Switching on the Lights: The Use of Optogenetics to Advance Retinal Gene Therapy

Photoreceptors carry a high number of mitochondria necessary to maintain their high metabolic rate. This allows them to respond rapidly to small changes in illumination. However, the metabolically demanding role of light reception appears to render photoreceptors particularly vulnerable to mutations; this is manifested in the many clinical forms of hereditary retinal degeneration (RD).¹ Remarkably, in many cases the remainder of the retinal circuitry remains capable of firing despite the death of the photoreceptor cells.^2,3 In this issue of Molecular Therapy, Caporale et al. exploit this residual capacity for signal transmission so as to endow retinal ganglion cells with the ability to take over the task of light detection from the photoreceptors in conditions of RD.⁴

The retina is a thin sheet of photosensitive neural tissue that lines the inside of the eyeball. It comprises three distinct layers of cells that vertically transmit the information carried by photons from the light-sensitive rod and cone photoreceptors via bipolar cells to the inner ganglion cells. Individuals with RD are progressively losing their rods and cones and face the prognosis of blindness. Many of these patients cannot hope to benefit from the ongoing clinical trials for gene replacement or pharmaceutical therapies, and a restoration of light sensitivity to the retina in its degenerated state may be the only possible treatment. The work by Caporale et al. was enabled by the growing field of optogenetics, which involves the heterologous expression of light-sensitive channels or receptors in a neuron, allowing control of neuronal firing by an external light impulse.⁵ When applied to retinal gene therapy, these tools allow non-light-sensitive retinal cells to fire directly in response to light without the need to be triggered by the photoreceptor cells.³ These latest results from the Flannery laboratory⁴ support the potential of optogenetics to address the ambitious goal of vision restoration.

Several groups in retinal research have used optogenetic tools from nature, such as channelrhodopsin2 (ChR2; ref. 3) and halorhodopsin (ref. 6). By contrast, Caporale et al.⁴ made use of an artificial optical switch that was engineered to generate large currents in neuronal cells.⁷ The switch, developed by Volgraf et al., is based on the ionotropic glutamate receptor (iGluR) family.⁸ In addition to mediating the principal excitatory currents of the central nervous system, the iGluR family carries a structurally ubiquitous ligand-binding domain that serves as a suitable template to achieve generic cell signaling. To the iGluR ligand-binding domain is incorporated a chemical called azobenzene which also serves as the carrier for the ligand glutamate; the entire moiety is termed maleimide-azobenzene-glutamate (MAG; see Figure 1). The light-dependent mechanics of the switch is achieved by the light-sensitive azobenzene, which undergoes isomerization from a trans to cis configuration in response to short wavelengths of light (≈380 nm). This azobenzene isomerization induces the molecular motion that allows glutamate (or its analogs) to be drawn into the pore of the receptor and the cation channel to open. This is the trigger for membrane depolarization, the necessary step for the neuron to fire. The geometry is reversed and the channel closed under longer-wavelength light (≈500 nm) and in conditions of darkness. Caporale et al. applied this light-dependent iGluR6 prototype (LiGluR6) to the ganglion cell and demonstrated its potential as a platform for further engineering to meet the goal of reversal of blindness.

Figure 1.

LiGluR6-mediated therapeutic strategy in the retina. (a) Light signal transmission through the healthy retina. Rods and cones of the photoreceptor cell layer are shown at the top receiving photons. Signals are transmitted by specific cells in the bipolar layer (RBCs or OFF-BCs, depending on the wavelength and intensity of light) to ganglion cells at the bottom. (b) Light signal transmission through a degenerated retina expressing LiGluR6. The photons of light trigger LiGluR6 via activation of maleimide-azobenzene-glutamate (MAG) and lead to opening of the channel pore, bypassing the degenerated photoreceptors and the bipolar cell layer. (c) Detailed view of LiGluR6. MAG binding under 380-nm light activates the receptor and opens its cation-selective channel, resulting in membrane depolarization. a, azobenzene; g, glutamate; LBD, ligand-binding domain; OFF-BC, OFF-bipolar cell; RBC, retinal bipolar cell.

