20.109ResearchProposal:RetinalImplants

Overview

Retinal implants have been identified as a possible remedy for blindness. However, in addition to the problems encountered by other forms of transplantation, retinal transplantation carries with it the additional challenges of not being able to graft tissue from the host in the case of optic nerve death and difficulty in connecting existing nerves to an implant. In a recent paper, Yagi et al. have developed a retinal implant that could restore vision to patients to the extent of a “twinkling stars in the sky” image. In our project, we could think about ways to make the implant higher resolution, or add color-detection capability.

Background

Humans have two kinds of cells that detect light: rods and cones. Rods contain rhodopsin, and are low-light photoreceptors located at the periphery of the retina. Meanwhile, cones may contain one of three different kinds of photopsin, one each for long, medium, and short wavelengths that detect approximately red, green, and blue light, respectively. Cones are concentrated at the center of the retina, while rods are located near the periphery; there are also a much smaller number of cones than rods. (Campbell book)

Recently, a number of groups have tried to engineer retinal implants. However, engineering of retinal implants encounters challenges in addition to those encountered by tissue engineering in general for a variety of reasons. Firstly, in cases of optic nerve death, there is nowhere to get graft tissue from. Even when the optic nerve is still functional, it can be difficult to connect the retina to the optic nerve, so that even if an engineered, functional retina is implanted into an eye with a functional optic nerve, signals might not be sent to the brain to decipher an image. (Yagi et al). In addition, some groups have reported a thickening (as a result of scar tissue or necrosis?) of the connected nerves post-implant, even when this surgery is successful (**source?!?-Yagi?).

Recent research by Yagi et al. has developed a retinal implant that does not require a living optic nerve in order to stimulate sight...

Research Problems and Goals

...however, the resolution of their implant was low (1024 x 1024) and they could only detect light (not color). Our goal would be to engineer an implant that could generate higher-resolution, color images.

Project Details and Methods

The optic nerve is a bundle of fibers; presumably, each fiber should correspond to a cell or region of cells in the retina. We could produce color vision by stimulating not just rod cells, but also cone cells.

Problems would be:

• 1. Which cell(s) corresponds to which part of LGN (relay center for visual information in the brain).
• 2. How we can make something small enough to get a high-enough resolution image...
• 3. Metal things are bad, because of possible bio-incompatibility. How we can put things in a bio-compatibile matrix, with all bio-compatible parts, so that we don't need to use any silicon/metal based things.
• 4. Efficiency.

Some ideas:

• 1. We can figure this out by taking a group of cells in a normal (working) eye, stimulate them all at once, and see which neurons fire, maybe by fluorescence (FRET), since there is a conformational change in the gated ion channels when the action potential threshold is reached, or by replacing groups of neurons with nerves that stimulate other cells, and seeing what happens to these other cells if we stimulate different groups of rod/cone cells.
• 2. This is a bit of a problem in silico, because there is a limit on how small we can make chips, etc. at the moment, but shouldn't be an issue if we use cells. The problem, though is that we will need to connect the cells with the LGN ==> 3
• 3. Possibly use intermediate nerve cells between rod/cone cells and LGN.
• 4. If we have to place every nerve/cell, that would be very very hard.

Predicted Outcome

• If everything works:
• If nothing works:

References

Yagi, T., Watanabe, M., Ohnishi, Y., Okuma, S., & Mukai, T. (2005). Biohybrid retinal implant: research and development update in 2005. Proc 2nd Inter IEEE EMBS Conference on Neural Engineering, 248-51.

Campbell & Reece 7ed.