Rat cardiomyocyte-derived H9c2 cells from the National Infrastructure of Cell Line Resources (Chinese Academy of Medical Sciences, Beijing, China) were cultivated in Dulbecco’s Modified Eagle Medium (DMEM), high glucose (Biological Industries, Israel) supplemented with 10% fetal bovine serum (FBS, Biological Industries, Israel), 1% penicillin-streptomycin (Corning, United States) at 37°C in a 5% CO₂ atmosphere. In vitro experiments were performed using H9c2 cardiomyocytes between passages 15 and 20, which were subcultured at a confluence of approximately 80%.

H9c2 cardiomyocytes (5,000/well) were cultured into 96-well plates for 24 h, and were exposed to a series concentration of evodiamine and rutaecarpine (Shanghai Yuanye Bio-Technology Co., Ltd., China) for another 24 h. Cell viability was measured using the cell counting kit-8 (CCK-8) solution assay (Biorigin Inc., China) at 450 nm. Subsequently, the absorbance values were applied to calculate the half maximal inhibitory concentration (IC₅₀ values) and select appropriate concentrations for further experiments. All experiments were performed independently in triplicate.

After incubation with different concentrations of evodiamine (5, 10, 25 μM) and rutaecarpine (60, 80, 100 μM) for 24 h, according to the manufacturer’s directions, the leakage of lactate dehydrogenase (LDH) and the activity of creatine kinase (CK) were determined using the commercial LDH and CK detection kits (Nanjing Jiancheng Bioengineering Institute, China), respectively. Additionally, the mitochondrial membrane potential and the intensity of calcium fluorescence were evaluated with the JC-1 and Fluo-3AM detection kit (Beyotime Biotechnology, China) through FACS Calibur flow cytometry detection (Becton, Dickinson and Company, United States).

The proteins of H9c2 cardiomyocytes from different groups were harvested and lysed with cold RIPA buffer (Beijing Solarbio Technology Co., Ltd., China), supplemented with a protease inhibitor cocktail for 15 min on ice, and the concentrations of the supernatant were measured with a BCA protein assay kit (Beijing Solarbio Technology Co., Ltd., China). Briefly, equal amounts (10 µg) of protein were separated via pre-cast 8% SDS-polyacrylamide gel and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore Inc., United States). After blocking with TBST containing 5% skim milk for 1 h at room temperature, the PVDF membranes were incubated overnight at 4°C with PRKG1 antibody (1:1,000, Proteintech Group, Inc., United States), cGMP antibody (1:1,000, Santa Cruz Biotechnology, United States), and GAPDH antibody (1:30,000, Proteintech Group, Inc., United States), followed by incubation with the appropriate secondary antibodies at room temperature for another 1 h. Ultimately, all the blots were visualized by SageCapture software (Beijing Sage Creation Science Compony, China), the levels of protein expression were normalized to that of GAPDH, and relative protein expression was quantified by utilizing Image-ProPlus 6.0 software (Media Cybernetics, United States). Western blots were performed at least three times.

Given the limitations of H9c2 cardiomyocytes, neonatal rat cardiomyocytes (NRCMs), considered common models for studying the morphological, biochemical, and electrophysiological characteristics of the heart (Chlopcikova et al., 2001), were obtained from 2-3 day-old Sprague–Dawley (SD) rats (Beijing Si Pei Fu Biotechnology, Certificate SCXK-2019-0010) after strict sterilization by the methodology used for isolation and cultivation in previous research, with some modifications (Sabri et al., 2003; Rafiq et al., 2006; Shukla et al., 2018). The apex of isolated heart tissue was digested repeatedly in the short term in a mixture of collagenase II (Biorigin Inc., China) and 0.25% trypsin (Gibco Life Technologies, China) with a magnetic stirrer at 37°C. The cells were incubated in DMEM, supplemented with 15% FBS and 1% penicillin-streptomycin for 1 h. There were fibroblasts adhering to the wall, and the supernatant was resuspended in 96-well plates at a density of 5 × 10⁵ cells/ml, while 100 μM 5-bromo-2-deoxyuridine (BrdU, Biorigin Inc., China) was added to the culture medium to inhibit fibroblast proliferation. These non-adherent cells were incubated at 37°C under humidified conditions of 5% CO₂ for 24 h, and the medium was replaced. On days 4 to 5 of culture, confluent monolayers of NRCMs with regular spontaneous contractility were used for the observation of cardiac toxicity induced by evodiamine and rutaecarpine (Frolova et al., 2016; Frolova et al., 2019).

