The Pseudotargeted Metabolomics Study on the Toxicity of Fuzi Using Ultraperformance Liquid Chromatography Tandem Mass Spectrometry

Fuzi is commonly used in traditional Chinese medicine. Clinical Fuzi poisoning cases have frequently been reported. Glycyrrhizae Radix is often used to alleviate Fuzi's toxicity. However, the poisoning mechanism of Fuzi and the detoxication mechanism of Glycyrrhizae Radix are still not clear. We identified the chemical components of Fuzi at different decoction times (0.5, 1, 2, 4, and 6 h) using ultrahigh performance liquid chromatography quadrupole time-of-flight mass spectrometry. A total of 35 compounds were detected in the Fuzi decoction, including diester alkaloids, monoester alkaloids, amino acids, phenolic acids, organic acids, glycosides, and sugars among others. The content of diester alkaloids (i.e., subaconitine, neoaconitine, and aconitine) in the Fuzi decoction decreased after 2 h of decoction time, while the content of monoester alkaloids (i.e., benzoyl aconitine and benzoyl subaconitine) reached a peak at 2 h. A total of 32 rats were randomly divided into four groups, including 8 cases in the low-dosage Fuzi decoction group A, 8 cases in the high-dosage Fuzi decoction group B, 8 cases in the Fuzi and glycyrrhizae decoction group C, and 8 cases in the control group D. The decoction was administered orally for 7 days. Then, a serum was obtained. The metabolites' changes were analyzed in serum metabolomics using liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Statistical analysis and pathway analysis were used to assess the effects of glycyrrhizae on the metabolic changes induced by Fuzi. The behavioral and biochemical characteristics indicated that Fuzi exhibited toxic effects on rats and their metabolic profiles changed. However, the metabolic profiles of the glycyrrhizae group became similar to those of the control group. These profiles showed that glycyrrhizae can effectively improve Fuzi poisoning rats. Our study demonstrated that the established pseudotargeted metabolomics is a powerful approach for investigating the mechanisms of herbal toxicity.


Introduction
Traditional Chinese medicine is generally considered natural and harmless [1][2][3]. Aconitum carmichaelii (called Fuzi in Chinese) has positive pharmacological effects on various diseases such as painful joints, collapse, rheumatic fever, bronchial asthma, syncope, edema, various tumors, and diarrhea [4][5][6]. Fuzi was first recorded in Shennong's Materia Medica. However, its medicinal application on a large scope was limited due to the high toxicity risk and narrow therapeutic range [7]. It is important to establish a standardized chemical method that ensures its safe use. e previous reports demonstrated that aconitine, hypaconitine, and neoaconitine were the main toxic components in Fuzi [8][9][10]. erefore, the strategy of reducing toxicity and increasing efficiency was developed to meet the needs of clinical applications. ere are a variety of chemical components in Fuzi, so we need to understand how to reduce the content of the toxic ingredients and increase the content of active ingredients.
A clear understanding of the mechanism of Fuzi at the selected decoction time is essential to evaluate its safety. Metabolomics is a comprehensive method for diagnosing diseases, discovering biomarkers, and identifying perturbed pathways [11][12][13][14]. Sui et al. investigated the mechanism of aconitine's potential neurotoxicity and nephrotoxicity using a UPLC-Q-TOF-based rat serum and urine metabolomics strategy [8]. Shen et al. evaluated the enhancement of the Glycyrrhizae Radix for the hepatic metabolism of hypaconitine using rat liver S9 [15]. Sun et al. developed an NMR-based metabolomics study of the effect of Glycyrrhizae Radix on the attenuation of toxicity in rats induced by Fuzi [16]. In another previous study, nontargeted metabolomics was applied, but it has the weakness of low sensitivity and poor reproducibility [17]. Targeted metabolomics has the advantage of high sensitivity and good reproducibility. However, it has the weakness of low coverage. erefore, pseudotargeted metabolomics is a promising tool for the high throughput elucidation of metabolic phenotypes [18].
First, nontargeted serum metabolites were used in full scan and data-dependent modes to generate as high of coverage serum metabolite profiling as possible. Secondly, serum metabolite ion pairs were constructed based on the characteristic fragmental ions and corresponding parent ions of the serum metabolites, which were monitored in an MRM mode in the pseudotargeted metabolomics method.
is method combines the advantages of nontargeted and targeted metabolomic approaches.
In a previous study, Luo et al. displayed the dynamic variation patterns of the Aconitum alkaloids in Fuzi during the decoction process [7]. However, there have been few comprehensive identifications of Fuzi components at different decoction times, as well as high sensitivity evaluation methods of the toxic mechanism and detoxification mechanism of Glycyrrhizae Radix.
In our study, we explored the relationship between the chemical components of Fuzi with the decoction time to identify the appropriate decoction time using ultrahigh performance liquid chromatography quadrupole timeof-flight mass spectrometry (UPLC-QTOF). Pseudotargeted rat serum metabolomics was used to evaluate the toxic mechanism of Fuzi and the detoxification mechanism of Glycyrrhizae Radix. is study was the first time where Fuzi poisoning was systematically interpreted, thus providing some information for the future clinical use of Fuzi.

