First-in-human controlled inhalation of thin graphene oxide nanosheets to study acute cardiorespiratory responses

Graphene oxide nanomaterials are being developed for wide-ranging applications but are associated with potential safety concerns for human health. We conducted a double-blind randomized controlled study to determine how the inhalation of graphene oxide nanosheets affects acute pulmonary and cardiovascular function. Small and ultrasmall graphene oxide nanosheets at a concentration of 200 μg m−3 or filtered air were inhaled for 2 h by 14 young healthy volunteers in repeated visits. Overall, graphene oxide nanosheet exposure was well tolerated with no adverse effects. Heart rate, blood pressure, lung function and inflammatory markers were unaffected irrespective of graphene oxide particle size. Highly enriched blood proteomics analysis revealed very few differential plasma proteins and thrombus formation was mildly increased in an ex vivo model of arterial injury. Overall, acute inhalation of highly purified and thin nanometre-sized graphene oxide nanosheets was not associated with overt detrimental effects in healthy humans. These findings demonstrate the feasibility of carefully controlled human exposures at a clinical setting for risk assessment of graphene oxide, and lay the foundations for investigating the effects of other two-dimensional nanomaterials in humans. Clinicaltrials.gov ref: NCT03659864.

were further processed using ImageJ; the lateral size of the graphene oxide flakes was manually measured by determining the longest Feret diameter in each flake, n being the total number of flakes analyzed.
Hydrodynamic diameter and surface charge (zeta potential) measurements.The hydrodynamic diameter and the zeta potential values of graphene oxide suspensions in Milli-Q water were measured with a ZetaSizer Nano ZS instrument (Malvern, UK).The results are reported as the average ± standard deviation of three measurements per sample.
UV-visible spectroscopy (UV-vis).Spectra of graphene oxide dilutions in Milli-Q water with concentrations ranging from 2.5 to 20 μg/mL were acquired using a Cary 50 Bio UV-vis spectrophotometer (Varian Inc., Agilent Technologies, UK). Measurements were performed at room temperature in a quartz cuvette (1 mL volume, 1 cm path length).Milli-Q water was used as a blank.
Fluorescence spectroscopy.Different concentrations of graphene oxide dispersions (25-200 μg/mL) were measured with a Cary Eclipse fluorescence spectrophotometer (Varian Inc., Agilent Technologies, UK). Spectra were acquired at room temperature, with λexc set to 525 nm.Milli-Q water was used as a blank.
Raman spectroscopy.Measurements were recorded by a DXR micro-Raman spectrometer (Thermo Fisher Scientific, UK) equipped with a 633 nm laser set to 0.4 mW.Calibration was performed on a polystyrene standard, the chosen objective was 50x, and the pinhole was set to 50 µm.Spectra were then recorded between 500 and 4000 cm -1 with a resolution of 2.5 cm -1 .All spectra were processed by background subtraction and normalisation by the G band intensity using OriginPro 8.5.1 software.
Fourier transform infrared spectroscopy (FTIR).Fourier transform infrared spectra were obtained with a Tensor 27 spectrometer (Bruker, UK), equipped with a 3000 Series High Stability Temperature Controller with RS232 Control (Specac, UK) and a MKII Golden Gate Single Reflection ATR system (Specac, UK) for measurements in ATR mode.The bottom plate of the Golden Gate ATR system was pre-heated at 60°C to allow the complete evaporation of water from the drops (typically 20 µL) of the original graphene oxide dispersions.Approximately 3 min after depositing the dispersions on the plate, the transmittance spectra of graphene oxide were recorded by acquiring 32 scans in the 4000-750 cm -1 range, with a resolution of 4 cm -1 .
Thermogravimetric analysis (TGA).The oxidation degree of graphene oxide materials was extracted from the degradation patterns measured with a TGA 4000 thermogravimetric analyser (PerkinElmer Ltd, UK).All measurements were carried out on 2 mg lyophilised material, in a nitrogen atmosphere (20 mL/min) at temperatures ranging from 25 to 900°C (10°C/min).

