Long-lasting LTP requires neither repeated trains for its induction nor protein synthesis for its development

Abstract

Current thinking about LTP triggered in the area CA1 of hippocampal slices is ruled by two "dogmas": (1) A single train of high-frequency stimulation is sufficient to trigger short-lasting LTP (1-3 h), whereas multiple trains are required to induce long-lasting LTP (L-LTP, more than 4 h). (2) The development of the late phase of L-LTP requires the synthesis of new proteins. In this study, we found that a single high-frequency train could trigger an LTP lasting more than 8 h that was not affected by either anisomycin or cycloheximide (two inhibitors of protein synthesis). We ascertained that the induction of this L-LTP made use of the same mechanisms as those usually reported to be involved in LTP induction: it was dependent on NMDA receptors and on the activation of two "core" kinases, CaMKII and PI3K. These findings call into question the two "dogmas" about LTP.

Figure 1. Induction of an L-LTP lasting for more than 8 h using a single train of high-frequency stimulation.

(A1, A2) Sketch showing the two independent synaptic inputs S1 and S2 to the same neuronal population. In each slice, two stimulating electrodes (S1 and S2) were put in place. S1 pathway was used to induce LTP, while S2 pathway acted as a control. (B, C) LTP induced by a single train of high-frequency stimulation (100 Hz) consisting of either 100 pulses (B, n = 6) or 15 pulses (C, n = 6). Each graph shows the time course of the fEPSP slope from 30 min before to 8 h30 after LTP induction. Sample fEPSP traces from individual experiments are shown in the insert for the S1 and the S2 pathways. They were recorded just before the LTP induction (thin traces) and at the end of the experiment (thick traces).

Figure 2. Previous NMDA activation is not required for the increased ease in inducing a long-lasting LTP.

The bars represent the fEPSP slopes obtained 2 h after LTP induction by a single train (100 Hz, 1 s) when dissection and recovery occurred during perfusion of varied ACSF solutions whereas normal ACSF was applied during LTP induction and LTP maintenance. Dissection and recovery of slices were carried out in ACSF either of normal composition (n = 3), without Ca⁺⁺ (n = 2), with 2.6 mM Mg⁺⁺ (n = 2) or added with 50 µM APV, an NMDA receptor inhibitor (n = 3).

Figure 3. Cycloheximide and anisomycin inhibit protein synthesis in hippocampal slices.

(A1–A2) Mice hippocampal slices were incubated in an interface chamber and CHX (300 µM), anisomycin (25 µM) or emetine (20 µM) was added for 85 min. [³H]-leucine (1 µCi/ml) was introduced for 1 hour during (A1) or 4 h after (A2) protein synthesis inhibitor application. We calculated the percentage of inhibition at each time point by determining the ratio between the TCA-precipitable radioactivity of each treated sample (set of 7 individual slices) and that of the corresponding control sample. (A3) Percentage of inhibition of protein synthesis during anisomycin (ANISO), cycloheximide (CHX) and emetine application and 4 h after ANISO or CHX application. (B1–2) Percentage of [³H]-leucine incorporation in slices submitted to forskolin and IBMX during 15 minutes. This percentage was calculated by determining the ratio between TCA-precipitable radioactivity of treated sample and that of the corresponding control sample.

Figure 4. Lack of effect of anisomycin and cycloheximide, two translational inhibitors of protein synthesis, on L-LTP.

In the two parts of the figure (A, B), either drug was applied starting 20 min before the induction of LTP on the S1 pathway till the end of the experiment. LTP was induced using a single high-frequency stimulation train (100 Hz, 1 s). In each part of the figure, the time courses of the fEPSP slope in the S1 pathway and S2 pathway (control), both in the presence and in the absence of the protein-synthesis inhibitor used, are displayed. Sample fEPSP traces from individual experiments are shown in the insert; they were recorded in the presence of the drug, before (thin traces) and 8 h30 after (thick traces) LTP induction. (A) Lack of effect of anisomycin at a dosage of 40 µM. (B) Lack of effect of cycloheximide at a dosage of 300 µM.

