- Email: [email protected]
Determination and validation of tetrodotoxin in human whole blood using hydrophilic interaction liquid chromatography–tandem mass spectroscopy and its application
Forensic Science International 217 (2012) 76–80
Contents lists available at SciVerse ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Determination and validation of tetrodotoxin in human whole blood using hydrophilic interaction liquid chromatography–tandem mass spectroscopy and its application Hwang Eui Cho b, Su Youn Ahn a,b, In Seop Son b, Sangwhan In a, Ran Seon Hong b, Dong Woo Kim a, Sang Hee Woo a, Dong Cheul Moon b,*, Suncheun Kim a,** a b
National Forensic Service, 305-348, Republic of Korea College of Pharmacy, Chungbuk National University, Cheongju 361-763, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 January 2011 Received in revised form 29 September 2011 Accepted 1 October 2011 Available online 22 October 2011
A sensitive analytical method was developed for the quantitative determination of tetrodotoxin (TTX), a powerful sodium channel blocker, in human postmortem whole blood. The sample mixture was cleaned up using cation exchange SPE catridge after protein precipitation by methanol and then separated on a PCHILIC (phosphorylcholine hydrophilic interaction liquid chromatography) column (150 mm 2.0 mm i.d., 5 mm) using a isocratic elution of 1% acetic acid and acetonitrile. The identification of TTX was performed on tandem mass spectrometry with electrospray ionization interface in positive ion mode. The retention time of voglibose (internal standard) and TTX was 5.1 and 6.0 min, respectively. TTX and internal standard (voglibose) were monitored and quantitated using the ion transitions: the respective precursor to product ion combinations, m/z 320/302 for TTX and m/z 268/92 for voglibose in the multiple reaction monitoring (MRM) mode. The recovery of TTX and voglibose was 61.4% and 62.8%, respectively and the good accuracy (97.7–103.9%), linearity (2–1200 ng/mL) and reproducibility were shown in this method. The limit of detection and limit of quantification were 0.32 ng/mL and 1.08 ng/mL, respectively. This method was applied in the case of three fishermen who were poisoned (including one death) by unknown fish on their boat in October 2010. In this case, the levels of TTX were 27.2, 30.0 and 29.7 ng/mL in heart blood, peripheral blood and serum of a victim, were 3.1 and 12.1 ng/mL in peripheral blood and 3.9 and 12.8 ng/mL in serum of two survivors, respectively. ß 2011 Elsevier Ireland Ltd. All rights reserved.
Keywords: Tetrodotoxin (TTX) Intoxication PC-HILIC Voglibose, Postmortem
1. Introduction Tetrodotoxin (TTX) is a neurotoxin contained in puffer fish of the family Tetraodontidae. TTX is one of the most potent nonprotein poisons found in nature, and no specific antidote or antitoxins to TTX are available [1,2]. TTX poisoning is most commonly induced by ingesting the improper prepared puffer fish [3]. It is an infrequent incident in Korea and accounts for around 50 deaths since 1990s, whereas the number of annual deaths in Japan from TTX poisoning is approximately 50 [4,5]. Deaths have also been reported in coastal regions such as Taiwan, Hong Kong and Southeast Asia [6]. TTX blocks voltage-gated Na+ channels, thus preventing depolarization and propagation of action potentials in
* Corresponding author. Tel.: +82 43 261 2819; fax: +82 43 275 6131. ** Corresponding author at: Central District Office, National Forensic Service, Daejon, Republic of Korea. Tel.: +82 42 862 8073; fax: +82 42 862 8074. E-mail addresses: [email protected] (D.C. Moon), [email protected] (S. Kim). 0379-0738/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2011.10.026
the brainstem, somatic motor, sensory, and autonomic nerves. It can cause ultimately skeletal muscle paralysis. Symptoms, including paresthesia, ataxia, dizziness, diarrhea, respiratory failure, nausea, vomiting, usually develop within 30 min of ingestion, and typical death usually occurs within 4–6 h. Despite fatal poisoning, TTX has potential medical use in treating brain disorders, ocular pain, excess pain in the large intestine and ileum, and relieving tension of the skeletal muscles, neuralgia, rheumatism, arthritis, and etc. On account of its potential lethality, more rapid and selective analytical method with less sample preparation constraints was required for draw legal conclusion. A few methods for detecting of TTX in biological samples have been reported. Some were based on precolumn derivatization using HPLC-FLD [7,8] or gas chromatography with mass spectrometry (GC/MS) [9–11]. But it had drawbacks of complex extraction procedure and time-consuming sample preparation. Kawatsu et al. developed the immunoaffinity chromatography method to detect TTX from urine samples, but monoclonal antibody against TTX is very expensive [12]. Recently, measurements of TTX by liquid chromatography–tandem mass
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
spectrometry (LC/MS/MS) are becoming more common due to the improved sensitivity and selectivity [9,13–15]. The previously reported methods have many constraints for extraction of TTX in the complex matrices like postmortem samples comprising blood, stomach contents and biles. For serum and urines samples, an extraction method using C18 cartridges and ultra filtration has been reported [9,15]. However, these methods have disadvantages – time-consuming steps, poor recovery and reproducibility. In addition, microcentrifuge filters are expensive and unsuitable for postmortem whole blood. Hence, we have developed a simple solid phase extraction method for cleanup of biological samples, which required minimal sample preparation step to apply LC-tandem mass spectrometric quantification. The optimal separation condition of the amphophilic analyte was accomplished by using HILIC (hydrophilic interaction liquid chromatography) prior to mass analysis. In PCHILIC, a hydrophilic column has been used for the analysis of polar compounds such as TTX that are difficult to retain and separate by reversed-phase HPLC. 2. Materials and methods 2.1. Reagents and materials 2.1.1. Reagents TTX (purity, >99%) was purchased from Sigma (St. Louis, MO, U.S.A.). Voglibose (internal standard, 99.9% purity) was obtained from CJ Pharm. Co. Ltd (Icheon, Korea). A cation exchange type SPE cartridges, adande:lTM PCX (3 mg/1 mL) cartridges were purchased from Shiseido (Tokyo, Japan). HPLC-grade acetonitrile and methanol were purchased from Fischer Scientific Co. (FairLawn, NJ, USA). Water was prepared using the Milli-Q water purification system (Millipore, Bedford, MA, USA). Analytical reagent-grade glacial acetic acid was purchased from Merck Company (Darmstadt, Germany). All other chemicals were the highest quality available and were used without further purification. 2.1.2. Samples Peripheral blood and serum samples of one victim and two patients being treated for TTX poisoning within 24 h after ingesting soup containing unknown fish were offered from a hospital. Blank blood and heart blood of a victim who already died before arrived at the hospital were offered from the National Forensic Service (NFS, Korea). Samples were stored at 4 8C until analysis. 2.2. Liquid chromatography The HPLC system consisted of an Agilent 1200 series with a degasser, a binary pump, a thermo-stated autosampler and a column oven set at 30 8C (Agilent Technologies, Palo Alto, CA, U.S.A.). Chromatographic separation was performed in a PC HILIC column (5 mm, 150 mm 2.0 mm i.d.; Shiseido, Tokyo, Japan) equipped with a guard column of HILIC cartridge (4 mm 2.0 mm i.d.; Phenomenex, Torrance, CA, USA). The mobile phase was composed of 1% acetic acid in acetonitrile/1% acetic acid in water (88:12, v/v, pH 4.1) and delivered isocratically at 0.4 mL/min.