X-ray photoelectron spectroscopy (XPS)
. XPS measurements of lyophilised graphene oxide samples were analysed using a Thermo Theta Probe XPS spectrometer with a monochromatic Al K-α source of 1486.68 eV.The spectra were acquired with PE of 40 kV, 0.1 eV step size and an average of 20 scans.
CasaXPS software (Casa Software Ltd, UK) was used for post-processing of spectra.The contribution of charge injected to insulating samples was corrected by calibrating all peaks according to the adventitious carbon C1s spectral component, set a binding energy of 284.6eV 3 .A Shirley background subtraction was applied to all spectra and Gaussian-Lorentzian (70:30) functions were used for fitting the functional groups, except for the asymmetric C-C and C=C peak, which was fitted using an asymmetric Lorentzian function.The full width half maximum (FWHM) value was constrained between 0.5 and 2 eV for all peaks, except for the π-π*.The following constrain regions were set for the binding energies: 284-285.secreted by bone marrow-derived macrophages exposed to graphene oxide was similar to control untreated cells using previously published method 4 ).S-graphene oxide was diluted to 1.3 mg/mL in sterile saline in aseptic conditions, aliquoted and stored at 4 o C until use.
Graphene oxide nanosheets (1.3 mg/mL) were aerosolized using a Schlick (Dusen-Schlick, model 970/S Untersiemau, Germany) compressed air nebulizer.Using two syringe pumps (TSE Systems, model 540200, Germany) the suspension was in-line diluted with high-performance liquid chromatography grade water and fed to the Schlick nebulizer.The suspension was transferred to a 5 mL syringe which was placed on the syringe pump and connected to the nebulizer.The compressed preheated (60 o C) airflow of the Schlick nebulizer was 12l pm.The aerosol was dried in a heated mixing glass tube (90 mm ID, length 550 mm), then diluted with high efficiency particulate air-filtered room air to the desired concentration, humidified to 50-60% relative humidity using an ultrasonic nebulizer (Omron Ultrasonic Nebulizer NE-U12, Japan).The aerosol was fed into a 200 L mixing chamber and delivered to the volunteer by an exposure mask placed over the mouth and nose.Temperature was kept constant throughout and relative humidity of exposure air was 50% maintained using fresh ultrapure water injected into a small side stream using an Omron ultrasonic nebulizer.Graphene oxides were delivered at an exposure concentration between 100 -300 µg/m 3 , with a target average concentration of 200 µg/m 3 (See Figure S1).This dose range was chosen based on our previous controlled exposure studies with dilute diesel exhaust, which were associated with impairment of a range of cardiovascular parameters without adverse effects 5,6 and with carbon and gold nanoparticles which did not alter cardiovascular parameters 7,8 .The concentration could be adjusted by altering the speed of the syringe pump delivering the suspension.The real-time mass concentration was measured by a tapered element oscillating microbalance (TEOM; Thermo Scientific, model 1400A, USA) as a guide for changing the speed of the pump.The concentration of the particles in the exposure was monitored and maintained by the exposure technicians.
Particle concentration in the aerosol was taken from the middle of the 200 L mixing chamber by a SS-tube.The particle characteristics measured were: particle mass (

