Homosynaptic and heterosynaptic inhibition of synaptic tagging and capture of long-term potentiation by previous synaptic activity

Abstract

Long-term potentiation (LTP) is an enhancement of synaptic strength that may contribute to information storage in the mammalian brain. LTP expression can be regulated by previous synaptic activity, a process known as "metaplasticity." Cell-wide occurrence of metaplasticity may regulate synaptic strength. However, few reports have demonstrated metaplasticity at synapses that are silent during activity at converging synaptic inputs. We describe a novel form of cell-wide metaplasticity in hippocampal area CA1. Low-frequency stimulation (LFS) decreased the stability of long-lasting LTP ["late" LTP (L-LTP)] induced later at the same inputs (homosynaptic inhibition) and at other inputs converging on the same postsynaptic cells (heterosynaptic inhibition). Significantly, heterosynaptic inhibition of L-LTP also occurred across basal and apical dendrites ("heterodendritic" inhibition). Because transient early LTP (E-LTP) was not affected by previous LFS, we examined the effects of LFS on the consolidation of E-LTP to L-LTP. The duration of E-LTP induced at one set of inputs can be extended by capturing L-LTP-associated gene products generated by previous activity at other inputs to the same postsynaptic neurons. LFS applied homosynaptically or heterosynaptically before L-LTP induction did not impair synaptic capture by subsequent E-LTP stimulation, suggesting that LFS does not impair L-LTP-associated transcription. In contrast, LFS applied just before E-LTP (homosynaptically or heterosynaptically) prevented synaptic tagging, and capture of L-LTP expression. Thus, LFS inhibits synaptic tagging to impair expression of subsequent L-LTP. Such anterograde inhibition represents a novel way in which synaptic activity can regulate the expression of future long-lasting synaptic plasticity in a cell-wide manner.

Figure 1.

LFS produces transient homosynaptic and heterosynaptic depression. A, Schematic of area CA1 in a mouse hippocampal slice, showing positions of two stimulating electrodes (S1 and S2) and a single recording electrode placed in the stratum radiatum. B, Sample fEPSP responses and analysis of interpathway PPF from a representative experiment. A lack of PPF demonstrates the independence of S1 and S2 inputs. C, LFS at 5 Hz for 3 min produced a transient depression in both homosynaptic (open squares) and heterosynaptic (open circles) pathways. D, Sample fEPSP responses and comparison of mean fEPSP slopes from homosynaptic (open squares) and heterosynaptic (open circles) pathways during baseline (a), immediately after 5 Hz stimulation (b), and 10 min after onset of 5 Hz stimulation (c). ^*Statistical significance, p < 0.05. Error bars indicate SE.

Figure 2.

Previous LFS impairs subsequent homosynaptically and heterosynaptically induced L-LTP. A, LFS did not affect subsequent E-LTP elicited by a single 100 Hz tetanus in the homosynaptic (open squares) or heterosynaptic (open circles) pathway (control, filled diamonds). B, LFS impaired subsequent L-LTP induced with four 100 Hz trains applied to the homosynaptic (open squares) or heterosynaptic (open circles) set of inputs (control, filled diamonds). C, Summary of LFS effects (homosynaptic, open squares; heterosynaptic, open circles) on one-train E-LTP (control, filled diamonds). Columns represent mean fEPSP slopes during baseline (a), 10 min after onset of LFS (b), and 60 min after LTP induction (c). D, Summary of LFS effects (homosynaptic, open squares; heterosynaptic, open circles) on four-train L-LTP (control, filled diamonds). Mean fEPSP slopes are shown from baseline (a), 10 min after onset of LFS (b), and 120 min after L-LTP induction (c). ^*Statistical significance, p < 0.05. Error bars indicate SE.

Figure 3.

Successful synaptic capture of L-LTP by weak LTP stimuli can be assayed by prolonged potentiation and a newly acquired immunity to depotentiation. A, E-LTP induced by one-train (weak) tetanus (S2) is input specific and decays to baseline within 120 min of induction (c, filled diamonds). E-LTP is sensitive to depotentiation and can be reversed to baseline values by LFS (b, open diamonds). B, When four-train (strong) tetanus is first established at one set of inputs (S1), one-train (weak) tetanus to S2 elicits potentiation that is persistent and stable at 120 min after tetanus (c, filled circles). The potentiation elicited by one-train tetanus (S2) is now resistant to depotentiation; after Dpt LFS, mean fEPSP slopes gradually recovered to potentiated levels (b, open circles). C, Summary of depotentiation data. Mean fEPSP slopes 55 min after depotentiation (open symbols) are compared with nondepotentiated controls (filled symbols) of both tetanus protocol groups (one train alone, diamonds; four train plus one train, circles). Error bars indicate SE. a, Pretetanus baseline.

