Mitochondrial damage mediates STING activation driving obesity-mediated atrial fibrillation

Abstract

Aims

Obesity is a significant risk factor for atrial fibrillation (AF), but the mechanisms by which obesity contributes to AF are not fully understood. Recent studies have indicated that the Stimulator of Interferon Genes (STING) signalling, mediated by mitochondrial damage, plays a crucial role in cardiac remodelling in various metabolic and cardiovascular diseases. This study aims to explore the role of STING in obesity-mediated AF and its potential mechanisms.

Methods and results

In this study, rats were divided into four groups: two groups received tail vein injections of AAV9-cTnT-STING siRNA and were fed either a normal diet or a high-fat diet (HFD) for 12 weeks; the other two groups received injections of AAV9-cTnT-NC siRNA and were similarly fed either a normal diet or a HFD. The atrial STING signalling, AF vulnerability, electrical remodelling, and substrate remodelling were assessed in all groups. Results showed that the induction of AF was increased in obese rats, accompanied by severe mitochondrial damage and upregulation of the STING inflammatory signalling cascade. STING activation was associated with atrial fibrosis, cardiomyocyte apoptosis, and substrate remodelling, including alterations in the gap junction protein CX40 and ion channels. Additionally, STING was linked to excessive calcium transfer from the endoplasmic reticulum to the mitochondria. Knockdown of STING prevented AF vulnerability and both electrical and substrate remodelling in obese rats.

Conclusion

Mitochondrial damage-mediated activation of the STING signalling pathway promotes obesity-induced atrial remodelling and the occurrence of AF.

Graphical Abstract

Graphical Abstract.

Translational perspective.

• In the atria of obese rats, STING expression is up-regulated.

• Cardiac-specific knockdown of STING can alleviate obesity-induced atrial remodelling and reduce the susceptibility to atrial fibrillation.

• STING is associated with abnormal calcium handling and mitochondrial calcium overload in the atria in the context of obesity.

Introduction

Atrial fibrillation (AF) is the most common sustained arrhythmia encountered in clinical practice, with a continuously rising global prevalence, posing a significant public health challenge.¹ Compelling evidence now highlights the importance of obesity in determining the risk of AF.² Although individuals with obesity often have other risk factors for AF, such as heart failure and obstructive sleep apnoea, population studies have identified obesity as a significant and independent risk factor for AF.³ The potential mechanisms by which obesity promotes AF are multifactorial and remain poorly understood. Recent evidence suggests that impaired mitochondria are particularly a substrate for AF in the context of obesity.⁴ Obesity-related increased mitochondrial reactive oxygen species (mtROS), and disrupted mitochondrial calcium homeostasis have been confirmed to impact the development of AF.^5,6 However, importantly, the role of mitochondrial DNA (mtDNA)-mediated sterile inflammatory in AF in the context of obesity remains unclear.

The stimulator of interferon genes (STING) is an endoplasmic reticulum-resident adaptor protein. Recent studies have revealed STING’s pivotal role in sterile inflammation initiated by endogenous cytosolic mtDNA.⁷ In the context of obesity, mtDNA can escape from damaged mitochondria into the cytosol, acting as damage—associated molecular patterns to activate the cytosolic DNA sensor cyclic GMP—AMP synthase (cGAS). This activation prompts STING to be activated on the endoplasmic reticulum and translocate to the Golgi apparatus, driving downstream inflammatory cascades and triggering the transcription of inflammatory components.^8,9 Therefore, STING is recognized as a central hub in the intracellular damage and inflammatory cascade.

It is well known that sterile inflammation is a common pathological feature shared by obesity and AF, and increasing evidence indicates that STING is closely related to cardiovascular and metabolic diseases.^10,11 Moreover, there is evidence that elevated levels of circulating mtDNA are present in the blood of AF patients,^12,13 and oxidative damage to mtDNA has been observed in the atrial muscle of AF patients.¹⁴ Therefore, investigating the role of mtDNA-mediated inflammatory cascades in obesity and other AF-related models is of great interest. In this study, we explored the hypothesis that STING drives obesity-mediated AF and investigated its potential mechanisms.

