The impact of graphene oxide sheet lateral dimensions on their pharmacokinetic and tissue distribution profiles in mice

1 1 2 3 The impact of graphene oxide sheet lateral dimensions on 4 their pharmacokinetic and tissue distribution profiles in mice 5 6 7 Dhifaf A. Jasima, Leon Newmana, Artur Filipe Rodriguesa, Isabella A. Vacchib, Matteo A. 8 Lucherelli b, Neus Lozanoa,c, Cécilia Ménard-Moyonb, Alberto Biancob, Kostas 9 Kostarelosa,c,* 10 11 a Nanomedicine Lab, National Graphene Institute and Faculty of Biology, Medicine & Health, 12 University of Manchester, AV Hill Building, Manchester M13 9PT, United Kingdom 13 14 b University of Strasbourg, CNRS, Immunology, Immunopathology and Therapeutic Chemistry, UPR 15 3572, 67000 Strasbourg, France. 16 17 c Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, 18 Bellaterra, 08193 Barcelona, Spain 19 20 21 22 23 24 25 26 27 28 29 * Correspondence to: 30 [email protected] 31 32 33 2 Abstract 1 2 Although the use of graphene and 2-dimensional (2D) materials in biomedicine has been 3 explored for over a decade now, there are still significant knowledge gaps regarding the 4 fate of these materials upon interaction with living systems. Here, the pharmacokinetic 5 profile of graphene oxide (GO) sheets of three different lateral dimensions was studied. 6 The GO materials were functionalized with a PEGylated DOTA (1,4,7,107 tetraazacyclododecane-1,4,7,10-tetraacetic acid), a radiometal chelating agent for 8 radioisotope attachment for single photon emission computed tomography (SPECT/CT) 9 imaging. Our results revealed that GO materials with three distinct size distributions, large 10 (l-GO-DOTA), small (s-GO-DOTA) and ultra-small (us-GO-DOTA), were sequestered by 11 the spleen and liver. Significant accumulation of the large material (l-GO-DOTA) in the 12 lungs was also observed, unlike the other two materials. Interestingly, there was extensive 13 urinary excretion of all three GO nanomaterials indicating that urinary excretion of these 14 structures was not affected by lateral dimensions. Comparing with previous studies, we 15 believe that the thickness of layered nanomaterials is the predominant factor that governs 16 their excretion rather than lateral size. However, the rate of urinary excretion was affected 17 by lateral size, with large GO excreting at slower rates. This study provides better 18 understanding of 2D materials behaviour with different structural features in vivo. 19 20 Key Words: Graphene oxide; functionalization; pharmacokinetics; nanomedicine, pharmacology 21 22 23 24 3 Introduction 1 2 Graphene and related flat-shaped materials exhibit outstanding properties 3 generated from their unique 2D geometry (1-3). These materials have attracted great 4 interest from different scientific disciplines (1-3). The biomedical applicability of graphene 5 has only been researched for the last decade (4). The available graphene surface area is 6 the largest for any material at the nanoscale, which provides a potential delivery platform 7 for maximum payload of therapeutic molecules and for bio-functionalization with imaging 8 probes (5-8). The electrical, electronic (9), mechanical, optical properties (8, 10, 11) and 9 flexibility of graphene based materials (2, 12) allow their use as biosensing platforms and 10 offer a great potential for use in regenerative medicine (13) and electroresponsive drug 11 therapy . It also can be useful to meet the requirements of the electroactive nervous and 12 cardiac systems and therefore provide a means of neuronal and cardiac drug delivery (14). 13 All these properties offer interesting possibilities of graphene materials following their 14 interaction with soft biological matter (11, 15). 15 Due to the great potential offered by using graphene materials for biomedical 16 applications, it is critical to understand their fate in vivo (16). Graphene oxide (GO) has 17 expanded the applications of graphene-based materials in biomedicine due to its 18 hydrophilicity and improved compatibility with biological systems. GO has been 19 administered intravenously (17-19), intraperitoneally, orally (20) and intravitreally (21) with 20 no reported toxic effects even after long exposure times (20, 21). After intravenous or 21 intraperitoneal administration, GO has been reported to accumulate in the mononuclear 22 phagocytic system, or as more commonly known the reticuloendothelial system (RES). 23 The spleen has been reported as the main site for the in vivo degradation of intravenously 24 injected functionalized graphene (22). GO materials have also been reported to 25 accumulate in the lungs (23, 24). Extensive urinary excretion of GO has been reported in 26 several studies after intravenous (i.v.) injection of functionalized GO sheets in mice (16, 27 20, 25-29). 28 Existing studies have used very different types of GO (30, 31) that can result in 29 significantly different biological interactions. These interactions will depend on the type of 30 surface functionalization, functional surface groups and dimensions of the GO sheets 31 (17). The most popular administration route used in the preclinical development of 32 materials for biomedical applications has been the intravenous (i.v.) route, which provides 33 100% bioavailability and therefore maximum information for tissue exposure and toxicity 34 (32). Many studies have used imaging for studying the biodistribution of graphene 35 materials by adding a labelling tag using single photon computed tomography (SPECT/CT) 36 with gamma (γ) emitting isotopes (28, 33), positron emission computed tomography 37 4 (PET/CT) with positron emitting isotopes (27, 34) or even fluorescence tags for optical 1 imaging (26, 34). 2 Previously we have demonstrated that the thickness of the graphene material plays 3 a major role in its tissue accumulation and excretion (27), however, no systematic 4 correlation with the GO sheet lateral dimension has been offered. In this study, three 5 highly purified and well characterised graphene oxide materials that differ only in their 6 lateral dimensions were functionalized with a chelating moiety for imaging purposes, 7 namely a DOTA molecule tethered to a polyethylene glycol (PEG) linker. The resulting 8 functionalized materials had three distinct sizes as well (l-GO-DOTA, s-GO-DOTA and us9 GO-DOTA). Then, whole-body imaging by SPECT/CT and full pharmacokinetic studies 10 were carried out following i.v. administration of the GO-DOTA materials, coupled with 11 analytical and histopathological analysis of critical organs. This study provided important 12 information on the future design of graphene materials for possible tissue targeting 13 applications. 