Photoreceptor Neuroprotection in RCS Rats via Low-Dose Intravitreal Sustained-Delivery of Fluocinolone Acetonide

Abstract

purpose. To study the neuroprotective effects of intravitreal fluocinolone acetonide (FA) in Royal College of Surgeons (RCS) rats.

methods. Five-week-old RCS rats were divided into four groups: 0.5 μg/d FA–loaded intravitreal drug-delivery implant (IDDI); 0.2 μg/d FA–loaded IDDI; inactive IDDI; and nonsurgical control. Electroretinography (ERG) and intraocular pressure (IOP) measurements were performed before surgery and weekly thereafter. Thicknesses of the retinal outer (ONL) and inner (INL) nuclear layers were evaluated at 9 weeks of age. ED-1–labeled activated microglia were counted. Total microglial cell counts were made by using Iba-1 antibody labeling.

results. At 9 weeks, control groups demonstrated an 80% reduction in ERG amplitudes (P < 0.001 for both groups). FA-treated groups demonstrated no statistically significant attenuation of ERG amplitudes at the end of the study, compared with the initial ERGs. Intraocular pressure (IOP) remained normal in all groups. ONL thickness in FA 0.2 μg/d–treated eyes was 2.1 ± 0.5 times greater than in nonsurgical eyes (P < 0.001) and 3.4 ± 0.7 times greater than in inactive IDDI-treated eyes (P < 0.0001). In FA 0.5 μg/d–treated eyes, ONL thickness was 1.5 ± 0.1 times higher than in nonsurgical controls (P < 0.05) and 2.4 ± 0.4 times higher than in inactive IDDI-treated eyes (P < 0.01). INL thickness was not different among groups. FA-treated eyes demonstrated significantly fewer activated microglia (P < 0.001) and overall number of microglia in the photoreceptor and outer debris zone layers (P < 0.001), compared with control groups.

conclusions. Chronic intravitreal infusion of FA is neuroprotective in RCS rats, preserves ONL morphology and ERG amplitudes and reduces retinal neuroinflammation. These findings may have a therapeutic role in human photoreceptor cell degenerations.

Retinitis pigmentosa (RP) affects one in four thousand individuals and is the fourth leading cause of visual disability in the United States after diabetic retinopathy, age-related macular degeneration (ARMD), and glaucoma. ¹ The family of diseases that are phenotypically recognized as RP consists of more than 150 genotypically distinct entities. ² In all cases, vision loss in RP is associated with photoreceptor cell death. Thus, therapeutic strategies that target common mechanisms of photoreceptor cell death among the multitude of RP genotypes may find broad clinical application. 

Although the role of apoptosis in photoreceptor cell death associated with RP and ARMD has been well documented, more recent studies in animals and humans have identified the role that microglia-mediated retinal neuroinflammation plays in photoreceptor cell loss. ^{3 4 5 6 7 8} These studies have shown that apoptosis and necrosis are coupled, not mutually exclusive, mechanisms of photoreceptor cell death in RP and ARMD. ^{9 10}  

Microglia are the tissue-resident macrophages within the retina and central nervous system (CNS). These cells are constitutively suppressed by endogenous cortisol under normal conditions, but become activated in the form of phagocytes and cytotoxic cells in the presence of bacteria, lipopolysaccharides, reactive oxygen species (ROS), and damaged cell membranes. ^{11 12 13 14} Activated microglia then migrate and recruit other microglia to the site of damage; kill the malfunctioning cells by the release of tumor necrosis factor-α, ROS, and proteases; and phagocytose cell debris. ^{15 16 17} The toxic extracellular milieu that is produced also results in the destruction of nearby healthy cells in a process termed bystander lysis, ^{18 19} thus creating a cycle of amplified cell damage and remediation events that accelerate photoreceptor cell loss. This pattern of neural degeneration, including oxidative cell damage, apoptosis, and inflammation, with aggressive invasion of microglial cells, has been identified in human RP, late-onset retinal degeneration, and ARMD. ³ The pathogenic role of microglia is well established in CNS neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis, Huntington’s disease, and acute spinal cord trauma. ^{20 21 22 23 24 25 26 27 28}  

At present, there are no clinically effective treatments for RP. We hypothesize, however, that by virtue of their antiapoptotic properties and anti-inflammatory microglial suppression, corticosteroids may block multiple pathways of photoreceptor cell death in the inherited retinal degenerations. The synthetic corticosteroid, fluocinolone acetonide (FA) has been approved by the U.S. Food and Drug Administration for chronic intravitreal delivery as part of a sustained-release system in patients with severe uveitis. ^{29 30 31} The purpose of this study was to determine whether chronic low-dose, sustained-release FA preserves photoreceptors in the RCS rat retinal degeneration model. 

