Laser-Induced Intrachoroidal Dexamethasone Drug Delivery System to Posterior Eye Segment

Abstract

Purpose.: To investigate the feasibility of laser-induced intrachoroidal dexamethasone (DEX) delivery as a potentially useful therapy for adjusting the most effective drug level to the posterior segment eye diseases.

Methods.: An implant was prepared by dissolving poly(DL-lactide) and DEX. In vitro release of DEX was evaluated at 7, 14, and 28 days by ELISA. In vivo, a DEX implant was inserted into a rabbit choroid, and 10, 50, or 200 burns of photocoagulation were applied at the implant lesion. After treatment, the vitreous humor was immediately aspirated and the DEX level was measured by liquid chromatography/mass spectrometry/mass spectrometry. Furthermore, the vitreous DEX level was measured at 1, 7, 14, and 28 days after implantation and 50 burns of photocoagulation. The toxicity of the laser-induced DEX implant was evaluated by ophthalmoscopy and light microscopy. Endotoxin-induced uveitis (EIU) was induced after DEX implantation and photocoagulation, and anti-inflammatory activities were evaluated by grading clinical signs, protein concentrations, and histopathologic studies.

Results.: Photocoagulation significantly increased the DEX release from the implant at 7 days in vitro. In vivo, the DEX implant exposed to 10, 50, and 200 burns of photocoagulation increased the vitreous DEX levels in a dose-dependent manner. The vitreous DEX level in the DEX implant applied to 50 burns of photocoagulation peaked 1 day after treatment. The laser-induced DEX implant showed no retinal abnormalities except the implantation site, and significantly inhibited the EIU.

Conclusions.: Laser-induced intrachoroidal DEX delivery controls the DEX level in the vitreous humor and effectively prevents the experimental uveitis.

Introduction

The synthetic glucocorticoid, dexamethasone (DEX), is widely used for the treatment of posterior segment eye diseases including uveitis. ^1,2 Posterior segment eye diseases require long-term treatment with DEX. It is difficult to deliver effective doses of DEX to the posterior part of the eye. Systemic administration necessitates large doses of DEX, resulting in general side effects such as hypertension, hyperglycemia, and gastric ulcer. ^3–7 Topical eye drops penetrate poorly into the posterior part of the eye, because of long diffusional path length, lacrimation, and corneal impermeability. ^3–7 Intravitreal injections needs to be repeated to maintain the therapeutic level of DEX and involve potential risks of glaucoma, cataract, retinal detachment, and endophthalmitis. ^3–7  

To circumvent these problems, sustained DEX drug delivery system would be helpful to increase site specificity, to reduce side effects in the surrounding tissues, and to prolong delivery time. ^3–7 Intravitreous sustained DEX delivery devices have been successfully used to treat posterior segment eye diseases. ^8–12 However, once the implant is inserted into the vitreous cavity through the pars plana, it is difficult to regulate the drug level corresponding to the severity of posterior segment eye diseases, and to adjust the most effective drug level to an individual patient. 

Photocoagulation has a long history of medical use, including ophthalmology. It has been reported that high intensity choroidal photocoagulation to create a chorioretinal venous anastomosis is effective for the treatment of central retinal vein occlusion. ^13–15 Therefore, photocoagulation applied at an intrachoroidal DEX implant may be useful to regulate adequate quantitative doses in the posterior eye segment. Transchoroidal drug delivery through the scleral approach offers a safe route for delivery by avoiding intraocular manipulation. ^16–23 Furthermore, laser photocoagulation is easily applied to an intrachoroidal implant at the posterior pole compared with an intravitreal implant at the pars plana. Here, we report that a laser-induced intrachoroidal sustained DEX delivery system controls the DEX level in the vitreous humor and effectively prevents experimental uveitis in rabbit eyes. 

