This is the accepted version of the journal article: Jasim, Dhifaf; Newman, Leon; Rodrigues, Artur Filipe; [et al.]. «The impact of graphene oxide sheet lateral dimensions on their pharmacokinetic and tissue distribution profiles in mice». Journal of Controlled Release, Vol. 338 (October 2021), p. 330-340. DOI 10.1016/j.jconrel.2021.08.028 This version is available at https://ddd.uab.cat/record/275369 under the terms of the license 1 2 3 4 The impact of graphene oxide sheet lateral dimensions on 5 their pharmacokinetic and tissue distribution profiles in mice 6 7 8 Dhifaf A. Jasima, Leon Newmana, Artur Filipe Rodriguesa, Isabella A. Vacchib, Matteo A. 9 Lucherelli b, Neus Lozanoa,c, Cécilia Ménard-Moyonb, Alberto Biancob, Kostas 10 Kostarelosa,c,* 11 12 aNanomedicine Lab, National Graphene Institute and Faculty of Biology, Medicine & Health, 13 University of Manchester, AV Hill Building, Manchester M13 9PT, United Kingdom 14 15 b University of Strasbourg, CNRS, Immunology, Immunopathology and Therapeutic Chemistry, UPR 16 3572, 67000 Strasbourg, France. 17 18 c Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, 19 Bellaterra, 08193 Barcelona, Spain 20 21 22 23 24 25 26 27 28 29 30 * Correspondence to: 31 kostas.kostarelos@manchester.ac.uk 32 33 1 1 Abstract 2 3 Although the use of graphene and 2-dimensional (2D) materials in biomedicine has been 4 explored for over a decade now, there are still significant knowledge gaps regarding the 5 fate of these materials upon interaction with living systems. Here, the pharmacokinetic 6 profile of graphene oxide (GO) sheets of three different lateral dimensions was studied. 7 The GO materials were functionalized with a PEGylated DOTA (1,4,7,10- 8 tetraazacyclododecane-1,4,7,10-tetraacetic acid), a radiometal chelating agent for 9 radioisotope attachment for single photon emission computed tomography (SPECT/CT) 10 imaging. Our results revealed that GO materials with three distinct size distributions, large 11 (l-GO-DOTA), small (s-GO-DOTA) and ultra-small (us-GO-DOTA), were sequestered by 12 the spleen and liver. Significant accumulation of the large material (l-GO-DOTA) in the 13 lungs was also observed, unlike the other two materials. Interestingly, there was extensive 14 urinary excretion of all three GO nanomaterials indicating that urinary excretion of these 15 structures was not affected by lateral dimensions. Comparing with previous studies, we 16 believe that the thickness of layered nanomaterials is the predominant factor that governs 17 their excretion rather than lateral size. However, the rate of urinary excretion was affected 18 by lateral size, with large GO excreting at slower rates. This study provides better 19 understanding of 2D materials behaviour with different structural features in vivo. 20 21 Key Words: Graphene oxide; functionalization; pharmacokinetics; nanomedicine, pharmacology 22 23 24 2 1 Introduction 2 3 Graphene and related flat-shaped materials exhibit outstanding properties 4 generated from their unique 2D geometry (1-3). These materials have attracted great 5 interest from different scientific disciplines (1-3). The biomedical applicability of graphene 6 has only been researched for the last decade (4). The available graphene surface area is 7 the largest for any material at the nanoscale, which provides a potential delivery platform 8 for maximum payload of therapeutic molecules and for bio-functionalization with imaging 9 probes (5-8). The electrical, electronic (9), mechanical, optical properties (8, 10, 11) and 10 flexibility of graphene based materials (2, 12) allow their use as biosensing platforms and 11 offer a great potential for use in regenerative medicine (13) and electroresponsive drug 12 therapy . It also can be useful to meet the requirements of the electroactive nervous and 13 cardiac systems and therefore provide a means of neuronal and cardiac drug delivery (14). 14 All these properties offer interesting possibilities of graphene materials following their 15 interaction with soft biological matter (11, 15). 16 Due to the great potential offered by using graphene materials for biomedical 17 applications, it is critical to understand their fate in vivo (16). Graphene oxide (GO) has 18 expanded the applications of graphene-based materials in biomedicine due to its 19 hydrophilicity and improved compatibility with biological systems. GO has been 20 administered intravenously (17-19), intraperitoneally, orally (20) and intravitreally (21) with 21 no reported toxic effects even after long exposure times (20, 21). After intravenous or 22 intraperitoneal administration, GO has been reported to accumulate in the mononuclear 23 phagocytic system, or as more commonly known the reticuloendothelial system (RES). 24 The spleen has been reported as the main site for the in vivo degradation of intravenously 25 injected functionalized graphene (22). GO materials have also been reported to 26 accumulate in the lungs (23, 24). Extensive urinary excretion of GO has been reported in 27 several studies after intravenous (i.v.) injection of functionalized GO sheets in mice (16, 28 20, 25-29). 29 Existing studies have used very different types of GO (30, 31) that can result in 30 significantly different biological interactions. These interactions will depend on the type of 31 surface functionalization, functional surface groups and dimensions of the GO sheets 32 (17). The most popular administration route used in the preclinical development of 33 materials for biomedical applications has been the intravenous (i.v.) route, which provides 34 100% bioavailability and therefore maximum information for tissue exposure and toxicity 35 (32). Many studies have used imaging for studying the biodistribution of graphene 36 materials by adding a labelling tag using single photon computed tomography (SPECT/CT) 37 with gamma (γ) emitting isotopes (28, 33), positron emission computed tomography 3 1 (PET/CT) with positron emitting isotopes (27, 34) or even fluorescence tags for optical 2 imaging (26, 34). 3 Previously we have demonstrated that the thickness of the graphene material plays 4 a major role in its tissue accumulation and excretion (27), however, no systematic 5 correlation with the GO sheet lateral dimension has been offered. In this study, three 6 highly purified and well characterised graphene oxide materials that differ only in their 7 lateral dimensions were functionalized with a chelating moiety for imaging purposes, 8 namely a DOTA molecule tethered to a polyethylene glycol (PEG) linker. The resulting 9 functionalized materials had three distinct sizes as well (l-GO-DOTA, s-GO-DOTA and us- 10 GO-DOTA). Then, whole-body imaging by SPECT/CT and full pharmacokinetic studies 11 were carried out following i.v. administration of the GO-DOTA materials, coupled with 12 analytical and histopathological analysis of critical organs. This study provided important 13 information on the future design of graphene materials for possible tissue targeting 14 applications. 15 16 Results 17 18 Preparation and characterisation of GO and GO-DOTA materials. GO was prepared 19 by a modified Hummers’ method under pyrogen-free conditions, as previously described 20 (35-37). Morphology of GO sheets was characterised by DLS, TEM, and AFM (Figure 1). 21 DLS revealed clear differences in the size distribution curves of the aqueous suspensions 22 of the three types of GO materials (l-GO, s-GO and us-GO). The sonication employed to 23 generate the different lateral size nanosheets (s-GO and us-GO) did not significantly 24 impact their surface charge (Figure 1A-B). 25 In order to perform biodistribution studies, the GO materials were functionalized with 26 the chelating agent DOTA, which was attached to a tetra-ethylene glycol [(PEG)4] 27 molecule bearing a free amine group [DOTA(PEG)4-NH2], as described in Scheme 1. The 28 functionalized chelating moiety [DOTA(PEG)4-NH2], is referred to as ‘DOTA’ for ease 29 throughout the manuscript. Although this process resulted in a decreased nanosheet 30 surface charge (Figure 1B), DOTA functionalization did not significantly impact the 31 colloidal stability of GO, in agreement with previous observations in physiological media 32 (28). DLS is not particularly suitable as a technique to determine the dimensions of non- 33 spherical particles (38) and we were not able to obtain reliable DLS data for the DOTA- 34 functionalized material. We emphasised on the structural characterisation of GO sheets 35 before and after DOTA functionalization, performed by TEM and AFM (Figure 1C-D). The 36 tested materials were composed of sheets with distinct differences in lateral dimensions 37 (i.e. the longest dimension observed of each 2D sheet), which are described in detail in 38 Table S1. 39 In summary, l-GO is comprised of sheets with lateral dimension between 1 and 35 40 μm, whilst s-GO sheets range between 30 nm and 1.9 μm. The smallest material (us-GO) 41 was characterised by a narrower size distribution, with lateral dimensions ranging from 10 42 nm to 550 nm. Upon functionalisation there was a reduction in the sheet lateral dimension, 43 particularly l-GO-DOTA compared to l-GO, where there was a 5-fold reduction (Table S1). 4 1 Nevertheless, the lateral dimensions of s-GO-DOTA and us-GO-DOTA sheets remained 2 markedly different from l-GO-DOTA and from each other. These differences could be 3 attributed to the ring opening reaction attacking cooperatively aligned epoxides present in 4 the GO sheet surface, which ultimately create fracture points (39, 40). AFM demonstrated 5 that the thickness of all GO materials had increased following DOTA functionalisation 6 (Table S1). This could be due to the addition of functional groups and has been observed 7 previously in several studies (20, 28, 41). 