ORIGINAL INVESTIGATION

MitoPark mice, an animal model of Parkinson’s disease, showenhanced prepulse inhibition of acoustic startle and no lossof gating in response to the adenosine A2A antagonist SCH412348

Steven M. Grauer & Robert Hodgson & Lynn A. Hyde

Received: 23 July 2013 /Accepted: 5 October 2013# Springer-Verlag Berlin Heidelberg 2013

AbstractRationale Psychoses are debilitating side effects associatedwith current dopaminergic treatments for Parkinson's disease(PD). Prepulse inhibition (PPI), in which a non-startlingstimulus reduces startle response to a subsequent startle-eliciting stimulus, is important in filtering out extraneoussensory stimuli. PPI deficits induced by dopamine agonistscan model symptoms of psychosis. Adenosine A2A receptorantagonists, being developed as novel PD treatments,indirectly modulate dopamine signaling in the basal gangliaand may have an improved psychosis profile which could bedetected using the PPI model.Objectives The aims of this study is to characterize PPI inMitoPark mice, which exhibit progressive loss of dopaminesignaling and develop a Parkinson-like motor phenotype, andassess standard and novel PD treatment effects on PPI inMitoPark mice, which more closely mimic the basal gangliadopamine status of PD patients.Results MitoPark mice displayed enhanced PPI as dopaminetone decreased with age, consistent with studies in intact micethat show enhanced PPI in response to dopamine antagonists.Paradoxically, older MitoParks were more sensitive to PPIdisruption when challenged with dopamine agonists such asapomorphine or pramipexole. Alternatively, SCH 412348, an

adenosine A2A antagonist, did not disrupt PPI in MitoParkmice at doses that normalized hypoactivity.Conclusion Use of MitoPark mice in the PPI assay to assessthe potential for PD treatment to produce psychoses likelyrepresents a more disease-relevant model. SCH 412348 doesnot differentially disrupt PPI as do dopamine agonists,perhaps indicative of an improved psychosis profile ofadenosine A2A antagonists, even in PD patients withdecreased dopamine tone in the basal ganglia.

Keywords Prepulse inhibition .MitoParkmice . Parkinson'sdisease . Psychosis . Adenosine A2A receptor antagonist .

Preladenant . SCH 412348

Introduction

Parkinson’s disease (PD) is a neurological disorder in whichdopamine-containing neurons in the substantia nigra region ofthe basal ganglia are progressively lost. The resulting reductionin dopamine signaling in the striatum produces a progressivemovement disorder that is characterized by symptoms oftremor, rigidity, and bradykinesia. While these motorsymptoms are the hallmark of the disease, patients with PDexperience non-motor and neuropsychiatric symptomsincluding sleep disturbances, compulsive and impulsivebehaviors, and autonomic dysfunction (Fernandez 2012).Patients with PD are particularly prone to psychosis, definedas delusions and in particular hallucinations, manifested mostfrequently in the visual domain (Fenelon et al. 2000; Sanchez-Ramos et al. 1996). The current standard of care for PDincludes dopamine agonists and levodopa replacement therapy,which increase dopamine tone in the basal ganglia. These samepsychoses symptoms have long been associated withdopaminergic therapy, and it is likely that pharmacological

Electronic supplementary material The online version of this article(doi:10.1007/s00213-013-3320-5) contains supplementary material,which is available to authorized users.

S. M. Grauer (*) : L. A. HydeNeuroscience Franchise, Merck Research Laboratories,2015 Galloping Hill Road, K-15-C209, Kenilworth, NJ 07033, USAe-mail: steven.grauer@merck.com

R. HodgsonDepartment of In Vivo Pharmacology (Neuroscience),Merck Research Laboratories, 2015 Galloping Hill Road,K-15-C209, Kenilworth, NJ 07033, USA

PsychopharmacologyDOI 10.1007/s00213-013-3320-5

interaction with the underlying disease pathology mayexacerbate them. Beyond the debilitating nature of these sideeffects, their occurrence is associated with an increased risk ofdementia and mortality (Friedman 2010). Novel, non-dopaminergic therapies for PD are currently in development(Poewe et al. 2012), and their psychoses liabilities areunknown. Characterization of novel mechanisms of actionsfor PD pre-clinically may be useful to gauge side effect liability,in particular, propensity for producing psychosis.

Prepulse inhibition (PPI) is a cross-species phenomenon inwhich the prior presentation of a weak, non-startle-elicitingprepulse reduces the startle response to a much stronger,startle-eliciting pulse and is thought to be important in filteringout extraneous or irrelevant stimuli (Braff et al. 1992). Patientswith schizophrenia exhibit reductions in PPI (i.e., respond justas much to a startle-eliciting pulse with the prepulse aswithout) (Swerdlow and Geyer 1998), and these deficitsmay contribute to the positive symptoms of the disease,which include delusions and hallucinations. A variety ofpharmacological manipulations can induce PPI deficits inrodents (Geyer et al. 2001). Since dopamine receptor agonistssuch apomorphine, selective D2 agonists such as quinpirole(Swerdlow et al. 1994), or dopamine releasers such asamphetamine (Swerdlow et al. 1990) can reduce PPI, andexiting PD treatments are designed to increase dopaminesignaling, this model could be useful to evaluate the potentialfor PD treatments to produce psychosis.

Rather than assess this potential to produce PPI deficits inintact animals, the existence of a transgenic mouse model ofPD offered the potential of improved translatability to thedisease state. Recent evidence has suggested a role formitochondrial dysfunction in PD (Bogaerts et al. 2008;Schapira 2008), and to model severe mitochondrialdysfunction in which dopamine neurons are selectivelytargeted, the MitoPark mouse was developed. MitoPark micehave selective deletion of the mitochondrial transcriptionfactor Tfam from midbrain dopamine neurons in thesubstantia nigra (SN), and they display a progressivedevelopment of cardinal motor symptoms characteristic ofPD, which include motor deficits tied to declining dopaminelevels which first manifest themselves in striatum and later incortex (Galter et al. 2010). While the phenotype of MitoParkmice has been characterized with regards to motor function,intraneuronal inclusions, loss of tyrosine hydroxylasestaining suggestive of declining dopamine signaling, andresponsiveness to levodopa (Ekstrand et al. 2007; Erkstrandand Galter 2009), these animals have not been assessed in thePPI model.

The aim of these studies was therefore twofold. The firstwas to perform an initial characterization of baseline PPI inMitoPark mice and compare this response to age-matchedlittermate controls across a range of ages. Second, using PPIdisruption as a simple psychosis-related index, pharmacology

studies were conducted in MitoPark and control mice toprofile compounds including the mixed D1/D2 dopamineagonist apomorphine, the PD standard of care pramipexole(a D2/D3/D4 dopamine receptor agonist), and SCH 412348, anon-dopaminergic adenosine A2A receptor antagonist. SCH412348 is a novel and potent A2A antagonist that demonstratesa high degree of selectivity over other adenosine receptorsubtypes and has shown efficacy in rodent PD modelsincluding reversal of haloperidol-induced catalepsy andpotentiation of L -dopa-induced rotations in 6-OHDA-lesioned rats (Hodgson et al. 2009). Moreover, SCH 412348is a structural analog of preladenant, a drug with positivephase II clinical efficacy. Of particular interest is whether thereis a differential sensitivity to pharmacologic disruption basedupon genotype of the mice, whether this varies dependingupon the age of the animals, and if adenosine A2A antagonistswould be devoid of PPI issues even in MitoPark mice thatmore closely mimic the symptoms and pathology of PD.

