Hyaluronic acid-based nanoparticles to deliver drugs to the ocular posterior segment

Abstract

Ocular posterior segment diseases such as uveitis, X-linked juvenile retinoschisis, or age-related macular degeneration usually result in progressive and irreversible vision loss. Although intravitreal injection is the main way to deliver drugs to the posterior eye, it still has shortcomings as an invasive operation. Nanocontrolled drug delivery technology is a promising option to avoid frequent injections. Due to the particularity of the human intraocular structure, drugs have unique pharmacokinetic characteristics in the eye. Various nanoparticles have been successfully investigated in experimental studies for vitreous injection, with advantages and drawbacks. Here, we introduce an ideal nanopolymer modifier to build nanodelivery systems in vitreous cavities. Hyaluronic acid (HA) is a natural polysaccharide with a broad molecular weight range, negatively charged surface, ligand–receptor binding capabilities, and hyaluronidase breakdown capability. Advances in CD44 receptor targeting for HA-based nanoparticles can improve mobility and penetration in the vitreous and retina, stabilize the nanoparticles, and regulate drug release. This review summarizes the intravitreal administration of nanoplatforms based on HA and the benefits of HA in drug delivery systems.

Keywords:

• Hyaluronic acid
• nanoparticles
• ocular drug delivery
• posterior segment disease

1. Introduction

Posterior segment illness is a leading cause of irreversible visual loss that must be addressed (Meza-Rios et al., 2020; Varela-Fernández et al., 2020). Although the eye is an easy-to-reach body part, drug administration to the posterior portion remains a challenge (Alshaikh et al., 2022). Topical application is a convenient and safe method for delivering ocular medications to the surface of the eye (Wei et al., 2023). Due to the various eye barriers, it is still challenging for topical administration to efficiently transfer drugs to the posterior of the eye. Intravitreal administration, compared with systemically and topically delivered agents, can directly reach intraocular lesions without systemic toxicity. Frequent injection to maintain the concentration of the drugs in the vitreous may lead to an increased risk of complications, such as increased intraocular pressure, endophthalmitis, vitreous hemorrhage, and cataract (Patel et al., 2022). The research and development of continuous-release formulations and intravitreal implants are two critical topics for reducing the quantity and frequency of vitreous cavity injections (Cabrera et al., 2019). To insert vitreous implants, invasive surgical operations are needed, and nonbiodegradable materials may require a second operation for removal. Nanocontrolled delivery technology can improve pharmacokinetics to prolong and monitor the release of medicines by increasing their solubility, bioavailability, and stability (Nayak & Misra, 2018; Cabrera et al., 2019).

Clinical trials have been launched for a controlled release system based on poly(lactic-co-glycolic) acid microspheres to deliver sunitinib malate via vitreous injection (Mann et al., 2018). The nanoparticle drug delivery system is a promising approach that has shown the potential to extend drug retention by three to six months (Bochot & Fattal, 2012). Nanoplatforms made of proteins, lipids, or polymers have been extensively researched for the delivery of medications, which can include peptides, proteins, nucleic acids, and small molecules. Choosing a suitable carrier to modify the physical and chemical properties of nanoparticles is an effective way to reach intraocular targets.

Hyaluronic acid (HA) is a natural linear polysaccharide that was first found in the vitreous humor of bovines. It is made up of consecutive disaccharide units of d-glucuronic acid and N-acetyl-d-glucosamine that are linked together by –1,4- and –1,3-glycosidic linkages (Figure 1) (X. Zhang et al., 2021). The linear polymer chain is highly variable in length and can reach 10⁷ Da in molecular size (Gallo et al., 2019; Allawadhi et al., 2022). HA has multiple applications in oncology (Heldin et al., 2020), orthopedics, and esthetic dermatology (Graça et al., 2020; Li, Jai et al., 2021). It has been a focus in ocular drug delivery since it is a natural substance found in the fluids in the eyes. Because of their excellent properties, including stickiness, viscoelasticity, biocompatibility, biodegradability, nontoxicity, and nonimmunogenicity, HA polymers are widely studied in ocular surface-related treatments and vitreous substitutes (Wang, Chi et al., 2021; Hynnekleiv et al., 2022). Ocular drops and vitreous substitutes containing HA and its derivatives have entered clinical testing (Lambiase et al., 2017; Beck et al., 2019; Hynnekleiv et al., 2022). As far as we know, the US Food and Drug Administration (FDA) has approved HA for clinical use through topical, intra-articular, intramuscular, and intravitreal injection. HA can be utilized not only as a carrier to deliver pharmaceuticals but also as a carrier to modify nanoparticles for the design of drug delivery systems. Although there is still a long way to go before nanodrug delivery systems to approach the clinical stage, HA-based nanodrug delivery systems have demonstrated superior performance in animal research. This review highlights research on HA polymers as nanomodifiers for intravitreal drug delivery to the ocular posterior portion.

Figure 1. Chemical structure of HA.

2. Pharmacokinetic characteristics in the posterior segment

The eyeball has a distinct anatomical structure and physiological properties as a specific organ of the human body (Figure 2A). For intravitreal injection, there are several main obstacles in the vitreous and retina that need to be overcome before reaching the target cells. Pharmacological formulation effectiveness is dependent on a detailed understanding of ocular barriers and pharmacokinetics. After intravitreal administration, the drug distribution into the vitreous is affected by drug diffusion, convection, and protein binding. Previous research has examined the effects of particle physicochemical properties, vitreous liquefaction, and vitreous proteins. Intravitreal pharmacokinetics include drug distribution in the posterior segment and elimination via the ocular–blood barrier.

Figure 2. The structure of the eyeball. A: The ocular posterior segment located behind the lens, containing the vitreous body, the retina, the optic disc, the choroid and most of the sclera. The lens and the part in front of it, called the anterior segment. B: Two elimination routes after intravitreal administration. The anterior clearance route refers to the path excluded via trabecular meshwork outflow channels (yellow arrows), while the posterior route is the exit through the iris, ciliary body and retina (black arrows). C: Schematic diagrams of the retina.

2.1. Drug distribution in the vitreous

2.1.1. Drug diffusion

The vitreous is actually a complex and exquisite network made up of collagen fibers and negatively charged HA polymer (Xu et al., 2013). When certain drug complexes or formulations penetrate the vitreous, the transparent gel-like structure of the vitreous humor may act as a barrier for the diffusion of particles. Surface charge and particle size are two critical issues to consider.

Rapid diffusion of small molecular drugs and proteins is possible due to the human vitreous mesh size being in the 500 nm range (Xu et al., 2013). With age, vitreous liquefaction leads to the appearance of pockets of liquid called lacunae (Bayat et al., 2020). The homogeneous vitreous humor undergoes phase separation, becoming a heterogeneous mixture of a stiffened phase composed of aggregated collagen fibrils and a loosened phase composed of dissociated HA fragments (Tram et al., 2021). Therefore, vitreous liquefaction in elderly individuals is unlikely to have much effect on diffusion. In fact, a homogeneous vitreous network with a mesh size of 500 nm is not a strong diffusion barrier. Vitreous fluids are negatively charged, which is attributed to the prevalence of negatively charged HA. Particles with a positive charge will gather in the vitreous cavity (Koo et al., 2012; Martens et al., 2013).

