Pleiotropic impact of constitutive fosB inactivation on nicotine-induced behavioral alterations and stress-related traits in mice
1 Laboratory of Molecular Psychobiology, Department of Psychiatry and Behavioral Sciences and 2 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY 10461, USA
* To whom correspondence should be addressed at: Laboratory of Molecular Psychobiology, Department of Psychiatry and Behavioral Sciences, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461, USA. Tel: +1 7184303124; Fax: +1 7184303125; Email: hiroi{at}aecom.yu.edu
Received January 9, 2007; Accepted February 4, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSSUPPLEMENTARY MATERIALREFERENCES |
|---|
Multiple genes are thought to influence both susceptibility to nicotine dependence and its comorbid behavioral traits in humans. However, which specific genes contribute to this pleiotropic effect is poorly understood. Previous rodent studies have shown that many addictive substances and stressful stimuli increase the expression of the transcription factor FosB in limbic and associated regions and that the protein products of fosB contribute to certain behavioral effects of cocaine and morphine. However, the role of this gene in nicotine-regulated behaviors and dependence-related behavioral traits is unknown. We tested the hypothesis that a constitutive level of FosB affects nicotine-regulated behaviors and comorbid behavioral traits using constitutive fosB knockout (KO) mice. Following repeated or prolonged nicotine administration, but not a single acute administration, KO mice were impaired in conditioned place preference, oral nicotine intake and motor suppression. In wild-type mice, repeated nicotine injections, but not a single acute injection, increased the expression of FosB and its truncated variant
FosB in the targets but not at the origins of the mesolimbic and nigrostriatal dopamine pathways; no detectable level of FosB/FosB was found in KO mice. In tasks designed to assess behavioral traits, KO mice exhibited more pronounced behavioral abnormalities when stress levels were high than when they were minimized. Our results suggest that the constitutive absence of fosB has a pleiotropic influence on the behavioral effects of repeated or prolonged nicotine administration and on stress-related behavioral traits in mice. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSSUPPLEMENTARY MATERIALREFERENCES |
|---|
Among the various types of substance dependence, cigarette smoking is one of the most alarming public health problems in the world today. Worldwide, an estimated 1.3 billion people smoke, and 5 million deaths annually are attributable to smoking (1). Fewer than 5% of those who quit for at least 1 day are able to stay tobacco-free for 3–12 months (2).
Nicotine is the dependence-inducing component of tobacco products. Smokers prefer regular to de-nicotinized cigarettes when given a choice (3–5). Nicotine-containing cigarettes are rated higher for subjective pleasure, satisfaction and preference than de-nicotinized cigarettes (4,6–8). Nicotine administered alone produces symptoms of dependence in both experimental animals and humans (9).
One intriguing aspect of substance dependence, including nicotine dependence, is that it is not inevitably triggered upon exposure to substances with a dependence potential. Some individuals easily develop dependence after exposures to substances with a dependence potential, but others never develop dependence even after repeated or chronic exposure (10,11). Susceptible individuals tend to exhibit comorbid traits such as novelty seeking, heightened anxiety, altered stress response and depression (11,12), which may predate the onset of daily smoking (13–18).
As studies with human twins have suggested that genetic factors affect the susceptibility to nicotine dependence (19), one interpretation of the overlap between dependence susceptibility and comorbid behavioral traits is that genes affect a specific set of behavioral traits, one of which is the susceptibility to dependence (11,20). From this perspective, the critical issue in understanding dependence is how specific genes constitutively affect not only dependence susceptibility but also specific comorbid behavioral traits. This genetic effect is consistent with the phenomenon of pleiotropy, in which a single gene affects many phenotypes. Although many genes have been proposed to affect susceptibility to nicotine dependence (21,22), the complex influences of genetic and environmental factors, together with the presence of genetically heterogeneous samples in each study population, have made it difficult to dissect out the specific genes that mediate this pleiotropic action in humans.
The expression of FosB, a Fos family transcription factor, and its truncated form FosB is increased in regions associated with the mesolimbic dopaminergic system and in other limbic brain regions by nicotine and stimulants in rodents (23–29). Previous rodent studies have also demonstrated that FosB and FosB contribute to some behavioral effects of cocaine and morphine (25,30–33). FosB and FosB are also induced in limbic and other related brain regions by external stimuli that evoke stress and alter the affective state (34,35). To ascertain the functional role played by this transcription factor in behavioral traits, as well as the behavioral effects of nicotine, we assessed the impact of a constitutive genetic deletion of fosB on these behaviors in fosB knockout (KO) mice. Our results suggest that a constitutive deficiency of FosB alters stress-related behaviors and the behavioral effects of repeated or prolonged nicotine administration.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSSUPPLEMENTARY MATERIALREFERENCES |
|---|
FosB KO mice do not exhibit conditioned place preference in response to repeated nicotine injections
We used the place conditioning paradigm to assess approach toward cues previously associated with nicotine (36–38). We first determined whether mice exhibit any pre-existing preference for either of the two large compartments of the apparatus on the pre-test day. Wild-type (WT) and KO mice spent indistinguishable amounts of time in the two compartments on a drug-free preconditioning day [genotype, F1,284 = 2.19, not significant (n.s.); compartment, F1,284 = 1.18, n.s.; interaction, F1,284 = 0.0002, n.s.] (Fig. 1A), indicating that our procedure was unbiased.
|
We then administered nicotine to the mice in this assay. Nicotine induces conditioned place preference (CPP) within an extremely narrow dose range in mice [see Agatsuma et al. (20) for discussion]. Because we used a wide range of doses to rule out a possible shift in the dose–response curve, half the doses did not induce either CPP or conditioned place aversion (CPA) following 3 days of conditioning (Fig. 1B). Owing to the nature of these data, a three-way analysis of variance (ANOVA) did not show a significant effect for the following comparisons: genotype (F1,200 = 1.52, n.s.), dose (F5,200 = 0.04, n.s.) or interaction between genotype and dose (F5,200 = 0.54, n.s.), genotype and compartment (F1,200 = 1.84, n.s.) or genotype, dose and compartment (F1,200 = 1.53, n.s.). However, we found significant effects for compartment (F1,200 = 8.03, P < 0.01) and interaction between dose and compartment (F5,200 = 5.59, P < 0.01). We therefore applied two-way ANOVAs separately to each dose. This analysis revealed a significant genotype x compartment interaction at 0.2 mg/kg (F1,28 = 6.09, P < 0.05) and 0.6 mg/kg (F1,22 = 4.67, P < 0.05). Newman–Keuls post hoc tests showed that these interactions resulted from CPP in WT mice but not in KO mice at 0.2 mg/kg. At 0.6 mg/kg, WT mice showed neither CPP nor CPA, but KO mice showed CPA. Both WT and KO mice showed CPA at 0.8 mg/kg (genotype, F1,70 = 0.15, n.s.; compartment, F1,70 = 23.13, P < 0.01; interaction, F1,70 = 1.99, n.s.) and 2.0 mg/kg (genotype, F1,24 = 0.002, n.s.; compartment, F1,24 = 25.90, P < 0.01; interaction, F1,24 = 1.88, n.s.).
