Next generation therapeutics for retinal neurodegenerative diseases

Abstract

Neurodegenerative diseases affecting the visual system encompass glaucoma, macular degeneration, retinopathies, and inherited genetic disorders such as retinitis pigmentosa. These ocular pathologies pose a serious burden of visual impairment and blindness worldwide. Current treatment modalities include small molecule drugs, biologics, or gene therapies, most of which are administered topically as eye drops or as injectables. However, the topical route of administration faces challenges in effectively reaching the posterior segment and achieving desired concentrations at the target site, while injections and implants risk severe complications, such as retinal detachment and endophthalmitis. This necessitates the development of innovative therapeutic strategies that can prolong drug release, deliver effective concentrations to the back of the eye with minimal systemic exposure, and improve patient compliance and safety. In this review, we introduce retinal degenerative diseases, followed by a discussion of the existing clinical standard of care. We then delve into detail about drug and gene delivery systems currently in preclinical and clinical development, including formulation and delivery advantages/drawbacks, with a special emphasis on potential for clinical translation.

Graphical Abstract

1. Introduction

Retinal neurodegenerative diseases are a leading cause of irreversible vision loss worldwide [1]. In glaucoma, prolonged elevation of intraocular pressure (IOP) leads to progressive death of ocular neurons, known as retinal ganglion cells (RGCs) [2]. In neovascular age related macular degeneration (AMD), leakage of fluid from blood vessels damages the retinal pigment epithelium (RPE), the cells that absorb ultraviolet light and scavenge reactive oxygen species, leading to loss of photoreceptors [3,4]. In diabetic retinopathy (DR) and diabetic macular edema (DME), hyperglycemic stress similarly leads to vascular leakage, RPE damage, and photoreceptor loss [5]. Inherited retinal degenerative diseases, such as retinitis pigmentosa (RP), are caused by genetic mutations in pathways that lead to the death of rod and cone photoreceptors [6].

Treatment of retinal diseases may include small molecule drugs, proteins, or gene therapies administered by various routes, including topical administration, peri- and intraocular injections and implants, or systemic administration. Of all approved ophthalmic therapies, over 75% are administered topically, and nearly 80% are small molecule therapeutics [7]. The ease of use and convenience of topical therapies is countered by the low intraocular bioavailability that requires high dosing frequency and related issues with patient compliance [8,9]. Further, topical eye drops and creams are largely used for diseases of the anterior segment. Approved treatments for diseases of the posterior segment include therapies that are delivered via intraocular injection or implants [7]. Intraocular injections subvert ocular barriers and deliver drugs directly to target tissue but require repeated physician visits for treatments and come with increased risk of infection and other complications. Surgical implants provide prolonged release of drug locally near the retina and require fewer physician visits, however, require more invasive insertion procedures and potential issues with excipient accumulation [10]. Systemic administration is less commonly used for treating the posterior segment, as it comes with increased risk of off-target side effects and results in lower accumulation of drug in the retina compared to local delivery.

The challenges associated with achieving safe and effective treatment of retinal neurodegenerative diseases have motivated the development of novel drug delivery methods to provide sustained therapeutic effects by overcoming ocular barriers to achieve increased efficacy or reduce toxicity. Herein, we describe recent advancements in the development of nanomedicine formulations of various kinds (nanoparticles, thermogels and hydrogels, liposomes, niosomes, viral and nonviral vectors), engineered materials (contact lenses, surgically implantable devices, and injectable devices, among others), and cell-based therapies administered topically or via intravitreal, intracameral, suprachoroidal, subretinal, or subconjunctival injection (Figure 1) [11–14]. We further highlight currently approved treatments for retinal neurodegenerative diseases, novel formulations which are currently undergoing clinical trials, and promising preclinical ocular delivery strategies with demonstrated translational potential (well-characterized systems with pharmacokinetic data and/or efficacy in relevant animal models). Advances in delivering small molecule and biologic therapies will be discussed, and in the case of RP, cell and gene-based therapies will be discussed. Many therapies covered herein evaluate efficacy and confirm intraocular bioavailability in an animal model with a larger eye, a step that is particularly necessary for topical treatments among others as a next step in translation. The ocular structure and morphology of rabbit and monkey eyes is much more like humans than those of smaller rodents [15]. Rabbits have a similar eye size and internal structure, with slightly smaller aqueous and vitreous humor volumes [16,17]. However, due to lower blink rates, corneal residence time and intraocular bioavailability may be altered compared to human, and these differences must be considered when determining clinical potential [17]. Though typically less accessible due to cost among other factors, non-human primates are often considered the closest model system to humans due to shared anatomy and genetic similarity [18,19].

Figure 1.

Ocular delivery routes and drug and nucleic acid formulation strategies. Adapted from [20] with permission and generated using BioRender.com.

2. Routes for achieving therapeutic delivery to the retina

2.1. Topical Administration

The primary modality for ophthalmic drug administration is topically instilled drops, largely due the ease of administration [21]. However, due to the low intraocular bioavailability of topically dosed drugs, many products must be administered multiple times per day to achieve the intended therapeutic effect. Repeated dosing increases systemic exposure, as up to 80% of a topical dose may reach systemic circulation via the conjunctiva and nasolacrimal drainage [22]. Pre-corneal and corneal factors primarily limit intraocular bioavailability of topically administered therapies, such as the tear film and blinking. The tear film is a mucus layer that not only protects the eye from pathogens, but dilutes the concentration of topically administered drugs and facilitates clearance from the ocular surface, reducing corneal residence time [23,24]. The use of penetration enhancers or mucoinert coatings, such as surfactants, polymers, or chelating agents, is a strategy for improving intraocular drug bioavailability by reducing clearance effects. By altering the integrity of the tear film and increasing the permeability of corneal epithelial cell tight junctions, increased drug accumulation is possible, though with the risk of cumulative toxicity [25]. Regardless, the fraction of the therapeutic that typically reaches the retina after topical administration is low, and transport may occur via three widely purported routes, including the corneal route, the transscleral/periocular route, and the uveoscleral route (Figure 2) [26,27]. Achieving therapeutic delivery to the retina via conventional eye drops has proven extremely challenging, and thus, there has yet to be much commercial success with topical eye drops for treating retinal diseases, though the development efforts continue [14].

Figure 2.

Schematic depicting different routes of diffusion of topically applied drugs to the posterior segment of the eye. Figure retrieved with permission from [27].

2.2. Peri- and intraocular injection and implants

The current standards for treatment of diseases of the posterior segment include intraocular injections and drug-releasing implants. Penetrating the globe with a needle or incision circumvents many anatomical and physiological barriers to topical delivery. Thus, therapeutically relevant drug concentrations can be reached in the retina with reduced risk of systemic side effects that may occur with topically administered therapies [22]. Intravitreal injections, particularly of anti-vascular endothelial growth factor (VEGF) biologics, are routinely performed as frequently as monthly [28]. However, there is a minor risk of severe complications, such as retinal detachment and endophthalmitis, leading the push towards longer-lasting injectable treatments [29]. Subconjunctival injections have been employed clinically for the administration of corticosteroids and antibiotics while providing high drug levels in both the anterior and posterior chamber [30,31]. As they are less invasive than intravitreal injections, there is further potential for application in the treatment of retinal neurodegenerative disease. More recently, subretinal injections of an adeno-associated virus-based gene therapy have been employed as a treatment for specific genetic mutations in patients with RP [14]. Subretinal injections provide access to the outer layers of the retina but are associated with increased risk of complications compared to intravitreal injections [32,33]. Suprachoroidal injections, which are generally considered less invasive while providing therapeutic delivery to the choroid and potentially the retina, are also being explored preclinically and in clinical trials involving small molecules, biologics, and engineered microneedles [34,35].

To reduce the frequency of dosing required, surgically implantable devices can provide sustained release to maintain therapeutic effect for longer intervals. While sustained release devices have the potential to improve patient adherence and clinical outcomes, there are potential drawbacks of requiring an invasive surgical procedure, as well as dose dumping and accumulation of excipient materials [36]. The polymer used may be either biodegradable or non-biodegradable, though non-biodegradable implants require a second procedure for removal. However, biodegradable implants require precise balancing of the polymer degradation properties and the drug loading to limit burst and/or non-uniform release of drug [37]. An alternative option that was recently approved for use in AMD treatment is a refillable implant [38,39].

2.3. Systemic Administration

Oral administration is by far one of the most common and convenient approaches, though achieving effective ocular drug delivery via the systemic route is very challenging [40]. To achieve efficacy, large doses are required, leading to potential off-target effects, side effects and drug toxicity. The blood-retina barrier (BRB) limits passage of a wide variety of molecules into the retina [40–42]. The BRB is comprised of similar structural features in the RPE, choroidal vasculature (choriocapillaris) and Bruch’s membrane. These form the outer layer, which, in conjunction with the inner barrier within retinal capillaries, restricts direct entry of molecules into the retina [43,44]. There are limited cases in which drugs are administered systemically for treating ocular diseases, such antimicrobials for the treatment of endogenous endophthalmitis or open-globe injury [45]. Of note, many systemic drug therapies, like corticosteroids and antineoplastic agents, administered for non-ocular indications are known to have ocular side effects, including secondary glaucoma [46,47].

3. Glaucoma

3.1. Background

Glaucoma was estimated to affect 80 million people worldwide in 2020 and is the leading cause of irreversible blindness [48,49]. Glaucoma is a group of progressive chronic diseases characterized by elevated intraocular pressure (IOP), damage to the optic nerve head, and RGC death [50]. The two main types of glaucoma are primary and secondary. Secondary glaucoma can be caused by surgical procedures, medications, and underlying medical conditions. For example, topical steroid usage can cause secondary glaucoma and associated symptoms [51]. Primary glaucoma is further divided into two main types: primary open angle glaucoma (POAG) and angle closure glaucoma [52]. POAG is the most common type of glaucoma, accounting for 75% of all glaucoma cases globally [53]. POAG and angle closure differ in their pathology that leads to ocular hypertension. In POAG, ocular hypertension is caused by decreased aqueous humor outflow due to resistance in the trabecular meshwork (Figure 3). This occurs gradually over time and often without the awareness of the patient until vision loss occurs. Angle closure glaucoma occurs more rapidly, as the drainage angle between the cornea and iris is suddenly closed by the lens pushing on the iris, resulting in a physical blockage of aqueous humor outflow. Finally, normotensive primary glaucoma, in which the IOP stays within normal range, is known to occur, but the pathology of RGC damage is unknown [50]. While the primary risk factor of POAG is ocular hypertension, IOP-independent risk factors such as vascular dysregulation are thought to be more important to the progression of normotensive primary glaucoma [54–56]. Herein, we will focus on the current standard of care and the emerging treatment options for both POAG and normotensive primary glaucoma.

Figure 3.

The typical pathway of aqueous humor outflow. Aqueous humor originates in the ciliary body and flows through the pupil to the trabecular meshwork and Schlemm’s canal, where it flows out. In the uveoscleral pathway, the aqueous humor takes the alternate pathway of flowing out through the ciliary muscle. Image reused with permission from [57].

3.2. Currently Approved Treatments

The current standards of care for the treatment of glaucoma involve lowering IOP through pharmacologic or surgical means [54–56]. The three main types of glaucoma medications are beta adrenergic antagonists, alpha adrenergic agonists, and prostaglandin analogs [58], all of which are administered primarily as eyedrops. While these medications typically reduce ocular hypertension in POAG when used as directed, issues with adherence and side effects necessitate new and more convenient treatment options.

3.2.1. Beta Adrenergic Antagonists

Beta adrenergic antagonists, or beta blockers, were once the first-line treatment for glaucoma [59,60]. Beta blockers work to reduce aqueous humor production by blocking nerve endings in the ciliary epithelium. This is done by either blocking both the β₁and β₂ adrenoreceptors (nonselective) or blocking only the β₁ adrenoreceptor (cardioselective). Nonselective beta blockers include timolol, metipranolol, levobunolol, and carteolol, while the only cardioselective beta blocker is betaxolol. However, beta blockers have both local and systemic side effects. The local side effects include ocular irritation and conjunctivitis, while the systemic side effects are more severe, including asthma, bradycardia, and systemic hypotension [61]. The systemic side effects are slightly reduced when using betaxolol due to its specific activity. In addition, beta blockers require administration up to two times a day [59].

3.2.2. Alpha Adrenergic Agonists

Alpha agonists decrease aqueous humor production by vasoconstriction of ciliary vessels [62]. The two main alpha agonists used for glaucoma treatment are brimonidine and apraclonidine. Apraclonidine is limited to short-term use after laser surgeries due to frequent allergies and short-term effects [58,63]. Brimonidine, in addition to reducing aqueous humor production, can also increase uveoscleral outflow [63]. Brimonidine results in longer-lasting effects and is therefore more frequently prescribed to patients with glaucoma for long term therapy. However, alpha agonists also have systemic and local side effects, including central nervous system depression, bradycardia, and hypotonia [58,63]. Local side effects include uveitis, contact dermatitis, conjunctivitis, and hyperemia, while systemic side effects include dizziness, fatigue, and hypotension.

3.2.3. Prostaglandin Analogs

Prostaglandin analogs, including latanoprost, bimatoprost, tafluprost, and travuprost, have emerged as the first line of treatment for glaucoma due to various advantages over beta blockers and alpha agonists. Prostaglandin analogs bind to receptors in the trabecular meshwork and Schlemm’s canal, causing muscle relaxation and increased aqueous outflow [64,65]. One major advantage of prostaglandins is that they can be administered once per day [58]. In addition, prostaglandins have minimal systemic side effects aside from headaches, though there are numerous local side effects, including conjunctival hyperemia, iris pigmentation, and eyelash growth. Notably, a sustained release implant containing bimatoprost, Durysta, was approved in 2020. Durysta consists of a biodegradable poly(lactic-co-glycolic acid) (PLGA) matrix and is inserted intracamerally to provide IOP reduction for up to four months [66]. However, Durysta cannot be reinjected since the implant does not completely degrade within the four months. Most recently, the FDA approved iDose, a removable implant designed to release travoprost, following promising Phase 3 results [67]. Inserted into the anterior chamber, iDose is composed of a titanium implant filled with 75 μg of a proprietary formulation of travoprost designed to be released via membrane-controlled diffusion. Phase 3 results showed significant IOP reduction for up to 3 months post-administration.

3.3. Current Outlook of Preclinical Research

There are a variety of topical and injectable sustained delivery systems that have been developed and tested preclinically for the treatment of glaucoma. To test potential therapeutics, a large range of animal models are employed, including normotensive animals and animals with pharmacologically or mechanically induced hypotension and/or damage to the RGCs and optic nerve. Many studies investigate IOP reduction in normotensive models [68–71]. In models of ocular hypertension, IOP elevation may be achieved using steroids [72], by injecting materials such as carbomer solution [73] or microparticles to physically block the trabecular meshwork [74], by injecting hypertonic saline into the episcleral vein [75], or lasering the trabecular meshwork [76]. In addition to evaluating IOP reduction, ocular hypertension is used to evaluate neuroprotection, as prolonged hypertension leads to progressive RGC death [75]. To evaluate neuroprotection independent of IOP, optic nerve crush can be used [77,78]. Promising preclinical studies utilizing a variety of administration routes are outlined below.

