OUP user menu

Blockade of dopamine D3 but not D2 receptors reverses the novel object discrimination impairment produced by post-weaning social isolation: implications for schizophrenia and its treatment

David J. G. Watson , Charles A. Marsden , Mark J. Millan , Kevin C. F. Fone
DOI: http://dx.doi.org/10.1017/S1461145711000435 471-484 First published online: 1 May 2012


Dopamine D3 receptors are densely expressed in mesolimbic projection areas, and selective antagonists enhance cognition, consistent with their potential therapeutic use in the treatment of schizophrenia. This study examines the effect of dopamine D3vs. D2 receptor antagonists on the cognitive impairment and hyperactivity produced by social isolation of rat pups, in a neurodevelopmental model of certain deficits of schizophrenia. Three separate groups of male Lister hooded rats were group-housed or isolation-reared from weaning. Six weeks later rats received either vehicle or the dopamine D3 selective antagonist, S33084 (0.04 and 0.16 mg/kg), the preferential D3 antagonist, S33138 (0.16 and 0.63 mg/kg) or the preferential D2 antagonist, L-741,626 (0.63 mg/kg) s.c. 30 min prior to recording; horizontal locomotor activity in a novel arena for 60 min and, the following day, novel object discrimination using a 2-h inter-trial interval. Isolation rearing induced locomotor hyperactivity in a novel arena and impaired novel object discrimination compared to that in group-housed littermates. Both S33084 and S33138 restored novel object discrimination deficits in isolation-reared rats without affecting discrimination in group-housed controls. By contrast, L-741,626 impaired novel object discrimination in group-housed rats, without affecting impairment in isolates. S33084 (0.16 mg/kg), S33138 and, less markedly, L741,626 reduced the locomotor hyperactivity in isolates without attenuating activity in group-housed controls. Selective blockade of dopamine D3 receptors reverses the visual recognition memory deficit and hyperactivity produced by isolation rearing. These data support further investigation of the potential use of dopamine D3 receptor antagonists to treat schizophrenia.

Key words
  • Dopamine
  • D2 receptors
  • D3 receptors
  • object recognition
  • post-weaning social isolation
  • schizophrenia


Schizophrenia is a debilitating, lifelong disorder affecting 1–2% of the population with positive (e.g. hallucinations and delusions), negative (social withdrawal and anhedonia) and cognitive symptoms. Although antipsychotics provide therapeutic benefit for positive symptoms, based on dopamine D2 receptor antagonism (Seeman et al. 2006), negative and cognitive symptoms are resistant to current therapy (Keefe et al. 2007). Since cognitive symptoms often precede positive symptoms, and their treatment response correlates better to functional outcome (Green, 2006; Young et al. 2009), there is a need to develop improved treatment for schizophrenia. One promising approach is to develop drugs with additional actions at mGluR2/3 (Patil et al. 2007), mGluR1 (Lesage & Steckler, 2010), 5-HT6 (Woolley et al. 2004), muscarinic M1 (Bradley et al. 2010) or dopamine D3 (Joyce & Millan, 2005) receptors.

Radioligand binding (Diaz et al. 1995; Herroelen et al. 1994), in-situ hybridization (Bouthenet et al. 1991; Suzuki et al. 1998) and imunohistochemistry (Diaz et al. 2000; Schwartz et al. 2000) show the dopamine D3 receptor is densely expressed in limbic areas [including ventral striatum, nucleus accumbens (NAc), globus pallidus, cerebellum and islands of Calleja] and primarily located in mesolimbic rather than nigrostriatal projections. Recent use of a preferential dopamine D3 PET agonist radioligand, [11C](+)-PHNO, in the presence and absence of a D3 antagonist, SB-277011, shows the receptor is more highly expressed than the D2 receptor in the ventral pallidum, substantia nigra, thalamus and habenula with almost exclusive D2 binding in the dorsal striatum of mice and baboons (Rabiner et al. 2009). In humans [11C](+)-PHNO shows preferential uptake in the ventral striatum and globus pallidus (Graff-Guerrero et al. 2010) compared to preferential uptake of the D2 antagonist, [11C]raclopride (Graff-Guerrero et al. 2010), in the dorsal striatum. Furthermore, D3 receptors are enhanced in ventral striatum of non-medicated schizophrenia patients (Gurevich et al. 1997), several single nucleotide polymorphisms in the DRD3 gene appear to influence age-of-onset of schizophrenia (Renou et al. 2007), D3 over-expression causes behavioural sensitization to psychostimulants (possibly related to the psychosis) (Richtand et al. 2001) and D3 receptors are strongly implicated in drug-seeking behaviour, a major risk factor for psychosis in adolescents and young adults (Heidbreder & Newman, 2010). Supporting a role of D3 receptors in pubertal events leading to psychosis, D3 receptors are developmentally regulated, being abundant in germinative neuroepithelial cells innervating the forebrain, and promoting neurogenesis (Diaz et al. 1997; Joyce & Millan, 2007; Winner et al. 2009). Finally, a 2-fold increase in D3 receptor mRNA levels occurs in peripheral blood lymphocytes in patients with schizophrenia (Ilani et al. 2001) which is proposed to be an index of receptor changes occurring in the CNS and a potential biomarker for the disorder.

