Spontaneous voltage oscillations in striatal projection neurons in a rat corticostriatal slice

Abstract

In a rat corticostriatal slice, brief, suprathreshold, repetitive cortical stimulation evoked long-lasting plateau potentials in neostriatal neurons. Plateau potentials were often followed by spontaneous voltage transitions between two preferred membrane potentials. While the induction of plateau potentials was disrupted by non-NMDA and NMDA glutamate receptor antagonists, the maintenance of spontaneous voltage transitions was only blocked by NMDA receptor and L-type Ca2+ channel antagonists. The frequency and duration of depolarized events, resembling up-states described in vivo, were increased by NMDA and L-type Ca2+ channel agonists as well as by GABAA receptor and K+ channel antagonists. NMDA created a region of negative slope conductance and a positive slope crossing indicative of membrane bistability in the current-voltage relationship. NMDA-induced bistability was partially blocked by L-type Ca2+ channel antagonists. Although evoked by synaptic stimulation, plateau potentials and voltage oscillations could not be evoked by somatic current injection--suggesting a dendritic origin. These data show that NMDA and L-type Ca2+ conductances of spiny neurons are capable of rendering them bistable. This may help to support prolonged depolarizations and voltage oscillations under certain conditions.

Figure 1. Cortical stimulation evoked long-lasting plateau potentials with repetitive firing in medium spiny neurons

A, scheme of a brain slice showing the connection stimulated. The stimulating electrode was placed in the cortical grey or white matter (pyramidal cell scheme), and the recording electrode on the dorsal neostriatum (neostriatal cell scheme). B, orthodromic response of a spiny neostriatal neuron to single subthreshold and suprathreshold stimulus in the cortex. Notice plateau potential with repetitive discharge after suprathreshold stimulus. C, a three times threshold cortical stimulus now evokes a plateau potential followed by several membrane potential oscillations, some of which exhibit the firing of action potentials. This behaviour is more commonly elicited with a stimulus train (see Fig. 4B). The time scale is different in B and C. D, a stimulus train can also evoke an arrhythmic sequence of voltage oscillations in pyramidal cells (sensorimotor cortex, layer 5). E, firing of a neostriatal neuron after a intracellular current injection at the soma (current stimulus not shown). F, firing of a neostriatal neuron during an spontaneous depolarization or ‘up-state’. Action potentials are clipped due to digitization procedures in E and F.

Figure 4. NMDA facilitates plateau potentials and membrane potential oscillations

A, a pair of synaptic responses was evoked with cortex stimulation. Only the second EPSP reached threshold (control). Addition of NMDA (25 µm) increased the duration of the second response and produced a plateau potential. Thus, NMDA mimicked an increase in stimulus strength (+NMDA). B, a train of orthodromic subthreshold stimuli were given at the cortex (arrow signals upward deflections of stimulus artifacts). The response was subthreshold to produce any plateau or oscillatory response (control). Superfusion with NMDA (25 µm) without a cortical stimulus did not produce oscillatory responses in most cases (+NMDA). However, the combined administration of a cortical stimulus and NMDA evoked membrane potential oscillations after a variable latency (two continuous traces at bottom). C, a 2 s ramp was given at the soma and the whole current response plotted against this voltage command. Control response shows firing of unclamped action currents but no crossings of the zero current axis. Addition of NMDA (20 µm) facilitated a negative slope conductance region, and a double crossing of the zero current axis, indicative of bistability (+NMDA). D, reduction of the electrotonic length with Ba²⁺, shifted the first crossing of the zero-current axis and the negative slope conductance region to the left in the voltage axis. This suggests that the origin of bistability is located on the dendrites. The graph also shows the inward rectifier block by Ba²⁺.

Figure 2. Evoked oscillation of the membrane potential after repetitive cortical stimulation

A, a pair of cortical stimuli (arrow) produces a plateau potential similar to that depicted in Fig. 1 (cf. time scale). B, repetition of the stimulus at 0.1 Hz produced a late subthreshold depolarization after the plateau potential produced by the third stimulus. C, continuous repetition of the same stimulus induced a suprathreshold late depolarization after the plateau potential evoked by the sixth stimulus. D, a continuous oscillation of the membrane potential followed this later plateau (see change in time scale, action potentials are clipped or incomplete due to digitization) without any further cortical stimulus. Several cells exhibited a tendency to slowly depolarize but continued to exhibit voltage oscillations for several minutes without overt cortical stimulation. E, an all point histogram of the membrane potential is shown for this cell. Notice Gaussian distributions around two different membrane potential values. F, a scatter plot for two-state membrane potential is shown for a sample of cells. This experiment was performed in the absence of NMDA. However, in other cells, NMDA greatly facilitated the appearance of membrane potential oscillations after the initial, evoked, plateau potential.

Figure 3. Plateau potentials are easily evoked after paired pulse stimulation

A, from top to bottom, the time interval between a pair of cortical stimuli was reduced without changing the stimulus strength. Notice a prolonged depolarization capable of sustaining repetitive firing after the briefer interval. B, the responses to a pair of subthreshold stimuli are shown. They were recorded at two different holding potentials. Paired-pulse facilitation is shown at −60 mV, and depression at −70 mV (top). However, if stimulus strength was increased, a prolonged depolarization could be elicited at −70 mV (bottom). This was never obtained by depolarizing the soma with current injections. C, a cell was hyperpolarized to −92 mV (top) while having voltage oscillations, as in Fig. 2D. Oscillations were reduced in both duration and amplitude to finally stop after this manoeuvre. An all point histogram (bottom) revealed bimodality of the membrane potential at −70 mV, but only a single peak at −92 mV.

Figure 5. AP5 blocks membrane potential oscillations

A, representative sweeps of spontaneous (in the absence of cortical stimulus or NMDA) membrane potential oscillations recorded in a neostriatal neuron (control). Action potentials are clipped or incomplete due to digitizing procedures. B, addition of AP5 (50 µm) in the bath saline blocked the membrane potential oscillations (+AP5). C, membrane potential oscillations could be restored after cortical stimulation and addition of NMDA (10 µm) during AP5 wash-out (AP5 wash-out + NMDA). D, AP5 could block voltage oscillations induced by cortical stimulation and NMDA in a sample of cells as this one (+AP5 traces at bottom).

Figure 7. L-type Ca²⁺ channels participate in the generation of synaptically evoked plateau potentials and membrane potential oscillations

A, a pair of suprathreshold cortical stimuli evoked a long-lasting plateau potential in a spiny cell (control). Notice sustained firing. Addition of the L-type Ca²⁺ channel antagonist, nitrendipine (5 µm), blocked most of the plateau potential (+nitrendipine). The subsequent addition of the NMDA receptor antagonist, AP5 (50 µm), blocked the remaining slow depolarization disclosing the underlying EPSPs (+AP5). B, membrane potential oscillations were evoked with cortical stimulation and NMDA (10 µm)(control). Addition of the L-type Ca²⁺ channel antagonist, nicardipine (5 µm), blocked the oscillations and made cortical stimulation ineffective for evoking them. C, superimposed current–voltage (I–V) plots from a medium spiny cell (control). Addition of NMDA (20 µm) induces a negative slope conductance region and bistability. A subsequent addition of nitrendipine (5 µm) to the bath saline abolished the negative slope conductance region induced by NMDA.

Figure 6. CNQX does not block membrane potential oscillations

A, membrane potential oscillations were induced by repetitive cortical stimulation plus NMDA (10 µm). B, once induced, membrane potential oscillations did not change in frequency or duration after the addition of CNQX (10 µm). C, box plots illustrate distributions of up-state duration before and during CNQX in a sample of neurons (n = 7).

Figure 8. L-type Ca²⁺ channel agonist, BayK 8644, increased the duration of membrane potential oscillations

A, membrane potential oscillations were evoked by cortical stimulation plus NMDA (20 µm; control). Addition of BayK 8644 (5 µm) increased the duration of up-state events (+BayK 8644 and B). C, the amplitude of these events was also increased after BayK 8644 (C).

Figure 9. GABA_A receptor antagonists enhance up-state duration

A, membrane potential oscillations were induced by repetitive cortical stimulation plus NMDA (10 µm). B, the duration of up-states was enhanced by the addition of GABA_A receptor antagonists to the bath saline (+picrotoxin, 5 µm). C, box plots illustrate distributions of up-state duration before and during the GABA_A receptor antagonists picrotoxin and bicuculline (5 µm) in a sample of neurons (n = 6).