14 15 Results 16 17 Preparation and characterisation of GO and GO-DOTA materials. GO was prepared 18 by a modified Hummers’ method under pyrogen-free conditions, as previously described 19 (35-37). Morphology of GO sheets was characterised by DLS, TEM, and AFM (Figure 1). 20 DLS revealed clear differences in the size distribution curves of the aqueous suspensions 21 of the three types of GO materials (l-GO, s-GO and us-GO). The sonication employed to 22 generate the different lateral size nanosheets (s-GO and us-GO) did not significantly 23 impact their surface charge (Figure 1A-B). 24 In order to perform biodistribution studies, the GO materials were functionalized with 25 the chelating agent DOTA, which was attached to a tetra-ethylene glycol [(PEG)4] 26 molecule bearing a free amine group [DOTA(PEG)4-NH2], as described in Scheme 1. The 27 functionalized chelating moiety [DOTA(PEG)4-NH2], is referred to as ‘DOTA’ for ease 28 throughout the manuscript. Although this process resulted in a decreased nanosheet 29 surface charge (Figure 1B), DOTA functionalization did not significantly impact the 30 colloidal stability of GO, in agreement with previous observations in physiological media 31 (28). DLS is not particularly suitable as a technique to determine the dimensions of non32 spherical particles (38) and we were not able to obtain reliable DLS data for the DOTA33 functionalized material. We emphasised on the structural characterisation of GO sheets 34 before and after DOTA functionalization, performed by TEM and AFM (Figure 1C-D). The 35 tested materials were composed of sheets with distinct differences in lateral dimensions 36 (i.e. the longest dimension observed of each 2D sheet), which are described in detail in 37 Table S1. 38 In summary, l-GO is comprised of sheets with lateral dimension between 1 and 35 39 μm, whilst s-GO sheets range between 30 nm and 1.9 μm. The smallest material (us-GO) 40 was characterised by a narrower size distribution, with lateral dimensions ranging from 10 41 nm to 550 nm. Upon functionalisation there was a reduction in the sheet lateral dimension, 42 particularly l-GO-DOTA compared to l-GO, where there was a 5-fold reduction (Table S1). 43 5 Nevertheless, the lateral dimensions of s-GO-DOTA and us-GO-DOTA sheets remained 1 markedly different from l-GO-DOTA and from each other. These differences could be 2 attributed to the ring opening reaction attacking cooperatively aligned epoxides present in 3 the GO sheet surface, which ultimately create fracture points (39, 40). AFM demonstrated 4 that the thickness of all GO materials had increased following DOTA functionalisation 5 (Table S1). This could be due to the addition of functional groups and has been observed 6 previously in several studies (20, 28, 41). 7 The samples were also characterised by X-ray photoelectron spectroscopy (XPS) 8 (Table S1 and Figure S1-3). The XPS survey table shows ~ 30-31% oxygen content in all 9 three starting materials. The oxygen content is slightly lower after functionalisation with 10 DOTA, with the introduction of 1.1-1.4% nitrogen content, due to the addition of the DOTA 11 molecule. Similarly, in the oxygen O1s high resolution spectra, there is a clear increase in 12 percentage of the carbonyl peak (Figure S2). This is due to the introduction of carboxylic 13 acids and amides. On the other hand, the carbon C1s high resolution spectra (Figure S1) 14 are more complex and revealed changes between the ratio of the carbon region and the 15 carbon-oxygen region. This is often the case after any kind of treatment of GO (42, 43) 16 due to the loss of some labile oxygenated functional groups. In particular, the component 17 attributed to C-O-C bonds decreased whereas the C-OH/C-N band increased. These 18 changes suggest that the reaction of epoxide ring opening occurred, thus generating C-OH 19 groups, alongside the introduction of amino groups from the DOTA compound, as 20 previously reported (44). However, the C-O-C peak did not disappear because of the ether 21 groups in the PEG chain. Two components could be identified in the nitrogen N1s high 22 resolution spectra, namely the amine/amide and the ammonium peaks at ~400 eV and 23 402 eV, respectively. Both amines and amides are present in the DOTA molecule, further 24 indicating the successful functionalisation of GO with the DOTA moiety (Figure S3). 25 Thermogravimetric analysis (TGA) also validated chemical functionalisation of GO 26 (Figure S4). GO starts to lose mass at temperatures even lower than 100°C due to 27 residual water molecules adsorbed on the GO sheets (37, 43, 45, 46).The two weight 28 losses above 150°C are due to the oxygenated species present on the surface of GO and 29 to the DOTA molecule in the case of the functionalized materials. The main weight loss 30 occurs at lower temperature (around 220°C) for the three GO-DOTA compared to the 31 starting materials (~240°C) was considered another indication of covalent functionalization 32 of the DOTA functionality. The %N obtained by XPS indicated that the three conjugates 33 have a similar DOTA loading. Taken together, these results show that the functionalization 34 of GO with the DOTA derivative was successful and did not cause significant reduction of 35 the starting material. 36 37 Efficiency and purity of radiolabeling of [111In]GO-DOTA. The efficiency of 38 radiolabelling of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us39 GO-DOTA) was compared to control [111In]DOTA in Figure 2A. The radiolabelling 40 efficiency of the three samples was ~ 70% at the application point. The samples were 41 purified by removal of unbound [111In]DOTA by centrifugation, reaching a purity of ~90% 42 after centrifugation (Figure 2B). 43 44 Stability of radiolabeling of [111In]GO-DOTA. The three samples had insignificant 45 variable stability in PBS at 370C up to 1 week as shown in Figure 2C, while the samples 46 were stable and retained their radiolabelling purity in 50% serum as shown in the same 47 figure and in Figure S5. The latter figure demonstrated that the three samples retained a 48 6 signal at the application point of the TLC compared to the control materials alone 1 [111In]EDTA and [111In]DOTA which moved to the solvent front. 