Materials and Methods

Preparation of the Intrachoroidal Implant

Poly(DL-lactide) (PLA), with an average molecular mass of 20 kDa was purchased from Wako Pure Chemicals Industries, Ltd. (Osaka, Japan). Dexamethasone was purchased from Sigma Chemical Co. (St. Louis, MO). 

The intrachoroidal implant was prepared by dissolving 50 mg DEX, 150 mg PLA, and 100 μg indocyanine green (ICG) (Dai-ichi Pharmaceutical Company, Tokyo, Japan) in 5 mL of dioxane. Indocyanine green was used to confirm the implant site in each fundus ophthalmoscopically. The resultant solution was kept at −80°C and lyophilized for 48 hours to obtain a homogeneous cake. The cake was compressed (SSP-10; Shimadzu, Kyoto, Japan) and made into a cylindrical implant. The cylindrical implant weighed 80 μg and was 0.2 mm in diameter and 0.5 mm in length (Fig. 1A). The implant had a DEX loading of 25% wt/vol. 

Figure 1

 

Photographs of dexamethasone (DEX) implant and implantation site. (A) Dexamethasone-loaded biodegradable implant (arrow). (B) Implantation site in a rabbit eye. The implant (*) is inserted into a choroidal pocket (arrows).

In Vitro Pharmacokinetics

Each of the eight implants were placed in 1 mL of PBS (pH 7.4) in a closed vial, and then 50 burns of an argon green laser was applied in four of eight implants, with 1000 mW in power, 50 μm in spot size, and 0.05 seconds in duration. 

On days 7, 14, and 28, approximately 1 mL of the released medium was removed and replaced with the same quantity of fresh medium. The amount of DEX released into the medium was measured by a competitive ELISA for corticosteroid using a commercial kit specific for DEX determination (Corticosteroid Elisa Kit; Randox Laboratories, London, UK). The experiments were performed according to the manufacturer's instructions. 

Briefly, the medium and conjugate were applied to a 96-well microplate and incubated at room temperature for 1 hour. After the plate was washed with wash buffer, substrate solution was added and the incubation was continued at room temperature for 20 minutes. After incubation, stop solution was added into the plate, and then the optical density of each well was measured at 450 nm. 

Surgical Procedure

All experiments were conducted according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and experimental animal guidelines in Iwate Medical University. 

Three-month-old pigmented rabbits (Kitayama Labes, Nagano, Japan), each weighing 2 to 2.5 kg, were used. The rabbits were anesthetized with ketamine hydrochloride (24 mg/kg) and xylazine hydrochloride (6 mg/kg). The ocular surface was then anesthetized with a topical instillation of 2% xylocaine. The pupil was dilated with topical 0.5% tropicamide. 

Under an operating microscope, a fornix-based conjunctival flap was made at the inferior site. A sclerotomy was made 3-mm parallel to and 6-mm posterior to the corneal limbus, and the bare choroid was exposed. Thereafter, a 25-gauge sideport knife (MANI, Tochigi, Japan) was slightly inclined to the exposed choroid in a tangential direction, and shallowly stabbed. The MVR blade was then moved ahead in parallel to the surface of the choroid at the position corresponding to the half of the choroidal thickness, and also moved in left and right directions, thereby creating a pocket consisting of an incision of 1 × 1 mm. After these procedures, the implant prepared as above was inserted into this pocket (Fig. 1B). The scleral and conjunctival wounds were sutured with 9-0 nylon and 6-0 silk, respectively. In all experiments, the choroidal detachment, retinal detachment, and subretinal hemorrhage were not shown ophthalmoscopically. 