77
10,000 g for 5 min. The supernatant solution was transferred and was evaporated under N2 at 50 8C. The residue was dissolved with 500 mL of 1% acetic acid in methanol. After the addition of 500 mL of acetonitrile to remove metallic cation and lipids in the solution, it was vortex-mixed again for 1 min, then centrifuged at 10,000 g for 3 min. The supernatant was loaded onto the cartridge (PCX cartridges, Shiseido, Japan), preconditioned with 1000 mL of methanol. After complete loading, the cartridge was washed with 1000 mL of 1% acetic acid and 1 mL of acetone followed by 1 mL of methanol. TTX was subsequently eluted with 1000 mL of 0.1 M HCl in methanol. And then, the eluate was evaporated under N2 at 50 8C. The dried residue was reconstituted in 100 mL of 1% acetic acid in methanol. 5 mL aliquots of the clear supernatant were injected into the tandem MS system. 2.5. Preparation of calibration standards and quality control samples Stock standard solutions of TTX were prepared by dissolving 1 mg of TTX in 10 mL of 1% acetic acid in methanol (v/v). The IS stock solution were prepared by dissolving 1 mg of volglibose in 10 mL of methanol. The standard solutions were kept at 4 8C in amber-glass vessels. Effective concentrations of calibration curves were 2.0, 4.0, 20.0, 40.0, 200, 400 and 1200 ng/mL, which were prepared by spiking 50 mL of the appropriate standard solution to 250 mL of drug-free whole blood. Working standard solutions were prepared by serial dilution of the stock solutions with 1% acetic acid in methanol to required concentrations. The IS working solution (2000 ng/mL) was also prepared by diluting its stock solution with methanol. Quality control (QC) samples were prepared by spiking the working standard solutions in drug-free whole blood for the evaluation of precision, accuracy, recovery and ion suppression to make the final concentrations at 0, 4.0, 20.0, 40.0, 200 and 400 ng/mL of TTX in blood, respectively. 2.6. Method validation The specificity of the method was determined by analyzing six different sources of postmortem whole blood to demonstrate the lack of chromatographic interference from endogenous blood components. The precision and accuracy of the method were determined by replicate analyses for batches of the QC sample (4.0, 20.0, 40.0, 200 and 400 ng/mL of TTX) on a day and six separate days. The different sources of blank postmortem whole blood in each QC sample set were used. The precision was calculated by the intra and inter-day percent relative standard deviation (RSD, %). The accuracy was measured as the percentage deviation from the nominal concentration. The recovery of the sample cleanup procedure was assessed by comparing the mean peak areas of the regularly prepared samples at five concentrations (4.0, 20.0, 40.0, 200 and 400 ng/mL) with the mean peak areas of spike-after-extraction blood samples in five replicates. To prepare the spike-after-extraction samples, blank whole blood was processed according to the sample preparation procedure as described above. All the eluate was mixed with the appropriate standard solutions of TTX at concentrations corresponding to the final concentration of the pretreated blood samples. After vortexing, the mixture was evaporated to dryness and the residue was reconstituted with 100 mL of 1% acetic acid in methanol (v/v). The effect of blood constituents over the ionization of TTX and IS was evaluated by comparing the chromatographic peak areas of analyte from the spike-after-extraction samples at five levels with the neat standard at the equivalent concentrations, whereas the IS was determined at single concentration of 200 ng/mL.
3. Results and discussion 3.1. Protein precipitation and SPE
2.3. Electrospray ionization mass spectrometry The blood concentration of TTX was quantified using liquid chromatography– mass spectrometry with a Sciex API 3200 QTRAP mass equipped with a Turbo Ion Spray interface to generate the positive ions [M+H]+. Characteristic mass fragments of the identified precursor ions for quantification were determined in multiple reaction monitoring (MRM) mode using precursor-product ion combinations. Unit mass resolution was used in Q1 and Q3. Optimal declustering potential (DP), 70 V; collision energy potentials (CE), 30 eV; collision exit potentials (CXP), 10 V and entrance potential (EP), 10 V were determined based on the relative intensities of selected product ions. A dwell time of 200 ms and inter-channel delay of 10 ms were used during the experiment. Source parameters were optimized as follows: ion spray voltage, 5500 V; curtain gas, 20; collision gas, high; ion source gas 1 and 2 at 40 and 80; interface heater, on; ion source temperature, 400 8C. Mass transitions m/z 320.3 ! 302.3 for TTX and 268.5 ! 92.0 for IS were used for quantification. Instrument settings, data acquisitions and processing were controlled by the software package Analyst (Version 1.5.1, Applied Biosystems). 2.4. Sample preparation 250 mL of H2O, 25 mL of internal standard (voglibose, 2000 ng/mL) and 1000 mL of methanol were added to a 250-mL aliquot of the samples in a 2-mL micro-tube. The sample mixture was vortex-mixed for 3 min thoroughly, then centrifuged at
The zwitterionic characteristic of TTX makes it extremely difficult to be retained in RP-HPLC and to extract from biological samples with organic solvents. A few different extraction procedures can be used with conventional detection methods. Derivatization can be performed in urine and serum samples for HPLC-FLD [7,8] and GC–MS analysis [9–11]. C18 cartridge column and ultra filtration can be applied to LC–MS/MS and LC–MS analysis [9,15–17]. However, these sample preparation approaches are still highly labor-intensive, time-consuming, and expensive. In some methods, poor reproducibility and recovery and high matrix effect was shown. In this study, a novel and rapid method for determining TTX from biological samples was developed. To find the most efficient protein precipitation condition, various materials – 5% trichloroacetic acid (TCA), 5% perchloric acid (PCA), ethanol, and methanol were tested. The protein precipitation by methanol was provided acceptable yields, minimal ion suppression and sufficient cleaning for LC–MS/ MS analysis. In addition, the method required less time-consuming
78
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
for evaporation to dryness so that facilitated more rapid sample preparation. To additionally clean up, various brands of cation exchange cartridges were tested. The most promising one was adande:lTM PCX (Shiseido, Tokyo, Japan). During the optimization it was found that the most efficient elution solution was 1 mL of 0.1 M HCl in methanol. The recoveries of spiked TTX from blood were above 52% in this case. Moreover, using the PCX cartridge as cation exchange resins and the optimal elution solvent enables to improve sensitivity and to prove to be robust by eliminating matrix effect from the blood samples. Thus, the extraction procedure described here will be possible to easily prepare whole blood samples. 3.2. Optimization of mass spectrometric and chromatographic conditions