Participants and eligibility criteria
Fifteen healthy volunteers were recruited by advertising the study by posters and e-mails in the hospital and university campus, as approved by local ethical review.The data from 14 subjects were included as one subject was unable to complete the exposure visits in the time-frame of the study.The target of 15 individuals was based on our previous controlled exposure studies with air pollutants based on changes to vascular reactivity and inflammatory cytokines in the blood, based on diesel exhaust exposure as there is no other controlled exposure study of a two dimensional material for comparisons.A 1-h exposure to diluted diesel exhaust produced an ~32% reduction in forearm bloodflow to 1 nmol/min bradykinin (~16±2 vs ~19±2.5 mL/100 mL tissue/min (±SD) for diesel exhaust vs filtered air control, respectively 9 .A 2-h exposure to diluted diesel exhaust produced a 12.5% increase of plasma TNF-α (0.99±0.07 vs 0.88±0.007pg/mL (±SD) for diesel exhaust vs filtered air, respectively 10 .Based on these figures, 12 and 10 volunteers, respectively, would be needed to detect these changes with significant of P<0.05 with an 80% power.Because no other study has tested the effects of an inhaled 2D material, as an additional precautionary step, the decision was taken not to increased group sizes beyond 15 for this study.
Interested volunteers were provided with a participant information sheet which they were asked to read and consider for at least 24 h before agreeing to be involved in the study.For study visits, participants abstained from alcohol for 24 h and from food and caffeine containing beverages for at least 12 h prior to the study visit.Participants were invited for an initial screening visit to ensure that they met the inclusion criteria (Table S1).Exclusion criteria included major or traumatic surgery within 12 weeks of screening, a history of and smoking or asthma, occupation with high exposure to air pollution or other inhaled irritant, acute respiratory illness within 3 weeks of enrolment, use of aspirin or antiinflammatory medication or vitamin and herbal supplements for the week prior to their study visit.
Women who were pregnant, lactating or taking contraceptive medication were also excluded from the study.We did not ask participants to wear a facemask outside of the study visits, as low compliance would have added an additional source of variability between participants (the study was run prior to the coronavirus pandemic before mask wearing became common in the UK).Additionally, even occupational facemasks have been shown to vary greatly in their removal of inhaled particles during different modes of activity 11 .Importantly, each volunteer acts as their own control and receives each exposure in a random order, minimising variation from both intrinsic biology and life-style factors.

Study Design
See Main Manuscript Figure 2. A screening visit was used to confirm eligibility criteria with the participant, followed by taking written consent and assignment of a participant code.Height, weight, heart rate, blood pressure and lung function were measured, and a 3-mL blood sample was taken for a full blood cell count.If parameters were within the normal range for young healthy individuals, participants were taken forward to full study days.Additionally, a graded cardio-respiratory exercise stress test on a bicycle ergonometer was performed to determine the workload required to generate a ventilation rate of 25 L/min/m 2 .
Two lateral dimensions of GO (maintaining all other physicochemical characteristics almost identical) were selected for the study: 'small' GO (s-GO) and 'ultrasmall' GO (us-GO).Both types of nanosheets have demonstrated neither acute, nor longitudinal adverse effects in our previous pre-clinical (rodent) studies 2 , contrary to 'large' GO sheets that were thus excluded from this work as a safety precaution.A double-blind randomised crossover study design was used for the study visits, whereby the order of exposures (filtered air, s-graphene oxide, us-graphene oxide) were randomised.All study visits were organised at least 2 weeks apart to allow a washout period between different exposures.The volunteer and clinician performing the study were blinded to the identity of the exposure group.All researchers involved with collating and analysing the raw data were blinded to the exposure group, with unblinding occurring only when ready for grouping by exposure.
Prior to exposures (t=0), heart rate, blood pressure and lung function were measured, and blood taken.Participants were asked to empty their bladders and then given a urine container to collect any urine over the course of the study visit.Participants were then taken to the exposure laboratory based at the Royal Infirmary of Edinburgh site for the duration of the study.An experienced research clinician and exposure technician were present throughout the exposure, with the same researcher and nursing support present during the rest of the protocol.
In the exposure laboratory (Figure S2), participants wore a face mask through which nanoparticles could be delivered by inhalation.Volunteers were asked to cycle at the workload required to increase respiratory rate to 25 L/min/m 2 (pre-determined by exercise testing at the screening visit) and rest alternatively for 15-min periods across the 2-h exposure.After exposure, the subject returned to the Clinical Research Facility for assessment of biological parameters.
Vital signs, lung function and blood collected pre-exposure (t=0), were repeated at t=2.25, 4 and at 6 h (ie 15 min, 2 h and 4 h after exposure).For ease of reading, the 2.25-h time point is referred to as t=2 throughout the manuscript).The ex vivo model of deep arterial injury was performed at 1-1.5 h post exposure, and forearm plethysmography performed at 2-4 h post-exposure (see below).A light lunch was provided that was identical for all volunteers and all study visits.As an additional safety measure, a shortened protocol (without the ex vivo thrombosis and plethysmography or 4-h measurements) was performed for first exposure of each group.The study visits for the subsequent volunteers with the full protocol were scheduled only after it was confirmed that there were no adverse events and no marked changes in blood biomarkers.Volunteers were compensated for their time and travel expenses, which was approved by the ethics committee.