2.1.2. Convection

In the vitreous cavity, convection is the movement of a portion of the fluid produced by the ciliary body toward the surface of the retina (Araie & Maurice, 1991; Ferroni et al., 2018). This procedure is carried out as a consequence of the temperature and pressure variations that exist between the anterior segment and the retina. Convection has little effect on the overall pharmacokinetics of low molecular weight (MW), freely diffusible drugs. In a three-dimensional rabbit eye model, convection could have a large effect on the time course of low diffusivity drugs (<1 × 1 0⁻⁵ cm²/s) (Stay et al., 2003; Park et al., 2005). It is thought that convection plays a large part in how nanoparticles move through the body. In older eyes, where the vitreous fluid has been broken down, diffusion may play the main role (Tavakoli et al., 2021).

Furthermore, since intraocular pressure affects convective movement, fluctuations in intraocular pressure may impact the intensity of convective movement (Park et al., 2005). Glaucoma usually causes high intraocular pressure because the outflow channels of the trabecular network in the anterior segment are blocked, and there may be more backward convective motion. Retinal detachment usually results in a decrease in intraocular pressure due to the formation of subretinal fluid, but convection is also increased because of the pressure on the retinal surface.

2.1.3. Protein corona

Protein corona formation is a constant, dynamic process in which proteins and other biomolecules in biofluids cover and surround nanoparticles (Ghazaryan et al., 2019; Mishra et al., 2021). The composition of the protein corona is affected by the biofluid environment and the physical and chemical properties of nanoparticles, such as the size, surface charge, hydrophilicity/hydrophobicity, and core components (Ke et al., 2017; W. Huang et al., 2021). Compared to hydrophilic nanoparticles, hydrophobic nanoparticles adsorb twice as many proteins (Yu et al., 2020). Most of the time, interactions in the protein corona are caused by hydrogen bonds, van der Waals forces, electrostatic forces, hydrophobic forces, etc. (Skalickova et al., 2021). The soft corona in a protein corona is composed of proteins connected by loosely connected bonds, whereas the hard corona is driven by strong bond connections. Healthy vitreous has an average concentration of 0.5 mg/mL protein, which is mostly albumin. Pathologically, proteins in the vitreous cavity may originate from regional excretion, filtering from blood vessels, and dispersion from nearby tissues. Protein corona development on the surface of nanoparticles is affected by differences in the structure and amount of proteins in the vitreous. Tavakoli et al. (2021) and Jo et al. (2016) found that liposomes, silica, and polystyrene nanoparticles can maintain their negatively charged surfaces in the vitreous by attracting negatively charged proteins, regardless of the surface charge of the nanoparticles. Jo et al. (2016) discovered that following the development of the protein corona, the diameters of 20 nm and 100 nm gold and silica nanoparticles in the vitreous cavity increased by approximately 5–8 nm. Tavakoli et al. (2021) measured the thickness of the protein corona on liposomes in the vitreous, and the hard corona was 2.2 nm and the soft corona was 2.5–3.1 nm. According to their report, neither changes in size nor surface charge prevented the nanoparticles from passing through the vitreous cavity. Nevertheless, the protein corona may affect how drugs are absorbed and distributed inside target cells, as well as how they are removed from cells (Wang, Chen et al., 2021). This may also slow the rate of intraocular clearance of the drug (Del Amo et al., 2017).

Some polymers have been demonstrated to have the potential to help reduce the formation of a protein corona, such as HA, dextran, poloxamine, and polyethylene glycol (PEG) (Skalickova et al., 2021). In fact, some scientists are developing nanoparticles that can form protein coronas with particular components to perform different biological functions. The protein corona may increase the possibility of nanoparticles interacting with cellular receptors and influencing subsequent biological reactions (Kelly et al., 2015). Yang et al. designed a stable protein corona composed of cyclic RGDyk peptide-modified bovine serum albumin around chitosan-based nanoparticles to reduce serum protein adsorption, which is a novel approach to improve delivery efficiency and therapeutic effectiveness (Yang et al., 2021). In addition to being utilized for therapy, alterations in the protein corona resulting from changes in the biofluid can be utilized as diagnostic tools in a variety of clinical disorders (Vidaurre-Agut et al., 2019).

2.2. Drug delivery to the posterior ocular tissues

The retina is composed of ten layers (Figure 2C), namely, the retinal pigment epithelium (RPE), photoreceptors, outer limiting membrane (OLM), outer nuclear layer, outer plexiform layer, inner nuclear layer, inner plexiform layer, ganglion cell layer, nerve fiber layer, and inner limiting membrane (ILM). The layer of the retina next to the vitreous border is the ILM, which is mainly made of mucopolysaccharides and collagen (Li, Wang et al., 2021). Three-dimensional networks with pore diameters between 10 and 25 nm have been found in rabbit ILM (Nishihara, 1991). Animal experiments show that the size range of biomolecules passing through the ILM is generally 60–90 kDa (Jackson et al., 2003). To our knowledge, no mesh diameter parameters for human ILM have been reported, and the size threshold for nanoparticles passing through ILM has not been properly defined. The morphology of ILM varies with age and localization in the retina (Peynshaert et al., 2019). Peynshaert et al. (2017) investigated bovine ILM and discovered that 40 nm negatively charged polystyrene beads could easily pass through ILM, while particles in the size range of 100 to 200 nm could not. In a subsequent investigation, they used transmission electron microscopy to explore the morphology of human ILM and discovered that the ILM in the posterior pole of the eyeball was the thickest and that ILM thickness gradually increased with age (Peynshaert et al., 2019). Notably, ILM breakage, inflammatory processes, or the stimulation of microglia may allow larger nanoparticles to enter the retina in retinal retinopathy (Bourges et al., 2003; Lou et al., 2021). In addition to the particle size, the surface charge of the nanoparticles is also one of the factors affecting the penetration of ILM. The components of ILM determine its negative charge properties, so it has a blocking effect on positively charged particles (Peynshaert et al., 2018, 2019). As a consequence, for human ILM of various ages and illness stages, the physical and chemical features of nanoparticles capable of overcoming ILM must be clarified further.

The retina’s primary glial cells, known as Müller cells, play an important role in the development of the retina’s normal structure and homeostasis (Yesilyurt et al., 2022). Phagocytosis by Müller cells can transport nanoparticles from the ILM to the outer layer of the retina. Koo et al. found that HA-modified human serum albumin (HSA) heterogeneous nanoparticles (344.8 ± 10.7 nm) could overcome the ILM physical barrier via clathrin-mediated endocytosis and active transport by Müller cells (Koo et al., 2012). The delivery of small molecules in retinal cells is dependent not only on passive diffusion but also on active transport. There are currently few publications on active transport membrane proteins of retina cells that are involved in medication delivery.

2.3. Blood–ocular barrier

When drugs are injected into the vitreous, they move into the ocular tissues and are then removed by the body’s systemic circulation. Vitreous drug concentrations are determined by the intravitreal volume of distribution and clearance (Del Amo et al., 2015). The primary parameters of intravitreal pharmacokinetics are volume of distribution and clearance, which represent drug distribution and elimination, respectively (Del Amo & Urtti, 2015).

