The deletion of STOP/MAP6 protein in mice triggers highly altered mood and impaired cognitive performances

Authors


Address correspondence and reprint requests to Marie-Pascale Martres, INSERM UMRS 952, CNRS UMR 7224, Université Pierre et Marie Curie, Physiopathologie des Maladies du Système Nerveux Central, 9 Quai Saint Bernard, case courrier 37, 75252 Paris cedex 5, France. E-mail: marie-pascale.martres@snv.jussieu.fr

Abstract

J. Neurochem. (2012) 121, 99–114.

Abstract

The microtubule-associated Stable Tubulie Only Polypeptide (STOP; also known as MAP6) protein plays a key role in neuron architecture and synaptic plasticity, the dysfunctions of which are thought to be implicated in the pathophysiology of psychiatric diseases. The deletion of STOP in mice leads to severe disorders reminiscent of several schizophrenia-like symptoms, which are also associated with differential alterations of the serotonergic tone in somas versus terminals. In STOP knockout (KO) compared with wild-type mice, serotonin (5-HT) markers are found to be markedly accumulated in the raphe nuclei and, in contrast, deeply depleted in all serotonergic projection areas. In the present study, we carefully examined whether the 5-HT imbalance would lead to behavioral consequences evocative of mood and/or cognitive disorders. We showed that STOP KO mice exhibited depression-like behavior, associated with a decreased anxiety-status in validated paradigms. In addition, although STOP KO mice had a preserved very short-term memory, they failed to perform well in all other learning and memory tasks. We also showed that STOP KO mice exhibited regional imbalance of the norepinephrine tone as observed for 5-HT. As a consequence, mutant mice were hypersensitive to acute antidepressants with different selectivity. Altogether, these data indicate that the deletion of STOP protein in mice caused deep alterations in mood and cognitive performances and that STOP protein might have a crucial role in the 5-HT and norepinephrine networks development.

Abbreviations used:
5-HT

serotonin

5-HTP

5-hydroxytryptophan

DA

dopamine

DAT

DA transporter

KO

knock-out

l-DOPA

3,4-dihydroxyphenylalanine

MAO

monoamine oxidase

MAP

microtubule-associated protein

NE

norepinephrine

NET

NE transporter

SERT

serotonin transporter

STOP

Stable Tubulin Only Polypeptide

TH

tyrosine hydroxylase

TPH2

tryptophan hydroxylase 2

VMAT2

vesicular monoamine transporter 2

WT

wild-type

A large body of evidence supports the hypothesis whereby changes to the neuronal architecture and/or synaptic efficacy may be implicated in psychiatric disorders (Mirnics et al. 2001). In particular, cytoskeletal alterations and abnormal microtubule dynamics may contribute to cerebral disconnectivity observed in these diseases (Morris et al. 2003; Callicott et al. 2005; Blackwood et al. 2007; Camargo et al. 2007). Consistently, the protein disrupted in schizophrenia 1 (DISC1) interacts with microtubule-associated proteins (MAP; Porteous and Millar 2006; Ross et al. 2006). In addition, dysbindin-1, the gene polymorphisms of which are associated with both bipolar disorders and schizophrenia, also interacts with and regulates microtubules (Talbot et al. 2006).

In line with this hypothesis, disruption in mice of the gene encoding the microtubule stabilizing protein STOP (stable tubule only polypeptide, also called MAP6) generates numerous schizoid-like characteristics (Andrieux et al. 2002). STOP knockout (KO) mice exhibit anatomical abnormalities in the brain (Powell et al. 2007), probable hypoglutamatergia (Andrieux et al. 2002; Brenner et al. 2007) and hyperdopaminergia in the limbic system (Brun et al. 2005; Bouvrais-Veret et al. 2008; Hanaya et al. 2008). At a behavioral level, STOP KO mice display hyperactivity, fragmentation of spontaneous activity, hypersensitivity to novelty, to mild stress and to locomotor effects of psychostimulants (Andrieux et al. 2002, 2006; Brun et al. 2005; Bouvrais-Veret et al. 2007, 2008). STOP KO mice exhibit deficits in sensorimotor gating (Fradley et al. 2005), maternal behavior and social interaction (Andrieux et al. 2002). They also display reduced learning and memory performance (Fradley et al. 2005; Bouvrais-Veret et al. 2007; Powell et al. 2007; Begou et al. 2008; Delotterie et al. 2010; Kajitani et al. 2010). Interestingly, some of these alterations are only manifested after puberty (Begou et al. 2007), mimicking the progression of symptoms in schizophrenic patients (Lieberman et al. 2001). Importantly, parts of these dysfunctions are improved by chronic treatment with typical (Andrieux et al. 2002; Brun et al. 2005) or atypical antipsychotics (Delotterie et al. 2010; Merenlender-Wagner et al. 2010), consistent with the hypothesis that STOP KO mice represent a pertinent model for some schizophrenia-like features (Desbonnet et al. 2009).

Interestingly, we recently reported that the deletion of the ubiquitous MAP6/STOP protein dramatically affects central serotonin (5-HT) neurotransmission, with a marked imbalance of 5-HT tone in brainstem versus forebrain areas (Fournet et al. 2010). Some 5-HT key-proteins are strikingly accumulated at the level of the 5-HT cell bodies in raphe nuclei and highly depleted in projection areas of serotonin neurons of mutant mice. The significant correlation observed between decreases in the density of the serotonin transporter (SERT) and 5-HT fiber lengths suggests an impaired axonal transport of SERT and/or development of 5-HT fibers in STOP KO mice. Furthermore, regional imbalance of 5-HT tone in mutant mice appears to be associated with a deficit in adult hippocampal neurogenesis and altered behaviors in anxiety- and depression-related paradigms (Fournet et al. 2010). Altogether, these data suggest that STOP KO mice might represent a pertinent experimental model for schizoaffective disorders. Indeed, recent literature data are inconsistent with the dichotomy existing between schizophrenia and mood disorders (Craddock and Owen 2005; Maier et al. 2006; Carpenter et al. 2009). For example, depression and mania are commonly detected in patients with schizophrenia, and suicide constitutes their first cause of mortality (Radomsky et al. 1999). Conversely, many depressive patients suffer from hallucinations (Murray et al. 2004). Moreover, epidemiological data show familiar co-aggregation of schizophrenia and depression (Cardno et al. 1999). Furthermore, genetic studies have reported several susceptibility genes common to psychiatric diseases, including those encoding the proteins DISC1, dysbindin-1, neuregulin 1 and catechol-O-methyltransferase (Craddock et al. 2006; Porteous and Millar 2006). In addition, distinct DISC1 mutations in mice induce phenotype related to either schizophrenia or depression (Clapcote et al. 2007).

To date, only contradictory and/or scattered data have been reported on anxiety and mood status of STOP KO mice. For example, mutant mice display a high degree of anxiety in the light/dark box test (Andrieux et al. 2002), but very little in the marble burying test (Delotterie et al. 2010), the elevated plus maze and open field tests (Fournet et al. 2010). These contradictory and heterogeneous data could be due to the fact that the above different studies were conducted on mice featuring a noninbred BALBc/129 SvPas-F2 genetic background.

