The clinical potential of optogenetic interrogation of pathogenesis

2.1 Structure and photophysics of microbial opsin

Optogenetics uses light to control protein conformation and empowers new ways to control molecular activity in live cells. Opsin-based optogenetics enabled accurate control of ion flux across the cell membrane through light-sensitive channels or pumps.^{7, 31-33} Successful implementation of opsin-based optogenetics requires (1) microbial opsins that transport ions upon exposure to light, (2) a strategy to express opsins in specific target cells, and (3) precise delivery of light to the target cells.³⁴ Channelrhodopsin-2 (ChR2), the opsin derived from the green freshwater algae Chlamydomonas reinhardtii, is a seven-pass transmembrane protein that belongs to microbial opsins. ChR2 binds a retinal cofactor that undergoes cis-trans conformational changes when absorbing blue light. Such a conformational change renders an increased cations influx (Na⁺ and Ca²⁺) through ChR2 and stimulates action potential in hosting excitable cells (e.g., neurons or muscles)^{35, 36} (Figure 1). ChR variants such as halorhodopsins (HRs) pump chloride ions into the cells upon light illumination, leading to optical induction of hyperpolarization and neuronal inhibition.³⁷ Genetic ChR expression can be virally driven (lentivirus or adeno-associated virus) in cells or animals. Cre-dependent ChR expression transgenic mice have also been developed (e.g., strain Ai32, the Jackson Laboratory) to facilitate the production of animals bearing the tissue-specific expression of ChR2.

2.2 The clinical accomplishment of opsin-based optogenetics

Current clinical trials of optogenetics mainly focus on treating blindness due to the easy delivery of light through the eyes. We discuss six records of clinical trials, five active and one completed, which target retinal diseases. Key information about each clinical trial is listed in Table 1.

The fundamental idea of using light to treat retinal diseases is to revamp defective visionary neurons with light sensitivity. Retinitis pigmentosa (RP) is a genetic disease that affects the retina and causes the breakdown of photoreceptors, affecting the peripheral vision first and eventually spreading to the central vision.³⁸ Stargardt's disease was a genetic disease with fatty material build up on the retina's central region (macula), resulting in the loss of central vision and light sensitivity.³⁹ AbbVie sponsored the earliest phase I/II clinical trials to determine the dose-dependent safety and tolerability of single-dose intravitreal RSO-001 to advanced RP patients (NCT02556736). Based on the results updated on 6 October 2022, of the 14 participants, the treatment resulted in 0% all-cause mortality or serious adverse events. Nine participants experienced grade 3, non-life threatening but medically significant side effects, including mild eye discharge, irritation, pain, and infection. A dose-escalation trial was carried out to evaluate the safety and tolerability of adeno-associated virus (AAV) as the transgene vehicle. This completed phase I/IIa clinical trial by Nanoscope Therapeutics (NCT04919473) evaluated patient tolerance to vMCO-I, a serotype 2 AAV carrying a multi-characteristic opsin (MCO) gene expression cassette. Eleven patients with advanced RP received intravitreal vMCO-I at high (3.5E11 viral genome per eye) or low (1.75E11 viral genome per eye) doses. vMCO-I infects and expresses polychromatic opsin, which can be activated by ambient light, in patients’ bipolar cells. All subjects had objective and subjective improvement in functional vision, including shape discrimination accuracy improved to greater than 90% in all subjects compared to baseline. Following the phase I/IIa trial (NCT04919473), Nanoscope Therapeutics started a phase IIb trial (NCT04945772) that addresses the efficacy and systemic adverse effects at a lower dose (dropping from 3.5E11-1.75E11 to 1.2E11-0.9E11). The same company has another ongoing phase IIa clinical trial that assesses the safety and efficacy of vMCO-010 for Stargardt's disease (NCT05417126).

Recently, GenSight Biologics reported promising efficacy results in the NCT03326336 trial, which showed partial visual function recovery in a blind RP patient. The patient received an intraocular injection of an AAV vector (5E10 viral genome per eye) encoding ChrimsonR, a red light (590 nm) responsive opsin. The red light was delivered through an engineered goggle to stimulate human retinal ganglion cells. During 84 weeks of treatment, the patient perceived, located, counted, and touched different objects. In addition, the patient's occipital neural activity assessment was also successfully correlated with functional visual recovery.

