Hybrid Optogenetic and Electrical Stimulation of Retinal Ganglion Cells for Artificial Vision

doi:10.21203/rs.3.rs-3938602/v1

Download PDF

Article

Hybrid Optogenetic and Electrical Stimulation of Retinal Ganglion Cells for Artificial Vision

https://doi.org/10.21203/rs.3.rs-3938602/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Millions of adults worldwide experience severe visual impairment due to photoreceptor loss from retinal diseases such as retinitis pigmentosa and macular degeneration. Retinal prostheses that provide artificial vision by stimulating the surviving retinal ganglion cells (RGCs) have emerged as a promising therapy. However, all clinically approved retinal prostheses that use electrical stimulation face the issue of electrical spread. As such, no device has been able to provide vision restoration beyond the threshold of legal blindness. Optogenetic approaches provide greater spatial precision, however, they have poor temporal fidelity compared to electrical stimulation. Questions remain over the safety of the high intensity light required. Here, we developed an opto-electrical hybrid approach that reduced optogenetic activation thresholds and increased the high frequency sensitivity of opsin expressing RGCs. Our findings suggest that incorporating hybrid stimulation in future retinal prostheses could dramatically improve high resolution vision.

Physical sciences/Engineering/Biomedical engineering

Biological sciences/Neuroscience

Retinal degeneration is the leading cause of blindness in industrialised countries¹. Retinal diseases that result in photoreceptor degeneration are some of the most prevalent retinopathies. Among them, age-related macular degeneration accounts for approximately 9% of blindness globally and the inherited disease, retinitis pigmentosa, has a prevalence of 1:4000^2,3.

Retinal prostheses attempt to overcome photoreceptor loss via exogenous stimulation of surviving higher-order neurons, e.g. retinal ganglion cells (RGCs) or bipolar cells. This stimulation is typically conducted through charge injection from surgically implanted electrode arrays. These prostheses have restored the ability of patients to perceive light, objects and movement. Implant recipients report improvements in their quality of life, ability to navigate their surroundings, social interactions, and ability to perform daily activities⁴. A significant limitation with all devices that have received US and European regulatory approval to date is their low spatial resolution, caused by issues such as electrical current spread. As a result, the highest visual acuity reported from a retinal prosthesis is only 20/460 via the PRIMA Implant (Pixium Vision, France)⁵. This is well below the threshold for legal blindness, which is 20/200. Furthermore, the implantation of electronic retinal prostheses requires complex and expensive surgeries and comes with potential complications, such as inflammation, making them suitable only for select groups of patients. Therefore, despite over ten devices reaching clinical trials, electronic retinal prostheses have not replicated the commercial or clinical successes of other neural prostheses, e.g. cochlear implants and deep brain stimulation⁶. We aimed to explore alternative strategies to overcome the challenges of restoring vision in cases of retinal degeneration.

In recent years, optogenetic based therapies have garnered significant attention. With optogenetics, light sensitive proteins (opsins) such as channelrhodopsin (ChR) are expressed on neurons, permitting neural circuit activation or silencing through the application of visible light. The exceptional level of spatial precision that comes with optogenetics has led to multiple therapeutic applications⁷. Given the unique optical accessibility of the retina, optogenetics has the potential to be implemented with less surgical invasiveness than electrical neural prostheses and thus is an attractive therapy for vision restoration.

There have been several studies exploring the restorative capacity of retinal optogenetic stimulation. In rd1 mice that are blind due to an inherited rod photoreceptor dystrophy, optogenetic activation of RGCs restored visually guided locomotion⁸. In nonhuman primate models, optogenetic stimulation of RGCs could drive signal transmission all the way to the primary visual cortex. Following optogenetic transduction, the pathway was amenable to optical modulation for over a year^9,10. In macaques, the spatial resolution of optogenetic RGC stimulation translated to a visual acuity of 20/249¹¹. In a promising clinical study, a 58-year-old patient who underwent adeno-associated virus (AAV) transduction of RGCs showed remarkable vision improvement. Prior to optogenetic therapy, the patient was only able to detect the presence of light. With optogenetic therapy, the patient gained the ability to identify and reach out to objects (~ 10 x 10 visual degrees object size) and execute visuomotor reaching tasks¹². With careful consideration of the AAV serotype, opsin variant and the transduction strategy, optogenetics has the potential to restore vision to an acuity of 20/72, according to a modelling study¹³.

However, optogenetic visual prostheses, suffer from two primary challenges. First, the light intensity required to activate opsins is significantly brighter than ambient light. Clinical investigations observing the effects of long-term exposure to intense light stimulation in neurons is limited and hence, there are unsolved concerns regarding long-term patient safety. The current safety threshold set by European parliamentary guidelines for blue light stimulation of the retina is 0.322 mJ / cm^− 2 / s¹⁴, with blue light being the driver to many channelrhodopsin variants. As many studies driving neural activity use optical intensities that exceed this prescribed retinal safety limit, it is important to seek alternative methods to reduce the intensity of light required for effective neuromodulation. Second, the temporal fidelity of responses to optogenetic stimulation is poor compared to electrical stimulation. Specifically, the amplitude and probability of responses are reduced during repeated stimulation¹⁵, potentially causing image fading in patients. Using ChR2 variants, RGCs can only reliably respond to optogenetic stimulation frequencies up to 20 Hz^16,17. This is far below the response rate of some RGCs which discharge at ~ 40 Hz in response to luminance changes¹⁸ or > 100 Hz in response to motion stimuli¹⁹. A key reason for this limitation is the rapid desensitization and relatively long inactivation time constant for most opsins, typically tens of milliseconds for most ChR2 variants^20,21. Ultrafast opsins with shorter inactivation time constants have been engineered to provide higher spike probabilities at stimulation rates up to 500 Hz. Unfortunately, they also require higher intensity light with the risk of phototoxic damage^22,23.

We hypothesized that a combination of electrical and optogenetic stimulation would improve the safety of optical stimulation by reducing optical power requirements to drive retinal responses. Additionally, this technique will retain the high level of spatial precision in neural activation that is characteristic of optogenetics. As a new frontier of optogenetic based therapies emerges, we envisage a technique such as an opto-electrical hybrid stimulation could contribute to improved delivery of artificial vision and serve as a pioneer for the intersection between gene therapy and bionics. The efficacy of hybrid stimulation has already been demonstrated in both the cochlea^24–27 and sciatic nerve²⁸. In the cochlea, hybrid stimulation reduced the electrical stimulation current required, thus significantly reducing the spread of activation compared to electrical-only stimulation²⁷. In addition, hybrid stimulation increased the probability that a cell would respond to repetitive high-frequency stimulation, improving temporal fidelity^26,27.

Here, for the first time within the visual system, we explore the potential of hybrid electrical and optogenetic stimulation as a novel approach to restore vision. Using a transgenic mouse model with ChR2-H134R expression in RGCs, we explanted the retinae to perform whole-cell patch clamp recordings of RGCs, and assessed their activation thresholds and temporal responses to repetitive high-frequency stimulation using hybrid, electrical or optogenetic stimulation. Furthermore, to study how RGCs in a degenerative retina, respond to hybrid stimulation at a population level, we recorded calcium transients in RGCs transduced with ChrimsonR and GCaMP7s in Royal College of Surgeon (RCS-p+) rats, an animal model of retinitis pigmentosa.

2.1 A transgenic mouse model to explore hybrid stimulation in the retina

To examine the feasibility of hybrid stimulation as a novel method for modulating RGC activity, we crossed mice expressing parvalbumin-Cre with mice expressing Cre-dependent ChR2(H134R)-EYFP to generate animals with ChR2(H134R) in parvalbumin expressing RGCs. We performed a histological survey of retinae from our ChR2(H134R) mouse line to determine the extent of ChR2(H134R) expression in the retina.

