Metabonomics coupled with integrated approach reveals the therapeutic effect of higenamine combined with [6]-gingerol on doxorubicin - induced chronic heart failure in rats

The combination of Aconiti Lateralis Radix Praeparata (ALRP) and Zingiberis Rhizoma (ZR) is one of the most typical representatives reecting the very essence of the theory of Chinese material medica compatibility, which has been used to treat cardiovascular disease for many years. Previously, we demonstrated that ALRP-ZR prevented doxorubicin (DOX)-induced chronic heart failure (CHF) in vivo. However, its active components are still unclear. This study was aimed to investigate the therapeutic effect and potential mechanism of higenamine combined with [6]-gingerol (HG/[6]-GR) against doxorubicin (DOX) - induced chronic heart failure (CHF) in rats. Therapeutic effects of HG/[6]-GR on hemodynamics indices, serum biochemical indicators, histopathology and TUNEL staining of rats were assessed. Moreover, a UHPLC-Q-TOF/MS-based serum metabonomic approach coupled with biochemical assay has been performed to identify the potential mechanisms of HG/[6]-GR on DOX-induced CHF. HG/[6]-GR had effects on promoting of hemodynamic indices, decreasing serum biochemical indicators, and alleviating histological damage of heart tissue. Serum metabonomics analyses indicated that the therapeutic effects of HG and [6]-GR were mainly associated with the regulation of eight metabolites and twelve pathways, which may be responsible for the therapeutic ecacy of HG/[6]-GR. Moreover, the results showed that HG/[6]-GR could substantially regulate the expression level of energy metabolism-related metabolites and pathways. Multivariate statistical analysis has provided new insights for understanding CHF and investigating the therapeutic effects and mechanisms of HG/[6]-GR, which inuencing the metabolites related to energy metabolism pathway. component analysis; OPLS-DA: Orthogonal partial least-squares discriminant analysis.

and multi-channel. Our previous studies have shown that the compatibility use of higenamine (HG, one of the active compounds of Aconiti Lateralis Radix Praeparata, ALRP) and 6-gingerol ([6]-GR, one of the active compounds of Zingiberis Rhizoma, ZR) inhibits doxorubicin (DOX)-induced CHF via promoting mitochondrial energy metabolism [6]. However, the potential mechanism of HG combined with [6]-GR (HG/ [6]-GR) for the treatment of CHF had not been comprehensively elucidated. It remains to be elucidated how HG/ [6]-GR can prevent and treat CHF by affecting mitochondrial energy metabolism.
Metabonomic is a comprehensive and systematic study of small molecule metabolites in biological samples or organs [7]. It can characterize changes of endogenous metabolites and their organic relations with physiological and pathological phenotypes after disturbance [8]. The metabolism of organism changes its dynamic balance due to the occurrence of disease, so it is helpful to understand the metabolic mechanism of organism by analyzing the composition of body uid through metabonomics and obtaining biomarkers changed by disease induction. One of the basic methods of metabonomics research is the combination of advanced modern analytical technology, pattern recognition and expert system [9]. In recent years, metabonomics has also been used to identify speci c biomarkers and evaluate the role of TCM in various diseases [10][11][12]. Therefore, the objective of current study was to use serum metabonomics analysis accompanied by biochemical and histopathological approaches to investigate and verify the metabolic pro les of blood metabolite spectrum caused by the development of CHF, as well as the treatment of HG/[6]-GR. Simultaneously, this study was expected to reveal the anti-CHF mechanism of HG/ [6]-GR in rats. given DOX hydrochloride injection in the doses of 2.5 mg/kg body weight twice a week for six times. Thus, the accumulative doses of DOX was 15 mg/kg body weight [13][14][15]. Hemodynamic indices were comprehensively assessed by a RM6240 multi-channel physiological signal acquisition system (Chengdu Instrument Factory, Sichuan, China) [16][17][18][19]. When the values of +dp/dt max were reduced to 50% of the control group, CHF model was successfully prepared.

