Degradation of Single-Layer and Few-Layer Graphene by Neutrophil Myeloperoxidase.

Biodegradability of graphene is one of the fundamental parameters determining the fate of this material in vivo. Two types of aqueous dispersible graphene, corresponding to single-layer (SLG) and few-layer graphene (FLG), devoid of either chemical functionalization or stabilizing surfactants, were subjected to biodegradation by human myeloperoxidase (hMPO) mediated catalysis. Graphene biodegradation was also studied in the presence of activated, degranulating human neutrophils. The degradation of both FLG and SLG sheets was confirmed by Raman spectroscopy and electron microscopy analyses, leading to the conclusion that highly dispersed pristine graphene is not biopersistent.

be interrogated. Along this line, biodegradation of graphene with oxidative enzymes secreted by cells like neutrophils are very relevant, since this type of cells is the first immune barrier intervening in the case of inflammation. Herein, we study the biodegradation of two types of water dispersible graphene corresponding to FLG and SLG. The biodegradability of these materials was assessed in vitro using recombinant human myeloperoxidase (hMPO) and ex vivo using freshly isolated neutrophils releasing hMPO extracellularly.
The intrinsic hydrophobicity of the graphene layers prohibits the production of stable graphene aqueous suspension. In the current study, we employed water dispersible FLG and SLG devoid of chemical functionalisation or surfactants to stabilise them in the aqueous media. FLG was produced by mechanochemical exfoliation of graphite thorough interaction with melamine by ball milling treatment in solvent free conditions. [12] After the treatment, water was added to the solid and melamine was removed using dialysis. FLG was stabilised in water or phosphate buffer at 0.1 mg/ml with a final concentration of melamine below 1 ppm ( Figure S1A). [12] The obtained FLG powder was characterised using TEM, Raman spectroscopy, thermogravimetric and elemental analyses ( Figure S2). FLG contains around 3.7% of oxygen, corresponding to the presence of few oxygenated groups as also observed by XPS. [12] Raman analysis revealed that the defects are mostly localised at the edges of FLG sheets. [12] Homogeneously dispersed SLG was instead obtained by oxidising (i.e. electron removal by air exposure) a graphenide (negatively charged graphene, KC 8 ) solution in tetrahydrofuran, mixing it with degassed water and evaporating the organic solvent to get water dispersed SLG ( Figure   S1B). [13] SLG was characterised using TEM, Raman spectroscopy, XPS and elemental analysis ( Figure S3). SLG is stable in water by a balance of weak inter-graphene sheet attractive forces and electrostatic repulsion due to spontaneous adsorption of OHions (zeta potential = -45 mV at neutral pH) on the graphene surface obtained from graphenide oxidation. [13] Raman analysis confirmed that SLG is a low-defect graphene due to presence of narrow D and G band linewidths and co-existence of few edge and sp 3 defects due to some functionalisation of the flakes with -H and -OH. [13] The elemental analysis also confirmed that SLG has about 3% of oxygen ( Figure   S3D).