As most current antipsychotics display high affinity for both D2 and D3 receptors, selective D3 antagonists could provide effective treatment of the positive symptoms without causing detrimental extrapyramidal side-effects (Millan et al. 2008c). Indeed, D3 antagonists, such as SB-277011-A, do little in locomotor models of positive symptoms, reducing neither phencyclidine- nor amphetamine-induced hyperactivity (Reavill et al. 2000). Prevalent D3 receptor expression in mesolimbic projection areas has led to considerable research of its role in learning the motivational significance of a stimulus, which may contribute to some deficits seen in schizophrenia (Heidbreder & Newman, 2010). Acute dopamine D3 receptor antagonists suppress NMDA antagonist-induced locomotor sensitization and inhibit apomorphine-reduced prepulse inhibition of acoustic startle (Park et al. 2005). Furthermore, chronic SB-277011-A administration reduces spontaneous activity of mesolimbic, but not nigrostriatal, dopamine neurons (Ashby et al. 2000), consistent with the potential that antagonists might have antipsychotic effects without producing extrapyramidal effects (Millan et al. 2008c). Furthermore, preclinical data suggests D3 antagonists may have pro-cognitive effects. For instance, D3 antagonists improve performance in animal paradigms including social recognition (Millan et al. 2007, 2008b) which is thought to have translational relevance to social cognition and visual learning and memory cognitive domains affected in schizophrenia (Young et al. 2009). Furthermore, D3 knockout mice show enhanced memory and learning in step-through passive avoidance (Micale et al. 2010) and take fewer trials to reach criterion in reversal learning, accompanied by an increase in c-fos activation in the prefrontal cortex (PFC) (Glickstein et al. 2005).

Development of new antipsychotics, requires robust, reliable animal models with good clinical predictive validity (Millan & Brocco, 2008). Rearing rats in social isolation from weaning is a neurodevelopment model producing a syndrome including irreversible behavioural, structural and neurochemical alterations, akin to the broad spectrum of changes seen in schizophrenia (Bianchi et al. 2006; Fone & Porkess, 2008; King et al. 2009b; Schubert et al. 2009; Jones et al. 2011). In particular, isolation produces hyperactivity in a novel area, reflecting mesolimbic dopaminergic hyperfunction; impaired prepulse inhibition of acoustic startle, reflecting a deficit in sensorimotor gating/pre-attention processing and alterations in social interaction, akin to negative symptoms. Isolation rearing impairs learning and memory, including visual learning, social cognition, reversal learning and attentional set-shifting (Bianchi et al. 2006; Fone & Porkess, 2008; King et al. 2009b), also resembling changes seen in patients with schizophrenia. Furthermore, isolated rats have reduced medial PFC volume (Schubert et al. 2009) and altered hippocampal microtubules which might account for loss of neuronal plasticity (Bianchi et al. 2009a). Finally, the model may have good predictive validity. For instance, the reduction in prepulse inhibition in isolation-reared rats is reversed by olanzepine (Bakshi et al. 1998) and α7 nicotinic receptor agonists (Cilia et al. 2005), while water maze impairments are attenuated by 5-HT6 receptor antagonists (Fone, 2008) and novel object discrimination deficits restored by mGluR2/3 antagonists (Jones et al. 2010a).

Only one previous study (Reavill et al. 2000) has evaluated whether a selective dopamine D3 receptor antagonist can reverse the behavioural deficits of a preclinical neurodevelopmental model of schizophrenia. The current study therefore compares the ability of the selective dopamine D3 receptor antagonist, (3aR,9bS)-N-[4-(8-cyano-1,3a,4,9b-tetrahydro-3H-benzopyrano[3,4-c]pyrrole-2-yl0-butyl]-(4-phenyl) benzamide (S33084) (Millan et al. 2000b), the preferential D3vs. D2 antagonist (N-[4-[2-[(3aS,9bR)-8-cyano-1,3a,4,9b-tetrahydro[1]-benzopyrano[3,4-c]pyrrol-2(3H)-yl)-ethyl]phenylacetamide, S33138 (Millan et al. 2008a), and the preferential dopamine D2vs. D3 antagonist (3-[4-(4-chlorophenyl)-4-hydroxypiperidinyl]methylindole (L-741,626) (Millan et al. 2000a, b) to restore the novel object discrimination deficits and locomotor hyperactivity produced by rearing rats in social isolation from weaning.

