Long-Term Protection of Genetically Ablated Rabbit Retinal Degeneration by Sustained Transscleral Unoprostone Delivery

Abstract

Purpose: To evaluate the long-term protective effects of transscleral unoprostone (UNO) against retinal degeneration in transgenic (Tg) rabbits (Pro347Leu rhodopsin mutation).

Methods: The UNO release devices (URDs) were implanted into the sclerae of Tg rabbits and ERG, optical coherence tomography (OCT), and ophthalmic examinations were conducted for 40 weeks. Unoprostone metabolites in retina, choroid/RPE, aqueous humor, and plasma from wild-type (Wt) rabbits were measured using liquid chromatography-tandem mass spectrometry. In situ hybridization and immunohistochemistry evaluated the retinal distribution of big potassium (BK) channels, and RT-PCR evaluated the expressions of BK channels and m-opsin at 1 week after URD treatment.

Results: The URD released UNO at a rate of 10.2 ±1.0 μg/d, and the release rate and amount of UNO decreased during 32 weeks. Higher ERG amplitudes were observed in the URD-treated Tg rabbits compared with the placebo-URD, or nontreated controls. At 24 weeks after implantation into the URD-treated Tg rabbits, OCT images showed preservation of retinal thickness, and histologic examinations (44 weeks) showed greater thickness of outer nuclear layers. Unoprostone was detected in the retina, choroid, and plasma of Wt rabbits. Retina/plasma ratio of UNO levels were 38.0 vs. 0.68 ng UNO*hour/mL in the URD-treated group versus control (topical UNO), respectively. Big potassium channels were observed in cone, cone ON-bipolar, and rod bipolar cells. Reverse-transcriptase PCR demonstrated BK channels and m-opsins increased in URD-treated eyes.

Conclusions: In Tg rabbits, URD use slowed the decline of retinal function for more than 32 weeks, and therefore provides a promising tool for long-term treatment of RP.

Retinitis pigmentosa (RP) is the leading cause of incurable inherited blindness caused by mutations in many different genes and resulting in rod photoreceptor cell death. During the course of rod photoreceptor cell death, marked increases of reactive oxygen species due to reduced oxygen utilization,^1–3 or decreases of rod neurotrophic factors,^4,5 are the main causes of gradual cone cell death. If the cones function well, patients with RP can read when the lighting is sufficient. Because clinically significant central vision loss is associated with cone functional deterioration,⁶ the prevention of cone loss during retinal degeneration has been targeted as a major goal of most therapeutic strategies for retinal degenerative diseases.⁵ Methods used to rescue the photoreceptors include genetic replacement, cell transplantation, and the use of exogenous neurotrophic factors.⁷ Because many different genes affect inherited retinal degeneration, and all or nearly all patients suffer rod cell loss, genetic replacement has been limited to a small number of patients.⁷ However, neurotrophic factors have been used to enhance cell survival following a wide variety of insults. These neurotrophic factors include brain-derived neurotrophic factor (BDNF),⁸ glial cell–derived neurotrophic factor (GDNF),⁴ and ciliary neurotrophic factor (CNTF).⁹ 

Isopropyl unoprostone (UNO), a prostaglandin metabolite analog, has been used clinically as an antiglaucoma agent, and is a 424.64-Da molecule with hydrophobic properties.¹⁰ Unoprostone also has been shown to provide photoreceptor protection against oxidative stress and light-induced retinal damage in rats¹¹ through big potassium (BK) channel regulation.¹² Several in vitro studies also have demonstrated the neuroprotective effects of UNO.^13,14 Topically administered UNO twice a day in patients with RP significantly improved the central 2° and 10° retinal sensitivity.¹⁵ The use of neurotrophic factors, as well as UNO, is applicable regardless of the gene mutation; therefore, the delivery method is the current focus for treatment optimization. 

Recent progress in the treatment of some retinal diseases, including AMD, makes it possible to administer some drugs by intravitreal injections, although repeated injections are required.¹⁶ Intravitreal injections of neurotrophic factors, including growth factors and cytokines, have been shown to rescue degenerating photoreceptor cells in animal models.¹⁷ However, limitations to their clinical usefulness include delivery to the appropriate site and the short half-life of the neurotrophic factors. Intravitreal sustained delivery of some neurotrophic factors also has been reported to rescue retinal cells.¹⁵ Although the surgical procedure was reported to be tolerable, the implants in the vitreous might induce adverse effects, such as cataract formation and increased IOP.¹⁸ Treatments inside the eyeball sometimes induce adverse side effects, such as retinal detachment and infection.¹⁹ We have previously reported a novel transscleral drug delivery device that consists of a drug-releasing semipermeable membrane and impermeable membranes acting as the drug reservoir. Because of the nonbiodegradable and one-way release properties of the device, a sustained release of the drug to the retina was achieved.²⁰ We previously reported the usefulness of this device using a laser-induced choroidal neovascularization model²¹ and a retinal degeneration model in rats.²² 

Here we demonstrated the protective effects of sustained transscleral UNO delivery by using the device in a retinal degeneration model of transgenic rabbits harboring the Pro347Leu rhodopsin gene mutation. 

