Vision Research 111 (2015) 134–141

Contents lists available at ScienceDirect

Vision Research

journal homepage: www.elsevier .com/locate /v isres

Photochemical approaches to vision restoration q

http://dx.doi.org/10.1016/j.visres.2015.02.0010042-6989/� 2015 Elsevier Ltd. All rights reserved.

q This work supported in part by US National Institutes of Health Grants PN2EY018241, P30 EY01730, R24 EY023937 and an unrestricted Grant from Research toPrevent Blindness – USA.⇑ Address: Department of Ophthalmology, University of Washington Medical

School, 325 9th Avenue, Seattle, WA 98104, United States. Fax: +1 206 543 4414.E-mail address: russvg@uw.edu

Russell N. Van Gelder ⇑Department of Ophthalmology, University of Washington School of Medicine, United StatesDepartment of Pathology, University of Washington School of Medicine, United StatesDepartment of Biological Structure, University of Washington School of Medicine, United States

a r t i c l e i n f o a b s t r a c t

Article history:Received 22 May 2014Received in revised form 6 January 2015Available online 11 February 2015

Keywords:Retinal degenerationPhotoswitchPharmacologyRetinal ganglion cells

Photoswitches are traditional pharmacologic agonists, antagonists, or channel blockers that are covalent-ly modified with an azobenzene derivative. Azobenzene undergoes wavelength-dependent isomerizationbetween cis and trans conformation. For some photoswitches, only one of these configurations isbiologically active, resulting in light-dependent activation or inhibition of function. Photoswitches thatfeature a quaternary ammonium coupled to the azobenzene moiety cause light-dependent neuronaldepolarization due to blockage of voltage-gated potassium channels. Two photoswitch strategies havebeen pursued. In the one-component strategy, the photoswitch is applied to native receptors; in thetwo-component strategy, the photoswitch is combined with virally-mediated expression of a geneticallymodified receptor, to which the photoswitch may covalently bind. The former approach is simpler but thelatter allows precise anatomic targeting of photoswitch activity. Acrylamide-azobenzene-quaternaryammonium (AAQ) is the prototypical first-generation one-component photoswitch. When applied to reti-nas with outer retinal degeneration, ganglion cell firing occurs in response to blue light, and is abrogatedby green light. In vivo, AAQ restored pupillary light responses and behavioral light responses in blind ani-mals. DENAQ is a prototypical second generation one-component photoswitch. It features spontaneousthermal relaxation so cell firing ceases in dark, and features a red-shifted activation spectrum. Interest-ingly, DENAQ only photoswitches in retinas with outer retinal degeneration. MAG is a photoswitched glu-tamate analog which covalently binds to a modified ionotropic glutamate receptor, LiGluR. When appliedtogether, MAG and LiGluR also rescue physiologic and behavioral light responses in blind mice. Together,photoswitch compounds offer a potentially useful approach to restoration of vision in outer retinaldegeneration.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction In theory, if one could recapitulate the native temporal firing

Acquired outer retinal degeneration associated with age-relatedmacular degeneration is the leading cause of blindness in the Unit-ed States and most of the developed world, while outer retinaldegeneration associated with retinitis pigmentosa is the leadinginherited cause of blindness. Blindness in both conditions resultsfrom irreversible loss of the rod and cone photoreceptors fromthe outer retina. The remaining retinal cell types – bipolar cells,horizontal cells, amacrine cells, Muller cells, and retinal ganglioncells – remain viable (although significantly rewired (Marc &Jones, 2003)).

pattern of each retinal ganglion cell in response to a dynamic visualscene, one could ‘restore’ vision to a retina devoid of rods andcones. Progress has been made on several approaches to this goal.Embryonic or inducible progenitor stem cells can be differentiatedto a retinal fate and transplanted; this approach restores some lightsensitivity in mouse models of retinitis pigmentosa (Lamba, Gust,& Reh, 2009). Second, external stimulation of retinal ganglion cellsby opto-electronic prosthetics has also advanced significantly inthe past decade, with Food and Drug Administration approval ofthe Argus II device, capable of restoring some visual phenomenawhen implanted in profoundly blind individuals (Ahuja et al.,2011; Humayun et al., 2012). Third, virally-mediated gene therapyapproaches using optogenetics, including channelrhodopsin andhalorhodopsin to excite and suppress retinal cells, respectively,have advanced in the past decade, and have also been shown torestore vision-like function to mice with advanced outer retinaldegeneration (Ivanova, Roberts, et al., 2010; Lagali et al., 2008;Tomita et al., 2007).

R.N. Van Gelder / Vision Research 111 (2015) 134–141 135

Collectively, these approaches offer substantial hope for visionrestoration in the foreseeable future. However, all share a commonfeature of effective irreversibility. Once stem cells are placed in thesubretinal space, or viral gene therapy vectors introduced into theeye, these alterations are effectively permanent. Similarly, it wouldbe challenging to replace an optoelectronic prosthesis once surgi-cally implanted in the eye. Given the barriers to demonstrationof efficacy and regulatory approval of any prosthetic or pharmaco-logic approach, the irreversibility of these treatments makes itera-tive progress in this field challenging.

