Biochemical and Behavioral Characterization of NovelMethylphenidate Analogs

M. M. SCHWERI, H. M. DEUTSCH, A.T. MASSEY,1 AND S. G. HOLTZMAN

Division of Basic Medical Sciences, Mercer University School of Medicine, Macon, Georgia (M.M.S., A.T.M.); School of Chemistry andBiochemistry, Georgia Institute of Technology, Atlanta, Georgia (H.M.D.); and Department of Pharmacology, Emory University School ofMedicine, Atlanta, Georgia (S.G.H.)

Received October 29, 2001; accepted January 10, 2002 This article is available online at http://jpet.aspetjournals.org

ABSTRACTAs part of a project to develop treatment agents for cocaineabuse, ()-threo-methylphenidate (TMP) and 11 analogs werecharacterized biochemically and behaviorally to assess theirpotential as anti-cocaine medications. The compounds con-tained aryl and/or nitrogen substitutions, and/or replacement ofthe ester function by an alcohol or ether. All of the analogs,except for the N-methyl-substituted compounds, showed in-creased inhibitory potency against 3H-()-2--carbomethoxy-3--(4-fluorophenyl)tropane 1,5-naphthalenedisulfonate([3H]WIN 35,428) ([3H]WIN) binding to the dopamine trans-porter, compared with TMP. In general, parallel results wereobtained for inhibition of [3H]dopamine ([3H]DA) uptake. Al-though compounds with N-substitutions were proportionallyless potent at blocking DA uptake than WIN binding (comparedwith the unsubstituted compounds), one such compound thatwas 6-fold more potent against [3H]WIN binding than [3H]DA

uptake did not attenuate inhibition by cocaine of synaptosomal[3H]DA transport. The compounds were significantly less potentin displacing [3H]citalopram binding from the serotonin trans-porter. In cocaine discrimination studies in rats, all but two ofthe analogs (both N-substituted) completely generalized withthe cocaine stimulus. Robust positive correlations were ob-served between potency in the drug discrimination assay andactivity at the dopamine transporter, but not the serotonintransporter. When tested for their ability to alter cocaine dis-crimination, four of the analogs (three of which had N-substitutions and shallow dose-response curves as cocainesubstitutes) actually enhanced cocaine discrimination, often atcombined doses of cocaine and test compound that wereinactive when given separately. Taken together, the resultssuggest that TMP analogs may have potential as substitutiontherapies for the treatment of cocaine abuse.

Illicit use of cocaine is a major public health problem world-wide. There is an urgent need for treatment agents that couldbe used to block the pleasurable effects of cocaine and/orreduce craving for the drug, without causing deleterious sideeffects. To this end, our research efforts have been directedtoward the synthesis of compounds that will interact with theDAT, the site in the brain thought to subserve the reinforcingeffects of cocaine (Ritz et al., 1987; Howell and Wilcox, 2001).Although cocaine is known to block norepinephrine and se-rotonin transport, as well as conductance through sodiumchannels (Reith, 1988), its abuse potential is generally attrib-

uted to its inhibition of the reuptake of the neurotransmitterDA at nerve terminals of the mesolimbic system. To increasethe probability that a proposed compound will target theDAT, our approach has been to use psychomotor stimulantagents (in this case, TMP) as the starting point for structuralmodification. TMP was selected for several reasons: 1) it hassomewhat more inhibitory potency at the DAT than at thenorepinephrine transporter or the SERT (Ritz et al., 1987); 2)it has several important chemical and structural propertiesin common with cocaine; and 3) based both on widespreadclinical experience with the drug as an oral treatment agentfor attention deficit disorder, and on several recent studies onits use as replacement therapy for cocaine addicts(Grabowski et al., 1997; Roache et al., 2000), TMP appears tohave low abuse potential and few serious side effects. The

Supported by a grant from the National Institute on Drug Abuse (DA06305)to H.M.D., M.M.S., and S.G.H., and by a Senior Scientist Award (DA00008) toS.G.H.

1Current address: Instructor, Division of Natural Sciences and Mathemat-ics, Macon State College, Macon, GA.

ABBREVIATIONS: DAT, dopamine transporter; DR, discrimination ratio; [3H]CIT, [3H]citalopram; [3H]DA, [3H]dopamine; [3H]WIN, [3H]WIN 35,428,3H-()-2--carbomethoxy-3--(4-fluorophenyl)tropane 1,5-naphthanedisulfonate; SERT, serotonin transporter; 3CTMP, ()-threo-3-chloro-methylphenidate; 3CTMPNMe, ()-threo-N-methyl-3-chloromethylphenidate; 4ATMP, ()-threo-4-aminomethylphenidate; 4FTMP, ()-threo-4-fluoromethylphenidate; 4MeTMP, ()-threo-4-methylphenidate; 4MeTMPNMe, ()-threo-N-methyl-4-methyl-methylphenidate; 3,4CTMP,()-threo-3,4-dichloromethylphenidate; TMP, ()-threo-methylphenidate; TMPNMe, ()-threo-N-methyl-methylphenidate; TMPNBn, ()-threo-N-benzyl-methylphenidate; TROHNBn, ()-threo-N-benzylritalinol; TROMeNBn, ()-threo-N-benzylritalinol methyl ether; ANOVA, analysis ofvariance.

0022-3565/02/3012-527535$7.00THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 301, No. 2Copyright 2002 by The American Society for Pharmacology and Experimental Therapeutics 4691/977159JPET 301:527535, 2002 Printed in U.S.A.

527

by guest on November 20, 2013

jpet.aspetjournals.orgD

ownloaded from

goal of this research is to create novel TMP analogs that willretain their selectivity for the DAT, and either serve asreplacement therapy for cocaine (full or partial cocaine ago-nists) or prevent its effects by directly blocking its binding tothe DAT (cocaine antagonists). Ideally, potential antagonistswould inhibit the binding of cocaine to its recognition site onthe DAT but not interfere with the binding and subsequentuptake of DA. Although this was once thought to be anunattainable goal, a large body of evidence now supports itsfeasibility (see references in Deutsch and Schweri, 1994).

Although the reinforcing effects of cocaine are not attrib-utable to its action on serotonin transport, serotonin canmodulate its subjective effects (Walsh and Cunningham,1997). Serotonergic systems affect stimulant-induced be-haviors in a somewhat complex manner. For instance,specific 5HT uptake blockers enhance cocaine discrimina-tion (Kleven and Koek, 1998, and references therein), butserotonergic agonists attenuate some cocaine-reinforcedbehaviors and self-administration of amphetamine (seereferences in Ritz et al., 1987). Therefore, because alter-ation of the SERT activity of the TMP analogs could influ-ence their utility as cocaine treatment agents, their inhib-itory potency against [3H]CIT binding to the SERT wasalso monitored.

