Journal of Hazardous Materials 414 (2021) 125471 Contents lists available at ScienceDirect Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat Ex vivo exposure to different types of graphene-based nanomaterials consistently alters human blood secretome Sandra Ballesteros a, Josefa Domenech a, Antonia Velázquez a, b, Ricard Marcos a, b, *, Alba Hernández a, b, * a Group of Mutagenesis, Department of Genetics and Microbiology, Universitat Autònoma de Barcelona, Spain b Consortium for Biomedical Research in Epidemiology and Public Health (CIBERESP), Carlos III Institute of Health, Madrid, Spain A R T I C L E I N F O A B S T R A C T Editor: Dr. S. Nan The biomedical applications of graphene-based nanomaterials (GBN) have significantly grown in the last years. Many of these applications suppose their intravenous exposure and, in this way, GBN could encounter blood cells Keywords: triggering an immunological response of unknown effects. Consequently, understanding the relationships be­ Graphene-based-nanomaterials tween GBN and the immune system response should be a prerequisite for its adequate use in biomedicine. In the Blood secretome present study, we have conducted a little explored ex vivo exposure method in order to study the complexity of Ex vivo exposure the secretome given by the interactions between GBN and blood cells. Blood samples from different healthy Cytokines Inflammation donors were exposed to three different types of GBN widely used in the biomedical field. In this sense, graphene oxide (GO), graphene nanoplatelets (GNPs), graphene nanoribbons (GNRs) and a panel of 105 proteins repre­ sentatives of the blood secretome were evaluated. The results show broad changes in both the cytokines number and the expression levels, with important changes in inflammatory response markers. Furthermore, the indirect soft-agar assay was used as a tool to unravel the global functional impact of the found secretome changes. Our results indicate that the GBN-induced altered secretome can modify the natural anchorage-independent growth capacity of HeLa cells, used as a model. As a conclusion, this study describes an innovative approach to study the potential harmful effects of GBN, providing relevant data to be considered in the biomedical context when GBN are planned to be used in patients. 1. Introduction (GNRs) that are still in a more preclinical phase. Although GBN applications look promising, one limitation to reach Nanomaterials have caught the attention of the scientific community clinical practice is that after its intravenous administration, GBN meet due to their unique physicochemical properties and multiple applica­ blood cells, potentially inducing immune perturbations (Boraschi et al., tions. To date, there is a growing literature about the potential 2018; Fadeel, 2019). Thus, previous works have reported differential biomedical applications of different nanomaterials in fields such as impacts of different forms of graphene oxide on primary human immune bioimaging, transplants, drug delivery, and diagnosis (Ramos et al., cell populations (Orecchioni et al., 2016, 2017). This has also been 2017). observed in the case of other nanomaterials. Hence, it is known that In the case of graphene-based nanomaterials (GBN), they have aluminum oxide nanoparticles induced an altered immune response in shown a significant potential for biomedical applications, including mice when intravenously injected (Park et al., 2016). Likewise, gene therapy, drug delivery, tumor therapy, and theragnostic (Han long-term exposures (up to 60 days) to subtoxic doses of palladium et al., 2019; Song et al., 2020). Additionally, graphene has been widely nanoparticles via intravenous injection alter the cytokine serum levels in used as a contrast agent in magnetic resonance imaging due to its treated Wistar rats, suggesting a possible impact of these nanoparticles magnetic properties (Lin et al., 2016; Loh et al., 2018). To date, the on the immune system after long-term exposures (Iavicoli et al., 2018). applications of graphene oxide are receiving more attention than other Different studies have been conducted to determine the potentially GBN such as graphene nanoplatelets (GNPs) or graphene nanoribbons harmful effects of GBN exposure, as recently reviewed (Madannejad * Corresponding authors at: Group of Mutagenesis, Department of Genetics and Microbiology, Universitat Autònoma de Barcelona, Edifici Cn, Campus de Bel­ laterra, Cerdanyola del Vallès, 08193 Barcelona, Spain. E-mail addresses: ricard.marcos@uab.cat (R. Marcos), alba.hernandez@uab.cat (A. Hernández). https://doi.org/10.1016/j.jhazmat.2021.125471 Received 23 July 2020; Received in revised form 16 January 2021; Accepted 17 February 2021 Available online 20 February 2021 0304-3894/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 et al., 2019). The reported results are heterogeneous depending on EDTA (1%). From each donor, 4 mL of blood were exposed to a final non- different factors including the used model (in vitro vs in vivo), and the cytotoxic concentration of 50 µg/mL of each type of GBN, lipopolysac­ evaluated target, among others. It is true that the results obtained using charides (LPS from Escherichia coli O55:B5; Sigma-Aldrich (Saint Louis, in vivo studies permit an easier extrapolation for human risk assessment; USA)) at 4 µg/mL (positive control), and equal amounts of vehicle nevertheless, such studies have inconveniences such as the cost and the (negative control) for 24 h at 37 ◦ C, and constant agitation at 80 rpm to associated ethical considerations. In this scenario, it has been proposed avoid sedimentation. After the incubation, blood was centrifuged at that ex vivo exposure using human blood represents a better model to 1600g during 10 min to obtain the serum. evaluate the biological interactions of nanomaterials than the use of in vitro studies involving mammalian cell lines (Cui et al., 2015). In fact, it 2.3. Determination of cytokine expression changes has been proved that polyethylene glycol nanoparticles