Figure 5. Plasticity-related proteins are synthesized neither during LTP induction nor during the recovery period.

For these experiments, the recovery period was extended to 4 h. LTP was triggered using a single high-frequency stimulation train (100 Hz, 1 s). Open circles correspond to the fEPSP slopes measured on the potentiated pathway, whereas filled circles represent the fEPSP slopes measured on the control pathway. (A) Anisomycin (40 µM) was applied starting 20 min before till 1 h after LTP induction (n = 4). (B) Slices were incubated in anisomycin (40 µM) during their recovery period till 1 h after LTP induction (n = 4).

Figure 6. NMDA receptors, α-CaMKII autophosphorylation and PI3K are required to induce a long-lasting LTP by a single train.

Each graph shows the time course of the fEPSP slope after LTP induction by a single train of high-frequency stimulation (100 Hz, 1 s). (A) APV (50 µM) was applied starting 20 min before till 20 min after LTP induction. Following washout of the drug, LTP was successfully triggered by a second inducing stimulus, 3 h after the first one (n = 3). (B) Dissection, recovery and recordings were carried out in ACSF containing either 1.3 mM Mg⁺⁺ and 2.5 mM Ca⁺⁺ (open circles, n = 16) or 2 mM Mg⁺⁺ and 2.2 mM Ca⁺⁺ (filled circles, n = 4). Other ACSF components were unchanged. (C) LTP was induced in α-CaMKII T286A mutants (filled circles, n = 3) or wild-type mice (open circles, n = 6). (D) Slices were treated with wortmannin (5 µM), a PI3K inhibitor, starting 1 h before LTP induction till the end of the experiment (filled circles, n = 4) or incubated in normal ACSF (open circles, n = 6).

Figure 7. LTP induced by a single train decreases when test stimulation frequency is increased.

Comparison of the time courses of the fEPSP slope after LTP induction by a single train of high-frequency stimulation (100 Hz, 0.25 s) when the test stimulation was delivered every 60 s (open circles, n = 5) or every 10 s (filled circles, n = 6). For clarity, only the data related to the tetanized pathways are shown.

Figure 8. The proportion of homomeric GluR1 AMPA receptors in the post-synaptic sites is not altered following LTP induction.

In these experiments, slices were treated with a 1-hour perfusion of NASPM (50 µM), an inhibitor of GluR2-lacking AMPA receptors. (A) NASPM was applied on a control pathway (n = 7). Baseline was calculated on the 20 minutes prior to the drug perfusion. (B) NASPM was added either 45 minutes (filled circles, n = 3) or 1 h45 (open circles, n = 4) after LTP induction by a single train of high-frequency stimulation (100 Hz, 1 s). Only data related to the tetanized pathways are shown. (C) Comparison of the decrease in the fEPSP slope induced by a 1-hour application of NASPM when the drug is applied on control synapses (open circles, n = 7) or on potentiated synapses (filled circles, n = 7). To facilitate the comparison, the fEPSP slope is expressed in percentage of the mean fEPSP amplitude measured (over a period of 20 min) before the application of the drug. In view of their similarity, the data related to the two time windows of application of NASPM during LTP maintenance (see B) were pooled.

Figure 9. Lack of effect of anisomycin on L-LTP induced by four trains of high-frequency stimulation.

Anisomycin (40 µM) was applied starting 20 min before the induction of LTP on the S1 pathway till the end of the experiment. LTP was induced using four high-frequency stimulation trains (100 Hz, 1 s, 5 min apart). The time courses of the fEPSP slope in the S1 pathway and S2 pathway (control), both in the presence and in the absence of the drug, are displayed. Sample fEPSP traces from an individual experiment are shown in the insert; they were recorded in the presence of the drug, before (thin traces) and 8 h after (thick traces) LTP induction.