The retina is comprised of five major neuronal subtypes, each organised within its laminar structure. Transmission of photic signals originates in the most superficial layers of the retina, the photoreceptors, before being di-synaptically relayed to the deepest layer, the ganglion cell layer (GCL), via bipolar cells in the outer nuclear layer. The cells within the GCL serve as the final outpost of visual information in the eye before it is transmitted to the brain via the optic nerve²⁹. These layers can be readily delineated with a Hoechst counterstain. Analysis of horizontal cryo-sectioned tissue revealed that expression of ChR2(H134R) was confined to the GCL (Fig. 1a-c) and within parvalbumin expressing RGCs (Fig. 1d-g). The proportion of ChR2(H134R) RGCs was estimated to be ~ 30–40% based on previous studies which quantified parvalbumin expression within the GCL^30,31.

RGCs are diverse and within the mouse over 40 subtypes have been defined based on their genetic, functional and morphological identities^32,33. More broadly, RGCs in the mouse can be segregated into three distinct populations based on their responses to light. The three types are 1) ON-cells that respond to brightness increments, 2) OFF-cells that respond to brightness decrements, and 3) ON-OFF cells that are responsive to both brightness increments and decrements³⁴. We back-filled cells with a fluorescent tracer during patch clamp recordings which permitted the reconstruction of the cells’ dendritic arbour (Fig. 1h, i). The depth within the inner plexiform layer (IPL) where the dendritic arbour stratifies was used to phenotype each cell into one of the three main classes. Within our experiments we found all three types expressing ChR2 and recorded from all three populations (Fig. 1j-l).

2.2 Hybrid stimulation reduces electrical and optical thresholds

The spread of activation due to electrically evoked activity depends on the amplitude of stimulation. Previous studies have demonstrated that by coupling electrical stimulation with optical stimulation, significantly less electrical current is required to activate the sciatic nerve or neurons in the cochlea^26,27,35. We wanted to determine if this hybrid approach could be translated to RGCs of the retina. In isolated wholemount retinae, we recorded responses to hybrid stimulation in single RGCs in a whole-cell current clamp configuration. RGCs were electrically stimulated intracellularly during patch recordings whilst 488 nm blue light pulses were delivered via a laser-coupled optical fibre, focused on the cell via the microscope objective. To ensure recordings were a direct result of optogenetic stimulation of RGCs, di-synaptic photoreceptor inputs to RGCs were sequestered through pharmacological synaptic blockade. For each cell, we first identified the thresholds for electrical and optical stimulation. We defined threshold as the stimulation amplitude required to elicit a 50% response probability during the 2 Hz pulse train (5 s train duration). The minimum stimulation level that would evoke 100% reliability was defined as supra-threshold (Fig. 2a).

Once supra-threshold levels were defined, 10 hybrid stimulation pulses were delivered at 2 Hz. Eleven linearly scaled steps between zero and supra-thresholds amplitudes for both electrical and optical stimulation were surveyed. Each optical step was paired with all 11 electrical steps in random order, creating 121 combinations (Fig. 2b, c). Optical pulse duration was 10 ms, electrical pulse duration was 3 ms. It has been reported that the timing of opto-electrical coupling impacts activation thresholds. Delaying the delivery of electrical pulses relative to the onset of the optical pulses reduces activation thresholds compared to simultaneous presentation of the stimuli^26,35. To explore this, hybrid stimulation was delivered with the electrical pulse being delivered with delays of 0 ms or 10 ms. With an optical pulse duration of 10 ms; 0 ms delay signifies simultaneous delivery and a 10 ms delay signifies sequential delivery. (Fig. 2b, d). The anatomical characteristics of our recorded RGC population are summarised in Table S1. No significant differences in response to stimulation were exhibited between the RGC sub-types (Table S2-4). Therefore, data from all RGCs were pooled and analysed collectively.

We quantified the efficacy of hybrid stimulation to reduce activation thresholds by plotting optical and electrical thresholds as functions of the intensity of electrical and optical stimulation, respectively. Threshold levels were normalized to the thresholds obtained with electrical or optical-only stimulation. Incorporating electrical stimulation could significantly attenuate optical thresholds compared to optical stimulation alone (Fig. 2e, (Two-way ANOVA, (3, 59) = 15, p < 0.0001)). In both the simultaneous and sequential conditions, the most apparent effect was seen when electrical stimulation was delivered at 50–70% intensity relative to threshold, where the optical threshold was approximately halved compared to optical stimulation alone (post hoc Tukey test, p < 0.0001). Delaying the electrical stimulus by 10 ms provided no further significant reduction in optical threshold (Two-way ANOVA, F (1,59) = 2.6, p = 0.1136).

2.3 The temporal bandwidth of hybrid stimulation surpasses electrical and optical only stimulation

We explored whether hybrid stimulation could enhance the response reliability of RGCs to higher stimulation frequency. We first determined the response probability thresholds of the RGCs by quantifying the number of action potentials evoked from electrical or optical stimulation at 2 Hz across a range of amplitudes (Fig. 3a, b). We coupled the near threshold optical stimulation with electrical stimulation at threshold amplitude (i.e. 50% response probability) with the rationale that we could leverage the high temporal fidelity that is characteristic of electrical stimulation, while reducing the effects of activation spread caused by supra-threshold electrical levels. Since light can be easily focused by the lens of the eye, our strategy for hybrid stimulation was to use light that was just above-threshold, combined with electrical stimulation at threshold intensities. In this way, we aimed to safely achieve a localized region of RGCs that are primed to be activated by stimulation pulses of relatively small current amplitudes, which may be delivered broadly across the retina. Light pulses were delivered with a 1 ms pulse width to determine if increases in temporal bandwidth could be achieved using lower power optical pulses (power is proportional to pulse length at a given light intensity).

We quantified how a RGC responded to supra-threshold optical-only stimulation, threshold electrical-only stimulation, or hybrid stimulation based on the number of action potentials evoked during a stimulus pulse train. Stimulation was delivered at 2, 20, 50 and 200 Hz for 3 s (Fig. 3c). The responses recorded under the three stimulation modalities at the four frequencies constituted one trial and each cell underwent a minimum of three trials (Fig. 3d, e).

For the 29 RGCs surveyed, response reliability under hybrid conditions was significantly higher throughout the 3 s stimulation pulse train compared to optical or electrical-only modalities. The anatomical characteristics of our population are summarised in Table S1. No significant differences in response to stimulation were exhibited between the RGC sub-types (Table S2-4). Therefore, data from all cell types were pooled and analysed collectively. From the onset of the stimulus, RGCs responded prolifically before the response rate decayed to a steady state for the duration of the stimulation. The decay in response is evident across all three stimulation modalities with the exception of 2 Hz stimulation. However, hybrid stimulation was able to sustain a significantly higher response rate at stimulation frequencies of 2–50 Hz (Fig. 3f, g, Table S5, S6). Our results suggest that hybrid stimulation is a viable method to enhance the temporal bandwidth of an optogenetics based approach for neural stimulation. Critically, from stimulus onset, hybrid stimulation achieved near-100% response rates at pulse trains up to 50 Hz, a significant enhancement from optical-only or electrical-only stimulation (Fig. 3f).

2.4 Hybrid Stimulation can activate RGCs in a degenerated retina with reduced optical power

We next determined if the advantages of employing hybrid stimulation that we observed in ChR2(H134R) RGCs are translatable to the degenerated retina. We adopted a Royal College of Surgeons (RCS-p+) rat model. Their genetic defect of progressive photoreceptor loss³⁶ is a faithful representation of the hereditary retinopathy retinitis pigmentosa³⁷. Previous studies in RCS rats have demonstrated that despite complete degeneration of photoreceptors, some of the RGC population is spared and remains functional. However, the physiology is altered with reports of RGCs requiring higher electrical depolarization thresholds for action potential propagation and reduced temporal bandwidth. At the late-stage of the disease, RGCs can become unresponsive to electrical stimulation due to the absence of inward sodium currents^38,39.

Therefore, to assess if hybrid stimulation can be implemented as a therapeutic, we needed to determine if hybrid stimulation can drive RGC activity in the degenerated retina at safe power levels and with greater spatial precision than electrical stimulation alone.

