Phototoxicity of Indocyanine Green under Continuous Fluorescent Lamp Illumination and Its Prevention by Blocking Red Light on Cultured Müller Cells

Abstract

Purpose.: To investigate the phototoxicity of persistent indocyanine green (ICG) under continuous visible light illumination and to determine whether blocking peak absorbance wavelengths of ICG is cytoprotective.

Methods.: Cultured quail Müller cells were exposed to 0 to 5 mg/mL ICG for 30 seconds or 10 minutes and then were cultured in a colorless medium for 24 hours with or without continuous fluorescent lamp illumination. Cells exposed to 5 mg/mL ICG for 10 minutes were cultured under illumination filtered through a dichroic mirror that blocks red to near-infrared, green, or blue wavelengths. After microscopic observation, cell viability and cell death were evaluated.

Results.: ICG exposure followed by illuminated culture induced severe morphologic changes in cells, significant reductions in cell viability, and increases in cell death from apoptosis compared with exposure to ICG or illumination alone or with no exposure. Although ICG exposure at higher concentrations caused cell damage in a dose- and time-dependent manner, an increase in cell viability was noted for cells exposed to lower ICG concentrations. Blocking red to near-infrared wavelengths prevented the decrease in cell viability and the increase in cell death in the culture exposed to ICG followed by illuminated culture.

Conclusions.: Continuous fluorescent lamp illumination enhanced the cytotoxicity of persistent ICG on Müller cells in a dose- and exposure time-dependent manner. Blocking peak absorbance wavelengths of ICG prevented photodynamic cytotoxicity of persistent ICG under continuous visible light illumination in vitro. This culture system could be used to study the mechanisms of prevention of unfavorable outcomes in ICG-assisted surgery.

Indocyanine green (ICG) is a photosensitive tricarbocyanine dye with a peak spectral absorbance at approximately 780 nm ^1,2 that is used for measuring cardiac output, plasma volume, and liver function ³ and for ophthalmic angiography. ^4,5 In addition to its conventional use as an intravenous contrast medium, ICG has been used in vitreoretinal surgery to visualize the transparent inner limiting membrane (ILM). ⁶ Thus, ICG staining makes it possible to remove the ILM efficiently and completely. ^7,8 Although ICG application has been accepted as a safe procedure, especially for macular surgery, ^9,10 several clinical reports have indicated ICG-related adverse outcomes ^11–13 such as atrophic changes in retinal pigment epithelium (RPE), poor improvement of visual acuity, and unexpected visual field defects. 

One possible cause of these adverse outcomes may be light irradiation from an endoilluminator on retinal tissues exposed to ICG. ^14,15 Gandorfer et al. ¹⁶ reported that irradiation of retinal tissue with a light-pipe light at wavelengths from 620 to 760 nm, which include the peak absorbance wavelengths of ICG, induced severe inner retinal damage in ICG-exposed cadaveric eyes, whereas the damage was mild under irradiation at wavelengths from 380 to 620 nm, which are only weakly absorbed by ICG. 

Clinical observations have shown that ICG persists on the retina after intravitreal injection for as long as several weeks and even up to several months. ^17–20 Therefore, persistent ICG may exert light-induced toxicity because of exposure to ambient light that contains its peak absorbance wavelengths. To our knowledge, the possible toxicity of persistent ICG irradiated with ambient light has not been investigated. 

Some in vitro studies have reported ICG toxicity in RPE cells, ^21–23 and ICG has been shown to damage neuroretinal elements. ²⁴ In particular, Müller cells are potentially vulnerable to toxic agents that bind to ILM because ILM consists of a basement membrane for Müller cells. ^25,26  

In this study, we established a cell culture system to analyze the effect of ambient light on ICG toxicity. This system, which uses a fluorescent lamp to simulate irradiation with ambient light, shows that the combination of ICG exposure and continuous fluorescent lamp illumination causes photodynamic cytotoxicity in cultured Müller cells. We successfully showed that the enhancement of persistent ICG cytotoxicity by illumination could be alleviated by blocking the range of peak absorbance wavelengths of ICG. 

