural remodeling, we hypothesized that PAGln could affect atrial electrophysiological properties. Surface ECG (lead II) recordings were performed on mice under light anesthesia. P waves, PR intervals, QRS durations and RR intervals were not significantly different among the four groups (Figure 3A and B).

Figure 3. PAGln intervention aggravates the susceptibility to atrial fibrillation (AF). (A) The electrocardiograph images of the four groups of mice, and (B) the statistical analysis of the P wave duration, PR, QRS and RR in the four groups of mice (n = 8). (C) The images of sinus node recovery time (SNRT) in the four groups of mice, and (D) the statistical analysis of SNRT in 100, 120, 150 ms pacing period in the four groups of mice (n = 8). (E) The statistical analysis of effective refractory period (ERP) in the four groups of mice (n = 8). (F) The images of AF induced by BURST stimulation in four groups mice. (G) The statistical analysis of ratio of AF induced by BURST stimulation in overpressure-induced HF mice (n = 12). (H) The statistical analysis of AF duration induced by burst stimulation in overpressure-induced HF mice. (n = 12). In B, D, F, and H data are expressed as mean ± SD, one-way ANOVA with Welch correction (unequal variance). In H data are expressed as percentage, Fisher’s exact test. *P < 0.05; ns, not statistically significant.

The electrophysiological parameters, SNRT, ERP and the incidence of AF were detected in Langendorff-perfused hearts. Compared with Sham group, the SNRT was significantly extended in overpressure-induced HF mice. Moreover, PAGln furtherly promoted the extension of the SNRT in overpressure-induced HF mice (Figure 3C and D). Sinus node function in HF mice was adaptively dysfunction, which increased the susceptibility to atrial arrhythmias such as AF. The left atrial ERP was significantly different between mice with HF that were treated with or without PAGln. Compared with Sham group, the left atrial ERP in HF group was significantly shorter, and PAGln further promoted shortening of the ERP in overpressure-induced HF mice (Figure 3E). PAGln aggravated the susceptibility to atrial arrhythmias. Consequently, we conducted an atrial arrhythmia induction experiment by using programmed electrical stimulation. In the Sham group and Sham + PAGln group, AF was induced in only one mouse in each group (1/12, 8.33%). However, the induction of AF in the HF group was 6/12 (50%) and was 10/12 (83.33%) in the HF + PAGln group. The induction rate of AF in the HF group was higher than in the Sham group, and PAGln further promoted increase in the induction rate of AF in HF mice; we also analyzed the AF duration induced by burst pacing, and PAGln promoted the increase of AF duration in overpressure-induced HF mice (Figure 3F–H).

In addition, we isolated the atrial myocytes from four groups mouse, and detected the action potential (AP) by whole-cell patch-clamp. The results showed that APD50 and APD90 was not changed in Sham mouse with or without PAGln, but APD50 and APD90 was significantly extended in overpressure-induced HF mice, and PAGln further extended the APD90 in atrial myocytes (Figure 4). Altogether, this data indicated that PAGln aggravated electrical remodeling of the atrium in pressure-overload induced HF mice.

Figure 4. The intervention with PAGln aggravate the prolongation of AP in atrial myocytes. (A) The images of APs in isolation atrial myocytes from four groups. (B and C) The statistical analysis of APD50, APD90 showed that APD50 and APD90 significantly extended in overpressure-induced HF mice, and PAGln furtherly extended the APD90 in atrial myocytes (n = 8). Data are expressed as mean ± SD, one-way ANOVA with Welch correction (unequal variance). *P < 0.05; ns, not statistically significant.

PAGln aggravated exacerbated the dysfunction in CX40 and calcium related channels in overpressure-induced HF mice

To further explore the effect of PAGln on atrial electrical remodeling, the expression levels of atrial connexin 40 (CX40) and calcium related channels protein (NCX1 and CACNA1C). Immunofluorescence staining for CX40 in atrium showed that the CX40 positive areas was significantly decreased in HF group than in the Sham group, and PAGln further promoted a decrease CX40 positive areas proportion in overpressure-induced HF mice (Figure 5A and B). Similarly, Western blotting of CX40 protein expression showed similar results. PAGln decreased the protein expression of CX40 in the atrium in overpressure-induced HF mice (Figure 5C). The expression of NCX1 and CACNA1C in atrium was also measured. Compared with the Sham group, the expression of CACNA1C was decreased and the expression of NCX1 was increased in the HF group. Moreover, PAGln further decreased the expression of CACNA1C and increased the expression of NCX1 in overpressure-induced HF mice (Figure 5D and E). Taken together, these data indicated that PAGln aggravated the disruption of CX40, CACNA1C and NCX1 expression and promoted the occurrence of AF in overpressure-induced HF mice.

