Restoring partial vision to a blind patient

Abstract

This paper reports an important breakthrough in partially restoring sight to a man who had lost his vision due to retinitis pigmentosa (RP), a heritable retinal degenerative disease that affects approximately 1 in 4000 people. Long considered an insurmountable challenge, a stellar team of vision scientists, engineers, basic biologists, and others, working together for many years, has enabled a man who had been legally blind for decades to begin distinguishing objects and navigating his environment¹.

Background

Retinitis pigmentosa (RP) is a heritable and largely incurable disease in which rods and cones, the retinal photoreceptors that respond to light, degenerate². Rods are sensitive to low levels of illumination, and their loss is manifest in the loss of night vision. Cones, which mediate high-resolution and color vision at higher levels of illumination, depend on rods for their survival and consequently die secondarily, leading to severely restricted vision and in some cases complete blindness. Although gene therapy to replace or repair a defective gene in rods is feasible for some forms of RP³, this approach is of limited applicability because RP can be caused by mutations in many different genes, and there is the need for genetic intervention prior to significant rod death. Despite the loss of rods and cones, however, RP patients still retain most of their retinal ganglion cells (RGCs), the neurons that integrate the signals that originate in the rods and cones and that become transformed through the retina’s intrinsic circuitry. RGCs transmit these signals through axons that course through the optic nerve to subcortical relay centers, the principal ones of which in turn send information to the areas of our cortex that enable us to distinguish objects, faces, places, text, and so on. Consequently, if RGCs can be modified to become directly sensitive to light, some vision, albeit of a lower quality than normal, might be possible. However, to realize this concept, numerous technical hurdles first have to be overcome, including optimization of a protein that will cause RGCs to become electrically active when illuminated; design of a viral vector suitable to transduce RGCs with a transgene encoding a light-gated ion channel; and design of an electro-optical device that captures continuous images of the outside world, transforms the images into a monochromatic array, and relays signals back to the eye to illuminate RGCs with light of an appropriate intensity and wavelength so as not to damage the retinal cells.

Main contributions and importance

This study represents the culmination of many years of interdisciplinary research. Important features of this work include:

(1) Gene therapy using an optimized viral vector to transduce RGCs.

(2) A gene encoding a genetically engineered light-sensitive ion channel (ChrimsonR). When illuminated by light of a particular wavelength and intensity (so as not to damage RGCs), the protein becomes activated (in much the same way as RGCs open ion channels in response to stimulation by second-order retinal bipolar cells), depolarizing RGCs to the point of being able to generate action potentials that are relayed through the optic nerve to the brain.

(3) Specially designed goggles that capture pixelated images of the outside world through a camera on the goggles.

(4) The relay of the image to an external processor whose output back to the goggles is a fine-grained array of near-infrared stimuli of an appropriate intensity that is projected onto the RGCs.

Using a carefully designed, well-controlled set of experiments, Drs. Sahel, Roska, and colleagues show that a patient who merely had light perception, i.e., had only been able to distinguish light from dark, could now identify and reach for objects when the goggles projected information about the outside world onto RGCs expressing the light-sensitive protein (see Figure 1). The study included data showing changes in the activity of the visual cortex indicative of visual processing. Control experiments showed that the patient showed no visual responsiveness when not wearing the goggles nor when the external processor was off. The psychometric assays provide a “real world” view of how this treatment might actually benefit patients – how it can actually improve their function in day-to-day living.

Figure 1. A man with retinitis pigmentosa who had been blind for decades points to a small object following gene therapy while wearing special goggles.

Photoreceptors are the cells in the back of the eye that respond to incoming light and then relay signals through the retina’s complex circuitry to retinal ganglion cells (RGCs). RGCs in turn send information about the outside world through the optic nerve to the visual processing centers of the brain. To restore vision, a gene that codes for a specialized protein, CrimsonR, was introduced into the man’s RGCs. When activated by light, CrimsonR opens a channel that enables electrically-charged particles to flow across the cell’s outer membrane, activating RGCs and causing signals to be conveyed through the optic nerve back to the brain. The goggles that the man is wearing contain a small camera that transmits signals to a computer, which in turn sends signals of the correct wavelength back to the eyes to activate CrimsonR in RGCs and provide the brain with information about the visual scene. A brief video of Dr. Sahel describing this discovery can also be found on YouTube: https://www.youtube.com/watch?v=VtlToPVqJp8. This non-copyright figure was kindly provided by Dr. Sahel.

Open questions

The present study reports the results of a single patient. Were there other pre-existing (or acquired) problems that influenced this patient’s recovery? In addition to RP, the patient exhibited divergence strabismus, probably as a secondary consequence of vision loss. The study originally included multiple subjects but was hampered by COVID-related restrictions. However, at a recent meeting of the Association for Research in Vision and Ophthalmology (ARVO)⁴, Dr. Sahel reported the outcome of a study with a second patient whose visual recovery was similar to that of the patient described in the Nature Medicine paper¹.

