Physicochemical Properties Affecting Retinal Drug/Coumarin-6 Delivery from Nanocarrier Systems via Eyedrop Administration

Abstract

Purpose.: To elucidate the effect of physicochemical properties of nanocarrier systems on drug delivery efficiency to the retina by eyedrop administration in mice, rabbits, and monkeys.

Methods.: Submicron-sized liposomes (ssLips) of different particle size, cholesterol content, surface charge, and multilamellar vesicles (MLV) were prepared by the hydration method. Fluorescence probe (coumarin-6)–incorporated liposomes, lipid emulsions, and FITC-labeled polystyrene particles were used to investigate their intraocular behavior after eyedrop administration, using epifluorescence microscopy in mice, rabbits, and monkeys.

Results.: Delivery efficiency of fluorescent probes to the mouse retina from dropped liposomes was extensively improved by reducing their particle size (<600 nm) and cholesterol content, whereas negligible improvement was observed in the case of MLV. Furthermore, FITC-labeled polystyrene particles and coumarin-6–incorporated lipid emulsions showed an insufficient effect on retinal delivery in mice even if their size was controlled at 110 nm. The highest accumulation of the fluorescent probe in the retina was observed around 30 minutes with any type of ssLip used, followed by the prompt disappearance of their fluorescence within 120 minutes in mice. Changes in the fluorescence intensity of coumarin-6 in rabbits and monkeys were observed in a manner similar to that described in mice. Retinal flat-mount images suggest that coumarin-6 incorporated in ssLip diffused from the iris and ciliary body side to the optic disc side in the retina after eyedrop administration.

Conclusions.: The delivery efficiency of coumarin-6 to the retina was altered depending on particle size, constituents, and rigidity. ssLips with appropriate features would be promising drug carriers for retinal delivery through eyedrops.

One of the greatest challenges in ocular drug delivery is to discover effective ways to administer drugs to the posterior segment of the eye to treat diseases of the retina, choroids, and vitreous body. Posterior segment diseases, including macular degeneration and diabetic retinopathy, are leading causes of vision impairment and blindness. Intravitreal injection is commonly used to deliver drugs directly to treat these diseases. ^1–3 Although direct intravitreal injection offers high concentrations of drugs in the retina, the short half-life of this method necessitates frequent administration, ⁴ which is accompanied by the risk of vitreous hemorrhage, retinal detachment, and endophthalmitis. ^5,6 Moreover, patients may not comply with such regimens. Systemic administration is one possible way to obviate these risks. However, large amounts of drug should be administered to maintain therapeutic concentrations because the tight junctions of the blood-retinal barrier restrict the entry of systemically administered drugs into the retina. ⁷ This leads to unexpected side effects because of the distribution and accumulation of drugs in all tissues of the body. ⁸ Thus, there is a pressing need for noninvasive and harmless delivery systems targeting the posterior segment of the eye. 

Topical administration using eyedrops is an alternative way to minimize side effects. Eyedrops are easy to use and do not interfere with vision. Corneal and conjunctival epithelia, along with tear film, serve as biological barriers to protect the eye from potentially harmful substances and drugs. As a consequence, conventional eyedrop formulations usually cannot effectively overcome these barriers. 

The use of colloidal drug delivery systems, such as liposomes, ^9,10 niosomes, ¹¹ nanoparticles, ¹² and nanoemulsions, ¹³ is considered a strategy to enhance the ocular bioavailability of eyedrop-administered drugs. In particular, liposomes can come into intimate contact with the ocular surfaces, thus working as barriers, and they can be used to protect drug molecules from metabolic enzymes at the tear/corneal epithelium interface. ¹⁴  

In a previous study, we reported for the first time the potential of submicron-sized liposomes (ssLips) as a novel system for delivering ocular drugs to the posterior segment of the eye, including the retina. ¹⁵ Briefly, fluorescence emission of coumarin-6 formulated into ssLips was obvious in the retinas of mice after eyedrop administration of ssLip. The rigidity of liposomal particles is an important factor in considering drug delivery efficiency. Several factors would affect the properties in colloidal drug delivery systems: particle rigidity, surface characteristics, constituent elements of particles, structural differences, and particle sizes. Crucial findings may be obtained by investigating the relationship among these factors and drug delivery efficiency to the retina. 

In this study, we tried to elucidate the most important factors of colloidal particles affecting delivery efficiency to the retina after eyedrop administration. Liposomes with different sizes and surface charges, lipid emulsions, and polystyrene particles were used as candidates for drug carriers. The feasibility of nanocarrier systems to retinal drug delivery was investigated using either rabbits or monkeys. 

