Mefloquine and Psychotomimetics Share Neurotransmitter Receptor and Transporter Interactions in Vitro

Mefloquine and psychotomimetics share neurotransmitter receptor and transporter interactions in vitro

Janowsky, A., Eshleman, A. J., Johnson, R. A., Wolfrum, K. M., Hinrichs, D. J., Yang, J., ... & Riscoe, M. K. (2014). Mefloquine and psychotomimetics share neurotransmitter receptor and transporter interactions in vitro. Psychopharmacology, 231(14), 2771-2783. doi:10.1007/s00213-014-3446-0

10.1007/s00213-014-3446-0 Springer

Version of Record http://cdss.library.oregonstate.edu/sa-termsofuse Psychopharmacology (2014) 231:2771–2783 DOI 10.1007/s00213-014-3446-0

ORIGINAL INVESTIGATION

Mefloquine and psychotomimetics share neurotransmitter receptor and transporter interactions in vitro

Aaron Janowsky & Amy J. Eshleman & Robert A. Johnson & Katherine M. Wolfrum & David J. Hinrichs & Jongtae Yang & T. Mark Zabriskie & Martin J. Smilkstein & Michael K. Riscoe

Received: 15 April 2013 /Accepted: 7 January 2014 /Published online: 2 February 2014 # Springer-Verlag Berlin Heidelberg (outside the USA) 2014

Abstract Results Hallucinogens and mefloquine bound stereoselectively Rationale Mefloquine is used for the prevention and treat- and with relatively high affinity (Ki=0.71–341 nM) to serotonin ment of chloroquine-resistant malaria, but its use is associated (5-HT) 2A but not 5-HT1A or 5-HT2C receptors. Mefloquine but with nightmares, hallucinations, and exacerbation of symp- not chloroquine was a partial 5-HT2A agonist and a full 5-HT2C toms of post-traumatic stress disorder. We hypothesized that agonist, stimulating inositol phosphate accumulation, with sim- potential mechanisms of action for the adverse psychotropic ilar potency and efficacy as the hallucinogen dimethyltrypta- effects of mefloquine resemble those of other known mine (DMT). 5-HT receptor antagonists blocked mefloquine’s psychotomimetics. effects. Mefloquine had low or no affinity for dopamine D1,D2, Objectives Using in vitro radioligand binding and functional D3,andD4.4 receptors, or dopamine and norepinephrine trans- assays, we examined the interaction of (+)- and (−)-meflo- porters. However, mefloquine was a very low potency antago- quine enantiomers, the non-psychotomimetic anti-malarial nist at the D3 receptor and mefloquine but not chloroquine or agent, chloroquine, and several hallucinogens and hallucinogens blocked [3H]5-HT uptake by the 5-HT psychostimulants with recombinant human neurotransmitter transporter. receptors and transporters. Conclusions Mefloquine, but not chloroquine, shares an in vitro receptor interaction profile with some hallucinogens and this neurochemistry may be relevant to the adverse neu- : : : : A. Janowsky A.: J. Eshleman R. A. Johnson D. J. Hinrichs ropsychiatric effects associated with mefloquine use by a M. J. Smilkstein M. K. Riscoe small percentage of patients. Additionally, evaluating interac- Research Service (R&D22), VA Medical Center, 3710 SW US Veterans Hospital Road, Portland, OR 97239, USA tions with this panel of receptors and transporters may be useful for characterizing effects of other psychotropic drugs : : : A. Janowsky A. J. Eshleman R. A. Johnson K. M. Wolfrum and for avoiding psychotomimetic effects for new pharmaco- Departments of Psychiatry and Behavioral Neuroscience, Oregon therapies, including antimalarial quinolines. Health and Science University, Portland, OR 97239, USA

A. Janowsky (*) Keywords Mefloquine . Chloroquine . Quinine . Malaria . The Methamphetamine Abuse Research Center, Oregon Health and LSD . Psychotomimetic . Neurotransmitter . Transporter . Science University, Portland, OR 97239, USA Serotonin receptor . Dopamine receptor e-mail: [email protected]

D. J. Hinrichs : M. K. Riscoe Department of Molecular Microbiology and Immunology, Oregon Introduction Health and Science University, Portland, OR 97239, USA

J. Yang : T. M. Zabriskie Mefloquine ([(R*,S*)-2,8-bis(trifluoromethyl)quinolin-4- Department of Pharmaceutical Sciences, College of Pharmacy, yl]-(2-piperidyl)methanol; Lariam™) is a synthetic derivative Oregon State University, Corvallis, OR 97331, USA of quinine ((R)-(6-methoxyquinolin-4-yl) ((2S,4S,8R)-8- vinylquinuclidin-2-yl)methanol) that has been used by mil- M. J. Smilkstein : M. K. Riscoe Department of Chemistry, Portland State University, Portland, lions of people worldwide to prevent and to treat symptoms of OR 97201, USA malaria following exposure to Plasmodium falciparum and 2772 Psychopharmacology (2014) 231:2771–2783

other Plasmodium species. The antimalarial mechanism of Gillespie et al. 2008), and 5-HT3 receptors (Thompson et al. action of mefloquine is not completely understood, but may 2007; Thompson and Loomis 2008). However, there is little include alteration of heme-iron transport and disposition across or no information available concerning the interaction of LSD the parasite digestive vacuole and cytoplasm, and inhibition of and related psychotomimetic agents with many of these re- cellular crystalline hemozoin formation (Haynes et al. 2012; ceptors (for a compendium of values and references, see http:// Combrinck et al. 2013). The drug is used in lower doses, once pdsp.med.unc.edu/kidb.php). weekly for prophylaxis, and in higher, more frequent doses to Drugs with differing chemical structures including LSD, treat acute infections. Side effects after prophylactic and, partic- 2,5-dimethoxy-4-methylamphetamine (DOM), dimethyltryp- ularly, therapeutic use has precluded more widespread use of tamine (DMT), and other hallucinogens share a number of mefloquine (Kennedy 2009). Although some side effects are common neurochemical effects that may be related to their common among antimalarials, psychotropic effects similar to psychotropic activity. For example, some hallucinogens bind those of mefloquine are not usually found in subjects taking to and stimulate specific 5-HT receptors (Nichols et al. 2002; other antimalarial agents such as chloroquine ((RS)-N'-(7- Rabin et al. 2002; Kanagarajadurai et al. 2009; for review see chloroquinolin-4-yl)-N,N-diethyl-pentane-1,4-diamine)), which Halberstadt and Geyer 2011) and LSD interacts with specific is in the 4-aminoquinoline class. Unwanted effects of mefloquine dopaminergic and noradrenergic receptors (Minuzzi and include sleep and dream disturbances (Toovey 2009), depression Cumming 2010). In contrast, many first generation antipsy-