Over the past 8 years, a new approach has been advanced forvision restoration. Compounds such as tetra-ethyl-ammoniumwhich block voltage-gated potassium channels have been knownfor over a half century to induce neuronal cell firing when admin-istered extracellularly (George & Johnson, 1961). By linking theseagents to a photo-isomerizable moiety (azobenzene), novel com-pounds have been produced that can activate neurons in a light-dependent fashion. When applied to retinas with outer retinaldegeneration, these compounds are capable of inducing light-de-pendent firing of remaining retinal ganglion cells. Such activationmay form the basis for reconstitution of vision in eyes with outerretinal degeneration.

2. Azobenzenes

The light-isomerizable moiety utilized in these compounds isazobenzene (Fig. 1). This class of compounds was among the first

Fig. 2. Structure of maleimide-azobenzene-quarternary ammonium (Mal-Azo-QA, or Msystem. (A) Structure of MAQ in trans and cis conformation. (B) Schematic of two-compowith light due to photoswitch activity. From: Mourot, Tochitsky, and Kramer (2013)) w

Fig. 1. Chemical structure of azobenzene in trans and cis conformations.

described in organic chemistry, having been characterized by Eil-hard Mitscherlich in 1834. Azobenzenes are produced by reductionof nitrobenzene; originally this was accomplished with iron filingsas catalyst (more recent methods utilize zinc). Azo dyes wereextensively used throughout the late 19th and early 20th centuryin industry to color clothing and foodstuffs.

The property of azobenzene and its derivatives that renders ituseful for photopharmacology is the photoisomerization of cisand trans isomers. The isomers can be interchanged with specificwavelengths of light. The trans-to-cis conversion utilizes shortwavelength light (typically ultraviolet for unsubstituted azoben-zene), while the cis-to-trans isomerization occurs under longerwavelength light (blue light for the unsubstituted azobenzene).The cis isomer is less stable than the trans and so cis-azobenzenewill thermally convert back to the trans in dark. Photoisomeriza-tion of trans to cis azobenzene is extremely rapid (on the order ofpicoseconds), while the reversion in dark of cis to trans takes hoursfor unsubstituted azobenzene. However, substitution of electron-withdrawing groups can alter both the thermal stability of thecis-form, as well as the action spectrum for isomerization(Banghart, Volgraf, & Trauner, 2006). Because of these propertiesazobenzene derivatives have been used in a variety of nanotech-nology applications ranging from surface holography, liquid crystaldisplays, and optoelectronics (Brehmer et al., 1997; Choi et al.,2007; Matharu, Jeeva, & Ramanujam, 2007; Nakatsuji, 2004;Oliveira et al., 2005).

The first successful application of azobenzene-mediated controlof a potassium channel blocker was the development of the SPARKchannel (Chambers et al., 2006). SPARK stands for synthetic photoi-somerizable azobenzene-regulated potassium channel. In thisapproach, Chambers et al. synthesized a modified Shaker potassi-um channel (including an EGFP tag for visualization). In one ver-sion, the Shaker channel was left in its native, potassiumconducting form (H-SPARK), while in another a point mutationconverted the channel to a non-specific cation channel capable ofconducting extracellular sodium (D-SPARK). The small moleculephotoswitch employed was maleimide-azobenzene-quaternaryammonium (MAL-AZO-QA), with the latter moiety responsible

AQ), the small molecule photoswitch used in the two-component SPARK channelnent blockade under trans conformation. (C) Example of change in cell conductivityith permission.

Fig. 3. Structure of acrylamide-azobenzene-quarternary ammonium (AAQ).

136 R.N. Van Gelder / Vision Research 111 (2015) 134–141

for channel blockade (Fig. 2). The maleimide moiety allowed forcovalent coupling of the compound to a cysteine on the cellsurface.

SPARK-expressing CHO cells showed no change in channel func-tion prior to addition of 300 lM MAL-AZO-QA, and light had noeffect on these cells without presence of both transgenic channeland azobenzene-based photoswitch. However, after transfectionand administration of MAL-AZO-QA, expected changes in the slopeof the current–voltage relationship of patched clamped cells forboth D-SPARK and H-SPARK were observed. Further, D-SPARK-ex-pressing cells showed hyperpolarization when cells were changedfrom 390 nm light to 505 nm light, and depolarization with theopposite lighting change. The opposite effect was seen with H-SPARK channels (i.e. depolarization with the UV ? blue-greenlighting change). Cells expressing GFP only or without MAL-AZO-QA showed no changes. When tested in mammalian hippocampalcells in culture, cells transfected with D-SPARK and treated withMAL-AZO-QA showed light-dependent action potential firing in vit-ro. Taken together, these results established that an azobenzene-based small molecule potassium channel blocker is capable ofinducing light-dependent cell firing in mammalian neurons.