Materials and Methods

Materials

L-α-Distearoyl phosphatidylcholine (DSPC) and egg phosphatidylcholine (EPC) were purchased from Nippon Oil and Fats Co., Ltd. (Tokyo, Japan). Dicetyl phosphate (DCP) and cholesterol were obtained from Sigma Chemical Co. (St. Louis, MO). Stearyl amine (SA) was purchased from Tokyo Kasei Kogyo Co., Ltd. (Tokyo, Japan). Coumarin-6 as a lipid marker was purchased from MP Biomedicals LLC (Illkirch, France). FITC-labeled polystyrene particle (Micromer-GreenF) was purchased from Corefront Co., Ltd. (Tokyo, Japan). Caprylate and caprate triglyceride (Triester F-810) were purchased from Nikko Chemicals Co., Ltd. (Tokyo, Japan). 4-(2-Hydroxyethyl)-1-piperazine-ethanesulfonic acid (HEPES) was purchased from Nakalai Tesque (Kyoto, Japan). Hanks balanced salt solution (HBSS) was purchased from Gibco BRL (Grand Island, NY). All other chemicals were commercial products of reagent grade. 

Preparation of Multilamellar Vesicles and Submicron-Sized Liposomes

Multilamellar vesicles (MLV), which were composed of DSPC, DCP, or SA, cholesterol, and coumarin-6, were prepared by the hydration method. The lipid composition of prepared liposomes is shown in Table 1. Values in parentheses represent the molar ratio in each composition. These materials were dissolved in a small amount of chloroform in a round-bottom flask and were dried in a rotary evaporator under reduced pressure at 40°C to form a thin lipid film. The film was dried in a vacuum oven for 12 hours to ensure complete removal of the solvent. Subsequently, the lipid film was hydrated at 70°C with HBSS-HEPES buffer by vortexing. The ssLips were prepared using an extruder (LipoFast-Pneumatic; Avestin, Inc., Ottawa, ON, Canada) with a size-controlled polycarbonate membrane (pore sizes of membrane filter: 0.1, 0.2, 0.4, or 0.8 μm; Whatman Japan KK, Tokyo, Japan). Extrusion was performed 41 times under nitrogen pressure (200 psi). The final DSPC and coumarin-6 concentrations in the resultant liposomal suspension were 20.4 μmol/mL and 0.143 μmol/mL, respectively. 

Table 1.

 

Characterization of Fluorescence-Labeled Particles

Samples	Lipid Composition (molar ratio)	Particle Size (nm)	Zeta Potential (mV)
MLV	DSPC/DCP/cholesterol = 8:2:1	6490.0	−113.7
ssLip 600	DSPC/DCP/cholesterol = 8:2:1	561.0	−76.9
ssLip 300	DSPC/DCP/cholesterol = 8:2:1	300.6	−67.8
ssLip 200	DSPC/DCP/cholesterol = 8:2:1	174.8	−58.0
ssLip 100	DSPC/DCP/cholesterol = 8:2:1	116.8	−53.9
ssLip 100 (8:2:1)	DSPC/DCP/cholesterol = 8:2:1	116.8	−53.9
ssLip 100 (8:2:4)	DSPC/DCP/cholesterol = 8:2:4	121.9	−54.6
ssLip 100 (8:2:8)	DSPC/DCP/cholesterol = 8:2:8	114.0	−53.5
ssLip-negative	DSPC/DCP/cholesterol = 8:2:1	116.8	−53.9
ssLip-neutral	DSPC/cholesterol = 8:1	128.0	−3.0
ssLip-positive	DSPC/SA/cholesterol = 8:0.2:1	134.4	25.9
Polystyrene particle		110.7	−49.8
Lipid emulsion		109.1	−53.8

 Zeta potential is the potential difference between the dispersion medium and the stationary layer of fluid attached to the dispersed particle. ssLip 100, ssLip 100 (8:2:1), and ssLip-negative are substantially the same.

Preparation of FITC-Labeled Polystyrene Particle Suspension

Polystyrene particles containing covalently bonded FITC were used in this study. To adjust the osmotic pressure of FITC-labeled polystyrene particle suspension (Micromer-GreenF; Corefront Co., Ltd.), glucose was added to the suspension. Final glucose and polymer concentrations were 50 mg/mL and 9 mg/mL, respectively. 