There are two clearance routes to eliminate drugs, the anterior route and the posterior route (Del Amo et al., 2017) (Figure 2B). In the anterior clearance route, the drug was eliminated by the aqueous humor flow via trabecular meshwork outflow channels. In the posterior route, drug clearance occurs through blood–ocular barriers composed of epithelial and endothelial cells bound by tight junctions, including the blood–aqueous humor barrier and the blood–retina barrier. The posterior surface of the iris, the iris’s capillary epithelium, and the nonpigmented ciliary forms the blood–aqueous humor barrier in the anterior segment. The blood–retinal barrier, located in the posterior part, is made up of the RPE and the endothelium of retinal capillaries (Del Amo & Urtti, 2015). The posterior clearance route depends on the permeability of the drugs, which is affected by the MW and lipophilicity of the drugs (H. M. Kim et al., 2021). Moreover, the effect of MW on the intravitreal half-life is more substantial than that of lipophilicity. Drugs with lower MWs have better permeability through the RPE, ILM, and blood–retinal barrier (H. M. Kim et al., 2020, 2021); in previous in vitro experiments, RPE was shown to be the main clearance route of low-MW drugs (Ramsay et al., 2019). Approximately 0.02 to 0.36 mL/h of small molecules are cleared by the RPE, whereas it is difficult to measure the clearance rate in retinal capillaries (Mannermaa et al., 2010). Proteins, as macromolecules, are eliminated via the anterior route instead of by posterior elimination (Del Amo et al., 2017; Ramsay et al., 2019). Hydrophilic drugs are thought to be mostly eliminated through the anterior route (aqueous humor) by passive diffusion, while lipophilic drugs are mostly eliminated through the posterior route (retina). Hydrophilic drugs are assumed to have a longer intravitreal half-life than lipophilic drugs with similar MWs because of the structural properties of the retina in the posterior route. Therefore, the combination of a high MW and low lipophilicity might be an effective strategy for extending the half-life of nanomedicines in intravitreal administration.

3. The advantages of HA modification in nanomaterials

Developing an effective, safe, and stable medication delivery system is a complex task. Nanoparticle size, surface charge, targeted ligands, and stability are important factors that affect intraocular distribution and elimination in intravitreal pharmacokinetics, but they are not the only factors (Huang & Chau, 2019b). The inherent features of HA, including adhesion, rheological properties, biocompatibility, degradability, and nonimmunogenicity, are well known (Valachová & Šoltés, 2021). In addition, HA has various features, including a broad range of MWs, a negatively charged surface, receptor binding properties, and chemical linkages that are susceptible to hyaluronidase hydrolysis (Figure 3). There are multiple functional groups in the HA molecule, including carboxyl, hydroxy, and N-acetyl groups, which allow it to be easily conjugated (Cai et al., 2019). HA might be an ideal polymer for modifying nanomaterials for drug delivery.

Figure 3. The advantages of HA modification in nanomaterials.

3.1. Various biological functions

The MW of HA determines the cell’s responsiveness to external signals as well as how it interacts with the cell matrix and membrane receptors on both resident and recruited cells. The MW and percentage of HA should be chosen carefully based on the therapeutic purpose and the type of carrier that will be used.

Both in vitro and in vivo studies have shown that the MW of HA determines its biological activity via its different effects on cellular signaling and function. Monslow et al. (2015) divided HA into high MW (with a MW higher than 1000 kDa), medium MW (250–1000 kDa), low MW (10–250 kDa), and oligo-HA (with a MW lower than 10 kDa) to investigate the biological activity in different tissues. In vivo, high-MW HA plays a positive role, such as inhibiting vascular injury with pathological significance, inhibiting cell migration and tumor growth in cancer, and enhancing immunosuppression in inflammation. Medium- and low-MW HA can act on vascular endothelial cells to inhibit angiogenesis. They can indirectly cause lung epithelial cells, chondrocytes, liver endothelial cells, and vascular smooth muscle cells to produce more inflammatory substances. Oligo-HA can enhance vascular tube formation, endothelial cell proliferation, and cytokine release. Stren and coworkers came to the same conclusion that small HA polymer fragments are angiogenic, inflammatory, and immunostimulatory (Stern et al., 2006). Inflammation is mainly induced by tetrasaccharide HA, whereas disaccharide HA may be able to inhibit it through Toll-like receptor 4 (Han et al., 2022). A wide range of different and often contradictory impacts on a variety of biological activities have been demonstrated in the literature, but this has not been explored in ophthalmology.

3.2. Drug clearance rate regulation

To modulate the drug release phase, high-MW HA with longer chains may create considerable entanglements and prevent water penetration into the mixture (Mayol et al., 2019). Due to the hydrophilicity of HA, water molecules surrounding the nanoparticles can act as an additional energy barrier to protect the polymer from degradation (Lei et al., 2021). A large particle size prevents the drug from being quickly removed from highly permeable blood vessels or the lymphatic system, and the HA coating can significantly increase the particle size and prolong the elimination process (Gan et al., 2013; Y. Hu et al., 2019). In addition, because of the variable MW, HA conjugation can be used to alter molecular size to contribute to the enhanced permeability and retention (EPR) effect in tumors (Ashrafizadeh et al., 2021). The EPR effect of nanoparticles, however, has only been found in the systematic injection of drug delivery for tumor therapy. For intravitreal injection, this phenomenon has not been described in the literature for the treatment of ocular pathologies.

3.3. Surface charge modification

HA is a negatively charged, hydrophilic polymer that attracts counterions and water while maintaining a net negative charge. The surface charge of a nanocarrier may change dramatically after HA modification. After being modified with varied ratios of HA, the surface charge of cationic nanoparticles will gradually change, and the cations will gradually neutralize or even revert to anions (Devoldere et al., 2019; Tan et al., 2021). Simultaneously, the zeta potential will change. The absolute zeta potential value is an important factor in evaluating nanoparticle stability. A higher absolute zeta potential means that the nanoplatform is more stable in suspension. Therefore, the modification of HA can affect the storage stability by affecting the surface charge of nanocarriers.

As mentioned in the Section 2.1, surface charge affects the biodistribution of nanoparticles in vivo by binding with biological tissues via electrostatic interactions or destabilization via dynamic exchange of nanoparticle components. It is known that the vitreous has a negative net charge, and the ILM is composed of negatively charged proteoglycans. Indeed, the membranes of retinal cells also consist of negatively charged phospholipids (Huang & Chau, 2019b). Due to electrostatic interactions, the vitreous humor has little effect on the movement of neutral and negatively charged particles, but it slows the movement of positively charged nanoparticles and polymers (Koo et al., 2012; Martens et al., 2013; Xu et al., 2013; Tavakoli et al., 2021). Koo et al. (2012) observed the distribution of various nanoparticles with similar sizes and different surface properties after intravitreal injection. They showed that highly cationic nanoparticles composed of polyethyleneimine or glycol chitosan aggregated with the vitreous and could not move to penetrate the vitreous, but anionic nanopolymers of the same size could cross the ILM. Xu et al. (2013) found that NH₂-coated cationic nanoparticles were trapped in the vitreous gel and even observed that cationic polyethyleneimine/DNA nanoparticles aggregated in the vitreous after injection. (Huang and Chau (2019a) found a similar result: when lipid nanoparticles with varied zeta potentials (ranging from –30 mV to +50 mV) were injected into the vitreous cavity, only nanoparticles with zeta potentials higher than +35 mV remained in the vitreous cavity for more than 24 h, whereas other particles could not be detected. Therefore, due to electrostatic interactions, the vitreous humor has little effect on the movement of neutral and negatively charged particles, but it slows the movement of positively charged nanoparticles and polymers (Tavakoli et al., 2021). With an intrinsic negative charge, the HA coating can reduce the positive charge of nanoparticles and even reverse the negative charge so that the nanocarriers can penetrate the vitreous successfully to reach the posterior tissue.

