Molecular Psychiatry (2013) 18, 543 -- 556 & 2013 Macmillan Publishers Limited All rights reserved 1359-4184/13 www.nature.com/mp

ORIGINAL ARTICLE A new perspective for schizophrenia: TAAR1 reveal antipsychotic- and antidepressant-like activity, improve cognition and control body weight

FG Revel1, J-L Moreau1, B Pouzet1, R Mory1, A Bradaia2, D Buchy1, V Metzler1, S Chaboz1, K Groebke Zbinden3, G Galley3, RD Norcross3, D Tuerck4, A Bruns1, SR Morairty5, TS Kilduff5, TL Wallace1,5, C Risterucci6, JG Wettstein1 and MC Hoener1

Schizophrenia is a chronic, severe and highly complex mental illness. Current treatments manage the positive symptoms, yet have minimal effects on the negative and cognitive symptoms, two prominent features of the disease with critical impact on the long-term morbidity. In addition, antipsychotic treatments trigger serious side effects that precipitate treatment discontinuation. Here, we show that activation of the -associated receptor 1 (TAAR1), a modulator of monoaminergic neurotransmission, represents a novel therapeutic option. In rodents, activation of TAAR1 by two novel and pharmacologically distinct compounds, the full RO5256390 and the partial agonist RO5263397, blocks psychostimulant- induced hyperactivity and produces a brain activation pattern reminiscent of the antipsychotic drug olanzapine, suggesting antipsychotic-like properties. TAAR1 agonists do not induce catalepsy or weight gain; RO5263397 even reduced haloperidol- induced catalepsy and prevented olanzapine from increasing body weight and fat accumulation. Finally, TAAR1 activation promotes vigilance in rats and shows pro-cognitive and antidepressant-like properties in rodent and primate models. These data suggest that TAAR1 agonists may provide a novel and differentiated treatment of schizophrenia as compared with current medication standards: TAAR1 agonists may improve not only the positive symptoms but also the negative symptoms and cognitive deficits, without causing adverse effects such as motor impairments or weight gain.

Molecular Psychiatry (2013) 18, 543--556; doi:10.1038/mp.2012.57; published online 29 May 2012 Keywords: antipsychotic; body weight; olanzapine; schizophrenia; TAAR1

INTRODUCTION combine the efficacy of current antipsychotics on the positive Schizophrenia is a chronic debilitating disorder affecting B1% of symptoms with improved effect on the negative and cognitive the population whose pathophysiological mechanisms remain symptoms, without causing motor or metabolic side effects.1,7,13 poorly understood.1--3 Aspects of schizophrenia are usually The development of antipsychotic drugs has typically focused on categorized into positive, negative and cognitive symptoms. The the DA D2 receptor as a target, considering treatment of psychotic positive (or psychotic) symptoms include hallucinations, delusions symptoms as the primary outcome.7,13 Engineering superior and thought disorganization that are caused by an excess of treatments will require addressing new biological targets, along (DA) in the mesolimbic system.4--6 These symptoms are with considering not only psychosis but also the mood and blocked by antipsychotic medication through inhibition of the DA cognitive aspects of the disease when evaluating efficacy.1,13--15 3,7,8 D2 receptors. In contrast, the negative symptoms (flattened Although the DA hypothesis of schizophrenia, which postulates a affect and emotions, anhedonia), and cognitive symptoms DA dysfunction, remains an important premise for the disorder, (impaired executive functions, attention and memory), which are several lines of evidence suggest a more complex picture in which the believed to result from a DA deficiency in the prefrontal cortex neurotransmission of other monoamines, glutamate and gamma- (PFC),4--6,8 are poorly responsive to current treatments. In addition aminobutyric acid, is also dysregulated5. Discovered in 2001,16,17 the to this limited therapeutic efficacy, current antipsychotic treat- trace amine-associated receptor 1 (TAAR1) has been demonstrated to ments produce undesirable side effects that compromise com- be an important modulator of the dopaminergic and serotonergic pliance with long-term treatment and quality of life.9--11 The first systems18--22 and potentially also of glutamatergic activity.18 TAAR1 is generation of drugs (typical antipsychotics, or neuroleptics) causes a G protein-coupled receptor that responds to so-called trace amines loss of emotional responsiveness and extrapyramidal motor (TAs), a subgroup of biogenic amines previously denoted as false disorders, such as tardive dyskinesia or dystonia, whereas the neurotransmitters.16,17,19--21,23 TAs such as b-phenylethylamine (PEA), second generation (atypical antipsychotics) engenders weight p- (pTyr) or are endogenous metabolites of gain, diabetes and increased cardiometabolic risks.9,11,12 As the amino acids with structural similarity to biogenic amines.22,24 cognitive and negative symptoms have a major impact on long- Although they are only found at low concentration in the brain, term morbidity,1,13 there is a critical need for treatments that abnormal levels of TAs have long been associated with various

1Neuroscience Research, Pharmaceuticals Division, F. Hoffmann-La Roche, Basel, Switzerland; 2Neuroservice, Domaine de Saint-Hilaire, Aix en Provence, France; 3Discovery Chemistry, Pharmaceuticals Division, F. Hoffmann-La Roche, Basel, Switzerland; 4Pharmacokinetics & Drug Metabolism, Pharmaceuticals Division, F. Hoffmann-La Roche, Basel, Switzerland; 5Biosciences Division, SRI International, Menlo Park, CA, USA and 6CNS Biomarker, Pharmaceuticals Division, F. Hoffmann-La Roche, Basel, Switzerland. Correspondence: Dr MC Hoener, F Hoffmann-La Roche, Department PCDFNB, Neuroscience Research, Pharmaceuticals Division, Building 70/331, CH-4070 Basel, Switzerland. E-mail: [email protected] Received 18 January 2012; revised 5 April 2012; accepted 9 April 2012; published online 29 May 2012 TAAR1: a promising target for schizophrenia therapy FG Revel et al 544 neuropathological disorders, including schizophrenia, major depres- (monkey and human TAAR1) as TAAR1 radioligands at a concentration 19,22--26 sion and Parkinson’s disease. In schizophrenic patients, urinary equal to Kd. and plasma PEA levels are increased, suggesting a role in the etiology 26--29 of the disease. Similar dysregulations of pTyr and tryptamine Electrophysiological recordings 25,30,31 have also been reported. Electrophysiological recordings in Xenopus oocytes expressing mouse The TAAR1 gene maps to locus 6q23, a region frequently 18,34 24,26 TAAR1 and human Kir3.1/3.2 were performed as reported. Visual associated with schizophrenia and bipolar disorder. In the whole-cell current clamp in brain slices were used to record firing activity mouse brain, Taar1 is expressed throughout the limbic and of VTA DA neurons and 5-HT DRN neurons as published previously.18,34 monoaminergic systems, including the ventral tegmental area 32 Horizontal slices (250 mm thick) were prepared from adult (3--6 months of (VTA) and dorsal raphe nucleus (DRN). Mice lacking Taar1 / / age) wild-type and Taar1 mice. All cells used for the statistical analysis (Taar1 mice) have no overt phenotype, however, they are displayed a stable firing activity for at least 30 min. hypersensitive to , showing enhanced locomotion and striatal release of DA, noradrenaline and (5- hydroxytryptamine (5-HT)) following an acute challenge.32,33 Selectivity screen Furthermore, the spontaneous firing rate of DA and 5-HT neurons RO5256390 and RO5263397 were tested in a selectivity screen at CEREP in the VTA and DRN, respectively, is augmented in Taar1/ mice, (Paris, France) as documented in Supplementary Methods. and only in wild-type mice does pTyr decrease this firing rate.18,32,34 These observations indicate that TAAR1 is an important Pharmacokinetic studies modulator of monoaminergic neurotransmission and suggest Pharmacokinetic analyses were performed as described in Supplementary therapeutic value in the context of neuropsychiatric disorders. Methods. Recently, engineering of two novel, selective and potent ligands 34,35 of mouse TAAR1, an antagonist (EPPTB or RO5212773) and an Measurement of locomotor activity (LMA) in mice agonist (RO5166017),18 has reinforced the therapeutic potential of TAAR1. Use of these ligands in vitro has revealed that TAAR1 LMA was evaluated as the number of horizontal beam breaks (horizontal activity) over 30 min as described,18,32 except where specified. Psychosti- interacts functionally with the D2 and 5-HT1A receptors in the VTA DA and DRN 5-HT neurons, respectively, to modify their mulants were dissolved in saline (0.9% NaCl þ 0.3% Tween 80). RO5256390 pharmacological properties.18,34,36 In vivo, TAAR1 agonism pro- and RO5263397 were dissolved in vehicle (H2O þ 0.3% Tween 80) and duced anxiolytic and antipsychotic-like effects in several mouse administered orally, except where specified. models.18 Notably, similar to the antipsychotic drug olanzapine, Cocaine experiments. C57BL/6J mice (n ¼ 8 per group) were administered RO5166017 reduces the DA-dependent hyperlocomotion trig- 1 1 gered either by cocaine or genetic deletion of the DA transporter, with vehicle, RO5256390 (0.3--3 mg kg ) or RO5263397 (0.3--3 mg kg ), placed into the activity monitor chamber for 30 min (habituation period), as well as hyperactivity induced by an N-methyl D-aspartate 1 (NMDA) antagonist.18 injected intraperitoneally with cocaine (20 mg kg ) and returned to the To further examine the therapeutic potential of TAAR1, we recording chamber for LMA monitoring (recording period). designed two agonists with high potency and selectivity at mouse, Phencyclidine (PCP) experiments. C57BL/6J mice (n ¼ 8 per group) were as well as rat, Cynomolgus monkey (Macaca fascicularis) and 1 human TAAR1. These new compounds, RO5256390 (full agonist) administered with vehicle, RO5256390 (0.03--3 mg kg ) or RO5263397 (0.003--1 mg kg1), and were placed for 30 min habituation into the and RO5263397 (partial agonist), show distinct pharmacological 1 signatures and display a promising profile in a range of paradigms recording chamber before receiving PCP hydrochloride (3.2 mg kg , and disease models in rodents and primates, suggesting not only intraperitoneally). Animals were then returned into the recording chamber antipsychotic-like properties but also pro-cognitive and antide- for immediate monitoring of LMA for 60 min. pressant potential. L-687,414 experiments. NMRI mice (n ¼ 8 per group) were administered with vehicle, RO5256390 (0.01--1 mg kg1) or RO5263397 (0.000031mg 1 1 MATERIALS AND METHODS kg ) 15 min before receiving L-687,414 (50 mg kg ) subcutaneously, and Animals were placed into the activity monitor chamber for 15 min (habituation) before LMA was recorded. Animal experiments at Roche (Basel, Switzerland) complied with the Swiss Federal and Cantonal laws on animal research and AAALAC regulations Cocaine experiments in association with olanzapine. C57BL/6J mice and received prior approval by the Cantonal Veterinary Office. Experiments (n ¼ 8 per group) were administered with RO5256390 (0.1 mg kg1)or with monkeys performed at Roche (Palo Alto, CA, USA) conformed with the RO5263397 (0.3 mg kg1) 30 min before receiving vehicle or olanzapine local Institutional Animal Care and Use Committee guidelines and AAALAC (0.3--1 mg kg1 orally), and were placed into the activity monitor chamber regulations. Sleep studies at SRI International were approved by SRI’s for 30 min (habituation period), injected with cocaine (20 mg kg1, Institutional Animal Care and Use Committee and were in accordance with intraperitoneally) and returned to the recording chamber for LMA the NIH guidelines. Details on animal provenance and housing are monitoring (recording period). The choice of doses for the TAAR1 com- provided in Supplementary Methods. pounds was based on prior dose-response studies with compounds from which a threshold dose for activity was selected for the combination study. Drugs All compounds were obtained from Sigma except where specified. (3R,4R)- L-687,414 experiments in association with olanzapine. NMRI mice 3-amino-1-hydroxy-4-methyl-2-pyrrolidinone (L-687,414), olanzapine, RO5256390, (n ¼ 8 per group) were administered with RO5263397 (0.001 mg kg1) RO5263397, EPPTB (RO5212773), [3H]RO5166017 ([3H](S)-4-ethyl-phenyl- 30 min before receiving vehicle or olanzapine (0.02--0.1 mg kg1 orally). amino)-methyl)-4,5-dihydro-oxazol-2-ylamine) and [3H]RO5192022 ([3H](S)- After 15 min, the animals were injected with L-687,414 (50 mg kg1, 4-2,4-difluoro-phenyl)-4,5-dihydro-oxazol-2-ylamine) were synthesized at subcutaneously), and were then placed into the activity monitor chamber Roche. for 15 min (habituation) before LMA was recorded for 60 min.