As both FLG (two to four layers) and SLG (monolayer) were obtained using completely different methods, their physicochemical properties are different, likely affecting their behaviour during the degradation process.  (Figure 1). After 25 h treatment, flat 2D nanosheets started to become porous, and numerous tiny nanopores were observed on the surface of FLG and their edges were highly crumpled ( Figure 1B). In addition, some sheets lost their 2D shape ( Figure 1C). Further significant changes were observed in the morphology of FLG sheets after 40 h, where highly porous residues were formed ( Figure 1D) along with completely degraded material ( Figure 1E). A few aggregated sheets were also found confirming the non-uniform degradation due to low-dispersibility of FLG in buffer ( Figure S4A). We noted that addition of H 2 O 2 alone did not affect the morphology of FLG even after 40 h (Figure 1F   Similar to the behaviour of FLG, significant differences in morphology were also evidenced for SLG after hMPO treatment (Figure 2). Treatment with the enzyme for 25 h drastically affected the structure of SLG (Figure 2B-C). Most of the sheets lost their flat shape leading to highly broken and porous sheets ( Figure 2C). Prolonging the treatment to 40 h increased further the damages ( Figure 2D-E), and most of the sheets were broken down to nanoscale fragment. Only a few partially degraded sheets were observed ( Figure S4B). Importantly, the changes in the morphology of SLG sheets were more drastic compared to FLG after hMPO treatment, which could be explained with a better aqueous dispersibility and the presence of only single-layer graphene sheets in the SLG sample. Control treatment with H 2 O 2 alone did not affect the structure of SLG, confirming that the degradation is due to the enzymatic action ( Figure 2F). Next, to complement the electron microscopy results, we employed Raman spectroscopy, which is a powerful tool to understand the changes in the structure, quantify the amount of defects and the oxidation level of graphene sheets. [14] We calculated the intensity ratio of D and G bands that indicates the number of defects on the surface of graphene. Initially, the D/G intensity ratio of  Figure 3A). [12] After incubation with hMPO for 25 h, the variation of this ratio was very small (D/G = 0.46). After 40 h, FLG sheets displayed two types of spectra. A series of averaged spectra presented well defined D and G bands (D/G = 0.41), while a second set of spectra were completely devoid of D bands and nearly missing G bands ( Figure 3A). The drastic decrease of these two bands is due to significant oxidation/degradation of FLG sheets, as observed by TEM.
Raman microspectrometric 2D mapping of FLG sample further confirmed the degradation ( Figure S5). Raman analyses of SLG sheets also showed a similar trend seen for FLG ( Figure 3B). SLG sheets dispersed in PBS had a D/G ratio of 1.5. [13,15] The hMPO treatment for 25 h resulted in an increase of D/G ratio to 1.8, confirming the increase of defects or oxidation of graphene sheets. [16] In addition, the intensities of D and G bands were reduced compared to control samples in PBS. Further, the D/G ratio slightly decreased to 1.42 after 40 h, and the intensities of D and G bands were significantly reduced compared early time points. It should be noted that the sample treated for 40 h showed two types of Raman signatures like FLG ( Figure 3B). One set of spectra present still defined D and G bands while another series of spectra evidenced the disappearance of D and G bands, explaining the presence of completely degraded graphene sheets as shown by TEM analysis (Figure 2E). 2D mapping analysis again supported the degradation of SLG ( Figure S6). This result also confirms that the degradation of SLG sheets by