have a different response when evaluated in vitro with 3T3 fibroblast and C1R lympho­ The Human XL cytokine Array Kit (R&D Systems Bio-techne, Prote­ blast cells, than when tested using whole blood as a more complex ome Profiler™) was used to determine changes in the cytokine’s system. Another advantage of using this human model is that the in­ expression induced by each GBN. Cytokine membranes were incubated clusion of different donors–with different genetic backgrounds-can with 350 µL of serum overnight at 4 ◦ C and constant agitation at 100 highlight key differences depending on the donor patient (Mann et al., rpm and then, manufacturer’s instructions were followed. Briefly, serum 2016). samples were diluted and mixed with a cocktail of biotinylated detection According to the above indicated, we have carried out a systematic antibodies. This mixture was then incubated with the array. The array study involving the analysis of the effects of the three GBN mainly used included membranes spotted in duplicate with capture antibodies. Any in biomedicine (GO, GNPs, and GNRs) after incubation in fresh whole cytokine/detection antibody complex present was bound by its cognate blood samples obtained from different donors. Changes in the blood cells immobilized capture antibody on the membrane. Streptavidin- secretome were analyzed, since it is known that it can reflect a potential horseradish peroxidase and chemiluminescent detection reagents were health problem (da Cunha et al., 2019). The analyzed proteins included added, and a signal was produced in proportion to the amount of hormones, cytokines, chemokines, interferons, colony-stimulating fac­ cytokine bound. Then, membranes were visualized using the GeneG­ tors (CSFs), growth factors, tumor necrosis factors (TNFs), and other nome XRQ enhanced chemiluminescence system (ECL, Cell Signaling bioactive molecules. All of them are biomarkers of pathophysiological Technologies, Danvers, MA). Finally, relative quantification of cytokine processes such as cell differentiation, invasion, metastasis, autophagy, expression was assessed using the ImageJ Protein Array (Carpentier, apoptosis, tissue organization, immune surveillance, angiogenesis, and 2009). Data were imported to String Database to generate functional cell-to-cell communication in cancer cells (Patel et al., 2014). Further­ cytokines association networks (Szklarczyk et al., 2015). more, and to understand whether the induced secretome alterations entail further functional implications, an indirect version of the soft-agar 2.4. Indirect soft-agar was also performed (Annangi et al., 2015; Vila et al., 2017). Serum from each donor after treatment with GO, GNPs, GNRs, LPS, 2. Materials and methods or vehicle were used for an indirect soft-agar assay as in Annangi et al. (2015) and Vila et al. (2017) with some modifications. Briefly, HeLa 2.1. Nanomaterials characterization cells used as cell model were collected and filtered through a 40 µm mesh to obtain single-cell suspensions. A suspension of 7 × 104 cells in GO, GNPs and GNRs were all obtained from Sigma-Aldrich (St. Louis, 1.4 mL of each serum was mixed in a 1:1:1 ratio with 1.2% bacto-agar USA). Transmission electron microscopy (TEM) was used to determine (DIFCO, MD, USA), and 2X DMEM supplemented with 20% of FBS, both the size and the morphology of the selected GBN on a JEOL JEM- 2% NEEA, 2% L-Glu 200 mM and 2% of penicillin-streptomycin. This 1400 instrument (Jeol LTD, Tokyo, Japan). Likewise, the characteriza­ mixture was enough to prepare triplicates of 2 × 104 cells per well by tion of both hydrodynamic size and zeta potential of the three selected dispensing 1.5 mL over a 0.6% base-agar (supplemented with 2X GBN were determined, after dispersion in 0.05% bovine serum albumin DMEM) in each well of a 6-well plate. Plates could sit for 30 min, and (BSA), by using dynamic light scattering (DLS) and laser Doppler then kept in the cell incubator. After 15 days of incubation, plates with velocimetry (LDV) methodologies, respectively. Such measurements cell colonies were stained with 1 mg/mL of INT solution (2-p-iodophe­ were performed on a Malvern Zetasizer Nano-ZS zen3600 instrument. nyl-3-p-nitrophenyl-5-phenyl tetrazolium chloride; Sigma, MO, USA). Layered graphene materials, GO and GNPs, were further characterized Finally, plates were scanned at high resolution for image analysis using to determine its thickness in a mechanical Profiler 7.0 P15 (KLA Tencor, the colony cell counter enumerator software OpenCFU (3.9.0), and the California, USA) with a stylus radius of 2 µm and an applied force of 2 NIST’s Integrated Colony Enumerator (NICE) software (NIST, Boulder, mg. For this metrological characterization, stock solutions of graphene CO). nanomaterials were diluted to 50 µg/mL in miliQ water and air-dried on a silicon substrate. 20 randomly selected GO or GNPs particles were 3. Results scanned at 20 µm/s, to determine their thickness. 3.1. GBN characterization 2.2. Whole blood exposure to graphene-based nanomaterials To complement the information obtained from the supplier, further According to the Nanogenotox protocol (Nanogenotox, 2011), GO (2 characterization of the used GBN was carried out. TEM images show the mg/mL) and GNPs (1 mg/mL) were dispersed in 1% bovine serum al­ morphology of the selected GBN after sonication (Fig. 1). As shown in bumin (BSA) in MilliQ water. The nanomaterials in the dispersion me­ Fig. 1A, GO is a single-layered nanomaterial. GNPs are nanoparticles of dium were sonicated for 16 min and a 10% amplitude sonication to different sizes and irregular shapes with higher electron density because obtain a final concentration of 1.9 and 0.95 mg/mL, respectively. GNRs of their multilayer conformation (Fig. 1A). Finally, GNRs present a were prewetted in 0.5% absolute ethanol and dispersed in 0.05% bovine graphene fibbers-like shape (Fig. 1A). To better characterize GNRs, in­ serum albumin (BSA) in MilliQ water. As before, GNRs in the dispersed formation on mean size and SD of the width was calculated