Safety Testing of Indocyanine Green and Trypan Blue Using Retinal Pigment Epithelium and Glial Cell Cultures

Abstract

purpose. Indocyanine green (ICG) and trypan blue have been advocated as vital stains for use during macular surgery. The safety of these agents was tested using a cell culture model.

methods. Human retinal pigment epithelium (RPE) and Müller cell lines were exposed to ICG over a range of concentrations up to 0.5%, and trypan blue up to 0.2%. Cells were exposed to each dye for 5, 15, or 30 minutes, rinsed, and incubated 24 hours. Cell viability was measured using a mitochondrial dehydrogenase-assay and fluorescent live–dead probe. Experiments were repeated using 0.5% and 1% ICG and 0.06% and 0.12% trypan blue, with follow-up at 0, 1, 5, and 15 days. ICG experiments were repeated in the presence of illumination from a xenon light-source channeled through a surgical endolight, and using reduced osmolarity solutions of 0.1%, 0.5%, and 1% (185 vs. 275 mOsM).

results. There was no clear relationship between cell viability and the concentration of the agent or duration of follow-up, except in RPE cells exposed to 1% ICG. These showed a linear (R ² 0.9952) decline in viability with time, with a significant reduction by day 15 (P = 0.016). RPE cells exposed to ICG and illumination were not significantly different from the negative control, but when illumination was combined with low osmolarity, viability was reduced (P = 0.0016). ICG and illumination reduced Müller cell viability (P < 0.0001 for both 185 and 275 mOsM). Müller cells incubated with 185 mOsM 1% ICG showed a significant reduction in viability (P < 0.0001) not seen with the 185 mOsM 0.5% or 0.1% solutions or in the low-osmolarity RPE groups.

conclusions. The combination of exposure to 0.5% ICG and the newer endoillumination light-sources can damage cultured Müller cells. Although the preparations of ICG most commonly used clinically did not produce significant damage, relatively small changes in ICG osmolarity and concentration did. This suggests that safety margins are not large. Trypan blue is safe in a cell culture model.

Trypan blue and indocyanine green (ICG) have been advocated as vital stains to enhance visualization of optically clear tissue during macular surgery. ^{1 2 3 4 5 6 7 8} By selectively staining ocular tissue, ^{8 9 10} these agents make structures such as the inner limiting membrane (ILM) ^{1 3 4 5 6 8 10 11 12 13} and epiretinal membranes ¹⁴ more evident to the surgeon. Although these chromophores are useful surgical tools, there has been a debate regarding the safety of ICG ^{11 12 13 15 16 17 18 19 20 21 22} and its effect on the retinal pigment epithelium (RPE). ^{12 15 23 24} Clinical studies have suggested functional visual loss after its use, ^{25 26} and many investigators have called for more safety testing. ^{11 12 13 15 18 21 27}  

The purpose of this study was to undertake safety testing of ICG and trypan blue using a cell culture model. Experiments were also undertaken to investigate reports that ICG and endoillumination combine to affect cell viability adversely, ^{21 23 28} as does the combination of ICG and low osmolarity. ²⁴ Unlike previous studies, experiments were undertaken on both RPE and Müller cell lines. This may be important, as studies suggest that it is not only the RPE, but also the neuroretinal elements that may be damaged by ICG. ¹⁹ In particular, profound structural changes have been shown in Müller cells after ICG-assisted macular surgery in cadaveric eyes. ²¹ Given the proximity of the Müller cell end feet to the vitreous surface and their integral association with the ILM, these cells are potentially vulnerable to damage from any neurotoxic agent that binds to the ILM or epiretinal membranes. 

Methods

A human RPE cell line (ARPE-19, passage 23; American Type Culture Collection, Manassas, VA) was cultured using established techniques. ²⁹ Cells were cultured in Ham’s F-10 medium (pH 7.4; Sigma-Aldrich, Poole, UK), supplemented with 2 mM glutamine, 25 mM HEPES, 10 IU/mL penicillin, 10 μg/mL streptomycin, and 15% heat-inactivated fetal calf serum (Sigma-Aldrich). Cells were grown to confluence in an incubator with a humidified atmosphere of 5% CO₂ 95% air at 37°C and kept in a confluent state for 24 to 48 hours before subculture. Cells were trypsinized and seeded at 5 × 10⁴ cells/well into 96-well flat-bottomed plates (TPP, Trasadingen, Switzerland) and 16-well chamber slides (Nunc Inc., Naperville, IL). 

