Cellular Uptake of Mildly Oxidized Nano-Graphene for Drug-Delivery Applications

: Graphene family materials (GFM) have large perspectives for drug delivery applications but their internalization in live cells is under investigation in a wide variety of studies in order to assess the best conditions for efficient cellular uptake. Here we show that mild oxidation of graphene nanoplatelets produces nano-graphene oxide (nGO) particles which are massively internalized into the cell cytoplasm. This remarkable uptake of nGO in NIH-3T3 cells has never been observed before. We performed vitality tests for demonstrating the biocompatibility of the material and analyzed the internalization mechanism under different oxidation degrees and concentrations. Moreover, we evaluated quantitatively, for the first time, the cell volume variation after nGO internalization in live cells through a label-free digital holographic imaging technique and in quasi real-time modality, thus avoiding the time consuming and detrimental procedures usually employed by electron-based microscopy. The results demonstrate that nGO formulations with a tailored balance between exposed surface and content of functional groups are very promising in drug delivery applications.


Introduction
Graphene is a two-dimensional (2D) material made of carbon atoms bonded together in a repeating pattern of hexagons. Its outstanding physical properties were reported for the first time in 2004 [1] and, since then, it attracts lots of interest for its unique properties such as high mechanical strength, specific surface area and high thermal and electrical conductivity [

2] [3] [4] [5] [6].
Different techniques are continuously developed and refined in order to obtain graphene family materials (GFM) with tailored properties. For instance, strong oxidizing agents can be used to produce an oxidized form of graphene, known as graphene oxide (GO) [7]. In this case, the oxygen-based functional groups reduce the thermal stability of graphene but are of crucial importance for promoting the compatibility with polar solvents [8]. Nevertheless, interactions between GFMs and proteins are better elucidated [43] [44]. Among them, strong π-π stacking non-covalent interactions, involving the π electrons of GFM and the aromatic residues of proteins in cell culture media, lead to the formation of protein corona complexes [44] [45].
Herein we show that mild oxidation of graphene nanoplatelets (GNP) results in high surface area nGO particles efficiently uptaken by cell. Compared to the above-mentioned works, we use here a simple oxidation technique that makes GNP biocompatible and bio-absorbable, thus avoiding iterative and harsh oxidation or conjugation procedures. We used bright and fluorescence field observation for qualitative characterization of the cellular uptake at different exposure times, and nGO concentration and oxidation degrees. Moreover, we developed a quantitative phase-contrast

Characterization of nGO
We studied the cellular uptake of mildly oxidized nGO under three different oxidation degrees, well below those usually reported in literature. The three different nGO materials are coded here as nGO1, nGO2, and nGO3 where the increasing number refers to a slight growth in oxidation extent. A commercial GO (GO-c) was used as reference material. For all samples three different concentrations in water were used ( Table S1). The nGOx samples were prepared starting from a commercial GNP using a modified Hummers' procedure [52], as detailed in Table S1. The nGOx samples were characterized by a multi-technique approach and Table 1 summarizes the obtained results. The energy dispersive X-ray spectroscopy (EDX) analysis indicated a progressive decrease of nGOx C/O atomic ratio with increasing amounts of permanganate oxidizing agent. The nature and the content of the oxygen-containing groups were assessed by FTIR spectroscopy (Figure 1a). FTIR analysis showed that the relative amount of hydroxyl, epoxide and carboxyl groups changed with the final oxidation degree of nGO. For example, the amount of hydroxyl groups was higher at lower oxidation degree, while epoxy groups prevailed at higher oxidation degrees. The presence of the peaks characteristic for hydroxyl, epoxy and carboxyl groups denotes that oxidation occurred on both edges and basal planes of the platelets [56]. We evaluated the intensities and the areas of defect (D) and graphene-like (G) bands by Raman analysis (Figure 1b). Upon oxidation, a progressive increase of the intensity ratio between the D-band and the G-band (ID/IG) was found going from GNP to nGO2 (Table 1), demonstrating a progressive breaking of the symmetry of the graphene lattice, due to edge effects, sp 3 -defects, vacancy sites or grain boundaries. A slight reduction was observed for nGO3, as at higher oxidation degrees the graphene sheets start to be dominated by the structurally disordered areas [57]. The average aromatic cluster size (L), related to the size of extended aromatic domains left undamaged upon oxidation, was evaluated by the empirical equation: where AG and AD are the areas of the G and D bands, respectively [52]. As shown in Table 1 Table 1). The analyzed dispersions were stable, and no precipitate was observed,  A further indication on the interaction of the organic components of complete DMEM with nGOx and GO-c was provided by ζ-potential analysis with laser doppler micro-electrophoresis (LDME). As reported in Table 1, ζ-potential of complete DMEM resulted to be -5. as reported in the next section.