(but see Schlagenhauf et al. 2009), hallucinations, and anxiety in chotic agents block dopamine D2 receptors, and second gen- a small but significant proportion of patients, and some of the eration antipsychotic drugs block D2 and (or) 5-HT2 receptors deleterious side effects have been reported to continue after (Miyake et al. 2012). Additionally, effects of these drugs on discontinuation of drug treatment (van Riemsdijk et al. 2005), neurotransmitter transporters vary depending on chemical suggesting that the drug has neurotoxic effects (see AlKadi 2007 structure, and may account for differences in the behavioral for review). The severe neuropsychiatric side effects of the drug properties of the drugs. However, despite its world-wide ther- have resulted in the very recent inclusion of a “black box” apeutic use and its unwanted behavioral effects, relatively little warning by the United States Food and Drug administration, is known about the receptor pharmacology of mefloquine. and discussion of the severity and frequency of the side effects We hypothesized that mefloquine’s pharmacological pro- has reached the popular press (MacLean 2013). file at neurotransmitter receptors and transporters in vitro Recent evidence suggests that mefloquine may cause oxi- would resemble the effects of some known psychotomimetic dative stress, alter neuronal morphology (Hood et al. 2010), agents, and tested this with in vitro models using recombinant and exert apoptotic effects that are intimately involved in neurotransmitter receptors and transporters. In addition, we neurotoxicity via interaction with non-receptor tyrosine kinase speculated that these results might make mefloquine a useful 2 (Pyk2) (Milatovic et al. 2011). tool to more fully characterize the pharmacological profile Additionally, a number of receptor-based neuropharmaco- required for psychotomimetic activity. Lastly, as in our previ- logical etiologies have been invoked to explain the psychiatric ous work (Kelly et al. 2009), we anticipate that exploiting effects of mefloquine. However, few — if any — of these neurotransmitter receptor and transporter interaction profiles pharmacological effects resemble those of other psychotomi- may add useful information with which to develop chemo- metic agents. For instance, mefloquine does not appear to therapeutic agents with potentially fewer psychotropic effects. interact with glutamate receptors (Caridha et al. (2008), but The results indicate that mefloquine has affinity for specific 5- the psychotropic effects of lysergic acid diethylamide (LSD) HT and dopamine receptors, and in assays of receptor func- may include indirect changes in the regulation of the N-meth- tion, is a partial agonist at the 5-HT2A receptor, a full agonist at yl-D-aspartate (NMDA) subtype of glutamate receptors 5-HT2C receptors, and an antagonist at the dopamine D3 (Marona-Lewicka et al. 2011), and the psychotomimetic, receptor. Additionally, mefloquine binds to and blocks recom- phencyclidine, is an NMDA receptor antagonist (Thomson binant human 5-HT transporters (hSERT), and may increase et al. 1985). In addition, there is literally no evidence synaptic 5-HT availability and stimulate specific aspects of concerning the interactions of LSD and other psychotomi- serotonergic function. Thus, mefloquine shares pharmacody- metics with gap junction elements, but mefloquine (25 μM) namic effects with other psychotomimetic agents. blocks a number of connexins (Cruikshank et al. 2004; Iglesias et al. 2008;Wangetal.2010) and is now commonly used as a research tool to block gap junction channels (Sarihi Materials and methods et al. 2012). Mefloquine also interacts with gamma- aminobutyric acid A (GABAA) receptors (Amabeoku and Materials Farmer 2005; Thompson and Loomis 2008), and interacts with peripheral benzodiazepine receptors (Dzierszinski et al. Racemic mefloquine, its individual enantiomers, and other

2002), adenosine A1 and A2A receptors (Weiss et al. 2003; anti-malarial agents were obtained from Walter Reed Army Psychopharmacology (2014) 231:2771–2783 2773

Medical Center or were synthesized, and structures verified in Gatch et al. 2011; Eshleman et al. 2013). For comparison of our laboratories (MZ, MR). Briefly, racemic mefloquine agonist affinities at [125I]DOI and [3H]ketanserin binding to

(BioBlocks) and (−)-3-bromo-8-camphosulfonic acid ammo- HEK-5-HT2A cells, the method was adapted from Sleight nium salt (Acros Organics) were mixed in aqueous methanol et al. (1996). In brief, well-washed membranes from HEK- to form the (+)-mefloquine·(−)-3-bromo-8-camphosulfonate h5-HT2A cells were resuspended in binding buffer (50 mM salt, which was further purified via recrystallization from Tris–HCl, pH 7.4 at 25°C, 5 mM MgCl2,10μM pargyline, aqueous methanol. Neutralization of this salt with 1 N NaOH, 0.1 % ascorbic acid). Assays were identical for each followed by recrystallization from aqueous methanol gave the radioligand and consisted of 50 μl membrane preparation, pure (+)-mefloquine. (−)-Mefloquine was obtained from the 25 μlradioligand,50μl of displacing compound, or buffer, above (−)-sulfonate filtrate after repeated recrystallizations as in a final volume of 250 μl. The reaction was incubated at previously described (Carroll and Blackwell 1974). 25°C for 1 h and terminated by filtration through FiltermatA LSD, DMT, DOM and cocaine were obtained from the filters (Perkin Elmer) presoaked in 0.05 % polyethylenimine.

National Institute on Drug Abuse Drug Supply Program (Be- Nonspecific binding was defined with 10 μM 5-HT. IC50 3 thesda, MD, USA) or from commercial sources. [ H]8-Hy- values were converted to Ki values (Cheng and Prusoff 125 3 125 droxy-2-(di-n-propylamino)tetralin (8-OH-DPAT), [ I]2,5- 1973). For [ H]ketanserin and [ I]DOI, the Kd values were dimethoxy-4-iodoamphetamine (DOI), [3H]7-chloro-3-meth- 0.52 and 3.62 nM and the final concentrations in the binding yl-1-phenyl-1,2,4,5-tetrahydro-3-benzazepin-8-ol assays were 1–1.5 and 0.04–0.06 nM, respectively. (SCH23390), [3H]N-(1-benzyl-2-methylpyrrolidin-3-yl)-5- For dopamine receptor binding assays, mouse fibroblast chloro-2-methoxy-4-(methylamino)benzamide (YM-09151- cells expressing the recombinant human D1 receptor at high 125 2, nemonapride), [ I]methyl (1R,2S,3S)-3-(4-iodophenyl)- density (LhD1 cells) and Chinese hamster ovary cells express- 8-methyl-8-azabicyclo[3.2.1]octane-2-carboxylate) (RTI-55), ing the recombinant human D2 or D3 receptor (CHOp-D2 or 3 3 3 [ H]dopamine, [ H]5-HT, and [ H]norepinephrine were pur- CHOp-D3) were obtained from Stanford Research Institute chased from Perkin Elmer Life and Analytical Sciences (Bos- (SRI). HEK cells co-expressing the human D4.4 receptor with ton, MA). Commonly used reagents were obtained from com- adenylate cyclase type 1 (HEK-D4.4-AC1 cells) were a gift mercial sources except where specified below. from Dr. Kim Neve (Portland, OR, USA). Radioligand binding assays for dopamine receptors were conducted as de- Tissue culture scribed previously (Toll et al. 1998;Eshlemanetal.2013).

Human embryonic kidney cells (HEK-293) were cultured and Recombinant human transporter binding and neurotransmitter transfected with the respective recombinant human receptor or uptake assays transporter using modifications of our previously described methods (Eshleman et al. 1999, 2013). HEK cells expressing the recombinant human dopamine transporter (HEK-hDAT), SERT (HEK- hSERT) or norepi- Receptor binding assays nephrine transporter (HEK-hNET) were used as described previously (Eshleman et al. 1999). Specific binding and up- Radioligand binding experiments were conducted by modifi- take were defined as the difference in binding or uptake cations of our previously described methods (Eshleman et al. observed in the presence and absence of mazindol (5 μM; 1999; Gatch et al. 2011; Eshleman et al. 2013) using some of HEK-hDAT and HEK-hNET) or imipramine (5 μM; HEK- the receptor and transporter characterization panels that have hSERT). For all receptor binding and uptake assays, three or been validated as part of the National Institute on Drug Abuse/ more independent competition experiments were conducted

Department of Veterans Affairs Interagency Agreement “In with duplicate determinations, unless the IC50 value for a drug Vitro Receptor, Transporter and Release Assays for NIDA was consistently greater than 10 μM, and then only two Medications Discovery and Abuse Liability Testing.” experiments were conducted.

