BK channels mediate a presynaptic form of mGluR-LTD in the neonatal hippocampus

Abstract

BK channels can control neuronal function, but their functional relevance in activity-dependent changes of synaptic function remains elusive. Here, we report that repetitive low-frequency stimulation activates BK channels through 12(S)HPETE, an arachidonic acid metabolite, produced downstream of postsynaptic metabotropic glutamate receptors (mGluRs) to trigger long-term depression (LTD) at CA3-CA1 synapses in hippocampal slices from P7-P10 mice. Activation of BK channels is subunit specific, as paxilline but not iberiotoxin blocked mGluR-LTD. Also, 12(S)HPETE does not change the electrophysiological properties of the BK channel when the BKα subunit is expressed alone but increases the channel open probability when the BKα is coexpressed with the β4-subunit. Our findings reveal an interaction between 12(S)HPETE and BK channels to regulate synaptic strength at central synapses and increase our understanding of the mechanisms underlying mGluR-LTD in the neonatal hippocampus that likely contribute to circuit maturation necessary for learning.

Keywords: BK channels; hippocampus; retrograde signaling; synaptic plasticity.

Fig. 1.

Arachidonic acid metabolite 12(S)HPETE is necessary to induce mGluR-LTD in the young hippocampus. (A) Low-frequency stimulation (LFS, 5 Hz, 3 min) induces a LTD of fEPSP (black circle) that was eliminated in the continuous presence of the mGluR antagonist MTEP (10 μM; open circle). Top, Representative traces of two consecutive fEPSP (100-ms interstimulus interval) before (black) and after (grey) LFS induction, and in the continuous presence of MTEP. Bottom, summarized graphs. (B) Representative traces (Top) and time course (Bottom) showing that bath application of the mGluR agonist DHPG (50 μM, 5 min) also induces LTD of fEPSP, an effect that was eliminated in the continuous presence of MTEP. Note that LFS- and DHPG-induced mGluR-LTD were accompanied by changes in PPR and coefficient of variation (1/CV², Right), suggesting a presynaptic mechanism involved. In the PPR and 1/CV² summarized plot, each open circle represents a single slice. (C) Representative traces (Top) and time course (Bottom) showing that blocking 12-lypoxigenase, and the production of 12(S)HPETE with two independent blockers (Baicalein and CDC), eliminates LFS-induced LTD. (D) Representative traces (Top) and time course (Bottom) showing that 12(S)HPETE (100 nM), but not its metabolite 12-HETE (100 nM), reduced fEPSP at CA3-CA1 synapses, an effect that was accompanied by changes in PPR and 1/CV². (E) Representative traces (Top) and time course (Bottom) showing that LFS-induced mGluR-LTD was unaffected by two different TRPV1 antagonists Capsazepine (CPZ, 10 μM) and AMG 9810 (AMG, 10 μM). (F) Representative traces (Top) and time course (Bottom) showing that LFS-induced mGluR-LTD is similar in wild-type (TRPV1^+/+) and TRPV1^−/− deficient mice. In all panels, data are presented as mean ± SEM, and averaged sample traces taken at times indicated by numbers are shown next to each summary plot. *P <0.05, **P <0.01, and the number of slices (s) and animals (a) are indicated in parenthesis.

Fig. 2.

Activation of BK channels containing the BKβ4 subunit is involved in mGluR-LTD. (A) Average traces (Top) and time course (Bottom) showing that LFS- and (B) DHPG-induced mGluR-LTD were unaffected by bath application of iberiotoxin (IBTx, 100 nM) but were eliminated in the continuous presence of Paxilline (Pax, 10 or 1 μM). (C) Average traces (Top) and time course (Bottom) showing that bath application of the 12(S)HPETE (100 nM) depresses fEPSP in the continuous presence of IBTx but had no effect in the presence of Pax. (D) Average traces (Top) and time course (Bottom) showing that bath application of NS11021 (3 μM), a BK channel agonist, depresses fEPSP in the continuous presence of IBTx but had no effect in the presence of Pax. NS1121-mediated depression of fEPSP was accompanied by changes in PPR and 1/CV, suggesting a presynaptic mechanism of action. (E) Average traces (Top) and time course (Bottom) showing that preincubation with NS11021 (3 μM) is able to block LFS- and DHPG-induced mGluR-LTD. (F) Average traces (Left) and time course (Right) showing that 12(S)HPETE-induced suppression of fEPSP occluded the NS11021-mediated depression of fEPSP at CA3-CA1 synapses (Top), and inversely, NS11021-induced depression blocks further effect of 12(S)HPETE (Bottom), suggesting a common mechanism of action. In all panels, data are presented as mean ± SEM, and averaged sample traces taken at times indicated by numbers are shown next to each summary plot, and the numbers of slices (s) and animals (a) are indicated in parentheses. *P < 0.05

Fig. 3.