Figure 5. PAGln intervention aggravates the decrease of protein expression of CX40 and calcium channel. (A) The images of CX40 immunofluorescence staining in four groups of mice, and (B) the statistical analysis of CX40 positive areas showed that the expression of CX40 in the four groups of mice (n = 8). (C) The images of CX40 Western blot and the statistical analysis of relative expression of CX40/GAPDH in overpressure-induced HF mice (n = 3). (D) The images of CACNA1C and NCX1 Western blotting, and (E) the statistical analysis of relative expression of CACNA1C and NCX1 in overpressure-induced HF mice (n = 3). Data are expressed as mean ± SD, one-way ANOVA with Welch correction (unequal variance). *P < 0.05; ns, not statistically significant.

PAGln exacerbated the accumulation of ROS generation and the activation of CaMKII signaling pathways in overpressure-induced HF mice atrium

To explore what the mechanism by which PAGln affects AF after reactive oxygen species (ROS) generation in atrium were measured. PAGln exacerbated the increase of the ROS generation in the HF mice atrium (Figure 6A). Compared with the Sham group, LDH, MDA levels were significantly increased, and SOD levels were significantly decreased in the HF group. Moreover, PAGln exacerbated the increase of plasma LDH, MDA levels and the decrease of plasma SOD levels in HF mice (Figure 6B–D). Considering the activation of the CaMKII signaling pathways caused by accumulation of ROS generation and the important role of CaMKII signaling pathways to AF, the CaMKII signaling pathways relative proteins were measured. The expression levels of phosphorylated CaMKII, phosphorylated PLB, phosphorylated RyR2 were significantly increased in HF + PAGln group (Figure 6E–H).

Figure 6. PAGln intervention aggravated the activation of CaMKII-signaling pathway. (A) The images of ROS stain and the analysis of ROS positive intensity in atrial tissue from the four groups of mice (n = 8). (B) The statistical analysis of the plasma of LDH in four groups of mice (n = 8). (C) The statistical analysis of the plasma of MDA in four groups of mice (n = 8). (D) The statistical analysis of the plasma of SOD in four groups of mice (n = 8). (E) The images of CaMKII pathway proteins and RyR2 Western blotting in four groups of mice. (F) The statistical analysis of relative level of p-RyR2, and (G) the statistical analysis of relative level of p-CaMKII, and (H) the statistical analysis of relative level of p-PLB, showed that the relative level of p-RyR2, p-CaMKII and p-PLB in overpressure-induced HF mice (n = 3). Data are expressed as mean ± SD, one-way ANOVA with Welch correction (unequal variance). *P < 0.05; ns, not statistically significant.

KN93 alleviated the susceptibility to AF by inhibiting the activation of CaMKII pathway

To further explore the mechanism of PAGln on AF after HF, we used KN93 in vivo, and the schematic diagram of the experimental groupings and procedure was shown in Figure 7A. The electrophysiological parameters, SNRT, ERP and the incidence of AF were detected through in Langendorff-perfused hearts. Compared with HF group, the SNRT was significantly extended in the HF + PAGln group, while KN93 shortened the SNRT (Figure 7B). the left atrial ERP was significantly shorter in the HF group, and PAGln further promoted the shortening of the ERP in overpressure-induced HF mice, while KN93 prolonged left atrial ERP (Figure 7C). The atrial arrhythmias induction experiment by using programmed electrical stimulation also showed that KN93 could alleviate the increase of induction rate of AF and the increase of AF duration caused by PAGln in overpressure-induced HF mice (Figure 7D–F). In addition, the results of ROS staining showed that KN93 alleviated the increase of ROS generation caused by PAGln in overpressure-induced HF mice (Figure 7G). The CaMKII signaling pathways relative proteins were also measured. The expression levels of phosphorylated CaMKII, phosphorylated PLB, phosphorylated RyR2 were significantly increased in the HF + PAGln group, while KN93 alleviated the increased expression of phosphorylated CaMKII, phosphorylated PLB, phosphorylated RyR2 (Figure 7H–K).