We next examined the effect of a single nicotine injection [0.2 mg/kg, subcutaneously (s.c.)] in this task (Fig. 1C). WT and KO mice did not differ (genotype, F1,66 = 1.07, n.s.) and showed no preference for either compartment (F1,66 = 0.13, n.s.). We found no interaction between genotype and compartment (F1,66 = 3.33, n.s.).
FosB KO mice exhibit lower levels of nicotine intake
When orally consumed by mice, nicotine accumulates and exerts physiological effects in the brain (39–44). We therefore assessed nicotine intake in a two-bottle choice paradigm (20,45,46), with nicotine solution in one bottle and tap water in the other. The total amount of nicotine intake was recorded every 3 days from days 3 to 15 (Fig. 2A). Overall, WT mice consumed more nicotine than did KO mice (genotype, F1,62 = 8.24, P < 0.01), and nicotine intake differed among concentrations (F2,62 = 8.68, P < 0.01) and across days (F4,248 = 3.84, P < 0.01). The overall difference between WT and KO mice depended on concentration (genotype x concentration, F2,62 = 8.74, P < 0.01) but not on day (genotype x day, F4,248 = 1.04, n.s.). Nicotine intake also varied depending on the interaction between concentration and day (F8,248 = 3.49, P < 0.01). The three-way interaction was not significant (F8,248 = 0.53, n.s.). Because we found a significant interaction between genotype and concentration, we applied two-way ANOVAs to determine the concentrations at which WT and KO mice differed. WT mice showed higher levels of nicotine intake than did KO mice at 50 µg/ml nicotine (F1,29 = 10.30, P < 0.01) with daily fluctuations (F4,116 = 3.99, P < 0.01) (Fig. 2A and B). The interaction between genotype and day was not significant (F4,116 = 0.53, n.s.). WT and KO mice did not differ at 12.5 or 25 µg/ml nicotine (Fig. 2A). Newman–Keuls post hoc tests showed a significant trend toward greater differences between WT and KO mice upon prolonged exposure.
|
Because KO mice consumed less nicotine at 50 µg/ml than WT mice, we tested whether this difference simply reflected less fluid intake in KO mice than WT mice. We analyzed the total fluid intake of WT and KO mice given free access to a water bottle and a bottle containing nicotine solution (Supplementary Material, Table S1). The main effects were significant for genotype (F1,83 = 7.93, P < 0.01) and concentration (F3,83 = 25.06, P < 0.01) but not for day (F4,332 = 2.28, n.s.). The interaction was significant for genotype x day (F4,332 = 4.28, P < 0.01) but not for concentration x day (F12,332 = 1.13, n.s.) or genotype x concentration (F3,83 = 1.99, n.s.). The three-way interaction was significant (F12,332 = 2.08, P < 0.05). We applied two-way ANOVAs to determine the concentrations at which WT and KO mice differed. WT and KO mice did not differ in total fluid intake at any concentration except for 25 µg/ml. Moreover, although not statistically significant, there was a trend for KO mice to consume more fluid per body weight than WT mice at 50 µg/ml (Supplementary Material, Table S1) (47). This was probably due to a slightly lower body weight in KO mice than WT mice, concomitant with indistinguishable levels of fluid intake between WT and KO mice in this concentration group (see Supplementary Material, Tables S2 and S3) (47). This analysis shows that the lower nicotine intake in KO mice at 50 µg/ml is not due to a reduced total fluid intake compared with WT mice.
Nicotine preference was analyzed using ratios of nicotine intake to the total nicotine and water intake. Overall, ratios were higher in WT mice than in KO mice (F1,83 = 8.12, P < 0.01), and ratios differed among concentrations (F3,83 = 100.44, P < 0.01) but not days (F4,332 = 1.66, n.s.) (Fig. 2C). Interactions were significant for genotype x concentration (F3,83 = 3.54, P < 0.05), genotype x day (F4,332 = 4.66, P < 0.01) and concentration x day (F12,332 = 5.57, P < 0.01). The three-way interaction was not significant (F12,332 = 0.46, n.s.). Because we found a significant interaction between genotype and concentration, we applied two-way ANOVAs to determine the concentrations at which WT and KO mice differed. WT mice showed a higher ratio than KO mice at 50 µg/ml (F1,29 = 22.07, P < 0.01) with daily fluctuations (F4,116 = 5.33, P < 0.01) (Fig. 2C and D). The interaction between genotype and day was not significant (F4,116 = 0.48, n.s.). WT and KO mice did not differ at 0, 12.5, or 25 µg/ml nicotine (Fig. 2C). Newman–Keuls post hoc tests showed that WT and KO mice differed on all but the first recording days (Fig. 2D).
FosB KO mice are normal in saccharin preference
To determine whether the impaired preference for nicotine can be generalized to a natural reward, we assessed saccharin preference in a two-bottle choice procedure (Fig. 3A). WT and KO mice had indistinguishable saccharin preference (F1,14 = 0.10, n.s.) and there was no daily fluctuation (F4,56 = 0.61, n.s.). We found no interaction between genotype and day (F4,56 = 0.08, n.s.). The result shows that fosB KO mice have normal preference for this natural reward.
|
FosB KO mice are normal in quinine aversion
To determine whether fosB KO mice might have an abnormal response to the bitter taste of the nicotine solution, we tested their response to quinine. Our pilot studies showed that 1 mM quinine induced avoidance equivalent to nicotine at 50 µg/ml (Figs 2C and 3B). WT and KO mice were indistinguishable in quinine aversion (F1,13 = 0.86, n.s.), but the aversion to quinine declined significantly over 15 days, resulting in a slight increase in ratios over days (F4,52 = 9.05, P < 0.01). There was no interaction between genotype and day (F4,52 = 2.21, n.s.). This result shows that the lower ratio of nicotine preference (i.e. a higher level of nicotine aversion) in KO mice compared with WT mice is not due to an altered aversion to bitter tastes.
