RESEARCH ARTICLE www.advancedscience.com Innate but Not Adaptive Immunity Regulates Lung Recovery from Chronic Exposure to Graphene Oxide Nanosheets Thomas Loret, Luis Augusto Visani de Luna, Alexander Fordham, Atta Arshad, Katharine Barr, Neus Lozano, Kostas Kostarelos,* and Cyrill Bussy* composite materials or coating. In the Graphene has drawn a lot of interest in the material community due to unique biomedical field, graphene has also been physicochemical properties. Owing to a high surface area to volume ratio and proposed for an array of applications. This free oxygen groups, the oxidized derivative, graphene oxide (GO) has includes the production of scaffolds for tis- promising potential as a drug delivery system. Here, the lung tolerability of sue regeneration[1] or the design of next- generation implants,[2] as well as the devel- two distinct GO varying in lateral dimensions is investigated, to reveal the opment of photo-thermal therapy, vaccine most suitable candidate platform for pulmonary drug delivery. Following platforms, or drug delivery systems.[3–5] In repeated chronic pulmonary exposure of mice to GO sheet suspensions, the this latter application, graphene oxide (GO), innate and adaptive immune responses are studied. An acute and transient the oxidized form of graphene, has stood influx of neutrophils and eosinophils in the alveolar space, together with the out due to two important features, namely a high surface area for high drug payload and replacement of alveolar macrophages by interstitial ones and a significant a surface rich in oxygen groups. Having sur- activation toward anti-inflammatory subsets, are found for both GO materials. face functional head-groups increases the Micrometric GO give rise to persistent multinucleated macrophages and possibility of further chemical functional- granulomas. However, neither adaptive immune response nor lung tissue ization, which allows the design of nanocar- remodeling are induced after exposure to micrometric GO. Concurrently, riers with greater specificity.[6] It also pro- milder effects and faster tissue recovery, both associated to a faster clearance vides superior colloidal stability compared to other graphene nanomaterials. Regard- from the respiratory tract, are found for nanometric GO, suggesting a greater less of the formulation, good colloidal sta- lung tolerability. Taken together, these results highlight the importance of bility is essential for the administration of dimensions in the design of biocompatible 2D materials for pulmonary drug drug delivery systems. delivery system. Whilst the potential benefits of GO to the biomedical field continue to be re- vealed, questions are concurrently raised regarding its safety profile and long-term 1. Introduction fate in the body. In several studies, limitations to GO biocom- patibility have been identified and associated to the physico- Graphene has been extensively investigated for multiple pur- chemical features of the tested materials. In particular, varia- poses, ranging from electronic devices to water filtration, through tion in the size of GO sheets, ranging from few nanometers to T. Loret, L. A. V. de Luna, A. Fordham, A. Arshad, K. Barr, K. Kostarelos, T. Loret, L. A. V. de Luna, A. Fordham, A. Arshad, K. Barr, K. Kostarelos, C. Bussy C. Bussy Nanomedicine Lab National Graphene Institute Faculty of Biology, Medicine and Health The University of Manchester The University of Manchester Manchester M13 9PL, UK Manchester Academic Health Science Centre T. Loret, L. A. V. de Luna, A. Fordham, A. Arshad, C. Bussy Manchester M13 9PT, UK Lydia Becker Institute of Immunology and Inflammation E-mail: kostas.kostarelos@manchester.ac.uk; Faculty of Biology, Medicine and Health cyrill.bussy@manchester.ac.uk The University of Manchester Manchester Academic Health Science Centre Manchester M13 9PT, UK The ORCID identification number(s) for the author(s) of this article N. Lozano, K. Kostarelos can be found under https://doi.org/10.1002/advs.202104559 Catalan Institute of Nanoscience and Nanotechnology (ICN2) CSIC and The Barcelona Institute of Science and Technology (BIST) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH. Campus UAB, Bellaterra, Barcelona 08193, Spain This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. DOI: 10.1002/advs.202104559 Adv. Sci. 2022, 9, 2104559 2104559 (1 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com micrometers, in thickness from single layer sheets to multi-layer did not induce significant inflammation or any other major platelets, in degree of surface oxidation, and in durability un- negative impact in rat lungs for up to 21 days.[30] Our group has der biological conditions were ascribed to potential safety limita- also previously investigated the pulmonary impact of GO.[31] tions. To address these concerns, comprehensive investigations Following single intranasal instillation in mice, we showed that of the biocompatibility of GO within the intended application,[7,8] micrometric GO sheets induced pulmonary inflammation with alongside the demonstration of its therapeutic efficacy, appears granulomas that did not lead to fibrosis, but persisted for up to critical when developing new GO based drug delivery systems. 90 days, whereas lungs exposed to smaller GO sheets recovered This is particularly important when the intended route of admin- fully within 28 days. These clear size dependent differences istration is the respiratory tract. Inhalation is not only a promis- between two GO types highlighted that materials designed with ing route of administration for nanomedicine drug formulation adequate lateral dimensions favoring better dispersibility could in difficult-to-cure infections,[9–11] but also a major route of unin- present safer profile in general and for intended application tended pulmonary exposure to engineered nanomaterials, which such as drug delivery. However, this study did not address the have raised serious safety concerns in the last 15 years.[12] potential impact of a chronic exposure or the possible activation In this respect, one of the main concerns around GO and other of the adaptive immunity after repeated exposure, although these graphene based materials comes from their proximity to the two features were deemed significant in the evolution toward chemical composition of other carbon nanostructures, namely pathological conditions for CNTs. Generating such data would be carbon nanotubes (CNTs) that have been reported to cause ma- extremely valuable toward the development of GO materials for jor lung issues. After inhalation, some CNTs were found to cause pulmonary drug delivery. persistent inflammation, impairment of immune response,[13,14] The aim of the present study was therefore to assess the extent formation of irreversible fibrotic lesions,[15–17] necrosis,[18] or of lung tolerability of GO when successive pulmonary exposures even pulmonary carcinogenesis.[19] In many of these reports, the were applied. Knowing that a single exposure to nanometric GO extent of chronic immune response as well as the activation of the sheets could be well managed by physiological processes,[32] we adaptive immunity caused by the tested CNTs were shown to be a hypothesized that nanometric GO sheets would still demonstrate hallmark in the progression toward pathological conditions.[14] In good tolerability due to a better clearance pattern and despite addition, it was highlighted that the exact physicochemical char- the repeated exposure regimen. On the contrary, micrometric acteristics of the tested materials (length, diameter, aspect ratio, GO sheets would trigger cellular and molecular pathways due rigidity, crystallinity, or nature of metal catalyst) and their abil- to their longer lung residency confirming their inadequacy as ity to agglomerate[20] and bio-persist in the lungs were strongly candidate platform for pulmonary delivery. To test this hypothe- associated to the extent of their impact.[21–23] Solutions to mit- sis, a thorough immunological evaluation of the impact of these igate these safety limitations have emerged from investigations two types of GO was performed, using two doses of endotoxin- in the chemical engineering and medicinal chemistry fields. Nev- free materials, and for up to 3 month after the last exposure. ertheless, recurrent reports of major pulmonary adverse effects This was completed by a detailed histological analysis and eval- for some CNTs have weakened the promotion of other CNTs for uation of the materials’ clearance patterns using confocal Ra- biomedical application. man microscopy. The findings confirmed the hypothesis: mi- With this in mind, GO sheets may present several benefits crometric GO sheets are not suitable platforms for pulmonary over CNTs and could be a safer option for the pulmonary delivery drug delivery as they induced widespread and long-lasting tis- of therapeutic molecules. Indeed, GO sheets do not possess a sue responses, despite noticeable recovery of the lungs and ab- fibre-shaped structure and can be produced as highly stable sence of adaptive immune response. In contrast, we found that colloidal suspensions that are free of endotoxins and metallic nanometric GO sheets were relatively well tolerated by lung impurities.[24,25] GO sheets have also been reported to either be tissues, which recovered more rapidly owing to a faster clear- eliminated rapidly or biodegraded under various physiological ance pattern involving innate immune cells only. Overall, these conditions.[26] Functionalized forms of GO were even used findings suggested that nanometric GO were eligible materi- successfully for intranasal immunization against influenza als for exploring their potential as pulmonary drug delivery sys- viruses in mice.[11] However, these material-related advantages tems. Nevertheless, we would recommend administering a sub- do not preclude the ability of GO to cause adverse effects in the threshold low dose of nanometric GO to ensure the absence respiratory tract. Duch et al. reported that GO induced severe of adverse effects, guaranteeing the safety of the drug delivery and persistent inflammation in mouse lungs over 21 days. They system. suggested that improved dispersibility to prevent agglomeration and low degree of oxidation were key to minimize the adverse 2. Results and Discussion effects.[27] Consistent with these findings, Li et al. found that GO with higher degree of oxidation induces higher levels of As mentioned above, GO physicochemical characteristics could lung inflammation.