Homosynaptic and heterosynaptic modes of endocannabinoid action at hippocampal CCK+ basket cell synapses

Abstract

Hippocampal interneuron subtypes exhibit remarkable diversity in their morphological, intrinsic and synaptic properties, converging to yield precision in the dynamics of firing during in vivo oscillatory activity (Somogyi & Klausberger, 2005). The basket cell, a type of interneuron specialized to innervate the perisomatic regions of principal cells, comprises two neurochemically distinct and mutually exclusive cell populations, which are either parvalbumin positive (PV+) or cholecystokinin positive (CCK+). Though morphologically similar in dendritic and axonal arborizations, profound immunocytochemical, physiological and synaptic differences exist between CCK+ and PV+ basket cell populations (Hefft & Jonas, 2005; Glickfeld & Scanziani, 2006). In response to afferent stimulation of the Schaeffer collaterals in vitro, PV+ basket cells are readily entrained to provide fast feedfoward GABAergic transmission to principal cell targets (Glickfeld & Scanziani, 2006). In contrast, afferent excitation of CCK+ basket cells is weaker and more transient, often requiring the integration of independent afferent pathways to engage CCK+ basket cells in feedback inhibition (Glickfeld & Scanziani, 2006). These differences fit with the suggestion that PV+ basket cells are well suited to provide rapid, reliable feedforward inhibition to pyramidal cells for sustaining intrinsic network oscillations (‘rhythm’). In contrast, CCK+ basket cells may act as an interface between the intrinsic hippocampal hardware and emotional centres of the brain; their extrinsic neuromodulation would confer emotional context to the spatial information being processed (‘mood’; Freund, 2003). A striking feature of CCK+ basket cells is their capacity to be modulated by endogenous endocannabinoids, which are generated in postsynaptic pyramidal cells upon depolarization, or through activation of G-protein coupled receptors (GPCRs) such as muscarinic (mAChRs) or metabotropic glutamate receptors (mGluRs). Once produced, endocannabioids are liberated to diffuse retrogradely across the synapse to bind CB1 receptors present on presynaptic CCK+ terminals, resulting in the inhibition of GABA release. In contrast, CB1 receptors are largely absent from PV+ basket cell terminals. This phasic mechanism, termed depolarization-induced suppression of inhibition (DSI), is well understood (Wilson & Nicoll, 2002), but less is known about whether endocannabinoids might also tonically modulate CCK+ interneurons. In the current issue of The Journal of Physiology, Neu et al. (2007) offer unprecedented resolution of how single CA1 CCK+ basket cells might be tonically modulated by endocannabinoids. In a series of technically demanding experiments, the authors performed whole cell recordings between pairs of neurochemically identified, monosynaptically connected CCK+ basket cells and CA1 pyramidal cells, thereby avoiding the confounding influence of PV+ cells and other cell types. The authors definitively showed that basal levels of endocannabinoids present tonically modulate GABAergic transmission at a subset of CCK+ basket cell terminals. Interestingly, the CB1 receptor-dependent tonic modulation was completely blocked by buffering calcium in the postsynaptic cell with 10 mm BAPTA, pointing to the postsynaptic cell as the sole source of endocannabinoid generation (Fig. 1). Apparently, basal levels of endocannabinoids released from nearby pyramidal cells cannot diffuse to the recorded CCK basket cell to tonically block presynaptic CB1 receptors. This homosynaptic mechanism contrasts sharply with the findings of Wilson & Nicoll (2001), who demonstrated through simultaneous pyramidal cell recordings that depolarizing one pyramidal cell caused a heterosynaptic depression of GABA release onto an adjacent non-depolarized pyramidal cell. Under different circumstances, Neu et al. (2007) also find evidence for heterosynaptic depression: with 10 mm BAPTA in the postsynaptic cell, exogenous activation of mAChRs induces presynaptic CB1 receptor activation, suggesting that mAChR activation induces ‘spill over’ of endocannabinoids from neighbouring pyramidal cells (Fig. 1). One alternative interpretation is that mAChR-induced release of endocannabinoids is relatively calcium-independent. Future experiments may distinguish between heterosynaptic and calcium-independent homosynaptic mechanisms by preventing GPCR-dependent endocannabinoid generation postsynaptically. In any case, regardless of whether GPCRs utilize a different route for endocannabinoid synthesis, global GPCR activation would be expected to strongly depolarize many CA1 pyramidal cells, which could then induce heterosynaptic effects via a classic calcium-dependent, DSI-like mechanism. Figure 1 Homosynaptic vs. heterosynaptic modes of endocannabinoid action Homosynaptic and heterosynaptic modes of endocannabinoid action pose intriguing questions regarding the impact of CCK+ basket cell-mediated inhibition in the context of other interneuronal networks, as well as their overall effect on excitability in the hippocampal network. How do tonic, DSI-like, and GPCR-mediated endo-cannabinoid signalling modes alter the balance of GABAergic inhibition? Are each of these modes associated with dynamic shifts in how CCK+ and PV+ basket cells shape the oscillatory activity of pyramidal cells? Finally, are these modes of endo-cannabinoid mobilization operative under in vivo conditions? Although the answers to these questions await further investigation, the findings of Neu et al. provide a useful framework for understanding how different conditions and possibly sources of endo-cannabinoids might be mobilized to alter CCK+ interneuron-mediated inhibition within the hippocampal network.