Role of Photobleaching Process of Indocyanine Green for Killing Neuroblastoma Cells

Abstract

Indocyanine green (ICG) is an FDA-approved near infrared (NIR) imaging agent for diagnosis and imaging guided surgery. It also exhibits phototoxicity under high-dose NIR irradiation, expanding its application as a photo-therapeutic agent. Since ICG’s efficiency as a type II photosensitizer has been controversial due to its low triplet state yield, other mechanisms have been explored. While claims of toxic decomposition products, accompanied by irreversible ICG photobleaching, were proposed as the main mechanism, evidences from systemic studies are lacking. In this work, we aimed to unravel the factors affecting ICG photobleaching and the associated photo-killing effect on neuroblastoma, one of the most common pediatric tumors but often escapes therapy. Specifically, we examined how albumin-induced ICG stabilization affects the ICG photobleaching process, and the effect of photobleached ICG on cell proliferation and viability of neuroblastoma cells. It was found that ICG photobleaching was significant only under aerobic conditions and was more efficient in solutions with higher concentration ICG monomers, which were stabilized from aggregates by the presence of BSA while increasing photobleaching and associated oxygen consumption. Photobleached ICG inhibited cell proliferation, indicating another effect of tumor treatment by ICG. Taken together, while enhanced photobleaching by BSA-bound ICG monomers may reduce the photodynamic effect targeting cellular components, the photoproducts directly contribute to tumor growth inhibition and assist in a secondary mechanism to stop tumor growth.

1. Introduction

Indocyanine green (ICG) has been widely used for medical diagnosis due to its favorable absorption and fluorescence in near infrared (NIR) range [1,2], which renders less interference with tissue autofluorescence and better imaging depth. Although ICG was generally considered a safe chemical reagent and has been approved by FDA for clinical use as an optical probe, its phototoxicity remains a controversial topic. Studies have shown that irradiation of cells incubated with ICG led to cell death, whereas no toxicity was observed when cells were incubated with the same concentration of ICG without irradiation [3,4]. The mechanism of ICG phototoxicity was proposed to be a traditional type II photosensitization process, in which its triplet state reacted with oxygen to produce singlet oxygen (¹O₂) [1]. The role of ¹O₂ was corroborated by the observation that ICG could kill colonic and melanoma cancer cells [3,4] and the addition of ¹O₂ quencher successfully reduced the killing effect in the presence of light [3]. However, detailed photophysical characterization of ICG determined that the triplet quantum yield was at 0.1%, a value too low to play a dominant role for this purpose [1], leading to the assumption that type I photosensitization may contribute largely to ICG’s phototoxicity.

In type I sensitization, the excited ICG reacts with surrounding chemical species to produce radicals or radical ions, which react with oxygen and ICG itself to yield products to cause cell death. ICG oxidation is accompanied by photobleaching while degrading into several photoproducts via the polymethine bridge to form 1,2-dioxetane rings that degrades into two carbonyls [5,6]. The decrease of fluorescence intensity was also observed when ICG decreased with light exposure, suggesting light-induced degradation [7,8]. This work aimed to unravel factors that affect ICG photobleaching, and consequently, contributing to cell killing. We quantified oxygen consumption during photobleaching to determine the impact of in situ oxygen during irradiation. Serum proteins, including lipoproteins and albumin, bind 98% of ICG as a mechanism to transport ICG in blood [6,9,10] with albumin having a binding constant of 10⁶ M⁻¹ [10]. In this work, we examined how albumin-induced ICG stabilization affects the ICG photobleaching process, and the effect of photobleached ICG on cell proliferation and viability.

Neuroblastoma (NB) is the most common extracranial solid tumor in children with a high mortality rate [11]. Surgery remains the best option due to the potential long-term side effects of chemotherapy and radiation-therapy [11]. During surgery, complete resection must be balanced with organ preservation, particularly for children when the tumor adheres to a vital organ, often leaving residual cancer cells (RCC) that increase the chance of recurrence and post-operative complications [11,12]. In this paper, we used NB as a model system to demonstrate that ICG can kill NB cells more effectively than normal fibroblast cells, and potentially be utilized as a bi-functional agent: an imaging probe under low-dose NIR irradiation, and a therapy reagent under high-dose irradiation to kill RCC owing to its unique phototoxic properties.

