Single exposure to aerosolized graphene oxide and graphene
nanoplatelets did not initiate an acute biological response in a 3D
human lung model
Barbara Drasler
a, Melanie Kucki
b, Flavien Delhaes
a, Tina Buerki-Thurnherr
b,
Dimitri Vanhecke
a, Daria Korejwo
b, Savvina Chortarea
b, Hana Barosova
a,
Cordula Hirsch
b, Alke Petri-Fink
a,c, Barbara Rothen-Rutishauser
a, Peter Wick
b,* aBioNanomaterials Group, Adolphe Merkle Institute, University of Fribourg, Chemin des Verdiers 4, 1700 Fribourg, SwitzerlandbLaboratory for Particles-Biology Interactions, EMPA - Swiss Federal Laboratories for Materials Science and Technology, Lerchenfeldstrasse 5, 9014 St.
Gallen, Switzerland
cDepartment of Chemistry, University of Fribourg, Chemin du Musee 9, 1700 Fribourg, Switzerland
a r t i c l e i n f o
Article history:
Received 24 November 2017 Received in revised form 3 May 2018
Accepted 4 May 2018 Available online 11 May 2018
a b s t r a c t
The increased mass production of graphene related materials (GRM), intended for a broad spectrum of applications, demands a thorough assessment of their potential hazard to humans and the environment. Particularly, the paramount concern has been expressed in regard to their interaction with the respi-ratory system in occupational exposure settings. It has been shown that GRM are easily respirable and can interact with lung cells resulting in the induction of oxidative stress or pulmonary inflammation. However, a comprehensive assessment of potential biological effects induced by GRM is currently hardly feasible to accomplish due to the lack of well-defined GRM materials and realistic exposure data. Herein, a 3D human lung model was combined with a commercial aerosolization system to study potential side effects of GRM. Two representative types of GRM were aerosolized onto the lung epithelial tissue surface. After 24 h post exposure, selected biological endpoints were evaluated, such as cell viability, morphology, barrier integrity, induction of (pro-)inflammation and oxidative stress reactions and compared with the reference material carbon black. Single exposure to all tested GRM at the two different exposure con-centrations (~300 and 1000 ng/cm2) did not initiate an observable adverse effect to the 3D lung model
under acute exposure scenarios.
© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Graphene and related materials (GRM) are defined as two-dimensional (2D) carbon-based nanomaterials (NM) that include diverse sheet-like carbon forms with considerable variations in their lateral dimensions, the number of layers and respective thickness, as well as in their surface chemical modifications [1e3]. Their common features are e.g. high electrical conductivity, good thermal and chemical stability, mechanical strength, unique optical behaviours, and a large surface area. Collectively, this makes GRM attractive candidates for a wide range of high-technology applica-tions, e.g. as materials for energy storage, photonics, and composite
materials, as well as for biomedical applications, e.g. in regenerative medicine, tissue engineering, biosensors, and cell imaging, respectively [4,5]. The mass production of GRM has considerably increased in the past several years [6] which requires a thorough determination of their safety profile [7,8] for workplace and con-sumer safety as well as for public acceptance. GRM-containing products are arriving on the market and according to the IDTe-chEx latest report it has been predicted that the total graphene market will grow immensely by 2020 [3,9,10].
However, concerns have been raised that this unique combi-nation of material properties can provoke unexpected side effects. Therefore a careful and comprehensive assessment of different exposure routes and application scenarios is required to assess the hazard of the materials. The respiratory system has been accepted as the most likely route of human exposure to airborne materials [3], yet also dermal exposure and ingestion might occur.
* Corresponding author.
E-mail address:peter.wick@empa.ch(P. Wick).
Contents lists available atScienceDirect
Carbon
j o u rn a l h o m e p a g e :w w w . e ls e v i e r . c o m / l o c a t e / c a r b o n
https://doi.org/10.1016/j.carbon.2018.05.012
The current knowledge about GRM as potential hazard for the lungs is incomplete and often contradictory. The reasons for this are manifold. First of all, physico-chemical characteristics can differ considerably between the individual GRM types and the term graphene has repeatedly been generalized to all GRM types (not only to the graphene form) [11]. Consequently GRM require type specific toxicological profile assessment. Also, the existing studies often lack thorough GRM characterization [2,3,9] which makes it impossible to re-evaluate the outcome. Hence such comparisons of the toxicological profiles within the GRM family can be problematic or even inaccurate. Secondly, the exposure doses used in the studies often do not reflect realistic human exposure scenarios to GRM. On one hand, there is no empirical evidence available on the respirable fraction of aerosolized GRM, i.e. the portion or the amount of GRM which potentially entersthe human respiratory tract, and on the deposition of GRM in the lungs. Hence, the doses applied in in vitro and animal in vivo studies can only partially be justified. On the other hand, applied GRM doses often considerably exceed the highest possible GRM deposition in the human respi-ratory tract; in particular in in vitro studies. Whilst high doses are sometimes desirable to observe in toxicological studies, such out-comes do not represent real-life conditions at inhalation [12]. In addition, due to their physico-chemical properties, GRM dose estimation can be challenging and has frequently not been adequately addressed. Due to the differences in thickness and lateral dimensions, dose metrics expressed in GRM mass units alone may not be sufficient for a reliable GRM dosimetry and to draw comparison among the studies [2,3,9].
Thefirst outcomes of existing in vivo animal studies on GRM toxicological profiles have been proven to be rather controversial [9]. GRM types with apparent resemblance in their physico-chemical properties showed acute toxicological responses in some experimental settings but not in others. The inconsistent outcomes indicate that already slight changes in the GRM proper-ties induce a significant change in initial response. Nevertheless, some outcomes of in vivo animal studies do suggest that different types of GRM are easily respirable and can persist in the lungs upon inhalation. It has been shown that GRM can closely interact with lung cells resulting in the induction of oxidative stress and pul-monary inflammation and can disturb immunological- and physiological-homeostasis [11e16], reviewed in Refs. [9,17].
Similarly, inconsistent toxicological profile of GRM upon inha-lation has been observed in existing in vitro studies too [11,12,18e20], and reviewed in Refs. [3,7,17,21e24]. In a vast ma-jority of studies, simplified monoculture cellular systems under submerged exposure conditions have been employed. The main pitfalls of such approaches are the lack of simulating air-liquid interface (ALI) conditions which are important for lung cell cul-tures and the lack of cell-cell communication between various lung cell types, e.g. between immune cells and epithelial cells. Both can be achieved with advanced multicellular co-culture models which have been proven to more closely simulate the real situation in the lung [25,26].
Therefore the aim of the study was to acquire a systematic un-derstanding of the biological impact of GRM with distinct physico-chemical properties at realistic doses in vitro, using a commercially available aerosolization system (Vitrocell® Cloud) for aerosol exposure of three dimensional (3D) human epithelial alveolar tis-sue barrier model, cultivated at ALI. For this purpose we selected and characterized two GRM types, GO and GNP, differing in the type of manufacturing process, lateral dimension, number of layers and carbon to oxygen (C/O) ratio. The outcome was benchmarked with carbon black (CB) [12,13,27]. A combination of relevant biological endpoints was assessed. Namely, the potential of aerosolized GRM to (i) affect cell viability based on the evaluation of membrane
ruptures and morphological alterations of the co-culture model, (ii) affect (pro)-inflammatory signalling via their ability to cause increased production of the (pro)-inflammatory cytokine tumour necrosis factor-alpha (TNF-
a
), the interleukin 1-beta (IL-1b
) involved in the inflammasome formation, and the chemokine interleukin (IL-8) contributing to the attraction of other immune cells at the site of inflammation, and (iii) induce oxidative stress as indicated by intracellular glutathione (GSH) depletion.2. Materials and methods
2.1. Chemicals, reagents and laboratory conditions
All chemicals were purchased from Sigma-Aldrich GmbH (Buchs, Switzerland), unless otherwise stated. Water in all the ex-periments refers to ultrapure deionized water of 180
U
S cm1 (Millipore AG, Switzerland). All cell culture reagents were pur-chased from Thermo Fisher Scientific (Reinach, Switzerland), unless otherwise specified. At every step during the cell cultivation and exposure, cells were kept in an incubator under controlled labo-ratory conditions (humidified atmosphere, 37C, 5% CO2). Uponexposures, the collected cell culture medium wasfirst kept at 4C
and stored at80C or 4C until further analysis. Cell
concentra-tions were determined using the Trypan blue exclusion method (0.4% Trypan blue solution).