3.3. Internal standard selection Choosing an appropriate I.S was an important aspect to achieving acceptable method performance, especially with HILIC–LC–ESI–MS/ MS. Use of stable isotope-labeled analogues as internal standard is highly recommended since matrix effect should not affect the relative efficiency of the ionization of the analyte and I.S, but there was no good isotope-labeled I.S commercially available for TTX. In this experiment, voglibose, structurally similar to TTX, was adopted as I.S, which has an almost identical retention time to that of the analyte. According to this, potential matrix effect for the target analyte and the I.S caused by co-eluted endogenous matrix components could be compensated, for their similar chromatographic and mass spectrometric properties. 3.4. Method validation
In order to achieve the quantitative determination of TTX in whole blood, the LC–ESI–MS/MS conditions were optimized to obtain sensitivity and signal stability during direct infusion of the analytes to electrospray ion source operated in positive ion mode at a flow rate of 10 mL/min. Variable mass spectrometric conditions (source temperature, ion spray voltage, collision energy, etc.) have been investigated. The MRM parameters were optimized to maximize the response for the analyte and I.S. The predominant peaks in the primary ESI spectra of TTX and I.S correspond to the [M+H]+ ions at m/z 320 and m/z 268, respectively. The primary product ions for TTX and IS scanned in Q3 after a collision with nitrogen in Q2 had m/z of 302, 92 (Fig. 1). The minor ions at m/z 284, 256 and 162 were assigned to [M+H2H2O]+, [M + H3H2O]+ and [2-aminohydroquinazoline]+, respectively [18,19]. The protein precipitation for blood preparation might lead to ion suppression when LC-ESI-MS/MS was applied. Therefore appropriate chromatographic column and suitable mobile phase are needed for accurate quantification of TTX in whole blood. This problem was overcome by PC-HILIC column instead of the reversed phase column [20]. The HILIC column gave several advantages. This column offers superior retention for very polar compounds that are difficult to retain under reversed-phase conditions. In HILIC mode, the weak mobile phase is organic solvent, thus, low aqueous-high organic mobile phase is used for the retention of polar TTX and voglibose, which resulted in a more efficient desolvation of analytes in the MS interface and a higher ion signal. The addition of modifiers (formic acid, acetic acid and ammonium acetate) was also studied since these have substantial effects on selectivity and efficiency. Consequently, we found significant improvements of the HPLC separation or peak shape of TTX by using acetic acid. Thus, the optimized mobile phase condition was acetonitrile/1% acetic acid (88:12, v/v). All analyte peaks were symmetrical in appearance. The column capacity factors (K0 ) for TTX and IS were calculated to be 5.5 and 4.5, respectively.
3.4.1. Specificity and selectivity The specificity and selectivity has been studied by using independent blood samples from six different deceased. A typical chromatogram is presented in Fig. 2. TTX and IS were not detected in the blank blood. These chromatograms revealed that there was no interfering peak derived from the endogenous components at the elution times of the TTX and IS. The TTX and IS eluted chromatographically at approximately 6.0 and 5.1 min, respectively. 3.4.2. Linearity of calibration curves and lower limit of quantification A linear regression analysis was performed using concentration (x) and peak-area ratios (y) to determine the correlation coefficient (r). Seven calibrators were used to generate a standard curve, in five replicates. The calibration curve was linear over the concentration range, 2.0–1200 ng/mL with r of 0.9997 0.0002. The mean equation of the regression line was y = (0.0113 0.0002)x (0.0049 0.0342). The relative standard deviation (RSD,%) of five replicate determinations of the seven points was in the range of 1.85–9.25%. The limit of detection (LOD) was 0.32 ng/mL defined by the concentration of analyte giving a signal-to-noise (S/N) ratio of 3. And the limit of quantification (LOQ) was 1.08 ng/mL defined by the lowest concentration in the linear range that can be detected with a variation within 13.9%. 3.4.3. Precision and accuracy The intra (n = 6) and inter-day (n = 6) precision and accuracy were summarized in Table 1. The inter-day assay variations were determined by duplicates of QC sample on six separate days. In both cases, the accuracy ranged from 97.7 to 103.9% at the concentrations investigated, and the RSDs were less than 3.4%. 3.4.4. Recovery and matrix effect In this study, the recoveries and matrix effect were evaluated by analyzing QC samples over a range of 4–400 ng/mL. The recoveries and matrix effect of TTX in blood samples were presented in Table 2.
Fig. 1. Product ion mass spectra of the TTX (A) and voglibose (B).
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
79
Fig. 2. Chromatograms of TTX (left) and voglibose (right) in blood. Blank blood (A), blood spiked with 2 ng/mL of TTX (LOQ) and 200 ng/mL of I.S (B).
Table 1 Precision and accuracy for the assay of TTX in blood. Nominal conc. (ng/mL)
4 20 40 200 400
Intra-assay (n = 6)
Inter-assay (n = 6)
Measured conc. (Mean S.D., ng/mL)
RSD (%)
Accuracy (%)
Measured conc. (mean S.D., ng/mL)
RSD (%)
Accuracy (%)
4.0 0.1 19.8 0.3 39.5 1.0 196.1 3.4 393.2 4.5
2.3 1.5 2.4 1.7 1.1
101.1 97.7 99.3 98.4 100.4
4.2 0.1 20.3 0.2 41.8 1.3 206.1 6.7 405.3 6.3
3.4 1.3 3.4 3.3 1.5
103.9 98.2 99.9 101.3 101.6
The recovery of the internal standard at 200 ng/mL was 62.8% (RSD 5.0%, n = 5) and it was similar with that of TTX. Average matrix effect values obtained were 98.3–111.2% at tested concentration, respectively. This method has higher selectivity for TTX detection in blood than other previously reported methods and provides the goodreproducible quantitative values.
3.4.5. TTX determination for samples The validated method was applied to sample analyses for one victim and two survivors with TTX poisoning. TTX was successfully detected in all samples and TTX level of the samples of them are shown in Table 3. In all cases, the levels of whole blood and serum were similar. In a victim, the levels of peripheral blood and heart
Fig. 3. Chromatograms of peripheral blood (A) and serum (B) of patient No. 1.