The SPARK system constitutes a ‘two-component’ photoswitch,requiring both transgenic expression of a modified receptor andaddition of the small molecule photoswitch. Subsequent researchhas pursued both this two-component approach, as well as aone-component approach which does not require genetically mod-ified channels.

3. One component photoswitches

The first milestone in development of one-component photo-switches was the synthesis of AAQ (Banghart et al., 2009; Fortinet al., 2008, 2011). Where MAQ had a maleimide group at its distalend, AAQ substituted an acrylamide (Fig. 3). This compound wascapable of photosensitizing wild-type voltage gated potassiumchannels. In vitro studies suggested that AAQ induced light-dependent blockade of a variety of Kv potassium channels (Fortinet al., 2008). Initially, it was thought the compound bound on the

Fig. 4. Reconstitution of retinal ganglion cell firing to rd1/rd1 mouse retina by addition ofmulti-electrode array raster, with each line representing one ganglion cell. Note strong fi(2012) with permission.

extracellular side of the receptor, where the acrylamide moietycould bind covalently to the receptor surface; however, subsequentstudies suggested that AAQ acts on the internal binding site anddoes not require covalent attachment to function (Banghart et al.,2009). Initial studies were performed on hippocampal slice prepa-rations (Fortin et al., 2008). Light-dependent cell firing wasobserved after addition of micromolar extracellular AAQ. Greenlight (530 nm) resulted in cell firing, while use of short wavelengthlight (390 nm) resulted in silencing of activity.

These characteristics suggested that AAQ might be used torestore light-dependent firing of retinal ganglion cells in animalswith outer retinal degeneration. Using isolated retina from rd1/rd1 mutant mice (with extensive outer retinal degeneration),Polosukhina et al. (2012) demonstrated that AAQ applicationrestored light-dependent firing activity to these mice as assayedby multi-electrode array recording (Fig. 4). Interestingly, the spec-tral sensitivity of AAQ-activated retina was reversed from that seenin hippocampus: short wavelength light (380 nm) activated cellfiring, while longer wavelength light (500 nm) abrogated firing.Retinas activated by 380 nm light showed firing activity of manyseconds on direct transition to dark, reflecting the slow thermalrelaxation of trans to cis conformation. The observed spectral sen-sitivity was reversed when retinas treated with GABAergic andglycinergic blockade were tested following AAQ application. In thiscase, retinal ganglion cell firing was reduced in 500 nm light, andminimally affected by 380 nm light. This result suggests that theprimary effect of AAQ on the retina is to light activate amacrinecells, which in turn inhibit retinal ganglion cell firing. Other inter-pretations are possible, however, including the possibility that thecis form of AAQ blocks channels present in retinal ganglion cells ofdegenerated retina.

The intensity of light required to activate firing was relativelybright, beginning at about 1015 photos/cm2/s and reaching mid-point at 1016 photos/cm2/s. In this study, the authors were alsoable to look at receptive fields of activated cells. They found a posi-tive ‘photoswitch index’ (a measure of relative firing rates in lightbefore and after administration of the photoswitch compound)when measuring cells within 200 lm of the light target, but didnot see positive switching for cells further away. This confirmsthe suggestion that AAQ effect reflects the local light environmenton the retina.

Several assays were employed to determine whether AAQ-mediated light-induced retinal ganglion cell firing resulted insignals interpretable by the central nervous system of blind mice.rd1/rd1;opn4�/� mice lack pupillary light responses (Panda et al.,2003). Following intravitreal administration of 1 lL of 80 mM

AAQ, as measured by multi-electrode array (MEA) recording. Upper panel in each isring under UV light and relative silencing under green light. From: Polosukhina et al.

Fig. 5. Structure and absorption spectra of second generation photoswitches. From: Mourot et al. (2011) with permission.

R.N. Van Gelder / Vision Research 111 (2015) 134–141 137

AAQ, partial pupillary light responses were restored toapproximately 35% of animals. In a second behavioral assay ofnegative phototaxis (modified from (Johnson et al., 2010)), inwhich time to turn from a bright light stimulus is measured, rd1/rd1;opn4�/� demonstrated no negative phototaxis, but the samemice showed significant light aversion following intraocular injec-tion of AAQ.

Taken together, these results demonstrated (1) azobenzene-based potassium channel blockers are capable of restoring light-dependent firing activity to retinal ganglion cells of animals blindfrom outer retinal degeneration, and (2) these signals are at leastcrudely interpreted by the brain and can influence physiologyand behavior.

AAQ has several drawbacks as an agent which make furtherdevelopment for potential human use problematic. First, AAQ isactivated by light of 380–400 nm. The human lens filters the vastmajority of these wavelengths and so achieving the bright intensi-ties necessary for AAQ activation at the level of the retina would beproblematic. Second, the slow relaxation of AAQ in dark wouldrequire switching between short and long wavelengths (with anattendant auxiliary visual stimulating device). Third, the presenceof the acrylamide moiety on AAQ raises the possibility of neuro-toxicity (although this has primarily been associated with themonomeric form of acrylamide (Spencer & Schaumburg, 1974a,1974b, 1975)).