Preparation of Lipid Emulsion

Lipid emulsion was composed of EPC, DCP, caprylate and caprate triglyceride (Triester F-810; Nikko Chemicals Co.), and coumarin-6. A mixture of EPC and DCP at a molar ratio of 8:2 was dissolved in a small amount of chloroform. Caprylate and caprate triglyceride and coumarin-6 were added as an oil core and a fluorescence marker of lipid emulsion, respectively. The lipid mixture was dried in a rotary evaporator under reduced pressure at 40°C to form a lipid film. The film was dried in a vacuum oven for 12 hours to ensure complete removal of the solvent. Subsequently, the lipid film was hydrated at 70°C with HBSS-HEPES buffer by vortexing. The obtained lipid emulsion was downsized with a probe-type sonicator for 10 minutes. Final EPC, caprylate and caprate triglyceride, and coumarin-6 concentrations were 20.4 μmol/mL, 22.4 mg/mL, and 0.143 μmol/mL, respectively. 

Physicochemical Properties of the Particles

The submicron-sized particle size was measured with an aliquot of the particulate suspension diluted with a large amount of distilled water by the dynamic light-scattering method (Zetasizer; Malvern, Worcestershire, UK). MLV particle size was measured by a laser diffraction size analyzer (LDSA-2400A; Tonichi Computer Applications, Tokyo, Japan). The zeta potential of particles was measured using a laser Doppler method (Zetasizer; Malvern). Each batch was analyzed in triplicate. 

Intraocular Behavior of Fluorescence-Labeled Particles in Mice

Unanesthetized male adult ddY mice (Japan SLC, Hamamatsu, Japan) weighing 30 to 35 g were used. The animals were fed a regular diet. All experiments were conducted in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The experimental protocol was approved by the Institutional Animal Care and Use Committee of Gifu Pharmaceutical University. A single dose of 3 μL fluorescence-labeled formulation was dropped onto the surface of the left eye. The contralateral eye was used as the control and received no treatment. The mice were then euthanatized 10, 30, 60, or 120 minutes after eyedrop administration. Both eyes were enucleated immediately, washed with an excess amount of saline, and fixed overnight in 4% paraformaldehyde at 4°C. Fixed eyes were immersed in 20% sucrose for 48 hours at 4°C and embedded in optimum cutting temperature compound (Sakura Finetek Co., Ltd., Tokyo, Japan). Samples were sliced with a cryostat (CM1850; Leica Instrument GmbH, Nussloch, Germany) into sections 10-μm thick and were placed onto slides under a coverslip. The retinal images, taken at distances between 375 and 625 μm from the optic disc of frozen sections, were observed using epifluorescence microscopy (BX50; Olympus, Tokyo, Japan) with an attached charge-coupled device (CCD) camera (DP30VW; Olympus) and fluorescence filters for coumarin-6 (U-WNIBA; Olympus). In the inner plexiform layer (IPL) and outer plexiform layer (OPL) at a distance between 475 and 525 μm (50 μm × 50 μm, IPL; 50 μm ×15 μm, OPL) from the optic disc, the fluorescence intensity of coumarin-6 was evaluated with appropriately calibrated computerized image analysis, using median density as an analytic tool (ImageJ software, developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). Fluorescence intensity of coumarin-6 was measured in the range of 0 to 255 as the mean density, using Image J at the constant area. Relative intensity indicates the value of a treated sample when the fluorescence intensity of an untreated sample is estimated as 1. To obtain the retinal flat-mount image, lenses were removed and retinas dissected from the eye after 30 minutes of dropped liposomes. Then various single images of different parts of the retina were taken with fluorescence microscopy (BZ-9000; Keyence, Osaka, Japan). Finally, the single images were combined to obtain a schematic representation of the retinal flat-mount image. 

Intraocular Behavior of Submicron-Sized Liposomes in Rabbits and Monkeys

Japanese White rabbits (Japan SLC, Hamamatsu, Japan) weighing 2.4 to 2.6 kg and cynomolgus monkeys (Keari, Osaka, Japan) weighing 3.8 to 4.2 kg were used. The animals were fed a regular diet. Doses (1×, 3×, 5×) of 35 μL ssLip were dropped onto the surface of the left eye every 5 minutes. The contralateral eye was used as the control and received no treatment. The animals were then euthanatized 10, 30, 60, 120, or 360 minutes after the first administration of the ssLip. Both eyes were enucleated immediately and washed with excess amounts of saline. The cornea and the lens were removed and fixed overnight in 4% paraformaldehyde at 4°C. Fixed eyes were immersed in 20% sucrose for 48 hours at 4°C and embedded in optimum cutting temperature compound. The samples were then sliced with a cryostat into sections 10-μm thick and were placed onto slides under a coverslip. Retinal images, taken at distances between 875 and 1125 μm from the optic disc of frozen sections, were observed using epifluorescence microscopy. In the IPL at a distance between 975 and 1025 μm (50 μm × 50 μm) from the optic disc, the medium fluorescence intensity of coumarin-6 was evaluated with ImageJ software according to the same evaluation method used for the mice. 