To address this aim, we delivered intravitreal injections of AAVs that encode the channelrhodopsin variant ChrimsonR and the genetically encoded calcium indicator GCaMP7s in RCS-p + rats (n = 3) (Fig. 4a and b). We utilized ChrimsonR as this opsin has improved kinetics compared to ChR2-H134R. More crucially the safety and long-term stability of ChrimsonR has been demonstrated in the retina in nonhuman primates and this has translated to its use clinically^9,12. To explore the effects of hybrid stimulation across a population of RGCs, we modified our hybrid paradigm and sequentially delivered electrical current epiretinally via extracellular stimulation of the GCL immediately after broad optogenetic illumination to our electrical stimulation region.

Following effective transduction, in whole isolated retinal flat mounts, we stimulated the retina with 3 levels of optical-only stimulation; 2 levels of electrical-only stimulation, and 4 hybrid electrical-optical combinations. We analysed the area of activation and the number of RGCs that could be activated by hybrid stimulation and compared these responses to electrical and optical-only stimulation modalities. Both electrical and optical stimulation were delivered at supra-threshold intensities. Stimulus evoked responses within RGCs were identified through the presence of calcium transients via confocal microscopy (Fig. 4c).

At the lowest levels of electrical-only and optical-only conditions (i.e. OL0/EL1 and OL1/EL0), only a few cells were activated. Predictably, an increase of intensity increased the number of cells activated (Fig. 4d). Hybrid stimulation with the lowest opto-electrical power (OL1/EL1) was able to activate more RGCs when compared to the supra-threshold amplitudes of the single modality controls (Fig. 4e). We believe that the enhancement in the total number of RGCs that could be activated is due to a reduction in activation threshold. This would be congruent with our findings of lower activation thresholds under hybrid conditions, observed in our whole-cell current-clamp assays (Fig. 2).

For a given area of activation, a comparable number of active cells were observed when using hybrid stimulation at low intensities and single modality stimulation at higher intensities (e.g. OL1/EL1 was comparable to OL2/EL0). This is indicated as a similar normalised percentage of active cells (Fig. 4e) and a similar active area along the diagonals in Fig. 4f. Delivering hybrid stimulation with the largest hybrid intensity tested (OL2/EL2) evoked activity in an area that surpassed the active area under the largest electrical-only (OL0/EL2) or optical-only (OL3/EL0) intensities tested. This is indicated by a percentage above 100% in the OL2/EL2 condition, as 100% is equivalent to the OL3/EL0 condition, and a larger percentage than the OL0/EL2 condition (182% in OL2/EL2 versus 99% in OL0/EL2).

At the turn of the century, restoration of vision via a retinal prosthesis became a viable possibility, but visual acuity remains poor. Optogenetic stimulation presents an attractive solution since the ability to focus light potentially allows a high level of spatial resolution for neural activation. However, it can be challenging to fully exploit these benefits because the poor temporal fidelity of many opsins precludes their use in therapeutic applications. A concerted effort has been made towards developing optogenetic constructs with faster kinetics. However, many of these opsins require high light intensities, with a concomitant negative impact on the welfare of implant recipients⁴⁰. By employing a hybrid opto-electrical strategy, we wanted to determine if such an approach could grant improved levels of spatiotemporal control while driving RGCs with reduced levels of irradiance.

Our hybrid stimulation approach reduced RGC activation thresholds, which is congruent with previous studies in the auditory system. In the mouse, cochlear implants were used to deliver hybrid stimulation to the cochlear spiral ganglion neurons. Using sub-threshold or near-threshold light, significantly less electrical current was required to activate these neurons^26,27. In turn, this resulted in less spread of activation in the cochlea, improving the spectral resolution and independence of stimulating channels^27,41. Perhaps more critically, our hybrid strategy permitted us to photoactivate opsin expressing RGCs with lower levels of irradiance. Electrical supplementation reduced optical thresholds by approximately half and this mechanism could potentially be leveraged with more red-shifted opsins. Such a change would improve the safety of long term light exposure in optogenetic therapies.

When we delivered hybrid stimulation to RGCs, response reliability also improved at higher stimulation frequencies when compared to optical-only and electrical-only methods. This again mirrors findings in the cochlea, in which it was shown that hybrid stimulation increased reliability, fidelity, and temporal precision of responses compared to optogenetic stimulation alone^24,26. Although the peak firing rates were still below the 300 Hz burst firing achieved by RGCs with light stimulation^42,43, the results reveal that hybrid stimulation enhances the reliability of responses to high frequency stimulation compared to optical-only methods. Our approach enhanced temporal fidelity during hybrid delivery at stimulation frequencies up to 50 Hz. We believe the limited enhancement could be attributed to a combination of the biophysical properties of ChR2-H134R and its temporal bandwidth when expressed on RGCs. Although ChR2-H134R has been shown to elicit near 100% response reliability during 40–60 Hz stimulation in cochlear spiral ganglion neurons^24,27, previous studies that focussed on RGC stimulation have shown a reduced response rate when light pulses were delivered at rates beyond 20 Hz^44–47. With the arrival of more robust opsins with faster recovery rates and lower levels of desensitization such as ChRmine or PsCatCh 2.0^8,48, replication of high frequency firing should be more attainable. Computational investigation of RGC control under the opsin ChRmine suggests that temporal control could be elicited at rates up to 280 Hz⁴⁹. If this theoretical temporal bandwidth is translatable to in vivo conditions, it could be potentially further enhanced through a hybrid delivery method. This would be a significant contribution to addressing the technical hurdles of meeting the spatial and temporal precision requirements in retinal prostheses. The scope of this proof-of-concept study only permitted survey of hybrid stimulation at single electrical and optical amplitudes that were delivered at threshold and supra-threshold levels, respectively, and with fixed pulse lengths and constant relative timing. Modest increases in the intensity of the electrical component of hybrid stimulation has been shown to improve the reliability of high frequency responses²⁷. Thus, future explorations of hybrid stimulation should encompass a broader range of electrical and optical amplitudes, pulse lengths and delays with faster and more photosensitive opsins, as well as stimulation frequencies. This would provide deeper insights into the level of spatiotemporal control and threshold attenuation in RGCs that stimulation under hybrid conditions could provide.

Clinically, electrical stimulation remains the dominant strategy to treat retinal degeneration^50–52. In parallel, optogenetic methods have also been developed to restore visual function in animal models and humans^12,53. Encouraged by the clinical evidence for vision restoration via optogenetics, along with recent in vivo demonstrations of improved spatiotemporal control of neurons in the cochlea, we wanted to survey the therapeutic potential of opto-electrical stimulation in a degenerated retina. Our hybrid approach comprised of extracellular electrical stimulation delivered through a single small electrode and broad application of light to evoke activity in ChrimsonR/GCaMP expressing RGCs in the RCS-p + rat.

Principally, the method of electrical stimulation to evoke neural activity is similar between intracellular and extracellular approaches. However, extracellular stimulation may activate a network of circuits that leads to activation of local inhibitory neurons. This might suppress temporal fidelity. The previously reported observation that evoking RGC activity with extracellular electrical stimulation required higher frequencies (~ 256 Hz versus ~ 64 Hz to elicit peak firing rate) and current amplitudes (µA versus pA)⁵⁴, likely indicates the influence of inhibitory circuits evoked from extracellular stimulation. Additionally, intracellular electrical stimulation of RGCs elicits higher spike rates compared to extracellular stimulation (25 spikes/sec versus 10 spikes/sec)⁵⁴. The reduced peak firing rates that can be elicited through extracellular stimulation may also be due to depolarization block, a phenomenon where overstimulation results in a decrease in activity. Typically, axons and not cell bodies are activated through electrical stimulation and this is consistent within the retina^55–57. The axon is vulnerable to depolarisation block and whilst the precise mechanisms are not fully understood, accumulation of extracellular potassium in the peri-axonal space is believed to be a significant contributor to perturbed action potential conduction during high frequency stimulation^58,59. Given the limitations on retinal response reliability with electrical-only stimulation methods at high frequency, there is a need to explore alternative methods to manipulate retinal neuron responses.

