Inducible forebrain-specific ablation of the transcription factor Creb during adulthood induces anxiety but no spatial/contextual learning deficits

Abstract

The cyclic AMP (cAMP)-response element binding protein (CREB) is an activity-dependent transcription factor playing a role in synaptic plasticity, learning and memory, and emotional behavior. However, the impact of Creb ablation on rodent behavior is vague as e.g., memory performance of different Creb mutant mice depends on the specific type of mutation per se but additionally on the background and learning protocol differences. Here we present the first targeted ablation of CREB induced during adulthood selectively in principal forebrain neurons in a pure background strain of C57BL/6 mice. All hippocampal principal neurons exhibited lack of CREB expression. Mutant mice showed a severe anxiety phenotype in the openfield and novel object exploration test as well as in the Dark-Light Box Test, but unaltered hippocampus-dependent long-term memory in the Morris water maze and in context dependent fear conditioning. On the molecular level, CREB ablation led to CREM up regulation in the hippocampus and frontal cortex which may at least in part compensate for the loss of CREB. BDNF, a postulated CREB target gene, was down regulated in the frontal lobe but not in the hippocampus; neurogenesis remained unaltered. Our data indicate that in the adult mouse forebrain the late onset of CREB ablation can, in case of memory functionality, be compensated for and is not essential for memory consolidation and retrieval during adulthood. In contrast, the presence of CREB protein during adulthood seems to be pivotal for the regulation of emotional behavior.

Keywords: BDNF; CREB; CREM; anxiety; learning; memory; mouse; tamoxifen induction.

Figure 1

Expression of CREB in hippocampus, nucleus accumbens and amygdala. (A) Hippocampal expression of CREB in (tamoxifen-treated) control animals with regular CREB expression and (B) in Creb^CamKCreERT2 mutant mice after induced ablation with tamoxifen in the adulthood with a definite loss of CREB expression in the pyramidal cell layer but not in putative GABAergic interneurons sparsely distributed in all hippocampal layers. (C) Expression of CREB in the nucleus accumbens (NAC) important for rewarding mechanisms in (tamoxifen-treated) control animals with regular CREB expression and (D) in Creb^CamKCreERT2 mutant mice after induced ablation with tamoxifen in the adulthood. (E) Expression of CREB in the basolateral amygdala (BLA) important for fear-related memory in control and (F) in Creb^CamKCreERT2 mutant mice after induced ablation with tamoxifen in the adulthood. N = 12 per genotype.

Figure 2

Locomotion and Novel object exploration in the openfield. (A) Distance moved and (B) velocity is unaltered in Creb^CamKCreERT2 mutant mice compared to controls in the openfield without object (first 10 min) and after introduction of the novel object. (C) Center time and (D) mean distance to walls is reduced in Creb^CamKCreERT2 mutant mice, reflecting anxiety-like behavior. (E) Latency to approach the novel object is enhanced (p = 0.059) and (F) the number of approaches is decreased in Creb^CamKCreERT2 mutant mice compared to controls. Black bars: controls, white bars: Creb^CamKCreERT2 mutant mice, * p < 0.05, N = 24 per genotype.

Figure 3

Anxiety-like behavior in the Dark-Light Box. (A) The latency to enter the light compartment is enhanced and the (B) number of exits is reduced in Creb^CamKCreERT2 mutant mice compared to controls. (C) Time in the bright compartment is unaltered in Creb^CamKCreERT2 mutant mice compared to controls. Black bars: controls, white bars: Creb^CamKCreERT2 mutant mice, * p < 0.05, *** p < 0.001, N = 24 per genotype.

Figure 4

Learning and Memory performance in Morris water maze and contextual fear conditioning. (A) Mutant mice with a loss of Creb in the adulthood do not show disturbed learning during acquisition in the Morris water maze as represented by the distance moved to reach the platform (repeated measurement ANOVA factor time: F_(11,132) = 18.975, p < 0.001). All data points represent the mean (±SEM), N = 8 controls and N = 7 mutants. (B) Induced loss of Creb in the adulthood does not affect spatial reference memory in the Morris water maze as represented by the time spent in the target zone vs. the means of the 3 other quadrants during the probe trial (two-way ANOVA factor place: F_(1,13) = 19.122, p < 0.001, factor genotype: F_(1,13) = 1.752, p = 0.208). All data points represent the mean (+SEM), N = 8 controls and N = 7 mutants. (C) Induced loss of Creb in the adulthood does not affect hippocampus-dependent associative learning in the fear conditioning paradigm. All data points represent the mean (+SEM) percent time spent freezing during context replacement 24 h after training. N = 14 controls, N = 10 mutants.

Figure 5

Analysis of CREM, ATF-1 and BDNF- mRNA level in hippocampus and frontal cortex. (A) Mutant mice with a loss of Creb at adulthood exhibit significant higher mRNA levels of CREM in the hippocampus and frontal cortex. (B) Induced loss of Creb did not affect ATF-1 mRNA level in hippocampus and frontal cortex. (C) Mutant mice express significantly less BDNF mRNA in the frontal cortex but its levels remain unchanged in the hippocampus. All data points represent mean (+SEM), stated as fold change of the control level. Black bars: controls, white bars: Creb^CamKCreERT2 mutant mice, ** p < 0.01, *** p < 0.001, N = 12 per genotype.