3.3.1. Topical Administration

The gold standard treatment option for glaucoma is a variety of eyedrops. Despite their popularity, the low bioavailability and rapid clearance of eyedrops limit their potency for lowering IOP. Therefore, many preclinical studies focus on approaches that increase ocular residence time, increase drug solubility, and/or increase drug penetration [79]. The use of various viscous hydrogel materials aims to increase ocular surface residence time, allowing for increased drug absorption [79]. Cheng et al. investigated chitosan-based hydrogels encapsulating latanoprost [72]. To prepare the hydrogel, chitosan and type A gelatin were dissolved in acetic acid. Glycerol 2-phosphate disodium salt solution was added to the resulting solution dropwise while on ice, followed by the addition of latanoprost. When the gel was characterized, the sol-gel temperature was found to be 34.2 ± 0.7°C, while the gelation time was 70.7 ± 7.2 seconds. In addition, scanning electron microscope (SEM) images showed the morphology of the gel to consist of a network of pores. A cell viability assay demonstrated that the hydrogel did not cause measurable cytotoxicity. Finally, an in vitro release study showed the hydrogel released a maximum of 50% of the drug, reaching this threshold at day 3. Upon installation in rabbits with triamcinolone acetonide-induced ocular hypertension, once weekly administration of the hydrogel was shown to reduce IOP comparable to latanoprost eyedrops for approximately 15–20 days [72].

Another approach that has been explored is incorporating nanoparticles into hydrogels for topical administration. Sun et al. developed a dual-control release system consisting of brimonidine encapsulated layered double hydroxide (LDH) nanoparticles embedded into a thermosensitive hydrogel [71]. To prepare the nanoparticles, MgCl₂, AlCl₃, and brimonidine were added to a solution of either NaOH (Bri@LDH (0)) or plus L-tartrate (Bri@LDH). After the solution was stirred, the nanoparticles were collected via centrifugation and lyophilization. To load the nanoparticles into a PLGA-polyethylene glycol (PEG)-PLGA gel, the amphiphilic copolymer was first dissolved in phosphate-buffered saline (PBS), after which the Bri@LDH nanoparticles were added, and the resultant solution was stirred. Characterization of the nanoparticles showed the diameter of the Bri@LDH (0) to be 86 nm, while the diameter of the Bri@LDH nanoparticles was 100 nm. In addition, Bri@LDH showed a higher loading of brimonidine compared to Bri@LDH (0), with drug loading of 25 μg/mg and 13.5 μg/mg, respectively. Based on this, Bri@LDH was loaded into the thermogel and selected for in vivo testing. The sol-gel transition temperature of the loaded thermogel was 38.6°C, while that of the blank thermogel was 37.3°C. To test the in vitro release, Bri@LDH powder was combined with PBS in a shaker and 0.5 mL aliquots were removed to be tested via high-performance liquid chromatography (HPLC) at different time points. To test the Bri@LDH/thermogel release, the hydrogel was formed and placed in a vial with 5 mL of PBS. Aliquots were removed for testing at different time points. These studies showed that while the nanoparticles alone exhibit a burst release, (reaching 100% release within 2 hours of the beginning of the test) the incorporation of the thermogel significantly extended the release of brimonidine, with over 80% being released over approximately 7 days. In rabbits, the Bri@LDH/thermogel was applied and an external contact lens was applied over the top to improve the spread. As a control, brimonidine eyedrops were applied. While a single dose of brimonidine eyedrops lasted about 30 hours, application of the Bri@LDH/thermogel plus contact lens reduced IOP for approximately 7 days [71].

Xu et al. developed a micelle-contact lens system loaded with both timolol and latanoprost [74]. To prepare the micelles, timolol base, latanoprost, and methoxy-PEG₂₀₀₀-polylactic acid₂₄₀₀ (mPEG₂₀₀₀-PLA₂₄₀₀) were dissolved in acetonitrile. After a uniform layer was formed under vacuum rotary evaporation, the micelles were hydrated with deionized water and filtered. To prepare the contact lens, hydroxyethyl methacrylate (HEMA) was combined with the micelle suspension and ethylene glycol dimethacrylate (EGDMA). The resulting solution was purged with nitrogen and photoinitiator 1173. Finally, the solution was poured into a mold and polymerized. The mean size of the micelles was 21 ± 6.1 nm, and the loading of timolol and latanoprost was determined to be 8.9% and 0.08% respectively. The contact lenses were characterized via the following traits: light transmission, oxygen permeability, and surface roughness. The light transmission of the micelle contact lenses was over 90% in the visible light range, indicating that vision would not be affected when the lenses were applied. A horizontal diffusion cell was used to test the oxygen permeability of the contact lenses. The highest oxygen permeability was measured in the commercial contact lens (24.6 ± 1.4 Barrer), while the micelle contact lens had a permeability of approximately 15 Barrer. However, the oxygen permeability of the micelle contact lens increased to 18 Barrer after the drug was released. Finally, the surface roughness was measured via atomic force microscopy (AFM). The blank contact lens had the lowest roughness of 1.1 nm, followed by the commercial contact lens with a roughness of 2.8 nm. The micelle contact lens had a roughness of 6.6 nm; however, this value significantly increased after drug release to 25.9 nm. In vivo trials showed that the contact lens lowered IOP similar to daily commercial eyedrops for over 7 days. In addition, the bioavailability (calculated using collected tear fluid at various timepoints) of the two drugs on the contact lens was much higher than the drugs in the eyedrops, confirming that more drug was available for intraocular absorption. While there are some limitations to fixed dose combinations, the relative ease of application may be a benefit [74]. While many drug-loading contact lens studies involve soaking a lens in a drug or nanoparticle solution, there are also studies that involve loading drug into the lens itself, such as this study by Ciolino et al., in which latanoprost was used during the fabrication of a drug-polymer contact lens [76]. To prepare the lens, latanoprost was combined with PLGA and ethyl acetate. The resulting solution was deposited onto dry polymerized methafilcon, after which spinning was used to evaporate the ethyl acetate. A lens shape was punched out and the lens was photopolymerized, which allowed the methafilcon to encapsulate the drug polymer. High (HD) and low dose (LD) versions were tested. The inner diameter (4.2 mm for LD and 2.8 mm for HD), the outer diameter (9.5 ± 0.3 mm for LD and 10.2 ± 0.2 mm for HD), and the surface area (57.4 ± 4.1 mm² for LD and 76.7 ± 2.2 mm² for HD) were measured. Both types of contact lenses were tested in glaucomatous monkeys for one week. The IOP remained below baseline values during the duration of lens wear, and recovered back to baseline after the contact lens was removed [76].

Kim et al. developed a hypotonic thermosensitive gelling eyedrop containing sunitinib for RGC protection [77]. Formulated below the critical gel concentration, the low ion content facilitated water absorption and concentration to form a gel on contact with the ocular surface. The delivery system was characterized in a previous study using two techniques, optical coherence tomography (OCT) and multiple particle tracking [80]. For OCT imaging, three formulations, 12% hypotonic F127 (12% Hypo), 12% isotonic F127 (12% Iso), and 18% isotonic F127 (18% Iso), were compared via application to a rat eye. They found that the 18% Iso formulation gelled immediately and unevenly, leading to rapid clearance after blinking. The 12% Iso formulation did not gel due to the low polymer concentration and lack of water absorption, and therefore was also cleared quickly upon blinking. In contrast, the 12% Hypo formulation formed an even gel layer and remained on the surface of the eye after blinking. Multiple particle tracking was studied by adding fluorescent probe nanoparticles to the formulations and applying them to a rat eye in vivo and visualized ex vivo. While both the 12% Hypo and 18% Iso formulations displayed low average mean square distance (<MSD>), the 12% Iso formulation resulted in a higher <MSD>, confirming the formulation did not gel. Surprisingly, topical administration of the hypotonic 12% F127 containing small molecule drugs led to therapeutic delivery to the retina [80]. In a neuroprotection study, 0.4% sunitinib malate was formulated in the hypotonic gel-forming drop (SunitiGel). When SunitiGel was administered to rats once weekly for 14 days (3 doses total), high concentrations of sunitinib were found in the posterior segment of the eye (choroid/RPE and retina) for up to 7 days after the last dose was administered. In addition, retinal ganglion cell (RGC) survival was increased significantly from 23% to 34% with once weekly treatment in a rat model of optic nerve crush [77].

Fedorchak et al. developed a brimonidine-encapsulating microsphere gelling eye drop [81]. To fabricate drug-loaded microspheres, PLGA was dissolved in dichloromethane (DCM), to which a 50 mg/mL solution of brimonidine tartrate was added. After sonication, the solution was sonicated in poly(vinyl alcohol) (PVA) for 1 minute. A 1% PVA solution was mixed with the resulting solution to allow the residual DCM to evaporate. The final microspheres were washed, lyophilized, and characterized via SEM and zetasizing for morphology and size, respectively. The diameter was found to be 7.46 ± 2.86 μm and the surface morphology was poreless. A thermoreversible hydrogel was formulated by combining poly(ethylene glycol) with n-isopropylacrylamide (NIPAAm) in DI water. After mixing, ammonium persulfate and tetramethylethylenediamine were added to induce polymerization. The mixture was refrigerated overnight, resulting in a gel to which the microspheres were added. The gel was characterized by the lower critical solution temperature (LCST), the degradation rate, and the swelling ratio. Absorption experiments showed the LCST to be 33.5°C for both the gel and the microsphere gel. The gel was stable at 37°C for a month without degradation, and the swelling ratio was measured to be 7.50 ± 0.04. New Zealand white rabbits were given either a single drop of the gel microsphere solution, topical brimonidine twice daily, or a single drop of a blank microsphere gel solution. The brimonidine gel microsphere drop led to an IOP reduction of approximately 15% that was sustained for up to a month. The change in IOP was comparable to twice daily topical brimonidine for a month, while the blank microsphere gel did not significantly effect the IOP [81].

3.3.2. Intravitreal Injection

Intravitreal injections are often employed for neuroprotection due to the proximity to the retina. Arranz-Romera et al. developed PLGA microspheres loaded with dexamethasone, melatonin, and coenzyme Q10 (CoQ10) for dual anti-inflammatory and neuroprotective properties [75]. To fabricate the loaded microspheres, the therapeutic agents and PLGA were ultrasonicated in methylene chloride, followed by homogenization in polyvinyl alcohol. Next, more polyvinyl alcohol was added, and the resulting solution was stirred to evaporate the organic phase. Finally, microspheres in the 20–38 μm range were isolated using two mesh sieves and lyophilized. Microspheres encapsulating individual drugs were prepared similarly. The microspheres were characterized for size, encapsulation efficiency, and in vitro drug release. The mean size of the multi-loaded microspheres (DMQ-MS) were larger than any of the individually loaded microspheres at 29.0 ± 1.9 μm, in addition to being larger than the empty microspheres (24.7 ± 0.8 μm). The encapsulation efficiencies (EE) were compared for each drug between the individually loaded and multi-loaded microspheres. Only the loading of dexamethasone was affected, with the EE decreasing from 97.5 ± 1.5% to 78.2 ± 0.4%. Dexamethasone exhibited a burst release of 4.4 ± 0.8 μg DX per mg of MSs followed by a controlled release rate of 0.6 ± 0.04 μg DX/mg MSs/day. Melatonin also had a burst release (12.9 ± 1.1 μg MEL/mg MSs) and then exhibited a biphasic release profile of 1.7 ± 0.3 μg MEL/mg MSs until day 14, followed by 0.7 ± 0.2 μg MEL/mg MSs until day 30. Finally, CoQ10 maintained a constant release rate of 0.6 μg for the duration of the experiment. Ocular hypertension was induced in rats via injection of hypertonic saline, and the microspheres were injected intravitreally. While the IOP did not decrease, there were significantly more RGCs after drug combination treatment compared to administration of individual drug MSs or compared to untreated controls (Figure 4) 74]).

Figure 4.

Retinal flat mounts of rodent eyes displaying RGCs. (Ai) Example of a retinal flat mount with higher magnification images of retinas with (Aii) untreated healthy RGCs, (B) induced ocular hypertension, (C) hypertension plus treatment with empty MSs, (D) hypertension plus treatment with mixture MSs, and (E) hypertension plus treatment with DMQ-MSs. Figure reprinted with permission from [75].

3.3.3. Suprachoroidal Injection and Implants

Suprachoroidal injection has recently been explored as a potential space that can be expanded to physically increase aqueous humor drainage and/or administer therapeutics [68]. Chae et al. explored the potential for expanding the suprachoroidal space using both a commercial hyaluronic acid gel and an optimized hyaluronic gel supplemented with thiol groups and polyethylene glycol diacrylate (PEGDA). To prepare the cross-linked hyaluronic acid (HA-XL), thiol-modified hyaluronic acid (HA-SH) was dissolved in Hank’s Balanced Salt Solution with varying amounts of PEGDA. The addition of the PEGDA prolonged the degradation time of the HA-XL when compared to the HA-SH, increasing from 33 days to 40 days. They found that suprachoroidal injection of the commercial hyaluronic acid reduced the IOP in normotensive rabbits by about 4 mmHg for approximately 40 days, while the HA-XL reduced the IOP by 6 mmHg for approximately 4 months [68].

Hai et al. utilized a polyzwitterionic molecule with known anti-fouling characteristics to develop a hydrogel using previously described methods [69,82]. To prepare the macromonomer polycarboxybetaine (PCB-OAA), the zwitterionic monomer, 2-((2-hydroxy-3-(methacryloyloxy)propyl) dimethyl ammonio) acetate (CB-OH), was dissolved in ultrapure water, to which 4,4’-Azobis (4-cyanopentanoic acid) (ACVA) was added and stirred. The resulting solution was purged with nitrogen and then stirred for 16 hours. After the addition of NaHCO₃ and ultrapure water, acrylic anhydride was added dropwise. The resulting solution, PCB-OAA, was lyophilized and redissolved in PBS. A Michael Addition reaction was used to crosslink the PCB-OAA with diothiothreitol (DTT), resulting in formation of a hydrogel [82]. The concentration of PCB-OAA was used to control the gelation time of the hydrogel, which was tuned to five minutes using 7 wt% of PCB-OAA. The gels were injected into normotensive New Zealand white rabbits, and the IOP was reduced by 4mmHg for the first 4 weeks post-administration, followed by a reduction of 2mmHg that lasted until day 56. Histological analyses and clinical evaluations confirmed there was no observed damage or ocular irritation. The authors posited that the anti-fouling and anti-microbial properties would contribute to safety of the therapeutic with long-term, repeated use [69].