Methods

Animals

Thirty-two homozygous recessive rdy albino RCS rats of both sexes, aged 5 weeks, weighing 150 to 180 g, were used in the study. All animals were bred from stock provided by the late Werner Noell. At this age, approximately 50% of their ERG response amplitudes are lost. ^{32 33} The animals were maintained and treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The animals were subdivided into four equal groups: (1) FA-loaded intravitreal drug delivery implants (IDDIs) with release rates of 0.5 μg/d; (2) FA-loaded IDDIs with release rates of 0.2 μg/d; (3) inactive IDDIs; and (4) nonsurgical control subjects. Animals from each litter used in the experiment were distributed evenly between the experimental groups. Electroretinograms were recorded before surgery and weekly after surgery. The animals were also observed with weekly funduscopy and handheld tonometer IOP measurements (Tonopen-XL; Reichert Ophthalmic Instruments, Depew, NY). 

At the conclusion of the study, 4 weeks after implantation surgery and at the age of 9 weeks, the animals were euthanatized via CO₂ asphyxiation followed by bilateral pneumothorax. Retinal histology was performed on both eyes of 20 animals (five animals from each group). Both eyes of the remaining 12 animals (three animals from each group) were used for microglial cell counting. 

FA Implant

Intravitreal implants were biocompatible, nonbiodegradable, FA-releasing devices. The FA implants were 2-mm long cylinders, 300 μm in diameter (Fig. 1A) , prepared from a 10:1 matrix of FA in polyvinyl alcohol (PVA) in a tube of polyimide, the ends of which were further coated in PVA and silicone to create a semipermeable membrane. The release rate was adjustable by altering the polyvinyl alcohol–silicone ratio. Release was determined in vitro by immersion in phosphate buffer at 37°C. The supernatant was change regularly to maintain sink conditions and assayed by reverse-phase high-performance liquid chromatography (HPLC). Release was determined by plotting cumulative release versus time and was linear over the experimental duration. Various FA implants have been developed and tested, with release rates ranging from 5.0 to 0.2 μg/d. The implants used in this study were physically the smallest available (enabling them to be implanted into a rat’s eye) and also had the slowest release rate in vitro. Our preliminary data (not shown in this article) has shown that these doses, as opposed to higher doses, did not cause toxic events such as elevation of the intraocular pressure and keratopathy. 

Each injector was constructed from a 1-mL syringe and a 26-gauge needle (Fig. 1B) . A 250-μm metal plunger was passed through the full length of the needle (Fig. 1Ba 1Bb) . The implant was loaded inside the needle and inserted into the vitreous with the thin metal plunger. The FA implants and the injectors for their intravitreal implantation were provided by Controlled Delivery Systems (product number NB283.137; Watertown, MA). 

Surgery

All IDDIs were placed within the right eye only. Procedures were performed under stereoscopic visualization by operating microscope (OPMC; Carl Zeiss Meditec, Thornwood, NY). Care was taken to maintain constant illumination levels and operative times for all procedures. In addition, nonsurgical control eyes were exposed to the light from the operating microscope for the same duration as an intravitreal implant procedure. Surgery was conducted under sterile conditions using 2% isofluorane general anesthesia, delivered via mask. The area around the operated eye was shaved and cleaned with 10% povidone iodine and 70% ethanol solutions. The surgical table was topped with a heating pad, to maintain the core temperature. At the end of the procedure, each animal received an intramuscular injection of butorphanol 0.1 mg/kg for the prophylaxis of pain. The entire procedure took approximately 1 hour. 

Intravitreal Implant Procedure

IDDIs were implanted in the vitreous cavity transsclerally, through the pars plana after conjunctival peritomy was performed at the 11-o’clock position with a specially designed injector. Immediately after implantation, the transscleral perforation was closed with a 50-μm-thick cyanoacrylate-fixed polyimide patch. At the end of the procedure, the conjunctiva was positioned over the polyimide patch and fixed with cyanoacrylate adhesive. 

Electrophysiology

Before testing, the animals were dark-adapted overnight and anesthetized with an intraperitoneal injection of ketamine 67 mg/kg and xylazine 10 mg/kg. Pharmacologic mydriasis was induced bilaterally with 1% tropicamide and 2.5% phenylephrine. Topical 0.9% saline was periodically applied to the corneas to prevent dehydration. Bilateral ERGs were measured in response to full-field 25 cd · s/m² white-flash stimulation, delivered via a contact lens-coupled, light-emitting diode, as previously described. ^{34 35} Recordings were made using platinum wire-loop corneal electrodes and an AC amplifier at a gain of 5000 and a band-pass between 10 and 100 Hz (CP511; Grass Telefactor, West Warwick, RI). Platinum reference needle electrodes were placed in the ears, and the abdomen was coupled to a ground electrode with conductive gel. Fifteen 1-ms flashes were delivered to both eyes simultaneously with constant, 20-second interstimulus intervals. During the ERG recording, response amplitudes and waveforms were monitored to confirm that there were no significant differences between the ERG responses within one recording session. 