To detect the influence of spontaneous contractility, the spontaneous beat frequency of the NRCMs was recorded after interventions with evodiamine (5, 10, 25 μM) and rutaecarpine (60, 80, 100 μM) for 15 min, 30 min, 1 h, 2 h, and 4 h, separately. Moreover, the cell viability and the LDH leakage of the NRCMs were detected using corresponding kits after 4 h of administration.

A hydrocortisone sodium succinate for injection (Tianjin Biochem Pharmaceutical Co., Ltd., China) was diluted with saline to a 20 mg/ml solution for use. The preparation of the 1.5 mg/ml thyroid suspension was made by dissolving oral thyroid tablets (Shanghai Zhonghua Pharmaceutical Co., Ltd., China) in carboxymethylcellulose sodium (CMC-Na, BioRuler Company, United States). In addition, the herbal materials called Euodiae Fructus Praeparata were purchased from Beijing Sanhe Pharmaceutical Co. Ltd (Beijing, China, Lot 12410101), and authenticated by Prof. Chunsheng Liu, Beijing University of Chinese Medicine, as the fruit of Tetradium ruticarpum (A. Juss.) T. G. Hartley. The decoction of Euodiae Fructus was boiled twice; 1 kg decoction pieces were decocted with water (1:10 volume) for 45 min the first time, before eight times the amount of water was added for another 30 min. Finally, the supernatants were combined and concentrated into a 0.525 g/ml decoction of Euodiae Fructus.

Adult male SD rats weighing 180 ± 10 g (Beijing Si Pei Fu Biotechnology, Certificate SCXK-2020-0033) were acclimatized for 3 days in the animal facility at Beijing University of Chinese Medicine. The rat models of Yang deficiency and those of Yin deficiency were gavage administered the with the decoction of Euodiae Fructus, whose potential cardiotoxicity was investigated to delineate the signaling mechanisms in vivo. All the animal experiments were conducted in accordance with approved guidelines specified by the animal ethics committee of Beijing University of Traditional Chinese Medicine (Beijing, China; No. BUCM-4-2021090302-3052).

The manufacture of rat models with Yang deficiency was achieved by an intragluteal injection of 20 mg/ml hydrocortisone sodium succinate (1 ml/kg), continued for 15 days, as in the previous work of our team (Zhang, 2013). Meanwhile, the rat models with Yin deficiency received gavage administration of 1.5 mg/ml thyroid suspension (10 ml/kg) for 15 days (Zhang et al., 2019). Simultaneously, the medication group received intragastric administration of the decoction of Euodiae Fructus (the low dose was 0.0583 g/ml, the high dose was 0.525 g/ml), based on the modeling of Yang and Yin deficiencies.

All rats were randomly divided into eight groups (n = 8/group): 1) the Yang-K group (treated with intragluteal injection of an equal volume of the sterilized saline); 2) the Yang-X group (received intragluteal injection of hydrocortisone sodium succinate 1 ml/kg); 3) the Yang-D group (administered intragluteal injection of hydrocortisone sodium succinate 1 ml/kg + the decoction of Euodiae Fructus 0.0583 g/ml); 4) the Yang-G group (administered intragluteal injection of hydrocortisone sodium succinate 1 ml/kg + the decoction of Euodiae Fructus 0.525 g/ml); 5) the Yin-K group (received gavage administration of water); 6) the Yin-X group (given gavage administration of thyroid suspension 10 ml/kg); 7) the Yin-D group (received gavage administration of thyroid suspension 10 ml/kg + the decoction of Euodiae Fructus 0.0583 g/ml); and 8) the Yang-G group (received gavage administration of thyroid suspension 10 ml/kg + the decoction of Euodiae Fructus 0.525 g/ml).