To examine the relation between laser intensity and vitreous DEX level, photocoagulation was simultaneously performed after the intrachoroidal implantation. Four rabbits each (four eyes) were applied to a circular argon green laser around the implant site (100 mW in power, 250 μm in spot size, and 0.2 seconds in duration) and then applied to 10, 50, and 200 burns of a high intensity laser at the implant lesion (800–1000 mW in power, 50 μm in spot size, and 0.05 seconds in duration), respectively. The high intensity laser burns were applied at the implant lesion to disrupt the retina and underlying the intrachoroidal implant. Most of experiments appeared no hemorrhage ophthalmoscopically, even though a few eyes showed minor hemorrhage, which resolved spontaneously within 1 week. After photocoagulation, 200-μL vitreous humor was immediately aspirated from each eye for the DEX level measurement. 

To further confirm the effect of the laser-induced DEX implant, the rabbits were divided into three groups of four rabbits each (four eyes). The rabbits in group 1 received the intrachoroidal implant with no DEX, while group 2 received the intrachoroidal DEX implant. Group 3 underwent the DEX implant in the same eye that was applied to 50 burns of the argon green laser. The vitreous humor (200 μL) was aspirated from each eye at 1, 7, 14, and 28 days after treatment for determination of the DEX concentration. 

In Vivo Release Study

The concentrations of DEX in the vitreous humor were determined by liquid chromatography/mass spectrometry/mass spectrometry (LC/MS/MS) using hybrid triple quadrupole/linear ion trap mass spectrometers (AB SCIEX 3200 QTRAP; AB SCIEX, Framingham, MA). The mass spectrometers were interfaced with an UPLC system (ACQUITY UPLC System; Waters, Milford, MA). 

Dexamethasone was extracted from the vitreous humor by the following procedure: 0.1 mL of the internal standard solution (propyl parahydroxy-benzonate in methanol: 1 μg/mL) and 3 mL of 0.2 M HCl were added to each sample. The mixture was centrifuged and the supernatant was collected. Dexamethasone was extracted twice with 3 mL of ethyl acetate. Ethyl acetate phases were then dried under reduced pressure with a centrifugal concentrator. Dried residues from the vitreous humor were reconstituted with 40% acetonitrile and 60% water containing 0.2% formic acid and 2 mM ammonium formate. 

UPLC (Waters) for vitreous humor samples containing the internal standard was performed on a C-18 column (2.1 × 100 mm, 1.7 μm; AQUITY UPLC BEH C18; Waters), using gradient elusion with 0.1% formic acid and 10 mM ammonium formate in solvent A (water) and solvent B (acetonitrile) at a flow rate of 0.2 mL/min. 

Mass spectrometric detection was accomplished by using positive ionization with electrospray ionization sources and scanning in the multiple reaction monitoring mode. The specific precursor product ion pairs used were a mass to charge ratio 393 → 91, and the LC/MS/MS retention time was approximately 13 minutes. Under these conditions, the detection limit for DEX was 0.1 ng/ml. 

Clinical Observations

To evaluate possible adverse effects of the intrachoroidal DEX implant and photocoagulation, treated eyes were observed under slit-lamp examination and indirect ophthalmoscopy at 7, 14, and 28 days. 

Histopathologic Examination

The treated eyes were enucleated after euthanasia at 28 days, and immersed in a mixture of 4% glutaraldehyde and 2.5% neutral buffered formalin for 24 hours. The cut specimens were dehydrated, infiltrated, embedded in paraffin, and sectioned with a microtome. Section was stained with hematoxylin and eosin. 

Uveitis Induction and Surgical Procedure

Twelve rabbits (twelve eyes) were divided into three groups, the intrachoroidal implant with no DEX in group 1, the intrachoroidal DEX implant in group 2, and the intrachoroidal DEX implant applied to 50 burns of the laser in group 3. 

After implantation or implantation/photocoagulation, endotoxn-induced uveitis (EIU) was simultaneously induced by lipopolysaccharide (LPS) from Escherichia coli (Sigma-Aldrich, St. Louis, MO) in all rabbit's eyes. One hundred nanograms of the LPS solution was injected intravitreally at the 12 o'clock position, 3-mm posterior to the limbus with a 27-gauge needle. ²⁴  

Clinical Observation

On days 7, 14, and 28, a masked observer used slit-lamp biomicroscopy and indirect ophthalmoscopy to evaluate the severity of inflammation. 