Once cells reached confluence, the growth medium was replaced with 100 μL of the test agent. The concentrations selected were designed to encompass those used clinically. Serial dilutions were prepared to simulate the situation that occurs when the vital stain is diluted into the vitreous volume and to look for dose-related effects. ICG was prepared as described previously. ^{1 30} Twenty-five milligrams of medical grade ICG (BD Biosciences, Cockeysville, MD) was dissolved in 0.5 mL distilled water. This was mixed until fully dissolved, then combined with 4.5 mL of a balanced saline solution (BSS; Alcon, Hemel Hempstead, UK) to produce a 0.5% (5 mg/mL) solution. This preparation was diluted with BSS to provide solutions with a final concentration of 0.5%, 0.25%, 0.125%, 0.0625%, and 0.03125% (n = 7 for each). Trypan blue was dissolved in BSS to give a final concentration of 0.2%, 0.1%, 0.05%, 0.025%, and 0.0125% (n = 8 for each). The osmolarity of each preparation was measured using a micro-osmometer (Advanced Instruments, Needham Heights, MA) and is shown in Table 1 . Exposure times were 5, 15, and 30 minutes for each concentration of agent. After this interval, the wells were rinsed three times with BSS and the growth medium was replaced. Cells were incubated for 24 hours and then viability was assessed. 

The experiments using cells exposed to 0.5% ICG for 5 minutes were repeated with irradiating white light provided by a standard, wide-angle, fiber-optic, endoillumination light-pipe (Alcon Laboratory, Ltd., Herts, UK). After exposure to ICG, wells were rinsed only once with BSS so that the monolayers were still stained with ICG. Each well was individually illuminated with the endolight powered by a medical 300-W xenon light-source (Keeler Instruments, Broomall, PA) set on full power. A xenon light source was selected over the more commonly used halogen light source, as it provided more stimulus for ICG excitation and more chance of detecting ICG-mediated phototoxicity. ³¹ Xenon light sources have recently been made available for use with some of the more commonly used vitrectors. The total lamp output was 5000 lumens (manufacturer’s data) and this was channeled into a fiber optic cable without the interposition of barrier filters. The fiber optic cable had an illumination transmittance of 45% and an angular spread of 8 numeric aperture (NA; manufacturer’s data). The wells were filled with BSS and the light-pipe immersed into each well and held for 1 minute, 5 mm above the cell monolayer (n = 12–24 for each group). After illumination the BSS was replaced with growth medium and cell viability was measured at 24 hours. Experiments were repeated on cells that had undergone illumination after incubation with BSS, and ICG but no illumination. 

To assess the potential for delayed toxicity, experiments were repeated with cells exposed to 0.5% and 1% ICG, and 0.06% and 0.12% trypan blue for five minutes each, with cell viability measured at days 0, 1, 5, and 15 (n = 24 for each concentration and time). The 1% ICG solution was prepared as per the 0.5% solution, except that 50 mg of ICG was dissolved into 0.5 mL of water for injection, instead of 25 mg. The osmolarity of this preparation was 282 mOsM. 

Cell viability was estimated using an MTT (3-(4 to 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay (Sigma-Aldrich). Cells were incubated at 37°C with 100 μL of filtered, 5 mg/mL MTT. After 4 hours, 100 μL of dimethyl sulfoxide (Sigma-Aldrich) was added to lyse the cells and solubilize the formazan reaction product. After 30 minutes, the plates were read in a microplate reader (MR5000; Dynatech, Guernsey, UK) at a test wavelength of 570 nm and reference wavelength of 630 nm. 

A qualitative assessment of cell viability was undertaken using a live–dead probe (Molecular Probes Inc., Eugene, OR). Live cells were identified using calcein-AM (CAM) and dead cells using ethidium homodimer (EH)-1. Titration experiments were conducted to determine the ideal concentration of reagents, as recommended by the manufacturer. Cells were viewed on confocal (LSM 510; Carl Zeiss Meditec, Jena, Germany) and fluorescence (Leitz, Wetzlar, Germany) microscopes. 