Serotonin receptors Inositol-1-phosphate (IP-1) formation

Human embryonic kidney (HEK-293) cells expressing the HEK-h5-HT2A or -h5-HT2C cells were used to determine recombinant human 5-HT1A (HEK-h5-HT1A), 5-HT2A agonist activation of the recombinant receptor as measured (HEK-h5-HT2A)or5-HT2C (HEK-h5-HT2C) receptors were by accumulation of IP-1 resulting from stimulation of the used. The methods for membrane preparation, [3H]8-OH- phospholipase C pathway. Assays were conducted as de- 125 DPAT binding to HEK-h5-HT1A,[ I]DOI binding to HEK- scribed previously (Gatch et al. 2011; Eshleman et al. 2013). h5-HT2A and HEK-h5-HT2C cell membranes and data analy- Agonist effects were normalized to maximal stimulation by 5- sis was conducted as previously described (Knight et al. 2004; HTabove basal, and antagonists were tested in the presence of 2774 Psychopharmacology (2014) 231:2771–2783

100 nM 5-HT and normalized to inhibition by 30 μM Data analysis ketanserin (5-HT2A)or1μMSB242084(5-HT2C). For radioligand binding, data were normalized to the binding in the absence of a competitive (mefloquine, etc.) drug. Three

Dopamine D2 and D3 mitogenesis assays or more independent competition experiments were conducted with duplicate determinations. GraphPAD Prism was used

CHOp-D2 and CHOp-D3 cells were maintained in alpha- to analyze the subsequent data, with IC50 values converted to MEM with 10 % fetal bovine serum (FBS; Atlas Biologicals, Ki values using the equation (Ki =IC50/(1+([drug*]/Kd Fort Collins, CO, USA), penicillin–streptomycin, and 200 μg/ drug*))), where [drug*] is the concentration of the labeled ml of G418. The assays were adapted from previous methods ligand used in the binding assays (Cheng and Prusoff 1973).

(Toll et al. 1998). The cells were seeded in 96 well plates at a The Kd values used in the equations are listed in Eshleman density of 10,000 cells/well. After ~60 h, the cells were rinsed et al. (2013). Differences in affinities were assessed by one twice and then incubated for 24 h at 37°C with serum-free way ANOVA using the logarithms of the Ki values for test alpha-MEM. Serial dilutions of test compounds were made in compounds. Tukey’s multiple comparison test was used to serum-free alpha-MEM. The medium was removed from the compare the potencies and efficacies of test compounds. plate and replaced with 100 μl of test compound. After 16 h For functional assays, GraphPAD Prism is used to calculate 3 (D3)or24h(D2), [ H]thymidine (0.25 μCi) in alpha-MEM either EC50 (agonists) or IC50 (antagonists) values using data supplemented with 10 % FCS was added to each well and the expressed as % 5-HT stimulation for IP-1 formation and % plates were incubated for 2 h at 37°C. The cells were quinpirole stimulation for mitogenesis assays. For functional trypsinized by addition of 1 % trypsin solution and harvested assays, one way ANOVA was used to assess differences in using a Tomtec 96-well harvester (Hamden, CT, USA) and efficacies using normalized maximal stimulation, and differ- radioactivity remaining on filters was counted using a Perkin ences in potencies using the logarithms of the EC50 values for Elmer microbeta scintillation counter. test compounds. Tukey’s multiple comparison test was used to compare test compounds with significance set at p<0.05.

Dopamine D4.4 adenylate cyclase assay Results

HEK-D4.4-AC1 cells were plated at a density of 375,000 cells perwellin48wellplatesinDMEMsupplementedwith5% 5-HT receptors FetalClone (HyClone, Logan, UT, USA), 5 % bovine calf serum and penicillin–streptomycin. After ~36 h, the medium In a relatively small percentage of patients, the symptoms of was changed to DMEM supplemented with 10 % charcoal- mefloquine intoxication resemble those of some hallucino- stripped FetalClone. The medium was removed ~18 h later. gens, including LSD, DMT and DOI. These drugs are rela- For agonist assays, 0.8 ml EBSS (116 mM NaCl, 22 mM tively potent agonists at various 5-HT receptors, and so we glucose, 15 mM HEPES, 8.7 mM NaH2PO4,5.4mMKCl, examined the ability of (+)- and (−)-enantiomers of meflo- 1.3 mM CaCl2, 1.2 mM MgSO4, 1 mM ascorbic acid, 0.5 mM quine to displace radiolabeled agonist binding from the re- IBMX [3-isobutyl-1-methyl-xanthine] and 2 % BCS, combinant h5-HT1A, 2A,and2C receptors. The results in Fig. 1 pH 7.4 at 37°C) was added, cells were incubated 20 min, and Table 1 indicate that the enantiomer that is active against agonists were added, and, after 20 min incubation, 10 μM the plasmodium parasite, (+)-mefloquine, has higher affinity forksolin was added in a final volume of 1 ml. After 20 min for the h5-HT2A receptor, but equal affinity for the h5-HT1A incubation with forskolin, the reaction was terminated by and h5-HT2C receptor as compared to the (−)-enantiomer. The aspiration, and 0.1 ml trichloroacetic acid was added. Plates enantiomers displayed similar rank orders of affinity across were incubated for 2 h on a rotator. Adenylate cyclase activity receptors, and had highest affinity for the [125I]DOI binding was measured using a cyclic AMP EIA kit (Cayman, Ann site on the recombinant h5-HT2A receptor and lowest affinity Arbor, MI, USA). Aliquots (9 μl) of each well were diluted to for the [3H]8-OH-DPAT binding site on the recombinant h5-

200 μl with EIA buffer from the kit, and 50 μl of the dilution HT1A receptor. Compared to the mefloquine enantiomers, was added to the EIA plate. After addition of tracer and chloroquine had higher affinity for the h5-HT1A receptor monoclonal antibody, the EIA plates are incubated for 18 h (p<0.05, one way ANOVA followed by Tukey’s multiple at 4°C. The reaction was aspirated, plates were washed 5× comparison test), similar affinity for the h5-HT2C receptor 300 μl with wash buffer, and Ellman’s reagent was added. (p>0.05), and lower affinity for the h5-HT2A receptor After 2 h, the plates were read at 410 nm. Basal cAMP was (p<0.05). (+)-Mefloquine and DMT had similar affinity for subtracted from all values. The maximal receptor effect is the h5-HT2A receptor (p>0.05). Psychotropic compounds defined with 1 μM quinpirole. LSD and DOM had higher affinity for the h5-HT2A and h5- Psychopharmacology (2014) 231:2771–2783 2775

Ta b l e 1 Affinities of Mefloquine and other drugs for recombinant human serotonin receptors

Drug h5-HT1A h5-HT2A h5-HT2C [3H]8OH-DPAT [125I]DOI [125I]DOI Ki (nM)±SEM

(+)-Mefloquine >9,100a 341±67 5,730±770 (−)-Mefloquine >10 μM 1,510±260 3,870±620 Chloroquine 6,000±1,600 >6,300a 6,200±2,200 5-HT 2.67±0.33 9.1±1 4.31±0.93 LSD 2.78±0.51 0.71±0.17 2.91±0.75 DMT 450±150 210±43 166±50 DOM 14,200±4,600 4.3±1.2 25.7±4.2 Ketanserin 6,510±320 7.6±1.9 117±34 Ro60-0175 6,680±430 13.1±2.1 7.3±1.6 SB242087 >9,300# 120±32 11.4±4.0 WAY 100635 1.21±0.16 331±54 3,000±1,100

Radioligand binding assays were conducted as described in the text. Each Ki value represents the mean of at least three independent experiments, each conducted with duplicate determinations 8OH-DPAT 8-hydroxy-2-(di-n-propylamino) tetralin, DOI 2,5-dimethoxy- 4-iodoamphetamine; DMT N,N-dimethyltryptamine; DOM 2,5- dimethoxy-4-methylamphetamine a If some experiments yielded IC50 or Ki values less than 10 μM and other experiments yielded IC50 or Ki values greater than 10 μM, the latter experiments were assigned a value of 10 μM and averages calculated. The actual value is greater than that average and no standard error is reported

5-HT receptor agonist-stimulated IP-1 formation

IP-1 formation by cells expressing h5-HT receptors was measured following treatment of cells with various drugs. The data in Fig. 2 and Table 2 indicate that the rank order of potency for psychotropic agents was LSD>>DOM>DMT at both h5-

