GABAergic synapse maturation: evidence of the instructive role of activity‐dependent BDNF release

Abstract

The study of the actions of brain-derived neurotrophic factor (BDNF) on central neurons and their synapses has revealed a constellation of observations ranging from neuronal survival and differentiation during early development to rapid effects on membrane conductance and intracellular Ca2+ levels in mature neurons (Amaral et al. 2007). Unfortunately, most studies in this abundant ‘BDNF literature’ are phenomenological and lack mechanistic insights, and even more importantly, direct evidence that the endogenous – native – protein is able to trigger an effect similar to that produced by application of recombinant BDNF. In addition, a few studies of BDNF actions generated long-standing controversial argumentations. Luckily, other BDNF effects have been consistently replicated, such as its role in the developmental maturation of GABAergic synapses. The paper by Kuczewski et al. (2008) in this issue of The Journal of Physiology presents direct evidence that endogenously released BDNF plays an instructive role in a new form of GABAergic synaptic plasticity in area CA3 of the neonatal hippocampus. The fact that BDNF was necessary for the maturation of GABAergic synapses was known from studies in the cortex of Bdnf overexpressing mice (Huang et al. 1999), but its source and mechanism of activity-dependent release remained unexplored. Now, a series of rigorous experiments provide solid evidence that glutamatergic activity causes Ca2+-dependent release of BDNF from postsynaptic CA3 neuron dendrites, which leads to an enduring potentiation of GABAAR-mediated synaptic currents. Despite the fact that BDNF actions on GABAergic synapses were described early on (Marty et al. 1997), the majority of studies focused on its role in excitatory synaptic transmission and plasticity (Poo, 2001); those studies revealed both pre- and postsynaptic sites of action, as well as indirect effects, generating a highly controversial landscape. Presynaptic mechanisms included modulation of probability of transmitter release via presynaptic Ca2+ levels, and modulation of synaptic vesicle mobilization, docking, depletion and recycling (Tyler et al. 2002), whereas postsynaptic mechanisms included modulation of NMDA receptor subunits or K+ channels and surface delivery of AMPA receptor subunits. On the other hand, the effects of BDNF on inhibitory synaptic transmission seemed more consistent (Tanaka et al. 1997), although the sites of action also spanned the synaptic cleft, e.g. modulation of presynaptic Ca2+ channel clustering (Baldelli et al. 2005) and postsynaptic modulation of GABAA receptors (GABAAR) (Brunig et al. 2001) or Cl− transport (Wardle & Poo, 2003). In the current study presented in this issue, Kuczewski et al. (2008) continue their exploration of the mechanisms by which spontaneous glutamatergic neuronal activity participates in the maturation of the neonatal (P1–P6) rat hippocampal network and uncover an enduring potentiation of GABAAR responses. This long-lasting potentiation of GABAAR-mediated synaptic activity (LLPGABA-A) was induced in CA3 pyramidal neurons by a brief and transient recovery following blockade of glutamatergic synaptic transmission, and was mediated by Ca2+-dependent postsynaptic release of endogenous – native – BDNF. Activity-dependent release of endogenous BDNF is necessary for LLPGABA-A induction (i.e. permissive role), and was demonstrated by its blockade with the extracellular BDNF scavenger TrkB-IgG. The BDNF effect was instructive, demonstrated by the rescue of LLPGABA-A induction by exogenously applied recombinant human met-BDNF under conditions that prevented endogenous BDNF release (e.g. the use of glutamate receptor antagonists or L-type Ca2+ channel block). Despite a well-designed experimental strategy, there are important issues that remain unclear, some from a technical standpoint and others from a conceptual perspective. For example, loading postsynaptic CA3 neurons with the Ca2+ chelator BAPTA prevented the increase in the frequency of spontaneous GABAAR-mediated synaptic currents. This observation seems counterintuitive, since only the one cell loaded with BAPTA was unable to release BDNF, and the frequency of spontaneous synaptic currents is likely to reflect the excitability of the entire neuronal network connected to the cell under recording. More conceptual voids are the mechanism(s) of expression/maintenance of LLPGABA-A (e.g. which steps take place between postsynaptic Ca2+ elevations and potentiation of GABAAR-mediated synaptic currents?) and whether there is a physiological consequence at the CA3 or hippocampal network level. It is immediately obvious that an enduring enhancement of GABAAR-mediated synaptic currents represents an important feature of the maturation of GABAergic inhibitory synapses. Certainly, more work will be necessary to fully comprehend the mechanisms by which endogenous BDNF is able to drive the activity-dependent developmental refinement of inhibitory synaptic networks, especially in a brain region like the hippocampus that – due to its recurrent and reentrant circuits that evolved for learning and memory – is always at the border of epileptiform activity.