Statistical Analysis

Data are presented as the mean ± SEM. Statistical comparisons were made using one-way ANOVA followed by Dunnett's test. P < 0.05 was considered to indicate statistical significance. 

Results

Intraocular Behaviors of Dropped Particles in Mice

To evaluate the particle size effect of dropped liposomes on intraocular behavior, liposomes of various particle sizes were prepared. Average particle sizes of MLV, ssLip 600, ssLip 300, ssLip 200, and ssLip 100 were 6490, 561.0, 300.6, 174.8, and 116.8 nm, respectively (Table 1). Coumarin-6 was incorporated into each liposome as a fluorescence marker for assessing intraocular behavior with epifluorescence microscopy (Fig. 1). Figure 2 shows epifluorescence microscopic images of the retina in mice 30 minutes after eyedrop administration of liposomes. Negligible fluorescence in the retina was observed when MLV were used as carriers (Fig. 2B). Fluorescence intensities were approximately the same as those observed for the untreated eyeball (Fig. 2A). However, the size reduction of liposomes induced an accumulation of coumarin-6, as shown in an increase in the fluorescence intensity in the retina of mice (Figs. 2C–F). The fluorescence of ssLips was distributed in the optic nerve fiber layer, IPL, OPL, and inner and outer segments except each nuclear layer. The fluorescence intensity in IPL and OPL of the retinal samples was clearly increased with decreasing ssLips particle size (Fig. 2G). Consequently, the ssLip with a 100-nm particle sizes showed the most effective delivery of coumarin-6 to the retina on eye dropping. 

Figure 1.

 

Scheme of coumarin-6–entrapped liposome. Liposomes are spherical vesicles with phospholipid bilayers. They can encapsulate drugs inside the cavity between the bilayers, depending on the hydrophilicity or the hydrophobicity of the drug. Coumarin-6 was incorporated into the lipid bilayers.

Figure 2.

 

Effects of various liposomal particle sizes on delivery efficiency to the retina in mice. Epifluorescence microscopic images of the retina in (A) untreated eye 30 minutes after eyedrop administration of (B) MLV, (C) ssLip 600, (D) ssLip 300, (E) ssLip 200, and (F) ssLip 100. (G) Changes in the measured median fluorescence intensity in the IPL and OPL after administration of liposomes of different particle sizes. Particle sizes of MLV, ssLip 600, ssLip 300, ssLip 200, and ssLip 100 were 6490.0, 561.0, 300.6, 174.8, and 116.8 nm, respectively. Data are shown as mean ± SEM (n = 4). *P < 0.05, **P < 0.01 versus each OPL. Scale bar, 50 μm.

To investigate the downsizing effect of carriers on retinal delivery, a different carrier was used as a candidate. FITC-labeled polystyrene particles and lipid emulsions of 110 nm were used as carriers in Figures 3A to 3C. As shown in Table 1, particle size of the polystyrene particles and lipid emulsions were 110.7 nm and 109.1 nm, respectively. The zeta potential of these carriers was similar to that of ssLip 100. FITC luminescence on the epifluorescence microscopic image of the retina was not detected after administration of the polystyrene particle, even though particle size and zeta potential were in a range similar to those for ssLip 100 (Fig. 3A). Fluorescence intensity of those samples was the same as that diluted 5000 times (Fig. 3D). Covalently linked FITC-fluorescence polystyrene was not delivered to the retina after eyedrop administration. In the case of lipid emulsion, weak fluorescence emission of coumarin-6 was detected after eyedrop administration (Fig. 3B). This result suggests that lipid emulsion was not efficient for the delivery of coumarin-6 to the retina. 

Figure 3.

 

Differences in delivery efficiency to the retina in mice induced by varying constituent elements of particles. Epifluorescence microscopic images of the retina 30 minutes after eyedrop administration of (A) polystyrene particles, (B) lipid emulsions, and (C) ssLip 100 in mice. (D) Fluorescence intensity of polystyrene particles and ssLip 100 diluted 5000 times. (E) Accumulated fluorescence intensity in the IPL after administration of nanoparticles. Data are shown as mean ± SEM (n = 4). **P < 0.01 versus untreated. Scale bar, 50 μm. Excitation and emission wavelengths were fixed at 490 and 520 nm, respectively.

