Abstract
Nanotechnology involves the creation and use of materials and devices at the size scale of intracellular structures and molecules. The aim of nanomedicine is to gain fundamental access to properties and behaviors at the molecular scale to achieve the comprehensive monitoring, control, construction, repair, defense, and improvement of human biological systems at the molecular level, using engineered nanodevices and nanostructures, operating massively in parallel at the single-cell level, performing “single-cell medicine,” ultimately to achieve medical benefit. Diverse applications of nanotechnology currently include: delivery of drugs, peptides, and genes; biomedical imaging; minimally invasive physiological monitoring; theranostics; prosthetics (including optogenetics and bionic retina); and cell-based therapy. Challenges to incorporation of nanotechnology in ophthalmology include: persistence of nanoparticles, biofouling, and safe manufacturing techniques.
Key words
bionic retina, extracellular matrix, nanoengineering, nanomedicine, nanoparticle, nanotechnology, physiologic monitoring, prosthetics, regenerative medicine, theranostics
Introduction
Nanotechnology provides an important new set of tools for the diagnosis and treatment of ocular diseases. Miniaturization of devices, chip-based technologies, and novel nanosized materials and chemical assemblies already provide novel tools that are contributing to improved healthcare in the 21st century and will impinge directly on ophthalmology. In this chapter, we review general principles of nanotechnology and nanomedicine as well as properties of nanomachines. We also consider specific and potential applications of nanotechnology to ophthalmology, including drug, peptide, and gene delivery; imaging; minimally invasive physiologic monitoring; prosthetics; regenerative medicine; and surgical technology. Finally, we consider obstacles to incorporation of nanotechnology into ophthalmology. Each of these topics has been reviewed in detail previously.
General Principles of Nanotechnology and Nanomedicine
Nanotechnology
Nanotechnology involves the creation and use of materials and devices at the size scale of intracellular structures and molecules. The systems and constructs deployed typically are on the order of <100 nm. (Recall that an average man is 1.6 meters [1.6 billion nanometers (nm)] tall; an erythrocyte is 7 µm [7 × 10 3 nm] wide; and a strand of deoxyribonucleic acid [DNA] is 2 nm wide.) Transformational opportunities in information storage, computation, and mechanical efficiency are available through nanotechnology.
Regarding information storage, Richard Feynman, who is credited with conceiving the field of nanotechnology, calculated that it was possible to write the entire 24 volumes of the Encyclopedia Britannica on the head of a pin. If one did not simply etch the letters on to the surface of the pin but used the interior of the material also, he calculated that one could fit all the information that humans had accumulated up to December 1959 (estimated at 10 15 bits) in a cube of material 1/200 inch wide, comparable to the size of a piece of dust. Today, through nanotechnology-based precision assembly of matter, storage densities of 10 11 bits per cm 2 have been demonstrated, which closely approximates Feynman’s vision. Efficient information storage is crucial for the complexity of biological systems, as each eukaryotic cell stores an enormous amount of information. A retinal pigment epithelial (RPE) cell has a diameter of approximately 4 × 10 –4 inches (1 × 10 –3 cm), and each cell stores the blueprint to create an entire human in DNA molecules (3 billion chemical basepairs, ~25,000 genes).
Regarding computation, Feynman also noted that biological systems do not simply store information, they create measurable outputs. The human brain has the capacity to make judgments, e.g., recognize a person’s face (even if shown at different distances, under different lighting conditions, at different angles), or play chess. Feynman reasoned that if computers could have as many computational elements as our brains, they could make judgments as well. Today sophisticated facial recognition can be accomplished with a powerful laptop computer (versus poorly in 1959, with a much larger computer), due to the development of microprocessors and sophisticated software. Defeating a grand master at chess, however, requires a supercomputer. A Cray XT5 supercomputer uses ~40 kW power/cabinet, and each cabinet measures ~81 × 23 × 57 inches (larger than a refrigerator) and weighs ~1530 lb (694 kg) ( http://www.cray.com/downloads/CrayXT5Blade.pdf ). It is remarkable that the “computer” in our cranium does not require the amount of rare elements, generate the heat, or have the energy requirements of a supercomputer. Thus, the evolution of our cognitive capacities from infancy to adulthood (derived from the interaction between a DNA template-guided, manufactured neuronal network and the external environment) is one demonstration that it is possible to develop nanoscale mechanical systems that create complex, measurable outputs.
Nanomachines are highly efficient. When organized in massively parallel structures, for example, nanomotors can generate large forces (e.g., muscles that move massive animals such as whales) or large electrical currents (e.g., those generated by the Hunter’s organ of electric eels). Nanomotors also can direct delicate processes such as ion transport and chromosomal migration during mitosis. Nanomachines are not only highly efficient, they typically have long operational half-lives and are mass-produced easily.
Nanomedicine
The aim of nanomedicine is the comprehensive monitoring, control, construction, repair, defense, and improvement of human biological systems at the molecular level, using engineered nanodevices and nanostructures, operating massively in parallel at the single-cell level, performing “single-cell medicine,” ultimately to achieve medical benefit. Integration of nanoscale technologies with the practice of medicine will alter profoundly our approach to the diagnosis, treatment, and prevention of disease. We will begin to diagnose and treat diseases at the single-cell level, for example, rather than just at the organ level.
General principles of nanotechnology as applied to nanomedicine include:
- 1.
Biomimicry: the approach that cells use to direct molecules within a cell and/or direct molecules/machines to the proper cells in the body.
- 2.
Size and location drive biocompatibility and biological efficacy.
- 3.
Engineer feedback control into therapeutic systems (e.g., therapeutic gene synthesis).
- 4.
Molecules as machines: engineer molecules to perform specific physical tasks, such as opening ion channels, to alter cell and organism behavior.
- 5.
“Pseudointelligence” resulting from intelligent design, e.g., self-assembly of extracellular matrix (ECM) molecules.
- 6.
Highly interdisciplinary undertaking: development of nanotechnologies typically involves expertise in biology, engineering, chemistry, and physics.
The functional properties of living systems arise not only from their component parts, but also from how these parts are assembled, which dictates interactions between the parts, the nature and flow of information within the system, and the outputs that the system produces. Thus, one concept from biology that may be important for development of nanomachines in medicine is that spatial control of the distribution of nanomachines directly affects the efficiency of the macromolecular assembly and nature of this assembly’s work product. Spatial control can be achieved through the use of membranes and anchoring molecules that place enzymes and substrates in proximity (e.g., as occurs in the endoplasmic reticulum for synthesis of ECM proteins, in the mitochondrial membrane for electron transport, and on the cell surface for ECM ligand integrin-mediated changes in intracellular signaling). This concept is exploited in the area of neural prosthetics, as described later in this chapter. Conversely, this engineering approach also permits segregation of molecules (e.g., segregation of lysosomal enzymes from the cytoplasm). Microelectromechanical systems (MEMS) or nanoelectromechanical systems (NEMS)-based techniques can be used to create engineered scaffolds (see next paragraph) that achieve this spatial control. One might also use a synthesized lipid bilayer or block co-polymer membranes to create artificial organelles as occurs in nature.
