Fitoterapia 105 (2015) 1–6

Contents lists available at ScienceDirect

Fitoterapia journal homepage: www.elsevier.com/locate/fitote

Effect of xanthotoxin (8-methoxypsoralen) on the anticonvulsant activity of classical antiepileptic drugs against maximal electroshock-induced seizures in mice Miroslaw Zagaja a, Daniel Pyrka b, Krystyna Skalicka-Wozniak c, Kazimierz Glowniak c, Magdalena Florek-Luszczki d, Michał Glensk e, Jarogniew J. Luszczki a,b,⁎ a

Isobolographic Analysis Laboratory, Institute of Rural Health, Jaczewskiego 2, PL 20-950 Lublin, Poland Department of Pathophysiology, Medical University, Ceramiczna 1, PL 20-150 Lublin, Poland c Department of Pharmacognosy with Medicinal Plant Unit, Medical University, Chodzki 1, PL 20-093 Lublin, Poland d Department of Public Health, Institute of Rural Health, Jaczewskiego 2, PL 20-950 Lublin, Poland e Department of Pharmacognosy, Medical University, Nankiera Square 1, PL 50-140 Wroclaw, Poland b

a r t i c l e

i n f o

Article history: Received 21 March 2015 Received in revised form 25 May 2015 Accepted 26 May 2015 Available online 28 May 2015 Chemical compounds studied in this article: Methoxsalen (PubChem CID: 4114) 8-Methoxypsoralen (PubChem CID: 4114) Xanthotoxin (PubChem CID: 4114) Carbamazepine (PubChem CID: 2554) Phenytoin (PubChem CID: 1775) Phenobarbital (PubChem CID: 4763) Sodium valproate (PubChem CID: 16760703) Keywords: Antiepileptic drugs Drug interactions Xanthotoxin Maximal electroshock-induced seizures Counter-current chromatography

a b s t r a c t The effects of xanthotoxin (8-methoxypsoralen) on the anticonvulsant activity of four classical antiepileptic drugs (carbamazepine, phenobarbital, phenytoin and valproate) were studied in the mouse maximal electroshock seizure model. Tonic hind limb extension (seizure activity) was evoked in adult male albino Swiss mice by a current (25 mA, 500 V, 50 Hz, 0.2 s stimulus duration) delivered via auricular electrodes. Total brain concentrations of antiepileptic drugs were measured by fluorescence polarization immunoassay to ascertain any pharmacokinetic contribution to the observed anticonvulsant effects. Results indicate that xanthotoxin (50 and 100 mg/kg, i.p.) significantly potentiated the anticonvulsant activity of carbamazepine against maximal electroshock-induced seizures (P b 0.05 and P b 0.001, respectively). Similarly, xanthotoxin (100 mg/kg, i.p.) markedly enhanced the anticonvulsant action of valproate in the maximal electroshock seizure test (P b 0.001). In contrast, xanthotoxin (100 mg/kg, i.p.) did not affect the protective action of phenobarbital and phenytoin against maximal electroshock-induced seizures in mice. Moreover, xanthotoxin (100 mg/kg, i.p.) significantly increased total brain concentrations of carbamazepine (P b 0.001) and valproate (P b 0.05), but not those of phenytoin and phenobarbital, indicating pharmacokinetic nature of interactions between drugs. In conclusion, the combinations of xanthotoxin with carbamazepine and valproate, despite their beneficial effects in terms of seizure suppression in mice, were probably due to a pharmacokinetic increase in total brain concentrations of these antiepileptic drugs in experimental animals. © 2015 Published by Elsevier B.V.

1. Introduction Xanthotoxin (8-methoxypsoralen, methoxsalen, Fig. 1) is a linear furanocoumarin, extracted from various medicinal plants, mainly from the seeds of Ammi majus L. (Apiaceae) commonly called the ‘Bishop's weed’ [1,2]. Together with psoralen and 5-methoxypsoralen (bergapten) they belong to the naturally occurring photoactive substances. These bioactive furanocoumarins are used in psoralen plus ultraviolet A radiation (PUVA) therapy for the treatment of cutaneous T-cell lymphomas and autoimmune disorders such as pemphigus vulgaris and scleroderma

⁎ Corresponding author at: Department of Pathophysiology, Medical University, Ceramiczna 1, PL 20-150 Lublin, Poland. E-mail addresses: [email protected], [email protected] (J.J. Luszczki).

http://dx.doi.org/10.1016/j.fitote.2015.05.020 0367-326X/© 2015 Published by Elsevier B.V.

[3–5]. They also have antitumor activity [6]. It has been reported that UVA plus furocoumarin derivatives exhibit anti-proliferative effects in human melanoma SK-MEL-28 and -37, Hela, HL-60, human leukemia K562 and Karpas 299T-lymphoma cells [7–10]. In a recent study, xanthotoxin (300 mg/kg, i.p.) produces anticonvulsant activity in mice [11]. The experimentally-derived ED50 values for xanthotoxin ranged between 219.1 mg/kg and 252.4 mg/kg in the maximal electroshock-induced seizure test [12]. Additionally, it was shown that imperatorin (8-isopentenyloxypsoralen), another natural furanocoumarin derivative, possesses anticonvulsant activity in preclinical studies by elevating the threshold for electroconvulsions [13] and enhancing the anticonvulsant action of carbamazepine, phenobarbital, phenytoin [14], and lamotrigine [15] in mice. Moreover, osthole (a natural coumarin derivative) suppresses seizure activity in the mouse maximal electroshock-induced seizure model [16].

2

M. Zagaja et al. / Fitoterapia 105 (2015) 1–6

Fig. 1. Structural formula of xanthotoxin.