We now report on the testing of TMP and 11 of itsderivatives, which contain amino, methyl, or halide sub-stituents on the aryl ring, methyl or benzyl substitutionson the piperidine nitrogen, and/or modifications of themethyl ester function. This is the first time, to our knowl-edge, that the effect of more radical structural changes(i.e., a disubstituted aryl ring, N-substitutions, and/or re-placement of the ester function by an ether or alcohol) havebeen examined. Previous reports correlating the biochem-ical and behavioral effects of MP analogs have been limitedto derivatives containing only monosubstituted aryl ringsor modification of the ester function (e.g., Patrick et al.,1981; Schweri et al., 1985; Gatley et al., 1996a; Thai et al.,1998). Preliminary characterization of these novel com-pounds was conducted, using both in vitro and in vivotechniques. The inhibitory potencies of the compoundsagainst [3H]WIN binding to the cocaine recognition site onthe DAT, [3H]CIT binding to the SERT, and [3H]DA uptakewere determined, if not assessed previously. Based on theresults, ()-threo-N-methyl-4-methyl-methylphenidate(4MeTMPNMe) was selected as a representative compoundfor further in vitro tests designed to examine the nature ofits interaction with the DAT, as well as its ability toattenuate the inhibition of DA uptake by cocaine. All of theanalogs were tested for cocaine-like discriminative stimu-lus effects in rats trained to discriminate between salineand 10 mg/kg cocaine; selected compounds were furthertested for their ability to block the discriminative effects ofcocaine. The drug discrimination assay is a proven animalmodel of the subjective effects of drugs in humans, withpredictive value for abuse potential (Holtzman, 1990). Pos-sible candidates for further study were identified, and thecombined data were evaluated with the aim of determininguseful characteristics and possible correlations betweenthe in vitro and in vivo results that might guide the designof future pharmacotherapies for cocaine abuse.

Materials and MethodsChemicals

Synthetic Compounds. The syntheses of ()-threo-methylphenidate(TMP), ()-threo-3,4-dichloromethylphenidate (3,4CTMP), ()-threo-4-fluoromethylphenidate (4FTMP), ()-threo-3-chloromethylphenidate(3CTMP), ()-threo-4-aminomethylphenidate (4ATMP), ()-threo-4-methylphenidate (4MeTMP), ()-threo-N-methyl-3-chloromethylpheni-date (3CTMPNMe), ()-threo-N-methyl-4-methyl-methylphenidate(4MeTMPNMe), ()-threo-N-methyl-methylphenidate (TMPNMe), ()-threo-N-benzyl-methylphenidate (TMPNBn), ()-threo-N-benzylritalinolmethyl ether (TROMeNBn), and ()-threo-N-benzylritalinol (TROHNBn)have been described previously (Deutsch et al., 1996; Froimowitz et al.,1997; Deutsch, 1998). 3CTMPNMe and TMPNMe were prepared as freebases; all other synthesized compounds were HCl salts. NMR analysissuggested that 3CTMPNMe contained approximately 10% of the ()-erythro enantiomers; this was likely due to epimerization that occurred atthe N-methylation step.

Other Chemicals. The TMP used in the in vitro assays was a giftof Ciba-Geigy (Basel, Switzerland). ()-Cocaine HCl used in the drugdiscrimination studies was obtained from the National Institute onDrug Abuse (Bethesda, MD); that used in the in vitro assays waspurchased from Sigma Chemical Co. (St. Louis, MO). Citalopramwas a gift of H. Lundbeck A/S (Copenhagen, Denmark). All radioiso-topes were purchased from PerkinElmer Life Sciences (Boston, MA);they are: [ring-2,5,6-3H]dopamine; [N-methyl-3H]citalopram, and[N-methyl-3H]WIN 35,428. All other drugs were purchased fromstandard commercial sources.

In Vitro Studies

Animals. Male Sprague-Dawley rats (Harlan, Indianapolis, IN),weighing 150 to 300 g, were anesthetized using CO2 gas and sacri-ficed by decapitation. Their brains were quickly removed and placedin ice-cold 0.32 M sucrose. Tissue preparations were prepared fromthose areas of the brain as required for the individual assays de-scribed below. This protocol was approved by the Institutional Ani-mal Care and Use Committee of Mercer University and is in accordwith the National Institutes of Health Guide for the Care and Use ofLaboratory Animals.

[3H]WIN Binding to the DAT. These assays were conductedusing rat striatal tissue as previously described (Deutsch et al.,1999). IC50 values (that concentration of drug which inhibits 50% ofspecific binding of the radioligand) and Hill coefficients were deter-mined by least-squares nonlinear regression analysis of sigmoidaldose-response curves using the GraphPad Prism (v.2) program(GraphPad Software, Inc., San Diego, CA).

[3H]CIT Binding to the SERT. The method used here wasmodified from binding assays described elsewhere (DAmato et al.,1987; Dutta et al., 1996). Cortical tissue from all brain areas exceptthe olfactory tubercles and medial cortex extending to approximately1 mm on either side of the interhemispheric fissure was combinedand homogenized in 20 volumes (based on wet weight) of 0.32 Msucrose, using 10 up/down strokes of a motorized Potter-Elvehjemhomogenizer. The homogenized tissue was centrifuged at 1,000g for10 min at 0C, and the resulting supernatant was further centri-fuged at 20,000g for 20 min. The resulting pellet (P2 fraction) wassuspended in 16 volumes of ice-cold assay buffer (25 mM sodiumphosphate buffer, pH 7.7 at 0C) using an Ultra-Turrax tissue ho-mogenizer. Binding was initiated by addition of 150 l of the P2suspension to samples containing 750 l of assay buffer, 50 l of thetest compound, 25 l of water or clomipramine (to define nonspecificbinding; final concentration, 1 M) and 25 l of [3H]CIT (finalconcentration, 2 nM). Test compounds were dissolved in water, di-luted dimethyl sulfoxide, or diluted methanol, with the concentrationof organic solvent in any assay tube limited to 0.3% by volume.Usually, a range of seven drug concentrations was used to generatethe dose-response curve; triplicate samples were run at each concen-

528 Schweri et al.

tration. The samples were incubated for 3 h at 0C. The reaction wasterminated by addition of 5 ml of ice-cold assay buffer to the assaytubes, followed by rapid filtration through glass microfiber filterstrips (presoaked in 0.05% polyethyleneimine) under reduced pres-sure using a Brandel 30-port cell harvester (Brandel Inc., Gaithers-burg, MD). Assay tubes were washed with an additional 5-ml aliquotof ice-cold assay buffer. The filters were transferred to scintillationvials, shaken vigorously for 30 min in the presence of 8 ml of Ready-Safe cocktail (Beckman Coulter, Fullerton, CA), and counted on aBeckman LS6000IC scintillation counter. Data were analyzed usingGraphPad Prism, as described above.

[3H]DA Uptake. The effect of test compounds on the accumula-tion of [3H]DA was determined using a crude synaptosomal prepa-ration of rat striatal tissue that was preincubated for 10 min in thepresence of the test compound, then exposed to 30 nM [3H]DA over a2-min period at 37C, as described previously (Deutsch et al., 1999).In those experiments in which the effect of 4MeTMPNMe on theability of cocaine to inhibit [3H]DA uptake was determined, both4MeTMPNMe and cocaine were present for the entire preincubationperiod. Studies to characterize the effect of 4MeTMPNMe on theMichaelis-Menten kinetics of [3H]DA uptake were performed underthe same conditions, except that the [3H]DA concentrations rangedfrom 25 to 800 nM. Km and Vmax were determined by nonlinearleast-squares regression fit of specific [3H]DA uptake versus free[3H]DA to a rectangular hyperbola function using the GraphPadPrism program (see above). The resulting values were then used togenerate the equations for the straight lines to allow plotting of thedata in the Lineweaver-Burk format.