Effects of nGO internalization on cell viability
To the best of our knowledge, the oxidized form of this commercial GNP has never been studied for cellular uptake applications. To investigate the effects on cell viability of nGOx with different oxidation degrees (nGO1, nGO2, nGO3) and at different concentration (20,50, and 100 µg/ml), we chose murine embryonic fibroblasts NIH-3T3 as a cell model to evaluate the cell response by MTT assay.
As shown in Figure 3, nGO1 and nGO2 at 20 μg/mL did not exhibit cytotoxicity, and the cell viability rate was more than 70%. Above 50 μg/mL, these materials exhibited obvious cytotoxicity effects, with a consequent decrease of metabolic activity. As the nGO1 and nGO2 concentration increased, the survival rate of cells decreased correspondingly. MTT results provided evidence that nGO3 was toxic to NIH-3T3 cells, independently of concentration.
However, as can be seen in the Figure 3, the viability of cells exposed to 100 μg/mL of nGO3 was higher than that relative to nGO1 and nGO2 at the same concentration. This outcome is in contrast with optical microscope observations (see Figure 6), which showed a round shape, typical of a dead  In case of 20 µg/mL (Figure 4) no significant effect after 3h incubation was observed. After 24 h, the cells with nGO1 and nGO2 presented a distinctive decoration of nGO clusters all around the nuclei (perinuclear region), as indicated by the red arrows. Correspondingly, the cells exposed to Nonetheless, the nGO uptake was not limited by the presence of the large nGO clusters outside cell membrane. Cells exposed to nGO3 showed a morphology change from elongated to spherical, typical of cell death, with evident features of low vitality, as demonstrated in Figure 3. The morphology of cells exposed to GO-c resulted to be almost unaltered. Similar results were obtained at 50 µg/mL concentration ( Figure S1). When exposed to nGO1 and nGO2, the cells were able to proliferate until forming a biofilm covering the well growth area, in spite of a slight decrease of their metabolic activity, as reported in Figure 3. nGO3 resulted to be cytotoxic also at 50 µg/ml, and only a slight nuclear decoration can be appreciated after 24h (see Figure S1b3). In case of GO-c no evidence of cell internalization was observed, since large scrolled-shaped sheets formed.
A conventional live/dead viability/cytotoxicity assay kit, reported in Supporting Information (Figures S3-S5), also tested the biocompatibility of nano-graphene at different oxidation degrees and different concentrations.  Bright field TEM observations of ultrathin NIH-3T3 cell sections corroborated the results obtained by optical microscopy. Figure 5 shows TEM images of the ultrastructural features of the NIH fibroblasts exposed to nGO2 at a concentration of 50 µg/ml for 48h. nGO was indeed internalized by cells and mainly located inside cytoplasm, around cell nucleus. In Figure 5c is possible to appreciate the nGO localization within the cytoplasm, as well as the dimension of the nanoparticle (300 nm) and the relative distance from the cell nucleus (150 nm). We also observed that, as the culture time increased, the amount of nGO2 inside NIH cells increased accordingly, and lots of nGO black particles appeared to stand in the cell cytoplasm around cell nucleus. Indeed, in order to evaluate cell internalization at longer exposure times we incubated the cells with nGO2 at a concentration of 50 µg/mL up to 96h. More massive cellular uptake was noticed, while cell adhesion and shape were preserved due to the nGO coronation process, which reduced the NIH-3T3 cellular morphological damage (Figures 6a-c).  As a further confirmation of the massive cellular uptake of nGO, leading to a peculiar accumulation of the particles into the cytoplasm volume, we evaluated also the cell morphology and the cytoskeleton assembly by staining the nuclei and the actin filament (see details in the Experimental Section). Figure 6 shows the typical bright field and fluorescence images for a control sample without nGO (Figure 6d, f), and for cells incubated with nGO2 at 20 µg/mL (Figure 6e, g), all fixed and labeled 72 h after seeding. The morphology of cells exposed to nGO2, appeared clearly perturbed compared to the control. The nucleus (labeled blue) was smaller and the actin filaments (labeled in green) were unstructured and poorly polymerized, probably due to the mechanical stress induced by the nGO internalized by the cells. The above reported outcomes suggest that nGO1 and nGO2 DMEM dispersions can be suitably used as nano-vectors for drug delivery applications, since exhibit low cytotoxicity and a significant internalization rate in cell cytoplasm [72].