Vitreous haze was graded by examination of the posterior pole and visibility of the optic nerve from 0 to 4 (0 = no inflammation, +0.5 = trace inflammation, +1 = mild blurring of retinal vessels and optic nerve, +2 = moderate blurring of optic nerve head, +3 = marked blurring of optic nerve head, +4 = optic nerve head not visible, or visibility of the posterior pole was prevented by corneal swelling). ²⁵  

Protein Measurement of Vitreous Humor

Four eyes in each group were used at days 7, 14, and 28 after treatment for protein measurement. The vitreous humor was collected from the eye of each rabbit, and centrifuged at 250g for 5 minutes, and then the supernatant was used for protein measurement. The protein content of the vitreous humor was determined using a Bio-Rad (Richmond, CA) assay kit with bovine serum albumin as a standard dilution reference curve, according to the manufacturer's recommendation. 

Histopathologic Study

The experimental eyes were enucleated after euthanasia on day 28, and assessed by histopathology. 

Statistical Analysis

Statistical analysis was conducted by Kruskal-Wallis test and post hoc test (Scheffe's F). Mann-Whitney U test was used for the analysis of in vitro and in vivo DEX release. P values less than 0.05% were considered as significant differences. 

Results

In Vitro Pharmacokinetics of DEX From the Implant

To examine whether photocoagulation increases the DEX level, we first investigated the effects of photocoagulation on DEX delivery in vitro. The cumulative release of DEX from the implant is plotted in Figure 2A. To summarize, in the absence of photocoagulation, the cumulative DEX release to the medium from our DEX implant was 29.4 ± 6.6, 49.3 ± 8.2, and 69.0 ± 0.3% at 7, 14, and 28 days, respectively. With photocoagulation, these values showed 47.8 ± 3.6, 70.1 ± 12.8, and 91.3 ± 0.1%, respectively. Although our implant gradually released DEX over 28 days even in the absence of photocoagulation, photocoagulation resulted in a significant increase in the DEX release rate after 7 days (P = 0.019 < 0.05). The DEX implant with photocoagulation showed an initial burst on day 7, and then the DEX was gradually released. This suggests that the laser photocoagulation effectively released DEX from the DEX implant in vitro. 

Figure 2

 

Dexamethasone release from the implant after laser photocoagulation in vitro and in vivo. (A) Profiles of in vitro release of DEX from the implant. The data are the mean ± SD (n = 4). *P < 0.05. The DEX implant applied to photocoagulation showed a significant increase in release rate at 7 days compared with the DEX implant without photocoagulation. (B) Dexamethasone concentration in the vitreous humor from the implant exposed to 10, 50, and 200 burns of argon green laser. The data are the mean ± SD (n = 4). *P < 0.05, **P < 0.01. Laser photocoagulation increases the vitreous DEX levels in a dose-dependent manner. (C) Time-dependent DEX release to the vitreous humor. The data are the mean ± SD (n = 4). *P < 0.05. The DEX concentration in the no–DEX-implant group was below the detection limit during the observation period. The level of the DEX in the laser-induced DEX-implant group at 1 and 7 days after treatment was significantly higher than that of the DEX-implant group.

In Vivo Pharmacokinetics of DEX in the Rabbit Eye

To examine the relation between laser intensity and rabbit vitreous DEX level, the photocoagulation was simultaneously performed after intrachoroidal DEX implantation to rabbit eyes and the vitreous DEX levels were measured by LC/MS/MS. The DEX concentrations in the vitreous humor of the implant exposed to 10, 50, and 200 burns were 39.4 ± 8.1, 174.2 ± 27.6, and 857.8 ± 80.2 ng/mL, respectively (Fig. 2B). The quantity of DEX released by the DEX implant exposed to 200 burns was 21.8- and 4.9-fold greater than that released by DEX implants exposed to 10 or 50 burns (P < 0.01 vs. 10 burns, P < 0.01 vs. 50 burns), respectively, while the DEX level released by the implant exposed to 50 burns was 4.5-fold greater than that released by the implant exposed to 10 burns (P < 0.05 vs. 10 burns). This indicates that laser photocoagulation increases the vitreous DEX levels in a dose-dependent manner. 