Materials and Methods

Cell Culture and Culture Medium

This study used QNR/K2 cells (American Type Culture Collection, Manassas, VA), an established cell line that was immortalized and cloned and that displays properties of Müller (astroglia) cells. The cells were confirmed to express glial fibrillary acidic protein by Western blot analysis (data not shown) and were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich, Poole, UK) supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% fetal bovine serum (FBS; JRH Bioscience, Lenexa, KS) at 39°C in a humidified atmosphere of 5% CO₂ and 95% air. The cells were trypsinized and subcultured to confluence in multiwell polystyrene plates. 

We applied a novel colorless culture medium consisting of Dulbecco's phosphate-buffered saline (PBS; Sigma-Aldrich) supplemented with 1% FBS, 1 mg/mL glucose, 1 mg/mL CaCl₂, 1 mg/mL MgCl₂, and antibiotics. 

ICG Exposure and Continuous Fluorescent Lamp Illumination

ICG (Daiichi Sankyo, Tokyo, Japan) was dissolved in distilled water to achieve a final concentration of 25 mg/mL. After adding balanced saline solution (BSS Plus; Alcon, Tokyo, Japan) to prepare a 5 mg/mL ICG stock solution, the ICG solution was further diluted in ICG dilution buffer; 80% balanced saline solution diluted with distilled water, to achieve the final ICG concentrations of 0.625, 1.25, and 2.5 mg/mL that were used for the experiments. The osmolarity of each ICG solution was approximately 242 mOsm, whereas that of the balanced saline solution was approximately 302 mOsm. ²¹ The cells were washed once with PBS and then exposed to the ICG solution for either 30 seconds or 10 minutes in the dark. After exposure, cells were immediately rinsed three times with the balanced saline solution and then cultured in colorless medium either in the dark within a metal box or under 2000 lx illumination from a daylight-colored fluorescent lamp (6500 K, Sunline; Hitachi, Tokyo, Japan) for 24 hours at 39°C in a humidified air in an incubator fitted with fluorescent lamp equipment (CPO₂-171; Hirasawa, Tokyo, Japan). 

For blocking of specific spectral wavelengths, the cells were cultured in 4-well plates. A red, green, or blue dichroic mirror (DM; 4 × 4 cm²; Koshin, Tokyo, Japan) was attached to the culture plates. Each of the DMs reflected wavelengths corresponding to its respective color and allowed wavelengths corresponding to the other colors to pass through. The cells were exposed to 0 or 5 mg/mL ICG for 10 minutes in the dark. The sides of plates were covered with aluminum foil to prevent the transmission of fluorescent lamp light. Four plates with or without DMs were incubated under illumination, and another plate was incubated for 24 hours in the dark. 

Measurement of Cell Viability and Cell Death

Quantitative assessment of cell viability was evaluated by measuring mitochondrial reductase activity using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS) assay (Promega, Madison, WI). Quantitative assessment of cell death was evaluated by a lactate dehydrogenase (LDH; Roche, Mannheim, Germany) assay that measures the activity of LDH released into the culture supernatant from dead cells. MTS and LDH assays were performed after the manufacturers' protocols. In some experiments, 2 μg/mL propidium iodide (PI) was added to the medium after incubation to detect dead cells. These samples were observed and photographed using a universal microscope (Biozero; Keyence, Osaka, Japan). 

Detection of Apoptosis

Apoptosis was detected by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) using fluorescein-conjugated deoxyUTP as the substrate (In Situ Cell Death Detection Kit; Roche). The cells were fixed with neutralized formalin and labeled according to the manufacturer's protocols. 

Spectrophotometric Measurements

The transmission spectra of the three DMs and the absorbance spectrum of the experimental ICG solution diluted with ICG dilution buffer were analyzed with a spectrophotometer (U-0080-D; Hitachi). 

Statistical Analysis

At least four samples in each group were measured for each experiment. Data from more than three groups were analyzed by nonrepeated-measures ANOVA with Dunnett's test for comparison with the control. Unpaired two-tailed Student's t-test was used to analyze the data from two groups, and the Student-Newman-Keuls test was used for multiple comparisons. P < 0.05 was considered significant. 