Materials
Thirty-two rats with successfully prepared CHF model were randomly assigned into ve groups of eight rats in each group: DOX group, DH positive group (50 μg/kg/d), HG group (5 mg/kg/d), [6]-GR group (5 mg/kg/d), and HG/[6]-GR compatibility group (10 mg/kg/d). Eight rats in the control group were given the same volume of normal saline. All rats were intraperitoneally injected with corresponding drugs once a day for seven consecutive days. It should be noted that CHF rats intraperitoneally injected with 5 mg/kg/d HG and [6]-GR showed a good therapeutic effect in our previous study [20]. Hemodynamic indices were assessed after the nal injection. All animals were sacri ced to collect serum samples and cardiac tissues for pharmacodynamic and serum metabonomic analysis.

Detection of pharmacodynamic indices
Serum biochemical indices, including BNP, NT-proBNP, LDH, CK-MB, and AST were determined on a Synergy hybrid reader (Biotek, Winooski, USA). In addition, the serum energy metabolism-related indices, including ATP, ATPase, NAD, NADH were also detected. BNP, NT-proBNP, LDH, and CK-MB were purchased from Shanghai MLBIO Biotechnology Co., Ltd. AST was obtained from Nanjing Jiancheng Bioengineering Institute. ATP, ATPase, NAD, and NADH were purchased from Shanghai Kanglang Biotech Co., Ltd. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was performed to indicate the cytotoxicity, cell damage and its recovery.

Preparation of serum metabonomics samples
Firstly, the serum samples of rats were thawed at 4°C condition. 200 μL of the serum was mixed with 600 μL of methanol to precipitate the protein. After centrifugation at 13,800 g, 4℃ for 10 min, the supernatant was transferred into a polypropylene tube and ltered via a syringe lter (0.22 μm) for obtaining the injection sample. Simultaneously, to assess the stability and reproducibility of serum metabonomics samples, the quality control (QC) sample was prepared by mixing all individual samples with10 uL aliquots in each.

Chromatography analysis
The serum samples were measured on an Agilent 1290 series UHPLC system (Agilent Technologies, Santa Clara, USA) coupled with a ZORBAX RRHD 300 SB-C18 column (100 × 2.1 mm, 1.8-μm, Agilent Technologies, Santa Clara, USA) for chromatography and separation. During the analysis, the setting conditions were set as follows: sample maintaining temperature, 4 °C; injection volume: 4 μL; column temperature: 30 °C; ow rate, 0.30 mL/min. The mobile phases were composed as solvent A (0.1% formic acid in acetonitrile), and solvent B (0.1% formic acid in water). The gradient elution was set as Table 1.  Data processing and multivariate data analysis After statistical analysis by MetaboAnalyst 4.0 (http://www.MetaboAnalyst.ca/) [21], the raw data were converted into "data_normalized.csv" format. Then, the normalized le in positive mode and negative mode were imported into the SIMCA-P program (version 14.1, MKS Umetrics) for multivariate analysis, respectively. Principal component analysis (PCA) was performed after concentration and normalization to check the overall metabolism of each sample group, and observe sample aggregation, dispersion and abnormal values. Next, orthogonal partial least-squares discriminant analysis (OPLS-DA) was used to identify the main difference variables that caused the aggregation and discretization. Subsequently, 100 iteration permutation tests were performed to avoid the over-tting of OPLS-DA. Potential biomarkers were selected according to the parameters of variable VIP > 1 and |Pcorr| > 0.58 from OPLS-DA. SPSS 23.0 software with the t-test was used to test the peak areas of differential metabolites and determine the differences of biomarkers between groups (p-value threshold was set at 0.05).