Lung function and vital signs
The participants were asked to rest in a sitting position for 15 min prior to measurement of vital signs and lung function.Lung function was measured by spirometry (Vitalograph Alpha III, UK), with the optimal breathing techniques that were demonstrated at the screening visit.Forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) were then measured, and a mean of two closely concurring consecutive runs were used.The participants were allowed to rest for a further 5 min, prior to measurement of blood pressure and heart rate by sphygmomanometry.

Vascular function
The clinical protocol was designed to include measurement of vascular function by venous occlusion plethysmography between t =4 and t=6 h.However, due to technical and staffing difficulties we were unable to obtain reliable data from sufficient volunteers to make meaningful conclusions, thus the data was omitted.Data analysis was performed after the collection of all study visits, thus the technique was a part of the protocol for all study visits, and subsequently the methods are outlined below.Blood flow was determined using mercury-in-silastic strain gauges placed around each forearm, as previously

Blood biomarkers
Blood was sampled before nanoparticle exposure (t=0) and at 2.25, 4 and 6 h.A 17-gauge cannula was inserted into a large antecubital vein of both arms, and flushed with sterile saline.First, 1 mL of blood was discarded and approximately 27 mL was then collected for analysis.EDTA-treated blood was used for measurement blood cell differentials, citrate-treated blood was used for coagulation markers (activated partial thromboplastin time, prothrombin time, fibrinogen) and clotted blood was used to collect serum for C-reactive protein (CRP) and cytokines (IL-6, TNFα).Blood measures were performed by the Clinical Biochemistry Unit at the NHS Royal Infirmary of Edinburgh by standard methodology.Cytokines were measured using ELISA (R&D Systems, UK), with limits of detection of 0.022 pg/mL for TNFα and 0.031 pg/mL for IL-6.Details of -omics analysis of blood factors are described below.Subsamples of blood and urine were frozen at -80 o C for biobanking.

Ex vivo thrombosis
The coagulability of blood was measured ex vivo using a model of thrombosis on deep arterial injury (Main Manuscript Figure 6).We have used this technique extensively in our clinical studies following exposure of volunteers to diesel exhaust 6,12 and testing of antithrombotic medication 13,14 .Blood was withdrawn from an antecubital vein via a pump set at a flow rate of 10 mL/min.The first 5 mL of blood was discarded, before the cannula was connected, using non-coagulation tubing (Masterflex Tygon, Cole Parmer, UK), to three sequential cylindrical perfusion chambers maintained at 37°C in a water bath.Strips of porcine aorta (Pel-freez, USA) were prepared by carefully removing the intima and a thin layer of media to act as a thrombogenic substrate, and mounted in the chamber according to physiological direction of blood flow.The rheological conditions in the first chamber simulate those of patent coronary arteries (low-shear rate, ∼212/s), whereas those in the second and third chambers simulate those of mildly stenosed coronary arteries (high-shear rate, ∼1690/s).The model thus acts as one of deep coronary arterial injury.Each chamber run lasted for 5 min after which saline was perfused over the strip to remove non-adherent blood.The porcine strips with thrombus attached were removed and fixed in 4% paraformaldehyde.Strips were cut into 8 cross-sections, wax-embedded, histologically sectioned and endogenous peroxidase activity was blocked using 3% hydrogen peroxide solution (Leica Microsystems GmbH, Wetzlar, Germany) for 5 min.Sections were then incubated at room temperature for 1-h with polyclonal rabbit anti-human fibrin(ogen) antibody (1.2 μg/mL, Dako, Glostrup, Denmark; Cat. No. A0080) and monoclonal mouse anti-human CD61 antibody (1.28 μg/mL, Dako; Cat.No.
A semi-automated slide scanner (Axioscan Z1; Zeiss, Jena, Germany) and image analysis software (QuPath 0.2.3) were used by a blinded researcher to quantify thrombus area.Digital images of each section were acquired at ×20 magnification.High-resolution classifiers based on colour were established to detect total thrombus area.Two blinded researchers quality assessed images and sections were discarded if there was evidence of poor vascular integrity, thrombosis forming at a disrupted surface layer, or that the thrombus was dislodged from the arterial strip.Strips with less than three sections per arterial strip were discarded.

