Consolidation of an olfactory memory trace in the olfactory bulb is required for learning-induced survival of adult-born neurons and long-term memory

Abstract

Background: It has recently been proposed that adult-born neurons in the olfactory bulb, whose survival is modulated by learning, support long-term olfactory memory. However, the mechanism used to select which adult-born neurons following learning will participate in the long-term retention of olfactory information is unknown. We addressed this question by investigating the effect of bulbar consolidation of olfactory learning on memory and neurogenesis.

Methodology/principal findings: Initially, we used a behavioral ecological approach using adult mice to assess the impact of consolidation on neurogenesis. Using learning paradigms in which consolidation time was varied, we showed that a spaced (across days), but not a massed (within day), learning paradigm increased survival of adult-born neurons and allowed long-term retention of the task. Subsequently, we used a pharmacological approach to block consolidation in the olfactory bulb, consisting in intrabulbar infusion of the protein synthesis inhibitor anisomycin, and found impaired learning and no increase in neurogenesis, while basic olfactory processing and the basal rate of adult-born neuron survival remained unaffected. Taken together these data indicate that survival of adult-born neurons during learning depends on consolidation processes taking place in the olfactory bulb.

Conclusion/significance: We can thus propose a model in which consolidation processes in the olfactory bulb determine both survival of adult-born neurons and long-term olfactory memory. The finding that adult-born neuron survival during olfactory learning is governed by consolidation in the olfactory bulb strongly argues in favor of a role for bulbar adult-born neurons in supporting olfactory memory.

Figure 1. Only spaced (but not massed) learning allowed long-term retention of an associative olfactory task and increased neurogenesis.

A. Experimental design. For the spaced conditioning (Ai) BrdU was injected 14 days prior to training which occurred over 5 days (4 trials per day). For the massed conditioning (Aii) BrdU was injected from day 17 to 14 prior to training which occurred during a single day (30 successive trials). In both the spaced and massed training, the mice had to learn to use an olfactory cue to find their reward. 5 days post training, retention of the task was assessed during 4 trials performed under the same conditions as during the learning phase. B. Behavioral performance was assessed by measuring the latency to find the reward in groups of conditioned (C) and pseudo-conditioned animals (PC) where the odor was pseudo-randomly associated with the reward. In the spaced training paradigm, latency decreased in the conditioned but not the pseudo-conditioned mice indicating that the conditioned animals had learned the task. 5 days post-training, the conditioned mice remembered the task (C versus PC, bilateral t-test, p<0.001) (Bi). In the massed training paradigm, the performance of the conditioned animals decreased from block 1 to block 7 (4 trials per block for B1 to B5 and 5 trials per block for B6 and B7) and differed from that of pseudo-conditioned animals. Massed conditioned animals did not remember the task 5 days post training (Bii). C. Adult-born cell counts. The density of BrdU-positive cells (Ci) in the granule cell layer of the OB was increased in the conditioned animals of the spaced group (Cii) but not the massed trained animals (Ciii) compared to their respective pseudo-conditioned groups. *:p<0.05; ***:p<0.001; ns: non-significant (p>0.05).

Figure 2. Neuronal differentiation in conditioned (C), pseudo-conditioned (PC) and naive animals.

A. Representative BrdU/NeuN double-labeled cell with orthogonal views. B. No difference in percentage of double-labeled cells was found between spaced and massed trained animals (group effect, F(3,8) = 0.231, p = 0.87) (Bi), nor between saline and anisomycin injected animals (group effect, F(3,5) = 0.512, p = 0.69) (Bii).

Figure 3. Anisomycin infusion in the olfactory bulb during the spaced learning blocked improvements in performance and increase in neurogenesis.

A. Experimental design. BrdU was injected 14 days before training. Animals underwent spaced olfactory associative learning and were infused after each training session with either anisomycin or saline. Naive untrained animals were similarly infused with anisomycin or saline. A retention test was performed 5 days post-training. B. Behavioral performance. Conditioned saline-infused (saline C) animals learned the task as shown by the decrease in latency and remembered it after 5 days. In contrast, conditioned anisomycin-infused (aniso C) animals did not show any change in latency. C. Trial by trial analysis of the learning curve further showed that anisomycin-infused animals returned to pre-training performance between each training session in contrast to saline-infused animals. However, they showed within-session learning. Black arrows symbolize post-training bulbar infusions. D. Adult-born cell counts. Conditioning increased BrdU-positive cell density in saline-infused animals. The infusion of anisomycin prevented this effect without affecting the basal rate of neurogenesis. *: p<0.05; **: p<0.01; ns: non-significant (p>0.05).

Figure 4. Anisomycin infusion did not alter locomotion, odor detection or exploratory behavior.

A. The videotracking assessment of locomotion on Day 5 of conditioning showed no difference between the saline- and anisomycin-infused animals. B. Spontaneous exploratory behavior of an odorized (+limonene: Lim) versus a non- odorized hole (mineral oil: MO) was measured during a non-reinforced trial in naive untrained animals. Animals differentially explored the two holes, indicating that they detected the odor. In this case, they spontaneously avoided the odorized hole. C. The total time spent by the animals exploring the two holes was also recorded as an index of their exploratory behavior and showed no difference between groups. *: p<0.05; ns: non-significant (p>0.05).