Analgesic potency of S-acetylthiorphan after intravenous administration to mice

Analgesic potency of S-acetylthiorphan after intravenous administration to mice

European Journal of Pharmacology, 243 (1993) 129-134 © 1993 Elsevier Science Publishers B.V. All rights reserved 0014-2999/93/$06.00 129 EJP 53332 ...

506KB Sizes 0 Downloads 12 Views

European Journal of Pharmacology, 243 (1993) 129-134 © 1993 Elsevier Science Publishers B.V. All rights reserved 0014-2999/93/$06.00


EJP 53332

Analgesic potency of S-acetylthiorphan after intravenous administration to mice D i d i e r M. L a m b e r t *, F r a n k M e r g e n , J a c q u e s H. P o u p a e r t a n d P i e r r e D u m o n t Laboratory of Medicinal Chemistry, School of Pharmacy, Catholic University of Louvain, Avenue E. Mounier 73, CMFA / UCL 73.40, B-1200 Brussels, Belgium

Received 5 April 1993, revised MS received 2 August 1993, accepted 3 August 1993

As hydrolysis in serum of acetorphan to acetylthiorphan (N-[(R,S)-3-acetylmercapto-2-bepzylpropanoyl]glycine) has been evidenced, both the neutral endopeptidase inhibition in vitro by acetylthiorphan and analgesic potency of acetylthiorphan after intravenous administration to mice in two analgesic models, the hot-plate and the tail-flick tests, were compared with those of thiorphan and acetorphan. Acetylthiorphan showed a decreased degree of neutral endopeptidase inhibition (IC50 = 316 ± 38 nM) compared to thiorphan (ICs0 = 1.8 __+0.2 nM). After intravenous administration followed by the hot-plate jump latency test, acetylthiorphan elicited a degree of analgesia equivalent to that with acetorphan but longer lasting. Like acetorphan and thiorphan, acetylthiorphan was devoid of analgesic activity in the tail-flick test. The results indicated that S-acetylation of the thiol function in acetylthiorphan ensures sufficient lipophilicity to permit crossing of the blood-brain barrier and that acetylthiorphan acts via a prodrug mechanism. Acetylthiorphan; Acetorphan; Thiorphan; Neutral endopeptidase; Hot-plate test; Analgesia

1. Introduction In brain, endogenous enkephalins are inactivated by at least three well-known enzymes: neutral endopeptidase (enkephalinase EC aminopeptidase N (EC, and dipeptidylaminopeptidase cleaving the Gly3-Phea,Tyrt-Gly e and Glya-Gly 3 amide bonds respectively (Hersh, 1981). The peptide fragments released are devoid of analgesic effects. In order to increase the physiological activity of enkephalins, several laboratories have synthesized inhibitors of the peptidases responsible for their degradation. There have now been developed neutral endopeptidase inhibitors, thiorphan from one laboratory, SCH 32615 and SCH 34729 from another (review by Rich, 1990). These compounds and related derivatives now serve as useful pharmacological probes or therapeutic agents in a wide variety of opioid-like effects including analgesia, increased locomotion, antidepressant (in the behavioral despair test) and antidiarrheal activity (Schwartz et al., 1985). These effects are prevented by opioid receptor antagonists such as nalox-

* Corresponding author. Tel. (32) 2 764 73 40, fax (32) 2 764 73 63.

one. Moreover, the lack of abuse potential is an attractive possibility (Knisely et al., 1989). Thiorphan, N-[( R , S )-3-mercapto-2-benzylpropanoyl]glycine (fig. 1), is a highly potent inhibitor of neutral endopeptidase but does not cross the blood-brain barrier to a significant extent and, except at very high doses, does not produce analgesia after peripheral administration (Roques et al., 1980). Acetorphan, a lipophilic prodrug of thiorphan, in which the carboxylic and thiol groups are derivatized with benzyl and acetyl residues respectively (fig. 1), is able to penetrate the brain and has been reported active both after intravenous and oral administration (Lecomte et al., 1986). These two chemical modifications drastically decrease the inhibition potency versus neutral endopeptidase in vitro. However, the conversion of acetorphan into thiorphan in vivo has been evidenced (Lecomte et al., 1986). Recently, the rapid hydrolysis in serum of the prodrug, acetorphan, into S-acetylthiorphan, a monoprotected form of thiorphan (keeping the carboxylic function free), has been demonstrated (Fourni6-Zaluski et al., 1992). In the present study, we investigated both the analgesic activity after parenteral administration and the ability to bind neutral endopeptidase enzyme of this metabolite, i.e. S-acetylthiorphan, N-[(R,S)-3acetylmercapto-2-benzylpropanoyl]glycine (fig. 1).