Materials and methods


Male Lister Hooded rats (Biomedical Services Unit, University of Nottingham, derived from Charles River stock) were weaned from dams on post-natal day (PND) 23–24 and half of each litter either reared in individual cages with minimal handling and weekly change in sawdust but no environmental enrichment, or in groups of 3–4 for 5 wk. All animals were housed in the same temperature- (21+2°C) and humidity- (40–60%) controlled room on a 12 h light/dark cycle (lights on 07:00 hours) and had auditory, olfactory and visual contact with conspecifics but no physical interaction when in isolation. All experiments were performed during the light phase with animals acclimatized to the behavioural suite for 30 min prior to testing. Food (standard animal chow, USA) and water was available ad libitum except during experimental procedures. All procedures were performed in accordance with the Animals (Scientific Procedures) Act, 1986 and local ethical committee (University of Nottingham) approval. All studies were conducted by a single experienced examiner who was unaware of the treatment. Behavioural apparatus was cleaned using 20% ethanol to remove odour cues.

Each dose–response study with a dopamine receptor antagonist was performed in a separate isolation-reared cohort of rats (n=8–11 per group). Within each study rats were randomly assigned to treatment groups after measuring the 60-min drug-free locomotor response in a novel area on PND 59, to ensure each isolation group had comparable basal responses prior to drug treatment. On PND 66–67 animals underwent a further assessment of novel arena-induced locomotor activity and novel object discrimination following a 2-h inter-trial interval, respectively.

Novel arena-induced locomotor activity

Twelve automated computerized infrared activity monitors were used to assess habitation in novel arena-induced locomotor activity, which reflects mesolimbic dopamine activity thought to be dysfunctional in schizophrenia. For all studies each rat was placed in a Perspex cage (39 cm×23.5 cm×24.5 cm) with wire mesh lids and an exchangeable floor, surrounded by photobeams to automatically detect horizontal ambulations (by consecutive break of two adjacent beams). The infrared recording equipment used in the S33084 study (Apple Macintosh IIcx computer running in-house designed software, Activity Monitor, Medical Physics, Nottingham University) (Clemett et al. 1998; King et al. 2009a) was replaced with more sensitive equipment having extra photobeams (Photobeam Activity System, San Diego Instruments, USA) explaining the lower count level recorded with S33084 compared to S33138 and L-741,626 studies. However, the activity boxes used by both recording systems were the same. Drugs or vehicle were administered 30 min prior to placing animals in activity chambers and cumulative beam breaks were recorded in 5-min epochs for 60 min. The order of testing was randomized across treatment groups and between days.

Novel object discrimination

To examine changes in recognition memory induced by isolation, a two-trial object discrimination task adapted from Ennaceur & Delacour (1988), routinely used in our laboratory (Bianchi et al. 2006, 2009b; King et al. 2009b; Lapiz et al. 2000), was performed.

On the test day, animals were re-acclimatized to the locomotor cage for 3 min, returned briefly to their home cage while two identical objects (8 cm high, 5 cm diameter water-filled, plastic bottles covered with white masking tape) were secured by Blu-Tack to the floor in opposite corners, 5 cm from the side and 10 cm from the back of each cage. The rat explored both objects for 3 min during the first familiarization trial and the time (s) spent exploring each object was recorded manually with stopwatches. Animals were returned to their home cage for an inter-trial interval of 2 h before being reintroduced to the cage with one of the original (familiar) objects and one novel object (identical white marked plastic bottle covered with four prominent 1.2 cm black horizontal stripes). During the second choice trial, exploration of each object was recorded again separately. The location of the novel object was varied in a pseudorandom order within groups. Exploratory behaviour was defined as sniffing, touching and direct attention to the object, with active vibrissae whilst the nose was within 1 cm of the object. Climbing on or chewing the object were not considered exploratory behaviours and not recorded. A trained observer performed behavioural scoring and during the learning process two observers simultaneously scored behaviour; the inter-rater reliability was excellent, having a highly significant (p<0.0001) linear correlation (R2=0.6988). In all experiments drugs were administered 30 min prior to the first trial.

Drug administration

S33138 (0.16 and 0.63 mg/kg) and S33084 (0.04 and 0.16 mg/kg) were dissolved in saline. L-741,626 (0.63 mg/kg) was dissolved in saline with lactic acid before being adjusted to neutral pH. All drugs were injected (l ml/kg s.c.) 30 min prior to each test. All drugs were provided by Institut de Recherches Servier.

Data analysis

The locomotor activity time-course was analysed with two-way ANOVA using treatment as main factor and time as repeated measure while the cumulative locomotor response was analysed by two-way ANOVA with LSD post-hoc tests. For object discrimination raw object exploratory times were compared during familiarization and choice trials and the pattern of exploration (novel vs. familiar for each rat) during the choice trial examined using multivariant repeated-measures (rm)-ANOVA (with object as the repeated measure) to determine any alteration in object exploration profile. As variation in exploration of individual objects between rats can confound interpretation, the choice trial raw data was converted to discrimination ratio [d2: (novel – familiar)/(novel+familiar)] and exploration time difference [d1: (novel – familiar)] values which were analysed by two-way ANOVA followed by LSD post-hoc test to determine any change in exploration pattern independent of actual object exploration times. Values are mean±s.e.m., and p<0.05 was considered significant.


