Artificial vision: testing of glutamate releasing nano-pore membranes and assessment of Müller cell glutamate uptake

Abstract

Purpose : A glutamate (glu) -based retinal prosthesis is envisioned to promote enhanced artificial vision, since its electrical counterpart fails to activate retinal ON and OFF channels adequately. The principle of artificial glu stimulation (stim) of retinal tissue was demonstrated previously. In this study we focus on implant-related matters: 1) Glu uptake-capability of Müller cells (MC) in the degenerated retina, 2) and testing of glu releasing nano-pore membranes.

Methods : All experiments were conducted in mouse models for retinal degeneration (rd1; age 15-30 d). 1) Retinal explants were exposed to different glu concentrations (bath-applied, n = 9 retinal explants). Outer retinal activity was captured by Ca²⁺ imaging. Movement of somas was assessed from Ca²⁺ imaging data as indicator for MC swelling, which is caused by glu uptake. 2) We developed a planar device for multi-site glu translocation at nano-scale. It consists of an ion-milled, nano-patterned silicon nitride membrane connected to a microfluidic system allowing controlled glu release through the membrane’s nano-pores. Retinal explants were placed on the membranes (n = 3, each tested with separate retina), outer retina interfacing the nano-pores. Glu (500 mM) was pulsed into a cavity under the nano-pores, ganglion cell (GC) activity was recorded by Ca²⁺ imaging. Temporal characteristics of the elicited activity were assessed in Ca²⁺ traces, extracted from imaging stacks and plotted with glu pulse time stamps. A diffusion time for glu through the nano-pores was estimated as difference between time stamps and start of GC responses (> δ pre-pulse baseline).

Results : 1) Glu concentrations ≥ 1 mM induced detectable, concentration-dependent retinal swelling, caused by MC glu uptake: 1 mM: 7.76 ± 2.33 µm, 3 mM: 14.52 ± 3.64 µm, 5 mM: 19.20 ± 3.60 µm. 2) Glu pulses elicited correlated GC responses. An average glu-membrane diffusion time was estimated at 7.41 ± 0.31 s (n = 3 membranes, m = 45 pulses, one-way ANOVA, p > 0.05).

Conclusions : MC of rd1 retina are capable of uptaking artificially applied, crucial for repetitive cellular responses in long-term glu stim applications. Nano-pore membranes were successfully tested: controlled glu release triggers network-mediated GC responses. The results are promising steps towards enhancing artificial vision with a naturalistic glu-based implant.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.