Ribozyme to Proliferating Cell Nuclear Antigen to Treat Proliferative Vitreoretinopathy

Abstract

purpose. A DNA-RNA chimeric ribozyme was developed that targets the mRNA of a cell cycle regulatory protein, proliferating cell nuclear antigen (PCNA). The hypothesis was that inhibition of PCNA, essential in DNA replication, would decrease the proliferation of cells that are involved in formation of granuloma after surgical procedures in the eye. The ability of intravitreous injection of this ribozyme to prevent or inhibit development of proliferative vitreoretinopathy (PVR) was tested in a dispase-induced rabbit PVR model.

methods. Rabbit genomic DNA encoding PCNA was cloned and sequenced. The cleavage of rabbit PCNA by the chimeric ribozyme was tested in vitro. Delivery of the ribozyme to rabbit retinal pigment epithelial (RPE) or fibroblast cells and its effects on proliferation of fibroblasts were examined. The stability of the ribozyme in vitreous fluid and serum was studied as well. In the dispase-induced rabbit model of PVR, the ability of the PCNA ribozyme to prevent or inhibit development of PVR and retinal detachment (RD) was tested. Experimental groups receiving intravitreous PCNA ribozyme, with or without a lipid vehicle, were compared with sham-treated control groups. Progression of PVR in rabbit eyes was followed by indirect ophthalmic examination and observations documented by fundoscopic photography, gross pathology, and histopathology.

results. The chimeric ribozyme targeted a specific sequence in the rabbit PCNA that was identical with that in the human. In vitro cleavage assays confirmed the ability of the ribozyme to cleave the mRNA of PCNA. The catalytic efficiency in vitro, calculated as k ₂/K _m ^app, was 0.26 μM⁻¹ min⁻¹. In vitro studies with fluoresceinated ribozyme indicated that lipid vehicles facilitated delivery of the ribozyme into cells causative of PVR (RPE and fibroblasts); however, the PCNA ribozyme decreased the proliferation of fibroblasts, with or without lipid vehicle. The ribozyme displayed good stability in vitreous fluid, whereas, it degraded quite rapidly in serum. In animal experiments, rabbits in sham-treated groups usually exhibited development of severe PVR characterized by focal traction or RD. Animals in the PCNA ribozyme–treated groups usually did not exhibit an RD. If they did have RD, it was small and localized, or focal tractions developed that did not progress to the degree that the sham-treated animal eyes did over the follow-up period. The in vivo use of a lipid delivery vehicle resulted in a precipitate; however, an effective naked ribozyme dose was identified that did not cause this side effect.

conclusions. In addition to validating the newly developed dispase PVR rabbit model, the results indicate that ribozyme targeted against the cell cycle agent PCNA is efficacious in the treatment or prevention of PVR in the rabbit eye. These experiments suggest that chimeric ribozyme targeted against PCNA may have a therapeutic or preventative role in humans.

Proliferative vitreoretinopathy (PVR) is the principal cause of failed retinal repair surgery, and occurs in 5% to 20% of all patients with retinal detachment (RD). ^{1 2} PVR is characterized by the migration and proliferation of cells that cause membranes to form on or beneath the retina or in the vitreous cavity. ³ The cells in the proliferative membranes include retinal pigment epithelial (RPE) cells, glial cells, fibroblasts, and inflammatory cells. ^{4 5 6 7 8 9 10 11} The aberrant proliferation leads to formation of neomembrane and contraction, which eventually causes tractional detachment of the retina. Recurrent detachment can result in permanent visual impairment. 

Current therapies for the treatment of PVR involve vitreoretinal surgical techniques in which scar tissue is mechanically peeled and the retina flattened with heavy liquids, gas, or silicone oil. Although the techniques have improved the reattachment rate to 90% or more in patients with PVR who experience RD, most patients have impaired vision because of macular involvement. ¹² A preventative treatment that inhibits onset of PVR in patients who have undergone retinal surgery would offer the best outcome for this subset of patients. 

There are several steps in the formation of neomembranes after injury. ^{3 13} They include cell division, growth, and migration, all of which are promoted by a number of growth factors. Cell division, in turn, is controlled by an array of proteins, including such cell division factors as cyclins, cyclin-dependent kinases, oncogenes, and tumor-suppressor genes. Targeting genes involved in the many checkpoints of cell division (cyclin and cyclin kinase-like proteins) is likely to be more effective at controlling cell proliferation than targeting genes that signal cell division (i.e., c-myc, c-myb), because there is redundancy at this level. ¹⁴ Although a suicide gene therapy strategy in which a herpes simplex virus thymidine kinase gene and antiviral therapy are used was tested against PVR, ^{15 16} agents that are cytostatic rather than cytotoxic for cells may be more effective in preventing proliferation without exacerbating inflammatory processes resulting in tissue injury. 

Ribozymes (Rzs) are small RNA molecules with endoribonuclease activity. They hybridize to complementary sequences of a particular target mRNA transcript through Watson-Crick base pairing and, under appropriate conditions, catalytically cleave a sequence-specific target. The chimeric Rz tested in our experiments is a chemically synthesized oligonucleotide consisting of natural DNA and RNA bases with specific 2′O-methyl modifications. ¹⁷ The presence of DNA with RNA imparts more stability to the Rz compared with those synthesized with RNA only. The binding arms of the molecule impart sequence specificity for the target mRNA. This type of oligonucleotide is not expected to have significant toxic side effects. ¹⁸  

For delivery in cell culture systems, lipid based formulations are often used to facilitate antisense oligonucleotide and Rz drug uptake. The lipid carrier must be optimized for the particular target cell type. However, uptake studies in cell culture are not predictive of formulation requirements in vivo. ¹⁹ In fact, free formulations are effective for delivering Rzs in vivo and in clinical administration. 

We used an Rz to target the cell cycle controlling gene PCNA to inhibit cell division for the prevention of PVR. PCNA is a critical cell cycle division factor in all cell types ^{20 21} ; therefore, delivery of therapeutic Rzs to the retina should transiently inhibit proliferation of a broad range of cells. Our experiments used the dispase rabbit PVR model to test this hypothesis. Although multiple models for PVR exist that involve introduction of exogenous cells such as fibroblasts or epithelial cells, ^{22 23} we used a model that induces PVR through the release of endogenous cells and factors. ²  

We describe our work verifying the sequence of rabbit PCNA and its enzymatic cleavage by the Rz targeted to it. Furthermore, we describe Rz uptake by RPE and fibroblasts, its stability in vitreous, and its ability to inhibit fibroblast proliferation in vitro. In addition, in vivo studies demonstrate that different formulations of Rz, with or without lipid vehicle, inhibit the development of PVR, resulting in RD in dispase-treated rabbits. 