Materials and methods

Animals

The breeding scheme for generating MitoPark mice has beendescribed in detail (Schapira 2008; Galter et al. 2010);however, it should be noted that these animals are generatedon a C57BL/6 background in which the DA transporter (DAT)promoter is used to drive cre recombinase expression, andthese animals are mated to mice that contain a loxP-flankedTfam gene. MitoPark mice used in these studies wereheterozygous for DAT-cre expression (DAT/DATcre) andhomozygous for the for the loxP-flanked Tfam gene(Tflam loxP /Tfam loxP ). Age-matched littermate controlsconsisted of mice that were heterozygous for both DAT-creexpression as well as Tfam-LoxP. Mice were bred for Merckat Taconic Farms (Hudson, NY, USA) and allowed aminimum of 1-week habituation in the facility prior to testing,which took place at ages ranging from 6 to 36 weeks. All micewere group housed (n =5 per cage) in an Association forAssessment and Accreditation of Laboratory Animal Care-accredited facility that was maintained on a 12-h light/darkcycle (lights on at 0700 h). Food and water were available adlibitum. Male mice were used exclusively in both PPI andlocomotor activity studies to control for sex-based variability,as female mice, regardless of genotype, tend to showincreased activity levels. Principles of laboratory animal carewere followed, and all studies were previously approved bythe Institutional Animal Care and Use Committee and wereperformed in accordance to the Guide for the Care and Use ofLaboratory Animals as adopted and promulgated by theNational Institutes of Health (Library of Congress ControlNumber 2010940400, revised 2011).

Psychopharmacology

Drugs

Apomorphine and pramipexole were obtained from Sigma(St. Louis, MO, USA) while SCH 412348 (Hodgson et al.2009) was prepared by Merck Research Labs (Rahway, NJ,USA). Apomorphine and pramipexole were dissolved inphosphate buffered saline; SCH 412348 was formulated in0.4 % methylcellulose in distilled water. All drugs wereadministered in a volume of 10 ml/kg by the indicated routeof administration. Dose calculations were based on activemoiety.

Procedures

Prepulse inhibition of startle response (PPI)

Each testing chamber (SR-LAB system; San DiegoInstruments, San Diego, CA, USA) consisted of a Plexiglascylinder (8.8 cm in diameter) mounted on a frame and held inposition by four metal pins attached to a base unit. Movementof the mouse within the cylinder was detected by apiezoelectric accelerometer attached below the frame. Aloudspeaker mounted 24 cm above the cylinder providedbackground white noise, acoustic noise bursts, and acousticprepulses. The entire apparatus was housed in a ventilatedenclosure (39×38×56 cm). Presentations of acoustic pulseand prepulse stimuli were controlled by SR-LAB softwareand interface system, which also digitized, rectified, andrecorded the responses from the accelerometer. Mean startleamplitude was determined by averaging 65 1-ms readingstaken from the beginning of the pulse stimulus onset. Forcalibration purposes, sound levels were measured with asound level meter (RadioShack), scale “A”, with themicrophone placed inside the Plexiglas cylinder.

Following 5 min of acclimation to background noise(65 dB), test sessions consisted of startle trials (Pulse Alone),prepulse trials [4, 8, 12, or 16 dB above background noiselevels (20 ms duration, broad-band noise) + Pulse, 100 msonset-to-onset], and no-stimulus trials. The Pulse Alone trialconsisted of 40 ms, 120 dB (A) pulse of broad-band noise.The no-stimulus trial consisted of background noise only(65 dB). Test sessions began and ended with 10 startle trials,with 60 trials in between, presented in pseudorandom order(12–20 s variable inter-trial interval). These include 10 startletrials, 10 no-stimulus trials, and 10 trials of each of theprepulse + pulse pairings (for an additional 40 trials). PPIwas defined as 100−[(startle amplitude on prepulse trials/startle amplitude on Pulse Alone trials) × 100] %.

Locomotor activity (LMA) in the open field

Spontaneous LMAwas measured in a Tru Scan Photo BeamActivity System (Coulbourn Instruments, Whitehall, PA,

USA) that used a three-plane tracking system with infraredsensors to record an animal's movement. Following thestated pretreatment interval, mice were placed into thearena (26.67 × 26.67 × 39.37 cm) and distance traveled(cm) was recorded in 12 5-min bins for a period of60 min. Data were expressed as a mean ± SEM for eachtreatment group. Testing occurred using normal ambientlighting in the procedure room.

Statistics

Startle responses for assessing startle reactivity to the pulsealone and for calculating PPI were the mean value of the mid-experiment startle levels (i.e., not including the first or last 10startle trials of the test session). While startle habituation wasevident within the initial few startle trials, there were nodifferences once levels equilibrated in any of the studies whencomparing middle and ending trials. PPI data were analyzedby repeated measures analyses of variance (ANOVA), withthe between-subjects factor being genotype and the within-subjects factors being prepulse intensity, and in the case of thepharmacology studies, drug treatment as well. To assessgenotype differences in responses during the no-stimulusand startle (pulse alone) trials, data were analyzed with t tests,but in the case of the pharmacology studies, data wereanalyzed with repeated measures ANOVAs with genotype asthe between-subject factor and treatment as the within-subjects factor. Since there were no genotype, drug, orgenotype × drug effects for the no-stimulus trials, these dataare not displayed. Following any significant interactions, posthoc comparisons were conducted using Fisher's protectedleast significant difference, with the threshold for alpha set at0.05.

In characterizing age-related changes in locomotoractivity (total distance traveled during the test session),data were analyzed with factorial ANOVAs with genotypeand age as between-subjects factors. For the pharmacologystudies, locomotor activity data were analyzed withrepeated measures ANOVA with genotype as the between-subjects factor and treatment as the within-subjects factor.Genotype differences in pharmacokinetic data were assessedwith t tests. Following any significant interactions, posthoc comparisons were conducted using Fisher's protectedleast significant difference, with the threshold for alphaset at 0.05.

For all ANOVA and post hoc analyses, only significanteffects are reported. Thus, if an effect is not reported, itcan be assumed that it was not significant. The only exceptionis for PPI data, where a main effect of prepulse wasalways significant in the ANOVAs with PPI increasing as theprepulse dB increased; thus, for simplification these data arenot discussed.

Psychopharmacology

Results

Recapitulation of the reported locomotor activity deficitin MitoPark mice

Spontaneous open field activity was recorded in five separategroups of male MitoPark mice and their age-matchedlittermate controls. Groups (n =5, each genotype and eachage) consisted of animals that were 6, 8, 12, 14, and 20 weeksold. Similar to what has been reported in the literature (Galteret al. 2010), MitoPark mice showed decreased activity whencompared to their age-matched controls, with the statisticalanalysis revealing significant genotype effect [F (1, 40) =114.79, p <0.0001], age effect [F(4, 40) = 29.68, p <0.0001],and genotype × age [F(4, 40) = 7.00, p <0.001] interaction. Apost hoc analysis indicated that activity deficits in theMitoParkmice became significant by 8 weeks of age, and the magnitudeof these deficits increased as the mice age (Fig. 1) likelyreflecting an age-related reduction in basal ganglia dopaminetone as per the literature.

