Danqi Pills (16120005) were purchased from TongrenTang (Beijing, China) and has a strict quality control by Pharmacopeia of the People’s Republic of China (Ministry of Health of the People’s Republic of China Pharmacopeia Committee, 2010) and definite clinical efficacy without known side effects. DQP is composed with the root of red-rooted salvia (S. miltiorrhiza Bge) and Panax notogenseng (Notoginseng Radix et Rhizoma). The fingerprint of DQP was analyzed by high-performance liquid chromatography (Supplementary Figure S1).

Medium 199 (M199, 10-060-CVR), matrigel (356231), and fetal bovine serum (FBS, 35-081-CV) were purchased from Corning (United States). Antibody against GAPDH (5174S) was purchased from Cell Signaling Technology (United States), antibodies against CD36 (ab64014), CPT1A (ab128568), VEGF-2 (ab10972), PPARα (ab3484) were obtained from Abcam (United States). CD31 (GB12063) was purchased from Servicebio (China). Secondary HRP-conjugated anti-mouse and anti-rabbit antibodies (Beijing TDY Biotech LTD, E009, and E011, respectively). Etomoxir (B1526025) was purchased from Aladdin (China). Trimetazidine (2010447) were purchased from Servier Pharmaceutical Company Limited (Tianjin, China). Calcinin (354216) was purchased from BD. Cell Counting KIT-8 was purchased from Japan (CCK-8, Dojindo Laboratories Inc., Kumamoto, Japan).

The animal experiments were performed according to the Care and Use Guide of Laboratory Animals published by the National Institutes of Health (NIH Publications No. 85-23, revised 1996). All procedures involving animals were approved by the Animal Care Committee of Beijing University of Chinese Medicine. SD Rats (220 ± 10 g) used in the experiment were purchased from Sbef (Beijing) laboratory Animal Science and Technology Company (Beijing, China). Rats were housed in specific pathogen free (SPF) class animal house of Beijing University of Chinese Medicine with 12 days and night cycles in temperature 22 ± 2°C, humidity 50–60%.

Ischemic heart model was prepared as previously described (Wang et al., 2015). In brief, 1% pentobarbital (50 mg/kg) was injected into intraperitoneal for anesthesia. After tracheal intubation, thoracotomy was performed between the 4th and 5th intercostal of rats and left anterior descending (LAD) coronary artery was ligated. After closing the chest, 0.1–0.2 ml lidocaine and 0.1–0.2 ml furosemide were injected into the abdominutesal cavity. The sham rats underwent the same procedures except that the LAD was not ligated. After the rats resumed spontaneous breathing, the trachea was pulled out. According to a random number method, rats (that had undergone surgery and survived) were divided into a sham group, a model group, a DQP group, and a trimetazidine group on the second day, with ten rats in each group. Among them, rats in the DQP group received a daily oral gavage of DQP solution at dose of 1.5 mg/kg for 28 days. Rat in the positive control group was treated with a daily gastric trimetazidine solution at dose of 6.3 mg/kg for 28 days. Rats in the sham group and the model group had the same volume of saline. 28 days after surgery, 1% pentobarbital was injected into the abdominal cavity for anesthesia. The heart was harvested and frozen in liquid nitrogen.

Echocardiography was used for measuring cardiac function. Parameters included ejection fraction (EF), left ventricular anterior wall; d (LVAW; d), left ventricular anterior wall; s (LVAW; s), Left ventricular end-diastolic diameter (LVED; d), left ventricular end-systolic diameter (LVED; s). The calculated Fractional shortening (FS%) is as follows: FS% = [(LVED; d-LVED; s)/LVED; d] × 100%.

Proteins were extracted by Radio Immunoprecipitation Assay (RIPA) lysate (supplemented with a protein phosphatase inhibitor cocktail) (Beijing PuLilai Gene Technology Co., Ltd., Beijing, China, lot number: P-1260-1). The protein concentration was measured by using a protein quantification kit (Beijing PuLilai Gene Technology Co., Ltd., Beijing, China, lot number: P1511-1) and a Spectra Max i3x microplate reader (Molecular Devices, United States). Then the sample (50 μg, protein) was added to a 8% SDS-PAGE gel for electrophoresis, kept at 80 voltage for 30 min and then at 120 voltage for 75 min by using the PowerPac Universal Power Supply (United States Bio-Rad). The proteins on the gel were transferred to PVDF membrane (GE Healthcare, United States, 10600023) and electrophoresis was performed at 250 mA for 90 min. The membrane was incubated with primary and secondary antibodies and then treated with Amersham ECL primary immunoblot detection reagent (GE Healthcare, United States) for 1 min at room temperature. Membranes were exposed to a Molecular Imager ChemiDoc XRS+ system (Bio-Rad, United States) and the bands on the membranes were observed and analyzed by Image Lab software.

Heart tissues were fixed in 4% paraformaldehyde for 72 h and then were embedded in paraffin and sectioned into 5 μm slices. Each group has four slides. The paraffin sections were subjected to haematoxylin-eosin (HE) and CD31 stainings. The comparison were made at similar location of the infarcted border heart tissues. The assessment was carried out by a researcher blinded to the interventions.

