Hyperfunction of the stress response system and novelty‐induced hyperactivity correlate with enhanced cocaine‐induced conditioned place preference in NCAM‐deficient mice

Several studies in humans and rodents suggest an association between impulsivity and activity of the stress response on the one hand and addiction vulnerability on the other. The neural cell adhesion molecule (NCAM) has been related to several neuropsychiatric disorders in humans. Constitutively NCAM‐deficient (−/−) mice display enhanced novelty‐induced behavior and hyperfunction of the hypothalamic‐pituitary‐adrenal axis. Here we hypothesize that NCAM deficiency causes an altered response to cocaine. Cocaine‐induced behaviors of NCAM−/− mice and wild‐type (+/+) littermates were analyzed in the conditioned place preference (CPP) test. c‐fos mRNA levels were investigated by quantitative polymerase chain reaction (qPCR) to measure neural activation after exposure to the cocaine‐associated context. NCAM−/− mice showed an elevated cocaine‐induced sensitization, enhanced CPP, impaired extinction, and potentiated cocaine‐induced hyperlocomotion and CPP after extinction. NCAM−/− showed no potentiated CPP as compared with NCAM+/+ littermates when a natural rewarding stimulus (ie, an unfamiliar female) was used, suggesting that the behavioral alterations of NCAM−/− mice observed in the CPP test are specific to the effects of cocaine. Activation of the prefrontal cortex and nucleus accumbens induced by the cocaine‐associated context was enhanced in NCAM−/− compared with NCAM+/+ mice. Finally, cocaine‐induced behavior correlated positively with novelty‐induced behavior and plasma corticosterone levels in NCAM−/− mice and negatively with NCAM mRNA levels in the hippocampus and nucleus accumbens in wild‐type mice. Our findings indicate that NCAM deficiency affects cocaine‐induced CPP in mice and support the view that hyperfunction of the stress response system and reactivity to novelty predict the behavioral responses to cocaine.


| INTRODUCTION
It is commonly accepted that individual vulnerability towards the addictive properties of drugs of abuse is determined by environmental and genetic factors. 1,2 In particular, studies in mice and humans suggest a link between specific personality and physiological traits such as novelty seeking and activity of the stress response and susceptibility to the effects of drugs of addiction. In humans, it has been reported that there is a strong correlation between impulsivity, mood disorders, novelty seeking, and substance abuse. [2][3][4] In mice, animals with high trait anxiety and activity of the stress response develop an enhanced cocaine-induced conditioned place preference (CPP) compared with mice with low trait anxiety and activity of the stress response. 5 Stress is considered one of the major environmental risk factors in addiction, suggesting that susceptibility to drugs of abuse depends on the activity of the hypothalamic-pituitary-adrenal (HPA) axis. Acute stressful events facilitate the expression of addictive behaviors, [6][7][8][9] and chronic stress potentiates cocaine-induced locomotor activity 10 and cue-induced relapse of cocaine self-administration 11

in rats.
We have shown that the stress response is enhanced in mice deficient for NCAM. 12,13 It has been proposed that altered NCAM function underlies an enhanced susceptibility for developing depression and anxiety disorders in mice and humans, 14,15 and several studies on humans have related altered expression and function of NCAM with the occurrence of schizophrenia and bipolar neuropsychiatric disorders. 16,17 NCAM belongs to the immunoglobulin superfamily of cell adhesion molecules and functions in homophilic and heterophilic cell recognition as well as proteolytic fragment-mediated cell interactions that influence signal transduction in distinct manners. 18 More specifically, NCAM is involved in cell migration, neurite outgrowth and targeting, axonal branching, synaptogenesis, and synaptic plasticity. 18 Noteworthy, constitutively NCAM−/− mice display two behavioral features hypothesized to predict an increased sensitivity towards the reinforcing effects of drugs of abuse, 19,20 namely, enhanced novelty-induced locomotion and anxiety as compared with NCAM+/+ control littermates. 12 Moreover, NCAM−/− mice show elevated activity of the stress response and reduced levels of glucocorticoid receptors in the hippocampus. 12 Here, we used the NCAM−/− mouse as a model to search for a link between behavioral and physiological characteristics supposed to predispose for addiction in humans and cocaine-induced conditioned behavior.