Materials and Methods

Animals

All animals were maintained and cared for in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and after receiving approval from the Institutional Animal Care and Use Committee of the Tohoku University Environmental and Safety Committee. New Zealand White (NZW) rabbits harboring the Pro347Leu rhodopsin gene mutation (Tg rabbits) were used at 5 weeks after birth, and were purchased from Kitayama Labs Co., Ltd. (Ina, Nagano, Japan). Rabbits were inbred for 1 week at Tohoku University. As controls, we used the same age-matched wild-type NZW rabbits (Wt rabbits). Thirty Tg rabbits and 51 Wt rabbits were used (Table 1). 

Table 1

 

In Vivo Study Demographics

 

Device Preparation

Sustained UNO release was achieved with a polymeric device using photo-polymerized tri(ethylene glycol) dimethacrylate (TEGDM) and poly(ethylene glycol) dimethacrylate (PEGDM). The release rate and activities of the UNO were determined to be approximately 10 μg/d. The devices were fabricated as reported previously.²² Briefly, each device consisted of a reservoir that could be loaded with a sustained release formulation of UNO and then sealed with a controlled release cover. Poly(ethylene glycol) dimethacrylate and TEGDM containing 1% 2-hydroxy-2-methylpropiophenone as a photo-initiator were used as starting materials. The reservoir was prepared by pouring the TEGDM pre-polymer into a microfabricated polydimethylsiloxane (PDMS) mold followed by UV (365 nm) light photo-polymerization for 40 seconds at an intensity of 11.6 mW/cm² (LC8; Hamamatsu Photonics, Shizuoka, Japan). The internal size of the reservoir was 10 mm length × 3.6 mm wide × 0.7 mm deep (Fig. 1). Unoprostone was dissolved in a mixture of 40% PEGDM/60% TEGDM (abbreviated as P40) having a UNO concentration of 500 mg/mL, and the mixture (5.7 μL) was poured into a reservoir and photopolymerized for 40 seconds. After loading the drugs, 3 μL of the P40 mixture without UNO was applied to the reservoir, and a PDMS mold was placed over the mixture, followed by UV curing for 4 minutes to provide a reservoir cover. The placebo device was prepared using PBS combined with P40 as a negative control drug formulation. 

Figure 1

 

(a) Photograph and schematic image of the URD. (b) Profile of UNO release from the URD during incubation in 1% Tween 20 in water. Values are means ± SD; n = 6.

In Vitro Release Study

The device was incubated in 1.5 mL 1% Tween 20 in water at 37°C. The collected solution was filtered and the amount of UNO that diffused out of the device was measured using HPLC (Prominence, Shimadzu, Tokyo, Japan). A YMC-pack ODS-AQ column (YMC) was used as a reversed phase analytical column. The mobile phase was H₂O/acetonitrile/methanol (2:3:4, vol/vol), and was delivered isocratically at 1 mL per minute. The chromatograms were monitored at 210 nm. The 1% Tween 20 in water was replenished during the course of the release study to ensure that the concentration of UNO was below 20% of its saturation value at all times. The results were determined by using a standard curve. 

Implantations

The rabbits were anesthetized with ketamine hydrochloride (90 mg/kg) and xylazine hydrochloride (10 mg/kg). Their ocular surfaces were anesthetized with a topical instillation of 0.4% oxybuprocaine hydrochloride. A 4 × 4-mm paralimbal conjunctival incision was made at the upper temporal limbus. The device was inserted between the conjunctiva and sclera and the front head of the device was placed just beside the optic nerve head. The device positions were confirmed after enucleation at the end of the experiments. The peripheral region was sutured onto the sclera with 7-0 silk to tightly fix the devices onto the sclera. The conjunctival sac was closed with 9-0 silk and antibiotic ophthalmic ointment was inserted into the conjunctiva. The devices were placed onto the left eyes. The right eyes remained untreated. 

