Expression, modulation and plasticity of AMPA receptor composition in retinal ganglion cells

Abstract

Retinal ganglion cells (RGCs) are the sole output cells of the retina. They integrate light signals from photoreceptors through synaptic inputs from On and Off bipolar cells. In the On pathway, light-evoked RGC responses are mediated by NMDA and AMPA-type glutamate receptors. Both the GluA2-containing, Ca2+-impermeable AMPARs (CI-AMPARs) and the GluA2-lacking, Ca2+-permeable AMPARs (CP-AMPARs) are expressed in RGCs, however specific functional roles for these receptor subtypes is incompletely understood. In this study I used an electrophysiological approach to understand how specific AMPAR subtypes contribute to visual processing at the bipolar-RGC synapse. First, I explored the regulation of the AMPAR subtypes by NMDARs at this synapse. I found that a brief light stimulus induced a rapid decrease in synaptic CI-AMPARs and a replacement with CP-AMPARs. This plasticity occurs though an increase in presynaptic activity, which activates perisynaptic NMDA receptors, leading to an increase in Ca2+ and a dynamin-dependent removal of CI-AMPARs. It is exclusive to the On pathway and occurs specifically within the intermediate range of rod pathway light intensities. These findings suggest that activation of perisynaptic NMDARs is an important mechanism to detect increases in synaptic activity and subsequently modulate AMPAR subtypes at the synapse. Next, I investigated the conditions for activation of CI- and CP-AMPARs in RGCs. I found that CI- and CP-AMPARs are differentially activated by high and low light conditions, such that for dim light intensities CP-AMPARs are activated exclusively and as the light intensity is increased CI-AMPAR responses are recruited. Together with the finding that enhanced spillover activates CI-AMPARs and that spontaneous EPSCs are exclusively mediated by CP-AMPARs, I suggest that CP-AMPARs are closest to presynaptic release sites and that CI-AMPARs are located more distally in the active zone. In addition, I find that the primary rod pathway is mediated exclusively by CP-AMPARs in RGCs and cone pathway responses are almost exclusively driven by CI-AMPARs. Next, I examine how NMDAR-mediated plasticity is expressed in the rod and cone pathways using mouse lines in which rod or cone photoreceptor input is absent. I find that plasticity is observed in rod driven RGCs and absent in cone driven RGCs. Finally, I establish that chronic changes in activity that occur on a 12-hour day/night cycle change the surface expression of AMPARs and that this has functional consequences for the expression of AMPAR plasticity. Together, this work establishes a functional role for AMPAR subtypes at the bipolar cell-RGC synapse and extends our knowledge of AMPAR regulation in the retina.