Mechanisms of amphetamine action illuminated through optical monitoring of dopamine synaptic vesicles in Drosophila brain

Amphetamines elevate extracellular dopamine, but the underlying mechanisms remain uncertain. Here we show in rodents that acute pharmacological inhibition of the vesicular monoamine transporter (VMAT) blocks amphetamine-induced locomotion and self-administration without impacting cocaine-induced behaviours. To study VMAT's role in mediating amphetamine action in dopamine neurons, we have used novel genetic, pharmacological and optical approaches in Drosophila melanogaster. In an ex vivo whole-brain preparation, fluorescent reporters of vesicular cargo and of vesicular pH reveal that amphetamine redistributes vesicle contents and diminishes the vesicle pH-gradient responsible for dopamine uptake and retention. This amphetamine-induced deacidification requires VMAT function and results from net H+ antiport by VMAT out of the vesicle lumen coupled to inward amphetamine transport. Amphetamine-induced vesicle deacidification also requires functional dopamine transporter (DAT) at the plasma membrane. Thus, we find that at pharmacologically relevant concentrations, amphetamines must be actively transported by DAT and VMAT in tandem to produce psychostimulant effects.

We compared larval crawling velocity in dVMAT null mutants to strains expressing either TH-GAL4 or Tdc2-GAL4 expression drivers (without the accompanying UAS-dVMAT transgene) in the dVMAT null background. Basal locomotion in all three strains fed vehicle (yeast alone) were low compared to wildtype (see Fig. 2a). dVMAT null larvae expressing TH-GAL4 alone exhibited no significant amphetamine-stimulated hyperlocomotion [p>.10; N=22 (vehicle), N=19 (amphetamine); independent sample t-test]. Tdc2-GAL4 expression resulted in a small increase in locomotion in response to amphetamine [t(53)=2.2, p=.032; N=26 (vehicle), N=29 (amphetamine); independent sample t-test]. However, this small amphetamine response matched that in the dVMAT null strain (Δ=0.03 mm/s for both groups), suggesting that the Tdc2-GAL4 did not have a direct effect on the response. Differences in average larval crawling speed between vehicle and amphetamine-fed groups are indicated above each respective genotype. (b) To verify that dVMAT-pHluorin functions as a monoamine transporter, we compared basal crawling velocity in wildtype versus dVMAT null mutants where either dVMAT or dVMAT-pHuorin expression was selectively rescued using the Tdc2-GAL4 expression driver. Rescue with the UAS-dVMAT transgene restored basal locomotion to levels not significantly different from wildtype [p>.05; N=36 (wildtype), N=20 (dVMAT Tdc Rescue); one-way ANOVA]. Rescue with the UAS-dVMAT-pHluorin transgene resulted in less efficient rescue of basal locomotion with a 29% reduction in basal locomotion compared to wildtype larvae [F(2, 91)=29.66, p<.01; N=38 (dVMAT-pHluorin Tdc Rescue); oneway ANOVA]. Every condition represents the mean ± SEM with all experiments conducted on ≥2 separate occasions. To characterize the drug delivery flow rate in our experimental system, we diluted a fluorescent green