Results

Culture Conditions

For evaluation of the effects of ambient light on ICG toxicity in ocular cell culture, a colorless medium was required to allow the entire spectral range emitted from the fluorescent lamp. However, commercially available DMEM usually contains phenol red, which absorbs some wavelengths of visible light. We therefore evaluated several culture media to establish a suitable culture system for this study. 

Initially, we attempted to culture Müller cells in phenol red-free DMEM-10% FBS. However, none of the cells survived in this culture medium under continuous fluorescent lamp illumination (data not shown). We next determined whether the observed cellular damage was induced directly by illumination or was due to a change in the culture medium caused by illumination. For this purpose, phenol red-free or phenol red-containing DMEM-10% FBS was first illuminated in the incubator for 24 hours in the absence of cells, and then cells were incubated in these media for 24 hours in the dark. Cell morphology was subsequently observed by phase-contrast microscopy. Although no morphologic changes were observed in cells in the phenol red-containing medium, cells in the phenol red-free medium were severely damaged and detached from the culture substrate (Figs. 1A, 1B, respectively). To clarify the relationship between the time of preillumination and cell damage, we further quantified cell viability and cell death in the culture in phenol red-free DMEM-10% FBS, which had been incubated under illumination for various times and subsequently cultured in the dark for 24 hours. Cell viability was evaluated using an MTS assay, and cell death was evaluated by measurement of LDH activity in the culture supernatant. Preillumination for >3 hours clearly induced a decrease in cell viability and an increase in cell death (Fig. 1C). These results indicated that the use of phenol red-free DMEM was not suitable for our study. 

Figure 1.

 

Light-induced toxicity of DMEM-based media and time course of cell death in a colorless PBS-based medium. DMEM-10% FBS, with or without phenol red, was preincubated without cells under fluorescent lamp illumination for 24 hours. Cells were cultured in the preilluminated DMEM-10% FBS with (A) or without (B) phenol red in the dark for 24 hours. Cell morphology was observed by phase-contrast microscopy. It was normal in the phenol red-containing medium (A), but the cells in the phenol red-free medium detached from the culture substrate (B). Scale bar, 400 μm. (C) Confluent cells were incubated in phenol red-free DMEM-10% FBS for the indicated periods under illumination and were further cultured in the dark for 24 hours. Cell viability and cell death were measured using the MTS assay (dotted line) and by assay of LDH activity in the culture supernatant (solid line), respectively. Optical density was measured at 490 nm and 690 nm for reference. Preincubation for 3 hours or longer with illumination induced severe cellular damage. (D) Cells were exposed to 0 or 5 mg/mL ICG for 10 minutes and were cultured in the colorless medium with or without illumination, and LDH activities at the indicated time in the culture supernatant were measured. Cells exposed to 5 mg/mL ICG followed by culture with illumination died most rapidly within 24 hours. The increase in cell death was slowest in the cultures without ICG exposure and incubated in the dark.

Thus, we decided to use a newly formulated colorless medium that consisted of 1% FBS, 1 mg/mL glucose, 1 mg/mL CaCl₂, 1 mg/mL MgCl₂, and antibiotics in PBS. Given that this colorless medium lacks nutrients such as amino acids and vitamins, we first tested how long the cells could survive in this medium with or without illumination and determined the appropriate time for evaluation of cellular damage by ICG phototoxicity. We examined the changes in LDH activity in the culture supernatant (Fig. 1D). LDH activity of the culture exposed to 0 mg/mL ICG without illumination gradually increased over 48 hours. LDH activity in the culture exposed to 5 mg/mL ICG without illumination was twice as high as that in the culture exposed to 0 mg/mL ICG without illumination at 6 hours. Thereafter, the increase in LDH activity was parallel for the two conditions. Illuminated cultures not exposed to ICG did not change LDH activity over the first 12 hours; rather, they induced more rapid increases in LDH activity thereafter compared with nonilluminated cultures not exposed to ICG. Illuminated cultures exposed to 5 mg/mL ICG induced the most rapid increase in LDH activity in the four conditions assayed over the first 24 hours, and LDH activity reached a plateau at 33 hours. After 33-hour illuminated incubation, measurements of cell viability and PI staining of nuclei in dead cells revealed no living cells in the ICG-exposed culture (data not shown). Thus, cell damage induced by a combination of ICG exposure and illumination was much more severe than damage induced by ICG exposure alone or by illumination alone. Based on these results, we chose 24 hours for evaluation of cytotoxicity by ICG exposure and/or illumination because, at this time, damage by either ICG exposure alone or illumination alone was evident and the change in LDH activity was relatively linear. 

