C60-Fullerenes: detection of intracellular photoluminescence and lack of cytotoxic effects

We have developed a new method of application of C60 to cultured cells that does not require water-solubilization techniques. Normal and malignant cells take-up C60 and the inherent photoluminescence of C60 is detected within multiple cell lines. Treatment of cells with up to 200 μg/ml (200 ppm) of C60 does not alter morphology, cytoskeletal organization, cell cycle dynamics nor does it inhibit cell proliferation. Our work shows that pristine C60 is non-toxic to the cells, and suggests that fullerene-based nanocarriers may be used for biomedical applications.


Background
Recent advances in materials science have fueled tremendous interest in numerous potential biomedical applications of various nanomaterials. For example, fullerene C 60 molecules are unique for their multi-functional uses in materials science and optics [1][2][3][4], and are considered for a variety of biological applications (reviewed in [5]), such as imaging probes [6], antioxidants [7][8][9] and drug carriers (taxol) [10]. Our laboratory is interested in exploring whether novel multifunctional nanoparticles can be designed for cancer therapy and diagnosis. Realization of such a goal requires a better understanding of the interactions between nanoparticles and cells and it is important to determine whether or not the particles by themselves impact cell growth and differentiation. We have chosen C 60 for initial studies because the established chemistries afford us the flexibility to couple various biologically interesting and relevant molecules.
However, some undesirable properties of C 60 present specific challenges. For example, due to its inherent hydrophobicity, C 60 is poorly soluble and naturally forms large micron-sized clusters in aqueous media. Therefore, organic solvents are routinely used for solubilization of C 60 [11] Consequently, cell biological studies with pristine C 60 have been limited.
Whereas chemical conjugation of C 60 to various water soluble molecules improves the overall aqueous compatibility, pristine C 60 is routinely dissolved in toluene [12,13], tetrahydrofuran (THF) [14] or other organic solvents, and then exchanged into water by extracting the organic phase with water. The resultant preparation is often referred to 'water soluble C 60 ' which is typically of light yellow color and is estimated to contain a few hundred micrograms of C 60 /ml [15]. It has been suggested that the aqueous C 60 is toxic to cultured cells and the toxic effects are due to per-oxidation of lipids in cell membranes [16][17][18][19]. Various groups have reported that C 60 (prepared using different methods) is not toxic [20][21][22][23][24] and some have attributed the toxicity of C 60 to the side chains present on the functionalized C 60 [25]. Possible mechanisms that might contribute to the observed toxicity of nano C 60 , include the solvent effects like atmospheric exposure of solvents such as THF (according to the manufacturer). Additionally, acquisition of ionogenic groups upon C 60 crystal formation in aqueous media via THF solvent exchange have been reported to contribute to the potential biological consequences [26]. In support of these possibilities, a recent study suggests that toxicity of THF-derived water soluble nano C 60 is abolished by removing THF by γ-irradiation. [27].
The conflicting data on cytotoxic effects of C 60 merits attention and requires a resolution if these materials are to become biologically useful. The following simple hypothesis may reconcile with the mutually contradictory data on the cytotoxic effects of pristine fullerenes. C 60 undergoes modifications during the preparation of water soluble C 60 , and such changes are responsible for the cytotoxic effects. Whereas the precise nature of such modifications is unknown at present, the hypothesis can be tested and the effects of C 60 can be unequivocally examined if C 60 can be applied to cells in such a way that obviates the need of preparing water soluble C 60 .
Studies presented in this manuscript examine the key issue of observed cytotoxic effects of C 60 in cultured normal and malignant breast epithelial cells. We have developed a new, yet simple, method to directly apply C 60 to cultured cells by modifying an established cell biological technique used in anoikis studies [28,29].
Although several key properties of fullerenes, such as the characteristic photoluminescence (PL) of C 60 are well characterized in solutions [30] and polymer complexes [31], few have examined such properties in cellular environment. Photoluminescence of crystalline C 60 occurs due to coupling of the vibrational modes of the lattice with electronic transitions and the PL signature of fullerene crystals may be useful to track the presence of C 60 . Results presented in this work demonstrate that unmodified C 60 crystals are taken up by cells and intracellular C 60 retains its optical properties, as determined by measurements of PL. Significantly, our studies reveal that C 60 prepared by a variety of methods up to 200 μg/ml is not toxic to a number of cell types.