Figure 7. The CaMKII inhibitor KN93 alleviated the susceptibility to AF by inhibiting the activation of CaMKII pathway. (A) Schematic diagram of the experimental groupings and procedure. (B) The statistical analysis of SNRT in 100, 120, 150 ms pacing period in five groups of mice (n = 8). (C) The statistical analysis of effective refractory period (ERP) in five groups of mice (n = 8). (D) The images of AF induced by burst stimulation in five groups of mice. (E) The statistical analysis of ratio of AF induced by burst stimulation in overpressure-induced HF mice (n = 12). (F) The statistical analysis of AF duration induced by burst stimulation in overpressure-induced HF mice (n = 12). (G) The images of ROS staining in atrial tissue from five groups of mice. (H) The images of CaMKII pathway proteins and RyR2 in five groups of mice. (I, J, and K) The statistical analysis of relative level of p-PLB/PLB, p-CaMKII/CaMKII, p-RyR2/RyR2 in five groups of mice (n = 3). Data are expressed as mean ± SD, two-way ANOVA followed by Bonferroni post hoc test. *P < 0.05; ns, not statistically significant.

KN93 alleviated the activation of CaMKII signaling pathway due to PAGln in Ang II-induced HL-1 cells

In vitro experiments, Ang II were used to induce a model of cellular hypertrophy. To further explored the effect of PAGln on the CaMKII signaling pathways, the CaMKII inhibitor, KN93 was used in Ang II-induced HL-1 cells. Consistent with previous findings in HF mice, the expression levels of phosphorylated CaMKII, phosphorylated PLB, phosphorylated RyR2 were significantly increased in the Ang II group, and PAGln further aggravated the increase in the expression levels of phosphorylated CaMKII, phosphorylated PLB, phosphorylated RyR2 in Ang II-induced HL-1. However, KN93 alleviated the increase in the expression levels of phosphorylated CaMKII, phosphorylated PLB, phosphorylated RyR2 caused by PAGln in Ang II-induced HL-1 (Figure 8). In conclusion, PAGln exacerbated the occurrence of AF after HF by aggravating atrial electrical remodeling and structural remodeling, and the mechanism is that PAGln aggravated the activation of CaMKII-signaling pathways to some extent by increasing ROS generation.

Figure 8. The CaMKII inhibitor KN93 alleviated the activation of CaMKII pathway in HL-1 cell treated with PAGln. (A) The images of ROS staining in Ang II-induced HL-1 cells treated with PBS, PAGln, KN93. (B) The images of p-CaMKII/CaMKII, p-PLB/PLB, p-RyR2/RyR2 Western blotting in HL-1 cells. (C) The statistical analysis of relative level of p-RyR2, and (D) the statistical analysis of relative level of p-CaMKII, and (E) the statistical analysis of relative level of p-PLB, showed that PAGln intervention promoted the relative level of p-RyR2, p-CaMKII and p-PLB increase in HL-1 cells, while the CaMKII inhibitor KN93 alleviated the increase of the expression of p-RyR2, p-CaMKII and p-PLB (n = 3). (The dose of reagent respectively: 1 μM Ang II, 1 μM KN93, 100 μM PAGln in vitro experiment.) Data are expressed as mean ± SD, two-way ANOVA followed by Bonferroni post hoc test. *P < 0.05; ns, not statistically significant. Ang II, angiotensin II; PAGln, phenylacetylglutamine; PLB, phospholamban; CaMKII, calcium/calmodulin-dependent protein kinase II; RyR2, ryanodine receptor 2.