The changes in the general status of different groups were observed immediately; the body weights and rectal temperatures of rats were measured 7 days and 14 days after treatment.

All rats per group were sacrificed on day 15 by anesthetization with an intraperitoneal injection of 10% chloral hydrate (3 ml/kg). After anesthesia, the rats were fixed in a supine position, and the ECG was recorded through a BL-420S biological function experiment system (Chengdu Taimeng Software Co., Ltd., China) to inspect the cardiac function.

Blood was collected from the abdominal aorta for different detection indexes, for which the plasma, serum and whole blood were prepared separately. To explore the cardiac injury, serum biomarkers, including lactate dehydrogenase (LDH), creatine kinase (CK), α-hydroxybutyrate dehydrogenase (HBDH), and aspartate aminotransferase (AST), along with the glucose and lipid metabolism involving glucose (GLU), triacylglycerol (TG), and cholesterol (CHO), were detected using the AU5800 automatic biochemical analyzer (Beckman Coulter, Inc., United States). With regard to the organ coefficients, the organs (including liver, kidneys, heart, spleen, and lungs) of each rat were dissected and weighed, and the hearts were removed for subsequent experimentation.

The content of triiodothyronine (T3) and the thyroid stimulating hormone (TSH) in serum, and cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) in plasma, were determined using related enzyme-linked immunosorbent assay (ELISA) kits (Wuhan Cloud-Clone Corp., China) in accordance with the manufacturer’s instructions. Using the hematology analyzer (Sysmex Corporation, Japan), the routine blood test was conducted on blood samples collected from rats in the different groups, measuring especially white blood cells (WBC), red blood cells (RBC), hemoglobin (HGB), platelets (PLT), neutrophil ratio (NEUT%), lymphocyte ratio (LYMPH%), and monocyte ratio (MONO%).

Normal saline was used for irrigating the hearts. Afterward, the hearts were fixed in 10% formalin for 24 h, embedded in paraffin, and sectioned transversely at 4 µm. The histopathological changes of myocardia for rats in different groups were investigated by haematoxylin-eosin (HE) staining.

Total proteins from the heart tissue of rats were homogenized and extracted, and the expression of PKG protein was examined through western blot analysis, according to related procedures of in vitro experimentation.

For the group with significant cardiotoxicity, the endogenous metabolites in the serum were investigated using the approach of ultra-high performance liquid chromatography quadrupole-exactive Orbitrap/mass spectrum (UHPLC-Q-Exactive Orbitrap/MS), as described previously (Liu et al., 2019).

Briefly, aliquots (100 μl) of plasma samples were mixed with 300 μl chromatographic acetonitrile. After centrifugation (13,000 rpm, 15 min, 4°C), the supernatant was transferred to a clean tube for analysis. For methodological investigations, the quality control (QC) samples were prepared from mixtures of 10 μl plasma in each sample.

Aliquots (2 μl) of experimental samples were eluted through an ACQUITY UPLC BEH C18 chromatographic column (2.1 mm × 100 mm, 1.7 µm, Waters Corporation, United States) in a Vanquish Duo UHPLC chromatograph (Thermo Fisher Scientific Inc., United States), using the mobile phases of eluents A (acetonitrile) and B (0.1% formic acid in water) at a flow rate of 0.3 ml/min.

Electron spray ionization was employed for detecting both positive and negative ions in the abovementioned plasma samples via a hybrid quadrupole Orbitrap mass spectrometer (Q Exactive, Thermo Fisher Scientific Inc., United States). The quadrupole scan range was set at mass-to-charge ratio (m/z) 100–1,200 Da, with the heated capillary temperature at 350°C, and the positive and negative spray voltages at 3.2 and 3.8 kV, respectively.