Results and discussion
To eliminate the use of toxic organic solvents for applying C 60 to cells, we have adapted methods routinely used in cell culture studies involving polymer coating of tissue culture dishes following solvent evaporation [28,29,32]. Colloidal suspensions of C 60 in methanol (0.2 mg/ml) were prepared by sonication as described in Materials and Methods and applied to tissue culture dishes as an uniform coating. The organic phase is allowed to evaporate in a tissue culture hood, which leaves behind a coating of C 60 on the dish. Cells are plated on to these dishes of C 60 . The C 60 plated using this technique requires minimal manipulation and does not contain harsh organic solvents in cell culture. We refer to this preparation of C 60 as 'methanol C 60 .'

1) Properties of methanol C 60
Sonication in methanol produces a uniform suspension of C 60 , which takes approximately 10-30 minutes to settle out of suspension. This slow rate of settling allows adequate time for recording of absorption spectra. Methanol C 60 is a light brown colored suspension, indicative of large crystals in supension, compared to purple suspensions of toluene C 60 which are known to contain significantly smaller sized crystals ( Figure 1A). To characterize the physico-chemical properties of methanol C 60 , we determined its spectral features and measured the particle sizes of the colloidal suspensions of C 60 in methanol. For example, C 60 has a characteristic triplet-triplet absorption spectrum at 350 nm [33][34][35]. The absorption spectra of C 60 in methanol was comparable with that prepared in toluene (λ max = 337 nm), which is more commonly used for suspending C 60 ( Figure 1B). C 60 exhibits a characteristic reddish orange PL signature in the solid state with a peak at 735 nm [31,36,37]. Methanol C 60 retained this key property that is dependent on the interstitial spacing between C 60 molecules in the crystalline structure with a broad peak around 750 nm ( Figure  1C). These spectral findings are consistent with the established behavior of C 60 , which exhibits slight shifts in the absorption and PL peaks dependent upon the temperature [36] and the solvent used to disperse C 60 [13]. Consistent with the properties described above, methanol C 60 suspensions, when applied to tissue culture substrata, exhibited readily detectable crystal sizes and marked PL when visualized by light microscopy (discussed in the next section). Together, these data suggest that C 60 remains adequately suspended in methanol and that the spectral characteristics are similar to those prepared in other organic solvents.
Particle size measurements confirm the stability of methanol-C 60 suspensions. Dynamic laser light scattering measurements show that toluene C 60 , used as a reference ( Figure 1D), yields uniformly sized particles with a mean size of 32.7 nm, consistent with published data [18,38]. Parallel measurements with methanol C 60 reveals two peaks at 106 nm and 342 nm size, which indicates heterogeneity in the particle size ( Figure 1D).
Transmission electron microscopy (TEM) was used to verify cluster sizes of fullerenes dried from methanol ( Figure  1E). Methanol C 60 clusters were observed in a wide range of sizes including large clusters in the micron range although many clusters smaller than 10 nm were observed. TEM micrographs corroborate particle size data obtained by dynamic light scattering which indicates the presence of a heterogenous mixture of variably sized clusters. Furthermore, following evaporation of methanol, the majority of fullerene clusters do not reaggregate, and have a range of sizes of tens of nanometers, although some larger clusters also exist. TEM data differ from that of the dynamic light scattering results in this regard since the light scattering apparatus accounts for the average of all sizes of fullerene clusters in solution.
Prolonged sonication of C 60 in various organic solvents is routinely employed to prepare solutions of C 60 [13,39]. As an additional measure to ascertain that suspension and sonication of C 60 in methanol has not introduced any modifications into the fullerene, we analyzed each preparation by matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry.
These analyses, performed in the positive ion mode, revealed a predominant species with a monoisotopic mass at 720.1 Da (theoretical mass of C 60 = 720.00 Da) indicative of C 60 preparations in methanol and toluene ( Figure 2). The observed mass is consistent with the formation of a positively charged C 60 ion by loss of an electron instead of gain of a proton. Interestingly, the same mass was observed upon analysis in the negative ion mode (data not shown). Each of the preparations contained a small amount of a species at 489.64 Da that was present in the original preparation of C 60 . In all cases, the principal component was pure C 60 with mass 720.1 Da. The method of preparation in methanol or water used in this study does not appear to significantly alter the structure of the C 60 .