In the present study, the behavioral performances of STOP KO mice on a homogeneous inbred C57BL6/129 SvPas-F1 genetic background in depression- and anxiety-related paradigms were carefully investigated. In addition, anatomical abnormalities and impaired neuronal plasticity previously reported in mutant mice led us to examine cognitive performances of mice on this new genetic background. The tissue levels and in vivo synthesis of 5-HT, norepinephrine (NE) and dopamine (DA) and the densities and activities of the 5-HT, NE, DA and monoamine transporters in various brain areas were also measured. Finally, the responses of these mice towards acute antidepressants of pure or mixed selectivity towards the 5-HT and NE transporters were investigated.

Materials and methods

Animals

STOP KO mice were generated as previously reported (Andrieux et al. 2002; Fournet et al. 2010). Homozygous wild-type (WT) and STOP KO mice were obtained by crossing heterozygous C57BL6 STOP with heterozygous 129 SvPas STOP to get inbred C57BL6 × 129 SvPas-F1 mice. Mice, genotyped as previously described (Andrieux et al. 2002), were kept under standard conditions, under a 12-h light/dark cycle (lights on at 07h30) and habituated to the animal holding room for at least one week prior to use. All experiments were conducted on WT and STOP KO females and/or males of the same litters, at 3–5 months of age, in accordance with the European Communities Council directive (86/809/EEC).

Drugs

Fluoxetine hydrochloride and 3-hydroxy-benzylhydrazine (NSD 1015) were purchased from Sigma-Aldrich (Saint Quentin-Fallavier, France). Desipramine hydrochloride, norepinephrine hydrochloride, reboxetine hydrochloride, serotonin hydrochloride and venlafaxine hydrochloride were from Tocris (Bristol, UK). Polyclonal VMAT2 antiserum was from Chemicon (Temecula, CA, USA). [3H]Citalopram (2.22–3.18 TBq/mmol), [3H]nisoxetine (2.22–3.18 TBq/mmol) and [125I]-IgG (74–370 kBq/μg, 9 kBq/ml) were from Perkin Elmer (Orsay, France).

Measurements of tissue levels and in vivo synthesis rate of 5-HT, NE and DA

The concentrations of endogenous 5-HT, NE, DA and of their precursors were determined by HPLC and electrochemical detection (Fournet et al. 2010; see Appendix S1).

Autoradiographic labelings of SERT, NET, DAT and VMAT2

Labelings of the 5-HT transporter SERT and the NE transporter (NET) were performed as reported (Fournet et al. 2010; Ordway et al. 1997; see Appendix S1). Immunoradiolabelings of the DA transporter (DAT) and of the vesicular monoamine transporter 2 (VMAT2) were conducted as previously described (Bouvrais-Veret et al. 2008; see Appendix S1). Densitometry measurements of numerized screens and films were performed with MCIDTM analysis software.

Uptake of [3H]-5-HT or [3H]-NE on synaptosomes or vesicles

Uptake of [3H]-5-HT or [3H]-NE was performed on synaptosomes prepared from brain stem and from cortex as previously reported (Bouvrais-Veret et al. 2008; see Appendix S1) or on vesicles prepared from cerebral cortex plus hippocampus according to Gras et al. (2008; see Appendix S1).

Behavioral tests

All experiments were conducted between 10:00 and 16:00 hours on females and/or males of both genotypes. With the exception of determination of coat state and locomotor activity, all tests were performed on independent groups of mice. Prior to each experiment, mice were allowed to habituate to the sound attenuated test room for at least 30 min before the task.

Locomotor activity

Horizontal (locomotion) and vertical (rearing) activities were assessed in transparent actimetry cages (20 x 15 x 25 cm), under five lux illumination, with automatic monitoring of photocell beam breaks, located at 1.5 cm (horizontal activity) and 6.5 cm (vertical activity) above the floor (Imetronic, Pessac, France).

Coat state

The coat state of mice of both genotypes and both genders was evaluated by a well-trained experimenter blind to genotypes. Assessment of the coat state took into account the whole body. A score was attributed to each mouse on a scale from 0 (dirty coat) to 20 (bright and clean coat).

Splash test

The splash test was adapted from Yalcin et al. (2008) and consisted of squirting a 10% sucrose solution on the dorsal coat of a mouse in its home cage, under 15 lux illumination. After applying sucrose solution, the latency before the first grooming episode and time spent grooming were recorded for 5 min.

Anhedonia test

Anhedonia was evaluated in mice individually housed in their holding room, using the two-bottle free choice test. Before the test, each mouse had access to two pipettes filled with tap water. When mice drank equally at the two pipettes, tap water of one pipette was replaced by a 4% sucrose solution. To prevent possible effects of side preference in drinking behavior, the positions of the two drinking pipettes were swapped each day for 4 days. No food or water deprivation was applied before the test. Preference for sucrose was expressed as the ratio (%) of sucrose consumption over total liquid (sucrose plus water) consumption.

Tail suspension test

Mice were suspended by the tail, using a paper adhesive tape, to a hook in a chamber of the apparatus (Bioseb, Vitrolles, France) under a 15 lux illumination. Their immobility time was automatically recorded during 6 min. To test the sensitivity of mice towards acute antidepressants, animals received intraperitoneally saline, fluoxetine, venlafaxine or reboxetine at various doses (100 μL/10 g body weight) and were tested 30 min later for their immobility.

Forced swimming test

The forced swimming test, adapted from Porsolt et al. (1979), was performed in a vertical glass cylinder (h = 30 cm; d = 15 cm) containing 20 cm of water maintained at 24 ± 1°C, under 15 lux illumination. Latency before the first episode of immobility, the total duration of immobility and the number of climbing attempts were recorded for 6 min. An animal was judged to be immobile when it remained floating passively, performing only slow motion to keep its head above the water. Climbing attempts were defined by upward-directed movements of the forepaws along the side of the container.

Elevated plus maze test

The elevated plus maze test was conducted under a 50 lux illumination, as reported (Fournet et al. 2010; see Appendix S1). The number of visits and the time spent in the arms, as well as the number of head-dipping (risk) were measured for 5 min.

Open field test

The open field test was performed in an arena (100 x 100 x 30 cm) with a central square (60 x 60 cm), located in a 100 lux illuminated room. Mice were introduced into a corner of the arena and allowed to freely explore the open field for 9 min. The total traveled distance in the arena, as well as the time spent, the traveled distance and the number of entries of mice with their four paws in the central square were measured.

Light/dark box test

The apparatus consisted of a box (50 x 30 x 30 cm) divided by an open door providing access to a white illuminated open area (300 lux) and a dark black enclosed area (5 lux). Mice were placed in the center of the dark area and latency to enter into the bright area, the number of visits (with four paws) and the total time spent in the bright area were measured for 9 min.

Marble burying test

The procedure for the marble burying test was adapted from Millan et al. (2001) with minor modifications. Mice were individually placed in transparent cages containing a 4 cm layer of sawdust and 12 identical glass marbles (d = 1.6 cm) evenly spaced throughout the cage (four rows of three marbles), under a 50 lux illumination. The number of marbles buried by each mouse in more than two-thirds of the sawdust was scored every minute for 10 min and then every 5 min up to 30 min.