8 The samples were also characterised by X-ray photoelectron spectroscopy (XPS) 9 (Table S1 and Figure S1-3). The XPS survey table shows ~ 30-31% oxygen content in all 10 three starting materials. The oxygen content is slightly lower after functionalisation with 11 DOTA, with the introduction of 1.1-1.4% nitrogen content, due to the addition of the DOTA 12 molecule. Similarly, in the oxygen O1s high resolution spectra, there is a clear increase in 13 percentage of the carbonyl peak (Figure S2). This is due to the introduction of carboxylic 14 acids and amides. On the other hand, the carbon C1s high resolution spectra (Figure S1) 15 are more complex and revealed changes between the ratio of the carbon region and the 16 carbon-oxygen region. This is often the case after any kind of treatment of GO (42, 43) 17 due to the loss of some labile oxygenated functional groups. In particular, the component 18 attributed to C-O-C bonds decreased whereas the C-OH/C-N band increased. These 19 changes suggest that the reaction of epoxide ring opening occurred, thus generating C-OH 20 groups, alongside the introduction of amino groups from the DOTA compound, as 21 previously reported (44). However, the C-O-C peak did not disappear because of the ether 22 groups in the PEG chain. Two components could be identified in the nitrogen N1s high 23 resolution spectra, namely the amine/amide and the ammonium peaks at ~400 eV and 24 402 eV, respectively. Both amines and amides are present in the DOTA molecule, further 25 indicating the successful functionalisation of GO with the DOTA moiety (Figure S3). 26 Thermogravimetric analysis (TGA) also validated chemical functionalisation of GO 27 (Figure S4). GO starts to lose mass at temperatures even lower than 100°C due to 28 residual water molecules adsorbed on the GO sheets (37, 43, 45, 46).The two weight 29 losses above 150°C are due to the oxygenated species present on the surface of GO and 30 to the DOTA molecule in the case of the functionalized materials. The main weight loss 31 occurs at lower temperature (around 220°C) for the three GO-DOTA compared to the 32 starting materials (~240°C) was considered another indication of covalent functionalization 33 of the DOTA functionality. The %N obtained by XPS indicated that the three conjugates 34 have a similar DOTA loading. Taken together, these results show that the functionalization 35 of GO with the DOTA derivative was successful and did not cause significant reduction of 36 the starting material. 37 38 Efficiency and purity of radiolabeling of [111In]GO-DOTA. The efficiency of 39 radiolabelling of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us- 40 GO-DOTA) was compared to control [111In]DOTA in Figure 2A. The radiolabelling 41 efficiency of the three samples was ~ 70% at the application point. The samples were 42 purified by removal of unbound [111In]DOTA by centrifugation, reaching a purity of ~90% 43 after centrifugation (Figure 2B). 44 45 Stability of radiolabeling of [111In]GO-DOTA. The three samples had insignificant 46 variable stability in PBS at 370C up to 1 week as shown in Figure 2C, while the samples 47 were stable and retained their radiolabelling purity in 50% serum as shown in the same 48 figure and in Figure S5. The latter figure demonstrated that the three samples retained a 5 1 signal at the application point of the TLC compared to the control materials alone 2 [111In]EDTA and [111In]DOTA which moved to the solvent front. 3 4 Pharmacokinetics and tissue distribution after i.v. administration. The biodistribution 5 and excretion of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us- 6 GO-DOTA) after i.v. administration was studied by SPECT/CT imaging and cut-and-count 7 γ-scintigraphy (Figure 3, Figure 4 and Figure S6-9). The levels in blood determined by γ- 8 counting are shown in Figure 3A. The curve demonstrated that pharmacokinetics for all 9 materials followed two-compartment first order kinetics. The pharmacokinetic parameters 10 are summarised in Table S2 and the values remaining in blood after 1h are compared to 11 24h in Table S3. All three materials were removed from blood very rapidly with only less 12 than 1% of the injected dose remaining in the blood after 24h. Dynamic SPECT/CT 13 imaging was carried out during the i.v. administration of the materials (Figure S6). The first 14 panel shows the 2 min phase of injection (movie can be played online), showing the 15 delivery from the tail vein and through the vena cava then through whole blood. All three 16 materials ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA) and the control 17 [111In]DOTA start to accumulate in the organs within the first hour as demonstrated in the 18 planer images in the second panel in the same figure. Organ accumulation is very evident 19 after 4h and 24h (last two panels, Figure S6). Time activity curves for each material are 20 presented in Figure S7, indicating a huge accumulation of the larger material ([111In]l-GO- 21 DOTA) in the lungs, while the other two materials accumulated mainly in the liver and 22 spleen. All three materials presented bladder and kidney signals at early time points. The 23 control sample [111In]DOTA was totally excreted after 1h with huge bladder signal 24 compared to the other materials (Figure S7). 25 These results were further confirmed by a separate experiment using 3D 26 SPECT/CT for better image resolution as demonstrated in Figure 3B and Figure S8 for 27 the first batch of mice and Figure S9 for a second batch of mice. The scale bars are 28 expressed in percentage of injected dose (%ID) per gram of tissue in Figure 3B and 29 Figure S9, while it is expressed in MBq in Figure S8 and Figure S10. These images show 30 minor intestinal signals in all mice. Expression in %ID per gram of tissue will show the 31 minute amounts in blood levels in light organs. The total remaining amounts in the whole 32 body after 24 h quantified from the SPECT images are presented in Table S4, with 50.8%, 33 36.5% and 38.1% for [111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA, 34 respectively. 35 The whole-body SPECT/CT imaging data were validated by a separate cut and 36 count experiment counting the %ID per whole organ of [111In]GO-DOTA or %ID per gram 37 of tissue measured by γ-scintigraphy in a separate experiment with at least 4 mice per 38 condition (Figure 3C).The data confirmed the same pattern of the SPECT/CT data with 39 lung accumulation for the largest material ([111In]l-GO-DOTA) and liver and spleen 40 accumulation for the all three materials ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 41 [111In]us-GO-DOTA). Some kidney and bladder signals were detected at early time points, 42 while minute intestinal signal was evident at the later time points. Very little signals were 43 detected in the control sample (Figure 3C, bottom row). 44 45 Urinary and fecal excretion. Urinary and faecal excretion was studied by collecting the 46 urine and faeces of injected mice to further examine the extent of elimination of the three 47 materials (Figure 4). Pooled urine samples (n = 4) were collected and counted for 6 1 radioactivity at different time points (Figure 4A). The [111In]us-GO-DOTA showed the 2 maximum excretion compared to the other two samples. Furthermore the samples were 3 analysed by Raman spectroscopy as shown in Figure 4B. The Raman signature of GO- 4 DOTA was detected in the urine of all three materials, confirming the excretion of the 5 material regardless of the lateral size. Pooled faecal samples were collected after 24h and 6 measured for radioactivity, all samples indicated faecal excretion (Figure 4C). 7 8 Histopathology after i.v. administration of the three materials. Tissue samples (lung, 9 liver, spleen and kidneys) were examined for histopathology using H & E (haematoxylin 10 and eosin) staining of paraffin embedded tissue sections for two animals per condition 11 (Figure 5 and Figure S10-15). Lungs of mice injected with all the GO materials 12 demonstrated an interesting distribution pattern (Figure 5 and Figure S10). All GO 13 materials were predominantly detected as agglomerates within the lumen of lung blood 14 vessels, which indicates the retention of GO in the lung capillaries as a size-dependent 15 phenomenon, with the l-GO-DOTA material showing the largest, and the most 16 agglomerates, as quantified in Figure S11 and shown in different regions in Figure 5 and 17 Figure S10. In the case of the mice spleen samples, no evidence of histopathology was 18 determined in the red pulp (second panel) and the white pulp (third panel) in any of the 19 samples as compared to the controls (Figures S12-13). Liver tissue (Figure S14) and 20 kidney tissue (both glomerular and tubular regions) (Figure S15) also demonstrated 21 healthy anatomical structures with no evident histopathology in any of the samples. 22 23 24 Discussion 25 26 To date there is no direct method to quantitatively measure the amount of GO in 27 physiological fluids (e.g. tissues, blood and urine) due to the background interferences 28 from the complex molecular composition of such biological fluids. Studies rely on labelling 29 of the material for quantification, though other methods can offer qualitative detection such 30 as TEM and Raman spectroscopy. In this work, we studied the pharmacokinetic profile of 31 three lateral sizes of thin GO sheets. The functionalisation of GO with PEGylated DOTA 32 was performed to allow the chelation of a radioactive metal [111In] for studying and 33 quantifying the tissue distribution of the materials. The functionalisation of GO with the 34 PEGylated DOTA was performed via epoxide opening. Because of the abundance of 35 epoxides on the GO surface, this strategy generally leads to higher levels of 36 functionalization compared to the derivatisation of hydroxyls or carboxylic acids (47). 37 Structural characterisation was carried out by TEM and AFM to reveal whether GO 38 sheets would have marked differences in their morphology. The materials were produced 39 with controlled lateral dimensions by tuning their lateral size by sonication as previously 40 reported (37, 48-50). After functionalization all three materials maintained their 2D 41 morphology, while the thickness increased from single to a few layers following DOTA 42 functionalization due to the presence of the functional groups and minor agglomeration 43 that led to slightly thicker sheets. These findings are consistent with previous reports of 44 functionalization of GO with DOTA (27, 28) PEG (20, 51), dextran (25) and bovine serum 45 albumin (52). Interestingly, we noticed that l-GO underwent a marked reduction in sheet 46 lateral dimension after DOTA functionalization. Surface functionalization of GO has also 47 been reported to reduce the lateral dimension of the sheets (28, 52). However, the lateral 7 1 dimensions of all functionalized GO-DOTA materials remained markedly different from 2 each other. 3 The analysis of the XPS spectra (Table S1 and Figures S1-3) showed that the 4 functionalization of GO with the DOTA derivative was successful, as demonstrated by the 5 detection of nitrogen in the GO-DOTA materials compared to their starting counterparts. 