Confirmation of isolation rearing syndrome

In all three drug studies isolation increased locomotor activity in a novel arena compared to that in vehicle-treated group-reared littermates, confirming the reproducibility of the neurodevelopmental changes. This also validated using this paradigm to assess the differential effect of dopamine D2 and D3 receptor antagonists on learning and memory deficits in the novel object task.

Neither social isolation nor drug treatment had any significant effect on rat weight in any of the three experiments (data not shown) confirming this cannot be a confounding factor on the marked behavioural changes observed.

Effects of dopamine antagonists on object discrimination deficits in isolation-reared rats

In all studies group-housed vehicle-treated rats readily distinguished novel from familiar objects, spending on average between 8–10 s more on the former. In contrast, vehicle-treated isolates failed to distinguish objects during the choice trial. Isolation rearing thus produced robust changes enabling the impact of drug pre-treatment on visual recognition deficits to be compared.

Acute pre-treatment with the preferential D3 antagonist, S33084, had no effect on the pattern of novel vs. familiar object exploration in group-housed controls but at the highest dose (0.16 mg/kg) restored preferential attention to the novel object in isolates (Fig. 1a, b). rm-ANOVA confirmed a significant effect of housing condition (F1,39=4.902, p=0.033), drug (F2,39=3.500, p=0.040) and a drug×housing interaction (F1,39=5.478, p=0.024) on the distribution of novel and familiar object exploration (Fig. 1a). Isolates given vehicle or the lowest S33084 dose spent equal time exploring the novel and familiar objects (Fig. 1a) but isolates given the highest dose spent significantly longer exploring the novel object (p<0.001) analogous to behaviour in vehicle- and drug-treated group-housed rats. Examination of the object exploration time difference (Table 1) showed isolates given vehicle or the lower dose of S33084 spent significantly less (p<0.01) time difference on the novel compared to the familiar object than either group-housed groups or isolates pre-treated with the highest S33084 dose but the time difference was restored with the higher dose in isolates. An identical pattern was seen with the derived discrimination ratio (d2); there being a main effect of housing (F1,39=15.393, p<0.001) and drug (F(2,39)=6.050, p=0.005) and a drug×housing interaction (F1,39=11.843, p=0.001), such that the highest S33084 dose restored discrimination in isolates without effecting group-housed controls (Fig. 1b).

Fig. 1

Effect of S33084 (0.04 and 0.16 mg/kg s.c. −30 min, as indicated) or vehicle on novel object discrimination in group-housed and isolation-reared rats. (a) Exploration of the novel and familiar objects (mean+s.e.m., s; n=8–9) during the choice trial, *** p<0.001 from novel object Bonferroni's test following ANOVA. (b) Discrimination ratio (novel – familiar)/(novel+familiar), *** p<0.001 from social/vehicle, +++ p<0.001 from social/0.16 mg/kg and ### p<0.001 from isolate/vehicle and isolate/0.04 mg/kg post-hoc test following ANOVA. ■, Novel; □, familiar.

View this table:

In the S33138 group rm-ANOVA showed a significant effect of drug treatment (F2,44=3.649, p=0.034) and a drug×housing interaction (F1,44=11.272, p=0.002) on the pattern of exploration of the novel and familiar objects in the choice trial (Fig. 2a). Only isolates receiving vehicle treatment spent an equal time exploring the novel and familiar objects and isolates given either dose of S33138 spent significantly longer (p<0.001) exploring the novel object, just like group-housed vehicle- or drug-treated rats. Further analysis of the derived difference (d1) exploration time (Table 1) or the discrimination ratio (Fig. 2b) confirmed an overall main effect of drug (F2,44=3.649, p=0.034 and F2,44=3.682, p=0.033, respectively) and a housing×drug interaction (F1,44=11.272, p=0.002 and F1,44=10.813, p=0.002, respectively) in the pattern of choice-trial object exploration, such that values for vehicle-treated isolation-reared rats were significantly less than that of any other group (Fig. 2b). Analysis of the pattern of object exploration showed both doses of S33138 reversed the loss in preferential exploration of the novel object produced by isolation, while the highest dose had no effect in group-housed controls.

Fig. 2

Effect of S33138 (0.16 and 0.63 mg/kg s.c. −30 min, as indicated) or vehicle on novel object discrimination in group-housed and isolation-reared rats. (a) Exploration of the novel and familiar objects (mean+s.e.m., s; n=9–10) during the choice trial, *** p<0.001 from novel object Bonferroni's test following ANOVA. (b) Discrimination ratio (novel – familiar)/(novel+familiar), ** p<0.01 from social/vehicle, ++ p<0.01 from social/ 0.16 mg/kg, ### p<0.001 from isolate/0.16 mg/kg and †† p<0.01 from isolate/0.63 mg/kg, post-hoc test following ANOVA. ■, Novel; □, familiar.

