Abstract
Purpose.:
To evaluate and compare the effects of the following dyes on human pigmented epithelial cells: indocyanine green (ICG), infracyanine green (IfCG), trypan blue (TB), bromophenol blue (BrB), patent blue (PB), and Brilliant Blue G (BBG).
Methods.:
ARPE-19 cells cultured in vitro were exposed to these dyes, and acute and chronic toxicity were evaluated. Cell viability was measured by colorimetry (MTT assay), morphology was observed by phase-contrast microscopy, membrane permeability (CMP) was evaluated by flow cytometry with propidium iodide (PI), and mitochondrial membrane potential (ΔΨm) was measured with 3,3′-dihexyloxacarbocyanine (DiOC6(3)).
Results.:
Each of the studied dyes exhibited toxicity after acute exposure at surgical doses. The presence of light often reduced cell viability, especially when measured 3 hours after incubation in the case of ICG, TB, BrB, and BBG. Morphologic changes were induced by ICG, IfCG, and BBG. Both CMP and ΔΨm were altered after exposure to surgical doses of ICG, TB, PB, and a fourfold surgical dose of BrB. Chronic exposure to residual amounts of some dyes was associated with reduced proliferation and even cell death.
Conclusions.:
It appears to be prudent to use the lowest possible dose of each dye, to minimize the risk of toxic effects. This precaution may be particularly important in the case of BrB, which should not be used in excess of 0.5%. In addition, abundant irrigation of the vitreous cavity after surgery to completely remove traces of dye may be of crucial importance, particularly in the case of ICG, in minimizing chronic toxicity.
One of the most important innovations in vitreoretinal surgery over the past 10 years has been the introduction of vital dyes to improve the visualization of preretinal tissues and membranes.
1 –4 Staining these structures facilitates the peeling of the internal limiting membrane (ILM) of the retina during vitrectomy and reduces surgical risks.
5 –9
In 2000, indocyanine green (ICG) became the first dye to be used to stain the ILM, and it is currently the most commonly used surgical dye in ophthalmology. Nevertheless, despite its routine use, it has generated controversial discussion over the past decade. ICG has been found to be associated with the risk of retinal damage,
10 –13 atrophy of the retinal pigment epithelium (RPE),
14 damage to the photoreceptors and RPE cells,
15,16 lower visual function outcome,
17 –22 loss of epiretinal cellular integrity,
23 and cellular toxicity,
24 –29 among other harmful effects. Recently, several groups have reported that ICG may persist in the ocular cavity, even 6 weeks after its application in surgery.
30 –32 Therefore, the need for the investigation of alternative dyes for vitrectomy has become apparent. Since then, several alternative stains have been introduced into vitreoretinal surgery and their number is constantly growing.
33 –37 However, it is not clear which dye would be ideal in terms of reduced toxicity, higher affinity, and minimal residual permanence.
Infracyanine Green (IfCG) is a dye with a chemical formula and pharmacologic properties similar to those of ICG. However, IfCG is synthesized without sodium iodine, which seems to represent a clear advantage, because it is believed that iodine damages the cornea and retina.
38 –40 On the other hand, iodide-free IfCG is not water soluble and has to be dissolved in a 5% glucose solvent.
41 –43
Trypan blue (TB) is used in microscopy for staining dead tissues. In ophthalmology, TB has preferential affinity for the epiretinal membrane (ERM).
44 –48 Although it may not enable ILM visualization as well as ICG, it does exhibit better biocompatibility.
49 –52
Bromophenol blue (BrB) is an acid-base indicator that has recently been proposed as a promising alternative biostain for vitrectomy, because it stains the ERM and ILM and has induced no damage in either in vitro or in vivo studies.
35,53,54
Patent blue (PB) exhibits moderate affinity for the ERM and low affinity for the ILM.
54 Little is known about the possible side effects of PB. Some toxic effects have been reported, but the data are conflicting.
55 –57
Brilliant blue G (BBG), or Coomassie blue, is commonly used for protein determination and gel electrophoresis.
58 –60 In 2006, this dye was reported to stain the ILM and to have no significant in vivo toxicity.