6 Moreover, nitrogen atoms were detected in the form of amines and amides (Figure S3), 7 which are present in the DOTA molecule. The higher intensity of the carbonyl peak in the 8 carbon high resolution spectra after functionalization (Figure S1) is due to the carboxylic 9 acids of the DOTA molecule and it is also an indication of the presence of amides. On the 10 other hand, the changes in the carbon high resolution spectra after functionalization are 11 more subtle. First, there is a decreased abundance of oxidised carbon atoms compared to 12 the graphitic carbon region, due to the loss of some labile groups after functionalization of 13 GO. The contribution of the epoxides (C-O-C) is reduced compared to the C-OH/C-N 14 band, which is coherent with the introduction of the DOTA moiety by opening of epoxides, 15 thus generating C-OH and introducing amino groups. Chemical functionalization of GO 16 was further evidenced by analysing the different GO derivatives by TGA under an inert 17 atmosphere (Figure S4). GO is thermally unstable and starts to lose mass at temperatures 18 even lower than 100°C due to residual water molecules adsorbed on the GO sheets. A 19 main weight loss was identified at 230°C that is due to the elimination of labile oxygen- 20 containing groups (53). All GO-DOTA materials revealed a lower thermal stability probably 21 due to functionalization with DOTA. Therefore, taken together these results confirm the 22 effective grafting of DOTA on the surface of the three GO samples. In addition, the 23 structure of GO is preserved and the functionalization strategy induced no significant 24 reduction, thus enabling to obtain stable suspensions in aqueous solutions. 25 Other studies have reported radiolabelling of GO materials with iodine resulting in 26 iodinated constructs that are unstable, while the high affinity of iodine for the thyroid gland 27 can be misleading with regards to 2D material biodistribution profiles (25). Radiolabelling 28 with 111In has been performed previously using physical adsorption of the DTPA 29 (diethylenetriaminepentaacetic acid) chelating agent on the surface of GO by π-stacking. 30 Such strategy however increased the thickness of the GO sheets dramatically (33). The 31 radiolabelling efficiency and stability of covalently bound DOTA to GO materials has been 32 tested before with high efficiency and stability (27, 28). In this study a similar strategy was 33 exploited using GO materials with three different lateral dimensions (l-GO-DOTA, s-GO- 34 DOTA and us-GO-DOTA). The radiolabelling efficiency was not affected dramatically by 35 the lateral size of the GO sheets. All samples remained stable up to 24h at 370C in 50% 36 serum. Similar to original samples without serum incubation, the control [111In]DOTA 37 travelled to the solvent front as compared to the three samples that remained at the 38 bottom of the TLC (Figure S5). This indicated that the [111In]DOTA was attached to the GO 39 materials with no interferences occurring from the serum proteins at the times that were 40 tested (reflecting the time predicted for the material will spend in blood after injection and 41 beyond that time). This avoids any possible conflictions in the tissue distribution that are 42 created from the detachment of the [111In]DOTA. Therefore the purity of material was 43 confirmed to be suitable for in vivo administration. Furthermore the in vivo stability was 44 confirmed by the Raman signal that coregistered with the radioactivity at the bottom of the 45 TLC strips in the urine of the mice after injection as compared to the control free label that 46 was predominately at the top of the TLC strips (Figure 4B). This was comparable to the 47 original samples before injection in Figure 2B this further confirmed the stability and 48 association of the label with the GO material even after injection. 8 1 Compared to our previous studies (27-29) where the small and thin GO-DOTA 2 material was only injected, in this study, three materials with comparable thickness and the 3 three different sizes were injected intravenously in mice. All three materials were removed 4 from blood within minutes, as shown by the first-phase distribution half-life (t1/2α). Although 5 no significant differences were seen in the pharmacokinetic data of the materials, the 6 [111In]DOTA control was removed from blood faster. In the second phase the [111In]us-GO- 7 DOTA material was more similar to [111In]DOTA control and remained slightly longer in 8 circulation in the second phase half-life (t1/2β). The amounts remaining in blood after 1h 9 and 24h were more elevated for the [111In]s-GO-DOTA and [111In]us-GO-DOTA compared 10 to [111In]l-GO-DOTA (Figure 3A, Figure S5,Table S2 and Table S3). Smaller 11 nanoparticles are well known to circulate longer (54, 55).The area under the blood 12 concentration time curve (AUC), steady state volume of distributions (Vdss) (indicating the 13 body overall and tissue exposure) and the clearance values are shown in Table S2. The 14 [111In]DOTA control was almost entirely removed from blood at the early hours after 15 administration (Table S3). On the contrary, the remaining amounts in the whole body after 16 24h for the three materials was much higher compared to control material (Table S4). 17 It is clear from our results that tissue distribution occurs very rapidly for all three 18 materials and is largely affected by the size of the graphene sheet (Figure 3B-C and 19 Figure S6-9). It is very evident that the large material ([111In]l-GO-DOTA) tends to 20 accumulate in the lungs early after injection, with reductions in the signal after 24h as seen 21 by both the time activity curve and the γ-counting experiment (Figure S7 and Figure 3C, 22 respectively). Larger nanomaterials tend to accumulate in the lungs after i.v. administration 23 due to the first capillary bed (55). This has also been demonstrated with graphene 24 materials (17, 56, 57). Some material also remained in the lung tissues after 24h as seen 25 in the SPECT/CT data, γ-counting and in the H & E sections/semi-quantification (Figure 26 3B-C, Figure S6-9 and Figure S11). Some thickening of the alveolar walls was also seen 27 especially in lungs from mice injected with l-GO-DOTA and s-GO-DOTA, indicating cellular 28 infiltration and constriction, probably as a result of the material entrapment in the lung 29 tissues. We observed a clear reduction of the lung signal after 24h. It is known that the 30 removal of nanomaterials from the lung is carried out by phagocytic uptake of the lung 31 macrophages. This is the main mechanism to remove the insoluble aggregated 32 nanoparticles into the micrometer-sized particles from the lung tissues. Particle-containing 33 macrophages may re-enter into the interstitium and be cleared by the lymphatics or other 34 organs (58). 35 The smaller materials ([111In]s-GO-DOTA and [111In]us-GO-DOTA) accumulated 36 mainly in the liver and spleen with slight variability between animals in maximum 37 accumulations in these organs. The [111In]us-GO-DOTA showed a slight reduction in the 38 hepatic signal after 24h compared to the [111In]s-GO-DOTA that remained the same. On 39 the other hand, both materials accumulated in the spleen at high concentration even after 40 24h (Figure 3B-C and Figures S6-9). It is well known that nanomaterials, including 41 graphene-based materials, get trapped in the mononuclear phagocytic system (RES) (16, 42 25, 34, 59, 60). This happens within the liver due to the non-continuous liver endothelia 43 with vascular fenestrations measuring 50–100 nm, leading to nonspecific accumulation of 44 nanoparticles within this range. In the spleen, interendothelial cell slits with a size range of 45 200–500 nm mediate the retention of particles >200 nm (55). Particle shape also accounts 46 for accumulation in certain tissues, with elongated nanoparticles mainly in the spleen, 47 while the more spherical ones accumulate in the liver (61). The total amount of material 9 1 remaining in the body of mice injected with the [111In]l-GO-DOTA was the highest 2 compared to the other two materials (Table S3), which can be explained by the fact that 3 larger and thicker materials are more difficult to be cleared from the body (27). 4 Though there was slight variability between animals, the bladder measurements 5 obtained from the γ-counting experiments are largely dependent on the urine content in 6 the bladder that can largely vary. But, it is clear from the data obtained herein that the 7 graphene sheets were excreted through urine regardless of their lateral dimension (Figure 8 3B-C, Figure 4A-B, Figures S6-9). The Raman signature of GO-DOTA was detected in 9 the urine of injected mice confirmed the presence of intact sheets undergoing urinary 10 excretion. The kidney and bladder profiles in Figure 3B-C and Figure S6 suggested that 11 the larger sheets ([111In]l-GO-DOTA) were excreted at a slower rates compared to the 12 other two materials ([111In]s-GO-DOTA and [111In]us-GO-DOTA). Though, there could be a 13 possibility of a very small fraction of small sheets in all the three GO-DOTA samples 14 because of population of size distributions, however previous studies by our group and 15 others gave direct evidence of the excretion of much larger sheets than the kidney 16 glomerular filtration cut off due to their thin and flexible nature (29, 62). All materials had a 17 delayed excretion profile compared to the control probe, which was eliminated almost 18 entirely instantly. Gamma counting of pooled urine samples can sometimes be misleading 19 due to huge loss on the surfaces of the metabolic cages. Therefore, comparing the 20 quantities in urine to the control ([111In]DOTA) chelator sample is more realistic as the 21 chelating agents are predicted to be excreted almost entirely with minimal tissue 22 accumulation (63). We observed that the quantities excreted were much higher for the us- 23 GO-DOTA compared to the other two samples ([111In]l-GO-DOTA and [111In]s-GO-DOTA) 24 and close to the excretion of the control [111In]DOTA. Overall, these results show that all 25 three materials were excreted and the smaller material was excreted faster and to a larger 26 extent. 