Pharmacokinetics of (synthetic) cannabinoids in pigs and their relevance for clinical and forensic toxicology

Pharmacokinetics of (synthetic) cannabinoids in pigs and their relevance for clinical and forensic toxicology

Toxicology Letters 253 (2016) 7–16 Contents lists available at ScienceDirect Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet Ph...

1MB Sizes 3 Downloads 63 Views

Toxicology Letters 253 (2016) 7–16

Contents lists available at ScienceDirect

Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet

Pharmacokinetics of (synthetic) cannabinoids in pigs and their relevance for clinical and forensic toxicology Nadine Schaefera,* , Jan-Georg Wojtyniakb , Mattias Kettnera , Julia Schlotea , Matthias W. Laschkec, Andreas H. Ewalda , Thorsten Lehrb , Michael D. Mengerc , Hans H. Maurerd, Peter H. Schmidta a

Institute of Legal Medicine, Saarland University, Building 80.2, D-66421 Homburg, Saarland, Germany Clinical Pharmacy, Saarland University, D-66123 Saarbruecken, Germany c Institute for Clinical & Experimental Surgery, Saarland University, Building 65/66, D-66421 Homburg, Saarland, Germany d Department of Experimental and Clinical Toxicology, Saarland University, Building 46, D-66421 Homburg, Saarland, Germany b

H I G H L I G H T S

G R A P H I C A L A B S T R A C T

 First study on the pharmacokinetics of JWH-210 and RCS-4 in pigs.  A three-compartment model described best pharmacokinetic data of THC, JWH-210, and RCS-4.  The allometrically upscaled THC pig model resulted in successful prediction of human exposure.  Pigs useful for prediction of human pharmacokinetics of synthetic cannabinoids.

A R T I C L E I N F O

Article history: Received 24 February 2016 Received in revised form 19 April 2016 Accepted 22 April 2016 Available online 23 April 2016 Keywords: Synthetic cannabinoids Tetrahydrocannabinol LC–MS/MS Population pharmacokinetic modeling Pigs Prediction of human pharmacokinetics

A B S T R A C T

Synthetic cannabinoids (SCs) are gaining increasing importance in clinical and forensic toxicology. They are consumed without any preclinical safety studies. Thus, controlled human pharmacokinetic (PK) studies are not allowed, although being relevant for interpretation of analytical results in cases of misuse or poisoning. As alternative, in a controlled animal experiment, six pigs per drug received a single intravenous dose of 200 mg/kg BW each of D9-tetrahydrocannabinol (THC), 4-ethylnaphthalen-1-yl-(1pentylindol-3-yl)methanone (JWH-210), or 2-(4-methoxyphenyl)-1-(1-pentyl-indol-3-yl)methanone (RCS-4). In addition, six pigs received a combination of the three drugs with the identical dose each. The drugs were determined in serum using LC–MS/MS. A population (pop) PK analysis revealed that a threecompartment model described best the PK data of all three cannabinoids. Central volumes of distribution were estimated at 0.29 L/kg, 0.20 L/kg, and 0.67 L/kg for THC, JWH-210, and RCS-4, respectively. Clearances were 0.042 L/min/kg, 0.048 L/min/kg, and 0.093 L/min/kg for THC, JWH-210, and RCS-4, respectively. The popPK THC pig model was upscaled to humans using allometric techniques. Comparison with published human data revealed that the concentration-time profiles could successfully be predicted. These findings indicate that pigs in conjunction with PK modeling technique may serve as a tool for prediction of human PK of SCs. ã 2016 Elsevier Ireland Ltd. All rights reserved.

* Corresponding author. E-mail address: [email protected] (N. Schaefer). http://dx.doi.org/10.1016/j.toxlet.2016.04.021 0378-4274/ ã 2016 Elsevier Ireland Ltd. All rights reserved.

8

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

1. Introduction Synthetic cannabinoids (SCs) have emerged on the drugs of abuse market sold via internet as herbal mixtures or incenses, often in unknown and unexpected compositions and concentrations. Due to the fact that many of these drugs have a more potent binding strength to cannabinoid receptors than D9tetrahydrocannabinol (THC) (Huffman et al., 2003; Makriyannis et al., 2001), a consumption can lead to strong psychoactive and unpredictable toxic effects. Numerous cases of intoxications have been reported with life-threatening conditions accompanied e.g. by tachycardia, somnolence, anxiety, nausea, vomiting, seizures, hyperglycemia, hypokalemia, agitation, hallucination, and acute psychosis resulting in admission to hospital (Harris and Brown, 2013; Hermanns-Clausen et al., 2013; Kronstrand et al., 2013; Schneir and Baumbacher, 2012), or even in death (Behonick et al., 2014; Kronstrand et al., 2013; Patton et al., 2013; Schaefer et al., 2013; Shanks et al., 2012). Therefore, SCs are gaining an increasing importance in clinical and forensic toxicology. For interpretation of analytical data of impaired or poisoned persons concerning e.g. time of intake or concentration at a particular time relevant for legal reasons, pharmacokinetic (PK) data are necessary. However, respective data of controlled studies are not available as these compounds are sold and consumed without safety pharmacological tests. Only data from biotransformation studies (Grigoryev et al., 2011; Hutter et al., 2012; Kavanagh et al., 2012), single case reports (Dresen et al., 2011; Hermanns-Clausen et al., 2013; Kronstrand et al., 2013), or self-experiments have been reported so far. A patient participating in an Intramural Research Board approved research study smoked an herbal incense that contained 17 mg/g JWH-018 and 22 mg/g JWH-073 and blood samples were taken after 19, 53, 107, and 199 min. Peak whole blood JWH018 and JWH-073 concentrations of about 5 ng/mL were detected after 19 min. After 199 min, concentrations had decreased below 1 ng/mL (Kacinko et al., 2011). In a self-experiment two subjects (one male and one female) smoked a cigarette containing 100 (volunteer one) and 150 (volunteer two) mg of the incense “Smoke”. This incense was found to contain JWH-018 and the smoked dose was equal to an approximately 50 mg/kg BM dose of JWH-018. Blood samples were drawn 5, 15, and 60 min as well as 3, 12, 24, and 48 h after the consumption. Peak serum JWH018 concentrations of about 10 ng/mL were found 5 min after administration and traces were still present after 48 h (Teske et al., 2010). In the third study, a self-experiment was conducted by oral administration of a gelatin capsule containing 5 mg of AM2201 and serum and urine specimens were collected for 11 days. A peak serum AM-2201 concentration of 0.56 ng/mL was determined 1 h and 35 min post-ingestion and the drug remained detectable for 5 days (Hutter et al., 2013). Nevertheless, due to the limited number of sampling points and a very small collective of subjects, these studies provide only insufficient information of SCs PK and should be supplemented by systematic studies, including a larger number of individuals. As systematic controlled human studies have not been performed, PK properties should be assessed in controlled animal studies. There is only one published controlled animal study providing PK properties of the SC WIN 55,212-2. The substance was administered as a 150 mg/kg intravenous (i.v.) dose to seven guinea pigs and plasma samples were obtained for 8 h (Valiveti et al., 2004). Experiments using small animals such as rodents are hampered by their little blood volume, not allowing for multiple blood sampling. As a consequence, a larger number of animals would be needed and complete kinetics could not be elucidated in the same animal. Pigs as a large mammalian species, however, allow for clarification of different issues in the same animal. They are suitable for extensive specimen sampling. Furthermore, pigs

