STOP null mice exhibit cognitive and social defects related to
schizophrenia
a,ba,ba,ba,bcMélina Bégou, Julien Volle, Jean-Baptiste Bertrand, Philippe Brun, Didier Job,
ca,ba,bcAnnie Schweitzer, Mohamed Saoud, Thierry d’Amato, Annie Andrieux, and Marie-
a,bFrançoise Suaud-Chagny
aInstitut Fédératif des Neurosciences de Lyon, 69677 Bron cedex, France.
bUniversité Claude Bernard Lyon 1, EA 4166, Centre Hospitalier le Vinatier, 69677 Bron cedex, France cInstitut National de la Santé et de la Recherche Médicale, U866; Grenoble Institut of Neurosciences, Team 1 ? Pathophysiology of Cytoskeleton ? ; CEA-Grenoble, iRTSV/GPC; Université Joseph Fourrier, BP170 , 38042 Grenoble, Cedex 9, France
Key words: schizophrenia, STOP null mice, cognitive defects, learning, memory, social interactions
Correspondence should be addressed to M.B
Centre Hospitalier le Vinatier EA 4166, Service du Pr Thierry d’Amato, 69677 Bron cedex, France
Tel: + 33 4 78 77 75 61
Fax: + 33 4 78 77 72 09
(melina.begou@u-clermont1.fr)
Abbreviations
DISC1: Disrupted In Schizophrenia 1
MAP6: Microtubule Associated Protein 6
NL: neuroleptics
PPI: Pre-Pulse Inhibition
STOP: Stable Tubule Only Polypeptide
WT: wild-type
1
Abstract
Recently, evidence has been accumulated that schizophrenia can arise from primary synaptic defects involving structural proteins and especially microtubule associated protein. Previous experiments have demonstrated that STOP gene deletion in mice leads to a phenotype mimicking some aspects of positive symptoms classically observed in schizophrenic patients. Here we have assessed social and cognitive functions in STOP null mice. Compared to wild-type mice, STOP null mice exhibited deficit in the non-aggressive component of social recognition, in short term working memory and in social and spatial learning, which, as in human schizophrenia, were poorly sensitive to a long term treatment with typical neuroleptics. Since, social and cognitive dysfunction have consistently been considered as central features of schizophrenia, we proposed that STOP null mice may provide a useful model to understand the neurobiological correlates of social and cognitive defects in schizophrenia and to develop treatments that better target these symptoms.
2
Introduction
Mohamed et Thierry : Pb entre symptômes cognitifs qui réferent à la désorganisation du comportement et de la pensée et ? cognitive impairment ou dysfunction ? tel qu’on le définit.
Schizophrenia is one of the most disabling and emotionally devastating illnesses of the human brain. Pathology is characterized by three broad types of symptoms, including psychotic symptoms, negative symptoms, and cognitive impairments. Although neuroleptics (NL) are very efficient to improve psychotic symptoms (Ereshefsky et al., 1990; Mueser and McGurk, 2004), there is no consensus that any of the currently available NL alleviate the negative symptoms and the debilitating cognitive dysfunction associated with schizophrenia (Sergi et al., 2007; Sernyak et al., 2003; Peuskens et al., 2005) suggesting different neuronal substrates for these symptoms (Brown and Pluck, 2000; Tamminga, 2006; Burns, 2006). Negative symptoms are deficit states in which basic emotional and behavioral processes are diminished or absent. Common negative symptoms include blunted affect, anhedonia and social withdrawal. In schizophrenia, cognitive dysfunction is not global and generalized but rather is specific and selective, including problems with attention and perception, problem solving, short- and long-term memory and in particular working memory. Of the many clinical features of schizophrenia, social dysfunction and disturbances in cognitive processes might represent the core features of the illness because they often occur prior to the first psychotic episode, persist and frequently deteriorate throughout the course of illness (Thaker and Carpenter, 2001; Chemerinski et al., 2002; Mueser and McGurk, 2004). Furthermore, social and cognitive dysfunction
are strongly associated with functional impairments, including community living and work. Being able to address these deficits in a preclinical model and consequently develop drugs that better target these symptoms could lead to better outcome and better quality of life for patients.