27 We and others have demonstrated previously that large, thin functionalized GO 28 sheets can cross the glomerular filtration barrier (GFB) although they have dimensions that 29 exceed the GFB cut off (16, 27-29, 62). However, this is the first time we demonstrate 30 urinary excretion of large GO sheets. Our histological data (Figure S16) resonates with 31 our previous results, illustrating the lack of damage to kidney regions. This suggested that 32 excretion of GO sheets occurred as a passive mechanism regardless of their size 33 (provided they are thin and flexible enough), possibly due to their morphological 34 reconfiguration of the sheets either by sliding, squeezing, rolling or folding (29). 35 The small intestinal signals in the SPECT/CT images at 24h and values in faeces 36 suggested slight faecal excretion, which could occur by biliary excretion, similarly to 37 previous findings with PEGylated GO (16), which bared PEG moieties functionalized GO- 38 DOTA materials. No damage to organs (spleen, liver and kidneys) was determined by our 39 histological H & E examinations for all three materials, consistent with many other previous 40 studies (16, 20, 64, 65). However, lung sections indicated some thickening of the alveolar 41 walls especially for the large material (Figure 5, S11 -12). However, all three GO-DOTA 42 materials were detected as agglomerates within the lumen of the lung blood vessels 43 (Figure S12) and the retention of material in the lung capillaries was a size-dependent 44 phenomenon. This retention and lung histopathological changes that can further lead to 45 fibrosis and lung damage at later time points have been observed in other studies using 46 large and non-functionalized materials (56, 66, 67) and requires further investigations. 10 1 This work suggests that the lateral dimension of the graphene material is critical to 2 determine the fate after i.v. administration. The multi-phase biodistribution profile of the 3 GO sheets studied here can be due to the wide size distribution of the GO sheets. We 4 envisage that the thin, flexible and small sheets tend to cross the glomerular filtration 5 barrier, while the larger and thicker sheets are preferentially entrapped in the lung. The 6 remaining smaller sheets were taken up by the liver and spleen cells. 7 In conclusion, GO was successfully covalently functionalized with the PEGylated 8 DOTA chelator with high chemical and radiolabelling stability regardless of the GO starting 9 lateral dimension. Intravenous administration of the GO-DOTA constructs led to rapid and 10 significant uptake by the lungs of the larger material, while the smaller materials targeted 11 the liver and spleen. All three radiolabelled GO-DOTA materials were excreted regardless 12 of their lateral dimension, indicating the possibility that the thickness rather than size may 13 be the major factor that governs the ability of GO to be excreted. Nevertheless, the lateral 14 dimension affected the rate of excretion. These findings provide further insight on the 15 kinetics and biodistribution of thin functionalized GO sheets with different lateral 16 dimensions after intravenous administration in mice. This has important implications in the 17 future design of graphene-based materials for biological applications and for targeting of 18 different organs. Indeed, large GO could be exploited for lung delivery, whereas smaller 19 dimensions could target the liver and spleen. Ultra-small GO is excreted to a larger extent, 20 which would make it attractive for imaging applications such as dynamic imaging where 21 the contrast agent needs to exit the body quickly after the imaging session, this facilitates 22 pharmacokinetic testing and reduces the number of animals required if used as a carrier 23 for image contrast agents (68). Due to the high surface area it could further be used to 24 deliver imaging probes and therapeutics simultaneously for thaeranostic applications. 25 Future studies are necessary to further assess the safety profile of graphene-based 26 nanomaterials and warrant direct quantitative detection techniques of the materials in the 27 complicated biological media for further validation. 28 29 30 11 1 Experimental 2 3 Chemical synthesis of GO. All reagents used in the production of GO were purchased from Sigma Aldrich 4 (UK), except water for injections, which was obtained from Dutscher Scientific (UK). GO was produced by a 5 modified Hummers’ method under endotoxin-free conditions (35-37). All glassware were depyrogenated prior 6 adding the mixture of 0.8 g of graphite flakes (Graflake 9580, kindly provided by Nacional de Grafite Ltda, 7 Brazil) with 0.4 g of sodium nitrate in a round-bottom flask that was placed in an ice bath, followed by the slow 8 addition of 18.4 mL of 99.999% sulphuric acid. When the mixture became homogenised, 2.4 g of potassium 9 permanganate was slowly added for 30 min. The dropwise addition of 37.5 mL of water for injections resulted 10 in a violent exothermic reaction, after which the temperature of the mixture was carefully maintained at 98oC 11 for 30 min. After the addition of 112.5 mL of water for injection, the reaction was stopped with12 mL of 30% 12 hydrogen peroxide, which reduced the unreacted potassium permanganate and manganese-based 13 intermediate products to manganese sulphate salts. The resulting graphite oxide suspension underwent a 14 series of exfoliation and centrifugation steps at 9000 rpm for 20 min, where the supernatant was replaced by 15 fresh water for injections, until the pH of the supernatant became neutral and a brownish gel-like layer was 16 formed on the top of the pellet. Warm water for injections was used to extract l-GO flakes from the graphite 17 pellet. In order to produce s-GO flakes, l-GO underwent a 5-min sonication step in a bath sonicator (VWR, UK) 18 operating at 80 W, which exfoliated and broke down the micrometre-sized flakes. The exfoliated suspensions 19 were centrifuged at 13000 rpm for 20 min at 20ºC, and the respective supernatants were carefully extracted, 20 containing only the small nanometre-sized flakes. For production of the us-GO flakes, the l-GO underwent a 21 similar process, involving a 4-hour sonication step followed by centrifugation at 13000 rpm for 1 h at 20 oC. 22 23 Preparation of GO-DOTA materials. NH2-PEG4-DOTA (CheMatech) (9 mg, 0.0129 mmol) was added to an 24 aqueous suspension of GO starting materials (us-GO, s-GO or l-GO) (9 mg, 1 mg/mL). The mixture was left to 25 react for 2 days under continuous stirring at room temperature. GO-DOTA was then directly dialyzed in MilliQ 26 water for 4 days. The final dialyzed GO-DOTA was stored at 7°C in water without further treatment. 27 28 Preparation of 111In labelled GO-DOTA and control DOTA materials. l-GO-DOTA, s-GO-DOTA, us-GO- 29 DOTA and control material DOTA(PEG)4-NH2 referred in this manuscript as ‘DOTA’ for ease, were diluted with 30 an equal volume of 0.2 M ammonium acetate buffer pH 5.5, to which 10-20 MBq of 111InCl3 (CURIUMTM 31 previously, Mallinckrodt RP) was added, in separate experiments. The indium was left to react with the GO- 32 DOTA and DOTA control for 60 min at 60oC with vortexing every 5 min after which the reaction was quenched 33 by the addition of 0.1 M EDTA chelating solution. The radiolabelling was carried out in several separate 34 experiments at different occasions and with different operators. 35 36 Radiolabelling efficiency of GO-DOTA materials. To determine the labelling efficiency of each final product, 37 aliquots were diluted five folds in PBS and then 1 μL spotted on silica gel impregnated glass fibre sheets 38 (PALL Life Sciences, UK). The strips were developed with a mobile phase of 50 mM EDTA in 0.1 M 39 ammonium acetate and allowed to dry before analysis. The strips were developed with a mobile phase of 25 40 mM EDTA in 0.1 M ammonium acetate and allowed to dry before analysis. This was then developed and the 41 autoradioactivity counted using a Fujifilm fluorescence image analyser (FLA-3000 series, Tokyo, Japan). The 42 immobile spot on the TLC strips indicated the percentage of radiolabeled GO-DOTA, while free 43 [111In]DOTA(PEG)4-NH2, referred throughout this manuscript as [111In]DOTA for ease was seen as the mobile 44 spots near the solvent front. 45 46 Radiolabelling stability of GO-DOTA materials. To determine the stability of the labeled [111In]GO-DOTA 47 materials, aliquots of each final product were diluted five-fold either in PBS or 50% mouse serum and then 48 incubated at 37˚C over 1 week. At different time points (0h, 24h, 48h and 7 days), 1 μL of the aliquots was 49 spotted on silica gel impregnated glass fibre sheets and then developed, and quantified as described above. 50 51 TEM analysis. 20 µL of sample (200 µg/mL) was dropped on a glow-discharged carbon-coated copper grid 52 (CF400-Cu) (Electron Microscopy Services, UK). Filter paper (Merck-Millipore, UK) was used to absorb the 53 unbound material. Samples were then observed with a FEI Tecnai 12 BioTWIN microscope (Techni, 54 Netherlands) with an acceleration voltage of 100 kV. Images were taken using a Gatan Orius SC1000 CCD 55 camera (GATAN, UK). Size using TEM distributions were performed as described previously (37, 49, 50). 56 12 1 AFM analysis. AFM was performed on freshly cleaved mica, treated with 40 μL of poly-L-lysine (Sigma- 2 Aldrich, UK) to present a positively charged surface, which facilitated the adhesion of the GO sheets. Aliquots 3 of 10 μL were then transferred of the respective suspensions onto the mica surface coated with poly-L-lysine 4 and left to adsorb for 2 min. Unbound GO sheets were then removed via gentle washing with 2 mL of MilliQ 5 H2O and left to dry at room temperature. During analysis, a Multimode atomic force microscopy (Bruker, UK) 6 was used in tapping mode in order to reduce damage to the samples for height (trace and retrace) and 7 amplitude. Scans were completed using an Otespa tapping mode tip (Bruker, UK) using the following 8 parameters: a scan rate of 1 Hz; lines per scan of 512; an integral gain of 1 and a proportional gain of 5; an 9 amplitude set point value of 150 mV was maintained constant between all measurements. Scan areas were 10 set at 2500 μm2, 400 μm2, 100 μm2, and 25 μm2. Post image processing was completed using the Bruker 11 Nanoscope Analysis software (version 1.4, Bruker, UK). Size and thickness distributions were performed as 12 described previously (37, 49, 50). 