4. Second generation one-component photoswitches

To address these limitations of AAQ, the Trauner laboratorysynthesized a series of ‘second generation’ one-componentphotoswitches by substituting the azobenzene moiety with more

electrophilic groups (Fig. 5). These substitutions result in substan-tial red-shifting of the action spectrum these compounds, broaden-ing of their activation spectrum, and rapid relaxation in dark(Mourot et al., 2011). One compound, with a diethylammoniumsubstituted for the acrylamide (DENAQ) had reasonable solubilityand was selected for further testing.

In vitro, DENAQ showed good activation of retinal ganglion cellsof rd1/rd1 mice with robust firing in response to light from 360 nmto over 500 nm (Fig. 6) (Tochitsky et al., 2014). DENAQ-activatedretina also showed higher sensitivity than AAQ, with light respons-es ex vivo starting in the mid 1013 photons/cm2/s range of intensity.Following a single intravitreal injection, light responses could beelicited from treated retinas for 3–4 days.

Surprisingly, DENAQ did not affect function of wild-type retinas,and did not appear to induce photoswitching in these retinas. Thesame lack of effect was seen ex vivo in retinas under glutamatergicblockade, isolating the retinal ganglion cells. This strongly suggest-ed that some aspect of retinal degeneration was creating an oppor-tunity for DENAQ to act, either through an upregulated channel orperhaps by increased access for the drug into retinal cells.Tochitsky et al. (2014) found that that the Ih conductance, mediat-ed by the HCN channel, was photoswitched in rd/rd retinas but notin retinas from wild-type animals. Use of HCN channel blockersivabradine and cilobridine attenuated or eliminated DENAQ-in-duced photoswitching, providing additional supporting evidencefor the HCN conductance being a major target of DENAQ activity.

In vivo, intravitreal injection of DENAQ in rd/rd mice resulted inaltered free range behavior (with increasing activity in light).Additionally, rd/rd mice injected with DENAQ were able to associ-ate light exposure with a noxious stimulus in a learning paradigm,performing equivalently to wild-type animals in this assay.

Fig. 6. Conferral of light sensitivity on retinal ganglion cells of rd1/rd1 mice by DENAQ. (A) Structure of DENAQ. (B) Multi-electrode array recording of rd1/rd1 retina prior toDENAQ administration. (C) Recording from rd1/rd1 retina post-administration of DENAQ. Note cessation of firing in dark. From: Tochitsky et al. (2014) with permission.

138 R.N. Van Gelder / Vision Research 111 (2015) 134–141

5. Two component photoswitches

Development has proceeded as well with two-component pho-toswitches. In this paradigm, treatment of blind retina entails bothgene therapy with a genetically modified cell surface receptor, andadministration of a photoswitch specific for that receptor. Unlikeintroduction of transgenic photoreceptors for treatment of blind-ness, this hybrid approach allows for native signal transductionand amplification mechanisms, and additionally has lower theore-tical chance of eliciting a long-lasting autoimmune response to for-eign protein. Given the central importance of glutamatergicsignaling in the retina (Brandstatter, 2002; Thoreson &Witkovsky, 1999), development of two component photoswitchesfor treatment of blindness has, to date, centered on glutamatereceptors and synthetic photoswitch ligands.

LiGluR is a recombinant ionotropic glutamate receptor (analo-gous to SPARK) with an attachment site for a synthetic ligand,MAG (maleimide-azobenzene-glutamate, Fig. 7) (Volgraf et al.,2006). When transfected into HEK293 cells, LiGluR had minimaleffect by itself. Similarly, application of MAG alone had no effecton untransfected 293 cells. However after conjugation of MAG inLiGluR-transfected cells, strong cationic currents were induced incells with light stimuli. Similar results were obtained when hip-pocampal neurons were transfected in vitro with LiGluR6 and con-jugated with MAG (Szobota et al., 2007). Even very high frequencylight (>80 Hz) resulted in matching action potentials.

To test the utility of the LiGluR/MAG two component system inthe treatment of blindness, Caporale et al. (2011) cloned the LiGluRreceptor into an adeno-associated viral vector (AAV) under controlof a synapsin promoter (driving expression in retinal ganglion cellsfollowing intravitreal injection), and applied virus to eyes of rd1/rd1 mice. Retinas harvested from these animals demonstratedlight-dependent firing in response to 380 nm light but not

500 nm light. Following intravitreal injection, MAG restored briskand complete pupillary light responses to gnat1�/�;cnga3�/�;op-n4

�/�mice (mice lacking rod transducing and cone cyclic-gated

nucleotide channel, and hence blind to light, as well as lacking mel-anopsin and non-visual signaling). Electrophysiologic visualevoked potentials (VEP) were restored to rd1/rd1 animals afterLiGluR viral infection and MAG at levels approximately 50% ofwild-type. Additionally, light avoidance, as measured in a water-maze paradigm, was restored in rd1/rd1 animals following treat-ment with LiGluR and MAG.