Potential metabolites identi cation and pathway analysis
Furthermore, a MassHunter Pro nder software (version B.06.00, Agilent, California, USA) was used to analyze the sample data for peak detection and alignment. Full scan mode was employed and the mass range was 80 to 1000 m/z. The online biochemical database HMDB database (http://www.hmdb.ca/) and METLIN (http://metlin.scripps.edu/) were used to identify the potential metabolites. MetaboAnalyst 4.0 was used for the pathway analysis. Finally, to identify and visualize the affected metabolic pathways, the biomarkers were put into MetaboAnalyst 4.0 based on the pathway library of Rattus norvegicus (rat).
In the present study, the bioactive components, possible biomarkers and potential mechanisms of HG and [6]-GR in the treatment of CHF induced by DOX were comprehensively elucidated using the serum metabonomics strategy.

Statistical analysis
All data were analyzed using SPSS 23.0 software program (Chicago, United States) and GraphPad Prism 8.2.0 software (GraphPad Software). The differences of data between groups were assessed by one-way analysis of variance (ANOVA). Values in the text were presented as mean ± SD. P < 0.05 was considered statistically signi cant. P < 0.01 was considered highly signi cant.

Hemodynamics indices
The therapeutic effects of HG/[6]-GR on heart function was evaluated by assessing the hemodynamics indices. Compared with the control group, DOX could substantially decrease the LVSP and +dp/dt max value while signi cantly increase the LVEDP and -dp/dt max value, indicating that the model of CHF was successfully prepared. However, compared with the DOX group, DH, HG, [6]-GR, and HG/[6]-GR could dramatically increase the levels of LVSP and +dp/dt max and decrease the LVEDP and -dp/dt max value.  Notes: Compared with the control group, ** P < 0.01; compared with the DOX group, ## P < 0.01; compared with the HG group, a P < 0.05, aa P < 0.01; compared with the [6]-GR group, bb P < 0.01.

Myocardial biomarkers
Serum levels of myocardial biomarkers were included in Figure 1. Compared with the control group, serum levels of BNP, NT-proBNP, LDH, CK-MB, and AST in the DOX group were signi cantly increased (P 0.01) while serum levels of ATP, ATPase, NAD, and NADH were decreased in the DOX group, which indicated the damage of heart function and energy metabolism disorder in DOX group. However, compared with the DOX group, HG and [6]-GR could reduce the serum concentrations of BNP, NT-proBNP, LDH, CK-MB, and AST, but increase the serum levels of ATP, ATPase, NAD, and NADH. Notably, these biomarkers were substantially changed in DH and HG/[6]-GR (P 0.01) group compared with the DOX group. Furthermore, HG/[6]-GR group was almost equal to the DH group, which markedly decreased the serum levels of BNP, NT-proBNP, LDH, CK-MB, and AST but increased the levels of ATP, ATPase, NAD, and NADH compare with HG or [6]-GR used alone (P 0.05, P 0.01). Thus, [6]-GR might enhance the therapeutic role of HG in the treatment of CHF.

Histopathological changes
The histopathological results showed the degree of damage in each group. After administration, the rats in DOX group had pathological changes such as widening and breaking of myocardial tissue space, vacuolar degeneration, edema, and necrosis of myocardial cells ( Figure 2B). Compared with the DOX group, the histopathology of HG and [6]-GR group was improved, but some rats still had widened and broken myocardial tissue space, vacuolation and degeneration of myocardial cells ( Figure 2D, 2E), while DH and HG/[6]-GR group showed signi cant improvement in cardiac pathology, less vacuolation, edema, necrosis, atrophy and other pathological changes of myocardial cells ( Figure 2C, 2F).
Detection of cardiomyocyte apoptosis TUNEL staining was used to detect the therapeutic effect of HG/[6]-GR on DOX induced cardiomyocyte apoptosis and its recovery. As shown in Figure 3, compared with the control group, the TUNEL positive proportion of cardiomyocytes in DOX treatment group increased signi cantly, indicating that DOX could cause cardiomyocyte apoptosis. In contrast, HG and [6]-GR used alone could reduce the apoptosis rate of cardiomyocytes in varying degrees. Moreover, HG/[6]-GR had a signi cant inhibitory effect on cardiomyocyte apoptosis, indicating that HG combined with [6]-GR had a synergistic anti-apoptotic effects. These results showed that HG/[6]-GR had a signi cant protective effect on CHF myocardial tissue. The results clearly showed that the HG/[6]-GR could effectively alleviate CHF in rats.