| Animals and husbandry
NCAM-deficient (NCAM−/−) and wild-type (NCAM+/+) male littermates were generated by heterozygous breeding and transferred from a pathogen-free breeding facility into a vivarium with an inverted 12:12 light:dark cycle (light off at 7:00 AM). They were maintained in groups of two to four siblings under standard housing conditions (22 ± 1 C, 40-60% humidity, food, and water ad libitum).
All mice were inbred on a C57BL/6J background (after at least 10 generations of backcrossing). C57BL/6J adult male mice were used to perform correlation analyses between their behavior in the cocaine-induced CPP test and NCAM mRNA levels in different brain regions. After two weeks of acclimatization, mice underwent the experiments that were performed between 9:30 AM

| Assay for cocaine-induced sensitization and CPP
The cocaine-induced CPP test assesses the rewarding properties of psychostimulants in rodents. 21 Also, it allows the evaluation of sensitization defined by the further increase of drug-evoked locomotor activity after repeated administration of the drug over consecutive experimental days. 22 The cocaine-induced CPP protocol was performed on five consecutive days comprising a preconditioning trial, three daily conditioning sessions, and a postconditioning trial 23 ( Figure 1A). The test was performed in a 50 × 50 × 40 cm box, laminated with rough, matted, light-gray resin. Each box consisted of two compartments (50 × 25 × 40 cm) separated by a white plastic wall with a cylinder containing two removable doors in the middle. Both compartments were illuminated by a white light bulb (10 lux). The two compartments differed in floor texture and landmarks on the walls. For the preconditioning trial, mice were placed in the cylinder connecting the two compartments and left exploring both compartments for 20 minutes. One NCAM+/+ mouse and one NCAM−/− mouse that spent more than 60% of the time in one compartment were excluded from further testing and analysis to avoid intrinsic preferences for one specific compartment. During the conditioning sessions on days 2 to 4, all mice underwent a trial during which they were first injected with vehicle solution (0.9% saline, ip) and then placed into the nonconditioned (CS−) compartment for 20 minutes. Four hours later, the second trial took place during which the mice were injected with either 10 or 0.5 mg/kg cocaine hydrochloride (diluted in saline solution, ip) and placed in the cocaine-conditioned (CS+) compartment for 20 minutes. During all conditioning trials, the door between the two compartments was closed to confine the mice to the compartment into which they had been placed. On day 5, mice performed the postconditioning trial to determine the time spent in the conditioned compartment as compared with the chance level (50%). Mice were placed in the cylinder and were free to explore both compartments for 20 minutes. The software EthoVision (Noldus, Wageningen, The F I G U R E 1 Enhanced cocaine-induced behavior in neural cell adhesion molecule (NCAM)−/− mice. A, experimental design of the cocaineinduced conditioned place preference (CCP) paradigm; B, distance moved (m) in the respective compartment after saline (SAL) or 10 mg/kg cocaine (10 mg) injections during the three conditioning days, *** P < .001 for the comparison between genotypes for the distance moved during the saline trials (Newman-Keuls post hoc test after mixed three-way ANOVA); C, time (%) spent in the cocaine-conditioned compartment was measured before (preconditioning) and after (postconditioning) the three conditioning days using 10 mg/kg cocaine, * P < .001, *** P < .05 (Newman-Keuls after significant effect of the interaction between genotype and trial calculated with a mixed two-way ANOVA); D, distance moved (m) in the respective compartment after saline (SAL) and 0.5 mg/kg cocaine (0.5 mg) injections during the three conditioning days, *** P < .001 for the comparison between genotypes for distance moved (Newman-Keuls post hoc test after mixed three-way ANOVA); E, time (%) spent in the cocaine-conditioned compartment as measured during the preconditioning and postconditioning trials using 0.5 mg/kg cocaine, ** P < .01 (Newman-Keuls after significant effect of the interaction between genotype and trial calculated with a mixed two-way ANOVA). The sample size was 15 NCAM+/+ and 13 NCAM−/− mice in B and C and 13 mice per genotype in D and E. § P < .05 for the comparison between cocaine-induced locomotion on days 3 and 1 regardless of genotype (Newman-Keuls post hoc test after significant effect of treatment analyzed with a mixed three-way ANOVA) Netherlands) was used to track the position of the mice and calculate distance moved and time spent in each compartment.