are closely related to the human species in terms of e.g. metabolism including cytochrome P450 (CYP) enzyme pattern (Anzenbacher et al., 1998), anatomical structures as well as physiological properties regarding e.g. cardiovascular, urogenital, and digestive systems (Bode et al., 2010; Swindle et al., 2012). In addition, they are omnivores, sensitive to a wide range of drugs and chemicals, and all routes of administration are possible using the pig (Bode et al., 2010; Svendsen, 2006; Swindle et al., 2012). Thus, alternatively to dogs and monkeys, pigs are increasingly used in preclinical toxicological testing of pharmaceuticals (Swindle et al., 2012) and they are a common model in pharmacological studies, especially to assess PK properties of substances (Mogi et al., 2012; Shimshoni et al., 2015; Sjögren et al., 2012). SCs are closely related to THC and PK of THC has already extensively been described in the literature using different animal models (Garrett and Hunt, 1977; Leuschner et al., 1986) and also in human studies (Huestis et al., 1992; Hunt and Jones, 1980; Lindgren et al., 1981; Ohlsson et al., 1982) after different routes of administration. Recently, Brunet et al. (2006, 2010) using a small number of animals and a non-compartmental PK approach developed a pig model, which is suitable for cannabinoid PK studies after i.v. administration. They suggested that the animal data can be compared to findings of controlled human studies (Huestis, 2002; Huestis et al., 1992). Therefore, the aim of the present study was to elucidate whether domestic pigs can be used for prediction of human PK of SCs. For this purpose, we determined in a first step the concentration-time profiles after i.v. administration of the two selected SCs 4-ethylnaphthalen-1-yl-(1-pentylindol-3-yl)methanone (JWH-210) and 2-(4-methoxyphenyl)-1-(1pentyl-indol-3-yl)methanone (RCS-4) to domestic pigs in comparison to that of THC. In a second step, we modeled the concentration-time profiles and assessed whether this model can predict published THC data in humans. 2. Materials and methods 2.1. Chemicals and reagents The used chemicals and reagents are listed in the Electronic Supplementary material (S1). 2.2. Animals As already described in a previous study (Schaefer et al., 2015), all experiments were performed in accordance with the German legislation on protection of animals and the National Institutes of Health Guide for the Care and Use of Laboratory Animals (permission number: 69/2013). Twenty eight domestic pigs (Swabian Hall strain; mean body weight 47.7  6.4 kg) were used for the study. The animals had free access to tap water and daily standard chow. They were kept fasting a night before the experiment with free access to water. 2.3. Surgical procedures The surgical procedures have already been described elsewhere (Schaefer et al., 2015) and are described in the Electronic Supplementary material (S2). 2.4. Study design The study included five different groups. Pigs of the groups 1– 3 (n = 6 each) received the respective drug by a single administration (200 mg/kg BW each), pigs of group 4 (n = 6) a

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

combination of the three drugs (200 mg/kg BW each) and pigs of group 5 (n = 4) a placebo solution (Polysorbat 80 and sodium chloride 0.9%). Analogous to a previous study (Schaefer et al., 2015), stock solutions of 100 mg/mL THC (A), 10 mg/mL JWH-210 (B), and 10 mg/mL RCS-4 (C) were first prepared in ethanol for i.v. drug administration to pigs of the groups 1–3. The appropriate volume of each solution (A: 76.4-104.8 mL, B: 736–1116 mL, or C: 748– 1216 mL) was used to obtain a 200 mg per kg body weight dose, respectively. For i.v. drug administration to pigs of group 4, appropriate volumes of stock solution A (75.2-99.2 mL) and stock solution D (752–992 mL; containing JWH-210 and RCS-4, 10 mg/mL each) were used to obtain the identical dose. The accurate volumes were measured using a calibrated pipette (with defined measurement uncertainty) that can be adjusted to the tenth of a ml. The volumes were fortified with about 1 mL Polysorbat 80 for solubilization, filled up with sodium chloride 0.9% to a volume of 10 mL, and administered into the jugular vein. Administration was conducted over a period of 30 s, followed by a 30 s washing step using 10 mL sodium chloride 0.9% (t = 0 min). Blood samples (about 10 mL each) were drawn before and 1, 2, 5, 10, 15, 30, 45, 60, 90, 120, 180, 240, 300, and 360 min after administration. Specimens were centrifuged at 1250g for 15 min to obtain serum. All samples were stored at 20  C until analysis.