ICG Staining of Müller Cells

We next determined the effect of illumination on ICG staining of the cells. The cells were exposed to 5 mg/mL ICG for 30 seconds or 10 minutes and then either assayed immediately or further incubated for 24 hours with or without illumination. Subsequently, changes of ICG staining were analyzed by bright-field microscopy. We observed persistent ICG staining of the cells in each experimental condition (Figs. 2A–F). Before the 24-hour incubation, the intensity of ICG staining depended on ICG exposure time (Figs. 2A, 2D). ICG staining persisted after 24-hour incubation without illumination, though the staining was of lower intensity (compare Figs. 2A and 2D with 2B and 2E). In the culture exposed to ICG for 10 minutes and incubated for 24 hours, individually stained cells were more easily distinguished than those in cultures before incubation (Figs. 2E, 2D, respectively). In contrast, ICG staining faded in the cells that were illuminated for 24 hours after ICG exposure (Figs. 2C, 2F). 

Figure 2.

 

Effect of time of ICG exposure and illumination on ICG staining, and effect of ICG exposure and illumination on cell morphology and cell death. (A–F) Cells were exposed to ICG for 30 seconds or 10 minutes and were immediately observed by bright-field microscopy for ICG staining (A, D, respectively) or were further incubated for 24 hours with (C, F, respectively) or without (B, E, respectively) illumination before observation. Arrows: persistent ICG staining immediately after ICG exposure (A, D). There were more cells with stronger staining in cultures exposed to ICG for 10 minutes than for 30 seconds. Staining persisted after 24-hour incubation without illumination (B, E) but faded with illumination (C, F). (G–N) Cells exposed to 0 or 5 mg/mL of ICG for 10 minutes, followed by culture for 24 hours with (+) or without (−) illumination, were analyzed for cell death by PI staining of nuclei (G–J) or for apoptotic cells by TUNEL (K–N). Merged images of fluorescence micrographs and phase-contrast micrographs are shown. Cells were exposed to 0 mg/mL (G, I, K, M) or 5 mg/mL (H, J, L, N) ICG for 10 minutes and were cultured for 24 hours without (G, H, K, L) or with (I, J, M, N) illumination. In ICG-exposed cultures without illumination (H, L) or in cultures not exposed to ICG without (G, K) or with (I, M) illumination, most cells maintained a flattened morphology with cellular processes, and few cells were PI positive (G–I) or TUNEL positive (K–M). In cells exposed to 5 mg/mL ICG, followed by culture with illumination, almost all cells were rounded and shrunken with PI-positive (J) or TUNEL-positive (N) nuclei. Scale bars: (A–F) 25 μm, (G–N) 100 μm.

Morphologic Changes and Detection of Dead Cells after ICG Exposure Followed by Continuous Fluorescent Lamp Illumination

The effect of illumination combined with ICG exposure on cell morphology and on the extent of cell death was then compared. We observed morphologic changes of the cells exposed to 0 or 5 mg/mL ICG for 10 minutes and detected dead cells by nuclear PI staining after 24-hour incubation with or without illumination (Figs. 2G–J). In cultures exposed to 0 mg/mL ICG or without illumination (Figs. 2G–I), most cells maintained a flattened morphology; only a few cells had a shrunken or rounded appearance with PI-positive nuclei. In cultures exposed to both 5 mg/mL ICG and illumination, many cells had shrunk, but others still adhered to the substrate with a flattened or an elongated shape. However, almost all these cells, including the adhered cells, exhibited nuclear PI staining (Fig. 2J). 