3.4. Targeting

Multiple HA receptors have been identified with multiple biological functions through hyaladherins, including cluster-determined 44 (CD44), receptor for hyaluronate-mediated motility, HA receptor for endocytosis, and lymphatic vessel endothelial hyaluronan receptor-1 (X. Zhang et al., 2021). Among them, CD44 has been investigated and applied most thus far. CD44 is a type I membrane glycoprotein with a transmembrane domain, a heavily glycosylated membrane-proximal region, a short cytoplasmic domain, and an N-terminal HA binding domain (Heldin et al., 2020). It can interact with other ligands, such as fibronectin, collagen, and matrix metalloproteinases. CD44 is not only critical for cell migration, inflammatory processes, lymphocyte localization, and cell proliferation but also helps with HA removal and healing after wounding or other types of damage (Garantziotis & Savani, 2019). CD44 can bind and internalize HA via van der Waals forces or H-bonds to inhibit inflammation, neovascularization, and cancer metastasis (Li, Sun et al., 2021). HA is a promising ligand for cancer-targeting nanoparticles due to the high expression of CD44 in cancers (Chen, Li et al., 2021; Gu et al., 2021; Li, Sun et al., 2021; D. Liu et al., 2021; Wang, Zhang et al., 2021).

CD44 receptors can be observed in retinal Müller cells, RPE cells, corneal epithelium, ganglion cells, and endothelial cells in eyes (Guter & Breunig, 2017; L. Chen et al., 2020). When there is inflammation, CD44 expression can increase in RPE cells, which makes them a good target for anti-inflammatory drugs (Crane & Liversidge, 2008; Laradji et al., 2021). Thus, HA’s receptor binding properties can be used to target the retina and to promote cellular uptake and transfection in gene therapy. In addition, the binding of HA with the CD44 receptor could slow clearance and prolong retention. As the MW of HA increased, cellular uptake increased as well until the MW reached 1000 kDa (Samuelsson & Gustafson, 1998; Wolny et al., 2010; Gan et al., 2013). Many previous studies have shown that the affinity of HA binding to CD44 becomes stronger as the MW of HA increases (Wolny et al., 2010; Guter & Breunig, 2017; Jiang et al., 2018; Sakurai & Harashima, 2019). Furthermore, HA with an MW of less than 10 kDa reversibly binds CD44, and the maximum binding affinity occurs at 1000 kDa (Jiang et al., 2018). Qhattal & Liu (2011) considered grafting density to be another essential determinant for CD44 affinity, in addition to the MW of HA. Increasing either HA chain length or grafting density can promote the binding of HA-coated liposomes to the CD44 receptor. The multivalent interaction of HA, on the other hand, is more than the sum of its monovalent units. The amount of bound HA was reduced as the receptor surface density decreased, while the binding stability remained the same. A cooperative phenomenon will occur when several receptor binding sites exist on the long HA chain of repeating disaccharides. Thus, the presence of one ligand changes the affinity for another, and the binding effects are not linear or easy to measure (Stern et al., 2006). Different cell types may have different levels of CD44 receptors, clusters, and turnover rates, which could make them interact with HA ligands in different ways.

3.5. Controlling drug release

Stimuli-responsive nanomedicines have the potential to lower toxicity while also increasing medication concentrations in specific organs or tissues. Based on the stimulus, drug delivery systems can be roughly classified into exogenous and endogenous stimuli-responsive systems (Choi et al., 2019). Exogenous stimuli include changes in light, heat, ultrasound, magnetic fields, and so on, whereas endogenous stimuli mostly include pH, redox compounds, and enzymes in the tumor or pathological environment. HA is frequently employed to build various types of stimulus-response drug delivery systems for the treatment of cancers (C. Chen et al., 2018; Ji et al., 2019; Lu et al., 2020; Zhang, Ren et al., 2020; Ashrafizadeh et al., 2021).

HA can be broken down by hyaluronidase, which substantially accumulates in some tumor tissues and organs (Y. Hu et al., 2019; Jung, 2020). When hyaluronidase is expressed specifically or significantly increased, as in cancers with high metabolic levels, HA can serve as a switch to control drug release (C. Hu et al., 2018; Lee et al., 2021). Once the HA-coated nanocarriers reach the targeted region, hyaluronidase in endosomes can breakdown the high-MW HA coating on the nanocarriers, allowing the smaller particles to penetrate deeper into the tissues (X. Yan et al., 2019). R. Liu et al. (2019) designed hyaluronidase-triggered nanoplatforms enveloping indocyanine green (ICG), bovine serum albumin, and gold nanoparticles in the HA shell, which can reduce the nonspecific distribution and drug release of nanoparticles in breast cancer lesions.

Although hyaluronidase activity was discovered in human vitreous (Schwartz et al., 1996), the impact on vitreous HA and exogenous HA polymer in vivo has not been deeply investigated. Xie et al. (2021) found that when the concentration of hyaluronidase was 16.7 U/mL, the hydrogel system composed of collagen II and HA had no degradation within 7 days in vitro. Dromel et al. (2021) discovered that when interpenetrating polymer network hydrogels with varied levels of HA were compared, gels with higher HA content degraded faster in the vitreous cavity. In a later study, 0.3 U/mL hyaluronidase was applied. The effects of hyaluronidase in the vitreous environment on the pharmaceutics of HA-based nanoparticles are still unknown. Considering the existence of hyaluronidase, it is necessary to investigate whether the percentage of HA in the nanocarriers will affect drug release and the degradation of the nanoplatform.

4. Delivery of HA-based nanoparticles to the ocular posterior segment

Various nanoplatforms (Figure 4) have been widely explored to deliver peptides, proteins, nucleic acids, and small molecules to target tissues via intravitreal injection. Table 1 summarizes the benefits and drawbacks of various nanoparticles investigated in experimental studies for vitreous injection. Investigations of HA as a nanomodifier to deliver drugs to the ocular posterior segment via intravitreal injection have yielded promising results. Table 2 lists the different nanocarriers, sizes of HA-coated complexes, MWs, and HA coating strategies that have been used to treat ocular posterior diseases in advanced research.

Figure 4. The schematic of nanoparticles.

Table 1. Advantages and drawbacks of nanoparticles for intravitreal administration.