One can produce nanomachines by assembling naturally occurring ones. M/NEMS-based engineering, however, permits the construction of small devices using computer-aided design and by the repeated application of a number of procedures, including oxidation, photolithography, etching, diffusion, sputtering, chemical vapor deposition, ion implantation, and epitaxy, as illustrated by the devices described later in this chapter. Control of features down to the submicron level permits production of mechanical structures at length scales ranging from 100 nm or less to greater than 1 cm. The ability to create complex microfabricated biomaterial substrates using these techniques enables one to define surface microarchitecture, topography, and feature size. By engineering the microenvironment, one can control individual cell responses utilizing structures at the micro- and nanoscale to alter cellular attachment and motility, attenuate the foreign-body response, simulate tissue organization, and promote cell differentiation.
A straightforward application of nanomachines to nanomedicine is the use of error-prone polymerase chain reaction and transposons to nanoengineer novel structures (e.g., viral capsids) as vectors for gene and drug delivery (as discussed later in this chapter). A somewhat less obvious application of nanomachines exploits their capacity for energy transduction. Motor proteins, for example, transduce the following forms of energy: (1) optical and electroosmotic; (2) optical and electronic; (3) chemical and electroosmotic; (4) chemical and electronic; (5) mechanical and electroosmotic; and (6) mechanical and chemical. By exploiting the transducing properties of nanomotors, one might be able to develop nanomachines that measure biological variables: oxygen tension, pH, glucose concentration, the redox state of a cell or subcellular organelle, temperature, intraocular pressure (IOP), or blood pressure. Alternatively, one can use nanomachines to induce energy sensitivity where it does not normally exist, e.g., creation of light-sensitive retinal ganglion cells (RGCs) or bipolar cells (described later in this chapter) or to provide therapeutic molecules to cells, e.g., wild-type alleles of retinal proteins to blind mutants (described later in this chapter). Finally, one might couple the measuring and therapeutic capacity of nanomachines to create devices that measure critical biological variables (e.g., redox state) for diagnostics and provide therapy (e.g., antioxidant therapy) at the appropriate time and place and in the appropriate quantity, a concept termed theranostics (discussed later in this chapter).
Properties of Nanomachines
Physical Properties
Feynman noted that, at the size scale of intracellular structures and molecules, materials acquire seemingly surprising properties that are predictable based on the principles of quantum physics. For instance, carbon becomes stronger than steel; gold melts at room temperature, and aluminum becomes highly explosive. Quantities such as weight and inertia are of relatively little importance.
Inhomogeneity of materials (e.g., metals versus plastics) might limit their utility. Magnetic properties on a very small scale are not the same as on a large scale. Lubrication is not needed if the machine is small enough because heat loss is rapid (due to the large surface area : volume ratio). Some of these properties produce unexpected results. For example, rapid heat loss might prevent gasoline from exploding, which would make a nanoscale internal combustion engine impossible.
The influence of gravity on the function of true nanomachines is negligible because their mass is that of atoms ( F g = Gm A m B / r 2 ). On the other hand, the distance between the elements is in nanometers. Because of van der Waals forces, parts of nanomachines might adhere to each other, which might not be desirable. Electrical resistance may be very large with nanocircuits, a feature that might be useful in biological systems. Another problem with nanocircuits is the inverse relationship between the size of the device and the amount of noise generated (Hooge’s rule). The development of stacked graphene sheets may provide a solution to this problem and facilitate the development of circuits much smaller than those in conventional silicone-based computer chips. However, when working on the scale of atoms, circuits might not be needed, and quantized energy levels might be manipulated for energy transfer according to the laws of quantum mechanics. This property is exploited in some nanoparticle-based imaging technologies.
Axial ratio is the ratio of the magnitude of the axes (e.g., length and width in the case of a two-dimensional object), the greater being divided by the lesser. Nanotubes and nanofibers have very large axial ratios. Synthetic methods exist to alter the length-to-volume ratio and thus the physical properties of the material. Conductivity, for example, is enhanced as the diameter of semiconducting nanotubes is reduced.
A key property of nanomaterials is that they are surface-rich objects in relation to volume. The following analogy illustrates this concept. A baseball filled with membrane surface-active enzymes would have only a minimal number of these extending to the surface of the ball, so the vast majority would be inactive. A pea filled with enzymes would have a greater proportion extending to the surface. A sphere 100 nm in diameter would have the majority of the enzymes it contains in an active state. This property results in a situation in which there is relatively limited unutilized surface area on to which one might attach molecules to the nanoparticle surface. The net effect is to create structures that accelerate many chemical reactions at their surfaces and act in some ways like very active enzymes.
Vacancy-engineered mixed-valence state cerium oxide (CeO 2 ) nanoparticles (nanoceria) illustrate the useful properties that materials can develop at the nanoscale. Alteration in the oxidation state of CeO 2 creates defects in its lattice structure via loss of oxygen or its electrons. As their size decreases, nanoceria (3–5 nm diameter) demonstrate formation of more oxygen vacancies in the crystal structure. As described later in this chapter, vacancy-engineered nanoceria may function as a highly effective treatment for ocular conditions associated with oxidative damage.
Manufacture
At each step in the process of manufacturing smaller and smaller machines, one must improve the accuracy of the equipment. Feynman speculated that if devices are built on the scale of 5–10 atoms, then it should be possible to mass-produce them such that they are perfect copies of each other. One useful outcome of a “nanomanufacturing” process is that the material costs of billions of these machines would be minimal since each is so small that minimal material is used. Nanomachines can have the capacity to repair and build themselves. Indeed, most nanoparticle structures are self-assembling under the proper thermodynamic conditions, allowing for production of a large number of virtually identical nanostructures.
Some properties of nanomachines are illustrated by the design and manufacture of an axon surgery platform. Using microtechnology, electrokinetic axon manipulation (i.e., dielectrophoresis), and cell fusion (i.e., electrofusion), Sretavan et al. developed a paradigm of direct axon repair involving the substitution of damaged axon regions with healthy segments from donor axons. This multidisciplinary group developed a multifunctional axon surgery platform that is ~1 mm 3 ( Figs. 38.1 and 38.2 ). The cutting device consists of a silicon nitride knife with an ultrasharp knife-edge mounted on to a silicon-based compliant knife suspension ( Fig. 38.1 ). The knife edge’s radius of curvature (~20 nm) is similar to the diameter of a single microtubule. Because the knife is manufactured from a silicon nitride membrane, it is nearly transparent, which permits visualization of axons during the cutting procedure. The mechanical compliance of the suspension can be varied to deliver sufficient force for cutting different tissues (e.g., single axons) or for harvesting specific cell populations from histologic tissue sections. The authors envision future improvements such as sensors as well as force-generating actuation mechanisms that automatically deliver a controlled cutting stroke, and they indicate that both piezoelectric and thermal expansion actuation mechanisms can deliver forces in the range needed for axon cutting. A femtosecond laser might also be used for axotomy. The goal is to develop a microcutting device with on-board sensing and actuation that can function as a semiautonomous instrument, requiring only initiating commands from the surgeon. Important limitations to the practical application of this invention remain. The authors estimate, for example, that ~20 seconds are required to repair a single axon using dielectrophoresis and electrofusion. Cutting and fusing multiple axons simultaneously might enable relatively rapid repair of several thousand axons.