Histology

Enucleated eyes were immersed in Karnovsky’s fixative overnight at 4°C. The eyes were rinsed with 0.01 M phosphate-buffered solution and prepared for sectioning by a transverse cut along the horizontal meridian. After gross preparation, the eyes were dehydrated in serial dilutions of alcohol, cleared with a xylene substitute (Pro-Par; Anatech Ltd., Battle Creek, MI) and embedded in paraffin containing DMSO (Fisher Scientific, Pittsburgh, PA). Serial 6-μm-thick whole-eye sections were obtained. Paraffin sections were mounted on poly-l-lysine-coated glass slides and stained using Harris’ hematoxylin and eosin. 

Microglia Cell Labeling

Enucleated eyes were kept in buffered 10% formalin for 15 minutes and then bisected sagittally. One half of the retina was fixed for an additional 4 hours at 4°C, and processed as a wholemounted retina. The wholemounts were used to obtain total numbers of microglia, both resting and activated, within different retinal layers. The other retinal half was embedded in optimal cutting temperature compound (Sakura Finetek USA. Inc., Torrance, CA) and flash-frozen in liquid nitrogen for cryosectioning to obtain transverse retinal sections. Orientation was preserved for each section. Retinal transverse sections were used to obtain numbers of activated microglia. 

Microglia Staining in Wholemounted Retina.

To stain both resting and activated microglial cells within the retinal wholemounts, three primary antibodies were used: Iba-1 (ionized calcium binding adapter molecule 1) antibody (Wako Chemicals USA. Inc., Richmond, VA), a marker of microglial cells ³⁶ ; ED-1 (cytoplasmic protein in monocytes and most macrophages) antibody (Serotec Ltd. Oxford, UK), a marker of activated microglial cells ³⁷ ; and ED-2 (protein in monocytes and most macrophages) antibody (Serotec Ltd.), a marker of macrophages. ³⁷ Goat anti-rabbit TRITC–conjugated antibody (Sigma-Aldrich, St. Louis, MO) was used as a secondary antibody. After fixation, the retinas were carefully removed and placed in PBS containing 1% Triton X-100 for 1 hour at room temperature. They were then incubated overnight at 4°C in a cocktail of Iba-1 antibody at 1:200 dilution and ED-1 antibody at 1:100 dilution, in PBS containing 0.1% Triton X-100 and 5% normal goat serum. After several washes, they were incubated in the secondary antibody solution for 4 hours at room temperature, washed again, and mounted with the inner limiting membrane (ILM) side down on glass slides and coverslipped with antifade medium (Vectashield; Vector Laboratories, Burlingame, CA). 

Microglia Staining in Transverse Retinal Sections.

To stain activated microglial cells in transverse retinal sections, we used the same three primary antibodies (Iba-1, ED-1, and ED-2). Goat anti-mouse IgG-FITC–conjugated antibody was used as a secondary antibody. Sections (10 μm) were cut on a cryostat (Leica Instruments GmbH, Nussloch, Germany), recovered on adhesive reagent (Vectabond; Vector Laboratories, Burlingame, CA)–coated slides and fixed for 1 minute in acetone. The slides were allowed to air dry and were stored at 4°C. The sections were rehydrated in Tris-buffered saline (pH 7.4) and the nonspecific staining blocked with 5% normal goat serum. They were then processed with the primary and secondary antibodies as already described. 

Data Analysis

ERG Analysis.

For each eye, after artifact rejection, ERG waveforms were averaged over 15 traces at each recording session. The a-wave amplitude was defined as the amplitude of the negative slope preceding the leading positive slope. The b-wave amplitude was defined as the peak-to-peak amplitude of the leading positive slope (from the lowest point of the a-wave to the highest point of the b-wave). The implicit times of the a- and b-waves were measured, defined as the latency of the most negative point (a-wave peak) and most positive point (b-wave peak) from stimulus onset. This was measured from each of the averaged ERG waveforms. The ERG b- to a-wave amplitude ratio was calculated for each averaged waveform for each time point. The group mean and standard deviations were computed for right and left eyes for every ERG parameter at each time point. 

Retinal Histology Analysis.

To evaluate the degree of retinal degeneration, we measured the ONL and INL thicknesses, according to the method of LaVail et al. ^{38 39 40} The ONL and INL thickness measurements were averaged from three entire retinal sections for each experimental eye. A microscope (Carl Zeiss Meditec) with a 330-μm diameter field was used. An eyepiece with a micrometer scale was used to perform the measurements. In each retinal section, 20 sets of three ONL measurements and three INL measurements were taken. Within each microscopic field, ONL and INL thicknesses were measured in the center of the field and 100 μm away from the center on each side. The first set of measurements was taken one microscopic field away from the optic nerve head (∼330 μm). Subsequent measurements were made from adjacent retina segments toward the ora serrata on both sides of the optic nerve head. Thickness measurements were averaged for the ONL and INL within each group to provide a single value for each retina for statistical comparison. To demonstrate the distribution of the ONL thicknesses across the retina within groups, measurements were plotted in a diagram. 

Microglial Cell Count Analysis.