Methods

Chimeric PCNA Rz

The PCNA Rz is the sodium salt of a 32mer oligonucleotide Rz composed of deoxy-, ribo-, and 2′O-methyl ribo-nucleotides linked by a phosphodiester backbone. The structural formula is: 5′dAdGdC-dCdCmU-mGmCrU-rGrAmU-rGmAmG-PPPP-mCrGrA-mArAmC-mCdAdG-dGdCdG-dC-3′ where dA, dC, and dG are natural deoxynucleotide bases; rA, rC, rG, and rU are natural ribonucleotide bases; mA, mC, mG, and mU are 2′O-methyl ribonucleotide bases; and PPPP is a four propanediol linker. The molecule self-pairs in a stem–loop region. The optimal binding arm length and loop configuration analysis was determined by us and designated PCN1. ²⁴ The optimal configuration was determined to be a 7 base stem 1 and 9 base stem 3, with a type A (thermal stable tetra loop) stem 2 configuration. The chimeric Rz recognizes a 15-base sequence in the human PCNA mRNA transcript (GenBank accession number NM_002592.1). It cleaves the message at the 3′ end of the C in the GUC target site, leaving the message untranslatable. Negative controls consisted of a formulation buffer of 5% dextrose in water (D5W), a scrambled Rz (sRz) that contained a sequence incapable of binding to the targeted region, even though it possessed an active catalytic loop, or heat inactivation of the active Rz (hiRz) for 18 hours at 110°C. 

Verification of the Sequence Identity of Rabbit PCNA

Rabbit DNA was cloned and sequenced to verify identity, using standard methods. ²⁴ In brief, PCR primers flanking the target site were used to amplify a fragment from rabbit genomic DNA extracted from VX-7 cells (ATCC, Dutch belted rabbit carcinoma). The primers of the inner primer pair, chosen from the most highly conserved regions between the known human and rat sequences, were as follows: forward 5′-TAGTGGCCACAACTCCGCCACCAT-3′ and reverse 5′-GGTCAGGTTGCGGTCGCAGCGGTA-3′. The resultant 223-base pair PCR fragment was inserted into the pCR2.1 TA cloning vector (InVitrogen, Carlsbad, CA) and sequenced. 

In Vitro Cleavage of Rabbit PCNA by Chimeric Rz

The sequence information obtained was used to synthesize the authentic rabbit substrate for testing in an in vitro cleavage assay. Synthetic RNA substrate was produced by Pharmacon Research, Inc. (Boulder, CO), and synthetic PCN1 Rz was synthesized by Proligo LLC (Boulder, CO). In vitro cleavage was tested in 2-hour time course reactions in 40 mM Tris (pH 7.5), 10 mM MgCl₂, and 2 mM spermidine, at 37°C. ²⁵ Reaction products were analyzed by polyacrylamide gel electrophoresis (PAGE) and quantified by phosphorescence image analysis (Molecular Dynamics, Sunnyvale, CA). The K _m (K _m ^app) and k ₂ were determined for the Rz by performing single-turnover kinetic experiments with Rz concentrations ranging from 62.5 to 1600 nM and a substrate concentration at 10 nM, with analysis as described earlier. The K _m ^app and k ₂ for the Rz was estimated for an [Rz]/Velocity versus [Rz] plot with R ² > 0.90. Catalytic efficiency was calculated as k ₂/K _m ^app. 

Isolation and Culture of Rabbit RPE Cell Explants and Fibroblasts

Rabbit RPE was isolated and cultured according to the method described by Brittis et al. ²⁶ Briefly, enucleated rabbit eyes were transferred to a biosafety hood, and all manipulations proceeded under sterile conditions. Two perpendicular cuts were made in the eye, allowing it to be laid flat in a tissue culture dish. The retina was gently teased from the back of the eye. The retinal tissue was washed once with Hanks’ buffered salt solution (HBSS), resuspended in 0.25% trypsin, and incubated at 37°C for 20 minutes. After incubation, the remaining cell aggregates were further mechanically dissociated by gentle pipetting. The cells were washed once with HBSS, plated in Dulbecco’s minimal essential medium (DMEM) containing 20% fetal bovine serum (FBS), and cultured at 37°C in a humidified incubator at 5% CO₂. 

New Zealand White rabbit RAB-9 fibroblasts were obtained from the American Type Culture Collection (Manassas, VA). They were cultured in DMEM containing 10% FBS, and trypsin (0.25%) diluted in HBSS was used to passage the cells as just described. 

In Vitro Uptake of PCNA-Targeted Rzs into Rabbit Fibroblasts and RPE Cell Explants

The Rz targeted to PCNA (PCNA Rz) was labeled with 5-(and 6-) carboxyfluorescein according to the manufacturer’s protocol ([5(6)]-FAM; Molecular Probes, Inc., Eugene, OR) by Trilink Biotechnologies Inc. (San Diego, CA). A synthetic lipid polymer (Lipofectin, 1 mg/mL; Life Technologies, Rockville, MD), DOTAP/cholesterol (4 mM; graciously supplied by Nancy Symthe-Templeton, National Institutes of Health, Bethesda, MD), and a commercial transfection reagent (GenePorter; Gene Therapy Systems, Inc., San Diego, CA) were lipid formulations used to facilitate entry of PCNA Rz into the cells. The Rz (500 μg/mL) was gently mixed and complexed separately with each type of lipid at 1:1 (vol/vol) for 15 minutes at room temperature. Rabbit fibroblast and RPE cell cultures at 70% to 80% confluency in eight-well chamber slides were washed and the culture medium replaced with 100 μL of synthetic medium (Optimem; Life Technologies). Naked Rz or Rz complexed to lipid was added to the medium at appropriate dilutions in a volume of 100 μL. This was added to the 100 μL already in each well to bring the total volume to 200 μL. The negative control consisted of medium alone. The slides were incubated at 37°C for 2 hours. Medium was aspirated from the wells, and the cells were washed with phosphate-buffered saline (PBS) before analysis by fluorescence microscopy. Triplicate wells were analyzed for each. Uptake was scored by two blinded observers who assigned the degree of positivity on a scale of + to +++ and negativity as −. 

Proliferation Assays to Test Inhibition of Rabbit Fibroblast Growth by PCNA Rz

Rabbit Rab-9 fibroblasts were seeded at 10⁴ cells/cm² into a 96-well flat-bottomed tissue culture plate and grown overnight in DMEM containing 10% fetal calf serum. The medium was carefully aspirated, and the monolayers were washed once with PBS solution. Serum-free culture medium (AIM-V; Life Technologies) containing (1) concentrated PCNA Rz (250 ng/well), (2) concentrated hiRz or sRz (250 ng/well), (3) no additives, (4) PCNA Rz (125 ng/well) diluted with DOTAP/cholesterol (1.5 μM), (5) hiRz or sRz (125 ng/well) diluted with DOTAP/cholesterol (1.5 μM), or (6) DOTAP/cholesterol (1.5 μM) was added to wells in triplicate, and the plates were incubated for 5 hours at 37°C. The wells were then carefully aspirated and 200 μL DMEM containing fetal calf serum, prewarmed to 37°C, was added to each well, and the plates were incubated an additional 48 hours at 37°C. At the end of the incubation period, cells were pulse labeled with ³H-thymidine (1 μCi/well, ICN Pharmaceuticals, Irvine, CA) for 2 hours before harvesting. The counts per minute (cpm) were determined on a scintillation counter (LS7500; Beckman Corp., Irvine, CA). The mean counts per minute of triplicate wells ± SE was calculated and probabilities determined by a two-tailed independent samples t-test (SPSS ver. 10.0; SPSS Sciences, Chicago, IL). 