Discussion

In this study, we utilized PAGln to intervene the overpressure-induced HF mice to explore the effect of PAGln on AF after HF. The major novel findings of this study are: (1) PAGln exacerbated cardiac dysfunction and enlargement of the left atrium in overpressure-induced HF mice; (2) PAGln exacerbated the atrial fibrosis and inflammation to promote maladaptive structural remodeling; (3) PAGln exacerbated electrical remodeling of the atrium in overpressure-induced HF mice; (4) PAGln mediated inflammation and fibrosis and induced AF at least partially by regulating CaMKII and RyR2 activation. PAGln plays an important role in AF after HF, and it may be a potential predictive marker and novel therapeutic target for the treatment of AF after HF.

AF and HF often coexist, and AF increases the risk of stroke, hospitalization for HF, and death.¹⁷ HF can promote AF through structural, ultrastructural, abnormal intracellular calcium handling, and neuroendocrine processes.^18,19 Notably, AF in HF patients is associated with atrial structural remodeling, including left atrial (LA) enlargement, increased LA pressure, and functional mitral regurgitation. These changes, along with stress induced by structural alterations and vasoconstrictive neurohormonal factors such as NT-pro-BNP and angiotensin II, contribute to atrial fibrosis, which is a structural foundation for AF.²⁰ Recent research has shown that treatments such as angiotensin II-converting enzyme inhibitors and other antifibrotic and anti-inflammatory agents can ameliorate atrial fibrosis in animal models of HF.²¹ Atrial fibrosis is increasingly recognized as a key player in AF pathogenesis and is supported by our previous findings in an obesity-induced AF animal model.^22–25 In the current study, we expanded on these observations and showed that PAGln exacerbates cardiac dysfunction, LA enlargement, atrial tissue fibrosis, and inflammation. Masson staining revealed a significant increase in collagen deposition in the myocardial tissue of PAGln-treated HF mice. In addition, the expression of mRNA associated with fibrosis, including CTGF, Col1A1 and Col3A1, was increased in these mice. PAGln exacerbated atrial inflammation in HF mice, as evidenced by increased inflammatory infiltration between cardiomyocytes and elevated mRNA expression of inflammatory markers such as IL-1β, IL-6 and TNF-α. During the pathophysiological progression of HF, increased pressure in the left atrium leads to atrial fibrosis and may also cause left atrial enlargement, and histopathological examination revealed hypertrophy of atrial myocytes. WGA staining further confirmed that the area of atrial myocytes was significantly increased in HF mice, with or without PAGln treatment. These findings indicate that PAGln exacerbates maladaptive structural remodeling in the atria of HF mice, providing a basis for the occurrence of AF.

A programmed electrical stimulation electrophysiological experiment confirmed that PAGln increased the vulnerability of AF in HF mice. During AF induction through BURST stimulation, PAGln significantly increased the occurrence rate and duration of stimulation-induced AF. In addition, PAGln shortened the atrial effective refractory period (ERP) in HF mice, which was also observed in AF patients and associated with increased AF.²⁶ Studies have reported that patients with persistent AF exhibit LA voltage abnormalities, primarily in areas exposed to high atrial stress.²⁷ In HF patients, remodeling of sinoatrial node (SAN) function is characterized by decreases in the intrinsic heart rate, prolonged corrected SAN recovery time (SNRT), and sinoatrial conduction time.²⁸ SAN dysfunction is a critical contributor to AF pathophysiology.²⁹ Our study revealed a prolonged SNRT in HF mice, which was exacerbated by PAGln. In HF patients, atrial electrical remodeling is attributed to abnormalities in electrical coupling due to dysfunctional connexins (CXs) proteins, such as CX40, and abnormal intracellular calcium handling, resulting in delayed afterdepolarizations (DADs) and increased atrial arrhythmogenic activity. A reduction in CX40 density impairs electrical coupling, leading to slowed conduction and the promotion of reentry.³⁰ Consistent with the findings of previous studies, the present study revealed a decrease in CX40 density in HF mice, and PAGln further exacerbated this reduction. Furthermore, PAGln increased the sodium/calcium exchange (NCX1) protein and decreased the L-type calcium channel protein (CACNA1C) in HF mice. Dysregulated calcium handling plays a crucial role in AF, leading to membrane depolarization and DADs.³¹