To accomplish their results, the authors achieved a high level of ganglion cell-specific LiGluR6 expression in the retinas of mice with retinal degeneration (rd1 mice). Tested ex vivo, electrical recordings show the retinas to be responsive to 380-nm light only when exposed to MAG, demonstrating the tight control of the switch. In vivo, photoresponses transmitted to the cortex were correlated in the behavior of entrained rd1 mice in a water-maze test. However, it is the restoration of the pupillary light reflex that sets this report apart from earlier studies of retinal optogenetic therapies. The pupillary light reflex refers to the contraction of the pupil as an involuntary response to light. This truly objective measure of non-image-forming photoresponse is mediated not only by the photoreceptors but also by a small subset of melanopsin-expressing ganglion cells.^9,10 The authors selected the ideal mouse model in which to test this response: triple-knockout mice that lack melanopsin in addition to essential components of the phototransduction machinery and therefore do not show a pupillary light reflex.¹¹ Caporale and colleagues show that the reflex is restored in the mutant mice if their ganglion cells express LiGluR6 and they have been treated with MAG.

It is important to note that targeting the ganglion cells may not take full advantage of the signal-processing power of the inner retina, where a lateral as well as a vertical transmission of information occurs.¹² Much of this processing takes place in the interneuron retinal bipolar cell layer before the light signal reaches the ganglion cells. Skipping the bipolar cells sacrifices to some extent two of their important processing roles: (i) discrimination of an object darker than background as compared with an object lighter than background and (ii) detection of the edges of an object whereby the center of the field is contrasted to an inhibitory surround. The absence of such processing would in theory lead to loss of resolution and noise reduction.

A more immediate obstacle to be overcome for clinical application of this strategy is the requirement that a sufficient level of the photoswitch MAG be delivered to meet the demands of newly turned-over iGluR6 channels. Encapsulated cell therapy may offer the best solution by allowing long-term administration of the glutamate analog in the eye.¹³ With this technology still under development, its incorporation may slow the progress of the LiGluR6 optical switch toward the clinic.

Nonetheless, LiGluR6 has a significant advantage over naturally occurring optical switches such as ChR2. The main deficiency of ChR2 is its high level of desensitization, which reduces the overall current that can be achieved.^5,14,15 In the future, channel modifications may improve such limitations; in the meantime the LiGluR6-MAG channel offers a higher level of channel conductance and can therefore achieve more adequate photon capture.

Mammalian opsin is a tough act to follow for those who endeavor to engineer photosensitive channels. For optogenetic retinal therapies to be translated to the clinic, specific hurdles must be overcome. Among these, the channels should be made receptive to more physiologically useful wavelengths, the efficiency of gene transfer to primate inner retinal cells needs to be improved, and the retinal cells that express the channel must be carefully selected so as to optimize visual processing while occupying regions of intact retinal circuitry.

Alternative strategies aimed at restoring the visual function lost in RD are being investigated, and these too have their own hurdles. Retinal regeneration through the manipulation of stem cells combined with tissue engineering has been the focus of intense research efforts and yielded some recent successes.¹⁶ However, the molecular mechanisms needed for appropriate stem cell differentiation remain largely unknown, as are the requirements for successful integration and survival of the derived cells following transplantation. The microelectronic retinal implant has also shown much promise since clinical trials commenced with the epiretinal chip (SecondSight) in February 2002 (ref. 17). The progress of these chips to full clinical use remains limited until improved resolution and full biocompatibility can be achieved.

As for the optogenetic approach, all these strategies require that layers of the inner retina and associated cell types maintain their architecture for a period of time following photoreceptor degeneration. Although this is broadly true in many instances of RD, it may nonetheless be of benefit to promote the parallel development of neurotrophic therapies that may optimally support the health of inner retinal cells. The study described here reminds us that if creative but methodical approaches are used to test strategies for vision restoration in advanced RD, the obstacles to the clinic will be overcome.