Thompson and colleagues showed that hybrid stimulation in the cochlea led to reduced spread of activation²⁷. Our calcium imaging approach did not permit us to make similar conclusions in our retinal study. The stimulation parameters we employed did not provide us with enough resolution to delineate how the boundaries of the area of activation could be scaled with stimulus strength. In clinically approved retinal prostheses, the electrode diameter is significantly larger than the 20 µm single electrode that was utilized in this study. The PRIMA implant which currently provides the highest level of visual restoration utilizes electrodes with 100 µm diameter, while less invasive suprachoroidal implants would utilize electrode diameters of 600 µm⁵¹. We speculate that hybrid stimulation might be particularly beneficial when large electrodes are used as these systems also have the largest electrical spread. Future explorations will factor in electrode diameter and method of delivery.

We observed that the number of RGCs and area of the retina activated by optical-only methods could similarly be activated under hybrid conditions, with reduced optical power when coupled with a low level of electricity. Critically, we were able to do this in a model of retinitis pigmentosa which positions our hybrid stimulation as a potentially viable therapeutic strategy. That is, where a low intensity electrical stimulus is delivered broadly throughout the retina alongside focused light. Optical intensities would be reduced with the supplementation of modest amounts of electrical stimulation. This would reduce the level of irradiance required to activate RGCs. Interestingly, we also identified that optical only stimuli activated cells that did not express ChrimsonR. On average 90%, 73% and 70% of cells detected as having a calcium response were found to express ChrimsonR at optical stimulation levels OL1, OL2 and OL3, respectively (Fig. S1). This could be due to a number of factors including transmission of signals from spared photoreceptors, light sensitive RGCs that contain melanopsin reaching optical threshold⁶⁰ or network activation of neighbouring RGCs. Alternatively, these cells may express low levels of ChrimsonR, so they were not detected during fluorescent imaging.

Although we could selectively activate remnant RGCs, the hybrid stimulation method employed in this study did not permit manipulation of RGCs at the subtype level or separate activation of ON or OFF pathways. The ON and OFF circuits can be disentangled as each pathway can be identified through their differential responses to electrical stimulation^61–64, via tunable nanoparticles⁶⁵, and genetic tools^66,67. More recently, with the development of AAVs with novel synthetic promoters, neuronal populations within the retina can be differentially targeted with a level of selectivity down to RGC subtype⁶⁸. As processing of visual features relies on the synergy of different retinal circuits, future explorations of RGC stimulation to restore visual perception should lean into the electrical and genetic tools that permit specific targeting of RGC subtypes and segregation of the ON and OFF pathways.

Vision restoration through retinal prostheses still faces many hurdles as multiple unresolved questions remain regarding the neural code of the retina. For example, how signals relating to different wavelengths of light are conveyed in RGCs is still poorly understood although some ground-breaking studies have begun to shed light on how colour perception can be delivered to patients with photoreceptor loss^69,70. As current devices are only able to provide achromatic vision, accurate object perception will be even more dependent on high spatiotemporal resolution.

Despite the limitations of our study, the use of an opto-electrical approach provided notable improvements over stimulation modalities that exclusively utilize electricity or optogenetics. Notably, there was improved spatiotemporal control when activating RGCs. We believe the utility of hybrid stimulation transcends retinal applications and could be harnessed in modulating neural activity in other organs. The issue of imprecise activation from electrical stimulation is not exclusive to retinal devices but persists as an issue in cochlear implants, vagus nerve stimulation therapies, management of cardiac arrhythmias, and treatments that utilize transcranial electrical stimulation^40,71–73

Considering implementation, patients could receive intraocular administration of AAVs encoding opsins. To project appropriately accurate light into the retina, GenSight Biologics have developed a device that uses specialised cameras, biomimetic mirrors and computer control. This has been used with success clinically¹². With regards to electrical stimulation of the retina, there are numerous ways this could be achieved^51,74. Surgical approaches for positioning electrodes can be epiretinal (on top of the inner limiting membrane), subretinal (between the pigment epithelium and INL) and suprachoriodal (into the sclera behind the choroid). Non-surgical approaches have also been developed and these include transcorneal and transorbital systems^75,76.

The benefits of reducing the invasiveness of device implantation should not be underestimated as recovery from surgical procedures can come with complications and have an adverse impact on the patient’s health and wellbeing⁵¹. Whilst the approach we took in this study involved broad light delivery and focal electrical stimulation, we want to emphasize that at the level of the cell, the significant finding surrounding hybrid stimulation is that we can reduce the optical and electrical requirements to elicit neural activity. That is, thresholds for photoactivation can be significantly reduced with a modest application of electrical stimulation. In alignment with our study, we suggest that this mechanism could be exploited in the degenerate retina. RGCs could be stimulated with near threshold depolarizing current which could be delivered by corneal or transorbital devices. In parallel, devices such as the GenSight goggles, could provide focal optogenetic stimulation. With reduced optical thresholds, response reliability will be significantly improved, which ultimately should improve the quality of artificial vision a patient receives.

4.1 Animals

All procedures performed in this study were in accordance with the ethical standards of the Animal Care and Ethics Committee of the University of Melbourne (Ethics ID# 21886).

Mice with RGCs expressing ChR2(H134R) were obtained as a breeding derivative of COP4*H134R/EYFP mice (Jax strain 012569:B6;129SGt(ROSA)26Sortm32(CAG- COP4*H134R/EYFP)Hze/J) and Cre-parvalbumin mice (Jax strain 008069:B6;129P2-Pvalbtm1(cre)Arbr). Male and female mice aged 7–11 weeks (n = 21) were used in this study and were sourced from St Vincent’s Hospital Biological Resource Centre, Australia.

For experiments pertaining to investigating hybrid stimulation in reducing the spread of activation, three Royal College of Surgeons (RCS-p+) rats (1 female, 2 male) aged 18–19 weeks were sourced from Ozgene, Australia. These rats were first anaesthetised via intraperitoneal injection of a mixture of ketamine (65 mg•kg^− 1) and xylazine ( 5 mg•kg^− 1). Depth of anaesthesia was confirmed via loss of response to a noxious toe pinch. Following this, a local anaesthetic eye drop (0.5% proxymetacaine) was administered. In each eye, 3 µl of a [1:1] cocktail of AAV2-Syn1-ChrimsonR-tdT (7.29 x 1013 GC•ml^− 1) and AAV2-Syn1-GCamP7s (1.06 x 1013 GC•ml^− 1) was introduced via intravitreal injection, in a method as previously described by Puthussery & Fletcher⁷⁷. In brief, a 32G needle is inserted through the sclera, approximately 1 mm posterior to the limbus, at a 45° angle to avoid damage to the lens. The needle was connected to a 10 µl Hamilton syringe (Model 701N) via silicon tubing to facilitate the injection. Post injection, an antibiotic eye ointment (tobramycin) was applied. Animals were then recovered for a minimum of 3 weeks to accommodate viral transduction before the animals were euthanized and dissection of retinal tissue occurred.

4.2 Retinal Preparation

Animals were anaesthetised via an intraperitoneal injection of a mixture of ketamine (100 mg•kg^− 1) and xylazine (10 mg•kg^− 1). After the depth of anaesthesia was confirmed via lack of response to a noxious toe pinch, the eye was enucleated and bathed in carbogenated Ames’ medium (Assay Matrix) supplemented with 23 mM NaHCO₃ and bubbled with 95% O₂/5% CO₂ at room temperature. Following the enucleation, the animal was humanely killed via an intracardiac injection of pentobarbital.

Post enucleation, an incision was made in the cornea, and the subsequent dissection was undertaken in Ames’ medium. The eyes were hemisected behind the ora serrata followed by lensectomy and vitrectomy. The sclera and retinal pigment epithelium were removed before radial incisions were made to flatten the retina. These retinal wholemounts were placed with ganglion side up in a recording chamber (Warner Instruments) whilst being superfused with the carbogenated Ames’ medium (2–3 ml/min). Visualization of the tissue was undertaken on a fixed-stage, upright microscope (Olympus cat# BX61WI).

