Hippocampal granule cells opt for early retirement.

Increased excitability and plasticity of adult-generated hippocampal granule cells during a critical period suggests that they may "orthogonalize" memories according to time. One version of this "temporal tag" hypothesis suggests that young granule cells are particularly responsive during a specific time period after their genesis, allowing them to play a significant role in sculpting CA3 representations, after which they become much less responsive to any input. An alternative possibility is that the granule cells active during their window of increased plasticity, and excitability become selectively tuned to events that occurred during that time and participate in later reinstatement of those experiences, to the exclusion of other cells. To discriminate between these possibilities, rats were exposed to different environments at different times over many weeks, and cell activation was subsequently assessed during a single session in which all environments were revisited. Dispersing the initial experiences in time did not lead to the increase in total recruitment at reinstatement time predicted by the selective tuning hypothesis. The data indicate that, during a given time frame, only a very small number of granule cells participate in many experiences, with most not participating significantly in any. Based on these and previous data, the small excitable population of granule cells probably correspond to the most recently generated cells. It appears that, rather than contributing to the recollection of long past events, most granule cells, possibly 90-95%, are effectively "retired." If granule cells indeed sculpt CA3 representations (which remains to be shown), then a possible consequence of having a new set of granule cells participate when old memories are reinstated is that new representations of these experiences might be generated in CA3. Whatever the case, the present data may be interpreted to undermine the standard "orthogonalizer" theory of the role of the dentate gyrus in memory.

Single cell analysis of activity-dependent cyclic AMP-responsive element-binding protein phosphorylation during long-lasting long-term potentiation in area CA1 of mature rat hippocampal-organotypic cultures.

Phosphorylation of the transcription factor cyclic AMP (cAMP)-response element-binding protein (CREB) has been implicated in long-term synaptic plasticity and memory, and its activation has been proposed to be required for the maintenance of long-term potentiation (LTP). The previously described temporal dynamics of CREB phosphorylation during the maintenance of LTP showed differences between experimental models. In the present study the level of CREB phosphorylation was evaluated in organotypic hippocampal slices from young adult rats (P25-30) after long-lasting LTP was induced. Immunohistochemistry and confocal imaging were used to determine the ratio between non-phosphorylated and phosphorylated CREB at a single cell resolution, revealing not only the temporal dynamics but also the extent of CREB phosphorylation. The activation of CREB after LTP-induction was compared with cAMP-activation after bath application of forskolin. An increase in cAMP by forskolin resulted in a persistent, uniform increase of the phosphorylated CREB (pCREB/CREB immunofluorescence ratio) in all hippocampal principal neurons. In contrast, the induction of long-lasting LTP in CA1 was accompanied by a local increase in the pCREB/CREB ratio. Both CREB activation and LTP induction in mature cultured slices required N-methyl-D-aspartate (NMDA) receptor activation. CREB phosphorylation continued to increase for 4 h during LTP maintenance. This sustained activation is in contrast to previous observations in acutely prepared slices and supports the hypothesis that CREB plays an important role during the late phases of LTP.