The raw data from the liquid chromatography-mass spectrometry (LC-MS) were manually phase-baseline corrected for peak area (PA) and retention time (RT) using the Mass Spectrometry-Data Independent Analysis software version 4 (MS-DIAL 4, http://prime.psc.riken.jp/compms/msdial/main.html) (Tsugawa et al., 2019; Tsugawa et al., 2020). Thereafter the multivariate data analysis was performed with SIMCA-P software (Version 14.1, Umetrics, Umea, Sweden), including principal component analysis (PCA) and the orthogonal partial least square-discriminate (OPLS-DA). Here PCA was a non-supervised approach to observe the distribution and outliers of the data set depicted in a scores plot based on orthogonal latent variables, which were obtained from the overall direction of maximum variance (Duan et al., 2018). Furthermore, owing to supervised algorithms, OPLS-DA was employed to extract the underlying variability in behavior characterizing the endogenous metabolomics. The evaluation methods of the OPLS-DA model were described by the Q² and R ² of the permutation respectively. The robustness of the model’s prediction ability is directly proportional to the Q² (0 < Q² < 1), while the R ² could represent the percentage of X and Y matrix information of the model interpretation (Triba et al., 2015; Li et al., 2018b; Jang et al., 2018; Plazas et al., 2019).

The most discriminant variables were selected in terms of variable importance in the projection (VIP) with significant statistical difference in the corresponding PA. On the one hand, discriminant metabolites (VIP >1.0) were collected according to related results of OPLS-DA. On the other, the statistical tests were exhibited by SPSS software. The normality of data, considered as an adjusted p-value > 0.05, was determined by a Kolmogorov–Smirnov test for each group. With regard to normal and homoscedastic variables, statistical significance was determined using a one-way ANOVA. Otherwise, the differences between groups were determined using the Kruskal–Wallis test, and the significance was considered as a p-value below 0.05.

Subsequently, corresponding metabolites were identified according to the Human Metabolome Database (HMDB, http://www.hmdb.ca/) (Wishart et al., 2018). As directly displayed in heatmaps for the content and correlation of identified metabolites, the cluster analysis was constructed using MetaboAnalyst 3.0 (http://www.metaboanalyst.ca/) (Chong et al., 2019), and the results of the pathway analysis for the differential metabolites in rats were visualized in a bubble chart, with the size of the bubble proportional to the importance of the pathway (Chong et al., 2018; Chong and Xia, 2020).

Compared with the control group, both evodiamine and rutaecarpine presented inhibitory effects in a dose-dependent manner for the cell viability of H9c2 cardiomyocytes. The IC₅₀ values of evodiamine and rutaecarpine separately were 42.82 ± 7.55 and 117.97 ± 9.69 μmol/L, and the related details are summarized in Table 1.

According to the results of Figure 2; Table 2, the leakage of LDH and the activity of CK were notably more highly dose-dependent in the high-dose evodiamine and rutaecarpine group than in the control group (p < 0.01). Similarly, the intensity of calcium fluorescence for H9c2 cells in the high-dose evodiamine and rutaecarpine group was obviously higher (p < 0.05). However, significant differences were only observed in the evodiamine-induced H9c2 cells compared to the control group (p < 0.01). These results indicate that evodiamine and rutaecarpine might change the permeability of the myocardial cell, the activity of the myocardial enzyme, the energy supply, and the calcium concentration, thereby inducing cardiotoxicity of H9c2 cardiomyocytes.

As presented in Figure 3, cGMP and PKG were downregulated in the H9c2 cardiomyocytes with evodiamine (5–25 μmol/L) and rutaecarpine (80–100 μmol/L), compared with the control group (p < 0.05), suggesting that the gene and protein expression levels of cGMP and PKG were significantly decreased in H9c2 cardiomyocytes under evodiamine and rutaecarpine (Supplementary Material).