Methods

A detailed version of the Methods is available in the Supplemental Material.

Ethical statement

This study was approved by the animal research subcommittee of our institutional review board (approval number: WDRM20240106B; date: 27 January 2024) and is in compliance with the guidelines of the National Institutes of Health for the care and use of laboratory animals.

Animals model

Sprague-Dawley rats (250–300 g) were provided by the Animal Experiment Centre of Renmin Hospital of Wuhan University. In this study, the rats were divided into four groups (with 20 rats per group): (i) Control (Con) group: rats received tail vein injection of AAV9-cTnT-NC (negative control) siRNA and were fed a normal diet for three months. (ii) Diet-induced obesity (DIO) group: rats received tail vein injection of AAV9-cTnT-NC siRNA and were fed a high-fat diet (HFD) for three months. (iii) Con + STING knockdown (KD) group: rats received tail vein injection of AAV9-cTnT-STING siRNA and were fed a normal diet for three months. (iv) DIO + STING KD group: rats received tail vein injection of AAV9-cTnT-STING siRNA and were fed a HFD for three months. The AAV9 adeno-associated viral vectors were generated by General BIOL (China). These vectors contained the cTnT cardiac promoter, which enabled specific KD of genes in cardiac tissue. Each rat received a tail vein injection at a dose of 2 × 10¹² vg/kg. The STING siRNA was used to knock down the STING gene, while scrambled siRNA served as the NC. The target sequence for STING gene KD was GCACATTCGGCAAGAAGAA, and the target sequence for the scrambled NC siRNA was TTCTCCGAACGTGTCACGT. The transfection and expression efficiency were detected by western blot.

Cell culture and treatment

HL-1 atrial myocytes were cultured in complete DMEM-F12 medium supplemented with 10% fetal bovine serum and 1% double antibiotics at 37°C in a 5% CO₂ environment. The cells were divided into four groups: (i) Ctrl group: transfected with NC siRNA; (ii) Palmitic acid (PA) group: transfected with NC siRNA and treated with 0.2 mM PA; (iii) Ctrl + STING KD group: transfected with STING siRNA; (iv) PA + STING KD group: transfected with STING siRNA and treated with 0.2 mM PA.

Electrophysiological properties assessment

As previously described,¹⁵ the rats were anesthetized and heparinized, and their hearts were rapidly excised and retrogradely perfused via the aorta with KH solution (in mM: 119 NaCl, 4 KCl, 1.8 CaCl2, 1 MgCl2, 1.2 KH2PO4, 25 NaHCO3, and 10 glucose) at a flow rate of 8 mL/min and a temperature of 37°C. In the isolated Langendorff-perfused hearts, left atrial epicardial activation mapping was performed using a 64-electrode multielectrode array. The images were acquired using EMapScope software (version 5.941) and were analyzed for left atrial conduction velocity and conduction heterogeneity.

Blebbistatin (10.26 µM) was administered through a drug infusion port to eliminate motion artefacts. Subsequently, the voltage-sensitive dye RH-327 (2.01 µM) and the calcium indicator Rhod-2 a.m. (0.89 µM) were introduced for optical mapping. The fluorescence signal of the transmembrane potential (Vm) was long-pass filtered (>700 nm) and captured using a CMOS camera. Digital images (150 × 150 pixels) were collected at a sampling rate of 0.8 kHz from a 2.2 × 2.2 cm field of view. The duration of the action potential at 90% repolarization (APD₉₀₎, the time to peak calcium transient, the duration of calcium transient at 90% recovery (CaTD90), and the calcium amplitude were recorded and analyzed using OMapScope software (version 5.8.2).

The atrial effective refractory period (AERP) was measured using a train of 10 S1 stimuli with a pacing interval of 167 ms, followed by a single premature S2 stimulus with a progressively decreasing interval. The AERP was defined as the longest S1S2 interval that failed to capture atrial activity. AF was induced using a burst pacing protocol. Persistent rapid and fragmented atrial electrograms with an irregular ventricular rhythm for at least 2 s were considered indicative of AF.

