Electrophysiological results from in vitro brain slices showed that synaptic potentiation induced by multiple theta burst stimulation was significantly reduced in Egr-1 knockout mice, suggesting the possible contribution of Egr-1 to memory related synaptic potentiation^[41]. Since it has been noted that significant and measurable levels of Egr-1 are detected after LTP-inducing stimulation, this suggests that Egr-1 may contribute, either directly or indirectly, to signaling pathways involved in synaptic potentiation. Although there was no difference in early synaptic potentiation in the hippocampus^[41], Egr-1 deletion affected early synaptic potentiation induced by theta burst stimulation in the amygdala. Future studies are clearly needed to investigate the synaptic mechanism for Egr-1 in early LTP (45~60 min).

Many studies focus on the interaction between early synaptic potentiation and long-term fear memory. These processes are obviously mediated by different time courses. In the case of synaptic plasticity in brain slices, most LTP lasts from several minutes to hours. For fear memory, behavioral responses usually last from days to weeks. It is likely that synaptic LTP in the amygdala and/or hippocampus contributes to the formation or induction of fear memory during early time periods, while late-phase plastic changes, including possible changes in cortical areas, may contribute to long-term storage of memory. Based on our present findings related to synaptic potentiation in the amygdala, we propose that defects in early potentiation may contribute to the early formation of long-term fear memory. Changes in potentiation in cortical areas, including the ACC, may contribute to long-term fear memory. We believe that future studies are needed to investigate detailed synaptic mechanisms for the loss of LTP in Egr-1 knockout mice. Since both NMDA receptors and L-type calcium channels are reported to contribute to synaptic plasticity^[61], it is of obvious importance for future studies to determine if Egr-1 may regulate the activity of NMDA receptors and L-type calcium channels in central neurons.

It has been widely reported that many physiological and pathological stimuli are able to activate immediate early genes in central nuclei. While the exact physiological functions of these immediate early genes remain to be investigated, activation of immediate early genes has proven to be useful for studying activity-triggered plasticity in the central nervous system. Since selective pharmacological antagonists for immediate early genes are not available, genetically manipulated mice become a good tool to investigate their physiological roles. CREB is probably the most thoroughly studied immediate early gene. The role of CREB in synaptic plasticity and its related physiological and pathological changes have been reported^[16]. In the present study, we found that Egr-1 plays a selective role in the late-form of auditory fear memory. Furthermore, we show that Egr-1 knockout mice have reduced anxiety-like behavior in the elevated plus maze paradigm. This suggests that Egr-1 is an essential component in the pathway from sensory input and interpretation of a fear-producing situation, to the manifestation of an emotional response.

Our findings that Egr-1 knockout mice have intact short term fear memory agrees with numerous studies that report a selective role for both Egr-1 and CREB in the late phase of LTP or learning and memory^{[16,20,21,42]}. A recent study using an Egr-1 antisense oligonucleotide to selectively knockout down expression in the hippocampus of rats reported that while Egr-1 was required for the reconsolidation of contextual memory, knockdown of Egr-1 before conditioning had no affect on behavioral responses 24 h later^[47]. Additionally, levels of Egr-1 expression in the anterior cingulate cortex were increased after exposure to the context 36 d, but not 1 d, after receiving multiple footshocks^[44]. Results from these and the present study suggest a role for Egr-1 in the reconsolidation of long-term memory.

In the present series of experiments, we employed the use of background contextual conditioning, i.e. the tone (conditioned stimulus) was the primary cue while the context remained static, as opposed to foreground conditioning which occurs in the absence of a conditioned stimulus so that the static contextual cues are brought into the foreground by becoming associated with the shock (unconditioned stiumulus)^[62]. Studies have shown that foreground contextual conditioning can activate Egr-1 expression in the amygdala^[43,57]. Additionally, lesion studies show that background contextual conditioning is hippocampus dependent^[62]. Since only background contextual conditioning was performed here, future studies should evaluate foreground conditioning in Egr-1 knockout mice.

Compared to traditional fear memory, the molecular mechanisms for trace memory are less investigated. A recent report used a selective hippocampal deletion of NMDA receptors to show that hippocampal structures are likely required for the formation of trace memory in mice^[50]. This study shows that trace memory is not affected by the deletion of Egr-1, while amygdala-dependent auditory fear memory was significantly reduced. Our results thus provide evidence that different signaling molecules from different regions of the central nervous system may contribute to trace vs classical fear memory as proposed previously^[50]. In summary, the present study provides evidence that the immediate early gene Egr-1, which is activated by a fearful conditioning shock, contributes to late auditory fear memory. These results show that Egr-1, in addition to CREB, may act as an important immediate early gene in emotional fear memory.

* * *

ACKNOWLEDGEMENTS: We would like to thank Ivan Wine for assisting in the design and layout of Fig. 1 and 2.

REFERENCES

1 Davis CM, Ambros-Ingerson J, Granger R, Wu J, Zabaneh R, Abdelnaby M, Lynch G. Triazolam impairs delayed recall but not acquisition of various everyday memory tasks in humans. Exp Neurol 1997; 148: 124-134.[PubMed abstract]

2 Davis M. The role of the amygdala in fear and anxiety. Annu Rev Neurosci 1992; 15: 353-375.[PubMed abstract]

3 LeDoux JE. Emotion circuits in the brain. Annu Rev Neurosci 2000; 23: 155-184.[PubMed abstract]

4 Maren S, Holt W. The hippocampus and contextual memory retrieval in Pavlovian conditioning. Behav Brain Res 2000; 110: 97-108.[PubMed abstract]

5 Frankland PW, Ohno M, Takahashi E, Chen AR, Costa RM, Kushner SA, Silva AJ. Pharmacologically regulated induction of silent mutations (PRISM): combined pharmacological and genetic approaches for learning and memory. Neuroscientist 2003; 9: 104-109.[PubMed abstract]

6 Malenka RC, Nicoll RA. Long-term potentiation —a decade of progress? Science 1999; 285: 1870-1874.[PubMed abstract]

7 Bredt DS, Nicoll RA. AMPA receptor trafficking at excitatory synapses. Neuron 2003; 40: 361-379.[PubMed abstract] (责任编辑：泉水)