The Correlation between Chemical Composition, as Determined by UPLC-TOF-MS, and Acute Toxicity of Veratrum nigrum L. and Radix paeoniae alba

The eighteen incompatible medicaments is an important theory in traditional Chinese medicine. The theory suggests that drugs in the eighteen incompatible medicaments can be toxic when used together. Veratrum nigrum L. and Radix paeoniae alba belong to the eighteen incompatible medicaments and have been prohibited for thousands of years. This study offers preliminary insight into the mechanism and chemical constituents responsible for the incompatibility and toxicity of these two agents. Specifically, we performed toxicology studies to identify and quantify the constituent substances of the two agents. Experiments revealed that acute toxicity increases when the dose of V. nigrum L. is higher than, or equal to, RPA. UPLC-TOF-MS analysis showed that, although the volumes of V. nigrum L. were the same, the content of some veratrum alkaloids changed significantly and had a trend toward a highly positive correlation (|r| ≥ 0.8) with toxicity. This suggests that the increased toxicity of the V. nigrum L. and RPA combination was due mainly to increased content of the special veratrum alkaloids. The cytotoxicity of veratridine in SH-SY5Y cells was decreased with increasing paeoniflorin concentrations. This study provides insight into the mechanism behind the incompatibility theory of TCM.


Introduction
Traditional Chinese medicine (TCM) is an important part of Chinese culture and makes significant contributions to the prosperity and health of the Chinese population. TCM has become more popular worldwide because of its efficacy and curative effects. However, it is important to ensure that these treatments are safe. Confirming the compatibility of TCM is currently the major method used to ensure its safety and efficacy [1]. TCM has formed a unique incompatibility theory over thousands of years, and the typical principles of prescriptions include the eighteen incompatible medicaments. This means that these specific agents can be toxic when used in combination [2]. However, some studies have used combinations of drugs in the eighteen incompatible medicaments to treat incurable diseases [2]. Therefore, it is important to determine whether these agents are incompatible when used in combination, and the reasons behind any incompatibility.
In this study, we used Veratrum nigrum L. (V. nigrum L.) and Radix paeoniae alba (RPA), which are two agents belonging to the eighteen incompatible medicaments whose concurrent use has been prohibited for thousands of years. Although there are no cases describing the compatibility of V. nigrum L. and RPA in modern Chinese medicine, fewer than 20 prescriptions using a combination of V. nigrum L. and RPA have been described to treat incurable diseases such as tumors, hemorrhoids, carbuncle, and breast carbuncle [3]. Therefore, it is necessary to determine whether the use of V. nigrum L. and RPA in combination should be prohibited, as well as the reasons for any incompatibility [4].
V. nigrum L. is the dried roots and rhizomes of Veratrum nigrum L. It has been used for medicinal purposes for thousands of years in China and was used in Europe during the Middle Ages [5] despite its well-known poisonous characteristics [4]. It is used to treat hypertension, stroke, excessive phlegm, and epilepsy. However, aqueous extracts 2 Evidence-Based Complementary and Alternative Medicine are toxic and irritate the digestive tract mucosa, nucleus nervi vagi, and central nervous system [6,7]. RPA, the dried root of Paeonia lactiflora Pall without bark, has been used as a medicinal herb in traditional Chinese medicine for centuries and exerts a wide range of pharmacological activities. In ancient pharmacology, RPA was used to calm liver wind, relieve pain, nourish blood, regulate menstrual functions, and suppress sweating [8]. In modern pharmacology, RPA decoctions could be used to treat rheumatoid arthritis, systemic lupus erythematosus, hepatitis, dysmenorrhea, muscle cramping and spasms, and long-standing fever [4,9,10].
In this study, UPLC/TOF MS, multivariate statistical analysis, and typical metabolomics approach were used to identify the key chemical markers that are responsible for the increased toxicity and the combination of V. nigrum L. and RPA. In addition, cellular and mouse acute toxicity assays were used to identify the toxic effects of this combination. The compatibility of V. nigrum L. and RPA has not yet been validated by modern science; therefore, it is important to provide preliminary insight into the mechanism of their reported incompatibility, as well as to identify and quantify their toxic chemical constituents.

Chemicals and Materials.
RPA from Panan County (Zhejiang, China; lot 100714) and V. nigrum L. from Huajia County (Changchun, China; lot 100701) were purchased from Anhui BBCA Tongling Chinese herbal medicine company. Formic acid (CNW Technologies GmbH) and acetonitrile (Fisher Scientific, Fair Lawn, NJ, USA) were both of chromatographic purity. Deionized water was prepared using a Millipore water purification system. Veratridine (D00114760) was purchased from Merck (Merck KGaA, Germany), and paeoniflorin was purchased from Sigma-Aldrich. The MTS cell proliferation assay kit was purchased from Promega. Roswell Park Memorial Institute 1640 (RPMI-1640) and fetal bovine serum (FBS) were purchased from Gibco BRL (Invitrogen, USA).