The cell viability of each group was apparently lower than in the non-medication group (p < 0.05). Additionally, compared with the PKG drug G1 group, only the 60 μmol/L rutaecarpine group was without significant inhibition of H9c2 cardiomyocytes, which means the combination of the PKG drug G1 with evodiamine or rutaecarpine could not have had an appreciable effect on the cell viability of H9c2 cardiomyocytes (Figure 4, Supplementary Material).

As shown in Figure 4, the PKG drug G1 could significantly reduce the leakage of LDH in the low-dose evodiamine and rutaecarpine groups of H9c2 cardiomyocytes, compared with the single agent group (p < 0.05). Meanwhile, treatment of the PKG drug G1 obviously improved the mitochondrial membrane potential in the group of 80 μmol/L rutaecarpine (p < 0.05), and there were no significant differences for the activity of CK and the intensity of calcium fluorescence between the combined group and the single agent group (Table 3). These results indicate that the PKG drug G1 might partially decelerate the cardiotoxicity of H9c2 cardiomyocytes caused by evodiamine and rutaecarpine.

As demonstrated by western blot analysis (Figure 4, Supplementary Material), compared with single compound groups, there was an increasing trend of protein expression of PKG in compatibility groups. Remarkably, the PKG drug G1 could greatly enhance the expression of PKG for H9c2 cardiomyocytes incubating with 80 μmol/L rutaecarpine (p < 0.05). The inhibitory effects of rutaecarpine (80 μmol/L) were antagonized in concentration-dependent ways by treatment with the PKG drug G1 at concentrations of 5 mol/L.

During the entire experiment in vivo, the weight of YANG-X, YANG-D, and YANG-G groups gradually decreased compared to the YANG-K group, while the YANG-D and YANG-G groups’ rectal temperatures increased compared with YANG-K and YANG-X, as presented in Figures 5A,B (p < 0.05) (7 days, 14 days, Supplementary Material). In the model of rats with Yin deficiency, there were significant differences in weight and rectal temperature following oral administration of Euodiae Fructus compared with the control group. The changes in general status demonstrate that the treatment of Euodiae Fructus can affect the physical status of rats with Yang or Yin deficiencies, resulting in weight loss and temperature elevation.

It was noticed that long-term exposure to Euodiae Fructus might induce changes in ECG for rats with Yang or Yin deficiency to different degrees. In particular, significant differences in heart rate, PR interval, QT interval, P-wave amplitude, R-wave amplitude, and ST-wave amplitude were observed in the high-dose groups compared with the corresponding control group and model group (Figures 5C–H, Supplementary Material). Namely, long-term and overdose exposure to Euodiae Fructus could cause ECG abnormalities for rats with Yang or Yin deficiencies, including marked prolongation of the ventricular depolarization period and shortening of the effective refractory period, hence disturbing the atrioventricular conduction, which could lead to cardiac arrhythmia.

The results of the alteration in serum myocardial enzymes manifest that the levels of LDH, CK, HBDH, and AST increased in the YANG-G group over the corresponding control group, with a statistically significant difference (p < 0.05). Similarly, a remarkable rise of HBDH in the YIN-D group, and LDH, CK, HBDH, and AST in the YIN-G group were also observed over the corresponding control and model groups (p < 0.05) (Table 6). Therefore, overdosage and unsuitable syndrome differentiation are associated with the elevation of myocardial enzymes induced by Euodiae Fructus in rats.

To assess whether Euodiae Fructus involves changes to the glycolipid metabolism of rats with Yang or Yin deficiency, levels of GLU, TG, and CHO were measured in rats exposed to Euodiae Fructus decoction for 14 days. As summarized in Table 6, the high-dose gavage administration for rats with Yang deficiency resulted in significantly changed GLU, TG, and CHO levels compared to the related model groups (p < 0.05), while rats with Yin deficiency indicated obvious disorders in GLU, TG, and CHO levels compared to the related control groups (p < 0.05). Euodiae Fructus could thus contribute to clinical efficacy for rats with Yang deficiency and metabolic abnormality for those with Yin deficiency.