2.2. Nanomaterials: (GO), (GNP) and CB
Graphene oxide (single layer graphene oxide, GO) and graphene nanoplatelets (GNP; HDPlas™ GNPs Grade 4) were obtained from Cheap Tubes (Battleboro, 112 Mercury Drive, VT05301, USA;http:// www.cheaptubes.com). Carbon black (CB) powder (Printex®85, Evonik-Degussa, Germany) was included in the study as a reference material. Material physico-chemical properties are summarized in Table 1. GO and GNP characterization data was recently published in Kucki et al. [28] whereas CB's physico-chemical properties were described by Figarol et al. [29].
GO samples showed good dispersion behaviour in aqueous suspension due to oxygen-functionalized groups, hence powder GO was dispersed in ultra-pure water (1 mg/mL) and sonicated in a water bath (2 3 min, vortexed before and after sonication). GO dispersions were prepared freshly before each cell culture experi-ment, whereas stock suspensions of GNP and CB were prepared in advance. Both materials were dispersed in sterilefiltered (0.22
m
m pores) 160 ppm Pluronic F-127 in ultra-pure water to a concen-tration of 1 mg/mL of GNP or CB, bath sonicated (30%, 37 kHz, Elmasonic P, Elma Schmidbauer GmbH, Singen, Germany) for 5 min and stored at 4C under sterile conditions. The non-ionic surfactant Pluronic F-127 facilitates the dispersion of GNP and CB, and it has been shown that in cell culture medium, Pluronic will be rapidly replaced by serum components [30]. Prior to the experiments, GNP and CB dispersions were additionally sonicated in a water bath (3 min) and thoroughly vortexed.2.3. Endotoxin detection
Endotoxin detection was performed by two different methods, the Limulus amebocyte lysate (LAL) Gel Clot Assay (Pyrogent™ Plus, Ref. No. N294-03, Lonza, Walkersville, USA), and the Endosafe®PTS (Charles River Laboratories, Charleston, USA). All tests were per-formed in the presence of the test materials and not with super-natants only. The GO, GNP and CB dispersions were prepared under the same conditions as for the cell-exposure experimental setup. A detailed description of the assays is provided in the Supplementary material. See also for further technical details [31].
2.4. Triple cell Co-culture model of the alveolar epithelial tissue barrier
The 3D triple cell co-culture (TCCC) model of the human alveolar epithelial tissue barrier has been previously established and char-acterized [32,33]. The model includes three representative cell types of the lung region, i.e. alveolar epithelial cells (human lung carcinoma cell line A549, ATCC®CCL-185™), and two types of pri-mary human derived immune cells, i.e. monocyte-derived macro-phages (MDM) and dendritic cells (MDDC). Monocytes were isolated from buffy coats provided by the Transfusion Blood Bank (Blutspendedienst SRK Bern AG, Switzerland) as described by Ref. [34] with the adaptation of using CD14 magnetic beads (MicroBeads, Miltenyi Biotec, Germany) for monocyte isolation. The primary monocytes were differentiated into MDDC by stimulation with the granulocyte macrophage-colony stimulating factor (GM-CSF, Miltenyi Biotec, Germany) and interleukin (IL)-4 (Bio-Techne AG, R&D Systems, Switzerland) and into MDM with macrophage colony stimulating factor (M-CSF, Miltenyi Biotec, Germany), all at concentration of 10 ng/mL in complete Roswell Park Memorial Institute (cRPMI)-1640 cell culture medium supplemented with 10% FBS, 1%L-glutamine, and 1% penicillin/streptomycin (all v/v;
referred as complete cell culture medium, cRPMI) for 6 days. The TCCC was cultivated on the polyethylene terephthalate (PET) membrane inserts for 12-well plates (BD FalconTM Cell Culture Inserts, pores with 3
m
m diameter, 8 x 105pores/cm2, surface area of 0.9 cm2, BD Biosciences, USA). The TCCC was grown at the air-liquid interface (ALI): the epithelial cells and MDM on the apical sides and MDDC on the basal sides of the membranes as previously described [32,33]. A detailed description is provided in the Supplementary material.2.5. TCCC exposure to nebulized GO and GNP
The TCCC grown at ALI were exposed to nebulized GO or GNP using the Vitrocell® Cloud exposure system (Vitrocell systems GmbH, Germany) consisting of a nebulizer (a vibrating mesh with a span of 4.0e6.0
m
m volumetric mean diameter; Aeroneb®Pro, Aerogen, IL, USA), an aerosolization chamber, a base module for 12 -well sized inserts and connected to the quartz crystal microbal-ance system (QCM). The QCM was placed in the base module instead of the membrane insert, hence was exposed to the aerosol in the same way as the cell cultures, and the deposited mass of theaerosolized material per area unit was monitored online as a function of exposure time [35].
Nebulization procedure: The PET inserts with TCCC grown at ALI were placed in the pre-heated base module (37C; each well containing 3.5 mL of phosphate-buffered saline (PBS) and exposed to aerosolized GO and GNP generated by the nebulizer. For each aerosolization, 2
m
L of sodium chloride (NaCl) water solution (1.5 mM) was added to 0.2 mL of the tested suspensions (0.25 and 1 mg GO or GNP/mL). Negative control TCCC samples were indi-vidually exposed to aerosolized NaCl (1.5 mM water solution) and Pluronic F-127 (160 ppm in water). After aerosolization and cloud settling, the TCCC were kept at ALI conditions in fresh cell culture medium (0.6 mL lower chamber) for 24 h post-exposure.2.6. GRM and CB deposition and characterization
The mass of deposited GO and GNP after nebulization was monitored online by the Quartz Crystal microbalance. Afterwards, GRM deposition was visualized using transmission and scanning electron microscopy (TEM and SEM, respectively). Surface coverage (in %) was estimated from automatically and randomly captured TEM images (Fig. S1). Briefly, single slot copper TEM grids (SF162-6, Plano GmbH, Wetzlar, Germany) were placed in the aerosolization chamber next to the Transwell®membrane inserts. After drying, they were analyzed in a Tecnai spirit TEM (ThermoFischer/FEI, Hillsboro, Oregon, US) operating at 120 kV, and images were recorded using a 4096x4096 Eagle CCD camera (FEI/ThermoFischer, Hillsboro, Oregon, US). 25 micrographs per treatment (GO and GNP at the two input concentrations, and the respective negative con-trols, i.e. nebulized 1.5 mM NaCl solution and Pluronic F-127 (40 and 160 ppm in water) were recorded in an automated systematic uniform random sampling procedure (more details in Supple-mentary material). For additional visualization of aerosolized GO and GNP and of CB dispersions, the material was aerosolized (GO and GNP) or pipetted (CB) over or onto SEM holders, i.e. poly-carbonate discs, mounted to aluminium by carbon tabs (12 mm diameter; Agar Scientific Ltd., UK). These were then coated with 4 nm platinum (High Resolution Sputter Coater 208, Cressington Scientific Instruments, Watford, UK) and imaged at 1.8 kV by Mira3 LM (Tescan, Brno, Czech Republic) FE-SEM, using a secondary electron detector.
For the evaluation of the effect of the non-2D reference material CB as well as positive controls for membrane rupture and for
Table 1
Characteristics of GO, GNP and CB.