2. Materials and Methods

2.1. Cell Culture, Proliferation Assay and Viability Test

KELLY neuroblastoma cells were cultured in Corning RPMI 1640 (Millipore Sigma, Burlington, MA) supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (Fisher Scientific), 5 mL reconstituted oxaloacetate-pyruvate-insulin (Millipore Sigma), 10% Foundation B fetal bovine serum (Gemini Bioproducts, West Sacramento, CA), and 2 mM L-glutamine (Fisher Scientific). KELLY cells were incubated 24 hours at 37°C in 5% CO₂ in humidity to allow surface adhesion before ICG administration. IR-125 laser grade indocyanine green (ICG) was purchased from Fisher Scientific (Waltham, MA). Bovine serum albumin (BSA) was purchased from Millipore Sigma (St. Louis, MO).

For proliferation and viability tests, cells were seeded in a 96-well plate at a density of 12,500 cells/cm², and were incubated for 24 hours with aqueous ICG, photobleached ICG, or no ICG (control). 20 μL of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS; Promega Corporation, Madison, WI) was added to each well and incubated for 3 hours for each data point. MTS in complete RPMI absent of cells served as a blank for each timepoint. Proliferation data was normalized to the timepoint before treatment. Viability was determined by comparing MTS absorbance of treated cells by normalizing the absorbance of untreated cells tested simultaneously.

2.2. Cellular uptake of ICG

1 mg/mL (1.3 mM) sterile filtered ICG in deionized (DI) water was diluted to 75 μM into complete RPMI medium and administered to cells. 75 μM ICG was chosen because it was determined early in our research that it was comparably effective as 100 μM in the study by Urbanska et al. to kill cells [4]. The dependence of ICG uptake on incubation time was determined by absorption spectra and the calibration curve based on the absorbance of ICG dilutions, which were added to the lysate of cells with pre-determined cell number by MTS assay (see Supplementary Data). The intracellular concentration per cell was estimated by measuring diameters of rounded cells from photomicrographs immediately after trypsinization, assuming a sphere.

2.3. Irradiation and dissolved oxygen measurement of ICG

ICG solutions were irradiated with an 803 nm laser (Lingyun Photoelectronic System Co. Ltd, Wuhan, China) with total output range of 0.2 to 5.0 W for dissolved oxygen (DO), photodegradation, and photobleaching experiments. To examine ICG phototoxicity, cells were irradiated using 780 nm LED array (Thorlabs, Newton, NJ) at 18.2 mW/cm². Each cell sample was irradiated in sets of six for each trial for statistical analysis. DO was measured using a YSI 55 DO meter (YSI Incorporated; Yellow Springs, OH). To eliminate the temperature effect on DO, samples were placed in an ice bath immediately before irradiation.

2.4. Microscopy and spectroscopy of ICG detection

Intracellular ICG was imaged using a Leica Dmi8 microscope (Leica Microsystems, Wetzler, Germany) equipped with a Cy7 filter (Chroma Technology Corp, Bellows Falls, VT) and Leica DFC9000 sCMOS camera. ICG absorption spectra of recovered ICG supernatant from cell lysate and prepared ICG solutions were obtained with a DU 800 UV-Visible spectrophotometer (Beckman Coulter, Brea, CA). Spectra associated with DO studies were obtained using a 1 mL sample removed from the solution before and irradiation. ICG fluorescence was measured using an LS55 luminescence spectrophotometer (Perkin Elmer, Waltham, MA) with excitation of 400 nm and 780 nm.

2.5. Statistics and Software

One-way analysis of variance (ANOVA) was used to compare variance of the data groups. Tukey’s method was used for pairwise multicomparison of data groups. XLSTAT software (Addinsoft; Paris, France) was used through Microsoft Excel. Excel F and Student t test functions were also used. All statistics were completed at the 95% confidence level (α=0.05).

3. Results

3.1. ICG Cellular Uptake

Cellular uptake of ICG by KELLY cells was demonstrated in the fluorescence image of cells after 24 hours incubation with 75 μM ICG (Fig. 1A), revealing ICG accumulation in the cytoplasm outside the nucleus. This is in agreement with Abels et al., who reported that ICG accumulated in the cytoplasm, but not in cell membrane, nucleus, lysosomes, or mitochondria [13]. Relative intracellular ICG concentrations was not statistically different after 22 h incubation (Figure S1), supporting ICG uptake reaches saturation and is estimated at 2100 (±100) μM by recovered cell lysate. This is comparable with intracellular ICG concentrations reported as large as 5250 μM in cells after 24 hours incubation [13]. 24 hours incubation was used consistently in the rest experiments.

Figure 1.