| Extinction and reconditioning to cocaine after extinction
A cohort of mice underwent the cocaine-induced CPP as described above followed by an extinction protocol. The extinction protocol comprised eight postconditioning trials (ie, mice were not injected and had access to both compartments) of 10 minutes, and each trial was performed on eight consecutive days. Whether mice could re-express the CPP after extinction was tested after one conditioning day as described in the following: On the morning, after the last extinction trial, mice underwent one conditioning trial during which they received a cocaine injection (5 mg/kg, ip) before being placed into the CS+ compartment without access to the CS− compartment. Four hours later, mice were injected with saline and confined to the CS− compartment. On the following day, mice were subjected to a 20-minute test trial during which they were free to explore both compartments. As a control, NCAM+/+ and NCAM−/− mice that had never been conditioned before (defined as "naïve") underwent a preconditioning trial and a 1-day conditioning session with one conditioning trial with 5 mg/kg cocaine and one conditioning trial with saline, followed by a postconditioning trial 24 hours later.

| Female-induced CPP test
For this test, the maze was similar to the one described for the CPP test with the only difference being that a transparent plastic beaker was placed in one corner of each compartment. Several holes were drilled into the beaker 1 cm from the bottom. Mice did three conditioning sessions as described for the CPP test. However, this time, the mice did not receive any injection, and an unfamiliar female mouse was placed in the beaker of one of the two compartments during the conditioning trial. Thus, one compartment was always associated with the presence of a female mouse. On day 5, mice underwent the postconditioning trial to determine the time spent in the conditioned compartment as compared with chance level (50%): Mice were placed in the cylinder and were free to explore both compartments for 20 minutes. The extinction protocol consisted of five trials performed on five consecutive days as described for the postconditioning trial, but this time, the duration was 10 minutes.

| Open field test
The open field was an enclosed arena (50 × 50 × 40 cm) made of wood laminated with light-gray resin and illuminated with white light (50 lux). Mice were placed in one corner of the arena and allowed to freely move for 20 minutes. Total distance moved, mean velocity, mean distance to the wall (a parameter for thigmotaxis) and time spent in the center (an imaginary inner square of 20 × 20 cm) were analyzed with EthoVision. A trained observer scored the behavior of the animals during the first 5 minutes of the test using the software The Observer (Noldus) as described. 24  as an endogenous reference. For the graphical presentation and statistical analysis, the mRNA amount for each sample was expressed as a percentage of the mean value of the NCAM+/+ group that was set to 100%.

| Corticosterone ELISA
Levels of corticosterone in the plasma were quantified using a commercially available kit (rat/mouse Corticosterone ELISA; Cusabio Technology LLC, Wuhan, China) according to the manufacturer's instructions. For the assay, the plasma samples were diluted 1:20 using the sample diluent provided in the kit. Concentrations are expressed as nanogram per milliliter.