BK channels containing β4 but not β2 subunit are directly activated by 12(S)HPETE. (A) Normalized I_tail-V relationship (I_tail(V)) for BK $\alpha$ channels before (white diamonds) or after (black diamonds) the addition of 12(S)HPETE (100 nM) to the cytoplasmic side in 0 [Ca²⁺]_i. The I_tail(V) data (mean ± SEM) were fitted using a Boltzmann function (Materials and Methods). Fit parameters were $V_h$ = 180 and 182 mV; $z_{\delta }$ = 1.02 and 0.98 before (n = 5) and after (n = 5) the addition of 12(S)HPETE, respectively. (B) Normalized I_tail(V) (mean ± SEM) for BK $\alpha +{\beta }_4$ channels before (white hexagons) or after the addition of 12(S)HPETE at concentrations of 0.001 nM (black squares); 1 nM (black circles) or 100 nM (black triangles) to the internal side. Fit parameters were $V_h$ = 233 mV and $z_{\delta }$ = 0.69, n = 7 (control); $V_h$ = 217 mV and $z_{\delta }$ = 0.78, n = 4 (0.001 nM); $V_h$ = 196 mV and $z_{\delta }$ = 0.81, n = 4 (1 nM); and $V_h$ = 183 mV and $z_{\delta }$ = 0.76, n = 5 (100 nM). (C) Average $V_h$ values obtained from the I_tail(V) curves plotted against 12(S)HPETE concentration (mean ± SEM). Data were fitted using a Hill equation where $V_h=V_{\mathit{max}}+\frac{\left(V_0-V_{\mathit{max}}\right)\times {\left[HPETE\right]}^n}{K^n+{\left[HPETE\right]}^n}$ where V_max is the half voltage in the absence of 12(S)HPETE, V_o is the half voltage at the indicated 12(S)HPETE concentration, K the dissociation constant, and n the Hill coefficient. 12(S)HPETE binding produces a leftward shift in $V_h$ (${\mathrm{\Delta }V}_h$) (n = 3 to 4). (D) Normalized I_tail-V (mean ± SEM) for BK $\alpha +{\beta }_2\mathrm{I}\mathrm{R}$ channels before (white diamonds) or after (black diamonds) the addition of 100 nM 12(S)HPETE to the cytoplasmic side in 0 [Ca²⁺]_i. Fit parameters were $V_h$ = 178 and 174 mV; $z_{\delta }$ = 0.99 and 0.90 before (n = 4) and after (n = 4) the addition of 12(S)HPETE, respectively. (E) Normalized I_tail(V) (mean ± SEM) for BK α + β4 channels, in the presence of 3 µM of Ca⁺², before (open circles) or after the addition of 12(S)HPETE at different concentrations. Data were fitted using a Boltzmann equation (Materials and Methods). The parameters were before the addition of 12(S)HPETE: V_h = 46.86 ± 1.12 mV, and zδ = 1.00 ± 0.04 e₀ (n = 3); 0.3 nM: V_h = 33.34 ± 1.31 mV, and zδ = 1.1 ± 0.06 e₀ (n = 3); 1 nM: V = −2.33 ± 1.25 mV, and zδ = 1.36 ± 006 (n = 3); 5 nM: V_h = −31.18 ± 1.99 mV, and zδ = 1.34 ± 0.12 (n = 3); 100 nM (n = 4): V = −30.33 ± 0.56 mV, and zδ = 1.77 ± 0.06 (n = 4). (F) Average V_h values obtained from the I_tail(V) curves plotted against 12(S)HPETE concentration (mean ± SEM) in the presence of 3 µM of Ca⁺². The V_half –[12(S)HPETE] data were fitted to a Hill equation as described in Fig. 3C.

Fig. 4.

Activation of BKβ4 is required for mGluR-LTD. (A) Average traces (Top) and time course (Bottom) showing that postsynaptic response to exogenous glutamate puff application (1 mM, 25 ms) was unaffected by bath application of BK channels agonist 12(S)HPETE (100 nM) and NS11021 (3 μM). (B) Average trace (Top) and time course (Bottom) showing that NS11021 induces a depression of isolated NMDAR-mediated currents in a similar way as to AMPAR-mediated currents. (C) Average trace (Left) and summarized graph (Right) showing that 12(S)HPETE (100 nM) decrease presynaptic neurotransmission (EPSC; Top) without affecting postsynaptic excitability (I_fAHP; Bottom) recorded at the same time from neonatal CA1 pyramidal neuron. (D) Individual traces (Top) and summarized graphs (Bottom) showing that 12(S)HPETE (100 nM) increased failure rate and decreased synaptic efficacy but had no effect on synaptic potency of minimal stimulation experiments in the CA1 area of the hippocampus, supporting a presynaptic mechanism of action. (E) Individual traces (Top) and time course (Bottom) showing that bath application of paxilline (10 μM) eliminated 12(S)HPETE-mediated increase in failure rate and decrease in synaptic efficacy, supporting that 12(S)HPETE effect is mediated by presynaptic BKβ4 channels. (F) Representative traces (Top) and time course (Bottom) showing that loading CA1 pyramidal neuron with paxilline (Pax; 10 μM) was not able to block NS11021- and 12(S)HPETE-induced synaptic depression. (G) Representative traces (Top) and time course (Bottom) showing that Pax in bath, but not intracellularly applied, is able to block LFS-induced LTD in the neonatal hippocampus. In all panels, data are presented as mean ± SEM, and averaged sample traces taken at times indicated by numbers are shown next to each summary plot and the numbers of cells (c) and animals (a) are indicated in parentheses. ns, no significant; * P < 0.05, ** P < 0.01.