13 14 XPS analysis. The three GO-DOTA aqueous suspensions were drop-casted on silicon wafers and dried 15 overnight. XPS analysis was carried out with a Thermo Scientific K-ALPHA monochromatic photoelectron 16 spectrometer with a basic chamber pressure of 10-8-10-9 bar and an Al anode as X-ray source (1486 eV). A 17 spot size of 400 µm was selected. Every sample was analyzed three times. The survey spectra are an average 18 of 10 scans taken with a pass energy of 200.00 eV and a step size of 1 eV. The high resolution spectra are an 19 average of 10 scans taken with a pass energy of 50.00 eV and a step size of 0.1 eV. The pass energy of 50.00 20 eV corresponds to Ag 3d5/2 line FWHM of 1.3 eV. A pass energy of 50.00 eV for the high resolution spectra 21 was applied because using lower pass energies has shown no improvement in FWHM for graphene materials 22 on Thermo Scientific K-ALPHA. An electron flood gun was turned on during analysis as charge neutralizer. 23 This electron flood gun was not totally efficient for charge compensation, thus an error due to charging was still 24 identified in some samples. 25 Data analysis casaXPS (2.3.18) software was used. A Shirley background subtraction and charge correction 26 were applied. No reference element was analysed, thus, the spectra were corrected to the C-C peak at 285.2- 27 285.3 eV, if needed. Error due to this charge correction was taken into account. A line-shape 70% 28 Gaussian/30% Lorenzian [GL(30)] is selected for all peaks beside for sp2 C peak. For sp2 C peak an 29 asymmetric line-shape was chosen [A(0.4, 0.38, 20)GL(20)]. FWHM was constrained to be the same for all 30 peaks, beside the pi-pi* peak and the water peak because these are broad signals. The peak position for each 31 bond was constrained to be the same in every spectra, with a low ratio of variability. Due to the difficulty to take 32 standards for carbon nanomaterials, database and reference articles were used as peak reference value. For 33 the carbon high resolution spectra almost every bond was considered individually, keeping into consideration 34 the error due to the proximity of the binding energy (BE) values and proportional to the overlapping ratio. Only 35 C-OH/C-N and C=OOH/C=ONR2 were cumulated under the same peak due to the higher proximity of the BE 36 values. For oxygen and nitrogen high resolution spectra all signals were assembled due to the high proximity 37 of the BE values instead. Also with this approximation a certain degree of error has to be taken into account 38 (69-71). 39 40 Thermogravimetric analysis. TGA was performed using a TGA1 (Mettler Toledo) apparatus from 30°C to 41 900°C with a ramp of 10°C/min under N2 using a flow rate of 50 mL/min and platinum pans. 42 43 Raman spectroscopy of GO, GO-DOTA samples in solution and GO-DOTA samples in urine. Samples 44 were prepared for analysis via drop casting 20 μL of GO (100 µg/mL) dispersion onto a glass slide. Samples 45 were left to dry for at least 2 h at 37°C. For the identification of GO in the urine, 5 μL of collected urine were 46 spotted on new silica gel impregnated glass fibre sheets and then developed as described above. The 47 developed TCL plates were then left for radioactivity decay for a month. The application points on each TCL 48 plate were screened for GO using Raman spectroscopy. As Raman is a qualitative method of detection and 49 urine components prove to be masking the GO signal at low concentrations, the 24h collected urine was only 50 chosen where there was the maximum amount of GO excreted for facilitating the detection of the signal. 51 Spectra were acquired using a micro-Raman spectrometer (Thermo Scientific, UK) using a λ = 633 nm laser at 52 0.4 mW with an exposure time of 25 s at a magnification of 50x. Spectra were averaged over 5 locations and 53 considered between 500–3500 cm-1. Post spectral processing included background correction and baseline 54 subtraction, followed by the calculation of the I(D)/I(G) intensity ratio, this was completed using 55 OriginProsoftware (version 8.5.1, Origin Lab, USA). 56 57 Animal handling procedures. Six- to eight-week-old C57BL6 mice (18.7 ± 2.4 g) were obtained from Envigo 58 (ex-Harlan) (Oxfordshire, UK), allowed to acclimatize for 1 week and were given access to food and water for 13 1 the duration of the experiments. All procedures were carried out in accordance with the UK Home Office Code 2 of Practice (1989) for the housing and care of animals in scientific procedures, in accordance with prior 3 approval from the UK Home Office. All mice were injected with a single injection of either l-GO, s-GO, us-GO, 4 l-GO-DOTA, s-GO-DOTA or us-GO-DOTA (2.5 mg/kg) or control materials such as dextrose 5% or 5 [111In]DOTA in 200 µL by either i.v. bolus injection or infusion through the tail vein. 6 7 SPECT/CT. Mice were subjected to anaesthesia via the inhalation of 2.5% isoflurane where oxygen was used 8 as a carrier gas set at a flow rate of 2 L/min. Each animal was then intravenously injected with the radioactive 9 materials (either [111In]l-GO-DOTA, [111In]s-GO-DOTA, [111In]us-GO-DOTA or [111In]DOTA) at a dose of 2.5 10 mg/kg (200 µL, 5-10 MBq). The injections were carried out by i.v. infusions through tail vein catheterisation in 11 order to enable the dynamic imaging. At 0-1h and 24 h SPECT/CT imaging was carried out using a Nano- 12 Scan® SPECT/CT scanner (Mediso, Hungary). SPECT images were obtained in 20 projections over 40-60 min 13 using a 4-head scanner with 1.4 mm pinhole collimators for the 3D pinSPECT images and a custom made 3 14 mm single pinhole collimator for the dynamic (time laps) and static 2D images. Due to practicalities of 15 radioisotope decay and large number of samples tested it was decided to run the fast 2D static scan at the 16 mid-point 4h time point rather than the full 3D pinSPECT scans that take much longer time. CT scans were 17 taken at the end of each SPECT acquisition using a semi-circular method with full scan, 480 projections, 18 maximum FOV, 35 kV energy, 300 ms exposure time and 1-4 binning. Acquisitions were done using the 19 Nucline v2.01 (Build 020.0000) software (Mediso, Hungary), while reconstruction of all images and fusion of 20 SPECT with CT images was performed using the Interview™ FUSION bulletin software (Mediso, Hungary). 21 The images were further analysed using VivoQuant 3.0 software (Boston, US) where the SPECT images with 22 scale bars in MBq were corrected for decay and for the slight differences in radioactivity in the injected doses 23 between animals. Images containing scale bars in %ID per gram were processed automatically by the 24 VivoQuant software after providing the injected dose in MBq. 25 26 Gamma scintigraphy. For more quantitative assessment of tissue biodistribution, a cut and count study was 27 carried out. Mice were anaesthetised by isoflurane inhalation. Each animal was injected in the tail vein with 28 one of the radiolabelled constructs either [111In]l-GO-DOTA, [111In]s-GO-DOTA, [111In]us-GO-DOTA or 29 [111In]DOTA 200 μL containing 2.5 mg/kg containing approximately 5-10 MBq. Mice were sacrificed at 1, 4 and 30 24 h after injection. Blood and all major organs and tissues were collected including, heart, lungs, liver, spleen, 31 kidneys, muscle, skin and bone. Urine and faeces were pooled and collected as well. Each sample was 32 weighted and counted on a γ-counter (Perkin Elmer, USA), together with a dilution of the injected dose with 33 dead time limit below 60%. The percentage injected dose per gram tissue was calculated, using four different 34 mice for each time point. 35 36 Histological analysis. Lungs, liver, spleen and kidneys were extracted from mice after 24 h and fixed with 4% 37 paraformaldehyde. This was followed by paraffin embedding of sections at known orientations. Sections of 5 38 µm were stained with H & E and imaged using a 3D Histech Pannoramic 250 Flash slide scanner. Images 39 were processed and analysed using Pannoramic Viewer (http://www.3dhistech.com/) and Fiji/ImageJ software 40 (version 1.5c; National Institutes of Health, Bethesda, MD). 41 42 Statistical analysis. Due to different sample sizes and the non-Gaussian distribution of the flake populations, 43 the TEM size distribution data were presented using boxplots and the median, minimum and maximum values 44 of each distribution were reported. A non-parametric test (Wilcoxon rank sum test) was performed using the 45 statistical package in MATLAB (version R2013a, MathWorks Inc., USA), in order to determine the statistical 46 significance of the difference between the lateral dimensions of the GO materials. Experiments were 47 completed with an n = 3-4 repeats and values are mean ± standard deviation (SD). Animal experiments were 48 completed with at least two repeats for each imaging and histological experiments and n = 4 repeats for the 49 quantitative experiments and values are mean ± standard error of the mean (SEM). One-way ANOVA 50 with Tukey’s posthoc test using IBM SPSS statistics 25 software were used when considering multiple 51 comparisons for the pharmacokinetic data. 52 53 54 55 14 1 Author Contributions 2 D.J. designed, planned, and led the study, performing all of the animal experiments, 3 imaging, histopathology, radiolabelling studies and data analysis, prepared the figures and 4 wrote the manuscript. L.N. contributed to the characterisation of GO and GO-DOTA using 5 TEM and AFM. He performed the Raman spectroscopy experiments and helped perform 6 the radiolabelling experiments. A.R. and N.L. prepared the initial GO and performed the 7 physicochemical characterisation of GO and GO-DOTA samples. A.R. also helped analyse 8 the lung histopathology data. I.V. and M. L.. performed the chemical functionalisation of 9 GO with DOTA and performed the XPS and TGA experiments and analysis. C.M. 10 designed and supervised the chemical reaction of GO with DOTA. A.B. provided 11 continuous guidance in the conceptual design of the work and data interpretation. K.K. 12 designed, planned, and discussed the findings, reviewed and edited the manuscript, and 13 overall supervised the work. D.J., L.N., A.R., I.V., M. L., C.M., A.B. and K.K. discussed the 14 findings and contributed to writing, reviewing, and editing the manuscript. 