Imbalances in gut microbial composition and gut metabolites can lead to the development and recurrence of AF.^32,33 For instance, trimethylamine N-oxide (TMAO), a gut metabolite, can destabilize atrial electrophysiology.³⁴ The gut metabolite, PAGln, which has recently garnered increased attention, is significantly elevated in conditions such as end-stage renal disease, coronary artery disease, and HF. PAGln has also been associated with major adverse cardiovascular events and has prognostic value in HF.^14,15,35 Dodd et al.’s research reported a role for the microbial porA gene in oxidative metabolism of phenylalanine (Phe) to PAA by the commensal Clostridium sporogenes, which promoted synthetic product of PAGln by liver and renal tissues. In this study, we found that C. sporogenes significantly increased in feces from HF mice.³⁶ Plasma PAGln levels are notably increased in AF patients and can predict the risk of AF.¹⁶ Nemet et al.’s study also showed that adverse CVD-related phenotypes observed with PAGln administration at physiological levels were attenuated with the β-blocker carvedilol, which indicated that PAGln might act through β-adrenergic receptors (β-AR) on the surface of cardiomyocytes.¹⁴ Despite the known detrimental effects of PAGln on CVDs, the exact mechanism remains unclear.

In humans and experimental animals, stimulation of β-AR has been indicated to have long-term deleterious effects on the heart, and β-AR activation have been found to be associated with cardiac remodeling and arrhythmias, including cardiac hypertrophy, fibrosis.³⁷ β-AR overactivation also led to ROS accumulation in myocardial tissue.³⁸ PAGln promoted the atrial electrical remodeling in HF mice possibly through exacerbating the activation of atrial β-AR. In addition, emerging evidence suggests that the accumulation of ROS, activation of the enzyme calmodulin kinase II (CaMKII), and alterations in calcium homeostasis medicated by ryanodine receptor 2 (RyR2) are key factors in the development of AF.^39,40 This study showed that PAGln exacerbated ROS accumulation and CaMKII activation in HF mice, leading to increased levels of phosphorylated-CaMKII, phosphorylated-phospholamban (p-PLB), and phosphorylated-RyR2. Furthermore, the specific CaMKII inhibitor KN93 mitigated the excessive activation of CaMKII signaling caused by PAGln, ultimately reducing the susceptibility to AF by alleviating maladaptive atrial structural and electrical remodeling. Taken together, these findings suggest that PAGln exacerbates AF occurrence following HF by inducing maladaptive atrial structural and electrical remodeling by increasing ROS accumulation and CaMKII activation, thereby leading to fibrosis, inflammation and disrupted calcium homeostasis. Thus, plasma PAGln levels may serve as predictive markers of AF in HF patients, and CaMKII inhibitors could be therapeutic targets for preventing AF in HF patients with increased plasma PAGln levels.

In conclusion, PAGln aggravated the development of maladaptive atrial structural and electrical remodeling and promoted the occurrence of AF in HF mice, by exacerbating the ROS accumulation in atrial tissue and activating CaMKII-signaling pathway.

Limitation

Although the findings suggest that PAGln excerbated susceptibility to AF after HF, there are certain limitations. KN92 is a negative control for KN93, and we did not establish a KN92 negative control group in this study, which may not be sufficient to truly clarify the role of CAMK II. However, KN93 mitigated the excessive activation of CAMK II signaling and alleviated maladaptive atrial structural and electrical remodeling. In addition, we have used Ang II to simulate the induced myocardial hypertrophy model in in vitro cell experiments, but we have not consistently used Ang II to induce myocardial hypertrophy in mouse models of heart failure in vivo animal experiments. During cardiac ultrasound examination, the mice were kept naturally awake or under mild anesthesia as much as possible to evaluate cardiac function more accurately. This study may have several limitations in the use of cardiac ultrasound in mice under anesthesia. This needs to be further improved in our follow-up research.