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
80
Table 2 Recovery and matrix effect for TTX in whole blood (n = 5). Nominal concentration (ng/mL)
Recovery (%) [mean S.D.]
RSD (%)
Matrix effect (%) [mean S.D.]
RSD (%)
4 20 40 200 400
51.5 0.01 53.7 0.03 54.4 0.03 57.8 0.03 61.4 0.04
2.4 6.5 6.1 5.2 5.8
106.4 0.04 107.7 0.14 98.3 0.07 111.2 0.05 105.1 0.10
4.2 13.1 7.6 4.9 9.5
Mean S.D.
55.8 0.03
5.2
105.7 0.08
7.9
Table 3 TTX concentrations in blood and serum of patients. Patient no.
a
1 2 3
Sex/age
M/47 M/46 M/48
Detected level (ng/mL) Heart blood
Peripheral blood
Serum
27.2
30.0 12.1 3.1
29.7 12.8 3.9
a Patient no.1 already died before arrived at the hospital. The heart blood was collected at autopsy and the peripheral blood and serum was collected at the hospital.
blood also were similar. The example chromatograms were presented in Fig. 3. 4. Conclusion A LC/ESI–MS/MS method was developed and validated for the determination of TTX in whole blood and serum. Compared with the analytical methods reported previously, the current method showed simpler and more reliable sample pretreatment and more rapid chromatographic time. This is the first application used internal standard for the determination of TTX in biological samples. The method in presented this paper can be used in qualifying and quantifying TTX in blood samples. Furthermore, this method showed acceptable linearity, precision, accuracy and recovery characteristics and was applied successfully to measure the toxin in postmortem samples for forensic studies and in blood samples for clinical toxicology. References [1] A.M.N. Ahasan, A.A. Mamun, S.R. Karim, M.A. Bakar, E.A. Gazi, C.S. Bala, Paralytic complications of puffer fish (tetrodotoxin) poisoning, Singapore Med. J. 45 (2004) 73–74. [2] T. Narahashi, Tetrodotoxin, Proc. Jpn. Acad. Ser. B 84 (2008) 147–154. [3] S.K. Chew, C.H. Goh, K.W. Wang, P.K. Mah, B.Y. Tan, Puffer fish (tetrodotoxin) poisoning: clinical report and role of anti-cholinesterase drugs in therapy, Singapore Med. J. 24 (1983) 168–171. [4] Tetradotoxin poisoning associated with eating puffer fish transported from JapanCalifornia, 1996, Morb. Mortal. Wkly. Rep. 45 (1996) 389–391.
[5] F.R. Chowdhury, H.A.M. Nazmul Ahasan, A.K.M. Mamunur Rashid, A. Al Mamun, S.M. Khaliduzzaman, Tetrodotoxin poisoning: a clinical analysis role of neostigmine and short-term outcome of 53 cases, Singapore Med. J. 48 (2007) 830–833. [6] C.K. How, C.H. Chern, Y.C. Huang, L.M. Wang, C.H. Lee, Tetrodotoxin Poisoning, Am. J. Emerg. Med. 21 (2003) 51–54. [7] T. Yasumoto, T. Michishita, Fluorometric determination of tetrodotoxin by high performance liquid chromatography, Agric. Biol. Chem. 49 (1985) 3077–3080. [8] M. Kono, T. Matsui, K. Furukawa, T. Takase, K. Yamamori, H. Kaneda, D. Aoki, J.H. Jang, M. Yotsu-Yamashita, Examination of transformation among tetrodotoxin and its analogs in the living cultured juvenile puffer fish, kusafugu Fugu niphobles by intramuscular administration, Toxicon 52 (2008) 714–720. [9] Y.H. Tsai, D.F. Hwang, C.A. Cheng, C.C. Hwang, J.F. Deng, Determination of tetrodotoxin in human urine and blood using C18 cartridge column, ultrafiltration and LC–MS, J. Chromatogr. B 832 (2006) 75–80. [10] S. Kurono, H. Hattori, O. Suzuki, T. Yamada, H. Seno, Sensitive analysis of tetrodotoxin in human plasma by solid-phase extractions and gas chromatography/mass spectrometry, Anal. Lett. 34 (2001) 2439–2446. [11] T. Noguchi, H. Narita, J. Maruyama, K. Hashimoto, Tetrodotoxin in the starfishes Astropecten polyacanthus, in association with toxification of a trumpet shell Boshubora Charonia sauliae, Bull. Japan. Soc. Sci. Fish. 48 (1982) 1173–1177. [12] K. Kawatsu, T. Shibata, Y. Hamano, Application of immunoaffinity chromatography for detection of tetrodotoxin from urine samples of poisoned patients, Toxicon 37 (1999) 325–333. [13] M. Horie, S. Kobayashi, N. Shimizu, H. Nakazawa, Determination of tetrodotoxin in puffer-fish by liquid chromatography–electrospray ionization mass spectrometry, Analyst 127 (2002) 755–759. [14] T. Nakagawa, J. Jang, M. Yotsu-Yamashita, Hydrophilic interaction liquid chromatography–electrospray ionization mass spectrometry of tetrodotoxin and its analogs, Anal. Biochem. 352 (2006) 142–144. [15] H.C. Jen, S.J. Lin, Y.H. Tsai, C.H. Chen, Z.C. Lin, D.F. Hwang, Tetrodotoxin poisoning evidenced by solid-phase extraction combining with liquid chromatography– tandem mass spectrometry, J. Chromatogr. B 871 (2008) 95–100. [16] M.A. O’Leary, J.J. Schneider, G.K. Isbister, Use of high performance liquid chromatography to measure tetrodotoxin in serum and urine of poisoned patients, Toxicon 44 (2004) 549–553. [17] H.-N. Huangb, J. Lin, H.-L. Lin, Identification and quantification of tetrodotoxin in the marine gastropod Nassarius by LC–MS, Toxicon 51 (2008) 774–779. [18] Y. Shoji, M. Yotsu-Yamashita, T. Miyazawa, T. Yasumoto, Electrospray ionization mass spectrometry of tetrodotoxin and its analogs: liquid chromatography/mass spectrometry tandem mass spectrometry, and liquid chromatography/tandem mass spectrometry, Anal. Biochem. 290 (2001) 10–17. [19] O.R. Pires Jr., A. Sebben, E.F. Schwartz, R.A. Morales, C. Bloch Jr., C.A. Schwartz, Further report of the occurrence of tetrodotoxin and new analogues in the Anuran family Brachycephalidae, Toxicon 45 (2005) 73–79. [20] X. Wang, W. Li, H.T. Rasmussen, Orthogonal method development using hydrophilic interaction chromatography and reversed-phase high-performance liquid chromatography for the determination of pharmaceuticals and impurities, J. Chromatogr. A 1083 (2005) 58–62.