Measurement of M1 Metabolite in the Retina, Choroid/RPE, Aqueous Humor, and Plasma

To evaluate the intraocular UNO distribution after UNO release device (URD) treatment, the concentration of M1, a primary metabolite of UNO, was measured by using liquid chromatography-tandem mass spectrometry (LC/MS/MS). One, 4, 12, and 24 weeks after URD implantation, the eyes were enucleated and frozen at −80°C. Plasma was also collected and frozen. The retina, choroid/RPE tissues, and aqueous humor were separated, diluted 1:9 (vol/vol) in plasma, and homogenized on ice. After centrifugation at 10,000g for 2 minutes at 4°C, supernatants were mixed with 200 μL of five volumes of propylene glycol and dried under a nitrogen stream, followed by the addition of 200 μL methanol/water (20:80, vol/vol), and finally followed by centrifugal ultrafiltration at 3000g for 2 minutes at 4°C. The final filtrate was used for LC/MS/MS (Shimadzu 10A; Shimadzu). Chromatographic separation was achieved by using a Develosil ODS-UG-3 (2.0 I.D. × 50 mm, 3 μm; Nomura Chemical Co., Ltd., Aichi, Japan) analytical column and an Inertsil ODS-3 (3.0 I.D. × 10 mm, 3 μm; GL Science Co., Ltd., Tokyo, Japan) precolumn with a gradient elution of two different mobile phases (mobile phase, acetonitrile:water:acetic acid [20:80:0.1, vol/vol/vol]; and mobile phase B, acetonitrile:acetic acid [100:0.1, vol/vol]). The flow rate was 0.25 mL per minute and the running time was 16.0 minutes. An AB SCIEX (API4000; Applied Biosystems, Foster City, CA, USA) was used to perform the MS/MS. The peak area ratios of M1 to the internal standard were used to construct a linear calibration curve using weighted regression analyses. The M1 concentration in the homogenates and plasma samples was determined by interpolation from the calibration curve. Remaining amounts of UNO in the URD removed from the eye after implantation were determined. The URD was homogenized in 1% Tween 20 in water, then the amount of UNO was measured by the same method as mentioned above using HPLC. 

Pharmacokinetic Parameters

Pharmacokinetic (PK) parameters were derived for each subject using noncompartmental analyses of retina and plasma concentration and time data using Win-Nonlin, version 5.2.1 (Model 200, NCA setting; Pharsight, Sunnyvale, CA, USA). Actual sample collection times and the concentration scatter plots were used for PK analyses. The devices were implanted onto the sclera by using the same methods as described above. One drop of 0.12% UNO was applied to the left eye of Wt rabbits. The sample collection times were 24, 72, and 168 hours for URD application and 0.5, 1, and 4 hours for topical applications. The parameters were calculated as the area under the retina (plasma) concentration-time curve (AUC_all), maximum retina (plasma) concentration (C_max), and time to reach the C_max (T_max), by the linear trapezoidal method.²³ 

Histologic Examinations

Optical coherence tomography (OCT) images were recorded every 4 weeks after treatment to 40 weeks. The animals were anesthetized with a mixture of ketamine hydrochloride (90 mg/kg) and xylazine hydrochloride (10 mg/kg). The pupils were dilated with 1% tropicamide and 2.5% phenylephrine (Santen, Osaka, Japan). A contact lens (radius of curvature of the central optic zone: 8 mm; diameter: 12.8 mm; power = ± 0 diopters [D]; center thickness [CT] = 0.2 mm; polymethyl methacrylate [PMMA]; Yunicon Co., Tokyo, Japan) were applied to the eyes. The RS-3000 Advance OCT system was used (Nidek, Aichi, Japan). The image quality on the real-time display was optimized by adjusting the ocular lens. The retinal thicknesses from the inner limiting membrane (ILM) to the RPE and outer nuclear layer (ONL) at the area 5 mm downward from the optic nerve (the region central to the optic disc, including the visual streaks) were calculated by averaging all of the points from end to end in the OCT image. 

Forty-four weeks after implantation, the eyes were enucleated and kept immersed for 24 hours at 4°C in a fixative solution containing 2.5% formalin, 1.25% glutaraldehyde, 65 mM phosphate buffer, and 5% acetic acid. Paraffin-embedded sections (5 μm thick) cut through the optic disc were prepared in the standard manner and stained with hematoxylin and eosin. Light microscopy images were photographed, and the thickness of the ONL from the optic disc was measured at 500-μm intervals on the photographs in a masked fashion by two observers. 