15 16 17 Acknowledgments 18 This work was supported by the EU 7th RTD Framework Programme, Graphene Flagship 19 project (FP7-ICT-2013-FET-F, Project no. 604391). This work was also supported by the 20 Agence Nationale de la Recherche (ANR) through the LabEx project Chemistry of 21 Complex Systems (ANR-10-LABX-0026_CSC). We are very thankful to the staff at St. 22 Mary’s hospital radiopharmacy for help in delivering the radioisotopes. We are also very 23 thankful to Mr Peter Walker at the University of Manchester histology facility for his advice 24 and assistance in the histology analysis. We would like to acknowledge the staff from the 25 EM Facility in the Faculty of Biology, Medicine and Health, particularly Dr Aleksandr 26 Mironov and Ms Samantha Forbes for their expert advice and assistance. We thank Dr 27 Nigel Hodson from the AFM Unit within the BioImaging Facility in the Faculty of Biology, 28 Medicine and Health (University of Manchester) for assistance and advice regarding the 29 AFM instrumentation. 30 15 1 References 2 1. K. Kostarelos, K. S. Novoselov, Graphene devices for life. Nature Nanotechnology 9, 744 (2014). 3 2. Y. Pan, N. G. Sahoo, L. Li, The application of graphene oxide in drug delivery. Expert Opinion on Drug 4 Delivery 9, 1365-1376 (2012). 5 3. P. Avouris, F. Xia, Graphene applications in electronics and photonics. MRS Bulletin 37, 1225-1234 6 (2012). 7 4. Z. Liu, J. T. Robinson, X. Sun, H. Dai, PEGylated nanographene oxide for delivery of water-insoluble 8 cancer drugs. Journal of Americal Chemical Society 130, 10876-10877 (2008). 9 5. K. Kostarelos, K. S. Novoselov, Exploring the interface of graphene and biology. Science 344, 261-263 10 (2014). 11 6. K. P. Loh, Q. Bao, G. Eda, M. Chhowalla, Graphene oxide as a chemically tunable platform for optical 12 applications. Nature Chemistry 2, 1015-1024 (2010). 13 7. L. Feng, Z. Liu, Graphene in biomedicine: opportunities and challenges. Nanomedicine 6, 317-324 14 (2011). 15 8. H. Shen, L. Zhang, M. Liu, Z. Zhang, Biomedical Applications of Graphene. Theranostics 2, 283 -294 16 (2012). 17 9. I. Calizo, I. Bejenari, M. Rahman, G. Liu, A. A. Balandin, Ultraviolet Raman microscopy of single and 18 multilayer graphene. Journal of Applied Physics 106, 043509 (2009). 19 10. A. C. Ferrari et al., Raman Spectrum of Graphene and Graphene Layers. Physical Review Letters 97, 20 187401 (2006). 21 11. A. Bendali et al., Purified Neurons can Survive on Peptide-Free Graphene Layers. Advanced Healthcare 22 Materials 10.1002/adhm.201200347, 929-933 (2013). 23 12. K. S. Novoselov et al., A roadmap for graphene. Nature 490, 192-200 (2012). 24 13. M. Zhou et al., Graphene oxide: A growth factor delivery carrier to enhance chondrogenic differentiation 25 of human mesenchymal stem cells in 3D hydrogels. Acta Biomaterialia 96, 271-280 (2019). 26 14. N. A. Kotov et al., Nanomaterials for Neural Interfaces. Advanced Materials 21, 3970-4004 (2009). 27 15. S. K. Seidlits, J. Y. Lee, C. E. Schmidt, Nanostructured scaffolds for neural applications. Nanomedicine 28 (London, England) 3, 183-199 (2008). 29 16. K. Yang et al., In Vivo Pharmacokinetics, Long-Term Biodistribution, and Toxicology of PEGylated 30 Graphene in Mice. ACS Nano 5, 516-522 (2011). 31 17. J. H. Liu et al., Effect of size and dose on the biodistribution of graphene oxide in mice. Nanomedicine 32 (London, England) 7, 1801-1812 (2012). 33 18. L. Zhan et al., Biodistribution of co-exposure to multi-walled carbon nanotubes and graphene oxide 34 nanoplatelets radiotracers. J Nanopart Res 13, 2939-2947 (2011). 35 19. G. Qu et al., The ex vivo and in vivo biological performances of graphene oxide and the impact of 36 surfactant on graphene oxide's biocompatibility. Journal of Environmental Sciences 25, 873-881 (2013). 37 20. K. Yang et al., In vivo biodistribution and toxicology of functionalized nano-graphene oxide in mice after 38 oral and intraperitoneal administration. Biomaterials 34, 2787-2795 (2013). 39 21. L. Yan et al., Can graphene oxide cause damage to eyesight? Chemical Research Toxicology 25, 1265- 40 1270 (2012). 41 22. C. M. Girish, A. Sasidharan, G. S. Gowd, S. Nair, M. Koyakutty, Confocal Raman Imaging Study Showing 42 Macrophage Mediated Biodegradation of Graphene In Vivo. Advanced Healthcare Materials 2, 1489- 43 1500 (2013). 44 23. K. Wang et al., Biocompatibility of Graphene Oxide. Nanoscale Research Letters 6, 8-8 (2011). 45 24. X. Zhang et al., Distribution and biocompatibility studies of graphene oxide in mice after intravenous 46 administration. Carbon 49, 986-995 (2011). 47 25. S. Zhang, K. Yang, L. Feng, Z. Liu, In vitro and in vivo behaviors of dextran functionalized graphene. 48 Carbon 49, 4040-4049 (2011). 49 26. K. Yang et al., Graphene in Mice: Ultrahigh In Vivo Tumor Uptake and Efficient Photothermal Therapy. 50 Nano Letters 10, 3318-3323 (2010). 51 27. D. A. Jasim et al., Thickness of functionalized graphene oxide sheets plays critical role in tissue 52 accumulation and urinary excretion: A pilot PET/CT study. Applied Materials Today 4, 24-30 (2016). 53 28. D. A. Jasim, C. Menard-Moyon, D. Begin, A. Bianco, K. Kostarelos, Tissue distribution and urinary 54 excretion of intravenously administered chemically functionalized graphene oxide sheets. Chemical 55 Science 6, 3952-3964 (2015). 56 29. D. A. Jasim et al., The Effects of Extensive Glomerular Filtration of Thin Graphene Oxide Sheets on 57 Kidney Physiology. ACS Nano 10, 10753-10767 (2016). 58 30. A. Bianco, Graphene: Safe or Toxic? The Two Faces of the Medal. Angewandte Chemie International 59 Edition 52, 4986-4997 (2013). 60 31. P. Wick et al., Classification Framework for Graphene-Based Materials. Angewandte Chemie 61 International Edition 10.1002/anie.201403335, 2-7 (2014). 62 32. R. P. Heaney, Factors influencing the measurement of bioavailability, taking calcium as a model. Journal 63 of Nutrition 131, 1344S-1348S (2001). 64 33. B. Cornelissen et al., Nanographene oxide-based radioimmunoconstructs for in vivo targeting and 65 SPECT imaging of HER2-positive tumors. Biomaterials 34, 1146-1154 (2013). 16 1 34. H. Hong et al., In Vivo Targeting and Imaging of Tumor Vasculature with Radiolabeled, Antibody- 2 Conjugated Nanographene. ACS Nano 6, 2361-2370 (2012). 3 35. D. A. Jasim, N. Lozano, K. Kostarelos, Synthesis of few-layered, high-purity graphene oxide sheets from 4 different graphite sources for biology. 2D Materials 3, 014006 (2016). 5 36. S. P. Mukherjee et al., Detection of Endotoxin Contamination of Graphene Based Materials Using the 6 TNF-α Expression Test and Guidelines for Endotoxin-Free Graphene Oxide Production. PLOS ONE 11, 7 e0166816 (2016). 8 37. A. F. Rodrigues et al., A blueprint for the synthesis and characterisation of thin graphene oxide with 9 controlled lateral dimensions for biomedicine. 2D Materials 5, 035020 (2018). 10 38. S. Bhattacharjee, DLS and zeta potential - What they are and what they are not? Journal of controlled 11 release : official journal of the Controlled Release Society 235, 337-351 (2016). 12 39. J. L. Li et al., Oxygen-driven unzipping of graphitic materials. Phys Rev Lett 96, 176101 (2006). 13 40. T. Sun, S. Fabris, Mechanisms for oxidative unzipping and cutting of graphene. Nano Lett 12, 17-21 14 (2012). 15 41. W. Zhang et al., Unraveling stress-induced toxicity properties of graphene oxide and the underlying 16 mechanism. Advanced materials (Deerfield Beach, Fla.) 24, 5391-5397 (2012). 17 42. A. M. Dimiev, L. B. Alemany, J. M. Tour, Graphene oxide. Origin of acidity, its instability in water, and a 18 new dynamic structural model. ACS Nano 7, 576-588 (2013). 19 43. S. Eigler, C. Dotzer, A. Hirsch, M. Enzelberger, P. Müller, Formation and Decomposition of CO2 20 Intercalated Graphene Oxide. Chemistry of Materials 24, 1276-1282 (2012). 21 44. I. A. Vacchi, C. Spinato, J. Raya, A. Bianco, C. Menard-Moyon, Chemical reactivity of graphene oxide 22 towards amines elucidated by solid-state NMR. Nanoscale 8, 13714-13721 (2016). 23 45. S. Stankovich et al., Synthesis of graphene-based nanosheets via chemical reduction of exfoliated 24 graphite oxide. Carbon 45, 1558-1565 (2007). 25 46. J. I. Paredes, S. Villar-Rodil, A. Martínez-Alonso, J. M. Tascón, Graphene oxide dispersions in organic 26 solvents. Langmuir : the ACS journal of surfaces and colloids 24, 10560-10564 (2008). 27 47. I. A. Vacchi, S. Guo, J. Raya, A. Bianco, C. Ménard-Moyon, Strategies for the Controlled Covalent 28 Double Functionalization of Graphene Oxide. Chemistry – A European Journal 26, 6591-6598 (2020). 29 48. M. Orecchioni et al., Molecular and Genomic Impact of Large and Small Lateral Dimension Graphene 30 Oxide Sheets on Human Immune Cells from Healthy Donors. Advanced Healthcare Materials 5, 276-287 31 (2016). 32 49. A. F. Rodrigues et al., Size-Dependent Pulmonary Impact of Thin Graphene Oxide Sheets in Mice: 33 Toward Safe-by-Design. Advanced Science n/a, 1903200. 34 50. L. Newman et al., Nose-to-Brain Translocation and Cerebral Biodegradation of Thin Graphene Oxide 35 Nanosheets. Cell Reports Physical Science 1, 100176 (2020). 36 51. W. Zhang et al., Unraveling Stress-Induced Toxicity Properties of Graphene Oxide and the Underlying 37 Mechanism. Advanced Materials 24, 5391-5397 (2012). 38 52. Y. Li et al., Surface Coating-Dependent Cytotoxicity and Degradation of Graphene Derivatives: Towards 39 the Design of Non-Toxic, Degradable Nano-Graphene. Small 10.1002/smll.201303234, 1544-1554 40 (2013). 41 53. I. Jung et al., Reduction Kinetics of Graphene Oxide Determined by Electrical Transport Measurements 42 and Temperature Programmed Desorption. The Journal of Physical Chemistry C 113, 18480-18486 43 (2009). 44 54. F. Alexis, E. Pridgen, L. K. Molnar, O. C. Farokhzad, Factors Affecting the Clearance and Biodistribution 45 of Polymeric Nanoparticles. Molecular Pharmaceutics 5, 505-515 (2008). 46 55. E. Blanco, H. Shen, M. Ferrari, Principles of nanoparticle design for overcoming biological barriers to drug 47 delivery. Nature Biotechnology 33, 941-951 (2015). 48 56. K. Wang et al., Biocompatibility of graphene oxide. Nanoscale Res Lett 6, 1-8 (2011). 49 57. S. K. Singh et al., Thrombus Inducing Property of Atomically Thin Graphene Oxide Sheets. ACS Nano 5, 50 4987-4996 (2011). 51 58. M. Geiser, Update on macrophage clearance of inhaled micro- and nanoparticles. Journal of Aerosol 52 Medicine Pulmonary Drug Delivery 23, 207-217 (2010). 53 59. K. Yang et al., The influence of surface chemistry and size of nanoscale graphene oxide on photothermal 54 therapy of cancer using ultra-low laser power. Biomaterials 33, 2206-2214 (2012). 55 60. H. Hong et al., In vivo targeting and positron emission tomography imaging of tumor vasculature with 56 (66)Ga-labeled nano-graphene. Biomaterials 33, 4147-4156 (2012). 