Materials and methods

Experimental animals and cell lines

Animal experiments were performed according to the Guide for the Care and Use of Laboratory Animals published by the US National Institute of Health (NIH publication No. 85-23, revised 2011) and were approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University. (Approval no. 20220301 A) All animals were housed in an environment with controlled light cycles (12 h light/12 h dark), temperature, and humidity, with free access to food and water. Experimental animals were 8-week old C57BL/6N male mice purchased from HFM Bio-Technology Company (Beijing, China), weight 22–25 g. TAC surgery created pressure-overload induced HF mice, as described previously.⁴¹ Briefly, after anesthetizing the mice with 3% pentobarbital sodium at 40 mg/kg, a 27 G needle was used to ligate the thoracic aorta. The sham surgery was only operated through thoracotomy without ligation. Postoperative analgesia (meloxicam, 4 mg/kg/d, intraperitoneal) and anti-infection (penicillin, 10000–20000 U/kg/d, intramuscularly) treatment continued for 3 days. At 24 h after the operation, PAGln (Sigma, USA, SMB00962) treated for consecutive 4 weeks through intraperitoneal injection, dose 100 mg/kg/d. KN93 (MCE, Shanghai, China) was administered for 4 consecutive weeks by intraperitoneal injection (a dose of 1 mg/kg/d). The control group was given an equal amount of saline.

The immortalized mice atrial myocytes HL-1 cell line was obtained from Pricella (Wuhan, China) and cultured in Dulbecco’s Modified Eagle’s Medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco, Grand Island, NY, USA) and penicillin-streptomycin (Gibco, Grand Island, NY, USA) in a humidified atmosphere containing 5% CO2 and 95% air at 37 °C. The stimulation of Ang II (MCE, Shanghai, China, HY-13948) at the concentration of 1 μM induced HL-1 hypertrophy for 24 h. In addition, KN93 (MCE, Shanghai, China, HY-15465), a CaMKII inhibitor, was added to cells at the concentration of 1 μM and PAGln at the concentration of 100 μM intervened HL-1 cell for 24 h. Subsequent experiments were carried out 24 h later.

Relative abundances of Clostridium Sporogenes

After 4 weeks surgery, the intestinal feces were collected from the Sham and HF groups. Approximately 250–500 mg of feces was added to individual tubes of Intestinal Microbial Genomic DNA Isolation Kit (TianGen, Beijing, DP712), and DNA was extracted according to manufacturer’s guidelines. In order to remove residual RNA interference, 10 μL RNase A was added to samples. And then DNA quantitation was performed with spectrophotometry, and A260/A280 range from 1.8 to 2.0. To amplify genomic DNA (gDNA), we used the SYBR Green qPCR Master Mix assay kit (Servicebio, Wuhan, G3328). Conditions were: 96-well Q-PCR plates, 20 μL reactions, primers at 500 nM each, 10 ng DNA per reaction. For fecal DNA samples, the relative quantity of C. Sporogenes was calculated according to the cycle number at the threshold crossing point using the standard curves.

Measurement of plasma PAGln and NT-proBNP

Blood samples were obtained and centrifuged for serum, which was used for subsequent measurement. The concentration of PAGln in serum samples was quantitatively determined by liquid chromatography-tandem mass spectrometry, as described previously.⁴² Plasma NT-proBNP levels were measured using an enzyme-linked immunosorbent assay kit (Elabscience, China, EL-M0834), following the manufacturer’s protocol.

Determination of ROS generation in atrium and measurement of plasma LDH, MDA, and SOD

Freezing heart slices were prepared and ROS stain kit (Sigma, Shanghai, D7008) was used to determinate total cellular ROS levels. Finally, the images of ROS stain were taken using fluorescence microscopy, and the experimental results were analyzed using ImageJ. And, blood samples were obtained by removing eyeballs and centrifuged for serum. The plasma of LDH, MDA and SOD were measured using spectrophotometry, following the manufacturer’s protocol (Servicebio, China, GM1120, GM1134, GM1133).

Echocardiography

Mice were anesthetized with 2.5% isoflurane, then transthoracic echocardiography (TTE) Vevo 2100 high-resolution imaging system (Visual Sonics), with 22–55 hMHz transducer was used to evaluate the cardiac function. Left ventricular ejection fraction (LVEF) were calculated as described previously.⁴¹ Left atrial (LA) anteroposterior diameter were measured using 2-dimensional imaging during three consecutive cardiac cycles.