Statistical analysis

Data are shown as the mean ± standard deviation from at least three independent experiments. All data were tested for normality using the Shapiro-Wilk test. Comparisons between two groups were performed using the unpaired Student’s t-test. For multiple group comparisons, one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test was conducted. P-value < 0.05 was defined as statistical significance. All statistical analyses were carried out using GraphPad Prism 10.

Result

STING expression is up-regulated in the atrial tissue of obese rats

We induced DIO rats by feeding them a HFD. The body weight of the DIO rats began to differ from that of the Con rats starting from the sixth week. At week 12, the average weight of the Con rats was 411.5 ± 25.5 g, compared with 661.8 ± 30.7 g for the DIO rats (Figure 1A). As compared to the Con group, the DIO group showed significantly increased levels of total cholesterol and triglycerides (Figure 1B), thereby supporting the phenotypic differences between the two groups. Considering that mitochondrial damage and mtDNA leakage into the cytosol are prerequisites for the mitochondrial-related inflammatory cascade, we used transmission electron microscopy to observe the mitochondria in the atrial tissue. The images showed an increased proportion of swollen mitochondria in the DIO group, with reduced relative density of mitochondrial cristae, partial cristae rupture, matrix dissolution, and calcium deposition in the mitochondria. In contrast, the mitochondria in the Con group were more intact, with abundant cristae and fewer damaged mitochondria (Figure 1C and D). Additionally, the DIO group showed a significant increase in cytoplasmic mtDNA in the atria compared with the Con group (Figure 1E), suggesting that a HFD induces mitochondrial damage and mtDNA leakage in the rat atria.

Figure 1.

STING expression is up-regulated in the atrial tissue of obese rats. (A) Changes in body weight of Con and DIO rats over time. (B) Changes in blood lipid levels after 12 weeks in both groups. (C) Representative electron microscopy images of atrial tissue, with red arrows indicating damaged mitochondria. (D) Quantification of mitochondrial damage. (E) Relative cytosolic mtDNA levels. (F) Western blotting showing the expression of cGAS and STING in atrial tissue between the two groups, with bar graphs showing quantification results. (G) Representative immunohistochemical images of atrial cGAS and STING (400X magnification), with bar graphs showing quantification results (H) mtROS levels in Ctrl and PA groups (200X magnification). (I) Representative fluorescence images of HL-1 Cells showing STING, P-Tbk1, P-Irf3 and the merged image (1000X magnification). *P < 0.05; **P < 0.01; ***P < 0.001.

Subsequently, we detected the expression of several common cytoplasmic DNA sensors, such as Z-DNA binding protein 1 (ZBP1), Absent in Melanoma 2 (AIM2), and cGAS in the atrial tissue. Our results showed that the expression of AIM2 and ZBP1 was low in the atria, with no significant difference observed between the DIO and Con groups (see Supplementary material online, Figure S1A). However, obesity significantly increased the expression of cGAS and its key downstream target STING in the atria (Figure 1F). Immunohistochemical analysis confirmed these findings, showing a significant upregulation of cGAS and STING in the atrial tissue of DIO rats compared with the control group (Figure 1G).

In vitro experiments were conducted using PA (the primary saturated fatty acid in high-fat diets) to treat HL-1 atrial myocytes. PA treatment led to mitochondrial dysfunction in HL-1 cells, characterized by increased mtROS and reduced mitochondrial membrane potential (Figure 1H and Supplementary material online, Figure S2A and B). Additionally, PA treatment resulted in the activation of the STING/TANK-binding kinase 1 (TBK1)/interferon regulatory factor 3 (IRF3) signalling cascade (Figure 1I). Overall, these results suggest a potential association between mitochondrial damage-related STING signalling and obesity-mediated atrial remodelling.