by medium were sonicated for 16 min and a 10% amplitude sonication, to measuring isolated fibers in random fields. The mean ± SD values of obtain a final concentration of 2.56 mg/mL. GNRs were 118 ± 39 nm (Fig. 1B). Fresh human whole blood was extracted from 8 healthy donors (4 Other measurements such as the hydrodynamic radius and Z-po­ males and 4 females, between 25 and 45 years old) and stabilized with tential for fresh GBN dispersions (100 µg/mL in 0.05% BSA) were 2 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 Fig. 1. Characterization of GO, GNPs, and GNRs. (A) TEM representative images of the selected GBN dispersed in water. (B) Size distribution histogram from TEM images of GNRs. (C) GBN average size and charge in distilled water by TEM and DLS. Data are represented as mean ± SD (n = 3). (D and D′ ) Metrological char­ acterization of GO and GNPs by Profiler 7.0. The analysis of a representative GO and GNPs particle is shown in D′ : white arrows indicate the scanner path over the selected GO or GNPs particle and the diagram of the measured profile for GO or GNPs particles is shown. The mean step height measured for GO and GNPs is indicated in the graph D. Data are represented as mean ± SEM (n = 20). determined (Fig. 1C). Characterization data in the Z-sizer show sizes of package are depicted in Fig. 2, showing the 105 cytokines measured 244.9 ± 7.4 nm and 243 ± 1.4 nm for GO and GNPs, respectively. The after the different GBN treatments, that showed no toxicity (cell viability hydrodynamic radius analysis for GNRs is not a reliable indicator to for 50 µg/mL GO, GNPs and GNRs treatments were 99.1 ± 1.6, conclude the size of these nanoparticles since as showing the z-potential 99.4 ± 1.2 and 96.7 ± 4.1, respectively). The Y-axis indicates the P value (− 1.64 ± 2.56) they present an important level of aggregation/ values, while the X-axis specifies the fold change with respect to the agglomeration. This value should tend to be around 30 or − 30 mV control. The dashed line represents the limit between the significant when there is no aggregation, but in the case of GNRs, it is very close to cytokines (red), which are those with a P value less than 0.05, and those 0, which indicates aggregation. Furthermore, the obtained poly- non-significant. As shown, GO exposure shows 31 cytokines signifi­ dispersion index (PdI) indicates that dispersions are homogenous since cantly deregulated, most of them overexpressed-only PDGF-AA and all three values tend to zero (0.354 ± 0.052 for GO, 0.3316 ± 0.024 for PDGF-BB (platelet-derived growth factors) are under-expressed-, the GNPs, and 0.3998 ± 0.020 for GNRs). In order to determine the thick­ most expressed being IL-6, CXCL1, CCL20, TNFA, and CCL3. Regarding ness of GO and GNPs, we analyzed the step height using the Profiler 7.0 GNPs exposure, the figure indicates 29 cytokines significantly deregu­ P15. A mean thickness of 89.42 ± 8.30 nm and 220.26 ± 33.68 nm was lated. Interestingly, those overexpressed are shared with GO, and those determined for GO and GNPs, respectively (Fig. 1D). These results agree under-expressed are PF4 and PDGF-BB. Finally, GNRs exposure shows with the single- and multi-layer conformation of GO and GNPs nano­ 33 significant cytokines, also sharing with the other two GBN those that particles and correlate with the TEM observations. Representative are overexpressed, except IL-6. Moreover, the under-expressed cytokines measurements for GO and GNPs particles are shown in Fig. 1D′ . Data are are ANG, APOA1, and RBP4. Supplementary Fig. 1 depicts the 51 cy­ represented as mean ± SEM. tokines that were significantly deregulated after LPS exposure (positive control). 3.2. Cytokines expression changes Fig. 3 shows all those cytokines showing significant changes of expression, highlighting the existing interactions among them. The color Volcano plots obtained by the R version 3.6.3 (2020-02-29) ggplot2 of the nodules represents the functions in which they are involved, 3 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 Fig. 2. Cytokine expression analysis after GBN exposure. Volcano Plot of secreted cytokines after the exposure to 50 µg/mL of GO (A), GNPs (B), GNRs (C). Each volcano plot is divided into two areas, the area above the dashed line represents the significant cytokines (in red) and the lower area depict the non-significant cytokines (in black). X axis depicts the log2 (Fold Change) and Y axis represents the –log10 (P value). Significance was calcu­ lated comparing cytokines expression from exposed samples to non-exposed ones by t-test (in red P-val­ ue < 0.05). The experiment was done with blood from 8 different donors. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 4 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 Fig. 3. STRING cytokine analysis of significantly dysregulated cytokines after GBN exposure Node colors represent functions in which they are implicated, and the networks express the different cytokines interactions. String Database Version 10.0 was used to generate the functional cytokines association networks. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) grouped in the inflammatory and the immune response, as well as cell Among them, Fig. 4B represents those cytokines which participate in the migration, and proliferation. Some cytokines are category-specific, for inflammatory response, being classified as pro-inflammatory cytokines example, TGFA, PDGFA or PF4 only participate in cell proliferation, and anti-inflammatory cytokines; and Fig. 4C highlights the most RETN only participates in the immune response, and APOA1 only par­ overexpressed cytokines after GO, GNPs, and GNRs exposures, showing ticipates in cell migration. Instead, other cytokines have a role in all CCL3 the highest increase. Instead, Fig. 5 displays the specific cytokines shown categories, like ENA-78, CCL2, IL1B, IL6 or TNF. It must be noted founded dysregulated after each GBN exposure. IL-34, LIF, PDGF-AA, that although the expression of ST2, TFF3, IL-31, RLN2, IL-16, TFR, MIF, ANG-1, and IL-31 are produced only by GO exposure whereas PF4, TGFA, and DBP is statistically significant, they do not appear in the ST2, TFF3, and IL-10 are specifically produced by GNPs exposure. STRING because they do not present interactions with other cytokines. Finally, 8 cytokines were found dysregulated exclusively after GNRs The cytokine expression patterns after the three GBN exposures were exposure. Among them, MMP-9, MIF, S EPINE1, MIG, and VCAM-1 depicted at the Venn diagram (Fig. 4A). Some cytokines are exposure- stand overexpressed compared to the non-exposed samples, while specific while overlapping areas show the commonly expressed ones. ANG, APOA1, and CD31 are under-expressed. 