2 3 Pharmacokinetics and tissue distribution after i.v. administration. The biodistribution 4 and excretion of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us5 GO-DOTA) after i.v. administration was studied by SPECT/CT imaging and cut-and-count 6 γ-scintigraphy (Figure 3, Figure 4 and Figure S6-9). The levels in blood determined by γ7 counting are shown in Figure 3A. The curve demonstrated that pharmacokinetics for all 8 materials followed two-compartment first order kinetics. The pharmacokinetic parameters 9 are summarised in Table S2 and the values remaining in blood after 1h are compared to 10 24h in Table S3. All three materials were removed from blood very rapidly with only less 11 than 1% of the injected dose remaining in the blood after 24h. Dynamic SPECT/CT 12 imaging was carried out during the i.v. administration of the materials (Figure S6). The first 13 panel shows the 2 min phase of injection (movie can be played online), showing the 14 delivery from the tail vein and through the vena cava then through whole blood. All three 15 materials ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA) and the control 16 [111In]DOTA start to accumulate in the organs within the first hour as demonstrated in the 17 planer images in the second panel in the same figure. Organ accumulation is very evident 18 after 4h and 24h (last two panels, Figure S6). Time activity curves for each material are 19 presented in Figure S7, indicating a huge accumulation of the larger material ([111In]l-GO20 DOTA) in the lungs, while the other two materials accumulated mainly in the liver and 21 spleen. All three materials presented bladder and kidney signals at early time points. The 22 control sample [111In]DOTA was totally excreted after 1h with huge bladder signal 23 compared to the other materials (Figure S7). 24 These results were further confirmed by a separate experiment using 3D 25 SPECT/CT for better image resolution as demonstrated in Figure 3B and Figure S8 for 26 the first batch of mice and Figure S9 for a second batch of mice. The scale bars are 27 expressed in percentage of injected dose (%ID) per gram of tissue in Figure 3B and 28 Figure S9, while it is expressed in MBq in Figure S8 and Figure S10. These images show 29 minor intestinal signals in all mice. Expression in %ID per gram of tissue will show the 30 minute amounts in blood levels in light organs. The total remaining amounts in the whole 31 body after 24 h quantified from the SPECT images are presented in Table S4, with 50.8%, 32 36.5% and 38.1% for [111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA, 33 respectively. 34 The whole-body SPECT/CT imaging data were validated by a separate cut and 35 count experiment counting the %ID per whole organ of [111In]GO-DOTA or %ID per gram 36 of tissue measured by γ-scintigraphy in a separate experiment with at least 4 mice per 37 condition (Figure 3C).The data confirmed the same pattern of the SPECT/CT data with 38 lung accumulation for the largest material ([111In]l-GO-DOTA) and liver and spleen 39 accumulation for the all three materials ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 40 [111In]us-GO-DOTA). Some kidney and bladder signals were detected at early time points, 41 while minute intestinal signal was evident at the later time points. Very little signals were 42 detected in the control sample (Figure 3C, bottom row). 43 44 Urinary and fecal excretion. Urinary and faecal excretion was studied by collecting the 45 urine and faeces of injected mice to further examine the extent of elimination of the three 46 materials (Figure 4). Pooled urine samples (n = 4) were collected and counted for 47 7 radioactivity at different time points (Figure 4A). The [111In]us-GO-DOTA showed the 1 maximum excretion compared to the other two samples. Furthermore the samples were 2 analysed by Raman spectroscopy as shown in Figure 4B. The Raman signature of GO3 DOTA was detected in the urine of all three materials, confirming the excretion of the 4 material regardless of the lateral size. Pooled faecal samples were collected after 24h and 5 measured for radioactivity, all samples indicated faecal excretion (Figure 4C). 6 7 Histopathology after i.v. administration of the three materials. Tissue samples (lung, 8 liver, spleen and kidneys) were examined for histopathology using H & E (haematoxylin 9 and eosin) staining of paraffin embedded tissue sections for two animals per condition 10 (Figure 5 and Figure S10-15). Lungs of mice injected with all the GO materials 11 demonstrated an interesting distribution pattern (Figure 5 and Figure S10). All GO 12 materials were predominantly detected as agglomerates within the lumen of lung blood 13 vessels, which indicates the retention of GO in the lung capillaries as a size-dependent 14 phenomenon, with the l-GO-DOTA material showing the largest, and the most 15 agglomerates, as quantified in Figure S11 and shown in different regions in Figure 5 and 16 Figure S10. In the case of the mice spleen samples, no evidence of histopathology was 17 determined in the red pulp (second panel) and the white pulp (third panel) in any of the 18 samples as compared to the controls (Figures S12-13). Liver tissue (Figure S14) and 19 kidney tissue (both glomerular and tubular regions) (Figure S15) also demonstrated 20 healthy anatomical structures with no evident histopathology in any of the samples. 21 22 23 Discussion 24 25 To date there is no direct method to quantitatively measure the amount of GO in 26 physiological fluids (e.g. tissues, blood and urine) due to the background interferences 27 from the complex molecular composition of such biological fluids. Studies rely on labelling 28 of the material for quantification, though other methods can offer qualitative detection such 29 as TEM and Raman spectroscopy. In this work, we studied the pharmacokinetic profile of 30 three lateral sizes of thin GO sheets. The functionalisation of GO with PEGylated DOTA 31 was performed to allow the chelation of a radioactive metal [111In] for studying and 32 quantifying the tissue distribution of the materials. The functionalisation of GO with the 33 PEGylated DOTA was performed via epoxide opening. Because of the abundance of 34 epoxides on the GO surface, this strategy generally leads to higher levels of 35 functionalization compared to the derivatisation of hydroxyls or carboxylic acids (47). 36 Structural characterisation was carried out by TEM and AFM to reveal whether GO 37 sheets would have marked differences in their morphology. The materials were produced 38 with controlled lateral dimensions by tuning their lateral size by sonication as previously 39 reported (37, 48-50). After functionalization all three materials maintained their 2D 40 morphology, while the thickness increased from single to a few layers following DOTA 41 functionalization due to the presence of the functional groups and minor agglomeration 42 that led to slightly thicker sheets. These findings are consistent with previous reports of 43 functionalization of GO with DOTA (27, 28) PEG (20, 51), dextran (25) and bovine serum 44 albumin (52). Interestingly, we noticed that l-GO underwent a marked reduction in sheet 45 lateral dimension after DOTA functionalization. Surface functionalization of GO has also 46 been reported to reduce the lateral dimension of the sheets (28, 52). However, the lateral 47 8 dimensions of all functionalized GO-DOTA materials remained markedly different from 1 each other. 