Material Graphene oxide (GO) Graphene nanoplatelets (GNP) Carbon black (CB), reference material Source Cheap tubes Inc. USA Cheap tubes Inc. USA Printex®85, Evonik-Degussa Preparation Modified Hummers method Commercial
microemulsion
NA
Starting material Natural graphite Natural graphite NA Size distribution/
Lateral dimension
1e10mm (SEM)a/
300e800 nmb Aggregate size 1e10mean ~5mm (SEM)a/mm,
1e2mm (AFM)b
Mean diameter: 42± 10 nm:
Specific surface area ND ND 184± 1 m2/gc
Number of layers/Thickness single to few layers/ 0.7e1.2 nmb
5-10 layers/ up to 5mma
NA
C/O ratio (XPS) 1.7± 0.1a 24.0± 2.5a NA
Raman ID/IG ratio 1.19± 0.08 (633 nm) 0.96± 0.02 (532 nm)a 0.37± 0.07 (633 nm) 0.17± 0.11 (532 nm)a NA Zeta potential in ultra-pure water (pH 6.4) 39.4 ± 1.3 mVa 62.6 ± 1.9 mVa 25 ± 9 mVc
Abbreviations: SEM, scanning electron microscopy; AFM, atomic force microscopy; XPS, X-ray photoelectron spectroscopy; ID/IG ratio, ratio of the intensity of D- Raman peak and G- Raman peak; NA, not applicable; ND, not determined; Data source:a[28];baccording to manufacturer;c[29].
oxidative stress, droplets of suspensions were added apically to the TCCC model cultivated at ALI. Aerosolization of CB was not possible due to the highly adhesive properties of the material in the nebu-lizer head and was avoided in order to prevent possible blockage of the aerosolization system. Therefore droplets of suspensions were added apically to the TCCC model cultivated at ALI (referred to as pseudo-ALI [36]) creating the closest possible conditions to the aerosolized material-exposed TCCC. Namely, 25
m
L of CB in Pluronic F-127 and 0.02% (v/v) Triton X-100 in PBS, for 24 h, with fresh cell culture medium (0.6 mL) in the lower chamber. A droplet (25m
L) of water and Pluronic F-127 (6 ppm in water, corresponding to that in the high CB samples) were added as negative controls.2.7. Membrane rupture/cytotoxicity
Membrane integrity of the TCCC was assessed 24 h post-exposure via the lactate dehydrogenase (LDH; cytosolic enzyme) release assay since the assay has been proven a reliable indicator of membrane rupture and cytotoxicity [37,38]. The cell culture me-dium was collected from the lower chambers of the exposed TCCC and analyzed in triplicates using the LDH cytotoxicity detection kit (Roche Applied Science, Mannheim, Germany) according to the manufacturer's manual. The absorbance of the colorimetric product was determined spectrophotometrically (Benchmark Microplate reader, BioRad, Cressier, Switzerland) at 490 nm with a reference wavelength of 630 nm. The absorbance values were expressed relatively to those of the respective negative controls (nebulized NaCl solution or a droplet of water).
Possible interference of the tested NM with the LDH assay was investigated at cell free conditions. GRM and CB cell free suspen-sions were diluted atfinal concentrations of 0e15
m
g/mL and LDH assay was performed according to the supplier's manual to analyze possible changes in optical density at 490 nm (OD490 nm), due to thepresence of the materials. The maximal possible deposition of material is 1
m
g/cm<sup>2</sup>, which equates to 1,5m
g/mL.This concentration range was chosen, to demonstrate effects up to ten times higher concentrations of material, than applied in this study. 2.8. Cell morphologyCellular morphology of the TCCC before and after 24 h exposure was observed by an inverted laser scanning confocal microscopy (LSM; Axio Observer. Z1, Zeiss, Germany).
First, each cell type was stained with different fluorophores (Vybrant®multicolor cell labeling kit, Invitrogen Molecular Probes) prior to the co-culture composition as previously described [39]. Briefly, MDDC and MDM were treated with vibrant dye DiI and DiD, respectively, (5
m
L for 1 mL of cell suspension) and incubated for 20 min in the incubator. Cells were centrifuged and washed with RPMI 1640 three times prior to seeding. In the meantime, the layer of A549 grown on the insert was treated with vibrant dye DiO (1.5m
L in 200m
L of RPMI 1640) and incubated for 20 min in the incubator. Cells were washed twice with RPMI 1640 for 10 min and finally the co-cultures were composed as previously [38] described. After material exposure and 24 h post-exposure the cell culture media was removed, TCCC were washed 3-times in PBS andfixed in 4% paraformaldehyde (PFA; in PBS) for 15 min at room temperature (RT), washed 3-times and stored at 4C. Then, cells were per-meabilized with 0.2% Triton X-100 (in PBS, v/v) for 15 min, washed (3-times PBS) and immersed in the blocking solution (0.3 M glycine and 1% bovine serum albumin in PBS), at RT for 20 min. Over a period of 1 h, the F-actin cytoskeleton was then stained with Rhodamine-phalloidin (Molecular Probes, Thermo Fisher scientific Inc.) at 0.132m
M and 40,6-diamidin-2-phenylindol (DAPI) at 1m
g/ mL in PBS for nuclear staining. After washing (3-times in PBS), thesamples were embedded in Glycergel (DAKO Schweiz AG, Baar, Switzerland) and visualized with the LSM. For each sample, representative images (z-stacks) were taken at three randomfields of view at both the apical and basal side of the membrane insert with TCCC. Image analysis and processing was performed using the software ImageJ [40].
2.9. (Pro)-inflammatory cytokine/chemokine quantification
The (pro-)inflammatory response of the TCCC 24 h post-exposure to GRM was assessed applying the enzyme-linked immunosorbent assay (ELISA) for the selected inflammatory markers: interleukin 8 (IL-8), interleukin 1-beta (IL-1
b
) and tumour necrosis factor alpha (TNF-a
); the combination of markers tested have been proven representative for evaluation of (pro-)in flam-matory response of the TCCC model to an external stimuli (including aerosolized NM) [38e40]. Positive control for the (pro-) inflammatory proteins was accomplished by treating TCCC from the lower compartment with lipopolysaccharide (LPS; from Escherichia coli; 1m
g/mL in cRPMI; Sigma Aldrich, Buchs, Switzerland), for 24 h. The amount of the proteins released in the cell culture medium (lower compartment) was quantified using the commercially available DuoSet ELISA Development Kit (R&D Sys-tems, Switzerland) according to the supplier's protocol.2.10. Oxidative stress
The intracellular reduced glutathione (GSH) content of the exposed TCCC was determined using the Glutathione Assay Kit (Cayman Chemical, MI, USA) according to the manufacturer's pro-tocol for the deprotonated cell lysates. Briefly, after 24 h post exposure, the TCCC cultivated on the membrane inserts were washed with PBS in order to eliminate cell culture medium con-taining extracellular GSH and proteins from the media. It was not possible to gain information pertaining to the oxidative GSH component (GSSG) as GSSG levels were below the limit of detection for the experimental set-up. Therefore, all the total reduced intra-cellular GSH values were presented relative to the protein expres-sion in the sample, quantified using the Pierce BCA Protein Assay kit (Thermo Fisher Scientific Inc., IL, USA). TCCC exposed apically to 100
m
L of L-Buthionine-sulfoximine (BSO; 200m
M; Sigma-AldrichGmbH, Buchs, Switzerland) for 24 h were used as a positive control. 2.11. Statistical analysis
Statistical analysis and graphical presentations were performed using the software Origin 2016 (OriginLab Corporation, North-ampton, MA, USA). A parametric one-way analysis of variance (ANOVA) was performed followed by the Tukey's multiple com-parison post hoc test. The data was considered significantly different if p< 0.05. The limits of detection of the experimental approach were defined as the concentration which gives an in-strument signal significantly different from the ‘blank’ or ‘back-ground’ signal. The concentration of the analyte (cytokines/ chemokines) was considered above the detection limit when giving a signal equal to the blank signal plus three standard deviations of the blank [41].
3. Results
3.1. Material characteristics
Both the primary characteristics of the as-obtained GRM and CB as well as that of the dispersions used and the deposited masses after aerosolization are summarized in Table 1. Micrographs
obtained by TEM and SEM revealed that no major alterations in morphology of GO and GNP occurred after the aerosolization pro-cess (Fig. 1; and in our previous study [28]). The mass deposition of the aerosolized GO and GNP at the two input concentrations (0.25 and 1 mg/mL), quantified via the online monitoring QCM system, revealed reproducible and controlled deposition resulting in two deposited dose ranges (referred as the “low” and “high” dose) (Table 1). Namely, the low deposition dose was 150e430 ng/cm2for GO or 150e500 ng/cm2 for GNP, and the high dose was 840e1020 ng/cm2for GO or 650e1100 ng/cm2for GNP, respectively (Table S1). For exposures of the TCCC to the reference material CB at the pseudo-ALI conditions, concentrations of CB dispersions were adjusted to yield the low and high ranges, i.e. 300 and 1000 ng/cm2. Determination of surface dose coverage after GRM aerosolization is essential for a valid interpretation of respective biological response. The surface coverage deposition was determined from automati-cally captured series of TEM images of aerosolized GO and GNP. The number of layers (thickness) of GO and GNP differed considerably: GNP formed much thicker stacks compared to GO (Fig. 1;Fig. S3). Consequently, surface coverage of GNP was considerably lower (below 1%; of the total surface) for both the low and high doses whereas the higher dose of GO resulted in 44% surface coverage (Table 2).