[28] In another study, the agglomeration of be advantageous in the prospect of developing drug delivery GO sheets in the lungs was revealed as the main cause of acute systems for pulmonary administration.[10,11] Nevertheless, some lung injury leading to chronic inflammation and pulmonary questions remain unanswered regarding GO’s safety profile. In fibrosis.[8] In contrast, Bengston et al. reported that GO could particular, the long-term impact and fate of these materials in the agglomerate in mouse lungs following oropharyngeal aspira- lungs is not well understood. To address these questions, mice tion and induce acute inflammation with hyperplasia, but not were exposed by oropharyngeal aspiration to 1 or 10 μg of GO fibrosis.[29] However, exposure to GO aerosols, which allow bet- materials once every 2 weeks for 4 weeks, for a total adminis- ter representation of what a real-world inhalation would cause, tered dose of 3 or 30 μg per animal. GO sheet suspensions that Adv. Sci. 2022, 9, 2104559 2104559 (2 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 1. a) Structural and morphological characterization of USGO and LGO. I,V) Height atomic force microscopy (AFM) images; II,VI) height mapping along the dotted lines in (I) and (V) representative height atomic force microscopy (AFM) images, with (II) and (VI) height mapping along the dotted lines present in (I) and (V) cross sections height AFM images; (III) representative scanning electron micrograph of USGO sheets with (IV) corresponding lateral dimension distribution; (VII) representative optical microscope image of LGO sheets with (VIII) corresponding lateral dimension distribution. b) Repeated exposure experimental design. Mice (n = 6) were exposed by repeated oropharyngeal aspiration to ultra-small graphene oxide (USGO), large graphene (LGO) oxide, multi-walled carbon nanotubes (MWCNTs), or control vehicle (sterile water for injection). Animals were exposed once every 2 weeks to 1 or 10 μg of nanomaterials, with a total of three exposures, to achieve a final dose of 3 μg (low dose scenario) or 30 μg (high dose scenario). On days 1, 7, 28, or 84 after the last exposure, mice were euthanized, bronchoalveolar lavages (right lung only) were performed, and lungs were collected for further analysis. Left lungs were collected for flow cytometry (n = 3) and histology study (n = 3). contained either nanometer-sized thin flakes (USGO) or larger 2.1. GO Caused a Transient Innate Immune Response and micrometer-sized (LGO) thin flakes (Figure 1a) were used. Ex- Preferentially Activated Macrophages toward Anti-Inflammatory cept the lateral size, the characteristics of the GO sheets were Subsets identical (i.e., thickness, surface characteristics (oxidation level, surface charge), lack of endotoxins, etc.) (Figure 1a; Table S1, Sup- 2.1.1. Influx of Granulocytes in the Airways porting Information).[32] Moreover, the GO materials were pro- duced following strict endotoxin-free conditions. For enhanced Repeated exposure to endotoxin-free GO sheets caused dose- comparison, reference, multi-walled CNTs (MWCNTs; Mitsui-7) dependent acute inflammation, with influx of neutrophils and were used as positive control due to their well-known pulmonary eosinophils in the alveolar space, irrespective of the sheet lateral adverse effects.[33,34] The MWCNTs were depyrogenated prior to dimensions (Figure 2a; Figures S1,S2, Supporting Information). administration in mice. A recovery time of 2 weeks between each Differences were observed between the two GO materials; but exposure was chosen to evaluate a potential activation of the adap- overall the response was significantly lower in GO exposed an- tive immunity (Figure 1b). The acute and chronic pulmonary imals compared to the MWCNT group. In the alveolar space, adverse effects as well as the mechanisms involved in the bio- eosinophils were found to persist for a longer period of time than logical response were investigated at days 1, 7 (1 week), 28 (1 neutrophils, as they were observed for up to 7 days in GO materi- month = 4 weeks), and 84 (3 months = 12 weeks) after the last als groups and for up to 28 days in MWCNT exposed animals. A exposure. peak in eosinophils was noted at day 7 for LGO and MWCNTs but Adv. Sci. 2022, 9, 2104559 2104559 (3 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 2. Innate immunity activation in response to repeated exposure to GO and MWCNTs. Whole lung samples and broncho-alveolar lavage fluids (BALF) were analyzed to evaluate innate immunity activation. a) Phenotyping of granulocytes in the whole lung (flow cytometry) and BALF (colorimetric staining). b) Gene expression (RT-qPCR) and protein levels (ELISA) in the whole lung. c) Phenotyping of macrophages and monocytes in the whole lung (flow cytometry) and BALF (colorimetric staining). For cell phenotyping in BALF, gene expression, and protein concentration, one-way ANOVA followed by Dunnett’s post-hoc test or Kruskall–Wallis followed by Dunn’s post-hoc test was used to evaluate significant differences compared to the negative control for each time-point (n = 6; * p < 0.05, ** p < 0.01, and *** p < 0.001). Two-way ANOVA followed by Dunnett’s post-hoc test was used to evaluate changing in the number of cells in the whole lung (n = 3; * p < 0.05, ** p < 0.01, and *** p < 0.001). Adv. Sci. 2022, 9, 2104559 2104559 (4 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com not for USGO. Noticeably, we observed a clear dose-response for observed increases in interstitial M2 macrophages that were sig- both GO types, but not for MWCNTs, although the recovery was nificant at day 1 for both USGO and LGO (Figure S4, Supporting faster at the lowest dose for MWCNTs. Differences were observed Information). Significant increases in anti-inflammatory alveo- between bronchoalveolar lavages (BAL) and whole lung analyses lar M2 macrophages were observed up to 7 days for the LGO but (Figure 2a; Figures S1,S2, Supporting Information), showing that not for the USGO compared to the control (Figure 2c). A clear evaluating the specific influx of granulocytes in the alveolar space dose-dependent increase in multinucleated macrophages, sug- (i.e., BAL) was more sensitive. gesting fusion of macrophages, was also observed in BAL flu- ids for up to 84 days in LGO exposed animals (Figure 2c). Fu- sion of macrophages with other macrophages or immune cells 2.1.2. Activation of Pro-Inflammatory Mediators in the Lungs is typically observed in presence of foreign bodies and aims to eliminate large particles more efficiently.[36] Their presence Concomitantly to the influx of granulocytes, a dose-dependent at 28 days for USGO and up to 84 days for LGO, in absence increase in pro-inflammatory mediators was noted at day 1 for of significant arginase expression (Figure 2b), suggests a shift all the materials (Figure 2b; Figure S3, Supporting Information). of macrophage phenotype over time and highlights the impor- Interestingly, we found similar pattern of responses between tance of macrophage plasticity for the regulation of the immune USGO and LGO, with a peak at day 1 in both cases. At the high- response.[37,38] In comparison to GO materials or the vehicle con- est dose, global inflammation at day 1 was slightly stronger for trol, repeated exposure to MWCNTs induced a greater activation USGO compared to LGO, as highlighted by a higher expression toward M2 alternative phenotypes (Figure 2c). MWCNTs caused in acute phase protein Serum Amyloid A3 (SAA3) (Figure 2b; also a higher and dose-independent increase in arginase positive Figure S3, Supporting Information). However, SAA3 levels were macrophages compared to the vehicle control. A significant in- back to normal after 7 days, highlighting the absence of strong crease in these macrophages was still observed 84 days after the chronic inflammation, and a limited systemic impact following last exposure to MWCNTs at the highest dose. exposure to both USGO and LGO. At the highest dose, there Acute pro-inflammatory responses, involving cells of the in- was also a significant increase in gene expression or protein lev- nate immunity, were shown in previous studies after single els of the pro-inflammatory cytokines IL-1𝛼, IL-6, and TNF-𝛼 for oro-pharyngeal aspiration to GO[29] or other graphene-based the two GO materials. Interestingly, an increased expression of materials.[39] However, the present study provides new evidences CXCL1 was observed at day 1 (statistically significant for LGO but that chronic exposure to endotoxin-free GO could preferentially not USGO), underlining possible involvement of non-immune activate lung macrophages toward alternative subsets. Similar cells in the inflammatory response to GO. Pro-inflammatory me- M2 activation has been previously reported after intraperitoneal diators, including IL-1𝛼, were still significantly activated at days injection of GO[40] or graphene quantum dots,[41] and after pul- 7 and 28 for LGO but not USGO, suggesting a faster tissue recov- monary exposure to CNTs.[14] ery for USGO. In comparison, MWCNTs elicited a stronger and more long-lasting response, involving significant chronic activa- tion of pro-inflammatory mediators. Several pro-inflammatory 2.2. GO Did Not Significantly Activate the Adaptive Immunity markers were indeed still significantly upregulated and produced after 28 days (i.e., SAA3, CXCL1, and TNF-𝛼) or 84 days (i.e., IL- The activation of macrophages toward anti-inflammatory sub- 1𝛼 and IL-6) (Figure 2b; Figure S3, Supporting Information). No- sets in GO exposed animals agreed well with the presence of ticeably, a clear dose-response was observed for MWCNTs, with granulomas[42] at the high dose (Table S2, Figure S5, Supporting significant increases in pro-inflammatory markers at days 7 and Information) or the activation of inflammatory mediators linked 28 at the high dose but not the low dose. This difference between to granulomas,[43] including osteopontin (SPP-1) until day 7, and doses could explain why a significant increase of neutrophils in IL-1𝛼 until day 28 in LGO exposed animals (Figure 2b). The pres- the alveolar space was observed until day 28 at the high dose only ence of granulomas, along with the transient apparition of small (Figure 2; Figure S2, Supporting Information). bronchus-associated lymphoid tissue (BALT) structures in some but not all animals exposed to GO (Table S2, Figure S5, Support- ing Information), could suggest an involvement of lymphocytes 2.1.3. Impact on Lung Macrophage Populations in the immune response caused by repeated exposure to GO. In whole-lung samples, a dose-dependent replacement of alveo- lar macrophages by interstitial ones was observed for all materials 2.2.1. Impact of GO on Lymphocyte Populations and Activation of (Figure 2c; Figure S4, Supporting Information). This aligns well Th1, Th2, and Th17 Pathways with the significant increase in the monocyte chemoattractant mediator MCP-1 found in the lungs of both GO and MWCNT There was no evidence of lymphocyte recruitment, as highlighted exposed animals (Figure 2b; Figure S3, Supporting Information). by the absence of significant increase in CD4+ or CD8+ cell pop- This replacement has been previously observed under acute and ulations in the whole lung, or the absence of increase in total chronic inflammatory conditions.