Pharmacological studies Pharmacological magnetic resonance imaging (phMRI) The cAMP and radioligand-binding assays were performed as described,18 Adult male Sprague-Dawley rats (250--300 g; n ¼ 8--9 per group) were using [3H]RO5166017 (mouse and rat TAAR1) or [3H]RO5192022 dosed orally with either vehicle (0.9% NaCl þ 0.3% Tween 80) or olanzapine

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited TAAR1: a promising target for schizophrenia therapy FG Revel et al 545 (6 mg kg1) in a first study, or with vehicle, RO5256390 or RO5263397 (1-- test session. Animals were left in the cylinder for 5 min during which 30 mg kg1) in a second study. Procedures for phMRI experiments are immobility time was recorded. detailed in Supplementary Methods.

Differential reinforcement of low-rate behavior Histology Adult male Cynomolgus monkeys (5--9 kg; n ¼ 6 per group) were trained to Staining of b-galactosidase activity was performed as reported.32 Counter- press a lever for food under a fixed ratio one schedule until they were able staining was made with nuclear fast red (Vectors laboratories, Burlingame, to earn 100 pellets during a 1-h training session. Monkeys were then CA, USA). Immunohistochemical detection of b-galactosidase was introduced to the differential reinforcement of low-rate behavior schedule performed with a chicken polyclonal antibody (catalog item ab9361, lot. where they were presented with a titrated delay period (5, 10, 30, 60, 100, 713063, Abcam, Cambridge, UK) as detailed in Supplementary Methods. 120, 160, 180 and 240 s) in which a premature response would not elicit a food reward and the interval was reset. Training continued until the MR relaxometry animals showed stable performance of two to six reinforcers during the 1-h MR relaxometry was used to measure body fat mass and lean mass content session under vehicle conditions. For RO5256390, animals received vehicle in conscious rats as detailed.37 Adult female Sprague-Dawley rats (B250 g (20% 1-methyl-2-pyrrolidinone in citrate buffer) or RO5256390 (0.3-- at the start of the experiment; n ¼ 8 per group) were housed individually 1.8 mg kg1 in vehicle, intramuscularly) 1 h before test. For RO5263397, under standard conditions with food (Kliba 3436, Provimi Kliba SA, animals received vehicle (Hypromellose (2910, 50 cps), polysorbate 80, 1 Kaiseraugst, Switzerland) and water ad libitum. The animals were benzyl alcohol in sterile H2O) or RO5263397 (1--10 mg kg in vehicle, p.o.) habituated to gavage by daily oral administration of vehicle (H2O þ 0.3% 90 min before test. The number of reinforcers obtained (number of correct Tween 80, 5 ml kg1) for 4 days. After an initial relaxometry measurement, responses), response rate (total number of lever presses (cor- the rats were treated per os (orally) (p.o.) daily for 2 weeks with either rect þ incorrect)/total session time) and inter-response time (time in vehicle or RO5263397 (3--10 mg kg1) in study 1, or with vehicle, seconds between lever presses) were analyzed and calculated as RO5263397 (1 mg kg1), olanzapine (2 mg kg1) or RO5263397 plus percentage of vehicle. olanzapine in study 2. Then body fat mass and lean mass contents were measured again and compared with the pretreatment values. In addition, food intake and body weight (BW) were assessed daily, and the cumulative Electroencephalography (EEG), core body temperature (Tb) and weight gain and food intake were determined. LMA telemetric studies Data recording and analysis were as reported.42 Three groups of eight male ± Haloperidol-induced catalepsy Sprague-Dawley rats (300 25 g) were implanted with recording devices for telemetric recordings of EEG, electromyography, Tb and LMA. Animals Haloperidol-induced catalepsy tests were performed as described in were acclimated to the handling procedures and were given two separate Supplementary Methods. 1 ml doses of vehicle p.o., one 7 days and the other 3 days before the first experimental day. Three studies were performed. For each study, a Object retrieval repeated measures design was used in which each rat received five The object retrieval paradigm was performed as reported previously38,39 separate oral dosings (2 ml kg1). In a first study, the dosings consisted of with minor variations. Briefly, 12 adult male Cynomolgus macaques were RO5256390 at three concentrations (1--10 mg kg1), a vehicle control 1 presented with a clear acrylic box (8 8 8 cm) with an open side baited (H2O þ 0.3% Tween 80) and zolpidem (30 mg kg ) during the middle of with a food treat (apple slice). Monkeys made attempts to obtain the treat the active (dark) period at Zeitgeber time 18 (Zeitgeber time 0 defined as during 17 trials consisting of 9 ‘easy’ food retrieval (food reward placed on light onset). In a separate experiment, rats received RO5263397 at three 1 the edge of the box; open side of box facing left or right), and 8 ‘difficult’ high concentrations (3--30 mg kg ), a vehicle control (H2O þ 0.3% Tween food retrieval (food reward placed deep within the box; open side of box 80) and caffeine (10 mg kg1)(RO5263397 study 1). In a third study, rats facing left or right). For RO5256390, animals were treated with vehicle received RO5263397 at three low concentrations (0.3--3 mg kg1), a vehicle 1 (20% 1-methyl-2-pyrrolidinone in citrate buffer) or RO5256390 (0.3-- control (H2O þ 0.3% Tween 80) and caffeine (10 mg kg )(RO5263397 3mgkg1 in vehicle) intramuscularly 60 min before test. For RO5263397, study 2). In studies with RO5263397, dosing occurred during the middle of animals received vehicle (Hypromellose (2910, 50 cps, Sigma-Aldrich, St Louis, the resting (light) period at Zeitgeber time 7. In all studies, the animals

MO, USA), polysorbate 80, benzyl alcohol in sterile H2O) or RO5263397 were then recorded for 6 h. Subsequent data analysis was performed as (0.3--10 mg kg1 in vehicle) p.o. 90 min before test. The percentage of described in Supplementary Methods. correct first reaches (first choice reaches in difficult trials) was analyzed. Statistical analyses Attentional set-shifting Data represent the mean±s.e.m. A P-value of 0.05 was accepted as The attentional set-shifting task was performed with adult male Long- 40,41 statistically significant. Electrophysiological data were analyzed with the Evans rats (225--275 g) as reported previously and described in Kolmogorov--Smirnoff test. Rodent behavioral data were analyzed with Supplementary Methods. All animals were treated subchronically with 1 Student’s t-test, Mann--Whitney test, one-way analysis of variance saline or PCP (5 mg kg bis in die (twice daily), intraperitoneally) for 7 days, (ANOVA), Kruskal--Wallis test or one-way repeated measures ANOVA, and experienced a washout period of 7 days before beginning habituation followed by Dunnett test, as appropriate. Catalepsy data were analyzed by and training to the set-shifting procedure. On the day of testing, rats were 1 Fisher’s exact test (RO5256390) or Mann--Whitney and Kruskall--Wallis tests given vehicle (H2O þ 0.3% Tween 80) or RO5256390 (1--10 mg kg , p.o.) followed by post-hoc Mann--Whitney tests. Monkey data were analyzed 1 h before the test session. with one-way repeated measures ANOVA or non-parametric repeated measures ANOVA, followed by Dunnett test. The phMRI data were Forced-swim test analyzed as detailed in Supplementary Methods. BW changes were On the day before the experiment, female Wistar rats (120--140 g; n ¼ 8 per analyzed by two-way ANOVA with post-hoc interaction contrasts compar- group) were placed into a glass cylinder (height, 40 cm; diameter, 18 cm) ing the linear time trend of each group against that of the vehicle and containing 16 cm of water at 23--25 1C for a 15 min training pretest forced olanzapine control groups followed by single comparisons with t-tests at swim session. Then animals received vehicle (H2O þ 0.3% Tween 80), each time point. Food intake and body composition data were analyzed RO5256390 (3--30 mg kg1) or RO5263397 (3--30 mg kg1) orally (5 ml with one-way ANOVA followed by Dunnett test. EEG data were analyzed kg1) immediately after training and B6 h later. On the next day, the rats using one- or two-way repeated measures ANOVA followed by paired were dosed again 1 h before they were placed in the glass cylinder for the t-tests.

& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 543 -- 556 TAAR1: a promising target for schizophrenia therapy FG Revel et al 546 RESULTS revealed a low affinity for all targets tested (X66-fold selectivity RO5256390 and RO5263397 are potent and selective TAAR1 for mouse TAAR1 Ki versus target Ki), including members of the agonists with distinct efficacy profiles monoaminergic system. The only exception was the imidazoline I2 RO5256390 ((S)-4-((S)-2-phenyl-butyl)-4,5-dihydro-oxazol-2-yla- receptor for which RO5263397 showed a selectivity ratio (Ki/Ki)of mine) and RO5263397 ((S)-4-(3-fluoro-2-methyl-phenyl)-4,5-dihy- 13-fold in favor of mouse TAAR1. In addition, neither RO5256390 dro-oxazol-2-ylamine; Table 1) were engineered by medicinal nor RO5263397 at concentrations up to 30 mM elicited cAMP chemistry using the same strategy as for RO5166017.18 Both production in HEK293 cells stably expressing mouse TAAR4 compounds exhibit high affinity at mouse, rat, Cynomolgus (Supplementary Table 1), the only other TAAR family member 43 monkey and human TAAR1 stably expressed in HEK293 cells, activated by TAs. Finally, pharmacokinetic analyses in mouse, rat where they potently activated cAMP production in all species and Cynomolgus monkey revealed favorable in vivo properties for (Table 1). In Xenopus oocytes co-expressing mouse TAAR1 with both compounds (Table 2). Thus, RO5256390 and RO5263397 are human Kir3 channels, both molecules evoked an inward current highly potent and selective TAAR1 agonists that are well suited for with potency comparable to that observed in the cAMP assays in vivo investigations in rodents and primates. Importantly, they (Table 1). However, the two compounds activated TAAR1 with display distinct intrinsic efficacies at TAAR1, RO5256390 being a different efficacies; while maximal cAMP levels reached by full agonist and RO5263397 a partial agonist. RO5256390 stimulation (79--107%) were comparable to that achieved by PEA (set as 100%), cAMP levels triggered by RO5263397 were lower (59--85%), suggesting partial agonism. TAAR1 agonism produces antipsychotic-like effects and The difference between the two compounds became even more potentiates the antipsychotic properties of olanzapine evident when tested on brain slices using patch-clamp recordings Recently, we demonstrated that RO5166017, a TAAR1 agonist with (Table 1; Supplementary Figure 1). RO5256390 behaved like the high efficacy and selectivity at mouse TAAR1, exhibits antipsycho- TAAR1 agonists pTyr and RO5166017,18 decreasing the firing tic-like properties in mice.18 Accordingly, we asked whether frequency of VTA DA and DRN 5-HT neurons, whereas RO5263397 RO5256390 and RO5263397 possess comparable properties. increased their firing rate, similar to the TAAR1 antagonist Similar to RO5166017 and to olanzapine,18 both compounds fully EPPTB,34 supporting a TAAR1 partial agonist profile for antagonized the locomotor-stimulating effects of cocaine, an RO5263397. Importantly, neither RO5256390 nor RO5263397 indirect DA agonist that blocks the DA transporter (Figure 1a). In altered the firing rate of VTA DA and DRN 5-HT neurons in slices addition, both RO5256390 and RO5263397 dose-dependently from Taar1/ mice (Supplementary Figure 1), demonstrating inhibited hyperlocomotion induced by two types of non- TAAR1-specific effects in the wild type. Of note, while the competitive NMDA receptor blockers, PCP and L-687,414 (Figures antagonistic efficacy of RO5256390 on the electrophysiological 1b and c).44,45 In these paradigms, the potencies of the new TAAR1 recordings was similar to the agonistic potency measured in the agonists were remarkably high, with a full effect of RO5263397 on cAMP assays, the agonistic measures for the TAAR1 partial agonist L-687,414-induced hyperlocomotion observed starting at 0.003 mg kg1 RO5263397 differed (Table 1). One explanation may be the relative (Figure 1c). These data confirm previous observations with inefficiency of the partial agonist RO5263397 to overcome TAAR1 RO516601718 and support the antipsychotic-like properties of constitutive activity or tonic activation by ambient levels of TAAR1 agonism. Interestingly, we further observed that TAAR1 agonist(s) in the preparation. agonists significantly potentiated the effect of olanzapine in Both compounds showed high selectivity for TAAR1, as blocking either cocaine- or L-687,414-induced hyperlocomotion determined from radioligand binding assays (CEREP) using 155 (Figures 1d--f). Similar results were obtained with RO5263397 in target proteins (Supplementary Table 1). For targets where potentiating the effects of another antipsychotic drug, risper- significant binding was seen at 3 mM, additional Ki determinations idone, on blocking cocaine-induced hyperlocomotion (Supple-

Table 1. Binding affinities, EC50 and IC50 values of RO5256390 and RO5263397 at primate and rodent TAAR1

Compound Parameter, assay, preparation Mouse Rat Human Monkey

a RO5256390 Ki, binding, HEK293 cells 4.4±1.6 2.9±0.8 24±116±6 b EC50, cAMP, HEK293 cells 2±0.9 5.1±1.7 16±12 16±13 (79±9%) (107±13%) (98±13%) (100±3%) c,d EC50, GIRK, Xenopus oocytes 18±6.0 ND ND ND (68±7%) d IC50, patch-clamp, VTA slices 15.2 ND ND ND d IC50, patch-clamp, DRN slices 9.8 ND ND

a RO5263397 Ki, binding, HEK293 cells 0.9±0.5 9.1±3 4.1±0.2 24±5 b EC50, cAMP, HEK293 cells 1.3±0.4 47±20 17±13 251±96 (59±4%) (76±7%) (81±9%) (85±8%) c,d EC50, GIRK, Xenopus oocytes 7.5±1.1 ND ND ND (64±2%) d EC50, patch-clamp, VTA slices 91.0 ND ND ND d EC50, patch-clamp, DRN slices 99.9 ND ND ND

Abbreviations: DRN, dorsal raphe nucleus; ND, not determined; VTA, ventral tegmental area. Values are given as nM (mean±s.e.m.). The results were obtained from at least three independent experiments. Values in parentheses represent the maximal efficacy with respect to that of b-phenylethylamine (EC50, cAMP) or p-tyramine (EC50, GIRK). aRadioligand [3H]RO5166017 for mouse and rat TAAR1, [3H]RO5192022 for human and Cynomolgus monkey TAAR1. bMillipore immunoassay for cAMP. c Current mediated by Kir3.1 and Kir3.2 co-expressed with TAAR1. dCurrent at 50 mV holding potential.

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited TAAR1: a promising target for schizophrenia therapy FG Revel et al 547 Table 2. Favorable pharmacokinetic properties of RO5256390 and RO5263397 in C57BL/6 mice, Wistar rats and Cynomolgus monkeys

Mouse Rat Monkey

Compound Parameter Units I.v. P.o. I.p. I.v. P.o. I.v. P.o. I.m.

RO5256390 Dose mg kg1 5.0 10 2.1 5.3 10 1.3 1.4 2.0 1 Cmax/dose ng ml 253 33.7 150 182 7.2 195 7.6 114 Tmax h 0 0.25 0.25 0.08 2.3 0.25 2.0 1.0 AUC/dose ng h1 ml1 323 128 193 226 53 620 35.1 580 T1/2 h 1.0 3.7 1.1 3.0 4.8 14 2.1 4.2 1 Vss lkg 4.4 6.4 11 CL ml min1 kg1 52 77 27 F % 40 60 25 5.7 94 Fu %324349 Brain/plasma Ratio 11 11

RO5263397 Dose mg kg1 1.9 4.0 1.6 1.0 3.2 1 1 1 1 Cmax/dose ng ml 564 167 252 222 45.7 1150 84.0 222 Tmax h 0.08 1 0 0.08 0.50 0.08 1.5 0.50 AUC/dose ng h1ml1 2048 867 820 508 425 2610 713 1550 T1/2 h 5.8 6.0 2.0 2.6 4.3 9.1 5.5 9.2 1 Vss lkg 3.6 6.1 3.8 CL ml min1 kg1 8.1 33 6.5 F %4240842759 1 1 1 1 Fu %28(5ngml)--51 (330 ng ml ) 47 (10 ng ml )--57 (1500 ng ml )60 Brain/plasma Ratio 2.6 2.4 1.3 18

Abbreviations: AUC, area under the plasma concentration versus time curve; Cmax, maximum concentration; CL, clearance; F, bioavailability; Fu, fraction unbound in plasma; i.m., intramuscular; i.p., intraperitoneal; i.v., intravenous; T1/2, terminal half-life; Tmax, time at which maximum concentration was observed; p.o., per os (orally); Vss, volume of distribution at steady state. RO5256390 showed low to moderate oral bioavailability, which was improved with i.p. and i.m. administration in the mouse and monkey, respectively, moderate to high volume of distribution at steady state, an elimination half-life of several hours and a high brain/plasma concentration ratio (11 more in the brain). RO5263397 showed moderate to high oral bioavailability, moderate to high volume of distribution at steady state, half-life of several hours and a high protein unbound fraction. Mouse parameters were derived from mean plasma concentrations for n ¼ 2 animals per time point. Rat and monkey parameters were derived from mean plasma concentrations for n ¼ 2 (rat) and n ¼ 3 (monkey) animals per dose route.

mentary Figure 2). Such potentiating effects appeared additive in six structures showed different responses to olanzapine and the cocaine challenge experiment (Figures 1d and e; Supplemen- TAAR1 agonists (Figure ; Supplementary Figure 3b). Perfusion in tary Figure 2), whereas they were synergistic in the L-687,414 the locus coeruleus, thalamus and core of the nucleus accumbens challenge experiment (Figure 1f). These observations highlight the decreased after olanzapine, but remained unchanged after TAAR1 potential of TAAR1 agonists as add-on treatment to current agonists. Similarly, olanzapine and RO5256390 did not alter antipsychotic medication to maximize therapeutic efficacy. perfusion in the striatum, whereas it was significantly augmented by RO5263397 at high dose. Interestingly, TAAR1 agonists did not significantly affect perfusion in the locus coeruleus (Supplemen- The brain activity pattern induced by TAAR1 agonism is similar but tary Figure 3b), where TAAR1 is neither expressed nor func- not identical to that of olanzapine tional.18,32 Finally, determination of the pattern match coefficient As TAAR1 agonists produced antipsychotic-like actions similar to (Figure 2d) confirmed that TAAR1 agonists and the atypical those of olanzapine in behavioral paradigms, we imaged and antipsychotic olanzapine alter neural activity in a similar, but not compared their effects on brain activity in living animals. For this, identical manner. Taken together, these imaging data further 1 the effects of olanzapine (6 mg kg p.o.) and different doses of support the antipsychotic potential of RO5256390 and RO5263397 1 RO5256390 or RO5263397 (1--30 mg kg , p.o.) administered to and its differentiation versus olanzapine, a representative of rats were assessed by phMRI based on continuous arterial spin current antipsychotic medication. labeling, as described previously.46,47 Continuous arterial spin labeling-phMRI provides a measure of regional blood perfusion (that is, diffusion of water from the capillary bed into the neuronal TAAR1 agonism does not increase BW and protects from tissue), which in turn is taken as a surrogate measure of local olanzapine-induced BW gain neuronal activity levels. Compared with vehicle-treated animals, A major side effect of antipsychotic medication is weight gain, the three compounds altered brain perfusion in a comparable which increases the risk of cardiovascular adverse events and manner (Figures 2a and b; Supplementary Figure 3a), with a diabetes, reduces the quality of life and causes treatment general tendency for increased perfusion in cortico-limbic areas discontinuation.9,10,48 Taar1 expression has been detected by (for example, the medial PFC and temporal cortex) and decreased reverse transcription-PCR in several peripheral organs,16,17,49--51 perfusion in more ventral, subcortical structures, such as the VTA, including the pancreas and gastrointestinal tract where TAAR1 substantia nigra, raphe nuclei, shell of the nucleus accumbens and activation might alter metabolic functions and BW regulation. entorhinal/piriform cortex (Supplementary Table 2). The magni- Using the Taar1/ mouse line in which Taar1 has been replaced tude of these changes differed with treatment and doses, but was by the LacZ reporter gene,32 we confirmed histologically the generally milder with TAAR1 agonists than with olanzapine, presence of Taar1 expression in several peripheral organs especially in the medial PFC and raphe nuclei (Figure 2b; (Supplementary Table 3). In agreement with the previous Supplementary Figure 3a; Supplementary Table 2). In contrast, reports,49--51 staining of b-galactosidase activity showed that

& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 543 -- 556 TAAR1: a promising target for schizophrenia therapy FG Revel et al 548 Taar1 is expressed in the islets of Langerhans in the pancreas, that, in contrast to many other drugs with antipsychotic proper- intestine and stomach (Figures 3a--c). No b-galactosidase activity ties,48 RO5263397 does not increase BW in rat but rather tends to was detected in tissues of control mice (Supplementary Figure 4). decrease it. Co-staining in the pancreas demonstrated that Taar1 is expressed Atypical antipsychotics are considered partly more effective by insulin-secreting b cells, but not by glucagon-secreting a than conventional neuroleptics on the negative and cognitive cells (Figures 3d and e; Supplementary Figures 5 and 6). These symptoms.13,52 However, they lead to greater weight gain and histological data show that Taar1 is expressed by peripheral metabolic disturbances which, eventually, preclude improvement organs associated with food absorption and control of glucose of the benefit-to-risk ratio.9,11,12 Among these, clozapine and metabolism. Although the function of TAAR1 in these organs olanzapine have the greatest propensity for inducing weight remains to be determined, it suggests a putative role for gain.9 As we determined earlier that TAAR1 agonism potentiates TAAR1 in the regulation of metabolic functions beside central the antipsychotic action of olanzapine, we examined whether it functions. would also modulate the effects of olanzapine on BW. Rats treated As TAAR1 agonists exhibit antipsychotic-like properties and daily with olanzapine (2 mg kg1, p.o.) experienced large increases display a brain activity profile reminiscent of olanzapine, and of BW and fat mass as compared with vehicle-treated animals especially considering the peripheral expression pattern of the (Figure 3g). However, co-treatment with a low dose of RO5263397 receptor, we addressed the question whether TAAR1 agonism (1 mg kg1, p.o.) prevented olanzapine from increasing both BW may alter BW like antipsychotic drugs. In rats, daily treatment with and fat accumulation (Figure 3g), whereas the same dose of RO5263397 for 2 weeks did not cause BW gain, as compared with RO5263397 did not alter significantly BW gain nor body fat mass vehicle (Figure 3f). Instead, BW stabilized and did not increase as when given alone. Similarly, RO5263397 also blocked the in control rats, presumably due to the observed reduction of food olanzapine-induced increase of food intake (Supplementary Figure intake (Supplementary Figure 7a). This translated into slight, but 7b). Overall, this experiment reveals that the TAAR1 partial agonist nonsignificant reductions of the lean and fat masses (Figure 3f; RO5263397 has a remarkable profile in concomitantly potentiating Supplementary Figure 7a), as shown by MR relaxometry measures the antipsychotic effects of olanzapine while minimizing its performed before and after treatment. This experiment shows metabolic-related adverse effects.

Figure 1. RO5256390 and RO5263397 inhibit cocaine and NMDA -induced hyperlocomotion in mouse and potentiate the effects of olanzapine. (a--c) Oral treatment with RO5256390 and RO5263397 significantly reduced hyperlocomotion induced either by the DA transporter blocker cocaine (20 mg kg1, intraperitoneal (i.p.)) (a) or by the NMDA receptor antagonists PCP (3.2 mg kg1, i.p.) (b) and L- 687,414 (50 mg kg1, subcutaneously) (c). ###Po0.001, #Po0.05 versus saline/veh; ***Po0.001, **Po0.01 and *Po0.05 versus stimulant/veh (n ¼ 6--8 per group in (a); 7--16 per group in (b); and 8--24 per group in (c)). Data represent the mean±s.e.m. in percentage of the stimulant/ veh group. In (b, c), data are combined from two (b) or three (c) experiments. (d, e) Both RO5256390 (0.3 mg kg1, p.o.) (d) and RO5263397 (0.3 mg kg1, p.o.) (e) potentiated the effect of subeffective doses of olanzapine on blocking hyperlocomotion induced by cocaine (20 mg kg1, i.p.; additive effects). (f) In contrast, RO5263397 (0.001 mg kg1, p.o.) and olanzapine had synergistic effects on inhibiting L-687,414- induced hyperlocomotion. **Po0.01 and *Po0.05 versus veh (n ¼ 8 per group). Data represent the mean±s.e.m. percentage inhibition of stimulant-induced LMA. For all panels, mice that did not receive stimulant were injected with saline. Veh, vehicle (water).

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited TAAR1: a promising target for schizophrenia therapy FG Revel et al 549

Figure 2. phMRI in rat brain shows antipsychotic profiles of RO5256390 and RO5263397. (a) Brain perfusion images of rats treated with RO5256390 (10 mg kg1, p.o.), RO5263397 (30 mg kg1, p.o.) or olanzapine (6 mg kg1, p.o.), after subtraction of vehicle controls. The left panel (atlas) shows the neuroanatomical templates used for quantification of signal, from the rostral pole (top) to the caudal pole (bottom). Brain areas are identified by numbers: 1, orbital PFC; 2, core of the nucleus accumbens (AcbC); 3, shell of the accumbens nucleus (AcbSh); 4, dorsal striatum; 5, entorhinal/piriform cortex; 6, temporal cortex; 7, medial PFC (mPFC); 8, septum; 9, insula; 10, bed nucleus of the stria terminalis; 11, ventral pallidum; 12, thalamus; 13, amygdala; 14, dorsal hippocampus; 15, hypothalamus; 16, perirhinal cortex; 17, ectorhinal cortex; 18, substantia nigra (SN); 19, VTA; 20, ventral hippocampus; 21, superior colliculus; 22, inferior colliculus; 23, periaqueductal gray; 24, raphe nuclei; and 25, locus coeruleus (LC). The color bar indicates perfusion differences versus vehicle-treated rats. Scale bar: 10 mm. (b) The effects of RO5256390 (10 mg kg1, p.o.) and RO5263397 (30 mg kg1, p.o.) on brain perfusion were similar to those of olanzapine (6 mg kg1, p.o.) in 10 brain areas. (c) In contrast, perfusion levels were affected differentially by the TAAR1 agonists and olanzapine in six brain structures. Data represent group-mean values after subtraction of vehicle controls (n ¼ 9 per group). Corresponding quantification plots for all the tested doses are shown in Supplementary Figure 3. (d) Pattern match coefficients (PMCs) between RO5256390, RO5263397 and olanzapine.

TAAR1 agonists reduce neuroleptic-induced catalepsy with haloperidol experienced catalepsy, which was partly pre- Extrapyramidal symptoms are major side effects of typical vented by the partial agonist RO5263397 at all doses tested antipsychotics caused by DA neurotransmission blockade in the (Supplementary Figure 8). In contrast, the full agonist RO5256390 striatum.53 In animals, neuroleptics (for example, haloperidol) was less effective, catalepsy being reduced significantly by dose induce comparable motor side effects defined as catalepsy, a 0.3 mg kg1 only. Thus, partial activation of TAAR1 seems prolonged state of fixed posture and muscular rigidity.54 In advantageous over full agonism for reducing neuroleptic-induced contrast, neither RO5256390 nor RO5263397 produced muscular extrapyramidal side effects. stiffness or any other signs of catalepsy. Moreover, because TAAR1 alters DA and 5-HT neurotransmission, which modulates extra- TAAR1 agonists enhance cognitive performance pyramidal symptoms,55 we examined whether TAAR1 agonists As TAAR1 modulates monoaminergic functions,18,19,21,32,34 which affect the cataleptic state caused by haloperidol. Rats injected regulate cognitive processes and mood states, we explored

& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 543 -- 556 TAAR1: a promising target for schizophrenia therapy FG Revel et al 550

Figure 3. RO5263397 prevents olanzapine-induced increase in BW and fat mass. (a--c) Staining of b-galactosidase activity in Taar1/ mice reveals Taar1 expression in the islets of Langerhans of the pancreas (a), stomach (b) and intestine (c). The pancreas and stomach were counterstained with nuclear fast red. (d, e) In the islets of Langerhans, b-galactosidase immunofluorescent labeling (red) localized to cells immunoreactive for insulin (green) (d), but not for glucagon (green) (e). Scale bars: 50 mm. (f) In rat, subchronic RO5263397 (RO;3 and 10 mg kg1, p.o. daily) dose-dependently reduced BW gain as compared with vehicle (veh; left panel) without significantly altering body fat mass (right panel). ***Po0.001, **Po0.01 and *Po0.05 versus veh (red: RO 10; black: RO 3). Contrasts of linear trends versus veh: P ¼ 0.0019 (RO3), P ¼ 0.0099 (RO10). (g) In contrast, subchronic treatment with olanzapine (2 mg kg1, p.o. daily; Olz) increased BW gain and fat mass. Co-administration of 1 mg kg1 RO5263397 p.o. (RO1), which did not alter weight gain alone, prevented the olanzapine- induced weight and fat mass increases. Weight gain: ***Po0.001, **Po0.01 and *Po0.05 versus veh; ###Po0.001, ##Po0.01, #Po0.05, Olz þ RO versus Olz þ veh. Contrasts of linear trends versus veh: P ¼ 0.046 (RO1), Po0.0001 (Olz), P ¼ 0.4 (Olz þ RO1). Fat mass changes: ***Po0.001 versus all other conditions. Treatments were given daily from day 1. For all panels, data represent the mean±s.e.m. (n ¼ 8 per group).