Cross-sectional assessment of PPI and startle responsein MitoPark mice

Initial assessment of PPI was performed in four different agedcohorts of MitoPark mice and their littermate controls, andthese included test naïve animals that were 7 (cohort A), 10(cohort B), 14 (cohort C), and 20 (cohort D) weeks of age(Fig. 2a). For all but the 20-week-old mice, there was nosignificant genotype or prepulse × genotype interactions. The20-week-old MitoPark animals, however, did displayincreased PPI relative to their littermate controls, with thestatistical analysis showing significant main effect of genotype[F (1, 18) = 26.61, p <0.0001] as well as a significantprepulse × genotype interaction [F (3, 54) = 4.79, p <0.005]with a larger genotype differences at the lower prepulses thanthe higher ones. A post hoc analysis showed that the increase

in PPI was evident at 4, 8, 12, and 16 dB prepulse intensities(Fig. 2b).

Likewise, startle response was unaffected inMitoPark miceages 7, 10, and 14 weeks of age, and yet the oldest 20-week-old MitoParks exhibited increased startle response comparedto littermate controls [t (18)=3.57, p <0.003] (Fig. 2c).

Longitudinal tracking of PPI and startle response in MitoParkcohorts

Given the appearance of increased sensory gating and startleresponse in the test-naïve 20-week-old cohort of MitoParkmice, additional studies were undertaken to try and answerseveral questions that arose from this observation. Would thisphenotype be replicated in other cohorts of animals? If so, atwhat age did this phenotype emerge? And finally, how robustwas this phenotype? Each of the aforementioned cohorts wasthus tracked longitudinally with PPI and startle responsemeasured at various time intervals. Significant increases inbaseline PPI were observed in MitoPark mice that were 15,19, and 23 weeks of age, cohorts B, C, and A, respectively(Fig. 3a). The increase in sensory gating was observed incohort B, 15-week-old MitoPark mice, with significantgenotype [F (1, 18)=5.94, p <0.03] and prepulse × genotype[F (3, 54)=3.95, p <0.02] interactions, and a post hoc testindicating the genotype differences was only at the 4 dBprepulse intensity. As for cohort C 19-week-old mice,statistical analysis revealed a significant main effect ofgenotype [F(1, 18)=5.56, p <0.03]. The increase in PPI incohort A 23-week-old MitoPark mice was modest, but therewas a significant main effect of genotype [F (1, 17)=5.28,p <0.04]. As a point of reference, the test sessions whichimmediately preceded the appearance of the increased PPIphenotype was 19 weeks for cohort A, 12 weeks for cohortB, and 16 weeks for cohort C where no genotype differenceswere observed (data not shown). The increased PPI phenotypepersists with repeated testing once it emerges (SupplementaryFig. S1).

None of the increases in PPI observed in cohorts A–C wasaccompanied by any significant alteration in startle response(Fig. 3b).

Pharmacology studies

Apomorphine differentially disrupts PPI in older MitoParkmice compared to controls

To ascertain whether MitoPark mice were more sensitive tothe disruptive effects of a direct dopamine agonist on PPI andwhether this would be age dependent, a dose of the dopamineagonist apomorphine was selected (1 mg/kg SC, 10 minpretreatment) that produced marginal effects on PPI in control

6 8 12 14 200

2000

4000

6000

8000

10000

12000 MitoPark MiceControl Mice

Age (Weeks)

Dis

tan

ce T

rave

led

(cm

)

* *

* *

Fig. 1 Locomotor activity declines in male MitoPark mice vs. age-matched littermate control as animals age. Data represent distancetraveled over a 60-min period and are expressed as a genotypic mean ±SEM (n =5 animals per genotype per cohort) for each of five independentcohorts. *Significantly different genotype effect (p <0.05)

Psychopharmacology

7 Wk (

Cohort A)

10 W

k (Cohort

B)

14 W

k (Cohort

C)

20 W

k (Cohort

D)

0

25

50

75

100%

PP

I

*

MitoPark

Control Mice

% P

PI

4 8 12 160

25

50

75

100

Control Mice (20 Wk)

MitoPark (20 Wk)

#

dB above Background

## #

Sta

rtle

Un

its

7 Wk (

Cohort A)

10 W

k (Cohort

B)

14 W

k (Cohort

C)

20 W

k (Cohort

D)

0

50

100

150

200

250

Control Mice

MitoPark

*

a b

c

Fig. 2 Baseline sensory gating of acoustic startle as measured by PPI in7-, 10-, 14-, and 20-week-old male MitoPark mice and age-matchedlittermate controls. Data are expressed as the mean percentage of prepulseinhibition for each of the four cohorts averaged over all four prepulseintensities (a), the mean percent prepulse inhibition at each prepulse

intensity for cohort D (b), and the mean basal startle response for all fourcohorts (c ) ± SEM (n =9–10 animals per genotype per cohort).*Significantly different genotype effect (p <0.05), #significantly differentgenotype × prepulse interaction (p<0.05)

15 Wk (Cohort B) 19 Wk (Cohort C) 23 Wk (Cohort A)0

25

50

75

100

% P

PI

* **MitoParkControl Mice

Sta

rtle

Uni

ts

15 Wk (Cohort B) 19 Wk (Cohort C) 23 Wk (Cohort A)0

50

100

150

200

250Control MiceMitoPark

a b

Fig. 3 Increased PPI developed in cohorts B, C, and A when trackedlongitudinally at 15, 19, and 23 weeks of age, respectively. Data representthe mean percentage of prepulse inhibition for each of the three cohorts

expressed as average value collapsed across all four prepulse intensities(a) and basal startle response (b) ± SEM (n=9–10 animals per genotypeper cohort). *Significantly different genotype effect (p <0.05)

Psychopharmacology

mice (data not shown) and profiled for its effects on sensorygating in MitoPark mice and age-matched littermate controlsin a cohort of mice that were 13 weeks of age, in a secondcohort that was 17 weeks of age, and in a third cohort that was24 weeks of age (Fig. 4a). Experiments were designed in sucha way that all animals received all treatment groups, using awithin-subjects counterbalanced approach, with a minimumof 48 h between test sessions, to allow for drug washout.

In the youngest 13-week-old animals (cohort B),apomorphine did not produce deficits in PPI in eithergenotype, with no significant drug effects or drug × genotypeinteractions. Further, as has been previously observed in miceat this younger age, there were also no significant differencesin PPI between MitoPark mice and controls. No significanteffects on startle response were observed in these animalseither. In those animals that were 17 weeks of age (cohortC), once again there was no effect of apomorphine on PPI ineither genotype with no significant drug effects or drug ×genotype interactions. However, as was observed previouslyin MitoPark mice this age and older, regardless of treatment,MitoPark mice showed improved PPI relative to the age-matched controls, as evidenced by a significant genotypeeffect [F(1, 18)=27.57, p <0.0001] and genotype × prepulseinteraction [F(3,54)=2.86, p <0.05]; post hoc test revealedthat although the genotype differences were larger at the lowerprepulses, MitoPark mice showed significantly greater PPI atall prepulses. With respect to startle response, there was asignificant genotype effect [F (1, 18)=6.80, p <0.02] withMitoPark mice, regardless of treatment, showing increasedstartle response, also in agreement with previous results.

Whereas there was no differential sensitivity to thedisruptive effect of apomorphine on PPI in the 13- or17-week-old animals, there was clear evidence of increasedsensitivity in MitoPark mice to apomorphine's effects on PPI

in 24-week-old mice (cohort D), with a significant drug effect[F (1, 18)=30.60, p <0.0001] and drug × genotype interaction[F (1, 18)=26.31, p <0.0001]. A post hoc analysis indicatedthat apomorphine significantly disrupted PPI in the MitoParks(p <0.0001) but not in the age-matched control animals(p > 0.78). There was also a drug × prepulse interaction[F (3, 54)=4.57, p <0.007]; a post hoc test showed thatalthough there was a tendency for the drug effects to be greaterat the lower prepulses, apomorphine significantly reduced PPIat all prepulses. For startle, these animals showed a significantmain effect of drug [F(1, 18)=34.05, p <0.0001] and drug ×genotype interaction [F (1, 18)=11.65, p <0.004] with the posthoc test indicating that vehicle-treated MitoPark showedsignificantly greater levels of startle response than vehicle-treated controls, and that MitoParks treated with apomorphineshowed reduced startle relative to vehicle-treated MitoParks(Fig. 4b). These results were replicated in a drug and testingnaïve cohort of mice (Supplementary Fig. S2).