Drug Discrimination Studies

Subjects. Male rats of Sprague-Dawley descent (Charles RiverLaboratories, Raleigh, NC) were used in the study. They weighed250 to 300 g when discrimination training began and were housed inpairs in a vivarium that had a 12-h light/dark cycle. Food and wateralways were available in the home cage. The vivarium was part of afacility that was accredited by the American Association for Accred-itation of Laboratory Animal Care. The care and testing of theanimals conformed to a protocol that was approved by the Institu-tional Animal Care and Use Committee of Emory University andwere in accord with the National Institutes of Health Guide for theCare and Use of Laboratory Animals.

Drug Discrimination. Rats were trained to discriminate be-tween i.p. injections of 10 mg/kg cocaine and saline in a two-choicediscrete-trial avoidance/escape procedure (Shannon and Holtzman,1976). Injections of cocaine or saline were given 15 min before asession, which was conducted in a standard testing chamber. Thebeginning of a trial was signaled by the onset of white noise andillumination of the house light in the chamber. Five seconds after atrial began, a constant electrical current (1.01.5 mA) was applied tothe grid floor of the chamber for 1.0 sec every 3.0 sec. To end a trial,the rat had to complete a two-response chain: press an observinglever in one wall of the chamber and then press one of two choicelevers mounted in the opposite wall. The choice levers were 15 cmapart and were separated by a 5.0-cm-wide Plexiglas partition thatextended from the floor to the ceiling. A response on the observinglever turned off the white noise and enabled a response on the correctchoice lever to extinguish the house light and end the trial. A trialwas recorded as correct if the rat pressed the choice lever appropriatefor what was injected beforthtee the session (i.e., cocaine or saline)right after it pressed the observing lever. A trial was recorded asincorrect (and did not end) if the rat pressed the choice lever that wasnot appropriate for what was injected before the session betweenpressing the observing lever and pressing the appropriate choicelever. A trial also ended if the rat did not complete either the corrector incorrect sequence of responses within 30 sec and was recorded asan incomplete trial. Trials were separated by a 50-sec period, duringwhich the chamber was illuminated by a red stimulus lamp. A

session consisted of 20 trials and, with fully trained rats, usuallylasted 1921 min.

Approximately half the animals were trained to press the leftchoice lever in training sessions that followed an injection of 10mg/kg cocaine and to press the right choice lever in training sessionsthat followed an injection of saline. The designation of choice leverswas reversed for the rest of the animals. Rats were considered to beunder the stimulus control of cocaine and saline when they com-pleted the response sequence of observing lever-appropriate choicelever in at least 18 trials of 20 in four consecutive training sessionsand two consecutive test sessions, half preceded by cocaine and halfby saline. Test sessions differed from training sessions in one re-spect: after a response on the observing lever, a response on eitherchoice lever ended a trial.

After rats met the criterion for acquisition of the discrimination,they all were tested first with graded doses of cocaine. Doses wereinjected i.p. 15 min before a session in a random sequence that alsoincluded saline. Other drugs were then tested in a nonsystematicorder in groups of four to five rats. They were injected i.p. 30 minbefore a session based upon the time course of action of methylpheni-date; each dose was administered once to each rat in the group. Drugtest sessions in which both choice levers were activated usually wereconducted twice weekly, 3 to 4 days apart. Training sessions in whichonly the pretreatment-appropriate choice lever was activated wereconducted three times each week to maintain stable discriminationperformance. Saline and cocaine were injected on alternate daysbefore a training session. If an animal failed to complete at least 18trials correctly in a training session, testing was suspended until ithad done so again in four consecutive training sessions.

Drugs. In all cases, doses refer to the free base. Cocaine wasdissolved in normal saline solution; 3CTMPNMe, 4MeTMPNMe, andTMPNMe were dissolved in one part dimethyl sulfoxide followed bythree parts distilled water; all other compounds were dissolved indistilled water. Drugs usually were injected in a volume of 1.0 ml/kgof body weight; however, twice this volume was used for some of thehigher doses.

Data Analysis. Data from stimulus-generalization tests are pre-sented as the average number of trials completed on the cocaine-appropriate choice lever in a 20-trial session; the remaining trials ofthe session were completed on the choice lever appropriate for thedrug vehicle. The dose of a drug that resulted in the selection of thecocaine-appropriate choice lever in 10 trials of a session (i.e., ED50)was calculated for individual rats. This was done by linear regressionof the ascending limb of the stimulus-generalization curve, usinglog10 dose and at least three points. In those instances when only twopoints defined the ascending portion of the curve, ED50 values werederived by simple interpolation. The individual ED50 values wereaveraged to obtain a group mean and 95% confidence limits. Slopesof stimulus-generalization curves also were determined by linearregression of the log-dose data. Comparisons among ED50 values andamong slopes were made by analysis of variance, followed, whereappropriate, by the Student-Newman-Keuls test for multiple com-parisons among all means. The alpha level was set at 0.05.

ResultsIn Vitro Studies. The potencies of the compounds as

inhibitors of [3H]WIN binding, [3H]CIT binding, and [3H]DAuptake in vitro are shown in Table 1, expressed as IC50values. The compounds show a hundred-fold variation inpotency against [3H]WIN binding, ranging from IC50 valuesof about 5 nM for both 3CTMP and 3,4CTMP to about 500 nMfor TMPNMe. A 200-fold range of potency was observedagainst [3H]DA uptake, from an IC50 of 7.0 0.6 nM for3,4CTMP to 1435 5 nM for TMPNMe. The most potentcompounds at the SERT are N-benzyl-substituted com-pounds in which the ester function has been converted to an

Methylphenidate Analogs as Treatments for Cocaine Abuse 529

ether (TROMeNBn; IC50, 166 8 nM) or an alcohol(TROHNBn; IC50, 204 9 nM); the least potent is 4ATMP,with an IC50 well over 10,000 nM.

All of the analogs are selective for the [3H]WIN over the[3H]CIT binding site (Table 1). The parent compound, TMP,exhibits well over 120-fold selectivity for the DAT. For themost part, N-substituted derivatives show the least separa-tion in binding affinities at the two transporters (e.g., 4MeT-MPNMe, TROMeNBn, and TROHNBn are only 8- to 9-foldmore potent against [3H]WIN binding than [3H]CIT binding),whereas derivatives with single substitutions on the aro-matic ring exhibit the greatest degree of selectivity (e.g.,3CTMP bound with more than 2000-fold affinity to the DATover the SERT).