Contents lists available at SciVerse ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Determination and validation of tetrodotoxin in human whole blood using hydrophilic interaction liquid chromatography–tandem mass spectroscopy and its application Hwang Eui Cho b, Su Youn Ahn a,b, In Seop Son b, Sangwhan In a, Ran Seon Hong b, Dong Woo Kim a, Sang Hee Woo a, Dong Cheul Moon b,*, Suncheun Kim a,** a b
National Forensic Service, 305-348, Republic of Korea College of Pharmacy, Chungbuk National University, Cheongju 361-763, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 January 2011 Received in revised form 29 September 2011 Accepted 1 October 2011 Available online 22 October 2011
A sensitive analytical method was developed for the quantitative determination of tetrodotoxin (TTX), a powerful sodium channel blocker, in human postmortem whole blood. The sample mixture was cleaned up using cation exchange SPE catridge after protein precipitation by methanol and then separated on a PCHILIC (phosphorylcholine hydrophilic interaction liquid chromatography) column (150 mm 2.0 mm i.d., 5 mm) using a isocratic elution of 1% acetic acid and acetonitrile. The identification of TTX was performed on tandem mass spectrometry with electrospray ionization interface in positive ion mode. The retention time of voglibose (internal standard) and TTX was 5.1 and 6.0 min, respectively. TTX and internal standard (voglibose) were monitored and quantitated using the ion transitions: the respective precursor to product ion combinations, m/z 320/302 for TTX and m/z 268/92 for voglibose in the multiple reaction monitoring (MRM) mode. The recovery of TTX and voglibose was 61.4% and 62.8%, respectively and the good accuracy (97.7–103.9%), linearity (2–1200 ng/mL) and reproducibility were shown in this method. The limit of detection and limit of quantification were 0.32 ng/mL and 1.08 ng/mL, respectively. This method was applied in the case of three fishermen who were poisoned (including one death) by unknown fish on their boat in October 2010. In this case, the levels of TTX were 27.2, 30.0 and 29.7 ng/mL in heart blood, peripheral blood and serum of a victim, were 3.1 and 12.1 ng/mL in peripheral blood and 3.9 and 12.8 ng/mL in serum of two survivors, respectively. ß 2011 Elsevier Ireland Ltd. All rights reserved.
Keywords: Tetrodotoxin (TTX) Intoxication PC-HILIC Voglibose, Postmortem
1. Introduction Tetrodotoxin (TTX) is a neurotoxin contained in puffer fish of the family Tetraodontidae. TTX is one of the most potent nonprotein poisons found in nature, and no specific antidote or antitoxins to TTX are available [1,2]. TTX poisoning is most commonly induced by ingesting the improper prepared puffer fish [3]. It is an infrequent incident in Korea and accounts for around 50 deaths since 1990s, whereas the number of annual deaths in Japan from TTX poisoning is approximately 50 [4,5]. Deaths have also been reported in coastal regions such as Taiwan, Hong Kong and Southeast Asia [6]. TTX blocks voltage-gated Na+ channels, thus preventing depolarization and propagation of action potentials in
* Corresponding author. Tel.: +82 43 261 2819; fax: +82 43 275 6131. ** Corresponding author at: Central District Office, National Forensic Service, Daejon, Republic of Korea. Tel.: +82 42 862 8073; fax: +82 42 862 8074. E-mail addresses: [email protected] (D.C. Moon), [email protected] (S. Kim). 0379-0738/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2011.10.026
the brainstem, somatic motor, sensory, and autonomic nerves. It can cause ultimately skeletal muscle paralysis. Symptoms, including paresthesia, ataxia, dizziness, diarrhea, respiratory failure, nausea, vomiting, usually develop within 30 min of ingestion, and typical death usually occurs within 4–6 h. Despite fatal poisoning, TTX has potential medical use in treating brain disorders, ocular pain, excess pain in the large intestine and ileum, and relieving tension of the skeletal muscles, neuralgia, rheumatism, arthritis, and etc. On account of its potential lethality, more rapid and selective analytical method with less sample preparation constraints was required for draw legal conclusion. A few methods for detecting of TTX in biological samples have been reported. Some were based on precolumn derivatization using HPLC-FLD [7,8] or gas chromatography with mass spectrometry (GC/MS) [9–11]. But it had drawbacks of complex extraction procedure and time-consuming sample preparation. Kawatsu et al. developed the immunoaffinity chromatography method to detect TTX from urine samples, but monoclonal antibody against TTX is very expensive [12]. Recently, measurements of TTX by liquid chromatography–tandem mass
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
spectrometry (LC/MS/MS) are becoming more common due to the improved sensitivity and selectivity [9,13–15]. The previously reported methods have many constraints for extraction of TTX in the complex matrices like postmortem samples comprising blood, stomach contents and biles. For serum and urines samples, an extraction method using C18 cartridges and ultra filtration has been reported [9,15]. However, these methods have disadvantages – time-consuming steps, poor recovery and reproducibility. In addition, microcentrifuge filters are expensive and unsuitable for postmortem whole blood. Hence, we have developed a simple solid phase extraction method for cleanup of biological samples, which required minimal sample preparation step to apply LC-tandem mass spectrometric quantification. The optimal separation condition of the amphophilic analyte was accomplished by using HILIC (hydrophilic interaction liquid chromatography) prior to mass analysis. In PCHILIC, a hydrophilic column has been used for the analysis of polar compounds such as TTX that are difficult to retain and separate by reversed-phase HPLC. 2. Materials and methods 2.1. Reagents and materials 2.1.1. Reagents TTX (purity, >99%) was purchased from Sigma (St. Louis, MO, U.S.A.). Voglibose (internal standard, 99.9% purity) was obtained from CJ Pharm. Co. Ltd (Icheon, Korea). A cation exchange type SPE cartridges, adande:lTM PCX (3 mg/1 mL) cartridges were purchased from Shiseido (Tokyo, Japan). HPLC-grade acetonitrile and methanol were purchased from Fischer Scientific Co. (FairLawn, NJ, USA). Water was prepared using the Milli-Q water purification system (Millipore, Bedford, MA, USA). Analytical reagent-grade glacial acetic acid was purchased from Merck Company (Darmstadt, Germany). All other chemicals were the highest quality available and were used without further purification. 2.1.2. Samples Peripheral blood and serum samples of one victim and two patients being treated for TTX poisoning within 24 h after ingesting soup containing unknown fish were offered from a hospital. Blank blood and heart blood of a victim who already died before arrived at the hospital were offered from the National Forensic Service (NFS, Korea). Samples were stored at 4 8C until analysis. 2.2. Liquid chromatography The HPLC system consisted of an Agilent 1200 series with a degasser, a binary pump, a thermo-stated autosampler and a column oven set at 30 8C (Agilent Technologies, Palo Alto, CA, U.S.A.). Chromatographic separation was performed in a PC HILIC column (5 mm, 150 mm 2.0 mm i.d.; Shiseido, Tokyo, Japan) equipped with a guard column of HILIC cartridge (4 mm 2.0 mm i.d.; Phenomenex, Torrance, CA, USA). The mobile phase was composed of 1% acetic acid in acetonitrile/1% acetic acid in water (88:12, v/v, pH 4.1) and delivered isocratically at 0.4 mL/min.