HT2A and h5-HT2C receptors. DMT was only a partial agonist at the h5-HT2A receptor (p<0.001, Tukey’smultiplecompar- Fig. 1 Displacement of radiolabeled agonist binding to h5-HT receptors ison test), but a full agonist at the h5-HT2C receptor (p>0.05). was conducted as described in Materials and methods. Assays were LSD and DOM approached full efficacy at both receptors. conducted with duplicate determinations, and were repeated at least three Interestingly, (+)-mefloquine was significantly more potent times. Ki values were derived from the Cheng–Prusoff correction and (about two orders of magnitude) at stimulating h5-HT2A and used radioligand Kd values as described in Materials and methods. DOM, DMT, and LSD are psychotomimetic drugs with relatively high affinity h5-HT2C receptors, as compared to (−)-mefloquine. Further, for 5-HT receptors and were included for purposes of comparison. The Ki both enantiomers were significantly more potent at stimulat- values are described in Table 1. 5-HT serotonin; LSD lysergic acid ing h5-HT receptors than h5-HT receptors (p<0.01, two- diethylamide; DMT N,N-dimethyltryptamine; DOM 2,5-Dimethoxy-4- 2C 2A methylamphetamine; 8OH-DPAT 8-hydroxy-2-(di-n-propylamino) tailed t-test). The EC50 values for (+)-mefloquine at the recep- tetralin; DOI 2,5-dimethoxy-4-iodoamphetamine tors (224 and 1990 nM at h5-HT2C and h5-HT2A receptors, respectively) indicated that it had similar h5-HT2C receptor potency as the psychoactive tryptamine, DMT (p>0.05, ’ HT2C receptors compared to the mefloquine enantiomers Tukey s multiple comparison test), but lower potency than (p<0.001). Additionally, LSD had high affinity for the the other psychotropic drugs at these receptors (p<0.01). 3 − [ H]8-OH-DPAT binding site on the h5-HT1A receptor (Ki= The (+)- and ( )-enantiomers had similar efficacy at the h5- 2.78 nM). Subsequently, experiments were conducted to de- HT2C receptor as compared to the full agonist, 5-HT (p>0.05). termine if mefloquine is an agonist at recombinant h5-HT However, (+)-mefloquine was only a partial agonist at the h5- receptors. HT2A receptor (p<0.001), and the (−)-enantiomer had almost 2776 Psychopharmacology (2014) 231:2771–2783

Ta b l e 2 Effects of mefloquine and other drugs on h5-HT2A- and h5- HT2C-receptor-mediated IP1 formation

Drug h5-HT2A h5-HT2C IP1 formation IP1 formation EC50 (nM)±SEM %stimulationa

(+)-Mefloquine 1,990±820 224±100 40.8±4.2 % 81.0±9.7 % (−)-Mefloquine >100 μM 12,000±3,700 <12 % 71±20 % Chloroquine >100 μM >100 μM <5 % <10 % effect 5-HT 51±14 1.52±0.58 99.7±1.6 % 93.2±3.0 % LSD 0.242±0.057 0.85±0.26 84.2±5.2 % 79±10 % DMT 227±64 96±34 41.0±6.2 % 96.1±9.3 % DOM 24.7±8.5 4.6±2.6 98.2±7.7 % 98±19 %

IC50 (nM)±SEM % inhibition Ketanserin SB242087 5.1±1.5 0.167±0.061 96.9±.2.1 % 93.7±5.5 %

Functional assays were conducted as described in the text. Each EC50 value represents the mean of at least three independent experiments, each conducted with duplicate determinations. IC50 values for antagonists are included to validate the assay 5-HT serotonin; IP1 inositol-1-phosphate; LSD lysergic acid diethylamide; DMT N,N-dimethyltryptamine; DOM 2,5-dimethoxy-4- methylamphetamine a Fig. 2 Stimulation of IP-1 formation was conducted as described in %stimulation— for each experiment, the maximal stimulation for each Materials and methods. Assays were conducted with duplicate determi- drug is normalized to the maximal stimulation above baseline for nations and experiments were repeated at least three times. a, b IP-1 serotonin formation is expressed as a percentage of IP-1 accumulation in response to the maximal effect of 5-HT. c Data are expressed as nM IP1 to allow comparison between HEK wild-type (HEK-wt), HEK-h5-HT2A and response in cells expressing the h5-HT2A or 2C receptor. As HEK-h5-HT2C cells. The drug concentrations were 1 μM5-HT, can be seen in Fig. 2c, there is little to no response to meflo- 100 μM (+)-mefloquine and 10 nM LSD. 5-HT serotonin; LSD lysergic quine in nontransfected, wild-type (wt) cells, as opposed to acid diethylamide; DMT N,N-dimethyltryptamine; DOM 2,5-Dimethoxy- transfected cells, in which mefloquine caused a 3- to 5-fold 4-methylamphetamine; DOI 2,5-dimethoxy-4-iodoamphetamine; IP1 inositol-1-phosphate increase above basal activity. To further verify that mefloquine is an agonist, we examined the concentration-dependent effect of the antagonist ketanserin on mefloquine-induced IP-1 ac- no effect on h5-HT2A receptor-mediated IP-1 formation. In cumulation in HEK- h5-HT2A cells, and the effect of the contrast to the other drugs, chloroquine had no effect on IP-1 antagonist SB202084 on mefloquine-induced IP-1 accumula- formation in either cell type. For comparison, the h5-HT2 tion in HEK- h5-HT2C cells. The data in Fig. 3 indicate that receptor antagonists ketanserin and SB242087 were tested. mefloquine is a partial agonist at h5-HT2A receptors as com- Both drugs potently inhibited 5-HT-mediated IP-1 formation pared to 5-HT and LSD, and that ketanserin dose dependently (Table 2). blocks the agonist effect of each drug. Additionally, meflo-

To verify that the effect of mefloquine on 5-HT receptor- quine is a full agonist at h5-HT2C receptors, as compared to mediated IP-1 accumulation was not a non-specific effect, we LSD and 5-HT, and the effects of all drugs on IP-1 accumu- characterized the effects of mefloquine and other drugs in lation are blocked in a dose-dependent and complete manner nontransfected HEK-293 cells. (+)-Mefloquine (100 μM), 5- by the antagonist SB202084. HT (1 μM), and LSD (10 nM) had no effect on IP-1 accumu- Sleight and coworkers (1996) examined the ability of drugs lation in nontransfected cells, but all produced a robust to displace antagonist ([3H]ketanserin), partial agonist ([3H]- Psychopharmacology (2014) 231:2771–2783 2777

4-bromo-2,5-dimethoxyphenylisopropylamine [DOB]), and 3 agonist ([ H]5-HT) radioligand binding to the 5-HT2A receptor. Additionally, Egan and coworkers (2000)used [3H]ketanserin [3H]DOB, [3H]5-HT and [3H]mesulergine to examine agonist affinity ratios (KL/KH)foranumberofago- nists at 5-HT2A and 5-HT2C receptors. The authors reported that agonists but not antagonists (Sleight et al. 1996) had higher affinity for [3H]agonist-labeled sites. We have now examined the ability of mefloquine and a number of other drugs to displace agonist and antagonist ligands in our cell lines (Fig. 4, Table 3). As indicated in Fig. 4 and Table 3,the agonists 5-HT, DOM and DMT have higher affinity for the agonist labeled site, however LSD did not have a signicantly higher affinity for the [125I]DOI-labeled site. Additionally, (+)-mefloquine did not have a statistically signficant higher affinity for the agonist-labeled site, although there was a large trend. (−)-Mefloquine tended to favor the antagonist-labeled site. As opposed to the previous report that antagonists do not have a significantly different affinity for agonist- or antago- Fig. 3 5-HT2A and 5-HT2C receptor antagonists inhibit (+)-mefloquine, nist–labeled h5-HT2A receptors (Sleight et al. 1996), the an- serotonin and LSD stimulated IP-1 formation. a h5-HT2A cells were tagonist MDL 100,907 had a significantly higher affinity for preincubated with ketanserin for 10 min before the addition of agonists the antagonist-labeled site. Thus, this assay did not allow us to (n=3–4). b h5-HT2C cells were preincubated with SB 242084 for 10 min contrast agonists versus antagonists based on differential af- before the addition of agonists (n=3) finities, and the question of agonist activity was addressed using functional assays described above.

