Cyclopamine

A highly sensitive quantification method for 12 plant toxins in human serum using liquid chromatography tandem mass spectrometry with a quick solid-phase extraction technique

Masaru Taniguchi (Conceptualization) (Methodology) (Validation) (Formal analysis) (Investigation) (Resources) (Writing – original draft) (Visualization), Tomiaki Minatani (Resources) (Writing – review and editing), Hitoshi Miyazaki (Visualization), Hitoshi Tsuchihashi

Abstract

Please cite this article as: Taniguchi M, Minatani T, Miyazaki H, Tsuchihashi H, Zaitsu K, A highly sensitive quantification method for 12 plant toxins in human serum using liquid chromatography tandem mass spectrometry with a quick solid-phase extraction technique, Journal of Pharmaceutical and Biomedical Analysis (2020), doi: https://doi.org/10.1016/j.jpba.2020.113676
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A highly sensitive quantification method for 12 plant toxins in human serum using liquid chromatography tandem mass spectrometry with a quick solid-phase extraction technique intra-day accuracies were 92.7%–116% and 91.6%–106%, respectively; and the inter- and intra-day precisions were below 14% and 11%, respectively. Also, the lower limits of detection and quantification were 0.0071–0.15 and 0.022–0.46 ng/mL, respectively, indicating the method’s high sensitivity. Finally, to confirm its feasibility, our method was applied to two model samples: (1) commercially available human serum and (2) pseudo poisoning serum via dilution of mouse serum with human serum. We were able to quantify α-chaconine at 0.84 ± 0.02 ng/mL in the serum (Case 1) and protoveratrine A at 0.15 ± 0.032 ng/mL in the pseudo poisoning serum (Case 2), demonstrating our method’s practicality. This is the first time that the 12 plant toxins in human serum were simultaneously quantitated. Our method can investigate accidental poisonings involving toxic plants, enabling prompt decisions on patient treatment.

Keywords: Plant toxins, Monolithic SPE column, Human serum, Validation, LC/MS/MS

1. Introduction

A wide variety of plants naturally produce various toxins, such as alkaloids, glycosides, proteins, terpenes, and steroids [1-5]. Drawing on their multiple and diverse biological properties, toxins play a significant role in protecting plants against various threats, such as bacteria, fungi, insects, and predators [6].
Most food poisonings involving plant toxins are generally caused by the accidental intake of toxic plants, which are often mistakenly ingested instead of edible plants. Moreover, some plant toxins are also used for suicide [7] and potentially in food terrorism [8]. When accidental poisonings that involve plant toxins occur, a quick and reliable identification of toxins in biological samples is quite important for clinical treatment [9]. A quick investigation into the causative substances helps minimize the number of victims in such accidental poisonings. To identify toxic compounds, various biological samples, such as blood (plasma and serum), urine, and saliva, are commonly used [10-13]. Among the biological samples, urine and saliva are noninvasively available and are often used for drug analysis in forensic toxicology [14, 15]. Although metabolites and their parent compounds are usually detected in urine, the urinary metabolites of some plant toxins have not been elucidated in humans. Moreover, urine is pooled in the bladder, preventing the urinary concentration of toxins from reflecting the toxic effects in the human body. Contrarily, a concentration of toxins in the blood is an indicator of the extent of their toxic effects. Therefore, a quantitative analysis of plant toxins in the blood (serum and plasma) is important to assess the victims’ poisoning levels. Since information on the level of plant toxins in the blood is scarce, it is necessary to collect quantitative data on plant toxins in the blood of poisoned victims.
Although sample preparation for analysis tends to be time-consuming, it is crucial to quantitating toxic blood concentrations. Several biological sample preparation techniques, such as deproteinization [16], liquid–liquid extraction (LLE) [17], and solid-phase extraction (SPE), have been reported [11]. Deproteinization commonly yields poor sample purification, and the extraction efficiency of LLE is greatly affected by the physical properties of analytes. Although SPE using a manifold is generally known to be effective for the purification of samples, it is not suitable for the quick processing of multiple samples.
Contrarily, the SPE method using MonoSpin®, a Monolithic SPE column, facilitates the quick processing of multiple samples. MonoSpin® has been recently used for analysis of drugs, plant toxins, and pesticides in biological samples [18-20]. A sample preparation using MonoSpin® is almost complete by centrifugation alone, hence the convenient handling of MonoSpin®.
In this study, we developed a highly sensitive quantification method for 12 plant toxins in human serum using liquid chromatography tandem mass spectrometry (LC/MS/MS) with the quick SPE with MonoSpin®. In this paper, we selected lycorine, galanthamine, protoveratrine A, protoveratrine B, veratramine, veratridine, jervine, cyclopamine, cevadine, α-solanine, α-chaconine, and solanidine as targeted analytes based on the statistical information of Japan’s Ministry of Health, Labor, and Welfare (Fig.1 and Table 1). In Japan, these toxins exist in major toxic plants and are often detected in accidental poisoning cases, which are commonly reported in other countries [5, 21-23]. To the best of our knowledge, no simultaneous quantification methods exist for the 12 plant toxins in human serum. Two extraction methods, namely, deproteinization and SPE with MonoSpin® C18, were compared and determined for sample preparation. The analytical conditions of LC/MS/MS were also optimized for the quick quantitation of the analytes. We further validated the quantitative method using SPE with MonoSpin® C18. Finally, we applied the quantitative method to Japanese human serum and pseudo poisoning serum to evaluate the feasibility of the method.