To assess the effect of ICG in combination with low osmolarity, MTT assays were repeated on cells exposed to ICG solutions with reduced osmolarity. Instead of dissolving ICG into 0.5 mL of water and then 4.5 mL of BSS, 0.1%, 0.5%, and 1% solutions were prepared by dissolving ICG into 2 mL water and 3 mL BSS to produce a solution of approximately 185 mOsM (range, 182–191; mean, 186 ± 4.5 SD). Cells were exposed to the test agent for 5 minutes, and viability was assessed at 24 hours (n = 24–32 for each group). Experiments using low osmolarity 0.5% ICG were repeated in the presence of endoillumination (n = 12). 

Control experiments were undertaken to determine the effect of hypo-osmotic solutions on cell viability. Cells were incubated with the test solution for 5 minutes, and viability was assessed at 24 hours with the MTT assay. Solutions of various osmolarities were obtained by mixing distilled water with BSS, as occurs in the preparation of ICG. These included a mix of 2 mL water and 3 mL BSS (osmolarity, 181 mOsM); 1 mL water, 4 mL BSS (242 mOsM); 0.5 water, 4.5 mL BSS (272 mOsM); 0.25 mL water, 4.75 mL BSS (286 mOsM); 0.125 water, 4.875 mL BSS (294 mOsM), and 0.063 mL water, 4.937 mL BSS (298 mOsM; n = 5 to 10 for each group). Results were compared with those from cells incubated with BSS alone. 

Negative (live-cell) controls were provided by incubating with BSS instead of the test agent. Dead cells were obtained by exposing cells to 30% methanol. For each test agent, the results obtained from the microplate reader were expressed as a percentage of the negative control. Using this system, values under 100% indicated that the concentration of formazan reaction product was less than that of the negative control, representing a reduced index of cell viability. 

Experiments were repeated using a Müller cell line (passage 57; gift of G. Astrid Limb, The Institute of Ophthalmology, London, UK) grown in Dulbecco’s modified Eagles medium containing l-Glutamax 1 (Invitrogen-Gibco, Paisley, Scotland, UK), supplemented with 2 mM glutamine, 10 IU/mL penicillin, 10 μg/mL streptomycin, and 10% heat-inactivated fetal calf serum (Sigma-Aldrich). The isolation and characterization of these cells is presented elsewhere. ³² Briefly, a spontaneously immortalized cell line was obtained from a 68-year-old female donor. Retina was vigorously pipetted and then trypsinized. Cells were filtered through a stainless-steel sieve, washed, and then grown to confluence. Müller cells were identified using phase-contrast microscopy and by immunostaining for glutamate synthetase, glial fibrillary acidic protein, α-smooth muscle actin, vimentin, cellular retinaldehyde binding protein, and epidermal growth factor receptor. Additional tests including electron microscopy and electrophysiology all confirmed the origin of these cells and are shown in the cited reference. 

Müller cells were passaged by rinsing them in Hank’s-buffered saline solution (Sigma-Aldrich) followed by immersion in one-fourth growing volume of 10× trypsin/EDTA solution (Sigma-Aldrich) for up to 5 minutes. Fresh medium was added so that the medium was returned to its original volume. The cells now in suspension were then split (usually 1:3 or 1:5, depending on cell density) to maintain their density at 60% to 80% confluence. 

As ICG and trypan blue were both chromophores in the blue-green region of the visible spectrum, experiments were conducted to determine whether these dyes interfered with the MTT assay of the blue formazan reaction product. Cells that had been incubated with trypan blue or ICG and then rinsed in the usual manner were placed into the microplate reader, without the addition of MTT. The optical density of these cells was compared with those incubated with BSS (n = 24). 

As noted by other investigators, ³³ we did not have any difficulty discriminating the round nucleolar staining pattern of dead cells labeled with EH-1 from the granular autofluorescence that may occur with higher concentrations of trypan blue. 

Experiments were conducted to determine whether the laboratory grade trypan blue used in the above experiments produced different effects on cell viability to the medical grade preparation used clinically. The laboratory grade preparation was chosen as it allowed a wider range of concentrations than the medical grade preparation that came premade as a 0.06% or 0.15% solution. Hence, concentrations higher than this could not have easily been prepared. Cells were incubated with 0.06% medical (Dorc, Zuidland, The Netherlands) or laboratory grade trypan blue for 5 minutes, and cell viability was measured at 24 hours, as in previous experiments (n = 24). 

Cells were defined as having reduced viability if the mean of at least three experiments using the MTT assay fell below two standard deviations of the negative control. Graphs showing the SD of the negative control show the SD for that experiment, rather than the smaller SD from the overall pooled data. Group comparisons were made using the independent t-test, with Welch correction where standard deviations differed significantly. Nonparametric (Mann–Whitney) tests were used if assumption tests (Kolmogorov-Smirnov) indicated that the groups were not sampled from populations with a Gaussian distribution. P ≤ 0.05 was considered significant. EH-1/CAM was used as an independent, qualitative test without statistical comparison. 

Results

Retinal Pigment Epithelium

The experiments that investigated a range of concentrations of ICG up to 0.5% and trypan blue up to 0.2% did not show reduced cell viability, in that no data point fell below the predetermined limit of two SD below the negative control (Figs. 1A 1B : results for 30-minute exposure). Three ICG results fell below 1 SD. These occurred in cells exposed to 0.125% and 0.25% solutions for 5 minutes and 0.125% for 15 minutes. These appeared to be isolated findings, as both higher and lower concentrations of ICG were within 1 SD of the negative control, as were those of cells exposed to these concentrations for 30 minutes. One trypan blue result fell below 1 SD of the negative control (0.0125% trypan blue for 30 minutes) but this was also an isolated finding, with no evidence of concentration dependent toxicity. 

In the experiments designed to detect delayed toxicity (Figs. 1C 1D) there was no clear relationship between cell viability and the duration of follow-up, except in cells exposed to 1% ICG. These showed a linear (R ² = 0.9952) decline in viability with time, so that the day-15 data point fell below 1 SD of the negative control. This reduction was statistically significant (P = 0.0158, t = 2.54). When stained with EH-1/CAM, these cells were not consistently different from the negative control. Cell damage was not noted in the other groups tested with EH-1/CAM. 

There was no measurable reduction in cell viability 24 hours after exposure to mixtures of BSS and water for injection, over a range of osmolarities. Although some of the data points were reduced, none fell below 1 SD of the iso-osmolar, BSS control. Further, this reduction in cell viability did not relate to the degree of hypo-osmolarity (Fig. 2 ; data points shown as unfilled circles). 

Cells incubated with 0.5% and 1% low (185 mOsM) osmolarity ICG showed a tendency for reduced viability at 24 hours when compared with cells incubated with BSS (Fig. 3 ; columns shown in white), but this difference was not significant (P = 0.056, t = 1.966 and P = 0.48, t = 0.7041, respectively). 

Cells exposed to ICG and then illuminated with an endolight (Fig. 4) were not significantly different from those of the negative control (P = 0.263, t = 1.139), the light-only group (P = 0.502, t = 0.695), or the ICG-only group (P = 0.793, t = 0.263). When these experiments were repeated combining a more hypo-osmolar solution of ICG (185 mOsM) and endoillumination, there was a significant reduction in viability compared with the negative control (P = 0.0016, t = 3.493) and light-only (P = 0.032, t = 2.308) groups, but not compared with the low-osmolarity ICG-only group (P = 0.4819, t = 0.7128). Cell viability in the 185 mOsM ICG-with-light group was significantly lower than the 285 mOsM ICG-with-light group (P = 0.0075, t = 3.124). 

The optical density of the cell monolayers previously stained with ICG or trypan blue without the addition of MTT, were not significantly different from those incubated with BSS (trypan blue P = 0.7406, t = 0.3323; ICG P = 0.7465, t = 0.3262). Experiments comparing medical and laboratory grade trypan blue did not show a significant difference (P = 0.2221, t = 1.242). 

Glial Cells

The studies of Müller cells incubated with a range of concentrations of ICG up to 0.5% and trypan blue up to 0.2% did not show reduced cell viability (Figs. 5A 5B) . Although no data point fell below the negative control minus 2 SD, four fell below 1 SD (0.5% and 0.125% ICG for 5 minutes; 0.25% ICG for 30 minutes; 0.1% trypan blue for 30 minutes). As in the RPE cell experiments, these appeared to be isolated findings, as there was no clear relationship with either concentration or duration of exposure. 

Experiments continued over 15 days showed no delayed toxicity. Cells exposed to 1% ICG, however, tended to have reduced viability compared with the 0.5% group (Fig. 5D) , particularly at day five, when they were 1.32 SDs below the negative control (P < 0.0001, U′ = 935.5). Some monolayers in this group showed scattered cells with positive nucleolar staining with the dead-cell probe EH-1 (Fig. 6) . As noted in the RPE experiments, the fluorescent live–dead probe appeared to be less sensitive than the MTT assay. The other cell cultures tested with the fluorescent live–dead probe showed no consistent pattern of damage. 