2 The analysis of the XPS spectra (Table S1 and Figures S1-3) showed that the 3 functionalization of GO with the DOTA derivative was successful, as demonstrated by the 4 detection of nitrogen in the GO-DOTA materials compared to their starting counterparts. 5 Moreover, nitrogen atoms were detected in the form of amines and amides (Figure S3), 6 which are present in the DOTA molecule. The higher intensity of the carbonyl peak in the 7 carbon high resolution spectra after functionalization (Figure S1) is due to the carboxylic 8 acids of the DOTA molecule and it is also an indication of the presence of amides. On the 9 other hand, the changes in the carbon high resolution spectra after functionalization are 10 more subtle. First, there is a decreased abundance of oxidised carbon atoms compared to 11 the graphitic carbon region, due to the loss of some labile groups after functionalization of 12 GO. The contribution of the epoxides (C-O-C) is reduced compared to the C-OH/C-N 13 band, which is coherent with the introduction of the DOTA moiety by opening of epoxides, 14 thus generating C-OH and introducing amino groups. Chemical functionalization of GO 15 was further evidenced by analysing the different GO derivatives by TGA under an inert 16 atmosphere (Figure S4). GO is thermally unstable and starts to lose mass at temperatures 17 even lower than 100°C due to residual water molecules adsorbed on the GO sheets. A 18 main weight loss was identified at 230°C that is due to the elimination of labile oxygen19 containing groups (53). All GO-DOTA materials revealed a lower thermal stability probably 20 due to functionalization with DOTA. Therefore, taken together these results confirm the 21 effective grafting of DOTA on the surface of the three GO samples. In addition, the 22 structure of GO is preserved and the functionalization strategy induced no significant 23 reduction, thus enabling to obtain stable suspensions in aqueous solutions. 24 Other studies have reported radiolabelling of GO materials with iodine resulting in 25 iodinated constructs that are unstable, while the high affinity of iodine for the thyroid gland 26 can be misleading with regards to 2D material biodistribution profiles (25). Radiolabelling 27 with 111In has been performed previously using physical adsorption of the DTPA 28 (diethylenetriaminepentaacetic acid) chelating agent on the surface of GO by π-stacking. 29 Such strategy however increased the thickness of the GO sheets dramatically (33). The 30 radiolabelling efficiency and stability of covalently bound DOTA to GO materials has been 31 tested before with high efficiency and stability (27, 28). In this study a similar strategy was 32 exploited using GO materials with three different lateral dimensions (l-GO-DOTA, s-GO33 DOTA and us-GO-DOTA). The radiolabelling efficiency was not affected dramatically by 34 the lateral size of the GO sheets. All samples remained stable up to 24h at 370C in 50% 35 serum. Similar to original samples without serum incubation, the control [111In]DOTA 36 travelled to the solvent front as compared to the three samples that remained at the 37 bottom of the TLC (Figure S5). This indicated that the [111In]DOTA was attached to the GO 38 materials with no interferences occurring from the serum proteins at the times that were 39 tested (reflecting the time predicted for the material will spend in blood after injection and 40 beyond that time). This avoids any possible conflictions in the tissue distribution that are 41 created from the detachment of the [111In]DOTA. Therefore the purity of material was 42 confirmed to be suitable for in vivo administration. Furthermore the in vivo stability was 43 confirmed by the Raman signal that coregistered with the radioactivity at the bottom of the 44 TLC strips in the urine of the mice after injection as compared to the control free label that 45 was predominately at the top of the TLC strips (Figure 4B). This was comparable to the 46 original samples before injection in Figure 2B this further confirmed the stability and 47 association of the label with the GO material even after injection. 48 9 Compared to our previous studies (27-29) where the small and thin GO-DOTA 1 material was only injected, in this study, three materials with comparable thickness and the 2 three different sizes were injected intravenously in mice. All three materials were removed 3 from blood within minutes, as shown by the first-phase distribution half-life (t1/2α). Although 4 no significant differences were seen in the pharmacokinetic data of the materials, the 5 [111In]DOTA control was removed from blood faster. In the second phase the [111In]us-GO6 DOTA material was more similar to [111In]DOTA control and remained slightly longer in 7 circulation in the second phase half-life (t1/2β). The amounts remaining in blood after 1h 8 and 24h were more elevated for the [111In]s-GO-DOTA and [111In]us-GO-DOTA compared 9 to [111In]l-GO-DOTA (Figure 3A, Figure S5,Table S2 and Table S3). Smaller 10 nanoparticles are well known to circulate longer (54, 55).The area under the blood 11 concentration time curve (AUC), steady state volume of distributions (Vdss) (indicating the 12 body overall and tissue exposure) and the clearance values are shown in Table S2. The 13 [111In]DOTA control was almost entirely removed from blood at the early hours after 14 administration (Table S3). On the contrary, the remaining amounts in the whole body after 15 24h for the three materials was much higher compared to control material (Table S4). 16 It is clear from our results that tissue distribution occurs very rapidly for all three 17 materials and is largely affected by the size of the graphene sheet (Figure 3B-C and 18 Figure S6-9). It is very evident that the large material ([111In]l-GO-DOTA) tends to 19 accumulate in the lungs early after injection, with reductions in the signal after 24h as seen 20 by both the time activity curve and the γ-counting experiment (Figure S7 and Figure 3C, 21 respectively). Larger nanomaterials tend to accumulate in the lungs after i.v. administration 22 due to the first capillary bed (55). This has also been demonstrated with graphene 23 materials (17, 56, 57). Some material also remained in the lung tissues after 24h as seen 24 in the SPECT/CT data, γ-counting and in the H & E sections/semi-quantification (Figure 25 3B-C, Figure S6-9 and Figure S11). Some thickening of the alveolar walls was also seen 26 especially in lungs from mice injected with l-GO-DOTA and s-GO-DOTA, indicating cellular 27 infiltration and constriction, probably as a result of the material entrapment in the lung 28 tissues. We observed a clear reduction of the lung signal after 24h. It is known that the 29 removal of nanomaterials from the lung is carried out by phagocytic uptake of the lung 30 macrophages. This is the main mechanism to remove the insoluble aggregated 31 nanoparticles into the micrometer-sized particles from the lung tissues. Particle-containing 32 macrophages may re-enter into the interstitium and be cleared by the lymphatics or other 33 organs (58). 