Stability Testing of PCNA Rz in Vitreous Fluid and Serum

Fresh vitreous was obtained with a 16-gauge needle by aspiration of rabbit eyes immediately after death. Ten microliters of Rz (1 mg/mL) were incubated at 37°C in 100 μL of vitreous fluid at time points between 0 and 60 (experiment 1), 0 and 180 (experiment 3), or 0 and 240 (experiment 2) minutes. Time points observed included 0, 1, 5, 10, 20, 30, 60, 120, 180, and 240 minutes. At the end of the incubation period, 10% sodium dodecyl sulfate in PBS was added to the sample, and the entire mixture was immediately frozen with liquid nitrogen. Reaction products were analyzed by polyacrylamide gel electrophoresis and quantified by phosphorescence image analysis (Molecular Dynamics). In experiment 1, human serum (100 μL) was also added to the Rz at time points between 0 and 30 minutes. In experiment 2, vitreous was incubated with the Rz for 240 minutes, then serum (Life Technologies) was added to Rz-vitreous samples for an additional 5, 10, and 30 minutes as a positive control (because it was determined in experiment 1 that the Rz was very sensitive to degradation by endonucleases present in the serum). 

Dispase-Induced PVR in the Rabbit Eye

PVR was induced in the right eyes of Dutch belted rabbits (Myrtle’s Rabbitry, Thompson Station, TN) using the dispase model described by Frenzel et al. ² Female Dutch belted rabbits weighing 1.3 to 1.5 kg were anesthetized with an intramuscular injection of ketamine (25 mg/kg) and xylazine (5 mg/kg). Topical proparacaine hydrochloride (0.5%; Alcon Laboratories, Fort Worth, TX), tropicamide ophthalmic solution (1% Mydriacyl; Alcon), and phenylephrine hydrochloride ophthalmic solution (2.5%; Bausch & Lomb Pharmaceuticals, Tampa, FL) were instilled on the ocular surface of right eyes only. Direct visualization and indirect ophthalmoscopy were performed; no ocular abnormalities were seen. The proteolytic enzyme dispase (0.07 U/0.1 mL; Roche Pharmaceuticals, Nutley, NJ) was diluted in PBS and injected into the vitreous directly over the posterior pole of the retina with a 30-gauge needle directed through the pars plana 2.5 mm posterior to the limbus supratemporally. The injection was performed under direct visualization with an indirect ophthalmoscope. To lower intraocular pressure, a 30-gauge needle was used to tap the anterior chamber immediately after intravitreous injection. 

Treatment and Monitoring of PCNA Rz- and Sham-Treated Animals by Indirect Ophthalmoscopy

Generally, rabbits were examined by indirect ophthalmoscopy 1 day after dispase injection and weekly thereafter by one of two masked observers. All examinations were performed on conscious rabbits with topical anesthetic and dilation. Rabbits were graded by indirect ophthalmoscopy using a modified grading system from Fastenberg et al. ^{27 28} and Sakamoto et al. ¹⁵ All eyes were scored on a scale of 1 to 6 (see Table 1 ). The scale was based on inflammation, hemorrhage, distortion, intravitreous haze and membranes, and formation of retinal breaks, traction, and detachments. For established disease experiments, Rz was injected on days 20 and 21 when minimal signs of PVR, including fibroglial proliferation, were present in most eyes. In the Rz prophylaxis studies, Rz was injected on day 7 before visible signs of PVR had developed, but the eyes were observed to have vitreous and preretinal hemorrhage. Initially a DOTAP/cholesterol vehicle was used for delivery of the PCNA Rz. Three control groups were tested: (1) 5% dextrose in water (D5W), (2) sRz that contained the same oligonucleotide bases as the experimental PCNA Rz, but in a different order, which would give it similar chemical properties but no sequence specificity; and (3) a hiRz (110°C, 18 hours), which on testing had 5% or less of the catalytic cleavage activity of the active PCNA Rz. 

Monitoring of Rabbit Eyes by Fundoscopic and Histopathologic Analyses

Rabbits were observed for progression of disease by fundoscopy at intervals (usually 7–10 days) after injection of dispase. A generous time frame was allotted to maximize the probability of the disease’s progressing in the dispase-treated groups. Fundoscopic evaluation was performed by two independent observers trained for the animal experiments. After 42 to 72 days, when the fundus grading stabilized, the rabbits were killed with an intravenous overdose of pentobarbital. For all procedures, rabbits were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

For histopathologic evaluation, eyes were enucleated and fixed in 10% buffered formalin for 1 week. They were evaluated grossly after sectioning in the vertical meridian at the inferior oblique muscle. Selected gross specimens were photographed before they were embedded in paraffin and sectioned. Hematoxylin and eosin staining was performed after dewaxing. 

Results

Molecular Identification and Characterization of Rabbit PCNA and Rz Target

Cloning of rabbit PCNA was performed by nested primers for use in the polymerase chain reaction (PCR). The primers were chosen from the most highly conserved regions between the known human and rat sequences (as given in the Materials and Methods section). Clones generated from the PCR product were identical at the human Rz target site, previously designated PCN1. ²⁴ These findings indicate that the chimeric Rz was capable of targeting the rabbit PCNA mRNA without any further modification. 

In Vitro Cleavage of Rabbit mRNA to PCNA by Chimeric Rz

The activity of the PCNA Rz was tested in an in vitro cleavage assay. When Rz was placed into the reaction mixture, both human (positive control) and rabbit PCNA substrates were cleaved by the Rz (Fig. 1) . Further studies indicated the substrate was cleaved efficiently. The K _m (K _m ^app) and k ₂ were determined for the PCNA Rz by performing single-turnover kinetic experiments with Rz concentrations ranging from 62.5 to 1600 nM and substrate concentration of 10 nM. The K _m ^app and k ₂ were 512 nM and 0.13/min, respectively, as estimated for an [Rz]/V versus [Rz] plot with R ² > 0.90. The catalytic efficiency, calculated as k ₂/K _m ^app, was 0.26 μM⁻¹/min⁻¹. 