Preparation of Decoctions.
Based on a previous study, the LD 50 of aqueous extracts of V. nigrum L. and RPA after intragastric administration were 2.566 g/kg and 160 g/kg, respectively [11]. In this study, the dose of V. nigrum L. was fixed at 2.566 g/kg, and the RPA dose varied from 0.2566 to 25.66 g/kg.
The doses of V. nigrum L. and RPA in each of 12 groups are shown in Table 1. Each group was extracted using deionized water (700 mL) for 1 h during microboiling under reflux. The extracts were filtered through three layers of gauze, and the drug extraction was then repeated. The filtrates were combined and were then concentrated to 100 mL at 60−70 ∘ C under reduced pressure. Samples were then shaken, calibrated, and stored at 4 ∘ C.
All decoctions were centrifuged at 13,000 rpm for 10 min using a Heraeus Labofuge 400R refrigerated centrifuge (Thermo Scientific, USA). The supernatants were then filtered through a 0.22 M aqueous microporous membrane and were stored at 4 ∘ C for UPLC-TOF-MS analysis. Groups were prepared in triplicate [12,13]. Mice were housed at a constant temperature of 25 ± 1 ∘ C with 50 ± 20% humidity, with a 12 h light/dark cycle and 10-15 air changes per hour. Mice were allowed free access to water and food during the experimental period.

Animals and Experimental
Mice were acclimatized to the facilities and environment for 3 days before the experiments. Two hundred and forty mice were randomized into 12 groups (A-L), with 10 mice per group. Food was removed from all animals 12 h before the experiments. Mice received decoctions at 0.4 mL/10 g via gavage and were observed for 14 days.

Mass Spectrometry.
Mass spectrometry was performed using a Waters SYNAPT mass spectrometer equipped with an electrospray ionization (ESI) source. Samples were injected twice: once in positive ESI mode and once in negative ESI mode. The data acquisition range was 100-1500 Da. The lock spray reference scan frequency was 20 s, with a reference cone voltage of 30 V. The MS source temperature was set at 100 ∘ C, and the desolvation temperature was 450 ∘ C with a gas flow of 900 L/h. The lock mass compound was leucine enkephalin (200 pg/ L), with an m/z of 556.2771 in the positive ion mode and 554.2615 in negative ion mode. The capillary voltages were set to 2.9 kV for ESI+ and 3 kV for ESI−. The cone voltage was 40 kV, and the collision energies were 6 V (trap) and 4 V (transfer), with 2.00 mL/min trap gas flow [16].

SH-SY5Y Cell Cultures and Cell Proliferation
Assay. SH-SY5Y cells (ATCC, Manassas, USA) were cultured in RPMI-1640 supplemented with 10% FBS. Cells were maintained at 37 ∘ C in an incubator with a saturated humidity atmosphere of 95% air and 5% CO 2 . They were cultured in a 96-well plate and treated with different drug combinations. An MTS cell proliferation assay was performed, and the OD at 490 nm was read using a VICTOR X plate reader (Perkin Elmer, USA).

Data Analysis.
The UPLC-TOF-MS data of all samples were analyzed using MassLynx4.1 software (Waters, Manchester, UK), and principal component analysis (PCA) was used for data analysis. Pearson correlation coefficients were used to identify relationships between the study parameters and mortality. The chemical markers in each group were identified using the V. nigrum L. and RPA chemical databases [12]. For data analyses, we compared the correlation between chemical composition and the acute toxicity of V. nigrum L. and RPA between the 12 groups. Experimental values are expressed as means ± standard deviations (SD). Statistical analyses were performed using two-tailed Student's -tests. A value of < 0.01 was considered statistically significant.

Results and Discussion
3.1. Mouse Acute Toxicity. Mice in the 12 groups were administered the decoctions by intragastric administration. Death occurred 5−10 min after administration. Toxicity was manifested predominantly as trembling, convulsions, and spasms. The major organs showed no obvious lesions at necropsy by the naked eye [6,17,18]. Mouse mortality is shown in Table 2.
Group L (with a V. nigrum L. dose of the LD 50 2.566 g/kg) had a mortality of 70%. Groups A−F were treated with the same dose of V. nigrum L., but the groups fed a lower ratio of RPA had a higher mortality than group L. Groups G−K had a higher proportion of RPA and lower mortality than group L, suggesting that toxicity was minimized. Groups B and F were the most toxic with mortalities of 100% and 95%, respectively. Overall analyses showed that mortality increased when the dose of V. nigrum L. was less than the dose of RPA. Conversely, toxicity decreased when the proportion of RPA increased. The trends in mortality are shown in Figure 1.
Female mouse mortality decreased with increasing proportions of RPA, whereas males mortality exhibited two peaks at groups B and F, which both had 100% (Figure 2).   Table 1.   Table 1.

