Rapid and simple determination of psychotropic phenylalkylamine derivatives in urine by direct injection liquid chromatography–electrospray ionization–tandem mass spectrometry

ABSTRACT

A direct injection liquid chromatography–electrospray ionization–tandem mass spectrometric method (LC-ESI-MS/MS) was developed and validated for the rapid and simple determination of 13 phenylalkylamine derivatives. Eight deuterium-labeled compounds were prepared for use as internal standards (ISs) to quantify the analytes. Urine samples mixed with ISs were centrifuged, filtered through 0.22 µm filters and then injected directly into the LC-ESI-MS/MS system. The mobile phase was composed of 0.2% formic acid and 2 mM ammonium formate in distilled water and 0.2% formic acid and 2 mM ammonium formate in acetonitrile. The analytical column was a Capcell Pak MG-II C₁₈ (150 × 2.0 mm i.d., 5 µm, Shiseido). Separation and detection of the analytes were accomplished within 10 min. The linear ranges were 5–750 ng/mL (ephedrine and fenfluramine), 10–750 ng/mL (3,4-methylenedioxyamphetamine, phendimetrazine, methamphetamine, 3,4-methylenedioxyethylamphetamine and benzphetamine), 20–750 ng/mL (norephedrine, amphetamine, phentermine and ketamine) and 30–1000 ng/mL (3,4-methylenedioxymethamphetamine and norketamine), with determination coefficients, R², ≥ 0.9967. The intra-day and inter-day precisions were within 19.1%. The intra-day and inter-day accuracies ranged from −16.0 to 18.7%. The lower limits of quantification for all the analytes were lower than 26.5 ng/mL. The applicability of the method was examined by analyzing urine samples from drug abusers (n = 30). Copyright © 2012 John Wiley & Sons, Ltd.

Abbreviations used

AP

amphetamine

ATS

amphetamine-type stimulant

BZP

benzphetamphetamine

EP

ephedrine

FFA

fenfluramine

KET

ketamine

MDA

3,4-methylenedioxyamphetamine

MA

methamphetamine

MDEA

3,4-methylenedioxyethylamphetamine

MDMA

3,4-methylenedioxymethamphetamine

NEP

norephedrine

NKT

norketamine

PDM

phendimetrazine

PT

phentermine

Introduction

Amphetamine-type stimulants (ATSs) are a group of synthetic substances structurally derived from phenylethylamine. ATSs such as amphetamine (AP), methamphetamine (MA), 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethamphetamine (MDMA) and 3,4-methylenedioxyethylamphetamine (MDEA) are psychoactive substances with stimulant or hallucinogenic properties (Glennon, 1989). MA is the most frequently abused drug. MDMA has been widely used as a recreational drug (Aronson, 2008; Cho, 1991). The increasing abuse of MA and MDMA has been a serious social problem in Korea (Supreme Prosecutors’ Office, 2004, 2011). Phendimetrazine (PDM), phentermine (PT), fenfluramine (FFA) and benzphetamine (BZP) are classified as appetite suppressants and have been used to treat obesity. These compounds are also controlled drugs that have a high potential for abuse and addiction (Jain et al., 1979; Musshoff, 2000; Weintraub et al., 1992). BZP is chemically related to AP. Tablets containing BZP in combination with trifluoromethylphenylpiperazine are abused for psychoactive effects. Ketamine (KET) has been abused since the mid 2000s as a party drug. Although KET is not as common as MDMA, its abuse is gaining popularity in Korea (Kim et al, 2006; Moore et al., 1997).

Rapid increases in abuse of ATSs and ketamines require convenient and rapid analytical detection methods in biological samples for the purpose of forensic toxicology. To date, several chromatographic methods have been reported, including gas chromatography–mass spectrometry (GC-MS; Kerrigan et al., 2011; Lu et al., 2010), capillary electrophoresis–mass spectrometry (Iio et al., 2003) and liquid chromatography–tandem mass spectrometry (LC-MS/MS) (Andersson et al., 2002; Fuh et al., 2006; Liu et al., 2010; Lu et al., 2010). LC-MS/MS technique is considered a powerful alternative to GC-MS. The main advantages of the technique are simplified sample preparation and direct measurement of hydrophilic analytes, eliminating hydrolysis and derivatization procedures. The direct injection of samples without labor-intensive sample preparation and sample loss leads to shortened time to results. Owing to its usefulness, a direct injection LC-MS/MS method has been reported for measuring opiates and ATSs (Andersson et al., 2008; Edinboro et al., 2005; Gustavsson et al., 2007).