| Statistical analysis
Paired measurements were tested with mixed two-or three-way ANOVA tests, always having "genotype" (NCAM+/+ and NCAM−/−) as the between-groups factor. The within-groups factors were "treatment" (saline and cocaine) and "day" (days 1-3) for distance moved during the conditioning trials, "preconditioning versus postconditioning trial" in the cocaine-induced CPP test, and "trial" (trials 1-8 or 1-5) for preference for the conditioned compartment during the extinction protocols. Distance moved in the open field test performed after the extinction protocol for assessing reinstatement was analyzed with mixed two-way ANOVA having treatment (saline and cocaine) as the within-groups factor. Expression of c-fos mRNA was analyzed with two-way ANOVA having "genotype" and "compartment" (saline and cocaine) as the between-groups factors. All  Figure 2A). Also, NCAM+/+ and NCAM−/− mice that had completed the extinction protocol ("NCAM+/+ post-extinction" and "NCAM−/− post-extinction") showed enhanced susceptibility to develop CPP after a single conditioning trial using 5 mg/kg cocaine when compared with mice that had not been conditioned before ("NCAM+/+ naïve" and "NCAM−/− naïve"). The three-way ANOVA (with "genotype" and "previous experience" as the between-groups factors and "trial" as the within-groups factor) detected a tendency for the interaction between genotype, previous experience, and trial on distance moved during the saline and cocaine conditioning trials,

| Conditioned responses and extinction are unaltered in NCAM−/− mice when a natural reward is used
To test whether NCAM−/− mice develop enhanced CPP compared with NCAM+/+ mice independently of cocaine, the behavior was tested in a spatial learning test as described. 26,27 In this test, mice develop a preference for a compartment that had been associated

| Enhanced c-fos expression induced by the conditioned compartment in NCAM−/− mice
To examine whether neural activation is associated with cocaineinduced behavior, c-fos expression was examined in the prefrontal cortex, nucleus accumbens, and cerebellum of NCAM+/+ and NCAM −/− mice. Mice underwent the CPP test using 10 mg/kg cocaine. One week after CPP, groups of NCAM+/+ and NCAM−/− mice were exposed for 30 minutes to the CS+ compartment (associated with 10 mg/kg cocaine injection during CPP) and another group to the CS− compartment (associated with saline injection during CPP) and sacrificed 10 minutes after termination of the test to measure c-fos mRNA expression by quantitative reverse transcription PCR ( Figure 4A). Overall, mice exposed to the CS+ compartment showed enhanced c-fos expression in the prefrontal cortex and nucleus accumbens compared with mice exposed to the CS− compartment