The origin of schizophrenia is still debated, but current data favor a model in which schizophrenia arises from defects in neuronal connectivity, principally caused by synaptic alterations (Mirnics et al., 2001; Owen et al., 2005). Recently, the protein coded by a gene (DISC1) known to be disrupted in a familial form of
schizophrenia has been characterized and shown to be involved in a variety of interactions with microtubule-related organelles or proteins, suggesting that connectivity disorders in schizophrenia can result from cytoskeletal alterations (Callicott et al., 2005; Morris et al., 2003).
Consistent with this hypothesis, Shimitzu et al. (2006) have found an association between schizophrenia and polymorphisms in the MAP6 gene which encodes a microtubule-associated protein. STOP (Stable Tubule
Only Polypeptide) null mice generated by disrupting the MAP6 gene (Bosc et al., 1996; Bosc et al., 2003)
satisfy this construct validity (Andrieux et al., 2002). This genetic animal model has been already shown to display a set of defects similar to those of schizophrenia disorders. At the behavioral level, STOP null mice have been shown to display neuroleptic-sensitive behavioral abnormalities thought to simulate some aspects of positive symptoms classically observed in schizophrenic patients (Andrieux et al., 2002; Brun et al., 2005). These mice exhibit increased basal locomotor activity during the dark phase of the light/dark cycle associated with purposeless and disorganized activity (Andrieux et al., 2002; Brun et al., 2005). Additionally, STOP null mice display a supersensitivity locomotor reaction to acutely stressful situation, such as a single saline injection and exposure to novelty and an increased locomotor stimulatory effect of psychomimetic drugs such as amphetamine (Brun et al., 2005). Some of these locomotor activity defects are present in juvenile mice, whereas others appear only in adulthood mimicking thus the life long evolution of the
3
pathology (Bégou et al., 2007). Interestingly, STOP null mice exhibit also severe perturbations of complex behaviors including impaired nesting and severe nurturing defects, anxiety-related behavior, dramatically reduced aggressive encounters, impairment in sensorimotor gating (pre-pulse inhibition-PPI), failing in object and odor discrimination and poor performance in a water-maze cued place task (Andrieux et al., 2002; Fradley et al., 2005; Powell et al., 2007; Bouvrais-Veret et al., 2007). These behavioral anomalies suggest that these mice could mimic some aspects of social dysfunction in schizophrenic patients and exhibit cognitive deficits in tasks that require learning and memory. Despite their complexity, these processes can be directly assessed in animals. Here, spatial working memory was tested in a Y-maze paradigm. Spatial and social learning were investigated using olfactory cues in a spatial learning olfactory guided test and a conspecific recognition test. Preliminary to learning ability assessment, as a control of STOP null mice olfactory perception skills, mice were tested in a hidden food retrieving test using familiar and palatable food. In parallel, the social functioning of STOP null mice was further characterized by measuring conspecific interactions in home-cage. Interestingly, long-term NL treatment has been shown to suppress hyperlocomotion disorders in STOP null mice (Brun et al., 2005). Because current NL treatments are known
to be poorly effective on cognitive dysfunction in human (Sergi et al., 2007; Sernyak et al., 2003; Peuskens et al., 2005), as a proof of concept, we tested the effect of a long-term NL treatment on the behavior of mice in the conspecific recognition test and in the spatial learning olfactory guided test. The spatial and social
learning abilities were compared in drug free and long term NL-treated wild-type (WT) and STOP null mice.
4
Materials and methods
A. Animals
Male STOP null mice (STOP-/-) and their wild-type (WT) littermates (STOP +/+) were generated as previously described (Andrieux et al., 2002). Mice were housed eight per cage in a temperature (22?1?C) controlled environment under a 12:12 light/dark cycle (light from 6:00 AM to 6:00 PM), with ad libitum access
to familiar food (Harlan pellets) and water unless otherwise specified. All animals used in this study arose
from the same colony (BALBc/129 Sv background). Different mice were allocated to each test unless otherwise specified. Drug free and long term neuroleptic (NL) treated animals (3- to 5-month-old, 25 to 30 g) were investigated in a random order for comparisons between genotypes. Experimenter was blind to genotype during testing. All tests took place during the light phase of the cycle between 9 and 12 AM except for spatial learning olfactory guided test conducted from 12 to 17 PM. In all experiments, animals were allowed to habituate to the testing room 1 h before the test.
