Mitochondrial Activity and Glutathione Injury in Apoptosis Induced by Unpreserved and Preserved β-Blockers on Chang Conjunctival Cells

Abstract

purpose. Quaternary ammonium ions have been demonstrated to induce apoptosis correlated with superoxide anion production in vitro. The purpose of this study was to further investigate the mechanisms of benzalkonium chloride (BAC), unpreserved and preserved β-blocker eye-drops–induced programmed cell death, with special attention to the roles of mitochondrial transmembrane potential and intracellular reduced glutathione.

methods. Chang conjunctival cells were incubated with different concentrations of unpreserved or preserved timolol (0.1%, 0.25%, and 0.4%), or carteolol (1% and 2%), or BAC (0.0001% to 0.01%) for 15 minutes, or for 15 minutes with a 24-hour recovery period in normal medium. Cellular viability (neutral red test), mitochondrial activity (rhodamine 123 test), intracellular reduced glutathione (monochlorobimane test), DNA condensation (Hoechst 33342 test), and reactive oxygen species (ROS) production (dichlorofluorescein diacetate and hydroethidine tests) were evaluated using microplate cold-light cytofluorometry.

results. A significant, concentration-dependent decrease in cellular viability was found with preserved β-blockers and with BAC alone, whereas unpreserved preparations did not show any toxicity. Only preservedβ -blockers induced chromatin condensation associated with an alteration of mitochondrial activity and a decrease of glutathione, suggesting an apoptotic phenomenon. BAC increased glutathione after 15 minutes, whereas a decrease was observed after a recovery period. ROS production was found with preserved formulations at significantly higher levels than those observed with unpreserved drugs.

conclusions. This in vitro study demonstrates that oxidative stress, evidenced by enhanced ROS production and mitochondrial injury rather than by cellular glutathione depletion, is a mechanism involved in apoptosis induced by preservative-containing eye-drops.

Cell death can follow two distinct pathways, apoptosis or necrosis. However, the early biochemical events that dictate the mode of cell death are still unclear. Various stimuli can induce programmed cell death. One of them is benzalkonium chloride (BAC), the most commonly used preservative in many available ophthalmic solutions. Apoptosis is a complex process characterized by cell shrinkage, chromatin condensation, and internucleosomal DNA fragmentation. ¹ Several protease families are implicated in apoptosis, the most prominent being caspases. ² Different models for caspase activation have been proposed: apoptosis induced by ligation of cell surface receptors, such as Fas or TNF-R, represents a pathway almost exclusively controlled by caspases; mitochondrial dysfunction, which occurs during apoptosis, can cause the release of cytochrome c from mitochondria into cytosol, where it binds to apoptotic protease activating factor (Apaf-)-1 and activates caspases; and perforin and granzyme B produced by cytotoxic cells can also activate the caspase cascade. ³ A role for oxidative stress in apoptosis has also been described. ^{4 5} In addition, depletion of glutathione (GSH) pools has been suggested to be part of the cell death effector machinery and accompanies ROS production during apoptosis in relevant systems. ⁶  

Several investigators have reported toxic side effects induced by BAC. ^{7 8 9} We recently showed that BAC was a strong proapoptotic agent in conjunctival cells ¹⁰ involving superoxide anion production. ⁵ Furthermore, we showed in an experimental model and in patients with glaucoma that preserved antiglaucoma drugs induce significant inflammatory and histopathologic changes in the ocular surface and in deeper ocular structures. ⁹ An increasing number of studies, both experimental and epidemiologic, have shown that glaucoma filtration surgery has progressively become less effective than initially described. ^{11 12 13 14} It has been reported in tissue cultures of human Tenon capsule fibroblasts that unpreserved timolol shows less toxicity than preserved β-blockers. ¹⁵ Animal studies have also confirmed that antiglaucoma drugs may induce toxic effects on the ocular surface, in part explained by the presence of BAC. ¹⁶ Our in vitro studies confirmed that quaternary ammonium ions may induce a large part of these effects. 