[35] However, we did not ob- lymphocytes in the alveolar space (Figure 3a; Figure S6, Sup- serve activation of macrophages toward pro-inflammatory sub- porting Information). There was also no activation of type 1 in- sets (M1), even at day 1 after the last exposure. Instead, we noted flammation, as no significant increases in IFN-𝛾 or the IL-12 an activation toward alternative phenotypes (M2) that was dose- family cytokines were measured in the lungs after exposure to dependent for GO but not for MWCNTs. At the high dose, we GO (Figure 3b; Figure S7, Supporting Information). In contrast, Adv. Sci. 2022, 9, 2104559 2104559 (5 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 3. Adaptive immunity activation in response to repeated exposure to GO and MWCNTs. Whole lung samples and broncho-alveolar lavage fluid (BALF) were analyzed to evaluate adaptive immunity activation. a) Phenotyping of dendritic cells and lymphocytes in the whole lung (flow cytometry) and BALF (colorimetric staining). b,d) Protein levels (ELISA) and c,e) gene expression (RT-qPCR) of inflammatory mediators associated to adaptive immunity. For cell phenotyping in BALF, gene expression and protein levels, one-way ANOVA followed by Dunnett’s post-hoc test or Kruskall–Wallis followed by Dunn’s post-hoc test was used to evaluate significant differences compared to the negative control for each time-point (n = 4–6; * p < 0.05, ** p < 0.01, and *** p < 0.001). Two-way ANOVA followed by Dunnett’s post-hoc test was used to evaluate changing in the number of cells in the whole lung (n = 2–3; * p < 0.05, ** p < 0.01, and *** p < 0.001). we found increases in protein levels of TNF-𝛼 (significant only volved in their secretion, including many cell types for TNF-𝛼, for LGO; Figure 2b;Figure S3b, Supporting Information) and and neutrophils or macrophages for IL-17A.[44,45] Moreover, we IL-17A (non-significant; Figure 3d) for both GOs at day 1 and did not measure significant increases in production or expres- the highest dose. Although these two cytokines can be released sion of type 2 inflammation mediators (e.g., IL4, Figure 3b,c) by Th1 or Th17 lymphocytes, respectively, other cells can be in- or immunoglobulin E (IgE) (Figure 3b,d) after exposure to GO, Adv. Sci. 2022, 9, 2104559 2104559 (6 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com although some eosinophils were found in the lungs (Figure 2a). alveolar or bronchial epithelium (Table S3, Figure S5, Support- Taken together, these findings support the idea that there was no ing Information). Moreover, no evidence of collagen deposition activation of the adaptive immunity or any allergy-like response was found in USGO or LGO exposed animals compared to the in GO exposed animals. In addition, despite a transient increase vehicle control (Figure 4). In contrast, MWCNTs induced colla- in M2 arginase positive macrophages (Figure 2c), we did not ob- gen deposition at 28 days irrespective of the dose applied (Fig- serve any significant increase in IL-10 or TGF-𝛽 (Figure 3b,c; ure 4f). Evidence of a time-dependent downregulation of colla- Figure S7, Supporting Information), suggesting that their pres- gen gene expression (statistically significant at 28 and 84 days; ence may not be related to a sustained Th2 pathway.[46,47] Interest- Figure 4a) was also found, suggesting a feedback regulation of ingly, we found an increase in non-mature resident dendritic cells the gene expression as the protein is secreted and the lung fibro- (CD11b+ DCs) at day 1, but only for LGO and at the highest dose, sis settled (Figure 4a; Figure S10, Supporting Information). De- and with no increase in conventional DCs (cDCs) (Figure 3a; spite the collagen deposition observed in the lungs of MWCNT Figure S8, Supporting Information). This absence of increase in exposed animals, no significant alteration of the gene expression cDCs also supports our conclusions that pulmonary exposure to for 𝛼-SMA and Vimentin (i.e., biomarkers of EMT process of- GO sheets does not significantly activate adaptive immunity.[48] ten associated to fibrogenesis[50,51] ) was found at any time-point (Figure 4a; Figure S10, Supporting Information). However, an alveolar and bronchial cell hyperplasia in tissue sections, and a 2.2.2. Comparison of GO and MWCNTs Impact on Adaptive decrease in matrix metalloproteinase gene expression in whole Immunity tissue lysates were identified in MWCNT exposed animals (Fig- ure 4a; Figures S5a,b, S10, Table S3, Supporting Information). In contrast to GO, lymphocytes were involved in the immune re- These last results likely contributed to the formation of fibrotic sponse to MWCNTs, used here as positive control for adverse ef- scars and further confirmed the negative impact of MWCNT on fects (Figure 3a; Figure S6, Supporting Information). A signif- lung tissue plasticity. icant increase in lymphocytes in BAL fluids and distinct BALT structures were found at every time-point for MWCNTs, with a peak at 7 days. Variations in specific populations were also ob- 2.3.1. Adaptive Immunity Activation and Progression toward served in the lungs, including significant increases in NK cells at Pathological Conditions day 1 and T helper populations (CD4+ ) at days 1, 7, and 28. We also found an increase in cDCs after repeated exposure to MWC- The absence of significant lymphocyte recruitment or chronic ac- NTs (Figure S8, Supporting Information), suggesting that DCs tivation of Th2 mediators following GO exposure and the oppo- maturation could have caused further activation of lymphocytes site following MWCNT exposure could explain both the rapid in- toward CD4+ Th2 subsets.[48] Moreover, we measured a signifi- flammation resolution found after repeated GO exposures and cant increase in expression or protein levels of the Th2 mediators the lung tissue remodeling observed after repeated MWCNT ex- IL-4 and IL-5 (Figure 3b–e) associated to a strong and chronic posures. Lung tissue remodeling is characterized by the occur- eosinophilia in the alveolar space (Figure 2a; Figure S2, Support- rence of structural changes in the lung parenchyma or in the ing Information). These results together with the increase in IgE airways, including changes of epithelial tissue structure and mor- and the observation of Charcot Leyden like crystals crystals[49] phology, or evidence of fibrosis.[52] Indeed, chronic Th2 response in the bronchia (Figure 3b,d; Figure S5a–c, Supporting Informa- with IL-4 secretion, as observed for MWCNTs, has been shown tion) were indicative of a Th2 response,[46] as previously reported to be critical for collagen deposition and the formation of fi- after single pulmonary exposure to MWCNTs.[14] Interestingly, brotic granulomas.[46,47,53] By promoting chronic eosinophilia, although we identified a dose-dependent secretion of innate or epithelial cell proliferation and differentiation through inhibi- adaptive inflammatory mediators into the lungs, the influx of im- tion of matrix metalloproteases and fibroblast activation, the mune cells, and the allergy-like response, especially the recruit- activation of lymphocyte CD4+ toward Th2 subsets is critical ment of eosinophils and lymphocytes in the BAL fluids, were not in maintaining the type 2 polarization.[46] This underline the dose-dependent (Figures 2,3). This leads to the conclusion that importance of evaluating adaptive immunity activation to pre- the innate immune response to MWCNTs was dose-dependent, dict long-term lung damages and evolution toward pathologi- whereas the adaptive immune response is independent from the cal conditions. In the present study, the formation of fibrosis dose, highlighting the greater health risk of these materials com- and progression toward granulomatous disease after MWCNT pared to GO. Indeed the absence of significant lymphocyte acti- exposure was further confirmed by the appearance of a late- vation and adaptive immune response to GO sheets, irrespective onset Th1 response at 84 days (e.g., increase of INF𝛾; Fig- of their lateral dimensions, constitutes a major difference com- ure 3b,d; Figure S7, Supporting Information) and the increased pared to MWCNTs. This important difference demonstrates the CD4+ /CD8+ ratio irrespective of the dose applied (Figure 4d).[54] relative immune tolerability of GO following repeated pulmonary A Th1 response would typically be expected to occur in the exposure and highlights the safer profile of the material. acute phase of the response to MWCNT exposure.[14] Never- theless, its absence at early time-points can be explained by 2.3. GO Did Not Cause Lung Tissue Remodeling and Showed the repeated administration regimen applied. Indeed, within Good Recovery the context of repeated exposures used here, it is rational- ized that when the last (third) exposure was applied, resolution In the present study, none of the endotoxin-free GO tested, re- of inflammation and healing processes that follow the previ- gardless of the dosage applied, caused any visible damages to the ous (second) exposure were still ongoing. Therefore the overall Adv. Sci. 2022, 9, 2104559 2104559 (7 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 4. Mechanisms associated to oxidative stress, cell death, and tissue remodeling in lungs exposed to GO and MWCNTs. 1, 7, 28, and 84 days after the last exposure, lungs were harvested and processed for analysis of a) gene expression in right lungs by RT-qPCR (n = 5–6), of oxidative stress I) pro-oxidant, II) anti-oxidant, programmed cell death, and tissue remodeling markers. b) TUNEL assay performed in left lung sections to investigate relative apoptosis over-time (n = 3). c) Number of non-immune cells counted by flow cytometry in digested left lungs (n = 3). d) Ratio of CD4+/CD8+ lymphocytes evaluated by flow cytometry in digested left lungs (n = 3), dotted blue lines represent the granulomatous disease threshold based on existing literature.