Delivery of PCNA-Targeted Rzs to Rabbit RPE Cell Explants and Fibroblasts with Lipid Vehicles

Delivery of a fluoresceinated PCNA-targeted Rz to rabbit fibroblasts and to cell explants of RPE was demonstrated when lipid vehicles were used. Figure 2 is representative of the delivery of fluoresceinated Rz, when complexed with DOTAP/cholesterol vehicle, to rabbit fibroblasts as visualized by fluorescence microscopy. Delivery appeared to be primarily cytoplasmic in both cell types, although nuclear localization of the PCNA Rz also has been observed in other cell types, such as in vascular smooth muscle cells. ²⁴ Higher fluorescence intensities (better deliveries) were observed when the Rz-DOTAP/cholesterol lipid formulation was used (positivity scale +++), compared with Rz complexed with synthetic lipid polymer (Lipofectin; Life Sciences) or another transfection reagent (GenePorter; Gene Therapy Systems) formulations (++ or +) in both rabbit cell types. At the limits of detection, very little to no delivery (positivity scale +/− to −) was seen in these cells when lipids were not used. 

Inhibition of Proliferation of Rabbit Fibroblasts by PCNA Rz

Marked reductions in the proliferation of rabbit Rab-9 fibroblasts were obtained when they were exposed to concentrated PCNA Rz alone or a PCNA Rz-DOTAP/cholesterol mixture, relative to the proliferation obtained when a sRz or a hiRz was used (Fig. 3) . The PCNA Rz caused significant reductions in the growth of rabbit fibroblasts, regardless of whether lipid vehicle was or was not present (P ≤ 0.0001). In addition, there was little difference in the proliferation of control cells in medium, with or without lipid, compared with the respective sRz- or hiRz-treated control cultures, with or without lipid. Data are the mean counts per minute ± SE of triplicate measurements. The probabilities were determined by a two-tailed independent-samples t-test. Confirmation of these findings (1.2- to 6-fold differences) was obtained in several more like experiments. Also, with a second colorimetric in vitro assay kit (CellTiter 96 AQ; Promega, Madison, WI), similar differences between the cell proliferation of Rz-treated and untreated fibroblasts (1.3- to 6.9-fold) were measured by assessing the relative metabolic activities as viable cell dehydrogenase (data not shown). 

Stability of PCNA Rz in Vitreous Fluid and in Serum

The stability of the chimeric Rz to PCNA was tested in serum and in freshly acquired vitreous fluid. Stability of the Rz in serum would indicate the fate of the Rz if it were to diffuse away from the intended site of delivery and into circulation, whereas assay of the stability in vitreous would indicate its fate when delivered at the needed therapeutic site. 

In experiment 1 (Fig. 4A) , the stability of the Rz in vitreous (0–60 minutes) was compared with that in serum (0–30 minutes). High levels of degradation products of the Rz were seen immediately in serum at all time points observed, indicating the apparent sensitivity of Rz to degradation by serum endonucleases. The half-life was less than 1 minute. No degradation products were visible when the stability of the Rz was assessed over 60 minutes in vitreous. 

In experiment 2, confirmation of the vitreous findings was again documented, but at time points of 1, 30, 60, 120, and 240 minutes (Fig. 4B) . Rz was stable in vitreous out to 240 minutes. In addition, at the end of the 240 minutes, a positive control serum was added to the Rz in vitreous fluid, and the sample was incubated an additional 5, 10, and 30 minutes. At all the time points tested after addition of serum, degradation of the Rz was seen on the gels. The remarkable stability of the Rz in vitreous indicates that the fluid through which it must travel to get to the PVR cell targets does not affect therapeutic efficacy. 

The properties of the Rz in vitreous demonstrate favorable pharmacokinetics for eye delivery and probable efficacy. In addition, the rapid degradation in serum provides a degree of safety to the treatment, because Rz would be expected to act only locally at the site of delivery and not outside that compartment. 

Pharmacokinetic studies were performed in vivo as part of the preclinical development of PCNA-targeted Rz (Habita C, unpublished data, 2001). In these experiments, 60 mg Rz was administered intravitreously, and the concentration, in intact or in degraded form, was evaluated at different time points (1, 4, 12, 72, and 96 hours) in the vitreous and the retina. At each time point, three animals were killed and the retina and vitreous analyzed by high-performance liquid chromatography to determine the amount of Rz present. These studies showed that the PCNA-targeted Rz was cleared in the vitreous but was also taken up and cleared in the retina. In vivo, the half-life in the vitreous was determined to be approximately 1 hour. No intact or degraded Rz was detected in the vitreous or the retina at 96 hours. 

Dispase-Induced PVR: Definition of Preventative and Established Disease Groups by a PVR Grading Scale Assignment

To validate the dispase model for PVR and confirm the findings of Frenzel et al., ² the right eyes of Dutch belted rabbits were injected with 0.07 U dispase in 0.1 mL to induce PVR. ² Rabbit eyes were examined 1 day after injection of dispase and then at regular, usually weekly, intervals, by indirect ophthalmoscopy. Eyes were given a PVR grading scale assignment (GSA) based on criteria including inflammation, hemorrhage, distortion, intravitreous haze and membranes, and formation of retinal breaks, traction, and detachments (see Table 1 ). ^{15 27}  

One day after dispase injection, observation by indirect ophthalmoscopy revealed inner retinal whitening and often dense preretinal hemorrhage overlying the medullary rays. All rabbits had light to moderate vitreous hemorrhage best visualized in the inferior fundus (GSA 1). At 1 week, resolution of inner retinal hemorrhage and some of the preretinal hemorrhage occurred. By 2 to 3 weeks, all retinal and preretinal hemorrhage had resolved with some residual inferior vitreous hemorrhage remaining (GSA 1, 2). Mild to moderate PVR was seen in most rabbits by 3 to 4 weeks; a small percentage exhibited focal areas of traction associated with moderate preretinal proliferation (GSA 2, 3). 

Treatment with PCNA Rz to Inhibit Progression of Established PVR Disease

Initial animal studies were performed with the DOTAP/cholesterol lipid delivery vehicle, because the in vitro uptake studies indicated this lipid facilitated the best Rz uptake by cells known to be responsible for PVR proliferation. 

In experiment 1, active PCNA Rz or D5W, and in experiment 2, active Rz or sRz was mixed with DOTAP/cholesterol. The final amount of Rz injected per eye was 32.5 μg, which was injected at day 21 in a 0.1-mL volume. Animals were examined at regular intervals, and eyes were given a PVR Grading Scale Assignment (GSA) according to the criteria described in the footnote to Table 1 . 