Pramipexole differentially disrupts PPI in agedMitoPark mice

The effect of the D2/D3 agonist pramipexole on PPI wasinitially investigated in control mice, and from this study adose of 1 mg/kg SC (10 min pretreatment) was selected, as itproduced only marginal levels of disruption (data not shown).Subsequent to this, this dose of pramipexole was profiled inMitoPark mice and age-matched controls that were 12 weeksof age, then in a second cohort that was 17 weeks of age, andin an older cohort that was 26 weeks old (Fig. 5a). Similar towhat was observed with apomorphine in younger mice, inboth the 12-week-old (cohort A) and 17-week-old (cohort B)mice, pramipexole was without effect with no significant drugor drug × genotype interactions. In addition, there were alsono significant differences in PPI between MitoPark mice and

% P

PI

13 Wk (Cohort B) 17 Wk (Cohort C) 24 Wk (Cohort D)0

20

40

60

80

100a b

Control- VehControl- ApoMitoPark- VehMitoPark- Apo

# *

Sta

rtle

Un

its

13 Wk (Cohort B) 17 Wk (Cohort C) 24 Wk (Cohort D)0

50

100

150

200

250 Control- VehControl- ApoMitoPark- VehMitoPark- Apo #

*

Fig. 4 Apomorphine differentially disrupts PPI in older MitoPark micecompared to age-matched control mice. Effect of apomorphine (1 mg/kgSC, 10 min pretreatment) on mean percentage of prepulse inhibition in13-, 17-, and 24-week-old MitoPark and control mice expressedat each of four prepulse intensities (a ) and basal startle response

(b ) ± SEM (n =9–10 animals per genotype per cohort, using awithin-subjects counterbalance design with a minimum 48-hwashout period between test sessions). *Significantly differentdrug × genotype interaction (p <0.05), #significantly differentgenotype effect (p <0.05)

Psychopharmacology

controls at these two ages. No significant effects on startleresponse were observed in the 12-week-old cohort, but therewas a significant drug effect [F(1, 18)=15.99, p <0.001] inthe 17-week-old cohort with pramipexole causing a small butsignificant decrease in startle response.

However, in the 26-week-old animals (cohort D),pramipexole differentially disrupted PPI in the MitoPark micewith a significant drug effect [F(1, 18)=13.61, p <0.002] anddrug × genotype interaction [F(1, 18)=24.35, p <0.0001] and apost hoc analysis indicating that the control animals did notdiffer as a result of drug treatment (p =0.39), but that PPI wassignificantly disrupted with pramipexole treatment in theMitoParks (p<0.0001). Startle response was also significantlyimpacted in this cohort, with a drug effect [F (1, 18)=8.83,p<0.009] indicating that pramipexole reduced startle responsesindependent of genotype. These results were replicated in a drugand testing naïve cohort of mice (Supplementary Fig. S3).

SCH 412348 does not differently disrupt PPI in agedMitoPark mice

Initial studies with the adenosine A2A antagonist SCH 412348in control mice (data not shown) did not identify a dose thatproduced deficits in PPI.We therefore selected a dose to profilein the MitoParks based upon an efficacy measure: reversal oflocomotor activity decreases that are observed in olderMitoPark mice (see below). The same dose of SCH 412348was then evaluated in the PPI assay (Fig. 6). In two separateaged cohorts (cohort B—28 weeks old; cohort G—34 weeksold), SCH 412348 (0.3 mg/kg PO, 60 min pretreatment)treatment did not produce significant drug effects [F(1, 16)=2.38, p >0.14] and [F(1, 16) =2.74, p >0.16], respectively, ordrug × genotype interactions [F(1, 16)=0.87, p >0.37] and[F(1, 16)=2.30, p >0.15], respectively. Further, both studiesshowed a preservation of the increased gating seen in older

% P

PI

12 Wk (Cohort A) 17 Wk (Cohort B) 26 Wk (Cohort D)0

20

40

60

80

100a b

Control- VehControl- PramMitoPark- VehMitoPark- Pram

*

Sta

rtle

Un

its

12 Wk (Cohort A) 17 Wk (Cohort B) 26 Wk (Cohort D)0

50

100

150

200

250 Control- VehControl- PramMitoPark- VehMitoPark- Pram

#

#

Fig. 5 Pramipexole differentially disrupts PPI in older MitoPark micecompared to age-matched control mice. Effect of pramipexole (1 mg/kgSC, 10 min pretreatment) on mean percentage of prepulseinhibition in 12-, 17-, and 26-week-old MitoPark and control miceexpressed at each of four prepulse intensities (a ) and basal startle

response (b ) ± SEM (n =9–10 animals per genotype per cohort,using a within-subjects counterbalanced design with a minimum48-h washout period between test sessions). *Significantly differentgenotype × drug interaction (p <0.05), #significantly different drugeffect (p <0.05)

% P

PI

28 Wk (Cohort B) 34 Wk (Cohort G)0

20

40

60

80

100a bControl- Veh

Control- '348MitoPark- VehMitoPark- '348

#

Sta

rtle

Un

its

28 Wk (Cohort B) 34 Wk (Cohort G)0

50

100

150

200

250Control- VehControl- '348MitoPark- VehMitoPark- '348

#

Fig. 6 SCH 412348 does not disrupt PPI in older MitoPark mice. Effectof SCH 412348 (0.3mg/kg PO, 60min pretreatment) onmean percentageof prepulse inhibition in 28- or 34-week-old MitoPark miceexpressed at each of four prepulse intensities (a ) and basal startle

response (b ) ± SEM (n =10 animals per genotype per cohort, usinga within-subjects counterbalance design with a minimum 48-hwashout period between test sessions). #significantly differentgenotype effect (p <0.05)

Psychopharmacology

MitoPark relative to control animals, regardless of treatmentconditions or prepulses, with significant genotype effects seenin both cohorts [F(1, 16)=5.76, p < 0.03 and F(1, 18)=20.66,p <0.0003]. Also, in the 28-week-old mice (cohort B), therewas a genotype × prepulse interaction [F (3, 48)=3.00,p <0.04] with the MitoPark mice tending to have greater PPIat the lower prepulses than the higher ones and a drug ×genotype × prepulse interaction [F (3, 48)=3.37, p <0.03]with no clear interpretation of this complex interaction beyondthe individual interactions already shown or described. Startleresponse was not significantly affected by drug treatment, butthere was a significant genotype effect in Cohort G 34 weeksold [F (1, 18)=4.75, p <0.05] indicating greater startleresponse in the MitoPark mice relative to control mice,regardless of treatment.

Effects of apomorphine, pramipexole, and SCH 412348to normalize locomotor activity in aged MitoPark miceto support that doses used in PPI studies are biologicallyrelevant and within the efficacious range

Utilizing the same doses that had been selected for the PPIpharmacology studies, the effects of apomorphine (1 mg/kgSC, 10 min pretreatment), pramipexole (1 mg/kg SC, 10 minpretreatment), and SCH 412348 (0.3 mg/kg PO, 60 min

pretreatment) were profiled in MitoPark mice and age-matched littermate controls (25–27 weeks of age) using thelocomotor assay. These studies were conducted using awithin-subjects counterbalanced design and a 48-h washoutbetween testing sessions. Using distance traveled in a 60-minsession as the functional endpoint, vehicle-treated MitoParkmice showed the expected (because of their age) hypoactivityrelative to vehicle-treated controls in all three experiments.