2. Materials and methods

2.1. Chemicals and reagents

Pooled human serum and serum from individual donors was were obtained from BioIVT (Westbury, NY, US) and Tennessee Blood Services (Memphis, TN, US). Galanthamine, protoveratrine A, veratramine, veratridine, jervine, cyclopamine, cevadine, α-solanine, and α-chaconine were purchased from PhytoLab GmbH & Co. KG (Vestenbergsgreuth, Germany). Lycorine hydrochloride and protoveratrine B were obtained from SigmaAldrich (Tokyo, Japan) and solanidine from ChromaDex (Los Angeles, CA, US). These analytes are categorized in Table 1 and their chemical structures are shown in Fig.1. Yohimbine-[13C1, D3] was purchased from IsoSciences (Ambler, PA, US) and used as the internal standard (IS). High-performance liquid chromatography (HPLC)-grade acetonitrile (ACN), methanol (MeOH), and 2-propanol (IPA) were purchased from Kanto Chemical (Tokyo, Japan). Ammonium formate and liquid chromatography–mass spectrometry (LC-MS)-grade formic acid (FA) were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Milli-Q® water (PURELAB® Ultra, ELGA LabWater, High Wycombe, UK) was used throughout the experiments.

2.2. Preparation of calibrants and quality control (QC) samples

Stock standard solutions of Veratrum alkaloids and IS were prepared using 0.1% of FA in MeOH, and the other toxins were dissolved in MeOH. A method development standard solution was prepared by mixing the stock standard solutions, and its concentration was adjusted to 10 µg/mL with 0.1% of FA in MeOH. All solutions were stored at –20°C until analysis. Mixed standard solutions for the preparation of calibrants and QC samples were made by diluting the stock standard solutions with MeOH. Calibrants and QC samples (i.e., spiked serum) were prepared by spiking 100 µL of the corresponding mixed standard solutions to 4.9 mL of the pooled human serum samples. The concentration ranges of calibrants and the concentrations of QC samples are presented in Table 2. An IS working solution (400 ng/mL) was prepared by diluting the IS stock solution with MeOH immediately prior to use.

2.3. Sample preparation

Analytes’ SPE was performed using the Monolithic SPE column (i.e., MonoSpin® C18) (GL Sciences, Tokyo, Japan). The column was conditioned with 0.2 mL of MeOH, followed by 0.2 mL of Milli-Q® water. Subsequently, it was centrifuged at 2,500 g for 1 min using a micro refrigerated centrifuge (Model 3740, Kubota Corporation, Tokyo, Japan), obtaining a conditioned column. The spiked or not-spiked human serum sample (0.4 mL each) was pipetted into a 1.5-mL microtube, and a 10-µL aliquot of IS working solution (400 ng/ml) was added to the sample. After vortexing for 10 s, the sample was centrifuged at 6,700 g for 1 min. The supernatant (0.2 mL) was applied to the abovementioned conditioned column, centrifuged at 2,500 g for 2 min. After the column was rinsed with 0.3 mL of Milli-Q® water, it was centrifuged at 2,500 g for 1 min. The analytes were eluted with 300 µL of 0.1% of FA in MeOH by centrifugation at 2,500 g for 2 min. The eluate was evaporated to dryness under a gentle nitrogen stream at 55°C. The residue was dissolved with 100 µL of MeOH. The reconstituted solution was filtered using an Ultrafree®-MC centrifugal device (0.22-µm pore size, hydrophilic PTFE, Merck KGaA, Darmstadt, Germany) by centrifugation at 10,000 g for 2 min. Finally, 10 µL of the filtrate was injected into the LC-MS/MS system.