61 –64
The purpose of the present study was to compare the in vitro effect on human RPE cells of a panel of the dyes most commonly used in vitrectomy. The effects of several concentrations of the dyes were examined, and the same experimental conditions were used for each of the six dyes, thus facilitating a more precise comparison of the specific effects of each dye.
The well-characterized human retinal pigment epithelial cell line ARPE-19
65 was purchased from the American Type Culture Collection (Manassas, VA). The cells were maintained in a 1:1 mixture of Dulbecco's modified Eagle's medium (BE12-709F; Lonza Ibérica, Barcelona, Spain) and Ham's F-12 (FG-0815; Biochrom AG, Berlin, Germany), supplemented with 10% fetal bovine serum (FBS; Lonza Ibérica), 2 mM
l-glutamine (BE17-605E, Lonza Ibérica), and 1% penicillin-streptomycin (DE17-602E; Lonza Ibérica) in a humidified 5% CO
2 atmosphere at 37°C. The cells were detached with 0.5% trypsin-0.2% ethylene-diamine-tetraacetic acid (EDTA; T4174; Sigma-Aldrich, St. Louis, MO) and subcultivated (1:3–1:4) twice a week.
First, we studied the proliferation characteristics of ARPE-19 cells to determine the appropriate plating density for each type of assay. For acute toxicity assays, the number of cells per well must be enough to allow detection (i.e., above the limit of detection of the assay method) in the event that cell death occurs. In contrast, for chronic toxicity assays, the cells should be seeded at a density that allows them to proliferate without experiencing contact inhibition forces due to confluence. To this end, we characterized the proliferation of cells seeded at a variety of initial concentrations (cells/well) in 96-well plates (data not shown). On the basis of these studies, we found that it was optimal to seed 104 cells/well for acute toxicity assays (subconfluent) and 3000 cells/well for chronic toxicity assays, since these densities allow the assays to be performed at least 72 hours later under nonconfluent conditions.
To evaluate the acute toxicity associated with vital dyes, we reproduced, as much as possible, the conditions used during vitreoretinal surgery. Thus, the ARPE-19 cells were seeded at 104 cells/well in 96-well plates and left to attach to the substrate until the next day. Then, the cells were exposed to 50 μL of each of the just-described solutions for 3 minutes. Afterward, the dyes were removed, and the cells were washed three times with DMEM-F12 and reincubated with fresh culture medium. Cell viability was measured 1.5, 3, and 24 hours later. All experiments were performed in quadruplicate and repeated three times. Results are expressed as the mean percentage ± SD of viable cells with respect to the control cells incubated without dye.
Since the cytotoxicity of dyes may be photosensitive, we also performed these assays in the presence of a maximum-power 35-W halogen xenon light with 520 lumens. This light source was placed at a distance of 10 mm from the cell cultures.
Residual dye can persist in the vitreous cavity after surgery if washing is not thoroughly performed. To emulate this condition, we evaluated the chronic toxicity of high dilutions of dye. Cells were seeded at 3000 cells/well in 96-well plates and left to attach until the next day. Then, they were incubated with 200 μL of a dye dilution, which was 1:500, 1:1000, or 1:2000 of the standard surgical concentration of each dye for 0, 24, 48, and 72 hours. Cell viability was subsequently evaluated. All experiments were performed in quadruplicate and repeated three times. The results are expressed as the mean proliferation rate ± SD of viable cells with respect to the control cells at the start of incubation (t = 0 hours).
Cell viability was assessed by using 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT; Sigma-Aldrich). MTT is reduced by mitochondrial and cytosolic dehydrogenases in living cells to a purple formazan dye that is spectrophotometrically detectable. After exposure to the dye, the cells were incubated at 37°C with 0.5 mg/mL MTT in culture medium. After 3 hours, the MTT solution was removed, and 100 μL/well dimethylsulfoxide was added. Optical densities were determined at 540 nm with a microplate reader (ELx800 Microplate Reader; BioTek Instruments, Winooski, VT).
IfCG can interfere with the measurement of formazan. This artifact was reduced by measuring the optical density immediately after MTT was added. The results reflected the effect of IfCG before the formazan reaction took place in the cells. These results were subtracted from the optical densities measured after 3 hours' incubation with MTT.
