To generate MI model, male C57BL/6 mice (8–10 weeks age, 20–25 g weight) were anesthetized with a mixture of xylazine (5 mg/kg) and ketamine (100 mg/kg) through intraperitoneal injection, and the left thoracotomy and left coronary artery ligation were performed as described elsewhere (16). Successful ligation was confirmed by both ST-segment elevation on the electrocardiogram and paleness of the myocardium. The infarct size was determined by triphenyltetrazolium chloride (TTC) staining at 6 h after surgery (17). Four weeks after MI, half of the mice with MI and an LV end-systolic volume index (LVESVI) >12 ml/m² were randomly selected for SVR.

Four weeks after MI or sham surgery, echocardiography was performed over a time course of 3 weeks (8 weeks for some mice) to assess the cardiac remodeling and function. Mice were anesthetized by inhaling 3% isoflurane and maintained anesthesia with 1.5–2% isoflurane by using a Vevo 2,100 system with a 30 MHz probe (Fujifilm VisualSonics, Ontario, Canada). During the operating process, mice were placed on the warming platform in the supine position and their heart rate was maintained between 400 and 550 beats/min (18, 19). M-mode images were recorded in the long-axis view of the LV to observe the cardiac morphology, and the two-dimensional parasternal short-axis views of LV were obtained at the papillary muscle level to guide the M-mode measurement of the LV diastolic posterior wall thickness (LVPWd, mm), LV systolic posterior wall thickness (LVPWs, mm), LV end-diastolic diameter (LVEDd, mm) and LV end-systolic diameter (LVEDs, mm). The Teichholz formula [V = 7D³/(2.4+D), V = volume, D = the internal diastolic or systolic dimension measured by echocardiography] was used to calculate the LV end-diastolic volume (LVEDV, ul) and LV end-systolic volume (LVESV, ul). The body surface area (BSA, m²) was calculated according to the Meeh-Rubner formula: BSA = k × (W^2/3)/10,000, k = 9.1, W = body weight (BW, g). LV ejection fraction (LVEF, %), LV end-diastolic volume index (LVEDVI, ml/m²) and LV end-systolic volume index (LVESVI, ml/m²) were calculated as follows: LVEF = ((LVEDV-LVESV)/LVEDV) × 100%, LVEDVI = (LVEDV/1,000)/BSA, LVESVI = (LVESV/1,000)/BSA. The right ventricle (RV) parameters were obtained as described by Wang et al. and the four-chamber view was obtained to measure the tricuspid annular plane systolic excursion of tricuspid valve (TAPSE) as described elsewhere (20–22).

At 3 weeks after SVR, speckle tracking was performed in sham, MI, and MI+SVR group. The two-dimensional (B-mode) parasternal long axis and short axis echocardiography were reserved for VEVO Lab (Fujifilm Visual Sonics). In short, three consecutive cardiac cycles with obvious R wave under B-mode ultrasound were selected for analysis. The endocardial and epicardial boundary were manually circled and the related parameters of endocardium and epicardium were measured both in longitudinal and circumferential axial plane. The measurement indexes included: left ventricular area change fraction (FAC), left ventricular global longitudinal strain peak (GLS), global radial strain peak (GRS) and global circumferential strain peak (GCS) (23).

LV hemodynamics were measured by using a Millar catheter system before the mice were sacrificed as described elsewhere (19, 24). To ensure the accuracy of the hemodynamic data as much as possible, we standardized the anesthetic dose of the mice to 2/3 of the normal anesthetic dose based on body weight (BW). During the procedure, the mice were placed on a 37°C thermostat plate to ensure a consistent body temperature for all mice. All mice had no major bleeding or fluid loss during the invasive assay and were not given additional rehydration.