Metabolic pro le analysis
The metabolic pro le analysis of serum samples was performed using UHPLC-Q-TOF/MS both in the positive and negative electrospray ionization (ESI) modes. PCA analysis was performed to assess alterations in the metabolome of each group. In the PCA score plot ( Figure 4A, 4B), the control groups and DOX groups were clearly divided into two clusters. In addition, the HG/[6]-GR and HG groups were signi cantly separated from DOX group and closer to the control group, especially the HG/[6]-GR group. Furthermore, to maximize the difference of metabolic pro les, OPLS-DA analysis was carried out subsequently (Figure 4 C, D). The results showed that the OPLS-DA models were veri ed by the class permutation and all these models had predictive ability with an R 2 Y (cum), and Q 2 (cum). The corresponding value had been marked in the Figure 4C-4F. The R 2 Y (cum) and Q 2 Y (cum) were 0.999 and 0.992 in ESI+ mode, 0.998 and 0.971 in the ESI-mode, respectively. Also, the OPLS-DA model was performed based on the DOX and HG/[6]-GR group ( Figure 4E, 4F). The DOX group could be clearly separated from the HG/[6]-GR group. The R 2 Y (cum) and Q 2 Y (cum) were 1 and 0.99 in the ESI+ mode, 0.997 and 0.963 in the ESI-mode, respectively. In addition, metabolic pro le analysis between the DOX and HG or [6]-GR group in the positive mode and negative mode was also performed ( Figure 5A-5D). Scatter plots of the control and DOX group, DOX and HG/[6]-GR group were shown in Figure 5E-5H, and scatter plots of the DOX and HG group, DOX and [6]-GR group were shown in Figure 5I-5L.
Identi cation and quanti cation of potential biomarkers Next, differential metabolites in CHF treatment were identi ed. The variables that substantially contributed to the clustering and identi cation were identi ed when their VIP values ≥ 1.0 and |p(corr)| values ≥ 0.58 in S-plots. Finally, eight potential metabolites were expressed at signi cant levels and identi ed as biomarkers for the treatment of CHF. The basic information of these potential biomarkers was summarized in Table 3 with their corresponding name, formula, mass (m/z), retention time (min), and ratio changes (signi cance). Next, the mechanism of action of HG/[6]-GR on DOX induced CHF and the changes of eight possible metabolites were assessed and discussed. Compared with the control group, DOX substantially decreased peak area of acetylphosphate, 3-carboxy-1-hydroxypropylthiamine diphosphate, coenzyme A, PE(O-18:1(1Z)/20:4(5Z,8Z,11Z,14Z)), oleic acid, anslysoPC(18:1(9Z)) ( Figure  6A-6E, 6G), but increase the peak area of PC(16:0/16:0) ( Figure 6F) and palmitic acid ( Figure 6H). Conversely, HG/[6]-GR could reverse these changes and decrease the peak area of PC(16:0/16:0) and palmitic acid. Notably, most of the metabolites indicated the formation of mitochondrial energy metabolism substrate. Besides, to determine the distribution and differences between groups, the clustering heat map and PCA were constructed based on the potential biomarker data ( Figure 7B). Overall, the results indicated that HG/[6]-GR had obvious therapeutic effects on DOX-induced CHF. Especially, the curative effect of HG/[6]-GR group was better than that of HG and [6]-GR used alone ( Figure 6).

