Optogenetics

When photoreceptors have degenerated, researchers face a choice of which surviving inner retinal neuron to target with an opsin. Three main options exist: RGCs (the output neurons of the retina), ON-bipolar cells (one synapse upstream of RGCs, still receiving rod/cone bipolar circuitry input), or dormant cone remnants (the soma of dead cones whose outer segments have degenerated but whose inner segments may survive for years). Each option carries different implications for signal quality, therapeutic window, and the type of visual computation that is preserved after rescue.

RGC targeting is technically straightforward and produces fast, light-driven spiking, but bypasses all inner retinal processing (no ON/OFF separation, no centre-surround receptive field organisation). Bipolar cell targeting is mechanistically superior because it leverages the preserved rod bipolar and cone bipolar circuitry: the light-driven signal from the opsin in the bipolar cell still passes through the bipolar-to-RGC synapse, partially reconstituting ON pathway processing. Cone remnant targeting aims to exploit long-surviving inner segments in slow-progressing dystrophies, with the goal of restoring near-natural photoreceptor signalling.

Measuring the quality of optogenetically restored vision in rodents requires careful instrument selection. The OptoDrum measures the subcortical optomotor reflex – a brainstem-level tracking response driven by the retino-olivary and accessory optic pathways. It directly reports whether the restored signal is strong enough to drive reflex head tracking, and at what spatial frequency threshold. This is the appropriate primary endpoint for most preclinical optogenetics studies. However, the OMR does not assess cortical visual processing: it cannot determine whether the animal can discriminate complex patterns, detect low-contrast textures, or make learned visual decisions. For those questions – relevant to translational studies modelling real-world visual rehabilitation – AcuiSee provides the complementary cortical operant endpoint.

Selecting the optimal opsin for a clinical-grade optogenetic therapy requires balancing several competing parameters: the spectral peak must match ambient illumination conditions (or the emission wavelength of a wearable light amplifier); the sensitivity (irradiance threshold) must be low enough to drive responses at safe, physiologically tolerable light levels; and kinetics must be fast enough to support temporal resolution at useful frame rates. Channelrhodopsin-2 (ChR2), the original optogenetic tool, absorbs maximally in the blue (λ ~470 nm) and requires high irradiance (>10¹⁵ photons cm^-2 s^-1) to drive reliable spiking – a potential safety concern for clinical use. Red-shifted variants (ReaChR, Chronos, ChrimsonR) reduce irradiance requirements. MW-cone opsins are G-protein-coupled and more sensitive at lower light levels but have slower kinetics. Opto-mGluR6, a fusion of ChR2 with the mGluR6 promoter driving expression in ON-bipolar cells, exploits the native ON-bipolar cell signalling cascade to amplify the response.

Preclinical benchmarking of these opsin variants requires a functional readout that is sensitive enough to detect acuity differences across irradiance levels and spectral conditions. The OptoDrum optomotor paradigm provides exactly this: spatial frequency thresholds can be measured across grating contrast and luminance conditions, constructing irradiance-acuity curves that reveal the functional irradiance threshold and the acuity ceiling for each opsin candidate.

Translating an optogenetic therapy from proof-of-concept to clinical vector dosing requires knowing the minimum effective dose – the AAV titre at which opsin expression in a sufficient proportion of target cells produces a detectable, sustained visual benefit. Too low a dose yields sub-threshold opsin expression: the retina may be partially transduced but the signal is too weak to drive OMR-detectable acuity. Too high a dose risks immune responses, off-target expression, and regulatory hurdles. Histological transduction counts (percentage of target cells expressing the opsin by immunofluorescence) provide structural dose-response data but do not translate directly to visual function: the functional dose-response relationship is non-linear and depends on the density of opsin-expressing cells relative to the spatial sampling mosaic of the inner retina.

The OptoDrum addresses this gap by providing a direct behavioural dose-function readout. Across a range of AAV titres, OptoDrum measures spatial acuity in blind mice after opsin delivery, constructing a dose-acuity curve that reveals the threshold titre for functional rescue and the plateau acuity at saturating doses. This non-invasive, repeatable readout can be obtained at multiple post-injection time points, enabling the researcher to also characterise the kinetics of functional recovery (the lag between vector injection and measurable visual restoration) and the durability of the benefit.

CRISPR-based delivery modalities introduce an additional dose-response dimension: the efficiency of genomic integration or targeted replacement, rather than episomal AAV expression, determines long-term opsin levels. Maddalena et al. (2023) validated that CRISPR-mediated optogenetic expression supports OptoDrum-measurable acuity recovery, establishing the paradigm that functional benchmarking by OMR is applicable regardless of the delivery vector.

Interpreting OptoDrum results in optogenetically rescued animals requires understanding both what the OMR measures and what it cannot measure. The OMR is a subcortical reflex mediated by the accessory optic system (AOS) and the nucleus of the optic tract (NOT): it detects whole-field motion, is biased toward low spatial frequencies, and is driven primarily by the superior temporal visual cortex-independent pathway. In a fully sighted mouse, the OMR acuity ceiling is approximately 0.5 cycles per degree. In an optogenetically rescued mouse where opsin is expressed in a sparse subset of bipolar cells or RGCs, spatial sampling is sub-optimal and the OMR threshold is correspondingly lower – typically 0.1 to 0.3 cycles per degree in the published literature, depending on opsin, cell type, and AAV dose.

This distinction matters for translational benchmarking. An OMR improvement above the blind baseline is genuine evidence that the opsin expression has re-established a functional retina-to-brainstem visual circuit. It does not demonstrate that the animal can perform pattern discrimination, navigate a complex visual environment, or detect fine spatial detail. These higher-order endpoints require cortical processing, which AcuiSee is designed to measure. A complete optogenetics functional battery for translational benchmarking would combine OptoDrum (reflex-level circuit integrity) with AcuiSee (operant cortical discrimination), analogous to the ERG-plus-visually-guided-behaviour battery in clinical trial design.

The sensitivity of the OMR assay also sets a practical floor: very low opsin expression, very sparse cell targeting, or opsins with high irradiance thresholds may not produce an OMR response even when electrophysiological recordings confirm measurable light responses in individual RGCs. This detection floor is determined by the minimum number of synchronously activated neurons required to drive the AOS reflex, and it is an important parameter to establish early in any optogenetics programme.

Clinical translation of optogenetic vision restoration requires a clear bridging argument: from in vitro opsin characterisation (expression levels, electrophysiology, kinetics) through in vivo rodent functional validation, to the prediction of therapeutic benefit in human participants. Regulatory and clinical development programmes need to demonstrate that the chosen opsin, at the proposed clinical dose, produces a functional signal above a reproducible threshold in the target cell population – and that this threshold corresponds to a meaningful visual gain. In human trials, the primary endpoint is typically a behavioural or psychophysical test (letter chart, spatial frequency discrimination, navigation task). The preclinical analogue is the OMR (OptoDrum) for reflex-level circuit integrity and an operant discrimination task (AcuiSee) for cortical-level benefit.

A critical challenge in translational benchmarking is the difference in spatial scale between rodent and human retinas: mouse inner retinal neurons are larger and less numerous, meaning the functional acuity ceiling in optogenetically rescued mice is inherently lower than what might be achievable in humans with a denser foveal bipolar cell mosaic. This species-scaling argument must be taken into account when using rodent OptoDrum data to predict clinical benefit. Nonetheless, the rodent OMR data serves an essential role in establishing proof-of-mechanism and dose-response parameters before primate or clinical studies.