6. Relative advantages and disadvantages of photoswitchactivation of retinal ganglion cells as a treatment for blindness

Taken together, results to date demonstrate that both one- andtwo-component photoswitch methods are capable of conferringlight-dependent firing activity on retinal ganglion cells in micewith near complete outer retinal degeneration. This activityappears sufficient to restore light-dependent physiologic activitysuch as pupillary light response, to mice from outer retinal degen-eration in vivo. Application of these techniques also restores behav-ioral light sensitivity, including negative phototaxis and light-dependent associative learning. It is important to note that, to date,no study has demonstrated restoration of form vision (as manifestby optokinetic reflex, for instance, or associative learning based ona true visual stimulus).

Functionally, photoswitch techniques have the same end-goalas opto-electronic prostheses (Ahuja et al., 2011; Barry, Dagnelie,& Argus, 2012; Humayun et al., 2012; Mandel et al., 2013; Rizzoet al., 2011; Stingl et al., 2013; Zrenner, 2013), namely convertinglight stimuli directly into retinal ganglion cell stimulation. The suc-cess of these implants in restoring some form vision bodes well for

Fig. 7. Structure of maleimide-azobenzene-glutamate (MAG, left) and proposed mechanism of action on ionotropic glutamate receptors. From: Volgraf et al. (2006) withpermission.

R.N. Van Gelder / Vision Research 111 (2015) 134–141 139

the potential of small molecule photoswitches to accomplish thesame. Both techniques have complementary strengths and weak-nesses. Opto-electronic devices can incorporate amplification cir-cuitry to boost signal in dim light, while photoswitch compoundswould require brighter light for the same effect. However, photo-switch compounds have the advantage of potentially conferringlight sensitivity on every retinal ganglion cell, without concernfor cross-talk induced by external electrical fields. Additionally,administration of photoswitch compounds and viral vectors willlikely occur by simple intravitreal injection, as opposed to the sig-nificant epiretinal or subretinal surgery required for implantationof an optoelectric prosthesis.

Gene therapeutic approaches offer another means for confer-ring light sensitivity on retinal ganglion cells, bipolar cells, ornon-degenerated photoreceptors (Busskamp et al., 2010, 2012;Ivanova, Hwang, et al., 2010; Ivanova, Roberts, et al., 2010;Ivanova & Pan, 2009; Tomita et al., 2007). Compared to one-com-ponent photoswitch these approaches allow more specific target-ing to subsets of remaining cells in the retina. For instance,targeting to cone pedicles may allow substantial retinal processingof signal and reconstitution of faithful action potential trains tovisual stimuli (Busskamp et al., 2010). Gene therapy approachesare also potentially permanent. However, to date channelopsinand halopsin techniques have required more light intensity thanis required for photochemical approaches. The permanence of genetherapy approaches may also be a liability as there is likely notopportunity for ‘upgrading’ to an improved channel. There are stilltheoretical concerns as well over the possibility of long-term uvei-tis resulting from high level expression of foreign channel proteinsin the retina, although this has not been observed in animal studiesto date.

Relative to each other, one-component and two-componentsystems are complementary in strengths and weaknesses. Theone component system is simpler, with lower potential for toxicitygiven the single component. However, targeting is limited by dis-tribution of the target of the photoswitched compound, whichmay encompass many cell types. Conversely, the two componentsystem allows precise targeting of the photoswitch, but requiresgene therapy in addition to photoswitch compound. From a practi-cal perspective, regulatory approval of a single component photo-switch is likely to be more expeditious than a two-componentsystem.

7. Frontiers in photochemical restoration of blindness

For one-component systems, a major challenge to translationfor human use is the solubility of photo-switch compounds. Futuredevelopment will be targeted at improving the solubility of thesecompounds, perhaps by using uncharged moieties rather thanthe charged, quaternary ammonium. Ultimately, a sustainedrelease delivery device, similar to a Retisert might be required tomaintain a steady and sufficient supply of photoswitch to the reti-na (Jaffe et al., 2000). Delivery of one-component compounds couldalso be improved by utilizing large cellular pores, an approachwhich has been used successfully to load the QAQ photoswitch (-containing two, charged quaternary ammonia) (Mourot et al.,2012).

A range of additional compounds, targeting other pathwaysincluding GABAergic and TRP channel pathways, may be renderedphotoswitchable as well (Kienzler et al., 2013; Reiter et al., 2013;Schonberger & Trauner, 2014; Stein et al., 2012, 2013). Some ofthese may have high specificity for certain visual processing path-ways. An ideal target might be an mGluR6-specific photoswitch,given its specificity for specific bipolar pathways (Snellman et al.,2008).

For two-component system, MAG’s activation at 380 nm is tooshort to be useful in a phakic human eye (van de Kraats & vanNorren, 2007). Red shifted MAG has now been synthesized andshown to be efficacious (Kienzler et al., 2013), which will alleviatethis shortcoming. Improvements in viral vector targeting throughdirected selection will help improve cell-type specificity followingintravitreal injection (Dalkara et al., 2013). And, as was the case forthe one-component switches, additional channels may be targetedfor two component function, potentially including metabotropicreceptors (Schonberger & Trauner, 2014) which would allow signalamplification, which in turn could improve sensitivity.