There have been continual demands for rapid drug tests for a large number of urine samples and multi-component analysis. LC-MS/MS technique was used for multi-component screening purposes to reduce the time of sample preparation and speed up the analysis (Björnstada and Beck, 2009; Nordgren and Beck, 2004). The aim of the study was to develop a rapid and simple method for simultaneous determination of 13 phenylalkylamine derivatives using direct injection LC-ESI-MS/MS. The method was validated and its applicability was confirmed by analysis of authentic urine samples.

Materials and methods

Chemicals

The reference compounds of AP, MDA, MA, MDMA, PT, NKT, KET, MDEA and FFA were obtained from Cerilliant (Austin, TX, USA). NEP, EP, PDM and BZP were obtained from Sigma-Aldrich (St Louis, MO, USA). The deuterated internal standards (ISs) AP-d₈, MA-d₁₁, MDA-d₅, MDMA-d₅, NKT-d₄, KET-d₄ and MDEA-d₅ were also obtained from Cerilliant. EP-d₃ was purchased from Sigma-Aldrich. Figure 1 shows the chemical structures and molecular weights of the analytes as well as the corresponding ISs. HPLC-grade acetonitrile was purchased from J. T. Baker (Phillipsburg, NJ, USA). The water was purified using a Direct-Q water purification system (Millipore, Bedford, MA, USA). Extrapure-grade formic acid was obtained from Sigma-Aldrich (Fluka, Switzerland). All other chemicals were of analytical grade. Nylon syringe filters with a 0.22 µm pore size were purchased from Restek (Bellefonte, PA, USA).

Figure 1. Chemical structure, abbreviation and molecular weight of the analytes.

Preparation of solutions

The working standard solutions (0.1, 1.0 and 10.0 µg/mL) of each compound were prepared by an appropriate dilution in methanol. The ISs were prepared in methanol to provide a working standard solution of 0.5 µg/mL. All of these solutions were stored at −20 °C before use.

0.2% formic acid solution and 2 m m ammonium formate in distilled water, and 0.2% formic acid and 2 mM ammonium formate in acetonitrile were used as the mobile phase. Each solution was sonicated for 20 min in an ultrasonic bath and filtered using 0.45 µm membrane filters (Millipore, USA).

Urine specimens

Drug-free urine samples were collected from laboratory staff. Urine samples (n = 30) collected from drug abusers were obtained from the Narcotics Departments at the District Prosecutors’ Offices in the Seoul metropolitan area. Samples were stored in a Hetofrig CL 410 deep freezer (Heto, Denmark) at −55 °C until analysis.

Sample preparation

Urine specimens (1 mL) including 30 μL of the IS solution (0.5 µg/mL) were centrifuged to obtain clear supernatants at 50,000 g for 5 min in a Sigma 3-30 K centrifuge. Aliquots (60 μL) of the centrifugated urine samples were filtered through a 0.22 µm Nylon syringe filter with a diameter of 13 mm. An aliquot (5 μL) of the sample solution was directly injected into the LC-ESI-MS/MS system for analysis.

Instruments

The HPLC system consisted of an Agilent 1200 series handheld control module, binary gradient pump, a vacuum degasser, an autosampler and a thermostatted column compartment (Palo Alto, CA, USA). The analytical column was a Capcell Pak MG-II C₁₈ (150 × 2.0 mm i.d., 5 µm, Shiseido, Japan). To protect the analytical column, a guard column (10 × 2.0 mm i.d., 5 µm) containing the same material as the analytical column was fitted between the analytical column and the autosampler. Mobile phases A and B were 0.2% formic acid and 2 m m ammonium formate in distilled water and 0.2% formic acid and 2 mM ammonium formate in acetonitrile, respectively. HPLC column flow rate was increased from 200 to 300 μL/min at 12 min. The B content of the mobile phase was increased from 10 to 60% at 12 min, followed by an additional increase from 60 to 90% at 13.5 min, and then an isocratic condition at 10% B for 2.5 min was used. The column was equilibrated at 10% B for 3 min.