Spontaneous alternation

This test was performed under 5 lux illumination in a Y-maze (details in Appendix S1). The number of entries in arms and the order of visits into the 3 arms were recorded during 5 min.

Novel object recognition task

This test was conducted in a vast arena under a 50 lux illumination (details in Appendix S1). After habituation to the arena and objects, each mouse was placed in the arena for 8 min in the presence of four identical new objects (sample phase). Mice were then removed and, after 10 min, 4 or 24 h, returned to the arena for 8 min (choice phase), with two objects from the sample phase and two novel identical objects. The time spent exploring familiar and novel objects was recorded.

Spatial Morris watermaze test

The spatial version of the Morris watermaze test was conducted as already reported (Weiss et al. 2007; see Appendix S1) in a pool filled with opaque water, under a 150 lux illumination. Mice learned the fixed position of the submerged platform using distal extra-maze cues. The time spent and the distance traveled to find the platform, the swimming speed and the proportion of successful trials were measured.

Statistical analyses

Data (means ± SEM) were subjected to factorial one-, two-, three-, or four-way anova, with genotype, sex, area, or treatment as between-group factors and time as within-group factor, using Statview analysis software. Significant main effects were further analyzed by post hoc comparisons of means using Fisher’s or Student’s t-test. The percentages of successful trials in the Morris watermaze test were compared using the Kolmogorov–Smirnov’s test. The kinetic parameters (Kd and Vmax) of 5-HT uptake by synaptosomes were determined by non-linear regression using GraphPad prism 5.0 software. The means ± SEM were then compared using Student’s t-test. For all tests, statistical significance was set at p < 0.05.

Results

The results of the present study were all obtained using WT and STOP KO mice on the homogeneous inbred C57BL6/129 SvPas-F1 genetic background. Accordingly, some characteristics of mutant mice on the former genetic background (BALBc/129 SvPas-F2) were determined on this new genetic background. When statistical analyses showed a significant effect of gender, it generally reflected a difference in the degree of variation between females and males. These gender effects are shown in Tables S1 and S2 and are not discussed further, with the exception of the few cases in which no effect was found between the two genders (e.g. for body weight, locomotor activity and coat state).

Body weights and locomotor activity of STOP KO mice

Up to 6–8 weeks of age, the body weight of STOP KO mice was significantly smaller compared with WT mice. Subsequently, the weight of STOP KO females was comparable to that of WT females, whereas the weight of mutant males remained significantly lower than that of WT males up to 20–24 weeks of age. In older males, no difference in body weight was observed between mutant and WT mice (Table S1 and Figure S1).

As already reported for STOP KO males on BALBc/129 SvPas-F2 genetic background, STOP KO mice were hyperactive compared with WT mice, displaying increased horizontal and vertical locomotor activities (Table S1 and Figure S2).

Depression-like behavior of STOP KO compared with WT mice

The depression-like status of mice of both genotypes was evaluated by means of five different tests, using females and males in equal proportion (Fig. 1).

Figure 1.

 STOP KO mice were more depressed than WT mice. (a) Coat state. Data represent the means ± SEM of scores of 14 females/14 males WT and 15 females/17 males STOP KO, over a 0–20 scale. ***p < 0.001, comparison between genotypes; ###p < 0.001, comparison between genders. (b) Splash test. Means ± SEM of the latency time before the first grooming and of the grooming duration for 12 females/14 males WT and 14 females/14 males STOP KO. ***p < 0.001, comparison between genotypes. (c) Anhedonia. Means ± SEM of the sucrose preference (% sucrose/total fluid consumption) over 4 days of 7 females/8 males WT or STOP KO. **p < 0.010 (repeated measures). (d) Tail suspension test. Means ± SEM of the immobility duration of 5 females/5 males WT and 5 females/6 males STOP KO. *p < 0.050, comparison between genotypes. (e) Forced swimming test. Means ± SEM of immobility duration, number of climbing attempts and latency to immobilize for 12 females/11 males WT and 13 females/13 males STOP KO. *p < 0.050; ***p < 0.001, comparison between genotypes.

Grooming behavior

A decreased grooming behavior is suggested to parallel careless appearance frequently observed in depressed patients (Willner 1991). In the present study, a decreased careless behavior of STOP KO mice compared with WT littermates was demonstrated.

Natural grooming was first assayed by careful examination of the coat state (Fig. 1a). Statistical analyses showed significant effects of gender and genotype (Table S1). The coat state did not differ between WT females and males, but was significantly improved in STOP KO females than in males (+ 53%, p < 0.0001). Moreover, coat state was no different between WT and STOP KO females, but was significantly worse in STOP KO than in WT males (−39%, p < 0.0001).

Subsequently, grooming abilities were measured directly using the splash test (Fig. 1b). Statistical analyses showed significant effects of both genotype and gender on latency and grooming (Table S1). The grooming behavior of STOP KO mice was impaired compared with that of WT mice: latency before the first grooming episode was increased (+228%, p < 0.0001) and the time spent grooming was significantly decreased (−34%, p < 0.0001) versus WT mice.

Anhedonia

In mice, a decreased consumption of palatable food or drink is interpreted as anhedonia and can be considered as a symptom of depression (Willner 1991).

Statistical analyses of total fluid consumption showed significant effects of genotype, time and of the time × gender interaction (Table S1). Over the 4-day assay, the mean total fluid consumed by STOP KO mice was decreased by 28% (p < 0.0001, repeated measures) compared with WT mice (not shown).

Statistical analyses of sucrose preference showed significant effects of genotype and gender (Table S1). The mean sucrose preference (ratio between sucrose consumption and total liquid consumption) of STOP KO versus WT mice was significantly decreased by 25% (p = 0.0063, repeated measures, Fig. 1c). The decreased preference of STOP KO mice for sucrose was probably not due to a defect in their taste and/or smell. Indeed, (i) in the task of olfactory discrimination, both mutant and WT mice displayed a distinct preference for the odor previously associated with sugar, 1 day after the 4-day training period; (ii) following food deprivation, mice of both genotypes actively searched for, found and ate the buried Froot Loop (Powell et al. 2007). Our data therefore likely suggest a decreased sensation of pleasure in mutant compared with WT mice.

Despair behaviors

The tail suspension and forced swimming tests represent two standardized paradigms for the assessment of despair behaviors, by analyzing immobility scores in inescapable aversive situations (Willner 1991). Our results clearly indicated an increased despair behavior in STOP KO mice compared with WT mice.

In the tail suspension test (Fig. 1d), statistical analysis of the immobility time showed significant effect of genotype (Table S1). The immobility of STOP KO mice was increased by 55% (p = 0.0151) compared with WT mice.

In the forced swimming test (Fig. 1e), statistical analyses indicated (i) significant effects of genotype, time and genotype × gender interaction on immobility time, (ii) significant effects of genotype and time on the number of climbing attempts and (iii) a significant effect of genotype on the latency before the first immobility episode (Table S1).

STOP KO mice displayed a higher despair behavior than WT littermates, characterized by an increased immobility (+15%, p = 0.0127; repeated measures), a highly decreased number of climbing attempts (−69%, p < 0.0001; repeated measures) and a decreased latency before the first immobility episode (−85%, p < 0.0001).

Overall data indicated that STOP KO mice exhibited a depression-like status, compared with their WT littermates.