[54] e) Semi-quantitative analysis of collagen-I deposition in lung sections at 28 and 84 days (n = 3). f) Lung sections stained with Masson Trichrome to evaluate collagen deposition at days 28 and 84 (n = 3). One-way ANOVA followed by Dunnett’s post-hoc test or Kruskall–Wallis followed by Dunn’s post-hoc test were used to evaluate statistical differences compared to the negative control (* p < 0.05, ** p < 0.01, and *** p < 0.001). Scale bar corresponds to 50 μm. biological response after the last exposure will be a combina- last exposure. Altogether, these combined responses (to each ex- tion of acute inflammatory (likely Th1 like response) and matura- posure) may explain the unconventional increase in specific Th2 tion/healing/remodeling phases (dominantly Th2 like response). markers measured here in the early phase of the response to the It may even be possible that long-lasting effects due to the first last MWCNT exposure (e.g., day 1) and the late-onset Th1 re- exposure have contributed to the outcomes measured after the sponse. Adv. Sci. 2022, 9, 2104559 2104559 (8 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com 2.3.2. Role of Macrophages in the Maintenance and Resolution of fects. Oxidative stress, immune cell death by piercing cell and the Inflammation lysozyme membranes,[61] as well as granulomatous inflamma- tion leading to fibrotic tissue and persistence in the lungs after In an inflammatory environment, the role played by pulmonary exposure have been extensively reported.[62] In the macrophages is important for both the maintenance and present study, both significant increases in gene expression of the resolution of the immune response. Macrophages activated the pro-oxidative heat shock proteins HSP70 and HO-1 at both to anti-inflammatory subsets can suppress inflammation me- days 1 and 7, and significant decreases in gene expression of diated by M1 macrophages through the secretion of IL-10 and the anti-oxidant markers GSH, SOD1, and NRF-2 at the earliest TGF-𝛽 to prevent installation of type 2 inflammation.[55] On time-points were found after repeated exposure to MWCNTs (Fig- the contrary, tissue repair macrophages activated in a Th2 envi- ure 4; Figure S9, Supporting Information). This is in agreement ronment were shown to be important in maintaining a type 2 with the previous studies mentioned above, highlighting the ox- inflammation.[46,47] In the present study, these two macrophage idative stress caused by these well-studied MWCNTs. Interest- population subsets were not distinguished but combined and ingly, significant inhibition in CYP1A1 was also identified (Fig- identified as M2, based on their arginase expression (Figure 2c). ure 4; Figure S9, Supporting Information), suggesting a possi- As a result, both GO sheets and MWCNTs were found to promote ble decrease in MWCNT detoxification processes, as reported for M2 macrophages, despite obvious differences in the immune other carbon substances including benzo[a]pyrene.[63] The down- response following exposure to these two categories of materials. regulation of CYP1A1 expression could also be attributable to the For MWCNTs, both M2 macrophages and Th2 mediators were presence of oxidative stress.[64] Finally, although there was a dose- increased. This suggests that macrophages were pushed toward independent increase in the number of apoptotic cells at all the the tissue repair subset. In contrast, considering the absence time-points after MWCNT exposure highlighting the toxicity of of a distinct Th2 response after exposure to GO, the results these materials (Figure 4b), there was a downregulation of two suggest that GO sheets were promoting a macrophage differ- regulators of apoptosis namely caspase-9 and -2 (Figure S11, Sup- entiation toward anti-inflammatory subsets rather than tissue porting Information). While evolution of mRNA levels should be repair phenotypes. In addition, in the absence of Th2 activation, interpreted with cautious because of post-translational regulation anti-inflammatory macrophages have been reported to regulate (i.e., pro-caspase vs active caspase), this decrease may explain the the immune response, preventing the development of chronic increase in non-immune cell number found in the lungs at day inflammation or collagen deposition in the airways.[46] This 7 after MWCNTs (Figure 4c). Such increase after MWCNT expo- immune regulation by anti-inflammatory macrophages would sure would be expected since Th2 immune response has been re- support our findings of a successful recovery after exposure to ported to cause hyperplasia of the epithelium.[46] In contrast, GO either LGO or USGO, and would explain lung tissue remodeling sheets irrespective of their dimensions did not cause sustained after exposure to MWCNTs (Figure 4; Figure S5, Supporting oxidative stress (Figure S9, Supporting Information) or sustained Information). apoptosis beyond day 1 at the highest dose (Figure 4b; Figure S11, Supporting Information). These two findings support the lack of 2.4. Biological Response to Pulmonary Exposure is adaptive immune response or lung tissue remodeling after GO Material-Specific exposure, and underline the safer profile of GO sheets compared to MWCNTs. GO sheets and MWCNTs present numerous physicochemical differences that can explain the discrepancies between the bio- 2.4.2. Role of Metallic Impurities in the Biological Response to GO logical responses to these materials. The two intertwined struc- and MWCNTs tural parameters that apply to both GOs and MWCNTs and have been associated before to adverse effects are the aspect-ratio The presence of metallic impurities in nanomaterials has been and rigidity/flexibility.[56] The MWCNTs used here were tubular associated to greater adverse effects in lungs. Pulmonary expo- nano-objects with a typical diameter of 49 ± 13.4 nm for a length sures to soluble metals or metallic particles have been shown ranging between 1 and 20 μm (median = 3.86 μm).[32,57] In con- to activate the adaptive immunity and trigger allergy-like Th2 trast, the GO sheets were thin 2D nano-objects (below 10 nm responses, causing tissue damages.[65,66] Carbon nanomaterials thickness, vast majority ≈1 nm) with lateral dimensions ranging containing metallic catalyst particles may also release those im- from 10–300 nm (for USGO) to 1–30 μm (for LGO), which gave purities in the form of ions after internalization by immune them a platelet-like shape.[58] The lateral dimensions of the LGO cells.[67,68] This release of metallic ions from carbon nanomate- sheets may qualify them as high aspect ratio materials to a sim- rials, especially iron, has been associated to greater toxicity.[69–72] ilar extent as the MWCNTs. However, LGO sheets appeared to Importantly, we did not detect any metallic impurities in our in- remain flexible nano-objects (see videos in ref. [59] in which the house GO sheets,[32] while the MWCNTs used here are known same materials were used) compared to MWCNTs that behave to have metallic residues (see Table 1 in ref. [73]). Therefore, like rigid needles.[60] the lack of metallic impurities in GO materials and their pres- ence in MWCNTs might explain the absence of Th2 response 2.4.1. Impact of the Aspect Ratio on the Biological Response to GO and lung tissue remodeling after repeated exposure to GO, and and MWCNTs the dose-independent allergy-like response to MWCNTs. Fur- ther investigations remain necessary to evaluate the impact of Long and rigid MWCNTs similar to those used here, owing to metallic impurities presence in the toxicity of these nanomate- their fibre-like aspect, were shown to cause specific adverse ef- rials. Nevertheless, such assessment is difficult since the use of Adv. Sci. 2022, 9, 2104559 2104559 (9 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com purification methods also alters the physicochemical characteris- amount of USGO interacting with BAL macrophages (from tics of the materials.[34] 52.6% at 1 day to 3.7% at 84 days; Figure 5b). This was sug- gesting that migration of GO laden macrophages, likely relocat- ing to draining lymph nodes or eliminated following mucociliary 2.5. GO Fate in the Lungs is Size-Dependent clearance mechanisms, may also play a significant role in USGO elimination from the respiratory tract alongside the reverse mi- 2.5.1. GO Clearance from the Lungs gration of GO laden neutrophils. In contrast, BAL macrophages in LGO exposed animals remained relatively unchanged for up When comparing LGO and USGO, most findings from the to 28 days after the last exposure (58.4% at 1 day, still 30.6% at present study were pointing toward the activation of similar cel- 28 days, then 15.1% at 84 days; Figure 5b). The persistence of lular and molecular mechanisms. Nevertheless, LGO sheets were LGO materials in lungs and BAL fluids could be explained by the inducing a longer pro-inflammatory reaction and a stronger anti- relative inability of macrophages to internalize large materials, inflammatory response compared to USGO sheets. The reduced which in turn leads to macrophage fusion and decreased elim- and shorten biological response to USGO exposure could be ex- ination. This aligns well with the findings of a greater amount plained by a faster clearance from the lungs of USGO compared of multinucleated macrophages after LGO exposure than after to LGO, as suggested by Raman spectroscopy-based images (Fig- USGO exposure, especially at the high dose (Figure 2c). More- ure 5a). Indeed, while there were clear evidences of remaining over, the persistence of GO laden BAL macrophages in the case materials in LGO exposed lungs after 84 days, it was difficult to of LGO exposed animals was consistent with the greater number identify remaining USGO materials at the same time point de- of M2 macrophages (Figure 2c) and granulomas (Table S2, Fig- spite their presence at earlier time points. These findings agree ure S5, Supporting Information), as well as the greater activation with our previous study in which a time-dependent and size- of specific inflammatory markers expressed in these granulomas dependent clearance of GO sheets from the lungs was also re- (e.g., IL-1𝛼; osteopontin, SPP-1) (Figure 2b). Indeed, macrophage ported after single intranasal administration.[31] fusion and activation toward M2 phenotypes, as well as the for- mation of granulomas, have been reported before when foreign bodies including discoid materials[76] cannot be internalized effi- 2.5.2. GO Internalization by Neutrophils ciently by macrophages.