As expected, regardless of treatment, there was a maineffect of genotype on locomotor activity with reduced activityin MitoPark mice [F (1, 18)=8.42, 21.23, and 8.58,respectively, p <0.01]. The effects of the test compounds werequite different depending on genotype (Fig. 7). Apomorphinefailed to reverse the hypoactivity observed in MitoPark mice,with no significant drug effects or drug × genotype interactionssuggesting that this dose was below that required for efficacy.The effects of pramipexole were paradoxical, with a significantdrug × genotype interaction [F(1, 18)=45.00, p <0.0001], andthe post hoc analysis showing a significant suppression ofactivity in control animals (p <0.0001) that was not due to atransition from activity to stereotypies, and a significantincrease in MitoParks (p <0.004), which exceeded theobserved activity levels in vehicle-treated controls; thissuggested that the dose of pramipexole was biologicallyrelevant and within the efficacy range required for improving

Control-Veh Control-Pram MP-Veh MP-Pram0

2000

4000

6000

8000

10000

12000

14000 27 Wk: Cohort A

*

*

Dis

tan

ce T

rave

led

(cm

)

Control-Veh Control-Apo MP-Veh MP-Apo 0

2000

4000

6000

8000

10000

12000

14000 25 Wk: Cohort E

Dis

tan

ce T

rave

led

(cm

)

Control-Veh Control-'348 MP-Veh MP-'3480

2000

4000

6000

8000

10000

12000

14000 26 Wk: Cohort A

Dis

tan

ce T

rave

led

(cm

)

*

Fig. 7 Effect of apomorphine (1 mg/kg SC, 10 min pretreatment), SCH412348 (0.3 mg/kg PO, 60 min pretreatment), or pramipexole (1 mg/kgSC, 10min pretreatment) to reverse hypoactivity in 25–27-week-old maleMitoParkmice. Data represent distance traveled over a 60-min period and

are expressed as a genotypic mean ± SEM (n =10 animals per genotypeper cohort, using a within-subjects counterbalance design with aminimum 48-h washout period between test sessions). *Significantlydifferent drug × genotype interaction (p<0.05)

Psychopharmacology

locomotor activity in MitoPark mice. SCH 412348 hadminimal effect on activity in control animals, but normalizedthe locomotor deficit in the MitoPark mice, with a significantdrug effect [F(1, 18)=21.32, p <0.0003] and drug × genotypeinteraction [F(1, 18)=7.95, p <0.02] with the post hoc analysisshowing no difference between the drug conditions in controlanimals (p =0.22), but a significant effect in theMitoParkmice(p <0.0001) suggesting this dose was within the efficacy rangefor MitoPark mice.

Discussion

Dopaminergic drugs such as levodopa and dopamine agonistshave long been viewed as the preferred treatment agents inpatients with early PD (Hristova and Koller 2000). Althoughthese drugs do not show clinical evidence of slowing diseaseprogression, they are effective in reducing the symptoms ofthis progressive movement disorder. Whereas psychosis hasbeen recognized as a potential consequence of PD even beforethe introduction of levodopa and dopamine agonists as atreatment agent, the widespread use of dopaminergic drugshas led to an increase in the reporting of this complication(Fenelon et al. 2006; Baker et al. 2009). Given the fact thatdopamine agonists have long been recognized for their abilityto produce psychosis in humans (Seeman 1987), theprevalence of such treatment liabilities is not a surprise.Symptoms are most frequently manifested in the visualdomain and range from minor non-disturbing illusions inwhich there are altered interpretations of real physical stimulito more debilitating occurrences of paranoid delusions anddisturbing visual hallucinations (Thanvi et al. 2005).Hallucinations are indeed common in PD and a recent meta-analysis of prospective studies estimates that up to 37 % ofpatients are affected by these complications (Fenelon andAlves 2010) with over 50 % of patients affected in late stagedisease (Williams and Lees 2005). Beyond the impact on thepatient, the prevalence of psychosis in PD has implications forthe course of clinical care since the appearance of thesesymptoms significantly impacts caregivers and increasedcaregiver stress leads to a greater likelihood of patients beingplaced in nursing homes (Friedman 2010). In addition,psychosis has been identified as a risk factor for thedevelopment of PD-associated dementia (Factor et al. 2003)as well as an increased rate of mortality (Forsaa et al. 2010).

It is believed that deficits in PPI represent an inability tofilter out unnecessary information, and these deficits may leadto cognitive fragmentation that underlies the psychosis seen indisorders such as schizophrenia and perhaps psychosis in PDas well. The pharmacology of rodent PPI deficits is wellcharacterized (Geyer et al. 2001) with dopamine agonistsdisrupting PPI, and antipsychotic medications that blockdopamine signaling reversing these deficits in rodents, as well

as improving PPI in schizophrenic patients (Swerdlow et al.2006). Progressive loss of nigral dopaminergic presynapticneurons can reduce striatal dopamine levels by 70–80% in PD(Dauer and Przedborski 2003), such that the use of intactanimals may not accurately model the disease state.

Since PD-associated psychosis likely stems from a complexinteraction of disease-related presynaptic dopaminergicde-afferentiation and chronic receptor stimulation bydopaminergic medications (Fenelon 2008), the use of animalswith intact dopamine systems to model side effect liabilitylikely fails to capture this interplay. As preclinical efficacymodels are considered most informative when they have somedegree of face, construct, or predictive validity, so shouldmodels of side effect liability. MitoPark mice, which exhibita progressive loss of dopamine cell bodies in the SN and anemergence of a corresponding Parkinsonian-like phenotype,should have utility in evaluating both efficacy and side-effectliability of both existing and novel treatments for PD.Specifically, MitoPark mice have selective deletion ofmitochondrial transcription factor A from midbrain dopamineneurons of the SNwhich leads to subsequent respiratory chainfailure and loss of dopamine cell bodies inferred from the lossof tyrosine hydroxylase (TH) staining (Ekstrand and Galter2009). MitoPark mice recapitulate many of the PD-associatedphenotypes at the histological, neurochemical, and behaviorallevel (Galter et al. 2010; Ekstrand et al. 2007; Ekstrand andGalter 2009), and are able to predict efficacy of variousdopaminergic therapeutics for PD. Here, we replicated theage-related progressive decline in motor activity that has beencharacterized in the literature suggesting that our MitoParkmice are similar to those reported elsewhere and likely havesimilar age-related progressive reductions in striataldopaminergic tone. While the progressive motor phenotypein MitoPark mice is well characterized, there are no publishedreports as to whether these animals exhibit any alterations inbaseline PPI response.