2.4. LC analytical conditions

Chromatographic separation was performed using the Nexera XR system (Shimadzu Corporation, Kyoto, Japan) with a Capcell Pak ADME column (2.1 mm i.d. × 150 mm, particle size: 3 µm, OSAKA SODA, Osaka, Japan) and a pre-column filter (Cat. No. 6010-55100, GL Sciences). The temperature of the column oven was 40°C. Mobile phases A and B contained 0.2 mol/L ammonium formate aqueous solution/Milli-Q® water/ACN (v/v/v, 10:85:5) and 0.2 mol/L ammonium formate aqueous solution/MeOH/ACN (v/v/v, 10:20:70), respectively. The gradient conditions were as follows: 0% B for 1 min, 0%–100% B (1–6.5 min, linear gradient), 100% B (6.5–11 min), and 0% B (11–15 min). The total flow rate was set at 0.3 mL/min.

2.5. MS analytical conditions

A QTRAP 4500 tandem mass spectrometer (Sciex, Framingham, MA, US), equipped with an electrospray ionization (ESI) source (Turbo V ionization source, Sciex), was used in the positive ionization mode. MS parameters were optimized by analyzing the representative analyte (α-Solanine) as follows: curtain gas, 20 psi (138 kPa); collision gas, 6 arbitrary units; ion source temperature, 500°C; ion source gas 1, 50 psi (345 kPa); ion source gas 2, 80 psi (552 kPa); and ion-spray voltage, 5,500 V. Selected reaction monitoring (SRM) parameters for each analyte were optimized by the infusion analysis of each standard solution using a syringe pump (Table 1). Data acquisition was performed using the scheduled SRM mode [24] (scheduled time: retention time for each analyte ± 0.5 min), and the data were processed using a built-in software Analyst (Sciex, version 1.6.2).

2.6. Extraction recovery and matrix effect

The extraction recovery of two sample preparation methods, namely, deproteinization and SPE with MonoSpin® C18, was compared. The detail of the procedures for the deproteinization and SPE were shown in Supplementary Materials. To calculate the extraction recovery, the peak area of each analyte obtained from a pre-spiked sample was divided by that obtained from a post-spiked sample (Equation 1). Then, the peak areas of analytes obtained by triplicate analyses were averaged. The matrix effect was also evaluated for the methods and determined using Equation 2, whereby the peak areas of the analytes obtained by triplicate analyses were averaged.

2.7. Method validation

The quantitative method was validated as per the requirements and criteria of the Guideline on Bioanalytical Method Validation in Pharmaceutical Development [25], where selectivity, linearity, inter- and intra-day accuracy, and precision of the method were evaluated. Selectivity was evaluated by analyzing 2 individuals and 4 different types of pooled human serum samples. Calibration curves (y=ax+b) were constructed using the internal standard method, and linear regression with a 1/x2 weighting factor was used, where “x” was the ratio of analyte concentration to IS concentration, “y” was the ratio of analyte peak area to IS area, “a” was a slope of the regression line and “b” was a yintercept. The intra- and inter-day accuracy and precision were evaluated by analyzing the QC samples (n = 5). The theoretical lower limit of detection (LLOD) and quantification (LLOQ) were determined by analyzing the calibrants and calculated according the formula: LLODs=3.29sd/a, LLOQ=10sd/a (sd= the standard deviation of y= intercepts of regression lines and a= slope of the calibration curve) [26]. In addition, the stability of the analytes in serum samples (25 ng/mL) were evaluated (n=3) under room temperature (25°C) for 1h.

3. Results and discussion

3.1. Selection of the analytes

Since 2009, 158 food poisoning accidents caused by toxic plants have been reported in Japan (in Japanese, https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/kenkou_iryou/shokuhin/syokuchu/04. html). The most common cases, which account over 50% of accidents, involved Amaryllidaceae plants, Veratrum plants, and Solanum tuberosum L. (Fig. 2). Amaryllidaceae plants contain toxic Amaryllidaceae alkaloids, with the typical toxicants being lycorine and galanthamine. Veratrum plants also have toxic Veratrum alkaloids, such as protoveratrine A and veratramine. Toxic glycoalkaloids, such as α-solanine and α-chaconine, exist in Solanum tuberosum L., especially in potato eyes and peel. Thus, we selected 12 plant toxins, which are mainly Amaryllidaceae plants, Veratrum plants, and Solanum tuberosum L. (Fig. 1).