High-fidelity nano-proteomics analysis of plasma samples.
Preparation of liposomal nanoparticles and enrichment of plasma proteins.HSPC:Chol:DSPE-PEG2000 (56.3:38.2:5.5)liposomes were prepared by thin lipid film hydration followed by extrusion, as previously described 15 .Lipids were dissolved in chloroform:methanol (4:1) and evaporated (150 rotations/min for 1 h under vacuum, 40°C) using a rotary evaporator (Buchi, Switzerland).Lipid films were hydrated at 60°C with ammonium sulphate (250 mM, pH 8.5) to produce large multilammer liposomes.Small unilamellar liposomes were then produced by extrusion through 800 nm and 200 nm polycarbonate filters (Whatman, UK) for 10 times each, and then 15 times through 100 nm and 80 nm extrusion filters (Whatman, UK) using a mini-Extruder (Avanti Polar Lipids, USA).
Proteins bound to the liposome nanoparticles were quantified by BCA Protein assay kit according to the manufacturer's instructions.
Eicosanoids were separated on a Hypersil GOLD C18 column (1.9 µm; 100 x 2.1 mm) (Thermo, UK) using a Shimadzu Nexera-X2 UHPLC system.The initial gradient conditions for analysis were 55% mobile phase A -45% mobile phase B. The percentage of mobile phase B was increased from 45% to 60% over 10 min, followed by 60% to 70% over 1 min, a linear increase to 100% between 11-18 min, held for 2 min before re-equilibration to the starting conditions over 5 min.The flow rate was 400 μL/min.The LC effluent was directed into an IonTurbo source of a Sciex QTRAP 6500 mass spectrometer.The instrument was operated in negative ion mode using the multiple reaction monitoring.
Eicosanoids were identified on the basis of their characteristic precursor/product ion pair transitions and matching retention time with authentic standards.Data were acquired and analysed using Sciex Analyst software v1.6.Concentrations of eicosanoids were determined by comparison to a calibration curve run in parallel for each compound and adjusted for recovery by reference to amounts of the appropriate internal standards.

General data and statistical analysis
Data were analysed using Excel 2010 (Microsoft, USA), R 3.2.2(R Foundation for Statistical Computing, Austria) and Prism 9.3 (Graphpad, USA).Data in table are presented as mean ± standard deviation, unless otherwise indicated.Continuous data are presented as means and standard deviation and statistical significance within groups and between groups were tested using two-way analysis of variance (ANOVA) with Tukey's Honest Significant Difference post-hoc test.Parametric assumptions (normal distribution and equal variances) were confirmed using the statistical packages above; where data was not normally distributed a non-parametric alternative (e.g.Kruskal-Wallis test) was used.
extrapolation from murine models to predict the pulmonary effects of longer-term exposure in humans.
Interestingly, a mouse model of asthma has demonstrated that GO can sensitise airway responsiveness to the agonist methacholine.However, there was no direct effect of GO at the time of the challenge, and responses did not align with markers of pulmonary inflammation 33 .Asthmatic individuals were excluded from our study and represent a potentially susceptible group for future investigations.