Quantitative evaluation of nGO uptake in live cells
A preliminary assessment of the cell thickness variation before and after nGO exposure was performed by a standard profilometer (see Supporting Information, Figure S6). For a quantitative evaluation of volume changes during nGO uptake in live cells, we used a DH system [73]. The technique allowed us to avoid the time-consuming and detrimental procedures usually employed for evaluating the GFM internalization, such as transmission electron microscopy, which requires cell fixing and laborious sample preparation. Moreover, DH is a label-free imaging technique based on a microscopy tool able to provide quantitative phase contrast images of the live cells, thus avoiding, also in this case, time-consuming sample preparations. Herein, we focused our attention on the cell uptake mechanism in case of nGO2 at 50 μg/ml concentration, since the latter appeared to be the best candidate for highly efficient drug delivery. Indeed, it possess the best compromise between exposed  The cells analysed in the DH system were cultured in a glass WillCo-dish, and a conventional CCD camera was used to record the intensity pattern of the hologram. The latter, after numerical processing, yielded access to the phase shift information Δφ, arising from the difference in refractive index between the specimen and the surrounding medium [74]: (2) where λ is the laser wavelength, nc(x,y) is the spatial refractive index of the cell, nm is the refractive index of the surrounding solution, and h is the cell height at position (x,y) in the field of view. It is well known that it is not possible evaluate the cells height without decoupling the contribution of the refractive index of the involved materials, thus we can measure phase volume changes only.
The cells were seeded in a Petri dish, and after 24 h incubation, a complete DMEM suspension of nGO2 (ultrasonicated, 50 µg/mL) was added and incubated for another cycle of 24 h. Afterwards, the cells were detached by incubation trypsin -EDTA and seeded in the WillCo-dish to be mounted into the DH system. We analysed the cells under both suspended and adherent conditions by recording the holograms just after seeding the cells into the WillCo-dish. Once DH analysis was completed, the cells were let to adhere on the bottom surface of the Petri dish in the successive three hours. By conventional optical microscopy, we observed that the nanoparticles clearly filled the cytoplasm and maintained the perinuclear localization (Figure 7d-e), analogously to the experiments shown above (Figure 4), where the cells were not detached after nGO uptake. This demonstrates that the cells, even after trypsinization, preserve the internalized material together with the adhesion functionalities. The image shows also how the DH technique can show clearly the graininess of the nGO2 internalized by the cytoplasm (see the red arrow). The phase map images show that the nanoparticles tend to fill the cytoplasm without penetrating the nucleus, in perfect agreement with the bright field images shown in Figure S2b4. However, the continuous overlap of adjacent cells in the field of view prevents an accurate image segmentation to select a single cell and evaluate the volume variation.
where S is the segmented region within each cell and Dj is the phase map reconstructed according to (2). Figure 8c reports the Gaussian distributions of the phase volumes of the cells before (red line) and after (blue line) the nGO2 internalization, confirming the significant volume increase. We measured an average increase of 7% of phase volume per cell (with 0.14% as standard error estimation), corresponding to a total internalization of nGO2 of about 5 µg/mL. We estimated this latter value by calculating an equivalent amount of internalized nGO if its nominal refractive index at our laser wavelength, nnGO = 2.4-1.0i [76], is used in equation (2), to calculate the nGO heights per each position (x,y) within the cell.
It is important to underline that the phase images in Figure 8a

Conclusion
We demonstrated how the mild oxidation of GNP followed by ultrasonication produces a material that promotes a highly efficient cellular uptake with massive internalization into the cytoplasm region.
We focused our attention on nano-graphene oxide at intermediate oxidation degree (nGO2) that provides the best results in terms of biocompatibility and cellular uptake thanks to the best

Experimental Section
Nanoparticles: Graphene Nanoplatelets (GNP, grade C, average lateral dimensions 2 µm, XG Science -Lansing, MI, USA). Commercial water dispersed GO (GO-c) and all reagents and solvents were purchased by Sigma Aldrich (Milan, Italy) and used without further purification.
Notes The authors declare no competing financial interest