34 The smaller materials ([111In]s-GO-DOTA and [111In]us-GO-DOTA) accumulated 35 mainly in the liver and spleen with slight variability between animals in maximum 36 accumulations in these organs. The [111In]us-GO-DOTA showed a slight reduction in the 37 hepatic signal after 24h compared to the [111In]s-GO-DOTA that remained the same. On 38 the other hand, both materials accumulated in the spleen at high concentration even after 39 24h (Figure 3B-C and Figures S6-9). It is well known that nanomaterials, including 40 graphene-based materials, get trapped in the mononuclear phagocytic system (RES) (16, 41 25, 34, 59, 60). This happens within the liver due to the non-continuous liver endothelia 42 with vascular fenestrations measuring 50–100 nm, leading to nonspecific accumulation of 43 nanoparticles within this range. In the spleen, interendothelial cell slits with a size range of 44 200–500 nm mediate the retention of particles >200 nm (55). Particle shape also accounts 45 for accumulation in certain tissues, with elongated nanoparticles mainly in the spleen, 46 while the more spherical ones accumulate in the liver (61). The total amount of material 47 16 References 1 1. K. Kostarelos, K. S. Novoselov, Graphene devices for life. Nature Nanotechnology 9, 744 (2014). 2 2. Y. Pan, N. G. Sahoo, L. Li, The application of graphene oxide in drug delivery. Expert Opinion on Drug 3 Delivery 9, 1365-1376 (2012). 4 3. P. Avouris, F. Xia, Graphene applications in electronics and photonics. MRS Bulletin 37, 1225-1234 5 (2012). 6 4. Z. Liu, J. T. Robinson, X. Sun, H. Dai, PEGylated nanographene oxide for delivery of water-insoluble 7 cancer drugs. Journal of Americal Chemical Society 130, 10876-10877 (2008). 8 5. K. Kostarelos, K. S. Novoselov, Exploring the interface of graphene and biology. Science 344, 261-263 9 (2014). 10 6. K. P. Loh, Q. Bao, G. Eda, M. Chhowalla, Graphene oxide as a chemically tunable platform for optical 11 applications. Nature Chemistry 2, 1015-1024 (2010). 12 7. L. Feng, Z. Liu, Graphene in biomedicine: opportunities and challenges. Nanomedicine 6, 317-324 13 (2011). 14 8. H. Shen, L. Zhang, M. Liu, Z. Zhang, Biomedical Applications of Graphene. Theranostics 2, 283 -294 15 (2012). 16 9. I. Calizo, I. Bejenari, M. Rahman, G. Liu, A. A. Balandin, Ultraviolet Raman microscopy of single and 17 multilayer graphene. Journal of Applied Physics 106, 043509 (2009). 18 10. A. C. Ferrari et al., Raman Spectrum of Graphene and Graphene Layers. Physical Review Letters 97, 19 187401 (2006). 20 11. A. Bendali et al., Purified Neurons can Survive on Peptide-Free Graphene Layers. Advanced Healthcare 21 Materials 10.1002/adhm.201200347, 929-933 (2013). 22 12. K. S. Novoselov et al., A roadmap for graphene. Nature 490, 192-200 (2012). 23 13. M. Zhou et al., Graphene oxide: A growth factor delivery carrier to enhance chondrogenic differentiation 24 of human mesenchymal stem cells in 3D hydrogels. Acta Biomaterialia 96, 271-280 (2019). 25 14. N. A. Kotov et al., Nanomaterials for Neural Interfaces. Advanced Materials 21, 3970-4004 (2009). 26 15. S. K. Seidlits, J. Y. Lee, C. E. Schmidt, Nanostructured scaffolds for neural applications. Nanomedicine 27 (London, England) 3, 183-199 (2008). 28 16. K. Yang et al., In Vivo Pharmacokinetics, Long-Term Biodistribution, and Toxicology of PEGylated 29 Graphene in Mice. ACS Nano 5, 516-522 (2011). 30 17. J. H. Liu et al., Effect of size and dose on the biodistribution of graphene oxide in mice. Nanomedicine 31 (London, England) 7, 1801-1812 (2012). 32 18. L. Zhan et al., Biodistribution of co-exposure to multi-walled carbon nanotubes and graphene oxide 33 nanoplatelets radiotracers. J Nanopart Res 13, 2939-2947 (2011). 34 19. G. Qu et al., The ex vivo and in vivo biological performances of graphene oxide and the impact of 35 surfactant on graphene oxide's biocompatibility. Journal of Environmental Sciences 25, 873-881 (2013). 36 20. K. Yang et al., In vivo biodistribution and toxicology of functionalized nano-graphene oxide in mice after 37 oral and intraperitoneal administration. Biomaterials 34, 2787-2795 (2013). 38 21. L. Yan et al., Can graphene oxide cause damage to eyesight? Chemical Research Toxicology 25, 126539 1270 (2012). 40 22. C. M. Girish, A. Sasidharan, G. S. Gowd, S. Nair, M. Koyakutty, Confocal Raman Imaging Study Showing 41 Macrophage Mediated Biodegradation of Graphene In Vivo. Advanced Healthcare Materials 2, 148942 1500 (2013). 43 23. K. Wang et al., Biocompatibility of Graphene Oxide. Nanoscale Research Letters 6, 8-8 (2011). 44 24. X. Zhang et al., Distribution and biocompatibility studies of graphene oxide in mice after intravenous 45 administration. Carbon 49, 986-995 (2011). 46 25. S. Zhang, K. Yang, L. Feng, Z. Liu, In vitro and in vivo behaviors of dextran functionalized graphene. 47 Carbon 49, 4040-4049 (2011). 48 26. K. Yang et al., Graphene in Mice: Ultrahigh In Vivo Tumor Uptake and Efficient Photothermal Therapy. 49 Nano Letters 10, 3318-3323 (2010). 50 27. D. A. Jasim et al., Thickness of functionalized graphene oxide sheets plays critical role in tissue 51 accumulation and urinary excretion: A pilot PET/CT study. Applied Materials Today 4, 24-30 (2016). 52 28. D. A. Jasim, C. Menard-Moyon, D. Begin, A. Bianco, K. Kostarelos, Tissue distribution and urinary 53 excretion of intravenously administered chemically functionalized graphene oxide sheets. Chemical 54 Science 6, 3952-3964 (2015). 55 29. D. A. Jasim et al., The Effects of Extensive Glomerular Filtration of Thin Graphene Oxide Sheets on 56 Kidney Physiology. ACS Nano 10, 10753-10767 (2016). 57 30. A. Bianco, Graphene: Safe or Toxic? The Two Faces of the Medal. Angewandte Chemie International 58 Edition 52, 4986-4997 (2013). 59 31. P. Wick et al., Classification Framework for Graphene-Based Materials. Angewandte Chemie 60 International Edition 10.1002/anie.201403335, 2-7 (2014). 61 32. R. P. Heaney, Factors influencing the measurement of bioavailability, taking calcium as a model. Journal 62 of Nutrition 131, 1344S-1348S (2001). 