Figure 3 a shows the preferential D2 antagonist, L-741,626 had a totally different effect on the pattern of choice-trial object exploration. L-741,626 was unable to restore discrimination in isolates and actually impaired discrimination in group-housed controls. rm-ANOVA showed significant effects of housing and drug (F1,39=19.259, p<0.001 and F1,39=19.347, p<0.001, respectively) and a drug×housing interaction (F1,39=6.607, p=0.014) on the choice-trial distribution of object exploration times (Fig. 3a). Only group-housed rats receiving vehicle spent significantly more time exploring the novel object (p<0.001). Analysis of the discrimination ratio showed L-741,626 reduced object discrimination in group-housed rats (Fig. 3b), ANOVA showing a main effect of drug (F1,39=23.686, p<0.001) and housing (F1,39=22.766, p<0.001) and a significant drug×housing interaction (F1,39=10.079, p=0.003). Thus the discrimination ratio of group-housed rats receiving L-741,626 was significantly less (p<0.001) than that in vehicle-treated group-housed rats and not different from isolates, irrespective of whether they had been given drug. Furthermore, an identical pattern was seen for the novel and familiar object time exploration time difference (Table 1), the value for group-housed rats given vehicle being significantly greater (p<0.001) than for any other treatments.

Fig. 3

Effect of L-741,626 (0.63 mg/kg s.c. −30 min, as indicated) or vehicle on novel object discrimination in group-housed and isolation reared rats. (a) Exploration of the novel and familiar objects (mean+s.e.m., s; n=10–11) during the choice trial, *** p<0.001 from novel object Bonferroni's test following ANOVA. (b) Discrimination ratio (novel – familiar)/(novel+familiar). *** p<0.001 from social/vehicle, post-hoc test following ANOVA. ■, Novel; □, familiar.

In all three studies drug changes in object exploration resulted from redistribution of time spent exploring the objects and not from reduced total cumulative exploration of both objects (Table 1) in either the familiarization (S33084 group: F2,39=0.164, p=0.849; S33138 group: F2,44=1.278, p=0.289; L-741,626 group: F1,39=0.026, p=0.873) or choice trials (S33084 group: F2,39=0.188, p=0.829; S33138 group: F2,44=0.058, p=0.944: L-741,626 group: F1,39=0.166, p=0.686), consistent with the effect being due to change in learning and memory and not non-specific motor impairment or loss of interest. Housing had no effect on total exploration during the choice trial in any group (S33084 group: F1,39=2.916, p=0.096; S33138 group: F1,44=0.276, p=0.602; L-741,626 group: F1,39=0.990, p=0.326). Housing did not affect total exploration during the familiarization trial in either the S33084 or S33138 groups (F1,39=0.073, p=0.788 and F1,44=0.030, p=0.864, respectively). There was a significant effect of housing in the L-741,626 group (F1,39=5.153, p=0.029) but only during familiarization, where isolates spent more time exploring both objects than their group-housed littermates (Table 1). There were no housing×drug treatment interaction effects on total exploration during either trial in any study.

Effect of dopamine antagonists on isolation-rearing induced hyperactivity

As expected isolation produced hyperactivity compared to group-housed controls in the novel arena. Irrespective of housing (isolation or group rearing) or treatment, rats showed a highly significant progressive reduction in activity over the hour in all studies (data not shown).

Two-way rm-ANOVA of the time-course of the locomotor response (data not shown) in rats given S33084 showed a housing×time-course interaction (F11,429=1.847, p=0.045) but no significant drug×time-course interaction. Visual analysis of the behaviour clearly suggested group-housed rats had lower activity, so the total cumulative locomotor activity was compared across the entire observation period.

Two-way ANOVA of the cumulative locomotor counts for the S33084 study showed a main effect of drug (F2,44=3.268, p=0.049) but no effect of housing nor any housing×drug interaction (Fig. 4a). Post-hoc analysis showed that isolates treated with either vehicle or the lowest S33084 dose exhibited significantly higher activity than group-housed rats treated with the highest S33084 dose. Isolated rats treated with 0.04 mg/kg S33084 were also more active (p<0.05) than vehicle-treated group-housed controls. In contrast, isolates treated with the highest S33084 dose were less active (p<0.05) than isolates given 0.04 mg/kg S33084, having counts indistinguishable from those in both drug- and vehicle-treated group-housed controls. The differences in total locomotor activity between the isolate/vehicle and group-housed/vehicle and the isolate/vehicle and isolate/0.16 mg/kg groups both just failed to reach significance (p=0.052 and p=0.054, respectively). Thus the highest dose of S33084 appeared to selectively reduce the hyperactive response in isolates but have no effect in group-housed littermates.