To further confirm the effect of the DEX implant exposed to laser burns, vitreous DEX levels were measured using LC/MS/MS in eyes treated with an implant containing no DEX, a DEX implant, and a laser-induced DEX implant. A set of 50 burns with an argon laser was chosen as the exposure condition because 0.15 to 4.0 μg/mL of DEX is known to be the effective dose for inhibition of inflammation. ^26–30 The DEX level in the vitreous humor at 1, 7, 14, and 28 days after treatment are shown in Figure 2C. The vitreous DEX concentration in the no–DEX-implant group was below the detection limit during the observation period. The vitreous DEX concentration in the DEX implant group at 1, 7, 14, and 28 days showed 1.48 ± 0.45, 1.65 ± 0.49, 0.92 ± 0.41, and 0.54 ± 0.42 ng/mL, respectively, while the vitreous DEX level in the laser-induced DEX-implant group at 1, 7, 14, and 28 days showed 85.0 ± 14.14, 11.58 ± 3.60, 1.97 ± 1.07, and 0.98 ± 0.22 ng/mL, respectively. In the laser-induced DEX-implant group, the DEX level at 1 and 7 days after treatment was significantly greater, by 57.4- and 7.0-fold, respectively, than that in the DEX-implant group (day 1; P = 0.020 < 0.05, day 7; P = 0.020 < 0.05), suggesting that the laser-induced DEX delivery system significantly increases the vitreous DEX level in a time-dependent manner in vivo. 

Clinical Observation

To evaluate possible adverse effects of the DEX implant and photocoagulation on ocular tissues, treated eyes were assessed by slit-lamp examination and indirect ophthalmoscopy. 

In all experiments, slit-lamp examination showed no significant infection, and the cornea, anterior chamber, and lens were clear during the observation period. 

Ophthalmoscopically, both the no–DEX-implant and the DEX-implant groups were shown a whitish lesion with hyperpigmentation, corresponding to the site of the implant at all time points (Figs. 3A, 3B). The clinical features of the retina in the DEX-implant group did not differ from those in the no–DEX-implant group. The laser-induced DEX-implant group was seen a scar formation at the site of the implant treated with photocoagulation during the observation period (Fig. 3C). Even though the retinas at the implant sites in all groups were degenerated, the retinas around the implant sites showed no abnormalities. 

Figure 3

 

Representative fundus photographs 4 weeks after treatment. (A) No-DEX implant, (B) DEX implant, (C) Laser-induced DEX implant. The eyes in the no-DEX implant (A) and the DEX implant (B) were shown a whitish lesion with hyperpigmentation at the implantation site (arrow). The eye in the laser-induced DEX implant (C) was observed a scar formation at the site of the implant (arrow). Even though the retinas at the implant sites in all groups, were degenerated, the retinas around the implant sites showed no abnormalities.

Histopathologic Examination

Possible adverse effects of the no-DEX implant, DEX implant, and laser-induced DEX implant on ocular tissues were evaluated by light microscopy. 

In all groups, the implants were still seen at the implantation sites on day 28 after treatment. The retinas in the no–DEX-implant group and the DEX-implant group were observed to have loss of photoreceptor outer segments and outer nuclear layer, corresponding to the sites of the implants (Figs. 4A, 4B). Morphologic findings were not significantly different between the no–DEX-implant group and the DEX-implant group at day 28. The retina and choroid in the laser-induced DEX-implant group were shown moderately degeneration at the implant site (Fig. 4C). Although the retinas at the implant sites in all groups were degenerated, the retinas, except the implant sites, showed no abnormalities. 