63 33. B. Cornelissen et al., Nanographene oxide-based radioimmunoconstructs for in vivo targeting and 64 SPECT imaging of HER2-positive tumors. Biomaterials 34, 1146-1154 (2013). 65 17 34. H. Hong et al., In Vivo Targeting and Imaging of Tumor Vasculature with Radiolabeled, Antibody1 Conjugated Nanographene. ACS Nano 6, 2361-2370 (2012). 2 35. D. A. Jasim, N. Lozano, K. Kostarelos, Synthesis of few-layered, high-purity graphene oxide sheets from 3 different graphite sources for biology. 2D Materials 3, 014006 (2016). 4 36. S. P. Mukherjee et al., Detection of Endotoxin Contamination of Graphene Based Materials Using the 5 TNF-α Expression Test and Guidelines for Endotoxin-Free Graphene Oxide Production. PLOS ONE 11, 6 e0166816 (2016). 7 37. A. F. Rodrigues et al., A blueprint for the synthesis and characterisation of thin graphene oxide with 8 controlled lateral dimensions for biomedicine. 2D Materials 5, 035020 (2018). 9 38. S. Bhattacharjee, DLS and zeta potential - What they are and what they are not? Journal of controlled 10 release : official journal of the Controlled Release Society 235, 337-351 (2016). 11 39. J. L. Li et al., Oxygen-driven unzipping of graphitic materials. Phys Rev Lett 96, 176101 (2006). 12 40. T. Sun, S. Fabris, Mechanisms for oxidative unzipping and cutting of graphene. Nano Lett 12, 17-21 13 (2012). 14 41. W. Zhang et al., Unraveling stress-induced toxicity properties of graphene oxide and the underlying 15 mechanism. Advanced materials (Deerfield Beach, Fla.) 24, 5391-5397 (2012). 16 42. A. M. Dimiev, L. B. Alemany, J. M. Tour, Graphene oxide. Origin of acidity, its instability in water, and a 17 new dynamic structural model. ACS Nano 7, 576-588 (2013). 18 43. S. Eigler, C. Dotzer, A. Hirsch, M. Enzelberger, P. Müller, Formation and Decomposition of CO2 19 Intercalated Graphene Oxide. Chemistry of Materials 24, 1276-1282 (2012). 20 44. I. A. Vacchi, C. Spinato, J. Raya, A. Bianco, C. Menard-Moyon, Chemical reactivity of graphene oxide 21 towards amines elucidated by solid-state NMR. Nanoscale 8, 13714-13721 (2016). 22 45. S. Stankovich et al., Synthesis of graphene-based nanosheets via chemical reduction of exfoliated 23 graphite oxide. Carbon 45, 1558-1565 (2007). 24 46. J. I. Paredes, S. Villar-Rodil, A. Martínez-Alonso, J. M. Tascón, Graphene oxide dispersions in organic 25 solvents. Langmuir : the ACS journal of surfaces and colloids 24, 10560-10564 (2008). 26 47. I. A. Vacchi, S. Guo, J. Raya, A. Bianco, C. Ménard-Moyon, Strategies for the Controlled Covalent 27 Double Functionalization of Graphene Oxide. Chemistry – A European Journal 26, 6591-6598 (2020). 28 48. M. Orecchioni et al., Molecular and Genomic Impact of Large and Small Lateral Dimension Graphene 29 Oxide Sheets on Human Immune Cells from Healthy Donors. Advanced Healthcare Materials 5, 276-287 30 (2016). 31 49. A. F. Rodrigues et al., Size-Dependent Pulmonary Impact of Thin Graphene Oxide Sheets in Mice: 32 Toward Safe-by-Design. Advanced Science n/a, 1903200. 33 50. L. Newman et al., Nose-to-Brain Translocation and Cerebral Biodegradation of Thin Graphene Oxide 34 Nanosheets. Cell Reports Physical Science 1, 100176 (2020). 35 51. W. Zhang et al., Unraveling Stress-Induced Toxicity Properties of Graphene Oxide and the Underlying 36 Mechanism. Advanced Materials 24, 5391-5397 (2012). 37 52. Y. Li et al., Surface Coating-Dependent Cytotoxicity and Degradation of Graphene Derivatives: Towards 38 the Design of Non-Toxic, Degradable Nano-Graphene. Small 10.1002/smll.201303234, 1544-1554 39 (2013). 40 53. I. Jung et al., Reduction Kinetics of Graphene Oxide Determined by Electrical Transport Measurements 41 and Temperature Programmed Desorption. The Journal of Physical Chemistry C 113, 18480-18486 42 (2009). 43 54. F. Alexis, E. Pridgen, L. K. Molnar, O. C. Farokhzad, Factors Affecting the Clearance and Biodistribution 44 of Polymeric Nanoparticles. Molecular Pharmaceutics 5, 505-515 (2008). 45 55. E. Blanco, H. Shen, M. Ferrari, Principles of nanoparticle design for overcoming biological barriers to drug 46 delivery. Nature Biotechnology 33, 941-951 (2015). 47 56. K. Wang et al., Biocompatibility of graphene oxide. Nanoscale Res Lett 6, 1-8 (2011). 48 57. S. K. Singh et al., Thrombus Inducing Property of Atomically Thin Graphene Oxide Sheets. ACS Nano 5, 49 4987-4996 (2011). 50 58. M. Geiser, Update on macrophage clearance of inhaled microand nanoparticles. Journal of Aerosol 51 Medicine Pulmonary Drug Delivery 23, 207-217 (2010). 52 59. K. Yang et al., The influence of surface chemistry and size of nanoscale graphene oxide on photothermal 53 therapy of cancer using ultra-low laser power. Biomaterials 33, 2206-2214 (2012). 54 60. H. Hong et al., In vivo targeting and positron emission tomography imaging of tumor vasculature with 55 (66)Ga-labeled nano-graphene. Biomaterials 33, 4147-4156 (2012). 56 61. M. Zhang et al., Radiolabeling, whole-body single photon emission computed tomography/computed 57 tomography imaging, and pharmacokinetics of carbon nanohorns in mice. International Journal of 58 Nanomedicine 11, 3317-3330 (2016). 59 62. S. Liang et al., In vivo pharmacokinetics, transfer and clearance study of graphene oxide by La/Ce dual 60 elemental labelling method. NanoImpact 17, 100213 (2020). 61 63. B. T. Kurien, N. E. Everds, R. H. Scofield, Experimental animal urine collection: a review. Lab Animals 38, 62 333-361 (2004). 63 64. B. Li et al., Influence of polyethylene glycol coating on biodistribution and toxicity of nanoscale graphene 64 oxide in mice after intravenous injection. Intenational Journal of Nanomedicine 9, 4697-4707 (2014). 65 65. L. Newman et al., Splenic Capture and In Vivo Intracellular Biodegradation of Biological-Grade Graphene 66 Oxide Sheets. ACS Nano 14, 10168-10186 (2020). 67 18 66. M. C. Duch et al., Minimizing Oxidation and Stable Nanoscale Dispersion Improves the Biocompatibility of 1 Graphene in the Lung. Nano Letters 11, 5201-5207 (2011). 2 67. L. Ma-Hock et al., Comparative inhalation toxicity of multi-wall carbon nanotubes, graphene, graphite 3 nanoplatelets and low surface carbon black. Particle and Fibre Toxicology 10, 23 (2013). 4 68. S. Macholl et al., High-throughput high-volume nuclear imaging for preclinical in vivo compound 5 screening(§). EJNMMI Research 7, 33 (2017). 6 69. A. Ganguly, S. Sharma, P. Papakonstantinou, J. Hamilton, Probing the Thermal Deoxygenation of 7 Graphene Oxide Using High-Resolution In Situ X-ray-Based Spectroscopies. The Journal of Physical 8 Chemistry C 115, 17009-17019 (2011). 9 70. L.-N. Zhou, X.-T. Zhang, W.-J. Shen, S.-G. Sun, Y.