Positive-stained microglial cells were counted with a microscope (model B-MAX 50; Olympus America, Melville, NY) with the appropriate filter sets: fluorescein isothiocyanate (FITC) filter (excitation, 460–500 nm; emission, 510–562 nm), and tetramethyl rhodamine isothiocyanate (TRITC) filter (excitation, 510–560 nm; emission, 573–647 nm). 

Microglial cell layers within the wholemounted retina were localized using deconvolution microscopy performed on a photomicroscope (Axiophot Triple-Camera Photomicroscope with Apotome module; Carl Zeiss Meditec). 

In retinal wholemount preparations, Iba-1 antibody–labeled microglial cell counts were made in six standardized fields at three defined levels within the retina: a layer at the level of the ganglion cells, a layer at the level of the IPL, and a layer at the level of the retinal photoreceptors. 

In retinal transverse sections, ED-1 antibody–labeled activated microglial cell processes were counted in two superior and two inferior retinal fields at the fourth retinal level of microglial cell accumulation, identified in the retinas of RCS rats between the RPE and retinal photoreceptors and occupying the outer debris zone. 

Statistics.

One-way ANOVA with concomitant t-test, when valid, were performed for both the 4-week postoperative ERG data and the averaged ONL and INL thickness data among treated and untreated eyes in all groups. Comparisons between right and left eyes for ERG amplitudes, implicit times and b- to a-wave ratios were performed within each group for all ERG time points, using the t-test. Within-group comparisons of the ERG preoperative to postoperative values were performed with the paired t-test. The degree of ERG amplitude loss was calculated for each group. Microglial cell counts were also averaged to calculate the mean and standard error for each microglial cell layer, among all groups. ANOVA and t-test were applied to perform the statistical comparison of microglial cell counts. 

Results

Postoperative Observations

Weekly ophthalmoscopic examinations during the 4-week postoperative period revealed no cataract or other media opacities, inflammatory changes, or retinal detachment. Weekly IOP measurements showed no statistically significant differences between the groups. 

Electroretinography

Figure 2shows weekly representative ERG traces for every experimental group. Figures 3 4 5show the ERG a- and b-wave amplitudes and b- to a-wave amplitude ratios among the groups over the course of the study. 

In the two control groups, end-point ERG amplitudes in rats at 9 weeks of age demonstrated pronounced reduction, compared with the initial ERGs (Figs. 3C 3D 4C 4D) . The nonsurgical control eyes showed a 73.4% ± 8.5% a-wave amplitude reduction and an 81.8% ± 4.1% b-wave amplitude reduction at the end of the study when the animals were 9 weeks of age (P < 0.001). The inactive IDDI group showed a 57.1% ± 2.9% a-wave amplitude reduction and a 74.1% ± 9.1% b-wave amplitude reduction at the end of the study (P < 0.001). No statistically significant differences were found between the right and left eyes during the study in both control groups. 

The ERG amplitudes did not change significantly from the baseline values in the two FA-treated eyes (Figs. 3A 3B 4A 4B) . ERG amplitude loss throughout the study in the eyes treated with FA 0.2 μg/d was 19.3% ± 11.2% of the initial amplitude for the a-wave and 11.6% ± 11.7% for the b-wave, which, for both ERG components, was not significantly different from the initial ERG values. In addition, in the FA 0.2 μg/d–treated group, the end-point ERG a-wave amplitudes were significantly greater in the right (treated) eyes than in the left (untreated) eyes (P < 0.05; Fig. 3B ). The ERG b-wave amplitudes in the right eyes were notably preserved, compared with the progressive reduction observed in left eyes 1 week after implantation (P < 0.05). These differences remained at this level of statistical significance until the end of the study (Fig. 4B) . Eyes treated with FA 0.5 μg/d showed intermediate preservation of the ERG amplitudes: a-wave amplitude loss was 45.4% ± 5.9% and b-wave amplitude loss was 42.4% ± 22.0%, which appeared to be a significant reduction (P < 0.05). No statistically significant differences in ERG amplitudes were found between treated and untreated eyes in this group during the course of the study. 

At 9 weeks of age, after 4 weeks of FA delivery, end-point ERG amplitudes in the FA-treated eyes were significantly greater than those in the nonsurgical control eyes and inactive IDDI-treated eyes (P < 0.001; Figs. 3 4 ). Although mean a- and b-wave amplitudes in the eyes treated with FA 0.5 μg/d were quantitatively lower than those of eyes treated with FA 0.2 μg/d, there was no statistically significant difference between them (P = 0.16 for the a-wave and P= 0.113 for the b-wave). No statistically significant difference was found between the inactive IDDI-treated eyes and the nonsurgical control subjects in ERG amplitudes (P = 0.115 for the a-wave and P = 0.787 for the b-wave). 