2) Growth of cells in presence of methanol C 60
Previous studies have suggested that water soluble nano-C 60 compromises the integrity plasma membrane, possibly due to lipid peroxidation [19]. To determine whether C 60 applied to cells by a different method would produce a similar toxic effect, we have tested the effects of methanol C 60 on cultured cells. First, we have examined whether C 60 crystals are taken up by cells.
Normal (MCF10A) and malignant (MDA MB 231 and MDA MB 435) breast epithelial cells were plated on either methanol-C 60 coated dishes or control dishes and cellular morphology of the attached cells was examined. The presence of methanol C 60 did not alter cell morphology or cell spreading and the PL signature of C 60 is retained under normal conditions of cell culture. Further, we found that crystalline C 60 is taken up by cells. To ensure that the nanoparticle is indeed internalized, the cells were trypsinized with trypsin to release them from the plate and replated on dishes coated with collagen I to enhance integrin-extracellular matrix interactions and cell spreading. Morphologically, cells cultured with methanol C 60 re-attached and spread like the control cells. The fullerene nanocrystals retained their reddish orange PL, under phase contrast ( Figure 3A) and bright field imaging used to ensure that the color of fullerenes is not due to an artifact of phase contrast.
The presence of intracellular C 60 crystals was verified via examination through multiple focal planes using confocal microscopy. Normal breast epithelial cells (MCF10A) cultured overnight on methanol C 60 were trypsinized, replated on collagen I, fixed in paraformaldehyde, extracted with 0.1% Triton X-100 and stained with FITClabeled phalloidin for counterstaining. C 60 crystals were readily evident by their characteristic reddish orange PL signature ( Figure 3B). Multiple crystals of C 60 of varying sizes were present in different focal planes, indicating their intracellular localization. Initial examination shows that intracellular C 60 does not interfere with cell spreading on ECM or alter microfilament reorganization following attachment to ECM. Untreated (control) cells, processed in parallel, on the other hand, do not exhibit orange PL. Similar results were obtained with MDA MB 231 and MDA MB 435 breast cancer cells (data not shown). Since cytoskeletal reorganization following integrin activation involves a series of complex signaling events beginning with integrin activation and orchestrated activation of Rho GTPases [40], our results suggest that treatment of C 60 is unlikely to interfere with the events following cell-ECM interactions.