Raman shift (cm -1 )
PBS hMPO is not completed even after 40 h unlike GO, that is entirely degraded within 24 h. [6,7] This kind of spectral changes was absent in control samples where FLG and SLG were treated only with H 2 O 2 ( Figure S7).
It is not surprising that graphene is more resistant to enzymatic degradation than GO, as the number of defects and functional groups that are determinant factors for this process are negligible in comparison to those in GO. Overall, the Raman analyses did not support a complete degradation/oxidation of graphene sheets in contrast to the TEM analyses. This could be due to the aggregation of graphene sheets in PBS, which resulted in displaying two kinds of Raman signatures even after 40 h treatment, and the technical limitations of Raman, which does not allow to evidence the presence of the nanoscale and porous fragments. Overall, both Raman and TEM analyses confirmed that some of FLG and SLG sheets were only partially degraded, likely due to low dispersibility or aggregation of graphene sheets into thicker sheets as shown in Figure S4. We reported earlier that also aggregated GO sheets could not be degraded by hMPO. [6] To gain more insights about the interaction between hMPO and graphene samples, we decided to conduct an electrophoresis analysis. As shown in Figure S8, SLG displays better interaction with cationic hMPO (arginine rich protein) over FLG because of the high negative surface charge of SLG (zeta potential = -45 mV). [13,17] Thus, the enhanced interaction between SLG and hMPO could be the driving force for the higher degradability of SLG over FLG sheets ( Figure   2). We previously noted a similar trend in the hMPO-driven biodegradation of GO samples characterised by different surface charge. [6] Then, we wanted to assess if hMPO-rich human neutrophils were able to biodegrade FLG and SLG. Our previous works have demonstrated that neutrophils are capable to degrade GO and CNTs extracellularly upon degranulation of oxidative enzymes. [7,8,18] The N-formyl-methionylleucyl-phenylalanine (fMLP) tripeptide is a potent activator of neutrophils and cytochalasin B (Cyto-B) enhances several fMLP-stimulated neutrophil responses, including aggregation, superoxide production, and degranulation. [19] Freshly isolated primary human neutrophils were activated with fMLP and Cyto-B and incubated with FLG and SLG. Neutrophils activated in this manner were added every day up to five days ( Figure 4A). Exogenous H 2 O 2 was not required, as these neutrophils already contain the complete system required for degradation. [20] The biodegradation of the FLG and SLG was analysed by confocal Raman microspectrometric mapping as shown in Figure 4B and 4C, respectively. FLG was significantly degraded after 5 days, as indicated by the significant reduction of D and G band intensities ( Figure 4B). occurred. [6,7] In addition, the peak enlargement, [21] leading to the lack of separation between the D and G bands for SLG (Figure 4C), together with the disappearance of the characteristic 2D band, further support a reduction of the degree of graphitization and the creation of defects on graphene sheets. Taken together, these results showed that activated primary human neutrophils can digest both FLG and SLG. SLG sheets were degraded at lower speed compared to FLG sheets by the neutrophils, likely due to the presence of minor number of defects and lower content of oxygenated functions on the surface of SLG sheets compared to FLG sheets. Since FLG was obtained by mechanical ball-milling and subsequent dispersion, whereas SLG sheets were synthesised by direct exfoliation in solution, a high quality graphene sheets is likely present in SLG. [13] In comparison to isolated hMPO, the cellular system deployed here seems to be more efficient in degrading FLG than SLG.
Overall, the treatment of graphene sheets either in vitro in the presence of recombinant hMPO or using activated neutrophils, suggested that degradation of pristine graphene sheets is also possible, similarly to degradation of other carbon materials like oxidised CNTs or GO. [4,6,7,18] Though FLG and SLG have only a slightly different oxygen content (~3.7% and 3%, respectively), they have different types of defects and functional groups at their edges. While FLG has some oxygenated groups like epoxides, carbonyl and carboxylates, [12] SLG has defects derived from the modification of the edges by protons and hydroxyls. Overall, comparing the degradation of few-layer graphene to that of single-layer graphene based on their oxygen content remain difficult to rationalise. The degradation is triggered by the generation of reactive intermediates of hMPO in the presence of H 2 O 2 and NaCl. In particular, the highly oxidant reactive intermediate hypochlorite (NaOCl, E o = 1.56 V) is the main species able to oxidise and degrade the graphene sheets. In the case of GO, we found that the total degradation process encompasses 24 h. [6,7] Since graphene does not have oxygenated functional groups and defects like GO, we can expect that it would be more difficult to degrade pristine graphene sheets.
Earlier studies dealing with in vivo and in vitro macrophage degradation of carboxyl functionalised graphene (graphene treated with HNO 3 ) revealed that the degradation of this type of graphene is possible within cells, though the mechanism of degradation was not proven. [22] In addition, our current study further reinforces the idea that pristine graphene is biodegradable, and shows for the first time that neutrophils are capable of extracellular digestion of graphene.
Although the in vitro degradation and neutrophil degradation of FLG and SLG have shown some differences, the current results add another important element to the safety assessment of GFMs, with potential implications for their in vivo use. Previously, Schinwald et al. showed that pulmonary exposure to pristine graphene nanoplatelets (average lateral dimension: 5 μm, average thickness: ~10 nm) significantly increased the number of neutrophils and eosinophils, 24 h postexposure compared to the control mice. [23] The acute inflammation reaction in the lungs was reduced after 7 days presumably due to the degradation of graphene platelets into smaller fragments. Notably, the material was too large to be completely phagocytosed by macrophages and it stands to reason that extracellular digestion of these materials may come into play, along with other clearance mechanisms. Our results show that pristine graphene sheets, from single-to few-layers of a few hundred nanometers in lateral size, can be degraded by the action of activated neutrophils, albeit at a slower rate than GO, [7] further supporting the potential of water dispersible graphene for biomedical applications.
To summarise, we have demonstrated that graphene can be degraded either by recombinant hMPO or by hMPO secreted by activated neutrophils. Two different graphene samples, SLG and FLG, showed different degradability behaviour due to their different physicochemical properties, resulting from differences in the synthesis methods. These results clearly demonstrate that pristine graphene with minimum defects can be degraded by hMPO-mediated oxidation, indicating that our immune system has strategies to degrade graphene materials.