57 61. M. Zhang et al., Radiolabeling, whole-body single photon emission computed tomography/computed 58 tomography imaging, and pharmacokinetics of carbon nanohorns in mice. International Journal of 59 Nanomedicine 11, 3317-3330 (2016). 60 62. S. Liang et al., In vivo pharmacokinetics, transfer and clearance study of graphene oxide by La/Ce dual 61 elemental labelling method. NanoImpact 17, 100213 (2020). 62 63. B. T. Kurien, N. E. Everds, R. H. Scofield, Experimental animal urine collection: a review. Lab Animals 38, 63 333-361 (2004). 64 64. B. Li et al., Influence of polyethylene glycol coating on biodistribution and toxicity of nanoscale graphene 65 oxide in mice after intravenous injection. Intenational Journal of Nanomedicine 9, 4697-4707 (2014). 66 65. L. Newman et al., Splenic Capture and In Vivo Intracellular Biodegradation of Biological-Grade Graphene 67 Oxide Sheets. ACS Nano 14, 10168-10186 (2020). 17 1 66. M. C. Duch et al., Minimizing Oxidation and Stable Nanoscale Dispersion Improves the Biocompatibility of 2 Graphene in the Lung. Nano Letters 11, 5201-5207 (2011). 3 67. L. Ma-Hock et al., Comparative inhalation toxicity of multi-wall carbon nanotubes, graphene, graphite 4 nanoplatelets and low surface carbon black. Particle and Fibre Toxicology 10, 23 (2013). 5 68. S. Macholl et al., High-throughput high-volume nuclear imaging for preclinical in vivo compound 6 screening(§). EJNMMI Research 7, 33 (2017). 7 69. A. Ganguly, S. Sharma, P. Papakonstantinou, J. Hamilton, Probing the Thermal Deoxygenation of 8 Graphene Oxide Using High-Resolution In Situ X-ray-Based Spectroscopies. The Journal of Physical 9 Chemistry C 115, 17009-17019 (2011). 10 70. L.-N. Zhou, X.-T. Zhang, W.-J. Shen, S.-G. Sun, Y.-J. Li, Monolayer of close-packed Pt nanocrystals on a 11 reduced graphene oxide (RGO) nanosheet and its enhanced catalytic performance towards methanol 12 electrooxidation. RSC Advances 5, 46017-46025 (2015). 13 71. C. Botas et al., Tailored graphene materials by chemical reduction of graphene oxides of different atomic 14 structure. RSC Advances 2, 9643-9650 (2012). 15 16 18 1 Figure Legends 2 3 Scheme 1. Synthesis of GO-DOTA. For the sake of clarity, electrostatic interactions between the 4 protonated amine in NH2-PEG4-DOTA and the carboxylate groups at the edges of GO or in the 5 PEG4-DOTA chain linked to GO are not shown. 6 Figure 1. Physicochemical characterisation of l-GO, s-GO and us-GO before and after DOTA 7 functionalisation. A) Dynamic light scattering (DLS) size distributions; B) Electrophoretic mobility 8 (ζ potential) mean surface charge data. Morphological and structural characterisation data is 9 shown, using C) TEM and D) AFM. 10 Figure 2: Radiolabelling efficiency and stability. A) Efficiency of radiolabelling of the three types 11 of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA) compared to control 12 [111In]DOTA after the radiolabelling reaction; values indicate the average of three independent 13 labelling repeats. B) Radiolabelling purity of the three samples after centrifugation and before 14 administration in the animals. C) Stability of radiolabelling in PBS and 50% serum up to 7 days 15 Figure 3: Biodistribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 16 [111In]us-GO-DOTA) compared to [111In]DOTA control at 1h and 24h. A) Blood profile up to 24h. 17 B) SPECT/CT images expressed as % of injected dose per gram of tissues (doses were decay 18 corrected at the second time point). From left to right (whole body maximum intensity projections 19 (MIP), sagittal, coronal and transverse views). Interactive 3D images of the MIPs are available 20 online. C) Organ distribution of the three materials as determined by γ-counting. Four animals per 21 group were used for A and C, while two animals were used for the imaging in B. 22 Figure 4: Excretion profile of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA 23 compared to control [111In]DOTA). A) Urinary excretion profile of the three materials at different 24 time points B) detection of the graphene material in the urine 24 hours post administration of mice 25 as demonstrated by radio-TLC and corroborative Raman spectroscopy, scale bars are 20 µm. C) 26 Faecal excretion of the three materials compared to the control. 27 Figure 5. Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on lung compared to control 5% 28 dextrose. Haematoxylin and eosin stained lung sections (5 µm thick) after injection of l-GO-DOTA, 29 s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h of two different mice for 30 each material. Scale bars for the images on the left are 100 µm while those on the right are 20 µm. 31 32 33 34 35 36 37 19 1 Figures 2 3 4 5 Scheme 1. Synthesis of GO-DOTA. For the sake of clarity, electrostatic interactions 6 between the protonated amine in NH2-PEG4-DOTA and the carboxylate groups at the 7 edges of GO or in the PEG4-DOTA chain linked to GO are not shown. 8 9 10 DOTA(PEG)4 O O HO HN DOTA HO O NH2 4 N N 4 O O O N HN O HO N N H O O OH OH OH O HN O OH OH OH OH 4 HO NH 4 OH O OH HO O HO OH O OH O DOTA N O H GO O NH DOTA GO-DOTA 11 12 13 14 20 1 Figure 1 2 3 A 35 C Large GO Small GO Ultra-small-GO 30 l-GO 25 s-GO % Intensity 20 GO us-GO 15 2 μm 500 nm 500 nm TEM 10 5 GO-DOTA 0 0 200 400 600 800 1000 2 μm 500 nm 500 nm Hydrodynamic diameter (nm) B l-GO s-GO us-GO D 0 5 nm 5 nm 5 nm -10 ζ-potential (mV) GO -20 1 μm -30 4 μm 0 nm 1 μm 0 nm 0 nm AFM -40 5 nm 5 nm 5 nm GO-DOTA -50 -60 -70 4 μm 0 nm 1 μm 0 nm 1 μm 0 nm GO GO-DOTA 4 5 6 Figure 1. Physicochemical characterisation of l-GO, s-GO and us-GO before and after DOTA 7 functionalisation. A) Dynamic light scattering (DLS) size distributions; B) Electrophoretic mobility 8 (ζ potential) mean surface charge data. Morphological and structural characterisation data is 9 shown, using C) TEM and D) AFM. 10 21 1 Figure 2 2 A B 86.3 ± 7.8% 93.9 ± 2.1% [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] s-GO-DOTA [111In] l-GO-DOTA [111In] us-GO-DOTA [111In ] l-GO-DOTA [111In] DOTA [111In] DOTA 67.4 ± 4.3% 77.6 ± 6.6% 77.7 ± 4.3% 88.7 ± 5.4% 90.6 ± 4.4% 92.2 ± 3.8% C PBS 50% Serum 120 120 % Radiolabelling at % Radiolabelling at 100 100 application point application point 80 80 60 60 [111In] l-GO-DOTA [111In] l-GO-DOTA [111In] s-GO-DOTA 40 [111In] s-GO-DOTA 40 [111In] us-GO-DOTA [111In] us-GO-DOTA 20 20 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Days post-centrifugation Days post-centrifugation 3 4 5 Figure 2: Radiolabelling efficiency and stability. A) Efficiency of radiolabelling of the three types 6 of GO ([111In] l-GO-DOTA, [111In] s-GO-DOTA and [111In] us-GO-DOTA) compared to control 7 [111In]DOTA after the radiolabelling reaction; values indicate the average of three independent 8 labelling repeats. B) Radiolabelling purity of the three samples after centrifugation and before 9 administration in the animals. C) Stability of radiolabelling in PBS and 50% serum up to 7 days. 10 22 1 Figure 3 2 3 A C 1hr 60 4hrs 100 Control [111In] DOTA %ID per g of tissues 24hrs [111In] l-GO-DOTA l-GO-DOTA [111In] l-GO-DOTA % ID in whole blood 40 s-GO-DOTA [111In] s-GO-DOTA 10 [111In] us-GO-DOTA us-GO-DOTA 20 1 0 0.1 %ID per g of tissues 60 [111In] s-GO-DOTA 0.01 0 500 1000 1500 40 Time (min) 20 B 1h 24 h 0 [111In] l-GO-DOTA 50 % ID/g 60 %ID per g of tissues [111In] us-GO-DOTA 40 [111In] s-GO-DOTA 20 0 [111In] us-GO-DOTA 60 %ID per g of tissues 40 [111In] DOTA 20 [111In] DOTA 0 0 % ID/g 4 5 6 Figure 3: Biodistribution of the three types of GO ([111In] l-GO-DOTA, [111In] s-GO-DOTA and 7 [111In] us-GO-DOTA) compared to [111In]DOTA control at 1h and 24h. A) Blood profile up to 8 24h. B) SPECT/CT images expressed as % of injected dose per gram of tissues (doses were decay 9 corrected at the second time point). From left to right (whole body maximum intensity projections 10 (MIP), sagittal, coronal and transverse views). Interactive 3D images of the MIPs are available 11 online. C) Organ distribution of the three materials as determined by γ-counting. Four animals per 12 group were used for A and C, while two animals were used for the imaging in (B). 13 14 15 23 1 2 3 4 Figure 4 5 A B l-GO-DOTA [111In ] s-GO-DOTA [111In] us-GO-DOTA [111In] [DOTA 111In] [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA 111In] [DOTA 40 1h 4h Solvent 24 h Front 30 % ID in urine 20 10 Application 0 Point Intensity (a.u.) Intensity (a.u.) Intensity (a.u.) Intensity (a.u.) Raman Spectroscopy 1000 1500 2000 1000 1500 2000 1000 1500 2000 1000 1500 2000 Raman Shift (cm-1) Raman Shift (cm-1) Raman Shift (cm-1) Raman Shift (cm-1) C Region of Interest 3 2.5 % ID in faeces 2 1.5 1 0.5 0 6 7 Figure 4: Excretion profile of l-, s- and us-GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 8 [111In]us-GO-DOTA) compared to control [111In]DOTA. A) Urinary excretion profile of the three 9 materials at different time points; B) detection of the graphene material in the urine 24 hours post 10 administration of mice as demonstrated by radio-TLC and corroborative Raman spectroscopy 11 (scale bars are 20 µm); C) faecal excretion of the three materials compared to the control. 12 13 14 15 16 24 1 2 3 Figure 5 4 Mouse1 Mouse2 Dex 5% l-GO s-GO us-GO 5 6 Figure 5. Effect of non-radiolabelled l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on lung 7 compared to vehicle alone (5% dextrose). Haematoxylin and eosin stained lung sections (5 µm 8 thick) after injection of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) 9 after 24h of two different mice for each material. Two magnifications were used, with scale bars on 10 the left in each panel showing 100 µm, while those on the right are indicating 20 µm. 11 12 13 25 1 Supporting Information 2 3 4 5 Name At% Name At% C=C 5.3 l-GO C=C 6.3 s-GO Name At% us-GO C=C 9.1 C-C 30.9 C-C 31.4 C-C 32.4 C-OH 13.4 C-OH 12.6 C-OH 15.8 C-O-C 38.2 C-O-C 38.6 C-O-C 32.5 C=O 4.6 C=O 3.9 C=O 2.9 C=OOH 5.2 C=OOH 5.0 C=OOH 5.4 CPS CPS CPS pi-pi* 2.4 pi-pi* 2.2 pi-pi* 1.9 296 294 292 290 288 286 284 282 296 294 292 290 288 286 284 282 296 294 292 290 288 286 284 282 Binding energy (eV) Binding energy (eV) Binding energy (eV) us-GO-DOTA Name At% l-GO-DOTA Name At% s-GO-DOTA Name At% C=C 8.1 C=C 7.4 C-C 41.0 C=C 10.8 C-C 41.8 C-OH, C-N 16.8 C-C 35.5 C-OH, C-N 16.0 C-O-C 25.1 C-OH, C-N 22.1 C-O-C 22.6 C=O 2.6 C-O-C 20.6 C=O 5.8 C=OOH, 5.2 C=O 5.2 C=OOH 5.3 CPS C=OOH, 4.8 CPS C=ONR C=ONR2 CPS pi-pi* 1.0 C=ONR2 pi-pi* 1.1 C-F contamination pi-pi* 1.0 R=C, H R=C,H R=C, H 296 294 292 290 288 286 284 282 296 294 292 290 288 286 284 282 296 294 292 290 288 286 284 282 Bindinge energy (eV) Binding energy (eV) Binding energy (eV) 6 7 8 Figure S1: C1s XPS spectra showing carbon high resolution spectra of all three GO materials with 9 their DOTA functionalized counterparts. 