4.3 Whole-cell patch clamping

Whole-cell patch clamping was conducted on RGCs expressing ChR2 (H134R) from mouse retina. RGCs with ChR2(H134R) were identified through their co-expression of EYFP, using a 40x water immersion objective (Olympus cat# N2667700) and a GFP filter (Olympus cat# U-MNIBA3). Cells were visualised with a monochrome digital camera (Olympus XM-10) and µManager software⁷⁸. Individual RGC responses were recorded under current clamp conditions, as described previously⁷⁹. Borosilicate micropipettes (8–10 MΩ) were prepared using a CO₂ laser puller (Sutter instruments Cat# P-2000) and filled with 6µL of fluorescent intracellular solution that consisted of (in mM) 125 K-Gluconate, 2 CaCl₂, 2 MgCl, 10 HEPES, 10 EGTA, 2 Mg-ATP, 0.5 Na₂-GTP and 0.125 Alexa Fluor 594. To obtain access to a cell, a small tear was made in the retinal inner limiting membrane with a micropipette (3–4 MΩ) controlled by a micromanipulator (Sutter instruments Cat# MPC-325). Upon successful tear and clear access to a ChR2 (H134R) RGC, a freshly prepared recording micropipette filled with intracellular solution was lowered into the extracellular bath. Prior to recording, the pipette resistance was compensated with a bridge balancing circuit on the amplifier (NPI, BA-1S). Membrane potentials were amplified (BA-1S, NPI), and digitised with 16-bit precision at 20 kHz (USB-6221, National Instruments). Instrument control and recording of membrane potentials were performed using custom MATLAB software.

When a whole-cell configuration was achieved, depolarising and hyperpolarizing current between − 200 pA and 200 pA was delivered in 20 pA current steps to obtain baseline intrinsic properties of the cell. The presence of rebound spikes during hyperpolarising current administration was also used to define OFF cells⁴². Following the recording of the cell’s intrinsic properties, the patched RGC in whole-cell configuration was prepared for optogenetic stimulation. To ensure RGC responses were a direct result of stimulating ChR2-H134R expressing RGCs and not from photoreceptor input, synaptic input to RGCs was eliminated by washing in a synaptic blocker cocktail consisting of (in mM) 0.01 NBQX, 0.05 D-AP5, 0.02 L-AP4, 0.1 picrotoxin and 0.01 strychnine. This blocked di-synaptic photoreceptor input to RGCs⁸⁰.

For optical stimulation, a 488 nm laser light source (Optotech) was used, projected down the optical axis of the microscope and focused to a spot size of 35 µm by the microscope objective at the sample plane. Electrical stimulation was delivered intracellularly via the patch microelectrode.

To determine electrical activation thresholds, 10 stimulus pulses at a 2 Hz repetition rate across a range of amplitudes was delivered to the cell. To calculate optical threshold levels, 10 light pulses at 2 Hz repetition rate was delivered at power levels between 0–12 mJ • cm^− 2 • s^− 1. The stimulus amplitude which elicited a response rate of 50% was defined as threshold and the minimum amplitude that could evoke 100% response reliability was defined as supra-threshold.

To study the impact of hybrid stimulation on stimulation threshold, electrical stimulation was delivered to RGCs (n = 11) using pulse durations of 10 ms and 3 ms, respectively. These stimulation pulses were delivered either simultaneous, or sequentially with electrical pulses being immediately after optical stimuli. Stimulation involving both optical and electrical modalities was delivered at 11 different amplitudes, including a zero-amplitude setting, in a random sequence at a frequency of 2 Hz. Therefore, this approach resulted in the evaluation of 121 unique stimulus combinations, with each combination being repeated 10 times. We defined the threshold as the amplitude that would elicit 50% response reliability according to a sigmoidal fitting.

To study the impact of hybrid stimulation on temporal response, optical and electrical stimulation were delivered using pulse durations of 3 ms and 1 ms, respectively. The stimulation was sustained for 3s at repetitive frequencies of 2, 20, 50 and 200Hz. For all temporal study experiments, electrical pulses were delivered immediately after optical stimulation.

4.4 Morphology Reconstruction and Cell Type Identification

Images of RGCs, backfilled with Alexa Fluor 594 hydrazide during recording experiments, were acquired on a confocal microscope (Olympus cat# FV1200). Following acquisition of the RGC morphology, the retained sample was counterstained with 100µM sulforhomadine-101 in Ames’ media for 5 mins to provide delineation of the retinal layers9. Z-stacks of images from the GCL to the inner nuclear layer were acquired at a resolution of 1024 x 1024 with a step size of 1µm. Adjustments to image contrast and brightness were performed on Adobe Photoshop 2023 or FIJI⁸¹ .

The diameter of the dendritic field was soma were measured using FIJI. To determine whether RFCs were ON, OFF or ON-OFF sub-types, the depth within the IPL where the dendritic field terminated was measured as previously described by Famiglietta Jr & Kolb⁸². Using the GCL/IPL boundary as reference point 0% and the INL/IPL boundary as 100%, RGCs with dendritic arbours that terminated at a depth of < 60% were deemed to be ON-cells.

4.5 Viral Vectors

Plasmids encoding ChrimsonR and GCaMP7s were procured by Addgene before being cloned and packaged into AAV2 vectors by VectorBuilder. pAAV-Syn-ChrimsonR-tdT was a gift from Edward Boyden (Addgene plasmid #59171; http://n2t.net/addgene:59171; RRID: Addgene_59171)¹⁵. pGP-AAV-syn-jGCaMP7f-WPRE was a gift from Douglas Kim & GENIE Project (Addgene plasmid # 104488; http://n2t.net/addgene:104488; RRID:Addgene_104488)⁸³.

4.6 Calcium Imaging

Calcium imaging was performed on retinal flatmounts which underwent AAV mediated transduction of ChrimsonR and GCaMP7s within the GCL in RCS rats. Using a confocal microscope (Olympus, FluoView FV1200) equipped with a 20x water-immersed objective, images were recorded at a frequency of 4.6–7.8 Hz. Each image covered an area of approximately 318 x 318 µm. Images covering different regions of the retina were taken in response to each stimulus, and subsequently stitched together using the MosaicJ plugin⁸⁴. Optical stimulation was delivered through an LED coupled optical fibre (595 nm, M595F, Thorlabs), with fibre tip cleaved. Each optical stimulus consisted of 10 optical pulses delivered at 50 Hz with a pulse duration of 10 ms. The optical stimulation was delivered with the fibre tip positioned on or near the retinal flatmount. In both instances, the optical fibre illuminated a broad region of the retina.

For electrical stimulation, we used an in-house 20 µm platinum-iridium wire insulated with a glass pipette as previously described by Yunzab and colleagues⁶⁴. The electrode was placed epiretinally on the surface of the retina. A silver/silver chloride wire was placed approximately 2 cm away from the stimulating electrode in the chamber and used as a distant electrical return electrode. Electrical stimulation was delivered using a Grapevine Neural Interface System and a Nano2 + Stim headstage (ripple neuro, USA). Each electrical stimulation consisted of a burst of 10 anodic-first biphasic pulses, with a pulse duration of 200 µs and inter-pulse-interval of 33 µs. For hybrid stimulation, electrical pulses were delivered immediately after each optical stimulation. For all stimuli (optical, electrical and hybrid), the stimulus was repeated 10 times and delivered every 2.1 s.

Cell activity was measured by first identifying the location of each ganglion cell body manually using the Region of Interest Manager in ImageJ (Fiji). A custom MATLAB script was used to measure the change in fluorescence (∆F/F) of each cell. For each image, the average cell response was viewed and the stimulus start time identified manually. For a cell to be considered responding to the applied stimulus it had to meet two conditions. Firstly, the ∆F/F signal was filtered to remove the DC component, and thresholded to the mean response plus 2 times the standard deviation. A cell had to reach the threshold at least 5 times within the identified stimulus window, as previously described⁸⁵. For the second condition, a Welch’s power spectral density estimate of the ∆F/F signal was obtained. A signal to noise ratio (SNR) was calculated by measuring the spectral density at 0.48 Hz and dividing by the spectral density at 2 Hz. A SNR above 5 indicated that the cell was responding at a frequency similar to the applied stimulus (Fig. 4c). This additional condition reduced the likelihood that the cell's response was spontaneous.