At present, there has appeared a new trend for combining drugs with various “healthy” dietary and nutritional supplements of natural origin. Some epileptic patients can take not only the antiepileptic drugs, but also some “natural” dietary compounds without clinicians' prescription and indications. In such a situation the patients treated with classical antiepileptic drugs can ingest “healthy” supplements prepared from medicinal plants and herbs, containing xanthotoxin or other coumarin derivatives, which can change the anticonvulsant activity of co-administered antiepileptic drugs. Taking into account the above-mentioned facts, it was of great importance to evaluate the effects of xanthotoxin on the protective activity of four classical antiepileptic drugs (carbamazepine, phenobarbital, phenytoin and valproate) against maximal electroshock induced seizures in mice. Additionally, total brain antiepileptic drug concentrations were measured to ascertain whether any observed effects were consequent to a pharmacodynamic and/or a pharmacokinetic interaction. 2. Materials and methods 2.1. Animals and experimental conditions Adult male Swiss mice (weighing 22–26 g) that were kept in colony cages with free access to food and tap water, under standardized housing conditions (natural light–dark cycle, temperature of 23 ± 1 °C, relative humidity of 55 ± 5%), were used. After 7 days of adaptation to laboratory conditions, the animals were randomly assigned to experimental groups consisting of 8 mice. Each mouse was used only once and all tests were performed between 08.00 and 15.00. The experimental protocols and procedures described in this manuscript were approved by the Second Local Ethics Committee at the University of Life Sciences, Lublin, Poland (License no.: 4/2012). All behavioral experiments were conducted in a blinded manner (observers: M.Z. and D.P.). 2.2. Drugs The following antiepileptic drugs were used in this study: carbamazepine (a gift from Polfa, Starogard Gdanski, Poland), phenobarbital (Polfa, Krakow, Poland), phenytoin (Polfa, Warsaw, Poland), and valproate (sodium salt — Sigma-Aldrich, Poznan, Poland). All drugs, except for valproate, were suspended in a 1% solution of Tween 80 (Sigma, St. Louis, MO, USA) in distilled water, while valproate was directly dissolved in distilled water only. All drugs were administered intraperitoneally (i.p.) as a single injection, in a volume of 5 ml/kg body wt. Fresh drug solutions were administered as follows: phenytoin — 120 min; phenobarbital — 60 min; and valproate and carbamazepine — 30 min, before electroconvulsions and brain sampling for the measurement of antiepileptic drug concentrations. The pretreatment times before testing of the antiepileptic drugs were based upon information about their biological activity from literature and our previous experiments [14,17]. Xanthotoxin (Fig. 1) used in this study was purified according to a previously described method [12]. For this purpose, a high-performance counter-current chromatography, equipped with either two analytical coils connected in series (PTFE tubing, 0.8 mm

i.d., 22 ml total volume) or two preparative coils (1.6 mm i.d., 137 ml total volume) was used. Ground, air-dried fruits (50 g) of Pastinaca sativa L. collected in Medicinal Plant Garden, Department of Pharmacognosy, Medical University in Lublin, Poland, in the summer of 2010 were extracted with dichloromethane under reflux. The crude extract was subjected to countercurrent separation. Different solvent systems: mixtures of n-heptane, ethyl acetate, methanol and water were tested in order to calculate partition coefficients. Finally, a mixture with the ratio of 1:1:1:1 (v/v/v/v), giving the K value = 0.92, was chosen as optimal. A rapid scale-up process from analytical to preparative was developed. The separation was performed in reversed-phase system. Collected fractions were analyzed using RP-HPLC-DAD system in the gradient of methanol and water. The yield of purified xanthotoxin from the injection of crude extract (200 mg) was 7.2 mg (98% purity). The standard of xanthotoxin was obtained from ChromaDex (Irvine, CA, USA). Xanthotoxin was suspended in a 1% solution of Tween 80 (Sigma-Aldrich, Poznan, Poland) in distilled water and administered i.p. at 60 min before electroconvulsions, based on our previous studies [11,12]. 2.3. Maximal electroshock seizure threshold test Electroconvulsions were produced by means of an alternating current (0.2 s stimulus duration, 50 Hz, 500 V) delivered via ear-clip electrodes by a Rodent Shocker generator (Type 221, Hugo Sachs Elektronik, Freiburg, Germany). The criterion for the occurrence of seizure activity was the tonic hind limb extension. To evaluate the threshold for maximal electroconvulsions, at least 4 groups of mice, consisting of 8 animals per group, were challenged with electroshocks of various intensities to yield 10–30%, 30–50%, 50–70%, and 70–90% of animals with seizures. Then, a current intensity–response relationship curve was constructed, according to a log-probit method by Litchfield and Wilcoxon [18], from which a median current strength (CS 50 in mA) was calculated. Each CS 50 value represents the current intensity required to induce tonic hindlimb extension in 50% of the mice challenged. Again, after administration of a single dose of xanthotoxin to 4 groups of animals, the mice were subjected to electroconvulsions (each group with a constant current intensity). The threshold for maximal electroconvulsions was recorded for 3 different doses of xanthotoxin: 50, 100 and 150 mg/kg. The experimental procedure has been described in more detail in our earlier studies [13–15,17,19]. 2.4. Maximal electroshock seizure test Electroconvulsions were produced by means of an alternating current (0.2 s stimulus duration, 50 Hz, 500 V) delivered via ear-clip electrodes by a Rodent Shocker generator (Type 221, Hugo Sachs Elektronik, Freiburg, Germany). The criterion for the occurrence of seizure activity was the tonic hind limb extension. The protective activity of various classical antiepileptic drugs (carbamazepine, phenobarbital, phenytoin, and valproate) was determined as their median effective doses (ED50 values in mg/kg) against maximal electroshock-induced seizures (fixed current intensity of 25 mA). The animals were administered with different drug doses so as to obtain a variable percentage of protection against maximal electroshock seizures, allowing the construction of a dose–response relationship curve for each antiepileptic drug administered alone, according to Litchfield and Wilcoxon [18]. Each ED50 value represents the dose of a drug required to protect half of the animals tested against maximal electroshock seizures. Similarly, the anticonvulsant activity of a mixture of an antiepileptic drug with xanthotoxin was evaluated and expressed as ED50 corresponding to the dose of an antiepileptic drug necessary to protect 50% of mice against tonic hind limb extension in the maximal electroshock seizure test. In the present study, carbamazepine was administered at doses ranging between 2 and 16 mg/kg, phenytoin at doses ranging between

M. Zagaja et al. / Fitoterapia 105 (2015) 1–6

8 and 16 mg/kg, phenobarbital at doses ranging between 20 and 40 mg/kg, and valproate at doses ranging between 150 and 300 mg/kg. This experimental procedure has been described in detail in our earlier studies [14,15,19].