Fig. 4

Comparison of the effect of (a) S33084 (0.04 and 0.16 mg/kg), (b) S33138 (0.16 and 0.63 mg/kg) and (c) L-741,626 (0.63 mg/kg) with vehicle (1 ml/kg s.c., 30 min prior to recording activity as indicated) in group-housed and isolation-reared rats on the total cumulative beam breaks (mean+s.e.m., n=8–10) recorded over 60 min under bright light (200 lx) in a novel open-field arena. Note that in all three drug study groups the isolation-reared rats were hyperactive compared to vehicle treated group-housed (social) rats. Although S33188 tended to reduce the cumulative counts in isolates this did not reach significance. (a) * p<0.05 from social/vehicle, + p<0.05 from social/0.16 mg/kg and # p<0.05 from isolate/0.04 mg/kg, (b) ** p<0.01 from social/vehicle, (c) * p<0.05 from social/vehicle, ++ p<0.01 from social/L741,626, Bonferroni post-hoc test following ANOVA. The infrared boxes used to record activity were replaced with more sensitive equipment following the study with S33084 accounting for the lower total cumulative counts recorded in this than the other two drug studies.

Two-way rm-ANOVA of the locomotor time-course in isolates given the preferential D3/D2 antagonist, S33138 (data not shown) found no significant interaction between housing condition or drug treatment, nor any housing×drug×time interaction, although once again vehicle-treated group-housed rats appeared to have the lowest activity level, especially during the last 30 min. Analysis of the total 60 min cumulative locomotor activity (Fig. 4b) showed a significant main effect of housing condition (F1,44=6.130, p=0.017) and a housing×drug interaction (F1,44=4.361, p=0.043). Post-hoc analysis showed that isolates were significantly more active (p<0.01) than group-housed vehicle-treated controls but there were no other significant differences between any treatment. Thus the total cumulative counts in isolates treated with either dose of S33138 were not significantly greater than that in vehicle-treated group-housed rats, consistent with the drug reducing the hyperactivity seen in isolates.

In the D2 receptor antagonist, L-741,626, study two-way rm-ANOVA showed a housing×time-course interaction (F11,429=8.887, p<0.001), such that group-housed rats appeared to exhibit lower activity than isolates, as in other drug studies, especially during the first 30 min. Comparison of the total locomotor counts (Fig. 4c) confirmed that vehicle-treated isolates had significantly higher counts than vehicle-treated (p<0.05) or L-741,626-treated group-housed rats (p<0.01). Isolates treated with L-741,626 did not have significantly lower activity than vehicle-treated isolates but their cumulative activity was not significantly greater than either group-housed groups suggesting that this drug also reduced hyperactivity produced by isolation rearing.


Many mixed dopamine D2/D3 receptor antagonists have been produced, but a few potent selective D3 antagonists have recently become available (Heidbreder & Newman, 2010; Sokoloff et al. 2006), and some have been examined in pharmacological animal models of schizophrenia (Gyertyan et al. 2008; Millan et al. 2011; Zhang et al. 2006). However, only one previous study has examined the effects of D3 antagonists in a neurodevelopmental model of this disorder using the prepulse inhibition of acoustic startle paradigm but no learning and memory task (Reavill et al. 2000). The spontaneous novel object discrimination paradigm used herein does not require training or reward for motivation but reflects the innate curiosity to explore novelty, making it a powerful tool to investigate the neurobiology of learning and memory (Dere et al. 2007; Winters et al. 2008). We have previously shown that isolation-reared rats have a robust deficit in novel object recognition (Bianchi et al. 2006; King et al. 2009b; Lapiz et al. 2003) when a 2- to 3-h inter-trial interval is used. However isolates can discriminate novel and familiar objects if a short (1 min) inter-trial interval is used (Lapiz et al. 2000), they do not spend more time attending the familiar object under such conditions as expected if they were unable to habituate to this object, but instead show enhanced speed of natural forgetting of the object compared to group-housed controls, indicative of a specific impairment in recognition memory. Lesion and microinjection studies implicate the critical importance of temporal cortical areas, such as perirhinal and entorhinal cortex and the globus pallidus (Winters et al. 2008), while with longer inter-trial intervals the amygdala and hippocampus become involved in a less well-defined way (Gaskin et al. 2003; Gould et al. 2002). A recent report resulting from the preceding MATRICS clinical evaluation was that the novel object discrimination task has translational relevance to visual learning and memory impairments seen in schizophrenia (Young et al. 2009). Time delay-induced impairments in object recognition are reversed by compounds acting through diverse pharmacological mechanisms, including histamine H3 antagonists (Pascoli et al. 2009), PDE4 (Rutten et al. 2006) and PDE5 (Prickaerts et al. 2005) inhibitors, 5-HT4 partial agonists (Moser et al. 2002) and 5-HT6 antagonists (Fone, 2008; King et al. 2004). In contrast, performance is impaired by CB1 receptor agonists (Kosiorek et al. 2003) and muscarinic receptor antagonists, such as scopolamine (Woolley et al. 2003). This pharmacological evidence justifies the validity and sensitivity of this task to compare the effects of dopamine receptor antagonists and evaluate their potential to treat cognitive deficits seen in schizophrenia.