Four different optical (O0, O1, O2, OL3) and three different electrical (E0, E1, E2) stimulus levels were used. The absence of stimulus was signified as ‘0’, the level at ‘1’ denotes a supra-threshold stimulus near threshold amplitude, ‘2’ was a supra-threshold level that was twice the amplitude of ‘1’, and ‘3’ was a supra-threshold level at least three times ‘1’. Suprathreshold stimuli were used to observe the impact of hybrid stimulation on a larger population of cells as this meant a larger number of cells were activated by the stimulus. We defined threshold as the amplitude required to evoke activity in at least one cell in a 318 µm x 318 µm imaging frame that had the electrical stimulation electrode within the field of view. This was assessed visually by reviewing the recording at the time of the experiment.

For each animal, the number of RGCs and area activated was quantified as a normalised percentage against the number of RGCs and area activated in the OL3 condition. The area of activated retina was measured by finding a 95% confidence area ellipse around the active cells using MATLAB script ‘plot_ellipse.m’. The area of this ellipse was then calculated. ChrimsonR cells were identified via visual identification of the tdTomato fluorescent reporter during confocal microscopy. The image threshold was determined using the MATLAB script ‘graythresh.m’ from Otsu’s method⁸⁶. The location of cells expressing GCaMP7s was then overlayed on the threshold image, and any cells co-expressing ChrimsonR were identified visually. The percentage of active ChrimsonR cells was calculated by firstly determining all cells that had a calcium response for a given stimulus (total active cells), and then determining which of these cells also expressed ChrimsonR (number of active ChrimsonR cells. The percentage of Active ChrimsonR Cells was then calculated by Eq. 1.

\(\%Active ChrimsonR Cells = \frac{Active ChrimsonR Cells}{Total Active Cells}\times 100\) [1]

The active cells and area activated were normalised to the response observed at the maximum optical only stimulus applied (O3). The percentage of ChrimsonR cells were min/max normalised.

4.7 Histology

The protocol for immunofluorescent labelling of the retina has been adapted from Hendrickson and colleagues⁸⁷. Retinae were fixed in 4% paraformaldehyde in 0.01M PBS for 24 hours at 4°C. Following fixation, they were washed in 0.01M PBS. To prepare retina sections, the tissue was placed in a [1:1] solution of OCT (Tissue Tek) and 30% sucrose in PBS for 16 hours. Retinae were then placed in a cryomold, embedded with OCT and samples were snap frozen in 2-methylbutane that has been chilled to -50°C for 2 minutes. Samples were then cut in the transverse plane on a cryostat (Leica #CM3050S) at a thickness of 20µm. Sections were mounted directly onto Superfrost Plus® slides and stored at -20°C.

To delineate the layers of the retina, sections were counterstained with Hoechst 33258 (Life Technologies #H3569, 1:1000). Sections were also immunolabelled with an antibody against GFP (Abcam #ab13970, 1:2000) to amplify the signal of the ChR2(H134R)-YFP cells and an antibody against mCherry (Rockland #600-401-P16) to amplify ChrimsonR signal in RCS-p + retina samples. Images from transverse sections were acquired on a wide field Leica Thunder Imager microscope using a 20x objective.

4.8 Data and Statistical Analysis

Our primary measure of the efficacy of a stimulation modality was the response rate, defined as the percentage of pulses that elicited action potentials. Action potentials were detected by threshold crossings at -20 mV, confirmed through visual validation of the stimulus artefact. For each stimulus level, we assessed the response efficacy by fitting a sigmoid function

\(f\left(x\right) = \frac{a}{1+{e}^{-b\left(x-c\right)}}\) [2]

in which f(x) is the response rate and x is the amplitude of electrical or optical stimulation. Here, coefficient b represents scaling factors that determine the saturation amplitudes, a represents the gain of the sigmoidal curve, and c represents the thresholds (50% of the saturation level).

For threshold analysis experiments, for each cell, we estimated optical thresholds at each level of electrical stimulation and vice versa. In our population analysis, hybrid threshold values were normalised to those observed under sole electrical or optical stimulation. Data from cells exhibiting over 50% response rate even at the lowest stimulation levels were excluded from the population analysis. We then categorized the threshold values into four groups based on normalised stimulation levels: 10–30%, 30–50%, 50–70% and 70–90%, facilitating comparative analysis. Two-way analysis of variance (ANOVA) with post hoc multiple Tukey comparisons was used to quantify the effects of hybrid stimulation on optical/electrical thresholds and response reliability (p < 0.05 was considered significant; GraphPad Prism).

For temporal analysis experiments, each cell was stimulated with the three stimulation modalities as described above at 2 Hz, 20 Hz, 50 Hz and 200 Hz repetition rates. The duration for each pulse train was three seconds and delivery of stimulation at the four frequencies and three stimulation modalities (12 conditions) constituted one trial. Cells included in this study underwent a minimum of three trials. The number of action potentials elicited during stimulation pulse delivery (e.g. 150 pulses during 3 sec of 50 Hz stimulation) was used as a measure of response reliability. The response reliability was expressed as a percentage and averaged (mean) across all trials. Two-way analysis of variance (ANOVA) with post hoc multiple Tukey comparisons was used to quantify the temporal fidelity of hybrid stimulation against optical-only and electrical-only stimulation across all four frequencies (p < 0.05 was considered significant; GraphPad Prism).

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgements

The authors wish to acknowledge the contributions and assistance of A. Brzostowska and S. Kwon for technical support with animal husbandry. This work was supported by an Ideas Grant from the National Health and Medical Research Council (GNT2002523). W.T. acknowledges the support from the Australian Research Council via DECRA Fellowship (DE220100302) and Linkage Grant (LP180100638). W.C.K. acknowledges the support provided by the National Health and Medical Research Council Dora Lush Postgraduate Scholarship (APP1190007). The Australian Regenerative Medicine Institute is supported by grants from the State Government of Victoria and the Australian Government. The Australian College of Optometry provided funding towards this project.

Author Contributions

W.C.K., E.K.B., M.R.I., R.T.R. and W.T. conceived and designed the research. W.C.K. performed all current clamp and histological experiments. W.C.K., E.K.B., J.M.B., performed intravitreal injection surgical experiments. E.K.B. and J.M.B. performed calcium imaging experiments. R.T.R provided ChR2-H134R mouse models. W.C.K., E.K.B. and W.T analysed the data. T.K., P.R.S., M.R.I., R.T.R. and W.T. supervised the study. W.C.K., E.K.B., M.R.I., R.T.R. and W.T interpreted the data and wrote the manuscript.