77
10,000 g for 5 min. The supernatant solution was transferred and was evaporated under N2 at 50 8C. The residue was dissolved with 500 mL of 1% acetic acid in methanol. After the addition of 500 mL of acetonitrile to remove metallic cation and lipids in the solution, it was vortex-mixed again for 1 min, then centrifuged at 10,000 g for 3 min. The supernatant was loaded onto the cartridge (PCX cartridges, Shiseido, Japan), preconditioned with 1000 mL of methanol. After complete loading, the cartridge was washed with 1000 mL of 1% acetic acid and 1 mL of acetone followed by 1 mL of methanol. TTX was subsequently eluted with 1000 mL of 0.1 M HCl in methanol. And then, the eluate was evaporated under N2 at 50 8C. The dried residue was reconstituted in 100 mL of 1% acetic acid in methanol. 5 mL aliquots of the clear supernatant were injected into the tandem MS system. 2.5. Preparation of calibration standards and quality control samples Stock standard solutions of TTX were prepared by dissolving 1 mg of TTX in 10 mL of 1% acetic acid in methanol (v/v). The IS stock solution were prepared by dissolving 1 mg of volglibose in 10 mL of methanol. The standard solutions were kept at 4 8C in amber-glass vessels. Effective concentrations of calibration curves were 2.0, 4.0, 20.0, 40.0, 200, 400 and 1200 ng/mL, which were prepared by spiking 50 mL of the appropriate standard solution to 250 mL of drug-free whole blood. Working standard solutions were prepared by serial dilution of the stock solutions with 1% acetic acid in methanol to required concentrations. The IS working solution (2000 ng/mL) was also prepared by diluting its stock solution with methanol. Quality control (QC) samples were prepared by spiking the working standard solutions in drug-free whole blood for the evaluation of precision, accuracy, recovery and ion suppression to make the final concentrations at 0, 4.0, 20.0, 40.0, 200 and 400 ng/mL of TTX in blood, respectively. 2.6. Method validation The specificity of the method was determined by analyzing six different sources of postmortem whole blood to demonstrate the lack of chromatographic interference from endogenous blood components. The precision and accuracy of the method were determined by replicate analyses for batches of the QC sample (4.0, 20.0, 40.0, 200 and 400 ng/mL of TTX) on a day and six separate days. The different sources of blank postmortem whole blood in each QC sample set were used. The precision was calculated by the intra and inter-day percent relative standard deviation (RSD, %). The accuracy was measured as the percentage deviation from the nominal concentration. The recovery of the sample cleanup procedure was assessed by comparing the mean peak areas of the regularly prepared samples at five concentrations (4.0, 20.0, 40.0, 200 and 400 ng/mL) with the mean peak areas of spike-after-extraction blood samples in five replicates. To prepare the spike-after-extraction samples, blank whole blood was processed according to the sample preparation procedure as described above. All the eluate was mixed with the appropriate standard solutions of TTX at concentrations corresponding to the final concentration of the pretreated blood samples. After vortexing, the mixture was evaporated to dryness and the residue was reconstituted with 100 mL of 1% acetic acid in methanol (v/v). The effect of blood constituents over the ionization of TTX and IS was evaluated by comparing the chromatographic peak areas of analyte from the spike-after-extraction samples at five levels with the neat standard at the equivalent concentrations, whereas the IS was determined at single concentration of 200 ng/mL.
3. Results and discussion 3.1. Protein precipitation and SPE
2.3. Electrospray ionization mass spectrometry The blood concentration of TTX was quantified using liquid chromatography– mass spectrometry with a Sciex API 3200 QTRAP mass equipped with a Turbo Ion Spray interface to generate the positive ions [M+H]+. Characteristic mass fragments of the identified precursor ions for quantification were determined in multiple reaction monitoring (MRM) mode using precursor-product ion combinations. Unit mass resolution was used in Q1 and Q3. Optimal declustering potential (DP), 70 V; collision energy potentials (CE), 30 eV; collision exit potentials (CXP), 10 V and entrance potential (EP), 10 V were determined based on the relative intensities of selected product ions. A dwell time of 200 ms and inter-channel delay of 10 ms were used during the experiment. Source parameters were optimized as follows: ion spray voltage, 5500 V; curtain gas, 20; collision gas, high; ion source gas 1 and 2 at 40 and 80; interface heater, on; ion source temperature, 400 8C. Mass transitions m/z 320.3 ! 302.3 for TTX and 268.5 ! 92.0 for IS were used for quantification. Instrument settings, data acquisitions and processing were controlled by the software package Analyst (Version 1.5.1, Applied Biosystems). 2.4. Sample preparation 250 mL of H2O, 25 mL of internal standard (voglibose, 2000 ng/mL) and 1000 mL of methanol were added to a 250-mL aliquot of the samples in a 2-mL micro-tube. The sample mixture was vortex-mixed for 3 min thoroughly, then centrifuged at
The zwitterionic characteristic of TTX makes it extremely difficult to be retained in RP-HPLC and to extract from biological samples with organic solvents. A few different extraction procedures can be used with conventional detection methods. Derivatization can be performed in urine and serum samples for HPLC-FLD [7,8] and GC–MS analysis [9–11]. C18 cartridge column and ultra filtration can be applied to LC–MS/MS and LC–MS analysis [9,15–17]. However, these sample preparation approaches are still highly labor-intensive, time-consuming, and expensive. In some methods, poor reproducibility and recovery and high matrix effect was shown. In this study, a novel and rapid method for determining TTX from biological samples was developed. To find the most efficient protein precipitation condition, various materials – 5% trichloroacetic acid (TCA), 5% perchloric acid (PCA), ethanol, and methanol were tested. The protein precipitation by methanol was provided acceptable yields, minimal ion suppression and sufficient cleaning for LC–MS/ MS analysis. In addition, the method required less time-consuming
78
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
for evaporation to dryness so that facilitated more rapid sample preparation. To additionally clean up, various brands of cation exchange cartridges were tested. The most promising one was adande:lTM PCX (Shiseido, Tokyo, Japan). During the optimization it was found that the most efficient elution solution was 1 mL of 0.1 M HCl in methanol. The recoveries of spiked TTX from blood were above 52% in this case. Moreover, using the PCX cartridge as cation exchange resins and the optimal elution solvent enables to improve sensitivity and to prove to be robust by eliminating matrix effect from the blood samples. Thus, the extraction procedure described here will be possible to easily prepare whole blood samples. 3.2. Optimization of mass spectrometric and chromatographic conditions