9

2.8. Prediction of human THC exposure The final THC pig model was upscaled to humans using allometric principles and the predicted human exposure was compared to actual findings obtained after i.v. administration to humans (Hunt and Jones, 1980; Lindgren et al., 1981; Ohlsson et al., 1982). Overall, 1000 simulations using the final THC PK model and fixed- and random-effects parameters were performed. The reported dosing regimens as well as the subjects’ body weight were included in the simulation scenarios. Median simulated plasma concentrations and corresponding 5th and 95th percentiles were plotted against time, and overlaid with the observed data. In addition, simulated areas under the curves (AUCs) were compared with reported AUCs. 2.9. Statistical tests An unpaired two-tailed Student’s t-test followed by an f-test was applied to compare serum drug concentrations after single and combined administration using GraphPad Prism 5.00 (GraphPad Software, San Diego, CA, USA). Data are given as mean and standard deviation. 3. Results 3.1. Concentration-time profiles

2.5. Sample preparation Specimens were prepared according to a published method described elsewhere (Schaefer et al., 2015). A description of the sample preparation is listed in the Electronic Supplementary material (S3). 2.6. LC–MS/MS LC–MS/MS conditions including instrumentation, chromatographic, and mass spectrometric conditions have already been described elsewhere (Schaefer et al., 2015) and can be seen in the Electronic Supplementary material (S4). 2.7. Population (pop) PK model development Population analyses, simulations and model evaluations were performed using non-linear mixed-effects modeling techniques (NONMEM version 7.3, ICON Development Solutions, Ellicott City, MD, USA), which allow estimation of population means (medians) for model parameters and quantify inter-individual variability (IIV) and residual (unexplained) variability. The first-order conditional estimation algorithm in NONMEM with the interaction option was used and IIV was modeled using exponential random effects models. Model selection was based on several analyses, including visual inspection of goodness-of-fit plots, precision of parameter estimates and the objective function value (OFV) provided by NONMEM. One nested model was considered superior to another when the OFV was reduced by 3.84 points (Chi2, p < 0.05, 1 of freedom). Body weight was incorporated as an exponential covariate in all PK models. For internal model evaluation, a visual predictive check (VPC) was performed based on 1000 simulations using the final PK models and fixed- and random-effects parameters. Median simulated serum concentrations and corresponding 5th and 95th percentiles were plotted against time, and overlaid with the observed data. R version 3.2.1 and higher (The R Foundation for Statistical Computing) was used for statistical analyses and generation of graphics.

The mean drug concentration-time profiles of the three substances in semi-logarithmic scale found in serum after single and combined administration indicated a triphasic decline as shown in Fig. 1A–C. These three phases consisted of a tissue distribution (a) phase, an elimination (b) phase, and a tissue release (g) phase. Based on that, half-lifes (t1/2) calculated for the a phase were 1.5 min for THC, 1.1 min for JWH-210, and 0.55 min for RCS-4. half-lifes of the b phase were 13 min for THC and 7.6 min for JWH-210 and RCS-4, respectively. Half-lifes of the g phase were found to be 121 min for THC, 122 min for JWH-210, and 137 min for RCS-4. After single administration, mean maximum concentrations (Cmax) of 1438  346 ng/mL for THC, 1600  362 ng/mL for JWH210, and 316  60 ng/mL for RCS-4 were reached immediately after administration (t = 1 min). The combined administration revealed Cmax of 1087  166 ng/mL for THC, 1474  254 ng/mL for JWH-210, and 266  89 ng/mL for RCS-4. Drug concentrations rapidly declined within the first hour. After 60 min, mean concentrations of 41.4  18.6 ng/mL for THC, 21.3  7.8 ng/mL for JWH-210, and 15.9  5.8 ng/mL for RCS-4 were observed after single administration. After combined administration, concentrations of 24.3  9.2 ng/mL for THC, 25.4  11.4 ng/mL for JWH-210, and 14.4  7.6 ng/mL for RCS-4 were determined. Afterwards, concentrations decreased more slowly. After 360 min, concentrations (Clast) of 7.1  4.3 ng/mL for THC, 3.1 1.1 ng/mL for JWH-210, and 3.8  1.1 ng/mL for RCS-4 were measured after single administration. After combined administration, Clast were 2.6  0.9 ng/mL for THC, 4.0  1.7 ng/mL for JWH-210, and 2.7  1.4 ng/mL for RCS-4. Concentrations determined after single and combined administration did not differ significantly except for Clast of THC (p < 0.05). As already published (Schaefer et al., 2015), specimens with concentrations exceeding the calibration range were diluted 1:10 and analyzed again. Neither THC nor SCs could be detected in the specimens of the four pigs, who had received the placebo solution. 3.2. popPK model development A three-compartment model best described the data of THC, JWH-210, and RCS-4 with first-order elimination processes. An

