Adult SD rats (250–300 g) were perfused by Langendorff system ex vivo. Details of the experiments were described previously (Gonon et al., 1998). Briefly, rats were anesthetized by pentobarbital (60 mg/kg). The heart was excised and perfused by a Langendorff apparatus at a constant pressure of 55 mmHg. The phosphate free Krebs-Henseleit buffer (Tian and Abel, 2001) was continuously gassed with 95% O₂/5% CO₂ (pH 7.4) and warmed by a heating bath. The heart temperature was continuously monitored and maintained at 37 ± 0.5°C. Ischemia was induced by the cessation of perfusion for 45 min then followed by reperfusion 60 min with PD168077 (TOCRIS, United States) or not, then LVEDP, LVDP, +dp/dt were harvested. The procedures for care and use of animals were approved by the Ethics Committee and Animal Care and Use Committee of Third Military Medical University. All applicable institutional and governmental regulations concerning the ethical use of animals were followed.

Adult mice cardiomyocytes (AMCs) and neonatal rat ventricular myocytes (NRVMs) were obtained as our previous work (Wang et al., 2010; Wang et al., 2017). For A/R model, NRVMs were obtained from neonatal rat hearts (<3 days) and grown in DMEM containing 10% FBS at 37°C in a humidified incubator with 95% atmosphere and 5% CO₂. After starvation for 12 h, cells were transferred into a hypoxic incubator containing 95% N₂, 5% CO₂ for 4 h; for reoxygenation, cells were rapidly transferred into a normoxic incubator containing 95% O₂ and 5% CO₂ for gradient time-point reoxygenation after anoxia with fresh DMEM and fetal bovine serum (Supplementary Figure S1A), then 8 h was chosen for A/R model.

Frozen section of heart tissue and slides containing NRVMs or AMCs were rinsed with PBS, then blocked with 10% normal goat serum. Next, the sections were incubated with DRD4 primary antibody (Sigma, Germany, 1:100 dilution) or GLUT4 primary antibody (Abcam, United Kingdom, 1:100 dilution) for 12 h at 4°C, then incubated with appropriate secondary antibody (Thermo, United States) for 2 h at 37°C, next, the DAPI and (or) DiO perchlorate (Solarbio, China, 5 × 10⁻⁶ M) were added for 10 s at room temperature and (or) for 15 min at 37°C. The immunofluorescence staining was visualized using Olympus confocal microscope at DAPI, DiO, 488 and 546 nm emission.

To measure the infarct size, perfused rat hearts were frozen at −20°C for 30 min and cut into slices (5–6 slices/heart), then incubated in a sodium phosphate buffer containing 1% 2,3,5-triphenyl-tetrazolium chloride (Sigma, Germany) for 15 min to visualize the unstained infarct region. The infarct areas and whole ventricle areas were determined by planimetry with ImageJ. The infarct size was calculated as infarct area divided by ventricle area.

For total protein, heart tissues and NRVMs were lyzed in radio immunoprecipitation assay (RIPA) buffer supplemented with protease inhibitor cocktail tablet and phosphatase inhibitor tablet (Roche, Germany). Protein content was measured according to the Pierce BCA protein assay kit (Sangon Biotech, China). For membranous protein, membrane protein in heart tissues and NRVMs were separated by the Mem-PER™ Plus Membrane Protein Extraction Kit (Thermo Fisher, United States) following the manufacturer’s protocol (Phuchareon et al., 2015). Then equal protein lysates were resolved electrophoretically on denaturing SDS-polyacrylamide gels, and transferred to 0.22 μm nitrocellulose membranes. After blocking in 5% milk in TBST, membranes were probed with the following primary antibodies: DRD4 (Sigma, Germany), cleaved caspase 3 (CST, United States), PMCA (CST, United States), GLUT4 (CST, United States), p-PI3K (CST, United States), PI3K (CST, United States), p-AKT (CST, United States), AKT (CST, United States), ATP1A1 (CST, United States) and GAPDH (Abcam, United Kingdom). The appropriate secondary antibody (LI-COR, United States) was used. The ATPase alpha-1 (ATP1A1) and plasma membrane-type Ca2⁺-ATPases (PMCA) as the loading control of membrane protein, while GAPDH as loading control of the total protein. The results were tested by Odyssey system and quantified by ImageJ.