3.2. Optimization of the analytical conditions of LC/MS/MS

In this study, the ADME column in which the adamantyl group was introduced as a functional group was used for LC separation. To quickly detect and determine the targeted analytes, the total run time of LC/MS/MS was set at 15 min (see SRM chromatograms for the analytes in Fig. 3). Although the separation of some compounds was somewhat difficult under this LC condition, mass spectrometric differentiation was achieved. Also, the peak shape of each analyte was satisfactory.
As presented in Fig. 1, the chemical structures of the plant-derived compounds are diverse, and thus, it was necessary to retain the high polar and non-polar compounds. The ADME column used in this study exhibited more hydrophilicity than the conventional octadecyl silyl (ODS) column, mainly due to surface modification using the adamantyl group [27]. The ADME column also contributed to the higher retention of highly polar compounds, such as some of the analytes in this study. Moreover, this column is based on reversedphase chromatography and can retain non-polar compounds. In this context, Capcell Pak ADME was found to be suitable for this study.
As presented in Fig. 3, lycorine and galanthamine have the same precursor ion (m/z 288), and thus, it was necessary to separate them chromatographically. In this study, these compounds were separated under the LC condition (resolution: 3.1), while the ODS column did not achieve complete separation (resolution: <1.5) under the same gradient condition. Incomplete chromatographic separations of the compounds will yield inaccurate data, such as crosstalk. Crosstalk refers to false detection observed in the compounds that have the same precursor ion, which is mechanistically caused by the inter-channel delay of the SRM transitions [28]. In effect, the crosstalk of lycorine and galanthamine was observed under the conditions of the ODS column, while it did not occur under the condition of the ADME column, demonstrating the latter’s superiority. The MS parameters were optimized to obtain the highest intensity of product ion for each analyte (Table 1). The protonated molecule [M + H]+ was selected as the precursor ion for all compounds, and two product ions were selected as qualifiers and quantifier ions for each compound.

3.3. Evaluation of the extraction procedures

In this study, a comparison was made on the two extraction methods. In the preliminary study, we applied the deproteinization method using MeOH, IPA, and ACN (see the results in Fig. S2). Unexpectedly, the extraction recovery of protoveratrine A and B was low, especially for MeOH and IPA. To investigate this, the scan analysis was conducted for protoveratrine A (Fig. S3). As presented in Fig. S3b, remarkable peaks at m/z 710 and 752 were observed in the extract from the spiked human serum by MeOH deproteinization. However, these peaks were not observed in the standard solution of protoveratrine A (Fig. S3a), suggesting the degradation of protoveratrine A during extraction: the unknown peaks at m/z 710 and 752 corresponded to the mass shift of 84 and 42 Da from the protonated molecule of protoveratrine A (m/z 794). This result strongly suggested that one or two ester bonds of protoveratrine A hydrolyzed during evaporation, as presented in Scheme S1a. The peak of m/z 710 is likely to be dideacetyl protoveratrine A [29], and the peak of m/z 752 may be mono-deacetyl protoveratrine A.
This deacetylation was accelerated under a weak alkaline condition in normal serum (pH: ca. 7.4). Contrarily, as presented in Scheme S1b, the equilibrium of the reversible reactions could be shifted to the far-left side (protoveratrine A) under an acidic condition. In other words, the extract evaporation under the acidic condition can inhibit the deacetylation of protoveratrine A. As presented in Fig. S3c, the use of 0.1% of FA in MeOH for deproteinization caused the peaks at m/z 710 and 752 to disappear. This result demonstrated that FA prevented the deacetylation of protoveratrine A, indicating its effectiveness in improving the recovery rate of protoveratrine A of the extract adjustment to the weak acidic condition.
As can be seen from Fig. S2, the highest extraction recovery rate, except for protoveratrine A and B, was obtained with IPA, and thus, 0.1% of FA in IPA was applied to the deproteinization of protoveratrine A. As presented in Fig. 4, the extraction recovery rates of protoveratrine A and B improved over 80% using 0.1% of FA in IPA as a deproteinization solvent. These results demonstrate the necessity of the FA addition to the deproteinization solvent when the solution containing protoveratrine A and B needs to be evaporated to dryness.
We employed the Monolithic SPE column (i.e., MonoSpin® C18) to compare the extraction recovery between SPE and deproteinization. The procedure for MonoSpin® C18 consists of sample loading, washing, and elution of the analytes, and these steps were performed by the column’s centrifuge operation. This procedure allowed easy and quick operation and multiple sample processing, improving analysts’ work efficiency. The total time for the completion of this SPE procedure was less than 15 min. Moreover, the total time for SPE clean-up and LC/MS/MS analysis was completed within 30 min.
As presented in Fig. 4a, the extraction recovery rate of the SPE method was over 80% for all analytes. Also, the effect of ion suppression was below 15% (Fig 4b). In addition, the extraction recovery rates and matrix effects of the SPE method were calculated for additional 5 different types of human serum samples (3 pooled and 2 individual serum) and the results are shown in Supplementary Materials (Fig. S4), showing satisfactory results. The detail of the procedures for the SPE and deproteinization were shown in Supplementary Materials.
Compared with the extraction recovery and matrix effect of the deproteinization method, almost no differences existed between the SPE and deproteinization methods (Fig. 4), meanwhile the SPE method is more effective than the deproteinization method in sample purification. In fact, the total ion current chromatogram background obtained from the pooled human serum by the SPE method was lower than that by the deproteinization method (data not shown). Consequently, we selected the SPE method for this study.