| Cocaine-induced hyperlocomotion and CPP correlate with novelty-induced locomotion and corticosterone levels
To test the hypothesis that enhanced cocaine-induced CPP of NCAM−/− mice relates to their elevated novelty-induced behavior F I G U R E 2 Reduced extinction in neural cell adhesion molecule (NCAM)−/− mice. A, time (%) spent in the cocaine-conditioned compartment during the extinction protocol (ie, eight test trials after the three conditioning days using 10 mg/kg cocaine), * P < .05, ** P < .01 between genotypes within each trial (Newman-Keuls post hoc test after mixed two-way ANOVA); B, distance moved (m) in the compartment after vehicle ("saline") or 5 mg/kg cocaine ("cocaine") injections during a single conditioning day in naïve mice ("NCAM+/+ naïve" and "NCAM−/− naïve") and mice that had undergone the extinction protocol ("NCAM+/+ post-ext." and "NCAM−/− post-ext."); * P < .05, ** P < .01, and *** P < .001 for the comparison between groups for distance moved during the conditioning trial with 5 mg/kg cocaine, and +++ P < .001 for the comparison between distance moved during the conditioning trial after saline versus the conditioning trial after cocaine injections within each group (Newman-Keuls post hoc test after mixed three-way ANOVA); C, time (%) spent in the cocaine-conditioned compartment 24 h after one conditioning day with 5 mg/kg cocaine injected after the extinction protocol, * P < .05 and *** P < .001 for the comparison between groups (Newman-Keuls post-hoc test after two-way-ANOVA). The sample size was 10 mice per genotype in A, and seven and eight mice per genotype in the naïve and post-extinction groups, respectively, in B and C and activity of the HPA axis, we sought to correlate cocaine- To examine whether cocaine-induced behavior correlates with NCAM expression in specific brain regions of wild-type mice, C57BL/6J mice underwent cocaine-induced CPP with 10 mg/kg as described above and, as expected, showed cocaine-induced hyperlocomotion and sensitization during the conditioning trials (effect of the interaction between treatment and day, F(1, 18) = 5.06, P = .018; Figure 7A) and developed CPP for the cocaine-associated compartment as tested in the postconditioning trial (paired t test t 9 = 3.30, P = .009; Figure 7B). Cocaine-induced locomotion during the third conditioning trial negatively correlated with NCAM mRNA levels in the hippocampus (Spearman r = −.73, P = .01; Figure 8A) F I G U R E 3 Female-induced conditioned place preference (CPP) is similar in both genotypes. A, time (%) spent in the femaleconditioned compartment was measured before (Before Cond.) and after (After Cond.) the three conditioning days. § § P < .01 as compared with chance level (50%, indicated by a dotted line) (Wilcoxon signed-rank test). B, time (%) spent in the female-conditioned compartment during the extinction protocol (ie, five test trials after the three conditioning days). The sample size was seven mice per genotype F I G U R E 4 Neural cell adhesion molecule (NCAM)−/− mice display enhanced c-fos mRNA levels induced by the conditioned compartment. (A) Experimental design, (B) c-fos mRNA level in the nucleus accumbens, and (C) prefrontal cortex, but not in the (D) cerebellum is enhanced in previously conditioned mice (injected with 10 mg/kg cocaine) exposed to the cocaine-conditioned (CS+) compartment compared with previously conditioned mice exposed to the saline-associated (CS−) compartment. § P < .05 and § § P < .01 as compared with genotype-matched mice exposed to the CS-compartment (Newman-Keuls post hoc test after two-way ANOVA). c-fos mRNA expression in the nucleus accumbens (B) and prefrontal cortex (C) induced by exposure to the CS+ compartment is enhanced in NCAM−/− versus NCAM+/+ mice. * P < .05 for comparison between NCAM+/+ and NCAM−/− mice exposed to the CS+ compartment (Newman-Keuls post hoc test after two-way ANOVA). Individual cfos mRNA levels analyzed in each brain region are normalized to the average levels in the NCAM+/+ group exposed to the CS− compartment. The sample size was five and six mice per genotype for the CS− and CS+ groups, respectively and nucleus accumbens (Spearman r = −.64, P = .047; Figure 8C).   Studies in humans and animals indicate that cocaine-induced CPP and extinction are modulated by the prefrontal cortex and nucleus accumbens, brain regions known to play an important role in relapse. 8,28 Based on these observations, we measured whether reexposure to the previously cocaine-conditioned compartment leads to functional activation of these regions, using the cerebellum as a negative control. 