Applications to Ophthalmology
Nanomedicine will foster revolutionary advances in the diagnosis and treatment of disease. Nanomedicine is likely to have a major impact on biopharmaceuticals (e.g., drug delivery, drug discovery), implantable materials (e.g., tissue regeneration scaffolds, bioresorbable materials), implantable devices (e.g., IOP monitors, glaucoma drainage valves ), and diagnostic tools (e.g., infectious disease diagnosis, genetic testing, imaging, IOP monitoring). Nanotechnology will bring about the development of regenerative medicine (i.e., replacement and improvement of cells, tissues, and organs), ultrahigh resolution in vivo imaging, microsensors and feedback devices, and artificial vision. “Regenerative nanomedicine,” a new subfield of nanomedicine, uses nanoparticles containing gene transcription factors and other modulating molecules that allow reprogramming of cells in vivo.
Delivery of Drugs, Peptides, and Genes
General Considerations Regarding Nanoparticles
Nanoparticles are colloidal carrier systems that can improve the efficacy of drug delivery by overcoming diffusion barriers, permitting reduced dosing (through more efficient tissue targeting) as well as sustained delivery ( Fig. 38.3 ). These features are attractive for drug treatment of chronic conditions such as glaucoma, uveitis, or retinal edema (due to venous occlusion or choroidal neovascularization [CNV]) as well as for treatment of intraocular tumors and other conditions associated with cell proliferation such as capsular fibrosis after cataract surgery, ocular neovascularization, and proliferative vitreoretinopathy. Nanoscale-engineered cell substrata (e.g., nanowires) and carbon nanotubes also can be used for gene and drug delivery.
Strategies in the design of nanoparticles for therapeutic purposes have been reviewed thoroughly by Petros and DeSimone, Moghimi et al. and Kompella et al. Particle size, shape, and surface properties influence nanoparticle biodistribution. Particle size, for example, affects whether the particle is internalized via phagocytosis, macropinocytosis, caveolae-mediated endocytosis, or clathrin-mediated endocytosis, which in turn results in exposure of the nanoparticle to different intracellular environments. The cell surface receptor, nucleolin, transports compacted polylysine DNA nanoparticles into cells and directly to the nucleus.
One can target nanoparticles to specific cells by attaching to the particle surface ligands/antibodies/peptides/aptamers for receptors/molecules that are abundant on the surface of the target cell/tissue. This approach can have complications. Receptor aggregation on the cell surface, for example, can induce unintended events, such as apoptosis. One can engineer the nanoparticle for a particular mode of intracellular entry depending on the choice of nanoparticle targeting molecules, e.g., cholesterol favors uptake via caveolin-mediated endocytosis, and transactivating transcriptional activator peptide favors macropinocytosis. Nanoparticle surface chemistry also can be manipulated to trigger cargo release under specific circumstances. For example, when exposed to a reducing environment such as is present in the cytosol, reductively labile disulfide-based crosslinks between the carrier and cargo are broken. Approaches for targeting nanoparticles to particular subcellular organelles, e.g., mitochondria or the nucleus, also have been developed.
Liposomes and polymer–drug conjugates are among the most frequently used nanoparticles for therapeutic purposes. Liposomes, which carry hydrophobic or hydrophilic cargo, can be coated with ligands that direct them to specific cell surface receptors for cell targeting as well as with polymers that prolong their half-life in the circulatory system. Poly(ethylene glycol) (PEG) can be conjugated with different molecules to enhance their solubility and stability in plasma and to reduce immunogenicity. Opsonization by immunoglobulin and/or complement proteins can lead to recognition of the nanoparticle as foreign and induce a hypersensitivity reaction. Coating a nanoparticle with albumin and/or PEG can create a hydrophilic surface that temporarily resists protein adsorption, thus prolonging the particle’s bioavailability. This approach allows for much longer drug circulation and concomitant lowering of therapeutic-level drug doses, which in turn can reduce many unintended side-effects.
Dendrimers are synthetic, highly branched polymers that have precisely controllable nanoscale scaffolding and nanocontainer properties, which in some senses mimic the properties of macromolecules such as DNA and ribonucleic acid (RNA). The diameter of poly(amidoamine) dendrimer ranges from 1.5 to 14.5 nm. As generation (G) number increases, the number of active terminal groups doubles. G3 dendrimers, for example, contain 32 terminal groups, and G4 dendrimers contain 64 terminal groups. In poly(amidoamine) dendrimers, full generations (e.g., G3) have terminal amine or hydroxyl groups while half-generation dendrimers (e.g., G3.5) have carboxylic acid terminal groups. Dendrimers have been explored as vehicles for controlled drug delivery, including cancer therapy, pilocarpine, gatifloxacin, and for vascular endothelial growth factor (VEGF) inhibition. Marano et al., for example, used a lipophilic amino acid dendrimer to deliver an anti-VEGF oligonucleotide into rats’ eyes with laser-induced CNV. The dendrimer–oligonucleotide conjugate inhibited CNV development for 4–6 months by up to 95%, whereas eyes injected with oligonucleotide alone showed no treatment benefit compared to vehicle-injected controls at these times. The dendrimer–oligonucleotide conjugate was well tolerated in vivo. Ideta et al. used dendrimer porphyrin encapsulated by a polymeric micelle to treat laser-induced CNV in rodents and found significant enhancement of photodynamic therapy efficacy with less light energy required for CNV occlusion.
Antibiotic Therapy
Typically, only a small fraction (<5%) of topically administered medications is biologically available due to limited ocular penetration and rapid clearance from the aqueous humor. Because dendrimers contain surface functional groups as well as void spaces within and between their branches, they can serve as delivery vehicles for therapeutic modalities such as carboplatin. Dendrimeric polyguanidilyated translocators (DPT) are nanosized dendrimers that translocate molecules across biological barriers efficiently. Durairaj et al. used a six-guanidine group-containing dendrimer to enhance gatifloxacin solubility (fourfold) and delivery to the anterior and posterior segment of rabbits. The DPT–gatifloxacin complexes (346 nm) enhanced tissue concentration in the conjunctiva (13-fold) and cornea (twofold). A single dose resulted in sustained aqueous humor levels ( t 1/2 = 9 hours), potentially allowing decreased frequency of administration (e.g., once-daily dosing). After multiple dosing, DPT–gatifloxacin achieved therapeutic levels in the vitreous humor for 12 hours (versus no drug levels detectable at 12 hours after topical gatifloxacin alone).
Antimetabolite Therapy
Shaunak et al. used anionic, polyamidoamine, generation 3.5 dendrimers to make novel water-soluble conjugates of d (+)-glucosamine and d (+)-glucosamine 6-sulfate with immunomodulatory and antiangiogenic properties, respectively. Dendrimer glucosamine inhibited Toll-like receptor 4-mediated lipopolysaccharide-induced synthesis of proinflammatory chemokines (i.e., macrophage inflammatory protein (MIP)-1α, MIP-1β, interleukin (IL)-8) and proinflammatory cytokines (i.e., tumor necrosis factor-α, IL-1β, IL-6) primarily from immature human monocyte-derived dendritic cells and monocyte-derived macrophages, but allowed upregulation of the costimulatory molecules CD25, CD80, CD83, and CD86. Dendrimer glucosamine 6-sulfate blocked fibroblast growth factor-2-mediated human umbilical vein endothelial cell proliferation, but not VEGF-mediated proliferation, and neoangiogenesis in human Matrigel and placental angiogenesis assays. When dendrimer glucosamine and dendrimer glucosamine 6-sulfate were used together in a validated, clinically relevant rabbit model of scar tissue formation after glaucoma filtration surgery, they increased the long-term success of the surgery from 30% to 80% ( p =.029). A clinical trial of this modality to reduce scarring after trabeculectomy, however, was not successful (P Khaw, R Ritch, personal communication).