FosB KO mice are impaired in motor suppression after repeated nicotine injections
Nicotine decreases locomotor activity by its action in the nucleus accumbens (48,49). We next examined the effect of constitutive FosB deficiency on this centrally mediated behavioral action of nicotine. WT and KO mice were indistinguishable from each other (genotype, F1,61 = 1.08, n.s.); both showed motor suppression in response to an acute nicotine injection (dose, F4,61 = 10.97, P < 0.01) (Fig. 4A). We found no significant interaction effect between genotype and dose (F4,61 = 0.62, n.s.). Newman–Keuls post hoc tests showed that both WT and KO mice exhibited statistically significant motor suppression at 0.6 and 0.8 mg/kg nicotine.
|
Because 0.8 mg/kg nicotine induced robust motor suppression, we used this dose to assess the effects of repeated nicotine injections (Fig. 4B). The overall main effects were significant for saline versus nicotine (treatment, F1,27 = 8.31, P < 0.01) and day (F10,270 = 16.89, P < 0.01) but not for genotype (F1,27 = 1.42, n.s.). Interaction was significant for treatment x day (F10,270 = 3.48, P < 0.01) but not genotype x treatment (F1,27 = 1.21, n.s.) or genotype x day (F10,270 = 0.91, n.s.). However, the three-way interaction among genotype, treatment and day was significant (F10,270 = 2.08, P < 0.05). Newman–Keuls post hoc tests showed that although WT and KO mice were indistinguishable on the first day, KO mice lost motor suppression more quickly than did WT mice (Fig. 4B).
FosB KO mice show normal blood levels of nicotine and cotinine
One possible explanation for the behavioral differences between WT and KO mice is that the two groups metabolize nicotine differently. To explore this possibility, we determined the levels of blood nicotine and its metabolite cotinine following four daily nicotine injections, at which point KO mice no longer exhibited motor suppression but WT mice still did (Fig. 4B). WT and KO mice showed indistinguishable levels of blood nicotine (F1,26 = 1.03, n.s., Fig. 5A) and cotinine (F1,22 = 0.10, n.s., Fig. 5B). There was a significant decline over the 30 min period in the levels of nicotine (F1,26 = 37.51, P < 0.01) but not of cotinine (F1,22 = 1.48, n.s.). No interaction was found between genotype and time for nicotine (F1,26 = 0.77, n.s.) or cotinine (F1,22 = 2.34, n.s.). These results rule out the possibility that the behavioral phenotype reflects different levels of blood nicotine between WT and KO mice.
|
Repeated nicotine exposure induces both FosB and
FosB in the nucleus accumbens and caudate-putamenHaving established that fosB is relevant to the behavioral actions of nicotine, we next determined the expression of FosB and
FosB in the target regions (nucleus accumbens and caudate-putamen) and origins [ventral tegmental area (VTA) and substantia nigra (SN)] of the mesolimbic and nigrostriatal dopamine pathways using western blotting (Fig. 6). We chose the injection regimen (three injections) and nicotine dose (0.2 mg/kg, s.c.) at which WT and KO mice differed in the place conditioning assay. Unlike voluntary oral nicotine intake, this paradigm controls both the amount of nicotine administered and the time interval between the last nicotine administration and sacrifice.
|
In WT mice, repeated nicotine injections increased the expression of FosB and
FosB in both the nucleus accumbens (FosB, F3,45 = 8.51, P < 0.01; FosB, F3,56 = 6.15, P < 0.01) and caudate-putamen (FosB, F3,45 = 3.51, P < 0.05; FosB, F3,55 = 5.88, P < 0.01) (Fig. 6A and B). Nicotine increased the expression of FosB at 1 and 6 h in the nucleus accumbens and at 0.5, 1 and 6 h in the caudate-putamen, whereas it increased expression of FosB only at 6 h in both the nucleus accumbens and caudate-putamen (Newman–Keuls post hoc tests). No detectable levels of basal or induced FosB or FosB were found in the VTA or SN (Fig. 6A).
We also used a single nicotine injection (0.2 mg/kg, s.c.), which did not induce CPP (Fig. 1C). Although there were some cases of increased expression in the nucleus accumbens and caudate-putamen, a single nicotine injection did not consistently increase FosB or FosB levels and, on a group basis, their increases failed to reach statistical significance in the nucleus accumbens (FosB, F3,54 = 0.22, n.s.; FosB, F3,54 = 1.98, n.s.) or caudate-putamen (FosB, F3,55 = 0.30, n.s.; FosB, F3,55 = 0.56, n.s.) (Fig. 6C and D). No detectable levels of basal or induced FosB or FosB were found in the VTA or SN following a single nicotine injection (Fig. 6C).
Our immunohistochemical analysis also showed that at the time point at which WT and KO mice differed in nicotine-induced motor suppression, nicotine (0.8 mg/kg, s.c., four injections, 30 min) increased the expression of FosB/FosB-like proteins in the core and shell divisions of the nucleus accumbens and throughout the caudate-putamen except for the dorsomedial quadrant in WT mice; no detectable levels of FosB/FosB were found in the VTA or SN (Supplementary Material, Figs S1 and S2).
KO mice did not show detectable levels of FosB/FosB in any of the regions examined following saline or nicotine injections (Fig. 6A and C and Supplementary Material, Fig. S1).
Taken together, these results show a regionally selective induction of FosB and FosB by repeated nicotine injections.