Phase-contrast micrographs were taken to illustrate changes in cell morphology after acute dye exposure. Cells were seeded at 2 × 105 cells/well into 12-well plates and allowed to adhere overnight. Then, the ells were exposed to 250 μL dye solution for 3 minutes, at a final concentration per well equivalent to surgical concentration. After 1.5 hours, cell morphology was observed with a phase-contrast microscope (Eclipse TS 100; Nikon, Tokyo, Japan) at a final magnification of ×100, and images were captured (ProgRes CapturePro, ver. 2.6; Imaging Planet, Goleta, CA).
Changes in cell membrane permeability (CMP) were evaluated by propidium iodide (PI) incorporation, whereas alterations to mitochondrial membrane potential (ΔΨm) were measured by 3,3′-dihexyloxacarbocyanine iodide (DiOC6(3)) labeling (Invitrogen-Molecular Probes, Eugene, OR). PI stains nuclear DNA and cytoplasmic RNA, but penetrates only the cells that have a permeable cell membrane, which is a manifestation cells as they die. In contrast, ΔΨm decreases before mitochondrial death, and DiOC6(3) is a ΔΨm-sensitive dye.
At 1.5, 3, and 24 hours after exposure to the surgical concentration of dye in the presence and absence of a halogen xenon light, the cells were incubated with 100 nM DiOC6(3) in culture medium for 20 minutes at 37°C. In this case, we also seeded 2 × 105 cells/well in 12-well plates and exposed them to 250-μL dye solutions to induce cytotoxicity. The cells were then washed and resuspended in fresh medium with 5 μg/mL PI. Fluorescence emission of at least 10,000 viable cells was analyzed by flow cytometry (Epics Elite ESP; Beckman Coulter Corp., Brea, CA).
ARPE-19 cells typically grow in stable monolayers as flattened cells, exhibiting a hexagonal shape, epithelial-like morphology, and functional polarity. These cells, like their in vivo counterparts, form tight junctions with transepithelial resistance. However, 1.5 hours after exposure to surgical doses of some dyes, the cells exhibited an altered morphology (
Fig. 3). We assayed only surgical concentrations of the dyes, because we had not observed significant differences between the three concentrations used in acute toxicity assays, with the exception of BrB.
In
Figure 3, we can see that ∼50% of the ARPE-19 cells assumed a more rounded shape after exposure to ICG. Cell monolayers exposed to IfCG retained a green film, even after they were rinsed with PBS. The only way to remove this film was by rinsing with the vehicle provided by the manufacturer. These cells presented alterations in their shape and organization. It appeared that exposure to IfCG altered the previous epithelial-like morphology, with cell groups forming a palisade, indicative of cell contact alterations.
In contrast, TB, BrB, and PB seemed to cause no observable changes in the morphology of the ARPE19 cells. Cell density appeared to be lower after exposure to TB and PB than in control cells, indicative of reduced proliferation. However, the cells that were incubated with a higher concentration (1%) of BrB showed significant morphologic changes. Thus, cells shrank, acquired a more rounded shape, detached from the plates, and exhibited a significant decrease in cell survival.
Finally, BBG-treated cells appeared to undergo mild morphologic changes, but their shape was not as evidently rounded as that of ICG-treated cells. However, some groups of cells were reminiscent of the palisade formed by the IfCG-treated cells.
Since the introduction of ICG in vitrectomy, it has been widely accepted that vital dyes represent useful agents to improve visualization and peeling of the ILM. However, over the past few years, controversial evidence has accumulated indicating that these dyes may have harmful effects. The discrepancies among these reports are due principally to the fact that each study used different manufacturers, conditions, dilution procedures, exposure times, illumination, and cell models, among other differences.
In this study, we ruled out these variables as possible causes for variations in results, by examining six vital dyes at the same time and under the same conditions. This experimental design allowed us to focus on the specific biological effects associated with each of the dyes and to determine which dye may be more appropriate in a given circumstance, on the basis of the characteristic effects and risks associated with each dye. The experimental conditions that we used mimic surgical conditions as closely as possible: We used short exposure times (3 minutes) and dye concentrations that are typically administered to patients during surgery. We also examined the effects of concentrations that were higher and lower than those routinely used in surgery, to determinate not only cell viability but also cell morphology and functional status. Finally, we assayed the effects of chronic exposure to highly diluted dyes to simulate the effect on RPE cells of residues of dye persisting in the vitreous cavity after surgery.
