Neurotoxic effects of indocyanine green cerebellar granule cell culture viability study

: The aim of this study was to examine neurotoxicity indocyanine green (ICG). We assessed viability of primary cerebellar granule cell culture (CGC) exposed to ICG to test two mechanisms that could be the first triggers causing neuronal toxicity: imbalance in calcium homeostasis and the degree of oligomerization of ICG molecules. We have observed this imbalance in CGC after exposure to 75-125 μΜ ICG and dose and application sequence dependent protective effect of Gadovist on surviving neurons in vitro when used with ICG. Spectroscopic studies suggest the major cause of toxicity of the ICG is connected with oligomers formation. ICG at concentration of 25 μ M (which is about 4 times higher than the highest concentration of ICG in the brain applied in in-vivo human studies) is not neurotoxic in the cell culture. concentration. For 125 μ M ICG concentration differences in toxicity of pure ICG and the mixture were observed only for 10 mM Gad; the percentage of living cells increased from 60% to control level. For cells exposed with method 2 the percentage of living cells was similar to that observed for the cells simultaneously exposed to 75 μ M ICG and Gad. For 125 μ M ICG the percentage of living cells changed minimally and for none of the used Gad concentration did it reach the control level. For cell exposure using method 3 at both 75 and 125 μ M ICG concentrations the percentage of living cells changed to the control value for the all Gad concentrations used in the experiment. These results showed that the sequence of application of the tested substances was important for neuron viability, and especially the first application of Gad and subsequently of ICG provided better protection of neuronal viability than the other methods.


Introduction
Indocyanine green is a dye of high absorption in near infrared wavelength range which reveals fluorescence. This dye was used in several clinical applications for many years and its physicochemical properties were very well characterized by many studies. ICG can be classified as a low-toxicity contrast agent because advert reactions after its intravenous injection are rather rare [1]. It is commonly used in ophthalmology for staining blood vessels on the retina [2,3]. It was also clinically utilized for testing the liver function because the dye is extracted from the body by methabolisation in the liver [4][5][6][7]. ICG-tracking method was also proposed and tested as a tool for estimation of the blood circulation parameters in humans [8][9][10][11][12][13][14]. Considering these applications of ICG, the toxicity tests were carried out on retinal cell cultures [15] and retina model ex-vivo [16] showing cytotoxic effects in a dose, patients and time-dependent manner. Reports of these tests were reviewed by Jackson [17]. All studies of acute toxicity and neural tolerance were carried out on female mice and female and male Wistar rats [18].
In several studies indocyanine green was tested as a contrast agent for brain perfusion studies in humans. It was shown that optical monitoring of the inflow and washout of the dye allows assessing cerebral blood flow [19][20][21][22] and cerebral blood volume [22][23][24]. Recently, it was shown that advanced optical techniques can be used in monitoring of the inflow and washout of ICG in brain tissues, which constitutes a good basis for development of a clinical useful method, easy to apply in Intensive Care Units (ICU) in monitoring of brain perfusion in critically ill patients [25]. Although most of these studies were carried out in healthy humans, the technique based on ICG injection was tested also in stroke patients [26][27][28] in whom the status of blood-brain barrier (BBB) was unknown. Therefore, the examination of ICG toxicity in neurons is of great importance for the development of the ICG-based techniques which can be used in brain tissue perfusion assessment and monitoring in critically ill patients treated in ICU. One of the aims of our study was to determine the minimal ICG concentration that expressed neurotoxicity after 30 min exposure ex-vivo in a cellular culture of neurons. We also observe the physicochemical changes of ICG at various concentrations by measuring absorbance and nuclear magnetic resonance (NMR) spectra.