-J. Li, Monolayer of close-packed Pt nanocrystals on a 10 reduced graphene oxide (RGO) nanosheet and its enhanced catalytic performance towards methanol 11 electrooxidation. RSC Advances 5, 46017-46025 (2015). 12 71. C. Botas et al., Tailored graphene materials by chemical reduction of graphene oxides of different atomic 13 structure. RSC Advances 2, 9643-9650 (2012). 14 15 16 19 Figure Legends 1 2 Scheme 1. Synthesis of GO-DOTA. For the sake of clarity, electrostatic interactions between the 3 protonated amine in NH2-PEG4-DOTA and the carboxylate groups at the edges of GO or in the 4 PEG4-DOTA chain linked to GO are not shown. 5 Figure 1. Physicochemical characterisation of l-GO, s-GO and us-GO before and after DOTA 6 functionalisation. A) Dynamic light scattering (DLS) size distributions; B) Electrophoretic mobility 7 (ζ potential) mean surface charge data. Morphological and structural characterisation data is 8 shown, using C) TEM and D) AFM. 9 Figure 2: Radiolabelling efficiency and stability. A) Efficiency of radiolabelling of the three types 10 of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA) compared to control 11 [111In]DOTA after the radiolabelling reaction; values indicate the average of three independent 12 labelling repeats. B) Radiolabelling purity of the three samples after centrifugation and before 13 administration in the animals. C) Stability of radiolabelling in PBS and 50% serum up to 7 days 14 Figure 3: Biodistribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 15 [111In]us-GO-DOTA) compared to [111In]DOTA control at 1h and 24h. A) Blood profile up to 24h. 16 B) SPECT/CT images expressed as % of injected dose per gram of tissues (doses were decay 17 corrected at the second time point). From left to right (whole body maximum intensity projections 18 (MIP), sagittal, coronal and transverse views). Interactive 3D images of the MIPs are available 19 online. C) Organ distribution of the three materials as determined by γ-counting. Four animals per 20 group were used for A and C, while two animals were used for the imaging in B. 21 Figure 4: Excretion profile of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA 22 compared to control [111In]DOTA). A) Urinary excretion profile of the three materials at different 23 time points B) detection of the graphene material in the urine 24 hours post administration of mice 24 as demonstrated by radio-TLC and corroborative Raman spectroscopy, scale bars are 20 µm. C) 25 Faecal excretion of the three materials compared to the control. 26 Figure 5. Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on lung compared to control 5% 27 dextrose. Haematoxylin and eosin stained lung sections (5 µm thick) after injection of l-GO-DOTA, 28 s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h of two different mice for 29 each material. Scale bars for the images on the left are 100 µm while those on the right are 20 µm. 30 31 32 33 34 35 36 37 20 Figures 1 2 3 4 Scheme 1. Synthesis of GO-DOTA. For the sake of clarity, electrostatic interactions 5 between the protonated amine in NH2-PEG4-DOTA and the carboxylate groups at the 6 edges of GO or in the PEG4-DOTA chain linked to GO are not shown. 7 8 9 10 11 12 13 14 DOTA(PEG)4 GO GO-DOTA O O O HO OH OH O NN NN OHO HO O HO O HO O N H OO NH2 HN HO OH OH O OH HN OH OH NH 4 O HN O DOTA O N H O DOTA O NH O DOTA 4 4 4 HO OH OH OH 21 Figure 1 1 2 3 4 5 Figure 1. Physicochemical characterisation of l-GO, s-GO and us-GO before and after DOTA 6 functionalisation. A) Dynamic light scattering (DLS) size distributions; B) Electrophoretic mobility 7 (ζ potential) mean surface charge data. Morphological and structural characterisation data is 8 shown, using C) TEM and D) AFM. 9 10 TEM GO-DOTA GO Large GO Small GO Ultra-small-GO 500 nm 500 nm 500 nm 2 μm 2 μm500 nm C A -70 -60 -50 -40 -30 -20 -10 0l-GO s-GO us-GO ζ-potential (mV) GO GO-DOTA B 0 5 10 15 20 25 30 35 0200 400 600 800 1000 % Intensity Hydrodynamic diameter (nm) l-GO s-GO us-GO AFM 5 nm 0 nm 5 nm 0 nm 5 nm 5 nm 0 nm GO GO-DOTA 5 nm 0 nm 1 μm 4 μm 1 μm1 μm 1 μm 4 μm 0 nm 5 nm 0 nm D 22 Figure 2 1 2 3 4 Figure 2: Radiolabelling efficiency and stability. A) Efficiency of radiolabelling of the three types 5 of GO ([111In] l-GO-DOTA, [111In] s-GO-DOTA and [111In] us-GO-DOTA) compared to control 6 [111In]DOTA after the radiolabelling reaction; values indicate the average of three independent 7 labelling repeats. B) Radiolabelling purity of the three samples after centrifugation and before 8 administration in the animals. C) Stability of radiolabelling in PBS and 50% serum up to 7 days. 9 10 20 40 60 80 100 120 012345678 % Radiolabelling at application point Days post-centrifugation 20 40 60 80 100 120 012345678 % Radiolabelling at application point Days post-centrifugation PBS 50% Serum [111In] l-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA C [111In] l-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA 67.4 ±4.3% 77.6 ±6.6% 77.7 ±4.3% 86.3 ±7.8% [111In] l-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA [111In] DOTA A B 88.7 ±5.4% 90.6 ±4.4% 92.2 ±3.8% 93.9 ±2.1% [111In ] l-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA [111In] DOTA 23 Figure 3 1 2 3 4 5 Figure 3: Biodistribution of the three types of GO ([111In] l-GO-DOTA, [111In] s-GO-DOTA and 6 [111In] us-GO-DOTA) compared to [111In]DOTA control at 1h and 24h. A) Blood profile up to 7 24h. B) SPECT/CT images expressed as % of injected dose per gram of tissues (doses were decay 8 corrected at the second time point). From left to right (whole body maximum intensity projections 9 (MIP), sagittal, coronal and transverse views). Interactive 3D images of the MIPs are available 10 online. C) Organ distribution of the three materials as determined by γ-counting. Four animals per 11 group were used for A and C, while two animals were used for the imaging in (B). 12 13 14 15 0 20 40 60 %ID per g of tissues 1hr 4hrs 24hrs 0 20 40 60 %ID per g of tissues 0 20 40 60 %ID per g of tissues 0 20 40 60 %ID per g of tissues AC B 0.01 0.1 1 10 100 0500 1000 1500 % ID in whole blood Time (min) Control l-GO-DOTA s-GO-DOTA us-GO-DOTA 1 h 24 h [111In] l-GO-DOTA 50 % ID/g 0 % ID/g [111In] l-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA [111In] DOTA [111In] DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] us-GO-DOTA [111In] DOTA [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] l-GO-DOTA [111In] s-GO-DOTA 24 1 2 3 Figure 4 4 5 6 Figure 4: Excretion profile of l-, sand us-GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 7 [111In]us-GO-DOTA) compared to control [111In]DOTA. A) Urinary excretion profile of the three 8 materials at different time points; B) detection of the graphene material in the urine 24 hours post 9 administration of mice as demonstrated by radio-TLC and corroborative Raman spectroscopy 10 (scale bars are 20 µm); C) faecal excretion of the three materials compared to the control. 