Nanoparticles	Advantages	Drawbacks	Ref
Lipid nanoparticles	Low or absence of in-vivo toxicity	Drug expulsion during storage (for SLN)	(Khosa et al., 2018; Costa et al., 2021; Navarro-Partida et al., 2021)
	Economic and solvent-free production techniques	Restricted loading capacity of hydrophilic drug (for SLN and NLC)
	Accessibility
	Possibility to be autoclaved or sterilized
	Great kinetic stability compared to liposomes and niosomes
	Condensing and protecting the genetic material
Liposomes	Biocompatibility	Visual blurring	(Bochot & Fattal, 2012; Akbarzadeh et al., 2013; Chaurasiya et al., 2022; Taléns-Visconti et al., 2022)
	Enhancing the stability of entrapped drugs	Inflammation caused by cationic surface charge
	Increasing the drug half-life in the vitreous	Possibility of aggregation during storage or in vivo when colloidal stability is poor
	Reducing the drug toxicity	High-cost of production
	Possibility of ligands attachment
Micelles	Enhancing the solubility and therapeutic efficacy of highly hydrophobic drugs	Ocular irritation	(Wu et al., 2016; Gote et al., 2020; Rodríguez-Acosta et al., 2021)
	Enhancing the stability of entrapped drugs	Limitation to large-scale production
	Accessibility and low-cost of production
	Enabling controlled drug release kinetics
	Ease of surface modification
Dendrimers	Intrinsic ability of the PAMAM dendrimers to target neuroinflammation-related cells	Cytotoxicity caused by cationic surface charge	(Khosa et al., 2018; Sathe and Bharatam, 2022)
	A well-defined shape, size, and narrow polydispersity
Human serum albumin	High drug loading capacity	Unstable	(Koo et al., 2012; Piazzini et al., 2019; Gao et al., 2020)
	Nonimmunogenicity
	Penetrating ILM by interaction with Müller cells
Gold nanoparticles	Biocompatibility	Need ligands to stabilize	(Kim et al., 2011; Masse et al., 2019; Bharadwaj et al., 2021)
	Accessibility
	Endowing precisely control over size and shape
	Antiangiogenic and anti-inflammatory
Niosomes	Biodegradability	Lower drug loading capacity compared to liposomes	(Hong et al., 2009; Gan et al., 2013; Yasamineh et al., 2022)
	Biocompatibility
	Nonimmunogenicity
	Chemically stable
	Encapsulation of both hydrophilic and hydrophobic drug molecule

Note: SLN, solid lipid nanoparticles; NLC, nanostructured lipid carriers; PAMAM, polyamidoamine; ILM, inner limited membrane.

Table 2. HA-based nanoparticles.

Nanocarriers	Drug	MW of HA	Size	Zeta potential	Type of HA coating	Disorder	Benefits of HA modification	Ref
Lipid nanoparticles	NA	10–100 kDa; 200–400 kDa	320 nm	‒25 mV	Electrostatic adsorption	Uveitis	Slowing down elimination in the retina and RPE/choroid, targeting to CD44-positive RPE cells	(Gan et al., 2013)
Solid lipid nanoparticles/protamine	Plasmid (pCMS–EGFP)	150 kDa; 500 kDa; 1,630 kDa	240–340 nm	+30 mV–+40 mV	Electrostatic adsorption	X-linked juvenile retinoschisis	Improving the cellular uptake and transfection, facilitating the release of DNA.	(Apaolaza et al., 2014)
Solid lipid nanoparticles/protamine	Plasmid (pCAG-GFP_CMV_RS1)	150 kDa	235 ± 28 nm	+31.4 ± 0.8 mV	Electrostatic adsorption	X-linked juvenile retinoschisis	Doing favor on the loosening of the plasmid in cytoplasmic, accelerating the diffusion process in the vitreous	(Apaolaza et al., 2015)
N,N′-cystaminebisacrylamide-4-aminobutanol (p(CBA-ABOL))/pDNA complexes	Plasmid (pGL4.13 and gwiz-GFP)	22 kDa;137 kDa; 2,700 kDa	647 ± 19 nm; 343 ± 4 nm; NA	‒22 ± 1 mV; −29.9 ± 0.5 mV; –37 ± 2 mV	Electrostatic adsorption	Retina dystrophies with a genetic cause	Stabilizing the nanoparticles with negative charges, improving the mobility in vitreous, increasing the cellular uptake via CD44	(Martens et al., 2015)
Solid lipid nanoparticles/protamine	Plasmid (pCMV-GFP_mOPS-RS1)	150 kDa	223 ± 20 nm	+31.5 ± 1.1 mV	Electrostatic adsorption	X-linked juvenile retinoschisis	Favor of the mobility of nano polymers through the vitreous and retina, facilitating the cellular uptake via CD44, improving the transfection via plasmid loosening	(Apaolaza et al., 2016)
Human serum albumin	Connexin43 mimetic peptide	120 kDa	252.70 ± 7.29 nm	‒43.97 ± 0.38 mV	Electrostatic adsorption	Retinal ischemia and inflammatory disorders	Enhancing the cellular uptake and retinal penetration via CD44, stabilizing the nanoparticles and alleviating the cytotoxicity, presenting more sustained and constant drug release	(Huang et al., 2017, 2018)
PolysiRNA polyplex	PolysiRNA targeting mouse VEGF-A	5 kDa	260.7 ± 43.27 nm	‒4.98 ± 0.47 mV	Electrostatic adsorption	Age-related macular degeneration	Overcoming the retinal barriers and cellular uptake by RPE cells via CD44 receptors	(Lee et al., 2016)
DOTAP/DOPE liposomes	Plasmid (pGL4.13 and gwiz-GFP)	1,600 kDa	159 ± 1 nm; 204 ± 4 nm	‒36 ± 1; −20 ± 4 mV	Electrostatic adsorption;Covalent adsorption	NA	Improving the mobility in vitreous, covalent adsorption of HA significantly promoting the transgene expression than by electrostatic adsorption	(Martens et al., 2017)
DOTAP/DOPE niosomes	Plasmid (pEGFP-C1)	1,900–3,900 kDa	286.20 ± 2.19 nm	‒41.1 ± 1.16 mV	Covalent adsorption	Genetic or acquired retina diseases	Targeting retina cells through interaction between HA and CD44, achieving remarkable transfection efficiency	(Qin et al., 2018)
TransIT-mRNA comlexes	mRNA (pEGFFP)	20 kDa; 200 kDa;>1,800 kDa	125–1,000 nm	‒40mV to +10 mV	Electrostatic adsorption	NA	Increasing the mobility in vitreous	(Devoldere et al., 2019)
Gold nanoparticles	NA	5 kDa	53.22 ± 3.36 nm; 90.01 ± 2.58 nm	‒24.73 ± 1.47 mV; ‒22.23 ± 1.77 mV	Covalent adsorption	Intraocular vascular disorders	Stabilizing the particles, enhancing the diffusion ability in vitreous, targeting to the deep retina tissues	(Apaolaza et al., 2020)
Light-activated liposomes	Indocyanine green	8–15 kDa	104 ± 25 nm; 68 ± 16nm	‒11.27 ± 0.21 mV; NA	Covalent adsorption	NA	Improving the stability and mobility in vitreous and plasma	(Kari et al., 2020)
Dendrimer PRHF/NLS/DNA complex; liposome	Plasmid (Flt23K)	10 kDa	200–650 nm	‒20 mV to +15 mV	Covalent adsorption	Age-related macular degeneration	Leading to higher cellular uptake due to the binding to CD44, improving the stability and mobility in vitreous, enhancing the penetration in retina tissues	(Tan et al., 2021)

Note: MW, molecular weight; HA, hyaluronic acid or hyaluronan; CD44, cluster-determined 44; RPE, retina pigment epithelium; GFP, green fluorescent protein; VEGF, vascular endothelial growth factor; DOTAP, 1,2-dioleoyl-3-trimethylammonium-propane; DOPE, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine; mRNA, messenger RNA; NLS, nuclear localization signal; NA, not accessible.