We further describe a series of experiments on the in vitro toxicity of BAC, preserved timolol (timolol BAC+), preserved carteolol (carteolol BAC+), preservative-free timolol (timolol BAC−), and preservative-free carteolol (carteolol BAC−). We used a human continuous conjunctival cell line previously used in toxicologic ocular studies in vitro. ^{5 10 17 18} This study was performed on live rather than dead cells, because labile markers can be significantly affected by the use of extraction techniques. We performed an original technique of cold-light cytofluorometry ¹⁹ validated in previous studies. ^{5 20 21} The purpose of this study, which is a continuation of our former work, was therefore to investigate in the same human continuous conjunctival cell line the immediate and delayed actions of different concentrations of preservative and antiglaucomatous preparations, to better understand the mechanisms of toxicity and to discriminate between apoptosis and necrosis in cell death. We evaluated cell viability, DNA condensation, cellular GSH production, and mitochondrial activity. Reactive oxygen species (ROS) production was also explored to extend our knowledge of iatrogenic ocular surface changes induced by antiglaucoma topical treatments. 

Materials and Methods

Conjunctival Cell Line

Wong Kilbourne–derived human conjunctival epithelial cells, an established cell line (Wong Kilbourne derivative of Chang conjunctiva, clone 1-5c-4, CCL-20.2; American Type Culture Collection [ATCC], Rockville, MD) were cultured in 75-cm² culture flasks (Nunc, Roskild, Denmark). The culture medium was Dulbecco minimum essential medium (DMEM; Eurobio, Les Ulis, France) supplemented with 10% fetal bovine serum (Dominique Dutscher, Brumath, France), 1% glutamine (Eurobio), 0.1% ampicillin (Panpharma, Fougères, France), and 2% kanamycin (Bristol-Myers Squibb, Paris, France). Cells from passages 6 to 17 (after ATCC initial passage 65) were used in all experiments. Cells were cultured under standard conditions (humidified atmosphere of 5% CO₂ at 37°C), and the medium was changed every 3 days. Normal culture development was assessed daily by phase-contrast microscopy. Confluent cultures were removed by gentle trypsin incubation, and cells were counted and then seeded into 96-well culture plates (5000 cells/well; Nunc). Cultures were kept at 37°C for 24 hours. After subconfluence was attained (culture surface covering nearly 70%), cells were exposed to the different formulations. 

Cell Treatments

Timolol BAC+ (with a fixed concentration of 0.01% BAC, as in the commercial solution; Transphyto, Clermont-Ferrand, France) and timolol BAC− (Transphyto) were tested at final concentrations of 0.1%, 0.25%, and 0.4% after dilution with culture medium; Carteolol BAC+ (with a fixed concentration of 0.005% BAC as in the commercial solution; Chauvin, Montpellier, France) and carteolol BAC− (Transphyto) were tested at final concentrations of 1% and 2% after dilution with culture medium; BAC (Transphyto) alone was tested at concentrations ranging from 0.0001% to 0.01%. All dilutions were realized in culture medium. The complete culture medium was used as negative control. Two incubation times were applied to control and treated cells: 15 minutes of treatment and 15 minutes with a 24-hour recovery period in normal culture medium to approach clinical conditions, as performed in our previous studies. ^{5 21}  

Experimental Procedures

Experiments were performed using a recent, previously validated fluorometric technology founded on the concept of cold-light fluorometry that allows fluorometric detection (280–870 nm) with high sensitivity (picograms to femtograms per milliliter) and specificity. ⁵ Cytotoxicity tests were performed on a microplate cytofluorometer (Fluorolite 1000 Dynex Thermobioanalysis; Dynex, Issy-Les-Moulineaux, France) according to European Center for the Validation of Alternative Methods (ECVAM) recommendations. Three cellular indicators were evaluated according to previously validated methods in the Chang cell line ^{5 21} and in other cell systems ^{20 22} : cell viability, cell proliferation, and cell metabolism, including free radical production, mitochondrial activity, and intracellular reduced GSH. ²³  

Cellular Viability Assay.

Briefly, cellular viability was evaluated by using a neutral red (Fluka, Ronkonkoma, NY) test with a fluorometric detection (excitation, 535 nm; emission, 600 nm). Neutral red was used at 50 μg/ml after exposure of cells to the different drugs. ⁵  

ROS Assays.

ROS production was detected with the 2′,7′-dichlorofluorescein diacetate (DCFH-DA; Molecular Probes, Eugene, OR) dye. This probe is a sensitive peroxide-specific fluorescence marker used on cells at 20μ M (excitation 490 nm; emission 535 nm). ^{5 20 22} O;2;;> was detected using hydroethidine (Molecular Probes) at 5μ M (excitation, 485 nm; emission, 600 nm). ^{5 24}  

Mitochondrial Activity Assay.