Upper panels: GO in dispersion (a; in water at 10
m
g/mL) and aerosolized at the low (b; 150e430 ng/cm2) and high (c; 840e1020 ng/cm2) ranges; lower panels: GNP in dispersion (d; in water at 10m
g/mL) and aerosolized at the low (e; 150e500 ng/cm2)and high (f; 650e1100 ng/cm2) of deposited masses (assessed by QCM;Table S1). Note: Due to the aerosolization process there can be NaCl crystals present along with the deposited GRM (noticeable as spherical spots in b).
3.2. Impact of GRMs on cell viability and lung cell morphology The TCCC wasfirst characterized and the three cell types, each represented in a different colour, were visualized by LSM. The presence of the alveolar epithelial layer (violet) on the upper side of
the insert with incorporated MDM (yellow) could be shown, while at the basal side MDDC (green) were observed (Fig. 2a). Biological effects of the aerosolized GRM and CB at the pseudo-ALI conditions on the TCCC were assessed after the post-exposure period (24 h). LDH assay was applied for evaluation of the exposure-induced membrane rupture andfluorescence staining for visualization of cell morphology of the TCCC. Single exposure to GRM did not induce membrane rupture of the TCCC cells in all the scenarios (Fig. 2). The release of LDH by potentially ruptured TCCC was measured in cell culture medium collected from the bottom com-partments of the experimental chamber (i.e. TCCC grown on the PET membranes). Therefore, an optical interference with GRM or CB can be excluded as the material is deposited on the apical side of the membrane. This is an advantage of the experimental setup used herein as generally GRM may cause strong interferences with such biological assays [42]. Besides, interference tests with all materials were performed to assure the absence of interference with the assay (Fig. S4). Exposure to the higher concentration of non-ionic Pluronic F-127, used as a dispersant for GNP and CB, did not elicit membrane rupture, under aerosolized or pseudo-ALI conditions.
In concordance to the observed absence of membrane rupture, no alterations in cell morphology, i.e. cell nuclei and F-actin cyto-skeleton, were observed 24 h post-exposure in all the scenarios (LSM imaging of thefluorescent stained TCCC samples;Fig. 3). 3.3. Impact of GRMs on the oxidative stress state and (pro-) inflammatory response
In the next step, the potential of GRM to induce oxidative stress and promote secretion of (pro-)inflammatory cytokines/chemo-kines (IL-8, IL-1
b
and TNF-a
) was assessed 24 h post-exposure us-ing the Glutathione Assay Kit and DuoSet®ELISA kits, respectively. Single exposures to all the tested materials did not reduce the total intracellular GSH content in TCCC extracts (Fig. 4A) or promote secretion of the tested (pro-)inflammatory markers with respect to the negative control co-cultures. Exposures to the higher concen-tration of non-ionic Pluronic F-127 (dispersant for GNP and CB;aerosolized or under the pseudo-ALI conditions) did not induce depletion of the total GSH levels and secretion of the (pro-)in-flammatory markers (also previously observed [42]). Secretion of all the three tested pro-inflammatory cytokines was induced by the bacterial endotoxin (LPS, 1
m
g/mL;Fig. 4B). Cytokines TNF-a
and IL-1b
were below the detection limits of the experimental set-up for all the exposure samples (Fig. 4C and 4D), respectively. Control experiments for interference were conducted to assure reliability of the ELISA results (Fig. S5).4. Discussion
Graphene and GRM have become the lodestar material within the last 10 years due to their outstanding properties which enables a broad variety of applications [5]. Therefore, it is not surprising that an increasing mass production of GRM is observed and that a plethora of various types of GRM are commercially available and will further increase in the future [3,11]. As for any emerging ma-terial or technology, a careful assessment and confirmation of its
Table 2
GO, GNP and CB dispersions and deposition characteristics.
Material Graphene oxide (GO) Graphene nanoplatelets (GNP) Carbon black (CB), reference material
Stock dispersion medium water Pluronic F-127
(160 ppm, in water)
Pluronic F-127 (160 ppm, in water) Endotoxin contamination
at 31mg/mL GRM (LAL gel clot)
<0.03 EU/mL 0.96 EU/mL 0.03 EU/mL Chromogenic LAL assay Endosafe®PTS, Charles River
(Sensitivity 0.01 EU/mL)
<0.010 EU/mL 0.015e0.021 EU/mL 0.021e0.028 EU/mL
Doses (range) Low/High Low/High Low/High
Deposited mass (by QCM) 145e430 ng/cm2 840e1020 ng/cm2 150e500 ng/cm 2/ 650e1100 ng/cm2 300 ng/cm2/ 1000 ng/cm2 Surface coverage (by TEM) 14.1± 2.3%/ 44.7± 16.6% 0.4± 0.2%/ 0.8± 0.4% NA
Abbreviations: LAL gel clot, Limulus amebocyte lysate assay; EU, enzyme unit; QCM, quartz crystal microbalance system; TEM, transmission electron microscopy; NA, not applicable.
Fig. 2. Characterization of the TCCC (a) and cell membrane rupture (b) assessed with the LDH release assay 24 h post-exposure. (a) Confocal laser scanning micrograph of the 3D co-culture model after 3D rendering: epithelial cells (violet), MDM (yellow), and MDDC (green). (b) The values are expressed as a fold increase over the respective negative controls, i.e. co-cultures treated with aerosolized NaCl water solution (1.5 mM) or a droplet of water (25mL) on the TCCC (pseudo-ALI conditions) (dotted line). Abbreviations and notes: GO, graphene oxide; GNP, graphene nanoplatelets; CB, carbon black; Triton X-100, 0.2% (v/v) in PBS, a detergent used as positive control for membrane rupture. Symbols on the box plots: Individual data values obtained from each biological repetition:B for GO,◊for GNP,, for CB, and x for positive control. Mean values are denoted with a line and statistically significant difference with # (p < 0.05; One way ANOVA, Origin 2016).
safety and sustainability is needed. To overcome the inconsistency in naming of GRM, a coherent approach describing the various forms of GRM is required [1]. A classification frame based on three physical-chemical descriptors specific for graphene and GRM was proposed by Wick [2] and co-workers and Kostarelos [4] and co-workers and has been already applied to facilitate the structure-activity relationships analysis of GRM in the context of their safety profile. The conclusions of a first literature mining of in vivo animal studies were that the lungs are the organ with the highest GRM accumulation, compared to the liver and spleen, and the side of reported adverse effects induced by respiratory exposure to GRM [9]. However, there is still an urgent need for more studies in this area in combination with improved GRM characterization. 4.1. GRM and CB physical-chemical and endotoxin characterization
The two application-relevant GRM (GO reviewed in Refs. [3,4,43] and GNP [44e46]) and CB as a reference material differ clearly from each other in regard of manufacturing, number of layers (thickness), C/O ratio, zeta-potential and atomic organi-sation of carbon (amorphous vs. hexagonal structure) (Tables 1 and 2). This allows afirst but rough correlation of the material prop-erties with biological response if one of these materials would be hazardous.
All three applied materials were thoroughly characterized including potential endotoxin contamination. According to US Food and Drug Administration (FDA) guidelines, the limit value for endotoxin content in materials intended for medical applications is 0.5 EU/mL [47]. The obtained endotoxin values determined by the LAL Gel Clot Assay and the Chromogenic LAL Assay were compa-rable among the assays for CB and, in part, also for GO (considering potential slight dilution errors and different limits of detection for the assays; Table 2). The endotoxin value of GNP could not be precisely determined by the LAL Gel Clot Assay as even low con-centration of 0.98
m
g GNP/mL led to clot formation and interference in form of assay enhancement could not be totally excluded (data not shown). Interference-enhancement control tests were techni-cally not possible in this case. Nevertheless, based on the obtained results the endotoxin values in GNP dispersion (31.25m
g/mL) can be estimated to be 1 EU/mL, whereas GO and CB showed low amounts of detected endotoxin at the same concentrations with both assays. Despite the slight endotoxin contamination of the material dispersions, the amount of endotoxin getting into contactwith cells after aerosolization and deposition can be estimated to be very low. Assuming an endotoxin value of 1 EU/mL in GNP dis-persions (31.25
m
g/mL GNP), only roughly 0.03 EU/cm2 were deposited at highest GNP deposition dose (1000 ng/cm2). For CB and GO the estimated deposited endotoxin doses were consider-ably lower (i.e. ~0.0005 EU/cm2and ~0.01 EU/cm2for the higher doses of GNP and CB, respectively). The absence of secretion of the (pro-)inflammatory markers, TNF-a
, IL-1b
and IL-8 (Fig. 4BeD), further supports that endotoxin contamination of the materials was minimal.4.2. Material deposition by VitroCell®cloud system
To assess potential adverse effect of GRM after inhalation, the well-established VitroCell®Cloud system combined with the hu-man TCCC, representing a reliable alveolar barrier model, was used [32,33]. This model has been validated for evaluation of respiratory hazard of different substances, including environmental particles or NM such as CNTs [42,48e51]. In addition, efficient aerosolization systems are available for spherical as well as fibre-shaped NM [50,51]. In the current study, the materials were suspended in pure water either with 160 ppm Pluronic F127 for GNP and for GO prior nebulization. The non-ionic detergent Pluronic has been reported to be biocompatible and frequently used in galenics [52]. Due to its weak binding to GRM, it is expected to be rapidly displaced by proteins or lipids from the surfactant layer of the epithelial cell culture [30].