Experiments were performed in accordance with French and European Economic Community guidelines for the care and use of laboratory animals.
B. Y-maze test
Spontaneous alternation behavior and exploratory activity were recorded in a Y-maze. The apparatus consisting of three walled arms (5 cm wide, walls: 15 cm) made of black painted wood was placed in a dark room illuminated only by an halogen lamp giving an uniform dim light in the apparatus (intensity of 10 Lux in the center of the maze). The start arm (20 cm long) and the two arms forming the Y (both 15 cm long) were radiating at an angle of 120? from each other. Without prior habituation to the apparatus, each mouse was placed into the start arm and allowed to move freely through the maze during an 8-min session. The arms were extensively cleaned with water between each animal change to avoid olfactive cues. The sequence of
arm entries was manually counted from video recordings that allowed the experimenter to be out of view of the animal. A mouse was considered to have entered an arm when all four paws were positioned in the arm runway. The number of arm entries was used as an indicator of locomotor activity. An alternation was defined as entries into all three arms on consecutive choices. The alternation score (%) for each mouse was defined as the ratio of the actual number of alternations to the possible number (defined as the total number of arm entries minus two) multiplied by 100 as shown by the following equation: % Alternation = [(Number of alternations)/(Total arm entries ? 2)] × 100. Mice that completed only 8 arm entries or less within 8 min were
excluded from further analysis (Sakaguchi et al., 2005).
C. Hidden food test
Retrieving capacities of WT and STOP null mice was assessed using familiar (Harlan pellet) and unfamiliar (raisin) food. Palatability of raisin was previously tested (data not shown) in comparison with other unfamiliar food i.e. chocolate cereal (Chocopops；), honey cereal (Mielpops；) and corn flakes. When WT and STOP
5
null mice had simultaneously free access to the different unfamiliar foods, they always ate at first the raisin. We then considered that raisin was the more palatable food between the different foods tested. In a preliminary experiment, several days before testing, unfamiliar food (a raisin cut into pieces of 10 to 15 mg) was placed overnight in the home cages of the mice. Observations of consumption were taken to ensure that the novel food was eaten by all mice. In a hidden food test, each mouse was proposed to retrieve in a 2-trial session either familiar food or unfamiliar palatable food out of sight. Hidden food test was performed in a transparent Plexiglas arena (10cm(w) x 20cm(l) x 13cm(h)) closed with a transparent Plexiglas cover. The floor of the arena was covered with bedding (2 cm). At the beginning of each trial, familiar food (a Harlan pellet) or unfamiliar palatable food (a raisin) was hidden beneath bedding at one end of the arena (in the middle of one of the shorter sides, 10 cm apart). Different mice were allocated to find familiar and palatable food. Bedding and food were changed between each trial and mouse. A 16-h food deprivated mouse was introduced at the opposite end of the arena and the time taken for the mouse to find the hidden food was recorded. The mouse, not allowed to eat, was then promptly returned to its home cage for a 30-min inter-trial period. If food was not found, after 5 min, it was considered that the mouse failed, the test was stopped and 300 s was then allocated for that trial. In each trial, scores were expressed as the percentage of mice succeeding in finding food and as the latency (mean?SEM) to do it.
D. Spatial learning olfactory guided test
In this learning task, changes of mice abilities in retrieving palatable food over successive trials were assessed. This test was performed in a transparent Plexiglas arena (26 cm(w) x 41 cm(l) x 20 cm(h)) tagged with visual cues (different shaped symbols) on each wall and closed with a transparent Plexiglas cover. The floor of arena was covered with bedding (2 cm). Mice previously allocated to the hidden food test were used. The test session was conducted the day after the end of a 4-day training session. In the training session, mice were given 3 trials each day in which a piece of raisin (10-15 mg) was placed visible on top of bedding at one end of arena (in the middle of one of the shorter sides, 10 cm apart). In each trial, a mouse was introduced into the opposite end and allowed to habituate and eat for 5 min before being returned to its home cage for a 1-h inter-trial period. Between each trial and mouse, bedding and raisin were changed. At the end of the training session, WT mice exhibited a criterion of 80% of correct responses (food found under cut-off) over 3 consecutive trials. In the test session, mice were food deprivated for 6 hours before testing and were given four successive 5-min trials. The first trial (T0) was conducted with a piece of raisin visible on top of the bedding while the three other trials (T1, T2 and T3) were conducted with a piece of raisin hidden beneath 1 cm of bedding. Other conditions, including the location of raisin, the starting position of mice and the inter-trial period were the same as described in the training session. For each mouse and trial, latency to find raisin was recorded. If the piece of raisin was not found after 5 min, the test was stopped and 300 s was then allocated for that trial. Scores were expressed for each trial as the percentage of mice succeeding in finding raisin and as the latency (mean?SEM) to do it.