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited TAAR1: a promising target for schizophrenia therapy FG Revel et al 551 whether TAAR1 agonists may also address the cognitive and TAAR1 agonism produces antidepressant-like effects negative symptoms of schizophrenia, for which available treat- Given the modulating role of TAAR1 on the neurotransmission of ments have insufficient efficacy.1,13 monoamines,18,19,21,32,34 which has a central role in controlling The effects of RO5256390 and RO5263397 on cognitive mood states, we next addressed the effects of RO5256390 and performance were first assessed in monkey using the object RO5263397 on depression-related behaviors. TAAR1 agonists were retrieval paradigm, a test of attention and response inhibition first evaluated in rat using the forced-swim test, the most widely mediated by the PFC.39,56 Cynomolgus macaques were presented used preclinical tool for assessing antidepressant activity.58 In this with a transparent acrylic box with one open side baited with a paradigm, the partial agonist RO5263397 dose-dependently food treat, and monkeys made attempts to obtain the treat in reduced immobility time compared with vehicle (Figure 5a), as ‘easy’ (treat close to the open side) and ‘difficult’ (treat deep within opposed to the full agonist RO5256390, which did not significantly box) trials. In the difficult trials, monkeys obtain the treat on the affect immobility time at any of the doses investigated. This first attempt B50% of the time. Drugs with cognition-enhancing suggests potential antidepressant-like properties for the TAAR1 properties improve accuracy.38 Similarly, both TAAR1 agonists agonists with lower intrinsic efficacy. enhanced performance during difficult trials (Figure 4a), with 66% Then, TAAR1 agonists were assessed in monkey using the and 67% of correct responses obtained with minimally effective differential reinforcement of low-rate behavior paradigm, a form doses of RO5256390 at 3 mg kg1 intramuscularly and RO5263397 of schedule-controlled behavior in which subjects are reinforced at 1 mg kg1 p.o., respectively. for withholding a response over a specified unsignaled delay The pro-cognitive effects of TAAR1 activation were also interval.59 In this paradigm, antidepressant drugs augment the assessed in rat using the attentional set-shifting paradigm, a number of reinforcers obtained and reduce response rates.59 In rodent analog of the Wisconsin Card Sorting task in human.40,41 Cynomolgus macaques, RO5256390 and RO5263397 produced Both tasks capture response inhibition, cognitive flexibility and response profiles reminiscent of antidepressant drugs (Figure 5b). attention (that is, executive function), behavioral domains Both compounds increased significantly the number of reinforcers subserved by the frontal cortex and often impaired in schizo- obtained, while RO5256390 further lowered the response rate and phrenia.4,15 In rat, subchronic administration of NMDA receptor lengthened the inter-response time. Taken together, these data antagonists such as PCP produces long-lasting neurochemical show that TAAR1 activation has antidepressant potential and may changes and performance impairments in set-shifting, similar to ameliorate negative symptomatology in disorders such as major the cognitive impairments observed in schizophrenia.40,41 Those depression or schizophrenia. deficits can be reversed by cognitive enhancers,40,41,57 but generally not by antipsychotic drugs, including olanzapine or clozapine.41 Here, animals were treated with PCP for 7 days, and The TAAR1 partial agonist RO5263397 increases wakefulness were trained after another 7-day washout period to discriminate As TAAR1 modulates monoaminergic activity, a critical component between various tactile and odor cues (that is, set-shifting). in the regulation of vigilance states,60 we investigated whether Consistent with previous reports,40,41,57 PCP induced a selective RO5256390 and RO5263397 alter sleep--wake activities, as impairment in extradimensional attentional set-shifting (Figure 4b) determined by telemetric recordings of the EEG, electromyogram, without affecting discrimination or reversal learning (data not Tb and LMA in Sprague-Dawley rats. The effects of the full agonist shown). This indicates that PCP treatment specifically alters RO5256390 were evaluated after administration in the middle of attention shifting without causing a generalized learning impair- the dark (active) period and compared with the standard hypnotic ment.57 Importantly, acute administration of RO5256390 before drug zolpidem. RO5256390 did not affect the amount of testing reversed the PCP-induced deficit in extradimensional wakefulness, non-rapid eye movement sleep (NREM) and REM attentional set-shifting performance at doses 1 and 3 mg kg1 over the 6 h subsequent to administration (Supplementary Figures (Figure 4b). 9a--c), nor did it modify the latency to sleep and the REM:NREM Taken together, these results demonstrate that TAAR1 activa- ratio (Supplementary Figures 9d--f). The measures of sleep--wake tion promotes executive function in primates and rodents, consolidation, NREM delta power, Tb and LMA were not suggesting potential benefits for the cognitive deficits in significantly altered either (data not shown). In marked contrast, schizophrenia. zolpidem produced the expected hypnotic effects (Supplementary

Figure 4. RO5256390 and RO5263397 show pro-cognitive effects. (a) In Cynomolgus macaques, RO5256390 (intramuscularly) and RO5263397 (p.o.) increased accuracy in the difficult trials of the object retrieval task. **Po0.01, *Po0.05 versus vehicle (veh; n ¼ 11 per group). (b) In the attentional set-shifting paradigm in rats, oral administration of RO5256390 fully reversed the executive function deficit in extradimensional shift induced by PCP (5 mg kg1 bis in die (twice daily) for 7 days, intraperitoneal). #Po0.05 versus saline/veh; *Po0.05 versus PCP/veh (n ¼ 8-- 12 per group). Data represent the mean± s.e.m. Numbers on the X-axes represent doses in mg kg1.

& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 543 -- 556 TAAR1: a promising target for schizophrenia therapy FG Revel et al 552 Figure 9) together with large reductions of LMA and Tb (data not most importantly, human, compared with RO5166017, a recently shown). These data indicate that TAAR1 activation with characterized mouse TAAR1 agonist.18 RO5256390 and RO5256390 does not significantly alter the sleep--wake activities, RO5263397 exhibit good selectivity for TAAR1 over a large panel at least when given during the active phase. of targets, including monoamine receptors and transporters. The Accordingly, the effects of the partial agonist RO5263397 were only exceptions are, for RO5263397, the imidazoline I2 receptor, then evaluated after administration in the middle of the light and for RO5256390, the 5-HT2B receptor, and to a lesser extent the period (the normal rest period for nocturnal rodents) in 5-HT2A, delta 2 opioid, imidazoline I1 and muscarinic M3 receptors. comparison with the standard wake-promoting compound, Interactions with imidazoline receptors are not thought to be caffeine. At this time of the day, RO5263397 (3--30 mg kg1, p.o.) essential for the biological activity of TAAR1 agonists as both strongly promoted wakefulness at the expense of sleep (Supple- RO5256390 and RO5263397 failed to alter the firing rate of VTA DA mentary Note 1; Supplementary Figures 10--12). This observation and DRN 5-HT neurons in brain slices from Taar1/ mice, despite was confirmed in a subsequent experiment with lower doses (0.3-- the presence of binding sites for I2 in the VTA and for both I1 and 1 61 3mgkg , p.o.) intended to determine the dose-dependency of I2 in the DRN. Along these lines, the firing frequency of DRN the wake-promoting effects (Figure 6; Supplementary Note 2; 5-HT neurons is reduced by harmane, a putative endogenous Supplementary Figures 11 and 13). Together, these two experi- of imidazoline-binding sites,61 whereas RO5256390 and ments with RO5263397 point to a robust and dose-dependent RO5263397 failed to produce such an effect in Taar1/ mice. increase of wakefulness over the 6 h subsequent to administration, Concerning the other receptors hit by RO5256390, they are while NREM and REM were reduced. RO5263397 increased the unlikely to contribute significantly to its biological activity as latency to REM, but not to NREM, and greatly reduced the RO5263397 produced similar effects without having significant REM:NREM ratio (Figure 6; Supplementary Figure 10). Caffeine affinity for those targets. promoted vigilance as expected with additional characteristic A major pharmacological difference between RO5256390 and increase in LMA and Tb that RO5263397 did not produce RO5263397 lies in their respective degree of intrinsic efficacy at (Supplementary Figure 14). Collectively, these results show that TAAR1. Whereas RO5256390 triggers maximal receptor activation TAAR1 partial activation by RO5263397 promotes vigilance when similar to pTyr, RO5263397 produces only partial activation. Both given during the light phase. agonists are similarly active in a number of experimental conditions but, in some paradigms, differences were observed (for example, patch-clamp, forced-swim test, EEG and haloperidol-induced DISCUSSION catalepsy). In patch-clamp recordings, TAAR1 full activation by RO5256390 attenuates neuronal firing in the VTA and DRN, similar to Current medications for schizophrenia and related disorders are 18 still unsatisfactory.1 Although antipsychotic drugs effectively pTyr or the full agonist RO5166017. In contrast, TAAR1 partial address psychotic symptoms, they show poor efficacy on negative activation by RO5263397 augments the firing frequency, as seen with the TAAR1 antagonist EPPTB or in brain slices of Taar1/ and cognitive symptoms, in addition to being associated with 34 severe motor, hormonal and metabolic side effects.3,9,11,53 Here, mice. Whether and where such differential effects occur in an we provide compelling evidence that TAAR1 agonism could be intact brain with its large and complex networks remain to be used for the treatment of conditions such as schizophrenia, determined. Although the in vivo effects of an agonist are easily bipolar disorder or major depression. Using the novel agonists predicted, those of a partial agonist are more challenging to RO5256390 and RO5263397, we reveal that TAAR1 activity anticipate, as its effects depend on the relative degree of receptor benefits a wide range of psychobiological domains, highlighting activation, which can vary according to the brain area considered, the importance of TAAR1 in fine tuning ‘normal’ brain functioning. thepresenceofendogenousagonist(s),thedegreeofTAAR1 constitutive activity, as well as the pathophysiological status. In vitro, previous observations suggested that TAAR1 is constitutively active RO5256390 and RO5263397 are TAAR1 agonists with distinct and and/or tonically activated by endogenous agonist(s),18,34 asituation improved pharmacological properties where partial agonism results in antagonist-like effects on firing Here, we report on two new potent and selective agonists of activity. In vivo, TAAR1 can be activated not only by endogenous TAs TAAR1 with improved pharmacological profiles and optimized and related amine metabolites (for example, synephrines) but also pharmacokinetic properties in rat, Cynomolgus monkey and, by DA and 5-HT that are partial agonists at TAAR1.17,43 Accordingly,

Figure 5. RO5256390 and RO5263397 show antidepressant-like properties. (a) Oral treatment with RO5263397, but not RO5256390, dose- dependently decreased immobility time in the forced-swim test in rat. ***Po0.001, **Po0.01 versus vehicle (veh) (n ¼ 8 per group). (b)In Cynomolgus macaques submitted to a differential reinforcement for low-rate (DRL) paradigm, RO5256390 (intramuscularly) significantly increased the number or reinforcers obtained (rings) at 1 mg kg1, decreased the response rate (squares) at 1 mg kg1 and increased the inter- response time (IRT; diamonds) at 1 and 1.8 mg kg1, with respect to veh. Similarly, RO5263397 (p.o.) increased the number of reinforcers at 10 mg kg1. Reinforcers: **Po0.01 versus veh. Response rate: ##Po0.01 versus veh. IRT: $Po0.05 versus veh (n ¼ 6 per group). Data represent the mean± s.e.m. Numbers on the X-axes represent doses in mg kg1.