5 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 Fig. 4. (A) Venn Diagram shows the number of cytokines significantly dysregulated after treatment with each GBN. The overlapping area indicates the number of cytokines commonly dysregulated between the three GBN. (B) Analysis of selected cytokines with important effects on the inflammatory response. Cytokines are classified as pro-inflammatory or anti-inflammatory cytokines. (C) Highly expressed cytokines in all GBN exposures. Those cytokines overexpressed more than 4-folds over the control were considered. Data represent the folds of overexpression of the cytokines expressed in 8 different GBN-exposed samples compared to the non- exposed ones (controls). The dashed lines (B and C) represent the baseline expression of controls. Fig. 5. Analysis of the cytokines over/under-expressed exclusively after the exposure to one of the GBNs analyzed. Controls are set as 100 (dashed lines represent the baseline expression of controls). *P < 0.05. **P < 0.01 (t-Student). 3.3. The soft-agar assay centrifugation. The results did not show any significant difference (P value > 0.05) in the number of colonies between the non-exposed The indirect soft-agar assay was performed to observe the functional samples and the GBN-exposed ones. Thus, we are confident that the effect of the secretome on cell growth ability. To measure this effect, observed effects are mainly due to the components of the secretome and HeLa cells (as a cell model) were exposed to serum (conditioned media) not to the effect of the residual NPs that can be there. obtained from the blood previously treated with GBN. Fig. 6A shows representative images of the HeLa colonies growing under the different 4. Discussion GBN exposures, as well as the negative and positive (LPS) controls. It can be appreciated an inhibitory effect on cells growing under the different Due to the increasing use of GBN in biomedical applications, its treatments, but mainly with GO. Likewise, Fig. 6B represents the number potential impact on immune cells is a fundamental area requiring of colonies growing with each GBN treatment. As observed in Fig. 6A, a further responses. As previously indicated, applications like imaging or high level of inhibition was observed with GO and GNPs and, although gene delivery make GBN to be in contact with the human immune sys­ no in a marked way, GNRs were also able to inhibit cell growth. When tem since the intravenous injection is one of the main routes of GBN the diameter of the colonies was evaluated, no significant differences for administration. Several studies are proposing GO as a magnetic reso­ any GBN were detected except for GNPs where a low but significant nance imaging (MRI) contrast agent (Campbell et al., 2019), and (P = 0.03) effect was observed. The diameter of the colonies induced by GO-Fe3O4 nanoparticles have been proposed for optical cancer detection LPS (positive control) was statistically significant regarding the negative (Gonzalez-Rodriguez et al., 2019). In addition, to GO, other GBNs have control (Supplementary Fig. 2). As a summary, our results show an also been proposed for such purposes. For instance, GNPs intercalated inhibitory effect on cell growth due to some of the components present with manganese (Mn2+) ions and functionalized with dextran in the secretome of whole blood cells after GBN exposure. It should be (GNP-Dex) were proposed as a contrasting agent for clinical MRI indicated that the potential remains of GBN in serum was not strictly (Kanakia et al., 2013). Furthermore, GNRs dispersed by carbox­ checked but, instead, we made an indirect soft agar with HeLa cells yphenylated substituent, and conjugated to aquated Gd3+ ions, have exposed to DMEM treated with GBN (conditioned media). Additionally, also been proposed as a high-performance contrast agent for applica­ this conditioned media was tested in HeLa cells without and after tions in MRI (Gizzatov et al., 2014). In the present study, we have 6 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 Fig. 6. Indirect soft-agar results of HeLa cells exposed to the serum extracted from blood samples previously treated with GO, GNPs, and GNRs. (A) Representative images corresponding to a section of the 35 mm soft-agar assay plates from each graphene. (B) Quantification of total colonies. Data represent mean ± SD-with controls set to 100%- of triplicates from 8 blood donors. ***P < 0.001 (t-Student). proposed an ex vivo experimental model to unravel potential 105 analyzed cytokines, we have found that 19% are commonly blood-nanomaterials interactions. It should be stated that in such deregulated in the three types of GBN. Among these proteins, most of context, our system could be a simple and sensitive method able to them have pro-inflammatory effects, such as IL1α, IL1β, CCL3, TGFα(A), clarify the impact of GBN in the blood system by determining molecular MCP3, TNFα, IL8,MCP1, MCSF, ITAC, ENA-78, CCL20, CXCL1, while changes at the secretome level. This could be a quick and easy way to others have anti-inflammatory effects, such as IL-1ra, TARC, PTX3. We determine biocompatibility between the different candidates with po­ would like to highlight the overexpression of TGFα and CXCL1, and, tential biomedical applicability. We have found that GBN exposure in­ TNFα, IL-1 (α, or β). In the first