Furthermore, we investigated mitochondrial function using rhodamine 123 (Molecular Probes), which is a specific cationic fluorescent probe well known in flow cytometry to measure mitochondrial transmembrane potential ²⁵ (mitochondrial activity, Δψm). The probe was used on cells at a final concentration of 5 μg/ml, as in flow cytometry. Rhodamine 123 was incubated with living cells for 30 minutes. After washing, a revealing solution (acetic acid-ethanol) was used before fluorometric detection (excitation, 490 nm; emission, 535 nm). 

Intracellular Reduced GSH Assay.

We also developed a monochlorobimane (Molecular Probes) test. This probe reacts specifically with reduced GSH through GSH transferase to form a fluorescent derivative. ^{20 26} It was used on cells at a final concentration of 100 μM in Ca²⁺- and Mg²⁺-free phosphate-buffered saline. Monochlorobimane was added directly to treated cells for 15 minutes before fluorometric detection (excitation, 360 nm; emission, 480 nm). 

Chromatin Condensation Assay.

Hoechst 33342 (Molecular Probes) is an intercalating dye that permits the determination of the total chromatin quantity variations and the degree of chromatin condensation. ^{27 28} It specifically reacts with the DNA at the levels of adenine and thymidine. ²² This probe was used on cells at a final concentration of 10 μg/ml (excitation, 360 nm; emission, 450 nm). Propidium iodide (Boehringer Mannheim, Mannheim, Germany) at 0.5 mg/ml was added to the Hoechst solution to control necrotic cells (this DNA probe reacts by intercalation and does not allow the Hoechst dye’s fixation by necrotic cells), as previously validated. ⁵  

Statistical Analyses

In all experiments, background fluorescence was determined on wells containing the dye solution but no cells and was deduced from all control and treated cells. All results were obtained in fluorescence units and are expressed as a percentage of the control value. Wells containing cells with complete culture medium but without any treatment were used as the control. Each concentration of drug was tested in six wells, and each experiment was realized in triplicate. Statistical comparisons were performed using a Wilcoxon test at a 0.05 level of significance (Statview IV for Windows; Abacus Concepts, Berkeley, CA). 

Results

Data after 15-minute cell exposure to drugs are depicted in the figures, and the 24-hour recovery data are presented in Table 1 . 

Cellular Viability Evaluation

Significant differences were found with preserved timolol formulations after 15 minutes of treatment, but no alteration of cellular viability was found with unpreserved timolol at any concentration (Fig. 1 ; Table 1 ). Cell viability decreased at 59%, 47%, and 40% of the control with 0.1%, 0.25%, and 0.4% timolol BAC+, respectively (P < 0.001 compared with timolol BAC− and with control for all values). The alteration of membrane integrity was irreversible: after a 24-hour recovery period in a normal medium (Table 1) , the toxicity observed with timolol BAC+ continued to be substantial and even increased, whereas cellular damages did not appear with unpreserved timolol (cellular viability, 103%, 103%, and 100%, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC−; 17%, 14%, and 14%, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC+; P < 0.001 compared with timolol BAC− and with control for all values). 

Similarly, carteolol BAC+ significantly decreased cell viability, whereas unpreserved carteolol did not induce any significant decrease in cell viability (93% of control with 1% carteolol BAC− and 98% with 2% carteolol BAC−; and 61% with 1% carteolol BAC+ and 53% with 2% carteolol BAC+; P < 0.001 compared with carteolol BAC− and control at the two concentrations; Fig. 1 ). After a 24-hour recovery period, carteolol BAC+ also showed alteration of membrane integrity (Table 1) but carteolol BAC− did not show any toxicity (cellular viability, 97% and 98%, respectively, with 1% and 2% carteolol BAC−; 19% and 15%, respectively, with 1% and 2% carteolol BAC+; P < 0.001compared with carteolol BAC− and with control for all values). 

Low concentrations of BAC (0.0001% and 0.001%) had little effect on this conjunctival cell line, but toxicity was found after 15 minutes of treatment at the concentrations of 0.005% and 0.01%, that are, respectively, the concentrations found in carteolol BAC+ and timolol BAC+ (Fig. 1) . Cellular viability significantly decreased after 0.005% and 0.01% BAC treatment with a mean level of 25% of the control (P < 0.001 compared with the control). After a 24-hour recovery period, the toxicity appeared at concentrations of 0.001% and higher, and increased in a concentration-dependent manner (Table 1) . Cellular viability decreased to 68% at 0.001% BAC and 28% at 0.01% (P < 0.001 compared with the control for the four concentrations). 