Besides bipolar cells, University Hospital, Basel, Switzerland, initiated a clinical trial (NCT05294978) in 2021 targeting cone cells to restore visions in patients with inherited retinal dystrophies. Cone cells are crucial in color vision, daytime activity, and faster bright light flash recovery. A niche is that light-sensitive cones remain alive but dormant before the disease progresses into the late stage, posing an excellent time window to re-sensitize them through optogenetics. However, the population of patients in the dormant stage is unknown. The primary aim of this study is to identify eligible patients worldwide (USA, China, Germany, Hungary, Italy, Switzerland, and the UK) through a multicenter ocular imaging study (EyeConic Study) for cone-optogenetics treatment. Currently, there are 1000 patients enrolled, and their eligibility will be determined via macular optical coherence tomography, which captures the pathogenic state of the retina, as well as deeper eye structure and vasculature, non-invasively.

3.1 Structure and photophysics of opsin-free photoactivatable proteins

Opsin-free optogenetics shares common prerequisites, such as photoactivatable proteins, transgene expression, and light delivery, as opsin-based technology. In contrast to fluorescent proteins, optical probes for localizing proteins, photoactivatable proteins serve as actuators to tune target molecules’ activity. Light stimulation of the fluorophore of these photoactivable proteins modifies PPIs, which expands the mode of action to modulate molecular activity and the scope of biological processes.

A photoactivatable protein's core components are the photosensory and effector domains. Photosensory domains contain chromophores, excitation of which converts energy stored in photons to the chemical potential that changes the conformation of host molecules. Chromophore excitation results in conformational changes that lead to protein association, dissociation, and uncaging or allosteric effects in host photoactivatable proteins. For example, blue light excitation of flavin mononucleotide (FMN) cofactor in Avena sativa light, oxygen, or voltage domain (AsLOV) results in its formation of an adduct with the proximal cysteine residue, which induces conformational changes in LOV. Cryptochrome uses flavin adenine dinucleotide (FAD) to mediate its light-regulated homo-oligomerization or heterodimerization. Phytochromes use phycocyanobilin (PCB) or biliverdin, red and far red light-sensitive chromophores, for conformational modulation. To date, crystal structures and light-sensitive chromophores of various photoactivatable proteins have been resolved (Figure 2, Table 2). The identified modes of action help the structure-guided design of innovative optogenetic systems. We will use recent studies to showcase each mode of action, such as photo-uncaging and association.

4.1 Spatiotemporal accuracy

Optogenetics uses light (photons) to modulate molecular activity. Compared to diffusive chemical ligands, photons can be spatially distributed in a user-defined manner. A simple optical objective can spatially focus a coherent photon flux (e.g., a laser beam) to span about only half wavelength of the photon, reaching the far-field diffraction limit. Fortunately, the wavelength of visible light (400–700 nm) is significantly smaller than the typical size of a cell, ranging from one micron for bacteria and tens of microns for mammalian cells (Figure 3). By systematically tuning the phase distribution of photons by devices like spatial light modulators, light can almost be distributed into any spatial pattern, for example, an array of focus, a doughnut-shaped (as in stimulated emission depletion microscopy or STED), sheet-shaped (as in light-sheet microscopy), defined by the user. Note that non-coherent light sources, such as those from a light bulb or LED, could not reach the diffraction limit because each photon has different propagation directions and phases.

4.2 Target specificity

Target specificity arises from the capacity to express the optogenetic protein in a tissue- and cell-type-specific manner. Recent work demonstrated the use of optogenetics in treating defective lower urinary tract (LUT) with a mouse model.⁴⁰ LUT's physiological role, storing and emptying urine, requires coordinated, counter-acting mechanisms of the detrusor and urethral sphincter muscles surrounding the bladder. Urination (emptying) involves the contraction detrusor and relaxation of the urethral sphincter muscles, whereas storing urine requires the opposite action from both types of muscles. However, pharmacological treatments (e.g., anticholinergics, adrenergic receptor agonists or botulinum toxin) or electronic nerve stimulation typically target upstream neurocircuits instead of muscle functions. These strategies have intrinsic drawbacks because neurocircuits regulating storing and voiding interact with other neural functions. Treatment could lead to side effects such as unwanted bowel movements or sexual functions. However, optogenetics could precisely express excitatory or inhibitory opsin proteins in the bladder smooth muscles and contract these muscles to empty urine by light. Similarly, the inhibitory opsin variant, such as HR, suppresses overactive balder symptoms, a frequent urinary bladder contraction regardless of voiding.