We further performed TUNEL of these cells to determine whether apoptosis was involved in the observed cell death (Figs. 2K–N). Only a few TUNEL-positive cells were found in cultures exposed to 0 mg/mL ICG or without illumination (Figs. 2K–M). In contrast, most cells exposed to 5 mg/mL ICG with illumination were TUNEL-positive (Fig. 2N). Quantification of the apoptotic cells in total cells in the culture was not possible because weakly adhered cells, including dead cells, had been washed out during the TUNEL procedure. 

Effects of ICG Exposure and Continuous Fluorescent Lamp Illumination on Cell Viability and Cell Death

To determine the effect of the exposure length and concentration of ICG on subsequent ICG-induced cell death, we examined the various ICG concentrations and exposure times of 30 seconds and 10 minutes. Cell viability (Fig. 3A) and cell death (Fig. 3B) under each experimental condition were compared with those in the control cultures not exposed to ICG. At 0 mg/mL ICG exposure, although illumination reduced cell viability to 55% and 57% and increased cell death to 172% and 160% in 30-second and 10-minute ICG exposure times, respectively, the values in cell viability and cell death were almost the same between ICG exposure times both in absence and the presence of illumination. 

Figure 3.

 

Effect of ICG concentration, exposure time, and subsequent illumination on cell viability and cell death. Cell viability (A) and cell death (B) were evaluated in cultures exposed to the indicated ICG concentrations for 30 seconds or 10 minutes, followed by 24-hour incubation with or without illumination. Cell viability and cell death were quantified by the use of MTS assays and LDH assays, respectively. The values in each experimental group were normalized to the control of 0 mg/mL ICG exposure. At high ICG concentrations, cell viability was decreased to varying extents, depending on ICG concentration and exposure time, and was lowest in the illuminated culture exposed to 5 mg/mL ICG for 10 minutes. At low ICG concentrations, cell viability was increased in the cultures exposed to ICG for 10 minutes, followed by incubation with or without illumination. Cell death increased in all cultures with increasing ICG concentration and was highest in the illuminated cultures treated with 5 mg/mL ICG for 10 minutes. In general, changes in cell viability and cell death were more evident in 10-minute exposure than 30 second-exposure. Cell death and cell viability were inversely correlated for each ICG concentration tested.

There was no significant difference in cell viability in cultures exposed to ICG for 30 seconds without illumination (Fig. 3A, black bars). In the cultures exposed to ICG for 10 minutes with or without illumination, lower ICG concentrations tended to increase cell viability whereas higher concentrations tended to decrease it (Fig. 3A, hatched and gray bars). 

The effect of higher ICG concentrations on cell death was almost opposite that on cell viability (Fig. 3B, hatched and gray bars). There was a dose-dependent increase in cell death in the cells exposed to ICG for 30 seconds without illumination, even though no significant difference in cell viability was observed under these conditions (Figs. 3A, 3B; compare black bars). A significant increase in cell death was detected at 5 mg/mL ICG concentration under all conditionstested. Although elevated cell viability at lower ICG concentrations was not always accompanied by decreased cell death, a decrease in cell viability and an increase in cell death at higher ICG concentrations were evident in all conditions except for 30-second ICG exposure without illumination. Therefore, exposure to higher ICG concentrations for a longer time induced cell damage, and this damage was enhanced by subsequent illumination (Figs. 3A, 3B; compare hatched and gray bars). 

Spectra of DMs, ICG Solution, and Fluorescent Lamp Light

To determine the possibility of preventing cytotoxicity of ICG by using a DM to block specific wavelengths of visible light, we measured the transmission spectrum of three different color DMs and ICG solution. Ranges of wavelength that were blocked more than 90% were as follows: by DM-red, 641 to 860 nm (red to infrared); by DM-green, 556 to 616 nm (green to orange); and by DM-blue, less than 516 nm (violet to blue) (Fig. 4A). The absorbance spectrum of 5 μg/mL ICG showed a peak at approximately 780 nm (Fig. 4B). DM-red blocked the peak range of the spectrum of the ICG solution. The relative energy of the fluorescent lamp light exhibited a double-peaked curve with two maxima and three peaks (Fig. 4C). ²⁷ Light energy in the wavelength range of ICG peak absorbance was low (Figs. 4B, 4C). Light energy in the spectrum of the fluorescent lamp light, at wavelengths greater than 641 nm, was also relatively low. Therefore, DM-red blocks less fluorescent lamp light energy than the other two DMs (Figs. 4A, 4C). 

Figure 4.

 

Analysis of the spectra of DMs, the ICG solution, and the light from the fluorescent lamp. The transmission spectra of the DMs used to block specific wavelengths of light (A) and the absorbance spectrum of a 5 μg/mL ICG solution (B) were measured using a spectrophotometer. (B) Bars under the horizontal axis show the range in which the DMs blocked more than 90% of transmission. The wavelengths blocked by DM-red overlapped with the range of the peak absorbance of ICG. The spectrum distribution of light from the fluorescent lamp light (C) is cited at http://www.hitachi-hll.co.jp/catalog/institution/index/s314-336.pdf, with permission from Hitachi, Ltd. The spectrum distribution indicates that the light energy was relatively low at wavelengths that overlapped with high absorbance of ICG.

Prevention of the Combined Cytotoxicity of ICG Exposure and Illumination by a Dichroic Mirror

The effects of DMs on light-induced cytotoxicity of ICG were examined in ICG-exposed cells cultured under illumination with or without a DM or in the dark. Cell morphology, cell viability, and cell death were assayed in culture after 10-minute exposure to 5 mg/mL ICG and 24-hour incubation. Most cells exposed to ICG and illuminated without a DM showed a change in morphology compared with ICG-exposed cells cultured in the dark (Fig. 5A). Some of the cells were shrunken and rounded, and flattened cells became less reflective or had vacuoles in their cytoplasm. When illuminated through DM-green or DM-blue (Figs. 5C, 5D, respectively), these changes were prevented to some extent. When illuminated through DM-red, cell morphology (Fig. 5B) was similar to that of the cells cultured in the dark (Fig. 5E). 

Figure 5.

 

Effect of DMs on cell morphology, cell viability, and cell death in the culture exposed to ICG in the absence or presence of illumination. Phase-contrast micrographs of cells exposed to 5 mg/mL ICG for 10 minutes followed by illuminated culture without DM (A), through individual DMs (B–D) or in the dark (E) for 24 hours. Cell viability (F) and cell death (G) were also measured, as described, for Figure 1. Values were normalized to those of cells illuminated without a DM. There were many shrunken cells and cells with vacuoles in the cultures illuminated without a DM (A) or illuminated through DM-green (C) or DM-blue (D). However, cells cultured with illumination through DM-red (B) did not show demonstrable morphologic changes, and their morphology was similar to cells cultured in the dark (E). Scale bar, 50 μm. Cell viability (F) and cell death (G) of the cells cultured in the dark was higher and lower, respectively, than cells cultured under illumination without a DM. Illumination through each DM had, to some extent, significant effects on cell viability and cell death compared with illumination without a DM. In particular, DM-red increased cell viability (F) and reduced cell death (G) more effectively than the other two DMs. The values for cell viability and cell death obtained for the DM-red culture were similar to those obtained for cells cultured in the dark.

When cultured in the dark, cell viability (Fig. 5F) and cell death (Fig. 5G) were higher and lower, respectively, than when the cultures were illuminated without DM. Culture illuminated through any of the DMs significantly elevated cell viability and decreased cell death compared with culture illuminated without a DM. In particular, of all DMs tested, illumination through DM-red provided the strongest cytoprotection against the phototoxic effects of persistent ICG. 

In the cells not exposed to ICG, the use of DMs did not significantly change cell viability or cell death, although incubation in the dark showed higher cell viability and lower cell death than it did in the four illuminated conditions (data not shown). 

Discussion

The present study was designed to investigate the toxicity of persistent ICG on cultured Müller cells under continuous fluorescent lamp illumination, which mimics ambient light. In this study, we demonstrated that fluorescent lamp illumination enhanced the cytotoxicity of persistent ICG and that the blocking of red to near-infrared wavelengths alleviated this light-induced cytotoxicity. 