11 12 13 14 15 16 0 10 20 30 40 % ID in urine 1 h 4 h 24 h 0 0.5 1 1.5 2 2.5 3 % ID in faeces A C DOTA [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] Solvent Front Application Point Raman Spectroscopy Region of Interest 1000 1500 2000 Intensity (a.u.) Raman Shift (cm-1) 1000 1500 2000 Intensity (a.u.) Raman Shift (cm-1) 1000 1500 2000 Intensity (a.u.) Raman Shift (cm-1) 1000 1500 2000 Intensity (a.u.) Raman Shift (cm-1) [111In] s-GO-DOTA B[111In ] l-GO-DOTA [111In] us-GO-DOTA [111In] DOTA 25 1 2 Figure 5 3 4 5 Figure 5. Effect of non-radiolabelled l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on lung 6 compared to vehicle alone (5% dextrose). Haematoxylin and eosin stained lung sections (5 µm 7 thick) after injection of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) 8 after 24h of two different mice for each material. Two magnifications were used, with scale bars on 9 the left in each panel showing 100 µm, while those on the right are indicating 20 µm. 10 11 12 13 l-GO Dex 5% s-GO us-GO Mouse1 Mouse2 32 1 2 3 Figure S7: Organ distribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 4 [111In]us-GO-DOTA) compared to control [111In]DOTA demonstrated by the time activity curve of the 5 corresponding planar SPECT data (n=2). 6 7 8 0 1 5 10 15 20 25 0 20 40 60 80 % ID/g Time (h) DOTA l-GO-DOTA s-GO-DOTA us-GO-DOTA 0 1 5 10 15 20 25 0 20 40 60 80 % ID/g Time (h) DOTA l-GO-DOTA s-GO-DOTA us-GO-DOTA 0 1 5 10 15 20 25 0 20 40 60 80 % ID/g Time (h) DOTA l-GO-DOTA s-GO-DOTA us-GO-DOTA 0 1 5 10 15 20 25 0 20 40 60 80 % ID/g Time (h) DOTA l-GO-DOTA s-GO-DOTA us-GO-DOTA Lung Liver Kidney Bladder [111In] l-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA [111In] DOTA 33 1 2 Figure S8: Biodistribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 3 [111In]us-GO-DOTA) compared to control [111In]DOTA at 1h and 24h in a first batch of mice. From 4 left to right (whole body maximum intensity projections (MIP), sagittal, coronal and transverse 5 views). The data here are normalized MBq for decay and dose differences between mice. 6 Interactive 3D images (of MIPs) are available online. 7 8 9 1 h 24 h [111In] l-GO-DOTA [111In] DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA 34 1 2 3 Figure S9: Biodistribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 4 [111In]us-GO-DOTA) compared to control [111In]DOTA at 1h and 24h in a second batch of mice. 5 From left to right (whole body maximum intensity projections (MIP), sagittal, coronal and transverse 6 views). SPECT/CT images expressed as % of injected dose per gram of tissues (doses were decay 7 corrected at the second time point). Interactive 3D images (of MIPs) are available online. 8 9 10 [111In] l-GO-DOTA [111In] DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA 35 1 2 Figure S10: Biodistribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 3 [111In]us-GO-DOTA) compared to control [111In]DOTA at 1h and 24h in a second batch of mice. 4 From left to right (whole body maximum intensity projections (MIP), sagittal, coronal and transverse 5 views). Data here are expressed in normalized MBq for decay and differences in injected doses 6 between mice. Interactive 3D images (of MIPs) are available online. 7 8 9 10 11 12 13 14 15 [111In] l-GO-DOTA [111In] DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA 36 1 2 3 4 5 6 7 8 Figure S11: Effect of l-GO-DOTA on lung. Haematoxylin and eosin stained lung sections (5 µm 9 thick) after injection of l-GO-DOTA after 24h of two different mice. Scale bars for the images on the 10 left are 100 µm while those on the right are 20 µm. These images are captured at different locations 11 compared to those in Figure 5. 12 13 14 l-GO Mouse1 Mouse2 37 1 2 3 4 Figure S12: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on lung compared to control 5% 5 dextrose. Haematoxylin and eosin stained lung sections (5 µm thick) after injection of l-GO-DOTA, 6 s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h of two different mice for 7 each material. Scale bars for the images on the left are 100 µm while those on the left are 20 µm. 8 9 10 l-GO s-G O us-GO 0.001 0.01 0.1 1 % infiltration Large GO Small GO Ultra-small GO 38 1 2 Figure S13: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on spleen structure compared 3 to control 5% dextrose. Haematoxylin and eosin stained spleen sections (5 µm thick) after 4 injection of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h 5 of first set of mice. No evidence of histopathology was determined in the red pulp (second panel) 6 and the white pulp (third panel) in any of the samples as compared to the controls. Each point 7 represents a different mouse. Scale bars for the images on the left are 100 µm while those in the 8 middle and last panels are 50 µm. The insets are 1 µm. 9 10 11 12 l-GO Dex5% s-GO us-GO 39 1 2 Figure S14: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on spleen structure compared 3 to control 5% dextrose. Haematoxylin and eosin stained spleen sections (5 µm thick) after 4 injection of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h 5 of second set of mice. No evidence of histopathology was determined in the red pulp (second 6 panel) and the white pulp (third panel) in any of the samples as compared to the controls. Scale 7 bars for the images on the left are 100 µm while those in the middle and last panels are 50 µm,. 8 The insets are 1 µm. 9 10 11 12 l-GO Dex5% s-GO us-GO 40 1 Figure S15: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on liver structure compared to 2 control 5% dextrose. Haematoxylin and eosin stained liver sections (5 µm thick) after injection of 3 l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h. No evidence 4 of histopathology was determined in the livers in any of the samples as compared to the controls. 5 Scale bars for the images are 100 µm. 6 7 8 9 10 l-GO Dex5% s-GO us-GO Mouse1 Mouse2 41 1 2 3 Figure S16: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on kidney structure compared 4 to control 5% dextrose. Haematoxylin and eosin stained kidney sections (5 µm thick) after 5 injection of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h. 6 No evidence of histopathology was determined in the kidney glomerular or tubular regions as 7 compared to the controls. Scale bars for the images are 50 µm. G: glomerulus, P: proximal 8 convoluted tubule and D: distal convoluted tubule. 9 10 11 Mouse1 l-GO Dex5% s-GO us-GO Glomerulus Tubules Glomerulus Tubules G PD P D G G P D G P D G P DG P D G P D PD Mouse2 G