5.1 Optical control of chimeric antigen receptor

Chimeric antigen receptors (CARs) are engineered recombinant receptors with an antibody-derived ectodomain that recognizes cancer-specific surface protein and intracellular signaling endodomains, including T-cell activation signal CD3ζ and costimulatory molecules for activating T cells or natural killer cells.^{41, 42} The United State Food and Drug Administration (FDA)-approved CAR-T therapy targeting B cell marker CD19 has successfully treated B-cell-related leukemia and lymphoma.⁴³ To increase the effectiveness of CAR-T therapy on solid tumors, the new generation of CARs called ‘TRUCKs’ incorporates inducible IL-12 production to enhance T cell activation, modulate the tumor microenvironment and recruit other immune cells.⁴⁴ However, CAR-T and TRUCK-T therapy face challenges in ‘on-target off-tumor' toxicity and cytokine-associated toxicity, leading to severe tissue damage and cytokine release syndrome with clinical symptoms such as nausea, fever, hypotension, vascular leakage, and life-threatening multiple organ failure.^{43, 44} Anti-inflammatory drugs and corticosteroids are used to manage the side effects of CAR-T therapies⁴⁵; nevertheless, adverse effects from high-dose corticosteroids and the cost of intensive care unit treatment needed for patients with severe symptoms or high disease burden urge better strategies for increasing safety without affecting efficacy.

Optogenetics can increase tissue specificity and alleviate the side effects of CAR-T therapies by providing spatiotemporal modulation. To date, two modes of action enable optogenetic control of CAR—optical induction of CAR expression or recombination of split CAR. The first strategy controls the expression of CAR by the light-inducible gene expression system LINTAD.⁴⁶ (Figure 4A,B) In this system, CRY2 is fused to the transcription activator VP64-p65-Rta (VPR) and kept in the nucleus by nuclear localization signals (NLS). CIB1 is fused to the deoxyribonucleic acid (DNA) binding domain LexA and AsLOV2-caged NLS, which stays cytosolic in the dark. Caging help lower the background expression as the basal level CRY2-CIB1 pair is high.⁴⁶ The expression of anti-CD19 CAR was tested in mice inoculated with tumor cells Nalm-6 and showed effectiveness in limiting the tumor size 21 days after 12-h blue-light illumination.⁴⁶ LINTAD system can achieve transient expression of CAR; for sustained expression of CAR, the system relies on the Cre-LoxP system and degradation of the expressed protein to achieve the temporal control of the expression.⁴⁶

The second strategy, for example, the light-switchable CAR (LiCAR),⁴⁷ uses light to reconstitute split CAR protein between the costimulatory molecule 4-1BB and activation signal CD3ζ through blue light-sensitive CRY2-CIBN or AsLOV2-based iLID⁴⁸ dimerizer system (Figure 4C,D). Combining with upconversion nanoplates, LiCAR reduced the tumour weight in mice with lymphoma after 14-day treatment with near-infrared (NIR) light.⁴⁷ By modulating the activity with light, LiCAR mitigated the cytokine release syndrome and alleviated ‘on-target off-tumour’ effects as the level of IL-6 is lower and the number of B cells higher in mice treated with LiCAR than those treated with traditional CAR-T.⁴⁷ By replacing AsLOV2 in the iLID system with circularly permutated LOV2 (cpLOV2), He and coworkers constructed cp-iLID. They demonstrated its applications in CAR activation by blue-light-inducible heterodimerization of split CAR.⁴⁹ Using a similar design, O'Donoghue and coworkers developed an optoCAR system that allows for precise modulation of the periodic activation of CAR. Intriguingly, expression of the T cell-activation marker CD69 was attenuated with a period of 25 min of light activation (20% duty cycle).⁵⁰ Light activation with a period shorter or longer than 25 min, even with the same integrated light input, resulted in higher CD69 expression, indicating that T cells temporally filter time-varying signals and the CD69 expression is gated through a band-stop filter.