Additionally, the neurotoxicity of ICG used in combination with Gd-based contrast agent which is typically used in MRI-based brain perfusion studies was assessed. The relationship between contrast agent kinetics in Gd-mediated MRI and ICG-based optical monitoring studies is of importance when both techniques are compared or will be used together. These two contrast based techniques may provide complementary information on structure of the lesion (MRI) and temporal fluctuation of perfusion in the area with limited blood flow. The possible application of ICG in brain ischeamia monitoring was recently presented by Obrig et al. [29,30].
Safety of application of ICG in case of broken down BBB remains an important question for further development of the ICG-bolus tracking methodology allowing for brain perfusion monitoring. Furthermore, the neurotoxicity of parallel or sequential application of ICG with Gd-based contrast agent (Gadovist) was not yet studied. In this paper we report on results of the study focused on neuronal ICG toxicity used alone and in combination with a rather nontoxic contrast agent -Gadovist [31], which forming a the mixture with ICG may change its properties and toxicity. These studies were performed using cerebellar granule cells (CGC) culture model and spectroscopies -absorption and NMR. Considering combined use of both contrast agents it is necessary to know how they should be administrated, mixed or separately in specific order. We decided to initially test two working hypotheses of neurotoxicity of ICG: imbalance of calcium homeostasis and degree of oligomerisation. We assumed that these could be the first triggers that lead to many further disturbances of metabolic pathways in neuronal cell and finally cell death. The studies of these two hypotheses are presented in this paper.

Cell culture
Primary cultures of cerebellar granule cells were prepared from 7-day-old Wistar rats according to the method of Schousboe [32], with slight modifications as described previously [33,34]. The use of rat pups was in accordance with the ethical rules and was approved by the Local Care of Experimental Animals Committee and with the European Communities Council Directive. In our studies we used Lokce25 solution as a standard medium for short time experiments for CGC culture. This model was used only for detection potential toxicity and for investigation of their mechanisms. Experiments were carried out in ionic medium containing less components as compared to growth medium with serum. This procedure allows to exclude possible interaction of the ICG with nutrients and/or serum compounds.

Induction and evaluation of neurotoxicity by ICG dye and gadobutrol contrast medium
The neurotoxicity of ICG and Gad evoked by 30 or 60 min exposure was induced at DIV 7 by replacing the basal growth medium containing Earle's salt and 10% fetal calf serum (BME) with Locke25 buffer (134 mM NaCl, 25 mM KCl, 4 mM NaHCO 3 , 2.3 mM CaCl 2 , 5 mM HEPES (pH 7.4) and 5 mM glucose) containing freshly prepared solution of ICG or Gad.
Non-treated ICG or Gad neuronal cell cultures incubated for 30-60 min in Locke25 medium were used as control for toxicity tests. The examined contrast media were tested either alone or combined, as required. After incubation at 37°C for 30 or 60 min, a short-term exposure to ICG-Gad mixture was terminated by two washes with Locke buffer. Then the cells were cultured in the original growth medium for 24 h. The cells were then fixed with 80% methanol, stained with 0.5 μg/ml propidium iodide, and the viable and dead neurons were counted using a fluorescence microscope (Zeiss-Axiovert, Germany). Cells that in the time of fixation were live differ morphologically from dead ones. Results were expressed as percentage of live cells in proportion to all cells.
In order to test the interactions between ICG and Gad, the concentrations of Gad were 0.1, 1 and 10 mM (0.1, 1 and 10 fold of the human dose) and those of ICG were 25, 75 and 125 μM [35].
We used three combinations of exposure: 1. Gad was added to Locke25 buffer simultaneously after ICG in all Gad and ICG concentration combinations; 2. Gad was added to the Locke25 buffer after 30 min exposition to the ICG; 3. Gad was added alone for 30 min, and ICG was added for further 30 min. Additionally, we tested the mortality of the neuronal cells after 60 min of incubation with the tested compounds. The viability of CGC was assessed 24h later.