FosB KO mice show a heightened locomotor response in an inescapable open field
Spontaneous locomotor activity in an inescapable open field is a rodent behavior that is positively correlated with individual variations in nicotine self-administration (50). Overall, KO mice showed a higher level of locomotor activity than WT mice when placed in an open field (F1,14 = 9.77, P < 0.01) and less reduction in locomotor activity across days (genotype x day: F2,28 = 10.35, P < 0.01) and time intervals (genotype x time interval: F5,70 = 8.34, P < 0.01) (Fig. 7A). The three-way interaction was also significant (F10,140 = 1.95, P < 0.05). Newman–Keuls post hoc tests showed that KO and WT mice initially showed indistinguishable levels of locomotor activity when they were placed in the open field, but KO mice showed significantly less reduction in locomotor activity across time intervals on the first 2 days than WT mice.
|
Activity in the central area of the open field is thought to reflect anxiety-related behavior (51,52). Overall, WT and KO mice showed indistinguishable levels of locomotor activity in the central area (F1,14 = 3.69, n.s.) (Fig. 7B). There was a decline in locomotor activity over days (F2,28 = 25.66, P < 0.01) and time intervals (F5,70 = 10.19, P < 0.01). However, WT and KO mice showed different rates of decline over days (genotype x day, F2,28 = 5.85, P < 0.01) and time intervals (genotype x time intervals, F5,70 = 5.29, P < 0.01). The three-way interaction was also significant (F10,140 = 2.34, P < 0.05). Newman–Keuls post hoc tests showed that KO mice showed delayed habituation on day 1 but not on other days.
In the marginal area, KO mice showed higher levels of locomotor activity than WT mice (F1,14 = 6.57, P < 0.05) (Fig. 7C). Locomotor activity declined over days (F2,28 = 31.34, P < 0.01) and time intervals (F5,70 = 38.12, P < 0.01). KO mice showed a delayed decline over time intervals compared with WT mice (genotype x time intervals, F5,70 = 6.71, P < 0.01). No other interaction was significant. Newman–Keuls post hoc tests confirmed that KO mice showed a slower rate of habituation within a session on the first 2 days.
Although KO mice traveled more in the central area than did WT mice on the first day (Fig. 7B), this might reflect a higher level of total locomotor activity in KO mice when compared with WT mice; an increased distance traveled results in an increased distance in the center as well as in the margin. Alternatively, KO mice might have shown a disproportionately higher level of activity in the center compared with WT mice. To critically evaluate these possibilities, we examined the ratio of distance traveled in the center to that in the marginal area of the open field (Fig. 7D). When analyzed this way, WT and KO mice did not differ (F1,14 = 1.46, n.s.), and distance traveled declined in the central area relative to the marginal area over days (F2,28 = 7.22, P < 0.01). A two-way interaction was significant between genotype and time interval (F5,70 = 3.68, P < 0.01), but not between genotype and day (F2,28 = 1.51, n.s.). The three-way interaction was not significant (F10,140 = 1.16, n.s.]. Newman–Keuls post hoc tests showed that WT and KO mice did not differ at any time point on any day.
Similar patterns were found when data were additionally analyzed for (i) the ratio of distance traveled in the center to that in the entire field (genotype, F1,14 = 1.29, n.s.; genotype x day, F2,28 = 1.84, n.s.; genotype x time interval, F5,70 = 5.79, P < 0.01; genotype x day x time interval, F10,140 = 0.97, n.s), (ii) the ratio of time spent in the center to that in the marginal area of the open field (genotype, F1,14 = 2.33, n.s.; genotype x day, F2,28 = 1.64, n.s.; genotype x time interval, F5,70 = 4.00, P < 0.01; genotype x day x time interval, F10,140 = 1.03, n.s) and (iii) the ratio of time spent in the center to that in the entire open field (genotype, F1,14 = 3.22, n.s.; genotype x day, F2,28 = 2.38, n.s.; genotype x time interval, F5,70 = 5.20, P < 0.01; genotype x day x time interval, F10,140 = 1.43, n.s.) (data not shown).
FosB KO mice show an exaggerated locomotor response in an inescapable open field following saline injections
An additional set of mice were tested for locomotor activity following saline injections. This procedure was added to alter the level of stress in this task. A saline injection in mice is a stressful event that elevates corticosterone levels and induces hyperactivity (53,54). KO mice showed a higher level of locomotor activity in the entire field compared with WT mice (F1,14 = 25.42, P < 0.01) (Fig. 7E). There was a gradual decline in activity over days (F2,28 = 14.29, P < 0.01) and time intervals (F5,70 = 42.48, P < 0.01). The rate of decline differed between WT and KO mice over days (genotype x day, F2,28 = 9.48, P < 0.01) but not time intervals (genotype x time interval, F5,70 = 0.21, n.s.). A two-way interaction was significant between day and time interval (F10,140 = 3.10, P < 0.01). The three-way interaction was also significant (F10,140 = 2.04, P < 0.05).
KO mice showed higher levels of locomotor activity in the central area of the open field than WT mice (F1,14 = 33.28, P < 0.01) (Fig. 7F). There was a gradual decline in locomotor activity over days (F2,28 = 35.59, P < 0.01) and time intervals (F5,70 = 7.35, P < 0.01). A two-way interaction was significant between genotype and day (F2,28 = 29.37, P < 0.01) and day and time interval (F10,140 = 5.42, P < 0.01), but not between genotype and time interval (F5,70 = 0.09, n.s.). The three-way interaction among genotype, day and time interval was significant (F10,140 = 2.33, P < 0.05).
Overall, WT and KO mice were indistinguishable in the marginal area (F1,14 = 0.02, n.s.) (Fig. 7G). There was a gradual decline in activity over time intervals (F5,70 = 45.17, P < 0.01), but not over days (F2,28 = 2.67, n.s.). The three-way interaction was significant (F10,140 = 4.82, P < 0.01). No two-way interaction was significant.
We next examined how saline injections altered locomotor activity in the central area relative to the marginal area, using the ratio of distance traveled in the center to that in the margin (Fig. 7H). Whereas WT mice clearly avoided the central area, KO mice traveled more in the center than in the marginal area on day 1. KO mice showed higher ratios than WT mice (F1,14 = 12.11, P < 0.01), and ratios changed over days (F2,28 = 11.47, P < 0.01). How WT and KO mice differed depended on day (F2,28 = 15.61, P < 0.01). No other effect was significant. Newman–Keuls tests showed that only on day 1 did KO mice consistently travel more in the center relative to the margin, compared with WT mice.
The same patterns were found when data were additionally analyzed for (i) the ratio of distance traveled in the center to that in the entire field (genotype, F1,14 = 15.64, P < 0.01; genotype x day, F2,28 = 16.10, P < 0.01; genotype x time interval, F5,70 = 1.62, n.s.; genotype x day x time interval, F10,140 = 1.14, n.s.), (ii) the ratio of time spent in the center to that in the marginal area of the open field (genotype, F1,14 = 11.18, P < 0.01; genotype x day, F2,28 = 16.26, P < 0.01; genotype x time interval, F5,70 = 0.48, n.s.; genotype x day x time interval, F10,140 = 0.90, n.s.) and (iii) the ratio of time spent in the center to that in the entire open field (genotype, F1,14 = 16.68, P < 0.01; genotype x day, F2,28 = 18.05, P < 0.01; genotype x time interval, F5,70 = 0.93, n.s.; genotype x day x time interval, F10,140 = 0.62, n.s.) (data not shown).