Anxiety-status of STOP KO compared with WT mice

The anxiety-status of STOP KO and WT mice was assessed thanks to four different tests, using females and males in equal quantity (Fig. 2). First, we assessed generalized anxiety by means of the elevated plus maze, open field and light/dark box tests, all based on a conflict between the natural tendency of mice to actively explore a new environment and their aversion for open space and light.

Figure 2.

 STOP KO mice were less anxious than WT mice. (a) Elevated plus maze. Means ± SEM of the % time spent and of the % number of visits into the open arms, and of the number of head-dipping (risk) of 12 females/14 males WT and 14 females/11 males STOP KO. ***p < 0.001, comparison between genotypes. (b) Open field. Means ± SEM of the time spent, the number of visits and the % distance traveled in the central square (C) for 7 females/11 males WT and 10 females/8 males STOP KO. *p < 0.050, comparison between genotypes. (c) Light/dark box. Means ± SEM of the time spent and of the number of visits into the light box (L) and of the latency before the first visit for 13 females/13 males WT and 13 females/9 males STOP KO. **p < 0.010, ***p < 0.001, comparison between genotypes. (d) Marble burying test. Means ± SEM of the number of marbles buried by 14 females/15 males WT and 11 females/17 males STOP KO. ***p < 0.001, repeated measures, comparison between genotypes.

In the elevated plus maze test (Fig. 2a), statistical analyses showed no effect of factors on the total time spent in the four arms, but significant effects of genotype and gender on the number of entries in the four arms, a significant effect of genotype on the % time spent and on the % entries into the open arms and significant effects of genotype and gender on the risk-taking (Table S1).

STOP KO mice showed a higher exploratory activity than WT mice, because their number of entries in the four arms was increased (+29%, p = 0.0139, Figure S2b). Mutant mice also displayed a lower anxiety-status, characterized by an increase of the % time spent (×4.32, p < 0.0001) and of the % entries (×2.75 p < 0.0001) into the open arms and by an increase of the risk-taking (number of head-dipping; +292%, p < 0.0001).

In the open field test (Fig. 2b), which corresponds to an inescapable situation, statistical analyses showed a significant effect of genotype on the total distance traveled, on the time spent, the number of visits and the % distance traveled into the central square (Table S1).

STOP KO mice were hyperactive, traveling greater distances in the open field than their WT littermates (+74%, p = 0.0012, Figure S2c). STOP KO mice also displayed a lower anxiety than WT mice, characterized by an increased time spent (+324%, p = 0.0218), number of visits (+178%, p = 0.0358) and % distance traveled (2.2-fold, p = 0.0443) in the central square.

In the light/dark box test (Fig. 2c), statistical analyses indicated significant effects of genotype and gender on the time spent in the light box, of genotype on the number of entries into the light box and on the latency before the first visit in the light box (Table S1).

STOP KO mice also displayed a lower anxiety than WT littermates, characterized by an increased time spent (+88%, p < 0.0001) and number of visits in the light box (+128%, p < 0.0001) and a decreased latency before the first visit to the light box (−48%, p = 0.0073).

In the marble burying test, mice bury objects viewed as dangerous. Statistical analyses showed significant effects of genotype and time on the number of buried marbles (Table S1). The number of marbles buried by STOP KO mice was decreased by 65% (p < 0.0001, repeated measures, Fig. 2d) over the 30-min session, compared with WT mice, suggesting that STOP KO mice did not consider marbles as anxiogenous objects.

Altogether, these data showed a reduced anxiety status of STOP KO compared with WT mice, in both escapable and inescapable situations.

Cognitive performances of STOP KO mice

Previous studies demonstrating anatomical and neuronal plasticity defects in STOP KO mice (Andrieux et al. 2002; Powell et al. 2007; Fournet et al. 2010) led us to question whether cognitive performance could be affected in STOP KO versus WT males, by means of three different tests (Fig. 3). First, the memory abilities of mice were determined using the spontaneous alternation test in the Y-maze (Kokkinidis and Anisman 1976), which assays very short-term memory, and in the object recognition task, which assays short- and long-term memories (Kokkinidis and Anisman 1976; Bevins and Besheer 2006).

Figure 3.

 STOP KO mice exhibited impaired memories. (a) Spontaneous alternation. Means ± SEM of the number of entries into the three arms and of the % spontaneous alternation of 9 WT and 10 STOP KO males. *p < 0.050, comparison between genotypes. (b) Object recognition test. Means ± SEM of the % time spent to explore the novel (N) and the familiar (F) objects, after an interval of 10 min, 4 h and 24 h between the sample- and choice-sessions, by 9–12 WT and 13 STOP KO males. Student’s t-test: **p < 0.010, ***p < 0.001, comparison between objects; #p = 0.050, ###p < 0.001, comparison between genotypes. (c) Morris watermaze-spatial version. Means ± SEM of the time spent and the distance swum to find the submerged platform by 10 WT and 10 STOP KO males over 5-day training. *p < 0.050, ***p < 0.001 (repeated measures). Bottom: Quadrant test. Means ± SEM of the % time spent in each quadrant by 10 WT and 10 STOP KO males. N: target quadrant. Student’s t-test: **p < 0.010, ***p < 0.001, comparison with the target quadrant; #p < 0.050, comparison between genotypes.

In the Y-maze test, analyses of data showed a significant effect of genotype on the total number of entries into the three arms of the Y-maze (Table S2), increased by 43% (p = 0.0186, Fig. 3a) in mutant mice. However, the spontaneous alternation of STOP KO mice did not differ from WT mice, indicating that their very short-term memory was preserved.

In the novel object recognition task, assays were performed at time intervals of 10 min, 4 and 24 h between sample- and choice-tests (Fig. 3b). Analyses showed a significant effect of genotype on the total time spent in exploring the four objects in all pre-tests and tests (Table S2), which was increased by +100–150% (p < 0.005, see Figure S2 for an example) in STOP KO mice. Moreover, the % preference of mutant mice for the novel objects was invariably 50%, in contrast to scores obtained by WT mice (p < 0.0001 after 10 min; p = 0.0045 after 4 h and p = 0.0049 after 24 h, Student’s t-test).

The above data demonstrated how short- and long-term memories of STOP KO mice were severely impaired and that poor performances were not due to a decreased exploratory behavior.

Finally, spatial memory was assessed using the Morris watermaze test (Fig. 3c). Statistical analyses showed significant effects of genotype and time on the time spent to reach the submerged platform and on swimming speed, as well as a significant effect of the time × genotype interaction on the distance covered to find the platform (Table S2).

WT mice displayed improved performance in reaching the platform between day 1 and day 5 (time: −25%, p < 0.0001; distance: −24%, p = 0.0028), indicating learning of the task. In contrast, the performance of STOP KO mice remained unchanged throughout the trials. The swimming speed of STOP KO mice was decreased by 5% (p = 0.0259, repeated measures) throughout the duration of the 5-day test, compared with WT mice (Figure S2). Furthermore, in contrast with WT mice which swam at a constant speed over the 5-day session, the speed of mutant mice slightly but significantly increased by 13% from day 1 to 5 (p = 0.0173). Finally, the proportion of successful trials was significantly lower for STOP KO mice (χ= 6.8, p < 0.010, not shown).