2. Materials and methods


2.1. Animals Male NMRI mice (20-25 g) raised at the University animal facility were housed at ambient temperature with free access to standard rodent diet and water. To diminish the influence of stress, the mice were housed in the experimentation room at least 2 h before the first injection. One treated mouse and one control mouse were evaluated alternately.

2.2. Synthesis of S-acetylthiorphan, N- { 2-[ (acetylthio )methyl]- 1-oxo-3-phenylpropyl} glycine



Acetylthiorphan was synthesized according to Ondetti, as described in a German patent (Ondetti and Pluscec, 1980) with minor modifications. Briefly, in a first step, 2-benzylpropenoic acid was coupled with glycine methyl ester, using N,N'-dicyclohexylcarbodiimide and triethylamine in ethyl acetate to yield N-(3phenyl-2-methylidene-l-oxopropyl)glycine methyl ester as a yellowish oil (yield, 69%, riD20 1.5325). This compound then underwent cleavage of its ester function in alkaline medium, yielding 80% of N-(3-phenyl-2-methylidene-l-oxopropyl)glycine as a whitish solid, prior to condensation with thioacetic acid, thus yielding the target compound N-{2-[(acetylthio)methyl]-l-oxo-3phenylpropyl}glycine (melting point: 104-105°C; mass spectra: m / e 296 [(MH)+]).

2.3. Drugs and chemicals Thiorphan was purchased from Sigma and naloxone hydrochloride (Narcan), from Dupont de Nemours. Acetorphan was synthesized as described by Roques et al. (1981). All compounds were stored at - 18°C. Solutions were always freshly prepared prior to use. For the intravenous route, dimethyl sulfoxide (DMSO, 99%,




d d d



Time (h) Fig. 2. Analgesic potency of acetylthiorphan (hatched area) compared to the reference enkephalinase inhibitors, thiorphan (black area) and acetorphan (white area) (i.v., 40 ~mol/kg). The results are expressed as mean jump latency+ S.E.M. (s). t-Test: * P < 0.1, ** P _<0.05, * * * P<0.01; * acetylthiorphan compared to acetorphan, ~' acetylthiorphan and acetorphan compared to thiorphan.

Aldrich Chemical Company) was used as vehicle. The administered volume was 2 ml/kg.

2.4. Neutral endopeptidase inhibition measurement Neutral endopeptidase inhibition was measured in the laboratory of Professor B.P. Roques as already described, using a purified enzyme preparation (Fourni6-Zaluski et al., 1984a). Briefly, after preincubation of samples of the purified enzyme (10/~1) in Tris-HCl buffer pH 7.4 (15 min, 25°C) with various concentrations of the tested inhibitor, [3H][D-AlaZ-LeuS]enkephalin (20 nM, final concentration) was added with four concentrations of unlabelled [D-Ala2-LeuS]enkephalin (between 0.3/~M and 10 IzM) in the mixture for 15 min at 25°C. The reaction was stopped by addition of 25 ml of 0.2 M HCI. The [3H][Tyr-D-AIa2Gly] metabolite was separated by Porapak column chromatography and the radioactivity was determined by liquid scintillation counting. Determinations were performed in triplicate.

2.5. Analgesic evaluation







Fig. 1. Structures of neutral endopeptidase inhibitors.

2.5.1. Hot-plate test This test was derived from that described by Eddy and Leimbach (1953). A glass cylinder 20 cm high and 9.5 cm in diameter was used to keep the mouse on the heated surface of the plate. The plate was heated to a temperature of 55 + 0.5°C, using a thermoregulated circulating water pump. A minimum of ten animals were used for each time point. Jump latencies and hindlimb paw licking times were recorded. The results are expressed as the mean + S.E.M. (s). Acetylthior-

131 200.

phan was compared under the same conditions to thiorphan (at 40 and 200/~mol/kg) and to acetorphan (at 40 and 200/~mol/kg). Groups of 10-19 mice were injected i.v. into a tail vein with vehicle, thiorphan, acetorphan or acetylthiorphan. Jump latencies and hindpaw licking times were recorded 20 rain, 1 h, 2 h, 3 h, 4 h, 5 h and 7 h later. The reversal of analgesia by naloxone was also tested. The protocol is presented in the legend to fig. 3.