3.3.4. Subconjunctival Injections and Implants

Subconjunctival injection of sustained delivery systems can be employed for long-lasting IOP lowering without the need for daily eye drops. Salama et al. loaded brinzolamide into PLGA nanoparticles prepared via an emulsification/solvent evaporation method [70]. Two types of PLGA (type A and type B) were used to create 40 different nanoparticle formulations. PLGA types A and B were defined as PLGA “with an ester terminal copolymer of DL-lactide and glycolide in a 50:50 molar ratio”. The difference in inherent viscosity was 0.55–0.75 dl/g for type A and 1.4 dl/g for type B. To formulate the nanoparticles, PLGA and varying amounts of Poloxamer (Plx) 188 and polysorbate 80 were dissolved in dichloromethane along with a fixed amount of brinzolamide. This solution was added dropwise to a solution of PBS containing varying amounts of Plx 188, polysorbate 80, and Brij 97. After emulsification, the remaining solvent was evaporated to isolate the nanoparticles. Upon characterization, eight nanoparticle formulations were selected for in vitro studies based on particle size and drug encapsulation efficiency. The nanoparticle diameters ranged from 300 nm (labeled A19 & B17) to 1,000 nm (labeled A16). Encapsulation efficiency also ranged from 6% (B11) to 100%(B16). Based on the in vitro release study results, nanoparticle types A19 (70% released over 200 days) and B11 (90% released over 200 days) were chosen to compare to subconjunctival injection of a brinzolamide nanosuspension in normotensive New Zealand white rabbits. The duration of drug release in vitro was not predictive of the in vivo behavior, with the A19 and B11 nanoparticles providing IOP reduction over 10 and 8 days, respectively. In comparison, the brinzolamide nanosuspension reduced IOP for 3 days [70]. While promising, a longer duration of action would likely be needed for clinical implementation.

Wang et al. prepared a hydrogel composed of layers of chitosan and sodium alginate to deliver timolol with levofloxacin for post-injection infection control [73]. The hydrogel ball consisted of an inner core surrounded by 5 alternating outer layers with and without timolol, with the outermost layer containing levofloxacin. To prepare the hydrogel core, chitosan (CS) in acetic acid was mixed with CaCl₂, while equal weight percentages of sodium alginate (SA) and zinc oxide-modified biochar (ZnO-BC) were mixed together. 5% w/v of timolol was added to the SA-ZnO-BC solution, after which the resulting solution was added dropwise to the CS solution. After an hour, the core was isolated by removing the excess liquid with a filter. The core was then dipped into a blank SA-ZNO-BC solution, then again into the CS solution to form the first outer layer. This was repeated with blank and drug-loaded solutions until all the outer layers were formed. Finally, the hydrogel ball was immersed in a 0.5% glutaraldehyde solution for hardening. Characterization showed that the diameter of the ball with 1 layer was 1.5 mm, the diameter with 3 layers was 1.7 mm, and the diameter with all 5 layers was 2 mm. The ball was shown to be temperature stable until 48 ± 4.4°C with a degradation time of ~12 days. For in vivo trials, the hydrogel balls were surgically implanted into New Zealand white rabbits with carbomer injection-induced hypertension, providing IOP reduction for 13 days [73]. However, for a clinically viable surgical implant, the duration of action would ideally need to be at least several months.

Another approach by Ng et al. involved the design of a polymer microfilm with poly(lactide)/poly (ε-caprolactone) (PLC) copolymer sandwiching a polycaprolactone (PCL)-PEG layer encapsulating timolol [83]. To prepare the drug layer, timolol was dissolved in dichloromethane, to which polymeric pellets consisting of 80:20 PLC/PCL-PEG were added. The polymer/drug solution was poured on top of a blank layer of PLC, after which another blank PLC layer was poured on top of the drug layer. The thicknesses of the layers were controlled to be 10 μm for the drug layer and 15 μm for the blank layers. The total weight percentage of drug in the total film was determined to be 20%. The film exhibited 0-order drug release at a rate of 2.7 μg timolol per day and released timolol continuously for 3 months. The microfilm (4 mm × 6 mm) was inserted 2 mm posterior to the limbus of monkeys after a surgical conjunctival incision, providing IOP reduction of more than 50% for ~4.5 months [83].

Hsueh et al. explored an injectable microcrystal formulation for the sustained release of sunitinib for RGC protection [78]. To create the microcrystals, a sunitinib-pamoate complex was formulated by dissolving sunitinib malate and pamoic acid disodium salt separately in ultra-pure water. The two solutions were combined and mixed, then kept in the dark for two hours, after which the precipitate was combined via centrifugation and further lyophilization. To create the microcrystals, the precipitate was combined with a 70% ethanol/water solution. The resulting solution was stirred for 2 hours to form the crystals. After the crystals were collected and washed, the crystals were characterized for shape, drug loading, and long-term stability. Crystal images showed the rod shape of the crystals, with a length of 25 ± 7 μm and a width of 10 ± 2 μm. The drug loading, determined via HPLC, was 50.5 ± 2.9% and the crystal size was found to be stable when stored at room temperature for 180 days. The pharmacokinetics of the crystal formulation and RGC survival was tested in rats. The crystals were suspended in a solution of 0.04% polysorbate 80 (PS80) and 0.5% methylcellulose and injected subconjunctivally. The concentration of sunitinib in the retina after 20 weeks was determined to be 69.2 ng/g. An optic nerve crush model was subsequently used to test the neuroprotective effects of the microcrystal formulation. The RGC survival in control animals was 24 ± 5.3%, as compared to rats treated with the microcrystal formulation at 3 weeks (38 ± 5.8%), 7 weeks (39 ± 7.0%), and 20 weeks (38 ± 6.9%) [78].

3.4. Clinical Trials

In addition to the wide variety of preclinical studies, there are many products for glaucoma currently in the clinical trial pipeline, including implants, contact lenses, and others. Allergan is developing a removable ocular insert in the shape of a ring designed to elute bimatoprost [84]. The ring is inserted in the upper and lower fornices and is designed to be removed after 6 months. A completed Phase 2a trial showed IOP reduction over 6 months, which was comparable to repeated topical administration of timolol [84]. Elios Vision has developed a procedure in which an excimer laser is used to create microscopic incisions in the trabecular meshwork, lowering the resistance of flow of aqueous humor. This technology is already approved for use in the European Union for adults with glaucoma and is beginning clinical trials in the United States [85]. MediPrint Ophthalmics has developed a bimatoprost-eluting contact lens that is currently in a Phase 2b trial [86]. Bimatoprost is printed onto the anterior surface of the contact lens to provide release over one week. PolyActiva recently announced the results of a Phase 2a trial for their PA5108 ocular implant [87–89]. Designed as a biodegradable implant carrying latanoprost, PA5108 resulted in a greater than 20% IOP reduction over 6 months. Additionally, the implant was found to be fully degradable with no toxicity and side effects. They are set to begin their Phase 2b trial in 2024. Finally, Nicox Ophthalmics recently began recruiting for their Phase 3b trial of NCX 470, a bimatoprost-based eye drop using their nitric oxide (NO) donating technology [90]. NO is a signaling molecule thought to cause a cascade of events eventually leading to IOP reduction. The combination of this technology with bimatoprost is designed to be a dual-arm technology to result in greater IOP reduction overall. Their Phase 2 and 3 results show that a 0.1% formulation of NCX 470 results in a 1.73 mmHg greater IOP reduction than latanoprost, and a significantly higher percentage of patients had an IOP reduction of 10 mmHg or more by month 3 of the trial [90].

4. Age-Related Macular Degeneration

4.1. Background

AMD is the third leading cause of blindness in the world after glaucoma and cataracts [91]. With a rise in the aging population, AMD was estimated to affect 196 million people in 2020 and is projected to affect 288 million by 2040 [92]. Once thought to be untreatable, AMD has become a manageable condition thanks to the approval of anti-VEGF treatments [93]. However, with a complicated and diverse pathology, AMD is still a target for many in-progress treatment options. AMD is generally characterized by atrophy of the macula caused by a variety of factors, such as choroidal neovascularization and geographic atrophy [94]. There are a large range of risk factors, including genetic predisposition, smoking, lipid metabolism, ageing, and oxidative stress [92]. There are two main forms of AMD, commonly known as dry and wet.

4.1.1. Dry AMD

Also known as early-stage AMD, dry AMD is primarily characterized by drusen, which are lipid deposits under the retina, and hyperplasia of RPE cells [95]. Dry AMD progresses in one of two ways: into advanced stage dry AMD, or wet AMD. About 90% of AMD cases progress to advanced stage dry AMD, while 10% advance to wet AMD [96]. Advanced dry AMD is characterized by geographic atrophy, in which a majority of the RPE cell layer and photoreceptors have died, leading to vision loss [97]. If atrophy crosses the central fovea, there will be a loss of central vision. In addition to the RPE and photoreceptor loss, geographic atrophy involves thickening of Bruch’s membrane and choriocapillaris atrophy. Compared to wet AMD, the pathophysiology of dry AMD is relatively unknown. While there are different types of geographic atrophy, drusen-associated geographic atrophy is most typical of dry AMD [95], in which large drusen, which contain RPE cell remnants, reside under the RPE layer without causing neovascularization [95]. By the time advanced dry AMD can be diagnosed, there is generally an irreversible level of RPE cell death (Figure 5) [98].

Figure 5.

The pathophysiology of AMD and therapeutic targets. Reused with permission from [97]. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; RPE, retinal pigment epithelium; ECM, extracellular matrix; mTOR, mammalian target of rapamycin; MMP, matrix metalloprotease; LDL, low density lipoprotein.

4.1.2. Wet AMD

Neovascular (wet) AMD is primarily characterized by the formation of new blood vessels in the choroid. The new vessels leak blood and serum, giving rise to the ‘wet’ characterization [92]. There are two subtypes of wet AMD, choroidal neovascularization (CNV) and polypoidal choroidal vasculopathy (PCV). In CNV, the new vessels push through the choroid to Bruch’s membrane and the RPE/photoreceptor layer of the retina [99]. This, combined with the leakage, can lead to macular damage and detachment, subsequently leading to vision loss. PCV also involves choroidal vascularization; however, the vasculature only passes through Bruch’s membrane while leaving the RPE intact. For this reason, the outcomes are much more favorable. The pathology is similar to CNV due to the underlying vascularization [99]. In addition to genetic and behavioral risk factors for AMD overall, oxidative stress in the RPE may lead to the promotion of pro-angiogenic factors. Alternatively, degeneration of choroidal vasculature may lead to early angiogenesis. Regardless of the pathway, there is a promotion of growth factors, namely VEGF, in wet AMD. In the eye, VEGF is responsible for angiogenesis as well as vascular permeability.

4.2. Currently Approved Treatments

4.2.1. Dry AMD

Until recently, there were no approved treatments for dry AMD. In 2023, two were approved for the treatment of geographic atrophy. The first one to be approved was Syfovre^™, an intravitreal pegcetacoplan injection designed to be administered every month to 2 months [100]. Pegcetacoplan is a complement inhibitor originally approved for paroxysmal nocturnal hemoglobinuria, but is used here to target the geographic atrophy in dry AMD patients [101,102]. This treatment was recently approved in February of 2023. Avacincaptad pegol intravitreal solution (IZERVAY^™), a complement C5 inhibitor, is a once-monthly formulation that was approved by FDA in 2023 for the treatment of geographic atrophy secondary to AMD. Avacincaptad pegol is a ribonucleic acid aptamer, covalently bound to an approximately 43-kilodalton (kDa) branched polyethylene glycol (PEG) molecule [103].

4.2.2. Wet AMD

The standard of care treatment for wet AMD is anti-VEGF therapy. Anti-VEGF therapeutics reduce the permeability and leakiness of the blood vessels, as well as inhibit neovascularization [98]. There are currently five FDA-approved anti-VEGF biologics, four of which (brolucizumab, aflibercept, ranibizumab, and pegaptanib sodium) are approved for the treatment of wet AMD [104]. Bevacizumab is monoclonal antibody approved for treatment of some forms of cancer but is often used off-label as a treatment for wet AMD. All of these drugs are administered intravitreally every 1–3 months. A fixed schedule was shown to present better outcomes than an irregular schedule of injections, and consistency was key to therapeutic benefit [104]. Recently, the FDA approved higher doses of aflibercept for AMD treatment to reduce the frequency of administration. The aflibercept dose was increased from 2 mg to 8 mg, and an updated dose schedule results in up to 16 weeks between doses [105]. The most recent significant clinical development in the treatment of wet AMD was the SUSVIMO Port Delivery System (PDS), which was approved by the FDA in 2021 [106,107]. The PDS consists of a ranibizumab-eluting implant that is implanted into the eye and provides up to 6 months of therapeutic effect before being refilled. The implant is made of polysulfone and contains a titanium mesh to control the diffusion of the drug. Phase 3 trials showed that at 40 weeks after the installation of the PDS, 80.7% of patients had a best-corrected visual acuity (BCVA) score of 69 or better, which is equivalent to 20/40; corresponding to slight nearsightedness. However, in late 2022, the PDS was voluntarily recalled by Genentech due to some of the implants not performing satisfactorily; however, patients with the PDS already installed can continue to receive refills.

While the availability of anti-VEGF therapies revolutionized the clinical management of AMD, there are various side effects and disadvantages with both the administration route and the therapeutic itself. While rare, intravitreal injections pose many risks, including endophthalmitis, retinal detachment, and IOP elevation [29]. Returning frequently to the clinic for injections can further be inconvenient and stressful for patients [108]. While there is concern with systemic exposure of anti-VEGF therapies, local injection largely minimizes the potential for systemic side effects [109]. The potential local side effects include intraocular inflammation, ocular hemorrhage, and RPE tears [29].

4.3. Current Outlook of Preclinical Research

There is a large variety of preclinical studies in progress, ranging from topical treatments to injectables. Since the standard of care for wet AMD typically involves intravitreal injections, there have been efforts to explore less invasive options such as topical treatments, as well as studies looking to improve on intravitreal formulas by developing more sustained release options. In addition, since there are currently limited treatments for dry AMD, there have been investigations into the delivery of neuroprotective agents for geographic atrophy. To test these new treatment options, various models are used. For dry AMD, sodium iodate can be injected to model geographic atrophy [110]. For wet AMD, choroidal neovascularization and the accompanying vessel leakage can be modeled with VEGF injections [111–113], a combination of VEGF and basic fibroblast growth factor (bFGF) injections [114], or laser photocoagulation [115,116].