Figure 4

 

Representative photographs of histologic sections from treated eyes 4 weeks after treatment. The eyes are same eyes shown in Figure 3A through 3C. (A) No-DEX implant, (B) DEX implant, (C) laser-induced DEX implant. In all groups, the implants were still seen at the implantation sites (*). The eyes in the no-DEX implant (A) and DEX implant (B) were shown loss of photoreceptor outer segments and outer nuclear layer, corresponding to the site of the implant. In the eye of the laser-induced DEX implant (C), the retina and choroid were observed moderately degeneration at the implant site. GCL, ganglion cell layer; INL, inner nuclear layer; Ch, choroid; Scl, sclera. Scale bar: 250 μm.

Effect of Laser-Induced Choroidal DEX Implantation on the Inhibition of Experimental Uveitis

Because the laser-induced DEX implant showed the effective DEX level in the vitreous and displayed minimal adverse effects on the ocular tissues, we investigated whether the laser-induced DEX implant was effective at preventing experimental uveitis. 

Ophthalmoscopically, severe vitreous exudates were noted within 7 days after LPS injection and vitreous opacity continued for 1 month in the no–DEX-implant group and the DEX-implant group. The laser-induced DEX-implant group inhibited development of vitreous opacity at all time points. 

The results of the clinical grading of vitreous opacity are shown in Figure 5A. The clinical grading in the no–DEX-implant group at 7, 14, and 28 days were 3.50 ± 0.57, 3.25 ± 0.50, and 2.50 ± 0.57, respectively, while in the DEX-implant group, they were 3.0 ± 0.81, 2.75 ± 0.50, and 2.0 ± 0.81, respectively. In the laser-induced DEX-implant group, they were 1.50 ± 0.57, 1.25 ± 0.50, and 0.75 ± 0.28, respectively. Clinical scores indicating the severity of uveitis were no significantly different between the no–DEX-implant and the DEX-implant group for all the following time points. However, the clinical scores of the laser-induced DEX-implant group were significantly lower those of the no–DEX-implant and the DEX-implant group during the observation period (day 7; P < 0.05 versus DEX implant, P < 0.01 versus no-DEX implant, day 14; P < 0.01 versus DEX implant, P < 0.01 versus no-DEX implant, day 28; P < 0.05 versus DEX implant, P < 0.01 versus no-DEX implant). Therefore, the data indicate that the laser-induced DEX implant effectively reduces vitreous opacity in EIU. 

Figure 5

 

Effect of DEX implant with laser photocoagulation on uveitis induction. (A) The clinical grading of vitreous opacity after the induction of uveitis. The data are the mean ± SD (n = 4). *P < 0.05, **P < 0.01. The clinical score of the laser-induced DEX-implant group was significantly lower than that of the no–DEX-implant and the DEX-implant group at all time points. (B) Protein concentration at 7, 14, and 28 days after the induction of uveitis. The data are the mean ± SD (n = 4). **P < 0.01. The eye in the laser-induced DEX-implant group suppressed the mean protein concentration at all time points compared with the eyes in the no–DEX-implant and the DEX-implant group.