Electroretinography Examinations

Electroretinography (Mayo, Aichi, Japan) amplitudes were recorded every 4 weeks after treatments to 40 weeks. Thirty minutes before the recording, the animals were anesthetized with a mixture of ketamine hydrochloride (90 mg/kg) and xylazine hydrochloride (10 mg/kg). The pupils were dilated with 1% tropicamide and 2.5% phenylephrine (Santen). Scotopic ERG amplitudes were recorded in the dark-adapted eyes by placing a golden-ring electrode (7.8 mm base curve; Mayo) in contact with the cornea. An identical reference electrode was placed in the mouth, and a ground electrode was attached to the ear. Stimuli were produced with a light-emitting diode stimulator (Mayo). The single white-flash stimuli of −3.523, −2.523, −1.523, −0.523, 0.477, and 1.477 log (cd*second/m²) were used. Photopic ERG amplitudes were recorded in the light-adapted eyes by the same methods as were used for scotopic ERGs. Single white-flash stimuli of −1.000, −0.050, 0.950, 1.477, and 2.000 log (cd*second/m²) were used. The amplitude of the a-wave was measured from the baseline to the maximum a-wave peak, and the b-wave was measured from the maximum a-wave peak to the maximum b-wave peak. 

In Situ Hybridization (ISH) and Immunohistochemistry

Paraffin-embedded blocks and sections of rabbit tissues for ISH were obtained from Genostaff Co., Ltd. (Tokyo, Japan). The rabbit tissues were dissected, fixed with tissue fixative (Genostaff Co.), embedded in paraffin by their proprietary procedures, and sectioned at 4 μm. For ISH, tissue sections were dewaxed with xylene, and rehydrated through an ethanol series, and then with PBS. The sections were fixed with 4% paraformaldehyde in PBS for 15 minutes and then washed with PBS. The sections were then treated with 12 μg/mL of Proteinase K in PBS for 30 minutes at 37°C, washed with PBS, refixed with 4% paraformaldehyde in PBS, washed with PBS, and placed in 0.2 N HCl for 10 minutes. After washing with PBS, the sections were acetylated by incubation in 0.1 M triethanolamine-HCl, pH 8.0, 0.25% acetic anhydride for 10 minutes. After washing with PBS, the sections were dehydrated through a series of ethanol. Hybridization was performed with probes at concentrations of 300 ng/mL in the Probe Diluent (Genostaff) at 60°C for 16 hours. After hybridization, the sections were washed in 5× HybriWash (Genostaff), equal to 5× saline-sodium citrate, at 60°C for 20 minutes, and then in 50% formamide, 2× HybriWash at 60°C for 20 minutes, followed by RNase treatment with 50 μg/mL RNaseA in 10 mM Tris-HCl, pH 8.0, 1 M NaCl, and 1 mM EDTA for 30 minutes at 37°C. Then the sections were washed twice with 2× HybriWash at 60°C for 20 minutes, twice with 0.2× HybriWash at 60°C for 20 minutes, and once with TBST (0.1% Tween 20 in Tris-buffered saline). After treatment with 1× G-Block (Genostaff) for 15 minutes at room temperature (RT), the sections were incubated with anti–digoxigenin-alkaline phosphatase conjugate (Roche Life Sciences, Indianapolis, IN, USA) diluted 1:2000 with 50× G-Block (Genostaff) in TBST for 1 hour at RT. The sections were washed twice with TBST and then incubated in 100 mM NaCl, 50 mM MgCl₂, 0.1% Tween 20, 100 mM Tris-HCl, pH 9.5. Color reactions were performed with nitroblue tetrazolium chloride/5-bromo-4-chloro-3′-indolyphosphate (NBT/BCIP) solution (Sigma-Aldrich Corp., St. Louis, MO, USA) overnight and then washed with PBS. The sections were counterstained with Kernechtrot stain solution (Mutoh Pure Chemicals, Tokyo, Japan), and mounted with CC/Mount (Diagnostics Biosystems, Pleasanton, CA, USA). The probes for BK channels were developed against the alpha subunit of the BK channel, and were targeted to three different positions in the sequence. 

For histochemical staining after ISH, sections were treated with 0.3% H₂O₂ in PBS for 30 minutes followed by Protein Block (Genostaff) using the avidin/biotin blocking kit (Vector Labs, Burlingame, CA, USA). The samples were incubated with 0.4 μg/mL biotinylated peanut agglutinin (PNA), anti-PKCα monoclonal antibody (Abcam, Cambridge, UK), anti-G₀ protein α rabbit polyclonal antibody (Novus Biologicals, Littleton, CO, USA), or anti-neurokinin B receptor rabbit polyclonal antibody (Novus Biologicals) at 4°C overnight. After washing with PBS, samples were reacted with peroxidase-conjugated streptavidin (Nichirei, Tokyo, Japan) for 5 minutes. Peroxidase activity was visualized by diaminobenzidine. The sections were mounted with CC/Mount (Diagnostics Biosystems). 