3

increase in the CS50 values was observed in the maximal electroshock seizure threshold test in mice (Table 1). 3.2. Effects of xanthotoxin on the protective action of classical antiepileptic drugs in the mouse maximal electroshock-induced seizure model

2.5. Measurement of total brain antiepileptic drug concentrations The measurement of total brain concentrations of various antiepileptic drugs was undertaken at the doses, which corresponded to their ED50 values from the maximal electroshock seizure test for all combinations investigated. Mice were killed by decapitation at times corresponding to the peak of maximum anticonvulsant effects for the antiepileptic drugs in the maximal electroshock seizure test. The whole brains of mice were removed from skulls, weighed, harvested and homogenized using Abbott buffer (1:2 w/v) in an Ultra-Turrax T8 homogenizer (IKA Werke, Staufen, Germany). The homogenates were centrifuged at 10,000 g for 10 min. The supernatant samples (75 μl) were analyzed by fluorescence polarization immunoassay for carbamazepine, phenytoin, phenobarbital or valproate content using a TDx analyzer and reagents exactly as described by the manufacturer (Abbott Laboratories, North Chicago, IL, USA). Total brain concentrations of antiepileptic drugs were expressed in μg/ml of brain supernatants as means ± S.E.M. of at least 8 separate brain preparations. 2.6. Statistics Both, CS50 and ED50 values with their 95% confidence limits were calculated by computer log-probit analysis according to Litchfield and Wilcoxon [18]. Subsequently, the respective 95% confidence limits were transformed to standard error of the means (S.E.M.) as described previously [20]. Statistical analysis of data from the electroconvulsive tests was performed with one-way analysis of variance (ANOVA) followed by the post-hoc Tukey/Kramer test for multiple comparisons. Total brain antiepileptic drug concentrations were statistically compared using the unpaired Student's t-test. Differences among values were considered statistically significant if P b 0.05. 3. Results 3.1. Influence of xanthotoxin upon the threshold for electroconvulsions Xanthotoxin (administered alone, i.p., 60 min prior to the test) dosedependently raised the CS50 values, necessary to produce tonic hind limb extension in 50% of animals. In this case, xanthotoxin at a dose of 150 mg/kg significantly elevated the CS50 value from 7.32 mA to 11.09 mA (by 51.5%; P b 0.001; Table 1). In contrast, the CS50 values for xanthotoxin, administered at doses of 50 and 100 mg/kg, did not reach statistical significance, although a slight (dose-dependent) Table 1 Effect of xanthotoxin upon the electroconvulsive threshold in mice. Treatment (mg/kg)

CS50 (mA)

n

Vehicle Xanthotoxin (50) Xanthotoxin (100) Xanthotoxin (150) F (3; 108) = 7.873; P b 0.0001

7.32 ± 0.44 8.26 ± 0.40 9.02 ± 0.62 11.09 ± 0.83⁎⁎⁎

32 32 24 24

Data are presented as median current strengths (CS50 values in mA ± S.E.M.), necessary to evoke seizure activity (tonic hindlimb extension) in 50% of animals tested. Xanthotoxin was given i.p., 60 min. prior to the test. Statistical evaluation of the data was performed using log-probit method and one-way ANOVA followed by the post-hoc Tukey–Kramer test for multiple comparisons. n — number of animals tested at those current intensities, whose convulsant effects ranged between 16% and 84%. However, log-probit method considered only those groups of animals, whose seizure effects ranged between 16% and 84%. This is the reason for number of animals (n) being different for various CS50 values. F — value of F-statistics from one-way ANOVA; P — value of probability from one-way ANOVA. ⁎⁎⁎ P b 0.001 vs. the control CS50 value (vehicle-treated animals).

The investigated classical antiepileptic drugs (carbamazepine, phenobarbital, phenytoin and valproate) administered alone exhibited anticonvulsant activity in the maximal electroshock seizure test in mice and their ED50 values are presented in Table 2. When xanthotoxin at doses of 50 and 100 mg/kg was co-administered with carbamazepine, it significantly enhanced the anticonvulsant effect of the latter drug against maximal electroshock-induced seizures by reducing its ED50 value from 13.97 to 10.38 and 5.01 mg/kg, respectively (Table 2). Oneway ANOVA followed by the post-hoc Tukey/Kramer test for multiple comparisons revealed that after administration of xanthotoxin at 50 and 100 mg/kg, the reduction of ED50 values of carbamazepine by 26% and 64%, attained statistical significance at P b 0.05 and P b 0.001, respectively (Table 2). In contrast, xanthotoxin at the lower dose of 25 mg/kg did not significantly potentiate the antiseizure activity of carbamazepine, although a 12% reduction in its ED50 value was observed in the maximal electroshock seizure test (from 13.97 to 12.28 mg/kg; Table 2). In the case of valproate, xanthotoxin at a dose of 100 mg/kg significantly enhanced the antiseizure action of valproate by reducing its ED50 value from 281.4 to 195.5 mg/kg (by 31%; P b 0.001; Table 2). Xanthotoxin at a lower dose of 50 mg/kg had no significant effect on the antielectroshock action of valproate, although a slight reduction in the ED50 value of valproate was observed (Table 2). Concerning phenobarbital and phenytoin, xanthotoxin at doses of 50 and 100 mg/kg did not affect significantly the anticonvulsant effects of the above drugs in the maximal electroshock seizure test (Table 2). In these cases, the xanthotoxin-induced reduction in ED50 values of phenobarbital and phenytoin did not attain statistical significance with one-way ANOVA (Table 2).