4.3.1. Topical Treatments

An option to improve transport and penetration upon topical administration is the use of cell penetrating peptides, or CPPs. CPPs have been shown to accumulate in cells and can be conjugated to different drug loads. Ge et al. conjugated penetratin, a known CPP, to lutein in a nanoemulsion gel for the treatment of dry AMD [110]. Lutein is a carotenoid and is a primary component of retinal macular pigment that has been shown to have antioxidant properties and to protect the retina from atrophy. An in situ nanoemulsion (NE) gel was developed using egg phospholipid and gellan gum, which could then be topically administered. To formulate the unmodified (NE) and penetratin-modified nanoemulsions (P-NE), lutein, vitamin E, and egg phospholipids were dissolved in an organic phase of medium-chain triglycerides and ethyl acetate. Separately, an aqueous phase consisting of 0.3% Poloxamer 188 was heated to 65°C. The two solutions were combined using high shear mixing, and the resulting mixture was homogenized then rotary evaporated to remove ethyl acetate, resulting in the NE formulation. Finally, stearyl-penetratin was added and the solution was incubated to form P-NE. To form the P-NE Gel, gellan gum was dissolved in water at 90°C for 20 minutes, after which P-NE was added. P-NE and NE were characterized via dynamic light scattering; the mean diameter of the NE and P-NE particles was ~110 nm. The P-NE gel was characterized for gelling ability, and 0.6% was found to form a gel while maintaining the spherical shape of the NE. Ocular distribution studies using fluorescently labeled NE in rats showed that the P-NE gel group showed the highest fluorescence in the posterior tissues, followed by the P-NE and then the NE. Dry AMD was induced in rats using sodium iodate, and the penetratin-modified gel (P-NE GEL) was compared to the gel (NE GEL) and the penetration-modified nanoemulsion (P-NE). Electroretinograms on the rats showed that the P-NE GEL group most significantly preserved retina and photoreceptor function compared to the control group, P-NE, and NE GEL group [110]. However, these findings need confirmation in an animal model with a larger eye size, due to inherent limitations of using small rodents for topical formulation development for treating the retina [117].

Yadav et al. loaded atorvastatin, a therapeutic approved for lowering high cholesterol, into solid lipid nanoparticle (SLN) to investigate if the drug could reach the posterior tissues upon topical administration [118]. Atorvastatin, in addition to statins in general, has been linked to reduction of choroidal neovascularization and the delay of late AMD [119]. To formulate the SLNs, a lipid phase consisting of Compritol^® 888 ATO, atorvastatin, and PEG₄₀₀ was heated to 75°C and combined with an aqueous phase consisting of Poloxamer 188 (P188) and Phospholipon 90H (P 90H). Various formulations with different ratios of Compritol^® 888 ATO/P 90H and P188/PEG₄₀₀ were synthesized. After the resulting mixture was stirred, it was homogenized then cooled to room temperature, forming the atorvastatin-encapsulating nanoparticles. An atorvastatin suspension (SUS) was formulated as a control, in which 0.25% (w/v) atorvastatin was added to water with 0.2% (w/v) sodium carboxymethyl cellulose and stirred. The mean diameter and entrapment efficiency of the various SLNs were found in order to fit the formulation ratios to a polynomial model to optimize encapsulation efficiency and particle diameter. The selected SLNs were characterized via size polydispersity index, drug content, entrapment efficiency, and in vitro release. The final SLN formulation had an average diameter of 256 ± 10.5 nm and a polydispersity index of 0.26 ± 0.02. TEM images showed small spherical nanoparticles with a surfactant layer surrounding the outside. In vitro release studies showed that approximately 60 percent of the encapsulated atorvastatin was released after 96 hours, while 100 percent of the drug in the control model was released within 20 hours. Ex vivo experiments with porcine eyes were performed to show the corneal permeability of the SLNs. 250 minutes after administration, approximately 50% of the SLNs had permeated, while only about 30% of the SUS had permeated in the same time frame. In vivo pharmacokinetic experiments were conducted on rabbits by administering the SLN drops 5 times in 5 minutes intervals and extracting the vitreous and aqueous humors at various time intervals. The SLNs resulted in maximum atorvastatin concentrations of 1.1 ± 0.2 μg/mL and 1.4 ± 0.2 μg/mL in the aqueous and vitreous, respectively, an hour after the last administration. On the other hand, the SUS formulation resulted in lower maximum concentrations of 0.7 ± 0.1 and 0.5 ± 0.1 μg/mL in the aqueous and vitreous respectively (t_max = 30 minutes for aqueous and 60 minutes for vitreous). In addition, the mean residence time of atorvastatin (measured using the lacrimal fluid from rabbit eyes with fluorescent SLNs and SUS) in the vitreous was 20.7 hours for the solid lipid nanoparticle system compared to 1.1 hours for the free drug system. While there were no experiments to demonstrate whether the drug levels achieved with the SLNs were therapeutic for preventing neovascularization, this study presents a potential non-invasive alternative to intravitreal injections [118].

Schopf et al. utilized the mucus-penetrating particle (MPP) technology to deliver loteprednol etabonate (LE), a corticosteroid, and KAL821, a tyrosine kinase inhibitor [111]. The MPP technology involves coating the surface of nanoparticles with mucoinert polymers to penetrate the mucus layer and increase intraocular drug absorption. They formulated LE-MPP and KAL821-MPP using a wet-milling approach with Pluronic F127 as the coating polymer. The LE-MPP had an average size of 240–350 nm, while the KAL821-MPP had an average size of 160 nm. A conventional particle (CP) formulation was also composed without the mucoinert coating by milling LE with the surfactant sodium dodecyl sulfate (SDS) (LE-CP) (240 nm in size). In vivo experiments were conducted in both rabbits and mini pigs. Single dose pharmacokinetic studies in rabbits showed increased levels of LE in the cornea and retina when comparing the LE-MPP to the LE-CP. Specifically, there was a 4-fold increase in the amount of LE in the cornea using LE-MPP, and a 5-fold increase of LE in the retina. To compare the effect of particle size on the pharmacokinetics, LE-MPP was compared to a micronized LE formulation (LE-Micro). The LE-MPP provided a 2-fold increase of LE in the cornea and a 3-fold increase in the retina when using the MPP formulation. To test the potential therapeutic effects in a wet AMD model, vascular leakage was induced by injecting VEGF into the mid-vitreous. KAL821-MPP were dosed topically compared to intravitreally injected triamcinolone acetonide (TA) and vehicle, and the VEGF was injected on day 3 after initiation of treatment. Rabbits in the vehicle group had a leakage score of 3, whereas those treated with the positive control of TA had leakage score of 0. Rabbits treated with either twice per day or 4 times per day administration of KAL821-MPP had a leakage score of only 0.5, demonstrating therapeutic effect in the posterior segment with topical therapy [111].

Finally, polymers can also be useful in aiding the transport and penetration of drugs. Huang et al. used polymers to develop an eye drop formulation with axitinib, a VEGF tyrosine kinase inhibitor [112], to treat wet AMD. VEGF tyrosine kinase inhibitors have been approved for cancer therapeutics; they inhibit VEGF receptors as well as platelet-growth factor receptor (PDGFR). To formulate the eye drop, axitinib was dissolved in glacial acetic acid to create the organic phase, which was combined with an aqueous solution of hydroxypropyl methylcellulose, caffeine, and (2-hydroxypropyl)-β-cyclodextrin. The resulting mixture was spray-dried and desiccated to form a powder. The powder was then rehydrated with water and filtered to create the eye drop solution. The axitinib content in the drops was found to be 8.5 ± 132.1% (w/v). The pH was 6.75, and the osmolarity was 580 ± 2.9 mOsM/kg, both of which are within acceptable ranges for topical products. Two in vivo pharmacokinetic studies were conducted, one in rabbits and one in monkeys. When rabbits were given a single dose, the concentration of axitinib stayed above the preset threshold of 100 times the IC₅₀ of axitinib in the retina and choroid (935 and 980-fold greater than the IC₅₀) up to 3 hours post one dose. On the other hand, the administration of a 2% axitinib suspension resulted in unquantifiable levels of axitinib in the retina. When 2 doses (4 hours apart) were administered to monkeys, the concentration of axitinib stayed above the preset therapeutic limit of 10 ng/g in the retina and choroid out to 24 hours, except at the 24-hour timepoint in the retina, where the concentration was 5.9 ng/g. Neovascularization and associated leakage was induced in rabbits by injecting VEGF locally. The eyedrops reduced leakage area by 94% compared to untreated controls, similar to 98% with bevacizumab injections (Figure 6). In addition, CNV was induced in rats via laser injury, after which different dosages of eyedrops were administered and compared to aflibercept injection. The resulting leakage area after aflibercept was similar to the 0.8% eyedrops. this study still demonstrated that the eye drop formulation can compare to the current standard of care [112].

Figure 6.

Fluorescein angiography of Dutch Belted rabbit eyes. Angiography for all three groups was performed before neovascularization was induced (naïve), as well as at day 2 and 4 post RNV induction with and without treatment with either a single intravitreal injection of Avastin (middle) or 4 topical doses/day of the ITRI AXN drops (bottom). Reprinted with permission from [112].

4.3.2. Intravitreal Injections

Mu et al. encapsulated bevacizumab in multivesicular liposomes (Bev-MVLs) for intravitreal injections [115]. To formulate the Bev-MVLs, an aqueous solution of bevacizumab, sucrose, and human serum albumin was added to a lipid solution consisting of chloroform, diethyl ether (1:1), 10 mg/mL 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 2 mg/mL 1,2-dipalmitoyl-sn-glycero-3-phosphoglycerol (DPPG), 8 mg/mL cholesterol, and 2 mg/mL triolein in a 1:1 ratio. The resulting mixture was homogenized and added dropwise to a second aqueous solution consisting of glucose and lysine. Finally, the mixture was homogenized and rotary evaporated to remove the organics. The encapsulation efficiency was 80.7 ± 4.4%, and the activity of the bevacizumab was preserved during encapsulation. The median size of the MVLs was ~30 μm, and the MVLs retained a uniform spherical shape. In vitro release studies resulted in 80–90% of bevacizumab being released by day 14. The Bev-MVLs were also shown to be stable in vitreous for about 3 days, after which erosion began; the MVLs burst open by day 7. In vivo pharmacokinetic studies in rabbits showed that the mean residence time of the bevacizumab-liposomes (Bev-MVLs) in the vitreous was 15 days, compared to ~10 days for the bevacizumab solution. Additionally, the concentration of bevacizumab in the vitreous was higher in the Bev-MVLs group than in the bevacizumab suspension group (207.3 ± 6.3 μg/mL vs. 163.1 ± 6.5 μg/mL at day 3), and the bevacizumab was quantifiable in the Bev-MVLs group out to day 56. Finally, laser induced CNV in rats was used to investigate the anti-angiogenic potential of the drug systems. 28 days after administration, they found that Bev-MVLs treatment led to significantly smaller CNV lesion size when compared to bevacizumab solution and the control [115].

In addition to lipid-based encapsulation techniques, there are other materials that can be used to enhance drug delivery, including surfactants and polymers. For example, Anand et al. used polyoxyethylene (10) stearyl ether (Brij L76), a surfactant, to create self-assembling amphiphilic carbon dots designed to encapsulate bevacizumab for the treatment of wet AMD [114]. To create the self-assembling carbon dots, Brij L76 was heated to 4 different temperatures for 3 hours, after which the solutions were dissolved in ethanol and centrifuged, resulting in the self-assembly of carbon dots (CDs). The CD/ethanol solution was added dropwise to sodium phosphate buffer containing bevacizumab (1 mg/mL) and stirred at room temperature. The solution was centrifuged and washed with a sodium phosphate buffer. The final pellet was dissolved in sodium phosphate buffer containing 1.0% PEG₃₀₀. Finally, sonication of the mixture resulted in the self-assembly of the carbon nanovesicles (BVZ@CNVs). TEM images were used to determine which formulation (based on different temperatures) was most suitable for further testing. The 200 °C group of BVZ@CNVs were selected since they formed uniform vesicles without aggregation. The vesicles had an average diameter of 100 nm. The loading efficiency of bevacizumab was 24.4 ± 3.9%, while the encapsulation efficiency was 56.4 ± 6.3%. In vitro testing showed evidence of a sustained release model, with only 45% of the bevacizumab being released by day 60. Cell viability was tested at various concentrations in both ARPE19 and human umbilical vein cells (HUVEC) and resulted in negligible toxicity in both cell lines at all tested concentrations. Pharmacokinetic studies in rabbits revealed that 21% of the bevacizumab dose was found in the vitreous at day 21 post-administration in the BVZ@CNVs group, while only 5.5% of the bevacizumab solution remained. CNV was induced in rabbits through VEGF and basic fibroblast growth factor (bFGF) (loaded into alginate hydrogels) injections. Seven days after intravitreal injections of bevacizumab carbon nanovesicles, relative vessel area was significantly lower than in animals treated with bevacizumab solution or empty nanovesicles. [114].

Varshochian et al. developed PLGA nanoparticles to load bevacizumab for wet AMD treatment [120]. To prepare the nanoparticles, PBS containing bevacizumab and human serum albumin was added to dichloromethane containing PLGA. After sonication, PVA in PBS was added and the solution was sonicated again. The resulting mixture was stirred overnight to evaporate the dichloromethane. Finally, the nanoparticles were collected via centrifugation and dispersed in PBS with 2% (w/v) mannitol. Dynamic light scattering revealed uniform (PDI: 0.17 ± 0.05) spheres (verified using SEM) with a diameter of 190 ± 29 nm. The PLGA nanoparticles had a high entrapment efficiency of 84.1% and a loading efficiency of 7.4%. In vivo pharmacokinetics in rabbits showed that the bevacizumab concentration after nanoparticle injection stayed above the minimum inhibitory concentration (MIC) in the vitreous for 56 days, while the concentration after free bevacizumab injection dipped below the MIC by day 30. The timing and magnitude of the C_max of bevacizumab in the vitreous after nanoparticle injection was reflective of the sustained release profile compared to free bevacizumab. Even though this study did not show how the nanoparticles affected the neovascularization in the retina, sustained release would be expected to prolong the therapeutic effect [120].

A study by Lima E Silva et al. investigated a 20-mer peptide (LRRFSTAPFAFIDINDVINF) derived from collagen IV for its potential to block angiogenesis [113]. This peptide, known as AXT107, was discovered previously via structure activity studies and a combination of computational and experimental methods [121]. AXT107 was also observed to form a gel upon injection into the vitreous in mice and rabbits. CNV was induced in mice using laser photocoagulation after treatment with 1 μg of AXT107, 40 μg of aflibercept, or a control scramble peptide was administered. AXT107 significantly prevented CNV area growth compared to the control and Aflibercept. When a combination of AXT107 and aflibercept was injected, it prevented CNV area growth more significantly than either treatment alone. To test the ability of AXT107 to reduce existing neovascularization, treatments were administered 7 days after laser photocoagulation. AXT107 still exhibited the greatest therapeutic effect by reducing neovascularization area when compared to aflibercept and a baseline control. In rabbits, the ability of AXT107 to reduce neovascularization leakage was tested. Three days after a 50 μg injection of AXT107, 10 μg of VEGF was injected. Seven days later (peak leakage in control condition), leakage was reduced by 60% compared to control. To compare AXT107 to aflibercept, 23 days after either 50 μg of AXT107 or 500 μg of aflibercept was injected, VEGF was injected. While aflibercept treatment resulted in a 69% reduction of leakage, AXT107 reduced leakage by 86%. Two weeks later, another round of VEGF was administered, after which the AXT107 still exhibited a therapeutic effect compared to the baseline and aflibercept [113].