4.1. Lipid nanoparticles

Lipid nanoparticles are colloidal carriers composed of biocompatible lipid molecules, including solid lipid nanoparticles (SLNs), nanostructured lipid carriers, and lipid–drug conjugate nanoparticles. The structure of lipid nanoparticles contains a lipid core and is stabilized by a surfactant shell (Figure 4). In comparison to other nanocarriers, lipid nanoparticle production methods can usually effectively prevent the presence of organic solvent residues and enable large-scale synthesis (González-Fernández et al., 2021). SLNs are aqueous colloidal carriers composed of a solid lipid core and liquid lipids and exhibit high surface area, low toxicity, and manufacturing stability (Wang, Gao et al., 2021). The drug delivery capacity of SLNs is limited by their crystalline properties. Drugs will be expelled if the crystalline lattice forms. Lipid nanoparticles made of a mixture of solid and liquid lipids are called nanostructured lipid carriers. Their loading capacity is larger and their half-life is longer than those of SLNs, making them more useful nanoparticles. As the second generation of lipid nanoparticles, they have improved drug loading capacity and a prolonged half-life (Khosa et al., 2018). In addition to proteins and nucleic acids, SLNs and nanostructured lipid carriers can encapsulate both hydrophilic and lipophilic drugs, although the hydrophilic drug loading capacity is restricted by the lipidic cores. Lipid–drug conjugate nanoparticles have been developed to overcome this limitation.

Lipophilic and hydrophilic medicines, genetic drugs, peptides, and vaccines can all be efficiently delivered via SLNs. SLNs are attractive nonviral vectors for gene therapy because they can effectively carry genetic material into the cytoplasm and nucleus (Suñé-Pou et al., 2019). Furthermore, compared to conventional viral delivery methods, they have a number of advantages for gene therapy, such as safety, low cost, high repeatability, and no limit on the size of the genetic material that can be used. Apaolaza et al. (2014) designed a drug delivery system made of SLNs, protamine, and HA to successfully deliver plasmid DNA for the gene therapy of X-linked juvenile retinoschisis. SLNs with HA modification may be more stable in biofluids than uncoated SLNs (S.-E. Lee et al., 2019). Dextran, similar to HA, is a negatively charged biocompatible polysaccharide with the capacity to enhance cellular uptake via the clathrin-mediated endocytosis pathway (Apaolaza et al., 2015). When Apaolaza et al. (2016) compared dextran- and HA-modified SLNs in vitro, they discovered that HA modification resulted in a sevenfold increase in transfection efficiency. Positively charged protamine and SLN, as well as negatively charged HA and DNA, bind to form complexes with a net positive charge due to electrostatic interactions. With the interaction of HA and CD44 or other specific receptors, caveolae/lipid raft-mediated endocytosis rather than clathrin-mediated endocytosis has been observed for HA-coated vectors. The HA molecule can smoothly release plasmid DNA in the absence of lysosomal enzymes when applied to transfect ARPE 19 cells and HEK-293 cell lines.

4.2. Liposomes

Liposomes are spherical vesicles with an aqueous core encapsulated in lipid bilayers (Taléns-Visconti et al., 2022) (Figure 4). Compared to other colloidal carriers, liposomes present several advantages in controlling drug release, biocompatibility, and low toxicity to decrease the number of intravitreal injections. Liposomes were found to protect labile molecules in the vitreous, such as peptides and oligonucleotides (Lajavardi et al., 2007). Liposome distribution is affected by the experimental animal species, disease model, and vitreous liquefaction (Bochot & Fattal, 2012). According to earlier studies, liposomes were observed to be removed from the vitreous by the anterior pathway and lymphatics (Camelo et al., 2007, 2008). However, liposomes’ lack of electrostatic or steric stabilization in the vitreous might cause some blurring effects and affect vision slightly until resorption. Despite this, liposomes as nanocarriers in ocular treatment are still worth considering. Both the liposome layer and the HA molecule can play protective roles to prolong the residence period and endow the nanoparticles with a negative charge for distribution in the vitreous. In addition, the high water-binding capacity of HA could enhance liposome stability (Kari et al., 2020).

Gan et al. (2013) designed and fabricated HA-modified liposome nanoparticles (HA-LCS-NPs) for posterior segment disease therapy. The lipid bilayer-coated nanoparticles shielded the positive charge of the core–shell and could move freely in the vitreous, while uncoated nanoparticles were blocked by the ILM. With HA modification, the HA-LCS-NPs possessed a lower zeta potential of approximately –25 mV, and the interaction of HA and CD44 receptors endowed them with RPE-targeting ability. Both endogenous HA and CD44 are upregulated in the experimental autoimmune uveitis eye. Endogenic HA can be found in the ILM, ganglion cell layer, inner plexiform layer, etc. (Gan et al., 2013). The findings revealed that the upregulated endogenic HA had no effect on the targeting capabilities of the modified nanoparticles (Cho et al., 2012). In addition, Tan et al. (2021) developed a core–shell liposome nanoplatform as an efficient carrier for targeted delivery of genes to the retina. In these liposomes, the outer shell layer targets retinal cells, and the core layer is made of the dendrimer PRHF/NLS (nuclear localization signal)/DNA complex. The loaded DNA may penetrate the retina, be collected by cells, bypass endosomes, and then be delivered into the nucleus through the nuclear localization signal. In the cytoplasm, HA-DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) and liposomes are critical in assisting endosomal escape.

Kari and coworkers evaluated the stability, loading capacity, and protein corona of liposomes coated with PEG and HA. The two coated nanoparticles had similar indocyanine green loading capacities and were both stronger than uncoated liposomes (Kari et al., 2020). Furthermore, they showed excellent stability in plasma and vitreous, whereas the uncoated liposomes were excluded for suboptimal stability. Following exposure to porcine vitreous, hydrophilic and negatively charged protein coronas formed on the liposomes coated with HA and PEG. Proteins that interact with collagen fibers in the vitreous channel bound to HA-coated liposomes more frequently. The HA coating bound more proteins that interact with collagen, causing the nanoparticles to move slightly slower than PEG-coated nanoparticles. However, the HA- and PEG-coated liposomes both diffused much faster than the untreated version.

Martens et al. (2017) synthesized electrostatic and covalent HA-coated liposomes to transfer plasmid DNA and assessed their efficacy in vitro in terms of vitreous mobility and transfection capability of retinal cells. Anionic and monodisperse HA-liposomes were acquired by using these two methods, enhancing their intravitreal mobility in the vitreous. On the other hand, covalent HA liposomes were more stable and provided an eightfold increase in transgenic expression compared to electrostatic coatings for gene therapy via intravitreal injection. This suggests that different coating strategies may lead to differences in HA presentation and hyaladherin avidity, bringing about a different affinity of CD44 receptors. The biological roles of various HA conjugations should be investigated further.