Competing Interests

The authors declare no competing interests

Jonas, J. B. et al. Visual Impairment and Blindness Due to Macular Diseases Globally: A Systematic Review and Meta-Analysis. Am. J. Ophthalmol. 158, 808–815 (2014).
Bull, N. D. & Martin, K. R. Concise Review: Toward Stem Cell‐Based Therapies for Retinal Neurodegenerative Diseases. STEM CELLS 29, 1170–1175 (2011).
Cross, N., Steen, C. van, Zegaoui, Y., Satherley, A. & Angelillo, L. Retinitis Pigmentosa: Burden of Disease and Current Unmet Needs. Clin. Ophthalmol. (Auckl., NZ) 16, 1993–2010 (2022).
Finn, A. P., Grewal, D. S. & Vajzovic, L. Argus II retinal prosthesis system: a review of patient selection criteria, surgical considerations, and post-operative outcomes. Clin. Ophthalmol. (Auckl., NZ) 12, 1089–1097 (2018).
Palanker, D., Mer, Y. L., Mohand-Said, S. & Sahel, J. A. Simultaneous perception of prosthetic and natural vision in AMD patients. Nat. Commun. 13, 513 (2022).
Allen, P. J. Retinal prostheses: Where to from here? Clin. Exp. Ophthalmol. 49, 418–429 (2021).
Bansal, A., Shikha, S. & Zhang, Y. Towards translational optogenetics. Nat. Biomed. Eng. 7, 349–369 (2023).
Chen, F. et al. Visual function restoration with a highly sensitive and fast Channelrhodopsin in blind mice. Signal Transduct. Target. Ther. 7, 104 (2022).
McGregor, J. E. et al. Optogenetic restoration of retinal ganglion cell activity in the living primate. Nat. Commun. 11, 1703 (2020).
Chaffiol, A. et al. In vivo optogenetic stimulation of the primate retina activates the visual cortex after long-term transduction. Mol. Ther. - Methods Clin. Dev. 24, 1–10 (2022).
Gauvain, G. et al. Optogenetic therapy: high spatiotemporal resolution and pattern discrimination compatible with vision restoration in non-human primates. Commun. Biol. 4, 125 (2021).
Sahel, J.-A. et al. Partial recovery of visual function in a blind patient after optogenetic therapy. Nat Med 27, 1223–1229 (2021).
Ferrari, U. et al. Towards optogenetic vision restoration with high resolution. PLoS Comput. Biol. 16, e1007857 (2020).
Richter, C. & Bruegmann, T. No light without the dark: Perspectives and hindrances for translation of cardiac optogenetics. Prog. Biophys. Mol. Biol. 154, 39–50 (2020).
Klapoetke, N. C. et al. Independent optical excitation of distinct neural populations. Nat Methods 11, 338–346 (2014).
Li, W., Ghaffari, D. H., Misra, R. & Weiland, J. D. Retinal ganglion cell desensitization is mitigated by varying parameter constant excitation pulse trains. Front Cell Neurosci 16, 897146 (2022).
Sengupta, A. et al. Red‐shifted channelrhodopsin stimulation restores light responses in blind mice, macaque retina, and human retina. EMBO Mol. Med. 8, 1248–1264 (2016).
Schmidt, T. M. et al. A Role for Melanopsin in Alpha Retinal Ganglion Cells and Contrast Detection. Neuron 82, 781–788 (2014).
Jepson, L. H. et al. High-Fidelity Reproduction of Spatiotemporal Visual Signals for Retinal Prosthesis. Neuron 83, 87–92 (2014).
Lin, J. Y. A user’s guide to channelrhodopsin variants: features, limitations and future developments. Exp. Physiol. 96, 19–25 (2011).
Lin, J. Y., Lin, M. Z., Steinbach, P. & Tsien, R. Y. Characterization of Engineered Channelrhodopsin Variants with Improved Properties and Kinetics. Biophys J 96, 1803–1814 (2009).
Wäldchen, S., Lehmann, J., Klein, T., Linde, S. van de & Sauer, M. Light-induced cell damage in live-cell super-resolution microscopy. Sci. Rep. 5, 15348 (2015).
Mager, T. et al. High frequency neural spiking and auditory signaling by ultrafast red-shifted optogenetics. Nat Commun 9, 1750 (2018).
Ajay, E. A. et al. Auditory nerve responses to combined optogenetic and electrical stimulation in chronically deaf mice. J. Neural Eng. 20, 026035 (2023).
Richardson, R. T. et al. Viral-mediated transduction of auditory neurons with opsins for optical and hybrid activation. Sci. Rep. 11, 11229 (2021).
Hart, W. L. et al. Combined optogenetic and electrical stimulation of auditory neurons increases effective stimulation frequencyan in vitro study. J Neural Eng 17, 016069 (2020).
Thompson, A. C. et al. Hybrid optogenetic and electrical stimulation for greater spatial resolution and temporal fidelity of cochlear activation. J Neural Eng 17, 056046 (2020).
Matarazzo, J. V. et al. Combined optogenetic and electrical stimulation of the sciatic nerve for selective control of sensory fibers. Front. Neurosci. 17, 1190662 (2023).
B., L., Barry, R., M., Paul & Ulrike, G. Retinal connectivity and primate vision. Prog. Retin. Eye Res. 29, 622–639 (2010).
Tae-Jin, K. & Chang-Jin, J. Morphological Classification of Parvalbumin-Containing Retinal Ganglion Cells in Mouse: Single-Cell Injection after Immunocytochemistry. Investig. Opthalmology Vis. Sci. 47, 2757–2764 (2006).
Kovács-Öller, T. et al. Spatial Expression Pattern of the Major Ca2+-Buffer Proteins in Mouse Retinal Ganglion Cells. Cells 9, 792 (2020).
Goetz, J. et al. Unified classification of mouse retinal ganglion cells using function, morphology, and gene expression. Cell Rep. 40, 111040 (2022).
Sun, W., Li, N. & He, S. Large‐scale morphological survey of mouse retinal ganglion cells. J Comp Neurol 451, 115–126 (2002).
Sanes, J. R. & Masland, R. H. The Types of Retinal Ganglion Cells: Current Status and Implications for Neuronal Classification. Annu. Rev. Neurosci. 38, 1–26 (2015).
Duke, A. R. et al. Combined optical and electrical stimulation of neural tissue in vivo. J. Biomed. Opt. 14, 060501-060501–3 (2009).
Adachi, K. et al. Optical Coherence Tomography of Retinal Degeneration in Royal College of Surgeons Rats and Its Correlation with Morphology and Electroretinography. PLoS ONE 11, e0162835 (2016).
Strauss, O., Stumpff, F., Mergler, S., Wienrich, M. & Wiederholt, M. The Royal College of Surgeons Rat: An Animal Model for Inherited Retinal Degeneration with a Still Unknown Genetic Defect. Cells Tissues Organs 162, 101–111 (1998).
Ren, Y.-M., Weng, C.-H., Zhao, C.-J. & Yin, Z.-Q. Changes in intrinsic excitability of ganglion cells in degenerated retinas of RCS rats. Int. J. Ophthalmol. 11, 756–765 (2018).
Chen, Z., Song, Y., Yao, J., Weng, C. & Yin, Z. Q. Alterations of Sodium and Potassium Channels of RGCs in RCS Rat with the Development of Retinal Degeneration. J. Mol. Neurosci. 51, 976–985 (2013).
Richardson, R. T., Ibbotson, M. R., Thompson, A. C., Wise, A. K. & Fallon, J. B. Optical stimulation of neural tissue. Healthc. Technol. Lett. 7, 58–65 (2020).
Azees, A. A. et al. Spread of activation and interaction between channels with multi-channel optogenetic stimulation in the mouse cochlea. Hear. Res. 440, 108911 (2023).
Margolis, D. J. & Detwiler, P. B. Different Mechanisms Generate Maintained Activity in ON and OFF Retinal Ganglion Cells. J. Neurosci. 27, 5994 (2007).
Uzzell, V. J. & Chichilnisky, E. J. Precision of Spike Trains in Primate Retinal Ganglion Cells. J Neurophysiol 92, 780–789 (2004).
Degenaar, P. et al. Optobionic vision?a new genetically enhanced light on retinal prosthesis. J. Neural Eng. 6, 035007 (2009).
Pan, Z.-H., Ganjawala, T. H., Lu, Q., Ivanova, E. & Zhang, Z. ChR2 Mutants at L132 and T159 with Improved Operational Light Sensitivity for Vision Restoration. PLoS ONE 9, e98924 (2014).
Tomita, H. et al. Channelrhodopsin-2 gene transduced into retinal ganglion cells restores functional vision in genetically blind rats. Exp. Eye Res. 90, 429–436 (2010).
Lin, C.-I. & Chiao, C.-C. Blue Light Promotes Neurite Outgrowth of Retinal Explants in Postnatal ChR2 Mice. eNeuro 6, ENEURO.0391-18.2019 (2019).