STING mediates atrial inflammation and damage associated with obesity

To investigate the role of STING in obesity-related atrial remodelling, we utilized AAV9 adeno-associated virus to create a myocardial STING KD rat model. Western blot analysis confirmed that STING KD reduced the expression of the STING/TBK1/IRF3 inflammatory cascade in the atria (Figure 2A). However, STING KD did not reduce body weight or lipid levels in obese rats (see Supplementary material online, Figure S2C and D). Considering that the ‘NACHT, LRR, and PYD domains-containing protein 3 (NLRP3)’ inflammasome is also an important downstream effector of STING and is associated with the pathogenesis of AF, we assessed the expression of NLRP3 signalling in the atrial tissue of four groups. The results showed no differences in ASC expression among the groups (Figure 2B). However, compared with Con rats, DIO rats exhibited significantly up-regulated expression of NLRP3, pro-caspase-1, pro-IL-1β, cleaved caspase-1, and IL-1β. In contrast, obese rats with STING KD showed significantly reduced expression levels of these proteins (Figure 2B). The immunohistochemistry results were consistent with these findings (Figure 2C). Overall, these results indicate that STING is a potential factor in the activation of NLRP3 in the atria of obese rats.

Figure 2.

STING mediates atrial inflammation and damage associated with obesity. (A) Western blotting showing the protein expression of the STING pathway, with bar graphs displaying the quantified results. (B) Western blotting showing the protein expression of the NLRP3 pathway, with bar graphs displaying the quantified results. (C) Representative immunohistochemical images of atrial NLRP3 (400X magnification), with bar graphs showing quantified results. (D) Representative images of TUNEL staining in the atria of each group (400X magnification). (E) Representative electron microscopy images of atrial tissue (15 000X magnification), with bar graphs showing quantification of mitochondrial damage. *P < 0.05; **P < 0.01; ***P < 0.001.

We further investigated the impact of STING on atrial damage. Compared with the Con group, the DIO group showed a significantly higher proportion of apoptotic cells in the atria (Figure 2D), with electron microscopy revealing disorganized, broken myofibrils and partial dissolution of sarcomeres (Figure 2E). In contrast, STING KD significantly improved obesity-induced apoptosis, with the atrvial myofibrils appearing more orderly. Moreover, STING KD can also alleviate obesity-induced atrial mitochondrial damage. (Figure 2E). In vitro experiments also showed that STING KD mitigated PA-induced HL-1 cells mitochondrial dysfunction (see Supplementary material online, Figure S2A and B).

STING KD reduces obesity induced atrial electrical remodeling

To determine the impact of obesity on atrial electrophysiological properties, we conducted electrophysiological assessments. S1S2 programmed stimulation revealed that compared with the Con group, the AERP was significantly shortened in DIO rats (Figure 3A and B). Under rapid atrial pacing, all DIO rats induced AF, while only 16.6% of Con rats did (Figure 3C and D). Moreover, the duration of AF was significantly prolonged in the DIO group compared with the Con group (Figure 3E). Optical mapping videos of conduction and phase showed that in the DIO group, electrical impulses propagated in a rotor-like manner, with multiple reentrant wavelets present, while conduction was uniform in the Con group (Figure 3F and Supplementary material online, Videos S1–S4). STING KD extended the AERP, reduced AF incidence, and shortened AF duration in obese rats. Additionally, under a predetermined stimulation protocol, we observed atrial action potential alternans only in the DIO group. This phenomenon was not observed in the other groups (see Supplementary material online, Figure S3A). In summary, these results indicate that the atrial electrical activity in obese rats is unstable, with obesity promoting the formation of reentrant substrates and increasing the susceptibility to AF. STING KD significantly reduces obesity-related AF susceptibility.

Figure 3.

STING KD reduces obesity-related AF susceptibility. (A) Representative tracing images of AERP measured by S1S2 stimulation. (B) AERP among groups. (C) Representative image of rapid burst-induced AF. (D) AF induction rate among groups. (E) AF duration among groups. (F) Screenshots of conduction and phase videos under optical mapping. **P < 0.01; ***P < 0.001.