3.4. SPE method optimization

Veratrum alkaloids were detectable in human blood at sub-ng/mL levels, as reported in previous studies [22, 23], and thus, the pg/mL -order LLOD of an analytical method is required for Veratrum alkaloids. Acute toxic blood or serum concentrations of Veratrum alkaloids are summarized in Table 3. To achieve such low LLOD, the SPE method was improved as follows. The amount of the loading sample for the spin column was changed from 50 to 200 µL, which achieved ca. fourfold lower LLOD. In this paper, we compared the amounts of the loading sample (100 or 200 µL) because matrix effects might be stronger when the amount of the loading sample increased. However, there were almost no differences in matrix effects between the amounts of the loading sample, except for galanthamine and α-chaconine. In addition, the extraction recovery did not change in tandem with the amount of the loading sample (Fig. S5), and thus, the optimal amount of the loading sample was determined to be 200 µL.

3.5. Method validation

To develop the quantitative method using the optimal SPE method, we validated its selectivity, linearity, precision, and accuracy. To identify the calibrants and QC samples, blank human serum was required, although only α-chaconine was detected in any serum sample at minimal concentration, which may be due to food intake. We determined the concentration of α-chaconine in pooled human serum using the standard addition method prior to the construction of its calibration curve. The extrapolated curve constructed using the standard addition method was linear over the concentration range (0–2.5 ng/mL), with the regression equation being y = 4822.5x + 3,576.4 (r2 = 0.999) (Fig. S1). The αchaconine concentration in the pooled human serum was 0.19 ± 0.038 ng/mL (n = 3) (Table S1). Based on this quantitation result, we prepared the calibrants and QC samples by adding a certain amount of α-chaconine standard solution to the aforementioned pooled human serum. Calibrants and QC samples for other analytes were also made using the pooled human serum, constructing their calibration curves. The method validation results are presented in Table 2. Each calibration curve exhibits good linearity (r2 > 0.991) over the corresponding range. The inter- and intra-day accuracies were 92.7%–116% and 91.6%–106%, respectively. The inter- and intra-day precisions were below 14% and 11%, respectively. The LLOD and LLOQ were 0.0071–0.15 and 0.022–0.46 ng/mL. As far as we know, there was no information on human blood concentration for Amaryllidaceae alkaloids in acute toxic cases, though the human blood concentration levels for glyco- and Veratrum alkaloids-induced acute toxicity were collectable (Table 3) [21-23]. The LLOQs for glyco- and Veratrum alkaloids in the present method were lower than the acute toxic blood or serum concentrations in the previous reports. Although our highly-sensitive method can quantitate serum concentrations of the other analytes as well as glyco- and Veratrum alkaloids, the present method will be applied to real serum samples when the acute toxic cases involved in Amaryllidaceae alkaloids will occur. Moreover, selectivity was evaluated using the 2 individuals and 4 different types of pooled human serum samples prepared using the SPE method. For targeted analytes and IS, except for α-chaconine, no endogenous interference peaks were observed at the corresponding retention times of each analyte in their SRM chromatograms, proving that the present quantification method is reliable for 12 analytes in human serum. As shown in Table S2, there was no change between each peak area before and after 1h-preservation at room temperature (25°C), demonstrating that the analytes were stable in serum under the conditions.