In experiment 1 (Table 1) , at day 21 after injection of dispase, all eyes had early signs of PVR characterized by fibroglial proliferation (GSA 1, 2). By day 35, three of four rabbits treated with D5W in lipid vehicle showed extensive RD or peripapillary RD (GSA 5), whereas four of four rabbits treated with Rz in lipid vehicle did not exhibit RD (GSA 1, 2, 3). In fact, when the Rz-treated group was observed at day 56, two rabbits exhibited only very mild progression of proliferation and two had no progression (GSA 2, 3). Figure 5 demonstrates the fundoscopic evaluation at day 35 of a normal eye (Fig. 5A) , a D5W-DOTAP/cholesterol–treated eye (Fig. 5B) , and a Rz-DOTAP/cholesterol–treated eye (Fig. 5C) . The optic nerve was clearly seen in the normal eye, whereas hyperproliferation (Figl 5B , arrow) was observed in the D5W-treated eye. The hyperproliferation was significant enough to lead to RD and verifies that PVR induced by dispase develops through endogenous cells. The Rz-treated eye, however, clearly showed the optic nerve site and a lesser degree of proliferation. In Figure 6 two photographs showing the gross eye disease at day 42 after injection of dispase in an eye (GSA 5) treated with D5W (Fig. 6A) or at day 56 after injection of dispase in an eye (GSA 3) treated with PCNA Rz (Fig. 6B) . The control-treated eye shows extensive formation of neomembrane (yellow material), whereas the Rz-treated eye is clear of neomembrane. 

In experiment 2 (Table 1) , treatment was also administered on day 21. Three rabbits received active Rz and three received sRz. Both were complexed to the lipid vehicle. The sRz sequence was used rather than lipid alone to compensate for the charge neutralization of the nucleic acid in lipid–nucleic acid complexes. The sRz is a scrambled version of the Rz with identical length and catalytic domain but with randomized arms to prevent binding to any specific target. Again, the results showed an inhibition, although not as marked, of progression of experimental PVR in the active Rz- versus the sRz-treated groups. Two of three rabbits in the sRz group had total RD (GSA 6), whereas one of three rabbits in the Rz-treated group had total RD. 

In these two experiments, all rabbits immediately showed significant precipitation or crystallization after injection of the treatment agents, including the group given DOTAP/cholesterol alone. Although it gradually resolved by 2 to 3 weeks after treatment, the potential for a precipitate to contribute to RD in the eye was deemed unacceptable. 

In experiment 3 (Table 1) , because of precipitation that was probably due to the lipid vehicle, a concentrated Rz formulation (10 mg/mL) was evaluated. Treatment was administered on day 20, and each eye received 1 mg PCNA Rz. One group received Rz alone, another received D5W, and another received no treatment, to assess the natural progression of PVR. In this experiment, none of the eyes exhibited vitreous precipitation of white material immediately after injection. Nine rabbits were included in the study for treatment. One of four rabbits in the concentrated Rz-treated group showed development of PVR to the extent that a focal RD was observed (GSA 4), whereas three others showed minimal to no proliferation (GSA 0, 2) at day 63. One animal showed improvement over time from a GSA of 2 to 0. In the D5W group, one of three rabbits exhibited focal RD (GSA 4), whereas the other two had moderate PVR (GSA 3). In the two rabbits that received no drug, one had moderate PVR (GSA 3) and the other, mild PVR (GSA 2). Overall, there was less proliferation in the Rz-treated eyes. In this experiment, a white coagulated substance was apparent at the next observation day, 8 days after injection of concentrated Rz (10 mg/mL). It gradually resolved and was not observed at death. The coagulation was considered an improvement over the precipitate noted earlier. 

Treatment to Prevent PVR from Developing in Dispase-Treated Eyes

Experiment 4 (Table 2) was started the same time as experiment 3 (Table 1) , but assessed the value of prophylactic treatment before development of PVR. Administration of active concentrated Rz was started 7 days after dispase induction of PVR. Six rabbits received 0.1 mL concentrated Rz (10 mg/mL) in D5W, and another six received D5W alone. By day 72, one of six rabbits treated with Rz developed a focal RD (GSA 4), three had traces of mild proliferation (GSA 2), and two had no visible proliferation (GSA 1). In the D5W group, one rabbit exhibited extensive RD (GSA 5), one had extensive PVR (GSA 3), and the other four had mild proliferation (GSA 2). 

In the latter two experiments (3 and 4) the severity of PVR decreased over time. This appeared to be consistent with loss of activity of the dispase used to induce PVR. For these experiments, dispase aliquots had been hydrated and frozen, whereas in the earlier experiments (1 and 2) aliquots were hydrated from lyophilized material. For this reason, we made the decision to induce PVR in the remaining animal experiments with freshly rehydrated dispase. 

In Vivo Observations of Precipitation and Coagulation with Lipid Vehicle and/or Rz: In Vitro Observations of Lipid and Ribozyme Precipitation or Coagulation in the Vitreous

The intravitreous precipitation observed in the first two experiments appeared to be due to the DOTAP/cholesterol vehicle. In the eyes of rabbits in experiments 3 and 4, no precipitate was observed, but rather a coagulative substance appeared and then resolved. However transient, the coagulative substance could interfere with vision and was also undesirable. For that reason, we conducted two small pilot experiments, one in vivo and one in vitro, to determine the maximum concentration of free Rz to place in the vitreous. 

In the in vivo experiment, six eyes were injected (0.1 mL) with different concentrations of Rz: 5, 6, 7, 8, 9, and 10 mg/mL. After 24 hours, eyes with between 6 and 10 mg/mL showed white coagulated substance, the degree of which increased with higher concentrations. The substance was not visible in the eye receiving 5 mg/mL. Four days later, none of the eyes had visible, persistent coagulation. 

From vitreous collected from the left eyes at death, Rz at 5, 6, 7, 8, 9, and 10 mg/mL was added (0.1 mL) to 1 mL of vitreous at room temperature in vitro. In agreement with the in vivo findings, a visually evident coagulated substance appeared at between 6 and 7 mg/mL and was present at all higher concentrations of Rz. A 5-mg/mL concentration was deemed the safest effective concentration to use in subsequent animal experiments in which 0.1 mL of injectate was used. 

Rabbit Experiments with Concentrated Active or Heat-Inactivated PCNA Rz at 5 mg/mL: Preventative and Established PVR Studies

Experiments 5 and 6 were started simultaneously with the same lot of dispase. The new dispase was more potent and caused greater proliferation than that seen in the eyes in previous studies (Table 3) . Experiment 5 tested prophylactic Rz treatment, which was administered on day 7 after injection of dispase. Experiment 6 tested treatment of established disease, and Rz treatment was administered on day 21. The control groups in each consisted of a hiRz that was determined later, in in vitro cleavage assays, to have 5% or less of the activity of the concentrated Rz. 

Experiment 5 was a preventative experiment in which 11 eyes were studied. Observation on day 56 indicated a differential between the Rz-treated group and the control group treated with hiRz (Table 3) . In the Rz-treated animals, only one of six showed development of total RD (GSA 6), whereas three of five of the hiRz-treated animals reached total RD (GSA 6). 