FosB KO mice show normal novelty preferences
Although WT and KO mice differed in a novel, inescapable open field, this phenotype could result from several trait differences. In fact, hyperactivity in mice is thought to reflect not only novelty exploration but also motor activity, anxiety, and stress reactivity (55–57). We therefore tested the hypothesis that KO mice respond more to novel stimuli than WT mice in paradigms with less stress and anxiety. We used two separate apparatuses to reduce the level of stress and anxiety. The availability of a choice has been shown to reduce stress (56). In the first version, mice were exposed to one of the two compartments for 3 days. On Day 4, we tested their choice between a non-habituated (i.e. novel) compartment and a familiar compartment (Fig. 8A). WT and KO did not differ (genotype, F1,26 = 0.43, n.s.), and both groups equally preferred a novel compartment (compartment, F1,26 = 64.01, P < 0.01; genotype x compartment, F1,26 = 0.44, n.s.). Their preference for a novel compartment was stable throughout the 15 min session (time, F2,52 = 0.03, n.s.) (Fig. 8B). No other interaction was significant.
|
Compared to an experimental apparatus, a home cage is thought to induce lower levels of anxiety and stress (20). In a second novelty task, two identical home cages were attached to each other and a partition door was placed to allow animals to move freely between the two cages. WT and KO mice did not differ from each other in preference to a novel cage (F1,14 = 1.35, n.s.). Both groups equally preferred an unfamiliar, novel home cage (F1,14 = 24.95, P < 0.01) (Fig. 8C). No interaction was found between genotype and compartment (F1,14 = 0.15, n.s.).
FosB KO mice show normal anxiety-related traits
Locomotor activity in the central area of an inescapable open field is thought to partly reflect the animal's anxiety-related behavior (51,52,57). We tested whether fosB KO mice show lower levels of anxiety-related behavior in an elevated plus maze and in a light–dark discrimination task (Table 1).
|
For the elevated plus maze (Table 1), WT and KO mice avoided the open arms equally [genotype:time, t(11) = 0.62, n.s.; frequency, t(11) = 0.52, n.s.].
For the light–dark discrimination task (Table 1), WT and KO mice equally avoided the brightly lit compartment (genotype: time, F1,20 = 1.39, n.s; frequency, F1,20 = 1.23, n.s; compartment: time, F1,20 = 65.69, P < 0.01; frequency, F1,20 = 11.10, P < 0.01). There was no interaction between genotype and compartment for time (F1,20 = 0.36, n.s) or frequency (F1,20 = 1.64, n.s.). Newman–Keuls post hoc tests showed that WT and KO mice did not differ in any of the parameters tested.
FosB KO mice have lower basal corticosterone levels than WT mice
Because cortisol is needed for the preference and subjective effects of nicotine in smokers (58) and the sensitizing effects of nicotine in rodents (59), we measured basal blood corticosterone levels. KO mice showed lower basal blood corticosterone levels than WT mice [t(19) = 2.22, P < 0.05] (Table 2).
|
FosB KO mice exhibit greater mobility in tail suspension test
We next examined immobility scores in the tail suspension test, a task thought to reflect an animal's affective state and stress reactivity (60–62). KO mice showed lower immobility scores (i.e. greater mobility) than WT mice [t(17) = 2.68, P < 0.05] (Fig. 9).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSSUPPLEMENTARY MATERIALREFERENCES |
|---|
Although human studies have provided evidence of a substantial constitutive genetic influence on the susceptibility to nicotine dependence and comorbid behavioral traits, the individual genes contributing to this association remain poorly understood due to the complexity of the influences of multiple genes. By delineating the extent to which fosB KO mice are abnormal in nicotine-regulated behaviors and multiple dependence-related behavioral traits, our results suggest that the constitutive inactivation of fosB results in pleiotropic, yet circumscribed, alterations in the behavioral effects of repeated or prolonged nicotine administration and in stress-related behavioral traits.
Although we used a constitutive KO mouse, it is unlikely that the observed behavioral phenotypes were influenced by compensatory alterations of FosB-related molecules or anatomical abnormalities. fosB KO mice do not exhibit any compensatory upregulation of other Fos family proteins (25,34,63,64). Moreover, no apparent anatomical abnormalities are detected in the dopaminergic neurons, striatal projection neurons, cortical structures, other brain structures or peripheral organs of the fosB KO mouse (25,34,63). Nor does the fosB KO mouse suffer from generalized behavioral impairments. FosB KO mice exhibit no neurological abnormalities in sensory responses to noise, pain or noxious olfactory stimuli (63). FosB KO mice are normal in the Morris water maze task, which measures vision, motor function and complex cognitive or memory processes (65). FosB KO mice are also normal in choice behaviors such as ethanol preference (47), olfactory discrimination (65), CPA (Fig. 1B), novelty preference (Fig. 8), elevated plus maze (Table 1) and light–dark discrimination (Table 1).
The constitutive absence of FosB might have caused subtle neurochemical abnormalities that alter the behavioral effects of nicotine in KO mice. These effects are not particularly disadvantageous, as a constitutive genetic impact is not expected to occur in isolation in humans, either; rather, the genetic impact is likely to occur through alterations in its downstream molecules in mice and humans (20).
Caution is still needed in ascribing the phenotypes of fosB deletion to the single gene deletion, however. One limitation on this interpretation is that the impact of a gene deletion must be understood in terms of its interaction with a specific genetic background. Studies have demonstrated that a single gene deletion produces different, even opposite, phenotypes in different genetic backgrounds (66–69). In fact, we found that a single injection of nicotine is sufficient to establish CPP in mice of the C3H/HeNTac genetic background (20) but not in fosB WT or KO mice (Fig. 1C). Our fosB WT and KO mice were generated after heterozygous mice had been back-crossed to the C57BL/6J mice up to seven times. On the basis of this number of back-crosses, we estimate that the WT and KO mice are homozygous with the C57BL/6J alleles at up to 
98% of loci unlinked to the fosB locus and the small remaining fraction of loci is mixed with the original backgrounds of 129S4/SvJae, 129X1/SvJ and Balb/c. However, the chromosomal regions linked to the fosB gene locus are likely to contain more allelic differences between WT and KO mice (70,71). The phenotypes reported here should be interpreted strictly in terms of the given genetic background composition.