dye in PBS buffer and determined the rate at which the solution equilibrated using identical experimental parameters to those employed to deliver drugs to fly brains (N=7 separate experiments). Quantitative imaging revealed that dye was first detected in the flow chamber 75.41±1.14 s after start of administration. λ ex =820 nm, λ em =460/50 nm FWHM.   dVMAT-pHluorin fluorescence intensities in presynaptic DA nerve terminals of MB-MV1 were measured following exocytosis due to treatment with 40 mM KCl buffers of differing pH. The pH calibration curve is best fit to a logistic dose response curve (in red, R 2 =0.96) revealing a pKa of 7.5 ± 0.2. Data were normalized to pH 8.4 in each of 4 separate experiments and pooled. (c) Changes in dVMAT-pHluorin fluorescence relative to the initial intensity at pH 7.5 were used to estimate vesicle intraluminal pH. We calibrated intensities to ionophore-containing buffers of known pH in ex vivo brain preparations from flies selectively expressing tetanus toxin light chain in DA nerve terminals to prevent vesicular exocytosis during treatment. The X-intercept of the pH calibration curve revealed the resting pH of the vesicles to be pH 5.8 (95% CI: pH 5.45-6.08). The curve (in red) was best fit with a logistic sigmoidal curve (Hill slope = 1, EC 50 = 6.1; blue lines indicate 95% confidence intervals for the curve fit). Data from 26 brains were used in calculating the curve with N≥3 independent experiments for each pH. For Panels b & c: Points represent means ± SEM. λ ex = 920 nm, λ em = 525/50 nm FWHM.   dVMAT blockade by (+)-CYY477 (1 µM, 10 min pretreatment, 25°C) prevented vesicle alkalization by treatment with 10 μM amphetamine in the same brains as measured by dVMAT-pHluorin brightening [F (2,9)=9.74, p=.006; one-way ANOVA with Bonferroni post-hoc analysis] but did not block subsequent alkalization by 100 μM CQ (p=.008 difference). Treatment with (+)-CYY477 also caused increased acidification as indicated by an overall decrease in dVMAT-pHluorin fluorescence (p=.019). Increases in fluorescence in MB-MV1 regions were normalized to final CQ (100 µM) changes (∆Fmax). Data is represented as mean intensities ± SEM in the MB-MV1 region from 4 separate experiments in all 3 conditions. ΔF/ΔF max Supplementary Figure 10. RP-HPLC chromatogram of (+)-CYY477. The chemical purity of (+)-CYY477 was determined by high-performance liquid chromatography (HPLC) using Phenomenex Gemini C18 column (5μ, 110Å, 250 × 4.6 mm) at 30°C and 60% acetonitrile in ammonium acetate aqueous solution (pH = 4.5) as the mobile phase. Figure 11. Chiral HPLC chromatogram of (+)-CYY477. The enantiomeric excess (ee) of (+)-CYY477 was determined by chiral HPLC using a ChiralPak AD-H column (250 × 4.6 mm) at room temperature and 20% 2-propanol in n-hexane (with 0.2% diethylamine) as the mobile phase.   Figure 13. 13 C-NMR spectra of (+)-CYY477. The 13 C-NMR spectra of (+)-CYY477 [150 MHz, CDCl 3 ; 20 mg (+)-CYY477 in 0.7 mL CDCl 3 containing 0.03% TMS] were recorded on a Bruker AVIII-600 FT-NMR spectrometer. Chemical shifts were expressed in parts per million (ppm) on the δ scale relative to the TMS internal standard.