case, TGFα is a protein described as the duces an immunological response, triggering important changes in the most prevalent and abundant in transformed cells and tumors. Together cytokine expression profile. This suggests that different sizes, forms, and with CXCL1, they appear overexpressed in non-small cell lung cancer physicochemical properties of GBN could impact human whole blood (Liu et al., 2017). On the other hand, different cytokines have been cells at the molecular level. related to inflammatory processes. TNFα and IL-1 (α, or β) are among the Studies evaluating the effects of GBN on the immune system using in most studied inflammatory cytokines in the tumor microenvironment vivo and in vitro approaches have reported controversial results (Dudek (Balkwill and Mantovani, 2012). TNFα has a key role in inflammation et al., 2016). For example, in vitro studies showed that GO could trigger and its overexpression has been related to many inflammatory diseases, IL-1β and TGF-β1 production in myeloid (THP-1) and epithelial like rheumatoid arthritis, ankylosing spondylitis, inflammatory bowel (BEAS-2B) cells (Wang et al., 2015). Similarly, a caspase-dependent disease and psoriasis (Kany et al., 2019). In addition, it is worth to IL-1β expression in primary human monocyte-derived macrophages highlight that CCL3 is the most overexpressed cytokine of our study. was observed after GO exposure (Mukherjee et al., 2018). Besides, CCL3 is an important inflammatory chemokine produced by a variety of although in vitro exposure to GO significantly increased the secretion of cells, including macrophages. It participates as a potent activator of both TNF-α by RAW-264.7 cells, no proinflammatory cytokine secretion was innate and adaptive responses. Specifically, CCL3 plays an important observed on primary immune cells (Feito et al., 2014). Furthermore, a role in the development, regulation and recruitment of leukocytes. recent study using an in vitro model of intestinal barrier showed a higher Previous studies of our group corroborate CCL3 overexpression after the uptake of GNPs than GO, and both graphene compounds inducing low exposure to nanopolysytrene nanoparticles (Ballesteros et al., 2020). levels of anti-inflammatory cytokines, suggesting a weak Similarly, other nanomaterials, like cooper or silver nanoparticles, anti-inflammatory response (Domenech et al., 2020). induce an inflammatory response where CCL3 levels are significantly Regarding in vivo studies, GO treatment significantly protected mice increased (Tang et al., 2019; Seiffert et al., 2016). from α-GalCer-induced lethality proposing its use as an adjuvant in Alternatively, both GO and GNPs induce secretion of IL-6, G-CSF, and immunotherapy (Lee et al., 2018). In the same way, experiments ex vivo PDGF-BB. In the case of IL-6, it is a very well-studied pro-inflammatory are controversial. Although it was reported that GO exposure caused cytokine. It has been considered as a key in tumorigenesis and metastasis significant changes in the lipid composition of the cell membrane of processes since it is up-regulated in many types of cancers, including primary human neutrophils freshly isolated from healthy human blood skin, breast, lung, esophageal, liver, pancreatic, gastric, colorectal, gy­ donors (Mukherjee et al., 2018); small and large GO sheets were necological, prostate, kidney, bladder and hematological cancers as well non-cytotoxic on peripheral blood mononuclear cells from healthy do­ as melanoma (Uciechowski and Dempke, 2020). Interestingly, and nors (Orecchioni et al., 2016). corroborating what we can observe in our study, IL-6, TNF-α, and G-CSF By using our ex vivo exposure model, and using whole blood samples, are a part of the inflammatory response triggered by GBN exposure. As we have detected that non-cytotoxic GBN exposure has an important in our study, an increase in the expression of IL-6, TNF-α, and G-CSF was impact on the whole blood secretome. It is important to remember that observed in tissues surrounding an implant using graphene materials we have not follow the standard procedure of selecting candidate cy­ (Guazzo et al., 2018). tokines; our approach use a wide panel of cytokines what, at priory, has Since the majority of the altered cytokines observed in our study play a higher chance of pick-up potential changes among the used set. Of the a role in the immune response and are secreted by monocytes/ 7 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 macrophages (da Cunha et al., 2019), we can consider these cells as the found important differences in the cytokine’s response due to the dose of most implicated in the response to GBN exposure. This would agree with nanoparticles. Furthermore, Orecchioni et al. (2017) demonstrated a the results showing that GO induced the production of TNF-α in mac­ different cytokine expression pattern depending on the immune cell rophages at the same levels that those induced by LPS exposure (Zhou population and the type of graphene oxide analyzed. Thus, rather than et al., 2012). These results agree with those showing that pristine gra­ the complexity, future studies must focus in analyzing different sizes and phene and GO causes an increased production cytokines IL-1α, IL-6, concentrations of graphene nanomaterials, as well as evaluating their IL-10, and TNFα by activated macrophages (Chen et al., 2012). effects on WBCs. If we focus on those cytokines exclusively altered in response to one As a conclusion, from our ex vivo model we have demonstrated that specific GBN, most of them are related to diseases associated with the selected GBN can trigger an immunological response that must be macrophage stimulation and inflammatory effects. For instance, if we seriously evaluated when directly intravenously applied to humans. focus on cytokines that appeared deregulated in GO exposure, we found Besides, we propose that the ex vivo exposure used in this study can be that the overexpression of angiopoietin-1 (ANG-1) correlated to pa­ useful to unravel potential