3) Cell survival in presence of pristine C 60
As discussed in the Introduction, there is a lack of consensus on the effects of C 60 on cell growth, and we have hypothesized that the apparent cytotoxic effects of the nanoparticle are due to the methods of preparation and application of C 60 to cells. Therefore, we have reassessed the effects of C 60 on cell proliferation using methanol C 60 and water soluble nano-C 60 prepared from toluene.
Several normal and malignant breast cancer cells were plated on tissue culture dishes pre-coated with various amounts (ranging from 10-200 μg (10-200 ppm) which corresponds to 13 nmoles to 277 nmoles) of methanol C 60 . Contrary to the published results which state that C 60 is toxic at 20 ppb [18], culturing cells with significantly higher (200 ppm) concentrations of C 60 did not adversely impact cell proliferation (Figure 4). The growth and proliferation of MCF10A ( Figure 4A), MDA MB 231 ( Figure  4B) was not affected by the presence of C 60 and no cytotoxic effects were observed. Similar results were obtained with MDA MB 435 and HepG2 cells (see Additional file 1). Lack of toxicity of C 60 on MDA MB 231 cells was further confirmed by 'live-dead' cell assays (Molecular Probes) ( Figure 4C). Further, cell cycle profiles of MDA MB 231 cells cultured with or without C 60 were essentially identical, indicating that the overall cell cycle parameters were unaltered ( Figure 4D), and no subG o -G 1 fractions (indicative of apoptotic populations) were evident in cells treated with C 60 (not shown).
Our finding that culturing cells with methanol C 60 does not inhibit cell proliferation is at variance with published results [16,18,19,41], and hence we investigated whether the different methods of preparation and application of C 60 would explain the differences in the effects of C 60 . We have prepared water soluble nano-C 60 from toluene, using the published protocols [12,13] and characterized the material. Nano C 60 prepared from toluene yielded 274 μg/ ml (274 ppm) of lightly yellow colored water-soluble C 60 . Absorption spectra ( Figure 5A) of nano C 60 are in agreement with established spectral properties of C 60 [33,35]. The particle size measurements of nano C 60 revealed the presence of crystals with an average size of 122 nm ( Figure  5B).
Culturing of MCF10A and HepG2 cells with up to 27.4 μg/ml (27.4 ppm) of water soluble nano C 60 derived from toluene had no effect on cell proliferation (Figures 5C &  D). The lack of cytotoxic effects was confirmed by two different assays (crystal violet staining and live-dead cell assays) and cell cycle analyses. The amounts of C 60 used in these experiments is comparable to those used in previous studies where extreme toxicity was reported with other water soluble nano C 60 preparations [18,19]. Thus, our findings with methanol C 60 and water soluble nano C 60 prepared from toluene demonstrate that cell proliferation is not inhibited by fullerenes and the nanoparticle does not exert toxic effects in cell culture.
Our efforts to increase the concentration of the nano C 60 in cell culture studies is limited by the maximum concentration of C 60 achievable in the water soluble preparation derived from toluene. Cell culture and proliferation in presence of other carbon nanomaterials, such as nanotubes, has also been successfully reported [42,43] and such findings are consistent with our data that show cell growth in presence of pristine C 60 is feasible. While several researchers (for example, see [44,45]) report that nanotubes indeed are cytotoxic, a recent publication [46] attributes such toxicity to, at least, in part to technical issues. This is analogous to our hypothesis that methods of preparation of C 60 accounts for the observed divergent cytotoxic effects of C 60 . Taken together, our data suggest that C 60 particles can be utilized for the design and development of multi-functional nanoparticles and the core nanoparticle is unlikely to adversely affect cell physiology.
An important finding of this study is that C 60 , when applied as methanol suspension, is non-toxic to a variety MALDI-TOF spectral analysis of C 60 preparations of cell types and does not interfere with cell proliferation. This finding is supported by cell proliferation assays, cell cycle analyses and vital stains. Further, cells continuously cultured with C 60 showed no defects in cell spreading and cytoskeletal organization, indicating the underlying cellmatrix interactions and signaling pathways are not adversely affected by C 60 . Our results are supported by other studies which show that C 60 , consistent with its well established electron acceptor properties, is a potent antioxidant [20,47]. This key finding differs from several published reports [16,18,19,41] which suggested that pristine nano C 60 is toxic. To reconcile with the cell type differences, we have employed several normal and malignant epithelial cells and tested their proliferation in presence of toluene-derived water soluble nano C 60 . Some investigators have reported weak toxicity of a preparation of polyvinyl pyrrolidine (PVP) and C 60 in cell culture and animal models compared to PVP alone [48,49]. However, it should be noted that the amount of C 60 used in those studies significantly exceeded that used in the present work and the method of preparation of C 60 is different.
Whereas several studies have examined the effects of C 60 on a variety of cells, few studies have examined whether fullerene crystals are taken up by the cells. Confocal microscopy of methanol C 60 -treated cells onto collagen matrices reveals intracellular C 60 nanocrystals of varying sizes in normal and malignant breast cancer cells ( Figure  3B). We believe that this is a first demonstration of intracellular pristine C 60 crystals using the PL signature as the reporter. The data shown in Figure 3B suggests that internalized C 60 retains its crystal structure as evident from its bright reddish orange PL. While we demonstrate of larger C 60 crystals in cells by confocal microscopy, smaller crystals (≤ 200 nm) may not be detectable by this technique.
Recent reports indicate the ability to detect fluorescence of Cellular uptake of methanol C 60 Figure 3 Cellular uptake of methanol C 60 . (A). Phase contrast image of a MDA MB231 cell which has internalized a C 60 cluster. Intracellular C 60 retains its PL signature. Scale bar is 20 μm. (B). Confocal microscopy of internalized C 60 aggregates (red) identified with arrows. Methanol C 60 -treated MCF10A cells were plated on collagen coated chamber slides, fixed, counterstained with FITC-phalloidin. A compiled 3-dimensional projection of optically sectioned z-stack is shown. Scale bar is 5 μm.
carbon nanotubes in cellular systems [50][51][52][53][54]. These findings suggest the possibility of detecting intracellular C 60 fluorescence, although the signal is generally weaker than the infrared signal of nanotubes. While other nanoparticles such as functionalized nanotubes [55,56] and gold nanoparticles [57] are reported to be internalized through endosomal pathways, the route of internalization of pristine C 60 is not known. Our data also suggest that the PL may be used as a reporting tool to estimate intracellular C 60 levels, provided the yield from the smaller crystals can be quantitatively measured.
In summary, our work describes a simple and rapid method for application of C 60 to cultured cells and to investigate the interactions of C 60 with cells. We provide evidence that pristine C 60 is taken up by normal and malignant cells and the intracellular C 60 retains its PL signature. Finally, we demonstrate that continuous culture of cells with C 60 is non-toxic and that cell adhesion, cytoskeletal reorganization following integrin activation and cell proliferation following treatment with C 60 remain unaffected. The reported toxicity of pristine C 60 is most likely due to incompletely understood solvent effects or to chemical modifications of the C 60 that may occur during preparation. A key implication of our research is that C 60 does not inhibit cell proliferation fullerene-based nanoparticles could possibly be utilized for biomedical applications without negative consequences from the fullerenes themselves.