4.3. Niosomes

Niosomes are amphiphilic vesicles with a nonionic surfactant and a bilayer structure similar to those of liposomes (Rezaie Amale et al., 2021) (Figure 4). The noisome formulation is determined by the synthesis process, the nonionic surfactant used, and the fabrication parameters (Verma et al., 2021; Aparajay & Dev, 2022). Compared with liposomes, niosomes are low-cost, less toxic, and more stable and accessible (Hong et al., 2009).

Niosomes coated with HA can be used to deliver drugs through the vitreous cavity, deliver genetic drugs, and successfully target RPE cells to treat retinal diseases. Y. Qin et al. (2018) fabricated cationic niosomes composed of Tween 80/squalene/1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) via ethanol injection, encapsulating the pEGFP-C1 plasmid. The ligand–receptor binding of HA and CD44 and the interaction between cation niosomes and the negatively charged cell membrane enhanced the RPE cellular uptake process. In addition, the cationic niosomes achieved lysosome escape and successfully released gene drugs via electrostatic adsorption with the negatively charged lysosomal membrane.

4.4. Micelles

Micelles are spherical vesicles that consist of amphiphilic molecules or block copolymers and form a hydrophobic core and a hydrophilic shell in aqueous fluid at certain concentrations (Ghosh & Biswas, 2021) (Figure 4). PEG–phosphatidylethanolamine, PEG–polypropylene oxide–PEG triblock copolymer, PEG–amino acids, and PEG–carbonates are amphiphilic polymers commonly used to synthesize micelles. The lipophilic core can consist of polymers such as polylactide, polycaprolactone, polylactide-co-glycolide, polyesters, poly(amino acid)s or lipids, while the outside hydrophilic layer can be composed of polyethylene glycol, polyoxazolines, chitosan, dextran, and HA (Ghosh & Biswas, 2021). The lipophilic core allows excellent encapsulation of hydrophobic drugs due to hydrophobic interactions, while the hydrophilic corona endows them with water solubility. Because drugs can bind to HA via disulfide bonds, HA is frequently used as a reduction-sensitive biopolymer in micellar studies. Due to their splendid stability and wide drug loading range, micelles modified with HA have been widely investigated for tumor therapy (W. Li et al., 2020; Zhang, Ren et al., 2020; Yan, Feng et al., 2022). Micelles can penetrate the cornea and pass through the sclera when applied by the topical route (Zhao et al., 2022). With excellent mucosal adhesion that can prolong the retention duration before the cornea, HA-modified micelles are usually applied in eye drop research (Beack et al., 2015). The Food and Drug Administration approved the first micelle-based cyclosporine eye drops for dry eye treatment in 2018.

The intravitreal injection of micelle-based drug delivery systems has been investigated in previous studies (Junnuthula et al., 2021; Li, Qian et al., 2021), but the synthesis of HA-modified micelles has seldom been published. Micelles and polymersomes formed by polyethylene glycol, polycaprolactone, and trimethylene carbonate were injected into the rabbit vitreous and exhibited a long retention time and effective drug delivery to the posterior segment (Junnuthula et al., 2021). Research by Sadeghi et al. reached the same conclusion (Sadeghi et al., 2022).

4.5. Dendrimers

Dendrimers, with their nanosize (2–100 nm), large multivalent surface area, well-defined polymeric architecture, and quantized building components, are among the most sophisticated nanodrug delivery tools (Alshaikh et al., 2022) (Figure 4). Dendrimer-based drug delivery has been investigated for direct application to the cornea as solutions or gels, intravitreal administration, or subconjunctival injections.

Neuroinflammation is a critical factor in the pathophysiology of several neurodegenerative illnesses and can provide a new target for treatment. The dendritic molecules formed by polyamidoamine (PAMAM) have their own unique ability to target cells associated with neuroinflammation (Dai et al., 2010). They can selectively localize in activated microglia in retinal neuroinflammatory diseases and provide sustained drug release (Iezzi et al., 2012; Kambhampati et al., 2015). However, the cationic charge of PAMAM dendrimers causes toxicity when they interact with negatively charged cell membranes, resulting in membrane disruption (Zhang, Pan et al., 2020). In CD44-overexpressing cancer cells, HA coating can shield the cationic charges of dendrimers and provide CD44 targeting ability (Zhang, Pan et al., 2020). In addition, the toxicity of PAMAM depends on concentration, generation, and time (Qin et al., 2020). In a previous study, generation-4 hydroxyl PAMAM dendrimers showed noncytotoxicity and nonimmunogenicity, as well as high levels of cellular uptake (Perumal et al., 2008). For PAMAM dendrimers beyond generation 5.0, the safety condensation limit was approximately 50 mg/mL, while it may be 75 mg/mL below generation 4.0 (Qin et al., 2020). Because of their ability to target neuroinflammation, PAMAM dendrimers will be promising nanocarriers to cure retinal inflammatory-related disease.

4.6. Human serum albumin

Human serum albumin (HAS) is a negatively charged, multifunctional, nonglycosylated extracellular plasma protein with an MW of 67 kDa and a length of 585 amino acids. Nonimmunogenic HSA, with its high drug delivery capacity, is an ideal platform for the delivery of a variety of agents, including peptide, genetic, and other chemical agents (Piazzini et al., 2019; Gao et al., 2020).

HSA nanoparticles have a superior ability to penetrate the ILM through interaction with Müller cells and were found to be well tolerated in the eye (H. Kim et al., 2009; Koo et al., 2012). According to Huang et al., after intravitreal injection, HA-coated HSA nanoparticles diffused freely in the vitreous body and selectively targeted the retina, showing prolonged retention and a sustained therapeutic effect (Huang et al., 2017, 2018). The HA-coated HSA nanoparticles had a particle size of 252.70 ± 7.29 nm, which is much larger than the pore sizes of the ILM and OLM (Huang et al., 2017). The endocytosis and exocytosis of Müller cells may be another penetration route to help nanoparticles pass through the ILM and OLM. Both uncoated and HA-coated HSA nanoparticles can penetrate the ILM and reach the ganglion cell layer, but only the HSA nanoparticles with HA coating can be delivered to the photoreceptor layer and the RPE via HA–CD44 interactions. The HA coating can slow the release rate of drugs and increase the residence time in the retina.

4.7. Gold nanoparticles

Gold nanoparticles (GNPs), different from gold particles, consist of a gold core and a surface coating and have a size range from 1 nm to 120 nm (Figure 4). GNPs are biocompatible, easy to make, and can be made in any shape or size (Masse et al., 2019). Different sizes and shapes endow GNPs with different chemical and physical properties, especially GNPs with a size range of 1–10 nm (Bharadwaj et al., 2021). GNPs have been widely investigated in cancer diagnosis and treatment, as well as in ophthalmic imaging techniques and therapies (Chen, Si et al., 2021; Hou et al., 2022).

During ocular administration, GNPs can be used as carriers for a variety of therapeutic drugs and can be delivered across the vitreous and ILM in the eye with surface modification. Previous studies suggested that GNPs could act as potent inhibitors of pathological angiogenesis (J.H. Kim et al., 2011; Roh et al., 2016; Darweesh et al., 2019; Singh et al., 2020). Roh et al. applied GNPs to inhibit laser-induced choroidal neovascularization via intravitreal injection and demonstrated that GNPs can inhibit the VEGF-induced activation of focal adhesion kinase (FAK), protein kinase B (Akt), and extracellular signal-regulated kinase (ERK)1/2 signaling in human umbilical vein endothelial cells (Roh et al., 2016). This conclusion is consistent with that of Kim et al. (H.M. Kim et al., 2011). They found that GNPs play a similar role in retinal neovascularization via the suppression of VEGFR-2 activation. Furthermore, GNPs can prevent RPE cell migration driven by VEGF or interleukin-1 (Karthikeyan et al., 2010).