Transporters is an agonist or antagonist at the hD2 or hD3 receptor, its effects on hD2 and hD3 receptor-mediated mitogenesis were Results in Table 4 indicate that the (+)- and (−)-enantiomers of characterized. Effects of quinpirole, a D2/D3 receptor agonist, mefloquine are very weak at the [125I]RTI-55 binding site of were compared to the effects of the other drugs. (+)-Meflo- the hNET and hDAT. However, (+)-mefloquine had similar quine had no agonist activity at the hD2 or hD3 receptors affinity to cocaine for the hSERT (p>0.05, two-tailed t-test). In (Fig. 5a and b; Table 6). LSD was a partial agonist at the D2 assays measuring the ability of drugs to block the uptake of receptor and a full agonist at the D3 receptor, as compared to 3 [ H]neurotransmitter by each transporter, (+)-mefloquine was quinpirole, with mid-nanomolar EC50 values. (+)-Mefloquine 3 equipotent with cocaine at blocking [ H]5-HT uptake by the also had no antagonist activity at the D2 receptor, but fully hSERT (p>0.05), and therefore could increase 5-HTavailabil- antagonized the effect of quinpirole at D3 receptors with an ity in the synapse. LSD and chloroquine had no effect on IC50 value of 3,210 nM (Fig. 5c and d; Table 6). The ability of 3 [ H]neurotransmitter uptake by any of the recombinant human drugs to inhibit hD4.4 receptor-mediated forskolin-stimulated transporters (Table 4). adenylate cyclase activity was also examined. LSD was a full agonist and inhibited 93±6 %, compared to the maximal Dopamine receptors inhibition by quinpirole, of forskolin-stimulated adenylate

cyclase activity with an EC50 value of 0.58±0.20 nM. How- The data in Table 5 indicate the effects of drugs on radioligand ever, (+)- and (−)-mefloquine and chloroquine had no effect binding to the recombinant dopamine hD1,hD2,hD3,and (data not shown). hD4.4 receptors. (+)-Mefloquine was selective, and blocked 3 [ H]YM 09151-2 binding to the hD3 receptor at a concentration that approaches its therapeutic concentration in blood Discussion

(Ki=1,960 nM), but had no effect on radioligand binding to the hD1,hD2,orhD4.4 receptors at concentrations up to The results of experiments described above indicate that both (+)- 10 μM. Chloroquine and (−)-mefloquine had no measurable and (−)-enantiomers of mefloquine interact with specific recom- effect on radioligand binding at any of the dopamine recep- binant human neurotransmitter receptors and transporters. These tors. LSD had moderate, mid-nanomolar affinity for the hD1, effects appear to differ from those of the anti-malarial agent, hD2,hD3,andhD4.4 receptors. To determine if (+)-mefloquine chloroquine, and from quinine, a structural analogue, and the 2778 Psychopharmacology (2014) 231:2771–2783

3 125 Fig 4 Displacement of [ H]ketanserin and [ I]DOI binding from h5-HT2A receptors by agonists and antagonist. Binding assays were conducted as described in Materials and methods. a (+)-Mefloquine, b LSD, c DOM, d 5-HT and e ketanserin concentration–response curves (n=4–9) effects of mefloquine are stereoselective. The relevance of these following a single i.v. dose, Dow et al. (2011) calculated a ratio of findings is particularly intriguing, given the far higher incidence 4.9/1,807, or 0.0027, with a whole brain concentration of about of serious mefloquine-induced psychotomimetic adverse effects 4 μM, and about 18 nM free brain mefloquine in mice. However, following high-dose therapeutic use, a setting in which 10-fold soldiers taking the drug prophylactically and who were killed in higher mefloquine concentrations have been reported (Simpson the line of duty had brain levels of 8.7–14 mg/kg (Jones et al. et al. 1999). Comparing free to whole mouse brain concentrations 1994). This should translate to about 100 to 135 nM free mefloquine. Pham and others (1999) measured mefloquine levels in brains from individuals who were taking the drug acutely Ta b l e 3 Comparison of Ki values of 5-HT2A agonists and antagonists at (750 mg, 37–70 h before death), and found 51.5 nmol/g tissue, [3H]ketanserin and [125I]DOI binding sites which should reach about 137 nM free drug. It is likely that 125 3 Ki [ I]DOI Ki [ H]ketanserin Ratio ketanserin/ humans taking repeated doses over time would have a much (nM) (nM) DOI Ki different ratio (more free drug), and so the whole brain/free ratio of 0.0027 may be a conservative estimate. Neuropsychiatric Agonists effects are reported by a relatively small percentage of patients 5-HT 3.734±0.92 12.6±1.8 3.37** who take mefloquine and so usual human brain concentrations (+)-Mefloquine 2,400±290 2,940±770 1.23 should not exceed the concentrations that have effects on trans- (−)-Mefloquine 4,260±840 3,280±1100 0.77 porters and receptors in vitro that are reported here. Tissue sample chloroquine 39,980±5,700 18,310±4,800 0.46* mefloquine levels in the 50–150 nM range are consistent with the LSD 0.72±0.17 0.475±0.094 0.66 finding that only a small number of patients, predisposed by DOM 2.06±0.52 13.7±6.0 6.65** dose, pharmacogenetics, liver function, size, co-ingestion of se- DMT 53±10 175±42 3.3* lective serotonin reuptake inhibitors, over-/underexpression of Antagonists some neurotransmitter signaling intermediates, etc., would devel- Ketanserin 4.1±1.7 0.77±0.26 0.19 op behavioral problems. MDL100,907 0.67±0.24 0.048±0.17 0.07** Because many psychotomimetic agents are agonists at h5- HT receptors, we hypothesized that mefloquine stimulates h5- Ki values were calculated from the IC50 values derived from the data in Fig. 4 HT receptors. The affinity of (+)-mefloquine for the [125I]DOI

Two-tailed t-test for each compound comparing the log Ki values for binding site of recombinant h5-HT2A receptors is similar to inhibition of [125 I]DOI and [3 H]ketanserin binding. 5-HT, DOM and the affinity of DMT (K =210 nM), a psychotomimetic, but 125 i DMT have higher affinity for the [ I]DOI binding site; chloroquine and DMT and mefloquine have much lower affinity than LSD and MDL100,907 have higher affinity for the [3 H]ketanserin binding site; and ketanserin, (+)-mefloquine, (−)-mefloquine and LSD have similar DOM at this receptor (Table 1). The Ki value for mefloquine affinity for both binding sites. n=3–8 independent experiments (341 nM), is below the minimum anti-plasmodial therapeutic *p<0.05; **p<0.01 concentration of the drug in blood (1.6 μM; see Schlagenhauf Psychopharmacology (2014) 231:2771–2783 2779