Reverse-Transcriptase PCR

To determine the effects of sustained UNO release on retinal gene expression, BK channel beta 2 and m-opsin expressions in the retina were examined by real-time RT-PCR. The eyes were enucleated at 1 week after treatment. The entire retina was homogenized and total RNA was extracted by using the RNeasy Plus Micro Kit (QIAGEN, Tokyo, Japan) and single-stranded cDNA was synthesized by SuperScript III First-Strand Synthesis SuperMix (Invitrogen, Tokyo, Japan) according to the manufacturer's instructions. Reverse-transcriptase PCR was performed using a LightCycler instrument (Roche, Meylan, France) and the LightCycler FastStart DNA Master SYBR Green I reagent kit (Roche). The PCR conditions were as follows: 95°C for 5 minutes; 95°C for 10 seconds at the denaturation temperature, annealing temperature (BK channel beta 2, 55°C; m-opsin, 59°C; β-actin, 59°C) was for 10 seconds; and 72°C for 10 seconds. The second step was repeated for 45 cycles (BK channel beta 2 for 55 cycles). The sequences of the PCR primer pairs were as follows: BK channel beta 2, 5′-TCTTGGATAAAAGGAAAACAGTCACA-3′ (forward) and 5′-GGCCAGTCCCAGGAGGAT-3′ (reverse); m-opsin, 5′-TCATCCCACTCAGCGTCATC-3′ (forward) and 5′-GCAAAGACCATCACCACCAC-3′ (reverse); β-actin, 5′-TGCGGGACATCAAGGAGAAG-3′ (forward) and 5′-TCGTTGCCGATGGTGATGA-3′ (reverse). All data were normalized to β-actin expression levels, thus providing the relative expression level by the comparative Ct method. 

Statistical Analyses

Experimental data were presented as means ± SDs. Statistical significance was calculated with Ekuseru-Toukei 2012 (Social Survey Research Information, Tokyo, Japan), using an unpaired t-test for normally distributed isolated pairs, and the ANOVA with Tukey's test for multiple comparisons. Differences were considered significant if P < 0.05 (*) or P < 0.01 (**). 

Results

In Vitro UNO Release

Unoprostone was pelleted with P40, loaded into the reservoir, and sealed with a P40-cover (Fig. 1a). The release rate estimated from the slope of the regression line was 10.2 ± 1.0 μg per day during the 24-week incubation period (Fig. 1b). The release profile showed a mild rapid release at the initial time, followed by a constant release with decreasing release rate. The release rates during 1 to 4 weeks, 4 to 12 weeks, 12 to 24 weeks, and 24 to 32 weeks are shown in Table 2. 

Table 2

 

In Vitro Release Rate of UNO

 

Effect of URD Treatment on IOP, ERG Results, and Histology in Tg Rabbits

Figure 2 shows the IOP during 40 weeks after URD treatments in Tg rabbits (6 weeks old). There were no significant differences in IOPs among the URD, placebo-URD, and nontreated groups. Figure 3 shows the time course of photopic ERG spectra and average amplitudes in Tg rabbits (6 weeks old). The spectrum shows that oscillatory potentials were clearly observed in the URD-treated group during 36 weeks after treatment, but not in the placebo-URD and nontreated groups. Although a gradual decrease of a- and b-waves was observed in the group, the URD treatment may have suppressed the decrease in the amplitudes of a- and b-waves at many points (Figs. 3d, 3e). There was no difference between placebo-URD treatment and nontreatment. In the scotopic ERG (Fig. 4), the URD treatment suppressed the decrease in the amplitudes of a- and b-waves at some points (Figs. 4d, 4e). The spectrum shows that oscillatory potentials were clearly observed in the URD-treatment group during 29 weeks after treatment, but not in the placebo-URD and nontreated groups. These results indicated that protection of cones, especially for the b-wave, was prominent and that the rescue effects were observed until 32 weeks after initiating the URD treatment. Similar protective effects were observed at different light stimulus intensities in photopic ERGs (Supplementary Fig. S1) and scotopic ERGs (Supplementary Fig. S2). Partial protective effects were observed at different light stimulus intensities in the photopic ERGs (Supplementary Fig. S3) and scotopic ERGs (Supplementary Fig. S4) in the older Tg rabbits (37 weeks old). 