3.6. Applications

To confirm its feasibility, the method was necessary to be applied to patient serum obtained from actual poisoning cases. Although two poisoning accidents which involved toxic plants occurred in Nagoya City during this study, the victims’ blood samples were not obtained because their symptoms were mild and they recovered quickly without a medical diagnosis. Alternatively, we applied the method to two model samples: (1) commercially available human serum, in which α-chaconine was expected to exist, and (2) pseudo poisoning serum made by 50-fold dilution of protoveratrine A-injected mouse serum with pooled human serum. The details of the animal study are described in the Supplementary Materials.

3.6.1. Model 1

Prior to the quantification of α-chaconine, the product ion scan was performed, with the spectra of α-chaconine in the Japanese single-donor serum (aged 45, male), purchased from BioIVT (Fig. S6b), and the reference standard α-chaconine (Fig. S6a) matching well, confirming the existence of α-chaconine in the Japanese serum. Quantitation of αchaconine in the Japanese serum was then performed, with its concentration being 0.84 ± 0.019 ng/mL (n = 3). To interpret this result, the α-chaconine serum concentration of poisoning or the detection window of α-chaconine after diet is necessary. Information on this point is scarce, except that from Hellenas et al., who reported that the average human serum concentration of α-chaconine is ca. 2 ng/mL 1 h after intake of mashed potato. They also reported the observation of specific symptoms via α-chaconine [30]. Compared with their finding, the result of this study (0.84 ± 0.02 ng/mL) might be due to a daily diet, including potatoes or their processed products. This result suggests that our method can detect even low concentrations of α-chaconine derived from the daily diet.

3.6.2. Model 2

Protoveratrine A (25 µg/kg) was administered intraperitoneally to mice (n = 4), and serum samples obtained via abdominal aorta were subjected to 50-fold dilution with the pooled human serum to obtain the pseudo poisoning serum. Our method was applied to this serum, enabling the detection of protoveratrine A with an average value of 0.15 ± 0.032 ng/mL (n = 4) in the pseudo poisoning serum (Table S3). Our method could quantify low concentration levels of protoveratrine A in the pseudo poisoning serum, demonstrating its practicality. The quantitative values slightly varied among the individuals (RSD: 21%), mainly because of individual differences in the distribution, metabolism, and excretion profiles of protoveratrine A. Also, the concentration was converted to the initial concentration in mice serum (7.6 ± 1.6 ng/mL, n = 4) (Table S3). To the best of our knowledge, this is the first time that protoveratrine A in mice serum was determined using the validated SPE-LC-MS/MS method.

3.6.3. Future perspectives

In this study, we were not able to obtain the actual poisoning serum. Nonetheless, our method can still be applied to real poisoning cases in the future. We will apply this method to the serum and report the results as case studies when we manage to obtain real poisoning serum. In addition, we will increase the number of targeted analytes in the method and extend them to other plant toxins, such as aconitine and atropine.

4. Conclusions

In this study, we developed a highly sensitive quantification method for 12 plant toxins in human serum by LC/MS/MS using a quick SPE technique. The total time for completing the SPE procedure was 15 min. To quickly identify and quantitate the analytes, the ADME column was used for LC separation, and the total runtime of LC/MS/MS was set at 15 min. Consequently, total time for SPE clean-up and LC/MS/MS analysis was completed within 30 min. The quantitative method was validated, and the results were satisfactory as follows: linearity (r2 > 0.99) over the corresponding range, inter- and intraday accuracy (92.7%–116% and 91.6%–106%, respectively), and inter- and intra-day precision (below 14% and 11%, respectively). The lower limits of detection and quantification were 0.0071–0.15 and 0.022–0.46 ng/mL, respectively, indicating the method’s high sensitivity.
Finally, our method was Cyclopamine applied to two model samples, demonstrating its practicality. Arguably, our method is useful for a quick and reliable quantification of the 12 plant toxins in human serum. It can also be used to investigate accidental poisonings involving toxic plants, enabling prompt decisions on patient treatment.