Platelet numbers and coagulation markers
There were no differences between exposure groups for any of the blood platelet numbers or coagulation markers reported in Extended Data Table ED4.GO can directly interact with blood to induce conformational changes to fibrinogen and activate complement (C3a, C5a) and intrinsic coagulation (prothrombin) pathways 34 .Direct exposure of platelets to GO can activate platelets and increase the occurrence of thrombus in pulmonary vessels after intravenous injection 35 .However, these effects were dependent on the degree of GO surface functionalisation, while other forms of graphene have minimal direct effects on haemolysis, platelet activation, prothrombin time and activated partial thromboplastin time 36 .It should be borne in mind that these experiments were performed with high concentrations of GO (>0.05 mg/mL) that are orders of magnitude higher than what would be expected to translocate to the circulation after inhalation 8,37 .Accordingly, five days inhalation of GO in rats also showed no effect on these measures of coagulation 23 .

Markers of inflammation in blood
GO can activate macrophages through internalisation and induction of complex cellular signalling mechanisms, which are dependent on the dimensions of GO.Smaller nanosheets may have a greater capacity to generate intracellular reactive oxygen species, whereas larger GO can induce necrosis and apoptosis through physical interactions with the cell membrane 38 .In the present study, there was no significant difference between exposure groups for any of the blood inflammatory cells counts or inflammatory cell biomarkers reported in Extended Data Table ED5.
Controlled exposure to diesel exhaust emissions induces a mild increase in blood neutrophils, IL-6 and TNFa 10,19,39 , although alterations in these biomarkers have not been consistent between studies.Levels of inflammatory cytokines may have been greater at time points after 6-h, although there was inconsistency across markers of inflammation 24-h after diesel exhaust exposure 10 .Markers of oxidative and genotoxic effects have been shown to be increased, and reduced six months after the installation of workplace filters, in a cohort of six workers in a graphene manufacturing facility 40 .The authors note that it was not possible to discriminate whether these effects could be attributed to nanomaterials or other chemical exposures, but these biomarkers may be suitable for biomonitoring (see also 41 ).We have previously demonstrated that spark-generated carbon black particles are unable to induce a systemic inflammatory response in healthy volunteers 7 .This is also in keeping with the available pre-clinical evidence for graphene materials, where pristine graphene did not directly release cytokines from peripheral blood monocytes 36 .Furthermore, inhalation of GO in mice induced only mild increases in circulatory inflammatory cells or cytokines 22,23 at high doses (>3 mg/m 3 ) that are likely to be associated with lung overload and do not extrapolate to anticipated real-life exposure scenarios in humans 42,43 .Previous studies have demonstrated that GO can induce an acute phase response in the liver of GO instilled mice 44 .The lack of effect of GO on C-reactive protein in the current study suggests that there was insufficient translocation of GO to the liver, or that the purity of GO may minimise the acute phase response, although the profile of the response at later time points remains to be confirmed.Studies making a direct comparison of our high purity materials and commercial sources of GO that are typically less pure and more heterogenous in their size distribution would be valuable.