Table 2 Effect of xanthotoxin on the protective activity of four classical antiepileptic drugs against maximal electroshock-induced seizures in mice. Treatment (mg/kg)

ED50 (mg/kg)

n

Carbamazepine + vehicle Carbamazepine + xanthotoxin (25) Carbamazepine + xanthotoxin (50) Carbamazepine + xanthotoxin (100) F (3; 116) = 21.61; P b 0.0001 Phenobarbital + vehicle Phenobarbital + xanthotoxin (50) Phenobarbital + xanthotoxin (100) F (2; 69) = 1.811; P = 0.1712 Phenytoin + vehicle Phenytoin + xanthotoxin (50) Phenytoin + xanthotoxin (100) F (2; 85) = 0.2798; P = 0.7566 Valproate + vehicle Valproate + xanthotoxin (50) Valproate + xanthotoxin (100) F (2; 85) = 10.47; P b 0.0001

13.97 ± 0.92 12.28 ± 0.96 10.38 ± 0.81⁎ 5.01 ± 0.74⁎⁎⁎

32 24 32 32

35.39 ± 2.79 32.71 ± 2.88 27.87 ± 3.03

24 16 32

13.26 ± 1.03 12.81 ± 0.91 12.21 ± 0.89

40 24 24

281.4 ± 13.63 240.7 ± 11.97 195.5 ± 14.76⁎⁎⁎

32 24 32

Results are presented as median effective doses (ED50 in mg/kg ± S.E.M.) of antiepileptic drugs, protecting 50% of animals tested against maximal electroshock-induced hindlimb extension. All antiepileptic drugs were administered i.p.: phenytoin — 120 min, phenobarbital — 60 min, carbamazepine and valproate — 30 min prior to the maximal electroshock seizure test. Xanthotoxin was administered i.p. at 60 min before the maximal electroshock seizure test. Statistical analysis of data was performed with log-probit method and one-way ANOVA followed by the post-hoc Tukey–Kramer test for multiple comparisons. n — number of animals tested at those doses, whose expected anticonvulsant effects ranged between 16% and 84%. However, log-probit method considered only those groups of animals whose seizure effects ranged between 16% and 84%. This is the reason that number of animals (n) is different for various ED50 values. F — value of F-statistics from one-way ANOVA; P — value of probability from one-way ANOVA. ⁎ P b 0.05 vs. the control (an antiepileptic drug + vehicle) ED50 value. ⁎⁎⁎ P b 0.001 vs. the control (an antiepileptic drug + vehicle) ED50 value.

4

M. Zagaja et al. / Fitoterapia 105 (2015) 1–6

3.3. Influence of xanthotoxin on total brain concentrations of classical antiepileptic drugs With fluorescence polarization immunoassay method, the total brain concentration of carbamazepine (5.01 mg/kg) administered alone significantly differed from that determined for the combination of carbamazepine (5.01 mg/kg) with xanthotoxin (100 mg/kg) (Table 3). In this case, xanthotoxin increased the total brain carbamazepine concentrations from 1.60 to 2.94 μg/ml (by 84%; P b 0.001; Table 3). Additionally, total brain concentrations of valproate, as measured using the same method, changed significantly after administration of xanthotoxin (100 mg/kg) (by 46%; P b 0.05; Table 3). In contrast, total brain concentrations of phenobarbital did not differ significantly when phenobarbital (27.87 mg/kg) was administered in combination with xanthotoxin (100 mg/kg) compared to when administered alone (Table 3). Similarly, total brain concentrations of phenytoin were not significantly affected when phenytoin was administered in combination with xanthotoxin (100 mg/kg) compared to when administered alone (Table 3). 4. Discussion Results indicate that xanthotoxin administered at sub-threshold doses (i.e., at doses that per se did not significantly affect the threshold for electroconvulsions in mice) enhanced the anticonvulsant action of carbamazepine and valproate, but not that of phenytoin and phenobarbital against maximal electroshock-induced seizures in mice. Xanthotoxin also increases the brain concentration of carbamazepine and valproate which could underlie the enhancement of antiepileptic drug effect by xanthotoxin. Further studies are needed to determine whether antiepileptic drugs also increase brain concentrations of xanthotoxin to levels that may exert anticonvulsant effect. Similarly, imperatorin (a natural furanocoumarin derivative) enhanced the anticonvulsant action of carbamazepine [14]. In contrast to the xanthotoxin, imperatorin increased the ED50 values of phenobarbital and phenytoin, but it did not affect the anticonvulsant activity of valproate in the mouse maximal electroshock-induced seizure model [14]. Results from the maximal electroshock-induced seizure threshold test indicated that xanthotoxin administered alone elevated the threshold for electroconvulsions and thus, it exerted the anticonvulsant action in mice. Additionally, we have reported that the experimentallyderived ED50 values for xanthotoxin ranged between 219.1 mg/kg and 252.4 mg/kg, when administered i.p. at four various pretreatment times (i.e., 15, 30, 60 and 120 min.) in the maximal electroshockinduced seizure test [12]. Generally, the maximal electroshockinduced seizure test in rodents is considered as a valid experimental model to detect the anticonvulsant action of drugs against generalized Table 3 Effects of xanthotoxin on total brain concentrations of antiepileptic drug in mice. Treatment (mg/kg)