Pathway analysis of CHF treatment
To explore the possible pathway of HG/[6]-GR and DOX intervention in CHF, the KEGG ID of endogenous metabolites was imported into the MetaboAnalyst 4.0 system for the pathway analysis and visualization. The results showed that CHF-related metabolites were responsible for energy metabolism pathway, including glycerophospholipid metabolism, biosynthesis of unsaturated fatty acids, fatty acid degradation, linoleic acid metabolism, alpha-Linolenic acid metabolism, glycosylphosphatidylinositol (GPI)-anchor biosynthesis, pantothenate and CoA biosynthesis, citrate cycle (TCA cycle), pyruvate metabolism, arachidonic acid metabolism, fatty acid elongation, fatty acid biosynthesis ( Figure 7A). The match status, p value, -log(p) and the impact of each metabolic pathway were listed in Table 4. In addition, the relationship among metabolic pathways and metabolites was shown in the Figure 7C. The recovery trend of metabolites showed that the therapeutic effect of HG/[6]-GR on heart was related to the above eight metabolic biomarkers and twelve metabolic pathways. These results were consistent with the biochemical parameters and histological examination.

Discussion
Cardiomyocyte energy metabolism, especially fatty acid and glucose metabolism, changes in CHF, and is considered to be a factor of heart function impairment in patients with HF [22]. Fatty acid β-oxidation is a process in which fatty acids decompose to produce ATP. In a series of steps of long-chain coenzyme A (COAs) entering mitochondria, COAs are converted into long-chain acyl coenzyme by carnitine palmitoyltransferase 1 (CPT1) [23]. Long chain acyl CoA can enter into β-oxidation of fatty acids. One acetyl CoA is generated from each cycle by this pathway as well as NADH and FADH2. The NADH and FADH2 produced by β-oxidation of fatty acids and the TCA cycling of the acetyl CoA are used by electron transport chain for producing ATP [24]. In addition, the metabolism of fatty acids is an important energy source under the conditions of hunger, starvation, infection and diabetic ketoacidosis. In the state of CHF, the mitochondrial fatty acids metabolism is signi cantly impaired. In return, inhibition of fatty acid metabolism can cause myocardial insu ciency [25,26].
The pharmacodynamic effects of HG/[6]-GR on CHF were systematically evaluated. Firstly, the multichannel physiological signal detection system was used to evaluate the CHF model. The results showed that + dp/dt max value had been reduced to 50% of the control group, indicating the successful preparation of CHF model. Secondly, the system was used to detect the therapeutic effect of HG/[6]-GR on CHF.
Surprisingly, HG/[6]-GR could signi cantly increase the + dp/dt max value of CHF rats, and their combination was similar to that of DH group. As serum BNP and NT-proBNP levels are the most widely used biomarkers in the diagnosis and treatment of HF, which are helpful for the diagnosis, differential diagnosis, risk strati cation, e cacy monitoring and prognosis evaluation of acute-HF (AHF) and CHF [27]. Serum LDH, CK-MB, and AST levels can be used to evaluate whether the myocardium is damaged [28]. These parameters were comprehensively detected in the current study. Our previous study has shown that HG in combination with [6]-GR can substantially increase the CPT-1 level decreased by DOX, which can relieve cardiomyocyte injury induced by DOX via regulating fatty acid metabolism in the TCA cycle based on cell metabonomics [20]. In the present study, serum metabonomics coupled with integrative pharmacology has further improved our understanding of the therapy of DOX induced CHF with HG/[6]-GR from several pivotal aspects. Furthermore, a UHPLC-Q-TOF/MS-based serum metabonomics approach was used to study serum metabolites changes in CHF. Moreover, we demonstrated the therapeutic effects of HG/[6]-GR against CHF in rats, which speci cally caused a signi cant restoration of their myocardial metabolic pro les. This alteration laid the foundation for further investigation into the key mechanisms of HG/ [6]-GR in the treatment of CHF. Eight metabolites were identi ed in the CHF treatment, including acetylphosphate, 3-carboxy-1-hydroxypropylthiamine diphosphate, coenzyme A, palmitic acid, PE(O-18:1(1Z)/20:4(5Z,8Z,11Z,14Z)), oleic acid, lysoPC(18:1(9Z)), and PC(16:0/16:0), which are distributed in twelve metabolic pathways. Most of the detected compounds are intermediates of energy metabolism. Among the changes of these potential metabolic pathways, the most