The LDH ELISA kit (Abcam, United Kingdom) and TNT ELISA kit (Thermo, United States) were used for measuring the LDH and TNT in perfusion liquid according to the manufacturers’ instructions.

TUNEL staining was performed by a in situ cell death detection kit (Roche, Germany) as described previously (Khan et al., 2013). Paraffin section (4 µm) were mounted on slides or slides containing NRVMs were conducted following manufacturer’s instructions. TUNEL positive cells were detected by Olympus confocal microscope.

NRVMs exposed to DRD4 agonist (PD168077, 10⁻⁵ M) during A/R injury were examined by PE Annexin V Apoptosis Detection Kit I (BD, United States) described previously (Schwartzenberg et al., 2007). According to manufacturer’s instructions, cells were stained with PE Annexin V and 7-amino-actinomycin (7-AAD) to detect early apoptotic cells (PE Annexin V+/7-AAD-events) and late apoptotic cells (PE Annexin V+/7-AAD + events), which were determined by flow cytometry (BD, United States).

After A/R or not, NRVMs were exposed to PD168077 (10⁻⁵ M) or (and) Wortmannin (10⁻⁸ M) for another 24 h in fresh DMEM. Next, 10 μl CCK8 (Dojindo, Japan) solution was added to each well and the plate was incubated for additional 2 h. Absorbance readings at 450 nm were obtained using a spectrophotometer (Thermo, United States). 10⁻⁸ M Wortmannin was chosen because cell viability was insusceptible when the Wortmannin dose was below 10^–7 independently (Supplementary Figure S1B).

The study was carried out according to the method described previously (Kaliman et al., 1995). Briefly, after A/R or not, NRVMs were incubated with PD168077 or (and) Wortmannin for 30 min, then glucose-free incubation was performed for 45 min in Krebs-Ringer phosphate buffer (Asano et al., 1992), next 100 μM 2-deoxy-D-[³H] glucose (0.5 ìCi/ml) was added. Transport was stopped 20 min later and measured by a liquid scintillation counter (Microbeta Trilux, Finland).

The data were expressed as mean ± SEM. Two-tailed Student t test for normally distributed data and Mann-Whitney nonparametric test for skewed data that deviate from normality were used to compare two conditions. One-way analysis of variance with Bonferroni post hoc test for normally distributed data and Kruskal-Wallis nonparametric test for skewed data were used to compare ≥3 means. Repeated measures ANOVA was used for repeated measures data. Differences with p < 0.05 were considered statistically significant.

Our previous works found DRD4 was expressed on kidney renal proximal tubule (RPT) cells (Chen et al., 2015; Zhang et al., 2016), so it was used as a positive control. As shown by immunofluorescence data (Figure 1A), the DRD4 was detectable in the RPT cells, NRVMs, AMCs and heart tissue. Meanwhile, this result was confirmed by western blot, which was observed that the DRD4 was expressed in RPT cells and WT rat heart but not the negative control, DRD4 ^−/− mice (Figure 1B). These results indicated that DRD4 was expressed on rat and mice cardiomyocytes.

To analyze the role of DRD4 on cardiomyocytes during I/R injury, the isolated rat heart was perfused by Langendorff system. After that, the TTC staining was performed. Compared with the I/R group, DRD4 activation by PD168077 markedly decreased myocardial infarction size (Figures 2A,B), profoundly improved cardiac function, evidenced by the lower LVEDP (Figure 2C), higher +dp/dt (%) and LVDP (Figures 2D,E). Moreover, under PD168077 administration, the damage in cardiomyocytes was alleviated by decreased LDH and TNT in perfusion fluid compared with I/R group (Figures 2F,G). These results demonstrated that DRD4 activation improved cardiac function and decreased the damage in cardiomyocytes during I/R injury.