10

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

Fig. 1. Mean concentration-time profile (n = 6 each) including standard deviation of A. THC after single (orange line) and combined (red line) administration, B. JWH-210 after single (light blue line) and combined (dark blue line) administration, and C. RCS-4 after single (pink line) and combined (purple line) administration determined in pig serum. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

exponent of 0.75 on each PK parameter described the data best and was incorporated. Parameters were precisely estimated with residual standard errors <22% (Table 1). Mild to moderate IIV was identified on clearances (CL) and volumes of distribution (V)

parameters (coefficient of variation <50%) as shown in Table 1. The differential equations and parameter calculations are provided in the Electronic Supplementary material (S5). The observed and predicted serum concentration-time profiles of all animals after

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

11

Table 1 Pharmacokinetic parameters of THC, JWH-210, and RCS-4 estimated from the final three-compartment pig model. Parameter

Structural Model Parameters Vcentral CL Q2 V2 Q3 V3 Variability IIV CL IIV V2 IIV Q2 IIV Q3 IIV V3 Proportional residual error (%)

Unit

THC

JWH-210

RCS-4

Estimate

RSE (%)

Estimate

RSE (%)

Estimate

RSE (%)

(L/KG) (L/min/KG) (L/min/KG) (L/KG) (L/min/KG) (L/KG)

0.2970 0.0428 0.0541 0.5650 0.0296 2.9000

7 9 21 18 18 11

0.2020 0.0482 0.0371 3.5900 0.0363 0.2650

10 7 8 9 14 17

0.6740 0.0936 0.1520 10.500 0.4550 1.5900

22 11 7 11 15 15

(%CV) (%CV) (%CV) (%CV) (%CV)

29 n.a. 46 35 27 24

21 n.a. 38 27 25 17

21 n.a. 15 12 n.a. 25

18 n.a. 38 49 n.a. 7

35 31 n.a. n.a. 36 21

21 23 n.a. n.a. 28 7

n.a.: not applicable; RSE: relative standard error; IIV: interindividual variability; CV: coefficient of variation; Q: intercompartmental clearances, V: volume of ditribution, CL: Clearance from central compartment.

single and combined administration are depicted in the Electronic Supplementary material (S6). The observed concentrations are in good agreement with the predicted ones. Almost every observed concentration-time profile is characterized by one or more transient rise/s of concentration in the descending part of the curve. Goodness-of-fit plots (Fig. 2A) showed that the data were well described by the model. Individual and population observed versus predicted serum concentrations were randomly distributed across the line of identity, indicating good descriptive properties. The VPC (Fig. 2B) showed a good descriptive performance with neither bias nor under- or over-estimation of the model variability. 3.3. Prediction of human THC exposure The human THC exposure was predicted based on the final popPK pig model. For all three studies, the concentration-time profiles were predicted adequately (Fig. 3A–C). All observed data points were within the 95% prediction interval. The AUC of Ohlsson et al. (1982) and Lindgren et al. (1981) was underpredicted in median 20% and 29%, respectively. The median AUC from Hunt and Jones (1980) was 19% overpredicted. 4. Discussion 4.1. Dosage In this study, a 200 mg/kg BW dose of THC, JWH-210, and RCS4 was administered intravenously, resulting in a total dose of 8.9  1.3 mg. Regarding THC, this dosage may be compared to a THC quantity that is reported to be consumed by frequent users (Brunet et al., 2006). In addition, in published human THC i.v. studies a concentration of 5 mg is usually administered (Lindgren et al., 1981; Ohlsson et al., 1982). The common drug users’ dose of JWH210 and RCS-4 is in the range of 0.5-8.0 mg, depending on the route of administration (www.eve-rave.ch/Forum/viewtopic.php? t=28044. 2015). Thus, a similar dose was chosen in the present study to allow for the comparability with already published human data and to assure that the animals remained under the influence of measurable substance concentrations, as confirmed in preliminary experiments (Schaefer et al., 2015). 4.2. Concentration-time profiles The concentration-time profiles of JWH-210 and RCS-4 observed in the present study are in rather good agreement with those of

Valiveti et al. (Valiveti et al., 2004), who administered intravenously 150 mg/kg of the SC WIN 55,212-2 to guinea pigs. JWH210 and RCS-4 concentrations dropped in the same manner and were in a similar scale. Moreover, JWH-210 and RCS-4 concentrations determined in the present study were partly in the same range as those described in single case reports (Dresen et al., 2011; Hermanns-Clausen et al., 2013; Kronstrand et al., 2013). Deviations may be explained by the fact that such levels have been obtained from clinical or forensic case work for which dosage, time, and frequency of consumption were mostly unknown. However, few human studies have been conducted as self-experiments or with a very small collective of participants. Kacinko et al. (2011) detected peak whole blood JWH018 and JWH-073 concentrations of about 5 ng/mL 19 min after an individual had smoked an herbal incense known to contain 17 mg/ g JWH-018 and 22 mg/g JWH-073. After 199 min, concentrations were below 1 ng/mL. In a study by Teske et al. (2010) two volunteers smoked a cigarette containing 100 (volunteer one) and 150 (volunteer two) mg of the incense “Smoke”, including JWH018 as the active ingredient. The smoked amount was equal to an approximately 50 mg/kg BM dose of JWH-018. Peak serum JWH018 concentrations of about 10 ng/mL have been reported 5 min after administration and traces were still present after 48 h. In the third published study, Hutter et al. (2013) conducted a selfexperiment by oral administration of a gelatin capsule containing 5 mg of AM-2201. The highest measured serum AM-2201 concentration was 0.56 ng/mL 1 h and 35 min post-ingestion. AM2201 was detectable for 5 days in this study. However, substantially lower substance concentrations were found in blood in these studies, although comparable doses had been administered. In these three studies SCs have been administered by pulmonary inhalation or oral ingestion. In our study, SCs have been given by i.v. injection, which bears the advantage that the i.v. administration results in a 100% bioavailability of the substance. This means that the whole amount of unchanged drug reaches the systemic circulation and can be exactly quantified to establish a PK model. On the contrary, if a substance is smoked, the bioavailability is less than 100%, depending on different issues such as frequency, depth of puffs, and amount of pyrolytic destruction. Oral ingestion of a substance also results in less than 100% bioavailability, because the whole amount of the substance usually does not pass biomembrans what is necessary to reach the circulation and finally the target location. In addition, the amount of unchanged parent compound can be reduced during the first liver passage, where most biotransformations take place. Concentrations determined after i.v. administration must therefore be expected to be different