Intravitreal ICG application for visualization of transparent ILM is a standard procedure for vitreoretinal surgery such as that on macular holes because ILM peeling is essential for successful anatomic closure. ^28,29 On the other hand, a meta-analysis report indicates that treatment of macular hole with ICG results in worse functional outcomes than treatment without ICG. ³⁰ Thus, intravitreal ICG application remains controversial. It is known that ICG has a dose-dependent toxicity ¹⁵ and that irradiation by light-pipe light enhances its toxicity. ^14,15 Moreover, it has been suggested that there is a risk of poor outcomes because of the effect of irradiation with ambient light on persistent ICG in ICG-exposed eyes. ^15,31  

A medium that transmits the entire range of visible light was required for our study. Initially, we attempted to culture the cells under fluorescent lamp illumination in DMEM without phenol red, which absorbs a specific range of visible light. ³² However, the cells could not survive in the phenol red-free medium, though they could survive in the phenol red-containing medium. It has been reported that riboflavin and tryptophan in a medium deteriorate by irradiation with fluorescent lamp light, ³³ and this might have caused the loss of competency of the medium sustaining the cells in our culture. In contrast, phenol red, which functions as an antioxidant, ³⁴ might have prevented the denaturation of these components, thereby allowing the cells to survive. In this study, the cells could be maintained for at least 48 hours in our colorless medium, even though this medium contained a small amount of amino acids or vitamins. For future evaluation of longer culture times, it will be necessary to prepare a new suitable colorless medium. 

ICG-related toxicity has been reported for various ICG concentrations (0.025%–0.5%), volumes (3 drops-2 mL), and exposure times (30 seconds-5 minutes) in clinical studies. ^{11,12,15,19,35,36} The actual intraocular ICG concentration during vitreoretinal surgery has been estimated. ¹⁵ However, the ICG concentration on the inner retinal surface could be higher than that in the vitreous or the value determined from theoretical calculations because of the accumulation of ICG at the vitreoretinal interface. ¹⁶ Thus, the actual intraocular concentration of ICG on the retina has remained controversial. There have been several in vitro studies of ICG toxicity using various ICG concentrations (0.01%–2.0%) and exposure times (5 minutes-3 hours). ^22,26,37 Based on these previous reports, we chose ICG concentrations between 0.625 and 5 mg/mL and ICG exposure times of 30 seconds and 10 minutes to examine the cytotoxicity of persistent ICG. 

Effects of ambient light may be influenced by illuminance, light wavelengths, and color temperature. Ashby et al. ³⁸ reported the effects of ambient light on myopia in chicks and adopted fluorescent lamp illumination (500–30,000 lx) as irradiation with ambient light. Although various fluorescent lamps that emit light of different color temperature are available, the CIE Standard D65 illuminant, which represents noon daylight (6504°K), is one of standard illumination conditions used in color science and engineering. ³⁹ We therefore chose the condition of fluorescent lamp illumination (2000 lx; daylight color, 6500°K) as irradiation with ambient light. 

In the time-course experiment of cell death, exposure to ICG without illumination induced rapid cell death in the early culture period but thereafter the death rate was slow, suggesting that although some cells died soon after ICG exposure, many cells survived. Hsu et al. ⁴⁰ reported that ICG exposure immediately damaged human RPE cells but that the surviving cells grew equivalently to the cells without ICG exposure. On the other hand, based on cell viability measurements, Jackson et al. ²¹ suggested that ICG exposure caused delayed toxicity in human RPE and Müller cells. Further investigation is necessary to clarify whether the surviving cells maintained the same properties as the cells not exposed to ICG. 

In contrast to ICG-exposed cells, cell death in illuminated cultures not exposed to ICG continued to increase over the incubation time. This result suggests that ICG exposure and illumination have different mechanisms of cytotoxicity. It is known that animal tissues absorb some wavelengths of visible light. ⁴¹ For example, cytochrome c oxidase in mitochondria absorbs near ultraviolet and visible light. ⁴² We therefore considered that gradual inactivation of molecules critical for cell survival, such as cytochrome c oxidase, might have caused the observed light-induced acceleration of cell death. 