The HPLC was coupled to an API 3200 QTrap triple-quadrupole mass spectrometer (AB Sciex, Foster City, CA, USA) equipped with a TurboIonSpray source. Electrospray ionization (ESI) was carried out in the positive mode using nitrogen as the nebulizing, turbo spray and curtain gas, with the optimum values set to 25, 45 and 70 (arbitrary units). The turbo gas temperature and the spraying needle voltage were set to 500 °C and 5500 V, respectively. The mass spectrometer was operated with unit (0.7 full width at half height) resolution for quadrupoles Q1 and Q3, respectively. Multiple reaction monitoring (MRM) detection was achieved using nitrogen as the collision gas (low, arbitrary units) with scheduled MRM transitions without increasing the MRM cycle time. Table 1 lists the monitored MRM transitions that were monitored, which included the fragmentations of the target molecules and the collision energy. Analyst 1.5.1 software was used for equipment control, data acquisition and analysis.

Table 1. Retention times, multiple reaction monitoring transitions and compound-dependent parameters for LC-ESI-MS/MS analysis of the analytes and internal standards

Compound	tRa (min)	Precursor ion	Product ion	DPb (V)	FPc (v)	EPd (V)	CEPe (V)	CEf (V)
^a tR, Retention time; ^b DP, declustering potential; ^c FP, focusing potential; ^d EP, entrance potential; ^e CEP, collision cell entrance potential; ^f CE, collision energy; ^g product ion underlined was used for quantification.
NEP	4.11	152.2	134.2g	16	4.5	14	13	4
		152.2	117.1	16	4.5	14	25	4
EP	5.17	166.1	91.1	21	5	14	41	4
		166.1	133.1	21	5	14	29	4
EP-d3	5.15	169.1	151.2	21	6.5	12	17	4
AP	6.00	136.2	91.1	21	8	14	23	4
		136.2	119.1	21	8	14	11	4
AP-d8	5.91	144.2	97.2	21	5	14	23	4
MDA	6.35	180.2	163.1	21	3.5	12	13	4
		180.2	105.0	21	3.5	12	31	4
MDA-d5	6.32	185.2	168.2	21	3.5	12	13	4
PDM	6.37	192.2	146.1	46	4.5	16	33	4
		192.2	91.1	46	4.5	16	45	4
MA	6.45	150.2	119.2	26	7.5	12	13	4
		150.2	65.0	26	7.5	12	55	4
MA-d11	6.35	161.2	97.2	26	7	14	25	4
MDMA	6.69	194.1	163.1	26	4	12	15	4
		194.1	105.1	26	4	12	33	4
MDMA-d5	6.67	199.2	165.2	26	4	12	17	4
PT	6.76	150.2	91.1	21	4.5	14	27	4
		150.2	133.2	21	4.5	14	13	4
NKT	6.93	224.1	125.1	26	3.5	18	33	4
		224.1	207.1	26	3.5	18	15	4
NKT-d4	6.91	228.2	129.1	31	3.5	14	29	4
KET	7.13	238.2	125.1	36	3	14	37	4
		238.2	179.1	36	3	14	21	4
KET-d4	7.10	242.2	129.1	31	3.5	14	37	4
MDEA	7.17	208.2	163.1	31	4	12	17	4
		208.2	105.1	31	4	12	35	4
FFA	9.44	232.2	159.1	36	3.5	16	33	4
		232.2	109.1	36	3.5	16	55	4
BZP	9.82	240.3	91.1	36	4	14	41	4
		240.3	119.0	36	4	14	21	4
MDEA-d6	7.14	214.3	166.2	26	4	14	17	4

Method validation

To evaluate selectivity, 10 different drug-free samples were analyzed for evaluation of potential interferences from the matrix ion suppression. The apparent response at the retention times of the analytes under investigation was compared with the response of analytes at the limit of quantification.

The potential for carryover was evaluated by injecting the highest point of the calibration curve, followed by blank urine, and the area of peaks present at the retention times of analytes under investigation was measured. The lower limit of detection (LLOD) and quantification (LLOQ) were calculated from the signal-to-noise ratios (S/N) 3 and 10, respectively, by analyzing 10 drug-free urine samples.