To evaluate the effect of liposomal rigidity of dropped liposomes on their delivery efficiency to the retina, a series of liposomes with different cholesterol contents (DSPC/DCP/cholesterol = 8:2:1, 8:2:4, or 8:2:8) were prepared (Table 1). Effects of liposomal cholesterol content on coumarin-6 delivery to the retina are shown in Figure 4. Fluorescence intensity of coumarin-6 in IPL decreased with an increase in liposomal cholesterol content. 

Figure 4.

 

Effects of various liposomal cholesterol contents on delivery efficiency to the retina in mice. Epifluorescence microscopic images of the retina in (A) untreated eye 30 minutes after eyedrop administration of (B) ssLip 100 (8:2:1), (C) ssLip 100 (8:2:4), and (D) ssLip 100 (8:2:8). (E) Changes in the accumulated fluorescence intensity in the IPL after administration of liposomes of different cholesterol contents. Data are shown as mean ± SEM (n = 4). **P < 0.01 versus ssLip 100 (8:2:1). Scale bar, 50 μm.

Effects of the surface charge of dropped liposomes on their delivery efficiency to the retina are shown in Figure 5. The positively charged ssLip (zeta potential +25.9 mV) was prepared by incorporating SA instead of DCP. The zeta potential of ssLip, which did not contain a charged additive (ssLip-neutral), was −3.0 mV. The particle sizes of these different charged ssLips were similar (Table 1). Figure 5N shows the changes in the medium fluorescence density per pixel in the IPL for coumarin-6–entrapped liposomal systems. The fluorescence density of coumarin-6 in the IPL gradually increased with time, peaking 10 to 30 minutes after administration of all ssLips. Fluorescence decreased thereafter to about one-third of the maximum intensity at 120 minutes after administration. Changes in fluorescence density occurred in the same manner regardless of the zeta potential of the liposomes, indicating no significant differences in intraocular behavior among these ssLips. 

Figure 5.

 

Effects of surface charge of ssLips on delivery efficiency to the retina in mice. Epifluorescence microscopic images of the retina after definitive intervals of eyedrop administration (0 minutes, A; 10 minutes, B, F, J; 30 minutes, C, G, K; 60 minutes, D, H, L; 120 minutes, E, I, M). (A) Untreated eye, (B–E) ssLip negative, (F–I) ssLip neutral, and (J–M) ssLip positive. (N) Changes in the accumulated fluorescence intensity in the IPL as a function of time after administration. The zeta potential of the ssLip-negative, ssLip-neutral, and ssLip-positive was −53.9, −3.0, and 25.9 mV, respectively. Data are shown as mean ± SEM (n = 4). There were no statistical differences at each time course. Scale bar, 50 μm.

To clarify the delivery route and distribution of the ssLip in the retina after administration onto the ocular surface, the retinal flat-mount images were observed at 30 minutes after administration of ssLip 100 (Fig. 6). Liposome-mediated fluorescence was obvious in the retina after administration of ssLip 100, whereas no fluorescence was observed in the untreated eyeball. In addition, the concentration gradient of coumarin-6 was clearly observed from the iris and ciliary body side to the optic disc side. This result demonstrated that ssLips were delivered from the ocular surface to the retina. 

Figure 6.

 

Retinal flat-mount images on epifluorescence microscopy. Each retinal image represents (A) untreated eye and (B) ssLip 100–treated eye 30 minutes after administration in mice. Scale bar, 500 μm.

Intraocular Behavior of ssLip in Rabbits and Monkeys

The pharmacokinetics and permeability of topically applied compounds, such as ssLips, in this study may vary significantly, depending on the geometry and size of the eye targeted. To assess the potential bioavailability and tissue distribution of ssLips in the human eye, we performed a series of experiments in which rabbit and monkey eyes, which are large and closely approximate the human eye, were used. Figure 7 shows the retinal images observed in rabbits after eyedrop administration of ssLip 100 with different dropping times. The fluorescence emission of coumarin-6 was also observed in the retina of the treated eyeball in the experiments using rabbits. The magnitude of accumulated fluorescence emission was increased with an increase in the number of drops. On the other hand, negligible fluorescence was observed in the retina of contralateral eyes administrated with ssLip at any time. These results clearly demonstrated that ssLips were delivered from the surface of the eye without delivery through a systemic route caused by nasolacrimal drainage. 

Figure 7.