The nebulized GO and GNP exhibited similar morphological structures compared to the shapes observed in suspension (Fig. 1; and [28]) indicating that the potential sheer forces during aero-solization procedure were not affecting the material properties [31].
The deposited masses of aerosolized GO and GNP (Table 2, Table S1), referred to as the low (up to 500 ng/cm2) and the high
(around 1000 ng/cm2) dose range, were considered as an acute rather than lifetime exposure scenarios. However, data on the realistic pulmonary load of inhaled GRM are almost non-existent [9]. There are only a few publications based onfield studies at the production sites of GRM that have reported on the number and mass concentrations of GRM particles in air for GNP only [44,45,52]. Also, there is no clear recommendation given from the exposure assessment side on the number of such evaluations required for a representative assessment of GRM concentrations with human
Fig. 3. Confocal laser scanning micrographs of the apical sides of the co-cultures exposed to the aerosolized GO (a, e) and GNP (b, f), CB (pseudo-ALI) (c, g), at the low (a, b, c) and high (e, f,g) deposited dose ranges, respectively. Negative control samples were nebulized with NaCl water solution (d) or a droplet of water (25mL) (pseudo-ALI conditions) (h). Fluorescence labelling: nuclei in cyan (DAPI), cytoskeleton (F-actin) in magenta (rhodamine-phalloidin). Projection views across xy- and xz-axes are shown.
exposure and health relevance. Hence, the interpretation was made on the realistic exposure data for CNT and CB measured in air at the manufacturing and development sites. As reported for CNT, a short term respiratory exposure (24 h) would result in the alveolar mass retention in the range of 100 ng/cm2[53,54], whereas full working lifetime (8 h per day, 5 days/week, 52 weeks; 1 mg/m3CNT aerosol concentration) between 10 and 50
m
g/cm<sup>2</sup> [54]. Regarding CB, there is very limited evidence on realistic occupa-tional exposures to CB particles (and not mixtures of CB with other compounds) and their impact on the human respiratory system. The currently accepted threshold limit value (TLV) for workplace CB particles is 3.5 mg/m3 [55]. However, an example of the mean concentration of realistic CB content at a workplace was measured to be more than 4-times higher (14.90 mg/m3) [55]. If these values are translated into the above-mentioned calculation for CNT and the normal breathing pattern parameters, an estimated alveolar mass retention of such CB particles would exceed 1m
g/cm3after a short-time exposure and could range even up to 700m
g/cm<sup>2</sup> after a life-time exposure (note: the values presented are only rough estimations). Indeed, Zhang and col-leagues observed a close association between exposure to CB and the levels of IL-1
b
, IL-6, IL-8, MIP-1b
, and TNF-a
in the serum of workers after one-year exposure at the workplace [55]. Although such translations of parameters from CNT to GRM are far from optimal, such comparisons are currently, to the best of our knowledge, considered as the closest estimations of a realistic hu-man exposure to GRM. In general, accurate dosimetry is a para-mount factor in inhalation toxicological studies [56]. A detailed understanding of deposited material and surface covering of cell layers or tissues is required for a precise dose-response assessment, comparison among studies, and data extrapolation across species or to realistic exposure doses [47]. Deposited masses of aerosolized GRM were assessed simultaneously with exposures of the TCCC model to the material using the online monitoring apparatus for assessment of deposited masses (QCM). QCM, directly coupled to the aerosolization system, displayed that two different inputFig. 4. Biochemical response of the 3D lung model after single exposures to aerosolized GO and GNP and CB at the pseudo-ALI conditions, assessed 24 h post-exposure. (a): The total intracellular GSH per total protein content. (b): Secretion of the (pro-)inflammatory chemokine IL-8. (c) and (d): Secretion of the (pro-)inflammatory cytokines TNF-aand IL-1b. The values under (a) are expressed as the total intracellular GSH content (inmM) normalized to the total protein content (mg/mL), whereas for the cytokines IL-8, TNF-aand IL-1b(b, c and d, respectively) in pg/mL. For TNF-aand IL-1ball but LPS-exposed TCCC were below the detection limits,i.e. at 40 pg/mL and 10 pg/mL, respectively (grey dotted line). Abbreviations and notes: GO, graphene oxide; GNP, graphene nanoplatelets; CB, carbon black; GSH, the total intracellular glutathione; BSO,L
-buthionine-sulfoximine, positive control for GSH; LPS, lipopolysaccharide, positive control for (pro-)
inflammation IL-8, interleukin 8; TNF-a, tumour necrosis factor alpha; IL-1b, interleukin 1-beta; NC neb, negative control TCCC exposed to the aerosolized NaCl water solution; NC p-ALI, negative control TCCC exposed at the air-liquid interface. Symbols on the box plots: Individual data values obtained from each biological repetition:B for GO,◊for GNP,, for CB,9 for negative control and x for positive control. Mean values are denoted with a line and statistically significant difference with # (p < 0.05; One way ANOVA, Origin 2016). (A colour version of thisfigure can be viewed online.)
concentrations of GRM resulted in two distinctive mass per surface area ranges (in the study referred to as the low and the high ranges (Table 1, Table S1). Low dose/high dose-dependent deposition patterns were confirmed by TEM-based image surface coverage analysis (Table 1,Table S1). Surface coverage by the high dose of GO was rather high: nearly 45% of surface was assumingly covered by the material (Table S1), whereas GNP at both the doses covered less than 1% of the total surface area. The exposed surface of the (nano) material to the cellular/tissue models has been considered as one of the most important metrics in inhalation (nano)toxicology studies, presumably correlated to lung inflammation [56,57]. Regarding GRM, close surface interaction with diverse mammalian cell types has been observed in vitro [58,59] hence, the important advantage of the experimental approach employed is the full control of the amount and of the portion of cell surface covered by aerosolized GRM.
4.3. Biological response: no acute adverse effects
Independent of the chemistry and shape of the three tested materials, single exposures did not induce significant acute response of the 3D lung model. Viability and morphology of the TCCC model were preserved within the 24 h post-exposure period (Figs. 1 and 2). GRM toxicity has been frequently correlated with the oxidative stress paradigm, based on both in vivo animal studies and in vitro mammalian cell studies [19,58,60] reviewed in Ref. [61]. In 24 h post-exposure period, the total intracellular GSH levels were not depleted upon exposure to the tested GRM and CB, indicating that potential formation of reactive oxygen species (ROS), as re-ported by Refs. [28,62], is not sufficient to manifest oxidative stress-based damages in the human lung model.