For all vision restoration techniques, one major hurdle is deter-mining the extent to which a non-native action potential series canbe interpreted by the vision processing centers of the brain. Thenormal retina performs a great deal of signal processing, includingdirectional motion detection, color surround processing, and wide-spread adaptation to lighting conditions, which are not intrinsicfeatures of ‘reanimated’ (and functionally reprogrammed) retinalganglion cells. While techniques that target upstream cells suchas bipolar cells may result in more native visual processing and

140 R.N. Van Gelder / Vision Research 111 (2015) 134–141

ultimately action potential firing, it still remains to be determinedif these signal sequences can be correctly interpreted as dynamicvisual scenes. It is conceivable that first-generation use of photo-switch compounds will also include a ‘visual processing unit’ thatwould interpret a visual scene and stimulate photoswitch-activat-ed retinal ganglion cells with particular light stimuli designed tomore faithfully reproduce normal retinal ganglion cell output. Inmaking such a processing unit user-tunable, it may be possibleto achieve better visual results. As an example, in primates the reti-nal ganglion cells subserving the fovea, while having very highresolution due to their one-to-one mapping the cones, are radiallydisplaced from the fovea proper (Curcio & Allen, 1990). As such, inresponse to a visual scene presented normally to the retina, directstimulation of retinal ganglion cells would likely mis-map perifo-veal stimulation to the center of vision, while producing a relativescotoma for central vision (due to the lack of retinal ganglion cellsin the fovea proper). This could be corrected with a device thatconverted a digital video image of the outside world to a raster-s-can of appropriate wavelength, which would displace the truefoveal image to be projected on the appropriate retinal ganglioncells. Similarly, the nature of color vision following photoswitch-mediated retinal reanimation is not apparent (as all color-oppo-nent ganglion cells would be stimulated equivalently), but againany detriment to vision might be mitigated by an altered retinalstimulation pattern to mimic native retinal firing. This challengewill be an active area for research (for all modes of vision restora-tion) for years to come.

8. Conclusion

In a relatively brief period of time since the first proof-of-con-cept that photoswitched channel blockers could induce light-de-pendent neuronal firing activity, the field has progressed todemonstration of partial reversal of blindness in animal modelsusing both one- and two-component systems. This approach iscomplementary to other approaches in vision restoration, butpotentially offers very high resolution (conceivably every ganglioncell), as well as (for the one-component system) the potential forsequential improvement in agents for a single patient. Significanthurdles with respect to demonstration of safety, drug delivery,and optimization of signaling to the visual pathways of the brainremain before this system can be effectively applied to humanblindness.

References

Ahuja, A. K., Dorn, J. D., Caspi, A., McMahon, M. J., Dagnelie, G., Dacruz, L., et al.(2011). Blind subjects implanted with the Argus II retinal prosthesis are able toimprove performance in a spatial-motor task. British Journal of Ophthalmology,95, 539–543.

Banghart, M. R., Mourot, A., Fortin, D. L., Yao, J. Z., Kramer, R. H., & Trauner, D. (2009).Photochromic blockers of voltage-gated potassium channels. AngewandteChemie, 48, 9097–9101.

Banghart, M. R., Volgraf, M., & Trauner, D. (2006). Engineering light-gated ionchannels. Biochemistry, 45, 15129–15141.

Barry, M. P., Dagnelie, G., & Argus, I. I. S. G. (2012). Use of the Argus II retinalprosthesis to improve visual guidance of fine hand movements. InvestigativeOphthalmology & Visual Science, 53, 5095–5101.

Brandstatter, J. H. (2002). Glutamate receptors in the retina: The molecularsubstrate for visual signal processing. Current Eye Research, 25, 327–331.

Brehmer, L., Kaminorz, Y., Dietel, R., Grasnick, G., & Herkner, G. (1997). Studies onpyroelectric response of polymers modified with azobenzene moieties. EXS, 80,155–165.

Busskamp, V., Duebel, J., Balya, D., Fradot, M., Viney, T. J., Siegert, S., et al. (2010).Genetic reactivation of cone photoreceptors restores visual responses inretinitis pigmentosa. Science, 329, 413–417.

Busskamp, V., Picaud, S., Sahel, J. A., & Roska, B. (2012). Optogenetic therapy forretinitis pigmentosa. Gene Therapy, 19, 169–175.

Caporale, N., Kolstad, K. D., Lee, T., Tochitsky, I., Dalkara, D., Trauner, D., et al. (2011).LiGluR restores visual responses in rodent models of inherited blindness.Molecular Therapy, 19, 1212–1219.

Chambers, J. J., Banghart, M. R., Trauner, D., & Kramer, R. H. (2006). Light-induceddepolarization of neurons using a modified Shaker K(+) channel and a molecularphotoswitch. Journal of Neurophysiology, 96, 2792–2796.

