Control of GABA Release at Mossy Fiber-CA3 Connections in the Developing Hippocampus

Abstract

In this review some of the recent work carried out in our laboratory concerning the functional role of GABAergic signalling at immature mossy fibres (MF)-CA3 principal cell synapses has been highlighted. While in adulthood MF, the axons of dentate gyrus granule cells release onto CA3 principal cells and interneurons glutamate, early in postnatal life they release GABA, which exerts into targeted cells a depolarizing and excitatory action. We found that GABA(A)-mediated postsynaptic currents (MF-GPSCs) exhibited a very low probability of release, were sensitive to L-AP4, a group III metabotropic glutamate receptor agonist, and revealed short-term frequency-dependent facilitation. Moreover, MF-GPSCs were down regulated by presynaptic GABA(B) and kainate receptors, activated by spillover of GABA from MF terminals and by glutamate present in the extracellular medium, respectively. Activation of these receptors contributed to the low release probability and in some cases to synapses silencing. By pairing calcium transients, associated with network-driven giant depolarizing potentials or GDPs (a hallmark of developmental networks thought to represent a primordial form of synchrony between neurons), generated by the synergistic action of glutamate and GABA with MF activation increased the probability of GABA release and caused the conversion of silent synapses into conductive ones suggesting that GDPs act as coincident detector signals for enhancing synaptic efficacy. Finally, to compare the relative strength of CA3 pyramidal cell output in relation to their MF glutamatergic or GABAergic inputs in adulthood or in postnatal development, respectively, a realistic model was constructed taking into account different biophysical properties of these synapses.

Keywords: GABAA-mediated post synaptic currents; activity-dependent plasticity; developing hippocampus; mossy fibres; presynaptic GABAB receptors; presynaptic GluK1 receptors; quantal analysis; realistic modelling.

Figure 1

AMPA receptors do not contribute to unitary postsynaptic currents evoked in CA3 principal cells by minimal stimulation of granule cells in the dentate gyrus. (A) Unitary synaptic currents evoked in a CA3 pyramidal cell at P4 with different stimulation intensities before (Control) and during application of GYKI 52466 (30 μM). Each trace is the average of 15–20 responses (including failures). (B) Peak amplitudes of synaptic currents represented in (A) are plotted as a function of stimulus intensities. Bars are SEM, dashed lines represent the averaged amplitude of GPSCs. (C) upper trace: MF-GPSCs evoked in a P3 principal cell by minimal stimulation of stratum granulosum in the dentate gyrus exhibit paired-pulse facilitation. Lower traces: MF-GPSCs recorded in the presence of GYKI were enhanced in amplitude by DNQX (50 μM), reduced by L-AP4 (10 μM) and abolished by PTX (100 μM). Each trace is the average of 30 traces. (D) Summary data obtained from 15 neurons. **p < 0.01; ***p < 0.001. Modified from Caiati et al., .

Figure 2

Short-term frequency-dependent facilitation of MF-GPSCs. (A) Upper trace: MF-GPSCs exhibiting strong paired-pulse facilitation in response to two stimuli (50 ms apart) delivered to the granule cells in the dentate gyrus (P3). Lower traces represent averaged responses (including failures) obtained at 0.05 and 0.33 Hz from the same neuron. (B) The mean amplitude of synaptic currents evoked in 6 pyramidal cells at 0.05 and 0.33 Hz (bars) is plotted against time. Note the slow build-up of facilitation of synaptic responses at 0.33 Hz that completely reversed to control values after returning to 0.05-Hz stimulation. Bars represent SEM. (C,D) as in (A,B) but for neurons recorded at P5–P6. (C,D) modified from Safiulina et al., .

Figure 3

Excitatory action of GABA exerted on a P3 CA3 pyramidal cells. (A) Stimulation of MF in stratum lucidum (arrow) evoked in a P3 CA3 pyramidal cell (recorded in cell-attached) an action potential that was abolished by picrotoxin (100 μM). (B) GDPs recorded from a P3 CA3 pyramidal neuron (whole cell under current-clamp conditions) before (control) and after bath application of Isoguvacine (3 μM). Note the increase in GDPs frequency in isoguvacine.

Figure 4

Presynaptic GABA_B and kainate receptors localized on MF terminals down regulate GABA release contributing to synapse silencing. (A) A “presynaptically” silent neuron before (left) and during bath application of CGP55845 (1 μM, right). Each trace is the average of 20 responses. Note the appearance of a synaptic current in the presence of CGP55845 (each trace is the average of 20 individual samples). (B) The peak amplitude of MF-GPSCs shown in A, before and during bath application of CGP55845, L-AP4 and PTX (closed bars) is plotted against time. (C,D) as in (A,B) but before and during bath application of UBP 302 (10 μM). Note that UBP 302 induced the appearance of the first response and enhanced the amplitude of the second one giving rise to an evoked GDP. (E) Each column in the graph represents the summary plot of peak amplitude, successes rate, PPR and inverse squared of CV of MF-GPSCs (expressed as percentage of controls) during CGP55845 (white columns; n = 10) or UBP 302 (black columns; n = 19). While amplitude and successes refer to both silent and non-silent neurons, PPR and CV⁻² only to non-silent cells. *p < 0.05; **p < 0.01; ***p < 0.001. (A,B) modified from Safiulina and Cherubini, .

Figure 5

Pairing switched on apparent “presynaptically” silent neurons. (A) Diagram of the hippocampus showing a CA3 pyramidal cell receiving a MF input. The stimulating electrode (stim) is in granule cell layer. (B) GDPs recorded from a CA3 pyramidal cell in current-clamp mode from the hippocampus at P2. On the right, a single GDP is shown on an expanded time scale (the intracellular solution contained QX 314 to block action potentials). The rising phase of GDPs (between the dotted lines) was used to trigger synaptic stimulation. (C) Peak amplitudes of MF-GPSCs before and after pairing (arrow at time 0) are plotted against time. The traces above the graph represent individual responses evoked in different experimental conditions. This synapse was considered “presynaptically” silent since it did not exhibit any response to the first stimulus over 48 trials (at 0.1 Hz) but occasional (two) responses to the second one. (D) Mean GPSCs amplitude and mean percentage of successes before (Control) and after pairing (n = 5). Modified from Kasyanov et al. (2004).

Figure 6

Amplitude distribution of stimulus evoked MF-GPSCs under conditions of low quantal content (extracellular Ca²⁺/Mg²⁺ ratio 0. 16). (A) Selected events, excluding failures, recorded from a P3 CA3 pyramidal neuron at −60 mV. (B) Histogram of GPSCs peak amplitude (289 events excluding failures) for the cell shown in A was fitted with the sum of four Gaussian functions with means of 21.7, 43.5, 65.2, 86.9 pA, respectively and a fixed SD of 6.5 pA; bin size = 5 pA.

Figure 7

Effects of GABAergic or glutamatergic synapses on a model neuron. Typical traces of somatic membrane potential are plotted during three simulations activating 30 synaptic inputs at an average frequency of 20 Hz (left) or 40 Hz (right), using GABAergic (top) or glutamatergic (bottom) synapses. The same stimulation pattern was used in both cases. The passive and active properties of the neuron were those used in a previous work, in a configuration that reproduced the experimental properties of CA3 cells showing firing adaptation (see Figure 9E in Hemond et al., 2008), adjusted to take into account the lower spike threshold (−45 mV) observed in early development in these neurons (Sivakumaran et al., 2009). The model files are available for download from the ModelDB section of the Senselab database (http://senselab.med.yale.edu).