obvious abnormality occurs in energy metabolism, which indicates that CHF is related to the disorder of energy metabolism in the heart. These nding are consistent with the previous studies [29][30][31]. Among these metabolites, acetylphosphate can phosphorylate biologically signi cant substrates in a way similar to ATP, promoting the origin of metabolism [32]. Coenzyme A is mainly involved in the metabolism of fatty acids and pyruvate, which can stimulate the tricarboxylic acid (TCA) cycle and provide 90% of the energy needed for the body's life [30]. Palmitic acid diets can cause lipotoxicity and energy metabolism imbalance in vivo and in vitro [33]. Speci cally, palmitic acid treatment can induce cardiomyocyte apoptosis, which is manifested by the appearance of apoptotic nucleus, the activation of caspase 3, the release of mitochondrial cytochrome C and the loss of mitochondrial cardiolipin [34]. Our results showed that DOX could substantially decrease the level of acetylphosphate and coenzyme A, but increase palmitic acid, indicating the damage to myocardial energy metabolism. Nevertheless, HG/[6]-GR could signi cantly reverse this change and affect the fatty acid metabolism and citrate cycle. As fatty acid metabolism is a notable mechanism for creating energy for the heart and a signi cant target for storing or creating energy for the heart [35,36], HG/[6]-GR may play a crucial role in the treatment of CHF by improving the energy metabolism function of myocardial mitochondria.
The comprehensive treatment has been advocated for thousands of years by TCM prescription, which is a special medical system to help ancient Chinese treat diseases. It is believed that multiple components of TCM can hit multiple targets and play a synergistic therapeutic effect [37]. Currently, researches have performed various of studies have con rmed that Chinese medicine can improve the symptoms of CHF in different degrees, and elaborated its mechanism of action. The compatibility of ALRP and ZR is commonly used in clinical practice in ancient and modern times. A number of studies have con rmed the objective truth of ALRP combined with ZR in the treatment of CHF from the perspective of mitochondrial energy metabolism, but the material basis and mechanism of its regulation of myocardial energy metabolism are still unclear. A various of studies have been done on the cardiotonic effect and potential mechanism of HG worldwide [38][39][40]. HG is a selective activator of beta2-adrenergic receptor, which plays a wide range of roles in blood vessels, bronchus and heart with positive inotropic effects [41,42]. [6]-GR is a novel AT1 antagonist, which can regulate blood pressure and enhance the heart function in the cardiovascular system [43]. In this study, HG from ALRP combined with [6]-GR in ZR were used to investigate the effect of HG/[6]-GR on serum metabolic markers of DOX-induced CHF rats, and to explore its possible mechanism from the perspective of metabonomics. From the perspective of mitochondrial energy metabolism, the potential mechanism of HG, [6]-GR and their compatibility in the treatment of CHF is more conducive to further study of ALRP, ZR and their compatibility in improving the cardiac function.
In this study, although the effectiveness and potential mechanism of HG combined with [6]-GR in the treatment of CHF have been elucidated by a comprehensive method, some limitations still exists: (a) this study indicates that HG/[6]-GR might play a role in the treatment of CHF by affecting myocardial energy metabolism, the gene and protein expression of related pathways have not been veri ed to con rm the target mechanism of HG and [6]-GR; (b) the present study only discussed the effect of HG and [6]-GR on several biomarkers of metabolic difference, but the speci c effect on other metabolites of tricarboxylic acid cycle remains to be elucidated; (c) in addition to myocardial energy metabolism, the causes of CHF include apoptosis and in ammation, whether HG and [6]-GR can play the role of CHF treatment through other channels remains to be studied; (d) the biomarkers of HG/[6]-GR affecting CHF have been discussed in serum metabolism level, how HG/[6]-GR affect these different biomarkers and what is the mode of action remains to be further studied. Therefore, although targeting mitochondrial energy metabolism is a promising strategy for the treatment of CHF, further studies are needed to con rm the potential bene cial effect of regulating these metabolic targets as a method for the treatment of CHF.