Figure 2

 

Intraocular pressure in Tg rabbits treated with URD, placebo-URD, or nontreated for 40 weeks. Values are means ± SD; n = 4.

Figure 3

 

Photopic ERG data at a stimulus intensity of 1.477 log (cd*seconds/m²). Representative photopic ERG spectra of Tg rabbits treated with URD (a), placebo-URD (b), and nontreated (c). Average ERG amplitudes of a- (d) and b-waves (e) in the groups treated with URD (blue rhombus), placebo-URD (red square), and nontreated (green triangle). Values are mean ± SD; n = 4. *P < 0.05, **P < 0.01: URD versus Placebo-URD; ⁺P < 0.05, ⁺⁺P < 0.01: URD versus nontreated (1-way ANOVA with Tukey's test). w, weeks.

Figure 4

 

Scotopic ERGs at a stimulus intensity of 1.477 log (cd*second/m²). Representative scotopic ERG spectra of Tg rabbits treated with URD (a), placebo-URD (b), and nontreated (c). Average ERG amplitudes of a- (d) and b-waves (e) in the groups treated with URD (blue rhombus), placebo-URD (red square), and nontreated (green triangle). Values are mean ± SD; n = 4. *P < 0.05, **P < 0.01: URD versus placebo-URD; ⁺P < 0.05, ⁺⁺P < 0.01: URD versus nontreated (1-way ANOVA with Tukey's test).

Optical coherence tomography examinations showed that the thickness from the ILM to the RPE layer and the ONL were significantly preserved in Tg rabbits (6 weeks old) treated with URD at 24 weeks after treatment compared with the nontreated groups (Figs. 5a, 5b). Figure 6 shows the histologic examination at 44 weeks after treatment of Tg rabbits (6 weeks old). The thicknesses from superior to inferior through the optic nerve were evaluated and significant preservation was observed in the URD-treated group compared with the nontreated group at approximately 7 mm downward from the optic nerve, close to the visual streaks where most cone cells are found. 

Figure 5

 

Thickness of retina measured by OCT. (a) Representative images of Tg rabbits treated with URD (a-1), placebo-URD (a-2), and nontreated (a-3) 24 weeks after treatment. The thickness from the ILM to the RPE (1) and ONL (2) was measured. The average thickness of ILM/RPE (b) and ONL (c) before (pre) and 24 weeks after treatment. Values are the mean ± SD; n = 4. RGC, retina ganglion cell.

Figure 6

 

The average thickness of the ONL at 44 weeks after implantation. Representative retinal cross sections of Tg rabbits treated with the URD (a-1), placebo-URD (a-2), and nontreated (a-3). Scale bars: 50 μm. Values are the mean ± SD; n = 4. *P < 0.05: URD versus placebo-URD (1-way ANOVA with Tukey's test).

Intraocular UNO Distribution

The UNO distribution in the retina, choroid/RPE, aqueous humor, and plasma was determined by measuring M1 metabolites using LC/MS/MS. The amount of M1 in the retina was 2.86, 0.90, and 0.29 ng/g at 1, 4, and 12 weeks after treatment (Fig. 7a). At 24 weeks, the amount of M1 in the retina was below the lower limit of quantification (< 0.5 ng/g). The amount of M1 in the choroid/RPE was 61.7, 7.57, 8.61, and 1.44 ng/g at 1, 4, 12, and 24 weeks after treatment (Fig. 7b). The amount of M1 in the plasma was 0.657, 0.500, 0.475, and 0.044 ng/mL at 1, 4, 12, and 24 weeks after treatment (Fig. 7c). The M1 was undetectable in all samples of the aqueous humor. The remaining amount of UNO in the removed devices after the experiments showed that the UNO release might be constant in vivo (Fig. 7d). The estimated release rate in vivo from the slope of the regression line in Figure 7d was 10.6 μg per day. 

Figure 7

 

Intraocular UNO distribution during 24 weeks of URD treatment in Wt rabbits. Concentration of M1 in the retina (a), choroid (b), and plasma (c). (d) Remaining amounts of UNO in the URD removed from the eye after implantation. Values are the mean ± SD; n = 4.