Preparation of the Decoctions.
Fuzi (60 g) was immersed in 200 mL water for 30 min and was then boiled for 0.5, 1, 2, 4, and 6 h, respectively. e supernatant was collected by filtration and centrifuged for 15 min at 4000 g. It was then concentrated to a final volume of 20 mL. e preparation of the Glycyrrhizae Radix decoction was the same as that for Fuzi. Glycyrrhizae decoction containing 3 g raw material per mL was obtained.

Animal Experiment.
irty-two male Wistar rats (180-220 g) were supplied by the Southern Medical University Laboratory Animal Center and were allowed to acclimatize in cages for 1 week before the experiment. e rats were randomly divided into four groups (n � 8/group) as follows: oral gavage with Fuzi at a dose of 30 g/kg (group A), oral gavage with Fuzi at a dose of 60 g/kg (group B), oral gavage with Fuzi and glycyrrhizae at a dose of 60 g/kg (group C), and oral gavage of the same volume of water to healthy controls (group D) as the other three groups. All the groups were given intragastric administrations twice a day for 7 days.

Collection and Preparation of Serum Samples.
Serum samples were collected from the retro-orbital venous plexus 7 days after administration. en, 300 μL acetonitrile was added into 100 μL serum and vortex-mixed for about 5 min for protein precipitation. e mixture was centrifuged at 14000 g for 15 min. Finally, 5 μL aliquots of the supernatant were used for analysis.

Chromatography and Mass Spectrometry
2.5.1. Fuzi Decoction Analysis. Chromatography separation of Fuzi decoction was performed on a Waters T3 column (2.1 mm * 100 mm, 1.8 μm) using a SCIEX ExionLC AD UPLC system (CA, USA). e column temperature was maintained at 40°C, and the mobile phase consisted of 0.1% formic acid in water (Phase A) and acetonitrile (Phase B) at a constant flow rate of 0.3 mL/min. e injection volume was 5 μL.
e 45 min binary gradient elution conditions were optimized as follows: linear gradient from 5% to 10% B (0.5-5 min), 10% to 95% B (5-35 min), 95% to 5% B (40-40.1 min), and then the column was returned to its starting conditions of 5% B for 5 min to allow for column re-equilibration. SCIEX X500R QTOF mass spectrometry (CA, USA) was used to analyze the components of the Fuzi decoction.

2.5.2.
Pseudotargeted Metabolomics Analysis. Chromatography separation of serum samples was performed on a Waters C18 BEH column (2.1 mm * 100 mm, 2 Evidence-Based Complementary and Alternative Medicine 1.7 μm) using a SCIEX ExionLC AD UPLC system (CA, USA). e column temperature was maintained at 40°C, and the mobile phase consisted of 0.1% formic acid in water (Phase A) and acetonitrile (Phase B) at a constant flow rate of 0.4 mL/min. e injection volume was 10 μL. e 11 min binary gradient elution conditions were optimized as follows: linear gradient from 5% to 95% B (1-7 min), 95% to 5% B (9.5-9.6 min), and then the column was returned to its starting conditions of 5% B for 1.5 min to allow for column re-equilibration. SCIEX 4000 QTrap mass spectrometry (CA, USA) was used to detect the metabolites of serum samples.

Data Processing and Statistical Analysis.
e components of Fuzi were identified by searching the SCIEX commercialization database using SCIEX OS software. e serum samples were analyzed using MultiQuant 3.0.3 software (CA, USA). Multivariate statistical analysis was performed on MetaboAnalyst 4.0 (Xia Lab at McGill University, Montreal, QC, Canada). Partial least-squares discriminant analysis (PLS-DA) was used to model all features of the four groups. All the metabolites with a significance threshold that satisfies the corrected p value cut-off of 0.05 in one-way ANOVA were considered as potential biomarkers.