Tsujinaka et al. developed self-aggregating sunitinib microparticles to block neovascularization [116]. To prepare the microparticles, PLGA 75:25 4A and PLGA-PEG5k were added to dichloromethane and combined with a sunitinib malate solution in dimethyl sulfoxide. After homogenization in a PVA solution, the solution was stirred to evaporate the dichloromethane. Subsequent filtration and lyophilization resulted in dry microparticles (MPs), which were then suspended in a sodium hyaluronate solution for in vivo trials. The MPs were measured to be 32 ± 9 μm via DLS. In vitro drug release studies demonstrated a slower release than a sunitinib drug suspension, with approximately 60% being released by one month post administration. In vivo studies included testing the efficacy of the MPs in a mouse CNV model, as well as pharmacokinetic studies in rabbits. CNV was induced in mice with a laser, and the MPs were injected intravitreally before damage induction. When compared to controls of empty MPs, sunitinib MPs significantly reduced mean CNV area out to 42 days post-administration. In addition, the MPs suppressed CNV growth for a longer time period than aflibercept, with aflibercept resulting in an increase of CNV by day 28 post administration. Finally, pharmacokinetic studies in rabbits showed sustained levels of sunitinib above the therapeutic threshold in the retina, RPE/choroid, and vitreous up to 6 months post administration, demonstrating the sustained release potential of this system [116].

4.4. Clinical Trials

In addition to preclinical studies, there are a few promising therapeutics in the clinical trial pipeline. One promising therapeutic in the clinical trial pipeline is RGX. RGX-314, developed by REGENXBIO, consists of subretinal and suprachoroidal delivery of gene therapy for wet AMD [122]. This is intended to be a one-time treatment in which the NAV AAV8 vector contains a gene that encodes for a monoclonal antibody fragment; this antibody is supposed to downregulate VEGF activity, thereby inhibiting neovascularization and leakiness. RGX-314 completed Phase 2 trials and is recruiting for Phase 3. Clearside Biomedical recently concluded a Phase 1/2a clinical trial of CLS-AX, which is an axitinib suspension intended to be administered via their suprachoroidal space microinjector [123,124]. The phase 2a trial showed that the treatment could be effective for up to 6 months, and axitinib was not found systemically or in off-target areas of the eye. EyePoint Pharmaceuticals recently concluded their Phase 2 study of EYP-901, a biodegradable intravitreal implant containing the tyrosine kinase inhibitor Vorolanib [125]. Their existing approved DURASERT^® technology, a non-degradable intravitreal implant, was modified to be biodegradable to through the removal of the polyimide coating. The results of their Phase 1 trial showed a stable BCVA score 6 months after a single injection, with no significant side effects or toxicity.

5. Diabetic Retinopathy and Diabetic Macular Edema

5.1. Background

DR is a complication of diabetes mellitus in which increased levels of blood sugar lead to retinal damage and progressive vision loss. Of about 422 million individuals living with diabetes worldwide, approximately 1/3 of them are diagnosed with DR, the leading cause of blindness in working-age adults (20–74) [126,127]. With the increasing incidence of diabetes cases worldwide, the burden of DR is expected to increase [128,129]. There are two types of DR, including non-proliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). Early stage NPDR is observable by retinal examination and characterized by retinal lipid exudates and microaneurysms in capillaries which can rupture, leading to hemorrhaging [130]. As these capillaries hemorrhage and disease conditions become increasingly severe, retinal ischemia and neovascularization occur, resulting in PDR. As neovascularization progresses, diabetic macular edema may develop in cases where there is fluid buildup in the macula either sub- or intra-retinally due to breakdown of the blood-retina barrier. DME presents in patients as light sensitivity, worsening of vision, and sometimes blindness [131].

The pathophysiology of DR heavily implicates hyperglycemia, as high circulating blood sugar levels lead to decreased pericyte coverage of retinal capillaries. Pericytes are essential in blood vessel formation, maintaining the integrity of the BRB by enveloping underlying endothelial cells, and modulating blood flow by assisting in vessel contraction and dilation [132,133]. Pericytes are particularly susceptible to hyperglycemic injury, with their loss leading to compromised integrity of the retinal capillaries and endothelial cell proliferation characteristic of NPDR microaneurysms [134]. NPRD progresses to PDR as large areas of capillary dysfunction lead to hypoperfusion. The resulting localized hypoxic ischemia leads to expression of VEGF in the retina, contributing to development of neovascularizing hyperpermeable vessels, onset of DME, tractional retinal detachment, and vitreous hemorrhage (Figure 7) [135].

Figure 7.

The clinical presentation of the eye of a patient with diabetic retinopathy shown by pseudo-colored fundus images (left) and fluorescein angiography (right). There is elevated leakage in NPDR, while PDR is characterized by regions of non-perfusion (NP) and further progression of vascular hyperpermeability. PDR is also characterized by the presence of neovascularization (NV), hard exudates (HE), and vitreous hemorrhage (VH). Reprinted with permission from [135].

There is evidence pointing to inflammation as another driving force in the progression of DR. It is reported that patients with NPDR have elevated levels of inflammatory cytokines (IL-1β, IL-6, IL-8, TNF-α and MCP-1) in ocular tissues. The prolonged immune response to hyperglycemic conditions leads to irreversible damage to the tissue [136]. How this chronic inflammation contributes to retinal neurodegeneration is a subject of study; however it is widely reported that inflammatory processes lead to widespread neuronal apoptosis, including the death of RGCs and reduced retinal nerve fiber thickness [137]. Neural changes may result in altered retinal function, indicated by irregular electroretinogram (ERG) readings even prior to clinical presentation of retinal microaneurysms in NPDR, allowing for the monitoring of disease progression [138].

5.2. Currently Approved Treatments

At present, there are few treatments for diabetic retinopathy. For early stage NPDR, management of blood sugar, blood pressure, and lipid intake are helpful for preventing the exacerbation of symptoms and disease progression [139,140]. A clinical trial conducted by the Diabetes Control and Complications Trial Research Group found that heavy intervention to maintain normal plasma glucose levels reduced the risks of NPDR progression [141]. While increased blood pressure in diabetes patients has not been directly correlated with onset and progression of NPDR, a study out of the United Kingdom Prospective Diabetes Study Group found that blood pressure control (means of 144/82 mmHg compared to 154/87 mmHg) reduced the risk of progression towards a more serious disease state [142]. Patients in the group with controlled blood pressure were also at decreased risk of needing retinal laser photocoagulation therapy, a frontline treatment for DR that is used to prevent further leakage from retinal vasculature [142].

As NPDR progresses towards PDR, retinal laser photocoagulation treatment is recommended to stop the growth of new vessels and prevent the progression towards DME [143]. Intravitreal anti-VEGF therapies (bevacizumab, aflibercept, and ranibizumab) are often used for treating DME. Ranibizumab was approved for use in treating DR with DME in 2015, and DR with or without DME in 2017 following the RISE and RIDE trials demonstrating that monthly intravitreal injections improved vision and reduced macular edema after 24 months [144]. The effect of aflibercept was studied in the PANORAMA trial with injection in PDR eyes without DME and found that two treatment regimens (every 16 weeks after 3 monthly injections or every 8 weeks after 5 monthly injections) provided significant improvement in disease state at 24, 52 and 100 weeks [145]. Aflibercept was approved for use in 2019 following these trials. Intravitreal faricimab, the first bispecific VEGF and angiopoietin-2 inhibitor, was approved for treating DME and AMD in 2022 following the YOSEMITE and RHINE trials. Faricimab was injected intravitreally bimonthly vs extended dosing intervals (on a patient-needed basis) and was compared to aflibercept bimonthly [146]. The study demonstrated that bimonthly faricimab compared similarly to bimonthly aflibercept with regard to visual acuity determined by EDTRS letter scores. Additionally, a flexible dosing interval with frequency up to once every 16 weeks provided similar therapeutic outcomes and non-inferiority of both regimens. Bevacizumab is also used off-label for treating DME [147].

In addition to anti-angiogenic treatments, corticosteroids have been injected intravitreally to address the inflammation, thus blocking the production of inflammatory cytokines. As there is a distinct correlation between levels of inflammatory mediator cytokines and levels of VEGF in patients with DME, reducing the local inflammation can treat neovascularization and the underlying pathogenesis [148]. A dexamethasone intravitreal implant (Ozurdex^®) was approved for use in 2009 treatment of DME following the MEAD trials, showing that a single implant allowed for treatment over a period of 6 months, reducing the frequency of intravitreal injection [149,150]. Over 3 years with an average of 5 total treatments, patients receiving the highest dose of dexamethasone (0.7 mg) intravitreal implant experienced significant increases in best-corrected visual acuity (BCVA) and in patients presenting with 20/40 vision with limited serious adverse events [151]. An intravitreal insert releasing fluocinolone acetonide (Iluvien^®) was approved for use in 2014 for treating DME following the FAME trials [152,153]. There was a significant improvement in BCVA in both the high dose (0.5 μg/day) and low dose (0.2 μg/day) treatments by letter score compared to baseline, as well as significant improvement in patients ending the study with 20/40 vision with a single yearly injection of fluocinolone acetonide. However, a common adverse event observed during this trial was patients presenting with IOP-elevation that was unable to be pharmacologically treated and required incisional glaucoma surgery to reverse, with more patients requiring this surgery in the high dose group [154].

5.3. Current Outlook of Preclinical Research

The current standard mouse and rat animal model for testing therapeutics in DR involves the intraperitoneal injection of streptozotocin (STZ) to induce a disease state that compares to type 1 diabetes in humans. STZ is a selective pancreatic β-cell antagonist that is administered as a single dose to cause necrosis of β-cells, which are responsible for the maintenance of normal blood glucose levels via insulin synthesis and release. Induction of a mild diabetic state in animals is characterized by a fasting blood glucose level concentration of at least 150 mg/dL after 72 hours and >250 mg/dL 3 weeks post-induction compared to a resting state of 80 mg/dL [155,156]. Relevant models of neovascularization are also employed to study the therapeutic benefits in treating DME. The leading induction model is a laser-induced rupture of Bruch’s membrane, leading to choroidal neovascularization (CNV) in mice, rats, and rabbits, among others [157].

5.3.1. Topical Administration

Incorporation of surface coatings to promote favorable corneal interaction and increased absorption of drug has been employed to increase ocular bioavailability for treatment of DR. Laddha et al. developed PLGA nanoparticles containing pioglitazone. Pioglitazone is a leading anti-diabetic medication and is a PPARγ (peroxisome proliferator-activated receptor-γ) agonist, leading to increased insulin sensitivity [158,159]. Both 75:25 and 50:50 PLGA were used to make nanoparticles with polysorbate 80 as a surface coating to promote corneal residence time. Particle formation and surface coating were established by dissolving PLGA and pioglitazone in a dichloromethane/methanol mixture, with the aqueous phase containing PVA and PS80. Dropwise addition of organic phase into aqueous under high-speed homogenization led to the formation of pioglitazone-PLGA nanoparticles with PS80 on the surface. 75:25 PLGA NPs were 163 nm (PDI 0.29, ζ-potential −10.8 mV, 7.9% loading efficiency), 50:50 PLGA NPs were 171 nm (PDI 0.28, ζ-potential −7.5 mV, 8.1% loading efficiency). Once-daily topical administration of 75:25 and 50:50 PLGA pioglitazone NPs led to a significant reduction in VEGF in the vitreous humor of STZ-induced diabetic rats as determined by ELISA, with the greatest effect observed in the highest dose group (6 mg/mL 50:50 PLGA NPs) with once-daily dosing for 4 weeks [158].

Srinivasarao et al. developed PCL nanoparticles with a Pluronic F68 coating to deliver triamcinolone acetonide (TA) [160]. They utilized a combination therapy of TA, which has been reported to cause posterior subcapsular cataracts with repeated long-term treatment, and pyrrolidine dithiocarbamate (PDTC), an NF-κb inhibitor that has proven effective in the prevention of cataract formation [160–162]. PDTC or TA NPs were generated by mixing either PDTC or TA in acetone with PCL with an aqueous solution of F68, evaporation of the acetone under reduced pressure. PDTC NPs were 158.9 ± 31.3 nm (PDI 0.01, ζ-potential −16.8 ± 5.8 mV) and TA NPs were 162.8 ± 36.2 nm (PDI 0.07, ζ-potential −21.6 ± 5.2 mV). They found that STZ-induced diabetic rats had higher lens carbonyl and malondialdehyde (MDA) content, both markers of oxidative stress, compared to non-diabetic control animals. Further, there was increased retinal HIF-1α and neovascularization area in diabetic rats compared to controls. Eye drops (25 μL) containing TA NPs or TA + PDTC NPs were administered 2x daily for 40 days. While TA alone provided a significant reduction in only neovascularization area, addition of PDTC NPs to the treatment regimen provided additional reduction in neovascularization, while also leading to reduction of lens and retinal oxidative stress markers (Figure 8) [160]. However, it is important that while addition of PDTC to TA treatment augments efficacy, combination therapies do have inherent drawbacks, including unknown drug-drug interactions or toxicity not apparent in single drug therapies. Further, the use of combination therapies should be accompanied by evaluation of formulation pharmacokinetics and safety, as these drugs may not always be synergistic [163].

Figure 8.

Characterization of retinal vasculature with Isolectin-B4 staining in healthy retinas (control) and 40 days after STZ induction and treatment with PBS, TA NP, or TA + PDTC NP in rats. Green arrows identify vascular tufts which are reduced following both TA NP and more so with TA NP + PDTC NP combination treatment. Scale bar; 200 μm. Reprinted with permission from [160].

Radwan et al. describe the delivery of apatinib, an FDA approved VEGF-2 inhibitor used to treat cystic carcinoma, formulated within hyaluronic acid (HA)-coated bovine serum albumin (BSA) nanoparticles [164]. Apatinib-BSA-NPs were prepared by combining a solution of BSA with apatinib in DMSO, desolvating by adding ethanol dropwise, and cross-linking using glutaraldehyde to form NPs. Coating with HA was performed by adding the apatinib-BSA-NP dispersion dropwise to aqueous solution of HA. Apatinib-BSA-NPs were 198.9 ± 3.5 nm (PDI 0.05, ζ-potential −29.2 ± 1.3 mV), whereas optimized HA-coated (2 mg/mL) particles were 222.2 ± 3.6 nm (PDI 0.22, ζ-potential −37.3 ± 1.8 mV), with an encapsulation efficiency of 69 ± 1%. NPs with a low concentration of HA (1 mg/mL) resulted in increased PDI, and coating with a high concentration of HA (4 mg/mL) caused a large increase in particles size, so the intermediate HA concentration (2 mg/mL) was chosen. Both coated and uncoated NPs were tested following either a single intravitreal injection or 2x daily topical administration (50 μL) for 14 days in STZ rats. Decreases in neovascularization and nearly normal retinal physiology (vasculature integrity, full retina and basement membrane thickness) by light microscopy and transmission electron microscopy analysis were observed in both coated apatinib NPs following both treatment routes compared to untreated controls. Furthermore, relative concentration of NPs in the retina were evaluated following fluorescent labeling and confocal microscopy. Following a single dose in either normal or diabetic rats, there was a significant increase in NP accumulation in diabetic compared to healthy rats. Additionally, HA coating promoted increased retinal uptake in diabetic rats compared to uncoated NPs, and intravitreal injection provided 3-fold increased accumulation compared to a single topical drop [164].