To evaluate the degree of loss of photoreceptor activity, we calculated the ERG b- to a-wave ratios as the relation of a- and b-wave amplitudes (Fig. 5) . The b- to a-wave ratio showed a dramatic elevation in the two control groups over the course of the study, but showed little change in the FA-treated groups. FA 0.2 μg/d–treated eyes demonstrated better preservation of the b- to a-wave ratio, compared with the eyes treated with FA 0.5 μg/d. For both treated groups, b- to a-wave ratios at the end of the study were significantly lower than in the control groups (P < 0.0001). A significant dose-related b- to a-wave ratio preservation was also observed in the fellow untreated eyes of all FA-treated animals. In these eyes, the ratio was significantly lower, compared with the IDDI-treated and nonsurgical control groups (P < 0.01). Untreated eyes in the FA 0.5 μg/d–treated group had better preservation of the b- to a-wave ratio than in the FA 0.2 μg/d group. There was no statistically significant difference in the b- to a-wave ratios between right and left eyes within groups. 

Figure 6shows the ERG a- and b-wave implicit times among the groups over the course of the study. End-point ERG implicit times at 9 weeks of age demonstrated different degrees of delay in all groups, compared with the initial ERG a- and b-wave timing. The end-point ERG a- and b-wave implicit time delays showed no statistically significant difference between the two control groups. There were also no differences in the implicit times between the two FA-treated groups. However, the two FA-treated groups showed significantly smaller a- and b-wave delays than in both control groups (P < 0.001). In the FA 0.2 μg/d–treated group, there was a statistically significant a-wave implicit time delay in the left (untreated) eye compared with the right (treated) eye at 3 and 4 weeks after surgery (P < 0.05). No statistically significant difference was found in any other group between the right and left eyes in a- or b-wave implicit time during the study. 

Of interest, we also observed a dose-related intermediate preservation of the ERG amplitudes in the fellow nonsurgical eyes of both FA-treated groups. Mean ERG amplitudes in left (untreated) eyes in the FA 0.5 μg/d group were significantly greater than those in the nonsurgical control animals (P < 0.03), but were not different from right or left eyes of the inactive IDDI group or left (untreated) eyes of the FA 0.2 μg/d–treated group. Quantitatively, mean end-point ERG amplitudes in the left eyes of the FA 0.2 μg/d–treated group were greater than those of nonsurgical control eyes and inactive IDDI group right and left eyes. However, these differences were not statistically significant. 

Histology

Figure 7shows photomicrographs of representative retinal histology for each study group, and Figures 8 and 9summarize ONL and INL thicknesses among groups. The ONLs of the eyes from the two control groups showed areas of one to three cell rows as well as areas where the ONL was absent. Pyknotic nuclei and cellular debris were prevalent in the ONL of these groups. These observations were more pronounced in the inactive IDDI-treated eyes. The ONL thickness showed no statistically significant difference between these two control groups (P = 0.09). Although not statistically significant, the observable differences in ONL thickness between the nonsurgical and inactive IDDI animals (Fig. 8A)suggest that the implantation procedure of an inactive intravitreal implant was associated with reduced photoreceptor viability. Eyes implanted with FA 0.2 μg/d demonstrated preserved ONL morphology, containing five to seven rows of photoreceptor cells; the ONL was 2.1 ± 0.5 times thicker than in nonsurgical control eyes (P < 0.001) and 3.4 ± 0.7 times thicker than in the eyes that received inactive IDDIs (P < 0.0001). Eyes implanted with FA 0.5 μg/d, had an intermediate preservation of the ONL, showing three to five cell rows; the ONL was 1.5 ± 0.1 times thicker than in nonsurgical control eyes (P < 0.05) and 2.4 ± 0.4 times thicker than in the eyes that received inactive IDDIs (P < 0.01). The ONL thickness of the FA 0.2 μg/d–treated eyes appeared significantly greater than in the FA 0.5 μg/d–treated eyes (P < 0.05). Left (untreated) eyes of the inactive IDDI-treated group showed significantly lower ONL thickness, than in the left eyes of all other groups (P < 0.05). ONL thicknesses of the left eyes of the FA-treated groups were not significantly different from each other or from the nonsurgical group. INL thickness was not significantly different between groups (P > 0.4). 

Microglial Cells in Wholemounted Retinas

Figure 10shows immunofluorescence micrographs of Iba-1– and ED-1–stained microglial cells in retinal wholemounts, within four layers, for all experimental groups. Inactive microglial cells with fine, filamentous processes were found within the ganglion and bipolar cell layers in the wild-type Sprague-Dawley rats. In the RCS rats, activated microglial cells with thickened, retracted processes were commonly found within four retinal layers (ganglion, bipolar, photoreceptor, and debris-zone layers). The nonsurgical control and inactive IDDI groups had higher microglial cell counts than did the groups that received the FA-containing IDDIs. Microglial cells assumed an engorged tuberous or sickle-shaped morphology in FA-treated eyes. ED-2 staining did not reveal the presence of macrophages in the neural retina in any experimental group. 

