Synaptic Organization of Perisomatic GABAergic Inputs onto the Principal Cells of the Mouse Basolateral Amygdala

Viktória K. Vereczki, Judit M. Veres, Kinga Müller, Gergö A. Nagy, Bence Rácz, Boglárka Barsy, Norbert Hájos

Spike generation is most effectively controlled by inhibitory inputs that target the perisomatic region of neurons. Despite the critical importance of this functional domain, very little is known about the organization of the GABAergic inputs contacting the perisomatic region of principal cells (PCs) in the basolateral amygdala. Using immunocytochemistry combined with in vitro single-cell labeling we determined the number and sources of GABAergic inputs of PCs at light and electron microscopic levels in mice. We found that the soma and proximal dendrites of PCs were innervated primarily by two neurochemically distinct basket cell types expressing parvalbumin (PVBC) or cholecystokinin and CB1 cannabinoid receptors (CCK/CB1BC). The innervation of the initial segment of PC axons was found to be parceled out by PVBCs and axo-axonic cells (AAC), as the majority of GABAergic inputs onto the region nearest to the soma (between 0 and 10 μm) originated from PVBCs, while the largest portion of the axon initial segment was innervated by AACs. Detailed morphological investigations revealed that the three perisomatic region-targeting interneuron types significantly differed in dendritic and axonal arborization properties. We found that, although individual PVBCs targeted PCs via more terminals than CCK/CB1BCs, similar numbers (15-17) of the two BC types converge onto single PCs, whereas fewer (6-7) AACs innervate the axon initial segment of single PCs. Furthermore, we estimated that a PVBC and a CCK/CB1BC may target 800-900 and 700-800 PCs, respectively, while an AAC can innervate 600-650 PCs. Thus, BCs and AACs innervate ~10 and 20% of PC population, respectively, within their axonal cloud. Our results collectively suggest, that these interneuron types may be differently affiliated within the local amygdalar microcircuits in order to fulfill specific functions in network operation during various brain states.