Sharma et al. describe a method for delivering the anti-angiogenic 5-fluorouracil (5-FU) via encapsulation in lipid nanocarriers with a chitosan coating to modulate drug release [165]. 5-FU NLCs were 163.2 ± 2.3 nm with a positive surface charge ζ-potential 21.4 ± 0.5 mV to promote interactions with negatively charged mucins on the corneal surface. Once daily topical administration (80 μL) for 20 days to STZ-induced diabetic rats followed by fundus imaging to visualize neovascularization showed that NLCs of 5-FU had a decrease in retinal vessel thickness compared to controls and an improvement relative to water-based 5-FU delivery [165]. The study showed potential for topically administering small molecule anti-angiogenic agents as an alternative to anti-VEGF biologics for treating DR, however future tests validating their head-to-head efficacy should be pursued, as well confirmation of delivery to the posterior segment in larger eyes.

5.3.2. Subconjunctival Injection

Rong et al. formulated a novel controlled insulin delivery system for managing DR with chitosan-coated insulin nanoparticles within a PLGA-PEG-PLGA hydrogel [166]. Various levels of retinal neuroprotection were observed following the single subconjunctival of the insulin hydrogel compared to an aqueous insulin, empty hydrogel, and controls at 4 weeks post-injection. STZ rats had scotopic b wave ERG amplitude and retinal thickness and morphology, including ONL thickness, resembling healthy controls, whereas aqueous insulin and blank hydrogels showed a decrease in retinal function by ERG and decreased retinal layer thickness. Further, the neuroprotective effects of the insulin hydrogel injection were evaluated by TUNEL assay, showing decreased levels of apoptotic cells in the retina of rats, decreased GFAP and VEGF levels and increased occludin levels compared to positive controls. These data accompanied with the ERG results show that a single subconjunctival injection of a nanosized insulin hydrogel can provide sustained structural and functional retinal neuroprotection; however confirmation of sufficient intraocular bioavailability of insulin following subconjunctival injection in a larger animal eye is necessary [166].

Pandit et al. developed a chitosan-coated PLGA nanoparticle containing bevacizumab for subconjunctival injection [167]. Nanoparticles were 222.3 ± 7.8 nm (PDI 0.19 ± 0.08) and were compared to subconjunctival injection of bevacizumab. In a pharmacokinetic study performed in rats, the delivery of bevacizumab nanoparticles led to increased accumulation in the vitreous humor and retina compared to commercial formulation. Further, there was lower systemic bevacizumab exposure over the course of the 6 weeks following injection. In total, the AUC_0–6 weeks was nearly double that of the commercial formulation in the retina and vitreous humor. Additionally, comparing the NP and commercial injections at 12 weeks following a single injection, there was a significant reduction in retinal VEGF levels with the NP formulation by ELISA, showing that the increased retinal and vitreous levels of bevacizumab translated to increased efficacy [167].

5.3.3. Intravitreal Injection & Implants

Qiu et al. developed nanoparticles of the PPARγ inhibitor, fenofibrate, for the management of dyslipidemia, while also providing anti-angiogenic properties [168,169]. PLGA nanoparticles were prepared by oil/water emulsification and then subsequent probe sonication. Various PLGA molecular weights (5, 18, 34 and 54 kDa) were employed to determine optimal drug loading, entrapment efficiency, and release profile. Encapsulation with the varying molecular weight PLGA led to similar sized particles (around 250 nm with PDI < 0.1) with nearly neutral surface charge. The high molecular weight PLGA (34 and 54 kDa) led to the highest drug loading (6 and 7.9%) and encapsulation efficiency (35 and 47%) respectively. The low molecular weight PLGA NPs had the fastest in vitro release rates (100% release within 1 week and 1 month), whereas the 34 kDa PLGA prolonged the release rate to 2 months. Since the 54 kDa PLGA led to a slow release with only ~70% released within 2 months, the 34 kDa PLGA NPs were selected for future in vivo testing, where fenofibrate was detected in the vitreous humor and retina 2 months following a single intravitreal injection. Injection of fenofibrate NPs provided a significant increase in scotopic a and b-wave amplitudes compared to STZ-induced and blank NP injected rats 2 months after injection. Further, fenofibrate NP injection led to a decrease in retinal VEGF and intercellular adhesion molecule 1 (ICAM-1), a marker for retinal inflammation. They also observed a decrease in total retinal thickness and retinal leukostasis as evidenced by adherent leukocytes to retinal vasculature. Increased retinal vascular health was further indicated by a significantly decreased CNV area in a laser-induced model in rats. Overall, they showed that a single intravitreal injection of the anti-inflammatory fenofibrate to treat inflammatory stress characteristic of DR [168].

Won et al. combined dexamethasone (anti-inflammatory) and bevacizumab in an intravitreal implant [170]. They designed a 3D-printed implant with an alginate hydrogel core containing dexamethasone (DEX) for rapid release and a PCL shell containing bevacizumab (BEV) for more sustained release. As inflammation and angiogenesis are highly intertwined processes, they hypothesized that the rapid release of DEX to alleviate inflammation associated with angiogenesis would act synergistically to improve the potency of the longer-duration release BEV [171]. To that end, in vitro release studies showed the implant provided burst DEX release in the first 3 days, with BEV release over the following 2 months. Intravitreally implanted drug-loaded rods were then compared to DEX or BEV injections alone in rabbit eyes and showed that while free DEX was fully cleared within 24 hours, the implant provided detectable DEX concentrations one week post implantation. Further, free BEV was detectable in ocular tissue 2 weeks following injection, while the implant provided measurable concentrations in ocular tissue for 2 months following implant insertion, greater than the approximate 1-week half-life of intravitreally injected bevacizumab [172]. In the rat CNV model, they demonstrated reduced IB4 expression and decreased CD45+ leukocytes in the implant group for up to 1 month, whereas the effects of intravitreal bevacizumab lasted for only 2 weeks. Continued reduction of CD45 is due to DEX loading in the drug rod (Figure 9) [170].

Figure 9.

Fluorescent microscopy images demonstrating the effects of intravitreal bevacizumab vs. a DEX-eluting rod (A) 2 weeks and (B) 4 weeks after CNV induction. Isolectin-B4 (IB4) staining for vasculature and CD45 staining for immune response. Intravitreal bevacizumab compared favorably with the DEX-eluting rod in reduction of neovascularization at 2 weeks, however the DEX-eluting rod provided prolonged reduction of CNV out to 4 weeks post CNV induction and reduction of immune response. Reprinted with permission from [170].

5.4. Current Clinical Trials

Two Phase 3 trials compared the efficacy of the anti-angiogenic brolucizumab to aflibercept in the KESTREL and KITE trials [173]. Variable dosing regimens demonstrated that 5 loading doses every 6 weeks followed by injections once every 12 weeks was non-inferior to aflibercept given as 5 loading doses every 4 weeks followed by injections once every 8 weeks. In general, patients receiving brolucizumab demonstrated significant increases in visual acuity by BCVA letter score with injections once every 3 months, providing an additional benefit aflibercept with less frequent dosing [173]. Additionally, OcuTerra Therapeutics sponsored the Phase 2 trial of a topical ophthalmic formulation of OTT166, a novel integrin inhibitor that acts to reduce angiogenesis, inflammation, and exudation characteristic of DR [174,175]. They analyzed the effect of twice daily dosing at either a high (5%) or low (2.5%) dose compared to vehicle. The results demonstrate that OTT166 was well tolerated in patients and, in 37% of treatment responders, there was a mean reduction of central retina thickness of 47.3 μm at 28 days (end of treatment) and 60.5 μm at 56 days (end of study). However, it was unclear why some patients responded to treatment and others did not, and with the short treatment duration, further experimentation is required [175]. Patient recruitment for less invasive routes of administration is underway. Bayer is sponsoring a trial for runcaciguat to be delivered orally to patients once daily [176,177]. Runcaciguat is a guanylate cyclase stimulator as a regulator of the nitric oxide cycle, blood pressure and protecting retinal vasculature from ischemic damage [177,178]. Oculis started the second stage of its Phase 3 trial for OCS-01, a 15 mg/mL topical dexamethasone suspension. Stage 1 of the Phase 3 trial showed a significant improvement in visual acuity measured by BCVA letter scores through the third month of treatment when compared to a control group [179]. Finally, AsclepiX Therapeutics concluded a Phase I/IIa trial studying the safety and bioactivity of AXT107, a peptide designed to block angiogenesis by inducing homeostatic mechanisms. Derived from collagen IV, AXT107 forms a gel in situ and was injected intravitreally at a low (0.1 mg), mid (0.25 mg) and high (0.5 mg) doses [180].

6. Retinitis Pigmentosa

6.1. Background

RP is a group of hereditary diseases affecting over 2 million people, resulting in dim light vision loss early in progression, and in many cases, complete blindness [181]. Patients often experience night blindness in early life and adolescence, with complete blindness onset by early adulthood. Disease progression in RP is driven by genetic mutations that result in the death of rod photoreceptors. The biochemical cause of RP is heterogeneous, with over 3,000 mutations in 70 genes related to visual transduction, the visual cycle, photoreceptor structure, folding, or transcription [182]. Mutations affecting Rho, USH2A, and RPGR account for roughly 15–30% of RP cases [183]. Within the same gene, a heterogeneous group of mutations has been identified, causing different rates of disease manifestation and progression [184]. Many of these mutations lead to an initial death of rod photoreceptors, followed by a gradual degeneration of cone photoreceptors [185]. The mechanisms of rod photoreceptor death are diverse; death can result from dysfunction in the RPE, rod outer segment, or ciliary structure [186]. Previous studies have indicated that regardless of the mutation causing rod degeneration, oxidative stress contributes to the gradual degeneration of cones [185,187]. Rod cells are highly metabolically active and greatly outnumber cones by about 20 fold in the human eye [188]. When rod degeneration occurs, there is a marked decrease in the consumption of oxygen, leading to higher concentrations of oxygen in the outer retina. The increased concentration of oxygen leads to the formation of reactive oxygen species (ROS) that cause oxidative damage and gradual degeneration of cone photoreceptors [185]. While the broad underlying genetic causes make the development of gene therapies to address the patient population challenging, the first ever approved gene therapy was designed to treat a specific mutation causing a severe form of RP [189]. Further, more generalizable therapies, including retinal implants, and small molecule pharmaceuticals have shown promise in clinical trials for vision preservation [190–192].

6.2. Currently Approved Treatments

While the first approved gene therapy was an exciting advancement for a small subset of RP patients, for most cases of RP, there is no treatment available. Since RP is an inherited disease, most care is focused on preserving vision for as long as possible through nutritional supplementation or visual aids. Gene therapy and other forms of personalized medicine have shown promise, but typically do not result in a complete return to vision and require repeated treatments [193]. Luxturna^™, the first clinically approved gene therapy for ocular disease, is prescribed for patients with RPE65-based retinal degeneration [194]. In a randomized open-label phase 3 trial, an adeno-associated virus (AAV) vector containing RPE65 under control of a modified promoter alongside was injected subretinally into patients with RPE65-mediated retinitis pigmentosa [194]. Of the interventional group, 65% passed a multi-luminance mobility test for maximal improvement in vision. Despite progress in personalized medicine, RPE65-based mutations are rare, and the use of Luxturna is limited to patients with the mutation early in disease progression.

Retinal prostheses are an alternative form of treatment that directly activate RGCs for neural stimulation without the need for photoreceptor activation, as the inner layer of the retina remains functional in RP [195,196]. The Argus II Retinal Prostheses System was the first FDA-approved vision device for retinal disease. A microelectrode is first surgically placed on the surface of the retina. Images are converted into electrical signals transmitted into the electrode to directly activate the inner retina via goggles [197,198]. However, a post-approval study revealed side effects were common. Of the 47 enrolled patients, 8 experienced conjunctival dehiscence, corneal melt, corneal opacity, and retinal tear [197]. The product has been discontinued as of 2023 due to financial difficulties. A second-generation technology, the Orion Visual Cortex Prosthesis System by Second Sight Medical Products, recently underwent feasibility testing in 6 RP patients [199,200]. Despite all the progress in RP clinical drug development, it remains an unmet medical need that requires new innovations to find alternative therapeutic strategies and cover a wider range of patient populations.

6.3. Current Outlook of Preclinical Research

Both disease-specific and symptomatic treatments are currently in preclinical or clinical evaluation. An emphasis has been placed on personalized medicine due to the heterogeneous pathogenesis and disease progression across RP diseases. Mutations in the rhodopsin gene are among the most common causes of RP. Because of their genetic malleability, mice and rats are the most common models with rhodopsin mutations [201]. Therefore, we review a selection of representative therapeutic strategies for the treatment of RP mainly in rodent models. Interventions and gene therapies that target the cause of photoreceptor loss directly hold great promise in the treatment of RP. In addition, cell therapies and therapies that aim to rescue vision by electrically stimulating the preserved inner retinal circuits with prosthetic implants have garnered significant attention in recent years. The out-of-pocket cost of emerging gene and cell therapies is counter-balanced by developing cheaper but effective options, which could help disadvantaged populations. Antioxidants and nutritional supplements have long been studied for the treatment of RP, but with controversial results.

6.3.1. Implants and Prosthetics

Encapsulated cell therapies (ECT) involve the implantation of a cell-based capsule, allowing for the production of therapeutic molecules directly in the target tissue [202]. There is potential in ECT, as the technology allows for the incorporation of any gene into the cell, leading to the controlled production of therapeutic or regulatory proteins [203]. Further, the cells can be programmed to release the therapeutic molecules at a controlled and stable rate [203]. Cells were placed in a semipermeable barrier that blocks immune response but allows for nutrient and therapeutic exchange. In a Phase 2 trial, ECT implantation of NT-501 provided consistent vitreous levels of ciliary neurotropic ciliary factor (CNTF) over a 2-year period showing the promising pharmacokinetics of this technology [203]. However, further trials to elucidate whether the promising CNTF levels translated to the protection of visual acuity in RP patients did not result in any therapeutic benefit despite effective delivery [204]. The depleting release of therapeutic over time and lack of a termination mechanism remain challenges that require further engineering for future consideration [205].