Targeted lipidomics to identify the effects of graphene oxide on eicosanoids
Eicosanoids represent a group of diverse mediators formed from the polyunsaturated fatty acid, arachidonic acid.Prostaglandins, leukotrienes, hydroxy-eicosatraenoic acids (HETEs) and epoxyeicosatrienoic/dihydroeicosaatrienoic acids (EETs/dHETs), formed via enzymatic activation of arachidonic acid, are recognised to be important mediators in the onset and progression of inflammation.
The free radical-mediated peroxidation of arachidonic acid leads to the production of isoprostanes, which are biomarkers of oxidative stress that is a hallmark of particle-induced cellular dysfunction.
Several studies have found that exposure to particulate air pollution in China is associated with inceases in a variety of proinflammatory eicosanoid levels in the blood of humans [45][46][47] .In mice, inhalation of nanoparticle-rich diesel exhaust has been shown to increase several eicosanoids (HETEs, HODEs, isoprostanes) in the lung lining fluid, plasma, liver and intestines; DE caused oxidative stress and dysfunction of anti-oxidant/anti-inflammatory high-density lipoprotein in the same model 48 .Although very few studies have investigated the effect of graphene materials on eicosanoids, graphene nanoplatelets altered arachidonic metabolism in a macrophage cell line at non-cytotoxic concentrations 49 , and low levels of GO modified arachidonic acid and eicosanoids in the brains of zebrafish.
Blood from a subset of participants (n=3) was used to collect preliminary data for eicosanoid profiling.Thirty five out of the fifty five eicosanoid species were detected in the plasma of volunteers, with eighteen species showing a significant difference between the graphene oxides and air (Extended Data Figure ED2 and Table ED6) prior to correction for multiple testing.Both s-graphene oxide and us-graphene oxide increased levels of several dHETs and HETEs, whereas greater levels of arachidonic acid, eicosapentaenoic acid and DHA were found after exposure to s-graphene oxide, but not usgraphene oxide.However, after adjusting for multiple comparisons, only six eicosanoids were significantly different from the air group with a p≤0.001: 14,15_dHET, arachidonic acid (AA) and docosahexaenoic acid (DHA) were greater for s-GO; 10_carboxy_LTB4, 5,6_dHET and 14_HDHA were greater for us-GO.14,15-dHET has been reported to impair neutrophil function 50 .Furthermore, dHETs are metabolites of EETs which have a variety of roles in the lung and cardiovascular system, including vasodilatation, and anti-thombotic and anti-inflammatory properties 51 .14,15-EET has been shown to provide some protection against cigarette smoke-induced lung injury 52 .However, there was no measurable change in 14,15-EET levels in response to GO suggesting that GO did not have overt effects on EET metabolism.Increased arachidonic acid formation would suggest that s-GO stimulates the activation of phospholipase-A2 via Ca 2+ mobilisation, increasing the availability of the substrate for other subsequent eicosanoid pathways.Nano-sized air pollution particles have been shown to increase arachidonic acid and downstream eicosanoids in mice, an effect that was accompanied by inflammation in the gastrointestinal tract (which is exposed to particles following mucocillary clearance from the lung) 53 .In the absence of alterations in downstream eicosanoids, neither of these eicosanoids are likely to represent rate-limiting steps, although it is possible that mobilisation of this substrate may have an influence on eicosanoid formation when other pathways are active, e.g. in the presence of a marked inflammatory response.Different mechanisms may be at play for the release of the fatty acids arachidonic acid and DHA as these are substrates leading to the generation of pro-inflammatory eicosanoids and pro-resolving mediators, respectively.10_carboxy_LTB4 is a metabolite of the inflammatory mediator LTB4, although LTB4 was not detectable in these plasma samples.5,6-dHET is also a metabolite 5,6-EETs which has been shown to dilate pulmonary blood vessels 54 , however, the latter was not affected by GO exposure.14_HDHA is a pathway marker for the pro-resolving mediator maresin 55 , suggesting the increased production of 14_HDHA could represent the initial stages of a counter response to inflammation, although maresin itself was not significant increased by either GO.
However, caution is required in drawing conclusions from a small number of volunteers, and given the small number of eicosanoid species that differed between exposures (6 out of 55 lipids in the panel at p≤0.001), differences in baseline (t=0) levels of eicosanoids, the small magnitude of differences, and the lack of a consistent pattern for specific graphene oxide sizes, we do not feel it is appropriate to speculate further.Nonetheless, the eicosanoid species identified in the present study, and their regulatory pathways, could be included or targeted in future omic-studies exploring the mechanisms by which graphene oxide, or other MNMs, induce downstream effects.