[36] Therefore, the large aspect ratio of the LGO sheets (i.e., platelet-like shape and large lateral dimensions) Going further in the understanding of the clearance process, we may be a disadvantageous feature in respect to their clearance by then identified that the recruited neutrophils, which were found resident and recruited macrophages,[56,77] or neutrophils, in com- in BAL fluid only at day 1 (Figure S1, Supporting Information), parison to USGO nanosheets. may play an essential role in the rapid elimination of USGO In addition to a cell-mediated clearance, the elimination of materials from the respiratory tract (Figure 5b). Indeed, Raman GO materials from the lungs (i.e., fast and constant for USGO, imaging coupled with differential staining of BAL cells from at slower pace for LGO) could also be attributed to mate- GO exposed lungs highlighted a greater interaction between rial degradation. Both neutrophils and macrophages have in- neutrophils and USGO than with LGO; 70% of neutrophils deed been reported to not only remove but also digest carbon in USGO condition were GO-positive, while only 20% of neu- nanomaterials.[26] Further research looking specifically at the trophils in LGO condition were GO-positive. This difference in degradation of GO in laden neutrophils and macrophages, in the interaction was associated to a difference in the relative amount lungs, or after migration in secondary locations such as draining of materials per neutrophil (total loading), reaching 15.6% for lymph nodes, are therefore warranted to reveal the final fate of USGO and only 3.2% for LGO. Material internalization by GO sheets after pulmonary exposure. This is a key requirement neutrophils has been previously reported in blood samples or for the clinical translation of promising pulmonary nanovectors. in vitro.[74] We therefore theorized that the greater interaction revealed by Raman could be evidence of a greater internalization 3. Conclusions of USGO sheets by the recruited neutrophils in comparison to LGO. In this hypothesis, the differences observed between In the present study, we comprehensively evaluated the lung tol- USGO and LGO could be attributed to a size-dependent in- erability of two types of endotoxin-free GO materials with vary- ternalization, with USGO nanosheets being more internalized ing lateral dimensions in mice after repeated pulmonary expo- due to smaller dimensions. Since neutrophils have the ability sure. The immune response and histology of the lungs were in- to reverse migrate quickly to lymph nodes or blood vessels as vestigated for up to 3 months after the last exposure. We found inflammation resolved,[75] a greater amount of USGO than LGO that regardless of their lateral dimensions, GO sheets caused could have been cleared from the lungs by reverse migration of a dose dependent acute immune response involving primarily GO laden neutrophils, reducing more quickly the overall lung cells of the innate immunity. At later time points, there was no burden for USGO in comparison to LGO. activation of the adaptive immunity or allergy-like response for any of the GO materials. Instead, there were clear evidences of immune response resolution and lung tissue recovery for both 2.5.3. GO Internalization and Clearance by Alveolar Macrophages GO materials, even at the highest dose. Lateral dimensions were found to influence the internalization of GO sheets by phago- When comparing different time points, Raman imaging of BAL cytic cells and their persistence in the lungs, which in turn affect cells also revealed a fast and continuous decrease in relative the pace of tissue recovery. While the nanometric GO materials Adv. Sci. 2022, 9, 2104559 2104559 (10 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 5. Size-dependent clearance of GO materials. a) Lung sections and b) broncho-alveolar lavage fluid (BALF) slides were analyzed by Raman spectroscopy imaging. BALF slides were then stained for cell phenotyping with Kwik–Diff staining usually used for blood cell morphology test to identify cells and evaluate material internalization by macrophages and neutrophils. c) High magnification images of BALF cells at day 1 after repeated exposure to USGO and LGO. d) Internalization of USGO and LGO in neutrophils at day 1. The proportion of neutrophils that have internalized GO (positive cells (%)), the average loading of GO-positive cells (%) and the total loading (%) are presented. e) Internalization of USGO and LGO in macrophages at days 1, 7, 28, and 84. The proportion of macrophages that have internalized GO (positive cells (%)), the average loading of GO-positive cells (%), and the total loading (%) are reported. At each time-point, t-test or Man–Whitney test were used to evaluate statistical differences (n = 3; * p < 0.05, ** p < 0.01, and *** p < 0.001). Adv. Sci. 2022, 9, 2104559 2104559 (11 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com were eliminated rapidly, consistent with a faster tissue recovery, Absorption Spectroscopy: Data were obtained using a Nanodrop 2000c the micrometric GO materials persisted in the lungs and caused spectrophotometer (ThermoFisherScientific), and a Hellma QS Quartz longer lasting effects such as granulomas. Taken together, these micro cuvette. GO samples were prepared in water (concentration range of 2.5–20 μg mL−1 ) and measured at room temperature. results highlight the ability of the immune system to resolve the Fluorescence Emission Spectroscopy: Fluorescence emission was as- sterile inflammation caused by repeated pulmonary exposures to sessed for GO samples at concentration range of 25–200 μg mL−1 using endotoxin-free GO materials, irrespective of their lateral dimen- a LS-50B Perkin Elmer Spectrofluorometer, at the excitation wavelength of sions. They also demonstrate that reducing lateral dimensions 525 nm, with both excitation and emission slits set at 20. improved the tolerability of GO in lungs owing to a better inter- Raman Spectroscopy: Raman spectra were acquired with a confocal nalization by phagocytes leading to a better clearance profile. This Raman microscope (WITec) at room temperature, using a 632 nm laser excitation and grating of 600 g nm−1 . Single Raman spectra were collected comprehensive analysis of GO lung tolerability will help guiding on several spots after irradiation with a power of 1 mW for 10 s and using the safe development of future pulmonary nanomedicines based a 50× objective. Samples were prepared by drop casting 20 μL onto glass on GO sheets, as well as providing wider stakeholders with key coverslip left to dry overnight. The data were analyzed with Origin software information regarding the pulmonary impact of GO-based nano- (OriginLab Corporation; https://www.originlab.com/). The ratio of D and materials. G band intensities (ID /IG ) was calculated without baseline correction (D: peak height intensity at ≈1340 cm−1 ; G: at ≈1580 cm−1 ). X-Ray Diffraction Spectroscopy: Data were acquired at the ICN2 X- 4. Experimental Section Ray Diffraction Facility with X’Pert MPD (Multipurpose diffractometer) equipped with a ceramic X-ray tube with Cu Κ𝛼 anode (𝜆 = 1.540 Å) as GO Materials: Biological-grade USGO and LGO materials, dispersed X-ray source and x’Celerator solid-state detector in the 2𝜃 scan range from in sterile water for injection, were produced by the modified Hummers’ 5° to 40°, by drop casting the GO samples. method, under strict endotoxin-free conditions as previously reported,[30] Zeta Potential: This was measured using a Zeta-sizer Nano ZS using graphite powder as starting material (Merck-Sigma). The two GO (Malvern Panalytical Ltd) equipped with disposable capillary cells at the materials were then fully characterized (Table S1, Supporting Information) ICN2 Molecular Spectroscopy and Optical Microscopy Facility. Water dis- in a similar fashion to what was previously described.[30] The structural persant settings for refractive index and viscosity, and automatic analysis and morphological properties of USGO and LGO sheets were determined were used for all GO measurements (20 μg mL−1 ). Each sample was mea- by atomic force microscopy (AFM), scanning electron microscopy (SEM), sured three times at room temperature. and optical microscopy (Figure 1a). First, AFM images were acquired us- Thermogravimetric Analysis: This was used to determine the weight ing an Asylum MFP-3D atomic force microscope (Oxford Instruments) loss of GO samples using a Pyris 6, Perkin-Elmer Ltd. A GO sample of operating in standard air-tapping mode and equipped with silicon probes 1–2 mg weighed into a ceramic crucible was analyzed from 25 to 995 °C (Ted Pella) with a resonance frequency of 300 kHZ; a nominal force of at 10 °C min−1 with a nitrogen flow of 20 mL min−1 . 40 N m−1 was used to characterize the surface. Samples were prepared X-Ray Photoemission Spectroscopy: Data were obtained using a Phoi- by drop casting 20 μL of GO suspension (100 μg mL−1 ) onto a freshly bos 150 (SPECS, GmbH) electron spectrometer equipped with a hemi- cleaved mica surface (Ted Pella) previously covered with 20 μL of poly- spherical analyzer, operating under ultrahigh-vacuum conditions, and with L-lysine 0.01% solution (Merck-Sigma), subsequently washed with water, an Al K𝛼 (h𝜈 = 1486.74 eV) X-ray source, at the ICN2 Photoemission Spec- and then dried overnight at room temperature. Images were processed troscopy Facility. Samples were prepared by deposition of 20 μg of GO ma- using Gwyddion software (http://gwyddion.net, version 2.56). To obtain terials onto 5 × 5 silicon wafers (Ted Pella) left to dry overnight. In order the lateral dimension distribution, analysis of height in AFM images was to estimate the photoelectron peak intensities, the CasaXPS software was performed using ImageJ software (National Institutes of Health, USA, used (Casa Software Ltd; http://www.casaxps.com). https://imagej.nih.gov). Second, SEM images were recorded on a Mag- Animal Exposure: C57BL/6J female mice (6 to 8 weeks old) were pur- ellan 400L field emission scanning electron microscope (Oxford Instru- chased from Envigo, UK. The experiment was randomized and mice were ments), which was equipped with an Everhart–Thornley as secondary elec- kept in groups of four in ventilated cages with ad libitum access to food trons detector, using an acceleration voltage of 20 kV and a beam current of and water under controlled environmental conditions (humidity, temper- 0.1 nA (Catalan Institute of Nanoscience and Nanotechnology, ICN2, Elec- ature, and light controlled environment). All procedures were conducted tron Microscopy Unit, Spain). The GO sample was deposited on an Ultra- after ethical