ICG cellular update. (A) Fluorescence image illustrating ICG distribution in live NB cells. Scale bar: 50 μm. (B) Irradiation time dependence of KELLY cell viability determined by MTS. The cells were cultured in control medium (yellow), with effects of ICG alone (gray), irradiation alone (orange), and the combination of ICG and irradiation (blue). (C) Absorption spectra of ICG in the lysates of cells with (blue) and without (orange) irradiation at 780 nm at a dose of 16.4 J.

3.2. Effect of ICG Irradiation on Cell Death

To determine cell response, 75 μM ICG was added to cell culture for 24 hours and irradiated for 5, 10, 15, and 20 minutes (equivalent light doses of 5.46 J to 21.84 J) after ICG removal from medium and washes, followed by 24 hours culture before cell viability was measured. As shown in Figure 1B, irradiation or ICG alone did not affect viability, but the combination of light and ICG decreased cell viability exponentially as light dosage increased. 5 minutes of irradiation (~5.46 J/cm²) resulted in 23 (±5)% viability. Less than 3% of cells were viable after 10 minutes irradiation. The irradiation was also found to induce the decrease of ICG absorption in the cell lysate (Figure 1C). It suggests that ICG photobleached in the cells.

3.3. Effect of Oxygen in Photobleaching

75 μM ICG was irradiated at 540 J/mL to examine the role of oxygen in photobleaching. Sealed aliquots of solution were either irradiated under ambient aerobic (Air) or anaerobic (N₂-bubbled) conditions, or not irradiated (control). No noticeable change was observed by eye under anaerobic conditions, while dramatic color change was observed under ambient aerobic conditions (Figure 2A). These are consistent with the corresponding spectral changes shown in Figure 2B. By 10 minutes irradiation under ambient aerobic condition, the characteristic ICG absorption at 701 nm and 778 nm was reduced by 92.7% and 90.0%, respectively. Under anaerobic condition, however, the peaks decreased only by 10.5% and 10.7%, respectively. It suggests that the presence of oxygen is essential for ICG photobleaching.

Figure 2.

Effect of oxygen on ICG photobleaching. (A) Visual representation of color change of a 75 μM ICG solution with irradiation under nitrogen and ambient air compared to the initial solution (control). The irradiation was at a dose of 540 J. (B) Absorption spectra of the solutions in (A): control (blue), N₂-bubbled (orange), and ambient air (grey). (C) Spectral changes at increased light dosages of 0 J, 135 J, 270 J, 405 J, 540 J.

ICG absorbance decreased with light dosage (Figure 2C). Note that the peak at 701 nm changed rapidly with light dosage though the samples were irradiated near 778 nm. These two peaks are associated with ICG aggregates and monomers, respectively, and their light responses are discussed in the later sections.

3.4. Effect of BSA on DO Consumption in ICG Photobleaching

75 μM ICG solutions in the presence and absence of 1% BSA (%w/v) were irradiated with 600 J/mL while stirring. DO consumptions for ICG and ICG-BSA solutions were found to be 0.37 (±0.02) mg/L and 1.29 (±0.06) mg/L, respectively (Figure 3A). Thus, DO consumption was 3.5 times as effective as the solution without BSA, and the difference was significant (P = 2.21×10⁻⁸). Corresponding absorption spectra are shown in Figure 3B. Noticeably, irradiation resulted in the appearance of a small peak at 500 nm, accompanied by the decrease of the ICG peaks at the longer wavelengths. The 500 nm peak is also evident in Figures 2B and 2C along with an isosbestic point around 550 nm, implying a new chemical species evolved from the photoreaction. ICG solution shows two absorption peaks, associated with monomer (at 778 nm) and small aggregates (at 701 nm), respectively [4]. Irradiation induced the decrease of the 778 nm and 701 nm peaks by 70.1% and 75.9%, respectively. In the presence of BSA, a peak maximized at 803 nm with a shoulder around 725 nm was observed. Irradiation induced the decrease of the peak and the shoulder by 92.5% and 92.2%, respectively. Therefore, ICG in the presence of BSA is more photoreactive.

Figure 3.

Spectral and Dissolved Oxygen Changes with ICG Concentration and Irradiation. (A) Changes in DO concentration after irradiation with and without BSA in aqueous solution. Error bars show standard error. (B) Absorbance changes of ICG with (black) and without (red) 1% BSA, before (solid lines) and after (dashed lines) irradiation. (C) UV-Vis spectra showing the changes in absorbance with respect to total ICG concentrations which are normalized to 1 M. Top: without BSA; bottom: with 1% BSA. (D) Fluorescence changes in ICG with (black) and without (red) 1% BSA, before (solid lines) and after (dashed lines) irradiation. The excitation wavelength was 400 nm (left) and 780 nm (right), respectively.