Iba-1 labeled retinal wholemounts demonstrated pronounced and statistically significant differences between groups in the photoreceptor layer microglial cell counts (Fig. 11) . Microglial cell counts within the photoreceptor layer/debris zone in FA 0.2 μg/d–treated eyes were approximately 35% lower than in the nonsurgical control eyes (P = 0.01) and approximately 40% lower than in eyes that received inactive IDDIs (P = 0.003). In the FA 0.5 μg/d–treated eyes, photoreceptor layer/debris zone microglial cell counts were approximately 38% lower relative to those in the nonsurgical control eyes (P = 0.004) and approximately 42% lower relative to eyes that received inactive IDDIs (P < 0.001). No statistically significant difference was found between the two groups that received the FA-loaded IDDIs in the microglial cell counts in the photoreceptor cell or debris zone layers. We observed no statistically significant differences in the four experimental groups in microglial cell counts within the ganglion and bipolar cell layers. Microglial cell counts of the untreated (left) eyes of the FA-treated animals (data not shown on figure) were not significantly different from those in the control groups. 

Microglial Cells in Transverse Retinal Sections

Figure 12shows representative photomicrographs of ED-1 antibody staining of activated microglia within the debris zone in the four experimental groups. Transverse ED-1 antibody-labeled retinal sections were used to count the number of activated microglial cell processes found within the retinal photoreceptor and outer debris zone layers. Figure 13shows counts of activated microglial cell processes in transverse retinal sections among the four experimental groups. ED-1–positive microglial cell processes in the FA 0.2 μg/d–treated eyes showed a fivefold decrease compared with the nonsurgical control eyes (P < 0.001) and a ninefold decrease compared with eyes that received inactive IDDIs (P < 0.001). In the FA 0.5 μg/d–treated eyes, the number of ED-1–labeled microglial cell processes in the debris zone was four times lower than that in the nonsurgical control eyes (P < 0.001) and seven times lower than in the eyes that received inactive IDDIs (P < 0.001). There were no differences between the two groups that received the FA-loaded IDDIs in the number of ED-1–labeled microglial cell processes. In the untreated (left) eyes (not shown in Fig. 13 ) of the FA-treated animals, the number of ED-1–positive microglial processes was approximately two times greater than in the right (treated) eyes and significantly lower than in the eyes of both control groups (P < 0.001). ED-2 staining was not observed within the neural retina. 

Discussion

Photoreceptor cell loss in the RCS rat similar to that in human RP occurs in association with the development of an outer debris zone. ^{41 42 43 44 45 46 47} In these animals, a null mutation in the gene encoding MERTK results in an impairment of Gas6-induced outer segment phagocytosis by RPE cells. ⁴⁸ By postnatal day 18, this impairment leads to the accumulation of a readily identifiable contiguous sheet of lipid material and proteins from shed rod outer segments. ⁴⁹ Concomitant with the development of this antigenic outer debris zone, a pattern of microglial activation and migration from the inner and middle retina toward the photoreceptor layer/outer debris zone becomes evident. ^{5 50} Subsequently, photoreceptors rapidly die. Microglial activation and photoreceptor phagocytosis have been demonstrated in patients with RP, late-onset retinal degeneration, and ARMD. ³  

In our study, FA, a synthetic corticosteroid delivered via a sustained-release IDDI, resulted in significant photoreceptor preservation in the RCS rat model, as measured by full-field ERG and quantitative histology. Untreated animals demonstrated a 50% reduction in ERG b-wave amplitudes and photoreceptors at 5 weeks of age and by 9 weeks of age, an 80% to 90% decrease in ERG b-wave amplitudes and a 70% to 80% loss in ONL cell counts was observed. We found that full-field ERG a- and b-wave amplitudes were preserved in all animals receiving FA implants. This preservation was statistically significant when FA-treated animals were compared to nonsurgical controls and animals that received inactive IDDIs. Animals treated with 0.2 μg FA/d demonstrated no statistically significant reduction in ERG b-wave amplitudes over the 4-week study. FA-treated eyes in both groups demonstrated significantly smaller ERG implicit time delays, compared with both control groups, which indicates preservation of signal transduction in these eyes. Additional evidence of the preserved photoreceptor activity in the FA-treated eyes was the ERG b- to a-wave ratio which demonstrated little if any change in these eyes during the study. Our treatment was started during the peak period of photoreceptor cell loss in the RCS rat at a point where 50% of the photoreceptors had already been lost. Consequently, the neuroprotective effects we observed in our model demonstrate that FA dramatically decelerated a degenerative process that had already begun. 

The lower dose of FA was reproducibly more effective in preserving ERG a- and b-wave amplitudes and retinal histology than the higher dose. This may relate to the dose-dependent interaction between the α- and β-glucocorticoid receptors, GRα and GRβ, respectively. Although GRα receptors account for the physiological action of corticosteroids, at higher doses, GRβ receptors exert negative feedback, dampening these physiological effects. Synthetic corticosteroids have been shown to activate GRβ, dampening GRα transactivation in a dose-dependent manner. ^{51 52} Thus, at higher doses, increased dampening by GRβ may reduce some of the GRα-mediated pharmacodynamic effects of glucocorticoids. 