Identification of Fuzi Components. All components of
Fuzi were represented as chromatographic peaks. e parent ions and their product ions were obtained for structural identification. To illustrate the identification of components, we took the m/z 646.3233 (t R � 20.72 min) as an example to be described as follows. Its molecular formula was speculated to be C 34 H 47 NO 11  e fragmentation pattern of aconitine is shown in Figure 1. e components of Fuzi decoction were identified and are listed in Table 1.

e Comparison of Fuzi Components at Different Decoction times.
e Fuzi decoction times were 0.5, 1, 2, 4, and 6 h, respectively. e content of the components was compared at different decoction times, as can be seen in Table 2. Hypaconitine, mesaconitine, and aconitine are diester alkaloids with strong toxicity. In our study, their content increased from 0 h to 1 h of decoction time and decreased after 2 h of decoction time. e results showed that diester alkaloids were thermally unstable. e content of benzoylhypacoitine, benzoylmesaconine, hydroxypurine, adenine, adenosine, and ferulic acid increased from 0.5 h to 2 h of decoction time and decreased after 2 h of decoction time. e content of neoandrographolide, p-coumaric acid, trigonelline, higenamine, and tuberostemonine increased from 0.5 to 1 h of decoction time. Phenprobamate, leucine, L(+)-arginine, L-valine, L-tryptophan, caffeic acid, vanillic acid, citric acid, amber acid, succinic acid, L-malic acid, salidroside, salidroside, guanosine, maltopentaose, D-galactose, D-(+)-mannose, nicotinic acid, 6-methyl coumarin, neoandrographolide, methyl 4-hydroxybenzoate, and norcantharidin can all be detected at the decoction time of 2 h. erefore, the Fuzi decoction at 2 h was used for the metabolomics study.   en, multivariate statistical analysis was carried out. LPS-DA was applied to model the four groups. As shown in Figure 3, the control group was well separated from groups A and B, suggesting that metabolic perturbation occurred significantly in the Fuzi group. Group C was closer to the control group, suggesting that glycyrrhizae could reduce the Fuziinduced metabolic perturbation. As shown in Figure 4, the  Evidence-Based Complementary and Alternative Medicine heatmap based on the intensity levels of the metabolites among the four groups was used to clearly characterize the serum metabolites' profile. All differentiated metabolites (p < 0.05) in one-way ANOVA were selected. e one-way ANOVA plot is shown in Figure 5. e differential metabolites were hexadecanol, 2-hydroxyphenylacetate, 4hydroxyphenyl acetate, 16-hydroxypalmitate, docosanoic acid, hexadecanal, hexadecanoic acid, hexadecenoic acid, icosapentaenoic acid, citrate, linoleate, N-acetyl-L-citrulline, N-acetyl-L-leucine, octadecatrienoic acid, octadecenoic acid, oleamide, phytanate, fluorocyclohexadiene, glycochenodeoxycholate-7-sulfate, suberic acid, taurochenodeoxycholate, and tetradecanoic acid. Pathway analysis was used to explore the metabolic pathway related to Fuzi toxicity. Linoleic acid metabolism, biosynthesis of unsaturated fatty acids, fatty acid biosynthesis, and the citrate cycle were disordered after the oral gavage of Fuzi ( Figure 6).