8 v


2.5.2 Tail-flick test The antinociceptive responses were determined according to the method of D'Amour and Smith (1941). The tail-flick latency was measured 1 h after i.v. injection. A minimum of eight mice per compound were used to study the analgesic responses to acetorphan and acetylthiorphan at 200/~mol/kg.



Without naloxone

With naloxone

Fig. 3. Naloxone reversibility of acetylthiorphan and acetorphan analgesia 1 h after i.v. administration. The left histograms present the jump latency values without administration of naloxone, the right ones present the jump latency values with preliminary subcutaneous (s.c.) treatment with naloxone 1 m g / k g . The controls (black area) received dimethyl sulfoxide i.v. 1 h before the test, then saline s.c. 30 min before the experiment; treated mice received acetorphan (white area) or acetylthiorphan (hatched area) at the dose of 200 izmol/kg i.v. 1 h before the test, then naloxone s.c. 30 rain before the experiment. The results are expressed as percentages of control (without naloxone pretreatment). Statistical analysis: * P < 0.1, ** P < 0.05, and * * * P < 0.01 as compared to control mice; <>P < 0.1, '><>P < 0.05, and ~ P < 0.01 illustrate the comparison between corresponding treatments with and without naloxone.

3. Results

3.1. Neutral endopeptidase inhibition measurement The potency of acetylthiorphan as neutral endopeptidase inhibitor was compared to that of thiorphan and acetorphan. The IC50 of thiorphan was 1.8 + 0.2 nM while acetorphan was 1000 times less potent. As expected, the ICs0 of acetylthiorphan (316 + 38 nM) lay between these two values.

3.2. Analgesic evaluation the hot-plate test are shown in table 1. First, dimethyl sulfoxide, the vehicle used in this study, did not itself exert significant antinociceptive effects. The jump latency observed with a parallel group without injection, 54 + 5 (n = 26), was never statistically significantly different at any vehicle point. The mean latency variation (43-53 s) of mice receiving the vehicle can be explained by the diurnal rhythm of response to this painful stimulus (Frederickson et al., 1977).

As in the neutral endopeptidase inhibition measurement, acetylthiorphan was compared to thiorphan and acetorphan in two well-known analgesic tests: the hotplate test and the tail-flick test. In the hot-plate experiment, the two values that can be recorded, i.e. hindlimb paw licking latency and jump latency, are of different relevance regarding analgesia (Frederickson et al., 1977). Jump latencies observed in TABLE l

Antinociceptive activity after i.v. administration of acetylthiorphan, thiorphan, acetorphan in the hot-plate test. The results are expressed as mean jump latency+ S.E.M. (n > 10). Compounds


Jump latency (s)



20 min







43 + 3 52 5- 5 a 53 + 4 a 64 + 8 c

45 + 5 56 + 6 50___5 67 + 6 b

49 + 5 58 + 6 52 + 5 44 ___3

46 + 3 64 + 5 c 47 + 5 51 5- 3

49+8 57+6 57+7 44+4

53+5 56+6 51+3 57:t:6

59 + 5 c 55 + 4 b 62 5- 7 c

64 + 4 c 55 + 7 69 + 9 b

58 + 3 51 + 6 65 5- 9

65 + 4 c 50 + 8 60 + 6 b

68+5 a 47+8 47+4

53+4 52+4 59+5

Vehicle Acetylthiorphan Thiorphan Acetorphan

40 40 40

44 + 49 + 43 + 53 +

Acetylthiorphan Thiorphan Acetorphan

200 200 200


4 3 5 3a

a Significantly different from vehicle group with P < 0.1. b Significantly different from vehicle group with P =<0.05. c Significantly different from vehicle group with P < 0.01.