The linearity of the method was confirmed over the concentration range of 5–750 ng/mL (EP and FFA), 10–750 ng/mL (MDA, PDM, MA, MDEA and BZP), 20–750 ng/mL (NEP, AP, PT and KET) and 30–1000 ng/mL (MDMA and NKT). Calibration was assessed using seven-point calibration curves at concentrations of 5, 10, 50, 100, 250, 500 and 750 ng/mL (EP and FFA), 10, 20, 50, 100, 250, 500 and 750 ng/mL (MDA, PDM, MA, MDEA and BZP), 20, 30, 50, 100, 250, 500 and 750 ng/mL (NEP, AP, PT and KET) and 30, 50, 100, 250, 500, 750 and 1000 ng/mL (MDMA and NKT). Each calibration point was run in triplicate to ensure the precision of the system. Linear regression analysis was performed on the peak area ratios of the analyte to IS vs the analyte concentrations.

Spiked quality control (QC) samples at three different concentrations were prepared by spiking distilled water with known amounts of each analyte (15, 40, 150 and 600 ng/mL). The intra-day precision and accuracy of the method was established by replicate analyses (n = 6) of the QC samples. The inter-day precision and accuracy was established by replicate analyses of the same QC samples on four different days. To determine the precision, the relative standard deviation (RSD) was calculated for the replicate measurements. The accuracy (%bias) was calculated from the degree of agreement between the measured and nominal concentrations of the spiked urine samples.

Results and discussion

Optimization of sample preparation and HPLC conditions

Urine is highly useful as a testing matrix. It is easily accessible, can be collected noninvasively and contains numerous potential biomarkers, including metabolites, salts, urea, proteins and nucleic acids. However, matrix components not removed during sample preparation may cause ion suppression or matrix effects. To minimize these difficulties, stable isotope-labeled ISs were used to normalize the effects of ion suppression. High-speed centrifugation at 50,000 g was applied for fast and efficient clean-up in the study.

The chromatographic separation of the analytes was optimized to improve the peak shape and retention characteristics. Optimization of the chromatographic condition was established as follows. In order to improve the separation capacity, 0.2% formic acid and 2 mM ammonium formate were added into the mobile phases. Isocratic conditions were used in the initial development of the HPLC method. As severe peak tailing was observed under the isocratic conditions, gradient conditions were investigated to overcome this drawback. Optimal gradient conditions for the separation of the compounds were obtained using mobile phase gradient elution. The mobile phase used was 2 mM ammonium formate and 0.2% formic acid in distilled water and 2 mM ammonium formate and 0.2% formic acid in acetonitrile. The retention time of the analyte is shown in Table 1. Separation and detection of the analytes was accomplished within 10 min. The run time of the assay was 13.5 min excluding a 3 min equilibrium cycle programmed to run before the subsequent injection. Since the retention time of the last eluting analyte was 9.82 min, the equilibrium cycle and the run time of the assay could be further reduced.

LC-MS/MS analysis

For all analytes, prominent [M + H]⁺ ions were selected as the precursor ion. The two fragment ions of the precursor ion provided the maximal response and were used for the LC-ESI-MS/MS quantitative measurement of the analytes. Analogous MS/MS fragmentations were observed for the deuterium-labeled ISs.

Instrumental conditions were optimized by a flow-injection method employing a methanolic solution of the analyte. Identification of analytes was achieved in the MRM mode using at least two characteristic transitions each, the respective analytes’ retention times and co-elution of stable isotopic analogues. Target peaks were confirmed by comparing the retention time with those of the reference standards, and the order of elution was clearly observed on the MRM chromatograms (Fig. 2). Table 1 lists the monitored MRM transitions, retention time and respective mass spectrometric parameters of the analytes and corresponding ISs.

Figure 2. Representative multiple reaction monitoring chromatograms including (A) drug-free urine without internal standards, (B) drug-free urine, (C) spiked urine containing 30 ng/mL of each analyte, and (D) drug positive urine samples.