Although the potency against [3H]DA uptake parallels thatfound for [3H]WIN binding, the correlation is not perfect(Pearson r 0.9380, p 0.0001). This modest deviation fromperfect linearity results in discrimination ratios (DRs; theratio of the IC50 of a given compound against [

3H]DA uptaketo its IC50 against [

3H]WIN binding) ranging from 1.3 for3,4CTMP to 7.7 for TROHNBn. The DR was calculated as apossible tool for use in the identification of compounds withpotential as partial or full cocaine antagonists. In theory,structural modifications that increase the DR of a compoundshould result in greater antagonist properties, because theyreflect the improved ability of the compound to block thebinding of cocaine to its recognition site, while leaving DAuptake largely unaffected. Interestingly, when the analogscontaining a substituent on the nitrogen were compared withunsubstituted analogs, it was found that the N-substituted

compounds had significantly higher DRs (5.0 0.8 versus3.3 0.5, one-tailed independent t test, p 0.05).

As a compound with one of the higher DRs in this series(6.6), 4MeTMPNMe was selected for further study to deter-mine the nature of its interaction with the DAT, as well as itsability to affect the inhibition by cocaine of DA uptake. In thefirst series of studies, the effects of 300 and 600 nM 4MeT-MPNMe on the Km and Vmax of [

3H]DA uptake were exam-ined. The saturation curves obtained in the presence andabsence of 4MeTMPNMe were analyzed according toMichaelis-Menten kinetics. Data from a representative ex-periment are shown as a Lineweaver-Burk plot in Fig. 1;results from three independent experiments are summarizedin Table 2. Results characteristic of a competitive inhibitorwere obtained; i.e., the apparent Km increased with increas-ing concentrations of 4MeTMPNMe, whereas the apparentVmax values did not differ significantly from the correspond-ing control values.

The ability of 4MeTMPNMe to block the inhibition of[3H]DA uptake by cocaine was examined using the methoddescribed by Simoni et al. (1993). In these experiments, theeffect of 200, 500, and 1000 nM 4MeTMPNMe on the IC50 ofcocaine was determined (Table 3). If cocaine and 4MeTMP-NMe competed for the same domain on the DAT, the IC50 forcocaine would be expected to rise by a predicted theoreticalamount. On the other hand, if cocaine and 4MeTMPNMeacted at two distinct and unrelated sites, the IC50 for cocainein the presence of added 4MeTMPNMe would be expected tochange very little from its value in the absence of 4MeTMP-NMe, whereas if the compounds acted at two overlapping but

TABLE 1Effect of compounds on [3H]WIN and [3H]CIT binding and [3H]DA uptake

Compound

Substitution [3H]WIN Binding[3H]DAUptakeIC50

b

Dis-crimi-nationRatioc

[3H]CIT BindingIC50 [

3H]CIT/IC50 [

3H]WINR1 R2 R3 R4 IC50a nH

a IC50d nH

dInhibition by

10 MCompounde

nM nM nM %

3CTMP H CO2CH3 Cl H 5.1 1.6 0.95 0.12 23.0 3.0 4.5 10,000 45.5 3.7 20003,4CTMP H CO2CH3 Cl Cl 5.3 0.7 2.08 0.05 7.0 0.6 1.3 1,064 62 0.99 0.01 93.3 2.2 213TROMeNBn Bn CH2OCH3 H H 17.8 1.1 1.12 0.13 88 14 5.0 166 8 1.03 0.09 9TROHNBn Bn CH2OH H H 23.7 3.3 1.08 0.06 181 32 7.7 204 9 0.85 0.01 94.1 3.2 94MeTMP H CO2CH3 H CH3 33.0 1.2 1.05 0.02 126 1 3.8 10,000 45.0 0.5 3004ATMP H CO2CH3 H NH2 34.5 4.0 0.96 0.09 114 10 3.3 10,000 7.9 2.6 3004FTMP H CO2CH3 H F 35.0 3.0 0.97 0.02 142 2 4.1 10,000 36.9 0.8 300TMPNBn Bn CO2CH3 H H 52.9 2.3 1.08 0.02 273 28 5.2 1,201 6 1.02 0.04 85.2 1.9 23TMP H CO2CH3 H H 83.0 7.9 0.90 0.09 224 19 2.7 10,000 19.6 2.5 1204MeTMPNMe CH3 CO2CH3 H CH3 140 9 1.02 0.03 920 145 6.6 1,109 40 0.78 0.08 74.2 3.2 83CTMPNMe CH3 CO2CH3 Cl H 160 18 0.96 0.04 441 89 2.7 10,000 33.3 6.1 60TMPNMe CH3 CO2CH3 H H 499 25 1.00 0.01 1435 5 2.9 10,000 25.2 1.2 20()-Cocaine 160 15 1.03 0.01 404 26 2.5 401 27 1.27 0.01 2.5Citalopram 9.2 2.3 1.00 0.07

a Mean S.E.M. from two or more independent experiments in which triplicate samples were run at each of six different concentrations of test compound.b Mean S.E.M. from two or more independent experiments in which duplicate samples were run at each of five different concentrations of test compound.c Discrimination ratio is IC50 for [

3H]DA uptake/IC50 for [3H]WIN binding.

d Mean S.E.M. from two or more independent experiments in which triplicate samples were run at each of seven different concentrations of test compound.e Mean S.E.M. from two or more independent experiments in which the inhibition due to 10 M compound was determined in triplicate.

530 Schweri et al.

nonidentical regions of the transporter, the IC50 for cocainein the presence of a fixed concentration of 4MeTMPNMewould be expected to exceed the theoretical value that wascalculated, assuming a straightforward competition for thesame binding site. The results show that, although the IC50for cocaine increased significantly as the concentration of4MeTMPNMe in the samples rose from 200 to 1000 nM, itnever differed significantly from the theoretical value thatwould be predicted if cocaine and 4MeTMPNMe were directlycompeting for the same site.

Drug Discrimination Studies. A total of 23 rats weretrained to discriminate between 10 mg/kg cocaine and salinein a median of 32 sessions (range: 1390). Doses of cocainefrom 1.0 to 10 mg/kg occasioned orderly increases in thenumber of trials completed on the cocaine-appropriate choicelever (Fig. 2). The group averaged more than 19 trials on thecocaine-appropriate choice lever at the training dose of co-caine and at the highest dose tested, 17.5 mg/kg. TMP alsooccasioned dose-dependent responding on the cocaine-appro-priate choice lever and was three times more potent thancocaine in this regard (Fig. 2; Table 4).

With the exceptions of 3CTMPNMe and TMPNBn, all of

the analogs of TMP occasioned dose-dependent increases inthe number of trials completed on the cocaine-appropriatechoice lever; at the highest dose they generalized completely(i.e., 18 trials to the cocaine-appropriate choice lever) oralmost completely with the training dose of cocaine (Fig. 3).3CTMPNMe, tested over a dose range of 0.1 to 3.0 mg/kg,occasioned completion of a maximum average of 10.6 trials onthe cocaine-appropriate lever at the highest dose, but withconsiderable variability among the five rats: 0, 3, 12, 18, and20 trials. The amount of 3CTMPNMe synthesized was notsufficient to test higher doses. TMPNBn (1.030 mg/kg) oc-casioned the completion of few trials to the cocaine-appropri-ate lever, the group average reaching only 6.0 at 30 mg/kg(Fig. 3), the highest dose available for testing. The results ofANOVA indicated a difference among the slopes of the stim-ulus-generalization curves for the 12 compounds listed inTable 4 that occasioned a group average of at least 10 trials

Fig. 1. Lineweaver-Burk analysis of the effect of 4MeTMP-NMe on [3H]DA uptake by rat striatal synaptosomes. Ac-cumulation of [3H]DA (25800 nM) was measured over a2-min period in an S1 fraction of rat striatal tissue that hadbeen mixed with 25 mM phosphate buffer, then preincu-bated for 10 min at 37C with water (controls), 300 nM4MeTMPNMe, or 600 nM 4MeTMPNMe before the addi-tion of the [3H]DA. In this representative experiment, theVmax values were 134.8 (12.5314.42), 134.2 (11.5215.31),and 152.5 (14.6715.84) pmol/(mg protein 2 min) and theapparent Km values were 180 (146214), 239 (157322),and 333 (304361) nM for the water, 300 nM 4MeTMP-NMe, and 600 nM 4MeTMPNMe treatment groups, respec-tively (95% confidence intervals are shown in parentheses).The results from three such experiments are presented inTable 2.