Experiment 6 was the established PVR disease study, in which 8 eyes were included for observation. By day 21, almost all rabbits had extensive PVR with at least focal traction; one eye with partial RD (rabbit 1) was included in the study. Several eyes (rabbits 1 and 3) had a significant, dispase-induced cataract that impaired the fundus view; these eyes were graded by gross pathology and by histopathology at the end of the experiment. The observation at day 42 indicated a more definitive difference between the Rz and the hiRz groups (Table 3) . In the Rz-treated group, there was no progression of proliferation, although cataract developed in all the eyes. Three of four eyes treated with Rz at day 21 did not further demonstrate progressive disease by day 42. One eye had partial RD, which was already present before Rz treatment (GSA 4 to 5). It is clearly seen in the wholemount by histopathology at day 42 (Fig. 7A) . The posterior lens capsule is disrupted and a cataract formed, probably as a result of dispase enzymatic digestion. Fibrous material emanates from the posterior aspect of the swollen lens and extends to the peripapillary retina (arrow). This led to a tractional peripapillary RD. The asterisks (Fig. 7A) depict an area of subretinal fibroglial proliferation that is not routinely observed in the dispase PVR model, unless a retinal break is present, as is the case here. This area is shown by high-power photomicroscopy (Fig. 7B) to contain fibroblasts and a small degree of lymphocytic infiltration (arrows). 

To summarize the animal experiments, the mean GSA of PVR on particular assessment days for each of the six experiments is listed in Table 4 . Because each of the experiments varied in group size and treatment conditions, a nonparametric Sign test was used to determine an overall one-tailed probability that the mean for each of the Rz-treated groups would be less than the mean GSA of the respective control group in each of six replications. In all cases, the PCNA Rz–treated eyes had lower averaged PVR scores than the respective sham-treated groups (P ≤ 0.016). To ensure that the small sample sizes did not skew the data, the median GSAs were also computed (P ≤ 0.03). In addition, if a nonparametric analysis by Mann Whitney test is run on the established disease experiments 1, 2, 3, and 6, a significant difference (P ≤ 0.05) is obtained between the Rz-treated groups and the control sham-treated groups. However, the same test performed on the preventative disease experiments 4 and 5, did not show a significant difference between the Rz- and sham-treated groups. 

Discussion

Pharmacologic and other adjunctive treatments have been chosen to target specific or nonspecific processes in development of PVR. ^{29 30 31} Targets have included nonspecific cell division processes as well as fibrin production. Cytotoxic agents and therapies such as 5-FU, daunorubicin, and radiation have exhibited marginal, often temporary, benefits in clinical trials. Other therapies have been designed to decrease fibrin production or inflammation by treatment with steroids or tissue plasminogen activator. ^{9 32} Heparin has been shown to reduce reproliferation in PVR, presumably by reducing cellular adhesion. ³³  

Many treatments that are effective in vitro are not effective in vivo because of insufficient drug delivery and/or inappropriate temporal exposure to the drug. New drug delivery devices made of biodegradable polymers for anti-inflammatory or antiangiogenic effect are being developed to overcome these difficulties. The feasibility of sustained-release delivery is yet to be determined. These factors make gene therapy a potentially important treatment for PVR and other diseases. 

To be effective, Rzs should be relatively stable to RNases so that rapid degradation does not occur. ¹⁷ The RNA-DNA chimeric Rz moiety used in this study was designed to be relatively more resistant to RNase degradation while maintaining catalytically active function. ^{24 25} Second, although lipid vehicles significantly enhanced Rz uptake in vitro, the remarkable stability of the Rz in vitreous fluid means that very little of it has to be delivered intravitreously to obtain an antiproliferative effect. Thus, naked Rz could be used for the PVR disease indication. Third, the Rz drug should target the site of action so that toxicity is avoided by using only small amounts of the drug. The drug was locally administered to limit systemic exposure and the quantity of drug needed to achieve the desired effect. In addition, the rapid degradation observed in serum meant that if the Rz traveled away from the site of delivery and into systemic circulation, little toxic effect would be expected. Thus, local delivery of PCNA-targeted chimeric Rz for antiproliferative therapy in the eye meets the necessary criteria to be useful therapeutically. 

The catalytic activities of DNA-RNA chimeric Rzs are usually lower than the RNA-Rz versions. ¹⁷ After the optimal binding arm and stem 2 configuration were determined, several chimeric versions of the Rz were synthesized to increase both the catalytic activity and the nuclease resistance of the Rz. Changes were made to the Rz in both primary sequence and backbone modifications. The Rz sequence was changed to more closely resemble the sequence of the native hammerhead. The backbone of the Rz was modified to incorporate 2′-O methyl nucleotides at positions not previously modified, as well as replacing some of the DNA bases. These changes served to increase the catalytic function of the chimeric Rz in vitro, by fivefold over that of the prototype chimera. ^{24 25}  

The PCNA gene is highly conserved between species, with 90% identity between human and rat. The specific sequence targeted, designated PCN1, is completely conserved between human, pig, rat, and mouse. ²⁴ Not unexpectedly, we confirmed that this site was identical with the PCNA Rz–targeted region in the rabbit. Therefore, the information obtained in vitro and in the PVR rabbit model is likely to be translatable to humans for clinical trials. 

Using this model for PVR, we demonstrated that 84% (54/64) of the eyes could be used for study after dispase induction of the disease. This is slightly lower than the 94% rate reported by Frenzel et al. ² when 0.05 to 0.07 U of dispase were injected. Some loss of assessable eyes in the dispase model was attributable to severe inflammation after injection of dispase. This led to cataract, corneal clouding, and/or corneal neovascularization in a few eyes. All eyes had dispase injected directly over the posterior pole of the retina, but it is possible that some dispase migrated to the posterior lens surface and caused disruption of the lens, worsening inflammation, cataract, and corneal clouding. In one case, an RD may have been created (see Fig. 7 ). Another factor may be nicking of the lens during intravitreous injections or anterior chamber paracentesis that was performed after each intravitreous injection. Despite cataract and other complications associated with dispase-induced of PVR, we find the model useful and possibly superior to the model in which exogenous fibroblasts are used for induction of PVR. 

Our early experiments focused on DOTAP/cholesterol–mediated delivery of Rz, which appeared to be effective, but there was a concern that precipitate within the vitreous cavity might contribute to retinal traction in some eyes. Despite the dissolution of this precipitate within 2 to 3 weeks in all eyes, the concentrated free Rz was considered the better alternative. Problems then were resolved with the concentration of Rz and appearance of a transient intravitreous coagulated substance. Both established disease studies and preventative studies (see tables) were performed with various controls including D5W, sRz, and hiRz. In the established disease studies, the number of eyes reaching total RD was less than the control groups. In the preventative studies, there was a definite slowing of progression of PVR, and the number of eyes reaching total detachment was less than in the control groups. 