In general, release of (pro-)inflammatory cytokines/chemokines from activated epithelial cells, MDM and MDDC is considered as the pivotal reaction of lung cells to interaction with deposited material [62e65]. Single exposures to the tested materials did not induce secretion of the three selected (pro-)inflammatory cytokines/che-mokines (TNF-
a
, IL-8, IL-1b
) compared to the negative control TCCC systems. It should be noted that the potential presence of endotoxin (LPS) in GRM samples has not been considered in several previous studies involving immune cells or other LPS-sensitive cell types. Therefore conflicting results have been published and comparison remains difficult. Recently, Mukherjee [31] and co-authors reported that TNF-a
production by human MDM exposed to different GRM samples under submerged conditions resulted at least partially from endotoxin contamination of the GRM. Nevertheless, some GRM did induce a slight increase in TNF-a
secretion based on their intrinsic material properties, especially their lateral dimensions [31].With respect to in vivo animal studies, lung inflammation in mice has been observed after an acute 1-day exposure to GNP [12,66]. It has been proposed that this effect was correlated with the lateral size of the nanoplatelets with lateral dimensions of >5
m
m inducing higher effects than those of 1e2m
m GNP [15]. After a single instillation, GNP remained in mouse lungs for 28 [11] and 90 [16] days and provoked a sub-chronic inflammatory response with the secretion peak of (pro-)inflammatory cytokines on day 14, including TNF-a
and IL-1b
[11]. A single inhalation exposure to GO induced minimal toxic and (pro-)inflammatory responses in rat lungs [14].For GO, administration in mouse lungs led to the generation of ROS and activation of (pro-)inflammatory and apoptotic pathways, resulting in severe and persistent lung injuries [67]. Similarly to GNP, GO toxicity and (pro-)inflammatory potential appeared to be strongly correlated with their lateral sizes both in vivo and in vitro. Larger GO sheets were proposed to adsorb to the plasma membrane
and activate nuclear factor-kappa B (NF-
k
B) pathways, whereas smaller ones were observed to be mostly taken up by macrophages [23]. The apparently conflicting results of different GRM upon inhalation [11e16,67] (and reviewed in Refs. [9,17]) can be attrib-uted to (i) the significant differences in animal vs human lung physiology [68], (ii) the significantly different properties of the GRM types used as well as the presence of biological contamina-tions such as endotoxins [9] and (iii) the different applied doses, including very high or even over dose regime.To allow a better correlation of the result among studies the inclusion of at least one benchmark material would allow to place the outcomes into the context of the existing literature [2]. Here, CB was used as a non-2D amorphous carbon benchmark material. Printex 85 [29] as well as Printex 90 [12,13,69], two types of CB, are reported not to induce an acute toxic response in vitro or in vivo.
Moreover, various types of carbon-based nanomaterials including carbon nanotubes (CNT) and cellulose nanocrystals (CNC), have already been extensively studied in the TCCC model [36,42]. This enables the assessment and comparison of the GRM in a much broader context. Single exposures to nebulized multi-walled CNT yielding deposited masses up to 0.4
m
g/cm<sup>2</ sup> did not elicit neither morphological alterations nor any cytotoxicity, oxidative stress or (pro-)inflammation reactions, yet individual CNT were observed to be present inside both MDM and A549 cells [42]. Similarly, CNC at a deposited dose of around 0.8m
g/ cm<sup>2</sup> did not induce any cytotoxic, morphological, oxidative stress or (pro-)inflammatory related responses. In contrast, the positive particle control (nebulized crystalline quartz, DQ12), elicited a (pro-)inflammatory response and a slight decrease in the intracellular GSH content at sub-lethal concentrations (i.e. deposited mass of around 0.2m
g/cm<sup>2</sup>), proving bio-logical sensitivity and responsiveness of the TCCC model [36].The absence of adverse acute toxicity in our study is in line with the general notion that GRM induce less undesired adverse effects than CNT [13]. If the lack of adverse responses of GRM can be further confirmed in vitro and or in vivo, GRM would be potential candidates to replace other, more critical materials, such as CNT in different applications such as nanofillers for composite materials [70] or biomedical applications [71e73]. The replacement or modification of critical substances towards non-critical substitutes without reducing the product performance underlines the success of the implementation of the safer-by-design concept. Therefore no-effects studies are as important as effects studies since this knowledge is part of an evidenced-based assessment of emerging materials.
5. Conclusion
Single exposures of the 3D human lung TCCC model to aero-solized GO and GNP at realistic human exposure scenarios did not affect cell viability or compromised cell morphology, nor did they induce the secretion of (pro-)inflammatory mediators or showed increased oxidative stress levels. Although GRM with highly distinct properties were used to assess their impact on biological effects, none of the addressed GRM properties, such as manufacturing, C/O ratio, or number of graphene layer did elicit an acute response. This was shown by combining an advanced human in vitro lung model with realistic particle application and realistic doses as afirst stage assessment of their human respiratory hazard. Only a cost and time effective monitoring of undesired side effects early in the development supports the growing global GRM pro-duction and the rapid development of high-technology and biomedical applications in a safe and sustainable way.
Acknowledgment
The research leading to these results has received funding from the European Union (EU) Seventh Framework Programme Gra-phene Flagship project (Grant Agreement No. 604391) and the EU Horizon 2020 Framework Graphene Flagship project Graphe-neCore1 (Grant Agreement No. 696656), from the NanoScreen materials challenge co-funded by the Competence Centre for Ma-terials Science and Technology (CCMX) as well as Swiss National Science Foundation (Grant 310030_169207). In addition, the Adolphe Merkle Foundation is acknowledged as well as the tech-nical support with TEM imaging by Ms Laetitia Haeni and the support with the 3D rendering of the TCCC by Dr. Dedy Septiadi. Appendix A. Supplementary data
Supplementary data related to this article can be found at https://doi.org/10.1016/j.carbon.2018.05.012.
References
[1] A. Bianco, H.M. Cheng, T. Enoki, Y. Gogotsi, R.H. Hurt, N. Koratkar, et al., All in the graphene family - a recommended nomenclature for two-dimensional carbon materials, Carbon 65 (2013) 1e6.
[2] P. Wick, A.E. Louw-Gaume, M. Kucki, H.F. Krug, K. Kostarelos, B. Fadeel, et al., Classification framework for graphene-based materials, Angew. Chem. Int. Ed. 53 (30) (2014) 7714e7718.
[3] Y. Volkov, J. McIntyre, A. Prina-Mello, Graphene toxicity as a double-edged sword of risks and exploitable opportunities: a critical analysis of the most recent trends and developments, 2D Mater. 4 (2) (2017).
[4] K. Kostarelos, K.S. Novoselov, Graphene devices for life, Nat. Nanotechnol. 9 (10) (2014) 744e745.
[5] A.C. Ferrari, F. Bonaccorso, V. Fal'ko, K.S. Novoselov, S. Roche, P. Boggild, et al., Science and technology roadmap for graphene, related two-dimensional crystals, and hybrid systems, Nanoscale 7 (11) (2015), 4598e4810. [6] W.C. Ren, H.M. Cheng, The global growth of graphene, Nat. Nanotechnol. 9
(10) (2014) 726e730.
[7] A. Bianco, Graphene: safe or toxic? The two faces of the medal, Angew. Chem. Int. Ed. 52 (19) (2013) 4986e4997.
[8] A. Bianco, M. Prato, Safety concerns on graphene and 2D materials: a Flagship perspective, 2D Mater. 2 (3) (2015).
[9] C. Bussy, D. Jasim, N. Lozano, D. Terry, K. Kostarelos, The current graphene safety landscape - a literature mining exercise, Nanoscale 7 (15) (2015) 6432e6435.
[10] http://www.idtechex.com/research/reports/graphene-2d-materials-and- carbon-nanotubes-markets-technologies-and-opportunities-2017-2027-000530.asp.
[11] E.J.L.G. Park, B.S. Han, B.S. Lee, S. Lee, M.H. Cho, J.H. Kim, D.W. Kim, Toxic response of graphene nanoplatelets in vivo and in vitro, Arch. Toxicol. 89 (9) (2015), 1157e1568.
[12] A. Schinwald, F.A. Murphy, A. Jones, W. MacNee, K. Donaldson, Graphene-based nanoplatelets: a new risk to the respiratory system as a consequence of their unusual aerodynamic properties, ACS Nano 6 (1) (2012) 736e746. [13] L. Ma-Hock, V. Strauss, S. Treumann, K. Kuttler, W. Wohlleben, T. Hofmann, et
al., Comparative inhalation toxicity of multi-wall carbon nanotubes, graphene, graphite nanoplatelets and low surface carbon black, Part. Fibre Toxicol. (2013) 10.
[14] S.G. Han, J.K. Kim, J.H. Shin, J.H. Hwang, J.S. Lee, T.G. Kim, et al., Pulmonary responses of sprague-dawley rats in single inhalation exposure to graphene oxide nanomaterials, BioMed Res. Int. 2015 (2015) 376756.
[15] J.R. Roberts, R.R. Mercer, A.B. Stefaniak, M.S. Seehra, U.K. Geddam, I.S. Chaudhuri, et al., Evaluation of pulmonary and systemic toxicity following lung exposure to graphite nanoplates: a member of the graphene-based nanomaterial family, Part. Fibre Toxicol. (2016) 13.