To further elucidate the role of STING in obesity-induced atrial reentrant substrates, we assessed atrial electrical conduction and APD in four groups of rats. Compared with Con, DIO rats exhibited prolonged atrial conduction times, reduced atrial conduction velocities, and increased indices of conduction heterogeneity (Figure 4A). At a stimulation frequency of 6 Hz, APD₉₀ was significantly shorter in DIO rats than in Con rats (Figure 4B). However, STING KD accelerated conduction velocity, reduced conduction heterogeneity, and prolonged APD₉₀ compared with DIO group (Figure 4A and B). These findings suggest that STING KD effectively prevents the formation of obesity-induced reentrant substrates. We further evaluated the expression of major ion channels that influence action potential. Compared with Con rats, DIO rats exhibited increased expression of the ultra-rapid delayed rectifier K⁺ channel (Kv1.5) and decreased expression of the α subunit of the L-type Ca²⁺ channel (Cav1.2) in the atria. STING KD prevented these changes. Additionally, DIO rats exhibited up-regulated expression of voltage-dependent Na⁺ -channel (Nav1.5). However, STING KD did not appear to affect Nav1.5 expression (Figure 4C).

Figure 4.

STING KD prevents obesity-induced reentrant substrate formation. (A) Activation time maps from optical and electrical mapping, with bar graphs showing the quantified results of conduction velocity and conduction heterogeneity. (B) Action Potential Duration at 90% Repolarization among groups, with bar graphs displaying the quantified results. (C) Western blotting showing the protein expression of the Nav1.5, Cav1.2 and Kv1.5 among groups, with bar graphs displaying the quantified results. *P < 0.05; **P < 0.01; ***P < 0.001.

STING KD attenuates obesity induced atrial structural remodeling

Myocardial fibrosis and abnormal connexin proteins can lead to localized slowing of conduction and increased electrical heterogeneity in myocardial tissue, which are significant causes of reentry. Therefore, we assessed the degree of atrial fibrosis and the expression of Connexin 40 (CX40) in four groups of rats. The expression of fibrosis-related proteins (collagen I and transforming growth factor beta 1) was significantly elevated in the DIO group (Figure 5A), Consistent with this, Masson’s trichrome staining revealed a significant increase in fibrotic areas in the atria of DIO rats compared with Con (Figure 5B). STING KD significantly reduced collagen deposition and the expression of fibrosis-related proteins in the atrial tissue of obese rats. Additionally, compared with the Con group, the expression of CX40 in the atria of DIO group was down-regulated, with significant lateralisation in spatial localisation (white arrow in Figure 5C), and these changes could be reversed by STING KD.

Figure 5.

STING KD attenuates obesity-induced atrial structural remodelling. (A) Western blotting showing the protein expression of collagen I, TGF-β and CX40, with bar graphs displaying the quantified results. (B) Masson’s trichrome staining of atrial tissue sections among groups. (C) Immunofluorescence of CX40 and α-actin in atrial tissue sections among groups. (D) Representative echocardiographic images of the left atrium and left ventricle, with bar graphs showing the quantified results of LAD and left ventricular function. *P < 0.05; **P < 0.01; ***P < 0.001.

Atrial enlargement increases the likelihood of conduction block and reentry, and is an independent risk factor for the development and progression of AF. We also assessed changes in cardiac structure. The results showed no statistically significant differences in left atrial diameter (LAD) among the four groups (Figure 5D), but the LAD in the DIO group showed a trend toward enlargement (P = 0.062), suggesting that a longer duration of HFD feeding might lead to atrial enlargement. Heart failure and hypertension are also risk factors for AF and are associated with obesity to some extent. However, in this experiment, we observed that there were no significant differences in left ventricular end-diastolic volume, left ventricular end-systolic volume, and left ventricular ejection fraction among the four groups (Figure 5D). Additionally, there were no significant differences in systolic and diastolic blood pressure among the four groups of rats (see Supplementary material online, Figure S3B). Therefore, in this experiment, the impact of obesity on atrial remodelling and the induction of AF is unlikely to be influenced by ventricular function and blood pressure.

