Homeostatic regulation of h-conductance controls intrinsic excitability and stabilizes the threshold for synaptic modification in CA1 neurons

Abstract

Key points: We determined the contribution of the hyperpolarization-activated cationic (h) current (Ih ) to the homeostatic regulation of CA1 pyramidal cells in vitro using chronic treatments (48 h) that either increase (picrotoxin) or decrease (kynurenate) neuronal activity. The h-conductance was found to be up- or down-regulated following chronic activity enhancement or activity deprivation, respectively. This bidirectional plasticity of Ih was found to subsequently alter both apparent input resistance and intrinsic neuronal excitability. Bidirectional homeostatic plasticity of Ih also determined EPSP waveform and EPSP summation tested at 5-30 Hz. Long-term synaptic modification induced by repetitive stimulation of the Schaffer collaterals was found to be constant across treatments in the presence of Ih but not when Ih was blocked pharmacologically. Thus, bidirectional homeostatic regulation of Ih stabilizes induction of long-term synaptic modification in CA1 pyramidal neurons that depends on EPSP summation.

Abstract: The hyperpolarization-activated cationic (h) current is a voltage-shock absorber, highly expressed in the dendrites of CA1 pyramidal neurons. Up-regulation of Ih has been reported following episodes of intense network activity but the effect of activity deprivation on Ih and the functional consequence of homeostatic regulation of Ih remain unclear. We determined here the contribution of Ih to the homeostatic regulation of CA1 pyramidal cell excitability. Intrinsic neuronal excitability was decreased in neurons treated for 2-3 days with the GABAA channel blocker picrotoxin (PiTx) but increased in neurons treated (2-3 days) with the glutamate receptor antagonist kynurenate (Kyn). Membrane capacitance remained unchanged after treatment but the apparent input resistance was reduced for PiTx-treated neurons and enhanced for Kyn-treated neurons. Maximal Ih conductance was up-regulated after chronic hyperactivity but down-regulated following chronic hypoactivity. Up-regulation of Ih in PiTx-treated cultures was found to accelerate EPSP kinetics and reduce temporal summation of EPSPs whereas opposite effects were observed in Kyn-treated cultures, indicating that homeostatic regulation of Ih may control the induction of synaptic modification depending on EPSP summation. In fact, stimulation of the Schaffer collaterals at 3-10 Hz induced differential levels of plasticity in PiTx-treated and Kyn-treated neurons when Ih was blocked pharmacologically but not in control conditions. These data indicate that homeostatic regulation of Ih normalizes the threshold for long-term synaptic modification that depends on EPSP summation. In conclusion, bidirectional homeostatic regulation of Ih not only controls spiking activity but also stabilizes the threshold for long-term potentiation induced in CA1 pyramidal neurons by repetitive stimulation.

Figure 1. Bidirectional homeostatic changes in excitability after chronic alteration of activity

A, spontaneous activity recorded in a control neuron (Cont, grey), a kynurenate‐treated neuron (Kyn, red) and a picrotoxin‐treated neuron (PiTx, blue). B, quantification of the net area of postsynaptic activity for each condition. C, intrinsic neuronal excitability measured in each condition. Left, individual traces for a depolarizing current pulse of 115 pA. Right, input–output curves for each condition showing the number of action potentials as a function of the injected current. Calibration bars, 200 ms, 20 mV. D, rheobase for each condition. E, comparison of the gain, measured as the slope of the input–output curve, in each condition. F, graph indicating the normalized spike number as a function of spontaneous excitatory activity induced by Kyn or PiTx treatments. The homeostatic regulation of excitability is indicated by both the decreased excitability when activity is increased by PiTx treatment and the increased excitability when activity is reduced by Kyn treatment.

Figure 2. Bidirectional regulation of apparent R_in

A, input resistance in control, Kyn‐ and PiTx‐treated CA1 neurons. Top, individual traces for a hyperpolarization of −125 pA for each condition (Cont, grey; PiTx, blue; Kyn, red). Middle, voltage measured at the steady‐state phase of the hyperpolarization as a function of the current injected for each condition. Bottom, input resistance for each condition, measured by fitting. Membrane potential: −64.8 ± 0.3 mV, n = 51 in control neurons; −64.3 ± 0.3 mV, n = 50 in Kyn‐treated neurons and −64.1 ± 0.4, n = 41 in PiTx‐treated cells (P > 0.1 for all comparisons). B, R _in in the presence of ZD‐7288 (10 μm). Membrane potential: −65.0 ± 0.6 mV, n = 26 in control; −64.5 ± 0.5 mV, n = 24 in Kyn; −63.3 ± 0.5, n = 21 in PiTx (P > 0.05 for all). C, effects of chronic treatments on membrane capacitance. Top, representative traces. Bottom, quantitative data. D, holding current injected to maintain the pyramidal neurons at a resting potential of −65 mV before and after ZD‐7288 application for each condition. E, normalized input resistance as a function of spontaneous excitatory activity. The homeostatic regulation of input resistance is indicated by both the decreased R _in when activity is increased by PiTx treatment and the increased R _in when activity is reduced by Kyn treatment.

Figure 3. Bidirectional regulation of h‐current

A, effects of ZD‐7288 on input resistance in Kyn‐ and PiTx‐treated neurons. Left, individual traces before and after ZD‐7288 application for each condition. The initial input resistance is identical for each condition. Right, percentage of input resistance increase after ZD‐7288 application. Membrane potential before ZD‐7288: −64.7 ± 0.4 mV in control, −64.2 ± 0.4 mV in Kyn‐ and −64.0 ± 0.4 mV in PiTx‐treated neurons (P > 0.1 for all). Membrane potential after ZD‐7288: −65.3 ± 0.6 mV in control, −64.3 ± 0.5 in Kyn‐ and −64.5 ± 0.6 in PiTx‐treated neurons (P > 0.1 for all comparisons). B, voltage‐clamp recording of h‐current in each condition. These traces were obtained by subtracting from control traces recorded after ZD‐7288 application to the control ones. C, activation curve of h‐current. Left, differences in the maximal conductance of I _h are observed between each condition. Right, normalized conductance allowed no changes to be seen in the biophysical properties of activation. D, h‐conductance as a function of spontaneous excitatory activity. Note the increased h‐conductance when activity is increased by PiTx treatment and the decreased h‐conductance when activity is reduced by Kyn treatment.

Figure 4. Role of I_h in the homeostatic regulation of cell excitability

A, number of action potentials fired for each current injection before and after ZD‐7288 application in control neurons. Inset, representative traces. Membrane potential before ZD‐7288: −65.2 ± 0.4 mV in control, −63.9 ± 0. 5 mV in Kyn‐ and −63.3 ± 0.6 mV in PiTx‐treated neurons (P > 0.01 for all comparisons). Membrane potential after ZD‐7288: −65.5 ± 0.7 mV in control, −64.7 ± 0.6 mV in Kyn‐ and −63.6 ± 0.5 mV in PiTx‐treated neurons (P > 0.05 for all comparisons). B, as in A with Kyn‐treated neurons. C, as in A with PiTx‐treated neurons. D, ZD‐7288 application reduced the divergence of the input–output curves (top, before ZD‐7288; bottom, after ZD‐7288).

Figure 5. EPSP waveform and EPSP summation is altered after homeostatic regulation of I_h

A, changes in field EPSP waveform after each treatment. Top, representative fEPSP evoked by stimulation of the Schaffer collaterals. Bottom, quantification of fEPSP half‐width for each treatment. B, simulated EPSPs generated by the injection of a simulated synaptic current in CA1 pyramidal neurons recorded in whole‐cell configuration. C, EPSP summation of simulated trains of EPSPs at 20 Hz for each condition. Top, representative traces. Bottom, quantification of the summation by normalizing the fifth EPSP to the first. D, differential EPSP summation before and after ZD‐7288 application for each condition. Representative traces for EPSP trains at 20 Hz before and after ZD‐7288 application. Note the maximal effect for PiTx‐treated neurons and the minimal effect for Kyn‐treated neurons.

Figure 6. Homeostatic regulation of I_h stabilizes induction of long‐term synaptic modification

A, induction of synaptic modification by 10 Hz stimulation. Left: time course of fEPSP slope before and after 10 Hz stimulation of synaptic afferents for each condition. Note the similar level of plasticity in each case. Right: time course of fEPSP slope before and after a second 10 Hz stimulation in the presence of ZD‐7288 application (1 μm). Note the selective induction of LTP in PiTx‐treated neurons. Top: representative examples of fEPSP slope changes for each condition before (light trace) and after (dark trace) 10 Hz stimulation (scale bars 0.1 mV, 0.5 ms). B, time course of fEPSP slope before and after 10 Hz stimulation of synaptic afferents for each condition in the presence of ZD‐7288 application (1 μm) in the same slices. Note the selective induction of LTP in PiTx‐treated cells. Top: representative examples of fEPSP slope changes for each condition before (light trace) and after (dark trace) 10 Hz stimulation (scale bars 0.1 mV, 0.5 ms). C, quantification of synaptic modification. Left, bar graph of fEPSP slope modifications after 10 Hz stimulation in the presence of ZD‐7288. Right, quantification of fEPSP slope modifications after 10 Hz stimulation. D, BCM curves in control and in the presence of ZD‐7288 for control, PiTx‐ and Kyn‐treated neurons. E, LTP induced by 10 Hz stimulation in whole‐cell configuration with or without functional h‐channels. Note the similar level of LTP for Kyn‐treated neurons but the differential effect for PiTx‐treated neurons. F, comparison of the ratio of fEPSP slope to presynaptic volley for each condition. Left, representative fEPSP for each condition. Note the difference in the volley amplitude compared to the fEPSP size. Right, plot of the fEPSP slope normalized to the presynaptic volley amplitude for each condition.

Figure 7. Interplay between regulation of I_h and synaptic plasticity: a hypothesis

A, the homeostatic regulation of EPSP summation compensates for the different potential for LTP expression. Kyn‐treated dendritic spines (top left) express a large number of AMPA receptors but very few h‐channels. Their potential for LTP is small and the EPSP summation in the 3–10 Hz range is high. PiTx‐treated dendritic spines (top right) express very few AMPA receptors but a large number of h‐channels. Their potential for LTP is therefore high but EPSP summation at 10 Hz is low. B, BCM curves with homeostatic regulation of I _h. Red curve, hypothetical curve obtained in Kyn‐treated neurons. Blue curve, hypothetical curve obtained in PiTx‐treated neurons. No difference is observed in each condition. C, BCM curves without homeostatic regulation. The curve obtained in PiTx‐treated neurons is shifted to the left and LTP induction is facilitated.