Brain concentration (μg/ml)

Carbamazepine (5.01) + vehicle Carbamazepine (5.01) + xanthotoxin (100) Phenobarbital (27.87) + vehicle Phenobarbital (27.87) + xanthotoxin (100) Phenytoin (12.21) + vehicle Phenytoin (12.21) + xanthotoxin (100) Valproate (195.5) + vehicle Valproate (195.5) + xanthotoxin (100)

1.60 ± 0.21 2.94 ± 0.23⁎⁎⁎ 12.07 ± 0.98 13.01 ± 1.19 1.27 ± 0.13 1.41 ± 0.15 95.0 ± 12.3 139.2 ± 13.2⁎

↑ 84%

↑ 47%

Data are presented as means ± S.E.M. of at least 8 separate determinations. Total brain concentrations of classical antiepileptic drugs were performed with fluorescence polarization immunoassay technique. Data were statistically verified by using the unpaired Student's t-test. All drugs were administered i.p. at doses corresponding to the ED50 value from the maximal electroshock-induced seizures. For more detail see the legend of Table 2. ↑ — increase in total brain concentration. ⁎ P b 0.05 vs. the respective control group (an antiepileptic drug-treated animals). ⁎⁎⁎ P b 0.001 vs. the respective control group (an antiepileptic drug-treated animals).

tonic–clonic seizures and partial convulsions with or without secondary generalization in humans [21]. Moreover, this experimental model of epilepsy is widely used for testing of new drugs and for selection of the agents with antiseizure activity in vivo [21]. Therefore, it was appropriate to use this test in the presented study to evaluate the anticonvulsant effects of classical antiepileptic drugs in combination with xanthotoxin. Pharmacokinetic studies revealed that xanthotoxin significantly elevated total brain carbamazepine and valproate concentrations in experimental animals. It is essential to note that pharmacokinetic interactions observed between xanthotoxin and carbamazepine and valproate seem to be entirely responsible for the observed anticonvulsant effects in the mouse maximal electroshock-induced seizure model. It was found that a 84% elevation of total brain carbamazepine concentrations by xanthotoxin (100 mg/kg) produced a 64% enhancement of the anticonvulsant action of carbamazepine in the mouse maximal electroshock-induced seizure model. The similar situation was observed for valproate and its interaction with xanthotoxin. It was found that a 46% increase in the total brain concentrations of valproate exerted a 31% increase in the anticonvulsant action of valproate in experimental animals after co-administration of xanthotoxin in the mouse maximal electroshock-induced seizure model. In the case of evaluation of total brain concentrations of phenobarbital and phenytoin in experimental animals, we found that xanthotoxin had no significant impact on total brain concentrations of these antiepileptic drugs. Lack of any significant changes in total brain concentrations of phenobarbital and phenytoin was correlated with no significant potentiation of the anticonvulsant activity of the tested antiepileptic drugs after administration of xanthotoxin in the mouse maximal electroshock-induced seizure model. This finding confirmed our hypothesis that xanthotoxin pharmacokinetically interacted with carbamazepine and valproate and the elevated total brain concentrations of carbamazepine and valproate were entirely responsible for the enhancement of the anticonvulsant action of these antiepileptic drugs in the mouse maximal electroshockinduced seizure model. It should be stressed that in the presented study, total brain concentrations of xanthotoxin were not estimated and thus, we do not know whether the tested antiepileptic drugs affect total brain concentrations of xanthotoxin or not. On the other hand, our observation reporting that xanthotoxin elevated the threshold for electroconvulsions and exerted anticonvulsant activity in the mouse maximal electroshock-induced seizure model indirectly confirm the fact that the coumarin penetrates through the blood–brain barrier and produces the antielectroshock action in mice. At present, it is unknown whether this effect was evoked by xanthotoxin alone or its active metabolite(s). To verify whether xanthotoxin undergoes metabolic transformation to some active metabolite(s) in mice or not, more advanced biochemical studies are required. Furthermore, this preclinical study was conducted on Albino Swiss mice and the observed pharmacodynamic/pharmacokinetic interactions of xanthotoxin with classical antiepileptic drugs may not be the same as in humans because of some species differences in the metabolism of furanocoumarins [22]. Perhaps, owing to the differences in the metabolic transformation of xanthotoxin in humans, doses of xanthotoxin exerting anticonvulsant effects would be low enough to be quite well tolerated by patients. To confirm, however, this hypothesis more advanced pharmacokinetic studies in humans are required. To explain the observed increase in the total concentrations of carbamazepine and valproate in brains of experimental animals, one should consider a fact that xanthotoxin can induce inhibition of multidrug resistance proteins such as P-glycoprotein, multidrug resistance proteins 1 and 2 (MRP1 and MRP2) or breast cancer resistance protein (BCRP) (ABCG2), whose normal physiological activity is related to the removal of drugs from the brain tissue. Of note, several antiepileptic drugs frequently used in the treatment of epilepsies are substrates of P-glycoprotein both, in rodents (gabapentin, lamotrigine, phenobarbital, phenytoin and topiramate) and humans (phenytoin, phenobarbital,