We also considered potential mechanisms that could underlie the observed ICG staining of cells and of fading after illumination. The ICG stain detected on cells after irrigation might have been derived from PBS-insoluble ICG precipitates in the media. ²³ On the other hand, at least some of this stain might have represented intracellular ICG because ICG has been shown to be incorporated into cells by a Na⁺-involved cotransporter and is retained in the cytoplasm. ^37,43,44 Moreover, ICG changes its molecular structure and forms aggregates whose absorbance properties vary, depending on their exposure to light, different temperature conditions and the time after ICG was dissolved in solvent. ⁴⁵ We propose that the fading of ICG staining observed after illumination was not caused by the dissolution of ICG into the medium but by the decomposition of ICG incorporated into the cells and/or changes in the light absorbance properties of ICG on the cell surface during incubation. Investigation of the changes in the chemical features of ICG during illuminated cell culture could provide a clue toward elucidating the mechanisms of photo-induced cytotoxicity of ICG. ^23,31,45  

Continuous fluorescent lamp illumination resulted in rapid fading of ICG staining and enhanced the cytotoxicity of persistent ICG. It has been reported that light exposure degrades ICG and lowers its peak absorbances. ⁴⁵ Engel et al. ³¹ investigated the mechanism of the photodynamic effects of ICG and reported that singlet oxygen produced by light-irradiated ICG led ICG to decompose into strongly cytotoxic products. Because these ICG decomposition products are almost colorless, in contrast to intact ICG, we considered that the same reaction occurred in our culture condition and caused both the fading of ICG staining and the strong cytotoxicity under illumination. There were many TUNEL-positive nuclei in the cells exposed to ICG after illuminated culture, suggesting that the combination of ICG exposure and illumination induces apoptosis. Furthermore, this combination caused a much greater incidence of cell death than the sum of the incidences of cell death induced by ICG exposure alone and illumination alone. Therefore, cytotoxic mechanisms of ICG exposure under illumination appear to be different from those in the dark. 

Cell exposure to high ICG concentrations caused both a decrease in cell viability and an increase in cell death in dose- and exposure time-dependent manners, similar to previous studies. ^26,37,46 Moreover, illumination enhanced these effects of ICG exposure. In contrast, ICG exposure at low concentrations induced an increase in cell viability that was not associated with a decrease in cell death, as shown in our previous study. ⁴⁷ In both the previous and the present studies, cell viability was quantitatively assessed by measurement of mitochondrial reductase activity, which is considered to represent the total number of metabolically active cells. Therefore, the increase in cell viability would be caused not by an increase in the number of viable cells but by enhanced metabolic activity in surviving cells. In early responses to stress, cells activate survival pathways to repair low-level damage. ⁴⁸ Therefore, it is conceivable that the surviving cells could have elevated their metabolic activity by activating survival pathways after ICG exposure at low concentrations. 

The reduction in ICG-induced phototoxicity by blocking red to near-infrared wavelengths is the main result of our present study. ICG adsorbed on human retinal specimens absorbs light of red to near-infrared wavelengths to the same extent as ICG solutions diluted for clinical use. ^2,49 Gandorfer et al. ¹⁶ reported that irradiation with wavelengths greater than 620 nm using endoillumination led to severe damage to the inner retina in cadaveric ICG-exposed eyes. In addition to their report simulating surgical conditions, we have demonstrated in culture that the wavelengths of light emitted from commonly used fluorescent lamps are harmful, as shown by the reduced light-enhanced cytotoxicity of ICG by blocking these wavelengths. 

Our results suggest the possibility that blocking red to near-infrared wavelengths may alleviate unfavorable functional outcomes in ICG-exposed eyes. These results, therefore, may have important clinical implications. Few reports mention the risk for postoperative exposure to ambient light on ICG-exposed eyes. ^15,31 Further in vivo studies under conditions similar to those used in the clinic are necessary to verify the effectiveness of blocking red to near-infrared wavelengths to reduce ICG toxicity in a clinical setting. 

At present, vital dyes such as trypan blue ^21,50,51 and brilliant blue G ^26,52 have been considered to be more biocompatible than ICG for ILM staining. Our culture system may also be useful for the evaluation of the phototoxicity of these vital dyes under continuous light exposure. 

Footnotes

 Supported by a grant from the National Defense Medical College.

Footnotes

 Disclosure: T. Sato, None; M. Ito, None; M. Ishida, None; Y. Karasawa, None