Thiorphan showed very weak activity: compared to vehicle, thiorphan was only active at 1 h with a P < 0.1 at 40 ~ m o l / k g and with a P < 0.05 at 200/zmol/kg. Somewhat higher activity was found for acetylthiorphan. Compared to vehicle, acetylthiorphan showed significant analgesic activity at 200/zmol/kg and at 40 txmol/kg. No significant differences were observed between acetylthiorphan and acetorphan at 200/~mol/ kg, except that acetylthiorphan was more active at 5 h (P < 0.05). At 40/~mol/kg, acetylthiorphan was significantly more active than acetorphan at 3, 4 and 5 h (fig. 2). In the hot-plate hindpaw licking test, no activity was evidenced for acetorphan and acetylthiorphan, consistent with results of previous studies on thiorphan and acetorphan (Roques et al., 1980; Lecomte et al., 1986). No significant differences were found compared to the vehicle values (14.9 + 1.4 s). The reversibility of analgesia by a prior subcutaneous administration of naloxone was investigated for acetorphan and acetylthiorphan. The data are shown in fig. 3. The analgesia was completely reversed by naloxone both with acetorphan and with acetylthiorphan, confirming an opioid mechanism of analgesia. In the tail-flick test, there were no significant differences in flick latency between treated (with acetylthiorphan or acetorphan) and control mice (table 2).

4. Discussion

Since the biotransformation of acetorphan to acetylthiorphan in serum has been demonstrated (Fourni6Zaluski et al., 1992), the present study was aimed at evaluating the analgesic activity of this metabolite in comparison with that of two well-established enkephalinase inhibitors, thiorphan and acetorphan. To our knowledge and despite the fact that the synthesis of acetylthiorphan has been described previously in a patent (Ondetti and Plusec, 1980), the analgesic potency of this compound has never been reported. In vitro measurements of neutral endopeptidase inhibition show that acetylthiorphan is a weaker inhibitor

TABLE 2 Lack of analgesia in the tail-flick test 1 h after administration of 200 ~tmol/kg acetylthiorphan and acetorphan. T h e results are expressed as m e a n flick latency+ S.E.M. (n __>8). Compounds

Dose (~,mol/kg)

Flick latency (s)

Vehicle Acetorphan Acetylthiorphan

200 200

3.2 + 0.4 3.6 + 0.8 3.6 + 0.6

than thiorphan but stronger than acetorphan. The result is in accordance with the mechanism of action of this class of inhibitors. Four functions are essential for efficient neutral endopeptidase inhibition: a mercapto function acting as ligand of the zinc ion, a carboxylic moiety, an aromatic residue accommodated by the hydrophobic pocket and an amide bond (Fourni6-Zaluski et al., 1984a; Gordon et al., 1983; Monteil et al., 1992; Roques et al., 1982, 1983). Acetylation of the mercapto function, which results in a markedly decreased tendency of sulfur to complex with class (b) metal ions, brings about a dramatic drop, about 175fold, of IC50 values: 316 nM for acetylthiorphan versus 1.8 nM for thiorphan. In comparison, the esterification of the carboxylic function of thiorphan by a methyl group decreases its inhibitory activity only 4-fold. Considering its low IC50 compared to thiorphan, the analgesic potency of acetylthiorphan therefore seems due to its in vivo hydrolysis to thiorphan. As can be seen in table 1, the activity of acetylthiorphan in the hot-plate jump test was of the same intensity as that of the i.v. injection of an equimolar dose of acetorphan and lasted even longer. The analgesia was completely reversed by naloxone. The well-known hyperalgesic effect of naloxone per se was confirmed (Frederickson et al., 1977; Roques et al., 1980; Lecomte et al., 1986) and is probably due to the blockade of an opioid-mediated response induced by the handling technique. Previous studies used the tail withdrawal test to investigate the analgesic potency of neutral endopeptidase inhibitors: neither thiorphan nor acetorphan exhibit any statistically significant activity even after i.c.v. administration (Roques et al., 1980; Lecomte et al., 1986). Our data from the tail-flick test (table 2) confirm these results. The lack of activity of neutral endopeptidase inhibitors in the tail-flick test is usually ascribed to the weak degree of analgesia elicited by these compounds, lower than with exogenous opioids, probably due to their selectivity versus one of the enkephalin degrading enzymes. Fourni6-Zaluski and collaborators, in order to overcome this limitation, have developed molecules such as kelatorphan, able to inhibit both neutral endopeptidase and aminopeptidase N (Fourni6-Zaluski et al., 1984b) and, more recently, 'mixed inhibitor-prodrugs' consisting of the chemical association of two inhibitors linked by an ester, thioester or disulfide bond (Fourni6-Zaluski et al., 1992). Even in the absence of an analgesic effect of dimethyl sulfoxide, the influence of this vehicle has to be considered. Many reports in the literature reach contradictory conclusions about the influence of DMSO on brain penetration of drugs (Broadwell et al., 1982; Iwen and Miller, 1986; Waiters et al., 1984). In the present study, thiorphan administered i.v. in DMSO solution had very low analgesic activity (one time,