Both strategies offered insights into designing light-inducible CAR and can be expanded by other optogenetic tools. In the split-CAR strategy, the CRY2-CIB1-based system showed less background activity and induced activity than the AsLOV2-based system,⁴⁷ suggesting the choice of optogenetic tool may change the system's efficacy. In addition to splitting at endodomain, photoactivable removal of inhibition at ectodomain is also suggested.⁵¹ Basal activity should be considered in selecting activation approaches. For example, CAR expression with photoactivable Cre-loxP system (TamPA-Cre),⁵² drug-dependent nuclear translocation was selected instead of AsLOV2-caged NLS.

5.2 Optical control of genome editing and gene expression

The Clustered Regularly Interspaced Short Palindromic Repeats or CRISPR-Cas systems correct genomic defects by precise modification of genomic sequence. The system can also be repurposed for transcriptional regulation by using a deactivated or ‘dead’ version of Cas9 nuclease (dCas9). Like optogenetic CAR, optogenetic control of CRISPR-Cas can be accomplished through distinct modes of action, including light-inducible recombination of split-Cas protein,^53-57 Cas expression,^58-62 and photo-uncaging.^{63, 64}

5.2.1 Optical recombination of split-Cas

The first demonstration of optogenetic control of CRISPR-Cas is the light-mediated recombination of split-Cas (Figure 5A,B). Both CRY2/CIB1 and VVD/Magnets pairs have been tested for fusion to N- and C-SpCas9, but only the Magnets-based split-Cas9 (paCas9) achieved decent indel mutations (60% efficiency of the full-length Cas9) and sequence insertion through homology-directed repair, indicating the relative orientation between photoactivatable protein and split Cas9 determines the efficacy of recombination.⁵⁷ The same group also developed light-inducible CRISPR-Cas12a (also known as CRISPR-Cpf1 from Lachnospiraceae bacterium), called paCpf1, by split Cpf1 between residues 730 and 731.⁵⁴

5.3 Optical control of gene expression

5.4 Optical control of mitochondrial dynamics

Mitochondria are membrane-bound organelles generating the majority of chemical energy through adenosine triphosphate (ATP) production. Mitochondria are also mobile organelles that exist in dynamic networks. Mitochondrial diseases are one of the most common types of inherited metabolic disorders, which affect one in 200 individuals, can present at any age, and affect any organ. Although genetic mutations that disrupt mitochondrial gene expression appear to be the most common cause of mitochondrial diseases, disorders of mitochondrial dynamics are emerging as a mechanism of disease.⁷⁵ Optic atrophy spectrum disorder is one such disease that often involves neuron loss in different tissues. A whole-exome sequence experiment revealed that such diseases involve mutation of SLC25A46, a gene that encodes a mitochondrial outer membrane protein. At the cellular level, a significant phenotype is the elonged hyperfused mitochondria, which show a significantly lower oxygen consumption rate and ATP production capacity. Inspired by the finding that lysosomes can regulate mitochondrial fission,⁷⁶ Qiu and coworkers developed optoMLC (mitochondria-lysosome contacts) that allows light-mediated recruitment of lysosomes to mitochondria and enhanced mitochondrial fission. After light stimulation, hyperfused mitochondria in SLC25A46-/- cells were significantly reduced, enhancing oxygen consumption rate and ATP production⁷⁷ (Figure 6).

6.1 Routes of transgene delivery and expression

Common to all gene therapies, optogenetics-based therapy requires the safe delivery and expression of transgenes in humans. Gene therapy has typically used viral vectors, such as adenoviruses, AAV, and lentivirus, to deliver therapeutic genes into patient tissues for long-term expression. Each type of vector has unique strengths and associated safety concerns. Adenovirus is significantly limited in human clinical trials because its poor stable expression span and high immunogenicity. The AAV would be a better choice with more stable expression and low immunogenicity.^78-81