M. Zagaja et al. / Fitoterapia 105 (2015) 1–6

lamotrigine and levetiracetam) [23–25]. Data on carbamazepine, levetiracetam and valproate are not clearly defined as yet [25,26]. The inhibitors of these proteins may contribute to the accumulation of antiepileptic drugs in the brain [27]. Therefore, it is highly probable that xanthotoxin as an inhibitor of multi-drug resistance proteins could increase carbamazepine and valproate contents in brains of experimental animals. Because furanocoumarins such as bergamottin from grapefruit juice are inhibitors of P-glycoprotein [28], it is highly likely that xanthotoxin can also inhibit P-glycoprotein activity. On the other hand, P-glycoprotein inhibition by xanthotoxin should theoretically increase the antiepileptic activity of phenytoin and phenobarbital, which did not occur in this case. Perhaps, these antiepileptic drugs (phenytoin and phenobarbital) possess higher affinity for P-glycoprotein than xanthotoxin and thus, they competitively block such transporters. Another fact, worthy of mentioning while explaining the results from this pharmacokinetic study is that xanthotoxin (methoxsalen) is both, a substrate and an inhibitor for CYP2A6 [29–31]. Methoxsalen inhibition of liver microsomal CYP2A6 occurs at low concentrations and is very rapid [29,30]. It is worth noting that valproate is also a substrate for this isoenzyme — CYP2A6 [32]. Therefore, it seems possible that the increase in the concentration of valproate in the mouse brain may be a consequence of the reduction of the isoenzyme's (CYP2A6) activity by xanthotoxin. The interaction of xanthotoxin with carbamazepine has been associated with a considerable increase in the total brain concentration of this antiepileptic drug. The underlying mechanism for this interaction is still unclear. It seems less likely to be related to metabolism, since carbamazepine is a substrate of CYP3A4 and coumarins are metabolized mainly by CYP2A6. On the other hand, coumarins (present in fruits), primarily bergamottin and other furanocoumarin dimers, are the major contributors to a pronounced inhibitory effect on CYP3A4 activity [33,34], the main enzyme metabolizing carbamazepine [35]. Also imperatorin increases the concentration of carbamazepine in the mouse brain, possibly by acting in the same way as xanthotoxin or bergamottin [14], reducing activity of one of the most important enzymes of the cytochrome P450 complex. Although the above-mentioned hypotheses are speculative, they can readily explain the observed changes in the brain carbamazepine and valproate contents after administration of xanthotoxin. Consequently, the most plausible mechanism is pharmacokinetic interactions, although pharmacodynamic interactions cannot be excluded. However, the influence of xanthotoxin administered singly, at 60 min before the maximal electroshock seizure test and brain sampling, may not be sufficient to affect the metabolism of the classical antiepileptic drugs. In comparison to imperatorin, xanthotoxin does not affect the anticonvulsant action of phenytoin and phenobarbital, which is probably caused by their different mechanisms of molecular activity. It has been suggested that imperatorin may complementary potentiate the anticonvulsant activity of carbamazepine, phenytoin and phenobarbital in experimental animals through imperatorin-induced irreversible inactivation of GABA-transaminase and subsequent increase in GABA content in the brain [14,36]. Moreover, imperatorin enhances GABA-mediated inhibitory neurotransmitter action through the interaction with benzodiazepine receptors [37]. According to Singhuber et al. [38] (oxy-) prenylated coumarin derivatives such as osthole, oxypeucedanin and imperatorin represent a new group of GABAA receptor modulators. It has been suggested that prenyl residues are essential for positive modulatory activity of these substances [38]. Additionally, the enhancement of inhibitory potential of GABA by these coumarins was not inhibited by flumazenil, indicating an interaction with a binding site distinct from the benzodiazepine binding site. Because xanthotoxin does not have any prenyl-rest, this may be the reason for substantial difference in the anticonvulsant activity between imperatorin and xanthotoxin in preclinical studies. One of the possible mechanisms of the anticonvulsant enhancement of carbamazepine and valproate by xanthotoxin is its action on potassium

5

channels. Xanthotoxin has been reported to block ATP-dependent potassium channels [39] and to modulate calcium-dependent potassium channels [40]. It is widely known that potassium channels are drug targets for therapeutic intervention in many diseases, including epilepsy [41,42]. To elucidate the exact molecular mechanisms of interaction of xanthotoxin in combination with classical antiepileptic drugs against maximal electroshock-induced seizures in mice, more advanced neurochemical studies are required. It is important to note that xanthotoxin, similarly to other furanocoumarins (including, osthole and imperatorin) exerted anticonvulsant properties in the mouse maximal electroshock-induced seizure model. At present, we do not intend to directly translate the results presented herein from preclinical studies to clinical conditions. However, xanthotoxin molecule may undergo structural modifications to chemically change its pharmacological profile in order to ameliorate the anticonvulsant properties of the compound. Generally, xanthotoxin may be considered as a substrate isolated from plants and medicinal herbs, which due to chemical modifications may gain some novel properties and indications for the clinical use. Of note, some bioactive compounds of natural origin can be readily isolated from plants and medicinal herbs, purified, concentrated and applied either as drugs or dietary supplements to cure various diseases or to support the healthy lifestyle [43,44]. Summing up, for the first time it has been found that xanthotoxin enhances the protective action of carbamazepine and valproate, but not that of phenytoin and phenobarbital against maximal electroshockinduced seizures in mice. Unfortunately, these favorable combinations of xanthotoxin with carbamazepine and valproate in the maximal electroshock seizure test were probably due to pharmacokinetic interactions that markedly increased total brain carbamazepine and valproate concentrations in animals. Hence, the application of xanthotoxin as an add-on drug to the treatment of epileptic seizures with carbamazepine, phenytoin, phenobarbital and valproate requires more advanced research, especially, in other seizure models and/or the application with other antiepileptic drugs. At present, it is too early to consider xanthotoxin as a credible add-on agent for human clinical use, however, xanthotoxin can be used as a natural dietary supplement by the patients with epilepsy. Despite no clinical reasons to take xanthotoxin (a skinsensitizing agent in PUVA therapy) as an anticonvulsant drug, this naturally occurring coumarin can be an ideal substrate for chemical modifications changing its structure in order to improve its anticonvulsant properties. In such a situation, xanthotoxin could be a potential source of bioactive anticonvulsant drugs in future. Disclosure of conflicts of interest The authors have no disclosures to declare. Acknowledgments This study was supported by grants from the Institute of Rural Health, Lublin, Poland (12120/2013) and the National Science Centre of Poland (N N405 617538). The authors are grateful for the generous gift of carbamazepine from Polpharma S.A. in Starogard Gdański, Poland. References [1] M. Purohit, D. Pande, A. Datta, P.S. Srivastava, Enhanced xanthotoxin content in regenerating cultures of Ammi majus and micropropagation, Planta Med. 61 (1995) 481–482. [2] H. Ekiert, E. Gomolka, Coumarin compounds in Ammi majus L. callus cultures, Pharmazie 55 (2000) 684–687. [3] S.E. Malawista, D. Trock, R.L. Edelson, Photopheresis for rheumatoid arthritis, Ann. N. Y. Acad. Sci. 636 (1991) 217–226. [4] H.W. Lim, R.L. Edelson, Photopheresis for the treatment of cutaneous T-cell lymphoma, Hematol. Oncol. Clin. North Am. 9 (1995) 1117–1126.