Figure 4.

Previous LFS does not affect transcription associated with L-LTP expression. Four-train (strong) tetanus was applied to S1 followed by one-train (weak) tetanus to S2. A, Application of Act D (25 μm) during four-train L-LTP induction prevented L-LTP expression in both S1 and S2 pathways (filled triangles). B, Act D applied during one-train tetanus (after four-train tetanus), did not affect L-LTP expression in either pathway (filled circles). C, D, LFS applied before four-train tetanus impaired L-LTP expression in S1 but did not affect L-LTP expression by one-train tetanus in S2. LFS was applied to the homosynaptic (diamonds) and heterosynaptic (squares) pathway, relative to four-train tetanus. One-train tetanus (S2) elicited stable potentiation (filled symbols) that was immune to depotentiation (open symbols). E, Summary of four-train plus one-train LTP data from the S2 pathway of all four treatment groups. Mean fEPSP slopes are taken during baseline (a) and 120 min after one-train tetanus (c). F, Summary of depotentiation data from LFS-treated groups, 60 min after one-train tetanus (b). Mean fEPSP slopes in S2 after depotentiation (open symbols) are compared with non-Dpt controls (filled symbols), within treatment groups (homosynaptic, diamonds; heterosynaptic, squares). ^*Statistical significance, p < 0.05. Error bars indicate SE.

Figure 5.

Previous LFS impairs synaptic capture of L-LTP expression and acquired immunity to depotentiation. A, B, Four-train (strong) tetanus was applied to S1 followed by one-train (weak) tetanus to S2. LFS was given before one-train tetanus in S2. A, Homosynaptic application of LFS (S2, naive pathway) before one-train tetanus impaired LTP expression in S2 (filled squares). LTP in S2 depotentiated to baseline levels (open squares). B, Heterosynaptic application of LFS (S1, potentiated pathway) before one-train tetanus impaired LTP expression in S2 (filled triangles), although mean fEPSP slopes remained elevated above baseline values. LTP in S2 recovered to potentiated levels after depotentiation (open triangles). C, Comparison of mean fEPSP slopes in S2 during baseline (a) and 120 min after one-train tetanus (c): one-train alone (filled diamonds), one train when paired with four trains (filled circles), and four train plus one train with LFS before one train (homosynaptic, filled squares; heterosynaptic, filled triangles). LFS significantly impaired L-LTP expression in S2 when applied homosynaptically (filled squares) or heterosynaptically (filled triangles) before one-train tetanus. Mean fEPSP slopes from LFS-treated groups did not differ significantly from one-train controls. D, Summary of depotentiation data from LFS-treated groups, 60 min after one-train tetanus (b). Mean fEPSP slopes in S2 after depotentiation (open symbols) are compared with non-Dpt controls (filled symbols), within treatment groups (homosynaptic, squares; heterosynaptic, triangles). E, One train of tetanus was applied to S1 followed by four trains of tetani to S2. L-LTP is expressed in both pathways (filled inverted triangles). LFS applied heterosynaptically (S2, naive pathway) relative to one-train tetanus in S1, significantly impaired L-LTP expression in S1 (open pentagons). Mean fEPSP values were at baseline levels 120 min after tetanus (b). ^*Statistical significance, p < 0.05. Error bars indicate SE.

Figure 6.

Immunity of four-train L-LTP to depotentiation is not affected by previous LFS. A, Four-train tetanus elicits L-LTP that is immune to depotentiation (filled diamonds). LFS applied homosynaptically (open squares) or heterosynaptically (open circles) did not affect the immunity of four-train LTP to depotentiation. B, Comparison of mean fEPSP slopes during baseline (a) and 55 min after depotentiating LFS (b). Mean fEPSP slopes in slices that received LFS before four-train tetanus (homosynaptic, open squares; heterosynaptic, open circles) recovered after depotentiation to levels that were comparable with slices that did not receive previous LFS (filled diamonds). Error bars indicate SE.

Figure 7.

LFS impairs subsequently induced L-LTP in a heterodendritic manner. A, Schematic of area CA1 in a mouse hippocampal slice, showing positions of two stimulating electrodes (one in SO and one in SR) and a single recording electrode placed in stratum radiatum. B, LFS in SO impaired subsequent L-LTP in SR (open squares). LFS applied to SO elicited a transient synaptic depression in both pathways. Mean fEPSP slopes recovered to baseline values (a) within 10 min of LFS onset (b). Four tetanus trains applied to SR after recovery elicited a transient potentiation that decayed close to baseline values by 120 min after tetanus (c). Mean fEPSP slopes in SR were significantly lower than those in control slices that did not receive LFS before four-train tetanus (filled diamonds). ^*Statistical significance, p < 0.05. Error bars indicate SE.