To deliver cargo to the desired tissue, gene therapists use tissue-specific promoters, a strategy applied to target muscle cells,⁸² tumour environments in colorectal cancer⁸³ and prostate cancer,⁸⁴ and neurons.⁸⁵ Such promoters demonstrate cell-specific expression and therapeutic potential, and they allow higher dosing than would be possible with more widely expressed transgenes. The promoter length must be considered for this technique since many delivery methods have relatively low size limits (∼5 kbp for adeno-associated virus).⁸⁶

The primary safety concerns when using these vectors are the risk of viral replication within the patient, viral integration into the patient genome (or integration at undesirable loci), and the patient's immune response, as also reviewed elsewhere.⁸⁶ Optogenetics-based therapies are expected to introduce foreign proteins into the patient's tissue, which might exacerbate the immunological response to viral transduction. Indeed, Maimon and coworkers recently reported that AAV-based ChR2 expression in the peripheral nervous system (PNS) requires co-administering immunosuppressants to prevent neural death.⁸⁷

6.2 Light source and delivery

Optogenetic proteins are sensitive to wavelengths of light from the near ultraviolet (UV) to the NIR. The required light wavelength and power depend on the specific protein(s) used. A combined set of light sources capable of blue (470 nm), orange (560 nm), and infrared (>700 nm) can activate the majority of currently available optogenetic proteins.⁸⁸ Light delivery poses a challenge concerning power requirement and correct positioning. With ample surface areas of tissues like muscle or the brain to be illuminated, demands of powering light sources become a limiting factor. The primary sources of light delivery are optic fibers and optical implants. While wireless optic sources offer opportunities for specific targeting and precise control, the power limit of these devices reduces the operational limit to 3 cm, a depth to be further optimized for human operation.⁸⁹

Most hospitals are already equipped with medical lasers and advanced optics for techniques such as photodynamic therapy.^{90, 91} Many of these light sources are much higher power than necessary for optogenetic therapies and can be easily filtered and coupled to fiber optic cables for tissue delivery. However, new devices must be manufactured and approved when surgical intervention is required. Clinical approval of these devices is a lengthy process, and establishing which regulations apply will require communication with regulatory bodies, especially if clinical trials recruit participants from multiple countries. The safety of surgically implanted devices must be carefully considered. Devices designed to illuminate internal organs with light must operate without generating excess heat, and implanted devices could complicate specific diagnostic techniques such as magnetic resonance imaging (MRI).⁹²

Improvement in light delivery systems can significantly improve the efficacy of optogenetic tools. A recent study focused on using a time-reversed ultrasonically encoded (TRUE)-based light delivery system to accurately focus light and avoid light scattering of the light 89. The TRUE technique could control neural firing rate much more efficiently than conventional light focusing techniques. However, the TRUE system can be ineffective if the system response time exceeds the decorrelation time of the tissue of interest. For example, decorrelation time can drop due to blood flow and muscle movements. Besides, the weight and size of light-delivering implants could significantly impede a patient's daily activities.

6.3 Photo and thermal toxicity from light stimulation

Senova and coworkers⁹³ studied the spatial distribution of light and its effects in large brain volumes of animal models. This study demonstrates that deep light penetration in the brain cortex is possible without detrimental photo and thermal damage to the organ. The authors reported that 1990 seconds of light administration with an irradiance of 100 to 600 mW/mm² and 5 ms pulse duration at 20, 40, and 60 Hz did not trigger non-physiological functional activation. Optimization of stimulation parameters is necessary to avoid potential damage to deep tissues. Assessment of phototoxicity of long-term photo-stimulation requires further investigation.

6.4 Bioavailability

Some photoactivatable proteins require exogenous cofactors, whose bioavailability becomes an essential measure for assessing the efficacy of any light-based therapy. Exogenous cofactors like plant-derived PCB are required for light absorption into various red-shifted photoactivatable proteins such as PhyB. PCB supply must be ensured via transgenic co-expression of PCB biosynthesis genes or by an external supply of synthetic PCB. Various PCB-providing dietary supplements like blue-green algae spirulina could be administered orally. However, the supplement dosage requirement for effective phytochrome activation remains to be validated.⁹⁴ Improved optogenetic tools should minimally depend on external cofactors or use endogenous cofactors in mammals such as biliverdin, FAD, FMN or riboflavin.