Our observations, in multiple cohorts of MitoPark micetracked longitudinally and assessed cross-sectionally,demonstrate a potentiation of PPI in MitoPark mice, with thisphenomenon emerging at 15–20 weeks of age. Thepotentiation of sensory gating in these animals as dopaminesignaling is lost is consistent with pharmacological studiesthat show antagonism of DA signaling can increase PPI inintact animals (Flood et al. 2011). Potentiation of PPI inMitoPark mice occurred several weeks after the developmentof the motor deficits, which manifests earlier at about 8–10 weeks of age. This delayed emergence may have anexplanation in the anatomical progression of dopamine cellloss. Dopamine signaling is mediated by two separatepathways: the mesolimbic system, which is comprised ofprojections from cell bodies in the ventral tegmental area(VTA) and their projections to ventral areas of the striatum,namely the nucleus accumbens (NA); and the nigrostriatal

Psychopharmacology

pathway, which projects from SN pars compacta to thecaudate putamen region of the striatum, which is more dorsal.Nigrostriatal projections, specifically targeted in the MitoParkmice, are of greater relevance to PD, while mesolimbicprojections play an important role in PPI. Reduction of DAinnervation in the caudate putamen detected with tyrosinehydroxylase immunohistochemistry is observed in theMitoPark mice from as early as 8 weeks in age. This decreasein TH staining of striatal DA fiber progresses over time in aventromedial direction, and eventually DA innervation isgreatly reduced in the mesolimbic pathway (e.g., NA) as well(Ekstrand and Galter 2009). Thus, it may not be unexpectedthat the PPI phenotype would emerge a few weeks after thefirst appearance of the hypoactivity phenotype. More in-depthimmunohistochemistry studies would be required to supportthis conclusion.

Given the observation that PPI increases in MitoPark miceas dopamine tone decreases, an interesting question arises: doParkinson’s patients show similar enhancements in PPI? Inaddressing this question, it is important to state that there areseveral key differences in how PPI is typically measured inpatients versus rodents. Whereas whole-body startle responsewas recorded in the above described rodent studies, the startleresponse in patients is usually assessed by measuring eyeblink reflex response either in response to supraorbital nervestimulation of a puff of air to the cornea, although it can alsobe elicited by sound or photic stimulation. While our studiesutilized an auditory prepulse, this may be replaced bysomatosensory stimuli for human studies. Reports ofabnormalities of auditory prepulse inhibition were morefrequent than those of somatosensory prepulse inhibition inPD (Valls-Sole et al. 2004), with deficits reported (Nakashimaet al. 1993). However, these abnormalities may be related tothe disturbance of the function of the reticular system, whichprojects from the substantia nigra and the striatum, and whichmodulates the blink reflex (Delwaide et al. 1990). Thegenerally accepted view of prepulse modification of startleresponse is that a descending system determines prepulseeffectiveness and the modification by the prepulse is relativelyindependent of the intrinsic reflex circuit characteristics.However, in both 6-OH-DA unilaterally lesioned rats, andpatients with Parkinson’s disease, a hyper-excitable trigeminalreflex blink is observed. Pairing an acoustic prepulse with thisaltered reflex response facilitated this reflex, resulting indeficits in PPI, but the opposite was observed in the respectivecontrol populations (Schicatano et al. 2000). This suggeststhat under pathological conditions, the reflex modification isbeing impacted by the intrinsic characteristics of the reflexcircuit itself, rather than by external modification via theprepulse.

Following the initial characterization of baseline PPI in theMitoPark mice, we next assessed the potential of twodopamine agonists, apomorphine and pramipexole, to produce

neuropsychiatric side effects, using disruption of PPI as asurrogate marker. Apomorphine, a mixed D1/D2 agonist, hasshown efficacy in treating the motoric symptoms of PD, butadverse events associated with its use include nightmares,visual hallucinations, and mild confusion (Hughes et al.1993). Although efficacious at improving PD symptoms inpatients, pramipexole, a partial or full agonist at D2, D3, andD4 receptors (Newman-Tacredi et al. 2002), has also beenshown to have an elevated risk of adverse events includinghallucinations (Etminan et al. 2003). For both of thesecompounds, preliminary studies were conducted in age-matched littermate controls to select a dose that hadonly marginal effects on PPI. Using a within-subjectscounterbalanced design, these doses were then probed inMitoPark and age-matched littermate controls that werebelow the age in which baseline PPI potentiation wasobserved as well as in older cohorts that would presumablyexhibit enhanced gating due to the underlying disease-relatedpathology. Neither apomorphine (1 mg/kg SC) norpramipexole (1 mg/kg SC) disrupted PPI in the youngercontrol or MitoPark mice, whereas both compounds producedsignificant gating deficits in older MitoPark mice, but not inthe age-matched control animals. Additional analysisdemonstrated that apomorphine significantly disrupted PPIin the older mice relative to the age-matched littermate cohortsthat received vehicle, whereas pramipexole did not, whichmay imply that apomorphine is truly disrupting sensorygating, whereas pramipexole is merely reducing the enhancedPPI levels observed in the MitoPark mice back to normallevels. This increased sensitivity to the disrupting effects ofdopamine agonists in older (15–20 weeks of age) MitoParkmice, whom paradoxically exhibited enhanced baseline PPI atthis age, could be the result of a compensatory up-regulationof postsynaptic dopamine receptors in response to decreasedpresynaptic dopamine tone. Although this has not beendemonstrated in MitoPark mice, PET studies in Parkinsonpatients have shown an upregulation of dopamine D2

receptors in drug-naïve patients with Parkin gene mutations(Scherfler et al. 2006), upregulation of putaminal D2 receptorsin early PD (Kaasinen et al. 2000), as well as increased D2

receptor binding after long-term treatment with antipsychotics(Silvestri et al. 2000), suggesting that this may be the case inMitoPark mice as well and thus providing a possibleexplanation of our pharmacological results.

Apomorphine or pramipexole were also assessed for theirability to reverse MitoPark hypoactivity to establish that thedoses used in the PPI studies were relevant with regard to aPD-related efficacy endpoint. The dose of apomorphine thatdisrupted PPI in our studies did not reverse the hypoactivity inolder MitoPark mice (25–27 weeks old), indicating itspotential for side effects may exceed its therapeutic utility.Pramipexole had an interesting profile, with a significantreduction in locomotor activity in control animals that was

Psychopharmacology

not the result of a heightened level of stereotypic behavior, butan increase in locomotor activity in MitoPark mice in excessof the level observed in vehicle-treated intact controls. Asimilar differential change in exploratory locomotor activityhas been reported in both intact mice (Chang et al. 2010) andrats (Chang et al. 2011), with an early locomotor suppression,believed to be D3 mediated, followed by a later D2-drivenhyperactivity, with higher doses producing shifts towardhyperactivity that occur earlier in time. If there is an alteredstriatal dopamine D2 receptor expression in theMitoPark miceas a result of decreased striatal dopamine input, than it wouldbe conceivable that these animals would respond to apramipexole challenge with a marked increase in activity, asany D2 influence would predominate. Further studies wouldbe required to explore this hypothesis. Conversely, the dose ofpramipexole used in the pharmacology studies detailed heremay be on the high end since the PPI-disrupting dose wentbeyond normalization of locomotor activity and producedhyperactivity in the MitoPark mice. Additionally,pharmacokinetic data (Supplementary Table S1) showed thatbrain levels of pramipexole were higher in the MitoPark micethan in control animals, even though plasma levels were notdifferent; this may be contributing to the greater sensitivity ofMitoPark mice to pramipexole. The reason for this genotype-related disparity is not immediately understood, althoughaltered D2 receptor density and/or binding affinity, both inthe striatum and in extrastriatal regions, may play a role(Rinne et al. 2004).