FosB and nicotine-regulated behaviors
Our observations suggest that constitutive fosB deficiency reduces the behavioral effects of repeated and prolonged nicotine administration, but not acute administration. First, consistent with most other studies demonstrating that nicotine induces CPP at a single dose or two close doses within a narrow dose range between 0.175 and 0.5 mg/kg (free base) in any given mouse strain (72–77), repeated nicotine injections at 0.2 mg/kg induced CPP in WT mice. In contrast, KO mice did not develop nicotine CPP following repeated nicotine injections, although KO mice, as well as WT mice, exhibited normal CPA at high doses. An acute, single nicotine injection was ineffective in inducing nicotine CPP in either WT or KO mice. Consistent with these observations, repeated injections of nicotine increased the expression of FosB and FosB in the nucleus accumbens and caudate-putamen of WT mice. In contrast, a single acute nicotine injection (0.2 mg/kg) failed to consistently increase FosB/FosB expression in WT mice. KO mice were devoid of detectable levels of FosB/FosB. Second, WT mice sustained their oral nicotine intake, but KO mice gradually reduced oral nicotine intake over 15 days; WT and KO mice did not differ on the first recording day. Third, fosB KO mice, upon repeated injections, lost the motor suppressant effects of nicotine more quickly than did WT mice, but a single, acute nicotine injection induced the same degree of motor suppression in WT and KO mice.
We found an interesting temporal gap between FosB/FosB induction and the occurrence of conditioning. Each conditioning session lasted for 0.5 h in the place conditioning paradigm. At 0.5 h, repeated nicotine injections did not increase levels of FosB or FosB in the nucleus accumbens, although FosB, but not FosB, was elevated in the caudate-putamen. Instead, expression levels of FosB and FosB were elevated at time points after conditioning was completed. One possible interpretation is that FosB in the caudate-putamen is functionally required for the CPP-inducing effects of nicotine. An alternative possibility is that the delayed induction of FosB/FosB in the nucleus accumbens or in the caudate-putamen is required for the consolidation of nicotine CPP or might cause alterations needed for expression of CPP on the drug-free test day, which occurs a day after the last nicotine injection. As the transcription factor FosB is likely to act as a modulator of its target molecules but not as a direct mediator of nicotine-triggered signaling, such lack of temporal association is not inconsistent with its presumed functional actions.
The mesolimbic dopamine pathway, projecting from the VTA to the nucleus accumbens, mediates many of nicotine's behavioral effects in rodents. Blockade of dopaminergic transmission in the nucleus accumbens attenuates intravenous nicotine self-administration and CPP in rats (78,79), and infusion of nicotine into the VTA establishes self-administration and CPP in mice and rats (80–82). Direct infusion of nicotine into the nucleus accumbens induces motor suppression and repeated infusion diminishes this effect (48,49). Systemic nicotine administration has been shown to increase the expression of Fos-related proteins in the nucleus accumbens; however, because these proteins were identified by an antibody raised against a sequence common to all Fos family members, their precise identity was unclear (83). Our immunoblotting analysis showed regionally selective induction of FosB and FosB using a specific antibody: repeated nicotine injections increased the expression of FosB and FosB proteins in the target regions (the nucleus accumbens and caudate-putamen) of the mesolimbic and nigrostriatal dopamine pathways, but not at the origins (the VTA and SN) (Fig. 6). Although immunoblotting is often not sensitive enough to detect regionally circumscribed induction of FosB/FosB proteins [see Perrotti et al. (27) for discussion], our immunohistochemical analysis avoided this technical limitation. Even at a high dose (0.8 mg/kg, s.c.), nicotine increased the expression of FosB/FosB in the nucleus accumbens and caudate-putamen (except for the dorsomedial quadrant), but not in the VTA or SN (Supplementary Material, Figs S1 and S2). It has been reported that repeated nicotine injections increased FosB/FosB in the nucleus accumbens but not in the caudate-putamen in rats (26). This apparent discrepancy probably reflects the fact that this previous analysis was limited to the dorsomedial corner of the caudate-putamen, where our analysis detected no upregulation of FosB/FosB. Although nicotine might primarily act on the VTA to induce CPP (80), this transcription factor is likely to mediate nicotine's behavioral effects in the nucleus accumbens and caudate-putamen.
The fosB gene produces FosB and FosB by alternative splicing. Both FosB and FosB dimerize with Jun family proteins and bind to AP-1 sites of DNA. However, because FosB lacks the C-terminal region needed for transactivation (84,85), only FosB can directly transactivate downstream genes (84–87). Nevertheless, FosB could indirectly regulate transcription by competing with FosB (84–86) and thus has been termed a ‘FosB antagonist’ (84), ‘dominant negative factor’ (85) or ‘trans-negative regulator’ (86). In fact, overexpression of FosB in the brain leads to regulation of many genes (32,88). FosB overexpression in the nucleus accumbens potentiates the CPP induced by another addictive substance, cocaine (30,32). Together with these observations, our previous demonstration that fosB KO mice show a potentiated CPP to repeated cocaine injections (25) might be interpreted as suggesting that FosB acts as an inhibitory modulator for cocaine CPP and FosB suppresses this inhibitory effect of FosB.
Given the inhibitory role of FosB in cocaine CPP, it is apparently paradoxical that fosB KO mice show reduced behavioral responses to nicotine in the place conditioning paradigm. One possible explanation is that this difference reflects the impact of different genetic backgrounds. The enhanced cocaine CPP was seen in fosB KO mice with the mixed 129S4/SvJae, 129X1/SvJ and Balb/c genetic backgrounds (25). In contrast, the present study used fosB KO mice with a predominantly C57BL/6J background (see Materials and Methods, Mice). However, this difference fails to fully account for the opposite effects of fosB deficiency on nicotine and cocaine CPP, because we have observed that fosB KO mice with the C57BL/6J genetic background are also more sensitive to cocaine in the place conditioning paradigm (data not shown).