Measurement of calcium concentration
Calcium ions were measured using fluorescent dye Fluo-3, and Fluo-3AM, a cell-permeant acetoxymethyl ester (AM) derivative of Fluo-3. The extracellular calcium level in Locke25, without cells and with cells culture, after addition of ICG, Gad or the mixture of both substances, was measured using 0.1 µM Fluo-3 and recorded by FLUOstar Omega (Germany) microplate reader at 485 nm excitation and emission with cutoff filter at 538 nm. The intracellular calcium concentration in CGC (1x10 6 /well) was assessed using Fluo-3AM. The neurons were loaded with 4 μM Fluo-3AM at 37°C for 30 min in the original growth medium. The loading was terminated by washing the cells three times with the Locke5 buffer. Changes in fluorescence after addition of the tested compounds (ICG, Gad or MK801 -an antagonist of NMDA receptors, were recorded every 1 min, over a 30 min period, using a microplate reader as described above.

Physicochemical studies of ICG and Gad
Visible-near infrared, Vis-nIR, and NMR spectroscopies were used for studies of interactions between ICG and Gad in water and Locke25 to reflect the CGC neurotoxicity examination. NMR spectroscopy allows monitoring protons in the molecule, and absorption spectroscopy monitoring changes in the structure of the molecules in solvent. NMR experiments in Locke25 were not collected because many signals from ICG were overlapped by high glucose signals.
Each sample was prepared directly from "stock" ICG in water in the spectrophotometer cuvette, just before the absorbance or NMR measurements. Vis-nIR absorption spectra were recorded using Perkin Elmer Lambda 950 spectrometer at 25°C, in polymethyl methacrylate (PMMA) 5mm path length cuvettes. The spectra were collected three times at 30 min intervals, in range from 500 to 900 nm with 1 nm data interval and scan speed of 110.5 nm/min.
The NMR spectra were collected at 25°C on a Varian Inova 400 NMR (Varian Inc., USA) spectrometer operating at a proton frequency of 399.96 MHz. In 1 H NMR examination, presaturation pulse sequence was applied. The measurement experiment consisted of 1024 repetitions. Pulse repetition time was set to 10s and line broadening of 0.5 was applied. NMR spectra of ICG were collected only in D 2 O. Trimethylpropionate sodium salt (TSP) was added to each sample as an internal standard. The measurement time was 4 h.
Statistical analysis was performed using non-parametric Mann-Whitney U test and pvalues lower than 0.05 was considered to be significant (Statistica v7, StatSoft Inc.).

Neurons viability study
On the seventh day of cell culture neurons were incubated with the solution of ICG in concentrations from 25 to 250 μM. The duration of the exposure of the neurons to the ICG was estimated on the basis of pharmacokinetic data described by [36]. In Fig. 1(a) the viability of the neuronal cells after 30 min exposure to ICG was presented. In a series of longer tests, after 24 h of exposure a slight decrease in neuron viability was observed at 50 μM ICG; at concentrations higher than 75 μM was statistically significant and drop from 61 to 34%, respectively as compared to 94% for control. Gad neurotoxicity, for 0.1, 1, 10 and 50 mM concentrations, was not observed in the course of the experiment ( Fig. 1(b)).

ICG oligomerisation study
For three ICG concentrations further absorption and NMR experiments were carried out. In NMR spectra ( Fig. 2(a)) differences in ICG structure can be observed: formation of oligomers and complexes of unknown structure (J-or H-aggregates) [37][38][39]. During NMR experiment the structure of the ICG molecules changed and we obtained spectrum with averaged signals of all ICG molecules in oligomers. These changes of ICG structure lead to the changes in absorption spectra of ICG (see Fig. 2(b)). Analyzing NMR spectra, it can be noted that signal intensities from oligomers, at 0.5, 1.0 and 1.3 ppm, increased with increasing concentration of ICG. It can be observed, that all signal intensities and shapes depend on ICG concentration and molecular aggregations. We have analysed the content of the monomers of ICG in the H 2 O solution by derivation of monomer to oligomers ratio from the absorbance signal. In this analysis the amplitude of the spectra at 775 nm was considered as a measure of concentration of the monomers and the absorbance at wavelength of 705nm was used as a measure of concentration of oligomers [40]. It can be noted that ICG at 25 μM concentration in water is almost monomeric. According to the study by Zhou et al. [39], for this concentration of ICG 90% of the ICG is present in monomeric form. This value was used as a reference for calculation of the percentage of monomer in our analysis. Only for 25 μM ICG in Locke25 a small increase in the monomer fraction was observed. The increase was minimal for dye concentrations of 75 and 125 μM in Locke25 as well as in water, after 30 as well as 60 min of the experiment. The monomer fraction decreased distinctly with increase of ICG concentration in H 2 O, whereas it decreased only minimally in Locke25 (Fig. 3).