According to the results of the organ coefficients in Table 6, the obvious differences of heart and kidney were not observed among different groups; however, there was a higher level of liver coefficient in groups of high-dose Euodiae Fructus (p < 0.05). The results reveal that an overdose of Euodiae Fructus might contribute to hepatic damage in rats, whether with Yang deficiency or Yin deficiency.

Aside from changes in glycolipid metabolism, rats with Yang or Yin deficiency also possessed differing content of T3, TSH, cAMP, and cGMP. The level of T3 in the YIN-G group was significantly higher than the corresponding control and model group. Hence, the intervention of Euodiae Fructus could increase cAMP/cGMP in rats with Yin deficiency significantly more than in the related control group (p < 0.05) (Table 7). These results suggest that an imbalance of hormone secretion and second messenger transcription might occur due to the irrational usage of Euodiae Fructus.

In addition, the levels of WBC, HGB, and NEUT% in the YANG-G group; of RBC, HGB, PLT, and NEUT% in the YIN-D group; and of WBC, RBC, HGB, PLT, NEUT%, and MONO% in the YIN-G group, were all different from the related model group with a statistically significant difference (p < 0.05), indicating continuous gavage with an overdose of Euodiae Fructus for 15 days could influence the blood routine levels of rats with Yang or Yin deficiencies.

As displayed in Figure 6A, obvious histological changes were not observed in the cardiac tissues of the YANG-K group, the YANG-X group, and the YANG-D group, as the cardiac muscle fibers were arranged neatly without inflammatory cell infiltration. In the YANG-G group, some myocardial fiber underwent hypertrophy and became uneven. In the YIN-D group, some cellular edema, break or necrosis, and obvious infiltration of inflammatory cells could be observed. Furthermore, pathological examination revealed that the myocardial fibers were in a disordered condition for the YIN-G group: the major lesions in the myocardial fibers were from degeneration and necrosis, inflammatory infiltration, and edema. These results establish that cardiac pathological injury in rats is associated with overdosage and unsuitable syndrome differentiation of Euodiae Fructus.

The inhibitory effects of Euodiae Fructus for the protein expression of PKG were concentration-dependent in rats with Yin deficiency, while the protein expression of PKG in heart issue was also lower in the YANG-G group than in the corresponding control and model groups, and statistically significant differences were observed among these groups (Figure 6B, Supplementary Material).

The untargeted metabolomics of cardiotoxicity induced by Euodiae Fructus in rats with Yin deficiency were evaluated; the serum samples of the YIN-K and YIN-G groups were determined using UHPLC-Q-Exactive Orbitrap/MS. According to the results of the multivariate data analysis in Figure 7, there was clear separation between the YIN-K and YIN-G groups, suggesting the metabolic profile might be different after continuous gavage of Euodiae Fructus for 15 days, and the details of PCA, OPLS-DA, and permutations are shown in the Supplementary Material.

Based on the limitation of the variables with VIP>1 and simultaneous significant difference, ultimately there were 3212 endogenous metabolites in total, with 2060 (64.13%) in the positive ion mode, and the remaining in the negative mode (35.87%). After the identification, 34 corresponding metabolites were highlighted as the most discriminant in the rats of the YIN-K and YIN-G groups, including D-proline, deoxycytidine, 5-hydroxyisourate, cytosine, uric acid, D-lysine, and so on (Supplementary Material).

The cluster analysis depicted in Figure 8A reveals that these discriminant metabolites were divided into two categories in a dendrogram, and there was close correlations or similar pathways for the metabolites in the same category. Furthermore, the results of the pathway analysis also pointed out that 10 metabolic pathways, including the purine metabolism, glycerophospholipid metabolism, glycerolipid metabolism, sphingolipid metabolism, and the phosphatidylinositol signaling system, as well as the arginine and proline metabolisms, were all strongly involved in the metabonomic signatures of rats exposed to Euodiae Fructus. This could induce cardiotoxicity in rats with Yin deficiency, and the most likely metabolic pathways and related discriminant metabolites are exhibited in Figures 8B,C and in the Supplementary Material.