Modification of cell-based encapsulation has shown promise in higher-order animal models for the treatment of RP [206]. Wong et al. constructed an alginate-collagen composite gel for delivery of glial-cell derived neurotrophic factor (GDNF) with production modulated by oral doxycycline (Dox) as a biosafety switch [206]. Gels were synthesized with 2 mg/ml collagen, 1.5% high MW alginate, and 5 × 10⁴ HEK293 cells/gel. Cells were equipped with a tet-on regulated pro-caspase 8 gene switch to allow for inducible apoptotic control with Dox and cell death was inducible with 20 pg/ml Dox. Gels were injected intravitreally and were surgically removable, and GDNF production was controllable with Dox. The therapeutic showed efficacy in Royal College of Surgeon rats (RCS), a prevalent animal model for hereditary retinal dystrophies [207]. The results demonstrated the prevention of photoreceptor apoptosis via TUNEL assay and preservation of scotopic and photopic electroretinogram (ERG) performance (Figure 10). Further implementation in higher-order animal models remains to be tested [206].

Figure 10.

(A) Apoptotic termination of CAC ECT gel in the eye upon oral delivery of Dox. Relative viability and appearance of gels retrieved from RCS rats treated with normal or 3 mg/ml Dox-containing drinking water for 48 h, starting from post-injection day 5. Left panel: Viability of cells in retrieved gels determined by MTS assay (mean − SD, n = 3 independent samples with duplicates for no Dox and 3 mg/ml Dox respectively). ***p < 0.0005 by Mann-Whitney. (B) Appearance and distribution of living, dead and apoptotic cells inside CAC ECT gels retrieved from animals receiving various dosages of Dox in a follow-up study. Top row: Gel morphology (n=3) Middle row: Live/Dead assay showing the distribution of living (green)and dead (red) cells in retrieved gels (n=3). Bottom row: TUNEL assay with DAPI nuclei counterstain (blue) showing the distribution of apoptotic cells (green) in retrieved gels (n=3). Scalebar, 100 μm. (C) Therapeutic outcomes of CAC ECT gel treatment on rats with inherited retinal degeneration. Representative scotopic (upper panel) and photopic (lower panel) electroretinogram waveforms showing the retinal function of dystrophic rats receiving one or two units of GDNF-secreting gel, gel control with non-functioning HEK293 cells, sham control, and no treatment at P56. Reprinted with permission from [206].

Organic semiconductors have been used as interfaces for neuronal photo-stimulation in retinal disorders including RP [208]. Conventionally, 2-dimensional devices composed of conjugated polymer (such as poly[3-hexylthiophene], P3HT) are supported onto a silk fibroin for implantation. When implanted in the subretinal space of RCS rats, a rescue of vision with sensitivity in the daylight was observed for up to 6–10 months. Maya-Vetencourt et al developed P3HT NPs about 300 nm in size and injected them subretinally in the RCS rat model to mediate light-evoked stimulation of retinal neurons and persistently rescue visual functions [208]. P3HT NPs spread over the entire subretinal space and did not migrate toward the inner retina and maintained subretinal distribution up to 8 months after injection. P3HT NPs were not endocytosed by neuronal cells (photoreceptors), but engulfed the cells, creating a close contact between the NP surface and the cell membrane. P3HT NPs displayed long-term rescue of visually evoked behavior in RCS rats compared to untreated rats, or rats injected with 300 nm glass beads. The potential for this high spatial resolution P3HT NPs may extend their applicability to earlier stages of RP and atrophic AMD [208].

6.3.2. Subretinal Delivery

Since RP is a genetic disease, much effort has been placed on gene delivery to correct the underlying genetic mutation. Delivery of gene therapies requires a vector to efficiently transfer nucleic acid into the cell, as nucleic acids alone show low cell transduction and rapid degradation. Numerous barriers to effective delivery from systemic circulation make local injection a more viable option [209]. Two classes of vectors are generally used: viral and nonviral [210,211]. Adeno-associated viral (AAV) vectors are the most studied for viral retinal therapies due to their well-characterized safety and efficacy. AAVs package a linear single-stranded DNA genome of up to about 5 kb and can effectively transduce non-dividing and post-mitotic cells, such as retinal photoreceptors [212]. Different AAV serotypes have different transduction efficiency and tropism profiles. Luxturna, the first clinically approved gene therapy, consists of an AAV2 vector that effectively transduces outer retinal components critical to RP disease progression, including rod photoreceptors and RPE cells. AAV-mediated retinal gene therapy was reported to induce intraocular inflammation and weakened efficacy after initial improvement. This could be attributed to targeted removal of transduced cells via anti-viral defense mechanisms. AAV has been shown to activate innate pattern recognition receptors (PRRs) such as toll-like receptor (TLR)-2 and TLR-9 resulting in the activation Type I interferons pathway. AAV can also induce capsid-specific neutralizing antibodies and T cell responses which both limit the therapeutic effect. However, viral gene vectors take advantage of the “immune privileged” retinal space, meaning that induction of immunogenic and inflammatory responses in the retina after injection of viral vectors are expected to be less severe compared to systemic delivery [213]. More detailed characterization of the immune responses following viral vector transduction in the eye of different species are comprehensively reviewed elsewhere [214–217].

The development of genome editing with CRISPR (clustered, regularly interspaced, short palindromic repeats)/CRISPR-associated protein 9 (Cas9) opened a new avenue for treatment. CRISPR-Cas9 can be used to create site-specific cleavage within DNA and mammalian cells to knock out or disrupt disease-causing genes, or knock-in genes by inserting new DNA sequences through homology-dependent and independent pathways [218,219]. Due to its large size, Yu et al. encoded CRISPR-Cas9 components (i.e. Cas9 and guide RNA (gRNA)) into dual AAV vectors and subretinal injection prevented retinal degeneration by preserving cone functions [220]. In three mice models of retinal degeneration, CRISPR-Ca9-mediated disruption of the Nrl gene, a rod fate determinant during photoreceptor development, resulted in rods acquiring partial cone features and prevented secondary cone degeneration [220]. McCullogh et al. later demonstrated that AAV-based CRISPR/Cas9 components could remove GUCY2D mutations, knock out the gene encoding retinal guanylate cyclase-1 (retGC1), and improve retinal function and structure in a macaque model [221]. Macaque eyes were injected subretinally with AAVs containing gRNAs and Cas9 RNA, and a 13% editing rate was seen in photoreceptors, demonstrating somatic gene editing in primates as a candidate for treating inherited retinal diseases. Importantly, there was minimal evidence of preexisting or Cas9-specific T-cell immune responses and inflammation within the eyes [221]. Keiichiro et al devised a new CRISPR-Cas9 technology for the homology-independent targeted integration (HITI), which allowed robust DNA knock-in in both dividing and non-dividing cells, in vitro and in vivo. In the RCS rat, HITI-CRISPR-Ca9 was used to insert a copy of the Mertk gene that is originally missing in the RCS rats [222]. The AAVs (serotype 8 or 9) were injected in the subretinal space of RCS rats 3 weeks after birth and analyses were performed between 7–8 weeks. Microscopic evaluation revealed that about 2 folds increase in the ONL thickness was observed in the HITI-treated group compared to untreated rats, or those receiving conventional CRISPR-Cas9 and homology-directed repair (HDR) DNA donor templates [222].

Leber congenital amaurosis type 10 (LCA10) is a form of severe retinal dystrophy with no FDA-approved treatment. The IVS26, an adenine to guanine point mutation that creates a novel splicing donor site, is the most common LCA10 causative mutation, resulting in complete inactivation of CEP290 which is required for phototransduction and outer segment regeneration [223]. Although gene augmentation or replacement is a promising strategy for treatment, the large size of the CEP290 coding sequence (~7.5 kilobases) exceeds the packaging capacity of AAV. To overcome the small loading size of AAV vectors, Editas Medicine targeted the CEP290 IVS26 mutation to correct the CEP290 splicing defect. AAV5 vector encoding Cas9 and CEP290 gRNA was delivered to photoreceptors following subretinal injection. Specifically, 1 × 10¹¹ to 1 × 10¹² vg/ml was injected in humanized CEP290 mice and resulted in editing rates of >10% 6 months, with productive editing rates reaching a maximum of 60.8 ± 30.2% with a 3 × 10¹² vg/mL dose. Further, non-human primates were injected with 100 μL AAV5 subretinally as a representation of the level of productive editing that would be observed in the human retina. Productive editing in non-human primates reached 27.9 ± 20.7% with the highest concentration (1 × 10¹² vg/mL), exceeding the expected 10% productive editing rate that would be the minimum threshold for efficacy. The natural tropism of AAV5 for photoreceptor cells and the use of the photoreceptor-specific GRK1 promoter limit the expression of the CRISPR/Cas9 system to only the therapeutic target tissue and cell type [223]. However, generally, cytotoxicity and immunogenic response are possible with the delivery of prokaryotic components to the retina, and off-target mutagenic effects may still occur as the Cas9 nuclease can tolerate up to 3 base pair mismatch between the gRNA and genomic target sequence [224,225]. Hence, the safety spectrum of genome editing in the scope of treating retinal degenerative diseases is yet to be characterized in the near future.

Non-viral gene delivery systems offer multiple advantages over viral vectors by controlling gene transfection kinetics and reducing toxicity, and immunogenicity [226]. A genomic DNA vector encoding mouse or human full-length rhodopsin genomic locus (endogenous promotors, all introns, and regulatory sequences) was compacted into nanoparticles (NPs) and delivered into rhodopsin knockout (RKO) mice [227]. The NPs were formulated by mixing the DNA with a diblock copolymer composed of 10 kD PEG and poly-l-lysine (CK30PEG) and decorated with the cell-penetrating peptide TAT. Subretinal injection resulted in long-term high levels of physiological transgene expression over a period of 5 months. The treatment partially rescued retinal electroretinography responses and retinal structure in RKO mice more significantly compared to complementary DNA (cDNA) encoding for the rhodopsin gene (reverse transcribed from rhodopsin mRNA) (Figure 11) [227].

Figure 11.

gDNA NPs delivery partially rescued retinal structure in RKO mice after 5 months of treatment. (A) Representative histology of H&E stained retinal cross-sections from gDNA- and cDNA-treated RKO mice at PI-5 months. WT and saline-injected RKO were used as controls. Scale bar, 200 mm. (B) Quantification of the number of rows in the ONL. Data measuring the number of rows in the ONL in gDNA-treated RKO mice compared to cDNA-treated RKO mice represent mean ± SEM, n = 3–5, *p < 0.05, **p < 0.01. ONH, optic nerve head; ONL, outer nuclear layer; INL, inner nuclear layer; IS, inner segment; OS, outer segment. Reprinted with permission from [227].

Oligonucleotides therapeutics represent a promising class of drugs, which can be used to interfere with RNA transcripts, modulate pre-mRNA splicing, induce mRNA knockdown, or block translation of disease-causing genes [228]. Small interfering RNA (siRNA) and antisense oligonucleotides (ASO) have been rationally designed to suppress retinal disease-causing genes. Their relatively small size (4–14 kD) and hydrophilic nature make them suitable for crossing retinal barriers by intravitreal injection, surpassing the need for the more invasive subretinal injection [229,230]. Many of the ASO used for the treatment of RP, including those in clinical development and trials, are intended for intravitreal injection (see section 6.3.3). Nonetheless, stable expression of short hairpin RNA (shRNA) using AAV vectors to induce RNA interference has been reported in the treatment of RP [231]. Mutations in the human GUCA1A gene expressing the Ca²⁺-binding protein GCAP1 are associated with autosomal dominant cone dystrophy. The authors first screened a range of short hairpin RNA (shRNA) in vitro to identify an efficient anti-GCAP1 shRNA, and then showed that AAV2/8-mediated delivery of the most potent shRNA could suppress the mutant GCAP1 in photoreceptors up to 1 year post subretinal injection in a mouse model expressing bovine GCAP1 [231].

mRNA delivery emerged as a new tool for the efficient introduction of therapeutic proteins or genome-editing nucleases. Non-viral delivery of mRNA can be used as an alternative to AAV-based delivery of CRISPR/Cas9, as these degradable particles can encapsulate larger cargo for direct protein production in non-dividing cells with better safety profiles [210,232]. Herrera-Barrera et al. have developed a peptide-guided lipid nanoparticle (LNP) that provided delivery of mRNA to the retina following subretinal or intravitreal injection [233]. Optimized peptides for delivery to the neural retinal were generated through phage display. The peptides were then validated by evaluating intracellular uptake by ARPE-19 and 661w cells. After intravitreal or subretinal injection in BALB/c mice, the peptide MH42 showed localization in the RPE with the highest preservation of photoreceptor inner and outer segments. Further, MH-42 conjugated LNPs improved transfection in the photoreceptors, glial cells, and RPE, whereas unconjugated LNPs only were able to transfect the RPE. Subretinal injection in NHP further recapitulated these results, demonstrating expression in the RPE, photoreceptors, and glial cells 48 hours after injection, highlighting the potential of targeted LNP-based nonviral strategies for gene therapy [233]. Devoldere et al compared the fate and delivery efficiency and distribution of green fluorescent protein (GFP)-expressing mRNA between subretinal and intravitreal injections [234]. Although subretinal injection of mRNA-complexed with lipoplexes resulted in evident GFP expression in the photoreceptor and RPE cell layers in mice, intravitreal injection of the same mRNA-lipoplex showed only minimal GFP expression. The authors also assessed the role of the inter-limiting membrane (ILM) in the mRNA delivery efficiency in ex vivo bovine explants. Interestingly, only when the ILM was removed, mRNA expression was apparent and seemed to predominantly colocalize with Müller cells after intravitreal injection. These findings highlight the importance of the ILM as a significant barrier to the non-viral delivery of mRNA in primates [234].

6.3.3. Intravitreal Delivery

Intravitreal injection bypasses the corneoscleral barriers, but still, the therapeutic agent needs to cross the inner retina to reach the outer retina for the management of RP. Importantly, the ILM separating the vitreous from the retina is an additional barrier to targeting the retina, with more serious implications in bigger eyes including primates. The structure of the ILM varies among species, as animals with larger eyes typically have a more robust ILM [235]. Therefore, extra attention is needed when designing gene therapies by intravitreal injection, since the results obtained in mice and rats may not translate effectively in primates due to the dramatic difference in the ILM structure. In addition, the relative lack of immune privilege in the vitreous represents another concern for many viral gene therapy vectors [235,236]. Thus, various delivery approaches are in development to achieve effective delivery with intravitreal injection [237].

Rosa et al. studied the effects of an intravitreal injection of stanniocalcin-1 (STC-1), a polypeptide associated with regulating cell death and survival, angiogenesis, and inhibition of oxidative damage [238,239]. STC-1 was injected biweekly in transgenic P23H pigs and outputs on retinal function and structure were monitored. They found that biweekly STC-1 from P15 to P90 led to an increase in cone function by photopic ERG, greater ellipsoid zone width, and larger outer nuclear layer thickness compared to saline injection. Increases in photoreceptor-associated gene expression (phosducin, recoverin, GNAT-2, MWL, and SWL) and decrease in markers of oxidative stress (protein carbonyl content and nitrotyrosine levels) were also observed following STC-1 treatment [238]. Further studies evaluating the mechanism of this photoreceptor protection are needed. Dalkara et al. engineered an AAV2 vector with a random 7-peptide long sequence optimized through directed evolution to overcome poor diffusion and transduction in the outer retina and ILM [240]. A library of 7mer peptides was screened for insertion into AAV2 and tested via GFP expression after intravitreal injection into mouse retinas. The library was then reduced to a single dominant sequence of LALGETTRP which showed pan-retinal gene expression. The vector was tested in male cynomolgus monkeys and right and left eyes were injected with 5 × 10¹² viral particles. The vector effectively transduced the outer retina from an intravitreal injection in primates after 3 weeks as determined by GFP expression [240]. The technology is currently being used in Phase 2 clinical trial [241].