References

[1] G. Lin, J.Y. Wang, N. Li, M. Li, H. Gao, Y. Ji, F. Zhang, H. Wang, Y. Zhou, Y. Ye, H.X. Xu, J. Zheng, Hepatic sinusoidal obstruction syndrome associated with consumption of Gynura segetum, J Hepatol 54(4) (2011) 666-73.
[2] L.M. Castello, S. Negro, F. Santi, I. Zanotti, M. Vidali, M. Bagnati, G. Bellomo, G.C. Avanzi, Accidental digitoxin intoxication: an interplay between laboratory and clinical medicine, Biochem Med (Zagreb) 22(3) (2012) 380-4.
[3] L. Polito, M. Bortolotti, M.G. Battelli, G. Calafato, A. Bolognesi, Ricin: An Ancient Story for a Timeless Plant Toxin, Toxins (Basel) 11(6) (2019).
[4] L. de Haro, P. Pommier, L. Tichadou, M. Hayek-Lanthois, J. Arditti, Poisoning by Coriaria myrtifolia Linnaeus: a new case report and review of the literature, Toxicon 46(6) (2005) 600-3.
[5] T. Minatani, H. Ohta, E. Sakai, T. Tanaka, K. Goto, D. Watanabe, H. Miyaguchi, Analysis of toxic Veratrum alkaloids in plant samples from an accidental poisoning case, Forensic Toxicology 36(1) (2017) 200-210.
[6] K. Welch, Editorial – Plant toxins, Toxicon 168 (2019) 140.
[7] M. Mizugaki, K. Ito, Y. Ohyama, Y. Konishi, S. Tanaka, K. Kurasawa, Quantitative analysis of Aconitum alkaloids in the urine and serum of a male attempting suicide by oral intake of aconite extract, J Anal Toxicol 22(4) (1998) 336-40.
[8] M.N.O. Sadiku, S.M. Musa, T.J. Ashaolu, Food Terrorism, International Journal of Trend in Scientific Research and Development Volume-3(Issue-4) (2019) 134-135. [9] M.A. Kirk, J. Cisek, S.R. Rose, Emergency department response to hazardous materials incidents, Emerg Med Clin North Am 12(2) (1994) 461-81.
[10] Y. Shin, J. Lee, J. Lee, J. Lee, E. Kim, K.H. Liu, H.S. Lee, J.H. Kim, Validation of a Multiresidue Analysis Method for 379 Pesticides in Human Serum Using Liquid Chromatography-Tandem Mass Spectrometry, J Agric Food Chem 66(13) (2018) 3550-3560.
[11] J. Carlier, J. Guitton, L. Romeuf, F. Bevalot, B. Boyer, L. Fanton, Y. Gaillard, Screening approach by ultra-high performance liquid chromatography-tandem mass spectrometry for the blood quantification of thirty-four toxic principles of plant origin. Application to forensic toxicology, J Chromatogr B Analyt Technol Biomed Life Sci 975 (2015) 65-76.
[12] T. Vasiljevic, G.A. Gomez-Rios, J. Pawliszyn, Single-Use Poly(etheretherketone) Solid-Phase Microextraction-Transmission Mode Devices for Rapid Screening and Quantitation of Drugs of Abuse in Oral Fluid and Urine via Direct Analysis in RealTime Tandem Mass Spectrometry, Anal Chem 90(1) (2018) 952-960.
[13] R. Chen, Z. Ning, C. Zheng, Y. Yang, C. Zhang, X. Ou, K. Chen, H. Yu, X. Wei, Q. Zhao, J. He, Simultaneous determination of 16 alkaloids in blood by ultrahighperformance liquid chromatography-tandem mass spectrometry coupled with supported liquid extraction, J Chromatogr B Analyt Technol Biomed Life Sci 1128 (2019) 121789.
[14] K. Zaitsu, M. Katagi, H.T. Kamata, T. Kamata, N. Shima, A. Miki, H. Tsuchihashi, Y. Mori, Determination of the metabolites of the new designer drugs bk-MBDB and bkMDEA in human urine, Forensic Sci Int 188(1-3) (2009) 131-9.
[15] M. Di Rago, M. Chu, L.N. Rodda, E. Jenkins, A. Kotsos, D. Gerostamoulos, Ultrarapid targeted analysis of 40 drugs of abuse in oral fluid by LC-MS/MS using carbon13 isotopes of methamphetamine and MDMA to reduce detector saturation, Anal Bioanal Chem 408(14) (2016) 3737-49.
[16] M. Kusano, K. Zaitsu, K. Taki, K. Hisatsune, J. Nakajima, T. Moriyasu, T. Asano, Y. Hayashi, H. Tsuchihashi, A. Ishii, Fatal intoxication by 5F-ADB and diphenidine: Detection, quantification, and investigation of their main metabolic pathways in humans by LC/MS/MS and LC/Q-TOFMS, Drug Test Anal 10(2) (2018) 284-293.