Neurotoxicity study of ICG and Gadovist mixture
The neurotoxicity study of ICG-Gad mixture was limited to concentrations of ICG 75 and 125 μM for which neurotoxicity of ICG has been observed. Figures 4(a)-4(c) shows the viability of cells after their exposure to ICG-Gad mixture. For 75 μM ICG the addition of Gad using method 1 increased the percentage of living cells from 60% of level with ICG alone to 85%, with a small increase with rising Gad concentration. For 125 μM ICG concentration differences in toxicity of pure ICG and the mixture were observed only for 10 mM Gad; the percentage of living cells increased from 60% to control level. For cells exposed with method 2 the percentage of living cells was similar to that observed for the cells simultaneously exposed to 75 μM ICG and Gad. For 125 μM ICG the percentage of living cells changed minimally and for none of the used Gad concentration did it reach the control level. For cell exposure using method 3 at both 75 and 125 μM ICG concentrations the percentage of living cells changed to the control value for the all Gad concentrations used in the experiment. These results showed that the sequence of application of the tested substances was important for neuron viability, and especially the first application of Gad and subsequently of ICG provided better protection of neuronal viability than the other methods. We have collected NMR spectra of ICG-Gad mixture in D 2 O and Vis-nIR absorption spectra of the mixture in Locke25. As described above, three methods of exposure of ICG-Gad mixture were used.
We have measured the spectra of absorption of ICG-Gad mixtures in Locke25 (Fig. 5). The results, presented as percentage of monomer of dye in the solution for 25 μM ICG and Gad concentrations from 0 to 10 mM were not significantly dependent on the exposure method. For ICG concentration of 25 μM no neurotoxicity was observed. The percentage content of monomers in 75 μM ICG for 0, 0.1 and 1 mM of Gad was not significantly different for all three exposure methods. Only for Gad concentration of 10 mM a significant rise in the monomers content was observed. Similar changes in the percentage of monomers were observed for ICG concentration of 125 μM. Analysis of the dependence of monomer content on toxicity revealed that for 49 ± 0.6% content of monomers neurotoxic effect can be observed but for 55 ± 5.7% no toxic effect was noted. The influence of Ca 2+ ions on the ICG structure in water was monitored for the lowest concentration of ICG -25 μM (Fig. 6). The addition of 2.3 mM Ca 2+ to the aqueous ICG solutions induced changes in its NMR spectra, related to changes in the dye structure or aggregation. There was no more predomination monomer on of oligomers, which occurred mainly at 125 μM ICG concentrations. The shapes of the visible absorption spectra of ICG in 2.3 mM Ca 2+ aqueous solution and in Locke25 are very similar to each other, in contrast to the dye spectrum in water only ( Fig. 6(a)). NMR spectra of ICG for both solvents, H 2 O/D 2 O and H 2 O/D 2 O with Ca 2+ , also were different. In water Ca 2+ of 25 μM ICG we observed neither NMR signals presented in high ICG concentration (oligomers) nor signals from monomers in water without calcium ions ( Fig. 6(b)). This suggests that other forms of ICG occurred in water with Ca 2+ ions. These forms could be similar to the forms which occur in Locke25.   