Histopathology

The hearts were excised and washed with saline solution, then isolated atrium were fixed in 4% paraformaldehyde for 24 h, then embedded in paraffin and sectioned (5-µm thick). Hematoxylin and eosin (H&E) staining was used for histopathology. Masson staining was used to evaluated interstitial fibrosis and the percentage of fibrosis was calculated using ImageJ. FITC-conjugated wheat germ agglutinin (WGA) stained for membranes was used to determine the cross-sectional area of the myocytes. A single myocyte was measured using a quantitative digital image analysis system (Image-Pro Plus 6.0).

CX40 immunofluorescence

The atrial tissue microsphere sections were dewaxed, and were perforated with four per thousand Triton solution for 30 min. Then, blocking with 5% BSA for 60 min, incubating overnight at 4 °C in primary antibody solution (CX40, 1:200, Abcam, ab313644). After washing off the primary antibody with PBS, the atrial tissue microsphere was incubated in the dark for 2 h with the corresponding secondary antibody (Alexa Fluor 488, 1:1000, Invitrogen, A-11094). The secondary antibody was washed off with PBS, counterstained with DAPI. The fluorescence intensity of the target proteins was detected using fluorescent microscope (Leica DM500), and analyzed by ImageJ software.

qRT-PCR

According to the manufacturer’s protocol, total RNA from the atrial tissue was isolated using a TRIzol reagent (Invitrogen, USA, 15596026). The detailed method of qRT-PCR was as described previously.²⁵ The relative changes were normalized to GAPDH mRNA using the 2^−ΔΔCT method. The sequences of the primers for each gene are listed in Table 2.

Table 2 Sequences of primers for qRT-PCR

Primer	Sequence	Access no.
GAPDH-F	5′-ACTCCACTCACGGCAAATTC-3′	NM_008084.3
GAPDH-R	5′-TCTCCATGGTGGTGAAGACA-3′
ANP-F	5′-AAGAACCTGCTAGACCACCTGGAG-3′	NM_008725.3
ANP-R	5′-TGCTTCCTCAGTCTGCTCACTCAG-3′
BNP-F	5′-TGCTGGAGCTGATAAGAGAAAA-3′	NM_008726.6
BNP-R	5′-GAAGGACTCTTTTTGGGTGTTC-3′
Collagen I-F	5′-AGGCTTCAGTGGTTTGGATG-3′	NM_007742.4
Collagen I-R	5′-CACCAACAGCACCATCGTTA-3′
Collagen III-F	5′-AAGGCTGCAAGATGGATGCT-3′	NM_009930.2
Collagen III-R	5′-GTGCTTACGTGGGACAGTCA-3′
CTGF-F	5′-GCTAAGTTCTGCGGGGTGT-3′	NM_010217.2
CTGF-R	5′-CATGGCCTTCCGTGTTCCTA-3′
Csp_qPCR_F	5′-TTGGCTCTGCACCGGGAATC-3′	Clostridium sporogenes
Csp_qPCR_R	5′-CTGCAAACGCCGTCCCTCTT-3′

Surface ECG and Langendorff perfusion

Mice were anesthetized with 3% pentobarbital sodium and surface ECG was recorded using subcutaneous electrodes connected to the PowerLab 8.1 station (AD Instruments), imitating surface ECG II leads for steadily recording 15 min. Langendorff-perfused hearts were prepared following previous methods.⁴¹ Programmed electrical stimulation and monophasic action potentials (MAPs) were then recorded. The paired platinum stimulating electrode is positioned on the right atrial basal surface. Recording electrode, a tungsten electrode with a stainless steel shielding layer is positioned on the left atrial surface to record MAPs. Sinus node recovery time (SNRT) was measured after a 30 s pacing train with a basic paced cycle length (PCL) of 150 ms, a stimulus amplitude of 2-fold stimulation threshold and a stimulus duration of 1 ms.⁴³ The SNRT was defined as the interval between the last stimulus and the onset of the first sinus return beat. S1–S2 pacing was used to determine the effective refractory period (ERP). The S1–S2 interval was shortened from 150 to 100 ms in 10 ms steps and 50 ms in 2 ms steps. The burst stimulation (50 Hz, 2 ms pulse width, 2 s duration, 5 s intervals, repeat 5 times) was performed to test for AF inducibility. AF was defined as a rapid irregular atrial rhythm with irregular RR intervals lasting at least 1 s. The duration of AF was measured from the end of burst stimulation to the first P wave detected after the atrial rhythm. LabChart 8.0 software was performed for data analyses offline.