health risks associated with GBN exposures. rameters of inflammation and bone destruction in inflammatory arthritis (Clavel et al., 2007); and IL-34 and IL-31 presenting an aberrant over­ CRediT authorship contribution statement expression in Inflammatory Bowel Diseases (Zwicker et al., 2015). Similar occurs when we focus on the cytokines altered in GNPs exposure, RM and AH planned the experiments. SB and JD carried out the we see that IL-10 stimulates the increased production of experimental part. SB and AV analyzed the data, carried out the sta­ pro-inflammatory cytokines in atherosclerotic lesions (Mannino et al., tistical analysis, and prepared tables/figures. SB, RM, and AH wrote the 2015); while ST2 overexpression has been correlated with Inflammatory final manuscript. Bowel Disease (Boga et al., 2016). In the same line, if we observe those cytokines altered in GNRs, MIF is related with TNFα and IL6 key cyto­ Declaration of Competing Interest kines in cancer-related inflammation (Kulbe et al., 2012); VCAM-1 increased expression (vascular cell adhesion molecule-1) is closely The authors declare they have no actual or potential competing associated with the progression of various immunological disorders, financial interests. including rheumatoid arthritis, asthma, transplant rejection, and cancer (Kong et al., 2018). Analyzing the different proteins deregulated with Acknowledgments the different types of GBN we cannot infer that one type of GBN is less biocompatible than another; we can only assume an equivalent risk for We wish to thank Mr. C. Valiente for his technical assistance. This all of them, according to our results. Maybe the exception is GNRs work was partially supported by the Spanish Ministry of Education and exposure that triggers the overexpression of MMP9, a protein involved Science [BFU2016-76831-R]. Sandra Ballesteros holds a fellowship from in the onset of cancer with important implications in the development of Universitat Autònoma de Barcelona [PIF-UAB]. angiogenesis and inflammation, and with an indirect impact on the evolution of lung cancer (Rivas-Fuentes et al., 2015). Therefore, special Appendix A. Supporting information attention should be paid to the effects induced by GNRs in future studies. In the same way, the comparison of ex vivo and in vivo data would help to Supplementary data associated with this article can be found in the confirm the goodness of the ex vivo approaches. online version at doi:10.1016/j.jhazmat.2021.125471. Besides the ex vivo system, another important feature of this study is the evaluation of the anchorage-independent growth to unravel the References functional impact of the secretome. The secretome combines all the cytokine expression changes and the consequent changes of cytokines’ Annangi, B., Bach, J., Vales, G., Rubio, L., Marcos, R., Hernández, A., 2015. Long-term interactions. Thus, any change in the anchorage-independent growth is exposures to low doses of cobalt nanoparticles induce cell transformation enhanced by oxidative damage. Nanotoxicology 9, 138–147. https://doi.org/10.3109/ reflecting the potential effects of the altered secretome on neighboring 17435390.2014.900582. blood cells. Indeed, we are showing here that changes in the secretome Balkwill, F.R., Mantovani, A., 2012. Cancer-related inflammation: common themes and produced by GBN can have an impact on the growth behavior of a model therapeutic opportunities. Semin. Cancer Biol. 22, 33–40. https://doi.org/10.1016/ cell line. Although the transforming potential of carbon nanotubes j.semcancer.2011.12.005. Ballesteros, S., Domenech, J., Barguilla, I., Cortés, C., Marcos, R., Hernández, A., 2020. compounds have been reported after chronic exposures (He et al., 2016; Genotoxic and immunomodulatory effects in human white blood cells after ex vivo Vales et al., 2016) modulating the ability of anchorage-independent exposure to polystyrene nanoplastics. Environ. Sci. Nano. https://doi.org/10.1039/ growth, as well as other hallmark biomarkers of cell transformation, D0EN00748J. Boga, S., Alkim, H., Koksal, A.R., Ozagari, A.A., Bayram, M., Tekin Neijmann, S., no data using graphene-based materials have been found in the open Alkim, C., 2016. Serum ST2 in inflammatory bowel disease: a potential biomarker literature. In this context, our results, showing that blood cells exposed for disease activity. J. Investig. Med. 64, 1016–1024. https://doi.org/10.1136/jim- to GBN are able to secrete compounds, look interesting and would need 2016-000062. Boraschi, D., Swartzwelter, B.J., Italiani, P., 2018. Interaction of engineered further studies to be demonstrated. nanomaterials with the immune system: health-related safety and possible benefits. One of the limitations of the study is that we have not assess the role Curr. Opin. Toxicol. 10, 74–83. https://doi.org/10.1016/j.cotox.2018.02.002. of concentration, or the effects associated to different sizes. Some studies Campbell, E., Hasan, M.T., Pho, C., Callaghan, K., Akkaraju, G.R., Naumov, A.V., 2019. Graphene oxide as a multifunctional platform for intracellular delivery, imaging, and like the one of Rodrigues et al. (2020), pointed out the importance of cancer sensing. Sci. Rep. 9, 416. https://doi.org/10.1038/s41598-018-36617-4. considering the size impact of graphene oxide nanomaterial (Rodrigues Carpentier, G., 2009. 〈http://image.bio.methods.free.fr/ImageJ/?Developpement-d-out et al., 2020). In this sense, we consider that the study of these effects in ils-d-analyse-et-de-traitement-d-images-avec-le-logiciel〉. Chen, G.-Y., Yang, H.-J., Lu, C.-H., Chao, Y.-C., Hwang, S.-M., Chen, C.-L., Lo, K.W., further works will provide relevant information. Additionally, it should Sung, L.Y., Luo, W.Y., Tuan, H.Y., Hu, Y.-C., 2012. Simultaneous induction of be mentioned that it is important to evaluate graphene effects at the autophagy and toll-like receptor signaling pathways by graphene oxide. Biomaterials level of WBCs subpopulations, since studies of our group have shown 33, 6559–6569. https://doi.org/10.1016/j.biomaterials.2012.05.064. interesting results in that sense. Thus, the studies of Rubio et al. (2020) Clavel, G., Bessis, N., Lemeiter, D., Fardellone, P., Mejjad, O., Ménard, J.F., Pouplin, S., Boumier, P., Vittecoq, O., Le Loët, X., Boissier, M.C., 2007. Angiogenesis markers and Ballesteros et al. (2020) demonstrated that nanopolystyrene nano­ (VEGF, soluble receptor of VEGF and angiopoietin-1) in very early arthritis and their particles triggered differential biological effects both in different human association with inflammation and joint destruction. J. Clin. Immunol. 124, hematopoietic cell lines and in blood cells analyzed in an ex vivo model 158–164. https://doi.org/10.1016/j.clim.2007.04.014. Cui, J., De Rose, R., Alt, K., Alcantara, S., Paterson, B.M., Liang, K., Hu, M., such as the used in the present study. Likewise, Ballesteros et al. (2020) Richardson, J.J., Yan, Y., Jeffery, C.M., Price, R.I., Peter, K., Hagemeyer, C.E., evidenced that the role of the concentrations must be considered as they Donnelly, P.S., Kent, S.J., Caruso, F., 2015. Engineering poly(ethylene glycol) 8 S. Ballesteros et al. Journal of Hazardous Materials 414 (2021) 125471 particles for improved biodistribution. ACS Nano 9, 1571–1580. https://doi.org/ Mukherjee, S.P., Kostarelos, K., Fadeel, B., 2018. Cytokine profiling of primary human 10.1021/nn5061578. macrophages exposed to endotoxin-free graphene oxide: size-independent nlrp3 da Cunha, B.R., Domingos, C., Stefanini, A.C.B., Henrique, T., Polachini, G.M., Castelo- inflammasome activation. Adv. Healthc. Mater. 7, 1700815 https://doi.org/ Branco, P., Tajara, E.H., 2019. Cellular interactions in the tumor microenvironment: 10.1002/adhm.201700815. the role of secretome. J. Cancer 10, 4574–4587. https://doi.org/10.7150/jca.21780. Nanogenotox, 2011. 〈http://www.nanogenotox.eu/files/PDF/Deliverables/nanogenotox Domenech, J., Hernández, A., Marcos, R., Cortés, C., 2020. Interactions of graphene %20deliverable%203wp4%20dispersion%20protocol.pdf〉. oxide and graphene nanoplatelets with the in vitro Caco-2/HT29 model of intestinal Orecchioni, M., Jasim, D.A., Pescatori, M., Manetti, R., Fozza, C., Sgarrella, F., barrier. Sci. Rep. 10, 2793. https://doi.org/10.1038/s41598-020-59755-0. Bedognetti, D., Bianco, A., Kostarelos, K., Delogu, L.G., 2016. Molecular and Dudek, I., Skoda, M., Jarosz, A., Szukiewicz, D., 2016. The molecular influence of genomic impact of large and small lateral dimension graphene oxide sheets on graphene and graphene oxide on the immune system under in vitro and in vivo human immune cells from healthy donors. Adv. Healthc. Mater. 5, 276–287. https:// conditions. Arch. Immunol. Ther. Exp. 64, 195–215. https://doi.org/10.1007/ doi.org/10.1002/adhm.201500606. s00005-015-0369-3. Orecchioni, M., Bedognetti, D., Newman, L., Fuoco, C., Spada, F., Hendrickx, W., Fadeel, B., 2019. Hide and seek: nanomaterial interactions with the immune system. Marincola, F.M., Sgarrella, F., Rodrigues, A.F., Ménard-Moyon, C., Cesareni, G., Front. Immunol. 10, 133. https://doi.org/10.3389/fimmu.2019.00133. Kostarelos, K., Bianco, A., Delogu, L., 2017. Single-cell mass cytometry and Feito, M.J., Vila, M., Matesanz, M.C., Linares, J., Gonçalves, G., Marques, P.A., Vallet- transcriptome profiling reveal the impact of graphene on human immune cells. Nat. Regí, M., Rojo, J.M., Portolés, M.T., 2014. In vitro evaluation of graphene oxide Commun. 8, 1109. https://doi.org/10.1038/s41467-017-01015-3. nanosheets on immune function. J. Colloid Interface Sci. 432, 221–228. https://doi. Park, E.J., Kim, S.N., Kang, M.S., Lee, B.S., Yoon, C., Jeong, U., Kim, Y., Lee, G.H., org/10.1016/j.jcis.2014.07.004. Kim, D.W., Kim, J.S., 2016. A higher aspect ratio enhanced bioaccumulation and Gizzatov, A., Keshishian, V., Guven, A., Dimiev, A.M., Qu, F., Muthupillai, R., altered immune responses due to intravenously-injected aluminum oxide Decuzzi, P., Bryant, R.G., Tour, J.M., Wilson, L.J., 2014. Enhanced MRI relaxivity of nanoparticles. J. Immunotoxicol. 13, 439–448. https://doi.org/10.3109/ aquated Gd3+ ions by carboxyphenylated water-dispersed graphene nanoribbons. 1547691X.2015.1122116. Nanoscale 6, 3059–3063. https://doi.org/10.1039/c3nr06026h. Patel, S., Ngounou Wetie, A.G., Darie, C.C., Clarkson, B.D., 2014. Cancer secretomes and Gonzalez-Rodriguez, R., Campbell, E., Naumov, A., 2019. Multifunctional graphene their place in supplementing other hallmarks of cancer. Adv. Exp. Med. Biol. 806, oxide/iron oxide nanoparticles for magnetic targeted drug delivery dual magnetic 409–442. https://doi.org/10.1007/978-3-319-06068-2_20. resonance/fluorescence imaging and cancer sensing. PLoS One 14, e0217072. Ramos, A.P., Cruz, M.A.E., Tovani, C.B., Ciancaglini, P., 2017. Biomedical applications of https://doi.org/10.1371/journal.pone.0217072. nanotechnology. Biophys. Rev. 9, 79–89. https://doi.org/10.1007/s12551-016- Guazzo, R., Gardin, C., Bellin, G., Sbricoli, L., Ferroni, L., Ludovichetti, F.S., Piattelli, A., 0246-2. Antoniac, I., Bressan, E., Zavan, B., 2018. Graphene-based nanomaterials for tissue Rivas-Fuentes, S., Salgado-Aguayo, A., Pertuz Belloso, S., Gorocica Rosete, P., Alvarado- engineering in the dental field. Nanomaterials 8, 349. https://doi.org/10.3390/ Vásquez, N., Aquino-Jarquin, G., 2015. Role of chemokines in non-small cell lung nano8050349. cancer: angiogenesis and inflammation. J. Cancer 6, 938–952. https://doi.org/ Han, S., Sun, J., He, S., Tang, M., Chai, R., 2019. The application of graphene-based 10.7150/jca.12286. biomaterials in biomedicine. Am. J. Transl. Res. 11, 3246–3260. Rodrigues, A.F., Newman, L., Jasim, D., Mukherjee, S.P., Wang, J., Vacchi, I.A., Bussy, C., He, X., Despeaux, E., Stueckle, T.A., Chi, A., Castranova, V., Dinu, C.Z., Wang, L., 2020. Size-dependent pulmonary impact of thin graphene oxide sheets in mice: Rojanasakul, Y., 2016. Role of mesothelin in carbon nanotube-induced carcinogenic toward safe-by-design. Adv. Sci. 7, 1903200 https://doi.org/10.1002/ transformation of human bronchial epithelial cells. Am. J. Physiol. Lung Cell. Mol. advs.201903200. Physiol. 