Analysis of the results obtained from Vis-nIR absorption and NMR spectroscopy led to the hypothesis that Gadovist has neuroprotective properties due to its prevention of interactions between calcium ions and the ICG molecules in the extracellular space. Therefore, we performed measurements of calcium ions level in the studied ICG-Gad mixture in the absence of cells, initially with Fluo-3, a fluorescent calcium ions indicator. Figure 7 shows the changes in the calcium indicator signal in ICG mixture with Gad in Locke25 solvent. Addition of 75 μM ICG to the Locke25 solvent decreased the fluorescence signal indicating calcium concentration to 50% of the initial value. Gad in the Locke25 solution, independently of its concentration had no effect on the calcium concentration. The mixture of ICG-Gad in the Locke25 solution changed the calcium concentration to the same level as only ICG. Based on this result we tested how the changes in extracellular calcium ions imbalance reflect on intracellular calcium homeostasis in CGC. At the same time, we tested changes in extracellular calcium concentration in the presence of CGC and 25, 75 and 125 µM ICG, using Fluo-3 as calcium indicator. Changes in intracellular calcium level were monitored with the fluorescent probe Fluo-3 AM. As presents in Application at 5 min of 25, 75 and 125 µM ICG solution resulted in a dose dependent rapid decrease in the measured Fluo-3 fluorescence signal from 100% to 62, 52 and 44%, respectively (see Fig. 8(a)), as mentioned above. Figure 8(b) shows that addition of 25-125 µM ICG resulted in the first 8 min of experiment a dose dependent increase in fluorescence, reflective intracellular calcium concentration at 35% above control level at maximal used concentration. After 8 min this increase is still above the control, but independent of ICG concentration. Simultaneous or 30 min prior application of 0.1-10 mM Gad did not reduce this increase (Fig. 8(c)). These outcomes indicate one of possible mechanisms (calcium homeostasis imbalance) of ICG neurotoxicity. To answer the question regarding the nature of observed intracellular increase in calcium concentration -either from intracellular calcium stores or from extracellular space or both -we used MK801, an antagonist of NMDA receptors, which blocks the main gate of calcium influx ( Fig. 9(a)). Additionally, we performed experiment using 45 Ca 2+ to monitor calcium uptake ( Fig. 9(b)). Results presented in Fig. 9(a) show that 0.5 µM MK801 has no effect on increase in intracellular calcium level after addition of 25-75 µM ICG. This indicates that the observed changes are connected to calcium release from intracellular stores. Similar results were obtained in experiments with radioactive calcium, presented in Fig. 9(b). In this case at all used concentrations of ICG the percent of radioactive calcium in CGC is less than in control. This phenomenon reflects the situation when ICG binds calcium outside the cells and less 45 Ca 2+ can be taken up by the neurons. This decrease should be dose dependent, but probably at the same time small amount of calcium was taken up by the cells (but it is still under control level). Addition of MK801 in two cases (25 and 75 µM ICG) confirmed it. To sum up, in our experimental condition ICG at 25-125 µM resulted in an increase in intracellular calcium concentration derived primarily from inner calcium stores. This effect supports previous finding that the Ca 2+ ions concentration is one of important factors influencing neurotoxicity of ICG.

Discussion
Study of neuronal viability in cell culture is the main method of evaluation of toxicity of tested compounds. In our experiments, the primary cerebellar granule cell culture model was used. CGC are the main, homogenous population of neurons in mammalian brain. Their development is well known and therefore culture of these neurons is the model of choice for the study of many cellular and molecular mechanisms, which lead to cell viability or mortality, especially to neurodegeneration and neuroprotection [41]. The culture consisted of 90-95% neurons and 5-10% astrocytes, which were necessary to keep the cell culture alive.