We also observed a dose-dependent crossover effect in efficacy. The fellow eyes of animals that received FA 0.5 μg/d demonstrated a relative preservation of ERG amplitudes, whereas the fellow eyes in the FA 0.2 μg/d–treated group did not. Perhaps, these crossover pharmacodynamic effects were observable in small mammals due to the small volume of distribution—particularly with high-potency molecules of low molecular weight such as steroids. 

FA significantly reduced the overall number of microglial cells in the photoreceptor cell layer of the RCS rats. In addition, the number of activated microglial cell processes found within the photoreceptor layer and outer debris zone were significantly reduced by FA. These treatment effects were greater in the FA 0.2 μg/d group than in the FA 0.5 μg/d group. In control Sprague-Dawley rat retinal wholemounts, microglia appeared to be quiescent and inactive and were present in the ganglion and bipolar cell layers only. In the RCS rats, our retinal wholemount preparations demonstrated that microglia were present in four distinct layers (at the levels of the ganglion cells, inner plexiform layer, retinal photoreceptors, and the debris zone) and that activated microglia were present at the debris zone and within the photoreceptor layer. These observations are consistent with previous reports that have described a phenomenon of microglial migration toward the subretinal space in eyes with retinal dystrophies. ^{5 6 7}  

Activated microglia appeared as amoeboid-shaped cells in the photoreceptor layer and debris zone but assumed a more engorged tuberous or sickle shape with significantly retracted processes in the FA-treated retinas. This microglial morphology was unique to eyes that received FA-releasing implants. Transverse retinal sections confirmed that there were far fewer activated microglial cells in the debris zone of the FA-treated eyes. We ruled out the presence of blood-borne macrophage infiltration of the retina by using the ED-2 antibody, specific to these cells. Macrophages were observed within retinal blood vessels via ED-2 antibody staining, but no staining was observed within the retinal neural tissue among the four experimental groups. 

Microglia are resident members of the dendritic immune system within the retina and CNS and are activated by bacterial cell wall lipopolysaccharides and gangliosides within damaged lipid membranes. ^{53 54 55} Damaged photoreceptor cell membranes provide the antigenic stimulus for microglial activation. ^{5 6 7 8} On activation, microglia assume a phagocytic phenotype, migrate toward the degenerating photoreceptors, and scavenge the debris. ^{3 5 6 50} In addition, activated microglia regulate the recruitment and activation of other microglia via chemotactic cytokines such as CCL-5 (RANTES), macrophage inflammatory protein (MIP)-1α and -1β, and macrophage chemoattractant protein (MCP)-1 and -3. ^{15 16 18 54} After assuming the activated phenotype, the microglia release cytotoxic free radicals such as NO and superoxide anion as well as proinflammatory cytokines such as TNF-α, IL-1, and IL-6, incurring further photoreceptor cell damage (bystander lysis). ^{15 16 17 18 56} Undampened, the release of these inflammatory and toxic agents induces a positive feedback cycle of microglia-mediated photoreceptor cell death that in turn exacerbates the bystander lysis of additional photoreceptors, thus coupling photoreceptor apoptosis and necrosis and accelerating disease progression. ^{49 57 58 59}  

The low-dose, sustained-release FA implant we studied in this work most likely inhibited the microglia-mediated neuroinflammatory responses and thus substantially reduced the bystander lysis of otherwise healthy photoreceptors. Glucocorticoids act via multiple pathways that may explain the functional and anatomic photoreceptor cell preservation that we observed. Corticosteroids antagonize apoptosis by inhibiting AP-1. ^{60 61} Dexamethasone-mediated inhibition of AP-1 has been shown to be neuroprotective in photoreceptors of the BALB/c light-induced photoreceptor degeneration model. ⁶¹  

In addition, natural and synthetic glucocorticoids suppress the activation of quiescent microglia into their phagocytic phenotype. Endogenous glucocorticoids play a crucial role in inhibiting microglial-associated neuroinflammation. ^{60 62 63} Corticosteroids suppress microglial activation by downregulating their production of the major histocompatibility complex (MHC) class II antigens TNF-α and IL-6. ^{11 12 13 60 62 64}  

Activated microglia also produce ROS such as NO and superoxide anions that are known to be toxic to photoreceptors. ^{56 57 58 65 66 67} Steroids inhibit microglial production of ROS by suppressing iNOS and reducing NO production. The antiapoptotic, neuroprotective effects of antioxidants in photoreceptors have been demonstrated in several in vitro studies. ^{58 59 68 69 70}  

Our findings demonstrate that FA is capable of inhibiting the toxic and unmitigated neuroinflammatory response within the RCS rat outer retina that is associated with photoreceptor apoptosis and necrosis. Although FA did not alter the underlying biochemical defect within our degenerative rat model, we believe that by suppressing retinal neuroinflammation, FA reduced the bystander lysis of healthy photoreceptors and greatly slowed the progression of vision loss. In light of our findings, we suggest that while retinal photoreceptor death in RP may originate from genetically impaired photoreceptor/RPE metabolism, it is greatly accelerated by an undampened process of neuroinflammation that leads to massive photoreceptor cell death. Treatment modalities that suppress neuroinflammation such as low-dose FA are likely to have a therapeutic role in RP and allied diseases. 