Human umbilical vein endothelial cells (HUVEC) were purchased from Promocell (C12206, German) and cultured in the endothelial cells medium (C-22011, German) at 37°C and 5% CO₂. EC were seeded 6000 cells/well in a 96-well plate. EC were stimulated with different concentration of DQP for 24 h and then treated with 50 μM H₂O₂ for 3 h to induce EC injury. 10 μl of CCK-8 was then added to each well. After 4 h, the optical density (OD) value of per well was determined with a microplate reader at 450 nm.

96 well plates were pre-coated with growth factor-reduced matrigel and the DQP-treated EC were seeded in well plates. After 2 h, EC were treated 50 μM H₂O₂ for 3 h. Then the diluted calcinin was added to the well for 10 min. The liquid was discarded. EC were washed twice with PBS and observed using an inverted fluorescence microscope. The results were analyzed by Image Pro-Plus software.

All data were presented as mean ± SD. Statistical analysis was performed with the SPSS program package (SPSS version 20.0) or GraphPad Prism 5. Statistical analysis was carried out using one way analysis of variance (ANOVA). The values of P < 0.05 were considered as statistically significant.

After ligation of LAD for 28 days, EF and FS were reduced by 60.5 and 71.8%, respectively, compared with sham group, indicating that ischemic model was successfully induced and cardiac functions were impaired in the model group. In DQP treatment group, EF was up-regulated by 44.0% and FS was up-regulated by 52.7% compared with model group. What’s more, LVID;s decreased in the DQP group (Figure 1A, Table 1). Positive control drug trimetazidine showed similar effects as DQP. In the sham group, the LV cardiac myocytes were arranged in an orderly pattern. In model group, myocardial cells in non-infarcted areas are hypertrophic, disordered, and infiltrated by inflammatory cells. DQP and Trimetazidine group significantly improved (Figure 1B). These results indicated that DQP had cardio-protective effects.

CD31 is a marker of angiogenesis and immunohistochemical staining showed that DQP could significantly increase the intensity of capillaries in the marginal area of infarction of the rat heart (Figure 2A). VEGF is the most potent and widely investigated proangiogenic growth factor. VEGF promotes angiogenesis by stimulating endothelial proliferation, migration, and capillary tube formation. The results showed that expressions of VEGF-2 were up-regulated by DQP treatment, indicating that DQP had angiogenic properties (Figure 2B). DQP also increased the expression of iNOS, decreased the expression of IL-1β and had no effect on LDAH (Supplementary Figures S3, S4).

CD36, also known as fatty acid translocase, imports fatty acids into cells. CPT1A is the key rate-limiting enzyme during β-oxidation of fatty acids. It promotes the transportation of fatty acids from cytoplasm into mitochondria for oxidation. PPAR-α could promote the expression of CPT1A (Li et al., 2015). Expressions of these proteins were quantified by western blotting. DQP could significantly up-regulate expression of CPT1A in the marginal area of the infarcted heart (Figure 3A). Expression of CD36 increased in the DQP group, although the difference was not significant (Figure 3B). Expression of PPAR-α was also significantly up-regulated by DQP (Figure 3C). The data demonstrated that DQP could promote β-oxidation of fatty acids by regulating PPAR-a-CD36-CPT1A axis under ischemic conditions.

Endothelial cell injury model was induced by incubation with 50 μM H₂O₂. Different concentrations of DQP were applied to EC injury model. The effects of DQP on H₂O₂-induced cytotoxicity were detected by CCK8 assay and OD values were measured in different groups of cells. As shown in Figure 4A, DQP treatment (400 and 600 μg/ml) provided protective effects against H₂O₂-induced injury.

The angiogenic capability of HUVEC was assessed using capillary-like tube formation assay on growth factor-reduced matrigel. The formation of capillary-like tube network by EC was compromised by H₂O₂ treatment (Figure 4B). Capillary-like tubes with higher number of branching points and longer capillary tube lengths were observed in DQP treatment groups, suggesting that DQP possessed pro-angiogenic ability.

Effects of DQP on key enzymes in fatty acids oxidation pathways in H₂O₂-stimulated cells were examined. Treatment of EC with H₂O₂ reduced expressions of CPT1A, CD36, and PPAR-α. Incubation of EC with DQP (400 and 800 μg/ml) up-regulated expressions of CPT1A, CD36, and PPAR-α (Figure 5). The data suggested that DQP could promote fatty acids oxidation in endothelial cells by activating key enzymes in β-oxidation pathway.

To investigate if DQP exerted angiogenic effects through fatty acids oxidation pathway, etomoxir, the specific inhibitor of CPT1A (Rupp et al., 2002), was used to treat H₂O₂-stimulated HUVEC with DQP. CCK8 assay showed that DQP protected EC against H₂O₂ induced cytotoxity. When endothelial cells were co-treated with etomoxir, proliferation of cells was reduced, as demonstrated by decreased OD value (Figure 6A). Furthermore, the pro-angiogenic ability of DQP was also compromised by blockade of CPT1A, as the capillary-like tubes formed by HUVEC were destructed by co-treatment with etomoxir (Figure 6B and Supplementary Figure S2). Meanwhile, the ATP production increased in response to DQP and inhibited by etomoxir (Figure 6C). These results suggested that DQP promoted tube formation ability of endothelial cells through fatty acids oxidation pathway.