Considering realistic concentration of the ICG in the brain tissue during passage of the dye we analyzed the recommended quantity of the ICG-PULSION (PULSION Medical Systems AG, Germany) which depends on the examination conditions and age of the subject. For various types of measurements (liver function testing, imaging in ophthalmology, cardiovascular hemodynamic measurements) the manufacturer recommends doses of 0.1-0.5 mg/kg per measurement and daily dose which should not exceed 5 mg/kg. Similar doses are recommended by other ICG manufacturers Akorn (IC GREEN, US) and Sigma-Aldrich (Cardiogreen, US). For these doses of ICG injected intravenously the molar concentrations of the ICG reaching the brain tissue can be estimated according to the algorithm described in detail in Liebert et al. [27]. These recommended doses of ICG administrated in adults correspond to 1.15-5.16 μM of ICG concentration in the brain tissue. Furthermore, it can be found in earlier publications that the typical ICG dose per measurement used in in-vivo experiments does not exceed 20 mg (corresponding to 2.58 μM concentration of ICG in the brain) [26,27,[42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59]. However it can be also found that some groups used 25 mg (3.83 μM in the brain) in a single dose [60][61][62][63][64] and even higher doses of ICG 0.5 mg/kg were reported in studies on humans [65][66][67]. The highest dose of ICG which was reported in human studies was 50 mg (7.68 μM in the brain) [68].
In the study by Keller et al. the process of ICG extravasation in the brain caused by disruption of the blood brain barrier was investigated [40]. It was shown on a rat model that the blood brain barrier disruption did not lead to accumulation of the ICG in the neuronal tissue. However, the neurotoxic effects caused by the dye were not analyzed in that study. The observed by Keller et al. disappearance of the ICG in the brain after period of 2-6 h after ICG injection in rats with blood brain barrier disruption may suggest that the direct contact between ICG and the neurons is limited to short periods of time. Therefore, our results showing neurotoxicity in-vitro within the 30 min incubation time may be used to estimate neurotoxic effects expected in-vivo after intravenous injection of the dye in patients with blood-brain barrier disruption.
Considering above discussed range of typical concentrations of ICG in the brain following intravenous injections we analyzed neurotixic effects for ICG concentrations of about 4, 10 and 25 fold the concentration in the brain related to the maximal reported dose of ICG injected in the human studies e.g. 25 μM, 75 μM and 125 μM, respectively. Experiments carried out in CGC showed that the doses related to the concentration of 75 μM become to be neurotoxic and the concentration of ICG of 25 μM in the brain is still safe for the neurons. In our study the neurotoxicity experiments were performed for ICG in ionic Locke25 medium without serum. In the presence of proteins many toxic agents are less toxic than in the Locke medium. Also ICG should be less toxic when bound to the albumins as it is expected in invivo applications in which the ICG is dissolved in blood [69][70][71]. Thus, the assessment of the toxicity limiting concentration which was done in the present study is reasonable.
It has been known that ICG decomposes into toxic waste materials under the influence of UV light, creating a number of still unknown substances [36]. It was reported that ICG has a toxic effect on retinal cells and that this effect depends on the ICG concentration [72,73]. ICG may pose a risk to specific tissues if it is able to leave the vascular space. However, if contained within the vascular space, ICG would very likely be with limited biological effect. Any use of ICG outside the vascular space should be carefully considered, and any condition that would allow for significant leak of ICG into an interstitial fluid space becomes a possible contraindication to its use [74].
Attempt to assess reasons of the neurotoxic effects of ICG have confirmed imbalance in calcium homeostasis after exposure of neurons to the dye. Changes observed in intracellular calcium level can in some but not all cases lead to cell death. It was found that for concentration of ICG 25 µM an increase in fluorescence can be observed without cell death. Such situation can be observed after stimulation of CGC with low doses of various toxic compounds [75]. We also observed correlation between changes in physicochemical properties of ICG based on oligomerization and the probability of death of neurons. However, it should be considered that the oligomerization effects may have limited influence on neurotixicity of the ICG in typical in-vivo applications in which it is bound to albumins after injection to blood stream [69][70][71].