Neurotrophic Factor Therapy
Nanoparticles can deliver growth and neurotrophic factors to cells. Intravitreal nanoparticle-based basic fibroblast growth factor (bFGF) delivery, for example, provides sustained retinal rescue in Royal College of Surgeons (RCS) rats. In RCS rats, the RPE cells have a mutation that prevents proper outer-segment phagocytosis, with secondary rod and cone photoreceptor degeneration. Some patients with retinitis pigmentosa (RP) have this same mutation. Sakai et al. prepared bFGF nanoparticles using gelatin isolated from bovine bone collagen and human recombinant bFGF. The nanoparticle diameter, assessed using dynamic light scattering, was ~585 nm.
Glaucoma, a leading cause of blindness worldwide, is associated with progressive RGC death and optic nerve atrophy. Intravitreal glial cell line-derived neurotrophic factor (GDNF)-loaded biodegradable (poly)lactic-co-glycolic acid (PLGA) microspheres provide sustained RGC protection in a rodent model of glaucoma. Microspheres (~8 µm diameter) containing GDNF were fabricated using a modification of a spontaneous emulsion technique. Since adeno-associated virus (AAV)-mediated GDNF secretion from glia delays retinal degeneration in a rat model of RP, it is possible that nanoparticle-mediated GDNF delivery can be applied to treating RP-like diseases.
Optic nerve injury (crush) has been treated using nanoparticles to deliver neurotrophic factors. Kim et al. demonstrated that human serum albumin nanoparticles (150 nm diameter) conjugated with brimonidine released brimonidine for 5 days and improved retinal ganglion cell density after injury by almost 400% compared to control 14 days after optic nerve injury in rats. Brimonidine was loaded into the hydrophobic pockets of the human serum albumin nanoparticles. The nanoparticles seemed to reduce amyloid-beta deposition in the retinal ganglion cell layer. In view of the pore size of the internal limiting membrane (10–20 nm) and the outer limiting membrane (3–4 nm), it seems likely that the nanoparticles penetrated the retina by a receptor mediated endocytosis mechanism, possibly the transforming growth factor-beta receptor.
Antioxidant Therapy
Age-related macular degeneration (AMD), RP, diabetic retinopathy, and retinopathy of prematurity are characterized, in part, by the presence of oxidative damage. As noted above, alteration in the oxidation state of CeO 2 nanoparticles creates defects in its lattice structure via loss of oxygen or its electrons. Chen et al. posited that engineered nanoceria can scavenge reactive oxygen intermediates because the large surface area-to-volume ratio at 5 nm diameter enables CeO 2 to regenerate its activity and thereby act catalytically. (Unlike nanoceria, most free-radical scavengers require repetitive dosing.) Chen et al. showed that intravitreal injection of nanoceria prevents light-induced photoreceptor damage in rodents, even if injected after the initiation of light exposure. Vacancy-engineered nanoceria also inhibit the development of and promote regression of pathologic retinal neovascularization in the Vldlr knockout mouse, which carries a loss-of-function mutation in the very low-density lipoprotein receptor gene and whose phenotype resembles a clinical entity known as retinal angiomatous proliferation ( Figs. 38.4 and 38.5 ). This regression occurs even if nanoceria are injected intravitreally after the mutant retinal phenotypes are established ( Fig. 38.6 ). Because nanoceria are a catalytic and regenerative antioxidant, a single injection has a prolonged effect (measured in weeks). Nanoceria inhibition of increased VEGF levels in this model may mean that CeO 2 nanoparticles will be effective in treating macular edema in diabetic eyes and CNV-induced retinal edema in AMD eyes.
C-60 fullerenes are cage-like structures (truncated icosahedron) of carbon atoms with antioxidant properties. Malonic acid C-60 derivatives (carboxyfullerenes) can eliminate superoxide anion and H 2 O 2 , and inhibit lipid peroxidation. Systemic administration of the C-3 carboxyfullerene isomer delayed motor deterioration and death in a mouse model of familial amyotrophic lateral sclerosis. It might be useful in the treatment of retinal diseases associated with oxidative damage.
Iron is an essential element for enzymes involved in the phototransduction cascade, in outer-segment disc membrane synthesis, and in the conversion of all- trans -retinyl ester to 11- cis -retinol in the RPE. Free Fe 2+ catalyzes the conversion of hydrogen peroxide to hydroxyl radical, which is a highly reactive species that causes oxidative damage (e.g., lipid peroxidation, DNA strand breaks). Increased intracellular iron causes oxidative photoreceptor damage. Polymeric nanoparticles can be used to chelate metals. Liu et al. showed that a chelator-nanoparticle system complexed with iron, when incubated with human plasma, preferentially adsorbs apolipoprotein E and apolipoprotein A–I, which should facilitate transport into and out of the brain via mechanisms used for transporting low-density lipoprotein. Iron accumulation in the RPE and Bruch’s membrane is greater in AMD eyes than in controls, including cases with early AMD and late stages of the disease (i.e., geographic atrophy, CNV). Some of this iron is chelatable. Although it is not proven that iron overload is a cause of AMD, iron chelation might have a therapeutic effect. Thus, the technology developed by Liu et al. might have utility for treating AMD eyes.
Immune-Suppressive Therapy
Cell-based therapy might be sight-restoring for patients with degenerative retinal diseases such as RP and AMD. An immune response to transplanted cells may depend upon the cell types included in the cellular therapy. Nanotechnology provides immune-suppressive therapy (local or systemic) in selected cases where its need is anticipated. In preclinical models, for example, nanoparticles are helpful in managing corneal allograft rejection. Yuan et al. manufactured 300-nm-diameter rapamycin-loaded chitosan/polylactic acid nanoparticles and demonstrated that they extended median allograft survival by 17% in rabbits compared with aqueous rapamycin eye drops. Topical chitosan particles were well tolerated in this study, but intraocular chitosan nanoparticles may not be well tolerated.