approval from the UK Home Office, under the Project License thin Lacey C grid; any excess of material was removed and dried overnight no. P089E2E0A. After at least a week of acclimatization, each animal was at room temperature. The lateral dimension distribution was obtained by exposed by oropharyngeal aspiration every 14 days to either 1 μg (low measuring the flakes using ImageJ software. Finally, optical images were dose) or 10 μg (high dose), to achieve a total cumulative dose of either 3 μg acquired with a Nikon Eclipse LV100 microscope in transmittance mode (low dose) or 30 μg (high dose) per animal after three exposures. Briefly, at a magnification of 50×. Only micrometer-sized flakes were visible un- animals were anaesthetized by inhalation of 4% isoflurane in 100% oxy- der the optical microscope, which resulted in the recording of images for gen and then held on a slanted board in order to deliver 30 μL of materials LGO, but no flakes were detectable for USGO (i.e., dimensions below the or vehicle (water for injection, Gibco, ThermoFisherScientific). The ani- light diffraction limit). The LGO flakes were measured using ImageJ soft- mals (6 animals per group) were kept for 1, 7 (=1 week), 28 (=4 weeks/1 ware. Further characterization techniques applied to these GO materials month), or 84 (=12 weeks/3 months) days after the last exposure. are detailed in the next section. Sample Collection: At days 1, 7, 28, and 84 after the last exposure, an- MWCNT Materials: MWCNTs (Mitsui, Japan, type Mitsui-7), kindly imals were euthanized by IP injection of pentobarbitone (100 μL). During provided by Prof. U. Vogel (National Research Centre for the Working En- the autopsy, left lungs were clamped and right lungs were washed with PBS vironment), were heated overnight in the oven at 160–180 °C for depyro- (Merck-Sigma) to collect BAL fluid (n = 6). After lavage, right lungs were genation. Characteristics of these materials were published before.[32–34] cut in small portions and then stored in tubes containing either 1 mL of MWCNTs were then dispersed in water for injection (Gibco, ThermoFish- RNAlater (n = 6) (Merck-Sigma) or 0.5 mL of RIPA Buffer (n = 6) (Merck- erScientific) containing 0.5% bovine serum albumin (BSA; Gibco, Ther- Sigma) for RNA or protein extraction respectively. Non-washed left lungs moFisherScientific) and submitted to sonication in a water bath for 5– were collected in tubes containing 1 mL of RPMI medium (Merck-Sigma) 7 min at 80 W (VWR essential, UK). and kept at 4 °C for flow cytometry (n = 3) or inflated and kept in formalin Endotoxin Assay: The potential contamination by endotoxins was eval- (Merck-Sigma) for histology (n = 3). uated according to Mukherjee et al.[78] using mouse primary bone marrow BAL Fluid Analysis: BAL fluids were centrifuged at 1500 rpm for 5 min derived macrophages, and all the materials (USGO, LGO, and MWCNT) at room temperature (Hettich GmbH). Supernatants were aliquoted and were tested negative (Figure S12, Supporting Information). stored at −80 °C (Brunswick, Eppendorf). Pellets were resuspended in PBS Adv. Sci. 2022, 9, 2104559 2104559 (12 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com (Merck-Sigma), cells were counted and cytospun at 600 rpm for 5 min with tions were normalized by the total protein content and were expressed in ≈100 000 cells per slide (Hettich GmbH). Slides were then fixed in 100% pg mg−1 of protein. ice-cold methanol for 10 min, air-dried and stored at −20 °C. Differential Gene Expression by Multiplex RT-qPCR: Prior to RT-qPCR analysis, lung cell staining was performed using Kwik-Diff (Shandon, ThermoFisherSci- samples were kept in RNAlater storage solution (Merck-Sigma) until pro- entific) following the provider’s instructions. The number of neutrophils, cessing. After removing of RNAlater and washing with PBS, lungs were eosinophils, mono- and multinucleated macrophages, and lymphocytes lysed in 1 mL of lysis buffer containing 2-mercaptoethanol (Gibco, Ther- were assessed using optical microscopy (AxioOberver, Zeiss). Colorimet- moFisherScientific). After homogenization using 5 mm stainless steel ric bright-field images (Figure 5c; Figure S1a, Supporting Information) beads and a TissueLyser LT (Qiagen) running at 50 Hz for 10 min, lysed were generated with a slide scanner (Pannoramic 250 Flash, 3DHistech lungs were centrifuged at 2600 g for 5 min (Hettich GmbH) to remove cell Ltd). The protein content in BAL supernatants was evaluated using a BCA debris. Supernatants were kept at −80 °C prior to extraction. Total RNA assay (Pierce BCA Protein Assay Kit, ThermoFisherScientific) following was extracted using spin cartridges containing silica membranes (Pure- provider’s instructions. Absorbance values were recorded at 562 nm using Link RNA Mini kit, Qiagen) according to the manufacturer’s instructions. a spectrophotometer (Varian Cary 50, Agilent) and concentration values Briefly, after addition of 1 volume of 70% ethanol (ThermoFisherScientific) were calculated using a standard curve. for 1 volume of sample to precipitate RNA, samples were loaded in the car- Cell Phenotyping Using Flow Cytometry: Harvested lungs were cut into tridge. After washing steps and addition of DNase, the extracted RNA was small parts and digested using 0.4 Wünsch units mL−1 of Liberase (Roche) eluted using 30 μL of RNase-Free Water (Gibco, ThermoFisherScientific). and 0.01 mg mL−1 of DNase I (Roche) in RPMI medium (Merck-Sigma). Total RNA concentration was measured and purity was calculated using After shaking for 30 min at 37 °C (MaxQ, ThermoFisherScientific), 5% fetal a Biophotometer Plus spectrophotometer (Eppendorf AG). RNA samples bovine serum (FBS; Gibco, ThermoFisherScientific) was added and the di- were kept at −80 °C until processing. For each sample, first-strand cDNA gested lungs were filtered through 100 μm filter meshes (Greiner Bio-one). was synthesized from 1 μg of extracted RNA using the High Capacity cDNA RPMI medium containing 5% of FBS (Gibco, ThermoFisherScientific) and Reverse Transcription kit (Applied Biosystems). Briefly, 10 μL of reverse 5 mm of EDTA (Gibco, ThermoFisherScientific) was added to prevent cell transcription Master Mix and 10 μL of sample were incubated for 10 min aggregation. Cells were centrifuged for 5 min at 2300 rpm, supernatants at 25 °C, followed by 2 h at 37 °C, and 5 min at 85 °C in a CFX96 real- were discarded, and red blood cells were lysed for 7 min using 500 μL of time PCR system (BioRad). The expression of pre-selected 48 genes (Table Red Blood Cell Lysing Buffer (Hybri-Max, Merck-Sigma Aldrich). After cen- S5, Supporting Information) was evaluated in lung samples by Multiplex trifugation at 2300 rpm for 5 min (Hettich GmbH), cell pellets were sus- RT-qPCR in a Biomark HD system (BioMark HD 96.96 IFC, Fluidigm) us- pended in 1 mL of RPMI medium (Merck-Sigma). Cells were counted using ing extracted cDNA (30 μl at 1 ng μL−1 ) and primers (10 μL at 50 μm) a haemocytometer and incubated using a Live/Dead UV viability dye (Ther- (Table S5, Supporting Information), according to the manufacturer’s in- moFisherScientific) for 20 min at room temperature for cell death evalua- structions. Samples were analyzed in duplicates. For every sample, ΔΔCt tion. After a washing step (centrifugation at 2300 rpm for 5 min and wash- values were calculated, normalized using house-keeping genes, and data ing with PBS), cells were suspended in 100 μL of staining buffer (3% FBS were finally expressed in fold change compared to the negative control. and 0.05% sodium azide (Merck-Sigma) in PBS). Non-specific sites were Histology Block Preparation: After sampling, inflated lungs were stored blocked with anti-mouse CD16/CD32 (Fc block, eBiosciences) for 5 min in formalin 10% (Merck-Sigma) for 24 h and then transferred to vials con- and cells were incubated with fluorochrome-conjugated antibodies (see taining ethanol 70% (ThermoFisherScientific). The lungs were embedded Table S4, Supporting Information) for 30 min at 4 °C for staining of spe- in paraffin, and sections of 5.0 μm of thickness were obtained using a mi- cific cell membrane receptors. Compensations, fluorescence minus one crotome (RM2255, Leica Biosystems). and controls were also prepared. After washing with PBS, cells were finally Histological Staining: For histopathological analysis, sections were fixed at room temperature for 20 min with 4% of PFA (Merck-Sigma) in stained with hematoxylin and eosin using an automatic stainer (XL au- PBS, washed again with PBS, and kept overnight at 4 °C in staining buffer tostainer, Leica Biosystems). Bright-field images were generated with a (3% FBS and 0.05% sodium azide in PBS). The next day, cells were washed slide scanner (Pannoramic 250 Flash, 3DHistech Ltd). Lung perimeter, with PBS and suspended in 100 μL of permeabilization buffer (0.5% tween pleural and bronchial thickness, and granulomatous and BALT areas of 20 with 1% FBS and sodium azide 0.01 in PBS). After incubation for 30 min each sample were identified and analyzed using CaseViewer (software ver- at 4 °C and a PBS washing step, 5% FBS in PBS was added for 20 min sion 2.4.0.11902, 3DHISTECH Ltd). Masson’s Trichrome staining of lung to block non-specific intracellular binding sites. Cells were washed with sections was also performed for the 28 and 84 days samples only. Level of PBS, suspended in 100 μL of permeabilization buffer, and then stained collagen deposition was determined by automatic pixel quantification of for 60 min at room temperature in the dark for intracellular staining with five images (region of interest) for each section using Python software (v. Arginase (Table S4, Supporting Information). After a final PBS washing 3.9.1). step, cells were suspended in 300 μL of staining buffer and analyzed by TUNEL Assay: Quantitation of apoptotic cells in the paraffin embed- flow cytometry using a BD LSR II (BD Biosciences), for evaluating variation ded lung sections was performed using the DeadEnd Fluorometric TUNEL of immune cell populations in the whole lung. Granulocyte, macrophage, System (Promega). As pre-treatment, sections were dewaxed and rehy- monocyte, dendritic cells, and lymphocyte populations and their subsets drated in series of ethanol solutions and water. For detection, sections were isolated using specific gating (see Figure S13, Supporting Informa- were then fixed in paraformaldehyde (PFA 4% in deionized water; Merck- tion). Sigma) and permeabilized with Proteinase K solution (20 μg mL−1 ; Merck- Cytokine Levels by ELISA: Harvested lungs were digested in 1 mL of Sigma) for 10 