Controlled Delivery of 5-Chlorouracil Using Poly(Ortho Esters) in Filtering Surgery for Glaucoma

Abstract

purpose. To evaluate the antimitotic and toxic effects of 5-chlorouracil (5-CU) and 5-fluorouracil (5-FU) and study their potential to delay filtering bleb closure in the rabbit eye when released by poly(ortho esters) (POE).

methods. Rabbit Tenon fibroblasts and human conjunctival cells were incubated with various 5-CU and 5-FU concentrations. Antiproliferative effects and toxicity were evaluated at 24 and 72 hours by monotetrazolium, neutral red, and Hoechst tests and cell counting. Mechanisms of cell death were evaluated using TUNEL assay, annexin V binding, immunohistochemistry for anti–apoptosis-inducing factor (AIF) and LEI/L-DNase II. Trabeculectomy was performed in pigmented rabbits. Two hundred microliters of POE loaded with 1% wt/wt 5-FU or 5-CU was injected into the subconjunctival space after surgery. Intraocular pressure (IOP) and bleb persistence were monitored for 150 days.

results. In vitro, 5-FU showed a higher antiproliferative effect and a more toxic effect than 5-CU. 5-FU induced cell necrosis, whereas 5-CU induced mostly apoptosis. The apoptosis induced by 5-CU was driven through a non-caspase–dependent pathway involving AIF and LEI/L-DNase II. In vivo, at 34 days after surgery, the mean IOP in the POE/5-CU–treated group was 83% of the baseline level and only 40% in the POE/5-FU–treated group. At 100 days after surgery, IOP was still decreased in the POE/5-CU group when compared with the controls and still inferior to the preoperative value. The mean long-term IOP, with all time points considered, was significantly (P < 0.0001) decreased in the POE/5-CU–treated group (6.0 ± 2.4 mm Hg) when compared with both control groups, the trabeculectomy alone group (7.6 ± 2.9 mm Hg), and the POE alone group (7.5 ± 2.6 mm Hg). Histologic analysis showed evidence of functioning blebs in the POE-5-CU–treated eyes along with a preserved structure of the conjunctiva epithelium.

conclusions. The slow release of 5-CU from POE has a longstanding effect on the decrease of IOP after glaucoma-filtering surgery in the rabbit eye. Thus, the slow release of POE/5-CU may be beneficial for the prevention of bleb closure in patients who undergo complicated trabeculectomy.

The use of antimitotic agents in filtering surgery for glaucoma is accepted in practice, especially in patients undergoing reoperations. The effect of various antimitotic agents (5-fluorouracil [5-FU], ^{1 2} mitomycin, ^{3 4} 5-fluorouridine, ^{5 6} fluoroorotate, ⁷ sodium butyrate, ⁸ octreotide, ⁹ cytosine arabinoside, ¹⁰ methotrexate, suramin ¹¹ and doxorubicin ⁴ ) on the proliferative ability of Tenon fibroblasts has been extensively studied. Among these, 5-FU and mitomycin C are the most frequently used in clinical practice. Topical application of 5-FU and mitomycin at the time of surgery has resulted in a significant reduction of filtering surgery failures in patients at high risk. ^{12 13} However, their use is associated with secondary adverse events. Mitomycin C has been associated with an increased incidence of postoperative delayed hypotony, ^{14 15} ciliary body toxicity, late postoperative endophthalmitis, and endothelial toxicity. ¹⁶ Intraoperative 5-FU improves the outcome of filtering surgery by reducing fibroblastic activity in Tenon capsule. Its first efficacy was shown in randomized controlled trials of postoperative subconjunctival 5-FU injections. ¹⁷ 5-FU, a synthetic pyrimidine analogue, inhibits the thymidylate synthetase that converts ribonucleotides to desoxyribonucleotides, thus inhibiting DNA synthesis. Because it acts selectively on the growth phase of the cell cycle, only cells in the synthesis phase are affected. The remaining cells can continue to proliferate after exposure to 5-FU. ¹⁸ Therefore, repeated applications are often required. The adverse effects of 5-FU include corneal epithelial toxicity, ciliary body toxicity, leaking blebs, endophthalmitis, scleral thinning, discomfort, and transient cataract. ^{19 20 21} The 5-FU solution used for ophthalmic application is an intravenous preparation developed for oncology treatments. No antimitotic molecule or formulation has been specifically developed for ophthalmic use. The properties required for an antimitotic agent in the ophthalmic setting may differ from those required in oncology. Fibroblasts proliferation during wound healing processes in the eye may be less aggressive than the proliferation of transformed cancer cells. Moreover, toxicity must be limited to the proliferating cells with minimal collateral damage to the surrounding tissues. With these objectives in mind, we hypothesized that an agent favoring apoptosis over necrosis may have beneficial effects on the control of postoperative wound healing after filtering surgery. Using a slow-release controlled-delivery system, such an agent could maintain a patent-filtering bleb over an extended period and could minimize potential side effects. 

During pathologic processes, cell death is essentially carried out through passive pathways as cell necrosis or through active and programmed cell death. ²²  

Caspase-dependent apoptosis is the most widely studied and characterized pathway of programmed cell death, ²³ but other non-caspase–dependent pathways have been found. The activation of caspase-dependent or caspase-independent pathways is mainly associated with the type of cells undergoing the death process but is also influenced by the apoptotic stimulus. ^{24 25}  

Apoptosis, whether caspase dependent or independent, is characterized by the formation of apoptotic bodies and phagocytosis of the dying cells by macrophages or neighboring cells. In necrosis, however, the plasma membrane is fragmented, and the cellular content reaches the intercellular space, where it produces a strong inflammatory reaction and tissue damage. 

Our group has developed a biodegradable viscous poly(ortho ester) (POE) polymer with an excellent biocompatibility profile. We have described the ability of this polymer to release 5-FU in the subconjunctival space for an extended period, ²⁶ maintaining filtration after trabeculectomy in rabbit eyes along with reduced toxicity. ^{27 28} In the present study, we have evaluated the antiproliferative and proapoptotic properties of 5-chlorouracil (5-CU), a synthetic pyrimidine analogue, and compared them with those observed with 5-FU. 

Materials and Methods

In Vitro Experiments: Comparison of Cell Death Mechanisms and Antimitotic Effects Induced by 5-FU or 5-CU

Hoechst Test and Neutral Red Test.

A human conjunctival cell line (Wong-Kilbourne derivative of Chang conjunctiva; American Type Culture Collection, Manassas, VA) was cultured under standard conditions (humidified atmosphere of 5% CO₂ at 37°C) in Dulbecco minimum essential medium (DMEM; Eurobio, Les Ulis, France) supplemented with 10% fetal bovine serum (Dominique Dutscher, Brumath, France), 0.5% penicillin/streptomycin, and 1% L-glutamine (Eurobio). Normal culture development was assessed daily by phase-contrast microscopy. Confluent cultures were removed by gentle trypsin incubation, and cells were counted. For cytotoxicity experiments, cells were seeded into 96-well culture plates (18,000 cells per well; Nunc, Roskilde, Denmark) until subconfluence (culture surface covering nearly 70%) was reached. 

Experiments were performed in microplate cold light fluorometry (Fluorolite 1000; Dynex, Issy-les-Moulineaux, France), as previously described. ^{29 30} According to the recommendations of the European Center for the Validation of Alternative Methods and to previously validated methods in the Chang cell line, ^{31 32} two cellular markers were evaluated, cellular viability and chromatin condensation. Cellular viability was evaluated with a neutral red (Fluka, Ronkonkoma, NY) test with fluorometric detection (excitation, 535 nm; emission, 600 nm). An intercalating dye (Hoechst 33342; Molecular Probes, Eugene, OR) was used that allows determination of the total chromatin quantity variations and the degree of chromatin condensation. ³³ It was used on cells at a final concentration of 10 μg/mL (excitation, 360 nm; emission, 450 nm). Propidium iodide (0.5 mg/mL; Roche Diagnostics Corp., Meylan, France) was added to the intercalating dye (Hoechst 33342; Molecular Probes) solution to control necrotic cells, as previously described. ³⁴ Supravital uptake of the intercalating dye (Hoechst 33342; Molecular Probes) combined with the extrusion of propidium iodide was proposed as an assay for apoptosis. ³³ Cell viability and chromatin condensation were evaluated after 24 and 72 hours incubation, with 0, 0.1, 1, 10, 100, and 1000 μg/mL 5-FU or 5-CU. The value of IC₅₀ (concentration inhibiting 50% of cell population) was used to compare 5-FU and 5-CU cytotoxicity. Dulbecco minimum essential medium was used as a control and was used to dilute the tested drugs. 

Isolation and Culture of Rabbit Tenon Fibroblasts.

A few pieces of Tenon capsule were isolated from rabbit eyes. Tenon capsule fibroblasts were dissociated and cultured in DMEM (Gibco BRL, Cergy Pontoise, France), supplemented with 10% fetal bovine serum (FBS; Dominique Dutscher), 1% penicillin-streptomycin (Eurobio, Les Ulis, France), and 1% glutamine (Eurobio) at 37°C in a humidified incubator saturated with 5% CO₂ and 95% air. 

Cells were seeded at a density of 2 × 10⁴ cells/well until they reached a subconfluent stage and were treated with 5-FU or 5-CU (0, 0.5, 5, 50, 500, 1000, or 5000 μg/mL) for 24 hours. Cell counting and monotetrazolium (MTT) were used to determine the effect on cell proliferation and survival. 

Colorimetric MTT Assay.

MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; Sigma; 1 mg/mL in PBS) was added to the wells, and plates were incubated at 37°C for 1 hour. Isopropanol was then added and mixed thoroughly to dissolve the dark blue crystals, and the plates were rapidly read on a microplate reader (model 450; Bio-Rad, Hercules, CA) using a test wavelength of 570 nm and a reference wavelength of 630 nm. 

Trypan Blue Exclusion Test.

To determine the effect of 5-CU and 5-FU on the proliferation of rabbit fibroblasts, cells were seeded onto 24-well plates and treated after 24 hours with different concentrations of 5-FU and 5-CU (0, 1, 10, 20, or 100 μg/mL), diluted in culture medium. After 24 hours, control and treated cells were washed with PBS, trypsinized, and stained with trypan blue 0.1% to 0.2% (Eurobio; solution at 0.4%), and the number of viable (uncolored) and dead (colored) cells was scored. The ratio between the number of uncolored cells and the total number of cells indicates the percentage of viable cells. 

Characterization of Cell Death Mechanisms

For these experiments, rabbit Tenon fibroblast cells were cultured in four wells (2 × 10⁴ cells/well; Labtek, Campbell, CA) and grown at 37°C in a humidified atmosphere containing 5% CO₂ and 95% air. Subconfluent cells were treated with 5 or 100 μg/mL 5-CU or 5-FU for 24 hours and then rinsed carefully with PBS. Control cells were run using the culture medium alone. All experiments were performed in triplicate. 

Annexin-V Binding.

For annexin-V binding, fibroblasts were rinsed, incubated in cold binding buffer for 5 minutes, and further incubated for 15 minutes on ice with 20 μL labeled annexin-V-FITC (Southern Biotechnology Associates Inc., Birmingham, AL) diluted in 200 μL of cold binding buffer, and rinsed with PBS (5 times, 5 minutes each). Slides were then mounted in mounting medium (Vectashield; Vector Laboratories, Burlingame, CA) containing DAPI. The slides were shed from light until examined. 

Terminal Transferase dUTP Nick End Labeling.

TUNEL assay was carried out as previously described (Roche Diagnostics Corp.). ³⁵  

Immunocytochemistry

Cell staining using the following antibodies was carried out as specifically described. Polyclonal L-DNase II antibody was diluted (1/100) in PBS containing 1% skimmed milk. ³⁶ Anti–apoptosis-inducing factor (AIF; Sigma, Saint-Quentin Fallavier, France) was diluted (1/100) in PBS containing 1% BSA. ³⁷ After incubation with the specific primary antibodies, the cells were rinsed (3 times, 5 minutes each) with PBS and then incubated for 1 hour with goat Alexa Fluor anti-rabbit IgG (Molecular Probes, Eugene, OR) diluted (1/250) in PBS containing 1% BSA and rabbit Alexa Fluor anti-goat (Molecular Probes) for 1 hour and washed in PBS (5 times, 5 minutes each). 

Statistical Analysis

Results were expressed as mean ± SEM. Statistical analyses were performed using the Mann–Whitney nonparametric test. 

Polymer Synthesis and Preparation of the Formulations

Materials.

The diketene acetal 3,9-diethylidene-2, 4, 8, 10-tetraoxaspiro[5,5] undecane (DETOSU) was furnished by A.P. Pharma (Redwood City, CA). Triethylamine (TEA), p-toluenesulfonic acid (p-TSA), and 1-decanol were purchased from Fluka (Chemie AG, Buchs, Switzerland). 1,10-Decanediol was purchased from Aldrich (Chemie, Steinheim, Germany), and DL-lactide was purchased from Purac. 5-CU was purchased from Sigma (Chemie AG). Solvents such as tetrahydrofuran (THF) and methanol were of analytical grade. 

Synthesis of Polymer.

1,10-Decanediol-lactate was synthesized by ring opening of lactide, as described previously. ³⁸ Thus, 7.205 g DL-lactide (50 × 10⁻³) and 8.713 g 1,10-decanediol (50 × 10⁻³) were introduced into a round-bottom flask sealed with a rubber septum under argon atmosphere. The mixture was vigorously stirred for 3 days at 160°C. Viscous diol-lactate was used in the polymerization without further purification. The synthesis of autocatalyzed POE₇₀LA₃₀ containing 30 mole percent lactic acid units in the polymer backbone has been previously described. ^{39 40} Briefly, POE was synthesized under anhydrous conditions (THF) by the acid catalyzed condensation of DETOSU (50 mmol) with 1,10-decanediol (20 mmol), 1,10-decanediol-lactate (15 mmol), and n-decanol (15 mmol) used as chain stoppers, as described previously. ³⁹ Low molecular-weight oligomers, unreacted monomers, and catalyst were removed by dissolution-precipitation using THF and methanol as solvent and nonsolvent. The precipitate was dried under vacuum at 40°C for 48 hours. Average molecular weights were determined by size exclusion chromatography using TSK polystyrene standards for the calibration curve. ³⁹  

Formulations.

Formulations were prepared under a laminar airflow hood. The added drug, 5-CU or 5-FU (Sigma, Buchs, Switzerland), had been gamma sterilized at 20 kGy and homogeneously dispersed in the aseptically prepared semisolid polymer under aseptic conditions at a concentration of 1% wt/wt. ⁴¹ The viscous mixture was conditioned in a 1-mL syringe; each sample was 200 μL (240 mg). 

In a preliminary investigation, the in vitro rate of release of 5-CU was determined according to the procedure described by Merkli et al. ⁴² and was evaluated at 8 μg/day over a period of up to 40 days. The 5-FU in vitro rate of release from POE has also previously been described by Heller et al. ⁴³ It was evaluated at 100 μg/day under the same conditions. 

In Vivo Experiments

Animals.

Pigmented Fauve de Bourgogne female rabbits, each weighing from 2 to 3 kg and 10 to 12 weeks of age, were used in this study (Elevage des Pins, Epeigne sur Deme, France). All experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Rabbits received intramuscular injections of general anesthesia (50 mg/kg ketamine and 15 mg/kg xylazine), and local anesthesia was obtained with oxybuprocaine 0.4%. At the end of the experiments, rabbits were killed by injection of lethal doses of pentobarbital. 

Protocols.

In this experiment, we evaluated the efficiency of a slow release of 5-FU or 5-CU in maintaining postoperative intraocular pressure (IOP) below its preoperative value. Twenty-four rabbits underwent trabeculectomy and were divided into four groups (n = 6), as follows: trabeculectomy alone without any subconjunctival injection (group 1); trabeculectomy with preoperative subconjunctival injection of 200 μL POE (group 2); trabeculectomy with preoperative subconjunctival injection of 200 μL POE containing 1% wt/wt 5-FU (group 3); and trabeculectomy with preoperative subconjunctival injection of 200 μL POE containing 1% wt/wt 5-CU (group 4). IOP and filtration bleb persistence were monitored for 34 days. 

Given that POE-5-FU had already been evaluated and that the aim of this study was to evaluate the potential of 5-CU delivered in a sustained manner to allow for a long-term maintenance of filtration, only the group of rabbits treated with POE-5-CU and their controls (trabeculectomy alone and POE alone) were followed up for 100 days. 

Glaucoma-Filtering Surgery Procedure

Trabeculectomy was performed as previously described. ²⁷ Briefly, a half-thickness, limbus-based, 4 × 4-mm scleral flap was extended anterior to the limbus and was flipped over. A 3-mm long limbal incision was made with a 45° blade that entered the anterior chamber. A piece of tissue containing inner sclera, trabeculum, and peripheral cornea, measuring approximately 3 × 1 mm, was excised. Peripheral iridectomy was performed. The scleral flap was sutured with two 10–0 nylon sutures. The conjunctiva was repositioned with a continuous suture (8–0 Vicryl; Ethicon, Ethnor, Neuilly, France). 

Subconjunctival Injections

Just before the last conjunctival suture was closed, a 0.9-mm diameter metallic needle (20 G) was inserted into the subconjunctival space, and injection of one of the following was delivered in a masked manner: 200 μL POE (group 2), POE/1% wt/wt 5-FU (group 3), or POE/1% wt/wt 5-CU (group 4), as previously described. ²⁷  

Clinical Follow-up

In the four groups, slit lamp (BQ900; Haag Streit, Harlow, Essex, UK) observations were performed at 5, 8, 15, 22, 27, and 34 days after surgery, to evaluate the filtering bleb persistence. At 1, 5, 15, and 34 days, IOP was measured using Goldmann applanation tonometry and was compared with the preoperative IOP. In groups 1, 2, and 4 POE/5-CU, we followed IOP evolution for an additional period of 66 days. 

Conjunctival redness, corneal transparency, and corneal or conjunctival epithelial defects were recorded on days 5, 15, and 34 for all four groups of animals. If an elevation of thinned conjunctiva was observed, the presence of a bleb was recorded. The extent of the bleb surface was not considered separately in the present evaluation. Vascularity of the bleb and leakage were also recorded. Clinical examination was performed in a masked manner by a single observer (MB). 

Histologic Examination

At 100 days after surgery, a silk suture was positioned at the trabeculectomy site for identification. The eyes were enucleated, fixed in Bouin solution, and embedded in paraffin for light microscopy. Anteroposterior sections were stained with hematoxylin-eosin to examine the conjunctiva, the iridocorneal angle, and the appearance of the bleb. Samples from the different groups of eyes were treated simultaneously to reduce potential biases that might have been induced by fixation procedure variations. Four eyes with trabeculectomy alone, four eyes with trabeculectomy treated with POE alone, four eyes with trabeculectomy treated with POE-5-CU, and four eyes of nonoperated control rabbits were studied. 

Histologic analysis of the filtration site was evaluated in a masked manner according to the grading system published by Perkins et al. ⁴⁴ with slight modifications, as follows: 0, no anatomic alterations; 1, minimal alterations demonstrating thickening of the conjunctiva but preservation of the conjunctival epithelium; 2, moderate alterations demonstrating thickening and increased cellularity of the conjunctiva; 3, as in grade 2 with increased collagen density and loss of fibril orientation in the stromal collagen, suggesting scar formation; 4, severe alteration with the loss of conjunctival epithelium, necrosis of the conjunctiva, total disorganization, and necrosis of the sclera underlying the treated area. Large vacuoles in the treated site were considered indications of functioning blebs. 

Results

In Vitro Experiments: Cell Death Mechanisms and Antimitotic Effects of 5-FU and 5-CU

Chromatin Condensation Evaluation.

After 24 hours of incubation with 5-FU or 5-CU at concentrations of up to 1 μg/mL, no significant difference in chromatin condensation was observed in subconfluent human conjunctival cells compared with control cells (without drug). 5-CU concentrations of 10 to 1000 μg/mL induced a significant increase in Hoechst fluorescence when compared with identical 5-FU concentrations (Fig. 1A ; P = 0.0079), indicating that 5-CU induces more apoptotic cell death than 5-FU. After 72 hours of exposure, a similar chromatin condensation pattern suggestive of apoptosis was observed in cells treated with 5-CU at 1 μg/mL to 100 μg/mL (Fig. 1B) . Interestingly, increasing 5-FU concentrations (>1 μg/mL) induced a dose-dependent significant decrease in chromatin condensation after 24 (P = 0.0079) or 72 (P = 0.02) hours of incubation (Figs. 1A 1B) . At very high concentrations (5 mg/mL), both 5-CU and 5-FU induced decreases in chromatin condensation at 24 or 72 hours (not shown). This test demonstrates that apoptotic death is induced on subconfluent conjunctival cells by 5-CU (up to 0.1 mg/mL) but not by 5-FU concentrations. 

Neutral Red Test.

After 24-hour incubation at concentrations up to 10 μg/mL, 5-FU and 5-CU induced similarly low cell viability decreases of Chang conjunctival cells (approximately 20%). At higher concentrations (100 μg/mL), 5-FU induced 55% cell death, whereas 5-CU concentrations of up to 1000 μg/mL induced only 35% cell death (Fig. 1C ; P = 0.0079). Incubation with 5-FU for 72 hours induced significantly decreased cell viability at a concentration as low as 1 μg/mL (Fig. 1D) . With 5-CU, however, only the highest concentration (4 mg/mL) induced a significant reduction in cell viability (not shown). 

Hoescht and neutral red test results can be interpreted to indicate that a different mechanism of cell death was activated by 5-FU or 5-CU. When the cells were incubated with 5-CU, an apoptotic cell death mechanism took place with an IC₅₀ of 2.5 mg/mL after 72 hours of exposure. Incubation with 5-FU, on the other hand, induced an apoptonecrotic cell death with an IC₅₀ of 7 μg/mL after 72 hours of exposure. 

Effect of 5-CU and 5-FU on Rabbit Fibroblast Viability

The toxicity of 5-CU on subconfluent Tenon capsule fibroblasts was reduced compared with that of 5-FU. After 24 hours of exposure, the highest 5-CU concentration (5000 μg/mL) induced only 33% cell death. Cultures incubated with 5-FU showed a level of 39.1% cell death with 5 μg/mL and 51% cell death with 50 μg/mL. Thus, 5-CU is 1000 times less toxic to nonproliferating rabbit conjunctival cells than 5-FU (Fig. 2) . 

Effect of 5-FU and 5-CU on Rabbit Fibroblast Proliferation

Cell cultures in the proliferating phase showed that 5-FU and 5-CU induced a dose-dependent inhibition after 24-hour incubation with the drugs. This inhibition was significant at concentrations of 10 μg/mL or greater. At a concentration of 10 μg/mL, 5-FU and 5-CU inhibited, respectively, 28% (P = 0.007) and 23% (P = 0.03) of cell proliferation. With 20 μg/mL, the inhibitions were, respectively, 40% (P = 0.0018) and 23% (P = 0.03), and with 100 μg/mL, they were 43% (P = 0.0014) and 25% (P = 0.009; Fig. 3 ). These experiments demonstrated that 5-CU is less antimitotic than 5-FU. Furthermore, increasing the concentration of 5-CU did not significantly increase its antimitotic activity. 

Analysis of Cell Death Mechanisms

TUNEL Assay and Annexin-V Binding.

As shown in Figure 4 , low (5 μg/mL) and high (100 μg/mL) doses of 5-CU induced nuclear fragmentation, an apoptotic morphology (Figs. 4B 4D) . However, these apoptotic cells were not labeled by the TUNEL assay, suggesting that a caspase-independent pathway might have been involved in this apoptotic process (Figs. 4A 4C) . When exposed to low doses of 5-FU, rare TUNEL-positive cells could be observed (Fig. 4E) . No TUNEL-positive cells and no apoptotic nuclei were observed when cells were treated with the higher dose of 5-FU (Figs. 4G 4H) . 

After exposure to 5-CU (high or low dose), a retracted cytoplasm, condensed nuclei (Figs. 5A 5C) , and diffuse annexin binding on the cell surface (Figs. 5B 5D , white arrows) are observed. When exposed to 5-FU, cells with enlarged nuclei, altered chromatin (Fig. 5E) , and annexin within the cell cytoplasm are observed. These findings indicate the enhanced membrane permeability of these cells (Figs. 5E 5F 5G 5H) . 

LEI and AIF.

Effectors of caspase-independent apoptosis are the AIF and the LEI/L-DNase II. ⁴⁵ After cellular stress, LEI is cleaved by serine proteases ³⁶ and is transformed into L-DNase II, an enzyme with a nuclear localization. In Tenon fibroblasts, LEI is localized in the cytoplasm and concentrates in the cell nucleus when the cell enters the apoptotic cycle. As shown in Figure 6 , when cells are exposed to low doses of 5-CU (5 μg/mL), translocated LEI can be observed in a few cells with condensed or fragmented nuclei (Figs. 6A 6B) . When the cultured cells are exposed to higher doses of 5-CU (100 μg/mL), LEI is observed in the nuclei of numerous cells (Figs. 6C 6D) , suggesting that a caspase-independent form of apoptosis is taking place. In cells treated with either the low or the high dose of 5-FU, there was no translocation of LEI to the nuclei (Figs. 6E 6F 6G 6H) . 

AIF is a caspase-independent apoptosis marker released from the mitochondria and located in the nuclei of apoptotic cells. ⁴⁶ As shown in Figure 7 , nuclear translocation of AIF is only observed in cells exposed to 5-CU (Figs. 7A 7C)and not in cells exposed to 5-FU (Figs. 7E 7G) . AIF concentration in the cell nuclei was observed only in cells with condensed nucleic morphologic signs for apoptosis (Figs. 7B 7D) . 

In Vivo Experiments

Intraocular Pressure.

As illustrated in Figure 8 , in group 1 eyes that underwent trabeculectomy alone (black line), the mean IOP decreased immediately after surgery and remained at nearly the same level until day 15 after surgery. By day 28, however, preoperative IOP levels were exceeded. In group 2 eyes receiving POE alone, (dark gray line, filled square), preoperative IOP levels were back to baseline level on day 8 after surgery. On the other hand, the mean IOP remained below its preoperative values at day 34 after surgery in group 3 eyes receiving POE/5-FU (black interrupted line) and in group 4 eyes receiving POE/5-CU (light gray disc line). At this time point, the mean IOP in the POE/5-CU–treated group (group 4) was 83% of the baseline level and only 40% in the POE/5-FU–treated group (group 3). 

Long-term follow-up of IOP showed that, approximately 100 days after surgery, IOP was still decreased in the POE/5-CU group compared with the controls (black and dark gray lines) and still inferior to the preoperative value. The mean long-term IOP, with all time points considered, was significantly (P < 0.0001) decreased in the POE/5-CU–treated group (6.0 ± 2.4 mm Hg) when compared with both control groups, the trabeculectomy alone group (7.6 ± 2.9 mm Hg), or the POE alone group (7.5 ± 2.6 mm Hg). 

Bleb Survival Curves.

Figure 9A 9B 9C 9Dshows the surgical sites after surgery. Four weeks after surgery, in group 1 eyes (trabeculectomy alone), a flat bleb is observed (Fig. 9C) . In group 4 eyes (trabeculectomy + POE IV/1% 5-CU), a high and extensive bleb is observed (Fig. 9B) . 

Figure 9Edisplays bleb survival curves in the four groups, from day 0 to day 34 after surgery. Fifty percent of blebs had collapsed by day 17 in group 1 eyes (trabeculectomy alone), with only 13.3% still patent by 34 days. In group 2 eyes (POE alone), 50% of blebs had collapsed by 21 days, and 81.8% had collapsed by 34 days. Group 3 eyes (POE/5-FU) showed that 80% were still patent at day 34. In group 4 eyes (POE/5-CU), all blebs remained patent by day 22 after surgery, and 80% survived after 1 month. At 100 days after surgery, blebs were still observed in 1 of 6 eyes in group 1 and group 2 eyes (16%), whereas in group 4 eyes (POE/5-CU), blebs were still present in 5 of 6 (83%) eyes. 

Clinical Examination

In the group of rabbits that underwent trabeculectomy alone, no conjunctival defects were observed during the follow-up period. A delay in conjunctival healing was observed in one eye in the POE alone group and in one eye in the POE/5-CU group. However, on day 34 after surgery, in both groups, no persistent epithelial or healing defects were observed. No persistent corneal edema was observed in any group at day 34. Conjunctival hyperemia totally regressed in all eyes treated with POE-5-FU and POE-5-CU at day 8, whereas in eyes treated with POE alone or trabeculectomy alone, these signs regressed in all eyes only at day 15. No persistent conjunctival redness was observed in any of the eyes at day 34. 

Interestingly, at day 100 after surgery, subconjunctival POE bubbles were still observed in the group of rabbits treated with POE alone, but they could not be detected at this time point in the group of eyes treated with POE-5-CU. 

Histologic Examination

In the group of rabbits that underwent trabeculectomy alone, histology scores were 2 in two eyes and 3 in the two other eyes (Fig. 10A , example of grade 3), with the presence of a bleb in 1 of the 4 eyes. In the group of eyes with “empty” POE, three eyes were scored as 2 (Fig. 11A , example of grade 2) and one eye was scored as 3. In the group of eyes treated with POE/5-CU, two eyes had a grade score of 1 and two eyes had a grade score of 2. In this group, a filtration site was seen in two eyes and blebs formation in 3 of the 4 examined eyes (Fig. 11B) . 

Discussion

Our results show that 5-CU may have advantages for the inhibition of undesired fibroblast proliferation taking place after surgery for glaucoma and results in filtration failure. Comparison of 5-CU and 5-FU cytotoxicity shows that 5-CU is approximately 1000 times less toxic on nonproliferative cells while it maintains a good antimitotic activity. Our results demonstrate that 5-CU has a more specific influence on proliferating cells. Furthermore, we observed that 5-CU induces apoptotic controlled cell death, whereas the influence of 5-FU is by apoptonecrosis. We observed that the 5-CU apoptotic mechanism is driven mainly by nuclear translocation of AIF and activation of the LEI/L-DNase II pathways, which are caspase-independent pathways. L-DNase II activation is observed primarily in TUNEL-negative apoptotic cells because this enzyme cleaves the DNA liberating 3′P ends, which cannot be labeled by terminal transferase (TUNEL reaction). ^{47 48 49} Although in 5-CU–treated cells annexin V binding to the cell surface is observed, in 5-FU–treated cells, increased cell membrane permeability induced intracytoplasmic accumulation of the annexin molecules. Moreover, in 5-FU–treated cells, enlarged nuclei with altered chromatin were observed, suggesting that necrosis was taking place. Usually, apoptotic programmed cell death does not induce an inflammatory reaction. Therefore, a drug inducing apoptosis rather than necrosis may have advantages for the preservation of patent-filtering blebs after glaucoma surgery. 

To use low doses of antimitotic agents and to inhibit the proliferative processes taking place during the postoperative wound healing period, a continuous controlled release of the drug is required. To this purpose, we have previously studied the ocular tolerance to POE for potential clinical uses. ²⁸  

In vitro, 1% 5-CU is released from POE intravenously at a daily rate of 8 μg/d for up to 40 days. Thus this slow, controlled, and continuous release can inhibit fibroblast proliferation over a period critical for the maintenance of a patent bleb after filtering surgery for glaucoma. 

In tissue culture, 5-CU at a concentration of 10 μg/mL inhibits 23% of the fibroblast proliferation. In vivo, we have found that 5-CU sequestered within POE maintains low IOP levels for up to 5 months after filtration surgery in the rabbit eye. Previous studies on animal models using other antiproliferative agents in a single-dose application have not yielded such promising long-term results. ⁵⁰ Single application of antiproliferative agents and subconjunctival injections are the only methods currently available. Earlier studies on 5-FU in tissue cultures showed that the inhibition of fibroblast proliferation was still observed after the discontinuation of treatment if exposure to the drug was sufficiently long or if its concentrations were sufficiently high. ⁸ Corneal epithelial cell toxicity induced by high concentrations of 5-FU limit their routine use in patients. Subconjunctival injections of 5-FU maintain an ID₅₀ level up to 24 hours at the injection site and 180° around it. However, the corneal levels are also high (20 μg/g at 1 hour, 5 μg/g at 24 hours). ⁵¹ With frequent topical administration, therapeutic levels can also be maintained. Unfortunately, they are accompanied by toxicity to the corneal epithelium. More recently, Chaudhry et al. ⁵² showed no significant benefit of a single perioperative 5-FU application over placebo. Given this, it became evident that slow-release devices are needed. Nonbiodegradable polyanhydride discs were used and improved the outcomes of filtering surgery. However, a high rate of complications was also observed and included surgical wound dehiscence, disc extrusion, and Tenon cysts. ⁵³ Fluoroorotate has been delivered during glaucoma filtration surgery using a liposomal formulation, but the liposomes are cleared rapidly from the eye. ⁵⁴ Hostyn et al. ⁵⁵ described a semisolid biodegradable implant with 5-FU that constantly released drug for 18 days and was biodegraded in less than 86 days. Topical 5-fluorocytosine, a prodrug of 5-FU, can be delivered to the subconjunctival space and adenoviral vector. ⁵⁶ In the present study, we found a high in vivo efficacy using POE loaded with 1% 5-CU. In this slow-release system, we obtained a drug release of 8 μg/d for up to 40 days without any evident corneal toxicity in the treated eyes. The reduced corneal toxicity may be the result of localized slow and controlled release of the drug and of the fact that 5-CU is less toxic to proliferating corneal epithelial cells. Moreover, our study shows that 5-CU–induced cell death is carried out through an apoptosis mechanism, whereas 5-FU tends to involve necrotic pathways. Histologic analysis of surgical sites 3 months after surgery showed that in POE/5-CU–treated eyes, functional filtering blebs were associated with preservation of the conjunctival epithelium and without any severe anatomic changes of the underlying sclera. We have previously demonstrated that POE biocompatibility is good in different ocular implantation sites. ²⁸ Sequestering the 5-CU within POE does not reduce the excellent biocompatibility of POE. Thus, development of the POE-based delivery system may allow the use of less toxic agents with enhanced efficacy. 

In conclusion, despite the fact that 5-CU is a weaker antiproliferating agent than 5-FU, its reduced toxic side effects to the epithelium and sclera allow for its enhanced efficacy and extended reduction of IOP after filtering surgery for glaucoma. The induction of an apoptosis mechanism by 5-CU is carried out mainly through a caspase-independent pathway involving LEI/L-DNase II. Using POE as a slow-release, biodegradable delivery system of 5-CU designed and tested specifically for the eye shows that it can be a promising alternative therapeutic application in patients at high risk for failed filtering surgery. 

 Submitted for publication July 20, 2007; revised December 23, 2007; accepted May 19, 2008.

 Disclosure: M. Berdugo Polak, P; F. Valamanesh, P; O. Felt, P; A. Torriglia, P; J.-C. Jeanny, P; J.-L. Bourges, P; P. Rat, P; A. Thomas-Doyle, P; D. BenEzra, P; R. Gurny, P; F. Behar-Cohen, P

 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.

 

Cell death and cell viability after 5-FU (gray columns) and 5-CU (black columns) exposure. As evaluated by Hoechst 33342 test (A, B), the chromatin condensation (apoptosis feature) of conjunctival cells (Chang) was significantly increased when exposed to 5-CU (black columns) compared with 5-FU (gray columns) at concentrations higher than 10 μg/mL after 24 hours (A) and higher than 1 μg/mL after 72 hours (B). Cellular viability (neutral red test) was significantly reduced for 5-FU compared with exposure to 5-CU after 24-hour (C) and 72-hour (D) incubations at similar concentrations. Results are expressed as a percentage of control values. *P < 0.05.

Figure 2.

 

Rabbit Tenon fibroblast cytotoxicity of 5-FU and 5-CU. Mitochondrial activity (MTT test) reflecting the number of living cells is presented after 24-hour incubation with 0.5 to 5000 μg/mL 5-FU (gray interrupted line) or 5- CU (black continuous line). MTT was markedly reduced at concentrations higher than 0.5 μg/mL of 5-FU, whereas only a mild reduction was observed with higher concentrations of 5-CU.

Figure 3.

 

Inhibition of rabbit fibroblast proliferation by 5-FU and 5-CU. Rabbit Tenon fibroblast cultures were incubated during 24 hours with 0 to 100 μg/mL 5-FU (gray interrupted line) or 5-CU (black continuous line). A reduction in the number of cells was observed for both drugs. This reduction was more prominent in 5-FU–exposed cultures with concentrations higher than 10 μg/mL.

Figure 4.

 

TUNEL analysis of fibroblasts treated with 5-FU or 5-CU. Rabbit Tenon fibroblasts were exposed to 5 μg/mL concentration (A, B, E, F) or 100 μg/mL concentration (C, D, G, H) of 5-CU or 5-FU. TUNEL-positive cells were looked for after 24 hours of exposure. (A) TUNEL with DAPI staining of cells exposed to a low dose of 5 μg/mL 5-CU. White arrow indicates one TUNEL-positive cell. (B) DAPI staining of the same cells. Arrowheads indicate fragmented nuclei suggestive of the presence of apoptotic cells that are TUNEL negative. (C) TUNEL with DAPI staining of cells exposed to 100 μg/mL 5-CU. (D) DAPI staining of the same cells (arrowheads indicate fragmented nuclei, suggestive of apoptosis). (E) TUNEL with DAPI staining of cells exposed to 5 μg/mL 5-FU. White arrow indicates TUNEL-positive cells. (F) DAPI staining of the same cells. (G) TUNEL with DAPI staining of cells exposed to 100 μg/mL 5-FU. (H) DAPI staining of the same cells. (I, J) No TUNEL-positive cells in control untreated cells. Scale bar, 20 μm.

Figure 5.

 

Analysis of annexin-V binding. Rabbit Tenon fibroblasts were exposed to 5 μg/mL concentration (A, B, E, F) or 100 μg/mL concentration (C, D, G, H) of 5-CU or 5-FU. Annexin-V binding was tested after 24 hours of exposure. (A) Annexin-V with DAPI staining of cells exposed to 5 μg/mL 5-CU. (B) Annexin V binding. White arrows indicate cells with diffuse binding of Annexin V on the cell surface suggesting phosphatidylserine exposure. (C) Annexin V with DAPI staining of cells exposed to 100 μg/mL 5-CU. (D) Annexin V staining of the same cells. White arrow indicates a cell with condensed nucleus and diffuse surface binding of annexin V. (E) Annexin V with DAPI staining of cells exposed to 5 μg/mL 5-FU. Higher magnification of the nucleus shows chromatin alteration. (F) Annexin staining of the same cells showing clear cytoplasmic accumulation of annexin V inside the cytoplasm, suggesting enhanced cell permeability. (G) Annexin V with DAPI staining of cells exposed to higher dose of 5-FU. (H) Annexin staining of the same cells showing clear cytoplasmic accumulation of annexin V within the cytoplasm, suggesting enhanced cell permeability (white arrows). (I, J) No annexin V binding in control cells. Scale bar, 20 μm.

Figure 6.

 

Fluorescent microphotographs of LEI/L-DNase II immunohistochemistry. Immunohistochemistry of LEI/L-DNase II and double staining with DAPI. LEI is located in the cytoplasm and is concentrated in the nuclei of some cells (arrows and high magnification) exposed to a low dose of 5-CU (A, B) and a high dose of 5-CU (C, D). No nuclear translocation of LEI/L-DNase II is observed in cells exposed to a low dose of 5-FU (E, F) or a high dose of 5-FU (G, H) or in control cells (I, J). Scale bar, 20 μm.

Figure 7.

 

Fluorescent microphotographs of AIF immunohistochemistry. Immunohistochemistry of double staining of AIF with DAPI, and DAPI alone. AIF is located in the cytoplasm mitochondria and is translocated to the nuclei of apoptotic cells treated with 5 μg/mL 5-CU (A, B) or 100 μg/mL 5-CU (C, D, white arrows). No nuclear location of AIF is observed in cells treated with a low dose of 5-FU (E, F), even when condensed nuclei are observed (high magnification), or in cells treated with a high dose of 5-FU (G, H) or in control untreated cells (I, J). Scale bar, 20 μm.

Figure 8.

 

Follow-up of mean IOP. IOP (mm Hg) measured with a Goldmann applanation tonometer before trabeculectomy (day 0) and during a 3-month postoperative follow-up period. POE/1% 5-FU (black interrupted line), POE/1% 5-CU (light gray line), trabeculectomy alone (black continuous line), or trabeculectomy + POE (dark gray line). IOP remained lower in 5-FU– and 5-CU–treated groups until the last examination. Results are expressed as IOP mean ± SD.

Figure 9.

 

Blebs survival curves. Slit lamp photographs of the rabbits. (A) Eight days after trabeculectomy with injection of POE-5CU showing the polymer under the conjunctiva (arrow). (B) Four weeks after surgery, well-shaped bleb could be observed in the trabeculectomy + POE IV/1% 5-CU rabbit (arrow). (C) Four weeks after surgery, a flat bleb is observed in the trabeculectomy alone rabbit. (D) Four weeks after surgery, a high and well-formed bleb is observed in a rabbit from the trabeculectomy+ POE/1% 5-FU rabbit (arrow). On Kaplan-Meier survival curves (E), postoperative filtrating blebs from group 1 (trabeculectomy alone; interrupted gray line) and group 2 (trabeculectomy + POE; interrupted black line) failed earlier and more frequently than those from group 3 (trabeculectomy + POE IV/1% 5-FU; continuous gray line) or group 4 (trabeculectomy + POE IV/1% 5-CU; continuous black line).

Figure 10.

 

Histology microphotographs. (A) Trabeculectomy site 3 months after surgery alone, showing a thickened conjunctiva (double arrow) with increased cellularity and dense and disorganized collagen fibrils in the sclera (star) compared with the control, nonoperated rabbit eye (B). Scale bar, 200 μm. (a, b) Low-magnification images of the surgical sites. Insets show the conjunctiva epithelium at high magnification (×80).

Figure 11.

 

Histology microphotographs. (A) Trabeculectomy site 3 months after surgery with POE alone, showing thickening of the conjunctiva (double arrow), increased cellularity but without sclera changes. Inset shows the conjunctiva epithelium with some goblet cells. (B) Trabeculectomy site 3 months after surgery with POE-5-CU, showing large vacuoles (stars) under a preserved conjunctival epithelium, suggesting a functioning bleb. Note that a cleft is observed in the sclera (arrow) without scar tissue around it. Insets show the epithelium at high magnification with numerous goblet cells. (a, b) Low magnification of the surgical sites. Scale bar, 200 μm.