Analysis of PK Parameters

Pharmacokinetic parameters, such as AUC, C_max, and T_max, were compared between topical eyedrops and URD treatment. A simple linear regression model fitted to the logarithm of the transformed PK parameters indicated that C_max in the plasma was 13.0 ng/mL and 8.39 ng/mL in eyedrops and URD treatment, respectively. The C_max in the retina was 16.8 ng/g (0.489 ng/g in the fellow eye) and 461 ng/g (0.328 in the fellow eye) in the eyedrops- and the URD-treatment groups, respectively. In the plasma and the retina, T_max was 0.5 hour for eyedrops- and 72.0 hours for URD-treatment groups, respectively. The AUC in the plasma was 19.1 and 952 ng/h/mL in the eyedrops- and URD-treatment groups, respectively. The AUC in the retina was 13.0 (0.707 in the fellow eye) and 36,200 (37.8 in the fellow eye) ng/h/g in the eyedrops- and URD-treatment groups, respectively. Using these values, the ratios of the retina AUC/plasma AUC were calculated as 0.681 (0.038 in the fellow eye) for topical application and 38.03 (0.040 in the fellow eye) for URD application (Table 3). 

Table 3

 

Area Under the Curve (AUC)

 

In Situ Hybridization and Immunohistochemistry

In situ hybridization showed that positive immunoreactivity of BK channel beta 2 was observed in the ganglion cell layer, inner nuclear layer, and photoreceptor layers in the retina (Fig. 8a). For double staining, anti-BK channel beta 2, anti-PKC-alpha, anti-G₀ alpha, anti-PNA, and anti-neurokinin were used for ISH of rod bipolar, cone ON-bipolar, cone cells, and cone-OFF bipolar cells (Figs. 8b–e). These data demonstrated that rod bipolar cells, cone ON-bipolar cells, and cone cells were localized with the BK channel beta 2 protein. 

Figure 8

 

(a) In situ hybridization for BK channel beta 2 of the eyeball in Wt rabbits treated with the URD for 1 week. Histochemical staining for (b) PKCα, (c) G₀ protein α, (d) PNA, and (e) neurokinin B receptor as a second staining after ISH for BK channel beta 2. Scale bars: 25 μm.

Reverse-Transcriptase PCR

Figure 9 shows the relative mRNA levels of BK channel beta 2 and m-opsin normalized to β-actin expression. The URD-treated group showed significantly higher levels of BK channel beta 2 expression than did the nontreated group in Tg rabbits. However, there was no difference between the two groups in the Wt rabbits. The Tg rabbits had higher BK channel expression than did the Wt rabbits. Similarly, the URD-treated group showed significantly higher levels of m-opsin expression than did the nontreated group in the Tg rabbits, but not in the Wt rabbits. The Tg rabbits showed lower m-opsin expression than did the Wt rabbits, which may be due to cone degeneration in the Tg rabbits. 

Figure 9

 

Reverse-transcriptase PCR for BK channel beta 2 (a) and m-opsin (b) in the retina in Wt and Tg rabbits treated with URD or nontreated. Values are the mean ± SD; n = 3. *P < 0.05 (Student's t-test).

Discussion

We showed a significantly slower decline of retinal function based on ERG, OCT, and histologic examinations in inherited retinal degeneration rabbit models treated with the URD (Figs. 3, 4). Comparisons of AUCs between URD and topical eyedrop applications indicated that the URD could deliver UNO locally to the retina with less relative systemic distribution (Table 3). This was also explained by the lack of detection of the UNO at the anterior chamber and by the low amounts of UNO in plasma for 24 weeks (Fig. 7). This local UNO distribution in the posterior segment might cause no significant differences in IOP between the URD-treated and control eyes (Fig. 2), although UNO is an ocular hypotensive agent for treatment of glaucoma. Unidirectional UNO release to the scleral side by URDs might contribute to the localized distribution.^20,22,24 The release rate in vivo (10.6 μg per day, from Fig. 7d) was almost the same as that observed in vitro (10.2 μg per day, from Fig. 1b), indicating the integrity of the release system. 

Preservation of cone functions has been one of the targets of the neuroprotective treatments for RP, because rod functions fail at the early stages of RP. Partially reprogramming the rod cells into cone cells by genetic inactivation of a cell fate switching gene was reported to preserve cone function.²⁵ Patients given vitamin A show significantly slower declines in cone amplitudes than do those in the control groups.⁷ Although rodents with inherited retinal degeneration have been used as model systems, rodents have small numbers of cone cells and show very small photopic ERG a-waves.²⁶ However, rabbits show large photopic a- and b-waves.²⁷ The relatively large eye balls also make it possible to examine the variable effects of treatments, such as intravitreal injections and drug delivery systems.²⁸ To examine the effects of these novel approaches, Tg rabbits imitating human RP are therefore useful.²⁹ 