Because EIU increases the protein level in the aqueous humor, ²⁴ we also investigated whether the laser-induced DEX implant was effective at reducing protein in the vitreous humor. The results of the protein concentration are shown in Figure 5B. The protein concentrations in the no-DEX implant at 7, 14, and 28 days were 40.5 ± 9.4, 33.5 ± 8.4, and 20.9 ± 6.5 mg/mL, respectively. With the DEX implant, they were 33.0 ± 5.8, 25.8 ± 4.8, and 16.5 ± 6.6 mg/mL, respectively, while with the laser-induced DEX implant, they were 8.4 ± 2.1, 5.4 ± 2.8, and 4.9 ± 1.7 mg/mL, respectively. The eyes in the laser-induced DEX-implant group significantly suppressed mean protein concentrations at all time points compared with the eyes in the no–DEX-implant and the DEX-implant group (day 7; P < 0.01 versus DEX implant, P < 0.01 versus no-DEX implant, day 14; P < 0.01 versus DEX implant, P < 0.01 versus no-DEX implant, day 28; P < 0.01 versus DEX implant, P < 0.01 versus no-DEX implant). The data also indicate that the laser-induced DEX implant significantly reduces the increased vitreous protein in EIU. 

Histopathologically, in the no–DEX-implant and the DEX-implant group, marked inflammatory cells were seen on the surface of the retina and in the vitreous, corresponding to the lesions of white exudates visible ophthalmoscopically (Figs. 6A, 6B). However, the laser-induced DEX-implant group had greatly reduced inflammatory cells on the surface of the retina and in the vitreous at 28 days after treatment (Fig. 6C). 

Figure 6

 

Representative photographs of histologic sections from treated eyes 4 weeks after the induction of uveitis. (A) No-DEX implant, (B) DEX implant, (C) laser-induced DEX implant. In eyes of the no-DEX implant (A) and the DEX implant (B), marked inflammatory cells were seen on the surface of the retina and in the vitreous. The eye of the laser-induced DEX implant (C) greatly reduced inflammatory cells on the surface of the retina and in the vitreous. Vit, vitreous; Ret, retina. Scale bar: 250 μm.

Discussion

In this study, we showed the feasibility of laser-induced intrachoroidal DEX implant in rabbit eye. Dexamethasone was released from the implant intrachoroidally and penetrated the retina and vitreous. The concentration of DEX was regulated by using photocoagulation, without showing severe adverse effects on the eye. Furthermore, the laser-induced intrachoroidal implant significantly inhibited uveitis ophthalmoscopically and histopathologically and the protein of vitreous humor was significantly reduced. 

This implant delivers its drug into the vitreous as the polymer degrades. In this study, we chose PLA with a molecular mass of 20 kDa as a relatively rapidly degradable polymer. The degradation of the matrix depends on its molecular weight and the release rate decreased with the increase of the molecular weight. ³¹ Further investigation may be needed to evaluate an adequate PLA molecular weight for long-term treatment. 

In the vitro study, laser-induced DEX delivery showed an increased release rate, with an initial burst and a second stage derived from diffusional release, whereas a DEX implant with no photocoagulation showed stable, long-term sustained release of the drug (Fig. 2A). Indocyanine green remained in the implant for 28 days, because it had been mixed into the implant homogeneously. Therefore, additional laser photocoagulation might be possible to increase the DEX level for adjusting the most effective drug level. In the present study, we used an argon green laser according to available references on high-intensity choroidal photocoagulation. ^13–15 However, it might be better to use a longer wavelength laser, because the ICG has greater absorption in the red wavelength. ^32,33 Further study is needed to assess the wavelength for laser treatment. 

In vivo, the DEX implant with no photocoagulation showed a quantity of intravitreal DEX that remained at a constant level for 4 weeks after treatment. The laser-induced DEX implant, on the other hand, showed an effective level for suppressing inflammation in in the vitreous at day 1, and then decreased the DEX level at day 7. On days 14 and 28, the DEX level of the laser-induced DEX implant was similar to that of the DEX implant without photocoagulation (Fig. 2C). 

Glucocorticoids such as DEX exert their anti-inflammatory effects by influencing multiple signal transduction pathways. ^34–37 By binding to cytoplasmic glucocorticoid receptors, corticosteroids in high doses increase the activation of anti-inflammatory genes, whereas at low concentrations they have a role in the suppression of activated inflammatory genes. ^35,38 Therefore, a drug-release profile that consists of an initial phase of high level of DEX, followed by a second phase of lower level, may continue to contribute to anti-inflammatory action of DEX for the duration of the implantation. Because DEX has physiological effects that are 30 times greater than hydrocortisone, ^39,40 even low concentrations of DEX during the second phase of drug delivery may be associated with significant biological activity in vivo, as suggested by inhibiting the experimental uveitis. 