6
E. Conspecific recognition test
In this test, WT and STOP null mice, used as residents, were individually housed for 1 week before testing to permit establishment of a home-cage territory. Heterozygous mice, used as intruders, were housed 6 per cage and were unfamiliar of mice used as residents. Intruder mice were deeply anaesthetized by an i.p. injection of urethane (1.8 g/kg; 10 mL/kg) at least 10 min prior to the experiment. Four trials lasting for 3 min were repeated with a 12-min inter-trial period. The first trial (T1) was used to assess home-cage social interaction. In the first three trials (T1 to T3), the same anaesthetized intruder (mouse 1) was used to assess habituation. In the fourth trial (T4), a novel anaesthetized intruder (mouse 2), representing a novel challenge, was used to assess discrimination. In each trial, the intruder was placed prone at the centre of each resident home cage (10cm(w) x 20cm(l) x 13cm(h)) and the exploration activity of the resident was immediately manually recorded for 3 min. The intruder was then promptly removed. Exploration activity was measured as the time spent by the resident in sniffing the intruder animal. Sniffing is defined as olfactory exploration and close contact (<1 cm) between the resident nose or vibrissae and the body of the intruder (Hunter and Murray, 1989; Yamada et al., 2000).
F. Long term neuroleptic treatment
Long term NL treated mice received haloperidol (Haldol；, Janssen-Cilag; 0.5 mg/kg/day) and
chlorpromazine (Largactil；, Aventis Pharma; 5 mg/kg/day) dissolved in the drinking water from weaning to the day of experiment, as previously described (Andrieux et al., 2002). Haloperidol and chlorpromazine are antipsychotic drugs, widely used in the treatment of schizophrenia symptoms. Their beneficial effects have been attributed to their ability to block dopaminergic transmission thanks to their dopaminergic D receptor 2
antagonist properties. Wild-type and STOP null mice long term treated with NL were compared to corresponding free drug animals in the conspecific recognition test and in the spatial learning olfactory guided test.
G. Statistical analysis
In the Y-maze test, the percentage of alternations and the total number of arms entered were expressed as mean ? SEM and comparisons between genotypes were analyzed with an unpaired Student’s t-test.
In the hidden food test, the latency to find the food was expressed as the mean ? SEM and data were analyzed with a two-way ANOVA, with genotype as main factor and trials as the repeated measures. An unpaired Student’s t-test was used for comparison between genotypes for each trial. A paired Student’s t-test
was used for comparison between trials for each genotype. The distribution of mice finding the food was expressed as the percentage of success and a chi-square test was used to compare genotypes, at each trial. In the spatial learning olfactory guided test, the latency to find the food was expressed as the mean ? SEM and data were analyzed with a three-way ANOVA, with genotype and treatment as main factors and trials as the repeated measures. Post-hoc Newman-Keuls comparisons were used to assess differences between trials within each genotype and differences between genotypes for each trial were assessed with an unpaired Student’s t-test. The distribution of mice finding the raisin was expressed as the percentage of success. A chi-square test was used to compare genotypes, at each trial.
7
In the conspecific recognition test, the time spent exploring the anaesthetized mouse at each trial was
expressed as the mean ? SEM. Data were analyzed as described in the spatial learning olfactory guided test.
8
Results
A. Y-maze test
Spatial working memory was assessed by testing the spontaneous alternation performance in a Y-maze. In STOP null mice, the percentage of spontaneous alternation (27.0 ? 3.6 %) was significantly lower (p<0.01, Student’s t-test) than that of WT mice (46.3 ? 3.4 %) (Fig 1a). The total number of arm entries did not change between genotypes (32.4 ? 3.5 and 32.0 ? 4.7 for WT and STOP null mice, respectively, p>0.05, Student’s t-
test) (Fig 1b). These results suggest that STOP null mice exhibit impairment in working memory.