Studies of experimental autoimmune uveoretinitis (EAU) demonstrate that nanoparticles can be used to modulate the inflammatory response in the retina and choroid. EAU is a T-cell-mediated autoimmune disease that targets the retina and related tissues and serves as a model for human autoimmune ocular diseases. Nanosuspensions of relatively insoluble glucocorticoids (developed using a high-pressure homogenization method) enhance the rate and extent of drug absorption as well as the intensity and duration of drug action, compared with conventional solutions and microcrystalline suspensions. Rats with EAU clear poly(lactic acid) (PLA) nanoparticles rapidly from the systemic circulation. As noted above, PEG can be used to modify the surface of the nanoparticles, which reduces opsonization and interactions with the mononuclear phagocyte system. Sakai et al. prepared polymeric nanoparticles with encapsulated betamethasone phosphate. These nanosteroid particles (~120 nm diameter) were composed of PLA homopolymer and a block copolymer of PEG. In vivo imaging of inflamed eyes of rats with EAU demonstrated greater nanoparticle accumulation and higher betamethasone concentration in eyes of PLA-PEG nanoparticle-treated rats versus PLA nanoparticle-treated rats. PLA-PEG nanosteroid-treated EAU rats also had lower clinical and histopathologic scores for ocular inflammation. The stronger therapeutic effect of PLA-PEG nanosteroids versus PLA nanosteroids may be due to prolonged blood circulation and sustained release in situ as well as due to targeting to inflamed eyes (the latter effect resulting from the small diameter of the nanoparticles). EAU also responds very well to intravitreal liposomal tacrolimus (mean diameter = 200 nm) with no side-effects on retinal function or systemic cellular immunity.
Gene Therapy
Nonviral Vectors
Viral vectors deliver genes efficiently but can be associated with risks such as immunogenicity and insertional mutagenesis. Nonviral vectors (e.g., polymers, lipids) and other methods (e.g., electroporation, nucleofection) have high gene-carrying capacity, low risk of immunogenicity, relatively low cost, and, possibly, greater ease of production. Nanoparticles can deliver genes efficiently to stem cells and have been explored as a means for gene delivery in the diagnosis and treatment of ocular disease. As viruses do, nanoparticles can use transactivating sequences that allow them to deploy the host cell machinery to manufacture therapeutic molecules in situ. Because these sequences can contain an upstream biomolecular control sensor, therapeutic molecules can be manufactured in situ under tight feedback control.
Electrostatic interaction of cationic polymers with negatively charged DNA/RNA molecules results in condensation of the material into particles ranging from 8 to 500 nm in diameter, protection of the genes from enzymes, and mediation of cellular entry. Complexes of cationic polymers and plasmid DNA, termed polyplexes, can have transfection efficiency comparable to adenoviral vectors. In addition to nanometer size, polyplexes have large vector capacity, are stable in nuclease-rich environments, and can have relatively high transfectivity for both dividing and nondividing cells. For example, nanoparticles compacted with a lysine 30-mer linked to 10 kDa PEG-containing cytomegalovirus-cystic fibrosis transmembrane conductance regulator (CMV-CFTR) cDNA were used successfully in a phase I/II clinical trial for the treatment of cystic fibrosis. Some particles, however, have low transfection efficiency, and the duration of gene expression can be short. When it occurs, toxicity is related to nanoparticle chemistry.
To some degree, compacted DNA nanoparticles can be targeted to different tissues in the eye through selection of an appropriate injection site (e.g., intravitreal injection can target the cornea, trabecular meshwork, lens, and inner retina; subretinal injection can target the outer retina and RPE). Nanoparticle size and charge influence migration through the vitreous cavity. Farjo et al. demonstrated that after subretinal injection of compacted lysine 30-mer DNA nanoparticles, gene expression is observed throughout the retina and not just at the site of the injection. By choosing cell-specific promoters, one can achieve additional specificity in the locus of gene expression. The rhodopsin promoter, for example, drives expression in rod photoreceptors, and the human red opsin promoter drives expression in cone photoreceptors. Interphotoreceptor retinoid-binding protein drives expression in both rods and cones. The vitelliform macular dystrophy promoter drives expression in RPE cells.
Cai et al. used a specific formulation of DNA nanoparticles consisting of single molecules of DNA compacted with 10 kDa PEG-substituted lysine 30-mer peptides containing the wild-type retinal degeneration slow ( Rds ) gene, peripherin/rds, to induce cone photoreceptor rescue in an animal model ( rds +/– ) of RP. After injection into the subretinal space, these particles did not induce a detectable immune response, cytotoxicity, or disruption of retinal function. These compacted plasmid DNA nanoparticles are small (8–20 nm), have rod or ellipsoid shape (depending on the counterion used), and have a large carrying capacity (at least up to 20 kilobases). PLGA nanoparticles can deliver genes to RPE cells in vitro and in vivo relatively efficiently and safely, and PLGA DNA nanoparticles can be associated with long-term gene expression. PLGA DNA nanoparticles tend to be larger than polylysine DNA nanoparticles, which may affect cellular uptake mechanism and delivery to the nucleus. PLGA DNA nanoparticles might be used to deliver therapeutic genes for conditions associated with RPE gene mutations, e.g., Best disease and a form of Leber congenital amaurosis.
Albumin has a highly charged amino acid content, which facilitates its action as a carrier for charged drugs and oligonucleotides. Albumin-derived nanoparticles that deliver plasmids containing genes for the Flt receptor (VEGFR1) which binds free VEGF penetrate keratocyte cytoplasm, and provide sustained inhibition of injury-induced corneal neovascularization.
Despite these promising results, concerns involving nanoparticle use remain. Although the immune response to polylysine-based nanoparticles seems to be less than that for capsid proteins, for example, the efficiency of gene transfer is not as high since most are degraded in the endosomal complexes. As a result, one may generate an immune response because one must use large numbers of nanoparticles to achieve therapeutic useful transfection. Also, the apparent low immunogenicity observed in murine models of RP may not be observed in human patients because the immune response to both nanoparticles and viruses varies from one species to another.
Viral Vectors
Critical issues for successful gene therapy include (1) vector uptake, transport, and uncoating; (2) vector genome persistence; (3) sustained transcriptional expression; (4) the host immune response; and (5) insertional mutagenesis and cancer. Virus-based gene therapy can induce immune responses, including innate, humoral, and cell-mediated, that are directed against the vector and/or the transgene product. Primary humoral responses directed against the vector can limit its capacity to deliver genes to the target cells as well as the ability to readminister the virus to the patient (e.g., when treating the fellow eye with a second surgical procedure). An immediate innate immune response and a secondary antigen-dependent response to intravenous administration of recombinant adenoviral vectors, for example, caused death in a patient with ornithine transcarbamylase deficiency. A humoral response against the transgene product may neutralize the therapeutic protein. A cell-mediated immune response against the vector or transgene product can eliminate the transduced cells. Two patients with hemophilia B, for example, developed vector dose-dependent transaminitis that limited hepatocyte-derived factor IX expression to less than 2 months due to CD8 + memory T cells that recognized AAV serotype 2 (AAV2) capsids and eliminated AAV2-transduced hepatocytes. The innate immune response can cause local and/or systemic toxicity and amplify a secondary antigen-dependent immune response. The likelihood of an immune response is influenced by the dose of viral particles, which in turn is influenced by the efficiency of vector uptake and gene expression, as well as by the specificity of targeting. If dendritic cells or antigen-presenting cells take up the vector, for example, an immune response is more likely.