³  Present affiliation: Department of Ophthalmology, Mayo Clinic, Rochester, Minnesota.

 Supported by the Ligon Research Center of Vision, Research to Prevent Blindness Career Development Award (RI), an unrestricted grant from Research to Prevent Blindness, and the National Institutes of Health Loan Repayment Program (RI).

 Submitted for publication September 4, 2008; revised December 24, 2008, and April 6, 2009; accepted August 17, 2009.

 Disclosure: I.V. Glybina, None; A. Kennedy, None; P. Ashton, pSivida Ltd. (E, I, P); G.W. Abrams, None; R. Iezzi, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) The IDDI positioned next to a ruler with 1-mm divisions. (B) Injector, constructed of a 1-mm syringe, a 26-gauge needle, and a 250-μm metal plunger passing through the needle, which enabled the IDDI loaded inside the needle to be pushed into the vitreous. (Ba, Bb) Two positions of the plunger: tip of the injector’s needle with the plunger inside it and tip of the injector’s needle with the plunger pushed outside, respectively.

Figure 2.

 

Representative ERG traces in the four experimental groups throughout the study. Week designations correspond to animals’ age.

Figure 3.

 

Mean (±SE) ERG a-wave amplitudes in the four experimental groups of RCS rats: (A) FA 0.5 μg/d; (B) FA 0.2 μg/d; (C) IDDI; (D) nonsurgical controls. OD, treated; OS, untreated. *Statistically significant difference between treated and untreated eyes (P < 0.05) in end-point ERG a-wave amplitudes. Week designations correspond to animals’ ages.

Figure 4.

 

Mean ERG b-wave amplitudes (±SE) in the four experimental groups of RCS rats: (A) FA 0.5 μg/d; (B) FA 0.2 μg/d; (C) IDDI; (D) nonsurgical control. Eyes treated with FA 0.2 μg/d demonstrated no significant reduction in ERG b-wave amplitudes over the 4-week study, whereas all other groups did. OD, treated; OS, untreated. (B) *Time points at which a statistically significant difference between treated and untreated eyes (P < 0.05) was observed. Week designations correspond to animals’ ages.

Figure 5.

 

Mean (±SE) ERG b- to a-wave ratios in treated (OD) and untreated (OS) eyes of the RCS rats throughout the study in the four experimental groups: b- to a-wave ratios from the ERG in the (A) treated and (B) untreated (fellow) eyes. Week designations correspond to animals’ ages.

Figure 6.

 

Mean (±SE) ERG a- and b-wave implicit times in treated (OD) and untreated (OS) eyes of the RCS rats throughout the study in the four experimental groups: ERG a-wave implicit times in the (A) treated and (B) untreated (fellow) eyes; ERG b-wave implicit times in the (C) treated and (D) untreated (fellow) eyes. Week designations correspond to animals’ ages.

Figure 7.

 

Representative retinal photomicrographs from 9-week-old RCS rats in the four experimental groups: (A) FA 0.5 μg/d; (B) FA 0.2 μg/d; (C) inactive IDDI; (D) nonsurgical control. ( ) ONL. Magnification, ×400.

Figure 8.

 

Mean (±SE) thickness measurements of (A) the retinal ONL and (B) the INL in the four experimental groups of 9-week-old RCS rats. *P < 0.05; **P < 0.001, compared with the nonsurgical control group.

Figure 9.

 

Mean measurements of ONL thickness in transverse retinal sections from the optic nerve head to the ora serrata of the 9-week-old RCS rats in the four experimental groups.

Figure 10.

 

Photomicrographs of Iba-1–labeled retinal microglial cells in the right eyes of the 9-week-old wild-type Sprague-Dawley (SD) rats and the four experimental groups of 9-week-old RCS rats. Columns correspond to the experimental groups and rows correspond to the retinal layers. Only two layers of microglial cells were visible in normal SD eyes. Abundant microglial cell processes were observed in the photoreceptor layer and the debris zone of RCS rats. FA treatment was associated with a reduction in the number of microglial cells in the photoreceptor/debris zone. Treatment also resulted in engorged sickle-shaped microglia. Magnification, ×400.

Figure 11.

 

Mean (±SE) microglial cell counts (per ×40 microscopic field) from Iba-1–labeled retinal wholemounts in right eyes of the 9-week-old RCS rats in the four experimental groups. *P < 0.01, compared with the nonsurgical control group.

Figure 12.

 

Photomicrographs of ED-1–positive staining in transverse retinal sections of right eyes of the 9-week-old RCS rats in the four experimental groups. A rrows: activated microglial cells.

Figure 13.

 

Mean activated microglial cell counts (per 40× microscopic field) from ED-1–labeled transverse retinal sections in the 9-week-old RCS rats in the four experimental groups. **P < 0.0001, compared with the nonsurgical control group.