Divergent cellular pathways of hippocampal memory consolidation and reconsolidation

Abstract

The reconsolidation of memories after their retrieval involves cellular mechanisms that recapitulate much of the initial consolidation process. However, we have previously demonstrated that there are independent cellular mechanisms of consolidation and reconsolidation in the dorsal hippocampus for contextual fear memories. Expression of BDNF was required for consolidation, while Zif268 expression was necessary for reconsolidation. Given the dichotomy between the obvious mechanistic similarity and notable dissociations between consolidation and reconsolidation, we sought to determine whether the separation at the level of gene expression reflected either parallel and independent upstream signaling pathways, or common upstream mechanisms that diverge by the level of transcriptional activation. Here we show that while consolidation and reconsolidation are commonly dependent upon NMDA receptor activation in the dorsal hippocampus there is a double dissociation between the effects of the MEK inhibitor U0126 and the IKK inhibitor sulfasalazine. Moreover, rescue experiments and western blot analyses show that there are functional NMDA receptor-ERK1-BDNF and NMDA receptor-IKKα-Zif268 pathways for consolidation and reconsolidation, respectively. Therefore, there are divergent pathways of hippocampal memory consolidation and reconsolidation, involving commonality at the cell surface, but separable downstream kinase cascades and transcriptional regulation.

Fig. 1

Schematic representation of the locations of injectors within the dorsal hippocampus. All injector tips across the behavioral experiments were located within the grey shaded area.

Fig. 2

Intra-dorsal hippocampal infusions of D-AP5 and U0126, but not sulfasalazine, impair contextual fear conditioning. A, Pre-conditioning D-AP5 impaired contextual freezing at 3 hr (STM), 24 hr (LTM) and 8 days (LTM2) after conditioning. B, Pre-conditioning U0126 impaired contextual freezing at the LTM and LTM2 tests, but not the STM test. C, Post-conditioning Sulfasalazine had no impact upon contextual freezing at either the STM or LTM test. Data presented as mean + SEM.

Fig. 3

Western blot analysis of conditioning-induced phosphorylation of ERK1 and IKKα in the dorsal hippocampus. Rats were either latently inhibited to the context prior to contextual fear conditioning (LI) or were subjected to fear conditioning alone (cond). Quantification of phospho-protein involved normalisation against the levels of total protein. Quantification of total protein involved normalisation against levels of actin. Representative bands for the group effects are presented. A, Contextual fear conditioning did not induce an increase in phospho-IKKα (pIKKα). B, Conditioning did not alter total IKKα (tIKKα) levels. C, Contextual fear conditioning resulted in an increase in phospho-ERK1 (pERK1). D, Conditioning did not alter total ERK1 (tERK1) levels. E, Infusion of D-AP5 into the dorsal hippocampus prior to conditioning attenuated the phosphorylation of ERK1 relative to PBS control. F, D-AP5 infusion had no effect on total ERK1 levels. Data presented as mean + SEM.

Fig. 4

Rescue of conditioning deficits with recombinant BDNF. A, D-AP5 and BDNF were infused into the dorsal hippocampus alone or in a mixed solution prior to contextual fear conditioning. While D-AP5 impaired contextual freezing at 3 hr (STM), 24 hr (LTM) and 8 days (LTM2) after conditioning compared to PBS control, co-infusion of BDNF rescued the deficits at each test. BDNF alone did not affect conditioning. B, U0126 and BDNF were infused into the dorsal hippocampus alone or in a mixed solution prior to contextual fear conditioning. While U0126 impaired contextual freezing selectively at the LTM and LTM2 tests compared to Vehicle control, co-infusion of BDNF rescued the deficits. BDNF alone did not affect conditioning. Data presented as mean +/− SEM.

Fig. 5

Intra-dorsal hippocampal infusions of D-AP5 and sulfasalazine, but not U0126, impair contextual fear memory reconsolidation. A, Pre-reactivation D-AP5 impaired contextual freezing at 3 hr (PR-STM), 24 hr (PR-LTM) and 8 days (PR-LTM2) after memory reactivation (LTM). B, Infusion of D-AP5 in the absence of memory reactivation had no impact upon subsequent contextual freezing. C, Post-reactivation sulfasalazine impaired contextual freezing at the PR-LTM and PR-LTM2 tests, but not the PR-STM test. D, Infusion of sulfasalazine in the absence of memory reactivation had no impact upon subsequent contextual freezing. E, Pre-reactivation U0126 had no impact upon subsequent contextual freezing. Data presented as mean + SEM.

Fig. 6

Western blot analysis of reactivation-induced phosphorylation of IKKα and ERK1 in the dorsal hippocampus. Rats were contextually fear conditioned and then either sacrificed without memory reactivation (NR) or subjected to memory reactivation (R). Quantification of phospho-protein involved normalisation against the levels of total protein. Quantification of total protein involved normalisation against levels of actin. Representative bands for the group effects are presented. A, Memory reactivation did not induce an increase in phospho-ERK1 (pERK1). B, Reactivation did not alter total ERK1 (tERK1) levels. C, Memory reactivation resulted in an increase in phospho-IKKα (pIKKα). D, Reactivation did not alter total IKKα (tIKKα) levels. E, Infusion of D-AP5 into the dorsal hippocampus prior to memory reactivation attenuated the phosphorylation of IKKα. relative to PBS control. F, D-AP5 infusion had no effect on total IKKα levels. G, Infusion of D-AP5 into the dorsal hippocampus prior to memory reactivation attenuated the upregulation of Zif268 relative to PBS control. H, Infusion of Sulfasalazine into the dorsal hippocampus prior to memory reactivation attenuated the upregulation of Zif268 relative to PBS control. Data presented as mean + SEM.