STING is associated with abnormal calcium handling in the atria

Consistent with action potential alternans, under a predetermined pacing protocol, we observed atrial calcium transient alternans only in the DIO rats, suggesting abnormal calcium handling in the DIO rat atria (Figure 6A). Our results showed no statistically significant differences in the time to peak calcium transient and the CaTD₉₀ among the four groups. However, compared with the Con group, the amplitude of atrial calcium transients was significantly reduced in the DIO group (Figure 6B). Although Ryanodine Receptor 2 (RYR2) plays an important role in calcium release from the endoplasmic reticulum, we did not observe differences in RYR2 expression levels among the groups (see Supplementary material online, Figure S3C). Therefore, we speculate that the reduced calcium transient amplitude may be related to dysfunction of L-type calcium channels. We further performed functional tests on the L-type calcium channels. The results showed a significant decrease in the density of L-type calcium current (ICa,L) across all tested potentials from −10 mV to +40 mV (Figure 6C–E), which might directly affect the process of calcium-induced calcium release. STING KD partially improved the reduced ICa,L density and calcium transient amplitude in obese rats.

Figure 6.

STING is associated with abnormal atrial calcium handling. (A) Calcium alternans images recorded in DIO rats during atrial tachycardia. (B) Quantitative results of calcium transient time to peak, CaTD90, and calcium transient amplitude among the groups. (C) Representative current traces of L-type calcium current (ICa,L). Inset: voltage protocol. (D) Average I-V (current-voltage) relationship of ICa,L. (E) Comparison of peak ICa,L at +10 mV for the four groups. *P < 0.05; **P < 0.01; ***P < 0.001.

Mitochondrial calcium regulation is a part of intracellular calcium handling, and the amplitude of atrial calcium transients is also influenced by mitochondrial calcium homeostasis. Excessive calcium transfer from the endoplasmic reticulum to the mitochondria can lead to mitochondrial calcium overload. Given the presence of calcium deposits in the atrial mitochondria of the DIO group, we examined the calcium transfer between the endoplasmic reticulum and mitochondria. The results showed that, compared with the Con group, the distance between the mitochondrial outer membrane and the endoplasmic reticulum was significantly reduced in the DIO group, with an increase in calcium deposit particles (Figure 7A). Additionally, the expression of Inositol 14,5-trisphosphate Receptor Type 1 (IP3R1) and Voltage-Dependent Anion Channel Protein 1 (VDAC1) was significantly up-regulated in the DIO group, while there was no difference in Glucose-Regulated Protein 75 (GRP75) expression (Figure 7B). In vitro experiments showed that, compared with the Ctrl group, PA treatment led to a significant increase in mitochondrial calcium concentration in HL-1 cells (Figure 7C), with a marked increase in the colocalization of VDAC1 and IP3R1 (Figure 7D). These findings support the overactivation of the mitochondrial calcium transfer complex and the manifestation of calcium overload under lipotoxic conditions. In both in vivo and in vitro experiments, STING KD reversed these changes, suggesting that STING is associated with abnormal calcium handling in the obese atria.

Figure 7.

STING is associated with calcium transfer from the endoplasmic reticulum to the mitochondria. (A) Representative electron microscopy images of atrial tissue from each group, with arrows indicating the contact between mitochondria and the endoplasmic reticulum (15 000X magnification), and bar graphs showing the quantified distance between the mitochondrial outer membrane and the endoplasmic reticulum. (B) Western blotting showing the protein expression of IP3R1, GRP75, and VDAC1, with bar graphs displaying the quantified results. (C) Representative fluorescence images of mitochondrial calcium in HL-1 cells from each group, with bar graphs showing the quantified results (400X magnification). (D) Representative images of VDAC1, IP3R1, and the merged image (1000X magnification), with scatter plots indicating the colocalization relationship between the two proteins. Bar graphs show the quantitative analysis of VDAC1 and IP3R1 colocalization (Pearson’s correlation coefficient). *P < 0.05; **P < 0.01; ***P < 0.001.