Supplementary Note 1: Chemical Synthesis and Characterization
Experimental procedures for the Synthesis of (+)-CYY477 Design and screening of (+)-CYY477 Tetrabenazine (TBZ) has been used as an antidyskinetic and antipsychotic medication for decades 1

SUPPLEMENTARY DISCUSSION
We extended our behavioral findings in rodents to more complex behavioral actions of amphetamine, using a drug self-administration procedure in rats. The average response rates conformed to a bell-shaped function of dose for both methamphetamine and amphetamine selfadministration, with comparable maximal response rates (0.086 ± 0.015 and 0.072 ± 0.012 responses/s, respectively; Figs. 1c & d and Supplementary Fig. 2). Rats self-administered cocaine at a maximum of 0.333 ± 0.143 responses/s (Fig. 1e). Thus, methamphetamine, amphetamine and cocaine treatments all reliably promoted self-administration. In contrast, neither (+)-CYY477 nor saline was self-administered at rates above those obtained when responses had no scheduled consequences (Supplementary Figs. 2a & b, compare open circles to points above EXT).
Acute pretreatment of rats with (+)-CYY477 decreased methamphetamine and amphetamine self-administration in a dose-dependent manner (Figs. 1c & d). Remarkably, at the highest (+)-CYY477 dose (0.01 mg/kg), amphetamine or methamphetamine self-administration was completely attenuated with response rates comparable to those obtained when responses had no scheduled consequences (Figs. 1c & d: points above EXT). Cocaine self-administration, in contrast, was unaffected by acute pretreatment with (+)-CYY477 (Fig. 1e). As a further control, we examined the effect of (+)-CYY477 on food reinforcement. (+)-CYY477 was equipotent in decreasing self-administration of methamphetamine and amphetamine and fully inhibited these responses at a dose that did not affect responses for food, suggesting that (+)-CYY477 specifically antagonized amphetamines without disrupting behavior maintained by other reinforcers (Supplementary Fig. 2c).
We also tested the ability of (+)-CYY477 to produce catalepsy and found that the (+)-CYY477 concentrations that blocked amphetamines' behavioral effects in our hyperlocomotion and drug self-administration assays were substantially lower than those required to induce catalepsy ( Supplementary Fig. 1b) -likely through vesicular monoamine depletion 6,7,8 . We also note that different dosages of (+)-CYY477 were needed to block amphetamines' action in mouse locomotion and rat self-administration assays. This is most likely due to different psychostimulant dosing regimens and routes of administration (i.p. versus i.v.) as well as the different species used. Regardless, in both paradigms, acutely pre-administered (+)-CYY477 reproducibly blocked the behavioral actions of amphetamines but not of cocaine, which acts independently of VMAT2.

Rodent Husbandry
Animal care and maintenance were provided by a contract from the National Institute on Drug  Germantown, NY) weighing ~300 g and approximately 8 weeks of age on receipt. Rats were singly housed with water and initially food available at all times. All subjects were acclimated to the animal colony for a minimum of one week before use. After acclimation, body weights of rats were maintained at ~320 g by adjusting daily food rations (Scored Bacon Lover Treats, Bio-Serv, Frenchtown, NJ).

Mouse Catalepsy
Each mouse was removed from its cage and injected with either (+)-CYY477, tetrabenazine, haloperidol, or vehicle and replaced in its cage. 30 min after injection, each mouse was again removed from its respective cage, and gently placed on the laboratory bench with its forepaws on a 0.5 cm diameter bar measuring 11 cm in length and positioned 4 cm above the bench.
Catalepsy was considered present when the forepaws remained on the bar, as that position is not a natural posture for a mouse to assume. The intensity of the cataleptic state was assessed by measuring the duration of that posture, which ended when the mouse first moved its forepaws off of the bar or moved its head in an exploratory manner. If the subject remained immobile for 100 s, it was removed and placed back in its cage. Observations were also made at 60, 90 and 120 min after injection. Treatments were administered to 4 mice no more frequently than once every

Mouse Locomotion Assay
Male Swiss-Webster mice (N=132) were placed singly in clear acrylic chambers ( To assess (+)-CYY477's selectivity of effect for drug self-administration, responses were maintained by different amounts of food in a separate cohort of subjects (N=6) using a procedure similar to the drug self-administration schedule described above. All conditions were identical to those used with drug self-administration except that food pellet presentations replaced the injections. Briefly, experimentally naïve subjects were trained with food reinforcement (20 mg grain-based food pellets, Bio-Serv) to press the right lever under an FR 5-response schedule of reinforcement. After training, the procedure was modified to a 5-component procedure analogous to that used for drug self-administration with different numbers of food pellets (0-4 pellets) delivered for completion of each FR 5 in successive components. Subjects were fed their daily food ration (~35 g of 1 g chocolate-flavored pellet, Bio-Serv) 150 min before sessions so that their response rates approximated those maintained by drug injections. Once performances were stable (as defined above), the effects of pre-session (+)-CYY477 injections (i.p.) were assessed by injecting (+)-CYY477 5-min before selected sessions. At least 72 hours separated tests of (+)-CYY477, which were conducted in a mixed order of doses. Response rates were determined by dividing responses by appropriate elapsed times (excluding time outs and 100 ms for each food presentation).