 

Changes in the fluorescence in the retinas of rabbits with different numbers of eyedrops. Epifluorescence microscopic images show the retinas 30 minutes after administration of ssLip 100. The number of eyedrops was 1× (A, B), 3× (C, D), and 5× (E, F) every 5 minutes. Eyes were enucleated 30 minutes after the first administration of ssLip 100. Left and right images represent the retina and its corresponding contralateral eye. Scale bar, 50 μm.

A time-course observation was carried out to clarify the intraocular behavior of this liposomal system in rabbits (Fig. 8A). The fluorescence from coumarin-6 around the ganglion cell layer (GCL) and IPL were significantly higher than that from the untreated eye. Results of the magnitude of green emission in the IPL of the retina supported the changes in retinal images (Fig. 8B). The fluorescence emission of coumarin-6 on the retinal images gradually increased with time, peaking 30 to 60 minutes after administration of ssLip 100. Statistical data analysis clearly shows the significant difference between the untreated eye and 30 to 60 minutes after administration of ssLip 100. Fluorescence decreased thereafter to about one-third the maximum intensity at 360 minutes after administration. 

Figure 8.

 

Time-course observation of the retina after eyedrop administration of a single dose of ssLip 100 in rabbits. (A) Epifluorescence microscopic images of the retina 10, 30, 60, 120, and 360 minutes after eyedrop administration. (B) Changes in the accumulated fluorescence intensity in the IPL after eyedrop administration in rabbits. Data are shown as mean ± SEM (n = 3). **P < 0.01 versus untreated. Scale bar, 25 μm.

Further studies were performed using monkeys to estimate the possibilities of the liposomal system for pharmaceutical applications in humans given that monkey eyes are similar to those of humans. Strong emission was observed in the retinal image of monkeys, as shown in Figure 9. There was a significant difference between the ssLip-treated eye and the untreated eye. These results suggest that the liposomal system used in this study was applicable to the monkey eye. 

Figure 9.

 

Changes in the fluorescence in the retina of monkeys with a 5× dose of eyedrops. Epifluorescence microscopic images show the retinas 30 minutes after the first administration of ssLip 100. Right and left images represent the retina in the contralateral eye (A) and the ipsilateral eye (B). (C) Changes in the accumulated fluorescence in the IPL after the first administration of ssLip 100. Data are shown as mean ± SEM (n = 3). **P < 0.01 versus untreated. Scale bar, 50 μm.

Discussion

In a number of review articles, the topical administration of drugs to the posterior segment of the eye is discussed. ^7,8,16 However, there have been few reports in which a drug carrier system noninvasively targets the retina. One of the primary problems about delivering drugs to the posterior segment of the eye is the presence of corneal and conjunctival barriers. As described previously, ssLips have potential for delivering ophthalmic drugs to the posterior segment of the eye, and the rigidity of liposomal particles is an important factor in considering delivery efficiency to the retina. ¹⁵ In the present study, we focused on the effect of particle properties—e.g., particle size, surface charge, constituents—on intraocular behavior after eyedrop administration in mice, rabbits, and monkeys. To compare the delivery efficiency of these ssLips to the retina, the magnitude of green emission in the IPL of the retina was quantified using ImageJ software. The IPL seems to be a good target for evaluating retinal delivery because it is located very close to the GCL. Retinal ganglion cell death is a common feature in many ophthalmic disorders, such as glaucoma, optic neuropathy, and retinal vein occlusions. ¹⁷  

Particle size is the most important factor in improving the drug delivery efficiency of carrier particles. As shown in Figure 2, delivery efficiency of coumarin-6 was extensively improved by reducing liposomal particle size to the submicron order. In ophthalmic delivery systems, nano-sized particles represent a greater surface area available for association between the cornea and the conjunctiva. ¹⁸ Kassem et al. ¹⁹ have reported that the mean residence time of drugs on the ocular surface increases as the particle size in the drug suspension decreases. The extensive delivery efficiency to the retina of coumarin-6 was attributed to the ability of liposomes to associate with the ocular surface tissues, partly because of their prolonged retention property on the ocular mucosa. The solubility of coumarin-6 to distilled water is extremely low; hence, DMSO solution was used as a control because its coumarin-6 concentration is similar to that of the liposomal formulation. No fluorescence was observed for the DMSO solution containing coumarin-6, in contrast to the liposomal formulations. ¹⁵ Therefore, coumarin-6 molecules dissolved in DMSO solution cannot deliver to the retina as a molecular state. 