12

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

Fig. 2. A. GAoodness-of-fit plots from the three-compartment pig model calculated for THC (upper panel), JWH-210 (lower left panel), and RCS-4 (lower right panel); observed (x-axis) vs. individual (left panel each) and population (right panel each) predicted (y-axis) concentrations with the line of identity (black dotted line) B. Visual predictive check for THC (upper panel), JWH-210 (lower left panel), and RCS-4 (lower right panel); the black dots represent observed concentrations, the black line median predicted concentrations, and the gray area the 95% prediction interval after 1000 simulations; the upper part of each panel illustrates concentration-time profiles in semilogarithmic (y-axis) and the lower part in logarithmic (x- and y-axis) scale.

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

13

Fig. 3. Prediction of A. human intravenous 5 mg THC dose of Ohlsson et al. (1982); B. human intravenous 5 mg THC dose of Lindgren et al. (1981), and C. human intravenous 2 mg THC dose of Hunt and Jones (1980); the black dots represent observed concentrations, the black line median predicted concentrations, and the gray area the 95% prediction interval. The insert is depicted in semi-logarithmic scale.

14

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

to concentrations measured after pulmonary or oral administration. To sum up, pulmonary or oral administration share the advantage that they reflect authentic user habits, but are associated with the disadvantage that the exact amount of substance reaching the circulation remains speculative. Concerning the question whether SC PK was affected by dosage and peak concentration of the respective substances, systematic studies were not available. Yet, the PK parameters AUC and Cmax are dosedependant and concerning linear PK (as assessed in the present study), they are proportional to the dosage. However, dose and concentration dependence does not apply to t1/2 in case of a linear PK. In addition, PK studies on THC indicated the applicability of our pig data to data obtained from human studies (see below). The third issue that has to be considered in this context is the applied anaesthetic isoflurane. Isoflurane was chosen instead of the common anaesthetic thiopental to avoid possible CYP enzyme interactions (Bovill, 1997). Nevertheless, an interaction of isoflurane with the PK of the studied substances due to effects on circulation might be possible. For example, influences on the PK of substances e.g. of lidocaine have been discussed in literature (Thomasy et al., 2005). As a matter of fact, a major influence of isoflurane seems unlikely, as the model fit of THC was not affected (see below). THC concentrations and concentration-time profiles found in the present study are in good agreement with those published from human studies (Huestis et al., 1992; Kelly and Jones, 1992; Lindgren et al., 1981). Huestis et al. (1992) conducted a randomized double blind study with six healthy male volunteers (with a history of marihuana use), who smoked within 11 min a marihuana cigarette containing 0 mg, 15.8 mg (low dose), or 33.8 mg (high dose) THC. They detected peak plasma levels of 50–129 ng/mL (mean 84 ng/mL) and 76–267 ng/mL (mean 162 ng/mL) about 8– 10 min after the beginning of the smoking period of low and high doses, respectively. Fifteen minutes later, concentrations had declined to 38–83 ng/mL (mean 52 ng/mL) and 32–162 ng/mL (mean 94 ng/mL), and 30 min later, to 8–26 ng/mL (mean 17 ng/mL) and to 10–51 ng/mL (mean 29 ng/mL) regarding the low and high dose, respectively. In comparison, we found THC concentrations of 126–315 ng/mL (mean 198 ng/mL) after 10 min, 60–245 ng/mL (mean 134 ng/mL) after 15 min, and 39–126 ng/mL (mean 69 ng/ mL) 30 min after administration. One pig exhibited very high THC concentrations leading to higher mean values, but overall our results are rather well comparable to those of Huestis et al. after high dose smoking (Huestis et al., 1992). Variations may be explained by the different routes of administration. As already discussed above, concentrations determined after smoking are expected to differ from those after i.v. administration due to lower systemic availability. Regarding THC, the bioavailability varies between 18 and 50% depending on consumption habits (frequent/ infrequent user) and issues such as smoking period and depth of puffs (Huestis, 2002). Considering the high dose (33.8 mg) smoked in the Huestis et al. (1992) and assuming a mean bioavailability of about 25% would lead to a total systemic dose of about 8 mg. This dose is consistent with the dose administered in our study. Nevertheless, it has to be taken into consideration that in our study a bolus i.v. injection was applied, resulting in a rapid delivery of the total drug dose into the bloodstream. On the contrary, the participants in the study by Huestis et al. (1992) smoked this dosage in a time period of 11 min. Taken this into account and bearing in mind the high inter-individual variabilities (Huestis et al., 1992), variations in concentrations are not surprising. Compared to Kelly and Jones (1992) who administered 5 mg THC in a 2 min i.v. infusion to four frequent and four infrequent users, our results are in rather good agreement. They determined mean THC concentrations of 437 ng/mL (frequent users) and 386 ng/mL (infrequent users) 2 min after administration, and 4 h later, they