Choi, S. W., Kawauchi, S., Ha, N. Y., & Takezoe, H. (2007). Photoinduced chirality inazobenzene-containing polymer systems. Physical Chemistry Chemical Physics:PCCP, 9, 3671–3681.

Curcio, C. A., & Allen, K. A. (1990). Topography of ganglion cells in human retina.Journal of Comparative Neurology, 300, 5–25.

Dalkara, D., Byrne, L. C., Klimczak, R. R., Visel, M., Yin, L., Merigan, W. H., et al.(2013). In vivo-directed evolution of a new adeno-associated virus fortherapeutic outer retinal gene delivery from the vitreous. Science TranslationalMedicine, 5, 189ra176.

Fortin, D. L., Banghart, M. R., Dunn, T. W., Borges, K., Wagenaar, D. A., Gaudry, Q.,et al. (2008). Photochemical control of endogenous ion channels and cellularexcitability. Nature Methods.

Fortin, D. L., Dunn, T. W., Fedorchak, A., Allen, D., Montpetit, R., Banghart, M. R., et al.(2011). Optogenetic photochemical control of designer K+ channels inmammalian neurons. Journal of Neurophysiology, 106, 488–496.

George, E. P., & Johnson, E. A. (1961). Solutions of the Hodgkin–Huxley equations forsquid axon treated with tetraethylammonium and in potassium-rich media. TheAustralian Journal of Experimental Biology and Medical Science, 39, 275–293.

Humayun, M. S., Dorn, J. D., da Cruz, L., Dagnelie, G., Sahel, J. A., Stanga, P. E., et al.(2012). Interim results from the international trial of second sight’s visualprosthesis. Ophthalmology, 119, 779–788.

Ivanova, E., Hwang, G. S., Pan, Z. H., & Troilo, D. (2010). Evaluation of AAV-mediatedexpression of Chop2-GFP in the marmoset retina. Investigative Ophthalmology &Visual Science, 51, 5288–5296.

Ivanova, E., & Pan, Z. H. (2009). Evaluation of the adeno-associated virus mediatedlong-term expression of channelrhodopsin-2 in the mouse retina. MolecularVision, 15, 1680–1689.

Ivanova, E., Roberts, R., Bissig, D., Pan, Z. H., & Berkowitz, B. A. (2010). Retinalchannelrhodopsin-2-mediated activity in vivo evaluated with manganese-enhanced magnetic resonance imaging. Molecular Vision, 16, 1059–1067.

Jaffe, G. J., Ben-Nun, J., Guo, H., Dunn, J. P., & Ashton, P. (2000). Fluocinoloneacetonide sustained drug delivery device to treat severe uveitis. Ophthalmology,107, 2024–2033.

Johnson, J., Wu, V., Donovan, M., Majumdar, S., Renteria, R. C., Porco, T., et al. (2010).Melanopsin-dependent light avoidance in neonatal mice. Proceedings of theNational Academy of Sciences of the United States of America, 107, 17374–17378.

Kienzler, M. A., Reiner, A., Trautman, E., Yoo, S., Trauner, D., & Isacoff, E. Y. (2013). Ared-shifted, fast-relaxing azobenzene photoswitch for visible light control of anionotropic glutamate receptor. Journal of the American Chemical Society, 135,17683–17686.

Lagali, P. S., Balya, D., Awatramani, G. B., Munch, T. A., Kim, D. S., Busskamp, V., et al.(2008). Light-activated channels targeted to ON bipolar cells restore visualfunction in retinal degeneration. Nature Neuroscience, 11, 667–675.

Lamba, D. A., Gust, J., & Reh, T. A. (2009). Transplantation of human embryonic stemcell-derived photoreceptors restores some visual function in Crx-deficient mice.Cell Stem Cell, 4, 73–79.

Mandel, Y., Goetz, G., Lavinsky, D., Huie, P., Mathieson, K., Wang, L., et al. (2013).Cortical responses elicited by photovoltaic subretinal prostheses exhibitsimilarities to visually evoked potentials. Nature Communications, 4, 1980.

Marc, R. E., & Jones, B. W. (2003). Retinal remodeling in inherited photoreceptordegenerations. Molecular Neurobiology, 28, 139–147.

Matharu, A. S., Jeeva, S., & Ramanujam, P. S. (2007). Liquid crystals for holographicoptical data storage. Chemical Society Reviews, 36, 1868–1880.

Mourot, A., Fehrentz, T., Le Feuvre, Y., Smith, C. M., Herold, C., Dalkara, D., et al.(2012). Rapid optical control of nociception with an ion-channel photoswitch.Nature Methods, 9, 396–402.

Mourot, A., Kienzler, M. A., Banghart, M. R., Fehrentz, T., Huber, F. M., Stein, M., et al.(2011). Tuning photochromic ion channel blockers. ACS Chemical Neuroscience,2, 536–543.

Mourot, A., Tochitsky, I., & Kramer, R. H. (2013). Light at the end of the channel:Optical manipulation of intrinsic neuronal excitability with chemicalphotoswitches. Frontiers in Molecular Neuroscience, 6, 5.