6

M. Zagaja et al. / Fitoterapia 105 (2015) 1–6

[5] R.M. Knobler, W. Graninger, A. Lindmaier, F. Trautinger, Photopheresis for the treatment of lupus erythematosus. Preliminary observations, Ann. N. Y. Acad. Sci. 636 (1991) 340–356. [6] M. Sumiyoshi, M. Sakanaka, M. Taniguchi, K. Baba, Y. Kimura, Anti-tumor effects of various furocoumarins isolated from the roots, seeds and fruits of Angelica and Cnidium species under ultraviolet A irradiation, J. Nat. Med. 68 (2014) 83–94. [7] J. Bartosova, K. Kuzelova, M. Pluskalova, I. Marinov, P. Halada, Z. Gasova, UVAactivated 8-methoxypsoralen (PUVA) causes G2/M cell cycle arrest in Karpas 299T-lymphoma cells, J. Photochem. Photobiol. B Biol. 85 (2006) 39–48. [8] V. Carneiro Leite, R. Ferreira Santos, L. Chen Chen, L. Andreu Guillo, Psoralen derivatives and longwave ultraviolet irradiation are active in vitro against human melanoma cell line, J. Photochem. Photobiol. B Biol. 76 (2004) 49–53. [9] L. Dalla Via, O. Gia, S. Marciani Magno, L. Santana, M. Teijeira, E. Uriarte, New tetracyclic analogues of photochemotherapeutic drugs 5-MOP and 8-MOP: synthesis, DNA interaction, and antiproliferative activity, J. Med. Chem. 42 (1999) 4405–4413. [10] G. Viola, D. Vedaldi, F. Dall'Acqua, E. Fortunato, G. Basso, N. Bianchi, et al., Induction of gamma-globin mRNA, erythroid differentiation and apoptosis in UVA-irradiated human erythroid cells in the presence of furocumarin derivatives, Biochem. Pharmacol. 75 (2008) 810–825. [11] J.J. Luszczki, M. Andres-Mach, M. Glensk, K. Skalicka-Wozniak, Anticonvulsant effects of four linear furanocoumarins, bergapten, imperatorin, oxypeucedanin, and xanthotoxin, in the mouse maximal electroshock-induced seizure model: a comparative study, Pharmacol. Rep. 62 (2010) 1231–1236. [12] K. Skalicka-Woźniak, M. Zagaja, K. Głowniak, J. Luszczki, Purification and anticonvulsant activity of xanthotoxin (8-methoxypsoralen), Cent. Eur. J. Biol. 9 (2014) 431–436. [13] J.J. Luszczki, K. Glowniak, S.J. Czuczwar, Time-course and dose–response relationships of imperatorin in the mouse maximal electroshock seizure threshold model, Neurosci. Res. 59 (2007) 18–22. [14] J.J. Luszczki, K. Glowniak, S.J. Czuczwar, Imperatorin enhances the protective activity of conventional antiepileptic drugs against maximal electroshock-induced seizures in mice, Eur. J. Pharmacol. 574 (2007) 133–139. [15] J.J. Luszczki, E. Wojda, G. Raszewski, K. Glowniak, S.J. Czuczwar, Influence of imperatorin on the anticonvulsant activity and acute adverse-effect profile of lamotrigine in maximal electroshock-induced seizures and chimney test in mice, Pharmacol. Rep. 60 (2008) 566–573. [16] J.J. Luszczki, M. Andres-Mach, W. Cisowski, I. Mazol, K. Glowniak, S.J. Czuczwar, Osthole suppresses seizures in the mouse maximal electroshock seizure model, Eur. J. Pharmacol. 607 (2009) 107–109. [17] J.J. Luszczki, R. Czernecki, K. Wojtal, K.K. Borowicz, S.J. Czuczwar, Agmatine enhances the anticonvulsant action of phenobarbital and valproate in the mouse maximal electroshock seizure model, J Neural. Transm. 115 (2008) 1485–1494. [18] J.T. Litchfield Jr., F. Wilcoxon, A simplified method of evaluating dose–effect experiments, J. Pharmacol. Exp. Ther. 96 (1949) 99–113. [19] J.J. Luszczki, E. Wojda, M. Andres-Mach, W. Cisowski, M. Glensk, K. Glowniak, et al., Anticonvulsant and acute neurotoxic effects of imperatorin, osthole and valproate in the maximal electroshock seizure and chimney tests in mice: a comparative study, Epilepsy Res. 85 (2009) 293–299. [20] J.J. Luszczki, N. Ratnaraj, P.N. Patsalos, S.J. Czuczwar, Isobolographic analysis of interactions between loreclezole and conventional antiepileptic drugs in the mouse maximal electroshock-induced seizure model, Naunyn Schmiedeberg's Arch. Pharmacol. 373 (2006) 169–181. [21] W. Loscher, C.P. Fassbender, B. Nolting, The role of technical, biological and pharmacological factors in the laboratory evaluation of anticonvulsant drugs. II. Maximal electroshock seizure models, Epilepsy Res. 8 (1991) 79–94. [22] Y.M. Lee, T.H. Wu, Effects of 5-methoxypsoralen (5-MOP) on arylamine Nacetyltransferase activity in the stomach and colon of rats and human stomach and colon tumor cell lines, In Vivo 19 (2005) 1061–1069. [23] E. Aronica, S.M. Sisodiya, J.A. Gorter, Cerebral expression of drug transporters in epilepsy, Adv. Drug Deliv. Rev. 64 (2012) 919–929.