still detected mean concentrations of 13 ng/mL (frequent users) and 5 ng/mL (infrequent users). In our study, THC concentrations obtained after 2 min laid between 444 and 967 ng/mL (mean 658 ng/mL) and after 4 h between 5.7 and 13 ng/mL (mean 8.8 ng/ mL). As far as the pig study by Brunet et al. (2006) is concerned, who applied a similar study protocol, higher THC Cmax have been determined in the present study. This could be due to the fact that Brunet et al. did not clarify the duration of drug injection (bolus or infusion) as well as the starting point (t = 0 min) of the experiment. Especially in the first few seconds after administration there is a very rapid decline in drug concentration, which can result in significantly different Cmax. Comparing the behaviour of the three substances under investigation, RCS-4 revealed much lower Cmax than JWH210 and THC. Moreover, a slighter decrease of RCS-4 concentrations has been observed. In addition, the CL of RCS-4 is about two times higher than that of JWH-210 and THC (Table 1). One possible explanation could be an extensive metabolism of RCS-4 resulting in a faster elimination and therefore lower Cmax and higher CL values of the parent compound. This issue should be clarified in future investigations by analyzing pig urine and tissue specimens, which were also collected during or at the end of the experiments. RCS4 also has the highest central and peripheral V (Table 1), which could be due to a high binding to tissue proteins and, therefore, a higher distribution to tissues. The fact that the concentrations of the three substances did not differ significantly after single or combined administration suggests that the PK of THC, JWH-210, and RCS-4 are not affected by each other if the compounds are administered simultaneously. This issue is important, as SCs are often consumed together or with THC. In almost every observed concentration-time curve a temporary increase of the blood concentrations at different time points was noticed in the elimination phase (Electronic Supplementary material; S6). This phenomenon could be explained by enterohepatic circulation (EHC). Further studies should allow for clarifying this by analyzing bile fluid, which was also collected in the present study, because substances underlying EHC processes pass the bile before they are distributed again into the bloodstream or excreted. This process can lead to an extended retention in the body resulting in a longer t1/2. The higher terminal t1/2 of THC, JWH210, and RCS-4 compared to t1/2a und t1/2b could indicate on the one hand an EHC. For THC this phenomenon is described in literature (Huestis, 2002). On the other hand, their high lipophilicity could implicate an extensive distribution and storage in adipose tissue (Schaefer et al., 2014). This mechanism can also lead to a prolonged t1/2, attributable to the release of the substance from adipose tissue into the bloodstream (Grotenhermen, 2003; Huestis, 2002). Regarding the terminal t1/2 of THC published in the literature, there is a huge variation from several hours to days (Grotenhermen, 2003; Huestis, 2002; Ohlsson et al., 1982) because of different study protocols, compartment models, and analytical methods used. Brunet et al. (2006) determined a mean terminal t1/ 2 of 10.6  2.2 h. In the present study a quite short THC terminal t1/2 of about 2 h was estimated. This underestimation is owed to the fact that Brunet et al. (2006) collected samples for 48 h, whilst in the present study samples have only been taken for 6 h. The data might have been in better agreement, if the duration of the experiment would have been extended with blood samples taken for a few days after administration. To conclude, the t1/2g determined in the present study reflects the elimination within the first six hours after administration. However, the real terminal t1/2 of THC could only reliably be estimated in a long-term elimination study. In our analyses it was also attempted to implement the EHC into the PK model as described elsewhere (Lehr et al., 2009). Unfortunately, the implementation did not improve the

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

description of the data, probably due to the uncontrolled release of the gall bladder of the narcotic animals versus a food controlled release in active humans. The determined V values (Table 1) of the three substances also indicate a distribution into deep compartments. The respective smaller Vcentral values compared to the peripheral V values suggest a rapid distribution from the central compartment into the peripheral compartments. Regarding forensic toxicology, storage in adipose tissue has to be considered by interpreting analytical data. Especially after chronic consumption longer windows of detection have to be taken into consideration due to accumulation and retarded release from adipose tissue. This has already been discussed for SCs and THC (Grotenhermen, 2003; Huestis, 2002; Kneisel et al., 2014).

4.3. Prediction of human THC exposure Prediction of human THC exposure using the final pig PK model revealed a good predictive performance, as compared with exposure from published studies (Hunt and Jones, 1980; Lindgren et al., 1981; Ohlsson et al., 1982). Slightly under- or overpredicted concentrations may have resulted from different study protocols e.g. different sampling points and duration of measurements. Concerning e.g. the study by Ohlsson et al., samples were taken for 48 h (Ohlsson et al., 1982). As illustrated in Fig. 3A, predicted concentrations at the end of the experiment laid outside the 95% confidence interval. This variation may be explained by the fact that the developed pig model was based on a sampling protocol of only 6 h. Nevertheless, concentrations at the early phase of the experiment have accurately been predicted (Fig. 3A). Another issue for the differences in predicted human exposure may be that in the compared human studies, heavy and light users participated. Lindgren et al. (1981) determined THC levels of 0.1– 3.0 ng/mL from the heavy users and below 0.2 ng/mL from the light users before the experiment started. In the studies of Ohlsson et al. (1982) and Lindgren et al. (1981) the subjects were requested to abstain from marihuana for at least 24 h prior to the experiment. Ohlsson et al. observed fluctuations in concentrations. They surmised that the heavy users did not abstain during the test period (Ohlsson et al., 1982). Human THC exposure could therefore be enhanced and this might also be one reason for the underpredicted concentrations.

5. Conclusion A popPK three-compartment model was successfully developed describing the serum-concentration-time profiles of THC, JWH-210, and RCS-4 in pigs after intravenous administration. The successful THC prediction of human exposure based on the pig PK model suggests that pigs in conjunction with PK modeling technique may serve as a tool for prediction of human PK of SCs. This may help in interpretation of clinical and toxicological findings in misuse and poisoning cases. Following this pilot study, use of the same administration route as in humans (e.g. smoking), of an extended sampling protocol, and of a higher number of animals would improve the predictive performance of the model. The inclusion of further species with different weight categories may also enable more precise predictions. Conflict of interest There are no financial or other relations that could lead to a conflict of interest.