If the visual resolution achieved by this approach increases, could the strategy used here be applied to a broader population of patients with eye diseases? With the current technology, can greater visual recovery be achieved with more time and experience, perhaps owing to neural plasticity in central visual processing centers? What is the ultimate “best resolution” that one can achieve with this approach?

Can the approach be improved upon? The current study converted the electrophysiologic properties of all transduced RGCs to a single type of light response, thereby losing the diversity of responses of different RGC subtypes that contribute to vision^5,6. Are there modifications to the present strategy that could allow for greater levels of visual recovery? One possibility is that two or more viral vectors encoding different light-sensitive ion channels could be used with appropriate promoters to enable different classes of RGCs (e.g., ON-center OFF-surround vs. OFF-center ON-surround) to respond to different features of the image. An alternative or additional strategy would be to use other RGC subtype-specific promoters to activate RGCs with transient vs. sustained responses. These strategies would be expected to increase visual resolution over the present approach, which stimulates all RGC subtypes indiscriminately, blurring together signals that normally convey more refined aspects of the visual scene. A related issue is whether projecting an image onto the retina is the best way to stimulate RGCs or can more advanced methods of stimulation be developed to increase the sensitivity to certain aspects of the visual environment (motion, edges, etc.)?

Is the crude vision that was restored here using optogenetics superior to what has been achieved thus far with retinal or cortical prostheses? And what is the relative potential of these different approaches? Are there other advantages of the present technology compared to retinal or cortical prostheses? The present technology is minimally invasive, requiring “only” an intravitreal injection. Retinal prostheses (e.g., Argus II) require ocular surgery to place an electrode array (activated through an external device) over the surface of the retina and, in addition to activating RGCs in a spatially organized fashion, also potentially stimulating axons emanating from all parts of the retina, thereby blurring the image received by the brain. A related, optogenetic approach has been shown in the nonhuman primate retina in vitro to enable RGC responses with high spatial resolution without simulating RGC axons⁷. Cortical implants require brain surgery to implant a device on the surface of the cortex. If these devices fail, replacement will involve new surgery.

Why was there such a long delay before visual responses could be detected? Is there a placebo learning effect? Prolonged inactivity of RGCs likely leads to altered structure of the axon initial segment and a reduction in myelin thickness. The general view of the panel was that visual recovery required plasticity, likely at the cortical level, in learning how to interpret the images after many years of disuse. Additionally, learning was presumably also needed because the visual signals arriving from the RGCs did not have the benefit of normal inner retinal signal processing. Another example of prolonged visual learning is the time required for patients with decades of vision loss due to cataracts to be able to interpret visual images after the cataracts are removed (Project Prakash, led by Dr. Pawan Sinha, MIT)^8,9.

What is the circuitry underlying the visual recovery? An earlier study by this group had shown that the viral vector used here in fact transduces other cells of the foveal region, in addition to RGCs. But is this a problem, or does it only add to the complexity of understanding the visual recovery?

Another issue is that of light and dark adaptation. As a result of photoreceptor mechanisms, our vision can function over about 10 log units of intensity. Can optogenetic vision do that?

Is there a conflict of interest? Dr. Sahel, Dr. Roska, and other investigators in this study are involved with GenSight Biologics, a company that has played a role developing and commercializing this technology. Although the panel members thought that this fact needed to be mentioned, they did not see it as a concern in view of the stellar reputations of the investigators and the government of France’s encouragement of such academic-commercial enterprises.

Is this study truly a landmark? Research in non-human primates by this group of investigators has foreshadowed the present study, and others are using similar methods to transduce bipolar cells to express light-sensitive ion channels. Nonetheless, the panelists as a whole strongly believe that the study was truly a landmark as the first clear demonstration of visual improvements in a person who had been virtually blind for decades.

Conclusion

Although based on a single case, this study in a leading journal is the first to report optogenetics-based visual recovery in a patient who had been virtually blind for decades. Because RP can arise from many different genetic mutations, developing gene therapy to address each one is a major challenge. The present approach bypasses the photoreceptors and avoids the need to address each mutation individually by enabling retinal ganglion cells to respond directly to light signals projected through an electro-optical device. By doing so, this approach can likely be applied broadly to most, and perhaps all, patients with RP and potentially other blinding conditions. Unlike retinal or cortical prostheses, this technology does not require major surgical intervention and lends itself to further improvements, e.g., simultaneous activation of different classes of RGCs to convey a more refined image of the outside world to the brain. The paper engendered lively discussion among the panelists, the majority of whom felt that, despite the precedents in animal studies, this study is a landmark by virtue of being the first to demonstrate the feasibility of using optogenetic approaches to restore at least some vision to blind human patients with photoreceptor loss.