Stereotyped initiation of retinal waves by bipolar cells via presynaptic NMDA autoreceptors

Rongwei Zhang:1 Xiaoquan Li:1 Koichi Kawakami:2 Jiulin Du:1
1:Institute of neuroscience 2:Division of Molecular and Developmental Biology, National Institute of Genetics 

Prior to the maturation of sensory pathways, spontaneous neuronal activities play critical roles in the formation of neural circuits and brain development. In immature retina, spontaneous patterned neural activities exhibit wave-like propagation among neighbouring retinal ganglion cells (RGCs), called retinal waves. After nearly three decades' intensive investigation, it has been known that retinal waves can propagate via the optic nerve to the central visual system, and guide the activity-dependent refinement of visual topographic maps. However, the origin and underlying mechanism of retinal waves, especially glutamatergic ones, are still mysterious to us. To investigate the origin of retinal waves, we first performed in vivo time-lapse two-photon calcium imaging in zebrafish larvae to monitor neuronal activities among the populations of bipolar cells (BCs), the primary excitatory presynaptic source of RGCs, and discovered for the first time that BCs themselves display glutamatergic retinal waves. To further explore the involved mechanism, we combined in vivo whole-cell patch-clamp recording, optogenetics, glutamate uncaging, and glutamate imaging, and demonstrated that BC waves are initiated at the BC ATs through a NMDA autoreceptor-dependent mechanism. Furthermore, the researchers found that BC waves can propagate to downstream RGCs and optic tectum (OT), the central recipient of RGC axons, suggesting that glutamatergic waves previously observed in RGCs and OTs are actually a mirror event of BC waves. In contrast to previous in vitro studies that showed a random pattern of retinal wave initiation, our in vivo work unexpectedly reveals that the temporal retina has the lowest threshold for wave initiation, though each region of the retina is capable of generating waves. This property results in the preferential initiation of retinal waves at the temporal retina, which may be the reason for more synaptic connections in this region and higher visual acuity for frontal environment. In conclusion, we discover that the ATs of BCs initiate glutamatergic retinal waves through a NMDA autoreceptor-dependent mechanism. Together with our previous findings (J Physiol 2010, 588:2557-2569; Neuron 2012, 75:479-489), it indicates that retinal waves can instruct synaptic reconstruction during development, leading to refinement of retinal circuits.