To characterize the spatial strategy of mice, the quadrant test was performed on the sixth day. This tests revealed how WT mice alone spent more time (Fig. 3c) and swam greater distances (not shown) in the target quadrant previously containing the platform compared with the three other quadrants (Student’s t-test: time : +17–19%, p = 0.0036–0.0004; distance: +13–18%, p = 0.0088–0.0004).

Taken together, the data obtained indicated that spatial learning skills and memory of STOP KO mice were highly impaired.

Tissue levels and in vivo synthesis of 5-HT, NE and DA

The depressive- and anxious-like status of STOP KO mice, and particularly their highly reduced climbing attempts in the forced swimming test (deemed to reflect a NE-related component, Detke et al. 1995), prompted us to measure tissue levels of three amines, 5-HT, NE and DA, in various areas of male WT and STOP KO mice (Table 1). We also measured the in vivo tryptophan hydroxylase 2 (TPH2) and tyrosine hydroxylase (TH) activities, respectively reflected by 5-HTP and l-DOPA accumulation after l-aromatic aminoacid decarboxylase inhibition. Statistical analyses of data showed significant effects of area and genotype on all studied parameters (Table S2 and Table 1).

Table 1.   High and parallel variations in tissue levels and in vivo synthesis of 5-HT and NE in STOP KO mice
AreaGenotype5-HTTPH2NEDATH
  1. Serotonin (5-HT), norepinephrine (NE) and dopamine (DA) tissue levels are expressed as means ± SEM in μg/g fresh tissue. For the determination of in vivo tryptophan hydroxylase 2 (TPH2) and tyrosine hydroxylase (TH) activities (means ± SEM in ng/g/30 min), mice received saline or 100 mg/kg NSD 1015, 30 min before killing. Respective saline 5-HTP and l-DOPA levels were subtracted from accumulated 5-HTP and l-DOPA. Number of male mice in parentheses. Ant- and Post-Cx, anterior and posterior cortices; BS, brain stem; CPu, caudate-putamen; Hipp, hippocampus; SN + VTA, substantia nigra and ventral tegmental area. Fisher’s test: ns, non-significant; #p = 0.063; *p < 0.050; **p < 0.010; ***p < 0.001, comparison between genotypes.

BSWT(7) 0.25 ± 0.02(7) 4.37 ± 0.39(8) 0.26 ± 0.02(5) 0.05 ± 0.02(8) 3.25 ± 0.55
KO(8) 0 35 ± 0.03(7) 5.53 ± 0.65(8) 0.40 ± 0.02(5) 0.04 ± 0.01(8) 2.84 ± 0.40
%+38%*+27%ns+54%***−5%ns−12%ns
SN + VTAWT(6) 0.86 ± 0.06(8) 12.0 ± 1.07(6) 0.50 ± 0.03(5) 0.37 ± 0.06(8) 13.4 ± 1.45
KO(8) 1.36 ± 0.09(6) 22.0 ± 1.18(8) 0.74 ± 0.04(8) 0.47 ± 0.06(6) 13.7 ± 1.11
%+58%***+83%***+46%***+28%ns+3%ns
HippWT(7) 0.54 ± 0.03(8) 6.95 ± 0.40(7) 0.30 ± 0.01(5) 0.04 ± 0.01(8) 4.16 ± 0.31
KO(8) 0.25 ± 0.03(8) 2.58 ± 0.28(8) 0.09 ± 0.01(5) 0.03 ± 0.01(8) 2.79 + 0.29
%−53%***−63%***−70%***−4%ns−31%**
CPuWT(8) 0.49 ± 0.02(7) 5.25 ± 0.36(8) 0.05 ± 0.007(7) 14.33 ± 0.37(7) 64.3 ± 4.46
KO(8) 0.38 ± 0.03(8) 5.42 ± 0.33(7) 0.03 ± 0.003(7) 11.27 ± 0.52(8) 56.7 ± 2.73
%−24%**+3%ns−36%*−21%***−12%ns
Post-CxWT(7) 0.33 ± 0.02(8) 3.51 ± 0.33(7) 0.24 ± 0.01(7) 0.25 ± 0.03(8) 7.00 ± 0.67
KO(7) 0.16 ± 0.01(8) 1.56 ± 0.16(7) 0.11 ± 0.01(7) 0.16 ± 0.03(7) 5.36 ± 0.55
%−52%***−56%***−54%***−35% #−24%ns
Ant-CxWT(8) 0.46 ± 0.02(7) 4.30 ± 0.20(8) 0.19 ± 0.01(7) 1.23 ± 0.04(7) 17.1 ± 1.78
KO(7) 0.28 ± 0.011(8) 3.40 ± 0.27(6) 0.10 ± 0.02(7) 1.16 ± 0.11(8) 13.4 ± 1.43
%−40%***−21%*−45%***−6%ns−23%ns

In STOP KO versus WT mice, 5-HT levels were significantly increased by 40–60% in the brain stem containing the raphe nuclei and in the substantia nigra plus the ventral tegmental area, but significantly decreased by 25–50% in terminal areas, in agreement with previous data published by our laboratory on mutants of the BALBc/129 SvPas-F2 genetic background (Fournet et al. 2010). NE levels were significantly increased by 50% in the brainstem and in the substantia nigra plus the ventral tegmental area, but significantly decreased by 40–70% in the four terminal areas. In contrast, DA levels were not modified, with the exception of a slight reduction of 21% in the caudate-putamen, in line with previous findings (Bouvrais-Veret et al. 2008). Furthermore, variations observed in 5-HT and NE levels in STOP KO mouse areas were highly correlated [r2 = 0.9580, F(1,4) = 91.15, p = 0.0007, not shown], whereas variations between 5-HT and DA levels displayed no correlation.

A trend toward increase and a significant 80% increase in the in vivo synthesis of 5-HT (TPH2 activity) were found in the brain stem and in the substantia nigra plus the ventral tegmental area of STOP KO mice, respectively. In contrast, TPH2 activity was significantly decreased by 20–60% in the hippocampus and in the posterior and anterior cortices, but not modified in the caudate-putamen, according to previous data from our group (Fournet et al. 2010). Finally, the in vivo synthesis of catecholamines (TH) was significantly decreased by 30% in the hippocampus, but remained unchanged in all other areas.

Densities of SERT, NET, DAT and VMAT2 in brain areas

Subsequently, the relative densities of the three monoamine transporters and of the vesicular monoamine transporter 2 were measured in various cerebral areas of WT and STOP KO mice (Table 2, Figures S3 and S4). Statistical analyses indicated significant effects of genotype and area on SERT, NET and VMAT2 densities and a significant effect of area on DAT densities (Table S2).