Results of MTT assays of cells exposed to mixed BSS and water over a range of osmolarities were similar to those of RPE cells and are shown together in Figure 2 . 

Cells incubated with 0.1% and 0.5% low (185 mOsM) osmolarity ICG did not show reduced viability (Fig. 3) , but those incubated with a 1% solution had significantly reduced viability compared with the negative control (P < 0.0001, t = 4.439). This difference was not significant when compared with the 275 mOsM 1% ICG solution (P = 0.1372, t = 1.510). 

The viability of cells incubated with ICG and then illuminated with an endolight (Fig. 7) was significantly reduced compared with the negative control (P < 0.0001, t = 5.982), light-only group (P < 0.0001, t = 5.919), and ICG-only group (P = 0.0321, t = 2.234). When these experiments were repeated combining a more hypo-osmolar solution of ICG (185 mOsM) and endoillumination, there was a significant reduction in viability compared with the negative control (P < 0.0001, t = 6.638) and light-only group (P < 0.0001, t = 4.838), but not compared with the low osmolarity ICG-only group (P = 0.0848, t = 1.775). There was no significant difference in the 185 and 275 mOsM ICG-with-light groups (P = 0.730, t = 0.347). 

Discussion

Trypan Blue

Trypan blue is a water-soluble acid dye that is used in the laboratory to stain collagen and to distinguish live cells that exclude the dye, from dead cells that do not. ³⁴ It has been used clinically as a retinal vital stain to highlight epiretinal membranes ^{14 35} and assist peeling of the ILM. ³⁶ The present study did not demonstrate reduction in cell viability 24 hours after RPE and glial cells were exposed to trypan blue up to 0.2%, or delayed toxicity with a concentration of 0.06% and 0.12%. Although there are early reports suggesting that high concentrations may damage the sclera and optic nerve, ³⁷ it has been used safely to stain the anterior capsular membrane, ³⁸ and early clinical reports suggest that it is safe as a macula vital stain. ^{2 35}  

Our findings are similar to a brief report by Stalmans et al., ³³ who used the same fluorescent live–dead probe and human RPE cells exposed to trypan blue. ²⁴ The absence of glial toxicity with 0.06% trypan blue is partly consistent with a study in which trypan blue was injected into the vitreous cavity of rabbit eyes. ³⁹ The investigators reported no toxicity with a 0.06% solution but in contrast to the present findings, a 0.2% solution produced retinal toxicity. This difference is not surprising, given important differences in methodology. Rabbits were killed 4 weeks after injection, at which time residual trypan blue was still evident. The contact time was therefore considerably more than in the present study (5–30 minutes), and the clinical application of this dye; typically no more than a few minutes. 

Indocyanine Green

Interaction of Indocyanine Green and Osmolarity.

There was no acute (day 1) damage in RPE or glial cells exposed to ICG preparations of up to 1%, when these were prepared with an osmolarity of approximately 275 mOsM. Although 0.5% and 1% ICG prepared at 185 mOsM showed a tendency for reduced RPE cell viability at 24 hours, this difference was not statistically significant. This tendency for reduced viability is similar to Stalman’s findings, ²⁴ although some investigators suggested that cell damage from ICG cannot be attributed to low osmolarity. ⁴⁰  

Our control experiments showed that RPE and glial cells were not measurably damaged by solutions prepared by mixing saline with distilled water over the range of osmolarities used clinically with ICG. However, glial cells incubated with 185 mOsM 1% ICG had a significant reduction in viability that was not evident with the 275 mOsM 1% ICG. Taken together, these findings suggest that it is the combination of low osmolarity and ICG exposure that produces glial cell damage, rather than each of these factors alone. These findings also suggest that glial cells are more vulnerable to osmotic damage than are RPE cells. Although speculative, it is possible that RPE cells are better able to tolerate osmotic stress because of their role in active fluid transport in vivo. 

Delayed Cell Damage.