[17] Y.S. Park, S.H. Kim, S.Y. Kim, Y.H. Kim, M.H. Lee, S.C. Yang, L.M. Shaw, J.S. Kang, Quantification of galantamine in human plasma by validated liquid chromatographytandem mass spectrometry using glimepride as an internal standard: application to bioavailability studies in 32 healthy Korean subjects, J Chromatogr Sci 50(9) (2012) 803-9.
[18] T. Saito, R. Umebachi, A. Namera, H. Oikawa, S. Miyazaki, Y. Fujita, S. Inokuchi, Identification and Quantification of Aconitines and Colchicine in Serum, Urine, and Plants using MonoSpin C18 and LC-MSMS, Chromatographia 78(15-16) (2015) 1041-1048.
[19] T. Saito, T. Fukushima, Y. Yui, S. Miyazaki, A. Nakamoto, A. Namera, S. Inokuchi, Monolithic spin column extraction and GC-MS for the simultaneous assay of diquat, paraquat, and fenitrothion in human serum and urine, Anal Bioanal Chem 400(1) (2011) 25-31.
[20] A. Nakamoto, M. Nishida, T. Saito, I. Kishiyama, S. Miyazaki, K. Murakami, M. Nagao, A. Namura, Monolithic silica spin column extraction and simultaneous derivatization of amphetamines and 3,4-methylenedioxyamphetamines in human urine for gas chromatographic-mass spectrometric detection, Anal Chim Acta 661(1) (2010) 42-6. [21] A. Nara, K. Saka, C. Yamada, T. Kodama, T. Takagi, Forensic analysis using ultrahigh-performance liquid chromatography-tandem mass spectrometry with solid-phase extraction of alpha-solanine and alpha-chaconine in whole blood, Forensic Toxicol 37(1) (2019) 197-206.
[22] Y. Gaillard, G. Pepin, LC-EI-MS determination of veratridine and cevadine in two fatal cases of Veratrum album poisoning, J Anal Toxicol 25(6) (2001) 481-5. [23] T. Grobosch, T. Binscheck, F. Martens, D. Lampe, Accidental intoxication with Veratrum album, J Anal Toxicol 32(9) (2008) 768-73.
[24] M. Dziadosz, J.P. Weller, M. Klintschar, J. Teske, Scheduled multiple reaction monitoring algorithm as a way to analyse new designer drugs combined with synthetic cannabinoids in human serum with liquid chromatography-tandem mass spectrometry, J Chromatogr B Analyt Technol Biomed Life Sci 929 (2013) 84-9.
[25] Guideline on Bioanalytical Method Validation in Pharmaceutical Development.pdf, 2013.
[26] L.A. Currie, Limits for qualitative detection and quantitative determination. Application to radiochemistry, Analytical Chemistry 40(3) (1968) 586-593.
[27] H. Li, L. Ai, S. Fan, Y. Wang, D. Sun, Rapid determination of 18 glucocorticoids in serum using reusable on-line SPE polymeric monolithic column coupled with LCquadrupole/orbitrap high-resolution mass spectrometer, J Chromatogr B Analyt Technol Biomed Life Sci 1065-1066 (2017) 79-86.
[28] B. Patel, B.N. Suhagia, A.G. Jangid, H.N. Mistri, N. Desai, Systematic evaluation of matrix effect and cross-talk-free method for simultaneous determination of zolmitriptan and N-desmethyl zolmitriptan in human plasma: a sensitive LC-MS/MS method validation and its application to a clinical pharmacokinetic study, Biomed Chromatogr 30(3) (2016) 447-58.
[29] H.L. Li, J. Tang, R.H. Liu, M. Lin, B. Wang, Y.F. Lv, H.Q. Huang, C. Zhang, W.D. Zhang, Characterization and identification of steroidal alkaloids in the Chinese herb Veratrum nigrum L. by high-performance liquid chromatography/electrospray ionization with multi-stage mass spectrometry, Rapid Commun Mass Spectrom 21(6) (2007) 869-79. [30] K.E. Hellenas, A. Nyman, P. Slanina, L. Loof, J. Gabrielsson, Determination of potato glycoalkaloids and their aglycone in blood serum by high-performance liquid chromatography. Application to pharmacokinetic studies in humans, J Chromatogr 573(1) (1992) 69-78.