15

Acknowledgements The authors thank Markus R. Meyer, Benjamin Peters, Dietmar Bregel, and the staff of the Institute for Clinical & Experimental Surgery at Saarland University for their support and help during the study. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. toxlet.2016.04.021. References Anzenbacher, P., Soucek, P., Anzenbacherova, E., Gut, I., Hruby, K., Svoboda, Z., Kvetina, J., 1998. Presence and activity of cytochrome P450 isoforms in minipig liver microsomes: comparison with human samples. Drug Metab. Dispos. 26, 56–59. Behonick, G., Shanks, K.G., Firchau, D.J., Mathur, G., Lynch, C.F., Nashelsky, M., Jaskierny, D.J., Meroueh, C., 2014. Four postmortem case reports with quantitative detection of the synthetic cannabinoid, 5F-PB-22. J. Anal. Toxicol. 38, 559–562. Bode, G., Clausing, P., Cervais, F., Loegsted, J., Luft, J., Nogues, V., Sims, J., Steering Group of the RETHINK Project, 2010. The utility of the minipig as an animal model in regulatory toxicology. J. Pharmacol. Toxicol. Methods 62, 196–220. Bovill, J.G., 1997. Adverse drug interactions in anesthesia. J. Clin. Anesth. 9, 3S–13S. Brunet, B., Doucet, C., Venisse, N., Hauet, T., Hebrard, W., Papet, Y., Mauco, G., Mura, P., 2006. Validation of large white pig as an animal model for the study of cannabinoids metabolism: application to the study of THC distribution in tissues. Forensic Sci. Int. 161, 169–174. Brunet, B., Hauet, T., Hebrard, W., Papet, Y., Mauco, G., Mura, P., 2010. Postmortem redistribution of THC in the pig. Int. J. Legal Med. 124, 543–549. Dresen, S., Kneisel, S., Weinmann, W., Zimmermann, R., Auwarter, V., 2011. Development and validation of a liquid chromatography-tandem mass spectrometry method for the quantitation of synthetic cannabinoids of the aminoalkylindole type and methanandamide in serum and its application to forensic samples. J. Mass Spectrom. 46, 163–171. Garrett, E.R., Hunt, C.A., 1977. Pharmacokinetics of delta9-tetrahydrocannabinol in dogs. J. Pharm. Sci. 66, 395–407. Grigoryev, A., Savchuk, S., Melnik, A., Moskaleva, N., Dzhurko, J., Ershov, M., Nosyrev, A., Vedenin, A., Izotov, B., Zabirova, I., Rozhanets, V., 2011. Chromatographymass spectrometry studies on the metabolism of synthetic cannabinoids JWH018 and JWH-073, psychoactive components of smoking mixtures. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 879, 1126–1136. Grotenhermen, F., 2003. Pharmacokinetics and pharmacodynamics of cannabinoids. Clin. Pharmacokinet. 42, 327–360. Harris, C.R., Brown, A., 2013. Synthetic cannabinoid intoxication: a case series and review. J. Emerg. Med. 44, 360–366. Hermanns-Clausen, M., Kneisel, S., Hutter, M., Szabo, B., Auwarter, V., 2013. Acute intoxication by synthetic cannabinoids – four case reports. Drug Test Anal 5, 790–794. Huestis, M.A., Henningfield, J.E., Cone, E.J., 1992. Blood cannabinoids: I. Absorption of THC and formation of 11-OH-THC and THCCOOH during and after smoking marijuana. J. Anal. Toxicol. 16, 276–282. Huestis, M.A., 2002. Cannabis (Marihuana)—effects on human behavior and performance. Forensic Sci. Rev. 14, 15–60. Huffman, J.W., Mabon, R., Wu, M.J., Lu, J., Hart, R., Hurst, D.P., Reggio, P.H., Wiley, J.L., Martin, B.R., 2003. 3-Indolyl-1-naphthylmethanes: new cannabimimetic indoles provide evidence for aromatic stacking interactions with the CB(1) cannabinoid receptor. Bioorg. Med. Chem. 11, 539–549. Hunt, C.A., Jones, R.T., 1980. Tolerance and disposition of tetrahydrocannabinol in man. J. Pharmacol. Exp. Ther. 215, 35–44. Hutter, M., Broecker, S., Kneisel, S., Auwarter, V., 2012. Identification of the major urinary metabolites in man of seven synthetic cannabinoids of the aminoalkylindole type present as adulterants in ‘herbal mixtures' using LC–MS/ MS techniques. J. Mass Spectrom. 47, 54–65. Hutter, M., Moosmann, B., Kneisel, S., Auwarter, V., 2013. Characteristics of the designer drug and synthetic cannabinoid receptor agonist AM-2201 regarding its chemistry and metabolism. J. Mass Spectrom. 48, 885–894. Kacinko, S.L., Xu, A., Homan, J.W., McMullin, M.M., Warrington, D.M., Logan, B.K., 2011. Development and validation of a liquid chromatography-tandem mass spectrometry method for the identification and quantification of JWH-018, JWH-073, JWH-019, and JWH-250 in human whole blood. J. Anal. Toxicol. 35, 386–393. Kavanagh, P., Grigoryev, A., Melnik, A., Simonov, A., 2012. The identification of the urinary metabolites of 3-(4-methoxybenzoyl)-1-pentylindole (RCS-4), a novel cannabimimetic, by gas chromatography-mass spectrometry. J. Anal. Toxicol. 36, 303–311. Kelly, P., Jones, R.T., 1992. Metabolism of tetrahydrocannabinol in frequent and infrequent marijuana users. J. Anal. Toxicol. 16, 228–235.