This study demonstrates that acute blockade of D3 receptors restores the object discrimination impairment produced by isolation rearing. In contrast, pre-treatment with a selective D2 antagonist impaired discrimination in group-housed controls and had no effect on recognition memory in isolates. In line with the current findings, the mixed dopamine D2/3 receptor antagonist, raclopride, impaired object discrimination in group-housed adult rats (Woolley et al. 2003) and, using a maze task discrete injection of dopamine D2 receptor antagonists, eticlopride, into the PFC impaired the ability of rats to switch set from a directional based to a visual-cue-related response solution (Floresco et al. 2006). Interestingly, the dopamine D3 receptor agonist, 7-OH-DPAT, impaired object discrimination in marmosets (Smith et al. 1999), suggesting that D3 and D2 receptors may have opposing actions in this task, but no previous study has reported the effect of a selective D3 antagonist on object discrimination. The current findings are consistent with the recent report that the preferential D3 receptor antagonist, S33138, can reverse delay-induced forgetting in novel object and social novelty recognition in adult rats (Millan et al. 2011). Furthermore, in rhesus monkeys chronically exposed to MPTP to produce cognitive impairment without motor disruption, S33138 (0.04 and 0.16 mg/kg p.o.) reversed the deficit in extra-dimensional shifts seen in an attentional set-shifting task of cognitive flexibility, and reversed age-related impairment in a delayed matching-to-sample task of working memory in old rhesus monkeys (Millan et al. 2011). These data strongly support the suggestion that dopamine D3 and D2 receptors have opposing effects on memory across species. This is consistent with passive avoidance task work where a scopolamine-induced memory disruption was antagonized by the modestly selective dopamine D3 antagonist, nafadotride, and reversal of scopolamine-induced disruption by apomorphine was antagonized by the D2 receptor antagonist, sulpiride (Sigala et al. 1997). Furthermore, SB-277011-A reverses the learning deficits in a water labyrinth induced by scopolamine and the anxiogenic benzodiazepine inverse agonist, FG-7142 (Laszy et al. 2005). In contrast, unlike haloperidol the preferential dopamine D3 receptor antagonists, SB-277011-A and A-437203 and the selective antagonist, AVE 5997 did not restore prepulse inhibition deficits in the neonatal ventral hippocampal lesion model of schizophrenia (Zhang et al. 2006), suggesting that D2 and not D3 receptor antagonism may mediate reversal of this sensorimotor gating deficit. Conversely, other groups found atypical antipsychotics and not haloperidol reversed the prepulse deficit in this lesion model (Lipska et al. 1995) which is thought to reflect post-synaptic dopamine supersensitivity in subcortical mesolimbic projection areas, while prefrontal cortical dysfunction may also be involved in the object recognition deficit seen in isolates. As drugs were given systemically in the current study, the precise anatomical substrate of the observed pro-cognitive effect of D3 receptor antagonists is unknown. Of note, however, neurons in the rostral and ventromedial shell of NAc, which exhibit high levels of dopamine D3 receptors, receive their dopaminergic innervation from the ventral tegmental area and project to the ventral pallidum and mediodorsal thalamus and from these areas neurons project as a potential polysynaptic loop to the entorhinal and prefrontal cortices (Canto et al. 2008; Sesack & Grace, 2010), regions both critical for performance of recognition memory in the novel object discrimination paradigm (Ennaceur et al. 1997). It is therefore interesting that acute administration of selective dopamine D3 receptor antagonist, SB-277011-A, increases extracellular levels of acetylcholine in the PFC (Lacroix et al. 2003, 2006). Furthermore, D3 and D2 receptor antagonists used herein have a differential effect on acetylcholine release in the PFC and reversal of social memory in the rat. S33084 and S33138 enhance cortical acetylcholine release as measured by microdialysis (Millan et al. 2007) and also reverse a scopolamine-induced impairment of social recognition in the rat, while the dopamine D2 receptor antagonist, L-741,626, had no effect on acetylcholine release and had amnesic properties (Millan et al. 2007). However, this does not prove any direct involvement of the PFC in these responses. Since microinjection of S33084 into the frontal cortex enhances social recognition in the rat (Loiseau & Millan, 2009), discrete microinjection studies and/or evaluation of whether the dopamine receptor antagonists can reverse delay-dependent or pharmacological-induced impairment in object recognition would clearly be of value to determine the specific brain regions and pathways involved in mediating the improvement in object recognition.