P<0.1). This fact argues against the influence of DMSO, at least at the concentration used, on the blood-brain barrier integrity. In terms of drug design, it is obvious that the presence of the free carboxylic function in acetylthiorphan instead of the benzyl ester in acetorphan brings about a drastic diminution of the lipophilicity and consequently is expected to decrease the ability of acetylthiorphan to cross the blood-brain barrier. In this connection, De la Baume et al. (1988) have studied the distribution of [3H]acetorphan in mice. The total radioactivity was distributed heterogeneously and decreased rapidly except in the kidney. The authors identified in this organ the main metabolite corresponding to thiorphan (about 60% of the recovered radioactivity) 30 min after i.v. injection, whereas acetorphan, thiorphan benzyl ester or thiorphan disulfide were not detectable. This experiment demonstrated the rapid hydrolysis of the thiol and benzyl esters of acetorphan after i.v. administration. Moreover, Fourni6-Zaluski et al. (1992) recently reported hydrolysis of the benzyl ester in the serum. These two studies indicate that cleavage of the ester function occurs in vivo. According to this hypothesis, the metabolite crossing the bloodbrain barrier should be acetylthiorphan. An additional point is that thiorphan itself exhibits, under the same conditions, rather weak activity 1 h after i.v. administration. This observation implies that a sufficient amount of thiorphan is able to reach the brain. Therefore, the higher degree of analgesia elicited by acetylthiorphan and acetorphan could be explained by increased cerebral penetration a n d / o r by protection of the thiol against metabolism. In the absence of the demonstration of a receptoror carrier-mediated mechanism of cerebral penetration for neutral endopeptidase inhibitors, our data indicate that the single protection of the thiol function ensures good bioavailability after systemic administration and sufficient lipophilicity to cross the blood-brain barrier. In conclusion, S-acetylthiorphan, a metabolite of acetorphan, presents analgesic properties similar to those of the parent drug: activity in the hot-plate jump latency test and inactivity in the hot-plate paw licking time or in the tail-flick test. Considering the lesser ability of S-acetylthiorphan to bind the neutral endopeptidase enzyme compared to thiorphan, its analgesic activity obviously seems due to a prodrug mechanism.

Acknowledgements The authors wish to thank Professors B.P. Roques and M.C. Fourni6-Zaluski from the Universit6 Ren6 Descartes (Paris, France) for the determinations of neutral endopeptidase inhibition, as well as for stimulating discussions.