Studies designed to detect delayed RPE and glial cell damage over a 2-week interval did not show any toxicity with 0.5% ICG. There was the suggestion of delayed toxicity, however, in RPE cells incubated with 1% ICG. These showed a consistent decrease in viability over time and were significantly lower than the negative control by day 15. The fluorescent live–dead probe failed to demonstrate a qualitative increase in cell death at this time point. This apparent discrepancy with the MTT assay cannot be used to exclude cell toxicity, because reports suggest that the MTT assay may be more sensitive than other measures of cell damage. ²³  

The apparent tendency for delayed RPE damage with 1% ICG cannot be attributed to the hypo-osmolarity of this solution alone, as the 0.5% solution had a marginally lower osmolarity and did not demonstrate cell damage. This supports the hypothesis that ICG can produce delayed toxicity, independent of any reduction in osmolarity. An alternative hypothesis is that increasing ICG concentration augments the damaging effect of hypo-osmolarity. This second hypothesis is consistent with the suggestion of acute (day 1) cell damage seen with high concentration, low-osmolarity ICG, and shown in Figure 3 . Although the preparation used for delayed (day 15) toxicity studies had a higher osmolarity (280 mOsM), it was nonetheless hypo-osmotic relative to physiologic saline. 

Müller cell viability also appeared to be lower in the 1% ICG group relative to the negative control and the 0.5% solution. However, this did not show the consistent decrease over time that was observed in RPE cells. The only data point that fell below 1 SD of the negative control occurred at day 5, but was not evident at day 15. Nonetheless, the fact that the 1% ICG values were lower than the 0.5% group at all three follow-up times suggests that there is a dose-related effect. Although 1% solutions are not used clinically, the presence of a dose-related effect is of clinical importance, confirming that the lower concentrations are likely to be safer. 

Interaction of ICG and Illumination.

There were some differences in the response of RPE and glial cells to light exposure after brief incubation with ICG, with glial cells showing a greater reduction in viability than RPE cells. The additional combination of illumination and low osmolarity resulted in a significant reduction in viability in both RPE and glial cells. Sippy et al. ²³ found reduced viability in a similar experiment using RPE cells and 0.1% ICG. The osmolarity of this solution was not stated but was approximately 247 mOsM. ²⁴ Exposure times were longer, with 10 minutes of illumination. Other researchers found similar results. ²⁸ Studies in human cadaveric eyes suggested that there was inner retinal damage when ICG was combined with irradiation beyond 620 nm. ²¹  

It was noted in our experiments that ICG was harder to rinse free than trypan blue, and there are several reports that ICG may persist in the eye several weeks or months after intraocular use. ^{11 41 42 43 44 45 46} The interaction of residual ICG and transmitted natural light focused on the fovea is not known. 

Strengths and Weaknesses.

One strength of this study was that two cell lines were used. Studies investigating the effect of ICG ^{23 24 28 47} and trypan blue ³³ on cells in culture have been undertaken almost exclusively with RPE cell lines. This may reflect the widespread availability of these cells and the clinical reports suggesting that the RPE is damaged by ICG. There are now also reports suggesting ICG-mediated damage in the inner retinal layers, especially in Müller cells, ²¹ and experimental studies showing functional damage in other neural tissue—namely, spinal root axons. ⁴⁸ Hence, it may be helpful to study neuroretinal cell lines. ⁴⁹ This particularly applies to Müller cells, given that their foot processes are integral to the ILM; the structure stained by ICG. ⁹  

One weakness of these ex vivo experiments is that they cannot fully replicate the situation that occurs clinically in humans. Cell culture studies provide a practical means of testing a wide range of concentrations and exposure routines, including those that would not be thought safe clinically. Testing concentrations beyond those used clinically is important in establishing the toxicity and safety margins of an agent. Cell culture also facilitates a general understanding of the cellular response to agents such as ICG and trypan blue. However, cell culture studies alone cannot be used to reach a conclusion on the clinical safety of these vital stains. Another potential weakness results from the large number of parameters that were investigated. This makes it more likely that some findings occurred by chance, more so in the experiments with dispersed data. 

Summary and Conclusions

In summary, this study found no evidence of cell damage in human RPE and Müller cell cultures incubated with trypan blue and ICG in the doses used clinically. However, glial cells were damaged when exposed to 0.5% ICG and xenon endoillumination. When concentrations were increased and osmolarities reduced beyond those used clinically, then cell damage was evident in both RPE and glial cells delayed RPE toxicity occurred with a 1% ICG solution, acute glial toxicity with low osmolarity 1% ICG. Glial cell viability was reduced when high-concentration ICG was combined with low osmolarity or light exposure. RPE cells were less vulnerable to low-osmolarity preparations, and these combinations did not significantly reduce viability. However, the combination of all three (high concentration, low osmolarity ICG, and illumination) resulted in significant reductions in viability. These findings can be used to draw three main conclusions. First, relatively small changes in ICG osmolarity and concentration may result in cell damage, suggesting that safety margins are not large. Second, 0.5% ICG may produce glial damage if combined with the newly introduced xenon light sources. Last, trypan blue is safe in a cell culture model. 

 Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2003.

 This work was part of a doctorate in philosophy (TLJ).

 Supported by research grants from Allerton Trust; Special Trustees of Guys and St. Thomas Hospitals; and the German Research Council Grant DFGHi758/1-1.

 Submitted for publication February 18, 2004; revised March 20, 2004; accepted March 22, 2004.

 Disclosure: T.L. Jackson, Keeler Instruments (F); J. Hillenkamp, None; B.C. Knight, None; J.-J. Zhang, None; D. Thomas, None; M.R. Stanford, None; J. Marshall, None

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

 

Table 1.

 

Osmolarity of Preparations Used in the Study

Agent (% weight/volume)	Osmolarity (mOsM ± SD)
Trypan blue
0.2	329 ± 2.3
0.1	319 ± 1.0
0.05	314 ± 1.5
0.025	310 ± 1.7
0.0125	308 ± 0.6
Indocyanine green
1.0	282 ± 1.0
0.5	276 ± 0.6
0.25	287 ± 0.6
0.125	294 ± 0.6
0.0625	300 ± 1.0
0.0313	301 ± 0.6
Balanced saline solution	302 ± 0.6
Water for injection	0 ± 0.0

Figure 1.

 

Data show results of MTT assays of RPE cell lines exposed to ICG and trypan blue. The y-axes show the optical density compared with the negative control (cells incubated with physiological saline solution). Values less than 100% reflect reduced cell viability. Horizontal dotted lines: the mean of the negative control ± 1 mean SD. Thus, data points falling below the lower horizontal line are below 1 SD of the negative control. (A, B) Cell viability in relationship to the concentration of agent. The exposure time was 30 minutes, with cell viability read at 24 hours (briefer exposure times are not shown). (C, D) Experiments designed to detect delayed toxicity. Two concentrations of agent were used for both ICG and trypan blue. (•) Higher concentrations; (□) lower concentrations. Note the linear (R ² = 0.9952) decline in viability over time of cells exposed to 1% ICG. Errors bars ± 1 SD.

Figure 2.

 

MTT assay of RPE and Müller cell lines 24 hours after exposure to solutions of varied osmolarity, for 5 minutes each. The serial dilutions of a balanced saline solution started with a mixture of 2 mL water and 3 mL saline (first data point on the left). In each subsequent dilution, the proportion of water was halved. Note that the x-axis is categorical and nonlinear. Data point on the right shows the result from cells incubated with saline alone. Errors bars, ± 1 SD.

Figure 3.

 

MTT assay of a RPE and Müller cell lines exposed to low-osmolarity (185 mOsM) preparations of 0.1%, 0.5%, and 1% ICG for 5 minutes each. Results are compared with those from cells exposed to a balanced saline solution (BSS; column 1). Error bars ± 1 SD.

Figure 4.

 

MTT assay of an RPE cell line exposed to 0.5% ICG with and without illumination by an endolight, and illumination alone. Exposure time was 5 minutes. Error bars ± 1 SD.

Figure 5.

 

Results of MTT assays of Müller cell lines exposed to ICG and trypan blue. As in Figure 1 , the y-axes show the optical density compared with the negative control. This is represented by the central horizontal dotted line ± 1 mean SD. (A, B) the exposure time shown was the maximum (30 minutes), over a range of concentrations, with cell viability read at 24 hours. Briefer exposure times not shown. (C, D) Experiments designed to detect delayed toxicity with the same symbols as those in Figure 1 . Errors bars ± 1 SD.

Figure 6.

 

Müller cell monolayer stained with a fluorescent live–dead probe, 5 days after exposure to 1% ICG for 5 minutes. Most of the cells stained green, due to uptake of calcein-AM. This is converted to the green fluorescent calcein by intracellular esterase activity, resulting in a cytoplasmic staining pattern. By contrast, scattered cells show positive red staining of their nucleoli by the dead-cell marker, ethidium homodimer-1. Bar, 100 μm.

Figure 7.

 

MTT assay of Müller cell line exposed to 0.5% ICG with and without illumination by an endolight and illumination alone. Exposure time was 5 minutes. Error bars ± 1 SD.

The authors thank Austin El Osta and Paul Constable for assistance with cell culturing.