ROS Production Evaluation

Increased hydroperoxide production was observed with both unpreserved and preserved timolol (Fig. 2) . This production increased up to 195%, 194%, and 180% of the control, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC+ (P < 0.001 compared with the control for all values) and to 153%, 145%, and 125%, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC− (P < 0.001 compared with timolol BAC+ and with control for all values). H₂O₂ production thus decreased with increasing concentrations of timolol. Carteolol BAC+ did not show any significant difference from the control or from unpreserved carteolol that showed no H₂O₂ production and even showed a decrease in H₂O₂ synthesis of approximately 70% at the two concentrations tested (P < 0.05 compared with the control). Significant H₂O₂ production was also observed with BAC used at concentrations ranging between 0.0001% and 0.01%. The maximum was observed with 0.005% BAC with H₂O₂ increased production at 215%; the minimum was 145% with 0.01% BAC (P < 0.001 for all concentrations tested compared with the control). This phenomenon was most likely caused by the decrease of cell viability at the highest concentration of BAC. 

O^⨪ ₂ Production Evaluation

Timolol BAC+ at concentrations of 0.1%, 0.25%, and 0.4% induced increases in O^⨪ ₂ production to 145%, 151%, and 118% of the control, respectively ((Fig. 3 ;P < 0.001 compared with the control for all values), whereas the increases observed with 0.1%, 0.25%, and 0.4% unpreserved timolol were only 106%, 113%, and 108% (P = 0.05 for 0.25% timolol BAC− compared with the control; P < 0.001 compared with timolol BAC+ for all values). O^⨪ ₂ production was also observed with unpreserved and preserved carteolol. This production increased to 150% and 164% with carteolol 1% and 2% BAC+, and to 114% and 123% with 1% and 2% carteolol BAC− (P < 0.001 for all unpreserved and preserved carteolol concentrations compared with the control, but P < 0.001 when carteolol BAC+ and BAC− were compared for all values). BAC alone induced O^⨪ ₂ synthesis from 123% with 0.005% BAC (P < 0.01 compared with the control) to 148% with 0.008% BAC (P < 0.001 compared with the control). 

Mitochondrial Activity Evaluation

Mitochondrial activity (Fig. 4 ; Table 1 ) profoundly decreased with timolol BAC+ at all concentrations (35%, 38%, and 37%, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC+), whereas this marker was unchanged with timolol BAC− after 15 minutes at the lowest concentration (98% for 0.1% timolol BAC−) and slightly but significantly decreased at 0.25% (84%; P = 0.05) and 0.4% (76%; P = 0.01). After 15 minutes of treatment and 24 hours of recovery (Table 1) , a more important decrease in mitochondrial activity was observed with timolol BAC+ (mitochondrial activity decreased to 35%, 28%, and 28% of the control, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC+) and unpreserved timolol (mitochondrial activity decreased to 90%, 72%, and 78% of control, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC−; P < 0.001 compared with timolol BAC+). Similar results were found with carteolol. Concentrations of 1% and 2% carteolol BAC+ induced a significant decrease in mitochondrial activity (mitochondrial activity, 58% at the two concentrations; P < 0.001 compared with the control) after 15 minutes, whereas with carteolol BAC−, this decrease occurred only at the concentration of 2% (mitochondrial activity, 80% of control; P < 0.01 compared with the control). After 24 hours of recovery, the decrease observed with 1% and 2% carteolol BAC+ reached 32% at the two concentrations (P < 0.001 compared with the control; Table 1 ). No significant decrease was observed with 1% carteolol BAC−, and a slight effect was found with 2% carteolol BAC− (84%; P < 0.05 compared with the control for this concentration). Mitochondrial activity also decreased from 58% to 42% with BAC at concentrations of 0.0001% and 0.01%, respectively (P < 0.001 for all points, compared with the control) after 15 minutes. After a 24-hour recovery period (Table 1) , a significant alteration of mitochondrial activity was found at concentrations of 0.005% BAC and higher. This decrease reached 35% to 37% with 0.005% to 0.01% BAC (P < 0.001 compared with the control), and nonsignificantly decreased at lower concentrations. 

Intracellular Reduced GSH Evaluation

A slight but significant decrease of intracellular reduced GSH (Fig. 5 ; Table 1 ) was observed with both timolol BAC+ (GSH decrease was 8%, 9%, and 10% of the control, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC+) and timolol BAC− (GSH decrease was 5%, 9%, and 7% of the control, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC−) after 15 minutes of treatment. After recovery in normal medium, the GSH level was still decreased except for 0.1% and 0.25% timolol BAC−. No modification of intracellular reduced GSH was found with carteolol BAC− after 15 minutes of treatment, but a significant decrease was observed after the 24-hour recovery period (GSH decrease, 8%; P < 0.001 compared with the control). A significant decrease appeared with preserved carteolol (20% and 28%, respectively, with 1% and 2% carteolol BAC+; P < 0.001 compared with the control) after 15 minutes of treatment and lessened, but was still significant, after 24 hours of recovery (the GSH decrease was 8% at the two concentrations tested; P < 0.001 compared with the control). In contrast, GSH was significantly increased with BAC alone after 15 minutes of treatment (from 8% with 0.0001% BAC to 25% with 0.01% BAC) but a decrease (from 2% with 0.0001% BAC to 9% with 0.008%) was observed after the recovery period. 

DNA Condensation Evaluation

Chromatin condensation, studied with the Hoechst test, revealed for timolol BAC+ a concentration-dependent increase in the Hoechst dye’s fluorescence at 109%, 124%, and 149%, respectively, with 0.1%, 0.25%, and 0.4% timolol BAC+ (P < 0.001 compared with the control for all tested concentrations) after 15 minutes of treatment, whereas unpreserved timolol did not induce any difference from the control (Fig. 6 ; Table 1 ). After a 24-hour recovery in normal medium, a decrease of 64%, 58%, and 58% of the control, with 0.1%, 0.25%, and 0.4% timolol BAC+, respectively (P < 0.001 compared with the control for all values), was observed (Table 1) , and a slight decrease was found with 0.1%, 0.25%, and 0.4% timolol BAC−, respectively, of 98%, 95%, and 88% (P = 0.05 for 0.4% timolol BAC− compared with the control; P < 0.001 compared with timolol BAC+ for all values). DNA condensation increased with 1% and 2% carteolol BAC+ (to 122% and 132%, respectively; P < 0.001 compared with the control) and a significant decrease appeared after a 24-hour recovery period (59% with both 1% and 2% carteolol BAC+; P < 0.001 compared with the control). Carteolol BAC−, as with timolol BAC−, did not show any significant difference from the control after 15 minutes of treatment, but a slight decrease was observed after 24 hours (mean fluorescence, 86% at 1% and 2%; P = 0.05 compared with the control; Table 1 ). 

BAC showed the same phenomenon with a concentration-dependent increase in fluorescence after 15 minutes of treatment, ranging from 102% with 0.0001% BAC to 304% with 0.01% BAC (P < 0.001 compared with the control at 0.005% and higher; Fig. 6 ). After a 24-hour recovery period, the fluorescence levels varied from 104% with 0.0001% BAC to 287% with 0.01% BAC (P < 0.001 compared with the control at 0.001% and higher; Table 1 ). 

Discussion

A wide spectrum of reactions has been extensively described after administration of antiglaucoma drugs, from infraclinical conjunctival inflammation and subconjunctival fibrosis to acute sight-threatening responses, including a rare but highly severe complication, pseudopemphigoid. ^{29 30} Other less severe adverse effects may currently be observed in patients with glaucoma, such as allergic conjunctivitis, punctate epithelial erosions, hyperemia, conjunctival folliculitis, and disturbances of the tear film. ^{31 32 33} Recent studies in humans have confirmed that the removal of BAC from a timolol preparation restores permeability and metabolism of corneal epithelium without reducing the drug’s efficiency in controlling IOP. ³⁴ Clinical tolerance of preserved and unpreserved antiglaucoma eye drops was also recently evaluated. Patient-reported adverse reactions and objective damage to the ocular surface were more frequent among patients treated with preserved formulations. ³⁵  

In in vivo models, BAC was shown to affect surface microvilli in rabbit and cat corneas ^{32 36} and to elicit an inflammatory response in rat conjunctiva. ^{37 38} In vitro, we recently showed that BAC was a proapoptotic agent in Chang conjunctival cells. ^{5 10 39} Cells die by necrosis after BAC treatment at high concentrations and by apoptosis at low concentrations. 