A lysine-conjugated polyethylene glycol (CK30PEG) nanoparticle to deliver plasmid DNA is in development by Copernicus Therapeutics [242]. The CK30PEG nanoparticles were injected subretinally or intravitreally in non-human primates and evaluated for distribution, tolerance, efficacy of gene transfer, expression levels, and any signs of immunogenicity with a GFP reporter gene under the control of 3 different RPE or photoreceptor-specific promoters. DNA concentration was held constant across each reporter (4.3 μg/μL), and NPs were rod-shaped, with a length of 200 nm and width of 8–11 nm, and near neutral ζ-potential (−3 – 1 mV) for each NP. GFP expression in the anterior and posterior chambers was measured, and only the DNA-NPs enhanced GFP expression in the posterior segment compared to naked NPs or saline. Subretinal injection led to 50% higher expression in the retina compared to intravitreal injection of DNA-NPs 15 days after injection, and there was no off-target GFP expression in the optic nerve or anterior chamber (cornea or lens). High expression of GFP was primarily detected in the quadrants where they were injected, whereas intravitreal-injected DNA-NPs led to expression in all four quadrants of the retina. Subretinal and intravitreal injections were well-tolerated, as indicated by little change in scotopic and photopic ERG b-wave amplitudes, no upregulation of pro-inflammatory cytokines, and no measurable accumulation in the brain [242]. Importantly, they found that the RPE was easily transfected preferentially to photoreceptors, and future studies will be employed to improve photoreceptor uptake of DNA-NPs.

Niosomes are nonionic surfactants used to deliver genes, benefiting from low-cost materials, and low toxicity [243]. Niosomes were shown to deliver plasmid DNA (pDNA) to RPE and outer retina cells in a rat model, motivating their further study as a low-cost alternative to viral delivery [244]. Transfection efficacy was low compared to viral vectors, but further work showed that modification of niosomal components increased efficiency [245]. Niosomes were composed of polysorbate 80 as a non-ionic surfactant, squalene as a helper lipid, and dioleoyl trimethylammonium propane (DOTAP) as a cationic lipid, and functionalized with hyaluronic acid, a glycosaminoglycan ligand for the CD44 receptor overexpressed in degenerating retinas [246,247]. The size of the cationic niosome increased from 134 nm to 154 nm after hyaluronic acid coating. This was further confirmed by the reversal of the ζ-potential from 41 mV before coating, to −31 mV after coating. pDNA loading into niosomes led to an increase in particle size (154.07 ± 0.40 nm and 286.20 ± 2.19 nm, respectively) with no change in surface charge. GFP transfection efficiency was observed in retinal layers following intravitreal injection in rats 3 days post-injection [245]. Cationic niosomes have shown promise for the delivery of genes in a rat model [248]. Niosome formulations were prepared using poloxamer 188, polysorbate 80, 2,3-di (tetradecyloxy) propan-1-amine cationic lipid with and without chloroquine diphosphate (CQ) incorporated in the niosome. CQ was described to play multiple roles in enhancing pDNA delivery efficiency, by enhancing lysosomal escape and protecting DNA from degradation [249]. The incorporation of chloroquine diphosphate resulted in a slight increase in particle size of 118.2 ± 1.5 nm as compared to the chloroquine-free niosomes (90.4 ± 0.7 nm). CQ-niosomes showed transfection efficacy with both 1 μL subretinal and 4 μL intravitreal injections at 25 ng DNA/mL into rat retinas in vitro as characterized by EGFP expression. However, intravitreal injections showed expression in only the GCL, INL, and outer plexiform layer as opposed to expression throughout the whole retina with subretinal injections. ONL transfection remains as a challenge for intravitreal nonviral delivery methods [245].

A novel antisense oligonucleotide (ASO) was designed to induce specific exon skipping in USH2A mutations, one of the most common causes of RP [250]. These mutations commonly reside on exon 13 of USH2A gene, leading to the formation of a premature stop codon after RNA splicing and a shortened version of the usherin protein. Six exon-13-skipping ASO constructs were synthesized and the most potent 2 ASOs s were injected in mutant zebrafish larvae. The ASOs resulted in successful exon skipping, normal protein production, and a complete restoration of retinal function as measured by ERG. The optimized lead to induce human USH2A exon skipping, namely QR-421a, showed efficient retinal uptake after intravitreal injection in wild-type C57BL/6J mice. QR-421a sustained a 50% exon 12 skipping rate for more than 6 months, equivalent of human USH2A exon 13. However, no efficacy data were obtained in a rodent or higher-order animals [250]. Intravitreal QR-421a is currently undergoing a phase 2/3 study to evaluate the efficacy safety and tolerability in subjects with RP due to mutations in exon 13 of the USH2A gene with early to moderate vision loss [251].

Compounds with antioxidative properties have been formulated as sustained-release microspheres to prevent retinal cell degeneration. Oveson et al. showed evidence that the antioxidant tauroursodeoxycholic acid (TUDCA) could slow the disease progression of RP in rd10 mice and mice subjected to prolonged light exposure [252]. Intraperitoneal injection with 500 mg/kg of TUDCA dissolved in 0.15 M NaHCO3 in rd10 mice every 3 days starting at postnatal day (P) 6, resulted in the preservation of cone cells. The TUDCA-treated mice had significantly greater scotopic and photopic ERG performance. Fernandez-Sanches et al. continued this work with TUDCA-loaded microspheres [253]. TUDCA was encapsulated in biodegradable PLGA microspheres and injected intravitreally into transgenic rats. 40 mg of TUDCA were suspended in 1 mL of PLGA solution in methylene chloride (20% w/v), resulting in a TUDCA:PLGA ratio of 1:5. The TUDCA-PLGA microspheres had a mean particle size of 22.9 ± 0.04 μm. Sustained release was observed over 28 days in an in vitro release assay where 40.5% of the total drug was released. TUDCA microspheres preserved cone photoreceptors as shown by cone arrestin staining with monthly injections. At P90, the maximum scotopic amplitudes obtained for a- and b-waves were 28% and 22% higher, respectively, in TUDCA-PLGA-microsphere treated eyes relative to a control [253].

RHO^P23H, a point mutation changing proline to histidine results in misfolding and complexing of rhodopsin with valosin-containing protein (VCP), is most common mutation in North American populations of RP [254]. VCP inhibition has previously been shown to slow photoreceptor death, motivating its reformulation for controlled release in a rat model [255]. Sen et al. reformulated VCP inhibitors for controlled intravitreal release to avoid poor solubility and rapid clearance by ocular barriers. Nanoformulations using self-assembling 5 kDa cholane- and cholesterol-PEG improved solubility and prolonged release of a VCP inhibitor over the course of 10 days. mPEG-cholane NPs were 54.8 ± 8.4 nm with ζ-potential −12.5 ± 1.6 mV, and mPEG-cholesterol nanoparticles were 31.9 ± 3.1 nm with ζ-potential −14.1 ± 1.5 mV. They found that exposure of explanted RHO^P23H retinas to the VCP inhibitor NPs prolonged photoreceptor survival and increased rhodopsin expression in culture, with the mPEG-cholane particles displaying increased efficacy. Further, intravitreal injection of VCP inhibitor in NPs was also tolerated as evidenced by fundus imaging performed in wild-type rats up to 2 weeks following injection [255].

6.3.4. Suprachoroidal Delivery

Suprachoroidal gene delivery is a potential solution as a less invasive, in-office alternative to subretinal injections that circumvents issues with bypassing the ILM [256]. Transscleral microneedles have shown promise in a nonhuman primate model [257]. Yiu et al. reported that a 700 mm-long 30-gauge microneedle inserted through the sclera led to RPE and photoreceptor transduction of GFP-encoding AAV8 in macaques. Suprachoroidal, subretinal, and intravitreal injections of 7 × 10¹² vg/eye showed that while suprachoroidal and subretinal injections transfected the retina for extended periods of time in the ONL and RPE layer, intravitreal injections did not show the same pan-GFP expression. Additionally, suprachoroidal injection showed expression throughout the entire retina while subretinal transduction was limited to the injection site (Figure 12) [257].

Figure 12.

Epifluorescence image entire globes showing different patterns of GFP transgene expression. (A-C) confocal fluorescence images showing GFP transgene expression (green) co-immunostained with anti-RPE65 (red) to label retinal pigment epithelium (RPE) along with DAPI (blue) to label cell nuclei in eyes after suprachoroidal (A), subretinal (B), and intravitreal (C) injections of AAV8-CMV-GFP. Scale bars, 100 μm. Reprinted with permission from [257].

Suprachoroidal delivery of AAVs expressing GFP have also shown less immunogenicity than intravitreal injections of the same vector in a macaque model [258]. Chung et al. found that intravitreal AAV8 injections resulted in increased serum levels of neutralizing antibodies compared to suprachoroidal injections due to increased systemic exposure of the AAV vector in the spleen and liver. Further, where suprachoroidal injection of the AAV in non-human primates led to widespread GFP transduction in the retina and across the blood-retina barrier in the RPE and sclera, both intravitreal and subretinal injections led to more focused expression only within the neural retina. These studies indicate that injection of viral vectors into the suprachoroidal space may lead to increased ocular retention and decreased systemic immunogenicity compared to injection of the same vectors via subretinal or intravitreal routes [258]. The difference is likely due to intravitreal AAVs being cleared through trabecular outflow into the systemic circulation, whereas suprachoroidal AAVs are primarily cleared through uveoscleral outflow which has a lower systemic exposure [259].

6.4. Current Clinical Trials

Oral N-acetylcysteine is currently in Phase 3 after promising results in a Phase 1/2 trial. The treatment is an agnostic approach to treating RP that is gene independent. Patients will be given 1800 mg twice a day versus a placebo, and measurement of ellipsoid zone will be compared to a baseline after 4 years of treatment [191]. There are many ongoing trials for treatment of RP targeting specific mutations. Most trials use AAV-based delivery strategies due to convenience and ease of manufacturing [260]. For treatment of X-Linked Retinitis Pigmentosa, MeiraGTx has concluded a subretinal injection Phase 1/2 clinical trial targeting RGPR [261]. Additionally, 4D Molecular Therapeutics is actively recruiting for an intravitreal administration of RGPR [262]. Other targeted genes include PDE6A, PDE6B, RLBP1, and Rho [263–267]. Editas Medicine reported results from a Phase 1/2 trial for LCA10 in which three out of 14 patients experienced BCVA improvement with no adverse effects [268]. OCUGEN has developed OCU400, a gene therapeutic designed to target NR2E3 and RHO. They are currently conducting their Phase 3 trial based on promising Phase 1/2 results in which 83% (10/12) of the patients receiving the treatment displayed stable or improved BCVA scores 6 months after treatment [269]. Finally, SparingVision is in the middle of a Phase I/II study (PRODIGY) for SPVN06, their new AAV gene therapeutic for RP patients with mutations in the RHO, PDE6A, or PDE6B gene. SPVN06 encodes for both Rod derived Cone Viability Factor (RdCVF) and Rod derived Cone Viability Factor (RdCVFL). Phase I/II is expected to be complete in 2025 [270].

7. Conclusion and Future Directions

Ocular neurodegenerative diseases involve the progressive degeneration of the neuronal network in the visual pathway, leading to irreversible vision loss and imposing a substantial burden on individuals and society as a whole. The structural complexity of the eye impedes most therapeutics from reaching the back of the eye at desired concentrations, while also posing delivery difficulties and nonspecific toxicities. The BRB further hinders the accumulation of drugs administered systemically at the target site. Given the progressive nature of retinal diseases, frequent dosing is often required, necessitating greater patient compliance. While topical eye drops are commonly prescribed, there is currently no FDA-approved topical formulation to treat posterior segment diseases. Intravitreal injections have emerged as a preferred option due to their proximity to the retina and potential for localized drug action. Intravitreal implants are designed to improve patient compliance by releasing therapeutic agents over several months, reducing the need for frequent clinic visits. However, repeated injections and implants carry the risk of serious complications such as retinal detachment and dose dumping. In addition to the above, intracameral, suprachoroidal, subretinal, and subconjunctival injections are being explored for their potential in delivering active agents effectively to the target site. As the incidence of retinal diseases rises in the elderly, the need for effective delivery options becomes paramount. To address the challenges associated with existing treatment options, several novel delivery strategies are being developed and discussed in detail in this review, highlighting their translational potential. Gene and stem-cell based therapies represent another innovative approach in treating retinal neurodegenerative disorders by restoring or preserving the visual function. The success of LNP-based vaccines for COVID-19 has opened up enormous opportunities to explore their use in retinal diseases. Viral/non-viral vectors and LNPs are used to deliver therapeutic genes directly into retinal cells, aiming to correct defective genes or promote neuroprotective mechanisms. Progress in delivering gene-based formulations holds great potential in the treatment and eradication of inherited retinal diseases. In addition to the above, the development of topical formulations that minimize dosing frequency and achieve effective concentrations posteriorly is crucial. The physicochemical characteristics of these formulations including particle size, surface charge, effective surface area, residence time, muco-inert coatings, and tear-film stability play a significant role in their fate. With controlled delivery systems such as implants, it is essential to understand the degradability, distribution, clearance, and stability of the polymers used, ensuring they do not accumulate in the body after the intended drug release is achieved. Furthermore, to expedite clinical translation, scalability of the manufacturing process must be carefully considered early in development to avoid setbacks in reproducibility and product quality. Repurposing existing drugs, and surface functionalization with cell-penetrating peptides and antibodies also offer great potential in advancing the next-generation therapeutics. The increasing popularity and accessibility of artificial intelligence and machine learning models can help accelerate the development programs. Collaborative efforts among researchers, clinicians, regulatory agencies, and pharmaceutical companies are essential in developing strategies that have the potential to transform the lives of patients, helping them see better and feel better.

• Retinal neurodegeneration arises from various factors, ranging from inherited genetic mutations, a complication of a preexisting condition, or injury to the eye

• The cornea and tear film are major barriers for topical ocular delivery, while the blood retina barrier limits some drugs from accumulating in the retina from systemic circulation

• Various administration techniques are used to sustain therapeutic delivery to specific ocular compartments, bypassing physical barriers to manage chronic retinal neurodegeneration

• Enhancing therapeutic penetration, targeting, and uptake to retinal cells using nanomedicine strategies enables treatment of inherited retinal diseases

• Studying new therapeutic agents in higher order animal eyes is imperative in demonstrating translational potential due to large anatomical differences among species

Data Availability

No data was used for the research described in the article.