[24] W. Loscher, H. Potschka, Role of drug efflux transporters in the brain for drug disposition and treatment of brain diseases, Prog. Neurobiol. 76 (2005) 22–76. [25] C. Luna-Tortos, M. Fedrowitz, W. Loscher, Several major antiepileptic drugs are substrates for human P-glycoprotein, Neuropharmacology 55 (2008) 1364–1375. [26] L. Moerman, L. Wyffels, D. Slaets, R. Raedt, P. Boon, F. De Vos, Antiepileptic drugs modulate P-glycoproteins in the brain: a mice study with (11)Cdesmethylloperamide, Epilepsy Res. 94 (2011) 18–25. [27] C. Brandt, K. Bethmann, A.M. Gastens, W. Loscher, The multidrug transporter hypothesis of drug resistance in epilepsy: proof-of-principle in a rat model of temporal lobe epilepsy, Neurobiol. Dis. 24 (2006) 202–211. [28] K. Iwanaga, S. Yoneda, Y. Hamahata, M. Miyazaki, M. Shibano, M. Taniguchi, et al., Inhibitory effects of furanocoumarin derivatives in Kampo extract medicines on P-glycoprotein at the blood–brain barrier, Biol. Pharm. Bull. 34 (2011) 1246–1251. [29] A.J. Draper, A. Madan, A. Parkinson, Inhibition of coumarin 7-hydroxylase activity in human liver microsomes, Arch. Biochem. Biophys. 341 (1997) 47–61. [30] L.L. Koenigs, R.M. Peter, S.J. Thompson, A.E. Rettie, W.F. Trager, Mechanism-based inactivation of human liver cytochrome P450 2A6 by 8-methoxypsoralen, Drug. Metab. Dispos. 25 (1997) 1407–1415. [31] L.L. Koenigs, W.F. Trager, Mechanism-based inactivation of P450 2A6 by furanocoumarins, Biochemistry 37 (1998) 10047–10061. [32] C. Xu, S. Goodz, E.M. Sellers, R.F. Tyndale, CYP2A6 genetic variation and potential consequences, Adv. Drug Deliv. Rev. 54 (2002) 1245–1256. [33] C. Ioannides, Pharmacokinetic interactions between herbal remedies and medicinal drugs, Xenobiotica 32 (2002) 451–478. [34] X.W. Chen, K.B. Sneed, S.Y. Pan, C. Cao, J.R. Kanwar, H. Chew, et al., Herb–drug interactions and mechanistic and clinical considerations, Curr. Drug Metab. 13 (2012) 640–651. [35] B.M. Kerr, K.E. Thummel, C.J. Wurden, S.M. Klein, D.L. Kroetz, F.J. Gonzalez, et al., Human liver carbamazepine metabolism. Role of CYP3A4 and CYP2C8 in 10,11-epoxide formation, Biochem. Pharmacol. 47 (1994) 1969–1979. [36] S.Y. Choi, E.M. Ahn, M.C. Song, D.W. Kim, J.H. Kang, O.S. Kwon, et al., In vitro GABAtransaminase inhibitory compounds from the root of Angelica dahurica, Phytother. Res. 19 (2005) 839–845. [37] O. Bergendorff, K. Dekermendjian, M. Nielsen, R. Shan, R. Witt, J. Ai, et al., Furanocoumarins with affinity to brain benzodiazepine receptors in vitro, Phytochemistry 44 (1997) 1121–1124. [38] J. Singhuber, I. Baburin, G.F. Ecker, B. Kopp, S. Hering, Insights into structure–activity relationship of GABAA receptor modulating coumarins and furanocoumarins, Eur. J. Pharmacol. 668 (2011) 57–64. [39] A. Szewczyk, J.R. De Weille, M. Lazdunski, 8-Methoxypsoralen blocks ATP-sensitive potassium channels and stimulates insulin release, Eur. J. Pharmacol. 216 (1992) 323–326. [40] M.C. Isoldi, E.A. Pereira, M.A. Visconti, A.M. Castrucci, The role of calcium, calciumactivated K+ channels, and tyrosine/kinase in psoralen-evoked responses in human melanoma cells, Braz. J. Med. Biol. Res. 37 (2004) 559–568. [41] B.S. Jensen, D. Strobaek, S.P. Olesen, P. Christophersen, The Ca2+-activated K+ channel of intermediate conductance: a molecular target for novel treatments? Curr. Drug Targets 2 (2001) 401–422. [42] O. Postea, M. Biel, Exploring HCN channels as novel drug targets, Nat. Rev. Drug Discov. 10 (2011) 903–914. [43] K.M. VanderMolen, G.R. Ainslie, M.F. Paine, N.H. Oberlies, Labeled content of two furanocoumarins in dietary supplements correlates with neither actual content nor CYP3A inhibitory activity, J. Pharm. Biomed. Anal. 98 (2014) 260–265. [44] S. Vazquez-Rodriguez, M.J. Matos, F. Borges, E. Uriarte, L. Santana, Bioactive coumarins from marine sources: origin, structural features and pharmacological properties, Curr. Top. Med. Chem. (2015), http://dx.doi.org/10.2174/1568026615666150427125916.