As shown Figure 3A, no fluorescence was observed in the retina after eyedrop administration of polystyrene particles. The lipid emulsions showed negligible emission of coumarin-6 after administration, whereas ssLips showed remarkable emission intensity. The delivery efficiency was different among those colloidal systems, even if both their particle size and zeta potential were approximately controlled as 110 nm and −50 mV, respectively. Given that polystyrene particles contain covalently bonded FITC, the absence of fluorescence in the retina suggested that polystyrene particles themselves could not reach the retina in our experiment. Amrite et al. ²⁰ reported that 20-nm polystyrene particles did not permeate the sclera-choroid-RPE in 24 hours. They also reported that intraocular tissues such as the retina and the vitreous did not have any quantifiable uptake of the polystyrene particles (particle size: 20, 200, 2000 nm) after subconjunctival administration. ²¹ Particle size is definitely a dominant parameter for effective delivery; however, another factor should be considered when choosing a colloidal carrier for retinal delivery. 

For the first step of the delivery, the colloidal particle should associate with the surface of the eye. The affinity between colloidal particles and cells may be a considerable factor in delivering drugs to the retina. It is well known that liposomes are composed of phospholipid, such as cell membrane (Fig. 1), and they show good affinity with the ocular barrier, such as the cornea and the conjunctiva. ²² Lipid emulsions are exploited commercially as a vehicle to improve the ocular bioavailability of lipophilic drugs. ^23,24 Naveh et al. ²⁵ noted that the intraocular pressure–reducing effect of a topically administered dose of pilocarpine-loaded lipid emulsions is prolonged compared with that of generic pilocarpine. Calvo et al. ²⁶ observed an improvement in indomethacin ocular bioavailability when the drug is incorporated in a lipid emulsion compared with commercial aqueous drops after topical application into the rabbit eye. Liposomes and lipid emulsion may have higher affinity than polystyrene particles. The lower affinity of polystyrene particles to cells may be the reason polystyrene particles do not show any fluorescence in the retina. Although both liposomes and lipid emulsions may have good affinity toward cells, coumarin-6 was not delivered effectively to the retina after administration of lipid emulsions (Fig. 3B). The excellent delivery efficiency of ssLips to the retina was not solely explained by the affinity to cells. Lipid emulsions are composed of an oil core surrounded by a phospholipid monolayer, whereas liposomes are composed of phospholipid bilayers. We assumed that the structural differences altered the stability of those colloidal particles and the release profile of hydrophobic material. The delivery efficiency of coumarin-6 might be attributed to the structural difference of liposome and lipid emulsion; in this experiment, the phospholipid bilayers of liposomes might have been better surroundings for coumarin-6 than the phospholipid monolayer of lipid emulsions. 

As shown in Figure 4, the delivery efficiency of coumarin-6 from liposomes decreased with an increase in liposomal cholesterol content. Cholesterol usually acts as a stabilizer for liposome formulation. In fact, the hardness of liposomes increased with increasing cholesterol content with which the unsaturated phospholipid (e.g., egg phosphatidylcholine [EPC]) was used as a lipid component of liposomes (EPC/DCP/cholesterol). However, cholesterol might have acted as a plasticizer in our formulation (DSPC/DCP/cholesterol) using saturated phospholipid of DSPC as a lipid component of liposomes. This trend has already been reported by Utsumi et al. ²⁷ In addition, we previously reported that atomic force microscopy images showed that DSPC/SA/cholesterol liposomes changed from spheres to ovals on mica surfaces with an increase in cholesterol content, indicating the decrease of liposomal rigidity with an increase in cholesterol content. ²⁸ The rigidity of DSPC/DCP/cholesterol liposomes may decrease with an increase in cholesterol content. We have reported that the efficiency of coumarin-6 delivery to the retina is higher for DSPC/DCP/cholesterol liposomes than for EPC/DCP/cholesterol liposomes. ¹⁵ These results supported the idea that the lower rigidity of liposomes is disadvantageous for retinal delivery. 

Several researchers have reported that the surface charge of liposomes affects the pharmacokinetics of drugs entrapped in liposomes. ^9,10 Law et al. ⁹ reported that the absorption rate of acyclovir in positively charged liposomes is slower than the absorption rate of acyclovir in negatively charged liposomes. Hathout et al. ¹⁰ reported that acetazolamide-entrapped positively charged MLV lowered intraocular pressure more than negatively charged MLV. In contrast to expectations, there were no significant differences in intraocular behavior among negatively, neutrally, and positively charged ssLips (Fig. 5). The positively charged liposome (DSPC/SA/cholesterol, 8:0.2:1) and the neutral liposome (DSPC/cholesterol, 8:1) exhibited an effective transfer of coumarin-6 to the retina. Although further investigations are required using several types of charged liposomes, ssLips with different surface charges might be potential carriers for retinal delivery. 