TABLE 2Effect of 4MeTMPNMe on kinetic parameters of [3H]DA uptakeKm and Vmax values (mean S.E.M.) are from three separate experiments, calcu-lated by nonlinear regression analysis of synaptosomal [3H]DA uptake as a functionof free [3H]DA concentration (see Materials and Methods for details).

[4MeTMPNMe][3H]DA Uptake Parameters

Apparent Kma Vmax

b

nM pmol/mg protein/2 min

0 157 11 153 21300 235 23* 159 29600 326 6** 175 27

a Significant effect of [4MeTMPNMe] on apparent Km by one-way ANOVA withrepeated measures (F2,4 41.04, p 0.002).

b Significant effect of [4MeTMPNMe] on Vmax by one-way ANOVA with repeatedmeasures (F2,4 7.43, p 0.045), but no significant difference between samples withand without 4MeTMPNMe by t test with Bonferroni correction.

* Significantly different from control by t test with Bonferroni correction (p 0.05).

** Significantly different from control by t test with Bonferroni correction (p 0.01).

TABLE 3Effect of 4MeTMPNMe on the inhibition of [3H]DA uptake by cocaine

[4MeTMPNMe]Cocaine IC50 (nM)

Observeda Theoretical (IC50)b,c

nM

0 454 38200 632 73 593 53500 861 7 801 75

1000 1208 58 1149 113a The observed IC50 values for cocaine differed in the presence of varying concen-

trations of 4MeTMPNMe (one-way ANOVA with repeated measures; F3,6 67.1, p0.00001). The observed IC50 for cocaine increased in the presence of increasingconcentrations of 4MeTMPNMe, with the following rank order as a function of[4MeTMPNMe]: 1000 nM 500 nM 200 nM 0 nM, by Scheffe multiplecomparisons.

b IC50 was calculated as described in Simoni et al. (1993) and expressed asmean S.E.M. from three separate experiments. For two inhibitors acting at thesame site, IC50 KCOC[1 [DA]/KDA [4MeTMPNMe/K4MeTMPNMe], where KCOCand K4MeTMPNMe are the equilibrium dissociation constants for cocaine and 4MeT-MPNMe, respectively, calculated from the Cheng-Prusoff equation for competitiveinhibitors (Cheng and Prusoff, 1973), using the IC50 values for cocaine and 4MeT-MPNMe obtained in each individual experiment. (Ki S.E.M. from the three exper-iments were 382 32 nM and 553 21 nM for cocaine and 4MeTMPNMe, respec-tively.) [DA] was 29.5 0.1 nM. KDA is the Michaelis constant for DA uptakedetermined under the same conditions (157 nM; see Table 2).

c No significant difference was found between the observed and theoretical IC50values for cocaine at any concentration of 4MeTMPNMe.

Methylphenidate Analogs as Treatments for Cocaine Abuse 531

to the cocaine-appropriate choice lever (F[11,259] 1.92, p 0.037). However, post hoc tests failed to identify differencesbetween specific compounds.

Comparison of the slopes of the stimulus-generalizationcurves obtained for the N-substituted compounds (8.5 1.4trials/log dose) with the slopes obtained for the compoundscontaining no N-substitution (13.2 1.5 trials/log dose) re-vealed that the average slope for the unsubstituted com-pounds was about 55% higher than the average slope for theN-substituted compounds, although the difference did notquite reach statistical significance (t9 2.196; p 0.056).The slope value for TMPNBn was not included in thesecalculations because a reliable slope could not be calculatedfrom the available data.

3,4CTMP was the most potent of the analogs, 8 times morepotent than TMP; TMPNMe was the least potent of thecompounds for which an ED50 could be calculated, 1/14 aspotent as TMP (Table 4). Thus, relative to the parent com-pound, the analogs spanned a potency range of 107-fold. Theinclusion of TMPNBn increases the potency range to morethan 200-fold. The relative potencies of TMP and the 10analogs for which ED50 values could be determined as co-caine-like discriminative stimuli correlated significantlywith their order of potency for displacing the binding of[3H]WIN and for inhibiting [3H]DA uptake in striatal prep-arations (Fig. 4): Correlation coefficients of 0.61 (p 0.048)and 0.75 (p 0.008) were obtained between the log ED50 forcocaine discrimination and the log of the [3H]WIN bindingIC50 or [

3H]DA uptake IC50, respectively. The correlationbetween the log ED50 for cocaine discrimination and the logof the DR approached significance (Pearson r 0.60; p 0.0534). No correlation was observed between the log of theED50 values for cocaine discrimination and the log of inhib-itory potencies against [3H]CIT binding to the SERT, basedon the four analogs for which [3H]CIT binding IC50 valuescould be calculated (Pearson r 0.534, p 0.46).

On the basis of stimulus-generalization curves, DR values,and chemical structures, four of the TMP analogs were se-lected for testing in combination with cocaine: 4ATMP,4MeTMPNMe, TROHNBn, and TMPNBn. 4ATMP is a morepotent compound than cocaine that required a 100-fold rangeof doses to define the ascending limb of the stimulus-gener-alization curve. 4MeTMPNMe is a less potent compoundthan cocaine that occasioned a low level of drug-appropriateresponding over a 30-fold dose range. The last three com-pounds had the highest DR values of all of the analogsstudied in this series, suggesting that a dose might be foundat which they could inhibit cocaine binding to the transporterbut not block transport of the substrate. The last two com-

Fig. 2. Stimulus-generalization curves for methylphenidate and cocainein rats trained to discriminate between 10 mg/kg cocaine and saline in atwo-choice discrete-trial avoidance/escape procedure. Drugs were admin-istered i.p. either 30 (methylphenidate) or 15 min (cocaine) before asession. Points are means based upon one observation in each of 5(methylphenidate) or 23 (cocaine) rats. They indicate the number of trialscompleted on the cocaine-appropriate choice lever in a 20-trial session;the remaining trials of the session were completed on the choice leverappropriate for saline (Veh). The upper and lower dashed horizontal linesindicate the minimum level of performance maintained in training ses-sions that followed an injection of 10 mg/kg cocaine or saline, respectively.