3.5. Effect of BSA on ICG Monomer Stability

Absorption spectra for ICG at 75 μM and its dilutions with and without 1% BSA were collected, and the normalized spectra by the length of light path and the total ICG concentration are shown in Figure 3C. In the absence of BSA, the peak intensities increased at 701 and decreased at 778 nm with increased ICG concentration, consistent with the formation of small aggregates (SA) from monomers. This is supported by the quantitative peak analysis and spectrum reconstruction as shown in Figure S2. In a 75 μM solution, 47.9% of ICG is monomeric and the remainder forms SA. Such concentration-dependent changes were minimal in the ICG-BSA solutions, suggesting BSA effectively stabilized ICG monomers. Taken together with the results in Figure 3B, BSA stabilization boosted the photoactivity of ICG monomers. In the absence of BSA, the monomers and the SA co-present. Although the irradiation was at 780 nm, the peak of SA also decreased. We surmise that the irradiation induced decrease of monomeric ICG shifted the equilibrium between monomers and SAs in the solution, hence, promoted the dissociation of SAs to monomers.

Fluorescence changes in ICG were also examined. As shown in Figure 3D (right), ICG emission from 780 nm excitation was enhanced by 6.61 times in the presence of BSA, and the peak maximum shifted from 808 nm to 825 nm. Meanwhile, the fluorescence intensity of photobleached ICG (IPB) in 1% BSA remained 1.53 times higher than that of ICG in DI water, while the maximum of IPB in water decreased to 0.74 times ICG in water. Interestingly, 400 nm excitation (Figure 3D, left) revealed an emission peak at 441 nm of IPB in the presence of BSA, and the intensity is 58.3 times that of IPB in water. However, the intensity of non-irradiated ICG, either with or without BSA, was negligible. Therefore, the emission peak at 441 nm is attributed to the photoreaction products of ICG monomers.

3.6. Cell Response to Photobleached ICG

Cell proliferation and viability were determined for KELLY cells incubated with photobleached ICG (ICG-I) in comparison to those with non-irradiated ICG (ICG-N). As shown in Figure 4A, cells with ICG-N proliferated similarly to control cells, while cells with ICG-I changed very little over the course of three days. F and t tests were used to analyze data before ICG was given to cells and compared to 3 days after ICG was removed. The variance was insignificant (p>0.75) between groups, and each group mean was insignificant (p>0.39) on the day ICG was removed (two sided). All groups differed significantly (p<0.0001) comparing the third day after treatment to before treatment (one sided). Both the means of the control and ICG-N groups were insignificantly different (p = 0.051) on day 3 (two sided) but both were significantly different (p =6.7×10⁻⁷, 3.0×10⁻¹⁰ respectively) from the ICG-I group (one sided). Therefore, while ICG-I showed no killing effect on cells, it inhibited cell proliferation.

Figure 4.

Effect of photobleached ICG on cell growth. (A) Proliferation data of KELLY cells administered with photobleached ICG (blue) and non-irradiated ICG (orange) in comparison to that of untreated control cells (grey). ICG solution was at 75 μM. (B) ICG concentration-dependent killing effect on KELLY (black) and fibroblast cells (red) under the same light dosage of 16.4 J.

Figure 4B established the effective administered concentration for killing 50% of cells (EC₅₀), which was 15.8 and 36.5 μM for KELLY and normal human fibroblast cells, respectively. By comparing the slopes, KELLY cells responded more rapidly than fibroblasts. EC₂₀ and EC₈₀ show the range of effective doses and how narrow the range of response is. For KELLY cells, these values are 11.6 and 21.6 μM (range=10 μM), respectively; for fibroblasts, 26.3 and 50.6 μM (range=24.3 μM).

4. Discussion

ICG is bi-functional, employed as both a molecular imaging probe and a phototherapeutic reagent for killing cancer cells. In the latter case, photobleached ICG plays a prominent role. It is supported by the observation of ICG bleaching in cells upon NIR irradiation, and the photobleached ICG effectively prohibited NB cell proliferation. This adequately compensates the low quantum yield of ICG to form triplet-state and singlet oxygen for killing tumor cells in cancer treatment.[2,3]

