Postinhibitory rebound spikes are modulated by the history of membrane hyperpolarization in the SCN

Abstract

The suprachiasmatic nucleus (SCN) of the hypothalamus regulates biological circadian time thereby directly impacting numerous physiological processes. The SCN is composed almost exclusively of gamma-aminobutyric acid (GABA)ergic neurons, many of which synapse with other GABAergic cells in the SCN to exert an inhibitory influence on their postsynaptic targets for most, if not all, phases of the circadian cycle. The overwhelmingly GABAergic nature of the SCN, along with its internal connectivity properties, provide a strong model to examine how inhibitory neurotransmission generates output signals. In the present work we show that hyperpolarizations that range from 5 to 1000 ms elicit rebound spikes in 63% of all SCN neurons tested in voltage-clamp in the SCN of adult rats and hamsters. In current-clamp recordings, hyperpolarizations led to rebound spike formation in all cells; however, low-amplitude or short-duration current injections failed to consistently activate rebound spikes. Increasing the duration of hyperpolarization from 5 to 1000 ms is strongly and positively correlated with enhanced spike probability. Additionally, the magnitude of hyperpolarization exerts a strong influence on both the amplitude of the spike, as revealed by voltage-clamp recordings, and the latency to peak current obtained in either voltage- or current-clamp mode. Our results suggest that SCN neurons may use rebound spikes as one means of producing output signals from a largely interconnected network of GABAergic neurons.

Figure 1

Hyperpolarization-induced rebound spikes in SCN neurons. A. Representative neuron exhibiting a rebound spike following application of a 1000 ms hyperpolarization from the -55 mV holding potential to -90 mV and return to −55 mV. B-D. Rebound currents recorded from a single representative neuron demonstrating that the kinetics of rebound currents are impacted by multiple parameters of the hyperpolarization step. At early durations (5 ms) rebound spikes only occur reliably at strong hyperpolarization strengths (A). Hyperpolarizations of increased duration (400 ms; B, and 1000 ms; C. trigger rebound spikes more reliably. In addition, for these longer durations, less intense hyperpolarization strengths are required to generate rebound spikes. Importantly, however, rebound spikes are more time-locked after longer hyperpolarization periods, as compared to shorter hyperpolarizations (C-D). E-G. Current-clamp recordings obtained from a single SCN neuron subjected to multiple injection regimens aimed at simulating the voltage-clamp conditions above. Rebound spikes can be observed following short (5 ms) current injections of -30 pA of hyperpolarizing current (E). Current injections of longer durations significantly increased the reliability of rebound spike appearance. In addition, longer current injections (400 vs. 1000 ms) typically triggered rebound spikes with faster kinetics (F and G, respectively). Arrows indicate the release from hyperpolarization.

Figure 2

Rebound spikes are influenced by the post-hyperpolarization return potential. Mean amplitude of rebound spikes plotted as a function of post-hyperpolarization return potential for three representative return plateau conditions (-55, -30 and -25 mV). Short hyperpolarizations (5 ms) are indicated in light grey with hyperpolarizations of 400 and 1000 ms indicated in black and dark grey, respectively. Note that the amplitude of rebound spikes was significantly larger for short duration hyperpolarizations, irrespective of the return holding voltage. This figure also clearly illustrates that, for 400 ms duration hyperpolarizations, the amplitude of the rebound spikes decreases in a quasi-linear fashion, while for long durations (∼1000 ms), rebound spike amplitude was not significantly altered irrespective of the return voltage. Vertical lines for each data point represent standard errors and asterisks indicate statistical significant (p < 0.05).

Figure 3

Hyperpolarization strength impacts the amplitude of rebound spikes. Mean amplitude (± SE) of rebound spikes for three conditions of hyperpolarizing steps (-65mV to -80 and -90mV). This graph illustrates that increased strengths of hyperpolarization tend to generate more vigorous rebound spikes.

Figure 4

Hyperpolarization duration positively influences the probability rebound spike occurrence. A. Reliability scores (see methods for details), a measure of the probability of spike occurrence, increases with longer hyperpolarization durations for all the return holding potentials investigated. Illustrated are representative data that were obtained from -60 mV and -90mV holding potentials. Note that spike probability increases linearly as a function of increased hyperpolarization duration (5 ms < 400 ms < 1000 ms). B. Data obtained with current-clamp recordings also demonstrate that longer hyperpolarizations increase the probability of rebound spike occurrence. Note the low probability of observing spikes in connection with hyperpolarizing current injections of 5 ms. This probability significantly increased following longer hyperpolarizations (400 and 1000 ms).

Figure 5

Latency to peak of rebound spike is dependent on the history of membrane hyperpolarization. Illustrated are latencies averaged across all magnitudes of hyperpolarization, for three representative durations (5, 400 and 1000 ms). Note that shorter hyperpolarizations lead to slower latency to rebound spike peak, whereas longer durations lead to significantly shorter latencies. Hyperpolarizations that exceed 400 ms do not significantly differ from those that approach 1000 ms.

Figure 6

Sodium channel antagonist lidocaine completely blocks rebound spike formation. Rebound spikes detected after 5 (top) and 1000 ms (bottom) hyperpolarization, when stepping a representative neuron from -55 to −90 mV, and returning to -55 mV. Rebound spikes are completely in the same neurons following bath application of lidocaine (50 μM). Arrow indicates where the post-hyperpolarization rebound spike should have been generated.