Wild type (n=8)
STOP -/- (n=8)
6040 (a)(b)
50 30
40
**3020
20% Alternation 10Number of arm entries 10
00
Figure 1: Y-maze spontaneous alternation performance in drug free WT and STOP null mice. The performance of WT and STOP null mice was expressed as the mean alternation ratios ? SEM (a) and the average number of arm entries ? SEM (b) for the overall 8-min session.
**p<0.01, unpaired Student’s t-test.
B. Hidden food test
Hidden food test was conducted to eliminate major olfactory impairments in STOP null mice. After a 16 h-food deprivation, each mouse was proposed to find in a 2-trial session either familiar food (a Harlan pellet) or unfamiliar palatable food (a raisin), hidden beneath the bedding. For each sort of food, the performance was expressed as the percentage of mice finding food and the latency to do it.
When familiar food was hidden, the percentage of mice succeeding to find food were similar between WT and STOP null mice (n=8/genotype, trial 1, 100 and 87.5 % and trial 2, 100 and 75 %, respectively) (p>0.05, trial 1 chi-square=1.07, df=1 and trial 2 chi-square=2.29, df=1). When latency was considered, a significant effect of genotype (F(1,14)=6.15, p<0.05) but not of trial was identified. The genotype x trial interaction was not significant (F(1,14)=3.38, p>0.05). Compared to WT mice, the latency to find hidden familiar food was higher in STOP null mice in the second trial (63.9 ? 16.2 and 186.3 ? 34.9 s respectively) (p<0.01, Student’s
t-test) but not in the first trial (72.3 ? 22.0 s and 128.4 ? 35.7 s, respectively) (p>0.05, Student’s t-test).
9
When the test was performed with palatable food, both genotypes were able to find hidden raisin. In WT and STOP null mice, the number of mice finding raisin under the cut-off increased over trials. Compared to WT mice, the percentage of STOP null mice finding raisin was significantly lower in first trial (p<0.05, chi-
When latency to find square=4.00, df=1), but not in the second trial (p>0.05, chi-square=2.25, df=1) (Fig 2a).hidden raisin was considered, a significant effect of genotype (F(1,16)=16.64, p<0.001) and trial (F(1,16)=27.57, p<0.001) was identified while the genotype x trial interaction was not significant (F(1.16)=0.01, p>0.05). Over trials, the latency to find hidden raisin was decreased in both WT (p<0.01, Student’s t-test) and STOP null mice (p<0.05, Student’s t-test). Compared to WT mice, the latency to find
hidden raisin was higher in STOP null mice in both trials (p<0.01, Student’s t-test) (Fig 2b). Totally, the
results suggest that in a second trial both genotypes succeed in finding hidden palatable food to the same extend under cut-off, but compared to WT mice, STOP null mice are slower to do it.
Wild type (n=9) STOP -/- (n=9)
100300** (a)(b)
25080
,200$** 60
# 150
40 Latency (s) 100 2050 $$
Percent of mice finding the hidden raisin 00Trial 1Trial 2Trial 1Trial 2
Figure 2: Hidden raisin test in drug free WT and STOP null mice.
In each trial, the performance of each genotype was expressed as the percentage of mice succeeding to find a raisin hidden beneath the bedding (a) and the latency to find it (mean;SEM) (b).
(a) #p<0.05, compared to corresponding WT mice (chi-square test).
(b) **p<0.01, compared to corresponding WT mice (unpaired Student’s t-test); $p<0.05 and $$p<0.01
compared to trial 1 (paired Student’s t-test).
C. Spatial learning olfactory guided test
Spatial learning was addressed as the ability of mice to locate hidden palatable food as rapidly as they located the visible one. Mice were tested on four consecutive trials (T0 to T3) to find a piece of raisin either visible (T0) or hidden beneath the bedding (T1 to T3).
When latency was considered, a significant main effect of genotype (F(1,35)=19.99, p<0.001) and trial (F(3,105)=17.24, p<0.001) but not treatment (F(1,35)=0.06, p>0.05) was identified. The genotype x trial
10