TABLE 4Potency of methylphenidate and analogs for producing cocaine-likediscriminative effects

CompoundED50 (95% Confidence Limits)

a

RelativePotencyb

mg/kg mol/kg

3,4CTMP 0.38 (0.20.73) 0.43 (0.131.36)c 7.984FTMP 0.26 (0.180.36) 1.03 (0.721.43)c,d 3.333CTMP 0.38 (0.200.73) 1.42 (0.752.73)c,d,e 2.424ATMP 0.39 (0.091.79) 1.57 (0.367.22)c,d,e 2.18TMP 0.80 (0.611.05) 3.43 (2.624.51)d,e,f 1.004MeTMP 1.15 (0.255.33) 4.65 (1.0121.6)e,f 0.743CTMPNMe 2.00* 7.10* 0.48Cocaine 2.58 (2.033.28) 8.45 (6.6610.8) 0.41TROMeNBn 8.28 (2.4627.9) 26.7 (7.9490.0)g 0.13TROHNBn 9.40* 31.6* 0.114MeTMPNMe 9.06 (3.2425.4) 34.7 (12.497.3)g 0.10TMPNMe 11.4 (3.4737.2) 46.1 (14.0151)g 0.07TMPNBn 30 92 0.04

a Dose resulting in selection of the cocaine-appropriate choice lever in an averageof 10 trials per session [n 5 for all drugs except TROHNBn (n 4) and cocaine (n23)].

b Potency relative to methylphenidate, based upon ED50 values (mol/kg).cg ED50 values that do not have a superscript in common are significantly

different from each other, p 0.05.* Estimated from group means and not included in the ANOVA of ED50 values.

Fig. 3. Stimulus generalization curves for methylphenidate and 11 analogsin rats trained to discriminate between 10 mg/kg cocaine and saline. Drugswere administered i.p. 30 min before a session. Points are means based uponone observation in each of five or four (TROHNBn) rats. Solid lines indicatecompounds that have a substitution on the ring nitrogen, and dashed linesindicate compounds that do not. The curve for methylphenidate (TMP) isreproduced from Fig. 2. Other details are the same as in Fig. 2.

532 Schweri et al.

pounds have N-benzyl substitutions, and one, TMPNBn, occa-sioned little cocaine-appropriate responding up to a dose of 30mg/kg. Administered 15 min before cocaine, each of the TMPanalogs tended to shift the cocaine stimulus-generalizationcurve upward and to the left (Fig. 5). This effect is particularlynotable with 3.0 mg/kg TROHNBn (Fig. 5, upper right) and 10mg/kg TMPNBn (Fig. 5, lower right), doses that occasionedlittle or no cocaine-appropriate responding when tested alone.

The ED50 of cocaine was lowered from 3.27 (2.105.08)mg/kg to 1.38 (0.375.14) and 1.43 (0.365.70) mg/kg afterpretreatment with 0.03 or 0.1 mg/kg 4ATMP, respectively. Itwas lowered from 2.60 (1.823.72) mg/kg to 2.25 (0.995.10)and 1.58 (0.435.79) mg/kg, after pretreatment with 1.0 or3.0 mg/kg 4MeTMPNMe, respectively. However, none of

these changes was statistically reliable, according toANOVA. Pretreatment with 3.0 mg/kg TMPNBn lowered thecocaine ED50 nonsignificantly from 2.72 (1.694.36) mg/kg toa group mean of 1.87 mg/kg.

ED50 values could not be calculated for cocaine after pretreat-ment with 10 mg/kg 4MeTMPNMe, 3.0 mg/kg TROHNBn, or 10mg/kg TMPNBn because the threshold for detection of cocainewas reduced so markedly. However, a reduction in ED50 can beinferred from the fact that pretreatment with each of the fourcompounds resulted in a flattening of the stimulus-generaliza-tion curve for cocaine. ANOVA confirmed significant differencesamong the three curves in the 4ATMP (F[2,15] 5.69, p 0.015)and TMPNBn (F[2,15] 14.32; p 0.001) series as well asamong the four curves in the 4MeTMPNMe series (F[3,20]

Fig. 5. Interactions of four analogsof methylphenidate with cocaine inrats discriminating between 10mg/kg cocaine and saline. Cocaine(n 79) was administered i.p. 15min before a session, and each an-alog (n 5) was administered i.p.30 min before a session. Pointsabove Veh represent the effect ofan injection of the indicated dose ofthe methylphenidate analog fol-lowed 15 min later by an injectionof saline, the vehicle for cocaine.Other details are the same as inFig. 2.

Fig. 4. The order of potency of methylphenidate and 10 analogs to produce cocaine-like discriminative effects in rats discriminating between 10 mg/kgcocaine and saline correlates significantly with their order of potency to inhibit [3H]WIN binding to the DAT (left) and [3H]DA uptake (right) in ratstriatal preparations. E, the N-substituted analogs; F, the analogs with no N-substitution. TMPNBn (open square with arrow) was not included in theregression analysis, because it did not generalize sufficiently to the cocaine stimulus to allow a valid calculation of its ED50 (see Table 4). It is shownhere to underscore its unique activity and allow comparison with the other N-substituted compounds.

Methylphenidate Analogs as Treatments for Cocaine Abuse 533

5.84, p 0.005); Students t test did the same for the slopes ofthe curves for cocaine alone and with TROHNBn (t10 5.32;p 0.001).

DiscussionThe methylphenidate analogs described here exhibit many

cocaine-like properties, both in vitro and in vivo, suggestingthat they may ultimately prove useful as substitution ther-apies for the treatment of cocaine addiction. Like cocaine,they potently inhibit both the binding of [3H]WIN to theDAT, as well as the uptake of [3H]DA by striatal synapto-somes. With the exception of 3CTMPNMe and TMPNBn, allof the compounds fully substituted for 10 mg/kg cocaine incocaine discrimination studies. The analogs with and withoutN-substitutions exhibit some unique differences that mayinfluence their usefulness as drug therapy for cocaine abuse.

All of the derivatives retained activity at the DAT in vitro,despite the various structural modifications made to themethylphenidate molecule. Substitutions on the aryl ring, orintroduction of an N-benzyl group (with or without simulta-neous modification of the ester function) increased the affin-ities of the resulting compounds for the WIN binding site,compared with the parent compound. On the other hand,introduction of an N-methyl substitution (with or withoutsimultaneous substitution on the aryl ring) caused up to a6-fold loss in affinity. Hill coefficients of approximately 1were obtained for all of the compounds in the WIN bindingassay, with the notable exception of 3,4CTMP, which had anH of 2, as reported previously (Deutsch et al., 1996). For themost part, parallel results were obtained for the inhibition of[3H]DA uptake by the various derivatives.

