Experimental animals

Male Sprague–Dawley rats weighing 300–350 g were procured from the Laboratory Animal Center at Guangzhou University of Chinese Medicine (GZUCM). The rats were housed under controlled light cycles (12 h of light, 12 h of dark) and maintained at a temperature of 23 ± 2 °C. Standard rat chow and water were provided ad libitum. Animal care and experimental protocols were approved by the Institutional Animal Care and Use Committee of GZUCM (Protocol Number: 20201124005).

Administration of acetylcholine (ACh) and CaCl2

The AF animal model was induced by the intravenous injection of Acetylcholine (ACh) and CaCl2, as previously described [24]. Briefly, CaCl2 (Sigma, Saint Louis, USA) was dissolved in saline, and ACh (Sigma, Saint Louis, USA) was added under protection from light. The AF rat model was established by daily intravenous injection of 1 mL/kg of the mixture (ACh, 66 μg/mL, and CaCl2, 10 mg/mL) for 10 consecutive days. Prior to injection, rats were anesthetized with 2% isoflurane and secured on an animal heating pad with adhesive tape on their abdomen.

EA treatment

Rats in the electroacupuncture (EA) group underwent EA treatment at the left ‘‘Xinshu’’ (BL15, 5 mm lateral to the spine at T5 level) or ‘‘Shenshu’’ (BL23, located 5 mm lateral to the posterior midline, on the level of the lower border of the spinous process of the 2nd lumbar vertebra) [38]. EA treatment was performed for 20 min per day for 10 consecutive days at 2 mA and 2 Hz, starting 10 min after ACh-CaCl2 administration, as previously described [39]. To account for potential effects of anesthesia or body fixation, rats in all groups were anesthetized with isoflurane (2%) and secured on the animal heating pad using adhesive tape in a standardize position on their abdomen. One needle (0.25 mm × 13 mm, Beijing Hanyi Medical Instruments Co Ltd, Beijing, China) was inserted at a 45-degree angle to a depth of 4–5 mm into the left BL15 or ‘‘Shenshu’’ (BL23), with the other needle inserted 5 mm away. The two needles were connected to the positive and negative poles of a stimulator (Master-8, AMPI, Israel), respectively. For the sham EA treatment, stainless steel needles were inserted 2–3 mm into the same points as the EA group, but without electrical current injection.

Preparation of isolated hearts

All rat hearts were isolated and perfused with the Langendorff system (Model: SGL, scope research institute of electrophysiology, China) as previously described [40]. After euthanization with an intraperitoneal injection of 2% pentobarbital sodium (0.3 ml/100g), the chest was opened, and the hearts were rapidly excised. The hearts were perfused with oxygenated (95% O2, 5% CO2) Tyrode's solution through the aorta until they beat rhythmically. Subsequently, the blood was flushed out, and the hearts were placed in the Langendorff system under continuous perfusion with oxygenated Tyrode's solution at 38 °C. The Tyrode's solution composition consisted of 134 mM NaCl, 4.5 mM KCl, 0.5 mM MgCl2, 2 mM NaH2PO4, 23 mM NaHCO3, 1.8 mM CaCl2, and 5.5 mM glucose, equilibrated with 95% O2 and 5% CO2 to maintain a pH of 7.4.

AF inducibility and ERP analysis

The electrocardiogram (ECG) signal from the isolated heart was recorded using an electrophysiological mapping system (Mapping Lab, EMS64-USB-1003, UK) [41]. In brief, a stimulation electrode was inserted into the left atrium of the heart, while two ECG recording electrodes were placed in the right atrium and left ventricle, respectively. The pacing current threshold was determined by a train stimulation, increasing the intensity from 1 to 3 mA with a duration of 2 ms. The minimum current required for successful pacing of the heart was considered the systole threshold for current intensity. The experimental stimulus current was then set at 2 times the systole threshold current amplitude. To assess AF inducibility, S1−S1 stimulations were applied with two stimulation steps at the frequency of 50 Hz. The same stimulation was repeated eight times to calculate AF inducibility and duration. To determine the effective refractory period (ERP) in all groups, a trained stimulation 1 (S1, 5 Hz) was repeated ten times to pace the heartbeat, which was considered to be the basic stimulation. The interval between S1−S1 was set to 100 ms. Stepwise decreasing electrical stimulation 2 (S2, 5 Hz) was applied, repeating S1−S2 ten times, with the interval of S1−S2 set to 60 ms and a reduction of 5 ms per cycle until S2 failed to induce a complete heartbeat signal trace. The last S1−S2 interval time was recorded as the ERP.