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited TAAR1: a promising target for schizophrenia therapy FG Revel et al 553

Figure 6. RO5263397 increases vigilance in rats. RO5263397 dose-dependently enhanced the cumulative time spent in wake (a) and decreased the cumulative times in non-REM (NREM) (b) and REM sleep (c) over the 6-h period following dosing in the middle of the resting phase (Zeitgeber time 6). RO5263397 did not alter significantly the latencies to NREM onset (d) and REM onset (e) and decreased the REM to NREM sleep ratio (f). Similarly, the positive control caffeine (10 mg kg1, p.o.; CAF) enhanced wakefulness and the latencies to NREM and REM, while decreasing the time spent in NREM and REM and increasing the REM to NREM ratio. ***Po0.001, **Po0.01 and *Po0.05 versus vehicle (veh) (n ¼ 8 per group). Numbers on the X-axes represent oral doses of RO5263397 (mg kg1). Data represent the mean±s.e.m. the net in vivo effects of RO5263397 depend on the relative local This finding is in agreement with a recent report by Espinoza concentrations of those molecules, resulting in stabilization of et al.,36 who found that haloperidol-induced catalepsy and striatal TAAR1 activity, with either agonistic-like or antagonistic-like c-Fos expression was significantly reduced in mice lacking TAAR1. consequences in situations where endogenous TAAR1 stimulation This suggests that RO5263397 effects are more likely due to its is insufficient or excessive, respectively.52 antagonistic properties, implying TAAR1 activation by endogen- ous TAs and/or constitutive receptor activation. How TAAR1 antagonism relieves catalepsy induced by D2 blockade is unclear, TAAR1 agonism to treat the positive symptoms of schizophrenia but may involve TAAR1 modulatory effects on monoaminergic Both RO5256390 and RO5263397 exhibit antipsychotic-like neurotransmission.5,18,21,55 Interestingly, neuroleptics like haloper- properties in rodent tests. Both compounds reduce hyperactivity idol were shown to enhance the activity of the aromatic L-amino- in mice challenged with cocaine, a psychostimulant that increases acid decarboxylase followed by an increased dopaminergic extrasynaptic DA, or with NMDA receptor blockers, such as PCP activity due to elevated PEA and possibly DA synthesis.63,64 Based and L-687,414. Along these lines, the TAAR1 agonist RO5166017 on our results the TAAR1 partial agonist RO5263397 may inhibit also inhibits locomotion in spontaneously hyperactive DA this effect induced by haloperidol. transporter knockout mice, a genetic model of hyperdopaminer- In line with the behavioral observations, the phMRI data show gia.18,62 Thus, the antipsychotic potential of TAAR1 agonism is that RO5256390 and RO5263397 produce brain activity signatures supported by the fact that various TAAR1 agonists with slightly reminiscent of atypical antipsychotics such as olanzapine or different pharmacological profiles all show potent activity in aripiprazole.65 This contrasts with neuronal activity patterns pharmacological and genetic models of schizophrenia. Interest- observed in schizophrenic patients and animal models. ingly, TAAR1 agonists are effective both in models based on either Imaging studies in schizophrenic patients have consistently increased DA activity or reduced glutamatergic NMDA activity,62 demonstrated dysfunctions of the fronto--temporal cortical areas, similar to typical and atypical antipsychotic drugs. In addition to whereas studies in rats using either neurodevelopmental or showing antipsychotic-like effects on their own, RO5256390 and pharmacological (PCP or amphetamine treatments) models of RO5263397 also add to the effects of suboptimal doses of schizophrenia have revealed abnormally decreased activity in the olanzapine on blocking cocaine- or L-687,414-induced hyperloco- temporal and PFCs associated with increased activity of the motion. Assuming that similar synergies occur in patients, the entorhinal/piriform cortex and shell of the nucleus accum- doses of olanzapine could be reduced to minimize its side effects bens.46,47 Altogether, these data suggest that in patients TAAR1 without altering therapeutic efficacy. agonists have the potential to normalize deficient brain activity Importantly, TAAR1 agonists are unlikely to cause extrapyrami- and generate antipsychotic effects. Interestingly, in some brain dal side effects as we did not observe any signs of motor rigidity areas, such as the thalamus or the locus coeruleus, TAAR1 agonists or dysfunctions following treatments with RO5256390 or and olanzapine altered brain activity in a different manner, RO5263397, in contrast to typical antipsychotic drugs. On the suggesting that in patients differential effects may occur for contrary, RO5263397 even partially prevented haloperidol-in- certain disease symptoms. duced catalepsy in rat, suggesting that TAAR1 partial activation Of note, some effects of atypical antipsychotics might occur via may mitigate extrapyramidal side effects caused by neuroleptics. indirect action on TAAR1. In a recent study, Karmacharya et al.66

& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 543 -- 556 TAAR1: a promising target for schizophrenia therapy FG Revel et al 554 demonstrated in mouse that TAAR1 is necessary for the enhancing A metabolic role for TAAR1: protection from olanzapine-induced effects of clozapine in the prepulse inhibition test, a paradigm that BW gain evaluates sensorimotor gating capacities. Deficits in prepulse In contrast to most antipsychotics, in particular those of the inhibition are observed in schizophrenic patients and animal atypical class,9,11,12 RO5263397 produces antipsychotic-like effects models indicative for schizophrenia, and antipsychotics improve without increasing BW and fat mass in rats, and additionally prepulse inhibition performance.67 This suggests that TAAR1 protects against olanzapine-induced BW and fat mass gains. To activation is disrupted in schizophrenia and that some anti- our knowledge, this is the first compound that associates psychotics may produce part of their therapeutic effects via antipsychotic-like features with protection from atypical antipsy- indirect modulation of TAAR1 activity. chotic-induced BW gain. Thus, association of RO5263397 to atypical antipsychotics-like olanzapine could maximize their therapeutic efficacy while reducing their metabolic-related side TAAR1 agonists have the potential to target the negative and effects, both by reducing the dose of antipsychotics and by a cognitive symptoms of schizophrenia direct physiological effect of TAAR1. Unfortunately, the mechan- Although current antipsychotic medications are generally effective isms by which antipsychotic drugs induce BW gain and metabolic on the positive symptoms of schizophrenia, they show little effects disturbances are still poorly understood.11,12,48 Antipsychotics on the cognitive and negative symptoms.1,7,13 Novel treatments generally affect a variety of neurotransmitter systems in the brain aim to target these life-long aspects that are key to the long-term and periphery, many of which have been implicated in the morbidity and functional outcome.1 regulation of food intake and energy homeostasis.11,72 The TAAR1 The cognitive and negative symptoms are believed to result agonist RO5263397 prevented olanzapine from increasing food from hypofunctioning of the PFC, a structure that has an intake, presumably via centrally mediated effects. In the mouse important role in executive functions and working memory. brain, TAAR1 is expressed in a variety of areas relevant for the Several lines of evidence suggest that hypoactivity of the control of food intake and energy homeostasis, including prefrontal DA system underlies the cognitive deficits.4--6,8 Atypical hypothalamic nuclei, the DRN, area postrema and the nucleus of drugs, such as clozapine and olanzapine, have small effects on the tractus solitarius.32 In addition, we now reveal that TAAR1 is cognitive function in patients68 that have been associated to expressed in peripheral organs directly relevant for the control of enhanced prefrontal DA levels and neuronal activation, contrast- energy homeostasis and feeding, specifically, the stomach, ing with typical antipsychotics.69,70 Our phMRI results in rats are intestine and b cells of the pancreas. Activation of TAAR1 in these consistent with such observations by showing that olanzapine organs may counteract the metabolic disturbances provoked by increases prefrontal and temporal perfusion levels. Similarly, olanzapine through modulation of digestion, food absorption and TAAR1 agonists enhance brain activity in these areas, suggesting insulin secretion. potential impact of TAAR1 activation on cognitive functions. This The distinct expression of b-galactosidase in the stomach, was substantiated by the demonstration of pro-cognitive effects intestine and the islets of Langerhans is well in agreement with of TAAR1 agonists in behavioral paradigms in monkeys and previous reports examining TAAR1 mRNA expression in human,16 rodents. Indeed, both compounds enhanced performance of rat17 and mouse49--51 tissues by reverse transcription-PCR. How- naive monkeys in the object retrieval task. In this test, ever, in marked contrast to the reports in human and rat, which performance is improved by cognition-enhancing drugs38,39 but screened for Taar1 expression in various tissues, we did not is greatly diminished after PCP exposure, correlating with reduced observe any expression in the heart, liver, kidney, lung, spleen or prefrontal DA function.56,71 In addition, RO5256390 reversed PCP- skeletal muscles. Direct detection of TAAR1 expression has proven induced impairments of rats in the attentional set-shifting task. remarkably challenging, mainly owing to its low expression levels Interestingly, RO5256390 shares this property with cognitive- and the absence of reliable antibodies.19 Reverse transcription- enhancing drugs,40 but not with antipsychotic drugs, including PCR is a particularly sensitive method to detect Taar1 expression. haloperidol, clozapine and olanzapine, which all fail to reverse the Yet, it does not provide histological resolution and may some- PCP-induced executive function deficits.41 Taken together, these times generate false positive results, particularly for genes results reveal a role for TAAR1 in modulating executive functions. expressed at low levels and for intronless genes (risk of DNA TAAR1 agonists show pro-cognitive effects, which, combined with contamination), both cases applying to TAAR1. In addition, as their antipsychotic-like properties, may represent a new treatment Taar1 expression has been reported in circulating blood cells of option for schizophrenia. human and mouse,73,74 contamination of tissue samples with this In addition to the cognitive alterations, prefrontal DA potential source of Taar1 mRNA cannot be excluded. The deficits may also contribute to the negative symptomatology. replacement of Taar1 by the LacZ reporter gene in the Taar1/ According to the phMRI data, RO5256390 and RO5263397 mouse line32 overcomes these limitations, enabling indirect increased neural activity in prefrontal areas, suggesting that localization of Taar1 expression with good sensitivity and TAAR1 agonists may have beneficial effects on negative symp- histological resolution. Finally, species differences in Taar1 toms. RO5263397, but not RO5256390, exhibited antidepressant- expression cannot be ruled out. like effects in the forced-swim test in rats, while both compounds showed a profile reminiscent of antidepressant drugs59 when evaluated in primates using the differential reinforcement of low- TAAR1, a receptor with a broad action spectrum rate behavior paradigm. Drugs from other therapeutic classes are Overall, our observations show that activation of TAAR1 produces usually devoid of such effects, although mixed results have been a remarkable variety of effects in psychiatric and cognitive models. reported for antipsychotics, including haloperidol and clozapine.59 This is likely to result from the close relationship between TAAR1 Of note, such studies were performed in rats, precluding direct and monoaminergic systems, in particular through TAAR1 comparison with our results in primates. Taken together, these interaction with the D2 and 5-HT1A receptors and potentially also data with RO5256390 and RO5263397 identify antidepressant-like with the DA transporter.18,21,32--34,36 Activation of TAAR1 may effects of TAAR1 activation and suggest a new therapeutic option result in stabilization of monoaminergic neurotransmission to fine- not only for the negative symptoms of schizophrenia, but tune DA and 5-HT activities. In contrast, inadequate TAAR1 activity potentially also for pathologies associated with depressive may remove such stabilization and contribute to pathological symptoms, such as major depression, bipolar disorders or monoaminergic activity levels. Indeed, various neuropsychiatric Parkinson’s disease. diseases have been associated with abnormal levels of TAs, such