Marshel, J. H. et al. Cortical layer–specific critical dynamics triggering perception. Science 365, (2019).
Bansal, H., Gupta, N. & Roy, S. Theoretical analysis of optogenetic spiking with ChRmine, bReaChES and CsChrimson-expressing neurons for retinal prostheses. J. Neural Eng. 18, 0460b8 (2021).
Weiland, J. D., Walston, S. T. & Humayun, M. S. Electrical Stimulation of the Retina to Produce Artificial Vision. Annu. Rev. Vis. Sci. 2, 273–294 (2016).
Ayton, L. N. et al. An update on retinal prostheses. Clin. Neurophysiol. 131, 1383–1398 (2020).
Tong, W., Meffin, H., Garrett, D. J. & Ibbotson, M. R. Stimulation Strategies for Improving the Resolution of Retinal Prostheses. Front. Neurosci. 14, 262 (2020).
Pan, Z.-H., Lu, Q., Bi, A., Dizhoor, A. M. & Abrams, G. W. Optogenetic Approaches to Restoring Vision. Annu. Rev. Vis. Sci. 1, 185–210 (2015).
Kotsakidis, R., Meffin, H., Ibbotson, M. R. & Kameneva, T. In vitro assessment of the differences in retinal ganglion cell responses to intra- and extracellular electrical stimulation. J. Neural Eng. 15, 046022 (2018).
Nowak, L. G. & Bullier, J. Axons, but not cell bodies, are activated by electrical stimulation in cortical gray matter I. Evidence from chronaxie measurements. Exp. Brain Res. 118, 477–488 (1998).
Nowak, L. G. & Bullier, J. Axons, but not cell bodies, are activated by electrical stimulation in cortical gray matter II. Evidence from selective inactivation of cell bodies and axon initial segments. Exp. Brain Res. 118, 489–500 (1998).
Fried, S. I., Lasker, A. C. W., Desai, N. J., Eddington, D. K. & Rizzo, J. F. Axonal Sodium-Channel Bands Shape the Response to Electric Stimulation in Retinal Ganglion Cells. J. Neurophysiol. 101, 1972–1987 (2009).
Shin, D. S. et al. High frequency stimulation or elevated K+ depresses neuronal activity in the rat entopeduncular nucleus. Neuroscience 149, 68–86 (2007).
Kameneva, T. et al. Retinal ganglion cells: mechanisms underlying depolarization block and differential responses to high frequency electrical stimulation of ON and OFF cells. J. Neural Eng. 13, 016017 (2016).
Do, M. T. H. & Yau, K.-W. Intrinsically Photosensitive Retinal Ganglion Cells. Physiol. Rev. 90, 1547–1581 (2010).
Muralidharan, M. et al. Neural activity of functionally different retinal ganglion cells can be robustly modulated by high-rate electrical pulse trains. J. Neural Eng. 17, 045013 (2020).
Hadjinicolaou, A. E., Cloherty, S. L., Hung, Y.-S., Kameneva, T. & Ibbotson, M. R. Frequency Responses of Rat Retinal Ganglion Cells. Plos One 11, e0157676 (2016).
Madugula, S. S. et al. Focal electrical stimulation of human retinal ganglion cells for vision restoration. J. Neural Eng. 19, 066040 (2022).
Yunzab, M. et al. Preferential modulation of individual retinal ganglion cells by electrical stimulation. J Neural Eng 19, 045003 (2022).
Begeng, J. M. et al. Activity of Retinal Neurons Can Be Modulated by Tunable Near-Infrared Nanoparticle Sensors. ACS Nano 17, 2079–2088 (2023).
Cronin, T. et al. Efficient transduction and optogenetic stimulation of retinal bipolar cells by a synthetic adeno‐associated virus capsid and promoter. EMBO Mol. Med. 6, 1175–1190 (2014).
Lu, Q. et al. AAV-mediated transduction and targeting of retinal bipolar cells with improved mGluR6 promoters in rodents and primates. Gene Ther. 23, 680–689 (2016).
Jüttner, J. et al. Targeting neuronal and glial cell types with synthetic promoter AAVs in mice, non-human primates and humans. Nat. Neurosci. 22, 1345–1356 (2019).
Yue, L., Castillo, J., Gonzalez, A. C., Neitz, J. & Humayun, M. S. Restoring Color Perception to the Blind An Electrical Stimulation Strategy of Retina in Patients with End-stage Retinitis Pigmentosa. Ophthalmology 128, 453–462 (2021).
Paknahad, J., Loizos, K., Yue, L., Humayun, M. S. & Lazzi, G. Color and cellular selectivity of retinal ganglion cell subtypes through frequency modulation of electrical stimulation. Sci Rep-uk 11, 5177 (2021).
Bortoletto, M., Rodella, C., Salvador, R., Miranda, P. C. & Miniussi, C. Reduced Current Spread by Concentric Electrodes in Transcranial Electrical Stimulation (tES). Brain Stimul. 9, 525–528 (2016).
R, H., Alexander. Current perspectives on the safe electrical stimulation of peripheral nerves with platinum electrodes. Bioelectron. Med. 3, 37–49 (2020).
Entcheva, E. & Kay, M. W. Cardiac optogenetics: a decade of enlightenment. Nat. Rev. Cardiol. 18, 349–367 (2021).
Palanker, D. Electronic Retinal Prostheses. Cold Spring Harb. Perspect. Med. 13, a041525 (2023).
Lee, S., Park, J., Kwon, J., Kim, D. H. & Im, C.-H. Multi-channel transorbital electrical stimulation for effective stimulation of posterior retina. Sci. Rep. 11, 9745 (2021).
Fujikado, T. et al. Effect of Transcorneal Electrical Stimulation in Patients with Nonarteritic Ischemic Optic Neuropathy or Traumatic Optic Neuropathy. Jpn. J. Ophthalmol. 50, 266–273 (2006).
Puthussery, T. & Fletcher, E. Extracellular ATP induces retinal photoreceptor apoptosis through activation of purinoceptors in rodents. J. Comp. Neurol. 513, 430–440 (2009).
Edelstein, A. D. et al. Advanced methods of microscope control using μManager software. J Biological Methods 1, e10 (2014).
Hamill, O. P., Marty, A., Neher, E., Sakmann, B. & Sigworth, F. J. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflügers Arch. 391, 85–100 (1981).
Guo, T. et al. Mediating Retinal Ganglion Cell Spike Rates Using High-Frequency Electrical Stimulation. Front Neurosci-switz 13, 413 (2019).
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat Methods 9, 676–682 (2012).
V., F., E. & Helga, K. Structural Basis for ON-and OFF-Center Responses in Retinal Ganglion Cells. Science 194, 193–195 (1976).
Dana, H. et al. High-performance calcium sensors for imaging activity in neuronal populations and microcompartments. Nat Methods 16, 649–657 (2019).
Thévenaz, P. & Unser, M. User‐friendly semiautomated assembly of accurate image mosaics in microscopy. Microsc. Res. Tech. 70, 135–146 (2007).
Tong, W. et al. Improved visual acuity using a retinal implant and an optimized stimulation strategy. J. Neural Eng. 17, 016018 (2020).
Otsu, N. A Threshold Selection Method from Gray-Level Histograms. IEEE Trans. Syst., Man, Cybern. 9, 62–66 (1979).
Hendrickson, A., Possin, D., Kwan, W. C., Huang, J. & Bourne, J. A. The temporal profile of retinal cell genesis in the marmoset monkey. J Comp Neurol 524, 1193–1207 (2015).

There is NO Competing Interest.

SupplementaryTablesandFigure.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Hybrid Optogenetic and Electrical Stimulation of Retinal Ganglion Cells for Artificial Vision

Status:

Version 1

Abstract

Figures

1. Introduction

2. Results

2.1 A transgenic mouse model to explore hybrid stimulation in the retina

2.2 Hybrid stimulation reduces electrical and optical thresholds

2.3 The temporal bandwidth of hybrid stimulation surpasses electrical and optical only stimulation

2.4 Hybrid Stimulation can activate RGCs in a degenerated retina with reduced optical power

3. Discussion

4. Methods

4.1 Animals

4.2 Retinal Preparation

4.3 Whole-cell patch clamping

4.4 Morphology Reconstruction and Cell Type Identification

4.5 Viral Vectors

4.6 Calcium Imaging

4.7 Histology

4.8 Data and Statistical Analysis

Declarations

Data Availability

Acknowledgements

Author Contributions

Competing Interests

References

Additional Declarations

Supplementary Files

Status:

Version 1