In the present study, we sought to investigate the toxicity of two widely used preserved and preservative-free β-blockers to confirm the toxicity of preserved solutions and to better understand the roles of superoxide anion, mitochondria, and reduced GSH in apoptosis. A cytotoxicity assay based on membrane integrity raised the possibility of irreversible and global cellular damage 15 minutes after application of BAC at the concentration found in eye drops. The damage was associated with a markedly high production of H₂O₂ and O^⨪ ₂ at BAC concentrations ranging from 0.0001% to 0.01%. Preservative-free and preserved β-blockers also induced significant H₂O₂ production, whereas no alteration of cell viability was observed with unpreservedβ -blockers. Significantly higher O^⨪ ₂ production was also found with preserved formulations, when compared with the unpreserved ones. 

Moreover, we recently showed that O^⨪ ₂ plays a role in the decrease of membrane integrity and in the increase of chromatin condensation observed with quaternary ammonium ions, suggesting that O^⨪ ₂ may induce apoptosis or at least participate in epithelial cell degeneration. ⁵ We thus confirmed the predominant role of O^⨪ ₂, rather than H₂O₂, in the decrease of membrane integrity observed with preserved β-blockers. This ROS production was shown to be harmful to the cells. ROS production is inhibited by various antioxidant defenses such as GSH, ascorbate, uric acid or plasma proteins. The tripeptide GSH is the most rapid and abundant weapon against ROS and regulates the redox state of many other cellular substances. ⁴⁰ After 15 minutes of treatment, BAC increased content of reduced GSH to detoxify the cell. After a recovery period, a decrease of reduced GSH was found with BAC and with allβ -blocker formulations (except 0.1% and 0.25% timolol BAC−), possibly due to inhibition of synthesis, the oxidation of GSH into GSH disulfide (GSSG), or a deficit in the GSH salvage pathway, but more probably due to an increased rate of GSH efflux. However, the actual contribution of reduced cellular GSH in initiating conjunctival cell apoptosis remains difficult to discern. Buthionine sulfoximine, well known to reduce cellular GSH levels, does not cause apoptosis in many cell types, indicating that GSH depletion alone may not trigger apoptosis. ⁴¹  

Data obtained from the rhodamine 123 assay clearly showed that the mitochondrial function of cells was affected by BAC and by preserved solutions of timolol and carteolol, after only 15 minutes. Timolol BAC− also altered mitochondrial activity, but this effect was always less important than with timolol BAC+. After 24 hours of recovery in a normal culture medium, no cell recuperation was observed, and transmembrane potential decreased with increasing concentrations of the drug. The toxicity of timolol BAC+ or carteolol BAC+ could therefore be explained by an apoptotic phenomenon. Several studies have shown that an early event of apoptosis was a decrease in mitochondrial transmembrane potential that was reflected by a loss of the cells’ ability to accumulate rhodamine 123. ^{42 43} All presently available data were compatible with the notion that mitochondrial transmembrane potential disruption marks a point of convergence of distinct apoptosis induction pathways and constitutes the point of no return of apoptosis. Before cells exhibited common signs of nuclear apoptosis, they underwent a reduction of the mitochondrial transmembrane potential that may have been due to the opening of mitochondrial permeability transition pores. ⁴⁴  

This event was associated with increased production of superoxide anions—mitochondria being the organelle where ROS is primarily produced. One conceivable scenario is that ammonium ions disturb cytoplasmic membrane, which may alter the calcium flux, leading to energy deficiency in mitochondria. Furthermore, the decrease in mitochondrial activity may cause the cytosolic efflux of cytochrome c where it binds to Apaf-1 and activated caspases. ^{3 45 46 47} Our results are also consistent with a recent study performed in serum-deprived fibroblast cultures that showed that oxidative stress is an early event in apoptosis. Mitochondrial DNA was damaged, and mitochondria significantly contributed to the oxidative stress associated with apoptosis by increasing the generation of peroxides. ⁴⁸ The consequences of mitochondrial dysfunction (collapse of the mitochondrial transmembrane potential, uncoupling of the respiratory chain, hyperproduction of superoxide anions, disruption of mitochondrial biogenesis, outflow of matrix calcium and GSH, and release of soluble intermembrane proteins) can entail a bioenergetic disturbance culminating in the disruption of plasma membrane integrity (necrosis) and/or the activation and action of apoptogenic proteases with secondary endonuclease activation and consequent oligonucleosomal DNA fragmentation (apoptosis). 

According to ECVAM recommendations, cellular DNA was evaluated using the Hoechst test. Hoechst 33342, a DNA fluorochrome, unlike propidium iodide, is not excluded by live or apoptotic cells. Brief exposure of cells to low concentrations of Hoechst 33342 leads to strong labeling of apoptotic cells. ⁴² In contrast, live cells, require much longer incubation with Hoechst to obtain a comparable intensity of fluorescence. Supravital uptake of Hoechst combined with exclusion of propidium iodide (to identify necrotic and late apoptotic cells) was proposed as an assay of apoptosis. ²⁷ The apoptotic phenomenon was confirmed with BAC and also with timolol BAC+ and carteolol BAC+ after 15 minutes. In fact, the fluorescence ratio (Hoechst-neutral red ratio) significantly increased in response to these drugs (Table 2) . Whereas the cytotoxicity was significant (decrease in membrane integrity with BAC, timolol BAC+, and carteolol BAC+), we found an increase in chromatin condensation related to an apoptotic mechanism after only 15 minutes. After a 24-hour recovery period, apoptosis observed with preserved β-blockers was followed by secondary necrosis, as suggested by the decrease of both cell viability and chromatin condensation. This is consistent with previous studies that show that at different concentrations, BAC can induce two distinct patterns of cell death: At low levels of cellular damage, the cells still have time to activate programmed apoptosis, whereas at higher levels of insult, the injury to the cells is too severe to allow a programmed mode of death, and cell therefore die by direct necrosis. ¹⁰  

BAC facilitated the course of programmed cell death and may have been partially responsible for the toxicity of preserved β-blockers, whereas unpreserved formulations produced very moderate cytotoxic changes. Thus, the apoptotic phenomenon observed in vitro on conjunctival cells after BAC or preserved β-blocker treatments is in agreement with in vivo observations of surface conjunctival epithelial cells collected by impression cytology. Fas antigen, a membrane receptor that plays a major role in induction of apoptosis, was found by flow cytometry in a significantly higher percentage of conjunctival cells in patients receiving topical antiglaucoma treatment than in normal eyes. Fas ligand and Apo 2.7 were also overexpressed by conjunctival cells in pathologic eyes. ⁴⁹  

A limitation in our conclusions concerns the conjunctival cell line in our model. These cells present some characteristics of conjunctival epithelium (e.g., desmosomes, microvilli, and absence of expression of HLA-DR) but features are not sufficient to directly extrapolate our findings to disease of the ocular surface in humans. In fact, the Chang epithelium is an immortalized monolayer from only one type of cell in which there is no tear film or well-defined mucus. Therefore, our model remains an experimental approach, even though it can serve as a basis for other in vivo or clinical studies, with which it is quite consistent. 

In our model, apoptosis can be identified with better assurance when more than a single viability assay is used. Thus, simultaneous assessment of membrane integrity together with chromatin condensation, mitochondrial transmembrane potential, ROS production, and reduced GSH evaluation offers a more certain means of identification of cell death than each of these methods alone. In the present study, the cytotoxicity of BAC, carteolol BAC+, and timolol BAC+ was clearly shown to be due to an apoptotic mechanism with alteration of mitochondrial activity and generation of ROS, rather than to variations in reduced cellular GSH. The observation of both ROS production and apoptosis on Chang conjunctival cells cultures after short-term incubation with preserved β-blockers and BAC should lead to a revision in the management of long-term ophthalmic treatments. Development of preservative-free solutions as well as pharmacologic research on antioxidant drugs could be promising in the near future for chronic diseases such as dry eye syndrome, allergy, and glaucoma. 

 Submitted for publication January 23, 2001; revised April 25, 2001; accepted June 5, 2001.

 Commercial relationships policy: N.

 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.

 

Tests Conducted In Vitro to determine the Toxicity of BAC, Unpreserved and Preserved β-blockers