described 5 .
The brachial artery of the non-dominant arm was cannulated with a 27-standard wire gauge steel needle under local anaesthetic.After a 30-minute baseline saline infusion, acetylcholine (endothelium-dependent vasodilator) at 5, 10, and 20 μg/min or sodium nitroprusside (an endotheliumindependent vasodilator) at 2, 4, and 8 μg/min were infused at a rate of 1 mL/min for 6 minutes at each dose.Vasodilators were obtained at clinical grade by the Royal Infirmary of Edinburgh Pharmacy.The two vasodilators were separated by 20-minute saline infusions and given in a randomized order, with the researcher blinded to the drug infusion.Expansion of the forearm was measured during venous occlusion and the gradient of the expansion of the forearm was used to determine blood flow, indicating of the ability of arteries to dilate in the presence and absence of vasodilators.Data was recorded on LabChart Reader software (ADInstruments) and exported to Microscoft Excel for analysis.
Nanoparticle-bound proteins (10 μg) were mixed with lysis buffer (10 μL) containing 5% SDS, triethylammonium bicarbonate (TEAB, 50 mM,pH 7.5) to allow protein solubilisation.Samples were reduced with dithiothreitol (5 mM), alkylated with iodoacetamide (15 mM) and dithiothreitol(5 nm)     added again to quench the alkylation reaction.Samples were centrifuged (14,000g, 10 min) to collect the protein lysates, then mixed with phosphoric acid (12%) and six-volume equivalents of S-trap binding buffer (90% aqueous methanol with TEAB (100 mM, pH 7.1)).Samples were added to a S-trap column and centrifuged (4000g, 2 min) to trap proteins in the columns.Pelleted proteins were washed four times with S-trap binding buffer, then digested with trypsin (0.1 μg/μL, 47 o C, 1 h).Peptide samples were extracted using digestion buffer (50 mM TEAB), 0.1% aqueous formic acid and 30 % aqueous acetonitrile containing 0.1 % formic acid.Finally, peptide samples were desalted by oligo R3 beads in 50% acetonitrile, dried using a vacuum centrifuge (Heto Speedvac) and stored at 4 o C until analysed.Samples were analysed by liquid chromatography mass spectrometry (LC-MS)/mass spectrometry (MS) using an UltiMate® 3000 Rapid Separation lipid chromatography platform (RSLC, Dionex Corporation, USA) coupled to a Q Exactive™ Hybrid Quadrupole-Orbitrap™ mass spectrometer (Thermo Fisher Scientific, USA).Data Analysis.To statistically compare the abundance of proteins identified in the liposomal coronas, mass spectrometry peak intensities were analyzed by importation of the DAW files into Progenesis LC-MS software (version 3.0; Nonlinear Dynamics) with automatic feature detection enabled.A representative reference run was selected automatically, to which all other runs were aligned in a pairwise manner.Automatic processing was selected to run with applied filters for peaks charge state (maximum charge 5) and a protein quantitation method with relative quantitation using Hi-N with N=3 peptides to measurements per protein.The resulting MS/MS peak lists were exported as a single Mascot generic file and loaded onto a local Mascot Server (version 2.3.0;Matrix Science, UK).The spectra were searched against the UniProt database using the following parameters: tryptic enzyme digestion with one missed cleavage allowed, peptide charge of +2 and +3, precursor mass tolerance of 15 mmu, fragment mass tolerance of 8 ppm, oxidation of methionines as variable modifications and carbamidomethyl as fixed modifications, with decoy database search disabled and ESI-QUAD-TOF as the selected instrument.Each search produced an XML file from Mascot and the resulted peptides (XML files) were imported back into Progenesis LC-MS to assign peptides to features.Data were filtered to present a 1% false discovery rate (FDR) and a score above 21 through the 'refine identification' tab of Progenesis QI toolbox.