Ta b l e 4 Interaction of mefloquine and other drugs with recombinant et al. 2011; Cmax=1,018 μg/l≈2.6 μM, but varies by prepara- human neurotransmitter transporters tion; see Wiedekamm et al. 1998). However, free concentra- Dopamine Norepinephrine Serotonin tions of mefloquine in brain would be much smaller, as transporter transporter transporter indicated above. Likewise, (−)-mefloquine also binds to the Drug A Inhibition of h5-HT receptor, but its affinity (K =1,510 nM) is lower than [125I]RTI-55 binding 2A i the affinity of the (+)-enantiomer, and much lower than any of Ki (nM)±SEM the psychotropic drugs that exert effects via the h5-HT2A a (+)-Mefloquine >6,400 5,300±710 229±39 receptor. Although the Ki value for (−)-mefloquine is also (−)-Mefloquine 6,500±930 5,280±59 2,900±310 close to the minimum therapeutic drug concentration, free Chloroquine >8,000a 1,052±81 >8,300a brain levels would again be minimal by comparison. Thus it Quinine >8,300a >5,500a >4,000a is possible that both enantiomers contribute to any side effects LSD >10 μM 5,600±260 >10 μM in patients presenting with neuropsychiatric symptoms by Cocaine 416±59 410±120 450±140 their action at h5-HT2A receptors. Both enantiomers of mef- 125 Drug B Inhibition of [3H]neurotransmitter uptake loquine also displaced [ I]DOI binding from recombinant IC50(nM)±SEM h5-HT2C receptors but the Ki values for the antimalarials are (+)-Mefloquine >10 μM>10μM341±64much higher. Many psychotomimetics, including LSD, DOM, − μ μ a ( )-Mefloquine >10 M>10M >6,500 and DMT also interact with the h5-HT2C receptor (Tables 1 Chloroquine >10 μM>10μM>10μM and 2). Thus, it is possible that some of the psychotropic Quinine >10 μM >10 μM 1,650±420 effects of mefloquine, like those of the psychotomimetic drugs LSD >10 μM >10 μM >10 μM LSD, DMT and DOM, are mediated by interactions with both Cocaine 272±71 194±21 295±44 h5-HT2A and h5-HT2C receptors (Nichols 2004; Passie et al. 2008; Halberstadt and Geyer 2011). Specific [125 I]RTI-55 binding to HEK-293 cells expressing each of the neurotransmitter transporters was assessed as described in the text. Each The 5-HT receptors are coupled to multiple signal trans- Ki value represents the mean of at least three independent experiments, duction systems (Gatch et al. 2011), and we examined the each conducted with duplicate determinations. Cocaine, a neurotransmit- ability of mefloquine and other drugs to stimulate IP-1 forma- ter transporter blocker, was included for purposes of comparison tion. The data in Table 2 indicate that (+)-mefloquine is a full RTI-55 methyl (1R,2S,3S)-3-(4-iodophenyl)-8-methyl-8- agonist at recombinant h5-HT receptors, as compared to the azabicyclo[3.2.1]octane-2-carboxylate; LSD lysergic acid diethylamide 2C effects of 5-HT, and stimulates IP-1 formation in a dose- a If some experiments yielded IC or K values less than 10 μM and other 50 i dependent manner. However, (+)-mefloquine is weaker than experiments yielded IC50 or Ki values greater than 10 μM, the latter experiments were assigned a value of 10 μM and averages calculated. LSD, DOM, or DMT. In addition, (+)-mefloquine stimulates The actual value is greater than that average and no standard error is h5-HT2A receptor-mediated IP-1 formation. However, (+)- reported

Ta b l e 5 Interaction of mefloquine and other drugs with recombinant human dopamine receptors

Drug D1 D2 D3 D4.4 [3H]SCH23390 [3H]YM 09151-2 [3H]YM 09151-2 [3H]YM 09151-2 Ki (nM)±SEM

(+)-Mefloquine >10 μM>10μM 1,960±230 >10 μM (−)-Mefloquine >10 μM>10μM>10μM>10μM Chloroquine >10 μM>10μM>10μM>10μM LSD 273±89 38.8±1.8 27.1±1.3 330±120 DA 4,300±1,000 1,710±150 61±20 630±180 SKF 38393 258±51 4,800±2,600 4,300±1,800 28,000±13,000 SCH23390 0.491±0.087 887±17 1,920±580 9,700±1,100 Quinpirole >10 μM 2,780±510 52.8±6.1 262±66 Butaclamol 4.1±2.0 2.04±0.40 4.0±1.3 255±63

Radioligand binding assays were conducted as described in the text. Each Ki value represents the mean of at least three independent experiments, each conducted with duplicate determinations. For purposes of comparison, agonists and antagonists for each receptor were included in the assays (D1 receptor: SKF 38393 and SCH 23390; D2,D3 and D4.4 receptor: quinpirole and butaclamol) SCH23390 7-chloro-3-methyl-1-phenyl-1,2,4,5-tetrahydro-3-benzazepin-8-ol; YM-09151-2 nemonapride N-(1-benzyl-2-methylpyrrolidin-3-yl)-5- chloro-2-methoxy-4-(methylamino)benzamide; LSD lysergic acid diethylamide; DA dopamine; SKF38393 1-phenyl-2,3,4,5-tetrahydro-1H-3- benzazepine-7,8-diol 2780 Psychopharmacology (2014) 231:2771–2783

Fig. 5 Mefloquine is an antagonist of quinpirole-stimulated mitogenesis agonists, the data are normalized to the maximal stimulation by in CHOp-D3 cells. Assays were conducted with duplicate determinations. quinpirole. c, d For antagonists, the data are normalized to the stimulation The number of independent experiments is given in Table 6. Mitogenesis by 30 nM quinpirole. quin quinpirole activity is expressed in terms of [3H]thymidine incorporation. a, b For

mefloquine is only a partial agonist at this receptor, as com- is also a partial agonist at the h5-HT2A receptor, while LSD pared to 5-HT. On the other hand the psychotomimetic, DMT, and DOM are full agonists. Interestingly, (−)-mefloquine