Drug (%)	Membrane Integrity Evaluation	Mitochondrial Activity Evaluation	Intracellular Reduced Glutathione Evaluation	Chromatin Condensation Evaluation
Timolol BAC−
0	100 ± 3	100 ± 14	100 ± 7	100 ± 6
0.10	103 ± 4 NS	90 ± 10 NS	98 ± 2 NS	98 ± 5 NS
0.25	103 ± 4 NS	72 ± 11*	99 ± 1 NS	95 ± 7 NS
0.40	100 ± 3 NS	78 ± 10, †	90 ± 1, ‡	88 ± 6, †
Timolol BAC+
0	100 ± 3	100 ± 14	100 ± 8	100 ± 6
0.10	17 ± 11, ‡	35 ± 9, ‡	90 ± 1, ‡	64 ± 5, ‡
0.25	14 ± 6, ‡	28 ± 8, ‡	91 ± 1, ‡	58 ± 2, ‡
0.40	14 ± 5, ‡	28 ± 9, ‡	92 ± 1, ‡	58 ± 4, ‡
Carteolol BAC−
0	100 ± 7	100 ± 10	100 ± 5	100 ± 3
1	97 ± 9 NS	85 ± 14 NS	92 ± 11, ‡	86 ± 10, †
2	98 ± 3 NS	84 ± 9, †	91 ± 1, ‡	86 ± 12, †
Carteolol BAC+
0	100 ± 7	100 ± 10	100 ± 9	100 ± 3
1	19 ± 9, ‡	33 ± 11, ‡	92 ± 12, ‡	59 ± 10, ‡
2	15 ± 5, ‡	32 ± 12, ‡	92 ± 2, ‡	59 ± 7, ‡
BAC
0	100 ± 6	100 ± 10	100 ± 7	100 ± 3
0.0001	100 ± 8 NS	97 ± 11 NS	98 ± 1 NS	104 ± 5 NS
0.001	68 ± 6, ‡	84 ± 11 NS	91 ± 1, ‡	155 ± 3, ‡
0.005	27 ± 13, ‡	37 ± 14, ‡	91 ± 1, ‡	241 ± 5, ‡
0.008	28 ± 10, ‡	35 ± 9, ‡	91 ± 1, ‡	257 ± 10, ‡
0.01	28 ± 9, ‡	35 ± 10, ‡	93 ± 1, ‡	287 ± 6, ‡

 Membrane integrity (neutral red test), mitochondrial activity (rhodamine 123 test), intracellular reduced glutathione (monochlorobimane test), and chromatin condensation (Hoechst 33342 test) evaluations (mean fluorescence: %/control ± SD) after 15 minutes of treatment and 24 hours of cell recovery. Probabilities are compared with results in control experiments.

*  P < 0.01.

^†   P < 0.05.

^‡   P < 0.001.

Figure 1.

 

Cellular viability (neutral red test) after 15 minutes of treatment with timolol, carteolol, or BAC. Results are expressed as percentages of control values. Timolol BAC−, unpreserved timolol; timolol BAC+, preserved timolol with a BAC concentration of 0.01% (for the three concentrations tested); carteolol BAC−, unpreserved carteolol; carteolol BAC+, preserved carteolol with a BAC concentration of 0.005% (for the two concentrations tested). ***P < 0.001 compared with control.

Figure 2.

 

ROS production (DCFH-DA test) after 15 minutes of treatment with timolol, carteolol, or BAC. See Figure 1 for description of formulations tested. *P < 0.05;*** P < 0.001 compared with control.

Figure 3.

 

O^⨪ ₂ production (hydroethidine test) after 15 minutes of treatment with timolol, carteolol, or BAC. See Figure 1 for description of formulations tested. *P < 0.05;*** P < 0.001 compared with control.

Figure 4.

 

Mitochondrial activity (rhodamine 123 test) after 15 minutes of treatment with timolol, carteolol, or BAC. See Figure 1 for description of formulations tested. *P < 0.05;** P < 0.01; ***P < 0.001 compared with control.

Figure 5.

 

Intracellular reduced GSH evaluation (monochlorobimane test) after 15 minutes of treatment with timolol, carteolol, or BAC. See Figure 1 for description of formulations tested. *P < 0.05;** P < 0.01; ***P < 0.001 compared with control.

Figure 6.

 

DNA condensation evaluation (Hoechst 33342 test) after 15 minutes of treatment with timolol, carteolol, or BAC. See Figure 1 for description of formulations tested. ***P < 0.001 compared with control.

Table 2.

 

Hoechst-Neutral Red Fluorescence Ratios in the Different Treatment Groups After a 15-Minute Exposure

Drug Concentrations (%)	BAC−	BAC	BAC+
	Timolol		Timolol
0.1	1.00		1.85
0.25	0.88		2.64
0.4	0.85		3.73
	Carteolol		Carteolol
1	0.82		2.02
2	0.87		2.47
0.0001		1.30
0.001		1.49
0.005		12.32
0.008		9.62
0.01		9.48

 Apoptosis is present at a Hoechst-Neutral red ratio > 1.