Rearing rats in social isolation from weaning produces a reproducible locomotor hyperactivity upon being placed in a novel, mildly aversive arena, such as an open field. Temporal analysis of this hyperactivity shows this is due to attenuation of habituation compared to group-housed littermates, rather than a generalized increase in basal ambulatory activity (Fone & Porkess, 2008; Fone et al. 1996; Gentsch et al. 1982; King et al. 2009b; Sahakian et al. 1977). In the current study, the isolation-induced hyperactivity was attenuated only by the highest dose of S33084 while the same dose had no effect on horizontal activity in group-housed littermates. Although isolates had higher total locomotor activity counts than group-reared controls during the 1-h observation period in the other two drug treatment studies, and there was no significant effect of either S33138 or L-741,626, both drugs also appeared to suppress activity in isolates as this was no longer significantly greater than seen in group-housed vehicle controls. Isolation-induced hyperactivity is thought to involve alterations in mesolimbic dopamine release in the striatum without any accompanying change in D2 receptor density (Del Arco et al. 2004) which may have translational relevance to the positive symptoms, such as psychotic agitation, seen in schizophrenia (Porsolt et al. 2010). D3 receptor levels are high in the rodent NAc (Joyce & Millan, 2005); however, it is unknown whether its expression or function there is altered by isolation rearing similar to that reported in schizophrenia (Gurevich et al. 1997), or whether alteration may contribute to the hyperactivity seen in isolates. Noteably the D3 receptor is co-localized with Trk B receptors in the NAc and D3 receptor expression is potently regulated by brain-derived neurotrophic factor (BDNF) during early development (Guillin et al. 2001; Sokoloff et al. 2006). Since corticolimbic BDNF protein and mRNA levels are modified in schizophrenia (Takahashi et al. 2000) and isolates (Scaccianoce et al. 2006) developmental changes in BDNF and D3 levels may be related to mesolimbic hyperdopaminergic tone. Preclinical data suggest that activation of the dopamine D3 receptor with low doses of agonists, such as 7-OH-DPAT, quinelorane and PD 128,907, inhibits spontaneous motor activity (Daly & Waddington, 1993; Millan et al. 2004), probably by activation of receptors in the NAc (Ouagazzal & Creese, 2000). By contrast, D3 antagonists, such as nafadotride, U99194A, and NGB 2904, enhance locomotor activity (Clifford & Waddington, 1998; Pritchard et al. 2007; Sautel et al. 1995) in the rodent but are without effect in D3 receptor knockout mice (Pritchard et al. 2007). Although such findings were initially controversial due to the lack of selectivity of the original compounds used (Boulay et al. 1999), more selective antagonists, like S33084, have confirmed that blockade of the D3 receptor does not reduce spontaneous locomotor activity in an open field (Millan et al. 2004), consistent with the current data: moreover, it produces a mild antiparkinsonian effect in the MPTP-lesioned marmoset without producing dyskinesia (Silverdale et al. 2004). By comparison, systemic administration of preferential dopamine D2 receptor antagonists, such as S23199 or L-741,626, suppress spontaneous locomotion (Millan et al. 2004) consistent with the observation in the current study that the latter drug tends to attenuate the hyperactivity seen in isolates. Higher doses of the D2 antagonists than utilized herein would reduce locomotion in group-housed rats (Millan et al. 2004) which would confound interpretation of novel object exploration. In a risk-based decision-making task where two levers were differentially rewarded, amphetamine-induced increased preference for risk-taking was prevented by pretreatment with the D2 antagonist, eticlopride, but enhanced by the D3 antagonist, nafadotride (St Onge & Floresco, 2009), consistent with these receptors producing opposing effects on several behaviours.

D3 antagonists attenuate hyperactivity in isolates suggesting there could be neurodevelopmental alteration in the distribution and or pre- and post-synaptic function of D3 receptors. Consistent with a functional link between D3 expression and spontaneous activity, the ratio of functional D3 to non-functional D3 mRNA splice variants was reduced in the substantia nigra and ventral tegmentum of rats showing a high locomotor response to a novel environment (Pritchard et al. 2006). However, isolation-induced changes in novel environment habituation correlate to altered expression of several immediate early-onset genes in the medial PFC (Levine et al. 2007), whose volume is reduced by this manipulation (Schubert et al. 2009), so changes in other areas could account for the actions of D3 antagonists. It is unlikely that S33084 attenuates activity in isolates by acting at D2 receptors since it had no effect on locomotion in group-housed rats, and this dose is specific to D3 receptors (Millan et al. 2000a, b).

In conclusion, selective blockade of dopamine D3 receptors reverses the visual recognition memory deficit and hyperactivity produced by rearing rat pups in isolation from weaning, supporting further investigation of the potential use of D3 receptor antagonists to treat some of the symptoms of schizophrenia.


We thank Miss Stacey Knapp and Mr Ian Topham for their technical assistance and Institut de Recherches Servier for funding this research.

Statement of Interest

The Institut de Recherches Servier (IRS) provided the financial support and the compounds utilized; M.J.M. is an employee of IRS.


View Abstract