Electrical response of retinal ganglion cells in an N-methyl-N-nitrosourea-induced retinal degeneration porcine model

Abstract

Retinal prosthesis is regarded as the treatment for vision restoration in the blind with retinal degeneration (RD) due to the loss of photoreceptors. A strategy for retinal prosthesis is to electrically activate surviving neurons. The retina's response to electrical stimulation in a larger RD model has not been studied yet. Therefore, in this study, we investigated electrically evoked retinal responses in a previously validated N-methyl-N-nitrosourea (MNU)-induced porcine RD model. Electrically evoked responses were evaluated based on the number of retinal ganglion cell (RGC) spikes via multichannel recordings. Stimulation pulses were applied to degenerative and wild-type retinas with pulse modulation. Compared to wild-type retinas, degenerative retinas showed higher threshold values of pulse amplitude and pulse duration. The rate of increase in the number of RGC spikes relative to stimulus intensity was lower in degenerative retinas than in normal retinas. In severely degenerated retinas, few RGCs showed electrically evoked spikes. Our results suggest that the degenerative porcine retina requires a higher charge than the normal porcine retina. In the early stage of RD, it is easier to induce RGC spikes through electrical stimulation using retinal prosthesis; however, when the degeneration is severe, there may be difficulty recovering patient vision.

Figure 1

Experimental setup and protocol for in-vitro multi-electrode array (MEA) recording. (A) The retinal patch was isolated from the visual streak area. The MEA covers 2 mm × 2 mm of the local area in the entire retina (white box in the visual streak area). (B) The isolated retinal patch was mounted on the MEA with the RGC layer facing the electrode array. (C) The experimental protocol.

Figure 2

Differences in the number of RGCs that respond to light or electrical stimulation based on presence or absence of MNU treatment. (A) Average population of light stimulus (LS)-responsive RGCs in the non-MNU-treated (normal), MNU-treated (non-severe RD), and MNU-treated (severe RD) groups. Error bars represent the standard error of the mean (SEM). (B) Average population of electrical stimulus (ES)-responsive RGCs in the three groups. Statistical difference among the three groups analyzed using ANOVA, post-hoc test, and Tukey’s HSD test (^n.sp > 0.05, *p < 0.05, and ***p < 0.001).

Figure 3

Response pattern of RGCs to electrical pulse amplitude and duration modulation. (A) A typical well-modulated RGC response pattern in raster presentation (upper row) and post-stimulus time as histograms (PSTH, lower row). The pulse amplitude of electrical stimulation was modulated into five steps from 10 to 30 μA. With each pulse amplitude, the pulse duration of electrical stimulation was also modulated into three steps from 0.5 to 2 ms. The time bin of PSTH was 20 ms. The red vertical dashed lines in the raster plot and PSTH indicate electrical stimulus onset. The red horizontal lines in the PSTH indicate 99% confidence level. (B) RGCs classified as well-modulated, poorly modulated, and unmodulated based on their response pattern to electrical stimulation modulation.

Figure 4

ES-responsiveness of well-modulated RGCs to pulse amplitude modulation (PAM) and pulse duration modulation (PDM). (A) Average relative RGC response plotted against pulse amplitude from 0.5 to 2 ms fixed duration for non-MNU-treated (left), MNU-treated non-severe RD (middle), and MNU-treated severe RD (right) groups. Error bars represent the SEM. The threshold level is plotted as red dashed line at 1.5 times. (B) Mean evoked RGC spike number plotted against pulse duration from 10 to 30 μA fixed amplitude.

Figure 5

Comparison of threshold pulse amplitude and threshold pulse duration in non-MNU-treated and MNU-treated groups. Values of threshold (A) pulse amplitude and (B) pulse duration calculated from individual RGCs. The boxes show the range from first to third quartiles. The bottom whisker span from the first quartile box down to 5% and the upper whisker span from the third quartile box up to 95%. The midline of the box represents the median value of thresholds. The filled square symbol represents average value of thresholds. The thin horizontal bars represent 1% and 99%. The minimum and maximum value of threshold are plotted as cross symbols.

Figure 6

Strength-duration (SD) curve. (A) Averaged SD curves of well-modulated RGCs from three different groups. (B) Rheobase (left) and chronaxie (right) values calculated from individual well-modulated RGCs. Each SD curve is fitted with hyperbolic functions. Statistical significance among the three groups analyzed using ANOVA (^n.sp > 0.05).