3.3. Internal standard selection Choosing an appropriate I.S was an important aspect to achieving acceptable method performance, especially with HILIC–LC–ESI–MS/ MS. Use of stable isotope-labeled analogues as internal standard is highly recommended since matrix effect should not affect the relative efficiency of the ionization of the analyte and I.S, but there was no good isotope-labeled I.S commercially available for TTX. In this experiment, voglibose, structurally similar to TTX, was adopted as I.S, which has an almost identical retention time to that of the analyte. According to this, potential matrix effect for the target analyte and the I.S caused by co-eluted endogenous matrix components could be compensated, for their similar chromatographic and mass spectrometric properties. 3.4. Method validation
In order to achieve the quantitative determination of TTX in whole blood, the LC–ESI–MS/MS conditions were optimized to obtain sensitivity and signal stability during direct infusion of the analytes to electrospray ion source operated in positive ion mode at a flow rate of 10 mL/min. Variable mass spectrometric conditions (source temperature, ion spray voltage, collision energy, etc.) have been investigated. The MRM parameters were optimized to maximize the response for the analyte and I.S. The predominant peaks in the primary ESI spectra of TTX and I.S correspond to the [M+H]+ ions at m/z 320 and m/z 268, respectively. The primary product ions for TTX and IS scanned in Q3 after a collision with nitrogen in Q2 had m/z of 302, 92 (Fig. 1). The minor ions at m/z 284, 256 and 162 were assigned to [M+H2H2O]+, [M + H3H2O]+ and [2-aminohydroquinazoline]+, respectively [18,19]. The protein precipitation for blood preparation might lead to ion suppression when LC-ESI-MS/MS was applied. Therefore appropriate chromatographic column and suitable mobile phase are needed for accurate quantification of TTX in whole blood. This problem was overcome by PC-HILIC column instead of the reversed phase column [20]. The HILIC column gave several advantages. This column offers superior retention for very polar compounds that are difficult to retain under reversed-phase conditions. In HILIC mode, the weak mobile phase is organic solvent, thus, low aqueous-high organic mobile phase is used for the retention of polar TTX and voglibose, which resulted in a more efficient desolvation of analytes in the MS interface and a higher ion signal. The addition of modifiers (formic acid, acetic acid and ammonium acetate) was also studied since these have substantial effects on selectivity and efficiency. Consequently, we found significant improvements of the HPLC separation or peak shape of TTX by using acetic acid. Thus, the optimized mobile phase condition was acetonitrile/1% acetic acid (88:12, v/v). All analyte peaks were symmetrical in appearance. The column capacity factors (K0 ) for TTX and IS were calculated to be 5.5 and 4.5, respectively.
3.4.1. Specificity and selectivity The specificity and selectivity has been studied by using independent blood samples from six different deceased. A typical chromatogram is presented in Fig. 2. TTX and IS were not detected in the blank blood. These chromatograms revealed that there was no interfering peak derived from the endogenous components at the elution times of the TTX and IS. The TTX and IS eluted chromatographically at approximately 6.0 and 5.1 min, respectively. 3.4.2. Linearity of calibration curves and lower limit of quantification A linear regression analysis was performed using concentration (x) and peak-area ratios (y) to determine the correlation coefficient (r). Seven calibrators were used to generate a standard curve, in five replicates. The calibration curve was linear over the concentration range, 2.0–1200 ng/mL with r of 0.9997 0.0002. The mean equation of the regression line was y = (0.0113 0.0002)x (0.0049 0.0342). The relative standard deviation (RSD,%) of five replicate determinations of the seven points was in the range of 1.85–9.25%. The limit of detection (LOD) was 0.32 ng/mL defined by the concentration of analyte giving a signal-to-noise (S/N) ratio of 3. And the limit of quantification (LOQ) was 1.08 ng/mL defined by the lowest concentration in the linear range that can be detected with a variation within 13.9%. 3.4.3. Precision and accuracy The intra (n = 6) and inter-day (n = 6) precision and accuracy were summarized in Table 1. The inter-day assay variations were determined by duplicates of QC sample on six separate days. In both cases, the accuracy ranged from 97.7 to 103.9% at the concentrations investigated, and the RSDs were less than 3.4%. 3.4.4. Recovery and matrix effect In this study, the recoveries and matrix effect were evaluated by analyzing QC samples over a range of 4–400 ng/mL. The recoveries and matrix effect of TTX in blood samples were presented in Table 2.
Fig. 1. Product ion mass spectra of the TTX (A) and voglibose (B).
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
79
Fig. 2. Chromatograms of TTX (left) and voglibose (right) in blood. Blank blood (A), blood spiked with 2 ng/mL of TTX (LOQ) and 200 ng/mL of I.S (B).
Table 1 Precision and accuracy for the assay of TTX in blood. Nominal conc. (ng/mL)
4 20 40 200 400
Intra-assay (n = 6)
Inter-assay (n = 6)
Measured conc. (Mean S.D., ng/mL)
RSD (%)
Accuracy (%)
Measured conc. (mean S.D., ng/mL)
RSD (%)
Accuracy (%)
4.0 0.1 19.8 0.3 39.5 1.0 196.1 3.4 393.2 4.5
2.3 1.5 2.4 1.7 1.1
101.1 97.7 99.3 98.4 100.4
4.2 0.1 20.3 0.2 41.8 1.3 206.1 6.7 405.3 6.3
3.4 1.3 3.4 3.3 1.5
103.9 98.2 99.9 101.3 101.6
The recovery of the internal standard at 200 ng/mL was 62.8% (RSD 5.0%, n = 5) and it was similar with that of TTX. Average matrix effect values obtained were 98.3–111.2% at tested concentration, respectively. This method has higher selectivity for TTX detection in blood than other previously reported methods and provides the goodreproducible quantitative values.
3.4.5. TTX determination for samples The validated method was applied to sample analyses for one victim and two survivors with TTX poisoning. TTX was successfully detected in all samples and TTX level of the samples of them are shown in Table 3. In all cases, the levels of whole blood and serum were similar. In a victim, the levels of peripheral blood and heart
Fig. 3. Chromatograms of peripheral blood (A) and serum (B) of patient No. 1.
H.E. Cho et al. / Forensic Science International 217 (2012) 76–80
80
Table 2 Recovery and matrix effect for TTX in whole blood (n = 5). Nominal concentration (ng/mL)
Recovery (%) [mean S.D.]