Conclusion
C 60 fullerenes are useful for several biological applications. Here we described a new and simple method of applying these materials to cells and shown that they are taken up by cells. Significantly, we demonstrate that unmodified C 60 fullerenes are not toxic to cells. This finding should clarify the issue of perceived toxic effects of fullerenes and enhance developing novel biomedical applications using these nanoparticles.

Materials and methods
Fullerene suspensions C 60 fullerenes (Sigma Chemical Co) were sonicated in methanol at 0.2 mg/ml using a water bath sonicator (Branson) for 30 minutes to create a suspended fullerene solution which is referred to as methanol C 60 . 'Water-soluble' nano C 60 suspensions were prepared from toluene using published procedures [12,13]. To prepare a 'nano-C 60 ' suspension from toluene 0.5 mg of C 60 was added per ml of toluene. The suspension was sonicated for 10 minutes in a water bath (Branson) until a uniform purple solution was obtained and all C 60 had been dissolved as determined by observation. Following sonication in tolu-Water soluble toluene nano C 60 also does not block cell proliferation Figure 5 Water soluble toluene nano C 60 also does not block cell proliferation. Absorption spectra (A) and particle sizes (B) of water soluble nano C 60 from toluene are consistent with those reported in literature. The peak absorption wavelengths are indicated by arrows in A and the average particle size of the water soluble C 60 is 122 nm. MDA MB 231 (C) and HepG2 (D) cells were cultured with 2.7 μg (dotted line) or 27.4 μg (dashed line) of water soluble toluene nano C 60 or were untreated (solid line) and cell proliferation was assayed by crystal violet staining method.
ene, an equal volume of deionized water was added to the toluene/C 60 suspension and an organic/water phase separation was observed. This solution was sonicated in a water bath until all the toluene had evaporated (no more purple solution left), typically requiring about 2-6 hours depending on batch quantity.