Degradation of graphene materials by hMPO
The stock solution of exfoliated graphene is a surfactant-free single-layer graphene (SLG) dispersed in water (0.1 mg/mL sheets were assumed to be more resistant to oxidation compared to GO [1] and carboxylated CNTs, [2] their degradation was carried out up to 40 h in comparison to 24 h previously used for GO. [1] Similar to the above experiment, 12.5 µg of few-layer graphene (FLG) [3]

Raman analyses
Raman analysis of in vitro degraded samples was performed using Raman spectra Renishaw inVia microRaman equipped with 532 nm laser and a Leica microscope. All samples were prepared by drop-casting 10 µL of respective samples on silicon wafer coated with SiO 2 (20 nm, TED Pella) and dried for 24 h at room temperature. All the spectra were recorded with 1% laser power (0.6 mW) using 100× objective lens. The results reported represent the averages of at least five individual spectra.
Raman microspectrometric 2D mapping of in vitro degraded samples was also performed using Raman spectra Renishaw inVia microRaman equipped with 532 nm laser and a Leica microscope. All samples were prepared by drop-casting 10 µL of respective samples on silicon wafer coated with SiO 2 (20 nm, TED Pella). The maps were recorded with 1% laser power (0.6 mW) using 100× objective lens. The scan areas were selected according to the different samples in order to avoid, as far as possible, the zones without nanomaterial. One µm was the measure of the distance between the spectra.
Raman analysis of ex vivo degraded samples was performed as previously described [4,5] using a confocal Raman microspectroscopy (WITec alpha300 system, Germany) with a laser of 532 nm wavelength set at an integration time of 0.5 s and 600× magnification. The scan area for each sample was adjusted to 100×100 µm. For determination of the intensities of the D-band (second order double resonant mode activated by defects, ∼1354 cm -1 ) and G-band (tangential C-C stretching modes, ∼1582 cm -1 ). The Raman spectrum of each sample is an average of 10000 spectra obtained from a scan size of 100 µm×100 µm.

TEM analysis
Six µL of each aliquot were deposited on carbon coated copper grid and dried under the lamp, and washed the grids in the Milli Q water for ~1 hour to remove the salts (from buffer). All the samples were analysed by a Hitachi H7500 microscope (Tokyo, Japan) with an accelerating voltage of 80 kV, equipped with a AMT Hamamatsu camera (Tokyo, Japan).

Gel electrophoresis
The scanning densitometry was performed with a GS-800™ alibrate Densitometer (BioRa ). The densitometry analysis was then performed using the ImageJ 1.48f software. Signal density was normalised to the hMPO density.

Degradation of graphene materials by neutrophils
Neutrophils were isolated from buffy coats of healthy human blood donors (Karolinska University Hospital, Stockholm, Sweden) as previously described. [6] The samples are completely anonymized and for this reason no specific approval is required. Briefly, neutrophils were isolated from healthy donors every day for 5 consecutive days by density gradient centrifugation using Lymphoprep (Axis Shield, Oslo, Norway) followed by gradient sedimentation in a 5% dextran solution and hypotonic lysis of residual erythrocytes. To study degradation, freshly isolated neutrophils (10 6 cells/mL) were incubated in phenol red-free RPMI-1640 culture medium (Sigma Aldrich) supplemented with 2 mM L-glutamine, 100 U/mL penicillin and 100 mg/mL streptomycin without serum and incubated in 5% CO 2 at 37°C with FLG or SLG at 20 µg/mL concentration in the presence of fMLP (10 nM) an cytochalasin B (5 μg/mL) (Sigma Aldrich) to trigger degranulation. [7] Freshly isolated, ex vivo activated neutrophils were added every day. Samples were collected at the indicated time-points and stored at -80 °C for further analysis.  confirming the few-layer nature of the graphene sample. [3] In addition, the thermal profile for cm -1 , ~1620 cm -1 , and ~2681 cm -1 , respectively. ID/IG = 1.50, and the full-width at half maximum (FWHM) of 2D band is 28 cm -1 , confirming the single-layer nature of the graphene sample. [8] (C) XPS profile of SLG as deposited film, without any annealing. The C1s X-ray photoelectron spectroscopy analysis of the SLG films shows minor widening on the high-energy side of the C1s peak confirming that there is only little functionalization of the graphene carbon sheets. Panel (D) displays the elemental analysis of SLG, calculated from the fitting of the XPS C1S peak, from which it can be concluded that the sample is mostly composed of carbon.  side) are shown. After treating with MPO for 40 h (bottom), we can observe a reduction in the G band intensity, which indicates degradation of the FLG sheets. D/G ratio is also increased compared to 0 h FLG sheets. Figure S6. Raman microspectrometric 2D mapping analysis of SLG samples at 0 h (top) and after treating with MPO for 40 h, where G band map images (left side) and D/G ratio map (right side) are shown. After treating with MPO for 40 h (bottom), we can observe a reduction in the G band intensity, which indicates degradation of the SLG sheets. D/G ratio is also increased compared to 0 h SLG sheets.