Astrocytes impose postburst depression of release probability at hippocampal glutamate synapses

Abstract

Many neurons typically fire action potentials in brief, high-frequency bursts with specific consequences for their synaptic output. Here we have examined short-term plasticity engaged during burst activation using electrophysiological recordings in acute rat hippocampal slices. We show that CA3-CA1 glutamate synapses exhibit a prominent depression of presynaptic release probability for approximately 1 s after such a burst. This postburst depression exhibits a distinct cooperativity threshold, is abolished by inhibiting astrocyte metabolism and astrocyte calcium signaling, and is not operational in the developing hippocampus. Our results suggest that astrocytes are actively involved in short-term synaptic depression, shaping synaptic activity during behaviorally relevant neural activity.

Figure 1.

A brief conditioning synaptic burst elicits PBD at hippocampal CA3–CA1 synapses. A, Field EPSP recordings illustrating the PBD elicited by a three-impulse, 50 Hz conditioning burst 500 ms before the test stimulation. Calibration: 20 ms, 100 μV. Above is a schematic illustration of the placement of stimulation (S) and extracellular recording electrode (E) in the CA1 stratum radiatum (S.R.). B, Comparison of PBD recorded using AMPA EPSCs (n = 7) and NMDA EPSCs (n = 7) as well as the decrease of 1/CV² for the AMPA EPSCs (n = 5). C, The PBD is maximal at conditioning (Cond)–test intervals of 0.1–1 s (conditioning burst consisted of 3 impulses at 50 Hz, n = 4–8 per bar). D, Increasing the number (Nr.) of stimuli in the conditioning burst beyond three at 50 Hz does not increase the magnitude of the PBD. The conditioning–test interval was 500 ms (n = 4–7 per bar). E, Whole-cell recordings of AMPA EPSCs reveal a cooperativity threshold for the PBD. The top traces illustrate a conditioning–test sequence using AMPA EPSCs evoked using very low, moderate, and high stimulation intensities. Calibration: 20 ms, 50 pA. The bottom graph summarizes five experiments the same as that illustrated on top (*p < 0.05).

Figure 2.

Developmental profile and astrocyte involvement in the PBD. A, Developmental profile of the PBD. Results from rats of different postnatal days were binned into 11 groups (n = 5–15 per group), and a sigmoid curve was fit to the data points. B, The PBD is blocked by the glial metabolic inhibitor FAc (10 mm FAc) and by BAPTA (50 mm) infused into the gap junction-coupled astrocyte network, but not by the gap junction blocker endothelin-1 (Et-1, 1 μm). The control PBD before FAc was 42 ± 5% (n = 11) and was reduced to −9 ± 6% (**p < 0.005) after FAc. The control PBD before BAPTA (43 ± 3%, n = 5) was recorded from a region in the slice without BAPTA-filled astrocytes and was compared with the PBD recorded from a region in the same slice with BAPTA-filled astrocytes (−4 ± 3%, n = 5, ***p < 0.001), as schematically indicated on top (scale bar indicates 50 μm). S, Stimulating electrode; E1, E2, extracellular recording electrodes.

Figure 3.

Astrocyte calcium signaling is critically involved in the generation of the PBD. The effect on the PBD is evaluated as relative change of control PBD in three different situations: “Network,” “Slice,” and “Uncoupled cell.” A, “Network”: BAPTA (50 mm) and Lucifer yellow were infused for 20 min via the patch pipette into a gap junction-coupled astrocyte exhibiting passive electrophysiological properties (scale bar indicates 20 μm). Control PBD before BAPTA infusion (45 ± 5%, n = 5) was significantly different from the PBD 20 min after BAPTA infusion (**p < 0.005). B, “Slice”: a BAPTA-containing patch pipette was simply placed into the slice for 20 min. Control PBD before placement of BAPTA-containing pipette in the slice (41 ± 7%, n = 5) was not significantly different from the PBD 20 min after (p > 0.05). C, “Uncoupled cell”: BAPTA infused for 20 min via the patch pipette into an uncoupled cell with nonpassive electrophysiological properties. Control PBD before patching (47 ± 4%, n = 3) was not significantly different from the PBD 20 min after patching (p > 0.05). Electrophysiological properties were tested using voltage steps of 10 mV from −100 mV to +50 mV (calibration: 50 ms, 100 pA). Top parts of the panels show schematic drawings depicting placement of extracellular recording electrode (E), intracellular recording electrode (I), and stimulation electrode (S) in the three different situations.