The opposite effects of fosB deletion on cocaine and nicotine CPP could reflect a difference in distribution patterns of FosB/FosB in the brain: whereas repeated cocaine administration upregulates FosB/FosB in the VTA, as well as in the accumbens and caudate-putamen (23–25,27,29), nicotine induces FosB/FosB in the nucleus accumbens and caudate-putamen, but not in the VTA or SN [present study, see also Marttila et al. (26)]. Although the exact reason for the contrasting roles of fosB in nicotine and cocaine CPP remains unclear, genetic inactivation of another transcription factor, cAMP-response element binding protein, potentiates the CPP induced by cocaine and blocks the CPP induced by nicotine (75,89,90). Distinct cascades of transcriptional regulation are likely to be involved in the behavioral response to these two addictive substances.
The complete absence of fosB gene products might represent an extreme case; however, relative expression levels of fosB gene products also have been associated with the behavioral effects of nicotine. Lewis inbred rats exhibit higher basal levels of fosB protein products in the nucleus accumbens than Fisher344 inbred rats (91) and Lewis, but not Fisher344, rats develop CPP following repeated nicotine injections (92).
It is inherently difficult to compare nicotine-regulated behaviors in mice and nicotine dependence in humans. However, our findings suggest that two specific aspects of nicotine dependence in humans might be influenced by constitutive FosB levels. Each individual smoker tends to maintain an extremely narrow, optimal blood nicotine concentration (9), and desired levels of nicotine vary widely among smokers (93,94). As genetic FosB inactivation influenced nicotine intake in mice, constitutive FosB levels might contribute to individual levels of smoking in humans. Moreover, cues associated with a substance with the potential for dependence are potent instigators of relapse in humans (95), and this is true for smoking (96–98). The robust effects of conditioned cues on nicotine self-administration have been demonstrated in rats and primates (99–101). The place conditioning paradigm is thought to be a model of approach behaviors toward sensory cues that are associated with a substance with a dependence potential (36,38). Because fosB KO mice were impaired in this task, this gene might also contribute to cue control of nicotine dependence.
FosB and behavioral traits
KO and WT mice did not differ in an inescapable open field at the beginning of each day (Fig. 7A), when the levels of novelty are considered to be highest. Similarly, in novelty tasks that involve a choice, KO mice showed normal approach responses toward novel stimuli (Figs. 8A and C). These results suggest that the constitutive absence of fosB does not affect the response to novel stimuli.
KO mice were impaired in locomotor habituation in an open field within and across sessions (Fig. 7A). Locomotor habituation in an open field is considered to reflect the animal's adaptive reaction to a novel environment (55). Repeated exposure to an open field induces lower rates of locomotor habituation and lower induction levels of FosB/FosB in the prefrontal cortex and caudate-putamen of DBA inbred mice compared with C57BL/6J mice (28). FosB expressed in these brain structures might therefore play a role in habituation. It should be noted, however, that a learning deficit seen as impaired locomotor habituation in KO mice does not fully account for the absence of CPP, as mice could develop normal CPP in the presence of a severe deficit in locomotor habituation (102). Although FosB expression is associated with the development of nicotine CPP and locomotor habituation, the mechanisms through which FosB contributes to these two types of behavioral alterations are not identical.
Although the habituation occurred more slowly in KO mice than in WT mice in the open field (Fig. 7A), this phenotype was not seen in the novelty preference test (Fig. 8B). The complex nature of the genetic contribution to the animal's response to novelty has also been demonstrated for other genes. Mice lacking the dopamine D4 receptor display normal locomotor activity in an open field but approach novel objects less frequently than WT mice (103). Complementing this finding, monoamine oxidase A KO mice show delayed locomotor habituation in a novel, inescapable open field but are normal in a novel compartment preference, compared with WT mice (20). Taken together, these results suggest that non-identical genetic bases exist for behaviors seen in various novelty tasks or that the so-called ‘novelty’ tasks include behavioral components other than novelty responses (104,105).
Spontaneous locomotor activity in the central area of the inescapable open field is thought to reflect, at least partly, an anxiety-related trait (51,52). However, the phenotypic difference between WT and KO mice seen in the center of the open field does not reflect an abnormality in anxiety-related behaviors. When we analyzed the ratio of center activity to margin activity, KO and WT mice were indistinguishable (Fig. 7D), suggesting that the slightly higher level of locomotor activity in KO mice in the central area does not reflect a selective increase in activity in that area but rather reflects generalized hyperactivity in the open field. Moreover, when mice had a choice of exploring or avoiding an anxiety-provoking situation (i.e. elevated plus maze and brightness–darkness discrimination task), fosB KO mice and WT mice were indistinguishable (Table 1).
Immobility in the tail suspension test is thought to be triggered by the hemodynamic stress that is inherent in this test (61). The immobility seen in this task is sensitive to antidepressants and thus might represent a behavior relevant to depression (61). However, this interpretation cannot be easily applied to the comparison of WT and KO mice for two reasons. First, the hyperactivity seen in KO mice under a stressful situation (Fig. 7E) might have non-selectively reduced the immobility score without altering the affective state of the mice. Second, the absolute affective states of the mice cannot be determined, and a relatively higher level of immobility in WT mice might simply reflect a less robustly positive affective state rather than ‘dysphoria’ or ‘depression’.
Our profiling of the behavioral traits in multiple tasks instead suggests a potentially common meta-phenotype of fosB KO mice: KO mice showed more pronounced abnormalities in tasks that include high levels of stress. First, when saline injections were given, KO mice showed a more exaggerated hyperactivity in the inescapable open field than did WT mice. Saline injections are a stressful event in mice and increase corticosterone levels (53). Unlike comparisons among different behavioral tasks, locomotor activity was measured under identical conditions except for the presence or absence of saline injections, thus allowing us to ascribe the phenotypic difference between the two conditions to the stress level. Second, KO mice showed lower immobility scores than WT mice in the tail suspension task (see Fig. 9), which includes stress reactivity (62). Third, KO mice had lower basal corticosterone levels than WT mice, complementing a reported association between high immobility scores and high basal corticosterone levels in diabetic mice (106). The observed association between a low basal corticosterone level and heightened locomotor activity in fosB KO mice is also seen in mice lacking corticotropin-releasing hormone receptor (107). Fourth, when a choice was provided (i.e. novel cage or compartment preference, elevated plus maze or brightness–darkness discrimination), KO and WT mice were indistinguishable. The level of stress is much lower, as measured by blood corticosterone levels, when mice are provided a choice than when they are not (56,108).