Cardiomyocytes are the main basis of cardiac function. Loss of them significantly worsen cardiac function. The TUNEL staining result showed that the apoptosis was enhanced under I/R injury, while this effect was diminished in DRD4 activation group (Figures 3A,B). Similarly, the expression of cleaved caspase 3, an apoptosis biomarker, was increased under I/R injury while reduced when DRD4 was activated by PD168077 (Figures 3C,D). Next, the effect of DRD4 on NRVMs was further determined. After A/R, the cell viability was blocked, while this effect was impaired by activation of DRD4 (PD168077, 10⁻⁵ to 10⁻⁷ M) (Figure 3E). Furthermore, the apoptosis cell was dramatically increased during A/R. However, this effect was mitigated after DRD4 activation (PD168077, 10⁻⁵ M) showed by flow cytometry (Figures 3F,H) and TUNEL staining (Figures 3G–I). These observations demonstrated that activation of DRD4 exert cardioprotection depended on decreasing apoptosis.

Since glycolysis becomes the main energy source under I/R injury, glucose uptake becomes more critical, insufficient glucose supply induces apoptosis in cardiomyocytes. As we known, GULT4 is more prone to translocation due to the I/R injury. It was observed that the membrane GLUT4 was dramatically increased during the I/R injury, and this effect was enhanced after DRD4 activation in heart (Figures 4A,B), the similar results are also found in NRVMs after A/R (Figures 4C,D). Next, GLUT4 translocation promoted by PD168077 was confirmed in NRVMs by immunofluorescence staining, which was co-localized with the membrane-specific DiO (Figure 4E). Furthermore, increased GLUT4 could effectively promote glucose transport monitored by [³H]-2-deoxy-D-glucose uptake analysis in vitro (Figure 4F). These results suggested that activation of DRD4 reduced apoptosis through promoting GLUT4 translocation.

As we know, GLUT4 translocation via the PI3K/AKT pathway was involved in I/R injury in cardiomyocytes (Matsui et al., 2001; Okumura et al., 2004; Cao et al., 2010; Ji et al., 2013; Chen et al., 2017). The current study found that p-PI3K and its downstream p-AKT were augmented during A/R injury, and this effect was enhanced by DRD4 activation but not the total PI3K or AKT (Figures 5A–D). Next, the PI3K inhibitor, Wortmannin, a covalent inhibitor of PI3Ks, was used to analyze the PI3K/AKT pathway does regulate the DRD4-induced GULT4 translocation. The p-AKT increase caused by activated DRD4 was weakened under wortmannin (10⁻⁸ M) treatment during A/R (Figures 5E,F), so did the GLUT4 on the membrane (Figures 5G,H). Moreover, increased GULT4 could effectively promoted glucose transport, while this effect was depleted by wortmannin, which was monitored by [³H]-2-deoxy-D-glucose uptake analysis (Figures 5I). These results demonstrated that activation of DRD4 induced GLUT4 translocation through the PI3K/AKT pathway.

Whether blocking the PI3K/AKT pathway blocked the activated DRD4 mediated apoptosis in cardiomyocyte. Firstly, Wortmannin counteracted the increased cell viability which is enhanced by DRD4 activation (Figure 6A). Next, the role of Wortmannin about apoptosis was determined by Flow cytometry (Figures 6B,C) and TUNEL staining (Figures 6D,E), both results showed that DRD4 activation decreased apoptosis under A/R injury. However, this effect was offset by Wortmannin. Meanwhile, the effect of wortmannin on DRD4 mediated apoptosis was verified by the apoptosis marker, including cleaved caspase-3, BAX, and BCL2. Activation of DRD4 decreased cleaved caspse-3 and BAX, while increased BCL2 during A/R injury, however, this protective effect was offset by Wortmannin (Figures 6F–I). These data indicated that DRD4 activation reduced apoptosis through PI3K/AKT mediated GLU4 translocation in vitro.