Adenosine A2A receptor antagonists have been the focus ofa concerted effort to develop non-dopaminergic treatments forPD given the co-localization of A2A and D2 receptors instriatopallidal indirect pathway of the basal ganglia, and theobservation that A2A and D2 receptors have an antagonisticreceptor interaction, such that antagonists at the A2A receptorwill decrease intracellular adenylyl cyclase activity to lowercAMP, mimicking the effect of D2 receptor agonists (Ferreet al. 1997; Schwarzchild et al. 2006). Numerous preclinicalstudies have demonstrated that A2A antagonists have efficacyin both rodent and non-human primate models of PD (Hickeyand Stacy 2012), and at least 25 clinical trials have beeninitiated over the last 8 years (Chen et al. 2013). Giventhe more localized and neuro-modulatory role of A2A

receptors, targeting these may provide for a more benignneuropsychiatric side effect profile, restoring the properbalance between A2A and D2 signaling in the indirect pathwayof the GP. Recent publications have shown that deletion ofstriatal A2A receptors in mice do not lead to deficits in PPI(Singer et al. 2013) and that A2A receptor antagonists do notdisrupt PPI in intact rats and mice, demonstrating an improvedPPI-related side effect profile versus dopamine agonists(Bleickardt et al. 2012).

In the experiments described here, SCH 412348 did notshow any evidence of disrupting PPI in either the control

animals, nor in the potentially more disease-relevant olderMitoPark mice, which were clearly more sensitive to thePPI-disruptive effects of apomorphine and pramipexole. Thissame dose of SCH 412348 normalized the hypoactivity seenin older MitoPark mice, without significantly impactinglocomotor activity in age-matched control animals; thisis in agreement with literature reports of A2A antagonistsalleviating motor deficits in older MitoPark mice (Marcellinoet al. 2010). These findings suggest while retaining efficacyagainst the motor symptoms of PD, A2A antagonists mayminimize the potential for hallucinations and delusionsassociated with standard dopaminergic therapy. Previousstudies have demonstrated that A2A receptor antagonists blockneuroleptic-induced motor side effects in rodents (Salamoneet al., 2008) and primates (Hodgson et al., 2010). While ourfindings indicate a lack of a psychosis-like side effectassociated with A2A receptor antagonism, they do not addressthe potential issue that A2A receptor antagonists may block theefficacy of antipsychotic medications. Given that a highproportion of PD patients experience visual hallucinationsand other psychotic episodes (Friedman 2010), there is anopportunity for the co-administration of PD and antipsychoticdrugs. Future studies assessing the possible disruption ofefficacy of antipsychotics by A2A antagonists and other PDtreatments would be warranted.

However, certain caveats remain. Previous studies in intactrats and mice (Bleickardt et al. 2012) have indicated that dosesof SCH 412348 up to 3 mg/kg PO, which is 10-fold higherthan a dose that shows efficacy in the reversal of haloperidolcatalepsy model, did not impair PPI. Since MitoPark miceappear to be more sensitive to PPI disruptions possibly due totheir altered dopamine signaling, it is possible that higherdoses of SCH 412348 may result in altered PPI in MitoParkmice. However, at a minimum, there is a therapeutic windowbetween efficacy and PPI disruption with A2A antagonistssince an efficacious dose of SCH 412348 that normalizedlocomotor activity did not disrupt PPI. This is in directcontrast to our data with the dopamine agonist apomorphinewhere PPI was disrupted at a dose that did not provide anyefficacy in MitoPark mice.

Our studies have demonstrated that older MitoPark mice,which model PD-related symptomology and pathology, notonly have improved baseline PPI response that is consistentwith a loss of dopamine signaling but that these miceare more sensitive to the PPI-disruptive effects of dopaminergicagents when using this behavioral endpoint to assessneuropsychiatric side-effect profile of PD treatments. Ourresults also demonstrate that SCH 412348, a selective andpotent adenosine A2A receptor antagonist, does not disruptPPI inMitoParkmice at a dose that normalizes themotor deficitin these animals and suggests that A2A antagonists may have animproved psychosis side-effect profile as compared to thecurrent PD standards of care.

Psychopharmacology

Acknowledgment The authors would like to thank Charles Joseph forthe pharmacokinetic analysis of plasma and brain exposures ofpramipexole and SCH 412348 in MitoPark mice and age-matchedcontrols.

Disclosure SMG, RH, and LAH are responsible for the work describedin this paper. All authors were involved in at least one of the following:conception, design, acquisition, analysis, statistical analysis, andinterpretation of data; and drafting the manuscript and/or revising themanuscript for important intellectual content. All authors provided finalapproval of the version to be published.

Conflict of interest SMG, RH, and LAH are employees of MerckSharp & Dohme Corp., a subsidiary of Merck & Co., Inc., who mayown stock and/or hold stock options in the Company.

References

BakerWL, Silver D,White CM, Kluger J et al (2009) Dopamine agonistsin the treatment of early Parkinson’s disease: a metal analysis.Parkinsonism Relat Disord 15:287–294

Bleickardt CJ, LaShomb AL, Merkel CE, Hodgson RA (2012)Adenosine A2A receptor antagonists do not disrupt rodent prepulseinhibition: an improved side effect profile in the treatment ofParkinson’s disease. Parkinsons Dis 2012:1–9

Bogaerts V, Theuns J, Van Broeckhoven C (2008) Genetic findings inParkinson's disease and translation into treatment: a leading role formitochondria? Genes Brain Behav 7:129–151

Braff DL, Grillon C, Geyer MA (1992) Gating and habituation of thestartle reflex in schizophrenic patients. Arch Gen Pyschiatry 49:206–215

Chang WL, Geyer MA, Buell MR, Weber M, Swerdlow NR (2010) Theeffects of pramipexole on prepulse inhibition and locomotor activityin C57BL/6J mice. Behav Pharmacol 21:135–143

Chang WL, Breier MR, Yang A, Swerdlow NR (2011) Disparate effectsof pramipexole on locomotor activity and sensorimotor gating inSprague–Dawley rats. Pharmacol Biochem Behav 99:634–638

Chen JF, Eltzschig HK, Fredholm BB (2013) Adenosine receptors asdrug targets—what are the challenges? Nat Rev Drug Discov 12:265–286

Dauer W, Przedborski S (2003) Parkinson’s disease: mechanisms andmodels. Neuron 39:889–909

Delwaide PJ, Pepin JL, Maertens de Noordhout A (1990) Startle reactionin parkinsonism: new evidence implying the reticular formation inpathophysiology of motor troubles. Mov Disord 5S:23

Ekstrand MI, Terzioglu M, Galter D, Zhu S, Hofstetter C, Lindqvist E,Thams S, Bergstrand A, Hansson FS, Trifunovic A, Hoffer B,Cullheim S, Mohammed AH, Olson L, Larsson NG (2007)Progressive parkinsonism in mice with respiratory-chain-deficientdopamine neurons. PNAS 104(4):1325–1330

EkstrandMI, Galter D (2009) TheMitoParkMouse—an animal model ofParkinson's disease with impaired respiratory chain function indopamine neurons. Parkinsonism Relat Disord 153:5185–5188

Etminan M, Gill S, Samii A (2003) Comparison of the risk of adverseeffects with pramipexole and ropinirole in patients with Parkinson’sdisease: a meta-analysis. Drug Saf 26(6):439–444

Factor SA, Feustel PJ, Freidman JH et al (2003) Longitudinal outcome ofParkinson’s disease patients with psychosis. Neurology 60:1756–1761

Fenelon G, Mahieux F, Huon R, Ziegler M (2000) Hallucinations inParkinson's disease. Prevalence, phenomenology and risk factors.Brain 123(4):733–745

Fenelon G, Goetz CG, Karenberg A (2006) Hallucinations in Parkinsondisease in the pre-levodopa era. Neurology 66:93–98

Fenelon G (2008) Psychosis in Parkinson’s disease: phenomenology,frequency, risk factors, and current understanding of pathophysiologicmechanisms. CNS Spectr 13(4):18–25