311, L538–L549. https://doi.org/10.1152/ajplung.00139.2016. Rubio, L., Barguilla, I., Domenech, J., Marcos, R., Hernández, A., 2020. Biological effects, Iavicoli, I., Fontana, L., Leso, V., Corbi, M., Marinaccio, A., Leopold, K., Schindl, R., including oxidative stress and genotoxic damage, of polystyrene nanoparticles in Lucchetti, D., Calapà, F., Sgambato, A., 2018. Subchronic exposure to palladium different human hematopoietic cell lines. J. Hazard. Mater. 398, 122900 https://doi. nanoparticles affects serum levels of cytokines in female Wistar rats. Hum. Exp. org/10.1016/j.jhazmat.2020.122900. Toxicol. 37, 309–320. https://doi.org/10.1177/0960327117702952. Seiffert, J., Buckley, A., Leo, B., Martin, N.G., Zhu, J., Dai, R., Hussain, F., Guo, C., Kanakia, S., Toussaint, J.D., Chowdhury, S.M., Lalwani, G., Tembulkar, T., Button, T., Warren, J., Hodgson, A., Gong, J., Ryan, M.P., Zhang, J.J., Porter, A., Tetley, T.D., Shroyer, K.R., Moore, W., Sitharaman, B., 2013. Physicochemical characterization of Gow, A., Smith, R., Chung, K.F., 2016. Pulmonary effects of inhalation of spark- a novel graphene-based magnetic resonance imaging contrast agent. Int. J. generated silver nanoparticles in Brown-Norway and Sprague–Dawley rats. Respir. Nanomed. 8, 2821–2833. https://doi.org/10.2147/IJN.S47062. Res. 17, 85. https://doi.org/10.1186/s12931-016-0407-7. Kany, S., Vollrath, J.T., Relja, B., 2019. Cytokines in inflammatory disease. Int. J. Mol. Song, S., Shen, H., Wang, Y., Chu, X., Xie, J., Zhou, N., Shen, J., 2020. Biomedical Sci. 20, 6008. https://doi.org/10.3390/ijms20236008. application of graphene: from drug delivery, tumor therapy, to theranostics. Colloids Kong, D.H., Kim, Y.K., Kim, M.R., Jang, J.H., Lee, S., 2018. Emerging roles of vascular Surf. B Biointerfaces 185, 110596. https://doi.org/10.1016/j.colsurfb.2019.110596. cell adhesion molecule-1 (VCAM-1) in immunological disorders and cancer. Int. J. Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., Huerta-Cepas, J.C., Mol. Sci. 19, 1057. https://doi.org/10.3390/ijms19041057. 2015. STRING v10: protein-protein interaction networks, integrated over the tree of Kulbe, H., Chakravarty, P., Leinster, D.A., Charles, K.A., Kwong, J., Thompson, R.G., life. Nucleic Acids Res. 43, D447–D452. https://doi.org/10.1093/nar/gku1003. Coward, J.I., Schioppa, T., Robinson, S.C., Gallagher, W.M., Galletta, L., Australian Tang, H., Xu, M., Luo, J., Zhao, L., Ye, G., Shi, F., Ménard-Moyon, C., Bianco, A., Ovarian Cancer Study Group, Salako, M.A., Smyth, J.F., Hagemann, T., Brennan, D. Fadeel, B., Kostarelos, K., Li, Y., 2019. Liver toxicity assessments in rats following J., Bowtell, D.D., Balkwill, F.R., 2012. A dynamic inflammatory cytokine network in sub-chronic oral exposure to copper nanoparticles. Environ. Sci. Eur. 31, 30. https:// the human ovarian cancer microenvironment. Cancer Res. 72, 66–75. https://doi. doi.org/10.1186/s12302-019-0214-0. org/10.1158/0008-5472.CAN-11-2178. Uciechowski, P., Dempke, W.C.M., 2020. Interleukin-6: a masterplayer in the cytokine Lee, S.W., Park, H.J., Van Kaer, L., Hong, S., Hong, S., 2018. Graphene oxide polarizes network. Oncology 98, 131–137. https://doi.org/10.1159/000505099. iNKT cells for production of TGFβ and attenuates inflammation in an iNKT cell- Vales, G., Rubio, L., Marcos, R., 2016. Genotoxic and cell-transformation effects of multi- mediated sepsis model. Sci. Rep. 8, 10081. https://doi.org/10.1038/s41598-018- walled carbon nanotubes (MWCNT) following in vitro sub-chronic exposures. 28396-9. J. Hazard. Mater. 306, 193–202. https://doi.org/10.1016/j.jhazmat.2015.12.021. Lin, J., Chen, X., Huang, P., 2016. Graphene-based nanomaterials for bioimaging. Adv. Vila, L., Marcos, R., Hernández, A., 2017. Long-term effects of silver nanoparticles in Drug Deliv. Rev. 105, 242–254. https://doi.org/10.1016/j.addr.2016.05.013. Caco-2 cells. Nanotoxicology 11, 771–780. https://doi.org/10.1080/ Liu, X., Chen, L., Tian, X.D., Zhang, T., 2017. MiR-137 and its target TGFA modulate cell 17435390.2017.1355997. growth and tumorigenesis of non-small cell lung cancer. Eur. Rev. Med. Pharmacol. Wang, X., Duch, M.C., Mansukhani, N., Ji, Z., Liao, Y.P., Wang, M., Zhang, H., Sun, B., Sci. 21, 511–517. Chang, C.H., Li, R., Lin, S., Meng, H., Xia, T., Hersam, M.C., Nel, A.E., 2015. Use of a Loh, K.P., Ho, D., Chiu, G.N.C., Leong, D.T., Pastorin, G., Chow, E.K., 2018. Clinical pro-fibrogenic mechanism-based predictive toxicological approach for tiered testing applications of carbon nanomaterials in diagnostics and therapy. Adv. Mater. 30, and decision analysis of carbonaceous nanomaterials. ACS Nano 9, 3032–3043. 1802368 https://doi.org/10.1002/adma.201802368. https://doi.org/10.1021/nn507243w. Madannejad, R., Shoaie, N., Jahanpeyma, F., Darvishi, M.H., Azimzadeh, M., Javadi, H., Zhou, H., Zhao, K., Li, W., Yang, N., Liu, Y., Chen, C., Wei, T., 2012. The interactions 2019. Toxicity of carbon-based nanomaterials: reviewing recent reports in medical between pristine graphene and macrophages and the production of cytokines/ and biological systems. Chem. Biol. Interact. 307, 206–-222. https://doi.org/ chemokines via TLR- and NF-κB-related signaling pathways. Biomaterials 33, 10.1016/j.cbi.2019.04.036. 6933–6942. https://doi.org/10.1016/j.biomaterials.2012.06.064. Mann, S.K., Dufour, A., Glass, J.J., De Rose, R., Kent, S.J., Such, G.K., Johnston, A.P.R., Zwicker, S., Martinez, G.L., Bosma, M., Gerling, M., Clark, R., Majster, M., Söderman, J., 2016. Tuning the properties of pH responsive nanoparticles to control cellular Almer, S., Boström, E.A., 2015. Interleukin 34: a new modulator of human and interactions: in vitro and ex vivo. Polym. Chem. 7, 6015–6024. https://doi.org/ experimental inflammatory bowel disease. Clin. Sci. 129, 281–290. https://doi.org/ 10.1039/c6py01332e. 10.1042/CS20150176. Mannino, M.H., Zhu, Z., Xiao, H., Bai, Q., Wakefield, M.R., Fang, Y., 2015. The paradoxical role of IL-10 in immunity and cancer. Cancer Lett. 367, 103–107. https://doi.org/10.1016/j.canlet.2015.07.009. 9