Nakatsuji, S. (2004). Recent progress toward the exploitation of organic radicalcompounds with photo-responsive magnetic properties. Chemical SocietyReviews, 33, 348–353.

Oliveira, O. N., Jr., Dos Santos, D. S., Jr., Balogh, D. T., Zucolotto, V., & Mendonca, C. R.(2005). Optical storage and surface-relief gratings in azobenzene-containingnanostructured films. Advances in Colloid and Interface Science, 116, 179–192.

Panda, S., Provencio, I., Tu, D., Pires, S., Rollag, M., Castrucci, A., et al. (2003).Melanopsin is required for non-image-forming photic responses in blind mice.Science, 301, 525–527.

Polosukhina, A., Litt, J., Tochitsky, I., Nemargut, J., Sychev, Y., De Kouchkovsky, I.,et al. (2012). Photochemical restoration of visual responses in blind mice.Neuron, 75, 271–282.

Reiter, A., Skerra, A., Trauner, D., & Schiefner, A. (2013). A photoswitchableneurotransmitter analogue bound to its receptor. Biochemistry, 52, 8972–8974.

Rizzo, J. F., 3rd, Shire, D. B., Kelly, S. K., Troyk, P., Gingerich, M., McKee, B., et al.(2011). Overview of the Boston retinal prosthesis: Challenges and opportunitiesto restore useful vision to the blind. Conference Proceedings: Annual InternationalConference of the IEEE Engineering in Medicine and Biology Society IEEEEngineering in Medicine and Biology Society Conference, 7492–7495.

Schonberger, M., & Trauner, D. (2014). A photochromic agonist for mu-opioidreceptors. Angewandte Chemie, 53, 3264–3267.

R.N. Van Gelder / Vision Research 111 (2015) 134–141 141

Snellman, J., Kaur, T., Shen, Y., & Nawy, S. (2008). Regulation of ON bipolar cellactivity. Progress in Retinal and Eye Research, 27, 450–463.

Spencer, P. S., & Schaumburg, H. H. (1974a). A review of acrylamideneurotoxicity. Part I. Properties, uses and human exposure. The CanadianJournal of Neurological Sciences Le journal Canadien des Sciences Neurologiques, 1,143–150.

Spencer, P. S., & Schaumburg, H. H. (1974b). A review of acrylamide neurotoxicity.Part II. Experimental animal neurotoxicity and pathologic mechanisms. TheCanadian Journal of Neurological Sciences Le Journal Canadien des SciencesNeurologiques, 1, 152–169.

Spencer, P. S., & Schaumburg, H. H. (1975). Nervous system degeneration producedby acrylamide monomer. Environmental Health Perspectives, 11, 129–133.

Stein, M., Breit, A., Fehrentz, T., Gudermann, T., & Trauner, D. (2013). Optical controlof TRPV1 channels. Angewandte Chemie, 52, 9845–9848.

Stein, M., Middendorp, S. J., Carta, V., Pejo, E., Raines, D. E., Forman, S. A., et al.(2012). Azo-propofols: Photochromic potentiators of GABA(A) receptors.Angewandte Chemie, 51, 10500–10504.

Stingl, K., Bartz-Schmidt, K. U., Gekeler, F., Kusnyerik, A., Sachs, H., & Zrenner, E.(2013). Functional outcome in subretinal electronic implants dependson foveal eccentricity. Investigative Ophthalmology & Visual Science, 54,7658–7665.

Szobota, S., Gorostiza, P., Del Bene, F., Wyart, C., Fortin, D. L., Kolstad, K. D., et al.(2007). Remote control of neuronal activity with a light-gated glutamatereceptor. Neuron, 54, 535–545.

Thoreson, W. B., & Witkovsky, P. (1999). Glutamate receptors and circuits in thevertebrate retina. Progress in Retinal and Eye Research, 18, 765–810.

Tochitsky, I., Polosukhina, A., Degtyar, V. E., Gallerani, N., Smith, C. M., Friedman, A.,et al. (2014). Restoring visual function to blind mice with a photoswitch thatexploits electrophysiological remodeling of retinal ganglion cells. Neuron, 81,800–813.

Tomita, H., Sugano, E., Yawo, H., Ishizuka, T., Isago, H., Narikawa, S., et al. (2007).Restoration of visual response in aged dystrophic RCS rats using AAV-mediatedchannelopsin-2 gene transfer. Investigative Ophthalmology & Visual Science, 48,3821–3826.

van de Kraats, J., & van Norren, D. (2007). Optical density of the aging human ocularmedia in the visible and the UV. Journal of the Optical Society America A OpticsImage Science and Vision, 24, 1842–1857.

Volgraf, M., Gorostiza, P., Numano, R., Kramer, R. H., Isacoff, E. Y., & Trauner, D.(2006). Allosteric control of an ionotropic glutamate receptor with an opticalswitch. Nature Chemical Biology, 2, 47–52.

Zrenner, E. (2013). Fighting blindness with microelectronics. Science TranslationalMedicine, 5, 210ps216.