In vivo ultrasound imaging and assessment

Vascular wall and hemodynamic function in the carotid and femoral arteries were studied using a Vevo 2100 Ultrasound Image system (Fujifilm Visualsonics, Canada), with an ultrasonic coverage of 22–55 MHz and a central frequency of 40 MHz [42]. Measurements were randomly obtained and labeled by one operator. Rats were anesthetized with 4% isoflurane in a chamber with 1 L/min of medical oxygen for 3–5 min, after which the isoflurane concentration was adjusted to 2% to maintain anesthesia during measurements. Hair on the legs and cervix of rats was removed using depilatory cream in the supine position. The probe was placed parallel to the rat's neck to locate the left and right carotid arteries. M-mode images were used to measure the diameters of the right and left common carotid arteries during relaxation and contraction, as well as the mean blood velocity (MV) of these arteries. Blood flow and diameter differences during relaxation and contraction were calculated as follows: blood flow (mL/min) = MV × π (vessel diameter/2)2; diameter difference (mm) = maximal diameter during vessel relaxation—minimum diameter during vessel contraction. Next, the probe was moved to the leg bone to locate the femoral artery and obtain measurements, including the diameters of the right and left femoral arteries during relaxation and contraction, and the MV of these arteries. The blood flow and diameter differences during relaxation and contraction of the femoral arteries were calculated using the same formulas as above.

Laser Doppler measurement

In vivo microcirculation of the lower limbs and kidneys was assessed using Laser Doppler flowmetry (PERIMED PeriFlux5000 system, Sweden), as previously described [43]. The rats were anesthetized with 2% isoflurane and placed on a heating pad at 37 °C to maintain their body temperature. The plantar surfaces of the lower limbs were positioned on a dark surgical towel after gentle removal of hair using hair removal cream. Tissue perfusions were quantified in regions of interest (ROI) within the upper 200–300 μm of the skin, and the mean Perfusion Unit (PU) was recorded. The same procedure was applied for PU assessment in the kidneys. The rats were placed in a dissecting pan with their ventral side exposed, and a cut was made from the middle of the body to the hindlimbs to expose the kidneys. The kidney surface was gently moistened with saline during the experiment. The laser scan head was adjusted to achieve the best image quality, and a constant distance was maintained for all measurements.

Enzyme-linked immunosorbent assay (ELISA)

Following the administration of the ACh and CaCl2 mixture and the 10 day EA treatment, all rats were anesthetized with an intraperitoneal injection of 2% pentobarbital sodium (0.3 mL/100 g) and dissected after the disappearance of the foot withdrawal reflex. Blood samples (3 ml) were collected from each rat through intracardiac puncture. D-Dimer and fibrinogen (FG) levels were analyzed by ELISA (Elabscience Biotechnology Co., Ltd. Shanghai, China) according to the manufacturer's instructions. Blood samples were transferred into empty tubes and centrifuged at 3000 g for 10 min. The liquid supernatant serum was collected and stored at − 80 °C for subsequent use.

Myograph

The carotid/femoral arteries were removed and placed in a 4 °C chilled Krebs solution (NaCl, 118 mM; sodium bicarbonate, 25 mM; glucose, 5.6 mM; potassium chloride, 4.7 mM; KH2PO4, 1.2 mM; MgSO4 7H2O, 1.17 mM; and CaCl2·2H2O, 2.5 mM). The residual blood and perivascular fat were carefully removed before the myograph experiments. Vascular endothelial function was evaluated using a myograph (No: DMT620, Danish Myo Technology, Aarhus, Denmark), as previously described [44]. The isometric force was recorded using a PowerLab/8SP data acquisition system (Software Chart, Version 5, AD Instrument, Colorado Springs, CO, USA). The isolated carotid/femoral arteries were cut into 2 mm rings and placed in 5 mL baths of myograph, equilibrated for 1 h in Krebs solution (37 °C) under a resting tension of 5.0 mN, and bubbled with 95% O2 and 5% CO2. Subsequently, phenylephrine (PE: 10–6 M, code: P1250000, Sigma, Saint Louis, USA) was used to precontract the vascular rings. Once the vessel contraction reached a plateau, acetylcholine (ACh: 10–9 to 10–5 M, endothelium-dependent vasodilator, No: A2661, Sigma, Saint Louis, USA) or the NO donor sodium nitroprusside (SNP: 10–9 to 10–5 M, endothelium-independent vasodilator, No: 71778, Sigma, Saint Louis, USA) was added to the chamber to produce a cumulative concentration–response vasorelaxation curve.

Electrocardiogram In vivo

The electrocardiogram (ECG) was conducted following the procedures described previously [45]. Briefly, the last injection of the mixed solution (ACh and CaCl2) was administered through the caudal veins of the rats. After a 5 min interval, the rats were anesthetized with 2% isoflurane and positioned in a supine position on an animal heating pad set at 37 ℃. Needle-like electrodes of a multichannel physiological recorder (PowerLab 16/35, AD Instruments Pty, Australia) were placed on both the right upper limb and one left lower limb (standard II lead). The rats were recorded once their respiration stabilized, and their heart rate reached 350 ± 25 beats/min. The ECG signals (5 mV, 1 k/s sampling rate) were analyzed using LabChart 8.0 software (AD Instruments Pty, Australia). Heart rate variability (HRV) analysis was conducted following established protocols [46, 47]. For frequency-domain analysis, the absolute power of different frequency bands commonly used in rats, including low frequency (LF) and high frequency (HF), was determined. The HF (0.75–2.5 Hz) peak is generally indicative of cardiac vagus nerve activity, while the LF (0.2–0.75 Hz) is often associated with a dominant sympathetic component.