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 26 1 2 3 4 5 Name At% Name At% Name At% C-O-C, 92.3 C-O-C, 94.7 C-OH l-GO C-O-C, 93.1 C-OH, s-GO C-O, us-GO (C=O)-OH (C=O)-OH (C=O)-OH C=O, 3.6 C=O, 5.3 C=O, 6.3 C=OOH C=OOH C=OOH H2O 1.7 H2O 1.6 H2O 1.4 CPS CPS CPS 538 536 534 532 530 528 538 536 534 532 530 528 538 536 534 532 530 528 Binding energy (eV) Binding energy (eV) Binding energy (eV) Name At% Name At% Name At% us-GO-DOTA C-O-C, 82.0 l-GO-DOTA C-O-C, 75.6 s-GO-DOTA C-O-C, 64.5 C-OH, C-OH, C-OH, (C=O)-OH (C=O)-OH (C=O)-OH C=O, 14.6 C=O, 19.6 C=O, 28.6 C=OOH C=OOH, C=OOH, C=ONR2 C=ONR2 C=ONR2 CPS CPS CPS H2O 3.4 H2O 4.8 H2O 6.9 R=C,H R=C,H R=C,H 538 536 534 532 530 528 538 536 534 532 530 528 538 536 534 532 530 528 Binding energy (eV) Binding energy (eV) Binding energy (eV) 6 7 Figure S2: O1s XPS spectra showing oxygen high resolution spectra of all three GO materials with 8 their DOTA functionalized counterparts. 9 10 11 12 13 14 15 16 17 18 19 20 27 1 2 3 4 5 6 7 Name At% Name At% l-GO-DOTA C-NR2, C=ONR2 58.8 s-GO-DOTA Name At% us-GO-DOTA C-NR2, C=ONR2 72.8 C-NR2, C=ONR2 76.4 + + C-N R3 41.2 C-N R3 27.2 + C-N R3 23.6 R=C,H R=C,H R=C,H CPS CPS CPS 408 406 404 402 400 398 396 408 406 404 402 400 398 396 408 406 404 402 400 398 396 Binding energy (eV) Binding energy (eV) Binding energy (eV) 8 9 10 Figure S3: N1s XPS spectra showing nitrogen high resolution spectra of all three GO starting 11 materials with their DOTA- functionalized counterparts. 12 13 28 1 100 l-GO L-GO DOTA-L-GO l-GO-DOTA 90 s-GO S-GO s-GO-DOTA DOTA-S-GO 80 us-GO US-GO Weight (%) DOTA-US-GO us-GO-DOTA 70 60 50 40 30 100 200 300 400 500 600 700 800 900 Temperature (°C) 2 3 Figure S4: TGA of the three starting GO materials and their functionalized conjugates. 4 29 [111In ] [111In] [111In] [111In] [111In] l-GO-DOTA s-GO-DOTA us-GO-DOTA EDTA DOTA 95.4 % 92.3 % 0h 90.8% 89.7 % 89.8% 92.2% 88.7% 1h 87.5% 92.2 % 89.0 % 86.5% 91.2% 48h 93.8% 95.0% 93.5% 1 2 3 Figure S5: Radiolabeling stability in serum. TLC strips of the samples [111In]l-GO-DOTA, [111In] - 4 GO-DOTA, [111In]us-GO-DOTA [111In]EDTA and [111In]DOTA control in 50% serum at different time 5 points. 6 7 8 30 1 Dynamic Static Planar 0 -2 min 2 min -1 h 4h 24 h [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] DOTA 2 3 Figure S6: Planar SPECT data showing the biodistribution of the three types of GO ([111In]GO- 4 DOTA, [111In]s-GO-DOTA and [111In]us-GO-DOTA) compared to control [111In]DOTA. A) The first 5 two columns show the dynamic videos of the injection phase (2min) and 1h dynamic profile, while 6 the two columns on the right are the planer static images at 4h and 24h. Dynamic videos are 7 available online. 8 9 31 DOTA 80 l-GO-DOTA s-GO-DOTA us-GO-DOTA Lung 60 % ID/g 40 20 [111In] DOTA [111In] l-GO-DOTA 0 [111In] s-GO-DOTA 0 1 5 10 15 20 25 Time (h) [111In] us-GO-DOTA DOTA 80 l-GO-DOTA s-GO-DOTA us-GO-DOTA 60 Liver % ID/g 40 20 0 0 1 5 10 15 20 25 Time (h) DOTA 80 l-GO-DOTA s-GO-DOTA us-GO-DOTA 60 Kidney % ID/g 40 20 0 0 1 5 10 15 20 25 Time (h) DOTA 80 l-GO-DOTA s-GO-DOTA us-GO-DOTA Bladder 60 % ID/g 40 20 0 0 1 5 10 15 20 25 1 Time (h) 2 3 4 Figure S7: Organ distribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 5 [111In]us-GO-DOTA) compared to control [111In]DOTA demonstrated by the time activity curve of the 6 corresponding planar SPECT data (n=2). 7 8 32 [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] DOTA 1h 24 h 1 2 3 Figure S8: 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 first batch of mice. From 5 left to right (whole body maximum intensity projections (MIP), sagittal, coronal and transverse 6 views). The data here are normalized MBq for decay and dose differences between mice. 7 Interactive 3D images (of MIPs) are available online. 8 9 33 1 2 [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] DOTA 3 4 Figure S9: Biodistribution of the three types of GO ([111In]l-GO-DOTA, [111In]s-GO-DOTA and 5 [111In]us-GO-DOTA) compared to control [111In]DOTA at 1h and 24h in a second batch of mice. 6 From left to right (whole body maximum intensity projections (MIP), sagittal, coronal and transverse 7 views). SPECT/CT images expressed as % of injected dose per gram of tissues (doses were decay 8 corrected at the second time point). Interactive 3D images (of MIPs) are available online. 9 10 34 [111In] l-GO-DOTA [111In] s-GO-DOTA [111In] us-GO-DOTA [111In] DOTA 1 2 3 Figure S10: 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). Data here are expressed in normalized MBq for decay and differences in injected doses 7 between mice. Interactive 3D images (of MIPs) are available online. 8 9 10 11 12 13 14 15 35 1 2 3 4 5 6 Mouse1 Mouse2 l-GO 7 8 9 Figure S11: Effect of l-GO-DOTA on lung. Haematoxylin and eosin stained lung sections (5 µm 10 thick) after injection of l-GO-DOTA after 24h of two different mice. Scale bars for the images on the 11 left are 100 µm while those on the right are 20 µm. These images are captured at different locations 12 compared to those in Figure 5. 13 14 36 1 2 Large GO Small GO Ultra-small GO 1 % in f ilt r a t io n 0 .1 0 .0 1 0 .0 0 1 O O O -G -G G l- s s 3 u 4 5 Figure S12: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on lung compared to control 5% 6 dextrose. Haematoxylin and eosin stained lung sections (5 µm thick) after injection of l-GO-DOTA, 7 s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h of two different mice for 8 each material. Scale bars for the images on the left are 100 µm while those on the left are 20 µm. 9 10 37 Dex 5% l-GO s-GO us-GO 1 2 3 Figure S13: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on spleen structure compared 4 to control 5% dextrose. Haematoxylin and eosin stained spleen 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 of first set of mice. No evidence of histopathology was determined in the red pulp (second panel) 7 and the white pulp (third panel) in any of the samples as compared to the controls. Each point 8 represents a different mouse. Scale bars for the images on the left are 100 µm while those in the 9 middle and last panels are 50 µm. The insets are 1 µm. 10 11 12 38 Dex 5% l-GO s-GO us-GO 1 2 3 Figure S14: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on spleen structure compared 4 to control 5% dextrose. Haematoxylin and eosin stained spleen 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 of second set of mice. No evidence of histopathology was determined in the red pulp (second 7 panel) and the white pulp (third panel) in any of the samples as compared to the controls. Scale 8 bars for the images on the left are 100 µm while those in the middle and last panels are 50 µm,. 9 The insets are 1 µm. 10 11 12 39 Mouse1 Mouse2 Dex 5% l-GO s-GO us-GO 1 2 Figure S15: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on liver structure compared to 3 control 5% dextrose. Haematoxylin and eosin stained liver sections (5 µm thick) after injection of 4 l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h. No evidence 5 of histopathology was determined in the livers in any of the samples as compared to the controls. 6 Scale bars for the images are 100 µm. 7 8 9 10 40 Mouse1 Mouse2 Glomerulus Tubules Glomerulus Tubules G G Dex 5% P P D D G D P l-GO G D P G P D s-GO G D P G G P P us-GO D D 1 2 3 4 Figure S16: Effect of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA on kidney structure compared 5 to control 5% dextrose. Haematoxylin and eosin stained kidney sections (5 µm thick) after 6 injection of l-GO-DOTA, s-GO-DOTA, us-GO-DOTA and 5% dextrose (negative control) after 24h. 7 No evidence of histopathology was determined in the kidney glomerular or tubular regions as 8 compared to the controls. Scale bars for the images are 50 µm. G: glomerulus, P: proximal 9 convoluted tubule and D: distal convoluted tubule. 10 11 41 1 Supporting Tables 2 Table S1: Summary of the physicochemical characteristics of GO materials used herein. Measured 3 sizes are expressed as means and ranges. Full characterisation was published in previous works 4 (37, 48-50). 5 Parameter Method l-GO s-GO us-GO Lateral TEM 10 μm [1 – 30 μm] 450 nm [0.2 – 1 μm] 122 nm [10 – 550 nm] Dimension AFM 21 μm [10 – 30 μm] 74 nm [29 – 369 nm] 69 nm [30 – 300 nm] Thickness AFM 1.8 nm [1 – 6 nm] 2 nm [0.5 – 12 nm] 1 nm [0.4 – 2.5 nm] 3229 837 nm 155.5 3.5 nm 54.6 ± 0.7 nm Colloidal DLS (PDI = 1.000 0.000) (PDI = 0.231 0.016) (PDI = 0.344 ± 0.008) properties ζ-potential -61.7 1.8 mV -55.0 0.8 mV -54.8 0.1 mV I (D)/I (G) Raman 1.30 0.04 1.35 0.02 1.34 0.03 Elemental XPS C: 68.6%, O: 31.4% C: 68.5%, O: 31.5% C: 69.8%, O: 30.2% composition Parameter Method l-GO-DOTA s-GO-DOTA us-GO-DOTA Lateral TEM 2 μm [0.2 – 5 μm] 354 nm [29 – 1434 nm] 160 nm [10 – 538 nm] Dimension AFM 0.36 μm [0.1 – 3.5 μm] 91 nm [30 – 800 nm] 49 nm [30 – 480 nm] Thickness AFM 1.9 nm [0.3 – 51.2 nm] 2.5 nm [0.2 – 34.3 nm] 2.6 nm [0.2 – 48.0 nm] I (D)/I (G) Raman 1.34 0.05 1.33 0.03 1.34 0.02 Colloidal ζ-potential -31.1 0.8 mV -36.7 0.4 mV -38.8 0.9 mV stability Elemental C: 71.4%, O: 27.5%, C: 69.6%, O: 29.0%, C: 68.9%, O: 29.7%, XPS composition N: 1.1% N: 1.4% N: 1.4% 6 7 8 9 10 11 12 13 14 15 16 17 18 19 42 1 Table S2: First-order pharmacokinetic parameters of all three [111In] GO-DOTA materials compared 2 to control [111In] DOTA post intravenous exposure. 3 t1/2α t1/2β AUC Vdss CL min h mg.h/L ml ml/h [111In] l-GO-DOTA 21.2 ± 15.5 7.2 ± 1.8 28.4 ± 11.8 14.5 ± 2.9 3.2 ± 1.3 [111In] s-GO-DOTA 22.9 ± 4.9 5.4 ± 3.2 22.9 ± 4.9 5.9 ± 1.6 3.1 ± 1.2 [111In] us-GO-DOTA 21.9 ± 5.8 9.8 ± 3.6 21.9 ± 5.8 13.2 ± 5.0 2.2 ± 0.8 [111In] DOTA 11.8 ± 1.5 9.4 ± 2.4 1.4 ± 0.2 10.8 ± 1.1 2.3 ± 0.5 4 43 1 2 Table S3: Remaining amounts of all three [111In] GO-DOTA materials (compared to [111In] DOTA 3 control) in the blood (µg per ml of blood), quantified by γ-counting (n=4 per condition). 1h 24 h Sample µg/ml µg/ml [111In] l-GO-DOTA 1.6 ± 3.7 0.41 ± 0.3 [111In] s-GO-DOTA 3.6 ± 2.6 0.41 ± 0.2 [111In] us-GO-DOTA 3.5 ± 4.1 0.42 ± 0.2 [111In] DOTA 0.1 ± 0.1 0.01 ± 0.0 4 5 6 44 1 2 Table S4: SPECT/CT quantification showing the remaining amounts of all three materials 3 compared to control in the body after 24h, n=2 per condition. Sample % ID after 24 h [111In] l-GO-DOTA 50.8 ± 17.4 [111In] s-GO-DOTA 36.5 ± 2.7 [111In] us-GO-DOTA 38.1 ± 5.3 [111In] DOTA 5.1 ± 1.9 4 45