Method validation

Method validation included the evaluation of selectivity, carryover, linearity, LLOD, LLOQ, precision and accuracy. The selectivity of the method was tested to assess the matrix effect and measured by comparing the chromatograms of different lots of drug-free urine and spiked urine. All drug-free urine lots were found to be free of interferences at the retention time of the target MRM transitions except for EP. The MRM analysis of drug-free urine using transition pair (m/z 166.1 → 148.2) shows the presence of strong interferences at the same retention time of EP. Therefore two MRM transition pairs (m/z 166.1 → 91.1 and m/z 166.1 → 133.1) were selected for peak identification and quantification of EP.

Table 2 summarizes the linear dynamic ranges, linearity, LLOD and LLOQ data. Seven-point calibration curves for the analytes were established from three replicates at each concentration. The determination coefficients (R²) were above 0.9967 for all 13 analytes, indicating good linear regression. The sensitivity of the method was evaluated by determining the LLOD and LLOQ for the analyte. The LODs and LOQs ranged from 0.9 to 7.9 and from 3.1 to 26.5 ng/mL for the analytes, respectively.

Table 2. Method calibration

Analyte	Concentration range (ng/mL)	Slope	y-Intercept	Linearitya (R^2)	LLODb (ng/mL)	LLOQ^c (ng/mL)
^a Linearity is described by the correlation coefficient for the calibration curve. ^b Lower limit of detection (LLOD) and ^c lower limit of quantification (LLOQ) were based on the concentration corresponding to a signal plus 3 and 10 standard deviations above from the mean of seven replicates of drug-free urine, respectively.
NEP	20–750	0.0286	−0.2137	0.9983	3.9	12.9
EP	5–750	0.0111	0.0170	0.9991	0.9	3.1
AP	20–750	0.0597	0.5895	0.9981	5.6	18.8
MDA	10–750	0.0638	0.2024	0.9992	2.3	7.8
PDM	10–750	0.0115	−0.0216	0.9990	2.3	7.8
MA	10–750	0.0309	0.1160	0.9992	2.1	6.9
MDMA	30–1000	0.0720	−0.0664	0.9991	7.9	26.5
PT	20–750	0.0479	0.2185	0.9984	5.3	17.6
NKT	30–1000	0.0680	0.4040	0.9986	6.9	23.1
KET	20–750	0.0565	0.4105	0.9982	3.4	11.4
MDEA	10–750	0.0800	0.0587	0.9988	2.0	6.6
FFA	5–750	0.0577	−0.3359	0.9980	1.2	4.1
BZP	10–750	0.1000	−0.6971	0.9967	2.0	6.6

In order to evaluate potential carry-over, the highest extracted calibrator was injected into the LC-ESI-MS/MS instrument, followed by an acetonitrile blank to determine if the results of the solvent blank injections were affected by the previous injection. It was found that there is no carry-over at the retention time of analytes and ISs. Subsequently, acetonitrile blanks were used throughout the sample sequence to confirm that no sample-to-sample contamination had occurred.

Table 3 summarizes the quantitative validation parameters. These data were within the acceptance criteria of ±20% of nominal concentration for low QC concentrations and ±15% of nominal concentration for middle and high QC concentrations. The intra- and inter-day precision and accuracy were obtained by analyzing seven replicates of three or four different spiked urine samples (15, 40, 150 and 600 ng/mL) for each analyte using the procedure described above. The intra- and inter-day precisions were within 19.1 and 12.8%, respectively, whereas the intra- and inter-day accuracies were between −13.9 and 18.7% and between −16.0 and 18.1%, respectively. The results were satisfactory when considering the complexity of the samples.