Nanoengineering of the viral capsid and transgene may provide a means to solve some of these problems. Recombinant AAVs (rAAVs) have been used successfully to treat preclinical models of human ocular disease and also have been used to treat humans with Leber congenital amaurosis. Modifications of the virus to improve clinical effectiveness illustrate some of the nanoengineering strategies that have been employed in this area. AAVs are small (4.7 kilobase carrying capacity), nonpathogenic, single-stranded DNA parvoviruses that can transduce dividing and nondividing cells. The capsid is critical for extracellular events related to the recognition of specific receptors, which influences cell tropism, as well as intracellular processes involving AAV trafficking and uncoating. In turn, the latter processes influence transduction kinetics and transgene expression efficiency. Due to previous exposure to various AAV serotypes, a significant proportion of the population harbors neutralizing antibodies that can block gene delivery. Because administration of low doses of viral vector might mitigate the severity of this problem, two nanoengineering techniques have been used to improve vector cellular tropism, transduction efficiency, and immunogenicity: directed evolution and site-directed mutagenesis. These are discussed below. Other nanoengineering devices (e.g., DNA transposons, bacteriophage recombinases ) may provide clinically useful means to achieve stable, safe DNA integration in the host genome and sustained transgene expression in the future.
Directed evolution of AAV capsids has generated vectors that are highly resistant to neutralizing antibodies. Maheshri et al. used error-prone polymerase chain reaction and a staggered extension process to generate an AAV2 library (>10 6 independent clones) with randomly distributed capsid mutations and then used high-throughput approaches (i.e., exposure of mutants to heparin affinity chromatography [wild-type AAV2 binds to heparan sulphate] or repeated amplification of AAV2 mutants that retain infectivity in the presence of serum containing neutralizing antibodies) to identify mutant AAV2 capsids with altered receptor-binding properties and the capacity to bind with very low affinity to neutralizing antibodies. This approach can be quite powerful. One mutagenesis and three selection steps generated mutant capsids, for example, with a threefold improved neutralizing antibody titer (versus wild-type capsid) and a ~7.5% infectivity at serum levels that completely neutralized wild-type infectivity. Directed evolution has been used to generate AAV variants that transduce Müller cells after intravitreal injection, which may provide a means to deliver growth factors to photoreceptors and RPE cells. These growth factors retard the progression of retinal degeneration in preclinical models of RP and possibly in human patients also.
Zhong et al. demonstrated that site-directed mutagenesis of surface-exposed tyrosine residues increases vector transduction efficiency 30-fold in vivo at one log lower vector dose compared to wild-type AAV2. The increased transduction efficiency is due to impaired capsid ubiquitination and improved intracellular trafficking to the nucleus. (Epidermal growth factor receptor protein tyrosine kinase [EGFR-PTK] signaling impairs AAV2 vector transduction by impairing nuclear transport of the vectors; EGFR-PTK can phosphorylate AAV2 capsids at tyrosine residues, and tyrosine-phosphorylated AAV2 vectors enter cells efficiently but do not transduce well, in part because the AAV capsids are ubiquitinated and then degraded by the proteasome. ) Thus, the T-cell response to AAV2 capsids seems to be manageable by using surface-exposed tyrosine mutant vectors. Another rate-limiting step in transduction efficiency, the conversion of single-stranded viral genome to double-stranded AAV DNA, has been overcome by deleting the terminal resolution site from one rAAV inverted terminal repeat, which prevents replication initiation at the mutated end, to generate self-complementary AAV (scAAV) vectors. (AAV has a tendency to package DNA dimers when the replicating genome is half the length of the wild type.)
Ocular Applications
Due to their relatively low immunogenicity, ability to target many nondividing cells, and capacity for sustained efficient therapeutic gene expression after a single treatment, rAAV vectors have been used to treat preclinical models of human retinal disease. Site-directed mutagenesis technology has been used to improve the treatment of degenerative retinal disease in these preclinical models. Vectors containing point mutations in surface-exposed capsid tyrosine residues in AAV serotypes 2, 8, and 9 display strong and widespread transgene expression in retinal cells after intravitreal or subretinal delivery. Petrs-Silva et al. demonstrated that tyrosine-to-phenylalanine capsid scAAV2 mutants showed much greater transduction efficiency (10–20-fold higher transgene expression) of the entire retina (including photoreceptors) after intravitreal injection compared to scAAV with wild-type capsids ( Fig. 38.7 ). Mutants of scAAV2, scAAV8, and scAAV9 also enhanced transduction of RGCs compared to wild-type AAV2 (e.g., 10 6 -fold reduction in the number of virus particles needed for RGC transfection with mutant scAAV2 compared to wild-type AAV2). Intravitreal delivery may offer an important clinical advantage over subretinal delivery. Subretinal virus delivery, which has been used in clinical studies, requires pars plana vitrectomy in the operating room and has a higher likelihood of complications (e.g., retinal tear) than intravitreal delivery, which can be done in an office under topical anesthesia. On the other hand, the subretinal space is a relatively immune-privileged site, which may reduce the likelihood of an immune response after repeat virus treatment. Li et al. demonstrated that a humoral immune response against AAV2 capsid proteins occurs after intravitreal but not after subretinal vector delivery. Subretinal injection of one of the mutant scAAVs also transduced Müller cells. These studies demonstrate two strategies for reducing the immune response to viral vectors via site-directed mutagenesis: increasing transduction efficiency, which permits lower doses of vector, and creation of multiple effective serotypes, which can be used sequentially for subsequent therapy.
Imaging
The use of nanomaterials for biomedical imaging has been reviewed. In addition to drug delivery, polyamidoamine dendrimer prototypes may be used as targeted diagnostic magnetic resonance imaging (MRI) contrast agents. Gold nanoparticles have been used to enhance tumor identification with computed tomography as well as for colorimetric biosensing. Colorimetric biosensing is based on the fact that the plasmon resonance frequency is a function of the average distance between the gold particles as well as their size, shape, and the dielectric properties of their environment. A plasmon is a quantum of plasma oscillation and can be viewed as the quantization of the oscillation of free electron density against the fixed positive charges in a metal. Gold nanoparticles (5 nm), for example, are orange-red but turn blue-purple when aggregated to larger nanoparticles. Colloids of gold nanoparticles impart vibrant colors to the stained-glass windows of some Gothic churches. If a gold nanoparticle–receptor complex is bound to a crosslinking molecule, nanoparticle clustering can occur with an associated colorimetric change. Superparamagnetic iron oxide (SPIO) nanoparticles have been approved for use by the US Food and Drug Administration as MRI contrast agents. Typically, an SPIO nanoparticle is composed of an iron oxide core coated with dextran, which renders it water-soluble. For in vivo long-term imaging, the label is internalized by endocytosis or phagocytosis with or without an excipient. These particles can have diameters of 60–150 nm and can be visualized using MRI. SPIO-labeled stem cells have been visualized in patients with brain trauma. Limitations of this approach include the dilution of signal associated with cell replication and/or migration. Also, no information is provided regarding the state of differentiation of the transplanted cells. Cytotoxicity (e.g., via iron-catalyzed generation of reactive oxygen species) might be a limitation of this approach. Quantum dots (Qdots) are light-emitting nanocrystals (2–10 nm) composed of atoms from groups II–VI (e.g., CdSe, ZnSe) or III–V (e.g., InP) of the periodic table. In contrast to SPIO nanoparticles, Qdots can be visualized with optical imaging (versus more complex MRI), including optical coherence tomography. In contrast to most organic dyes and fluorescent proteins, Qdots have durable fluorescence intensity, which helps one to distinguish the signal from background autofluorescence, and a broad excitation/narrow emission spectrum, which permits analysis of multiple cell targets with a single excitation wavelength. Qdots can be used to monitor survival, distribution, and differentiation of stem cells in vivo. Limitations can include signal dilution (due to cell proliferation and/or migration) as well as cytotoxicity. In some cases, toxicity might arise from oxidative degradation of Qdots with subsequent Cd release and mitochondrial damage. Coupling SPIO nanoparticles or Qdots to antibodies that recognize molecules such as components of the complement system or molecular constituents associated with AMD-induced changes in Bruch’s membrane might provide a means to image the biochemical and/or structural abnormalities associated with AMD. This capacity may help one assess the effectiveness of AMD treatments that target early molecular changes of the disease.