Discussion

This study investigated the impact of STING on AF in obesity. We provide new evidence for the following: (i) HFD induced obesity promotes the activation of cGAS/STING signalling in rat atria. (ii) STING KD can improve obesity-induced atrial remodelling, prevent reentrant substrate formation, and reduce AF susceptibility. (iii) In the obesity model, STING is associated with abnormal calcium handling in atrial myocytes. Our findings suggest that specific interventions targeting STING signalling can effectively prevent the progression of AF in obesity.

Mitochondria are involved in a variety of important cellular biological processes. Recent studies have revealed the key role of mitochondria in regulating sterile inflammation. Under external stimuli, mtDNA leaks from damaged mitochondria into the cytosol, recognized by cytosolic DNA sensors and triggers inflammatory cascades.⁹

Several cytosolic DNA sensors have been identified in the cytoplasm and have been shown to play roles in various cardiac diseases.¹⁶ Our results showed that cGAS, rather than ZBP1 or AIM2, may play the predominant role in the atria of obese rats. ZBP1 and AIM2 are expressed at very low levels in the atria, even in the obese rats. In contrast, cGAS and its key downstream STING signalling are significantly up-regulated in the atria of obese rats. Therefore, we consider the cGAS/STING signalling pathway as a potential target for obesity-related atrial remodelling.

STING is considered a key protein driving chronic inflammation. Upon activation by cGAS, STING translocates from the endoplasmic reticulum to the Golgi apparatus, where it recruits and phosphorylates TBK1 and IRF3. Phosphorylated IRF3 then migrates to the nucleus, binds to promoters, and initiates the transcription of inflammatory signals, leading to the production of various pro-inflammatory cytokines such as IL-1β, TNF-α, and interferons.^17,18 These inflammatory cytokines can directly affect the abnormal calcium release from the sarcoplasmic reticulum and the expression of connexin proteins in cardiomyocytes, and also promote cardiomyocyte apoptosis and the progression of atrial fibrosis.^19–21 Moreover, in addition to the classical cascade, NLRP3 acts as an effector molecule downstream of STING.^22,23. The role of NLRP3 in atrial inflammatory remodelling has been widely recognized. In diet-induced obese mice, the activation of the NLRP3 inflammasome in the atria promotes atrial arrhythmias.^24,25 Our results indicate that in the atria of obese rats, knocking down STING effectively prevents the activation of TBK1/IRF3 and NLRP3 signalling, suggesting that STING is an important initiator protein in obesity-related atrial inflammation.

In terms of electrophysiological properties, compared with Con rats, obese rats exhibited shortened APD and AERP, along with slowed and heterogeneous atrial electrical conduction. These changes increased the likelihood of electrical signal retention in local atrial regions, which favoured the formation of reentrant circuits, thereby leading to the persistence of AF.²⁶ Indeed, we observed multiple reentrant wavelets in the atria of obese rats during AF, with electrical impulses propagating in a rotor-like manner, supporting the arrhythmogenic pathological features induced by obesity. STING KD effectively prevents the formation of reentrant substrates through multiple aspects. For example, STING KD can prevent obesity-induced abnormal expression of the Cav1.2 and the Kv1.5, which might directly affect the APD. Additionally, STING KD significantly attenuates the progression of atrial fibrosis and the remodelling of CX40.