Light spectroscopy
Fullerene suspensions were characterized by UV/Vis absorption (Beckman DU7500 spectrometer) and fluorescence spectroscopy. Photoluminescence (PL) measurements were made using a Safire 2 multifunctional monochromator based microplate reader (Tecan Instruments). Because methanol C 60 suspensions settle rapidly, spectra were recorded within 10 minutes of sonication.

Particle sizing
Size measurements of the colloidal fullerene suspensions prepared from methanol and toluene were carried out using a light scattering Zetasizer Nano-S light scattering instrument (Malvern Instruments, Southboro, MA). Sonicated methanol C 60 suspensions were immediately measured to prevent settling of the particles. Recording of the spectra was routinely completed within 10 minutes of sample sonication.

Transmission electron microscopy
Transmission electron microscopy was done on fullerene clusters dried from methanol onto formvar grids. A Phillips TEM Transmission Electron Microscope (model 400, 120 keV) was used and a sample of C 60 in methanol was dried onto a formvar grid for observation of the clusters.

MALDI-TOF
An Esquire MALDI-TOF mass spectrometer (Bruker Daltonics Instruments, Billerica, MA) was used to measure the masses of molecular species present in the various C 60 preparations. Solutions containing C 60 were mixed with equal volumes of saturated matrix solution (10 mg αcyano-4-hydroxycinnamic acid per mL of 0.05% trifluoroacetic acid and 25% CH 3 CN). Mass spectra were recorded in positive and negative ionization modes using the reflectron mode and calibrations were performed using a peptide mass calibration kit supplied by Bruker Daltonics.

Cell culture
Methanol C 60 suspensions were prepared and immediately applied to 12-well tissue culture dishes based on a protocol used for anoikis assays [29,32]. Following application of the suspensions, methanol was allowed to evaporate from the culture dishes while standing open in a sterile hood. Cells were plated onto the coated dishes and cultured in regular growth media in a tissue culture incubator. Cell proliferation was measured using crystal violet assays [58]. Culture dishes were rinsed with phosphate buffered saline (PBS) and stained in crystal violet stain (0.25% w/v in 50% methanol) for 10 minutes. Following rinsing of the dishes to remove excess stain, the dishes were air-dried, the protein-bound dye was solubilized in 50% methanol and the absorbance was recorded at 540 nm [59]. Each sample was measured in triplicate and the experiments were repeated at least twice. For some experiments, cell proliferation was assessed with a live-dead cell assay kit (Molecular Probes) containing calcein AM and ethidium dyes. Fluorescence microscopy was used to determine cell viablity by examining ratios of green (viable) to red (dead) cells.

Flow cytometry
Cell cycle profiles were determined by flow cytometry using established protocols [29,60]. Cells were trypsinized and fixed in 70% ethanol for at least 24 h at 4°C, stained with propidium iodide and subjected to flow cytometric analysis on a BD FACStar instrument. The DNA content of cells in various phases of cell cycle was determined by Modfit program.

Light and confocal microscopy
All cells lines were incubated with 200 μg of C 60 from the methanol preparation for 24 hours at 37 0 C. Following incubation, cells were extensively washed with PBS to remove adherent extracellular fullerene clusters, trypsinized, and replated on collagen I coated (5 μg/cm 2 ) chamber slides [32]. Samples were either directly viewed by phase contrast microscopy using an Olympus microscope or processed for confocal microscopy. Light microscopy images were recorded with a standard white light source without a UV filter. For confocal microscopy preparation, samples were fixed in 4% paraformaldehyde, extracted with 0.5% Triton X-100, incubated with FITClabeled phalloidin (Molecular Probes) to visualize actin cytoskeletal filaments, and mounted with the anti-fade kit (Molecular Probes) [32,60]. Samples were viewed on a Zeiss LSM 510 confocal microscope. Detection of C 60 was accomplished by excitation at 458 nm and the use of a long pass filter for λ > 650 nm. Images were optically sectioned and the projections of the compiled z-stack were imported into Adobe Photoshop (version CS2).