Nanowires are structures with a diameter of less than 100 nm and indefinite length. Semiconductor nanowires have unique electronic properties and sizes comparable with biological structures involved in cellular communication, thus making them promising nanostructures for establishing active interfaces with biological systems. Cells can be grown on nanowire arrays, which can measure electrical functions in different parts of the same cell. Ophthalmic applications might include monitoring RGC survival/physiology in glaucoma patients or photoreceptor/RPE survival/physiology in transplant recipients.
Minimally Invasive Physiologic Monitoring
We have reviewed this issue extensively and recapitulate these observations here. Continuous measurement of critical biophysical properties can give insight into disease pathogenesis and the efficacy of a given treatment modality. An assembly of such monitors could be useful for telemedicine and remote patient monitoring, particularly for patients with chronic diseases such as glaucoma (e.g., detect elevated IOP, progressive optic nerve atrophy), diabetic retinopathy (e.g., detect macular edema, retinal neovascularization), and AMD (e.g., detect subretinal fluid/retinal edema associated with CNVs). Ideally, the technology would permit repeated (if not continuous) noninvasive monitoring of preselected biomarkers. These platforms should function with minimal power (e.g., rather than incorporating a battery, utilize continuously available power sources) and provide accurate, precise information over an extended period of time (e.g., years). Recall that IOP is a dynamic quantity that fluctuates from moment to moment and diurnally (higher while asleep). It may be that nocturnal IOP (and blood pressure) measurement is more critical to glaucoma management rather than the daytime measurement performed in an outpatient setting. When IOP is measured over 24 hours, the results often lead to a change in glaucoma management. Thus, continuous, minimally invasive IOP measurement could be an important innovation in glaucoma management. Finally, intraocular monitoring devices must be small because there is little unutilized space within the eye.
IOP measurement typically is done with a Goldmann tonometer. This device does not measure IOP directly but instead measures the force required to applanate a corneal surface whose circular area is 7.35 mm 2 ([3.06/2] 2 π). As a result, changes in corneal thickness as well as changes in corneal tension at different IOPs (Laplace’s law for a thin-walled sphere: T = ( P × R )/2 h , where T = tension in the wall, P = pressure difference across the sphere wall, R = sphere radius of curvature, and h = thickness of the wall) can affect the accuracy of IOP measurements done with the Goldmann tonometer.
A number of nano-based IOP monitoring systems have been developed, but since this chapter is focused on retinal disease, they will only be briefly described. One noninvasive approach to IOP monitoring involves the use of a wireless contact lens. Leonardi and coworkers developed a disposable silicone soft contact lens with an embedded sensor.
An alternative approach that may be better suited for more invasive monitoring relies on capacitive pressure sensors (versus strain gauges).
Chen et al. developed a passive, biocompatible, micro-machined pressure sensor, based on the concept of a Bourdon tube (a thin-walled tube with an elliptical cross-sectional shape that can measure pressure quite accurately).
Challenges to the approaches mentioned above include short range (due to low signal-to-noise ratio), limited stability (e.g., due to variable mechanical contact with the cornea), high profile (despite use of MEMS technology), and high manufacturing costs. Rizq et al. developed a piezoresistive IOP sensor that is implanted into the suprachoroidal space. The device is micro-machined and is fabricated with associated electronics (interface circuit, radiofrequency powering, and reverse telemetry) with full onboard electronics to provide active readout and superior signal-to-noise ratio, range, and accuracy. This device measured IOP accurately in human cadaver eyes. Dresher and Irazoqui designed a compact, ultra-low-power operational amplifier that can be used to record IOP. This CMOS operational amplifier can be incorporated with a wireless IOP monitoring system. It has a power consumption of 736 nW, chip area of 0.023 mm 2 , and output impedance of 69 Ω to drive low-impedance loads. The authors envision implantation of the device into the suprachoroidal space and have designed and fabricated a high-frequency transmitter integrated circuit that has sufficient signal-to-noise ratio margin for a high data rate transmission wirelessly. There is a possibility for choroid/retinal damage with suprachoroidal insertion and a likely need for a method to secure the device (e.g., sutures, glue) to ensure reproducible measurements. In addition it is not clear what the maximum distance between the sensor and the receiver antenna will be, which might place logistical constraints on the use of the device for continuous IOP monitoring.
Coupling Diagnostics and Therapeutics
Theranostics
Theranostics refers to a process in which diagnosis of a disease state, individualized for a particular patient (even to particular cells within a patient), is coupled with therapy that is targeted precisely in its amount, nature, and location. Prow et al., for example, developed a biosensor DNA tethered to a magnetic nanoparticle. The biosensor uses an enhanced green fluorescent protein (EGFP) reporter gene driven by an antioxidant response element (ARE). The ARE is activated by oxidative stress and enhances the expression of genes downstream to it. Exposure of the cells to hyperoxia drives the expression of EFGP. This engineered nanoparticle penetrates endothelial cells (preferentially dividing cells), and after subretinal injection, these biosensor nanoparticles report the activation of the ARE in diabetic rat RPE. The antioxidant biosensor could provide a means for clinicians to identify patients likely to need therapy (e.g., babies with retinopathy of prematurity who will need laser photocoagulation or other treatment) at a time before clinical manifestations of severe disease are evident. By coupling a therapeutic gene (e.g., catalase, peroxidase, superoxide dismutase) to the ARE (in addition to a reporter gene such as EGFP), one creates a combined diagnostic–therapeutic device that enables endothelial cells (or any cells that take up the nanoparticle) to “treat themselves” in the setting of oxidative damage. Therapeutic possibilities are limited primarily by one’s knowledge of the pathways involved in disease pathogenesis. One could couple genes to manage angiogenesis, for example pigment epithelial-derived factor, and to serve as a neurotrophin in eyes with AMD-associated CNV and geographic atrophy. Attractive features of this approach to gene therapy are (1) one uses the cell’s endogenous metabolic machinery to drive the expression of exogenous genes that will enhance the cell’s capacity to manage environmental stress; and (2) individual cells titer their own therapeutic enzyme/drug concentration in relation to their “toxic” exposure and endogenous capacity to respond to environmental stress. The concept and some experimental results are shown in Fig. 38.8 .