BrB has not been used widely as a dye in ophthalmology. The few available studies have not revealed any toxic effects in vitro,
35 or in the retina or lens at low concentrations (0.2%). However, at this concentration, the dye does not stain very intensely.
53,54 In contrast, we found that concentrations ≤0.25% (concentrations used in the clinical setting) led to reduce RPE cell viability. Moreover, this reduction was enhanced significantly in the presence of light (
Fig. 2), but surprisingly, was not accompanied by visible changes in culture morphology. This finding suggests that the observed effects of the BrB dye may be due to reduced MTT metabolism, consistent with our data regarding the status of the membranes (
Table 2).
The situation is radically different when higher concentrations of BrB were used to increase the intensity of the staining. Thus, after exposure to 1% BrB, massive cell death was found and there were no signs of recuperation according to microscopy and flow cytometry. The mean PI fluorescence intensity initially revealed two distinct populations; later, one of these subpopulations disappeared, suggesting a cellular explosion that is typical of necrosis. This subpopulation may represent the cells that had exhibited reduced viability. It is likely that these cells died by necrosis, since no changes were observed in DiOC6(3), a marker of apoptotic processes. For these reasons, we highlight the importance of careful management of concentrations of this dye in particular, since high concentrations may have significant toxic effects.
As occurred with other dyes, small traces of BrB were found to delay the rhythm of cell growth. It would be interesting to evaluate whether BrB induces arrest of the cell cycle at G0–G1 via increased expression of p21.
Useful applications of BBG in the area of ophthalmology were discovered not long ago.
61 Since then, its security has been demonstrated in both in vitro and in vivo studies.
61 –63 However, a decrease in cell viability, similar to that which we observed, with the same exposure time (3 minutes) and with similar concentrations, has been reported.
69 This reduction has been attributed to a cytostatic effect,
67 since it has been shown that BBG acts as a P2X7 receptor antagonist, leading to reduced cell growth.
74 This feature may contribute to a beneficial postoperative effect by reducing the formation of fibrous material.
In other studies of BBG, the influence of light on BBG's effects was not examined. We found BBG to be mildly phototoxic and to induce variations in the morphologic aspect of cultures. However, the cells appeared to be healthy, and no significant changes were observed by flow cytometry (
Table 2). Thus, our findings corroborate the reports that BBG does not lead to apoptosis or necrosis and provide further support for the idea that reduced cell viability, as measured in MTT assays, represents a delay in cell growth rather than the induction of death.
In contrast to other studies of the chronic toxicity of BBG, we found only a slight reduction in growth rate associated with this dye.
69 The reduced effect could have occurred because BBG is more hydrosoluble than ICG and IfCG; it would thus penetrate less into the cells and be more easily washed away, leaving less residues after surgery.
Results obtained in vitro cannot necessarily be extrapolated to in vivo clinical situations. Nevertheless, we consider that the in vitro finding that all dyes have the capacity to induced at least growth arrest, strongly suggests that caution should be exercised when using these dyes in the clinical setting. Residual amounts of dye can be retained in the vitreous cavity, even after its irrigation and this can lead to chronic cytotoxic effects, especially with low water-soluble dyes (i.e., hydrophobic dyes such as ICG). Thus, the vitreous cavity should be copiously and thoroughly irrigated after staining, to minimize the risk of chronic cytotoxicity.
Finally, our results suggest that the dyes routinely used in ophthalmic surgery may exert toxic effects, not only by altering cell viability, but also by perturbing cell morphology and membrane status, since the dyes were found to lead to reduced or altered functional and metabolic capacity, with substantial consequences for cell behavior in vitro. In light of these studies, we consider it prudent to reduce the duration of exposure to these six dyes during surgery to the extent possible and thus minimize the risk of inducing toxic effects.
Disclosure:
M.-C. Morales, None;
V. Freire, None;
A. Asumendi, None;
J. Araiz, None;
I. Herrera, None;
G. Castiella, None;
I. Corcóstegui, None;
G. Corcóstegui, None