Some individuals having lost their eyesight further to degeneration of their photoreceptor cells can now benefit from an artificial retina. Fixed onto or beneath the retina, this implant allows them to perceive light signals once again. Several systems are currently in evaluation, and research is continuing with the aim of improving their performance and visual perception in patients.
Difficulty4 sur 5
Understanding the principle of the artificial retina
Located in the ocular fundus, the retina is made up of light-sensitive cells - photoreceptors - and a neuron network. Photoreceptors transform light signals into electric signals which travel to the brain via the optic nerve. Failure of the photoreceptors therefore impairs eyesight and may lead to blindness.
Artificial retina to replace photoreceptors. These are basically implants (measuring 3 x 3 mm) fixed onto or beneath the retina, made up of electrodes which electrically stimulate retinal neurons. The first devices tested in the 1990s comprised 16 to 20 electrodes. They now contain up to 1,500 electrodes. However, the patients' visual perception is not directly related to the number of electrodes.
Several devices are currently in evaluation. Three have already obtained CE marking from the European authorities - Argus II (Second Sight, United States), Retina Implant (AG, Germany) and IRIS II (Pixium Vision, France). Argus II has also been given the green light from the US agency (FDA) and is marketed in France in the context of the Forfait innovation (national innovation funding). The three existing devices yield fairly similar results. Research is being pursued with a view to improving the performance of these different systems.
How does the artificial retina function?
Two of the existing devices (Argus II and Iris II) are epiretinal implants: these are placed on the surface of the retina, in contact with the ganglion cells. The third (Retina Implant) is placed beneath the retina, where the photoreceptor cells are located. In all cases, surgical intervention is required in order to position the implant in contact with the retina.
For epiretinal models, cameras are incorporated into a pair of spectacles. These transmit visual information to a microcomputer placed in the patient's pocket or belt. This microcomputer converts the visual information into electric signals and transmits them via radio waves to a receptor placed on the eye. This then translates them into electric currents which travel along the wire into the eye, to reach the implant consisting of electrodes and fixed onto the retina. The Argus II implant contains 60 electrodes, and the Iris II model comprises 150 electrodes. Clinical trials on the latter are starting, following the conclusive results obtained with IRIS I comprising 49 electrodes.
The German implant (Retina Implant) operates in a different way. It comprises 1,500 electrodes and an equal number of light-sensitive diodes. The diodes transform light into electric currents which are amplified by an electronic circuit before being released by the electrodes. This system does not require spectacles, although the electric circuit connected to the photodiodes requires a power source. This necessitates the presence of a cable which runs from the chip beneath the retina, outside the eye, and behind the ear. The cable is very complicated to fit. In this device, each electrode does not produce an image pixel, which could explain why patients do not achieve the visual performance expected for an implant with so many electrodes.
Who is able to benefit from an artificial retina?
The artificial retina device is intended for blind individuals whose photoreceptor cells have degenerated, but whose retinal nerve cells and optic nerve are still functional. This corresponds to patients with retinitis pigmentosa and age-related macular degeneration (AMD).
Retinitis pigmentosa is a genetic disease which affects approximately 1 in 4,000 births (approximately 30,000 cases in France). Patients live with a gradually shrinking visual field, then lose their eyesight, generally after the age of 50.
AMD is an age-related disease. The central visual field is affected, which makes it difficult to read or recognize faces. A quarter of individuals aged over 75 are affected. Despite recent progress in the management of certain complications, no treatments are currently able to cure these two diseases.
At present, the implants are fitted in patients in the terminal stages of retinal degeneration, hence the functional capacity of the tissue remains very limited. Earlier intervention would yield better results. This should, however, be demonstrated by clinical studies.
What are the current results?
In France, approximately twenty individuals suffering from retinitis pigmentosa have already undergone implantation of the Argus II system, thanks to the Forfait innovation (national innovation funding) by the Ministry of Health which provides funding for the device (approximately EUR 90,000 per unit). All of the patients, with ages ranging from 29 to 73 years, had a satisfactory surgical outcome, and the majority are in the rehabilitation phase. In total, 36 patients will receive the implant, in three hospitals: CHNO des Quinze-Vingts in Paris, Bordeaux University Hospital and Strasbourg University Hospital.
These patients are managing to perceive light signals. Some of them are able to walk alone, identify a door or window in a room, visualize crosswalks or even follow a line on the floor. Some of these patients are even able to read words in large white font on a black background, on a computer screen, or even short sentences. However, there are major differences in the results obtained. This could be partly explained by the variability in degree to which the retinal tissue has been generally preserved from one patient to another.
In all cases, the patients do not regain their previous eyesight. They need to learn a new way of seeing, by associating and ordering the perceived light signals. This requires several hours of exercises each day. The current resolution proposed by the Argus II implants is 60 pixels, and that for the new IRIS II implants is 150 pixels.