Table 3. Intra- and inter-day precision and accuracy

Analyte	Nominal concentration (ng/mL)	Intra-day (n = 3)		Inter-day (n = 4)
		Precisiona (%RSD)	Accuracyb (%bias)	Precision (%RSD)	Accuracy (%bias)
^a Expressed as the coefficient of variance of the peak area ratios of analyte/internal standard. ^b Calculated as [(mean calculated concentration – nominal concentration)/nominal concentration] × 100.
NEP	40	7.4	−5.0	3.8	−11.4
	150	5.5	−13.9	2.4	−13.3
	600	9.6	2.6	1.9	−3.2
EP	15	19.1	−0.8	4.9	−4.5
	40	7.3	−3.1	2.2	−9.8
	150	5.5	−3.4	2.1	−5.0
AP	40	8.6	−3.1	3.3	−16.0
	150	4.7	7.7	1.8	5.6
	600	6.3	7.1	6.6	4.7
MDA	15	17.3	−10.0	3.9	−15.8
	40	10.2	−2.5	3.4	−9.4
	150	1.9	0.7	1.8	−0.6
PDM	15	5.7	−4.1	3.8	3.8
	40	4.5	−10.4	12.8	−6.1
	150	4.6	−3.2	4.2	−7.3
MA	15	16.8	0.3	6.4	−10.1
	40	13.5	−1.6	8.1	−6.6
	150	4.6	4.8	4.6	1.7
MDMA	40	12.1	4.1	3.6	−4.9
	150	8.8	0.3	4.0	−4.3
	600	7.6	6.0	8.0	−0.2
PT	40	14.1	12.9	6.3	16.1
	150	13.0	11.2	4.1	14.6
	600	0.5	9.7	0.7	9.3
NKT	40	8.4	−4.0	1.6	−11.1
	150	9.7	2.9	3.5	2.2
	600	4.6	8.0	3.7	3.6
KET	40	10.2	−3.9	3.7	−12.8
	150	5.5	5.8	4.2	2.3
	600	3.9	1.6	3.0	−0.2
MDEA	15	18.1	−0.9	8.4	−9.2
	40	6.8	2.7	3.9	−6.4
	150	3.3	3.2	4.0	−0.3
FFA	15	7.0	13.4	2.5	14.8
	40	5.1	−8.3	1.7	−13.6
	150	8.6	−9.0	3.7	−14.7
BZP	15	1.5	18.7	3.7	18.1
	40	4.7	−6.4	3.0	−11.2
	150	8.4	−8.1	4.0	−11.1

Application of the method to the urine samples of drug abusers

In order to demonstrate the applicability of the developed method to real forensic samples, the method was used to analyze the urine samples from drug abusers. Figure 2(D) shows a representative chromatogram of urine samples obtained from drug abusers. A total of 30 urine samples were analyzed and quantified. The results are reported as the number of analyzed samples that tested positive to NEP, EP, AP, MA, PT and PDM (Table 4). MA was the most frequently abused drug in association with its major metabolite AP. Its abuse was occasionally related to multi-drug consumption in combination with commonly used prescriptions or over-the-counter drugs. Among the 11 samples containing MA and AP tested, multiple drug use included EP, NEP, PT and PDM in seven samples. As it is widely used in over-the-counter cough and cold preparations, EP and its metabolite, NEP, could easily be detected in urine samples. PT and PDM have been widely used as anti-obesity drugs. Among the urine samples, 13 were positive for PT and four for PDM. These results show that the method is suitable for rapid and simultaneous determination of several types of abused drugs.

Table 4. Quantitative results of the analytes in urine samples obtained from drug users

Analyte	Analyzed samples	Positive samples	Range (ng/mL)	Mean (ng/mL)
^a Numbers in parentheses refer to urine samples above the highest calibrator concentration (750 ng/mL) of each analyte.
NEP	30	17(6)a	13.3–556.1	148.3
EP	30	14(10)	7.5–680.4	193.1
AP	30	11	26.3–242.9	91.1
MA	30	11(1)	66.1–524.5	235.4
PT	30	13(12)	228.1	—
PDM	30	4	54.7–389.3	226.0

Conclusions

In this study, a rapid and simple LC-ESI-MS/MS method for the determination of 13 analytes in urine was developed and validated. High-speed centrifugation at 50,000 g produced cleaner supernatants and reduced the interference from the chemical background noise. The direct injection was crucial step to decreasing sample preparation time and sample loss. The direct injection analysis was also successfully applied to LC-ESI-MS/MS with higher simplicity and sensitivity for all analytes. The method seems to be advantageous to forensic toxicologists, who are frequently confronted with volume-limited samples and need to detect multiple drug use. The applicability was proven by analyzing 30 authentic urine samples.

Acknowledgments

This study was supported in part by a grant (M10640010000-06 N4001-00100) from the National R&D program of Ministry of Education, Science and Technology and National Research Foundation of Korea. There is no conflict of interest to declare.