Autonomic nerve activity recording

The cardiac sympathetic and vagus nerve activity were recorded following established procedures [48]. In brief, the rats were placed in a supine position on a warm heating pad and anesthetized with 4% isoflurane for induction and 2% for maintenance. An incision was made at the neck midline to expose the left carotid artery, and the trachea was separated with a glass needle. This allowed access to the cervical sympathetic nerve medial to the carotid and the vagus nerve lateral to the carotid. The epineurium was carefully peeled after fixing the sympathetic and vagus nerves with a glass needle. A platinum-iridium alloy wire cuff electrode (Kedou Brain Computer Technology, China) was used to hook onto the cardiac sympathetic or cervical vagus nerve, with a referred electrode inserted subcutaneously into the ipsilateral neck. The signal of nerve activity was recorded using a recording system (Bio-Signal Technologies, USA) and analyzed with Neuroexplorer 5.0 software (Plexon, USA). The signal was band passed at 100–1000 Hz.

Virus constructs and chemogenomic manipulation

For the chemogenetic manipulation experiments of paraventricular nucleus of the hypothalamus (PVN) neurons, rats were anesthetized with 2% isoflurane and placed in a surgical stereotactic apparatus (RWD Instruments, China). The skull was leveled using bregma and lambda landmarks, and injections were made into the PVN at an angle of 6°: ± 1.2 mm from the bregma, 2.05–2.25 mm lateral from the midline, and 7.95 mm ventral to the skull. Adeno-associated virus (AAV) expressing hM3D (Gq) (pAAV-SYN-HA-hM3D (Gq)2A-mCherry-3FLAG, Viral titer: 1.22E + 13) obtained from Obio Technology, Shanghai, was used. Bilateral viral injections of 100 nL virus were administered with two 1 μL Hamilton Syringes connected to a micro pump (RWD Instrument, China). The rat body temperature was maintained with a heating pad set at 36℃. To allow for viral expression in the PVN, clozapine N-oxide (CNO, APExBIO) intraperitoneal injection was performed 60 min before EA treatment at 3 weeks after the initial viral injection. Artery imaging baseline was recorded before CNO administration and after virus injection. Rats were injected with 1 mg/kg CNO for hM3Dq activation, while control rats were injected with the same amount of vehicle. Experimenters were blinded to viral identity at the time of viral injection and to rat identity until after the experiments were completed.

Immunofluorescence

Rats were transcardially perfused with 0.9% NaCl followed by 4% paraformaldehyde in 0.1 M PB (pH 7.4; 4 °C) after deep anesthesia with an intraperitoneal injection of 2% pentobarbital sodium (0.3 ml/100 g). The brains were immediately removed and postfixed with 4% paraformaldehyde overnight. The tissues were then dehydrated in 15% and 30% sucrose in 0.1 M PBS at 4 °C for 48 h. OCT-embedded blocks were sectioned to a thickness of 40 μm. Sections from each group were rinsed in 0.01 M PBS three times and blocked for 2 h with blocking liquid (5% goat serum and 0.2% Triton-100 in 0.01 M PBS) at room temperature. The sections were then incubated with rabbit anti-c-Fos antibody (1:1000, mAb#2250, CST) in blocking buffer overnight at 4 °C. The following day, the free-floating sections were washed with 0.01 M PBS three times and incubated with the secondary antibody Alexa Fluor 488 goat anti-rabbit (1:500; ab150077, Abcam) at room temperature for 2 h. DAPI was applied for nucleus staining for 5 min. After being washed in 0.01 M PBS, the brain slices were cover-slipped. Images were captured using a microscope (Axio Imager.A2, ZEISS, Germany), and analysis was performed using NIH Image J software (Bethesda, MD, USA).

Statistical analyses

Statistical analyses were conducted using SPSS 21.0 (Chicago, IL, USA) and GraphPad Prism version 5.0 (San Diego, CA, USA). All data were expressed as mean ± SEM. Statistical differences were determined using analysis of variance (one-way or two-way ANOVA), and Bonferroni multiple comparison test was used for paired comparisons between groups when the data met a normal distribution. For data that did not meet a normal distribution, Dennett’s tests were employed for pairwise comparison between groups. Experimental values of relaxation were calculated relative to the maximal changes in the contraction produced by PE, respectively, taken as 100% in each tissue. Curves were fitted to all data using nonlinear regression, and the half-maximum response (pEC50) of each drug, expressed as -log molar (M), was used for comparison of potency. The number of experiments is indicated by “n,” and p value < 0.05 was considered statistically significant.