In contrast to their activity against [3H]WIN binding to theDAT, the compounds were much less potent as inhibitors of[3H]CIT binding to the SERT. Comparison of the IC50 valuesof the analogs in the two assays showed that the derivativesranged from 8- to more than 2000-fold less potent against[3H]CIT binding than against [3H]WIN binding (Table 1).Cocaine exhibited much less selectivity than the methyl-phenidate derivatives for the DAT: its IC50 against [

3H]WINbinding was only 2.5-fold less than its IC50 against [

3H]CIT.In agreement with previous reports (Ritz et al., 1987; Gatleyet al., 1996b), TMP had little inhibitory activity at the SERT.N-Benzyl modification increased affinity for the [3H]CITbinding site, and simultaneous conversion of the ester func-tion to an alcohol or methyl ether increased it even more(10-fold). In contrast to the effect of ring halogenation on[3H]WIN binding, where 3-Cl and 3,4-Cl modifications pro-duced equivalent increases in potency, 3,4CTMP is muchmore potent than 3CTMP against [3H]CIT binding. Althoughdichloro substitution increased potency at the SERT com-pared with unmodified TMP, it did not raise the nH to 2, as itdid against [3H]WIN binding. Because the [3H]WIN and[3H]CIT binding assay conditions are so similar, this findingsuggests that the nH obtained for this compound in the[3H]WIN assay is valid and not an artifact caused by its highlipophilicity.

In general, the in vitro interactions of these agents withthe DAT are predictive of their activity in the in vivo cocainediscrimination studies (Fig. 4), consistent with the generallyaccepted notion that inhibition of DA uptake is the chiefinternal cue for identification of cocaine-like drugs (Kleven

and Koek, 1998, and references therein; Howell and Wilcox,2001). In accord with this are the positive correlations ob-served between the logs of the [3H]WIN binding IC50, [

3H]DAuptake IC50, or DR, and the log ED50 for cocaine discrimina-tion. However, results from in vitro assays were not perfectpredictors of potency in vivo. Although the low in vivo po-tency of the N-Me-substituted compounds reflects their lowpotency in vitro, this was not the case with the N-Bn-substi-tuted compounds. TROMeBn, for instance, was approxi-mately 5 times more potent than TMP against [3H]WINbinding and two and a half times more potent against [3H]DAuptake than TMP; yet it was almost 8-fold less potent thanTMP as a substitute for cocaine. TMPNBn, the low potency ofwhich in the cocaine discrimination studies precluded a cal-culation of its ED50, was comparable with TMP in the[3H]WIN binding and [3H]DA uptake assays. Five of the sixN-substituted compounds lie above the regression line,whereas all of the compounds without an N-substitution fallon or below the line (Fig. 4). As a group, the N-substitutedcompounds had shallower stimulus-generalization curves,leading to higher ED50 values than would have resulted hadtheir slopes resembled those of the unsubstituted com-pounds. N-substituted analogs are more hydrophobic thantheir secondary amine counterparts. It is possible, therefore,that the shallower dose-response curves of this group areattributable to differences in onset of action caused by phar-macokinetic factors, such as absorption, binding to plasmaproteins, and subsequent transfer across the blood-brain bar-rier. There was no correlation between the potencies of com-pounds to substitute for cocaine and their inhibitory poten-cies against [3H]CIT binding to the SERT.

As a group, the N-substituted analogs also have DR valuesthat are approximately 50% higher than those obtained forcompounds without substitutions at the piperidinyl nitrogen.The DR is an empirical measure utilized in our laboratory asa preliminary screen to identify potential cocaine antago-nists. Because it represents the ratio of the IC50 of a com-pound against [3H]DA uptake to its corresponding IC50against [3H]WIN binding, high values would be expected tobe associated with compounds that are more effective atblocking cocaine binding than DA uptake. For example, acompound that inhibited 80% of [3H]WIN binding sites whileblocking only 10% of [3H]DA uptake would theoretically havea DR value of 36 (Deutsch et al., 1999). The DR is utilizedonly as one means of comparison of compounds assayed un-der rigorously maintained experimental conditions: evensmall variations in the assay conditions can drastically alterthe DR value (Rothman et al., 1993; Xu et al., 1995). Al-though the DR values of the tertiary amines were signifi-cantly higher than those for the unsubstituted analogs, theirmean absolute value (5.0 0.8) nevertheless was quite lowrelative to the value theoretically required for a compound toexhibit predominantly antagonist-like characteristicsagainst cocaine; this suggests that these compounds willhave greatest utility as substitution therapy for cocaine.

In an effort to learn more about the interaction of thesecompounds with cocaine, both in vitro and in vivo studieswere conducted. In vitro, the effect of 4MeTMPNMe on co-caine inhibition of [3H]DA uptake into synaptosomes wasexamined. 4MeTMPNMe was selected because of its rela-tively high DR (6.6) and its shallow dose-response curve inthe cocaine discrimination assay. In the absence of cocaine,

534 Schweri et al.

4MeTMPNMe acted as a classic competitive antagonist of[3H]DA uptake. When tested in combination with cocaine,4MeTMPNMe appeared to act additively with cocaine at thesame or similar site(s) to block the transport of [3H]DA; it didnot prevent access of cocaine to its binding site on the trans-porter in such a fashion as to raise the IC50 of cocaine overand above that predicted for the combination of two inhibi-tors acting at the same site.

When examined in vivo in combination with cocaine, mostof the analogs tested appeared to potentiate cocaine discrim-ination, even at doses that by themselves had little or nococaine-like discriminative stimulus activity. It is unlikelythat these compounds potentiate effects of cocaine in vivo byinterfering with its metabolism, because even the nonesteri-fied analogs (which would not be expected to compete formetabolism) caused this effect. It is conceivable that thepotency of these compounds against serotonin transport un-derlies their potentiation of the discriminative stimulus ef-fects of cocaine. Although it is generally acknowledged thatinhibition of DA transport is the internal cue responsible forthe discrimination of cocaine by rats (Cunningham and Cal-lahan, 1991), serotonin transport inhibitors increase the sen-sitivity for detection of this cue, even though they cannot bythemselves substitute for cocaine (Kleven and Koek, 1998,and references therein). Indeed, TROHNBn, which potenti-ated cocaine discrimination by 8-fold when rats were pre-treated with approximately 0.3 ED50 of it, was one of themost potent analogs against [3H]CIT binding, whereas ap-proximately the same relative dose of 4ATMP, which hasvirtually no activity at the SERT, caused only a 3-foldincrease in cocaine discrimination. On the other hand,synergism with cocaine has also been reported for 1-{2-[bis(4-fluorophenyl-)methoxy]ethyl}-4-(3-phenylpropyl)-piperazine (GBR-12909) and 4-chlorobenztropine, structur-ally diverse dopamine uptake inhibitors with little activity atthe SERT (Tolliver et al., 1999; Holtzman, 2001).

In summary, structural modification of TMP yields com-pounds with high to moderate potency as inhibitors of theDAT, and varying degrees of activity at the SERT, that mayprove useful as substitution therapy for cocaine abuse. Fur-ther systematic studies, several of which are under way inour laboratories, are needed to investigate and exploit theunique (although sometimes conflicting) properties identifiedin this study for the N-substituted analogs. Their shallowdose-response curves may augment their utility as substitu-tion therapy by providing a broader therapeutic dosing rangebefore the threshold for full cocaine-like character is reached.If further chemical modification can generate compoundswith even higher DR values, clinically useful partial ago-nists/antagonists may result. On the other hand, if the left-shift of the cocaine discrimination curve seen with theseagents reflects their ability to potentiate the rewarding ef-fects of cocaine, their appeal would diminish.