Table 2.   Parallel variations of SERT and NET densities in STOP KO brain areas
Coronal levelAreaSERT (%)NET (%)DAT (%)VMAT2(%)
  1. Values are expressed as percent variations over respective WT values (5–6 mice for each transporter and each genotype). IA, interaural, coronal level according to Franklin and Paxinos (1997). Acc, nucleus accumbens; BLA or BMA, basolateral or basomedial amygdala; Cg Cx, cingulate cortex; CPu, caudate-putamen; DAT, dopamine transporter; DR, dorsal raphe nucleus; DRI, dorsal raphe intermediate; Hipp, hippocampus; IP, interpeduncular nucleus; LC, locus coeruleus; l or mSept, lateral or medial septum; MEnt Cx, medial entorhinal cortex; MnR, median raphe nucleus; Mot Cx, motor cortex; NET, norepinephrine transporter; PrL Cx, prelimbic cortex; RS Cx, retrosplenial cortex; Sens Cx, somatosensory cortex; SERT, serotonin transporter; SN, substantia nigra; Vis Cx, visual cortex; VMAT2, vesicular monoamine transporter 2; VTA, ventral tegmental area. Two-way anova followed by Student’s t-test: #p = 0.056; *p < 0.050; **p < 0.010; ***p < 0.001.

Locus coeruleus
IA = −1.72 to −1.54
DRI+93***+39**  
LC+64*−10ns  
Raphe
IA = 0.80 to −0.44
RS Cx−83***−74***−2ns−11ns
DR+71***+62***+2ns−3ns
MnR+86***+110***−10ns−4ns
IP+86**+40* −15ns
MEnt Cx−85***−51***−4ns−18***
Substantia nigra
IA = −0.08 to 0.88
RS Cx−82***−71***+11ns−6ns
Vis Cx−83***−75***+7ns 
Hipp−68***−56***+3ns−16*
SN+59***−16ns+1ns+11ns
VTA+75***+34*−3ns+2ns
IP+84*+31* +12ns
MEnt Cx−75***−51***−1ns−26**
HippocampusRS Cx−63***−51***+6ns−1ns
IA = 1.98 to 2.74Mot Cx−68***−52***+17ns−5ns
Sens Cx−60**−50***+12ns−9ns
Hipp−33*−57***+5ns−6ns
BLA−5ns−17ns+3ns−10*
BMA−15ns−11ns −12*
Striatum
IA = 4.78 to 5.34
Cg Cx−69***−40***0−8ns
Mot Cx−59***−26*+14ns0
Sens Cx−42***−31**+21ns 
CPu−57***−24***−4ns−5ns
Acc−36*−24**+1ns−10#
lSept−9ns−24* −10ns
mSept−26ns−26** −17*
Prefrontal cortex
IA = 5.78 to 6.38
Cg Cx−73***−49***  
PrL Cx−71***−42***  
Mot Cx−49***−37**  

In STOP KO mice, parallel marked variations of SERT and NET density were noted (Table 2). Compared with WT mice, mutants displayed SERT and NET densities 60–110% higher in the dorsal and median raphe nuclei, the substantia nigra (except NET) and the ventral tegmental area, but 30–85% lower in all the other areas. In contrast, DAT density did not differ significantly in any of the studied areas in STOP KO versus WT mice. Finally, VMAT2 density was significantly decreased by 10–25% in some areas of mutant mice, namely the hippocampus and medial entorhinal cortex (Table 2). The results on SERT and DAT densities were in agreement with data previously found in mice on the first BALBc/129 SvPas-F2 genetic background (Bouvrais-Veret et al. 2008). In contrast, the decreases of VMAT2 density were not previously found in mice on the first genetic background (Bouvrais-Veret et al. 2008), probably because of the use of a less selective antibody.

In STOP KO mice, the variations of SERT were correlated with both 5-HT fiber length (Figure S4a) and variations of NET density (Figure S4b). In the latter case, the slope of the linear regression higher from 1 (p = 1.473 ± 0.1048) suggests that deletion of STOP elicits a greater imbalance of 5-HT than of NE tone. Finally, the observation that NET density did not vary in STOP KO mice in the locus coeruleus, which contains NE somas, nor in the substantia nigra, which contains DA somas, remains intriguing.

The variations of SERT, NET and VMAT2 density were functional

5-HT and NE uptakes were measured on synaptosomes from the brain stem, which contains the 5-HT and DA somas, and from the posterior cortex (Fig. 4).

Figure 4.

 The variations of SERT, NET and VMAT2 densities in STOP KO mice were functional. (a) Uptake of serotonin (5-HT) by synaptosomes prepared from the brain stem and the posterior cortex of WT and STOP KO mice. Means ± SEM of data from six pools for each genotype, in two independent experiments. (b) Uptake of norepinephrine (NE) by synaptosomes prepared from the brain stem and the posterior cortex of WT and STOP KO mice. Means ± SEM of data from three pools for each genotype, performed in triplicate. (c) Uptake of 5-HT by synaptic vesicles prepared from cortex plus hippocampus. Means ± SEM of data from six pools for each genotype, in two independent experiments. Student’s t-test: *p < 0.050; **p < 0.001, comparison between genotypes.

The kinetic parameters of the 5-HT uptake (Fig. 4a) indicated that the maximal rate was increased by 50 % in the brain stem of STOP KO mice (1683 ± 37 vs. 2528 ± 62 fmol/mg prot/min, p < 0.001, in WT and STOP KO mice, respectively). In contrast, maximal rate was decreased by 66% in the posterior cortex of these mice (486 ± 12 vs. 163 ± 5 fmol/mg prot/min, p < 0.001, in WT and STOP KO mice, respectively). However, 5-HT affinities for SERT were not modified (brain stem: 48.4 ± 3.2 vs. 48.5 ± 3.6 nM; cortex: 29.0 ± 2.6 vs. 29.1 ± 3.5 nM, in WT and STOP KO mice, respectively). Interestingly, the Kd value for 5-HT was significantly 2-fold lesser (p < 0.001) in the cortex than in the brain stem, in both WT and STOP KO mice. This increased affinity (decreased Kd) of cortical SERT for 5-HT may be the consequence of its association with chaperone proteins different from that found near DA and 5-HT somas.

In the same way, NE uptake (Fig. 4b) was significantly increased by 8085% (p < 0.01) in synaptosomes from the brain stem and decreased by 4080% (p < 0.05) in cortical synaptosomes.

Finally, to measure the activity of VMAT2, we prepared synaptic vesicles from total brain minus the brain stem and cerebellum, rich in white matter, and minus the striatum, rich in DA terminals. The 5-HT uptake by synaptic vesicles was significantly decreased by 4050% (p < 0.05, Fig. 4c) at the two 5-HT concentrations tested. This result suggested that vesicular stores of both 5-HT and NE could be lower and consequently less released in the synaptic space in STOP KO than in WT mice.

Altogether, these data confirmed the regional 5-HT imbalance already found in STOP KO mice on the first genetic background (Fournet et al. 2010) and suggested that the NE tone was parallely affected. Moreover, the important variations in the density of SERT, NET and VMAT2 had functional consequences.

Sensitivity of STOP KO mice towards acute antidepressants

The sensitivity of WT and STOP KO mice towards acute antidepressants was investigated using the tail suspension test (Fig. 5). Thirty minutes before the test, we administered saline or fluoxetine, a selective inhibitor of SERT, venlafaxine, a mixed inhibitor of SERT and NET and reboxetine, a selective inhibitor of NET. Statistical analyses showed significant effects of genotype, gender and treatment on immobility after antidepressant administration (Table S2).

Figure 5.