Isolation of atrial cardiomyocytes and AP measurement

After 4 weeks Sham and TAC surgery, left atrial myocytes were isolated. Briefly, mice were anesthetized with 3% sodium pentobarbital and heparinized (100 U intraperitoneally). The hearts were removed rapidly and retrograde perfused with Ca²⁺-free Tyrode’s solution (mM): 126 NaCl, 5.4 KCl, 1.0 MgCl₂, 0.3 Na₂HPO₄, 10 D-Glucose, 10 HEPES, pH 7.4 through the aorta at 37 °C for 5 min, then, perfused with the same solution containing collagenase type II (0.6 mg/mL, Sigma), protease (0.1 mg/mL, Sigma) and 0.1% bovine serum albumin for another 5–10 min. the left atrial appendage was dissected and stored in Kraftbruhe (KB) buffer (mM): 85 KOH, 50 K-glutamate, 30 KCl, 20 taurine, 1.0 MgCl₂, 10 HEPES, 10 D-glucose, 0.5 EGTA, pH 7.4. Single myocytes were obtained by gently triturating the left atrial appendage and were stored in KB buffer at 4 °C until use. For recording action potential (AP), single left atrial myocytes were superfused with standard Tyrode’s solution. The pipettes with tip resistance of 3–6 MΩ were filled with solution containing (mM): 110 K-aspartate, 20 KCl, 8.0 NaCl, 1.0 MgCl₂, 1.0 CaCl₂, 4.0 Mg-ATP, 0.1 EGTA, 10 HEPES, pH 7.2 with KOH. AP was performed using the EPC-9 amplifier (List Instruments, Germany). Data were analyzed with the Pulse-fit software interface (Version 8.31, HEKA Co. Germany), and were measured at 50% (APD50) and 90% (APD90) repolarization.

Western blotting

Atrial tissue and HL-1 cell extracted total protein, and protein concentrations were determined by the BCA (Servicebio, China, G2026). Western blotting determined the proteins expression level of CX40 (1:1000, Abcam, ab313644), NCX1 (1:1000, ABclonal, A5583), CACNA1C (1:1000, Invitrogen, MA5-27717), Phospholamban (PLB, 1:1000, Abcam, ab219626), p-PLB-Ser16 (1:1000, Abcam, ab15000), CaMKII (1:1000, Cell Signaling Technology, 4436), p-CaMKII-Thr286 (1:1000, Cell Signaling Technology, 12716), RyR2 (1:1000, Invitrogen, PA5-104444), p-RyR2-Ser2808 (1:1000, Invitrogen, PA5-105712) as described previously.²⁵ The specific band was visualized by enhanced chemiluminescence detection and analyzed using ImageJ software.

Statistics

All the experimental results were analyzed using SPSS (IBM, Armonk, NY) or GraphPad Prism 8.0 software (GraphPad Software, United States). Data are expressed as the mean ± SD or percentage. The unpaired Student’s t test (two-tailed) was used for statistical analysis of two groups. For data with normal distribution, when there were more than two groups, one-way ANOVA (equal variances) or one-way ANOVA with Welch correction (unequal variance) was employed; when analyzing two independent variables, two-way ANOVA was employed. For data with non-normal distribution, the nonparametric statistical analysis was used with Wilcoxon (Mann–Whitney U) test. Categorical data were analyzed using Fisher’s exact test. P < 0.05 was considered statistically significant.

Authors’ contributions

Hui Fu and Wei Shuai contributed equally to draft and review the manuscript. Bin Kong, Jun Zhu, Xi Wang, Yanhong Tang participated in data collection, analysis and interpretation. He Huang and Congxin Huang designed the study and supervised the project. All authors contributed to the article and approved the submitted version.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data that support the findings of this study, including the full set of images of raw immunoblot data and staining presented, is deposited in Mendeley Data, V1, doi: 10.17632/m35k5c9ftb.1.

Additional information

Funding

This work was supported by grants from the Technical Innovation Project of Hu Bei Province of China (Grant No. 2016ACA153) and the Fundamental Research Funds for the Central Universities (Grant No. 2042021kf0119).