Prosthetics: Molecules as Machines (e.g., Light-Sensitive Ion Channels), Abiotic–Biotic Interfaces
Induced Photosensitivity
The use of molecules as machines will revolutionize neural prosthetics. The application of this concept to induced photosensitivity has been reviewed in detail elsewhere. Here we recapitulate this analysis. Although rewiring of inner retinal circuits and inner retinal neuronal degeneration occur in association with photoreceptor degeneration in RP, it is possible to create visually useful percepts by stimulating RGCs electrically. Use of light-sensitive ion channels, rather than electrodes, to stimulate RGCs provides an alternative approach to retinal cell stimulation. Induced light sensitivity has the potential for noninvasive neuronal stimulation with high spatial resolution.
Channelopsin-2 is a light-gated ion channel derived from green algae and is sensitive to blue light. When its attached chromophore, all- trans retinaldehyde, undergoes reversible photoisomerization, channelopsin-2 undergoes a conformational change that alters its permeability to mono- and divalent cations. In contrast to mammalian rhodopsin, which loses its chromophore after 11- cis retinal-all- trans retinal isomerization, channelopsin-2 remains attached to its chromophore after all- trans to 11- cis retinal isomerization. Thus, there seems to be no need to provide chromophore to the transduced cells. The complex of channelopsin-2 and all- trans retinal is termed channelrhodopsin-2 (ChR2). Using a rAAV delivery system (rAAV serotype-2) in rd1 mice, which have a null mutation in a cyclic GMP phosphodiesterase (PDE6b), and in RCS rats, which have a mutation in the tyrosine kinase, Mertk, ChR2 expression can be achieved in inner retinal neurons (primarily ON and OFF RGCs). Each of the latter mutations causes some form of RP in humans. ChR2 converts these neurons into cells that respond to light with membrane depolarization. In addition, intraocular injection of rAAV2-ChR2 can restore the ability of the animals to encode light signals in the retina and transmit them to the visual cortex.
Recombinant AAV2 vectors have been used to deliver channelopsin-2 to retinal cells (primarily ganglion and amacrine cells) in wild-type adult mice stably for up to 18 months. Up to 20% of ganglion cells were infected (with normal morphology), and a sufficient number of functional ChR2 channels were maintained to drive robust ganglion cell membrane depolarization and spike firing in response to light. These investigators estimated that, at high viral concentrations, approximately 40% of all A-II amacrine cells were labeled. In mammals, rod signals are related through rod bipolar cells to A-II amacrine cells. These signals are coupled on to ON and OFF cone pathways by gap junctions and glycinergic synapses, respectively. Thus, the ability to target A-II amacrine cells with this vector may enable recovery of both ON and OFF light responses in RP retinas. Restoration of both ON and OFF pathways probably will be important for achieving good contrast sensitivity and proper spatial and temporal signal processing. Recombinant AAV2-mediated transfection of retinal neurons in nonhuman primates (marmoset) via intravitreal injection results in functional expression of ChR2 in all retinal neurons, but preferentially ganglion cells (all major types). Regional variations in transfection efficiency seemed to correlate with the thickness of the inner limiting membrane. This potential barrier for rAAV2-mediated intravitreal gene delivery could, in principle, be overcome by internal limiting membrane peeling, a standard technique in vitreoretinal surgery, or by enzymatic digestion of the inner limiting membrane.
Recently a new optogenetic tool has been bioengineered. Opto-mGluR6 is a chimeric all-retinal G-protein-coupled-receptor (GPCR) consisting of the intracellular domains of the ON-bipolar cell-specific metabotropic glutamate receptor, mGluR6, and the light-sensing domains of melanopsin. Melanopsin, which also belongs to the GPCR protein family, is a blue-light-sensitive retinal photopigment present in a subpopulation of photosensitive retinal ganglion cells that control the pupillary light reflex as well as the circadian rhythm by signaling to the suprachiasmatic nucleus. The light-induced isomerization of this retinal chromophore is reversed while bound to the opsin. This property makes melanopsin highly resistant to bleaching by strong light and allows successive light activation without a response rundown. Melanopsin is activated by moderate daylight as opposed to high intensity light (e.g., bright sunlight on a snow-covered field) required for activating the ion channel ChR2. The mGlu6 receptor normally mediates light responses in ON-bipolar cells by coupling glutamate signals from the photoreceptors to TRPM1 cation channels. Because it is a GPCR, mGluR6 greatly amplifies glutamate signal via the intracellular G-protein coupled second messenger cascade. These features enable opto-mGluR6 coupling to bipolar cell-specific preexisting G-protein complexes, which include regulators of G-protein signaling essential for fast signal kinetics. When illuminated, opto-mGluR6 causes cellular hyperpolarization as it does in situ . At light intensities that just begin to stimulate ChR2 rAAV (e.g., 10 15 photons/cm 2 /sec), opto-mGluR6 rAAV exhibits a maximal conductive response.
Zhang et al. showed that expression of halorhodopsin (HaloR), a yellow light-activated chloride ion pump from halobacteria, in inner retinal neurons converts them into OFF cells. In these experiments, HaloR was ~20-fold less sensitive to light than ChR2. HaloR and ChR2 coexpressing cells can produce ON, OFF, and ON–OFF responses, depending on the illumination wavelength. Experiments in these preclinical models indicate that kinetics of ChR2- and HaloR-mediated light responses are compatible with temporal processing requirements of visual information in the retina. A current limitation of this approach is that ChR2 and HaloR both exhibit low light sensitivity, with threshold activation light intensities ~5–6 log units higher than those of cones. Furthermore, the light intensity operating range of microbial rhodopsins is 2–3 log units, compared to normal retinal dynamic range of 10 log units. Doroudchi et al. achieved stable, specific expression of ChR2 in ON bipolar cells using a rAAV vector packaged in a tyrosine-mutated capsid. Light levels that elicited visually guided behaviors were within the physiologic range of cone photoreceptors. There was no evidence of induced inflammation or toxicity. As indicated above, signal convergence from bipolar cells on to RGCs may mean that targeting ChR2 to rod bipolar cells will provide increased light sensitivity as well as higher spatial resolution, but this approach may be compromised by the alterations in synaptic circuitry that accompany photoreceptor degeneration.
Greenberg et al. reconstructed an excitatory center and antagonistic surround by targeting humanized ChR2 to the somata and enhanced HaloR to the dendrites of RGCs ( Figs. 38.9 and 38.10 ). This approach to the deployment of optical neuromodulators retains crucial information processing (edge detection) while being independent of the state of inner retinal circuit remodeling during degeneration. Fusion of the humanized ChR2 to ankyrin G polypeptide localized this opsin to the soma and proximal dendrites because ankyrins couple sodium channels to the spectrin-actin network. Fusion of enhanced HaloR to PSD-95 protein targeted this opsin to the dendritic regions in RGCs. As a result, Greenberg and coworkers nanoengineered RGCs with differential spatial and spectral photosensitivity. Depending on which opsin is fused to ankyrin G and which to PDS-95, both ON and OFF-center ganglion cells could be created. Because this approach generated nonphysiologic center surround dimensions, Greenberg et al. preprocessed the visual image with gaussian blurring, such that when convolved with the dimensions of the soma and dendrites, the gaussians approximated the relative dimensions of the ganglion cells’ center and surround receptive fields. Thus, imaging processing enabled extraction of edge information. These data and the above considerations indicate that ChR2/HaloR-based RGC prosthetics will require image preprocessing to perform light amplification, dynamic range compression, and local gain control operations.