[16] E. Park, S. Lee, K. Lee, Y. Choi, B. Lee, G. Lee, et al., Pulmonary persistence of graphene nanoplatelets may disturb physiological and immunological ho-meostasis, J. Appl. Toxicol. 37 (3) (2017) 296e309.
[17] V.C. Sanchez, A. Jachak, R.H. Hurt, A.B. Kane, Biological interactions of graphene-family nanomaterials: an interdisciplinary review, Chem. Res. Toxicol. 25 (1) (2012) 15e34.
[18] N.V.S. Vallabani, S. Mittal, R.K. Shukla, A.K. Pandey, S.R. Dhakate, R. Pasricha, et al., Toxicity of graphene in normal human lung cells (BEAS-2B), J. Biomed. Nanotechnol. 7 (1) (2011) 106e107.
[19] L. Horvath, A. Magrez, M. Burghard, K. Kern, L. Forro, B. Schwaller, Evaluation
of the toxicity of graphene derivatives on cells of the lung luminal surface, Carbon 64 (2013) 45e60.
[20] S.C. Reshma, S. Syama, P.V. Mohanan, Nano-biointeractions of PEGylated and bare reduced graphene oxide on lung alveolar epithelial cells: a comparative
in vitro study, Colloids Surfaces B Biointerfaces 140 (2016) 104e116. [21] A.B. Seabra, A.J. Paula, R. de Lima, O.L. Alves, N. Duran, Nanotoxicity of
gra-phene and gragra-phene oxide, Chem. Res. Toxicol. 27 (2) (2014) 159e168. [22] X. Wang, M.C. Duch, N. Mansukhani, Z. Ji, Y.-P. Liao, M. Wang, et al., Use of a
pro-fibrogenic mechanism-based predictive toxicological approach for tiered testing and decision analysis of carbonaceous nanomaterials, ACS Nano 9 (3) (2015) 3032e3043.
[23] J. Ma, R. Liu, X. Wang, Q. Liu, Y. Chen, R.P. Valle, et al., Crucial role of lateral size for graphene oxide in activating macrophages and stimulating pro-inflammatory responses in cells and animals, ACS Nano 9 (10) (2015) 10498e10515.
[24] G. Lalwani, M. D'Agati, A.M. Khan, B. Sitharaman, Toxicology of graphene-based nanomaterials, Adv. Drug Deliv. Rev. 105 (2016). Part B:109-44. [25] L. Müller, M. Riediker, P. Wick, M. Mohr, P. Gehr, B. Rothen-Rutishauser,
Oxidative stress and inflammation response after nanoparticle exposure: differences between human lung cell monocultures and an advanced three-dimensional model of the human epithelial airways, J. R. Soc. Interface 7 (Suppl 1) (2010) S27eS40.
[26] F.R. Paur HRC, J. Teeguarden, H. Fissan, S. Diabate, M. Aufderheide,
W.G. Kreyling, O. H€anninen, G. Kasper, M. Riediker, B. Rothen-Rutishauser, O. Schmid, In-vitro cell exposure studies for the assessment of nanoparticle toxicity in the lungda dialog between aerosol science and biology, J. Aerosol Sci. 42 (10) (2011).
[27] C.W. Lam, J.T. James, R. McCluskey, R.L. Hunter, Pulmonary toxicity of single-wall carbon nanotubes in mice 7 and 90 days after intratracheal instillation, Toxicol. Sci. 77 (1) (2004) 126e134.
[28] M. Kucki, P. Rupper, C. Sarrieu, M. Melucci, E. Treossi, A. Schwarz, et al., Interaction of graphene-related materials with human intestinal cells: an in vitro approach, Nanoscale 8 (16) (2016) 8749e8760.
[29] A. Figarol, J. Pourchez, D. Boudard, V. Forest, C. Akono, J.M. Tulliani, et al., In vitro toxicity of carbon nanotubes, nano-graphite and carbon black, similar impacts of acid functionalization, Toxicol. Vitro 30 (1) (2015) 476e485. [30] P. Cherukuri, C.J. Gannon, T.K. Leeuw, H.K. Schmidt, R.E. Smalley, S.A. Curley, et
al., Mammalian pharmacokinetics of carbon nanotubes using intrinsic near-infraredfluorescence, Proc. Natl. Acad. Sci. Unit. States Am. 103 (50) (2006) 18882e18886.
[31] S.P. Mukherjee, N. Lozano, M. Kucki, A.E. Del Rio-Castillo, L. Newman, E. Vazquez, et al., Detection of endotoxin contamination of graphene based materials using the TNF-alpha expression test and guidelines for endotoxin-free graphene oxide production, PLoS One 11 (11) (2016), e0166816. [32] B.M. Rothen-Rutishauser, S.G. Kiama, P. Gehr, A three-dimensional cellular
model of the human respiratory tract to study the interaction with particles, Am J Resp Cell Mol. 32 (4) (2005) 281e289.
[33] B. Rothen-Rutishauser, L. Muller, F. Blank, C. Brandenberger, C. Muhlfeld, P. Gehr, A newly developed in vitro model of the human epithelial airway barrier to study the toxic potential of nanoparticles, Altex-Altern. Tierexp. 25 (3) (2008) 191e196.
[34] A. Lehmann, C. Brandenberger, F. Blank, P. Gehr, B. Rothen-Rutishauser, A 3D model of the human epithelial airway barrer, in: T. Maguire, E. Novik (Eds.), Methods in Bioengineering Alternative Technologies to Animal Testing, Artech House, Boston, London, 2010, pp. 239e260.
[35] H.R. Paur, S. Mulhopt, C. Weiss, S. Diabate, In vitro exposure systems and bioassays for the assessment of toxicity of nanoparticles to the human lung, J. Verbrauch Lebensm 3 (3) (2008) 319e329.
[36] C. Endes, O. Schmid, C. Kinnear, S. Mueller, S. Camarero-Espinosa, D. Vanhecke, et al., An in vitro testing strategy towards mimicking the inha-lation of high aspect ratio nanoparticles, Part. Fibre Toxicol. 11 (40) (2014) 40. [37] M.J.G.P. Clift, B. Rothen-Rutishauser, Nanotoxicology: a perspective and dis-cussion of whether or not in vitro testing is a valid alternative, Arch. Toxicol. 85 (7) (2011) 723e731.
[38] L.T.A.F. Farcal, L. Di Cristo, B.M. Rotoli, O. Bussolati, E. Bergamaschi, A. Mech, N.B. Hartmann, K. Rasmussen, J. Riego-Sintes, J. Ponti, A. Kinsner-Ovaskainen, F. Rossi, A. Oomen, P. Bos, R. Chen, R. Bai, C. Chen, L. Rocks, N. Fulton, B. Ross, G. Hutchison, L. Tran, S. Mues, R. Ossig, J. Schnekenburger, L. Campagnolo, L. Vecchione, A. Pietroiusti, B. Fadeel, Comprehensive in vitro toxicity testing of a panel of representative oxide nanomaterials: first steps towards an intelligent testing strategy, PLoS One 10 (5) (2015 May 21), e0127174. [39] E.V.D. Durantie, L. Rodriguez-Lorenzo, F. Delhaes, S. Balog, D. Septiadi,
J. Bourquin, A. Petri-Fink, B. Rothen-Rutishauser, Biodistribution of single and aggregated gold nanoparticles exposed to the human lung epithelial tissue barrier at the air-liquid interface, Part. Fibre Toxicol. (2017) 14.
[40] M. Abramoff, P.J. Magalhaes, S.J. Ram, Image processing with Image, J. Biophot. Int. 11 (7) (2004) 36e42.
[41] J.N. Miller, J.C. Miller, Statistics and Chemometrics for Analytical Chemistry, 6 ed., Pearson Education Limited, Harlow, UK, 2010.
[42] S. Chortarea, M.J. Clift, D. Vanhecke, C. Endes, P. Wick, A. Petri-Fink, et al., Repeated exposure to carbon nanotube-based aerosols does not affect the functional properties of a 3D human epithelial airway model, Nanotoxicology 9 (8) (2015) 983e993.
[43] C. Chung, Y.K. Kim, D. Shin, S.R. Ryoo, B.H. Hong, D.H. Min, Biomedical ap-plications of graphene and graphene oxide, Accounts Chem. Res. 46 (10) (2013) 2211e2224.