Identification and Quantification of the Chemical Components of V. nigrum L. and RPA.
To explain the cause of acute toxicity of the combined use of V. nigrum L. and RPA, we compared the chemical composition of the 12 groups using a supervised orthogonal partial least squared discriminant analysis (OPLS-DA). After Pareto scaling with mean centering, the data from both the positive and negative ion modes were displayed as scores plots (Figure 3). The scores plots clearly revealed that the samples were clustered Scores comp [1]   Negative ion mode. The same score plots represent the same group. Both the positive and negative ion modes showed that the samples were clearly clustered into 12 groups. Replicates of the 12 groups could be distinguished easily.
into 12 groups; replicates of same group were comparable, but the 12 groups could be distinguished easily. This suggests that the changes in chemical compositions of the 12 groups were consistent, and the experiment was reproducible. UPLC-TOF-MS is a rapid, specific, and sensitive method used to identify and quantify individual components. Representative BPI chromatograms of the 36 samples in the positive and negative mode ESI are shown in Figure 4.
The contents of some special alkaloids change significantly in different proportion decoctions. To determine which constituents contributed to the differences in toxicity   Tables 3 and 4. observed with the 12 decoctions, statistical analyses were performed using Pearson's correlation, and the correlation coefficient ( ) described the degree of linear correlation between two variables: total mouse mortality and the chemical composition of the 12 decoctions. The values were between −1 and +1; > 0 indicates that two variables were positively correlated, whereas < 0 indicates a negative correlation. The larger the absolute value, the higher the correlation; therefore we defined = 0.90-1.00 as extremely related and = 0.80-0.89 as highly related. We extracted the relevant chemical component data when ≥ 0.8 and identified the individual components using the RPA and V. nigrum L. chemical composition databases.
According to Pearson's correlation coefficient, 131 compositions had an extremely positive relationship ( ≥ 0.9) with toxicity. However, only 14 of these could be identified and all came from V. nigrum L. In addition, 35 components were highly positively correlated ( ≥ 0.8) to acute toxicity, and these constituents contributed most of the toxicity identified in the 12 decoctions. Therefore, we hypothesized that these components might play important roles in the acute toxicity of this drug combination. We identified that these compounds were all veratrum alkaloids that were detected in each sample in various ionized forms. The chemical components with ≥ 0.8 are shown in Table 3 [7,[19][20][21][22][23][24].
3-Veratroylgermine, jervine, veratramine, germanitrine, germidine, and germerine were the main chemical components that were associated with acute toxicity, with Pearson's correlation coefficients of 0.9779, 0.9661, 0.9589, 0.9581, 0.9402, and 0.9393, respectively. These compounds are all veratrum alkaloids, and the changes in their content are shown in Figure 5. The other 29 components that were highly positively correlated ( ≥ 0.8) with acute toxicity are shown in Figure 6.

Cell Proliferation Assay.
To further validate the toxicity of the combination of V. nigrum L. and RPA on neurons, we used veratridine and paeoniflorin, the major pharmacologically active components of V. nigrum L. and peony for in vitro experiments. The effect of these two chemical components on   (Figure 8(c)). Therefore, paeoniflorin could offset the cytotoxicity caused by veratridine in SH-SY5Y cells [25].

Discussion
The eighteen incompatible medicaments (Shi Ba Fan) is a well-known traditional Chinese medicine (TCM) theory. It basically listed eighteen pairs of TCM herbal medicine that are not to be used in combination as they may cause fatal consequences. This principal has been used as guidance for general TCM practices for hundreds of years with little chemical comprehension. In this work, for the first time, we used the UPLC-TOF MS coupled with multivariate statistical analysis in an effort to gain scientific understanding for the causes of the incompatibility. We previously reported that the LD 50 of aqueous extracts of V. nigrum L. and RPA were 2.566 g/kg and 160 g/kg, respectively, after intragastric administration [11]. V. nigrum L. extract was severely toxic, whereas RPA had minimal toxicity. Although the amount of V. nigrum L. in each group was unchanged, the mortality changed significantly with altered ratios of V. nigrum L. to RPA. When the doses of