weakly stimulates h5-HT2C receptors, but appears to have no

Ta b l e 6 Dopamine D2 and D3 receptor-mediated mitogenesis effects on h5-HT2A receptor-mediated IP-1 formation. Thus, the psychotropic effects of a racemic mixture of mefloquine Drug D2 mitogenesis D3 mitogenesis analogues may reside with the (+) enantiomer. The finding EC50 (nM)±SEM (n) % stimulation* that mefloquine is a partial to full agonist at h5-HT2C receptors is interesting, but could be the result of receptor over- (+)-Mefloquine NC (7) NC (4) expression, which can uncover partial agonist properties LSD 18.1±7.4 (6) 48±14 (7) (Gazi et al. 1999). However, the other psychotomimetic drugs 47.9±6.2 % 104.4±6.1 % that were examined in these assays had properties that were Quinpirole 27.7±8.0 (7) 2.31±0.47 (4) consistent with previous reports (see http://pdsp.med.unc.edu/ 99.1±2.0 % 99.4±5.1 % kidb.php), suggesting that the finding is not an artifact of cell Dopamine 56±13 (3) 1.0±0.40 (3) 110±0 % 117.5±2.5 % expression levels. Multiple reports detail the differences in the affinity of IC50 (nM)±range % inhibition agonists and antagonists at displacing agonist and antagonist (+)-Mefloquine NC (6) 3,210±470 (6) radioligands from 5-HT receptors (Sleight et al. 1996; Egan 97.3±2.7 % et al. 2000). Our data (Table 3,Fig.4)fromtheuseofagonist Butaclamol 0.028±0.011 (4) 0.78±0.22 (4) and antagonist radioligands are interesting but do not appear 98.1±1.9 % 96.4±2.8 % to decisively define agonists or antagonists. In agreement with the previous report, the agonist, 5-HT, had signficantly higher Assays were conducted in duplicate as described in Materials and 125 methods. Data represent the mean±SEM. (n) is the number of indepen- affinity for the [ I]DOI-labeled site. In addition, DOM, and dent experiments; NC the data did not converge. For purposes of com- DMT (Egan et al. 2000), had signficanlty higher affinity for parison, the dopamine D2-like receptor agonist, quinpirole, and the en- the agonist-labeled site. However, the agonist LSD did not dogenous agonist dopamine were included. The dopamine D2-like recep- have higher affinity for the [125I]DOI-labeled h5-HT receptor antagonist, butaclamol, was included for comparison, when investi- 2A gating the effects of drugs on antagonism of quinpirole (30 nM)-stimu- tor. Interestingly, LSD had a relatively low KL/KH ratio in the lated mitogenesis previous report (Egan et al. 2000). Differences in the findings Psychopharmacology (2014) 231:2771–2783 2781 could be explained by potential differences in the concentration drugs against the receptor panel described here could help to of guanine nucleotides across assays, as well as differences in identify therapeutic candidates with fewer psychotomimetic the definition and use of full and partial agonists. The over- effects. expression of receptors that has been shown to uncover partial agonist properties of antagonists (Gazi et al. 1999)couldalsobe Acknowledgements We thank Yuan Chou for technical expertise in the important since Ki values across agonist and antagonist binding mitogenesis assays. This work was supported by a grant from the Na- sites are being compared. However, the question of agonist tional Institute on Drug Abuse [1P50 DA018165], and NIH/VA Inter- agency Agreement [ADA 12013], a V.A. Merit Review [1I01BX000939- activity for mefloquine and other drugs used here are examined 01] and the V.A. Research Career Scientist Program (AJ), and by the Bill in assays of receptor function, as opposed to radioligand bind- and Melinda Gates Foundation (TMZ). ing, and indicate significant agonist activity in vitro. The (+) enantiomer of mefloquine also interacts with hSERT Conflict of interest There are no conflicts of interest. The authors have 125 full control of primary data and will allow the journal to review their data (Table 4). (+)-Mefloquine displaced [ I]RTI-55, a cocaine if requested. analogue, from hSERT with a Ki value of 229 nM, and blocked 3 [ H]5-HT uptake with an IC50 value of 341 nM. The (−) enantiomer and quinine were much weaker at blocking transporter References function, and chloroquine was essentially inactive. These data, and the results of experiments describing the effects of mefloquine at 5-HT receptors, above, suggest that mefloquine in- AlKadi HO (2007) Antimalarial drug toxicity: a review. Chemotherapy 53:385–391 creases 5-HT availability in the synapse and also directly stimu- Amabeoku GJ, Farmer CC (2005) Gamma-aminobutyric acid and lates 5-HT receptors. Thus, the combined pre- and post-synaptic mefloquine-induced seizures in mice. Prog Neuropsychopharmacol effects on serotonergic function are substantial and could play a Biol Psychiatr 29(6):917–921 role in its psychotomimetic effects. Chloroquine also interacted Caridha D, Yourick D, Cabezas M, Wolf L, Hudson TH, Dow GS (2008) 125 Mefloquine-induced disruption of calcium homeostasis in mamma- weakly (Ki=1,052 nM) with the [ I]RTI-55 binding site on the lian cells is similar to that induced by ionomycin. Antimicrob hNET, but had no effect on transporter function. At pharmaco- Agents Chemother 52(2):684–693 logically relevant concentrations the anti-malarial drugs did not Carroll FI, Blackwell JT (1974) Optical isomers of aryl-2- affect radioligand binding to, or the function of the hDAT piperidylmethanol antimalarial agents. Preparation, optical purity, and absolute stereochemistry. J Med Chem 17(2):210–219 (Table 4). Likewise, none of the hallucinogens/psychotropic Cheng Y, Prusoff WH (1973) Relationship between the inhibition con- agents that were tested had an effect on the dopamine transporter. stant (Ki) and the concentration of an inhibitor which causes 50 per Because some psychotomimetic agents interact with dopa- cent inhibition (I50) of an enzymatic reaction. Biochem Pharmacol mine receptors, we examined the effects of (+)- and (−)-meflo- 22:3099–3108 Combrinck JM, Mabotha TE, Ncokazi KK, Ambele MA, Taylor D, quine in dopamine D1,D2,D3,andD4.4 receptor binding assays Smith PJ, Hoppe HC, Egan TJ (2013) Insights into the role of heme using cells that express the respective recombinant human in the mechanism of action of antimalarials. ACS Chem Biol 8(1): receptors (Table 5). (+)- but not (−)-mefloquine displaced 133–137 Cruikshank SJ, Hopperstad M, Younger M, Connors BW, Spray DC, radioligand binding from the D3 receptor while neither enan- Srinivas M (2004) Potent block of Cx36 and Cx50 gap junction tiomer displaced radioligand binding from the D1,D2,orD4.4 channels by mefloquine. Proc Natl Acad Sci U S A 101(33):12364– receptors. LSD and known dopamine receptor ligands also 12369 bound to D1 and D2-like receptors with affinities that resemble Dow GS, Milner E, Bathurst I, Bhonsle J, Caridha D, Gardner S, Gerena those described in previous reports (Neve and Neve 1997). L, Kozar M, Lanteri C, Mannila A, McCalmont W, Moon J, Read KD, Norval S, Roncal N, Shackleford DM, Sousa J, Steuten J, Because LSD is an agonist at dopamine D2-like receptors, we White KL, Zeng Q, Charman SA (2011) Central nervous system examined the effects of LSD and mefloquine on dopamine D2 exposure of next generation quinoline methanols is reduced relative and D3 receptor-mediated signal transduction (Fig. 5). The to mefloquine after intravenous dosing in mice. Malar J 10:150 Dzierszinski F, Coppin A, Mortuaire M, Dewailly E, Slomianny C, antimalarial had no efficacy at stimulating dopamine D2-and D -receptor-mediated mitogenesis, as compared to LSD and Ameisen JC, DeBels F, Tomavo S (2002) Ligands of the peripheral 3 benzodiazepine receptor are potent inhibitors of Plasmodium quinpirole. It was a weak D3 receptor antagonist with an IC50 falciparum and Toxoplasma gondii in vitro. Antimicrob Agents value of 3.21 μM. Dopamine disposition and function play Chemother 46(10):3197–3207 integral roles in reinforcement behaviors (Glimcher 2011), and Egan C, Grinde E, Dupre A, Roth BL, Hake M, Teitler M, Herrick-Davis so in contrast to other psychotomimetics, mefloquine would not K (2000) Agonist high and low affinity state ratios predict drug intrinsic activity and a revised ternary complex mechanism at sero- be expected to have reinforcing properties. tonin 5-HT2A and 5-HT2C receptors. Synapse 35:144–150 Mefloquine’s behavioral effects differ from those of many Eshleman AJ, Carmolli M, Cumbay M, Martens CR, Neve KA, of the psychotomimetic drugs that were examined here. How- Janowsky A (1999) Characteristics of drug interactions with recom- ever, stimulation of serotonin receptors is shared by all of the binant biogenic amine transporters expressed in the same cell type. J Pharmacol Exp Ther 289(2):877–885 drugs, and is part of the etiology of the behavioral effects that Eshleman AJ, Wolfrum KM, Hatfield MG, Johnson RA, Murphy KV, are shared by these drugs. Thus, profiling of new antimalarial Janowsky A (2013) Substituted methcathinones differ in 2782 Psychopharmacology (2014) 231:2771–2783