References Broadwell, R.D., M. Saleman and R.S. Kaplan, 1982, Morphologic effect of dimethylsulfoxide on the blood-brain barrier, Science 217, 164. D'Amour, F.E. and D.A. Smith, 1941, A method for determining loss of pain sensation, J. Pharmacol. Exp. Ther. 70, 74. De la Baume, S., F. Brion, M. Dam Trung Tuong and J.C. Schwartz, 1988, Evaluation of enkephalinase inhibition in the living mouse, using [3H]-acetorphan as a probe, J. Pharmacol. Exp. Ther. 247, 653. Eddy, N.B. and D. Leimbach, 1953, Synthetic analgesics. II: Dithienylbutenyl- and dithienylbutylamines, J. Pharmacol. Exp. Ther. 107, 385. Fourni6-Zaluski, M.C., E. Lucas, G. Waskman and B.P. Roques, 1984a, Differences in the structural requirements for selective interaction with neutral metalloendopeptidase (enkephalinase) or angiotensin enzyme. Molecular investigation by use of new thiols inhibitors, Eur. J. Biochem. 139, 267. Fourni6-Zaluski, M.C., P. Chaillet, R. Bouboutou, A. Coulaud, P. Ch6rot, G. Waksman, J. Costentin and B.P. Roques, 1984b, Analgesic effects of kelatorphan, a new highly potent inhibitor of multiple enkephalin degrading enzyme, Eur. J. Pharmacol. 102, 525. Fourni6-Zaluski, M.C., P. Coric, S. Turcaud, E. Lucas, F. Noble, R. Maldonado and B.P. Roques, 1992, 'Mixed inhibitor-prodrug' a new approach toward systemically active inhibitors of enkephalin-degrading enzymes, J. Med. Chem. 35, 2473. Frederickson, R.C.A., V. Bergers and J.D. Edwards, 1977, Hyperalgesia induced by naloxone follows diurnal rhythm in responsivity to painful stimuli, Science 198, 756. Gordon, E.M., D.W. Cushman, R. Tung, H.S. Cheung, F.L. Wang and N.G. Delaney, 1983, Rat brain enkephalinase: characterization of the active site using mercaptopropanoyl amino acid inhibitors, and comparison with angiotensin-converting enzyme, Life Sci. 33 (Suppl. 1), 113. Hersh, L.B., 1981, Solubilization and characterization of two rat brain membrane-bound aminopeptidases active on Met-enkephalin, Biochemistry 20, 2345. Iwen, P.C. and N.G. Miller, 1986, Enhancement of ketoconazole penetration across the blood-brain barrier of mice by dimethyl sulfoxide, Antimicrob. Agents Chemother. 30, 617. Knisely, J.S., P.M. Beardsley, M.D. Aceto, R.L. Balster and L.S. Harris, 1988, Assessment of the abuse potential of acetorphan, an enkephalinase inhibitor, Drug. Alcohol Dependence 23, 143. Lecomte, J.M., J. Costentin, A. Vlaiculescu, P. Chaillet, H. MarcaisCollado, C. Llorens-Cortes, M. Leboyer and J.C. Schwartz, 1986, Pharmacological properties of acetorphan, a parenterally active 'enkephalinase' inhibitor, J. Pharmacol. Exp. Ther. 237, 937. Monteil, T., M. Kotera, L. Duhamel, P. Duhamel, C. Gros, N. NoEl, J.C. Schwartz and J.M. Lecomte, 1992, Importance of the amide bond of thiorphan in the inhibitor-enkephalinase docking process demonstrated with some thiorphan isosteres, Bioorg. Med. Chem. Lett. 2, 949. Ondetti, M.A. and J. Pluscec, 1980, Mercaptoacylpeptide, Verfahren zu ihrer Herstellung und ihre Verwendung als ACE-Inhibitoren und bei der Behandlung von Hochdruck (Squibb & Sons, Ger. Offen 3.012.140). Rich, D.H., 1990, Peptidase inhibitors, in: Comprehensive Medicinal Chemistry: The Rational Design, Mechanistic Study and Therapeutic Application of Chemical Compounds, Vol. 2: Enzymes and other Molecular Targets, ed. C. Hansch (Pergamon Press, Oxford) p. 391. Roques, B.P., E. Lucas-Soroca, P. Chaillet, J. Costentin and M.C. Fourni6-Zaluski, 1980, The enkephalinase inhibitor thiorphan shows antinociceptive activity in mice, Nature 288, 286.

134 Roques, B., J.C. Schwartz and J.M. Lecomte (Soci&6 Civile Bioprojet, Paris, France), 1981, Aminoacid Derivatives and their Therapeutic Applications, European Patent 0038758 - U.S. Patent 4000513009. Roques, B.P., M.C. Fourni6-Zaluski, D. Florentin, G.Waksman, A. Sassi, P. Chaillet, H. Collado and J. Costentin, 1982, New enkephalinase inhibitors as probes to differentiate 'enkephalinase' and angiotensin-converting-enzyme active sites, Life Sci. 31, 1749.

Roques, B., E. Soroca, P. Chaillet, J. Costentin and M.C. Fourni6Zaluski, 1983, Complete differentiation between enkephalinase and angiotensin-converting enzyme inhibition by retro-thiorphan, Proc. Natl. Acad. Sci. USA 80, 3178. Schwartz, J.C., J. Costentin and J.M. Lecomte, 1985, Pharmacology of enkephalinase inhibitors, Trends Pharmacol. Sci. 6, 472. Waiters, A., V. Jackson-Lewis and S. Fahn, 1984, Little effect of dimethyl sulfoxide on blood-brain barrier to dopamine, Experientia 40, 859.