Stressful stimuli increase expression of FosB/FosB proteins in diverse brain regions, including the prefrontal cortex, ventral orbital cortex, nucleus accumbens, caudate-putamen, septum, endopiriform nuclei and amygdala (34,35). Our results show that nicotine also increases FosB/FosB proteins in two of these structures, the nucleus accumbens and caudate-putamen. These structures might be sites at which constitutive levels of FosB act as a determinant for both the behavioral effects of nicotine and stress reactivity.
FosB KO mice had a lower basal level of corticosterone and exhibited a downward shift in the dose–response curve of nicotine CPP and intake. Similarly, studies have demonstrated that a certain basal level of cortisol and corticosterone are needed for the subjective effects of smoking in humans (58) and locomotor sensitizing effects of nicotine in rodents (59), respectively. Corticosterone is also needed for self-administration of other addictive substances such as amphetamine and cocaine in rats (109,110). More work is needed to test the possibility that FosB contributes to cue control of nicotine dependence and individual rates of nicotine preference/intake in the presence of a certain level of corticosterone.
In summary, our study shows that a constitutive deficiency of fosB alters the behavioral effects of repeated or prolonged nicotine administration and stress-provoked behaviors. Our results strengthen the suggestion that a single gene could have a pleiotropic influence on the behavioral effects of nicotine and related behavioral traits (11,20,45). Given that nicotine dependence is thought to be influenced by multiple genes, many genes might exert such a pleiotropic effect on susceptibility to distinct aspects of nicotine dependence and comorbid traits. Moreover, as the protein products of fosB are thought to transcriptionally regulate a large number of target genes, our study provides a foundation to explore many potential therapeutic molecular targets for nicotine dependence.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSSUPPLEMENTARY MATERIALREFERENCES |
|---|
Mice
Wild-type (WT) and fosB KO mice were generated by crossing non-sibling fosB heterozygotes. Age-matched littermates were used at 2–5 months of age for all experiments except for the analysis of nicotine-induced motor suppression, for which 2–8-month-old KO mice and WT littermates were used. This mouse line originally had a mixed genetic background of 129S4/SvJae, 129X1/SvJ and Balb/c (65) and was subsequently back-crossed to C57BL/6J mice for up to seven generations. Up to five mice were housed per cage and were maintained under a 14 h light:10 h dark cycle with food and water available ad libitum. Animal handling and use followed a protocol approved by the Animal Care and Use Committee of Albert Einstein College of Medicine, in accordance with NIH guidelines.
Drugs
(–)-Nicotine hydrogen tartrate salt (Sigma Chemical Co., St Louis, MO, USA) dissolved in 0.9% saline was injected s.c. at a volume of 2.0 ml/kg. For oral intake, (–)-nicotine free-base solution (99%, Sigma Chemical Co.) was dissolved in tap water. All doses are expressed as free base.
Behavioral analyses
Place conditioning paradigm
The apparatus included two large compartments (each 24.5 cm x 18 cm x 33 cm) that differed in wall color and ceiling light intensity [see Agatsuma et al. (20) for details]. These compartments were separated by guillotine doors from a third, central compartment (13 cm x 18 cm x 33 cm) that allowed animals to move freely between the two large compartments. The apparatus was designed such that on a preconditioning day animals showed no baseline preference for either compartment.
Male fosB KO mice (n = 5–17 for each dose) and WT littermates (n = 6–20 for each dose) were used for this analysis. The place conditioning procedure was carried out over 5 days. On day 1 (the preconditioning day), animals were placed in the central compartment of the apparatus and were allowed to move freely for 15 min between the two large compartments through the open guillotine doors. On each of days 2–4 (conditioning days), animals received two conditioning sessions 5 h apart. Each session consisted of administering either saline or nicotine hydrogen tartrate salt (0.025, 0.05, 0.2, 0.6, 0.8 or 2.0 free base mg/kg, s.c.) (n = 5–20 mice per genotype per dose) followed by confinement (by closed guillotine doors) for 30 min to one of the two large compartments. We previously demonstrated that pairings of saline injections with both compartments of this apparatus do not produce preference or aversion to either compartment (20). Approximately equal numbers of animals received pairings with nicotine in each compartment. Each animal received one nicotine injection and one saline injection each day; the order of nicotine injections was also counterbalanced so that approximately half of each dose group received nicotine in the morning sessions and the other half in the afternoon sessions. On day 5 (the test day), animals were placed in the center compartment with no injections and allowed to move freely for 15 min between the two large compartments through open guillotine doors. An observer who was blinded to genotype and treatment recorded the amount of time each animal spent in each of the two large compartments.
To assess the acute effects of nicotine, a separate group of male mice (WT, n = 22; KO, n = 13) received one pairing with nicotine hydrogen tartrate salt (0.2 free base mg/kg, s.c.) and one pairing with saline on day 2, and animals were tested for preference with no injections on day 3. Otherwise, the procedure was identical to the standard method described above.
Nicotine oral intake
A two-bottle choice paradigm was used to study nicotine-dependent drinking behavior, as previously described (20,45,46). Male fosB KO mice (n = 7–12 for each concentration) and WT littermates (n = 10–19 for each concentration) were used for this analysis. Mice were housed individually in home cages at the onset of this experiment. A water bottle containing a single concentration (0, 12.5, 25 or 50 free base µg/ml) of alkaline (–)-nicotine solution was given in parallel with a bottle containing tap water for 15 days. Each animal was tested at a single concentration of nicotine. Each bottle and each mouse were weighed and fresh solution was given between 1:00 and 3:00 p.m. every 3 days. Nicotine consumption, expressed in mg/kg over 3 days, and the ratio of nicotine preference to total fluid intake were calculated for each 3-day period.
Taste response tests
Male fosB KO mice (saccharin, n = 8; quinine, n = 7) and WT littermates (saccharin, n = 8; quinine, n = 8) were tested for response to the sweet taste of saccharin (3.275 mM) and the bitter taste of quinine (1 mM) in the two-bottle choice procedure, as described above. The ratios of saccharin preference or quinine aversion to total fluid intake were calculated.
Nicotine-induced motor suppression
Locomotor activity was measured in four sets of automated open fields made of transparent Plexiglas (26 cm x 26 cm x 38.5 cm), each with two sets of 34 pairs of photocell beams (Truscan, Coulburn Instrument, PA, USA). Illumination was provided by the fluorescent light on the ceiling of the room, resulting in an intensity of