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited TAAR1: a promising target for schizophrenia therapy FG Revel et al 555 19,22,24--28,75,76 as PEA and pTyr, suggesting possible roles in the 11 Baptista T, Zarate J, Joober R, Colasante C, Beaulieu S, Paez X et al. Drug induced 26,28,29,75 etiology of those diseases. weight gain, an impediment to successful pharmacotherapy: focus on In addition, we now reveal that TAAR1 modulation alters the antipsychotics. Curr Drug Targets 2004; 5: 279--299. sleep/wake pattern, suggesting that vigilance may be affected by 12 Casey DE, Haupt DW, Newcomer JW, Henderson DC, Sernyak MJ, Davidson M endogenous levels of TAs, whose levels depend on monoaminer- et al. Antipsychotic-induced weight gain and metabolic abnormalities: implica- gic neuron activity and catabolism rate.19,24 Interestingly, only the tions for increased mortality in patients with schizophrenia. J Clin Psychiatry 2004; TAAR1 partial agonist RO5263397 increased wakefulness in our 65Suppl 7: 4--18; quiz 19--20. studies, contrasting with the full agonist RO5256390 that did not 13 Carpenter WT, Koenig JI. The evolution of drug development in schizophrenia: past issues and future opportunities. Neuropsychopharmacology 2008; 33: alter the sleep/wake parameters, at least when dosed during the 2061--2079. active phase. How exactly TAAR1 activity impacts alertness 14 Kirkpatrick B, Fenton WS, Carpenter Jr WT, Marder SR. The NIMH-MATRICS remains to be determined, but it is likely to relate to TAAR1 consensus statement on negative symptoms. Schizophr Bull 2006; 32: 214--219. modulation of monoaminergic activity. 15 Green MF, Nuechterlein KH, Gold JM, Barch DM, Cohen J, Essock S et al. Approaching a consensus cognitive battery for clinical trials in schizophrenia: the NIMH-MATRICS conference to select cognitive domains and test criteria. Biol CONCLUSION Psychiatry 2004; 56: 301--307. TAAR1 agonists represent novel and differentiated options versus 16 Borowsky B, Adham N, Jones KA, Raddatz R, Artymyshyn R, Ogozalek KL et al. Trace amines: identification of a family of mammalian G protein-coupled existing therapies for treating schizophrenia. They display receptors. Proc Natl Acad Sci USA 2001; 98: 8966--8971. antipsychotic-like properties as potent as current antipsychotics 17 Bunzow JR, Sonders MS, Arttamangkul S, Harrison LM, Zhang G, Quigley DI et al. without inducing major side effects. Their combination with Amphetamine, 3,4-methylenedioxymethamphetamine, lysergic acid diethyla- current antipsychotics could improve therapeutic efficacy while mide, and metabolites of the neurotransmitters are agonists of reducing side effects, in particular, antipsychotic-induced weight a rat trace amine receptor. Mol Pharmacol 2001; 60: 1181--1188. gain and extrapyramidal symptoms. Finally, TAAR1 agonists show 18 Revel FG, Moreau JL, Gainetdinov RR, Bradaia A, Sotnikova TD, Mory R et al. significant potential for improving cognitive and negative TAAR1 activation modulates monoaminergic neurotransmission, preventing symptoms. These results in rodents and non-human primates hyperdopaminergic and hypoglutamatergic activity. Proc Natl Acad Sci USA raise new hopes for improving the condition of schizophrenic 2011; 108: 8485--8490. 19 Grandy DK. Trace amine-associated receptor 1-family archetype or iconoclast? patients. Proof-of-concept studies in humans are now awaited to Pharmacol Ther 2007; 116:355--390. determine whether these findings translate into therapeutic 20 Lindemann L, Hoener MC. A renaissance in trace amines inspired by a novel GPCR advances. family. Trends Pharmacol Sci 2005; 26: 274--281. 21 Miller GM. The emerging role of trace amine-associated receptor 1 in the functional regulation of monoamine transporters and dopaminergic activity. J CONFLICT OF INTEREST Neurochem 2011; 116: 164--176. All authors are employed by F. Hoffmann-La Roche, with exception of AB who is 22 Narang D, Tomlinson S, Holt A, Mousseau DD, Baker GB. Trace amines and their employed by Neuroservice, and SRM and TSK who are employed by SRI International. relevance to psychiatry and neurology: a brief overview. Bull Clin Psychopharmacol Research of AB, SRM and TSK was supported in part by F. Hoffmann-La Roche. 2011; 21: 73--79. 23 Sotnikova TD, Caron MG, Gainetdinov RR. Trace amine-associated receptors as emerging therapeutic targets. Mol Pharmacol 2009; 76: 229--235. ACKNOWLEDGEMENTS 24 Burchett SA, Hicks TP. The mysterious trace amines: protean neuromodulators of synaptic transmission in mammalian brain. Prog Neurobiol 2006; 79: 223--246. We acknowledge the continuous support by F. Hoffmann-La Roche and we thank C 25 Davis BA. Biogenic amines and their metabolites in body fluids of normal, ´ Andre, G Chiu, J Basile, H Dao, V Graf, M Maco, P Mortas, M-T Miss, S Debilly, C Waelti, psychiatric and neurological subjects. J Chromatogr 1989; 466: 89--218. S Schoeppenthau, T Bielser and M Brecheisen for excellent technical assistance; 26 Berry MD. The potential of trace amines and their receptors for treating V Dall’Asen, MS Gruyer, C Rapp and J-P Parys for the pharmacokinetic in vivo neurological and psychiatric diseases. Rev Recent Clin Trials 2007; 2: 3--19. studies and C Flament for the bioanalytical analysis; J Rodefer for the attentional set- 27 Potkin SG, Karoum F, Chuang LW, Cannon-Spoor HE, Phillips I, Wyatt RJ. shifting experiment; S Lazic for help with statistics; G Trube for valuable discussions Phenylethylamine in paranoid chronic schizophrenia. Science 1979; 206: 470--471. and the Xenopus oocyte experiments; A Harmeier, L Steward, B Kuennecke, M von 28 O0Reilly RL, Davis BA. Phenylethylamine and schizophrenia. Prog Neuropsycho- Kienlin, S Sewing, S Raab and AC Harrison for valuable discussion and input on the pharmacol Biol Psychiatry 1994; 18: 63--75. manuscript. 29 Sandler M, Reynolds GP. Does phenylethylamine cause schizophrenia? Lancet 1976; 1: 70--71. 30 Sullivan JL, Coffey CE, Basuk B, Cavenar JO, Maltbie AA, Zung WW. Urinary REFERENCES tryptamine excretion in chronic schizophrenics with low platelet MAO activity. 1 Insel TR. Rethinking schizophrenia. Nature 2010; 468: 187--193. Biol Psychiatry 1980; 15: 113--120. 2 Lewis DA, Lieberman JA. Catching up on schizophrenia: natural history and 31 Boulton AA. Some aspects of basic psychopharmacology: the trace amines. Prog neurobiology. Neuron 2000; 28: 325--334. Neuropsychopharmacol Biol Psychiatry 1982; 6: 563--570. 3 Carpenter Jr WT, Buchanan RW. Schizophrenia. N Engl J Med 1994; 330: 32 Lindemann L, Meyer CA, Jeanneau K, Bradaia A, Ozmen L, Bluethmann H et al. 681--690. Trace amine-associated receptor 1 modulates dopaminergic activity. J Pharmacol 4 Abi-Dargham A, Moore H. Prefrontal DA transmission at D1 receptors and the Exp Ther 2008; 324: 948--956. pathology of schizophrenia. Neuroscientist 2003; 9: 404--416. 33 Wolinsky TD, Swanson CJ, Smith KE, Zhong H, Borowsky B, Seeman P et al. The 5 Carlsson A, Waters N, Holm-Waters S, Tedroff J, Nilsson M, Carlsson ML. trace amine 1 receptor knockout mouse: an animal model with relevance to Interactions between monoamines, glutamate, and GABA in schizophrenia: new schizophrenia. Genes Brain Behav 2007; 6: 628--639. evidence. Annu Rev Pharmacol Toxicol 2001; 41: 237--260. 34 Bradaia A, Trube G, Stalder H, Norcross RD, Ozmen L, Wettstein JG et al. The 6 Heinz A, Romero B, Gallinat J, Juckel G, Weinberger DR. Molecular brain imaging selective antagonist EPPTB reveals TAAR1-mediated regulatory mechanisms in and the neurobiology and genetics of schizophrenia. Pharmacopsychiatry 2003; dopaminergic neurons of the mesolimbic system. Proc Natl Acad Sci USA 2009; 36Suppl 3: S152--S157. 106: 20081--20086. 7 Webber MA, Marder SR. Better pharmacotherapy for schizophrenia: what does the 35 Stalder H, Hoener MC, Norcross RD. Selective antagonists of mouse trace amine- future hold? Curr Psychiatry Rep 2008; 10: 352--358. associated receptor 1 (mTAAR1): discovery of EPPTB (RO5212773). Bioorg Med 8 Abi-Dargham A, Laruelle M. Mechanisms of action of second generation Chem Lett 2011; 21: 1227--1231. antipsychotic drugs in schizophrenia: insights from brain imaging studies. Eur 36 Espinoza S, Salahpour A, Masri B, Sotnikova TD, Messa M, Barak LS et al. Functional Psychiatry 2005; 20: 15--27. interaction between trace amine-associated receptor 1 and dopamine D2 9 Allison DB, Casey DE. Antipsychotic-induced weight gain: a review of the receptor. Mol Pharmacol 2011; 80: 416--425. literature. J Clin Psychiatry 2001; 62Suppl 7: 22--31. 37 Kunnecke B, Verry P, Benardeau A, von Kienlin M. Quantitative body composition 10 Allison DB, Mackell JA, McDonnell DD. The impact of weight gain on quality of life analysis in awake mice and rats by magnetic resonance relaxometry. Obes Res among persons with schizophrenia. Psychiatr Serv 2003; 54: 565--567. 2004; 12: 1604--1615.

& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 543 -- 556 TAAR1: a promising target for schizophrenia therapy FG Revel et al 556 38 Ballard TM, Knoflach F, Prinssen E, Borroni E, Vivian JA, Basile J et al. RO4938581, a 58 Cryan JF, Mombereau C. In search of a depressed mouse: utility of models for novel cognitive enhancer acting at GABAA alpha5 subunit-containing receptors. studying depression-related behavior in genetically modified mice. Mol Psychiatry Psychopharmacology 2009; 202: 207--223. 2004; 9: 326--357. 39 Rutten K, Basile JL, Prickaerts J, Blokland A, Vivian JA. Selective PDE inhibitors 59 O0Donnell JM, Marek GJ, Seiden LS. Antidepressant effects assessed using rolipram and sildenafil improve object retrieval performance in adult cynomolgus behavior maintained under a differential-reinforcement-of-low-rate (DRL) operant macaques. Psychopharmacology 2008; 196: 643--648. schedule. Neurosci Biobehav Rev 2005; 29: 785--798. 40 Wallace TL, Callahan PM, Tehim A, Bertrand D, Tombaugh G, Wang S et al. 60 Datta S, Maclean RR. Neurobiological mechanisms for the regulation of RG3487, a novel nicotinic alpha7 receptor partial agonist, improves cognition and mammalian sleep-wake behavior: reinterpretation of historical evidence and sensorimotor gating in rodents. J Pharmacol Exp Ther 2011; 336: 242--253. inclusion of contemporary cellular and molecular evidence. Neurosci Biobehav Rev 41 Rodefer JS, Nguyen TN, Karlsson JJ, Arnt J. Reversal of subchronic PCP-induced 2007; 31: 775--824. deficits in attentional set shifting in rats by sertindole and a 5-HT6 receptor 61 Smith KL, Jessop DS, Finn DP. Modulation of stress by imidazoline binding sites: antagonist: comparison among antipsychotics. Neuropsychopharmacology 2008; implications for psychiatric disorders. Stress 2009; 12: 97--114. 33: 2657--2666. 62 Gainetdinov RR, Mohn AR, Caron MG. Genetic animal models: focus on 42 Morairty SR, Hedley L, Flores J, Martin R, Kilduff TS. Selective 5HT2A and 5HT6 schizophrenia. Trends Neurosci 2001; 24: 527--533. receptor antagonists promote sleep in rats. Sleep 2008; 31: 34--44. 63 Berry MD, Juorio AV, Li XM, Boulton AA. Aromatic L-amino acid decarboxylase: 43 Lindemann L, Ebeling M, Kratochwil NA, Bunzow JR, Grandy DK, Hoener MC. Trace a neglected and misunderstood enzyme. Neurochem Res 1996; 21: amine-associated receptors form structurally and functionally distinct subfamilies 1075--1087. of novel G protein-coupled receptors. Genomics 2005; 85: 372--385. 64 Hadjiconstantinou M, Neff NH. Enhancing aromatic L-amino acid decarboxylase 44 Alberati D, Moreau JL, Mory R, Pinard E, Wettstein JG. Pharmacological evaluation activity: implications for L-DOPA treatment in Parkinson’s disease. CNS Neurosci of a novel assay for detecting glycine transporter 1 inhibitors and their Ther 2008; 14: 340--351. antipsychotic potential. Pharmacol Biochem Behav 2010; 97: 185--191. 65 Nordquist RE, Risterucci C, Moreau JL, von Kienlin M, Kunnecke B, Maco M 45 Large CH. Do NMDA receptor antagonist models of schizophrenia et al. Effects of aripiprazole/OPC-14597 on motor activity, pharmacological predict the clinical efficacy of antipsychotic drugs? J Psychopharmacol 2007; 21: models of psychosis, and brain activity in rats. Neuropharmacology 2008; 54: 283--301. 405--416. 46 Bruns A, Kunnecke B, Risterucci C, Moreau JL, von Kienlin M. Validation of cerebral 66 Karmacharya R, Lynn SK, Demarco S, Ortiz A, Wang X, Lundy MY et al. Behavioral blood perfusion imaging as a modality for quantitative pharmacological MRI in effects of clozapine: involvement of trace amine pathways in C. elegans and M. rats. Magn Reson Med 2009; 61: 1451--1458. musculus. Brain Res 2011; 1393: 91--99. 47 Risterucci C, Jeanneau K, Schoppenthau S, Bielser T, Kunnecke B, von Kienlin M 67 Geyer MA, Krebs-Thomson K, Braff DL, Swerdlow NR. Pharmacological studies of et al. Functional magnetic resonance imaging reveals similar brain activity prepulse inhibition models of sensorimotor gating deficits in schizophrenia: a changes in two different animal models of schizophrenia. Psychopharmacology decade in review. Psychopharmacology 2001; 156: 117--154. 2005; 180: 724--734. 68 Bilder RM, Goldman RS, Volavka J, Czobor P, Hoptman M, Sheitman B et al. 48 Boyda HN, Tse L, Procyshyn RM, Honer WG, Barr AM. Preclinical models of Neurocognitive effects of clozapine, olanzapine, risperidone, and haloperidol in antipsychotic drug-induced metabolic side effects. Trends Pharmacol Sci 2010; 31: patients with chronic schizophrenia or schizoaffective disorder. Am J Psychiatry 484--497. 2002; 159: 1018--1028. 49 Regard JB, Kataoka H, Cano DA, Camerer E, Yin L, Zheng YW et al. Probing cell 69 Gessa GL, Devoto P, Diana M, Flore G, Melis M, Pistis M. Dissociation of type-specific functions of Gi in vivo identifies GPCR regulators of insulin secretion. haloperidol, clozapine, and olanzapine effects on electrical activity of mesocor- J Clin Invest 2007; 117: 4034--4043. tical dopamine neurons and dopamine release in the prefrontal cortex. 50 Regard JB, Sato IT, Coughlin SR. Anatomical profiling of G protein-coupled Neuropsychopharmacology 2000; 22: 642--649. receptor expression. Cell 2008; 135: 561--571. 70 Robertson GS, Matsumura H, Fibiger HC. Induction patterns of Fos-like 51 Ito J, Ito M, Nambu H, Fujikawa T, Tanaka K, Iwaasa H et al. Anatomical and immunoreactivity in the forebrain as predictors of atypical antipsychotic activity. histological profiling of orphan G-protein-coupled receptor expression in J Pharmacol Exp Ther 1994; 271: 1058--1066. gastrointestinal tract of C57BL/6J mice. Cell Tissue Res 2009; 338: 257--269. 71 Jentsch JD, Redmond Jr DE, Elsworth JD, Taylor JR, Youngren KD, 52 Kim DH, Stahl SM. Antipsychotic drug development. Curr Top Behav Neurosci Roth RH. Enduring cognitive deficits and cortical dopamine dysfunction 2010; 4: 123--139. in monkeys after long-term administration of phencyclidine. Science 1997; 277: 53 Casey DE. Tardive dyskinesia and atypical antipsychotic drugs. Schizophr Res 1999; 953--955. 35: S61--S66. 72 Casey DE, Zorn SH. The pharmacology of weight gain with antipsychotics. J Clin 54 Sanberg PR, Bunsey MD, Giordano M, Norman AB. The catalepsy test: its ups and Psychiatry 2001; 62Suppl 7: 4--10. downs. Behav Neurosci 1988; 102:748--759. 73 D0Andrea G, Terrazzino S, Fortin D, Farruggio A, Rinaldi L, Leon A. HPLC 55 Lucas G, Bonhomme N, De Deurwaerdere P, Le Moal M, Spampinato U. 8-OH- electrochemical detection of trace amines in human plasma and platelets and DPAT, a 5-HT1A agonist and ritanserin, a 5-HT2A/C antagonist, reverse expression of mRNA transcripts of trace amine receptors in circulating leukocytes. haloperidol-induced catalepsy in rats independently of striatal dopamine release. Neurosci Lett 2003; 346: 89--92. Psychopharmacology 1997; 131: 57--63. 74 Nelson DA, Tolbert MD, Singh SJ, Bost KL. Expression of neuronal trace 56 Jentsch JD, Taylor JR, Elsworth JD, Redmond DE, Jr. Roth RH##Altered frontal amine-associated receptor (Taar) mRNAs in leukocytes. J Neuroimmunol 2007; cortical dopaminergic transmission in monkeys after subchronic phencyclidine 192: 21--30. exposure: involvement in frontostriatal cognitive deficits. Neurosci 1999; 90: 75 Sabelli HC, Javaid JI. Phenylethylamine modulation of affect: therapeutic and 823--832. diagnostic implications. J Neuropsychiatry Clin Neurosci 1995; 7:6--14. 57 Rodefer JS, Murphy ER, Baxter MG. PDE10A inhibition reverses subchronic 76 Sabelli HC, Fawcett J, Gusovsky F, Javaid J, Edwards J, Jeffriess H. Urinary PCP-induced deficits in attentional set-shifting in rats. Eur J Neurosci 2005; 21: phenyl acetate: a diagnostic test for depression? Science 1983; 220: 1070--1076. 1187--1188.

Supplementary Information accompanies the paper on the Molecular Psychiatry website (http://www.nature.com/mp)

Molecular Psychiatry (2013), 543 -- 556 & 2013 Macmillan Publishers Limited