BSA is a known ICG carrier and stabilizer in blood. We demonstrated that in the absence of BSA, ICG is present as a mixture of monomer and SA, denoted by two absorption peaks. The two forms of ICG can interconvert (Fig. 3C), suggesting the presence of a predominant oligomeric species. Zhou et al. reported similar observation [14]. Using a monomer-plus-single-oligomeric-species model, a tetramer model was found appropriate by performing a least-squares fit based on concentration-dependent absorption spectra of ICG [14]. Compared to SAs, the monomers are more photoreactive leading to bleached products, indicated by the new absorption and emission peaks. In the presence of BSA, ICG is predominantly in the monomer form. This is ascribed to stronger BSA-ICG interaction than ICG-ICG interaction, evidenced by the bathochromic shift and dramatic intensity change observed in both absorption and fluorescence spectra. Similar enhancement was observed when BSA was replaced by silk proteins (data not shown). It reveals that ICG inherently interacts with proteins strongly, likely due to its molecular structure that enables effective hydrogen bonding, electrostatic interaction and π-π interaction with various amino acids, the protein building blocks. BSA-stabilized monomers were more effectively photobleached to the photoproducts to reduce cell proliferation.

Oxygen is a requisite for ICG photobleaching. We surmise that ICG photobleaching is resulted from reaction of excited-state ICG with surrounding proteins or ICG itself. It is known that ICG binds to albumin’s site IIA pocket [15], which is dominated by hydrophobic residues and includes multiple electron donor residues, such as Lys195, Lys199, Trp214, Arg218, Arg222, His242, Arg257, and His288 [16]. The excited state of ICG is highly oxidative and undergoes single electron transfer reactions with these residues to produce ICG anion radicals, which further react with O₂ to generate superoxide anion radicals and other reactive oxygen species. These species can cause degradation of cellular components by damaging different biomolecules including proteins, DNA, and lipids leading to cell destruction and death, and kill both normal and tumor cells. They can also react with conjugated hydrocarbons such as those within ICG and other organic dyes, leading to photobleaching by disrupting conjugated bonds [6,17]. Note that the photobleached ICG reduced cell proliferation but didn’t induce apoptosis. With our observations that the ICG-mediated killing effect is more prominent for KELLY cells than normal fibroblast cells, the photobleaching products presumably prohibited tumor cell proliferation selectively through pathways, such as suppressing genes or kinases overexpressed in tumor cells and associated with cell growth and cell cycle [18–20]. For instance, α,β-unsaturated aldehydes in the photoproducts may act as soft electrophiles targeting cysteine residues as soft nucleophiles in a 1,4 Michael addition adduct reaction, or its carbonyl group acts as a hard electrophile to target harder nucleophiles like those in lysine, histidine, or nitrogen in pyrimidines and purines of nucleic acids. These reactions can cause cross-linking of protein residues or proteins to nucleic acids, resulting in the loss of their functions [21]. It is conceivable that the overexpressed kinases or genes in tumor cells are more susceptible to these reactions due to competitive equilibrium between the concentration of resultant adducts from the higher concentration of reactants and the concentration of deactivated species. This mechanism requires further investigation and will be undertaken in a next step.

The presence of BSA boosted both DO consumption and photobleaching. It provides further evidence that ICG monomers play essential roles in the reactions [4,14]. Therefore, the control of tumor cell proliferation by photobleached ICG limited the number of tumor cells in growth, giving rise to a highly potent overall killing effect of ICG even though its quantum yield to form triplet-state and singlet oxygen is lower than many other photosensitizers [1,2]. ICG responses were effectively controlled by oxygen presence, light dose, total concentration, and the presence of protein-binding.

Since ICG requires oxygen to induce cancer cells apoptosis and proliferation prohibition, its effect is limited in solid tumors due to the hypoxic environments [20–22]. However, its application in open surgery is unrestricted ascribing to the presence of ambient oxygen. During surgery, ICG can be effectively used as a safe optical probe at a low light dose to guide the surgery due to its favorable absorption and fluorescence in NIR range. When solid tumor is removed, residual cancer cells remain light-up and can be traced and killed by an increased light dose in a precisely defined area owing to the photoreactions induced by ICG. This is particular useful in children’s surgery, because organ preservation is as critical as complete removal of residual cancer cells when they adhere to a vital organ [11,12]. We have shown that ICG can kill NB cells more effectively than normal fibroblasts. Thus, the ICG guided imaging and light-activated treatment approach can be a useful tool in future surgery.

Highlights.

• Photobleached ICG inhibited neuroblastoma cell proliferation.

• ICG requires and consumes dissolved oxygen for photobleaching in cells

• BSA enhances photobleaching by preventing monomers from aggregating

• Monomers are more photoreactive than aggregates

• Lower ICG doses preferentially kill neuroblastoma cells over fibroblast cells