RSD (%)
Matrix effect (%) [mean S.D.]
RSD (%)
4 20 40 200 400
51.5 0.01 53.7 0.03 54.4 0.03 57.8 0.03 61.4 0.04
2.4 6.5 6.1 5.2 5.8
106.4 0.04 107.7 0.14 98.3 0.07 111.2 0.05 105.1 0.10
4.2 13.1 7.6 4.9 9.5
Mean S.D.
55.8 0.03
5.2
105.7 0.08
7.9
Table 3 TTX concentrations in blood and serum of patients. Patient no.
a
1 2 3
Sex/age
M/47 M/46 M/48
Detected level (ng/mL) Heart blood
Peripheral blood
Serum
27.2
30.0 12.1 3.1
29.7 12.8 3.9
a Patient no.1 already died before arrived at the hospital. The heart blood was collected at autopsy and the peripheral blood and serum was collected at the hospital.
blood also were similar. The example chromatograms were presented in Fig. 3. 4. Conclusion A LC/ESI–MS/MS method was developed and validated for the determination of TTX in whole blood and serum. Compared with the analytical methods reported previously, the current method showed simpler and more reliable sample pretreatment and more rapid chromatographic time. This is the first application used internal standard for the determination of TTX in biological samples. The method in presented this paper can be used in qualifying and quantifying TTX in blood samples. Furthermore, this method showed acceptable linearity, precision, accuracy and recovery characteristics and was applied successfully to measure the toxin in postmortem samples for forensic studies and in blood samples for clinical toxicology. References [1] A.M.N. Ahasan, A.A. Mamun, S.R. Karim, M.A. Bakar, E.A. Gazi, C.S. Bala, Paralytic complications of puffer fish (tetrodotoxin) poisoning, Singapore Med. J. 45 (2004) 73–74. [2] T. Narahashi, Tetrodotoxin, Proc. Jpn. Acad. Ser. B 84 (2008) 147–154. [3] S.K. Chew, C.H. Goh, K.W. Wang, P.K. Mah, B.Y. Tan, Puffer fish (tetrodotoxin) poisoning: clinical report and role of anti-cholinesterase drugs in therapy, Singapore Med. J. 24 (1983) 168–171. [4] Tetradotoxin poisoning associated with eating puffer fish transported from JapanCalifornia, 1996, Morb. Mortal. Wkly. Rep. 45 (1996) 389–391.
[5] F.R. Chowdhury, H.A.M. Nazmul Ahasan, A.K.M. Mamunur Rashid, A. Al Mamun, S.M. Khaliduzzaman, Tetrodotoxin poisoning: a clinical analysis role of neostigmine and short-term outcome of 53 cases, Singapore Med. J. 48 (2007) 830–833. [6] C.K. How, C.H. Chern, Y.C. Huang, L.M. Wang, C.H. Lee, Tetrodotoxin Poisoning, Am. J. Emerg. Med. 21 (2003) 51–54. [7] T. Yasumoto, T. Michishita, Fluorometric determination of tetrodotoxin by high performance liquid chromatography, Agric. Biol. Chem. 49 (1985) 3077–3080. [8] M. Kono, T. Matsui, K. Furukawa, T. Takase, K. Yamamori, H. Kaneda, D. Aoki, J.H. Jang, M. Yotsu-Yamashita, Examination of transformation among tetrodotoxin and its analogs in the living cultured juvenile puffer fish, kusafugu Fugu niphobles by intramuscular administration, Toxicon 52 (2008) 714–720. [9] Y.H. Tsai, D.F. Hwang, C.A. Cheng, C.C. Hwang, J.F. Deng, Determination of tetrodotoxin in human urine and blood using C18 cartridge column, ultrafiltration and LC–MS, J. Chromatogr. B 832 (2006) 75–80. [10] S. Kurono, H. Hattori, O. Suzuki, T. Yamada, H. Seno, Sensitive analysis of tetrodotoxin in human plasma by solid-phase extractions and gas chromatography/mass spectrometry, Anal. Lett. 34 (2001) 2439–2446. [11] T. Noguchi, H. Narita, J. Maruyama, K. Hashimoto, Tetrodotoxin in the starfishes Astropecten polyacanthus, in association with toxification of a trumpet shell Boshubora Charonia sauliae, Bull. Japan. Soc. Sci. Fish. 48 (1982) 1173–1177. [12] K. Kawatsu, T. Shibata, Y. Hamano, Application of immunoaffinity chromatography for detection of tetrodotoxin from urine samples of poisoned patients, Toxicon 37 (1999) 325–333. [13] M. Horie, S. Kobayashi, N. Shimizu, H. Nakazawa, Determination of tetrodotoxin in puffer-fish by liquid chromatography–electrospray ionization mass spectrometry, Analyst 127 (2002) 755–759. [14] T. Nakagawa, J. Jang, M. Yotsu-Yamashita, Hydrophilic interaction liquid chromatography–electrospray ionization mass spectrometry of tetrodotoxin and its analogs, Anal. Biochem. 352 (2006) 142–144. [15] H.C. Jen, S.J. Lin, Y.H. Tsai, C.H. Chen, Z.C. Lin, D.F. Hwang, Tetrodotoxin poisoning evidenced by solid-phase extraction combining with liquid chromatography– tandem mass spectrometry, J. Chromatogr. B 871 (2008) 95–100. [16] M.A. O’Leary, J.J. Schneider, G.K. Isbister, Use of high performance liquid chromatography to measure tetrodotoxin in serum and urine of poisoned patients, Toxicon 44 (2004) 549–553. [17] H.-N. Huangb, J. Lin, H.-L. Lin, Identification and quantification of tetrodotoxin in the marine gastropod Nassarius by LC–MS, Toxicon 51 (2008) 774–779. [18] Y. Shoji, M. Yotsu-Yamashita, T. Miyazawa, T. Yasumoto, Electrospray ionization mass spectrometry of tetrodotoxin and its analogs: liquid chromatography/mass spectrometry tandem mass spectrometry, and liquid chromatography/tandem mass spectrometry, Anal. Biochem. 290 (2001) 10–17. [19] O.R. Pires Jr., A. Sebben, E.F. Schwartz, R.A. Morales, C. Bloch Jr., C.A. Schwartz, Further report of the occurrence of tetrodotoxin and new analogues in the Anuran family Brachycephalidae, Toxicon 45 (2005) 73–79. [20] X. Wang, W. Li, H.T. Rasmussen, Orthogonal method development using hydrophilic interaction chromatography and reversed-phase high-performance liquid chromatography for the determination of pharmaceuticals and impurities, J. Chromatogr. A 1083 (2005) 58–62.