Challenges facing research
What progress is expected? The ultimate goal is to offer a bionic eye which makes it possible to recognize faces, to read or even walk around with complete autonomy. In order to perform these complex tasks, studies suggest that a minimum resolution of 600 to 1,000 pixels would be sufficient.
However, the performance of the German implant which comprises 1,500 electrodes is very limited. This demonstrates that it is not enough to multiply the number of electrodes to increase the resolution. This is limited by the retinal surface upon which glial cells form an insulating layer between the neurons and the implant. This makes it difficult for the electric currents transmitted by each electrode to focus on a limited group of neurons. Moreover, the signals are diffused in such a way that the stimulation produced by two adjacent electrodes overlaps. To improve the resolution, it is essential for each electrode to receive and emit a specific electric signal, and for it, in turn, to stimulate a zone independent from its neighbors, in an extremely focused manner. In other words, point-by-point rather than diffuse coding of visual information is desired.
Procure clearer images for patients
This is what the Institute of Vision has been working on, in partnership with Stanford University. The researchers used an implant developed by Stanford, fitted with return electrodes enabling this individual processing of the signal emitted by each electrode. This high-tech version was tested on the retinas taken from blind primates and placed in culture. The researchers measured the activity of the ganglion cells according to the stimulation emitted. They thus showed that individual ganglion cells may be specifically activated by a single electrode. This experiment demonstrates the very high spatial resolution of this new implant. The prototypes were manufactured in France, by the Pixium Vision firm which is continuing to pursue the development of this "new-generation" PRIMA implant. This is a sub-retinal implant which does not require a power cable. It is limited to a plate comprising photoelectric diodes which are stimulated by infrared light signals transmitted by spectacles. The model, initially tested with 200 to 600 electrodes, could ultimately contain 1,000 to 2,000 electrodes, or even more. The company is planning to perform the first implant in humans in the course of 2016.
The researchers from the Institute of Vision are also developing 3D implant technology in collaboration with the ESIEE (Paris Chamber of Commerce and Industry Institute of Higher Scientific Education). Instead of being flat, these implants have contours, in the form of wells housing the retinal neurons. This enables electric stimulation to be focused on the neurons introduced into the wells. This implant technology has been successfully tested on animals, showing that the neurons are indeed effectively introduced into the wells. Modeling of the currents demonstrates the increase in the spatial resolution of these 3D devices for stimulating independent pixels. Ultimately, these 3D implants should make it possible to produce a very fine resolution of electric stimulation.
Improving rehabilitation protocols
Another possible area for improvement concerns the rehabilitation methods used to train patients to get the most out of their implant. Hospital teams are working on these methods. For example, clinical practitioners hope to improve results by teaching patients how to reprogram the implant themselves. The Streetlab firm, which has constructed an artificial street at the Institute of Vision, is currently very involved in developing these rehabilitation protocols.
Real-time data integration
With the prospect of ultra-high-performance artificial retinas, mathematicians from the Institute of Vision are studying the process for transforming visual information into electric stimulation. The expected increase in the number pixels in the implants will require longer calculation times for the images, hence increased energy consumption necessitating a large battery. To circumvent these problems, the Institute of Vision has developed a new imageless visual information sensor, so as to limit the burden of the visual information processed. This sensor has already been incorporated into the spectacles designed by Pixium Vision.
Increasing electrode performance
The implants, which are well tolerated, are currently made from polymers widely used in medicine, with metallic electrodes. However, these materials can degrade over time, and electrode efficacy could be further improved. The researchers are therefore studying new materials, such as graphene and diamond. The latter, being extremely stable, is endowed with excellent semiconductor properties and extreme biocompatibility. In collaboration with CEA-List, the researchers from the Institute of Vision have confirmed the biocompatibility of the material with retinal cells, both ex vivo and in animal studies. They are now evaluating the electric properties of diamond in terms of neuron stimulation. This notably involves determining whether it is possible to reduce the electric charge sent into the tissue.
Optogenetics, another way of restoring vision
A technique currently in development could soon rival the artificial retina: optogenetics. This strategy involves introducing a light-sensitive protein able to create an electric current into the retinal neurons through the cell membrane. These proteins exist in the natural state, and are known as opsins, such as channelrhodopsin-2 derived from an algal species or halorhodopsin derived from a type of bacteria. Expression of these proteins may be obtained in a nerve cell by introducing their gene, via a viral vector. Hence, the neurons in a "blind" retina can become sensitive to light and transmit an electric signal to the brain once again.
Conclusive trials have been conducted in blind mice, and on the retinas of primates. These initial results show that the ganglion cells are able to express opsins after intravitreal injection of the viral vector. These cells are excited by very short stimulations (a few milliseconds) with intense light. The Institute of Vision is currently working on developing an asynchronous camera, built into spectacles, able to transmit only certain wavelengths, not toxic to the eye, and able to stimulate opsins. An American firm (Retrosense) has announced the start of a clinical trial (on one patient). In France, the Gensight Biologics firm is also expected to start clinical trials in the course of 2017.