min. Next, sections were fixed again in PFA 4% and left RIPA buffer (Merck-Sigma) supplemented with EDTA-free protease in- in the equilibration buffer for 10 min. After equilibration, sections were hibitor (complete Mini, Roche), and homogenized with 5 mm stainless incubated with 50 μL of the TdT reaction mix, covered with plastic cov- steel beads using a TissueLyser LT system (Qiagen), operating at 50 Hz erslips, and placed in a humidified chamber at 37 °C away from light for for 10 min. Cell lysates were then centrifuged for 5 min at 2600 g (Het- 60 min. After reaction, sections were washed with the Stop Reaction buffer, tich GmbH), and supernatants were collected and stored at −80 °C until PBS, then stained with DAPI (Merck-Sigma), and mounted with coverslips analysis. After measurement of total protein contents using a BCA assay using mounting media. Imaging was performed with an epifluorescence (Pierce, ThermoFisherScientific), the concentrations of the cytokines IL- microscope (Axio Oberver, Zeiss) and the corresponding filter for detect- 1𝛼, IL-1b, IL-6, TNF-𝛼, MCP-1, GM-CSF, IL-17A, IL-23, IL-12p70, IFN-𝛾, ing fluorescein emission. The quantification of apoptosis was calculated IFN-𝛽, IL-27, and IL-10 were evaluated using the multiplex Mouse Inflam- based on the number of positive cells relative to the negative control. mation Panel (13-plex, v-plate, Biolegend) according to manufacturer’s in- Immunohistochemistry: Lung tissue remodeling and fibrosis were structions. Fluorescence intensities were measured using a BD FACSVerse evaluated in lung sections by immunohistochemical quantification of col- flow cytometer (BD Biosciences) and concentration values were extrapo- lagen I (rabbit anti-mouse collagen-I, #34710, Abcam). Briefly, the lung lated from standard curves. IgE and IL-4 concentrations were evaluated sections were first dewaxed with Histo-clear (HistoChoice clearing agent, using single ELISA kits (Biolegend). For each sample, cytokine concentra- Merck-Sigma), followed by rehydration in series of ethanol solutions and Adv. Sci. 2022, 9, 2104559 2104559 (13 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com water. Antigen retrieval was performed with citric acid bath (pH = 6) Acknowledgements and microwave heating for 10 min. After cooling down at room temper- ature, the sections were permeabilized with 0.01% Triton X-100 (Merck- A.F. and A.A would like to acknowledge studentships co-funded by the Sigma) in PBS (Merck-Sigma) for 10 min. Blocking was performed with Lloyds Register Foundation funded International Consortium on Nan- heat-inactivated normal donkey serum (10%, Merck-Sigma) in 1% BSA otechnology (LRF-ICON) and the UK Research and Innovation Engineer- (Gibco, ThermoFisherScientific) solution for 2 h. The sections were then ing and Physical Sciences Research Council (UKRI-EPSRC) Centre for Doc- gently drained with a pipette to remove the excess blocking buffer, and toral Training programme Graphene NOWNANO (EP/L01548X/1). This primary antibodies in PBS solution (Merck-Sigma) with 1% BSA (Merck- work was also supported by the European Union (EU) 8th Framework Sigma) were applied; sections were incubated overnight at 4 °C. The fol- Programme for Research and Technological Development, Graphene- lowing day, sections were washed with 0.01% Triton X-100 in PBS and Flagship project (H2020-FET-GrapheneCore2—#785219 and H2020-FET- then incubated at room temperature for 1 h with secondary antibodies GrapheneCore3—#881603). The authors also thank Dr. Adrian Esteban (donkey anti-rabbit Alexa fluor 555, #31572, ThermoFisherScientific) sus- Arranz, Dr. Miguel Arellano, Irene Rebollido, and Angeliki Karakasidi for pended in PBS solution with 1% BSA, away from direct light. After incuba- their contributions to the synthesis and characterization of the GO ma- tion, sections were gently washed in water, dried, and then mounted with terials used in the present study. Matthew Drummond and Abbie Dodd mounting medium (ProLong Gold Antifade Mountant with DAPI, Molecu- are acknowledged for their contributions to the tissue collection. In addi- lar Probes, ThermoFisherScientific) and glass coverslips. The fluorescence tion, the authors would like to acknowledge the following groups at ICN2 intensity of each section was measured by epifluorescence microscopy (Catalan Institute of Nanoscience and Nanotechnology, Barcelona, Spain): (Axio Oberver, Zeiss), and semi-quantify using ImageJ software (US Na- Advanced Electronic Materials and Devices led by Prof. Jose Garrido for tional Institutes of Health, Bethesda). the AFM and Raman instrumentation and assistance, and Supramolecular Evaluation of Clearance Using Raman Spectroscopy Imaging: Deparaf- NanoChemistry and Materials led by Prof. Daniel Maspoch for the UV–vis finized lung sections and cytospun BAL cells fixed with ice-cold methanol equipment and assistance. The ICN2 is funded by the CERCA programme, were scanned by Raman microscopy (XploRA Plus, HORIBA) using a laser Generalitat de Catalunya, and is supported by the Severo Ochoa Centres excitation wavelength of 638 nm and a grating of 600, with 3 μm of distance of Excellence programme by the Spanish Research Agency (AEI, grant no. between each point. GO was identified based on characteristic D (≈1340 SEV-2017-0706). At the Faculty of Biology, Medicine and Health, The Uni- cm−1 ) and G (≈1580 cm−1 ) bands (Figure S14, Supporting Information). versity of Manchester, UK, the authors would like to thank the staff from Only G bands were used to evaluate the interaction/internalization be- the Biological Services Facility, and the following staff members for expert cause less interferences were observed with biological tissues around 1580 advice and assistance in using core facility and their provision of techni- cm−1 . The internalization of GO sheets in neutrophils and macrophages cal support services: Peter Walker and Grace Bako (Histology Facility), Dr. was evaluated by overlapping the maps of GO intensities with the BAL Gareth Howell (Flow Cytometry Facility), Claire Morrison (Genomic Facil- cell bright-field images (the authors counted 30 to 100 neutrophils and ity), Dr. Peter March (Bioimaging Facility), and Dr. N. Hodson (Bio-AFM around 150 macrophages per sample; n = 3). Only interaction of GO with Facility). The Bioimaging Facility microscopes used in the present study neutrophils and macrophages was reported, as the authors did not ob- were purchased with grants from the UKRI Biotechnology and Biological serve GO sheets interacting with the other cells present in BAL fluids (i.e., Sciences Research Council (UKRI-BBSRC), the Wellcome Trust, and the eosinophils, lymphocytes). The number of GO-positive cells (interacting University of Manchester Strategic Fund. with GO) and GO-negative cells (not interacting with GO) was recorded. For each GO-positive cell, the relative quantity of material internalized was evaluated based on the number of GO-positive pixels counted, divided by the total cell surface area (in pixel), and expressed as average loading in Conflict of Interest GO-positive cells (%). The total loading (%) per cell type was then cal- culated by multiplying the percentage of GO-positive cells by the average The authors declare no conflict of interest. loading in GO-positive cells (%). Statistical Analysis: Data are expressed as mean ± standard deviation. Statistical analysis was performed using Graphpad Prism 8.0 (GraphPad Software Inc., San Diego, CA). For flow cytometry and TUNEL, two-way Author Contributions ANOVA followed by Dunnett’s post-hoc test was used to evaluate statis- T.L. and L.A.V.L. contributed equally to this work. T.L.: Conceptualiza- tical differences compared to the control (n = 3; * p < 0.05, ** p < 0.01, tion (lead); investigation (lead for biological analyses); formal analysis and *** p < 0.001). Two-way ANOVA followed by Dunnett’s post-hoc test (lead); methodology (lead); visualization (lead); writing—original draft was also used for the histopathological analysis to evaluate differences (lead); writing—review and editing (equal). L.A.V.L.: Conceptualization compared to the negative control or differences between different time- (lead); investigation (lead for biological analyses); formal analysis (lead); points (n = 3; p < 0.05:* , p < 0.01:** , p < 0.001:*** ). For BALF, gene expres- methodology (lead); visualization (lead); writing—original draft (lead); sion, protein concentration, and collagen deposition via immunostaining writing—review and editing (equal). A.F.: Investigation (supporting); or Masson Trichrome, one-way ANOVA followed by Dunnett’s post-hoc visualization (supporting); writing—review and editing (equal). A.A.: test or Kruskall–Wallis followed by Dunn’s post-hoc test, depending on Investigation (supporting). K.B.: Methodology (lead); software (lead). data normality, was used to evaluate significant differences compared to N.L.: Investigation (lead for materials); resources (lead for GO materials); the negative control at each time-point (n = 4–6; p < 0.05:* , p < 0.01:** , p < formal analysis (lead for GO materials); visualization (lead for materials); 0.001:*** ). For Raman analysis, at each time-point, t-test or Man–Whitney supervision (supporting); writing—review and editing (equal). K.K.: test, depending on data normality, were used to evaluate statistical dif- Conceptualization (lead); resources (lead for GO materials); funding ferences between the LGO and USGO (n = 3; p < 0.05:* , p < 0.01:** , acquisition (lead); supervision (lead); writing—review and editing (equal). p < 0.001:*** ). As responses to MWCNT exposure were extremely posi- C.B.: Conceptualization (lead); investigation (supporting); resources (lead tive for most endpoints tested, an additional comparison between LGO, for MWCNT materials); funding acquisition (lead); supervision (lead); USGO, MWCNTs, and the negative control was performed separately to visualization (supporting); writing—review and editing (equal). avoid false negative results. Supporting Information Data Availability Statement Supporting Information is available from the Wiley Online Library or from The data that support the findings of this study are available from the cor- the author. responding author upon reasonable request. Adv. Sci. 2022, 9, 2104559 2104559 (14 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Keywords [24] C. Bussy, H. Ali-Boucetta, K. Kostarelos, Acc. Chem. Res. 2013, 46, 692. adaptive immunity, chronic exposure, graphene oxide, innate immunity, [25] C. J. Bullock, C. Bussy, Adv. Mater. Interfaces 2019, 6, 1900229. lungs, nanomaterials [26] C. Martín, K. Kostarelos, M. Prato, A. Bianco, Chem. Commun. 