Mouse eyes are very small, and their ocular barriers may be weaker than those of larger animals. The mean thickness of the cornea, the main absorption barrier, in mice, rats, and rabbits is approximately 110 μm, 160 μm, and 350 μm, respectively. ²⁹ Therefore, experiments using other animals are required to confirm the penetration of coumarin-6 to the retina. As shown in Figures 7 and 8, fluorescence emission of coumarin-6 was clearly observed in the retinas of rabbits. The rabbit eye is approximately 70% to 80% the size of the human eye in terms of axial length, diameter, corneal thickness, scleral thickness at the limbus, scleral thickness at the equator, and scleral surface area. ³⁰ In addition, the fluorescence emission of coumarin-6 was also observed in the retinas of monkeys (Fig. 9). Eyes of mammals such as rabbits, pigs, and monkeys are similar to those of humans. For these reasons, our findings are valuable for designing ocular drug delivery systems that target the human retina. 

Data shown in Figure 8 confirm that ssLip moved gradually to the retina within 60 minutes of administration. The disappearance of fluorescence at 360 minutes suggests two possible phenomena: clearance of the liposomal particles from the retina and collapse of the liposomal structure and resultant diffusion of coumarin-6 molecules in the retina. Episcleral and choroidal circulations play a significant role in clearing drugs from the retina after subconjunctival administration. ³¹ Amrite et al. ^20,21 reported that clearance of subconjunctivally administered 20-nm polystyrene particles can enter systemic circulation through the local intraocular circulation after uptake by the conjunctival or episcleral blood vessels. This disappearance of fluorescence might be attributed to the diffusion of ssLip itself or of coumarin-6 molecules to these periocular circulatory systems. 

Absorption of liposomes after topical administration to the surface of the eye is assumed to occur primarily through three routes: systemic, corneal, and noncorneal pathways. ³² However, negligible fluorescence in the retina of the contralateral eye (Figs. 7B, 7D, 7F) exhibited no contribution of systemic delivery from nasolacrimal drainage. It may be that liposomes are delivered to the retina by way of a transscleral pathway. As shown in Figure 2, the fluorescence emission in OPL proportionally increased with a decrease in particle size that corresponded well to the results shown in IPL. If liposomes were delivered by transscleral pathway, the concentration gradient of coumarin-6 from the scleral side to vitreous side of the retina may be observed. Additional evidence exists for the inability of small 100-nm liposomes small PLGA nanoparticles to permeate the sclera. ³³ Therefore, the movement of intact liposomes to the retina by transscleral pathways is unlikely. Based on the entire eye image after the administration of ssLip, we previously considered that the delivery of liposomes to the posterior segment of the eye may occur primarily by noncorneal pathways—liposome access through the tissues involving the trabecular meshwork, iris root, and pars plana. ¹⁵ However, we cannot rule out the possibility of the transscleral route. Further investigations will be needed in the future. The coumarin-6 concentration gradient from the iris and ciliary body to the optic disc was observed in the retinal flat-mount image (Fig. 6B). All four petal-shaped images exhibit similar fluorescence images depicting a fluorescence gradient from the ocular surface to the retina. This finding suggests that liposome-mediated fluorescence was distributed homogeneously in the dorsal, ventral, temporal, and nasal retina after eyedrop administration of ssLip. 

In the animal study using mice, rabbits, and monkeys, fluorescence emission of coumarin-6 was obvious in the retina when submicron-sized liposomes were topically administered as eyedrops. The magnitude of fluorescence in the retina was closely related to the particle size of liposomes. Polystyrene particles and lipid emulsions showed an insufficient effect compared with that of liposomes on retinal delivery, even if particle size and zeta potential were in a similar range. The liposomal surface charge was not affected by the intraocular behavior of liposomes by eyedrop administration. Epifluorescence microscopy of the retinal flat-mount image revealed that the delivery of liposomes to the retina might have occurred by diffusion from ocular surface involvement to the iris ciliary body side. These mechanisms will be investigated further with active pharmaceutical ingredients. 

Footnotes

 Supported by a Grant-in-Aid (21390011) from the Ministry of Education, Culture, Sports, Sciences and Technology (Monbukagakusho) of Japan.

Footnotes

 Disclosure: Y. Inokuchi, None; K. Hironaka, None; T. Fujisawa, None; Y. Tozuka, None; K. Tsuruma, None; M. Shimazawa, None; H. Takeuchi, None; H. Hara, None