In our studies, the amplitudes of photopic a- and b-waves in the URD-treated eyes were significantly higher than those of the placebo-URD–treated and nontreated eyes at a high light stimulus intensity of 1.477 log (cd*second/m²) (Fig. 3). The amplitudes of the postphotoreceptor components in Tg rabbits are reported to be significantly larger than those of the Wt rabbits at lower stimulus strengths of 0.2 and 0.7 log (cd*seconds/m²).²⁷ This may be why significant differences in the amplitudes of photopic a- and b-waves were observed at higher light stimulus intensities. The amplitudes of rod a- and b-waves in the URD-treated eyes were also higher than those of the placebo-URD–treated and nontreated eyes (Fig. 4). The results of ISH showed that almost no rod photoreceptor cells expressed BK channels, but rod bipolar cells that are near the neurons of rod photoreceptor cells showed BK channel expression (Fig. 8). It was reported that BK channels modulate pre- and postsynaptic signaling at the level of A17 amacrine cells and rod bipolar cells.³⁰ The supportive effects of these cells might rescue the function of rod photoreceptors. Electroretinographic examinations using BK^−/− mice at high scotopic and low mesopic conditions showed a significant reduction of b-wave amplitude.³¹ These reports agree with our results that BK channels were expressed in PKC-positive rod bipolar cells and that the amplitudes of ERG b-waves in some points were partially preserved in the URD-treated eyes at early stages of retinal degeneration. 

Notably, BK channels were expressed in cone and cone ON-bipolar cells (Fig. 8). Optical coherence tomography (Fig. 5) and histologic examinations (Fig. 6) showed the preservation of retinal thicknesses around the visual streaks where numerous cone cells exist.³² Reverse-transcriptase PCR results (Fig. 9) showed increased m-opsin expression in the retina with increased BK channel expression in Tg rabbits. These results suggested that URD treatment enhances the protection of cone cells through the activation of BK channels in cone or cone ON-bipolar cells. In addition, rod-derived neurotrophic factors could also enhance cone survival. Léveillard et al.³³ reported that factors secreted from rod cells were essential for cone viability. Because most genes associated with RP are not expressed by cone cells, the loss of cones is thought to be indirect. Morimoto et al.³⁴ reported that transcorneal electrical stimulation promoted survival of photoreceptors and improved retinal function in Tg rabbits. They suspected that increased secretion of some unknown neurotrophic factors in the retina might be related to retinal neuroprotection. Unfortunately, there have been no reports that identified the molecules and/or receptors. 

A limitation of this study is the lack of data on the mechanism of the protective effects of UNO on retinal degeneration. Pathogenesis of RP includes extremely complex processes with unknown mechanisms. To further support the URD-mediated retinal protection mechanism, protein/gene expression analyses related to apoptosis and survival in the retina as well as RPE and choroid are needed in the future. 

Additionally, the reason why the protective effect was prolonged for 32 weeks despite almost no detection of the M1 metabolite in the retina remains unknown. Although an in vitro release study was not performed after 32 weeks (Fig. 1b), the sustained release could have continued after 32 weeks until the initially loaded UNO (2.85 mg) was depleted (approximately 40 weeks). Because the M1 metabolite was detected in the choroid at 24 weeks, some retinal neuroprotection could have occurred. Tanito et al.³⁵ reported that delayed loss of cones was related to impaired choroidal circulation. The choroidal layer in patients with RP is thin when compared with those of controls.³⁶ Unoprostone was reported to increase the blood flow through BK channels,³⁷ therefore the increased choroidal circulation might support the photoreceptor survival hypothesis. An evaluation of choroidal circulation using laser speckle flowgraphy³⁸ should, however, be performed in the future. 

In conclusion, we showed the functional preservation of retinal function by URD treatment and a possible involvement of BK channels in this preservation. Most animal model studies have reported the preservation of the ONL thickness, but not a functional improvement.^9,39 Therefore, URD is a promising device for treatment of RP. 

Acknowledgments

The authors thank Kaori Sampei for help with the RT-PCR. 

Supported by Grant KAKENHI (15H03015) from the Ministry of Education, Culture, Sports, Science, Technology (MEXT), and Grants 15ek0109073h0001 and 16ek0109073h0002 from the Japan Agency for Medical Research and Development (AMED). 

Disclosure: N. Nagai, None; E. Koyanagi, None; Y. Izumida, None; J. Liu, None; A. Katsuyama, None; H. Kaji, None; M. Nishizawa, None; N. Osumi, None; M. Kondo, None; H. Terasaki, None; Y. Mashima, R-Tech Ueno (E); T. Nakazawa, None; T. Abe, None