Gadobutrol has not toxic effects in doses used for clinical examinations. This was investigated by Takahashi [76], Ray [77], Wack [31]. Studies on primary cultured rat cortical neurons indicated toxic influence of gadolinium chloride on cell viability and connected it with dysfunction of the mitochondria followed by production of free radicals, decrease in ATP level and release of cyt C which led to cell death [78]. Cortical neurons are considered to be more sensitive than granular cells. Gadolinium chloride cannot be compared with gadobutrol used in our studies because Gd 3+ in gadobutrol was specially surrounded by other molecules to avoid the toxic properties of gadolinium.
In the present studies we did not observe any toxic effect of Gad on primary neuronal culture even at a very high concentration of 50 mM. Moreover, we observed that the mixture of Gad and ICG reveals less neurotoxic effect on the CGC than the ICG itself. This effect is dose dependent and depends on the sequence of Gad administration.
The results of our studies showed that the mechanism of protection through Gad is still unclear and needs further investigation. This is not simply stabilization of calcium level, because no effect has been observed on extra-or intracellular calcium concentration, even in cells preincubated with 10 mM Gad ( Fig. 7 and Fig. 8(c)). Calcium ions control many fundamental metabolic pathways in the neuron, i.e. catabolic decomposition of lipids and proteins. Through effect on lipids and the state of proteins in membranes, calcium can modify the function of intracellular organelles or the cytoskeleton, bind in a directly to some proteins or interact on their phosphorylation or polymerization [79]. Destabilization of extra-and/or intracellular calcium homeostasis may lead to disturbance in calcium-dependent processes in the cell, which can result in their death.
The study of the combined use of Gad and ICG was motivated by the possible sequential or combined application of both contrast agents in humans for brain perfusion assessment. Such tests may allow to understand differences in pharmacokinetics of both compounds and to use Gad-and ICG-tracking methods as complementary tools for clinical brain perfusion imaging (based on tracking Gad in MRI) and monitoring (based on tracking ICG by NIRS). It was found that the Vis-nIR spectrum of Gad within the observed spectral range 500-900 nm, was not detectable with the equipment applied. In analogy, the ICG cannot be found in NMR spectra because the spectral components related to composition of ICG overlap with many chemical components of the tissues. These facts lead to a conclusion that the comparison of pharmacokinetics of these contrast agents cannot be done without application of both modalities (MRI and NIRS). However, the tests of neurotoxicity of the mixture of ICG and Gad suggest that use of both contrast agents in typical doses is rather safe even in patients with disturbances of blood-brain barrier and these contrast agents can be used in combination or in sequence without increase of the risk of neurotoxicity.

Conclusion
It was found that indocyanine green reveals toxicity for neuronal cell cultures in concentrations higher than 75 μM. Results of our experiments showed that this toxicity is related to imbalance in extra-and intracellular calcium homeostasis. Our studies confirmed that this phenomenon is the first trigger of ICG neurotoxicity. Spectroscopic measurements and neurotoxicity data suggest that oligomers are more toxic to neurons than monomeric ICG molecules. We conclude, that degree of oligomerization is the second factor influencing toxicity of ICG. Our results suggest that considering neurotoxicity the combined use of Gadovist and ICG, 30 min after the use of Gadovist the ICG dose should be administered. In this sequence of application Gadovist has a neuroprotective effect on neuronal cells.
Results of our study showed that ICG and Gadovist can be administrated in animals and humans in their typical diagnostic concentrations with no toxic effects on neuronal cells. This conclusion suggests that application of both contrast agents is safe for patients with broken down blood-brain barrier in which extravasation of the compound in the brain can be expected. Also ICG used in combination or in sequence with Gadovist in their diagnostic concentrations has no toxic effects on neuronal cells. Nevertheless, one has to consider the potential toxicity of the studied mixture for other organs, e.g. kidneys or liver.