In conclusion, by molecular cloning techniques, we showed sequence identity at a site in the rabbit PCNA gene that had previously been shown to be efficiently cleaved by chemically modified Rzs in human, pig, and rat cells. Therefore, a rabbit model of PVR could be used to investigate the ability of the PCNA-targeted Rzs to inhibit the progression of disease. Several cell types commonly contributing to proliferative damage in the eye were shown to take up fluorescein-tagged Rzs in cell culture experiments. On Rz exposure, a decrease in fibroblast proliferation also was demonstrated. The chemically modified Rzs were shown to be extremely stable in vitreous extracted from the rabbit eye, while having a limited but useful half-life in cell lysate (that being most indicative of intracellular half-life). It should be noted that our studies do not address whether the in vivo mode of action of these Rzs is specific to PCNA mRNA. As such, future studies are needed to demonstrate with certainty that other cellular mRNAs are excluded from the Rz’s action. 

From our studies, we conclude that animal experiments in the dispase model of PVR showed inhibitory and prophylactic effects in Rz-treated eyes. The ability of this chimeric Rz to prevent or lessen the degree of damage due to aberrant proliferation of cells in the eye could be a significant advancement in the treatment of PVR. This type of therapy may be especially well suited for use in patients facing vision loss due to the failure of traditional surgical treatment. ³⁴ Other surgical procedures, such as trabeculectomy, in which scar formation leads to adverse outcomes, may also benefit from prophylactic treatment with PCNA-targeted Rzs. Overall, the data indicate that PCNA-targeted Rz is an effective treatment to prevent damage due to hyperproliferation in injured eyes. 

 Supported in part by Grants SBIR R43 and R44 EY12743 from the National Institutes of Health (JMR) and by Research to Prevent Blindness (MWW), The St. Giles Foundation of New York (MWW), and the R. Herbert and Alma S. Manweiler Memorial Research Fund at the University of Colorado Foundation.

 Submitted for publication December 6, 2001; revised June 6, 2002; accepted June 10, 2002.

 Commercial relationships policy: F, C (NM, CAK); C (MWW); E (EA, RT, JRB, JMR); N (PB, DBP, SBR).

 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.

 

Table 1.

 

Development of PVR in Rabbit Eyes after Treatment with Chimeric PCNA Rz, sRz or D5W

Rabbit	Treatment, †	PVR Grading Scale Assignment*
				Days after Dispase Injection
				1	7	14	21	28	35	42	49	56	63
Experiment 1, ‡
1	Rz	1	1	1	2	2	3	3	3	3
2	Rz	1	1	1	1	1	1	2	2	2
3	Rz	1	1*	1	2	2	3	3	3	3
4	Rz	1	1	2	2	2	2	2	2	2
5	D5W	1	1	1	2	5	5	5
6	D5W	1	1	2	2	4	5	5
7	D5W	1	1	1	2	2	3	3
8	D5W	1	1	1	1	2	5	5
Experiment 2, ‡
1	Rz	1	1	1	2	2	5	6
2	Rz	1	1	1	2	2	4	4
3	Rz	1	1	2	3	3	4	4
4	sRz	1	1	2	3	4	4	4
5	sRz	1	1	2	2	5	6	6
6	sRz	1	1	1	2	3	6	6

		Days after Dispase Injection
				1	9	16	20	28	35	41	48	56	63
Experiment 3, §
1	cRz	1	2	2	2	2	2	2	1	2	2
2	cRz	1	2	1	2	2	2	2	2	2	2
3	cRz	1	2	1	1	2	1	1	0	0	0
4	cRz	1	2	1	2	2	4	4	4	4	4
5	D5W	1	1	2	2	2	1	2	4	4	4
6	D5W	1	1	1	1	2	2	2	3	3	3
7	D5W	1	1	1	1	2	1	3	4	3	3
8	none	1	2	1	1	2	1	1	3	3	3
9	none	1	1	1	1	2	1	1	2	2	2

 Rz, PCNA ribozyme; cRz, concentrated PCNA ribozyme; sRz, scrambled PCNA ribozyme; D5W, 5% dextrose in water; none, no treatment agent.

*  PVR grading scale: 0, normal eye; 1, vitreous hemorrhage/preretinal gliotic changes; 2, focal traction with little to mild preretinal proliferation; 3, moderate to extensive PVR with focal traction; 4, severe PVR with focal RD of medullary ray; 5, extensive RD; peripapiliary retinal detachment; 6, total RD, with or without retinal tears.

†  Adapted, from Fastenberg DM, Diddie KR, Dorey K, Ryan SJ. The role of cellular proliferation in an experimental model of massive periretinal proliferation. Am J Ophthalmol. 1982;93:565–572; Fastenberg DM, Diddie KR, Sorgente N, Ryan SJ. A comparison of different cellular inocula in an experimental model of massive periretinal proliferation. Am J Ophthalmol. 1982;93:559–564; and Sakamoto T, Kimura H, Scuric Z, et al. Inhibition of experimental proliferative vitreoretinopathy by retroviral vector-mediated transfer of suicide gene. Can proliferative vitreoretinopathy be a target of gene therapy [see comments]? Ophthalmology. 1995;102:1417–1424.

‡  All experimental and control treatments, administered on day 21, included a lipid DOTAP/cholesterol vehicle. The final amount of PCNA ribozyme in the intravitreal injectate was 32.5 μg in 100 μL. In experiment 1, the D5W-treated animals were killed on day 42 when a clear difference between the experimental and the control groups was evident. Follow-up of the experimental group continued until day 56. In experiment 2, all animals were killed on day 42.

§  Treatment with concentrated ribozyme was administered on day 20 without addition of lipid vehicle. Follow-up continued to day 63.

Figure 1.

 

In vitro cleavage of synthetic RNA substrate containing the PCNA target site by the PCNA-specific Rz. Phosphorescent image of the 5′-³²P labeled PCN1 substrate 20mer and its 8mer cleavage product, as visualized on a polyacrylamide gel. Effective cleavage is demonstrated for human (positive control, lane 3) and rabbit (experimental, lane 4) substrates, as visualized by the appearance of appropriate 8mer cleavage product. Negative controls were the control human PCN1 sequence (lane 1) and the rabbit PCN1 conserved sequence (lane 2) substrates incubated with formulation buffer without Rz. Lanes 3 and 4 demonstrate cleavage of human and rabbit substrates, respectively.

Figure 2.

 

Uptake of fluoresceinated PCNA Rz by rabbit fibroblasts when exposed to Rz in the presence of DOTAP/cholesterol vehicle. Localization appears to be primarily cytoplasmic.

Figure 3.

 

Inhibition of proliferation of rabbit Rab-9 fibroblasts caused by PCNA Rz. The reduction in [³H]-thymidine uptake at 48 hours caused by either the PCNA Rz mixed with DOTAP/cholesterol lipid vehicle (Rz+lipid) or concentrated PCNA Rz (Rz) alone are shown compared with cells treated with sRz or hiRz (having ≤5% of the activity of the experimental Rz) control with or without lipid vehicle. The proliferation of untreated cells in medium, with or without lipid vehicle, is also shown and did not statistically differ from either of the respective sRz or hiRz control results. With or without delivery vehicle, the PCNA Rz caused a statistically significant reduction in growth compared with all the controls (P ≤ 0.0001). Data represent the mean counts per minute ± SE of triplicate measurements.