29 Our data show that exposure to the cocaine-associated compartment induced higher c-fos mRNA expression in the nucleus accumbens and prefrontal cortex when compared with the exposure to the saline-associated compartment. Since mice had been equally often exposed to both the saline-and cocaine-conditioned compartments, the enhanced c-fos mRNA expression induced in mice exposed to the "conditioned compartment" compared with mice exposed to the "saline compartment" appears to be caused by the association to the previous cocaine injections. Thus, not only cocaine injection, as reported, 30 but also a cocaine-associated stimulus appears to be sufficient to functionally activate the nucleus accumbens and prefrontal cortex, similarly to the ventral tegmental area. 31 Cocaine-induced sensitization and conditioned place preference (CPP) in C57BL/6J mice. A, distance moved (m) in the respective compartment after saline (SAL) or 10 mg/kg cocaine injections (10 mg/kg) during the three conditioning days. * P < .05 and *** P < .01 for comparison between days within the cocaine-conditioning trials; +++ P < .001 for comparison between saline-and cocaine-conditioning trials performed on the same day (Newman-Keuls post hoc test after two-way ANOVA for repeated measurements with treatment and day as between groups factor); B, time (%) spent in the cocaine-conditioned compartment was measured before (Pre-cond.) and after (Post-cond.) the three conditioning days using 10 mg/kg cocaine, ** P < .01 (paired t test). The sample size was 10 mice caused by NCAM deficiency in the prefrontal cortex. This hypothesis would imply that the elevated amount of c-fos mRNA induced by the conditioned compartment in the prefrontal cortex of NCAM−/− mice was a secondary effect of NCAM depletion in the hippocampus and nucleus accumbens. Alternatively, NCAM in the prefrontal cortex modulates, indeed, cocaine-induced behavior, but no correlation could be found because NCAM is weakly expressed in the prefrontal cortex and almost exclusively in interneurons. 35,36 Although the aim of the present study was not to investigate the molecular and cellular mechanisms by which NCAM deficiency leads to enhanced behavioral responses to cocaine, several mechanisms can be brought forward based on previous studies. It has been suggested that cell adhesion molecules, such as SynCAM 1 37 or NrCAM, 38 regulate addiction liability because of their pivotal role in synaptic plasticity. 39 Similarly, NCAM had long been known to affect synaptic plasticity and ablated NCAM impairs hippocampal long-term F I G U R E 8 Cocaine-induced sensitization and conditioned place preference (CPP) negatively correlate with NCAM mRNA levels in the hippocampus and nucleus accumbens of C57BL/6J mice; neural cell adhesion molecule (NCAM) mRNA expression in the hippocampus negatively correlates with cocaine-induced hyperlocomotion on conditioning day 3 (A) and time spent in the CS+ compartment during the postconditioning trial (B). NCAM mRNA expression in the nucleus accumbens negatively correlates with cocaine-induced hyperlocomotion on conditioning day 3 (C). A tendency was detected for a negative correlation between c-fos mRNA levels in the nucleus accumbens with time spent in the CS+ compartment during the postconditioning trial (D). No correlation was detected between NCAM mRNA levels in the prefrontal cortex (PC) and cocaine-induced hyperlocomotion on conditioning day 3 (E) and time spent in the CS+ compartment during the postconditioning trial (F). NCAM mRNA expression is indicated as the arbitrary unit (au) calculated for each data point as the relative difference to the group mean for each brain region. The sample size was 10 mice potentiation. 18  It is possible that the same molecular mechanisms underlying the enhanced cocaine responses of NCAM−/− mice, for instance an altered expression of dopamine receptors, are also responsible for the enhanced novelty-induced behavior and elevated activity of the HPA axis 44 that have been described for NCAM−/− mice. [12][13][14] For instance, it has been shown that corticosterone affects dopamine neurotransmission specifically when the dopamine transporter is pharmacologically blocked by cocaine, thus enhancing the effects of cocaine. 45,46 Also, a 30-minute restraint enhances cocaine-induced CPP due to cholinergic and glutamatergic activation of dopaminergic neurons in the ventral tegmental area. 47 Since stress is known to affect glutamatergic transmission, 48,49 it is plausible to assume that several molecular mechanisms are altered under NCAM deficiency, leading to a complex behavioral phenotype affecting the stress response, novelty-induced behavior, and susceptibility to drugs of addiction. Our present data indicate that activity of the HPA axis and novelty-induced behaviors correlate with cocaine-conditioned addictive-like responses at the individual and group levels, supporting the view that subjects with an enhanced novelty-seeking and activity of the stress response are more at risk to develop maladaptive addictive behaviors upon exposure to drugs.
In conclusion, the present study has uncovered a correlational link between NCAM expression, coping strategy in response to novelty and stressors, and risk to develop addictive-like behavior in mice.
Since polymorphisms in the NCAM gene have been related to risk for alcoholism in humans 50 and low expression of polysialylated NCAM has been associated with the risk for alcoholism in mice, 51 it is tempting to propose that NCAM deficiency or low levels of NCAM expression lead to a general predisposition to the effects of several drugs acting on different molecular targets, such as gammaaminobutyric acid receptors for alcohol, and serotonin-norepinephrine-dopamine reuptake for cocaine. Since stress downregulates NCAM expression in the brain, 10