The RPE, which is the outer blood–retina barrier, is between the choroid and the retina. The RPE has tight junctions of the nonleaky type and has low permeability to many compounds. ⁴¹ Our study revealed that the retina and choroid, corresponding to the site of the implant, were moderately degenerated, especially laser-induced intrachoroidal implant. Therefore, we speculate that DEX may pass through the RPE to the retina. In addition, molecules larger than 40 kDa cannot diffuse across the internal limiting membrane (ILM) of the retina, precluding intravitreous drug delivery. ^42–44 However, DEX has a small molecular weight of 392 Da, and accordingly it may permeate to the vitreous through the ILM. 

It has been reported that glucocorticoid concentration was higher in the retina and choroid than in the vitreous in an intravitreal release system. ^45,46 Presumably, DEX may not distribute by simple diffusion in the eye, and the clearance of the drug in the retina and choroid may be much slower than in the vitreous. Accumulation of the drug in the posterior eye segment has to be taken into consideration in intraocular drug delivery systems. 

In this study, the intrachoroidal implantation triggered no visible choroidal detachment, subretinal hemorrhage, and retinal detachment. Laser photocoagulation also showed no choroidal hemorrhage in most of experiments. In order to avoid such complications, a delicate and deliberate manner is important during surgical procedure. Our study showed some retinal damage corresponding to the implantation site. It is important to choose an implantation site that will result in minimal visual disturbance. Indocyanine green has been reported to cause undesirable retinal damage. ^47,48 In this study, to minimize unnecessary damage to the retina and choroid, 0.05% ICG was used and did not show severe toxicity in both of clinical and histologic observation. 

It has been reported that DEX doses for retinal toxicity range from 440 to 4000 mg in rabbit eyes. ⁴⁹ Our study used only 20 μg DEX in the implant and the laser-induced intrachoroidal DEX implant did not show severe toxicity as assessed through clinical and histologic examination (Figs. 3, 4). However, this does not guarantee the long-term effects. There is also a risk that locally implanted drug may have systemic effects through bloodstream. It has been reported that effective DEX concentrations for suppressing various inflammatory processes range from 0.15 to 4.0 μg/mL, ^26–30 though no study of effective DEX doses on the posterior eye segment has been conducted. Further investigation may be needed to establish the ideal concentration of DEX in the posterior eye segment. 

In the present study, the laser-induced DEX implant showed the protective effect in EIU, although DEX implant with no photocoagulation did not inhibit EIU sufficiently. Therefore, the laser-induced DEX implant may be effective to adjust the DEX level corresponding to the severity of EIU. The endotoxin-induced model used in this study is not an exact representation of clinical uveitis, but the inflammatory response to endotoxin closely resembles the acute phase of clinical uveitis. ²⁴ The reduced severity of inflammatory changes observed in histopathologic examination and of clinical manifestations in the inflamed eye (Figs. 5A, 6) was the result of significant inhibition of vascular and cellular inflammatory responses. The suppression of inflammatory responses was also suggested by significantly low levels of proteins (Fig. 5B). 

In conclusion, we have demonstrated a novel methodology for laser-induced intrachoroidal sustained drug delivery system in a rabbit model. Additionally, this therapy effectively prevented an experimental posterior segment eye disease. This new drug delivery system may be useful in adjusting the most effective medication to an individual patient. 

Acknowledgments

The authors thank Masayuki Taira and Takashi Nezu for excellent technical assistance. 

Disclosure: M. Murata, P; A. Sanbe, None; J.W. Lee, None; H. Nishigori, None