Abnormal calcium handling is a key mechanism in inflammation related AF.²⁷ Mitochondrial calcium homeostasis is an essential component of intracellular calcium processing. Imbalances in mitochondrial calcium uptake and release can disrupt intracellular calcium cycling and promote arrhythmogenic Ca²⁺ alternans.^28,29 Wu et al. performed an enrichment analysis of sequencing data from atrial samples of obese and normal-weight patients, revealing that genes encoding ‘Ca²⁺ input into mitochondria’ were significantly enriched in obese atrial samples, while the process of ‘Ca²⁺ ion input into the cytosol’ was down-regulated.³⁰ This finding is similar to ours. Our results indicate that in the atrial of obese rats, the contact between mitochondria and the endoplasmic reticulum is more intimate, calcium in the endoplasmic reticulum seems to be more inclined to be released via IP3R1 and transferred to the mitochondria through VDAC1 in the mitochondrial outer membrane. Recent studies have shown that overactivation of the IP3R1-GRP75-VDAC1 calcium transfer complex in the atria of diabetic mice leads to mitochondrial calcium overload, promoting cardiomyocyte apoptosis and atrial remodelling.^31,32 Notably, recent studies have shown that mitochondrial calcium overload mediated by the IP3R1-GRP75-VDAC1 complex promotes the opening of the mitochondrial permeability transition pore, leakage of mtDNA into the cytoplasm, and activation of the cGAS/STING signalling pathway.³³ Interestingly, although STING activation is generally considered a downstream event of mitochondrial calcium overload and injury, our results show that knocking down STING can alleviate obesity-induced mitochondrial calcium overload and mitigate mitochondrial damage. This suggests a presumed feedforward loop between STING activation and mitochondrial injury, which amplifies atrial inflammatory signals and forms a vicious cycle akin to ‘AF begetting AF.’

Despite the clinical application of STING-targeted therapy in the cardiovascular field still being in its early stages, preclinical evidence has provided a feasible framework for its translation. For instance, the nitrofuran derivative C-176 can specifically palmitoylate the Cys91 site of the STING protein, effectively blocking the activation of downstream inflammatory signalling. Evidence suggests that C-176 can inhibit vascular inflammation and macrophage activation in atherosclerotic mice, thereby attenuating the formation of atherosclerosis.³⁴ Additionally, another inhibitor, H-151, has been shown to improve cardiac inflammation, fibrosis, and cardiac function in myocardial infarction mice.³⁵ The dose-response and time-effect relationships of these STING inhibitors in obesity and other AF risk models, as well as the optimal dosing regimens based on different disease stages, need to be further determined in future studies.

Limitation

Some limitations should be acknowledged. The mechanism by which STING regulates calcium handling in atrial myocytes remains unclear. Although inflammatory factors can affect calcium handling in cardiomyocytes,³⁶ STING-mediated calcium signalling regulation is unlikely to be an isolated effect of inflammatory signalling. STING is located on the endoplasmic reticulum, and its unique position may enable it to directly affect calcium release and transport. For example, evidence suggests that in macrophages and monocytes, STING can directly bind to IP3R1, triggering the release of Ca²⁺ from the endoplasmic reticulum, leading to coagulation during sepsis.³⁷ Additionally, in renal cell carcinoma, STING binds to the mitochondrial membrane protein VDAC2 and inhibits the transfer of Ca²⁺ ions from the endoplasmic reticulum to the mitochondria.³⁸ These studies demonstrate that in different cell types and environments, STING can bind to various Ca²⁺ channels to exert unique biological functions, which are not dependent on inflammatory cascades. Therefore, the role of STING in calcium handling in atrial myocytes needs to be confirmed in future studies. Secondly, we selectively targeted STING in the myocardium to avoid potential compensatory changes that might occur following systemic knockout. This is evident in obese mice with systemic STING deficiency, which exhibit reduced body weight and plasma free fatty acid levels.³⁹ However, we did not observe improvements in body weight or lipid profiles in obese rats with myocardial STING KD. Nevertheless, we cannot rule out the potential impact of reduced STING in the ventricles on the atria. Given the potential confounding effects of increased risk of AF following thoracotomy, we chose tail vein injection of AAV9 instead of direct injection into the atria during thoracotomy. Although echocardiography showed no significant differences in ventricular function among the groups, we cannot entirely exclude the potential influence of ventricular changes. Selective gene manipulation in the atria would help further clarify these effects.

Conclusion

In conclusion, our study showed that STING is a critical mediator in the development of AF in the context of obesity. Targeting STING may emerge as a promising therapeutic strategy for the prevention and treatment of obesity-related AF.

Supplementary material

Supplementary material is available at Europace online.

Funding

This work was supported by the National Natural Science Foundation of China (nos. 81970277 and 82170312).

Data availability

The data that support the findings of this study are available from the authors upon reasonable request.