Acknowledgments

We thank Monica Stafford and Adam Eason for assistance in thein vitro assays, and Christine Engels for assistance in the drugdiscrimination assays.

ReferencesCheng Y-C and Prusoff WH (1973) Relationship between the inhibition constant (Ki)

and the concentration of inhibitor which causes 50 per cent inhibition (I50) of anenzymatic reaction. Biochem Pharmacol 22:30993108.

Cunningham KA and Callahan PM (1991) Monoamine reuptake inhibitors enhancethe discriminative state induced by cocaine in the rat. Psychopharmacology (Ber-lin) 104:177180.

DAmato RJ, Largent BL, Snowman AM, and Snyder SH (1987) Selective labeling ofserotonin uptake sites in rat brain by [3H]citalopram contrasted to labeling ofmultiple sites by [3H]imipramine. J Pharmacol Exp Ther 242:364371.

Deutsch HM (1998) Structure-activity relationships for methylphenidate analogsand comparisons to cocaine and tropanes. Med Chem Res 8:9199.

Deutsch HM, Collard DM, Zhang L, Burnham KS, Deshpande AK, Holtzman SG,and Schweri MM (1999) Synthesis and pharmacology of site-specific cocaine abusetreatment agents: 2-(aminomethyl)-3-phenylbicyclo[2.2.2]- and -2[2.2.1]alkane do-pamine uptake inhibitors. J Med Chem 42:882895.

Deutsch HM and Schweri MM (1994) Can stimulant binding and dopamine transport bedifferentiated? Studies with GBR 12783 derivatives. Life Sci 55:PL115PL120.

Deutsch HM, Shi Q, Gruszecka-Kowalik E, and Schweri MM (1996) Synthesis andpharmacology of potential cocaine antagonists. 2. Structure-activity relationshipstudies of aromatic ring-substituted methylphenidate analogs. J Med Chem 39:12011209.

Dutta AK, Xu C, and Reith MEA (1996) Structure-activity relationship studies ofnovel 4-[2-[bis(4-fluorophenyl)methoxy]ethyl]-1-(3-phenylpropyl)piperidine ana-logs: synthesis and biological evaluation at the dopamine and serotonin trans-porter sites. J Med Chem 39:749756.

Froimowitz M, Deutsch HM, Shi Q, Wu K-M, Glaser R, Adin I, George C, andSchweri MM (1997) Further evidence for a dopamine reuptake pharmacophore.The effect of N-methylation on threo-methylphenidate and its analogs. Bioorg MedChem Lett 7:12131219.

Gatley SJ, Meehan SM, Chen R, Pan D-F, Schechter MD, and Dewey SL (1996a)Place preference and microdialysis studies with two derivatives of methylpheni-date. Life Sci 58:PL346PL352.

Gatley SJ, Pan DF, Chen RY, Chaturvedi G, and Ding YS (1996b) Affinities ofmethylphenidate derivatives for dopamine, norepinephrine and serotonin trans-porters. Life Sci 58:PL231PL239.

Grabowski J, Roache JD, Schmitz JM, Rhoades H, Creson D, and Korszun A (1997)Replacement medication for cocaine dependence: methylphenidate. J Clin Psycho-pharmacol 17:485488.

Holtzman SG (1990) Discriminative stimulus effects of drugs: relationship to poten-tial for abuse, in Testing and Evaluation of Drugs of Abuse (Adler MW and CowanA, eds), pp 193210, Wiley-Liss, Inc, New York.

Holtzman SG (2001) Differential interaction of GBR 12909, a dopamine uptakeinhibitor, with cocaine and methamphetamine in rats discriminating cocaine.Psychopharmacology 155:180186.

Howell LL and Wilcox KM (2001) The dopamine transporter and cocaine medicationdevelopment: drug self-administration in nonhuman primates. J Pharmacol ExpTher 298:16.

Kleven MS and Koek W (1998) Discriminative stimulus properties of cocaine: enhance-ment by monoamine reuptake blockers. J Pharmacol Exp Ther 284:10151025.

Patrick KS, Kilts CD, and Breese GR (1981) Synthesis and pharmacology of hydroxy-lated metabolites of methylphenidate. J Med Chem 24:12371240.

Reith MEA (1988) Cocaine receptors on monoamine transporters and sodium chan-nels. Natl Inst Drug Abuse Res Monogr 88:2343.

Ritz MC, Lamb RJ, Goldberg SR, and Kuhar MJ (1987) Cocaine receptors ondopamine transporters are related to self-administration of cocaine. Science (WashDC) 237:12191223.

Roache JD, Grabowski J, Schmitz JM, Creson DL, and Rhoades HM (2000) Labora-tory measures of methylphenidate effects in cocaine-dependent patients receivingtreatment. J Clin Psychopharmacol 20:6168.

Rothman RB, Becketts KM, Radesca LR, De Costa BR, Rice KC, Carroll FI, andDersch CM (1993) Studies of the biogenic amine transporters. II. A brief study onthe use of [3H]DA-uptake-inhibition to transporter-binding-inhibition ratios forthe in vitro evaluation of putative cocaine antagonists. Life Sci 53:PL267-PL272.

Schweri MM, Skolnick P, Rafferty MF, Rice KC, Janowsky AJ, and Paul SM (1985)[3H]Threo-()-methylphenidate binding to 3,4-dihydroxyphenylethylamine up-take sites in corpus striatum: correlation with the stimulant properties of ritalinicacid esters. J Neurochem 45:10621070.

Shannon HE and Holtzman SG (1976) Evaluation of the discriminative effects ofmorphine in the rat. J Pharmacol Exp Ther 198:5465.

Simoni D, Stoelwinder J, Kozikowski AP, Johnson KM, Bergmann JS, and Ball RG(1993) Methoxylation of cocaine reduces binding affinity and produces compoundsof differential binding and dopamine uptake inhibitory activity: discovery of aweak cocaine antagonist. J Med Chem 36:39753977.

Thai DL, Sapko MT, Reiter CT, Bierer DE, and Perel JM (1998) Asymmetricsynthesis and pharmacology of methylphenidate and its para-substituted deriva-tives. J Med Chem 41:591601.

Tolliver BK, Newman AH, Katz JL, Ho LB, Fox LM, Hsu K Jr, and Berger SP (1999)Behavioral and neurochemical effects of the dopamine transporter ligand 4-chlo-robenztropine alone and in combination with cocaine in vivo. J Pharmacol ExpTher 289:110122.

Walsh SL and Cunningham KA (1997) Serotonergic mechanisms involved in thediscriminative stimulus, reinforcing and subjective effects of cocaine. Psychophar-macology (Berlin) 130:4158.

Xu C, Coffey LL, and Reith MEA (1995) Translocation of dopamine and binding of2-carbomethoxy-3-(4-fluorophenyl)tropane (WIN 35,428) measured under iden-tical conditions in rat striatal synaptosomal preparations. Inhibition by variousblockers. Biochem Pharmacol 49:339350.

Address correspondence to: Dr. Margaret M. Schweri, Division of BasicMedical Sciences, Mercer University School of Medicine, 1550 College Street,Macon, GA 31207. E-mail: schweri_mm@mercer.edu

Methylphenidate Analogs as Treatments for Cocaine Abuse 535