Figure 4.

 

Stability of the PCNA Rz in serum and in vitreous. (A) In the first experiment, stability of the Rz in serum occurred at 0, 1, 5, 10, 20, and 30 minutes (left lanes 2–7). Degradation product was noted at all time points. In vitreous no degradation occurred at any of the time points between 0 and 60 minutes (right lanes 2–7). The Rz was placed in lane 1 of each and labeled C for control. (B) In the second experiment, stability of the Rz in vitreous was visualized at 1, 30, 60, 120, and 240 minutes (lanes 2–6). At 240 minutes, serum was added to the Rz–vitreous fluid samples and incubated an additional 5, 10, and 30 minutes (lanes 7, 8, and 9, respectively). Degradation products were noted only in the samples containing the serum. Control lane 1: Rz only; lane 10: vitreous only; lane 11: vitreous+serum.

Figure 5.

 

Fundoscopic photographs show PVR progression in the rabbit dispase model visualized at day 35 after injection of dispase. (A) Normal eye; (B) sham-treated eye given D5W and DOTAP/cholesterol lipid vehicle; (C) Rz+DOTAP/cholesterol-treated eye. Hyperproliferation (long arrow) was easily seen in the D5W-treated eye. The optic nerve (short arrow) was clearly visible in the normal and Rz-treated eyes.

Figure 6.

 

Gross pathology of (A) a postdispase injection eye treated with D5W or (B) a day 56 postdispase injection eye treated with PCNA Rz. The control-treated eye showed extensive formation of neomembrane (yellow material), whereas the Rz-treated eye was clear of neomembrane.

Table 2.

 

Prevention of Dispase-Induced PVR with Chimeric PCNA Ribozyme

Rabbit	Treatment	PVR GSA
				Days after Dispase Injection
				2	7	8	21	28	36	43	65	72
Experiment 4
1	cRz	1	1	1	1	1	1	2	2	2
2	cRz	1	1	1	1	0	1	1	1	1
3	cRz	1	1	1	1	0	0	0	1	1
4	cRz	1	1	1	2	2	2	2	4	4
5	cRz	1	1	1	1	1	2	1	1	2
6	cRz	1	1	1	1	0	0	1	2	2
7	D5W	1	1	1	1	2	2	2	3	3
8	D5W	1	1	1	1	2	2	2	2	2
9	D5W	1	1	1	1	2	1	1	2	2
10	D5W	1	1	1	2	3	4	4	5	5
11	D5W	1	1	1	1	2	1	2	1	2
12	D5W	1	1	1	1	2	2	2	2	2

 Treatment was administered prophylactically on day 7 before PVR was fully developed. Data are GSAs. The grading scale and abbreviations are as in Table 1 .

Table 3.

 

Preventing the Development of Dispase-Induced PVR in Rabbit Eyes

Rabbit	Treatment	PVR GSA
				Days after Dispase Injection
				2	7	16	37	42	56
Experiment 5
1	cRz	1	1	1	1	1	1
2	cRz	1	1	2	3	3	3
3	cRz	1	1	3	4	4	4–5
4	cRz	1	1	3	3	3–4	4
5	cRz	1	1	2	3–4	3–4	3–4
6	cRz	1	1	5	6	6	6
7	hiRz	1	1	2	6	6	6
8	hiRz	1	1	5	6	6	6
9	hiRz	1	1	2	2–3	3	3–4
10	hiRz	1	1	2	2–3	3	3–4
11	hiRz	1	1	P†	P	P	6

		Days after Dispase Injection
				2	7	16	21	37	42
Experiment 6
1	cRz	1	no	3	4–5	P*	4–5
2	cRz	1	no	3	2	2	P 2–3*
3	cRz	1	no	2	2	P*	2
4	cRz	1	no	2–3	3	3	P 2–3*
5	hiRz	1	no	3	3	6	6
6	hiRz	1	no	3	3	6	6
7	hiRz	1	no	2–3	4	4	4
8	hiRz	1	no	3–4	3–4	4	4

 Treatment was administered on day 7 (exp. 5) or on day 21 (exp. 6) with concentrated ribozyme, cRz (5 mg/mL) or heat-inactivated concentrated ribozyme, hiRz; no, not observed. Data are GSAs after treatment with concentrated PCNA-targeted Rz on day 7, or in established PVR disease after treatment on day 21.

*  No view, pathology (P) at end of experiment used for final PVR GSA score.

Figure 7.

 

Histopathology of an eye after treatment with concentrated PCNA Rz, at day 42 after injection of dispase. (A) A gross whole mount of the eye with GSA 4 to 5 showed partial RD, which existed before treatment with Rz. This eye did not exhibit any further progressive disease after treatment. Fibrous material (arrow) connected the posterior pole of the lens to the retina and likely was responsible for the detachment. Proliferation was also evident grossly as a pink flocculent material wedged between the retina and the sclera ( ). (B) High-power photomicroscopy of the latter area described in (A) demonstrates fibroglial proliferation with a degree of lymphocytic infiltration (arrows).

Table 4.

 

Summary of Animal Experiments with the Mean PVR GSA

Experiment	Treatment	Lipid Vehicle	n	Treatment Day/Assessment Day	Mean GSA
1	Rz (32.5 μg/eye)	+	4	21/42	2.5
	D5W	+	4	21/42	4.5
2	Rz (32.5 μg/eye)	+	3	21/42	4.66
	sRz	+	3	21/42	5.33
3	cRz (1 mg/eye)	−	4	20/63	2.0
	D5W	−	3	20/63	3.3
4	cRz (1 mg/eye)	−	6	7/72	2.0
	D5W	−	6	7/72	2.66
5	cRz (0.5 mg/eye)	−	6	7/56	3.67
	hiRz	−	5	7/56	5.0
6	cRz (0.5 mg/eye)	−	4	21/42	2.88
	hiRz	−	4	21/42	5.0

 Averaged mean scores of the GSA of Rz- and sham-treated rabbit eyes with dispase-induced PVR. Two independent assessments were made by trained observers. The probability that the mean GSA for each of the Rz-treated groups could be less than the mean GSA of the respective control groups in each of six separate replications was computed by a one-tailed Sign test. In six of six instances the mean was lower (P ≤ 0.016).

The authors thank Andy Weiss and Peter Gutenberg for technical support, Barry Broswick for obtaining the fundoscopic photographs of the rabbit eyes, Thomas Nicholas for statistical analysis, MaryJo Garascia who processed the eye tissues for histopathologic analyses, and Cellia Habita for critical reading of the manuscript.