Seven weeks after the 1st surgery, the mice were sacrificed by overdosing pentobarbital (150 mg/kg, i.p). BW, heart weight (HW) and tibia length (TL) were measured. The heart was fixed in 4% paraformaldehyde and embedded in paraffin according to standard protocols. Azan-Masson staining was used to evaluate myocardial fibrosis (25). Immunofluorescence staining was performed to evaluate the regeneration of cardiomyocytes (CMs) and vessels. Three to five different slides of the same sample were observed and averaged. Paraffin-embedded blocks of heart tissue waxes were cut into 4 mm tissue slices and then deparaffinized and subjected to antigen retrieval. Tissue sections were then rinsed three times with PBS for 5 min each and permeabilized for 20–25 min with 0.1% Triton X-100. After permeabilization, samples were rinsed three times with PBS for 5 min each and blocked with 2% bovine serum albumin (BSA) for 1 h at room temperature. Then, samples were incubated for 16–18 h at 4°C with primary antibodies [anti-Ki67 (1:600, Abcam, USA), anti-Ph3 (1:600, Invitrogen, USA), anti-Aurora B (1:600, Abcam, USA), anti-cTnI (1:200, Proteintech, USA), anti-αSMA (1:200, Abcam, USA)]. After incubation, samples were rinsed three times with PBS for 5 min each and stained with secondary antibodies (Alexa Fluor 488, Abcam, 1:200; Alexa Fluor 555, Abcam, 1:200) for 1 h at room temperature. DAPI sealing solution (Beyotime, USA) was used to stain the nuclear and seal the slices.

Total RNA was obtained from mouse heart tissue with a total RNA isolation system (Omega, Norcross, GA, USA) and then converted to cDNA using oligo (dT) primers with PrimeScriptTM RT Master Mix (Takara Bio Inc., Shiga, Japan). Quantitative reverse transcription–polymerase chain reaction (qRT-PCR) was performed by using SYBR Green PCR Master Mix (Takara Bio Inc., Shiga, Japan) and a Light Cycler 480 System (Roche, Germany). The sequences of the primers are listed in Supplementary Table 1.

A total of 140 mice were used in this study, of which 40 were used in the sham group and 100 were used for MI surgery at the beginning. Seventy-five mice survived after MI surgery. Sixty-two were eligible for SVR surgery after echocardiography certification. Of these 62 mice, five were randomly selected for sampling and PCR experiments. Then, 30 mice were randomly selected for SVR surgery (MI+SVR group) and the remaining 27 mice were kept as MI group. Five mice died during SVR surgery due to mislabeling or excessive blood loss. Ten mice in each group were selected for continuous observation of echocardiographic indices and statistics, and the remaining mice were measured at individual time points to ensure successful surgery. Three weeks after SVR, 6 sham mice, 7 MI mice and 5 SVR mice were randomly selected for whole transcriptome sequencing to better investigate the molecular biological mechanism of SVR (not involved in this study at this time, data are not shown), while 5 SVR mice were kept under observation until 8 weeks after SVR surgery, and the hearts of these mice were not stained or subjected to PCR experiments. For the remaining mice, 10 of each group were randomly selected for speckle tracking measurements, 5 mice for invasive hemodynamic testing, 10 mice for HW and TL measurements, 5 mice for histological staining (while 10 MI mice for Masson staining), and 3 to 5 mice for PCR of target genes in myocardial tissue.

The data were analyzed with GraphPad Prism 7.0 software (GraphPad Software, Inc., CA, USA). All quantitative data are presented as the mean ± standard error of the mean (SEM). Comparisons among multiple groups were performed using one-way analysis of variance followed by Bonferroni's correction or paired Student's t-test for multiple post hoc comparisons. Survival analysis was performed using the log-rank (Mantel-Cox) test. A P-value < 0.05 was considered statistically significant.

To induce sufficient LV remodeling, the mouse left coronary artery was ligated 1 mm below the left atrium, and a significant ST segment elevation and infarct size larger than 40% were used as quality control criteria (Figures 1A,B). Four weeks after MI, 75% of mice survived and exhibited severe LV remodeling, as manifested by a 100% incidence of LVA, a significantly increased LV volume, a markedly reduced LVEF, and significant upregulation of natriuretic peptide type A (Nppa) and natriuretic peptide type B (Nppb) gene expression (Figures 1C–G).

Four weeks after MI, half of the mice with MI underwent SVR (Figure 2A). A time-course monitoring of echocardiography was performed in Sham, MI and MI + SVR groups (Figure 2B). The results indicated that the baseline data were similar between the MI and MI + SVR groups at 4 weeks after MI (Figures 2C–K), but that the LV posterior wall thickness exhibited a modest increase in SVR-treated mice but no significant change in untreated mice with MI (Figures 2B–D). Furthermore, the LV volume decreased and systolic function index of LVEF increased in a time-dependent manner in SVR-treated mice, whereas there was a gradual increase in LV volume and a decrease in LVEF in the untreated MI group (Figures 2B,E–G). However, no significant effect of SVR on right ventricular dimension and function was noted, as reflected by the pressure gradient across the pulmonary valve (PV maximum pressure), right ventricular free wall thickness, and TAPSE (Figures 2H–K).