transporter and receptor interactions. Biochem Pharmacol Minuzzi L, Cumming P (2010) Agonist binding fraction of dopamine D2/ 85(12):1803–1815 3 receptors in rat brain: a quantitative autoradiographic study. Gatch MB, Forster MJ, Janowsky A, Eshleman AJ (2011) Abuse liability Neurochem Int 56(6-7):747–752 profile of three substituted tryptamines. J Pharmacol Exp Ther Miyake N, Miyamoto S, Jarskog LF (2012) New serotonin/dopamine 338(1):280–289 antagonists for the treatment of schizophrenia: are we making real Gazi L, Bobirnac I, Danzeisen M, Schüpbach E, Langenegger D, progress? Clin Schizophr Relat Psychoses 6(3):122–133 Sommer B, Hoyer D, Tricklebank M, Schoeffter P (1999) Neve KA, Neve RL (1997) Molecular biology of dopamine receptors. In: Neve Receptor density as a factor governing the efficacy of the dopamine KA, Neve RL (eds) The dopamine receptors. Humana Press, Totowa D4 receptor ligands, L-745,870 and U-101958 at human recombi- Nichols DE (2004) Hallucinogens. Pharmacol Ther 101:131–181 nant D4.4 receptors expressed in CHO cells. Br J Pharmacol 128(3): Nichols DE, Frescas S, Marona-Lewicka D, Kurrasch-Orbaugh DM. 613–620 (2002) Lysergamides of isomeric 2,4-dimethylazetidines map the Gillespie RJ, Adams DR, Bebbington D, Benwell K, Cliffe IA, binding orientation of the diethylamide moiety in the potent hallu- DawsonCE,DourishCT,FletcherA,GaurS,GilesPR,Jordan cinogenic agent N,N-diethyllysergamide (LSD). J Med Chem AM, Knight AR, Knutsen LJ, Lawrence A, Lerpiniere J, Misra 45(19):4344–4349 A,PorterRH,PrattRM,ShepherdR,UptonR,WardSE,Weiss Passie T, Halpern JH, Stichtenoth DO, Emrich HM, Hintzen A (2008) SM, Williamson DS (2008) Antagonists of the human adenosine The pharmacology of lysergic acid diethylamide: a review. CNS A2A receptor: Part 1. Discovery and synthesis of thieno[3,2- Neurosci Ther 14:295–314 d]pyrimidine-4-methanone derivatives. Bioorg Med Chem Lett Pham YT, Nosten F, Farinotti R, White NJ, Gimenez F (1999) Cerebral 18(9):2916–2919 uptake of mefloquine enantiomers in fatal cerebral malaria. Int J Clin Glimcher PW (2011) Understanding dopamine and reinforcement learn- Pharmacol Ther 37(1):58–61 ing: the dopamine reward prediction error hypothesis. Proc Natl Rabin RA, Regina M, Doat M, Winter JC (2002) 5-HT2A receptor- Acad Sci U S A 108(Suppl 3):15647–15654 stimulated phosphoinositide hydrolysis in the stimulus effects of Halberstadt AL, Geyer MA (2011) Multiple receptors contribute to the hallucinogens. Pharmacol Biochem Behav 72(1-2):29–37 behavioral effects of indoleamine hallucinogens. Sarihi A, Mirnajafi-Zadeh J, Jiang B, Sohya K, Safari MS, Arami MK, Neuropharmacology 61:364–381 Yanagawa Y, Tsumoto T (2012) Cell type-specific, presynaptic LTP Haynes RK, Cheu KW, Chan HW, Wong HN, Li KY, Tang MM, Chen of inhibitory synapses on fast-spiking GABAergic neurons in the MJ, Guo ZF, Guo ZH, Sinniah K, Witte AB, Coghi P, Monti D mouse visual cortex. J Neurosci 32(28):13189–13199 (2012) Interactions between artemisinins and other antimalarial Schlagenhauf P, Adamcova M, Regep L, Schaerer MT, Bansod S, Rhein drugs in relation to the cofactor model—a unifying proposal for HG. (2011) Use of mefloquine in children- a review of dosage, drug action. Chem Med Chem 7(12):2204–2226 pharmacokinetics and tolerability data. Malar J 10:292–302 Hood JE, Jenkins JW, Milatovic D, Rongzhu L, Aschner M (2010) Schlagenhauf P, Johnson R, Schwartz E, Nothdurft HD, Steffen R (2009) Mefloquine induces oxidative stress and neurodegeneration in pri- Evaluation of mood profiles during malaria chemoprophylaxis: a mary rat cortical neurons. Neurotoxicology 31(5):518–523 randomized, double-blind, four-arm study. J Travel Med 16(1):42–45 Iglesias R, Locovei S, Roque A, Alberto AP, Dahl G, Spray DC, Scemes Simpson JA, Price R, ter Kuile F, Teja-Isavatharm P, Nosten F, E. (2008) P2X7 receptor-Pannexin1 complex: pharmacology and Chongsuphajaisiddhi T, Looareesuwan S, Aarons L, White NJ signaling. Am J Physiol Cell Physiol 295:C752–60 (1999) Population pharmacokinetics of mefloquine in patients with Jones R, Kunsman G, Levine B, Smith M, Stahl C (1994) Mefloquine acute falciparum malaria. Clin Pharmacol Ther 66(5):472–484 distribution in postmortem cases. Forensic Sci Int 68(1):29– Sleight AJ, Stam NJ, Mutel V, Vanderheyden PM (1996) Radiolabelling 32 of the human 5-HT2A receptor with an agonist, a partial agonist and Kanagarajadurai K, Malini M, Bhattacharya A, Panicker MM, an antagonist: effects on apparent agonist affinities. Biochem Sowdhamini R (2009) Molecular modeling and docking studies of Pharmacol 51(1):71–76 human 5-hydroxytryptamine 2A (5-HT2A) receptor for the identi- Thompson AJ, Lummis SC (2008) Antimalarial drugs inhibit human 5- fication of hotspots for ligand binding. Mol Biosyst 5(12):1877– HT3 and GABAA but not GABAC receptors. Br J Pharmacol 153(8): 1888 1686–1696 Kelly JX, Smilkstein MJ, Brun R, Wittlin S, Cooper RA, Lane KD, Thompson AJ, Lochner M, Lummis SC (2007) The antimalarial drugs Janowsky A, Johnson RA, Dodean RA, Winter R, Hinrichs DJ, quinine, chloroquine and mefloquine are antagonists at 5-HT3 re- Riscoe MK (2009) Discovery of dual function acridones as a new ceptors. Br J Pharmacol 151(5):666–677 antimalarial chemotype. Nature 459(7244):270–273 Thomson AM, West DC, Lodge D (1985) An N-methylaspartate receptor- Kennedy K (2009) Army scales back use of anti-malaria drug. Army mediated synapse in rat cerebral cortex: a site of action of ketamine? Times, March 24:2009 Nature 313(6002):479–481 Knight AR, Misra A, Quirk K, Benwell K, Revell D, Kennett G, Toll L, Berzetei-Gurske IP, Polgar WE, Brandt SR, Adapa ID, Rodriguez Bickerdike M (2004) Pharmacological characterisation of the L, Schwartz RW, Haggart D, O'Brien A, White A, Kennedy JM, agonist radioligand binding site of 5HT(2A), 5HT(2B) and Craymer K, Farrington L, Auh JS (1998) Standard binding and 5HT(2C) receptors. Naunyn-Schmeideberg’s Arch Pharmacol functional assays related to medications development division test- 370:114–123 ing for potential cocaine and opiate narcotic treatment medications. MacLean D.S. (2013) Crazy Pills. The New York Times. August 7. NIDA Res Monogr 178:440–466 Available at http://www.nytimes.com/2013/08/08/opinion/crazy- Toovey S (2009) Mefloquine neurotoxicity: a literature review. Travel pills.html?_r=0 Med Infect Dis 7(1):2–6 Marona-Lewicka D, Nichols CD, Nichols DE (2011) An animal model of Van Riemsdijk MM, Sturkenboom MC, Pepplinkhuizen L, Stricker BH schizophrenia based on chronic LSD administration: old idea, new (2005) Mefloquine increases the risk of serious psychiatric events results. Neuropharmacology 61(3):503–512 during travel abroad: a nationwide case-control study in the Milatovic D, Jenkins JW, Hood JE, Yu Y, Rongzhu L, Aschner M (2011) Netherlands. J Clin Psychiatry 66(2):199–204 Mefloquine neurotoxicity is mediated by non-receptor tyrosine ki- Wang Y, Denisova JV, Kang KS, Fontes JD, Zhu BT, Belousov AB. nase. Neurotoxicology 32:578–585 (2010) Neuronal gap junctions are required for NMDA receptor- Psychopharmacology (2014) 231:2771–2783 2783

mediated excitotoxicity: implications in ischemic stroke. J Authorship contributions Neurophysiol 104(6):3551–3556. Weidekamm E, Rüsing G, Caplain H, Sörgel F, Crevoisier C (1998) Lack of bioequivalence of a generic mefloquine tablet with the standard Participated in research design: Janowsky, Smilkstein, Hinrichs, and product. Eur J Clin Pharmacol 54(8):615–619 Riscoe Weiss SM, Benwell K, Cliffe IA, Gillespie RJ, Knight AR, Lerpiniere J, Contributed new reagents or analytical tools: Yang, and Zabriskie Misra A, Pratt RM, Revell D, Upton R, Dourish CT (2003) Performed experiments and conducted data analysis: Johnson, Discovery of nonxanthine adenosine A2A receptor antagonists for Wolfrum, Yang, and Eshleman the treatment of Parkinson's disease. Neurology 61(11 Suppl 6): Wrote or contributed to the writing of the manuscript: Janowsky, S101–S106 Smilkstein, Riscoe, Eshleman, and Zabriskie