Horizontal slices of mouse retina expose horizontal cells and their properties

In most neurons, the post-synaptic response is constructed from a sum of individual quantal currents that each arise from release of a presynaptic vesicle. Characterizing the properties of quantal release events is essential for understanding how information about changing light levels transduced in photoreceptor outer segments is subsequently encoded for transmission across the first synapse in the retina. To study responses of second-order retinal neurons, investigators have often used thin vertical slices cut perpendicular to the retinal surface to expose cell bodies of bipolar and horizontal cells (HCs). However, the large dendritic arbors of HCs are typically truncated in these preparations and HC somata can be difficult to visualize in vertical slices. There have therefore been few published recordings from mouse HCs. To visualize HC somata and keep their processes intact, Feigenspan and Babai (2015) developed a preparation in which they slice the retina horizontally along a plane parallel to the retinal surface. To do so, they embed small pieces of mouse retina in a block of low melting point agarose and then slice the retina into thin sections with a vibratome. Cutting through the inner nuclear layer exposes HC somata that can then be targeted by whole cell recording. Fig. 1 shows an example of such a horizontal slice created from a mouse retina containing fluorescently-labeled HCs.

Figure 1.

Fluorescent image of a horizontal retinal slice prepared from a CB60Tg(GAD1-EGFP)G42Zjh/J transgenic mouse showing somata and extensively branched axon terminals of HCs. (Image courtesy of A. Feigenspan and N. Babai.)

HCs from horizontal retinal slices exhibited numerous spontaneously occurring miniature excitatory post-synaptic currents (spEPSCs) arising from the release of glutamate-filled vesicles by presynaptic photoreceptors. The frequency of spEPSCs was diminished by flashes of bright light. Light levels used in this study bleached most of the rod photopigment and so most of the light-sensitive spEPSCs were due to release from cones. The decline in release caused by light-evoked hyperpolarization of cones indicates that many “spontaneous” events were in fact evoked by Ca²⁺ entering through L-type Ca²⁺ channels that are open when cones are depolarized in darkness.

Similar to non-mammalian HCs (Cadetti et al., 2005), Feigenspan and Babai found that quantal amplitudes ranged widely. Careful analysis of high quality recordings suggested that unitary quantal amplitudes of spEPSCs were quite small, averaging only ~3 pA, but they also observed substantially larger events up to 65 pA. Kinetics of large and small events were similar to one another and there were multiple peaks in the amplitude histograms suggesting that large events may arise from coordinated multivesicular release by cones, akin to ribbon synapses of rod bipolar and hair cells (Singer et al., 2004; Glowatzki and Fuchs, 2002). Events were generally larger in darkness suggesting that evoked release triggered by Ca²⁺ channel openings in darkness may increase the likelihood of multiquantal release. The small amplitude of unitary events may reduce quantal noise during sustained release in darkness whereas large multiquantal events may help discriminate changes in evoked release from spontaneous release. Feigenspan and Babai noted that a few spEPSCs persisted even when release from cones was strongly inhibited by bright light. Does this spontaneous release derive equally from rods and cones? Is it Ca²⁺-dependent or Ca²⁺-independent? And what is the function of spontaneous release at tonically active photoreceptor synapses?

The authors showed that mouse HCs share many properties with other species including similar ion channels, intensity-response functions, light response kinetics, and substantial contributions from AMPA receptors. However, AMPA receptor inhibition did not completely block light responses and a selective kainate (KA) receptor agonist potentiated HC light responses suggesting a role for KA receptors in mouse HCs. There is molecular and immunohistochemical evidence for KA receptors in HCs of various species, but the only previous physiological evidence for KA receptors in HCs comes from human retina (Yang, 2004; Shen et al., 2004). KA receptors contribute little to light responses of lower vertebrates but many OFF bipolar cells also signal through KA receptors in mammalian retinas (Puthussery et al., 2014; Borghuis et al., 2014; Lindstrom et al., 2014) suggesting that perhaps KA receptors have a greater role in mammalian than non-mammalian retinas.

While the present study focused on release from cones, dark-adapted retinal slices could also be used to study rod-mediated responses where, compared to cones, intracellular Ca²⁺ stores play a greater role (Babai et al., 2010) and a larger portion of post-synaptic responses involve non-quantal actions of glutamate (Cadetti et al, 2008). Th use of the horizontal slice preparation can also be combined with genetic tools in mice and opens up numerous possibilities for future studies on the roles of HCs in visual processing by the outer retina.