2019, 55, 5540. Received: October 14, 2021 [27] M. C. Duch, G. R. S. Budinger, Y. T. Liang, S. Soberanes, D. Urich, S. Revised: December 21, 2021 E. Chiarella, L. A. Campochiaro, A. Gonzalez, N. S. Chandel, M. C. Published online: February 15, 2022 Hersam, G. M. Mutlu, Nano Lett. 2011, 11, 5201. [28] R. Li, L. M. Guiney, C. H. Chang, N. D. Mansukhani, Z. Ji, X. Wang, Y.-P. Liao, W. Jiang, B. Sun, M. C. Hersam, A. E. Nel, T. Xia, ACS Nano 2018, 12, 1390. [29] S. Bengtson, K. B. Knudsen, Z. O. Kyjovska, T. Berthing, V. Skaug, [1] Y. Qian, J. Song, X. Zhao, W. Chen, Y. Ouyang, W. Yuan, C. Fan, Adv. M. Levin, I. K. Koponen, A. Shivayogimath, T. J. Booth, B. Alonso, A. Sci. 2018, 5, 1700499. Pesquera, A. Zurutuza, B. L. Thomsen, J. T. Troelsen, N. R. Jacobsen, [2] A. E. Jakus, E. B. Secor, A. L. Rutz, S. W. Jordan, M. C. Hersam, R. N. U. Vogel, PLoS One 2017, 12, e0178355. Shah, ACS Nano 2015, 9, 4636. [30] Y. H. Kim, M. S. Jo, J. K. Kim, J. H. Shin, J. E. Baek, H. S. Park, H. J. [3] I. De Lázaro, S. Vranic, D. Marson, A. F. Rodrigues, M. Buggio, An, J. S. Lee, B. W. Kim, H. P. Kim, K. H. Ahn, K. Jeon, S. M. Oh, J. H. A. Esteban-Arranz, M. Mazza, P. Posocco, K. Kostarelos, Nanoscale Lee, T. Workman, E. M. Faustman, I. J. Yu, Nanotoxicology 2018, 12, 2019, 29, 13863. 224. [4] T. Yin, J. Liu, Z. Zhao, Y. Zhao, L. Dong, M. Yang, J. Zhou, M. Huo, [31] A. F. Rodrigues, L. Newman, D. Jasim, S. P. Mukherjee, J. Wang, I. Adv. Funct. Mater. 2017, 27, 1604620. A. Vacchi, C. Ménard-Moyon, A. Bianco, B. Fadeel, K. Kostarelos, C. [5] C. Xu, H. Hong, Y. Lee, K. S. Park, M. Sun, T. Wang, M. E. Aikins, Y. Bussy, Adv. Sci. 2020, 7, 1903200. Xu, J. J. Moon, ACS Nano 2020, 14, 13268. [32] A. F. Rodrigues, L. Newman, N. Lozano, S. P. Mukherjee, B. Fadeel, [6] M. De Sousa, L. A. Visani De Luna, L. C. Fonseca, S. Giorgio, O. L. C. Bussy, K. Kostarelos, 2D Mater. 2018, 5, 035020. Alves, ACS Appl. Nano Mater. 2018, 1, 922. [33] J. Nikota, A. Banville, L. R. Goodwin, D. Wu, A. Williams, C. L. Yauk, [7] L. Mao, M. Hu, B. Pan, Y. Xie, E. J. Petersen, Part. Fibre Toxicol. 2016, H. Wallin, U. Vogel, S. Halappanavar, Part. Fibre Toxicol. 2017, 14, 37. 13, 7. [34] A. J. Taylor-Just, M. D. Ihrie, K. S. Duke, H. Y. Lee, D. J. You, S. Hussain, [8] B. Li, J. Yang, Q. Huang, Y. Zhang, C. Peng, Y. Zhang, Y. He, J. Shi, W. V. K. Kodali, C. Ziemann, O. Creutzenberg, A. Vulpoi, F. Turcu, M. Li, J. Hu, C. Fan, NPG Asia Mater. 2013, 5, e44. Potara, M. Todea, S. Van Den Brule, D. Lison, J. C. Bonner, Part. Fibre [9] J. C. Bonner, Expert Rev. Respir. Med. 2011, 5, 779. Toxicol. 2020, 17, 60. [10] S. Azarmi, W. H. Roa, R. Löbenberg, Adv. Drug Delivery Rev. 2008, 60, [35] A. J. Byrne, S. A. Mathie, L. G. Gregory, C. M. Lloyd, Thorax 2015, 70, 863. 1189. [11] C. Dong, F. Wang, L. M. Shekhtman, M. M. Danziger, J. Fan, R. Du, [36] A. K. McNally, J. M. Anderson, Adv. Exp. Med. Biol. 2011, 713, 97. J. Liu, L. Tian, H. E. Stanley, S. Havlin, Proc. Natl. Acad. Sci. U. S. A. [37] Y. Okabe, R. Medzhitov, Nat. Immunol. 2016, 17, 9. 2021, 118, e1922831118. [38] T. Hussell, T. J. Bell, Nat. Rev. Immunol. 2014, 14, 81. [12] G. Oberdörster, V. Castranova, B. Asgharian, P. Sayre, J. Toxicol. Envi- [39] J. R. Roberts, R. R. Mercer, A. B. Stefaniak, M. S. Seehra, U. K. Ged- ron. Health, Part B 2015, 18, 121. dam, I. S. Chaudhuri, A. Kyrlidis, V. K. Kodali, T. Sager, A. Kenyon, S. A. [13] L. A. Mitchell, J. Gao, R. V. Wal, A. Gigliotti, S. W. Burchiel, J. D. Mc- Bilgesu, T. Eye, J. F. Scabilloni, S. S. Leonard, N. R. Fix, D. Schwegler- donald, Toxicol. Sci. 2007, 100, 203. Berry, B. Y. Farris, M. G. Wolfarth, D. W. Porter, V. Castranova, A. [14] J. Dong, Q. Ma, Front. Immunol. 2018, 9, 1120. Erdely, Part. Fibre Toxicol. 2016, 13, 34. [15] A. A. Shvedova, E. R. Kisin, R. Mercer, A. R. Murray, V. J. Johnson, [40] M. J. Feito, R. Diez-Orejas, M. Cicuéndez, L. Casarrubios, J. M. Rojo, A. I. Potapovich, Y. Y. Tyurina, O. Gorelik, S. Arepalli, D. Schwegler- M. T. Portolés, Colloids Surf., B 2019, 176, 96. Berry, A. F. Hubbs, J. Antonini, D. E. Evans, B.-K. Ku, D. Ramsey, A. [41] B.-C. Lee, J. Y. Lee, J. Kim, J. M. Yoo, I. Kang, J.-J. Kim, N. Shin, D. J. Maynard, V. E. Kagan, V. Castranova, P. Baron, Am. J. Physiol.: Lung Kim, S. W. Choi, D. Kim, B. H. Hong, K.-S. Kang, Sci. Adv. 2020, 6, Cell. Mol. Physiol. 2005, 289, L698. eaaz2630. [16] A. A. Shvedova, E. Kisin, A. R. Murray, V. J. Johnson, O. Gorelik, S. [42] L. W. Locke, E. D. Crouser, P. White, M. W. Julian, E. G. Caceres, A. Arepalli, A. F. Hubbs, R. R. Mercer, P. Keohavong, N. Sussman, J. Jin, C. Papp, V. T. Le, W. Sadee, L. S. Schlesinger, Am. J. Respir. Cell Mol. J. Yin, S. Stone, B. T. Chen, G. Deye, A. Maynard, V. Castranova, P. A. Biol. 2019, 60, 84. Baron, V. E. Kagan, Am. J. Physiol.: Lung Cell. Mol. Physiol. 2008, 295, [43] I. Huizar, A. Malur, Y. A. Midgette, C. Kukoly, P. Chen, P. C. Ke, R. L552. Podila, A. M. Rao, C. J. Wingard, L. Dobbs, B. P. Barna, M. S. Kavuru, [17] R. R. Mercer, A. F. Hubbs, J. F. Scabilloni, L. Wang, L. A. Battelli, S. M. J. Thomassen, Am. J. Respir. Cell Mol. Biol. 2011, 45, 858. Friend, V. Castranova, D. W. Porter, Part. Fibre Toxicol. 2011, 8, 21. [44] S. Mukhopadhyay, J. R. Hoidal, T. K. Mukherjee, Respir. Res. 2006, 7, [18] C.-W. Lam, Toxicol. Sci. 2004, 77, 126. 125. [19] T. Kasai, Y. Umeda, M. Ohnishi, T. Mine, H. Kondo, T. Takeuchi, M. [45] N. I. Lorè, A. Bragonzi, C. Cigana, Cytokine Growth Factor Rev. 2016, Matsumoto, S. Fukushima, Part. Fibre Toxicol. 2016, 13, 53. 30, 19. [20] K. Fujita, M. Fukuda, S. Endoh, J. Maru, H. Kato, A. Nakamura, N. [46] R. L. Gieseck, M. S. Wilson, T. A. Wynn, Nat. Rev. Immunol. 2018, 18, Shinohara, K. Uchino, K. Honda, Inhal. Toxicol. 2015, 27, 207. 62. [21] R. R. Mercer, J. F. Scabilloni, A. F. Hubbs, L. A. Battelli, W. Mckinney, [47] T. A. Wynn, K. M. Vannella, Immunity 2016, 44, 450. S. Friend, M. G. Wolfarth, M. Andrew, V. Castranova, D. W. Porter, [48] M. Merad, P. Sathe, J. Helft, J. Miller, A. Mortha, Annu. Rev. Immunol. Part. Fibre Toxicol. 2013, 10, 33. 2013, 31, 563. [22] K. Donaldson, C. A. Poland, F. A. Murphy, M. Macfarlane, T. Cher- [49] E. K. Persson, K. Verstraete, I. Heyndrickx, E. Gevaert, H. Aegerter, nova, A. Schinwald, Adv. Drug Delivery Rev. 2013, 65, 2078. J.-M. Percier, K. Deswarte, K. H. G. Verschueren, A. Dansercoer, D. [23] K. B. Knudsen, T. Berthing, P. Jackson, S. S. Poulsen, A. Mortensen, N. Gras, P. Chanez, C. Bachert, A. Gonçalves, H. Van Gorp, H. De Haard, R. Jacobsen, V. Skaug, J. Szarek, K. S. Hougaard, H. Wolff, H. Wallin, C. Blanchetot, M. Saunders, H. Hammad, S. N. Savvides, B. N. Lam- U. Vogel, Basic Clin. Pharmacol. Toxicol. 2019, 124, 211. brecht, Science 2019, 364, eaaw4295. Adv. Sci. 2022, 9, 2104559 2104559 (15 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com [50] N. Rout-Pitt, N. Farrow, D. Parsons, M. Donnelley, Respir. Res. 2018, [66] A. Ambrosi, C. K. Chua, B. Khezri, Z. Sofer, R. D. Webster, M. Pumera, 19, 136. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 12899. [51] K. S. Duke, J. C. Bonner, Wiley Interdiscip. Rev.: Nanomed. [67] D. Elgrabli, W. Dachraoui, C. Ménard-Moyon, X. J. Liu, D. Bégin, S. Nanobiotechnol. 2018, 10, e1498. Bégin-Colin, A. Bianco, F. Gazeau, D. Alloyeau, ACS Nano 2015, 9, [52] G. Liu, A. M. Philp, T. Corte, M. A. Travis, H. Schilter, N. G. Hansbro, 10113. C. J. Burns, M. S. Eapen, S. S. Sohal, J. K. Burgess, P. M. Hansbro, [68] R. Kurapati, S. P. Mukherjee, C. Martín, G. Bepete, E. Vázquez, A. Pharmacol. Ther. 2021, 225, 107839. Pénicaud, B. Fadeel, A. Bianco, Angew. Chem., Int. Ed. 2018, 57, [53] A. J. Pagán, L. Ramakrishnan, Annu. Rev. Immunol. 2018, 36, 639. 11722. [54] Y. Shen, C. Pang, Y. Wu, D. Li, C. Wan, Z. Liao, T. Yang, L. Chen, F. [69] C. Ge, Y. Li, J.-J. Yin, Y. Liu, L. Wang, Y. Zhao, C. Chen, NPG Asia Mater. Wen, EBioMedicine 2016, 8, 302. 2012, 4, e32. [55] Q. Ma, Front. Immunol. 2020, 11, 1060. [70] E. Aldieri, I. Fenoglio, F. Cesano, E. Gazzano, G. Gulino, D. Scarano, [56] J. J. Moon, B. Huang, D. J. Irvine, Adv. Mater. 2012, 24, 3724. A. Attanasio, G. Mazzucco, D. Ghigo, B. Fubini, J. Toxicol. Environ. [57] D. W. Porter, A. F. Hubbs, R. R. Mercer, N. Wu, M. G. Wolfarth, K. Sri- Health, Part A 2013, 76, 1056. ram, S. Leonard, L. Battelli, D. Schwegler-Berry, S. Friend, Toxicology [71] R. M. Silva, K. Doudrick, L. M. Franzi, C. Teesy, D. S. Anderson, 2010, 269, 136. Z. Wu, S. Mitra, V. Vu, G. Dutrow, J. E. Evans, P. Westerhoff, L. [58] A. Bianco, H.-M. Cheng, T. Enoki, Y. Gogotsi, R. H. Hurt, N. Koratkar, S. Van Winkle, O. G. Raabe, K. E. Pinkerton, ACS Nano 2014, 8, T. Kyotani, M. Monthioux, C. R. Park, J. M. D. Tascon, J. Zhang, Carbon 8911. 2013, 65, 1. [72] D.-K. Lee, S. Jeon, J. Jeong, I. J. Yu, K. S. Song, A. Kang, W. S. Yun, J. [59] S. Vranic, A. F. Rodrigues, M. Buggio, L. Newman, M. R. H. White, D. S. Kim, W.-S. Cho, Nanomaterials 2020, 10, 379. G. Spiller, C. Bussy, K. Kostarelos, ACS Nano 2018, 12, 1373. [73] S. Søs Poulsen, N. R. Jacobsen, S. Labib, D. Wu, M. Husain, A. [60] C. Købler, S. S. Poulsen, A. T. Saber, N. R. Jacobsen, H. Wallin, C. L. Williams, J. P. Bøgelund, O. Andersen, C. Købler, K. Mølhave, Z. O. Yauk, S. Halappanavar, U. Vogel, K. Qvortrup, K. Mølhave, PLoS One Kyjovska, A. T. Saber, H. Wallin, C. L. Yauk, U. Vogel, S. Halappanavar, 2015, 10, e0116481. PLoS One 2013, 8, e80452. [61] W. Zhu, A. Von Dem Bussche, X. Yi, Y. Qiu, Z. Wang, P. Weston, R. [74] S. Keshavan, P. Calligari, L. Stella, L. Fusco, L. G. Delogu, B. Fadeel, H. Hurt, A. B. Kane, H. Gao, Proc. Natl. Acad. Sci. U. S. A. 2016, 113, Cell Death Dis. 2019, 10, 569. 12374. [75] S. De Oliveira, E. E. Rosowski, A. Huttenlocher, Nat. Rev. Immunol. [62] J. E. Rayahin, R. A. Gemeinhart, in Macrophages, Springer, Berlin 2016, 16, 378. 2017, pp. 317–351. [76] N. Doshi, A. J. Swiston, J. B. Gilbert, M. L. Alcaraz, R. E. [63] V. M. Arlt, A. M. Krais, R. W. Godschalk, Y. Riffo-Vasquez, I. Mri- Cohen, M. F. Rubner, S. Mitragotri, Adv. Mater. 2011, 23, zova, C. A. Roufosse, C. Corbin, Q. Shi, E. Frei, M. Stiborova, F.-J. H105. Van Schooten, D. H. Phillips, D. Spina, Toxicol. Sci. 2015, 146, 213. [77] P. Decuzzi, M. Ferrari, Biophys. J. 2008, 94, 3790. [64] R. Zangar, Toxicol. Appl. Pharmacol. 2004, 199, 316. [78] S. P. Mukherjee, N. Lozano, M. Kucki, A. E. Del Rio-Castillo, L. New- [65] K. A. Roach, A. B. Stefaniak, J. R. Roberts, J. Immunotoxicol. 2019, 16, man, E. Vázquez, K. Kostarelos, P. Wick, B. Fadeel, PLoS One 2016, 87. 11, e0166816. Adv. Sci. 2022, 9, 2104559 2104559 (16 of 16) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH