Granulocyte colony-stimulating factor attenuates myocardial remodeling and ventricular arrhythmia susceptibility via the JAK2-STAT3 pathway in a rabbit model of coronary microembolization

Each rabbit was sedated with ketamine and xylazine (25 and 3.75 mg·kg^− 1 im), intubated, and ventilated with supplemental oxygen (2–4%). During the procedure, the rabbits were anesthetized with continuous intravenous infusion of ketamine and xylazine (5 and 4.5 mg·kg^− 1·h^− 1). The right common carotid artery was exposed and punctured, a guidewire was inserted into the ascending aorta under X-ray fluoroscopy, and a catheter was introduced into the left coronary artery for non-selective coronary angiography. For rabbits in the CME, G-CSF, and AG490 groups, autologous thrombus particles (5 mg, prepared using 1 ml of ear vein blood from each rabbit, dried into a blood clot at 37 °C, ground, and sieved with a 38-μm screen mesh to make an autologous microthrombus) were injected via the catheter into the LAD artery to establish the CME model. For rabbits in the Sham group, an equal volume of normal saline was injected. The rabbits in the G-CSF group received subcutaneous injections of G-CSF (10 μg/kg/d) beginning at 2 h after surgery and then daily for 6 days [14]. The rabbits in the AG490 group received an intraperitoneal injection of AG490 (a specific JAK2 inhibitor, 0.25 mg/kg/d) 2 h after surgery and 30 min before each subcutaneous injection of G-CSF (10 μg/kg/d) for 6 days [15].

Bodyweight and heart rate were measured before and 2 weeks after surgery using an electronic scale for animals and a multi-channel electrophysiology recorder, respectively.

Echocardiography was performed before and 2 weeks after surgery, as previously described [16]. Each rabbit was anesthetized with ketamine and xylazine (25 and 3.75 mg·kg^− 1 im). Echocardiography was carried out using a Vivid 7 ultrasound instrument (GE Healthcare, Chicago, IL, USA) and an S12 transducer. The following parameters were recorded: parasternal left ventricular end-diastolic dimension (LVED, mm), left ventricular end-systolic dimension (LVES, mm), interventricular septum dimension (IVS, mm), left ventricle posterior wall dimension (LVPW, mm), fractional shortening (FS, %), and left ventricular ejection fraction (LVEF, %). The average of three independent measurements was used for the analysis.

The electrophysiology study was performed using a modification of a method described previously [17]. Two weeks after surgery, each animal was anesthetized with ketamine and xylazine (25 and 3.75 mg·kg^− 1 im) and buprenorphine (0.03 mg·kg^− 1 sq), intubated, and ventilated with isoflurane (1–2%, Fi_O2 0.5). Heart rate, QRS interval, and QT interval (lead II) were recorded using a multi-channel electrophysiology recorder. The right jugular vein and carotid artery were identified and separated and heparinized. Then, 4-French, quad-polar electrode catheters were inserted under X-ray guidance into the right and left ventricles through the jugular vein and carotid artery, respectively. The catheters were connected to a multi-channel electrophysiology recorder, and stimuli were applied at twice the diastolic threshold with a pulse width of 2 ms. The evaluated parameters were ventricular effective refractory period (VERP), VERP dispersion, and occurrence of ventricular tachycardia (VT) or ventricular fibrillation (VF). VERP was defined as the shortest S1-S2 interval that elicited an electrogram response in the absence of absolute refractoriness. VERP dispersion was defined as the maximum difference between VERPs measured at the left and right ventricular apex as well as the mitral valve and tricuspid annulus. The presence/absence of VT/VF was determined after the application of S1S1 stimuli (40-ms duration) at the right ventricular apex for 10 s, repeated three times at intervals of 3 min.

After completion of the electrophysiology experiments, the animals were sacrificed with an overdose of pentobarbitone sodium (111 mg·kg^− 1, i.v.), and the heart was excised and washed in 0.9% pre-cooled (4 °C) saline. A portion of the left ventricle containing the papillary muscles was fixed in 10% neutral formalin for 24 h, embedded in paraffin, and sectioned to produce 10 consecutive sections with a thickness of 3.5 μm.

Paraffin-embedded sections were dewaxed using xylene, rehydrated in graded ethanol, and washed in water. The sections were stained with hematoxylin and eosin using standard techniques. The slides were washed in water, graded alcohol, and xylene. After mounting in resinene, 10 fields were randomly selected from each section and observed under an inverted biological microscope (DMI3000B, Leica, Wetzlar, Germany). The ratio of the microinfarcted area to the total area (%) was calculated using the Image-Pro Plus (IPP) 6.0 (Media Cybernetics, Rockville, MD, USA) image analysis software.

Standard techniques were used to stain paraffin-embedded sections with Wiegert’s iron hematoxylin (10 min), plasma stain (acid fuchsin, Xylidine Ponceau, glacial acetic acid, and distilled water; 5 min), phosphomolybdic acid (5 min), and aniline blue (15 min). Collagen fibers are stained blue, cardiomyocytes are stained red, and nuclei are stained blue-black. Sections were observed under the DMI3000B microscope. IPP6.0 was used to calculate the ratio of the area of collagen in the ventricular interstitium (excluding collagen around blood vessels) to the total area.

The expression levels of phosphorylated Cx43 (p-Cx43), total Cx43 (t-Cx43), and G-CSFR in the anterior ventricular wall were evaluated. The paraffin sections were dewaxed, blocked with goat serum, and incubated overnight at 4 °C with goat anti-p-Cx43 (1:300, SC-25165, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-t-Cx43 (1:2500, AB0016, Sicgen), or anti-G-CSFR (1:200, SC-323899, Santa Cruz Biotechnology, Santa Cruz, CA, USA) primary antibody. The sections were incubated with mouse anti-goat secondary antibody labeled with horseradish peroxidase (HRP) (ZSGB Biotechnology Co., Ltd., Beijing, China) for 20 min at 37 °C. The sections were developed using streptavidin and diaminobenzidine. Negative controls were incubated with phosphate-buffered saline instead of antibodies. After mounting with resinene, the sections were evaluated under high magnification (100× and 400×).

Protein samples were extracted from left ventricular myocardium using Protein Extraction Kit (ZSGB Biotechnology Co., Ltd., Beijing, China). The protein concentrations were determined using the bicinchoninic acid protein assay (Beyotime Biotechnology, Jiangsu, China). Equal amounts (50 μg) of protein were subjected to 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, blocked with 5% non-fat milk in Tris-buffered saline (TBS; 1 h at room temperature), and incubated overnight at 4 °C with a primary antibody: anti-p-Cx43 (1:300, SC-101660, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-t-Cx43 (1:300, BA1727, Boster Biological Technology Co., Ltd., California, USA), anti-G-CSFR (1:300, SC-9173, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-t-JAK2 (1:300, SC-278, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-p-JAK2 (1:300, SC-21870, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-t-STAT3 (1:300, SC-8019, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-p-STAT3 (1:300, SC-8059, Santa Cruz Biotechnology, Santa Cruz, CA, USA), or anti-β-actin (1:1000, Cell Signaling Technology, Danvers, MA, USA). β-actin was used as internal control. After three washes in TBS-Tween-20, the membranes were incubated with the appropriate secondary antibody labeled with HRP (1:3000, ZSGB Biotechnology Co., Ltd., Beijing, China) in the dark for 1 h at room temperature. Protein bands were developed by electrochemiluminescence, and their intensities were analyzed using the ImageJ software (National Institutes of Health, Bethesda, MD, USA).

One rabbit from the AG490 group died during the experiments, and 39 rabbits completed the study. Before surgery, there were no significant differences in body weight among groups. Bodyweight at 2 weeks after surgery was significantly lower in the CME group than in the Sham (P < 0.01), G-CSF (P < 0.01), and AG490 (P < 0.05) groups (Fig. 1a). There were no significant differences in body weight among the Sham, G-CSF, and AG490 groups at 2 weeks after surgery (Fig. 1a). Heart rate did not differ significantly among groups after surgery (Fig. 1b).

Before surgery, there were no significant differences among the groups in any of the echocardiography parameters (Fig. 1c–h). At 2 weeks after surgery, the CME group had a significantly higher LVED and LVES (P < 0.01) and a significantly lower LVEF and FS (P < 0.01) than the Sham and G-CSF groups (Fig. 1c, d, g, h). Moreover, LVED, LVES, LVEF, and FS exhibited no significant differences between the G-CSF and Sham groups or between the CME and AG490 groups (Fig. 1c, d, g, h), indicating that G-CSF alleviated the effects of CME and that AG490 blocked the effects of G-CSF. IVS and LVPW were similar among the four groups (Fig. 1E, F).

Representative electrophysiology data are shown in Fig. 2a–b. QRS duration and QT interval did not differ significantly between groups (Fig. 2c, d). VERP was significantly shorter in the CME group than in the Sham and G-CSF groups (P < 0.01) and significantly shorter in the AG490 group than in the Sham group (P < 0.01) (Fig. 2e). Notably, there were no significant differences in VERP between the Sham and G-CSF groups or the CME and AG490 groups (Fig. 2e), suggesting that G-CSF alleviated the effects of CME and that AG490 blocked the effects of G-CSF. VERP dispersion was significantly higher in the CME and AG490 groups as compared with the Sham and G-CSF groups (P < 0.01), with no significant differences between the Sham and G-CSF groups or the CME and AG490 groups (Fig. 2f). Similarly, the VT/VF induction rate and VT/VF mortality rate were significantly higher in the CME and AG490 groups than in the Sham and G-CSF groups (P < 0.01), with no significant differences between the former two groups or latter two groups (Fig. 2g, h). Taken together, the results suggest that G-CSF restored the changes to be similar to those of the Sham group, while AG-490 (a JAK2 inhibitor) blocked the effects of G-CSF.

HE-stained sections of the myocardium demonstrated a normal arrangement of myocardial cells in the Sham group. The CME and AG490 groups exhibited multiple circular regions of microinfarction with inflammatory cell infiltration, but these abnormalities were less prominent in the G-CSF group (Fig. 3a). The microinfarcted area was significantly larger in the CME, G-CSF, and AG490 groups than in the Sham group (P < 0.01), but significantly smaller in the G-CSF group than in the CME or AG490 groups (P < 0.01; Fig. 3a). Staining with Masson trichrome revealed that interstitial collagen deposition was significantly increased in the CME, G-CSF, and AG490 groups compared with the Sham group (P < 0.01), but interstitial collagen content was significantly lower in the G-CSF group than in the CME or AG490 groups (P < 0.01; Fig. 3b). Again, G-CSF alleviated the effects of CME, and that AG490 (a JAK2 inhibitor) blocked the effects of G-CSF.

Immunohistochemistry showed that in the Sham group, the p-Cx43 protein was distributed mainly at the longitudinal junctions between cells where intercalated discs are located (Fig. 4a). By contrast, p-Cx43 showed a disordered distribution in the CME and AG490 groups, with substantial lateralization of p-Cx43 and only limited expression at the longitudinal junctions between myocytes (Fig. 4a). The distribution pattern in the G-CSF group was between that of the Sham group and the two other groups (Fig. 4a). The p-Cx43 expression level was significantly lower in the Sham group than in the other groups (P < 0.01) and substantially higher in the G-CSF group than in the CME or AG490 groups (P < 0.01) (Fig. 4a). The cellular localization of t-Cx43 in each group was similar to that of p-Cx43 (Fig. 4b). There were no significant differences among groups in the t-Cx43 expression levels (Fig. 4b). G-CSFR expression was much higher in the G-CSF group than in the other groups (P < 0.01; Fig. 4c). G-CSFR expression was slightly higher in the CME group than in the Sham and AG490 groups (P < 0.05; Fig. 4c). Those results suggest that G-CSF alleviated the changes in Cx43 expression induced by CME and that AG490 (a JAK2 inhibitor) blocked the effects of G-CSF.

Western blot experiments were performed to measure the protein expression levels of p-Cx43, t-Cx43, G-CSFR, p-JAK2, t-JAK2, p-STAT3, and t-STAT3 (Fig. 5). The western blot data for p-Cx43, t-Cx43, and G-CSFR were consistent with the immunohistochemistry results. There were no significant differences among groups in the expression levels of the t-JAK2 and t-STAT3 proteins (Fig. 5). The expression levels of p-JAK2 and p-STAT3 were significantly higher in the G-CSF group than in the Sham group (P < 0.01), but these elevations were attenuated in the AG490 group (Fig. 5).