16

N. Schaefer et al. / Toxicology Letters 253 (2016) 7–16

Kneisel, S., Teske, J., Auwärter, V., 2014. Analysis of synthetic cannabinoids in abstinence control: long drug detection windows in serum and implications for practionners. Drug Test Anal. 6, 135–136. Kronstrand, R., Roman, M., Andersson, M., Eklund, A., 2013. Toxicological findings of synthetic cannabinoids in recreational users. J. Anal. Toxicol. 37, 534–541. Lehr, T., Staab, A., Tillmann, C., Trommeshauser, D., Schaefer, H.-G., Kloft, C., 2009. A quantitative enterohepatic circulation model. Clin. Pharmacokinet. 48, 529–542. Leuschner, J.T., Harvey, D.J., Bullingham, R.E.S., Paton, W.D.M., 1986. Pharmacokinetics of delta9-tetrahydrocannabinol in rabbits following single or multiple intravenous doses. Drug Metab. Dispos. 14, 230–238. Lindgren, J., Ohlsson, A., Agurell, S., Hollister, L., Gillespie, H.K., 1981. Clinical effects and plasma levels of delta9-tetrahydrocannabinol in heavy and light users of cannabis. Psychopharmacology (Berl.) 74, 208–212. Makriyannis, A., Deng, H., W.I.P. Organization, 2001. Cannabimimetic indole derivatives. Mogi, M., Toda, A., Iwasaki, K., Kusumoto, S., Takehara, H., Shimizu, M., Murayama, N., Utoh, M., Yamazaki, H., 2012. Simultaneous pharmacokinetics assessment of caffeine, warfarin, omeprazole, metoprolol, and midazolam intravenously or orally administered to Microminipigs. J. Toxicol. Sci. 37, 1157–1164. Ohlsson, A., Lindgren, J.E., Wahlén, A., Agurell, S., Hollister, L., Gillespie, H.K., 1982. Single dose kinetics of deuterium labelled delta1-tetrahydrocannabinol in heavy and light cannabis users. Biomed. Mass Spectrom. 9, 6–10. Patton, A.L., Chimalakonda, K.C., Moran, C.L., McCain, K.R., Radominska-Pandya, A., James, L.P., Kokes, C., Moran, J.H., 2013. K2 Toxicity: fatal case of psychiatric complications following AM2201 exposure. J. Forensic Sci. 58, 1676–1680. Schaefer, N., Peters, B., Bregel, D., Kneisel, S., Auwärter, V., Schmidt, P.H., Ewald, A.H., 2013. A fatal case involving several synthetic cannabinoids. Toxichem Krimtech 80, 248–251. Schaefer, N., Peters, B., Bregel, D., Maurer, H.H., Schmidt, P.H., Ewald, A.H., 2014. Can JWH-210 and JWH-122 be detected in adipose tissue four weeks after single oral drug administration to rats? Biomed. Chromatogr. 28, 1043–1047.

Schaefer, N., Kettner, M., Laschke, M.W., Schlote, J., Peters, B., Bregel, D., Menger, M. D., Maurer, H.H., Ewald, A.H., Schmidt, P.H., 2015. Simultaneous LC–MS/MS determination of JWH-210, RCS-4, delta9-tetrahydrocannabinol, and their main metabolites in pig and human serum, whole blood, and urine for comparing pharmacokinetic data. Anal. Bioanal. Chem. 407, 3775–3786. Schneir, A.B., Baumbacher, T., 2012. Convulsions associated with the use of a synthetic cannabinoid product. J. Med. Toxicol. 8, 62–64. Shanks, K.G., Dahn, T., Terrell, A.R., 2012. Detection of JWH-018 and JWH-073 by UPLC-MS-MS in postmortem whole blood casework. J. Anal. Toxicol. 36, 145–152. Shimshoni, J.A., Britzi, M., Sobol, E., Willenz, U., Nutt, D., Edery, N., 2015. 3-Methylmethcathinone: pharmacokinetic profile evaluation in pigs in relation to pharmacodynamics. J. Psychopharmacol. 29, 734–743. Sjögren, E., Bredberg, U., Lennernäs, H., 2012. The pharmacokinetics and hepatic disposition of repaglinide in pigs: mechanisitic modeling of metabolism and transport. Mol. Pharm. 9, 823–841. Svendsen, O., 2006. The minipig in toxicology. Exp. Toxicol. Pathol. 57, 335–339. Swindle, M.M., Makin, A., Herron, A.J., Clubb Jr, F.J., Frazier, K.S., 2012. Swine as models in biomedical research and toxicology testing. Vet. Pathol. 49 (2), 344–356. Teske, J., Weller, J.P., Fieguth, A., Rothamel, T., Schulz, Y., Troger, H.D., 2010. Sensitive and rapid quantification of the cannabinoid receptor agonist naphthalen-1-yl(1-pentylindol-3-yl)methanone (JWH-018) in human serum by liquid chromatography-tandem mass spectrometry. J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 878, 2659–2663. Thomasy, S.M., Pypendop, B.H., Ilkiw, J.E., Stanley, S.D., 2005. Pharmacokinetics of lidocaine and its active metabolite, monoethylglycinexylidine, after intravenous administration of lidocaine to awake and isoflurane-anesthetized cats. Am. J. Vet. Res. 66, 1162–1166. Valiveti, S., Hammell, D.C., Earles, D.C., Stinchcomb, A.L., 2004. Transdermal delivery of the synthetic cannabinoid WIN 55,212-2: in vitro/in vivo correlation. Pharm. Res. 21, 1137–1145.