Fenelon G, Alves G (2010) Epidemiology of psychosis in Parkinson’sdisease. J Neurol Sci 289:12–17

Ferre S, Fredholm BB, Morelli M, Popoli P, Fuxe K (1997) Adenosine–dopamine receptor–receptor interactions as an integrativemechanism in the basal ganglia. Trends Neurosci 20:482–487

Fernandez H (2012) Nonmotor complications of Parkinson disease.Cleve Clin J Med 79(2):S14–S18

Flood DG, Zuvich E, MarinoMJ, Gasior M (2011) Prepulse inhibition ofthe startle reflex and response to antipsychotic treatments in twooutbred strains in comparison to the inbred DBA/2 mice.Psychopharmacology (Berl) 215(3):441–454

Forsaa EB, Larsen JP, Wentzel-Larsen T, Alves G (2010) What predictsmortality in Parkinson’s disease? A prospective population-basedlong-term study. Neurology 75:1270–1276

Friedman JH (2010) Parkinson’s disease psychosis 2010: a review article.Parkinsonism Relat Disord 16:553–560

Galter D, Pernold K, Yoshitake T, Lindqvist E, Hoffer B, Kehr J, LarsonNG, Olson L (2010) MitoPark mice mirror the slow progression ofkey symptoms and L-DOPA response in Parkinson's disease. GenesBrain Behav 9:173–181

Geyer MA, Krebs-Thomson K, Braff DL, Swerdlow NR (2001)Pharmacological studies of prepulse inhibition models ofsensorimotor gating deficits in schizophrenics: a decade in review.Pyschopharmacology 156:117–154

Hickey P, Stacy M (2012) Adenosine A2A antagonists in Parkinson’sdisease: what’s next? Curr Neurol Neurosci Rep 12(4):376–385

Hodgson RA, Bedard PJ, Varty GB et al (2010) Preladenant, a selectiveA(2A) receptor antagonist, is active in primate models of movementdisorders. Exp Neurol 225(2):384–390

Hodgson RA, Bertorelli R, Varty GB et al (2009) Characterization of thepotent and highly selectiveA2A receptor antagonists preladenant andSCH 412348 [7-[2-[4-2,4-difluorophenyl]-1-piperazinyl]ethyl]-2-(2-furanyl)-7H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-amine] in rodent models of movement disorders and depression. JPharmacol Exp Ther 330(1):294–303

Hristova AH, Koller WC (2000) Early Parkinson’s disease: what is thebest approach to treatment? Drugs Aging 17(3):165–181

Hughes AJ, Bishop S, Kleedorfer B, Turjanski N et al (1993)Subcutaneous apomorphine in Parkinson’s disease: response tochronic administration for up to 5 years. Mov Disord 8(2):165–170, J Neurol Sci 289:12-17

Kaasinen V, Ruottinen HM, Nagren K, Lehikoinen P, Oikonen V, RinneJO (2000) Upregulation of putaminal dopamine D2 receptors in earlyParkinson’s disease: a comparative PET study with [11C]racloprideand [11C]N-methylspiperone. J Nucl Med 41(1):65–70

Marcellino D, Lindqvist E, Schneider M et al (2010) Chronic A2A

antagonist treatment alleviates parkinsonian locomotor deficiencyin MitoPark mice. Neurobiol Dis 40(2):460–466

Nakashima K, Shimoyama R, Yokoyama Y, Takahashi K (1993)Auditory effects on the electrically elicited blink reflex in patientswith Parkinson’s disease. Electroencephalogr Clin Neurophysiol 89:108–112

Newman-Tacredi A, Cussac D, Audinot V, Nicolas JP et al (2002)Differential actions of antiparkinson agents at multiple classes ofmonaminergic receptor. II. Agonist and antagonist properties atsubtypes of dopamine D2-like receptor and α1/α2-adrenoceptor.JPET 303:805–814

Poewe W, Mahlknecht P, Jankovic J (2012) Emerging therapies inParkinson’s disease. Curr Opin Neurol 25(4):488–459

Rinne JO, Iivanainen M, Metsahonkala L, Vainionpaa L, Paakkonen L,Nagren K, Helenius H (2004) Striatal dopaminergic system in dopa-

Psychopharmacology

responsive dystonias: a multi tracer PET study shows increased D2receptors. J Neurol Transm 111(1):59–67

Salamone JD, Betz AJ, Ishiwari K, Felsted J,Madson L,Mirante B, ClarkK, Font L, Korbey S, Sager TN, Hockemeyer J, Muller CE (2008)Tremorolytic effects of adenosine A2A antagonists: implications forparkinsonism. Front Biosci 13:3594–3605

Sanchez-Ramos JR, Ortoll R, Paulson GW (1996) Visual hallucinationsassociated with Parkinson's disease. ArchNeurol 53(12):1265–1268

Schapira AH (2008) Mitochondria in the aetiology and pathogenesis ofParkinson's disease. Lancet Neurol 7:97–109

Scherfler C, Khan NL, Pavese N et al (2006) Upregulation of dopamineD2 receptors in dopaminergic drug-naïve patients with Parkin genemutations. Mov Disord 21:783–788

Schicatano EJ, Peshori KR, Gopalaswamy R, Sahay E, Evinger C (2000)Reflex excitability regulates prepulse inhibition. J Neurosci 20(11):4240–4247

Schwarzchild MA, Agnati L, Fuxe K, Chen JF, Morelli M (2006)Targeting adenosine A2A receptors in Parkinson’s disease. TrendsNeurosci 29:647–654

Seeman P (1987) Dopamine receptors and the dopamine hypothesis ofschizophrenia. Synapse 1(2):133–152

Silvestri S, Seeman MV, Negrete JC, Houle S et al (2000) Increaseddopamine D2 receptor binding after long-term treatment withantipsychotics in humans: a clinical PET study. Psychopharmacology152:174–180

Singer P, Wei CJ, Chen JF, Boison D, Yee BK (2013) Deletion of striataladenosine A2A receptor spares latent inhibition and prepulse

inhibition but impairs active avoidance learning. Behav Brain Res242:54–61

Swerdlow NR, Mansbach RS, Geyer MA, Pulvirenti L, Koob GF, BraffDL (1990) Amphetamine disruption of prepulse inhibition ofacoustic startle is reversed by depletion of mesolimbic dopamine.Psychopharmacology 100:413–416

Swerdlow NR, Braff DL, Taaid N, Geyer MA (1994) Assessing thevalidity of an animal model of deficient sensorimotor gating inschizophrenic patients. Arch Gen Psychiatry 51:139–154

Swerdlow NR, Geyer MA (1998) Using an animal model ofdeficient sensorimotor gating to study the pathophysiologyand new treatments for schizophrenia. Schizophr Bull 24:285–301

SwerdlowNR, Light GA, Cadenhead KS, Sprock J, HsiehMH, Braff DL(2006) Startle gating deficits in a large cohort of patientswith schizophrenia: relationship to medications, symptoms,neurocognition, and level of function. Arch Gen Psych63(12):1325–1335

Thanvi BR, Lo TCN, Harsh DP (2005) Psychosis in Parkinson’s disease.Postgrad Med J 81:644–646

Valls-Sole J, Munoz JE, Valldeoriola F (2004) Abnormalities of prepulseinhibition do depend on blink reflex excitability: a study inParkinson’s disease and Huntington’s disease. Clin Neurophysiol115:1527–1536

Williams DR, Lees AJ (2005) Visual hallucinations in the diagnosis ofidiopathic Parkinson disease: a retrospective autopsy study. LancetNeurol 4:605–610

Psychopharmacology