[44] L.M. Lo, D. Hammond, I. Bartholomew, D. Almaguer, W.A. Heitbrink, J. Topmiller, Engineering Controls for Nano-scale Graphene Platelets during Manufacturing and Handling Processes, Department of health and human
services Centers for Disease Control and Prevention National Institute for Occupational Safety and Health, Branch DoARaTEaPH, 2011, 2011. Contract No.: EPHB Report No. 356-12a.
[45] W.A. Heitbrink, L.M. Lo, K.H. Dunn, Exposure controls for nanomaterials at three manufacturing sites, J. Occup. Environ. Hyg. 12 (1) (2015) 16e28. [46] J.H. Lee, J.H. Han, J.H. Kim, B. Kim, D. Bello, J.K. Kim, et al., Exposure monitoring
of graphene nanoplatelets manufacturing workplaces, Inhal. Toxicol. 28 (6) (2016) 281e291.
[47] G.M.C.J. DeLoid, G. Pyrgiotakis, S.V. Pirela, A. Pal, J. Liu, J. Srebric, P. Demokritou, Advanced computational modeling for in vitro nanomaterial dosimetry, Part. Fibre Toxicol. 12 (32) (2015).
[48] M.J.D. Clift, C. Endes, D. Vanhecke, P. Wick, P. Gehr, R.P.F. Schins, et al., A comparative study of different in vitro lung cell culture systems to assess the most beneficial tool for screening the potential adverse effects of carbon nanotubes, Toxicol. Sci. 137 (1) (2014) 55e64.
[49] I. Tomasek, C.J. Horwell, D.E. Damby, H. Barosova, C. Geers, A. Petri-Fink, et al., Combined exposure of diesel exhaust particles and respirable Soufriere Hills volcanic ash causes a (pro-)inflammatory response in an in vitro multicellular epithelial tissue barrier model, Part. Fibre Toxicol. (2016) 13.
[50] A.G. Lenz, E. Karg, B. Lentner, V. Dittrich, C. Brandenberger, B. Rothen-Rutishauser, et al., A dose-controlled system for air-liquid interface cell exposure and application to zinc oxide nanoparticles, Part. Fibre Toxicol. 16 (6) (2009) 32.
[51] C. Brandenberger, C. Mühlfeld, Z. Ali, A. Lenz, O. Schmid, W.J. Parak, et al., Quantitative evaluation of cellular uptake and trafficking of plain and poly-ethylene glycol-coated gold nanoparticles, Small 6 (15) (2010) 1669e1678. [52] A.C.K.B. Sabuncu, S. Qian, M.W. Stacey, A. Beskok, Dispersion state and toxicity
of mwCNTs in cell culture medium with different T80 concentrations, Colloids Surfaces B Biointerfaces 78 (1) (2010) 36e43.
[53] C. Bussy, H. Ali-Boucetta, K. Kostarelos, Safety considerations for graphene: lessons learnt from carbon nanotubes, Accounts Chem. Res. 46 (3) (2013) 692e701.
[54] A. Erdely, M.M. Dahm, M.K. Schubauer-Berigan, B.T. Chen, J.M. Antonini, M.D. Hoover, Bridging the gap between exposure assessment and inhalation toxicology: some insights from the carbon nanotube experience, J. Aerosol Sci. 99 (2016) 157e162.
[55] R.D.Y. Zhang, X. Zhang, Y. Niu, T. Meng, Y. Li, H. Duan, P. Bin, M. Ye, X. Jia, M. Shen, S. Yu, X. Yang, W. Gao, Y. Zheng, Reduced pulmonary function and increased pro-inflammatory cytokines in nanoscale carbon black-exposed workers, Part. Fibre Toxicol. 11 (73) (2014 Dec 14),https://doi.org/10.1186/ s12989-014-0073-1.
[56] E.D. Kuempel, L.M. Sweeney, J.B. Morris, A.M. Jarabek, Advances in inhalation dosimetry models and methods for occupational risk assessment and expo-sure limit derivation, J. Occup. Environ. Hyg. 12 (sup1) (2015) S18eS40. [57] T. Stoeger, C. Reinhard, S. Takenaka, A. Schroeppel, E. Karg, B. Ritter, et al.,
Instillation of six different ultrafine carbon particles indicates a surface area threshold dose for acute lung inflammation in mice, Environ. Health Perspect. 114 (3) (2006) 328e333.
[58] W. Zhang, L. Yan, M. Li, R. Zhao, X. Yang, T. Ji, et al., Deciphering the underlying
mechanisms of oxidation-state dependent cytotoxicity of graphene oxide on mammalian cells, Toxicol. Lett. 237 (2) (2015) 61e71.
[59] L. Ou, B. Song, H. Liang, J. Liu, X. Feng, B. Deng, et al., Toxicity of graphene-family nanoparticles: a general review of the origins and mechanisms, Part. Fibre Toxicol. 13 (1) (2016) 57.
[60] S. Mittal, V. Kumar, N. Dhiman, L.K.S. Chauhan, R. Pasricha, A.K. Pandey, Physico-chemical Properties Based Differential Toxicity of Graphene Oxide/ reduced Graphene Oxide in Human Lung Cells Mediated through Oxidative Stress, vol. 6, 2016, p. 39548.
[61] A. Jarosz, M. Skoda, I. Dudek, D. Szukiewicz, Oxidative stress and mitochon-drial activation as the main mechanisms underlying graphene toxicity against human cancer cells, Oxid. Med. Cell Longev 2016 (2016), 5851035. [62] J.T.E. Russier, A. Scarsi, F. Perrozzi, H. Dumortier, L. Ottaviano, M. Meneghetti,
V. Palermo, A. Bianco, Evidencing the mask effect of graphene oxide: a comparative study on primary human and murine phagocytic cells, Nanoscale 5 (22) (2013) 11234e11247.
[63] A.V. Tkach, N. Yanamala, S. Stanley, M.R. Shurin, G.V. Shurin, E.R. Kisin, et al., Graphene oxide, but not fullerenes, targets immunoproteasomes and sup-presses antigen presentation by dendritic cells, Small 9 (2013) 1686e1690. [64] M. Orecchioni, D. Bedognetti, F. Sgarrella, F.M. Marincola, A. Bianco,
L.G. Delogu, Impact of carbon nanotubes and graphene on immune cells, J. Transl. Med. (2014) 138.
[65] K. Donaldson, L. Tran, L.A. Jimenez, R. Duffin, D.E. Newby, N. Mills, et al., Combustion-derived nanoparticles: a review of their toxicology following inhalation exposure, Part. Fibre Toxicol. 2 (1) (2005) 10.
[66] A. Schinwald, F. Murphy, A. Askounis, V. Koutsos, K. Sefiane, K. Donaldson, et al., Minimal oxidation and inflammogenicity of pristine graphene with resi-dence in the lung, Nanotoxicology 8 (8) (2014) 824e832.
[67] M.C. Duch, G.R.S. Budinger, Y.T. Liang, S. Soberanes, D. Urich, S.E. Chiarella, et al., Minimizing oxidation and stable nanoscale dispersion improves the biocompatibility of graphene in the lung, Nano Lett. 11 (12) (2011) 5201e5207.
[68] M. Bracken, Why animal studies are often poor predictors of human reactions to exposure, J. R. Soc. Med. 102 (3) (2009).
[69] K. Driscoll, J. Carter, B. Howard, D. Hassenbein, W. Pepelko, R. Baggs, et al., Pulmonary inflammatory, chemokine, and mutagenic responses in rats after subchronic inhalation of carbon black, Toxicol. Appl. Pharmacol. 136 (2) (1996) 372e380.
[70] X. Hu, QZ, Health and ecosystem risks of graphene, Chem. Rev. 113 (5) (2013) 3815e3835.
[71] Z. Liu, J. Robinson, X. Sun, H. Dai, PEGylated nanographene oxide for delivery of water-insoluble cancer drugs, J. Am. Chem. Soc. 20 (130) (2008) 10876e10877.
[72] X. Sun, Z. Liu, K. Welsher, J. Robinson, A. Goodwin, S. Zari, et al., Nano-gra-phene oxide for cellular imaging and drug delivery, Nano Res. 1 (3) (2008) 203e212.
[73] K. Yang, J. Wan, S. Zhang, Y. Zhang, S. Lee, Z. Liu, In vivo pharmacokinetics, long-term biodistribution, and toxicology of PEGylated graphene in mice, ACS Nano 5 (2011) 516e522.