Discussion
A simple, efficient, and sensitive method was established to identify the components of Fuzi decoction at different decocting times using the X500R QTOF system. A total of 35 compounds were found, including 3 diester alkaloids, 2 monoester alkaloids, 3 other alkaloids, 3 base compounds, 6 amino acids, 4 phenolic acids, 3 organic acids, 3 glycosides, and 3 sugars. ere were also 5 other categories examined (1 vitamin, 1 coumarin, 1 lactone, 1 ester, and 1 anhydride). e results showed that the diester alkaloids in Fuzi gradually increased and reached the peak in 1∼2 h and then decreased significantly. e monoester alkaloids also gradually increased and reached the peak at 2 h and then decreased significantly. Diester alkaloids with high toxicity are thermally unstable. ey can be transformed into monoester alkaloids and further transformed into other alkaloids. is suggests that Fuzi can effectively reduce toxicity after a decocting time of more than 2 h, thus providing useful information for the clinical use of Fuzi.
When used in clinical settings, Fuzi could cause cardiotoxins, neurotoxins, nausea, palpitations, dizziness, vomiting, arrhythmia, hypotension, asystole, shock, coma, and neuron apoptosis among others [19,20]. Aconitine, mesaconitine, and hypaconitine are the pharmacological and toxic components in Fuzi [21,22]. In order to explore the effect of Fuzi on serum metabolites, the Fuzi decoction at 2 h was used for the metabolomics study. In clinical settings, glycyrrhizae could alleviate the side effects of Fuzi. In this study, it was used to validate the potential biomarker related to the toxicity of Fuzi.
In our study, pseudotargeted metabolomics was used to investigate the effect of glycyrrhizae on Fuzi-induced toxicity. e pseudotargeted method combines nontargeted and targeted analysis, which has proven to be a high-quality and information-rich method [23]. From analyzing the metabolomics study, the levels of 22 differential serum metabolites became abnormal (as seen in the Supplementary Materials section (available here), where there is a box plot chart of 22 metabolites), and the metabolite profiles of 22 candidate biomarkers were obtained from the quantitative analysis of the subjects. e figure was obtained using GraphPad Prism, and the names of the metabolites are shown in the box plot. e box plot consists of the median (i.e., horizontal line) and the interquartile range, and the whiskers indicate the minimum and maximum values unless there are outliers, in which case the whiskers extend to a maximum of 1.5 times the interquartile range. e 17 serum metabolites, including tetradecanoic acid, taurochenodeoxycholate, suberic acid, phytanate, octadecatrienoic acid, N-acetyl-L-leucine, N-acetyl-L-citrulline, linoleate, icosapentaenoic acid, hexadecenoic acid, hexadecenoic acid, glycochenodeoxycholate-7-sulfate, docosanoic acid, isocitrate, 16hydroxypalmitate, 4-hydroxyphenyl acetate, and 2hydroxyphenylacetate, were upregulated in groups A and B. eir contents were significantly higher than that of group D. However, their content in group C was closer to that of group D. is indicated that they were all close to the normal level after glycyrrhizae intervention. e 5 serum metabolites, including oleamide, octadecenoic acid, hexadecanal, fluorocyclohexadiene, and 1-hexadecanol, were downregulated in groups A and B. eir contents were significantly lower than that of group D. However, the content in group C was closer to that of group D. It also indicated that they were all close to the normal level after glycyrrhizae intervention. erefore, the 22 differential metabolites were related to Fuzi. When Fuzi treatment was combined with the administration of glycyrrhizae, the concentrations of the 22 differential metabolites returned close to their normal levels. In previous studies, glycyrrhizae could delay the absorption of Fuzi or accelerate the metabolism of aconitine, mesaconitine, and hypaconitine to reduce the toxicity of Fuzi [24][25][26][27][28]. Other mechanisms may also be involved [29]. Sun et al. investigated the effect of glycyrrhizae in the attenuation  S_25  S_27  S_22  S_21  S_23  S_26  S_24  S_30  S_31  S_32  S_28  S_29  S_5  S_7  S_6  S_12  S_15  S_13  S_14  S_9  S_16  S_17  S_8  S_10  S_11  S_1  S_2  S_19  S_20  S_18 S_3 S_4 Figure 4: e hierarchical clustering heatmap of the potential biomarkers.
-log10 (p)  Evidence-Based Complementary and Alternative Medicine of toxicity in rats induced by Fuzi using the NMR-based metabonomics method [30]. In our study, after pathway analysis, many metabolic pathways, including linoleic acid metabolism, biosynthesis of unsaturated fatty acids, fatty acid biosynthesis, and citrate cycle, were seriously impacted by Fuzi. Glycyrrhizae could regulate the disrupted citrate cycle (i.e., the central metabolic energy pathway) and fatty acid metabolism.

Conclusions
In our study, a total of 35 components in Fuzi were identified using UPLC-QTOF. Pseudotargeted metabolomics was used to detect the effects of glycyrrhizae on Fuzi-induced toxicity in rats. e results showed that amino acids and organic acids were significantly altered by Fuzi administration in rats. Glycyrrhizae could mitigate these metabolic changes, indicating that glycyrrhizae administration could reduce the toxicity of Fuzi at the metabolic level. e toxicity of Fuzi could be reduced at more than 2 h of decoction time. Our results demonstrate that glycyrrhizae reduces toxicity at the metabolic level through a series of pathways, such as linoleic acid metabolism, biosynthesis of unsaturated fatty acids, fatty acid biosynthesis, and citrate cycle.

Data Availability
e datasets generated during the present study are available from the corresponding author on reasonable request.

Ethical Approval
All experiments were performed in accordance with institutional guidelines and approved by the Health Authorities and Ethics Committees of the Hospital of Traditional Chinese Medicine of Zhongshan.

Consent
All subjects signed the informed consent prior to being included in the study.

Conflicts of Interest
e authors declare that they have no conflicts of interest regarding this publication.