In addition, the topical application of GNPs plays an anti-inflammatory role in endotoxin-induced uveitis (Pereira et al., 2012). Thus, with intrinsic antiangiogenic and anti-inflammatory therapeutic properties (Ba Fakih et al., 2020; Di Bella et al., 2021; McCarrick et al., 2021; Kosuge et al., 2022), GNPs have the potential to be widely used in the treatment of various neovascular-related retinopathies caused by VEGF and other inflammation-related ocular diseases. In addition to targeting, modification with HA can stabilize GNPs and maintain their penetration capacity in cells and tissues (Li, Ren et al., 2021). Apaolaza and colleagues modified GNPs with HA (HA-GNPs) to deliver drugs via an intravitreal route. GNPs were blocked by ILM in the vitreous, whereas HA-GNPs were successfully delivered to the photoreceptor layer of the retina (Apaolaza et al., 2020). HA modification endows GNPs with a negative charge and prevents protein corona formation (Liang et al., 2022).

4.8. Gene nanomedicine

Retinal gene therapy is a promising method since the eye is easily accessible and protected from systemic circulation, which decreases the risk of unwanted side effects. Gene therapy involves the cellular uptake of new genes to restore or increase gene expression for the purpose of treating dystrophic diseases with a genetic cause. The loading vectors can be divided into viral vectors (e.g. lentivirus, adenovirus, adeno-associated viruses) and nonviral vectors (Santana-Armas & Tros de Ilarduya, 2021). Among nonviral vectors, nanoplatforms are promising carriers to deliver gene medicines because of their nontoxicity and availability (Patil et al., 2019; Yan, Lin et al., 2022) (Figure 4).

Small interfering RNA (siRNA) therapeutic drugs downregulate certain disease-related mRNAs, providing a novel therapeutic modality to treat ocular posterior segment diseases. Lee et al. designed a poly-siRNA polyplex by siRNA polymerization (poly-siRNA) followed by coating with branched polyethylenimine (bPEI) and HA to suppress choroidal neovascularization formation via intravitreal administration (Lee et al., 2016). In this study, anti-vascular endothelial growth factor (anti-VEGF) monosiRNAs were polymerized into polysiRNAs by disulfide-bonded bPEI complexes. When the positively charged bPEI and negatively charged poly-siRNA were combined at a ratio of 1:1, a balance was achieved that enabled coagulation into nanosized particles and the subsequent release of poly-siRNA into the cytoplasm. When naked genetic medicines enter the vitreous cavity, they are usually broken down and removed quickly (Zhang, Rombouts et al., 2020). Zhang et al. found that the half-lives of naked mRNA and plasmid DNA in biological fluids such as human serum, human ascites, and bovine vitreous are approximately 1–2 min and 1–4 h, respectively (Zhang, Rombouts et al. 2020). In addition, compared to the half-life of naked genetic material in biological fluids, genetic material encapsulated with lipid nanoparticles had a considerably longer half-life. The anionic poly-siRNA polyplex can be well preserved against decomposition in the vitreous, penetrate the ILM by Müller cell-mediated transcytosis, and enter RPE cells via the CD44 receptor-mediated endocytosis pathway. More importantly, siRNAs of twenty-one nucleotides or longer were discovered to reduce choroidal neovascularization via Toll-like receptor 3 (TLR3) (Kleinman et al., 2008). The bPEI-HA double coating effectively inhibited the interaction between TLR3 and the encapsulated siRNA; thus, the anticipated anti-VEGF siRNA mechanism played a main role in therapy. Indeed, the developed polyplex coated with bPEI and HA can serve as an effective carrier to deliver poly-siRNA to the retina.

Martens et al. (2015) made an HA polymer coating on positively charged plasmid gene medicines, which consisted of plasmid DNA and N,N′-cystaminebisacrylamide-4-aminobutanol, to provide a hydrophilic and negatively charged surface. HA-modified polyplexes with different MWs (22 kDa, 137 kDa, and 2700 kDa) were constructed, and the resulting nanoparticles had a negative surface charge and normal drug loading ability. The mobility of the polyplexes with 2700 kDa HA did not increase, most likely due to their large size and aggregation. Both 22 kDa and 137 kDa HA-coated polymers maintained vitreous mobility and were able to induce transfection. In particular, ARPE-19 cells were more efficiently transfected with 22 kDa HA-coated polymers. As HA is an established ligand for the CD44 receptor, the researchers compared the affinity for hyaladherins with free HA. The HA polymer with a lower MW did not have a strong binding ability to hyaladherins, so free HA did not affect the modified nanoparticles. Preincubation of the CD44 receptor with free HA substantially inhibited the adsorption of 137 kDa HA-coated particles. In addition to siRNA and plasmid DNA, the encapsulation of mRNA by the Trans IT complex followed by coating with HA was investigated (Devoldere et al., 2019). The HA coating had no effect on the cellular uptake and transfection efficiency of the target gene. Moreover, the mRNA complex mobility in the vitreous was effectively increased after HA coating.

5. Conclusion

The nanodrug delivery system has great potential for controlled drug release, prolonged retention time and reduced administration frequency. Intravitreal injection is a vital treatment for posterior ocular diseases. We summarize the special pharmacokinetic characteristics of vitreous injection based on the anatomical structure of eyes. Previous studies have shown that the pharmacokinetics of nanoparticles after injection are related to their physicochemical properties and biological properties, such as their nanostructure, size, surface charge, hydrophilicity, and ligand receptor binding. Various nanoparticles have been successfully investigated in experimental studies for vitreous injection, with advantages and drawbacks (Table 1). In addition, some nanoparticles have shown intrinsic therapeutic effects or retinal cell targeting properties, which can be prioritized, such as PAMAM dendrimers and GNPs.

HA is biocompatible, biodegradable, nontoxic, and nonimmunogenic and has the ability to bind to receptors. Because of its outstanding biological properties, HA and its derivatives are widely utilized in the biological and pharmaceutical fields. The extensive application of HA in nanodrug delivery platforms for intravitreal injection is highlighted here. The following HA functions and features may play a critical role in the platform: (1) a wide range of molecular weights; (2) modulating drug release and clearance rate; (3) altering the surface charge of nanoparticles; (4) targeting capabilities due to ligand–receptor interactions; and (5) being catalyzed to degrade by hyaluronidases. The effect of the interaction between HA and the enzyme on drug release is poorly known and warrants further investigation. HA-based nanodrug delivery systems, which can carry a variety of medications, have shown promising results in animal studies but require more research and development before they can be used in humans.

Ethical approval statement

NA.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

This work is supported by the Chengdu Science and Technology Program (2021-YF09-00024-SN), 135 project for disciplines of excellence-Clinical Research Incubation Project, West China Hospital, Sichuan University (No. 2021HXFH026).