 STOP KO mice were hypersensitive to acute antidepressant effect in the tail suspension test. Means ± SEM of the % inhibition of immobility by fluoxetine, venlafaxine and reboxetine of 3–5 females and 3–5 males per genotype, per treatment and per dose. Dashed lines indicated the 50% inhibition by antidepressants. Post hoc Fisher’s test: *p < 0.050, ***p < 0.001, comparison between genotypes, #p < 0.050, ###p < 0.001, comparison between saline and antidepressant dose.

STOP KO mice were hypersensitive to acute antidepressant administration, whatever their selectivity. Indeed, low doses of the three antidepressants elicited a significant higher inhibitory effect on the immobility of STOP KO than that of WT mice (10 mg/kg fluoxetine: 76% vs. 31%, p = 0.0102; 5 mg/kg venlafaxine: 75% vs. 31%, p = 0.0182; 2 mg/kg reboxetine: 69% vs. 7%, p = 0.0001, Fig. 5).

Discussion

The data obtained in the present study demonstrated how STOP KO mice on the new inbred C57BL6/129 SvPas-F1 genetic background clearly exhibited depressive-like phenotype, reduced anxious-like phenotype and hypersensitivity to serotonin and/or norepinephrine reuptake inhibitor acute treatments. Mutant mice also displayed preserved very short-term memory, but impaired short- and long-term memories, as well as spatial learning skills and memory. In addition, 5-HT and NE tissue levels were parallely increased in monoaminergic cell body areas and depleted in all forebrain projection areas. In comparison, DA tissue levels were only marginally decreased in some projection areas. The majority of these characteristics had been previously found in BALBc/129 SvPas-F2 STOP KO mice (see Introduction for references), indicating that these phenotypes were strong features which did not depend on the two genetic backgrounds studied. Nevertheless, the generation of STOP KO mice on the new inbred F1-genetic background yielded a more homogeneous and clearcut phenotype. We hypothesize that behavioral alterations observed in STOP KO mice were at least partly linked to impaired monoaminergic neurotransmissions induced by STOP deletion (Brun et al. 2005; Bouvrais-Veret et al. 2008; Fournet et al. 2010; the present study).

Parallely altered 5-HT and NE tone in STOP KO mice

Here, we showed that 5-HT and NE tissue levels were parallely increased in the raphe nuclei, the substantia nigra and the ventral tegmental area and decreased in all studied forebrain projection areas in STOP KO versus WT mice. Such changes, associated with marked parallel variations in SERT and NET densities, may elicit strong but opposite consequences on 5-HT and NE extracellular levels in midbrain versus forebrain areas. Indeed, mice over-expressing SERT showed increased 5-HT tissue levels but decreased brain extracellular 5-HT (Jennings et al. 2006), whereas SERT KO mice were reported to have decreased endogenous 5-HT levels in forebrain and markedly increased brain extracellular 5-HT (Fabre et al. 2000; Kim et al. 2005). Accordingly, it was assumed that 5-HT and NE extracellular levels might be decreased in the midbrain where SERT and NET were up-regulated and increased in the forebrain where SERT and NET were down-regulated. However, further studies should be undertaken to verify this hypothesis.

In the substantia nigra plus the ventral tegmental area, containing the DA cell bodies and highly innervated by 5-HT and NE terminals, an accumulation of 5-HT and NE markers was observed. The strong alterations of 5-HT and NE tones in these areas may represent an adaptive response of 5-HT and NE neurons toward abnormal signaling from DAergic cell bodies and/or result from abnormal 5-HT and NE traffic/innervation inducing accumulation of 5-HT and NE markers in these DA areas.

Interestingly, DA tissue levels were only marginally decreased in some projection areas and consistently with the absence of any variation of DAT densities, as already reported on F2 Balbc/SV129 STOP KO mice (Bouvrais-Veret et al. 2008). Finally, as hippocampus contains dense noradrenergic but only sparse dopaminergic innervations (Deltheil et al. 2008), the observed reduction of TH activity is most probably due to NE rather than DA neurotransmission alterations.

The discrepancies observed between 5-HT and NE versus DA neurotransmission alterations may underline a differential, but crucial, role of MAP6 proteins in the establishment of the corresponding networks.

The depressive phenotype of STOP KO mice was probably linked to reduced tissue 5-HT in terminals

The increased depressive-like status of STOP KO mice was established by their concordant performance in five different tests, that is, coat state, splash test, sucrose preference, tail suspension and forced swimming tests, confirming previous sparse observations (Delotterie et al. 2010; Fournet et al. 2010). The role of each monoamine (5-HT, DA and NE) in depressive-like behaviors has been extensively documented, with a major role for 5-HT. Indeed, decreased 5-HT tissue levels in mutant rodent lines are associated with a depressive-like phenotype: both SERT KO mice and rats show increased depressive-related behaviors in various tests (Holmes et al. 2003; Lira et al. 2003; Alexandre et al. 2006; Olivier et al. 2008; Popa et al. 2008), although discrepancies have been reported between studies and in different mouse strains (Kalueff et al. 2010). However, mice with reduced brain monoamines, such as heterozygous VMAT 2 mutants (Fukui et al. 2007) or depleted 5-HT such as TPH2 KO mutants (Beaulieu et al. 2008; Savelieva et al. 2008) also exhibit increased depressive-like behavior. Paradoxically, transient exposure to SERT inhibitors during development produced depressive syndrome in adult mice (Ansorge et al. 2008) and Plasmacytoma Expressed Transcript 1(PET-1) KO, another 5-HT depleted mouse mutant, did not differ from paired WT in the tail suspension and the forced swimming tests (Schaefer et al. 2009). In STOP KO mice, both the endogenous 5-HT depletion in 5-HT terminal areas and the decreased VMAT2 activity might induce the depressive-like status of STOP KO mice. Nevertheless, one cannot totally discard the effects of NE and DA neurotransmission dysfunctions in this phenotype, although DAT and NET KO mice show reduced immobility in despair tests, indicating a reduced helplessness (Spielewoy et al. 2000; Dziedzicka-Wasylewska et al. 2006; Perona et al. 2008).

The reduced anxious phenotype of STOP KO mice might result from increased SERT in somas

STOP KO mice exhibited reduced anxiety-like phenotype in four different tests, that is, elevated plus maze, open field, light/dark box and marble burying tests, extending previous sparse observations (Delotterie et al. 2010; Fournet et al. 2010). A large body of reported data supports the idea that 5-HT neurotransmission may play a key role in anxiety. For example, 5-HT depleted mice such as SERT-over-expressing mice display reduced anxiety (Jennings et al. 2006) and the 5-HT depleted PET-1 KO mice also exhibit reduced anxiety in the elevated plus maze and the light/dark box tests, but not in the marble burying test (Schaefer et al. 2009). Conversely, mice with increased 5-HT extracellular levels, such as SERT KO mice and antidepressant-treated mice, exhibit strong anxiety-like phenotype (Holmes et al. 2003; Lira et al. 2003; Kalueff et al. 2007; Line et al. 2010). In the same line, a strong anxiety has also been observed in MAOA/B KO mice (Chen et al. 2004). In summary, extracellular 5-HT concentration correlates with anxious status, that is, low concentration being associated with less anxiety and vice versa. Thus, the increase of SERT in somas of STOP KO mice might lead to low extracellular concentration of 5-HT and thereby to low anxiety-like status.