To determine how long the SVR-induced improvement of LV remodeling persisted, we extended the observation period to 8 weeks after SVR. No significant changes in LV volume or LVEF were noted at 8 weeks after SVR compared to 3 weeks after SVR (Supplementary Figures 1A–E).

Speckle-tracking of echocardiography and invasive LV hemodynamic analysis were used to evaluate heart function 3 weeks after SVR. As shown in Figure 3A, myocardial motion in the sham group showed spiral morphology with a strong strain capacity in both the longitudinal and circumferential views, whereas in the MI group, the myocardial motion trajectory was irregular, and the strain capacity was weak. In contrast, mice with MI subjected to SVR had a much better myocardial trajectory and strain capacity than untreated mice. Quantitative analysis showed that the GLS, GCS, and GRS of the LV were markedly impaired in both the endocardium and epicardium in the MI group, while SVR reversed these trends (Figures 3B–D). Similarly, the LV FAC was also altered in the MI group, whereas SVR reversed this change to a certain extent (Figure 3E). These results collectively indicate that LV systolic function improves in response to SVR.

We further analyzed the strain in 6 myocardium segments on the longitudinal and circumferential plane of the LV (Supplementary Figures 2A–H). The strain of the myocardium segments in the MI group was significantly impaired, which was reversed by SVR (Supplementary Figure 2A). Quantitative analysis showed that the radial strain of almost all segments on the longitudinal axis improved significantly after SVR, except the posterior and anterior apexes, which were the area of the LVA (Supplementary Figure 2C). SVR significantly improved the longitudinal strain of the endocardium and epicardium in posterior middle segment, as well as the longitudinal strain of the epicardium in anterior middle segment (Supplementary Figures 2D,E). The strain variation of the circumferential segments was not as uniform as that of the longitudinal segments. SVR markedly improved the radial strain in posterior wall and inferior free wall, the endocardial strain in anterior free wall and lateral wall, and the epicardium strain in posterior wall segment (Supplementary Figures 2F–H). LV hemodynamic analysis revealed that the mice with MI had a lower LV systolic pressure (LVSP), a lower maximum change rate of the LV pressure (LV dp/dt max and dp/dt min), lower LV contractility and a longer LV exponential relaxation time constant (τ) than in the sham group (Figures 4A–E), while the SVR group had a larger LV dp/dt max, a larger dp/dt min, greater LV contractility and a shorter τ than untreated mice with MI (Figures 4C–E). These results suggest that SVR can improve both LV systolic and diastolic function.

To further consolidate the beneficial effect of SVR on LV hemodynamic, we performed pressure-volume catheterization at 4 weeks after SVR, and noted a smaller end-systolic elastance and larger end-diastolic stiffness, LV end-diastolic pressure and τ in the MI group than in sham group, and these changes were partially reversed by SVR (Supplementary Figures 3A,B).

Three weeks after SVR, the hearts of SVR mice were markedly smaller than that of untreated mice with MI (Figure 5A). Quantitative analysis indicated that the SVR group had a smaller HW/ BW ratio and HW/ TL ratio, less myocardial fibrosis and lower expression of the Nppa and Nppb genes than untreated MI mice (Figures 5B–E). The typical longitudinal views of plication part are shown in Supplementary Figure 4.

Seven weeks after MI, very few or no proliferating CMs were visible in the border zone of the heart, as indicated by staining for Ki67, Aurora B and pH3 (Figures 6A–C). In contrast, SVR led to a significant increase in CMs proliferation around the plication zone after 3 weeks, with the SVR group exhibiting a more than 2-fold increase in CM proliferation compared to the 7-week MI group (Figures 6A–C). These findings indicate that SVR enhanced the proliferative capacity of CMs. In addition, we found more alpha smooth muscle actin (α-SMA)-positive vessels and higher expression levels of vascular endothelial growth factor-α (VEGFα) around the plication zone in SVR-treated mice than in untreated mice, suggesting that SVR increases the vessel density in the heart after MI (Figures 7A–C).