Implication of 5-HT receptors in the expression of anxiety has also been extensively explored. For example, 5-HT1A antagonist (Griebel et al. 2000), 5-HT2A/C and 5-HT2B agonists (Masse et al. 2008) and 5-HT3 partial agonist (Delagrange et al. 1999) have shown anxiolytic properties. Mice over-expressing 5-HT1A receptors show decreased anxiety-like behavior (Kusserow et al. 2004), whereas KO of the 5-HT1A receptor elicits an opposite effect (Ase et al. 2001). Interestingly, the forebrain-specific conditional rescue of 5-HT1A gene in 5-HT1A KO mice demonstrated that expression of the 5-HT1A receptor during early postnatal period is crucial for the establishment of normal anxiety-like behavior in the adult (Gross et al. 2002). Finally, recent data underline the crucial role of 5-HT1A autoreceptor in the innate anxiety (Richardson-Jones et al. 2011), in agreement with the over-expression of 5-HT1A receptors in the midbrain of STOP KO mice (Fournet et al. 2010).

Indeed, although numerous neurotransmitters, particularly the GABAergic system, have been implicated in the regulation of anxiety level (Belzung 2001), we propose that the reduced anxious-like phenotype of STOP KO mice might be mainly linked to increased SERT in the raphe nuclei containing the 5-HT somas. However, a potential dysfunction of the hypothalamo-pituitary-adrenal axis (Lanfumey et al. 2008) should also be taken into account to explain the marked alterations of STOP KO behaviors reported here. Finally, comorbidity between depression and anxiety is the rule in both humans and animal models (Kessler et al. 2005). In our case, the unusual association of a depression-status with lower anxiety in STOP KO mice could be due to the imbalance in the 5-HT tone between somas and terminals.

STOP KO mice were hypersensitive to acute antidepressants

STOP KO mice exhibited increased sensitivity towards the three different antidepressants tested, that is, fluoxetine, a specific 5-HT reuptake inhibitor, venlafaxine, a mixed 5-HT and NE reuptake inhibitor and reboxetine, a specific NE reuptake inhibitor. These data were consistent with the hypothesis that both 5-HT and NE systems are strongly impaired in STOP KO mice. Moreover, this hypersensitivity of mutant mice to acute antidepressant treatments suggests a lower competition between the antidepressant and the extracellular 5-HT and NE for their binding to SERT and NET, respectively. This could be the case notably at the level of 5-HT somas where SERT and NET densities were increased, potentially leading to decreased extracellular monoamine levels.

STOP KO mice exhibited impaired cognitive performance

Cognitive performance of STOP KO mice of the new genetic background was also characterized in three different tests involving several parameters, ranging from attention, motivation, learning to memory. This study showed that STOP KO mice exhibited preserved very short-term memory in the spontaneous alternation test. This result was not consistent with previous data (Delotterie et al. 2010), possibly because of different genetic backgrounds and/or experimental conditions. In contrast, short- and long-term memories of STOP KO mice were severely impaired in the novel object recognition test, as already reported (Powell et al. 2007). Finally, STOP knockout mice showed neither spatial learning, nor spatial memory in the Morris watermaze.

To date, the numerous studies aimed at examining cognitive performance of STOP KO mice demonstrated lower performance in nearly all tests undertaken. This strong phenotype was probably not related to the genetic background, but rather to numerous dysfunctions reported in STOP KO adult mice, including alterations of synaptic plasticity, such as long-term potentiation and depression (Andrieux et al. 2002), neuronal plasticity such as neurogenesis (Fournet et al. 2010), synaptic vesicular pools (Andrieux et al. 2002), synaptic protein levels (Eastwood et al. 2007), hypoglutamatergia (Brenner et al. 2007) and also dysfunctions of monoaminergic and nicotinic neurotransmissions (Brun et al. 2005; Bouvrais-Veret et al. 2007, 2008; Fournet et al. 2010). Indeed, monoamines including DA (Nieoullon 2002; El-Ghundi et al. 2007; Morice et al. 2007), NE (Franowicz et al. 2002; Brown and Silva 2004) and 5-HT (Gonzalez-Burgos and Feria-Velasco 2008; Jenkins et al. 2010) are known to mediate a variety of cognitive processes, such as attention, learning, memory and adaptative behaviors.

The role of microtubule-associated proteins in monoaminergic dysfunctions of STOP KO mice

Altogether, the results of this study indicate a crucial role of MAP6/STOP protein in the establishment of monoaminergic neuronal networks, underlining the role of the cytoskeleton effectors in these mechanisms. Accordingly, neurotransmission defects have recently been reported in mice deleted for microtubule-associated proteins, such as DISC1 and dysbindin 1. DISC1 mutant mice exhibit dysfunctions affecting information processing (Niwa et al. 2010), working (Kvajo et al. 2008) and spatial (Pletnikov et al. 2008) memories. Interestingly, mice with the above mutations also show abnormalities in postnatal mesocortical dopaminergic maturation (Niwa et al. 2010), as well as decreased DA contents in the prefrontal cortex and hippocampus (Ayhan et al. 2010). These DA dysfunctions elicit behavioral impairments such as prepulse inhibition deficit, spontaneous and psychostimulant-induced hyperactivity, higher methamphetamine-induced DA release (Niwa et al. 2010). However, DISC1 misexpression elicits also depressed behavior (Clapcote et al. 2007; Ayhan et al. 2010), despite unchanged 5-HT and NE levels (Ayhan et al. 2010). In the same way, the deletion in mice of dysbindin-1, another protein interacting with microtubules, alters DA receptor trafficking (Ji et al. 2009) and models some aspects of schizophrenia (Talbot 2009). Thus, DISC1 and possibly dysbindin-1 misexpressions seem to impact more selectively the DAergic neurotransmission. In contrast, STOP deletion dramatically influences 5-HT and NE neurotransmissions, but moderately affects the DA neurotransmission (Brun et al. 2005; Bouvrais-Veret et al. 2008). The importance of the impact of the lack of MAP6 protein on specific neurotransmitters (5-HT > NE > DA) could be related to their more or less early expression onset during brain ontogenesis.

Conclusion

The deletion of MAP6 protein induced marked alterations of depressive-like and anxious-like status and highly impaired memory performance in adult mice. We propose that this phenotype was partly the result of decreased 5-HT levels in terminals (increased helplessness), associated with SERT over-expression in the 5-HT somas (decreased anxiety). Taken together, the data obtained suggest that the microtubule-stabilizing MAP6 protein plays a crucial role in the development and/or maintenance of the 5-HT, and likely of the NE, neuronal networks. Additionally, STOP KO mice might be a pertinent and useful experimental model for assays of novel therapeutic agents for schizoaffective disorders.

Acknowledgements

The authors wish to thank Elise Morice-Poret and Gbassay Serra for their helpful discussions and improvement of our manuscript, Johanne Germain for her expertise in the evaluation of mouse coat state, Dominique Divers for genotyping, Stéphane Baton and Julien Boivin for their technical assistance. This study was supported by grants from INSERM and Université Pierre et Marie Curie. Vincent Fournet is the recipient of fellowships from the MENESR (France). The authors reported no biomedical financial interests or potential conflicts of interest.

Ancillary